Time-Series Trends of Trace Elements in an Ice Core from Antarctica by SIDDIK SINAN KESKIN M. S., Nuclear Engineering, The University of Michigan (1990) B. S., Nuclear Energy Engineering, Hacettepe University, Rep. of Turkey (1986) Submitted to the Department of Nuclear Engineering in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY in Nuclear Engineering at the Massachusetts Institute of Technology September 1995 © 1995 Massachusetts Institute of Technology. All rights reserved. Signature of Author... ....................... "_ Department of Nuclear Engineering July 6, 1995 F/e Certified by....... -Dr. D Ilhan Olmez Principal Researc Scientist, Nuclear Reactor Laboratory Th - a An 79 i S s se Certified by......... sor J Prof. Chester C. Langway, Jr. ...... 6 i uperv ............... Department of Geology at SUNY, Buffalo (Emeritus) Thesis Reader Accepted by .................................... Prof. Jeffrey P. Freidberg Department of Nuclear Engineering Chairman Comjittee on Graduate Students OF T.A Ec,-CHNOLOYUSF i' . OF TECHNOLO.G APR 2 2 1996 TIME-SERIES TRENDS OF TRACE ELEMENTS IN AN ICE CORE FROM ANTARCTICA by SIDDIK SINAN KESKIN Submitted to the Department of Nuclear Engineering on July 6, 1995 in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Nuclear Engineering ABSTRACT Trace element measurements were made by instrumental neutron activation analysis on stratigraphically dated ice core samples from Byrd Station, Antarctica, to determine the concentration levels of natural and anthropogenic substances. Sampling was continuous between 1926 A.D. and 1989 A.D. and selective between 1711 A.D. and 1926 A.D. Twenty-one elements with concentrations above the detection limits were determined. The time period between 1969 A.D. and 1989 A.D. showed an enhanced impact on the Antarctic ice sheets from natural sources in the form of marine and crustal aerosols. A disturbed ocean-atmosphere interface due to El Niflo Southern Oscillation (ENSO) events seems to be a candidate especially for the enhanced marine aerosol deposition in Antarctica. Time-series trend of the concentration of deposited aluminum, which is mainly a crustal aerosol related element, shows a strong negative correlation with the time-series trend of annual average total column ozone concentrations homogenized between the 600 S and 900 S latitudes from the Total Ozone Mapping Spectrometer (TOMS) ozone data set. Although the time period is not long enough to draw a strong conclusion (1979-1989), the special role of crustal origin clay minerals on cloud nucleation dynamics might be a factor in the heterogeneous stratospheric ozone depletion chemistry through polar stratospheric cloud dynamics, assuming some troposphere-stratosphere mixing of these aerosols. The correlation of antimony and arsenic enrichments with known or suspected volcanic events was established. These marker elements was shown to be useful especially for the identification of specific historical volcanic events with low sulfur emissions. Although a clear anthropogenic impact was not observed, concentrations of arsenic, chromium, and zinc, which might come from both natural and anthropogenic sources, indicated an increase after 1960's. Principal component factor analysis indicated a possible transition-metal (especially manganese and iron) catalyzed bromine chemistry cycle, which has been suggested as the cause of tropospheric surface-level ozone depletion observed in Greenland. Calculated snow-to-air scavenging ratios indicated more efficient scavenging for crustal aerosols followed by marine and volatile elements. A new method was developed for direct air content determination in small deep ice core samples through the measurement of enclosed argon gas by instrumental neutron activation analysis. Thesis Supervisor: Dr. Ilhan Olmez Title: Principal Research Scientist, MIT Nuclear Reactor Laboratory, and Department of Nuclear Engineering Thesis Reader: Prof. Chester C. Langway, Jr. Title: Professor of Geology, SUNY, Buffalo (Emeritus) BIOGRAPHICAL SKETCH Name: SIDDIK SINAN KESKIN Born: August 31, 1964; Artvin, Republic of Turkey Education: Doctor of Philosophy in Nuclear Engineering, Massachusetts Institute of Technology. September 1995. Thesis Title: Time-Series Trends of Trace Elements in an Ice Core from Antarctica. Master of Science in Nuclear Engineering, The University of Michigan, May 1990. Bachelor of Science in Nuclear Energy Engineering, Hacettepe University, Republic of Turkey, June 1986. High School, Haydarpasa Teknik Lisesi, Republic of Turkey, June 1982. Employment: Arcelik, Kocaeli, Republic of Turkey, Summer 1981 Arcelik, Kocaeli, Republic of Turkey, Summer 1980 Super Elektrik, Istanbul, Republic of Turkey, Summer 1979 Publications and Presentations: Keskin, S.S., I. Olmez, and C.C. Langway, Jr., "Time-Series Analysis of Chemical Trends in a Dated Ice Core from Antarctica", (Abstract), American Nuclear Society, Winter Meeting, November 13-17, 1994, Washington, D.C. Keskin, S.S., C.C. Langway, Jr., and I. Olmez, "Trace Elements in Dated Ice Cores From Antarctica and Greenland", paper presented at the Second International Conference on Managing Hazardous Air Pollutants, July 13-15, 1993, Washington, D.C. Keskin, S.S., X. Huang, I. Olmez, and C.C. Langway, Jr., "Trace Elements in a Dated Ice Core From Antarctica", (Abstract), American Nuclear Society, National Meeting, June 7-12, 1992, Boston, MA. Huang, X., S. Keskin, I. Olmez, and G.E. Gordon, "Automobiles:Possible Sources of Metals Other Than Lead in the Urban Atmosphere", (Abstract), American Nuclear Society, National Meeting, June 7-12, 1992, Boston, MA. Olmez, I., X. Huang, S. Keskin, G.E. Gordon, and J. Ondov, "New Markers and Source Composition Library for Motor Vehicle Emissions", American Chemical Society, National Meeting, April 5-10, 1992, San Francisco, CA. ACKNOWLEDGMENTS I would like to thank to the Turkish Ministry of Education for giving me the chance to come to the United States for my graduate study and providing financial support during my early years. Further financial support during this research came from the different projects that I have participated in as a research assistant for the Environmental Research and Radiochemistry Division. I am grateful to my thesis supervisor Dr. Ilhan Olmez for his trust, encouragement, and constant support during these years. He has been a good friend and a serious teacher who was always able to convince me to volunteer to do more. His vision and will to pursue research in diverse fields has always impressed me as a colleague. I am also grateful to my thesis reader Professor Chester C. Langway, Jr. for his cooperation and help during this research. He opened the way to this research by providing the samples used and letting me use his laboratory at SUNY, Buffalo. His tremendous experience in the field of glaciology was always an assurance. I also thank to my thesis committee members Prof. Otto K. Harling, Prof. Norman C. Rasmussen, Prof. Jacquelyn C. Yanch, and Prof. Sidney Yip for their critiques and suggestions during this research. I would like to thank Thomas J. Langway, Kazuo Osada, and Ronald Wentzek of the Ice Core Laboratory at SUNY, Buffalo for their help during sample preparation in their laboratory. Special thanks to the MITR-II operations staff for their help during the sample irradiations. Thanks are also due to the NRL machine shop staff for their help. The Nuclear Reactor Laboratory and Nuclear Engineering Department staff deserve credit for their assistance during my stay at MIT. Particular thanks to Rachel Morton for being a resource about computational issues. I would like to mention my sincere appreciation to my friends and colleges Michael Ames, Jianmei Che, Jec-Kong Gone, Xudong Huang, and Rena Lee at the Environmental Research and Radiochemistry laboratory for their friendship and help during all these years. It was a close family away from home that I always enjoyed to be a part of. Amory Wakefield deserves special thanks for her help in data handling and grammatical corrections at various stages. Karl Sebelius was also a help in some of the data handling tasks. Special thanks are also due to visiting scientists Namik K. Aras, Jack Beal, Richard D. Fink, Gulen Gullu, and Francis Pink for their friendship and helpful discussions. I would like to thank Prof. Otto K. Harling, Prof. Alan F. Henry, and Prof. Mujid S. Kazimi for their help and trust during my early years at MIT. My countrymen and friends Huseyin Akcay, Ulfet Atav, Tamer Bahadir, Erol Cubukcu, Oguz Ertugrul, Mehmet Gozcu, Tugrulbey Kiryaman, Volkan Kubali (and his notebook computer), Ata Mugan, Cetin Ozbutun, Sadettin and Asuman Ozturk, Yuksel and Asuman Parlatan, Ilker and Sibel Tari, Mehmet Tombakoglu, Ibrahim Ucok, and Serhat Yesilyurt deserve particular thanks for their friendship and support during my stay in this country. My former and current housemates Patrick Donnally, Salvatore Galea, Eric M. Jordan, Matthew Marjanovic, Diane Ronan, and Kaoru Takase also deserve special thanks for their friendship. Prof. Osman K. Kadiroglu and Prof. Yalcin Sanalan of Nuclear Energy Engineering at the Hacettepe University deserve very special thanks for introducing me to this field and preparing me for the future. As this is the formal end of my career as a student, I would like to mention my appreciation to all of my previous teachers who did their best to prepare me for life as a member of humanity. Finally, I am very much grateful to my mother Sennur Keskin, my father Fikret Keskin, my sisters Aysu Keskin and Ayse Keskin, and my brother M. Gokhan Keskin for their continuing love and support throughout my life. God's guidance and their existence made my life joyful and fulfilling. To show my sincere appreciation, I am dedicating this thesis to them. TABLE OF CONTENTS ....................................... A B STR A C T ....................................................................... ............................. BIOGRAPHICAL SKETCH...................................................... ........................................... ACKNOW LEDGM ENTS........................................ ............................ TABLE OF CONTENTS................................ LIST O F FIG U RES.............................................. ...................................................... LIST O F TA BLES............................................... ....................................................... 1. IN TR O DU C TIO N .................................................................. 1.1. 1.2. 1.3. 1.4. 1.5. 2. 3. 2 3 4........ 5 7 9 ........................ 11 11 ................................................. B ackground...................................... A Brief History of Antarctic Research....................................14 ..... 15 NBY-89 Ice Core Research Program................................ Scope of Research................................................... 19 Basic Theory of the Instrumental Neutron Activation Analysis.............23 EXPERIM ENTAL................. .................................................................. 26 26 26 29 30 2.1. ........ Sample Selection and Preparation.............................. ........ ..................... 2.1.1. Selection....................................................... ........... ............ 2.1.2. Preparation............................................. ..... 2.1.3. Pre-concentration by Freeze-drying........................ 2.2. ... 33 Instrumental Neutron Activation Analysis.............................. ........... ............ 33 2.2.1. Irradiations............................................. ....... 35 2.2.2. Gamma-Ray Spectroscopy.............................. 2.2.3. Blank Correction..............................................38 2.3. Air Content Determination.........................................40 RESULTS AND DISCUSSION.............................................44 3.1. Data Quality Assessm ent..................................................................... 44 3.2. Time-Series Trends.............................................................. 3.2.1. 3.2.2. 3.2.3. 3.2.4. 3.2.5. ................... 51 ........ 53 The Marine Aerosols................................... Possible El Nifio Events-Marine Aerosols Connection............. 60 Crustal Aerosols and Their Possible Role in Polar Atmosphere....67 The Volcanic Aerosols................................................................. 78 ...... 90 The Anthropogenic Aerosols................................ 5 3.3. Source Characterization........................................ .......... .......... 96 ..... 96 3.3.1. Source-Receptor Relationship............................. Factor Analysis...............................98 3.3.2. Principal Component 3.4. Air/Snow Partitioning.......................................... ......... .......... 107 3.4.1. Current State of Knowledge............................ 3.4.2. Scavenging Ratios............................... 3.5. 4. 5. Air Content Analysis................................................................................ 107 112 16 SUMMARY AND SUGGESTIONS FOR FUTURE WORK.....................1.......19 4.1. Summary.........................................................119 4.2. Suggestions for future work........................................... REFERENCES............................... 122 125 APPENDIX A. Elemental concentrations and related statistical information....... 135 APPENDIX B. Crustal and marine Enrichment Factors................................. APPENDIX C. Data tables for the samples obtained from volcanic eruption horizons and a Rare Earth Elements anomaly horizon............ 183 APPENDIX D. Basic theory of the Principal Component Factor Analysis...........188 APPENDIX E. A brief review of the heterogeneous stratospheric ozone ....... depletion chemistry............................ ...... . 163 204 LIST OF FIGURES Figure 1-1. Map of Antarctica and major ice coring sites.................................. Figure 1-2. Illustration of (n,y) reaction followed by positron decay..........................23 Figure 2-1. Schematic illustration of ice sample surface cleaning setup....................30 Figure 2-2. Schematic illustration of the freeze-drying system............................. Figure 2-3. An example gamma-ray spectrum for ice core samples after ten minutes irradiation............................................... .......... 16 32 .......... 37 Figure 2-4. Illustration of air-free ice production system...................40 Figure 2-5. Illustration of Air standard vial volume determination........................ Figure 2-6. Illustration of ice sample selection for air content measurement............42 Figure 3-1. Sketch of sampling system for cross-profiling sample preparation............45 Figure 3-2. Elemental concentrations measured in cross-profiling samples..............47 Figure 3-3. Sea-salt sodium (Nass) concentrations time-series trend.......................54 Figure 3-4. Sodium and chlorine concentrations time-series trends.........................55 Figure 3-5. Chlorine-to-sodium ratio time-series trend................................ Figure 3-6. El Nifio events after 1926................................. Figure 3-7. The power spectrum of the sea-salt sodium time-series data.................64 Figure 3-8. Sea-salt sodium time-series trend together with 4.2 year period fit 41 ... 59 61 ............ based on the analysis with The Linear Least Squares Spectral Analysis (LLSSA) method.................................................................... Figure 3-9. 65 Sea-salt sodium time-series trend together with the top eight period fit based on the analysis with The Linear Least Squares Spectral Analysis (LLSSA ) m ethod................................................................................. 66 69 ......................... Figure 3-10. Aluminum concentrations time-series trend.............. Figure 3-11. Aluminum and scandium concentrations time-series trends...................70 Figure 3-12. Scandium crustal Enrichment Factors time-series trend.........................71 Figure 3-13. Annual average total column ozone (600 S-900 S) and aluminum concentrations time-series trends................................. Figure 3-14. ...... 75 Antimony and arsenic crustal Enrichment Factors time-series trends........80 81 Figure 3-15. Antimony and arsenic concentrations time-series trends...................... Figure 3-16. Lanthanum, Cerium, and Samarium crustal Enrichment Factors time-series trends..................................................... ........................................... 86 Figure 3-17. Excess antimony concentrations time-series trend.................................92 Figure 3-18. Excess arsenic concentrations time-series trend................................. 93 Figure 3-19. Excess zinc concentrations time-series trend................................... 94 Figure 3-20. Excess chromium concentrations time-series trend...............................95 Figure 3-21. Schematic of source-receptor relationship........................... .... 97 Figure 3-22. Factor Scores results for the first factor............................... 101 Figure 3-23. Factor Scores results for the second factor............................................102 Figure 3-24. Factor Scores results for the third factor............................... Figure 3-25. Factor Scores results for the fourth factor................................... 104 Figure 3-26. Air-to-snow scavenging ratios for the elements Al, Fe, La, Na, Zn, Sb, and As .................................... .... 103 115 Figure A-1. Cumulative percent data versus overall relative uncertainty..................136 Figure D- 1. Sample vector space with three factors and five data column vectors.....202 LIST OF TABLES Table 1-1. Individual studies, investigators, affiliation, and current status of science studies program for the NBY-89 ice core................................18 Table 2-1. Ice core samples selected for the study of trace elements time-series trends................................................. Table 2-2. ......... .......... 27 Ice core samples selected to locate a previously observed rare earth elements enrichment peak around 89 meters depth...............................28 Table 2-3. Ice core samples selected from volcanic eruption horizons....................28 Table 2-4. Ice core samples selected for quality control purpose............................28 Table 2-5. Elements scanned in ice core samples by INAA.................................. 38 Table 2-6. Container bag-blank concentrations..................................39 Table 2-7. Air-free ice production time chart........................................41 Table 3-1. Concentration results for cross-profiling samples.............................. Table 3-2. Median uncertainties on the concentrations of elements with significant observation frequency............................... 46 ...... 51 Table 3-3. Periodicities and their power values in sea-salt sodium data set..............63 Table 3-4. Average crustal Enrichment Factor results for potentially crustal origin elements............................ Table 3-5. ........................ 68 The crustal Enrichment Factors of the enriched elements in suspected volcanic eruption horizons................................ Table 3-6. ..... 83 Percent average concentration changes for Sb, As, Zn, and Cr relative to the 1711-1901 time period................................................... 91 .... 96 Table 3-7. Calculated natural background concentrations....................... Table 3-8. Factor analysis results for related parameters....................................... 100 Table 3-9. Factor loadings (after varimax rotation) and estimated communalities ....105 Table 3-10. Air-to-snow scavenging ratios for some elements............................... 113 Table 3-11. Air content measurement results for samples and blanks.....................1...16 Table A-1. Net elemental concentrations measured in individual NBY-89 ice core samples................................................. .......... .......... 140 Table A-2. Statistical summary information for the data in Table A- .................... 149 Table A-3. Depth-age relationship for NBY-89 ice core..........................................152 Table A-4. Net elemental concentrations calculated in annual intervals for NBY-89 ice core........................................ Table A-5. ........ Most probable values calculated from elemental concentration distribution functions ............................................. Table B-1. ................................................ 173 Elemental concentrations for the samples obtained from 184 volcanic eruption horizons ............................. Table C-2. Crustal Enrichment Factors for the samples obtained from volcanic eruption horizons................................... Table C-3. ........ 185 Elemental concentrations for the samples obtained from a REE anomaly horizon...................................... Table C-4. 164 Enrichment Factor calculation results with respect to marine sodium................................... Table C-1. 162 Enrichment Factor calculation results with respect to crustal alum inum ........................................ Table B-2. 153 ........ 186 ........ 187 Crustal Enrichment Factors for the samples obtained from a REE anomaly horizon...................................... 1. INTRODUCTION 1.1. Background A fuller understanding of the interaction between the land, ocean, and atmosphere related processes has improved considerably within the last few decades (e.g., Charlson, 1987; Harriss et al., 1988; Budd, 1991; Schneider, 1994). At the same time, an awareness of the impact of natural and anthropogenic substances on these systems has been developed. For example, important clues regarding the influence of large scale volcanic emissions on atmospheric albedo and, therefore, on short term climate variations were recently obtained from various in-situ and satellite measurements (e.g., Handler, 1989; Michelangeli et al., 1989). Similarly, more substantial information is reported on the impact of anthropogenic emissions, such as chlorofluorocarbon (CFC) compounds, and their relation to changes within the atmosphere, by depleting stratospheric ozone (e.g., Molina and Molina, 1987; Toon et al., 1989; Molina, 1991). In addition, a number of characteristic elemental signatures have been identified for aerosol source areas for terrestrial and extraterrestrial materials (e.g., Phelan-Kotra et al., 1983; Lowanthal and Rahn, 1985; Olmez and Gordon, 1985; Olmez et al., 1986; Olmez et al., 1988; Huang et al., 1994). Dynamic characteristics of environmental processes and the delicate balance between the environment and anthropogenic activities creates a broad spectrum of scientific, interdisciplinary research programs, which cover land, ocean, and atmosphere related processes. However the lack of long-time observations is an important limiting factor for the better characterization and formulation of these processes. The amassing of surficial environmental data over long periods, except for temperature and perhaps pressure and CO2, is quite recent. Reliable data with long collection histories are mostly limited in space and duration. At the same time, determination of changes in atmospheric composition over long time-units has become more important , especially as these changes concern and are related to ecological and climatic changes as well as influencing life- forms. Some examples are increasing acid rain formation, stratospheric ozone depletion, greenhouse warming, and increasing toxic trace substances in the environment (e.g., Powers and Rambo, 1981; Lindqvist, 1985; Ramanathan, V., 1988; Abbatt and Molina, 1993). A renewed interest has evolved in recent years to obtain historical data for environmental parameters from specific sources, such as polar ice sheets, sediments, and tropical sea corals, which preserve the information within their stratigraphic layers. The Antarctic ice sheet has an average thickness of 2300 meters, with a maximum known thickness close to 5000 meters. Two thirds of the land on which the ice sheet rests is above sea level, mainly in East Antarctica, while most of the West Antarctic land is below sea level (Walton, 1987). The Antarctic and Arctic ice sheets contain a wide range of paleodata entrapped in stratigraphic layers. The layered sequences contain records of changing climatic conditions and deposits of atmospheric constituents. These polar ice sheets are sinks for atmospheric trace elements and other chemical species which were transported by wind systems from various global sources (Wolff and Peel, 1985; Barrie, 1986; Shaw, 1989; Davidson, 1989; Peel, 1989). Recognizing this potential, a substantial number of researchers have already obtained valuable concentration profiles of trace element and other chemical species in several ice cores from polar locations (Langway et al., 1974; Petit et al., 1981; Herron, 1982; Boutron and Patterson, 1983; Palais and Legrand, 1985; Angelis et al., 1987; Saigne et al., 1987; Ivey and Davies, 1987; Boutron and Wolff, 1989; Buat-Menard, 1990; Olmez et al., 1993). The research presented in this thesis made on a polar ice core from the Antarctic ice sheet was designed to measure the pre-industrial ( 18th and early 19t" centuries) and modem era (late 19" and 20" century) annual average multi-element concentrations. The core was meticulously studied by others and reliably dated by multi parameter crosscorrelation (Langway et al., 1994). The main scientific objective here was to assess the impact of anthropogenic and natural processes on atmospheric elemental concentrations and identify some of these sources. The study would result to further identify the overall atmospheric elemental concentration budget of the Southern Hemisphere for both natural and anthropogenic contributions and their probable sources areas. Although anthropogenic impact on the environment is as old as humanity itself, the industrial revolution starting in the early nineteenth century can be considered the beginning of large scale contributions into the natural environment. As part of the study, snow strata that covered 278 years (1711 AD to 1989 AD) was investigated to observe whether or not there is a significant difference in the Antarctic ice sheet elemental concentrations due to anthropogenic impact. In addition, included in this time-period were significant Earth events, such as, volcanic emissions, variations in global air-mass movements, and changes in Earth's vegetation cover, which might have had an effect on the elemental composition of the atmosphere which could be investigated using trace element patterns. Even though there have been previous research studies to determine the trace element concentrations in polar ice sheets (e.g., Peel and Wolff, 1982; Boutron and Patterson, 1983; Wolff and Peel, 1985; Boutron and Patterson, 1987; Girlach and Boutron, 1992; Suttie and Wolff, 1992), most involved a relatively small number of trace element measurements over discontinuous time-intervals. They were also limited by the specific earlier analytical techniques used, the physical quality of the firn or ice core samples used, and an evident lack of a close interdisciplinary and multiparameter approach to an integrated study program. Of great importance, most of the previous data available prior to this investigation does not reveal continuity over long time-periods. As stated earlier, long records are necessary to provide an adequate understanding of the impact of various emission sources on the atmosphere. Since changes in chemical composition trends in the atmosphere are extremely slow, detecting these changes requires numerous measurements and observations over an extended period of time. The results of this study will hopefully fill an existing gap in our knowledge. Although the expressions "ice core elemental concentrations" and "atmospheric elemental concentrations" are used interchangeably here because of their relationship to each other, this relationship is not straightforward by any means. This matter will be discussed in section 3.4, which analyzes and explains the air-to-snow transport of aerosols by various processes and mechanisms. An auxiliary goal in this study was to develop a new method for determining the air content in very small ice samples. A refined method to measure air content (and for that matter argon by INAA) would be useful to study the relationship between the flow patterns and air content changes of glacier ice with depth (as well as dating purposes). At great depths in ice sheets (>1000 m), the thickness of an annual ice layer is often quite small (generally <5 cm and less) due to plastic deformation and stretching of layers under high pressure (Herron, 1982b). It is often difficult, if not impossible, to obtain seasonal variations in parameters in these thin ice layers using the somewhat standard stratigraphic dating parameters as discussed in section 1.4. In addition, even if the seasonality of stratigraphic dating parameters are detectable, analyzing tens of thousands of samples from long ice cores are often not feasible. Under these circumstances, non-stratigraphic dating methods, briefly mentioned in section 1.4, are desired for ice core dating. Consequently a methodology was developed to measure air content within a few minutes by direct ice phase irradiation and Argon gamma-ray spectrum analysis of 2 to 3 gram samples. 1.2. A Brief History of Antarctic Research The development of Antarctic science is a perfect example of human courage and curiosity to learn more about the unknown. In depth information about this development can be found in the Antarctic literature (e.g., Walton, 1987; Fogg, 1992). Fogg suggested that the beginning of Antarctic science can be considered as early as 1699 when Edmond Halley, a British geophysicist, sailed as far as 520 South in the Southern Ocean to measure magnetic variation for determining longitude in open seas (Fogg, 1992). However, the significant scientific expeditions in the continent started at the beginning of the twentieth century. In the early years efforts were mainly in the form of surveying and geological expeditions using dog sledges as the main transportation vehicles. The start of aerial flights in 1928 opened a new chapter for these expeditions to map the continent. Official collaboration between nations started with the first International Polar Year of 1882-1883. A second Polar Year in 1932-1933 followed the first one. In 1951 a special committee (Comit6 Special6 de l'Ann6e G6ophysique Internationale) was set up for the coordination of international scientific efforts in polar regions. Its first meeting (Paris, July 1955) was devoted mainly to deciding on the distribution of field stations and coordinating scientific and logistic objectives (Walton, 1987). This cooperation between nations resulted in many achievements in geology, geophysics, glaciology, biology, and atmospheric sciences and still heralds a high potential for the future advancement of international scientific endeavors. The first ice core recovered in Antarctica was obtained during the NorwegianBritish-Swedish expedition of 1949-1952 on the Maudheim ice shelf. This nearly 100 meter-long ice core was drilled with modified rock drilling equipment (Fogg, 1992). In 1968 a U.S. team reached bedrock and recovered a 2164 meter-deep ice core at Byrd Station in West Antarctica using a mechanical drilling rig, which is still the only ice core that has completely penetrated the Antarctic Ice Sheet (Ueda and Garfield, 1969). In addition, the U.S. teams drilled numerous other ice cores between 1956 and 1968. In 1972 a Soviet team started drilling a deep ice core at Vostok Station in East Antarctica, which has a total ice sheet thickness of 3700 meters. A total of 2083 meters of core has been recovered (Legrand et al., 1988b). At Vostok a thermal drilling technique (invented by the Byrd'68 team) was used which relies on electrical heating of the corer annulus and removing melted water quickly before refreezing. The glaciostatic pressure is balanced by a mix of 224 trimethylpentane and diesel fuel (developed by Byrd'68 team) which has a density close to that of ice and a lower freezing point than the surrounding temperature. They also drilled another core at Komsomolskaya, which is 872 meters long, in 1983. A joint US and French team drilled a 905 meter deep ice core near Dome C in 1977-1978. In addition to the deeper ice cores mentioned, a number of shallow ice cores have been obtained by a number of national teams from different locations. Figure 1.1. is a general location map of ice coring sites. 1.3. NBY-89 Ice Core Research Program The NBY-89 ice core was augured in November 1989 by a three member team of engineers from the Polar Ice Core Office (PICO) at the University of Alaska, with logistic support from the U.S. Navy, at a location 1 km upstream from the temporary Byrd Station Surface Camp (NBY-89, or Navy Byrd-89, the name designated by the U.S. Navy), West 5 90OW o*E SIce Figure 1-1. Map of Antarctica and major ice coring sites (from Palais and Legrand, 1985) Antarctica (80' 01' S, 1190 31' W, 1530 m elevation) (Langway, 1992). The ice core is 164 m deep and 10 cm in diameter. Core recovery was continuous with over 99% of the profile obtained. In addition, two 10 m-deep ice cores were augured about 14 km (NBY2) and 29 km (NBY-3) upstream from the main core. Two meter-deep pits were hand dug at each drilling location. Four members from the Ice Core Laboratory, SUNY at Buffalo, formed the science field team. The seven member science and drilling team spent ten days at Byrd station and vicinity to accomplish these tasks (Langway, 1992). Megascopic stratigraphic features, bulk density, and acidity (by Electric Conductivity Measurement, ECM) of the core were measured continuously in the field. A significant goal of this shallow ice core drilling operation was to investigate the near surface layers, using contemporary state-of-the-art laboratory techniques to overlap and extend to the surface the important environmental records obtained in the original deep ice core drilling program in 1968 (Ueada and Garfield, 1969). All of the 1968 ice core studies began 88.4 meters (reference horizon dated 1340 AD) below the 1968 snow surface, where continuous core recovery first began, due to limitations in the then existing drilling technology. As a result, more than 500 years of ice core chronology (top 88.4 m) from the 1989 A.D. surface at Byrd Station was not available for study prior to the undertaking of this project. The full research program for the NBY-89 ice core include the laboratory measurement and analysis of variety of physical and chemical properties performed by several research groups as listed in Table 1.1. The age of the ice core at the bottom was determined to be 1360 years (Langway et al., 1994). correlating the ECM, 8180, Dating was done by cross- and ionic chemistry stratigraphic records and with well recognized reference horizons of volcanic origin and cross-checked with the density profile and a total f3 curve for the upper firn layers. The overall dating accuracy is estimated as better than ±2 years. Table 1.1 Individual studies, investigators, affiliation, and current status of science studies program for the NBY-89 ice core (Updated from Langway, 1992). I Institute Investigator(s) Study Samplinga Status State University of New York at Buffalo C.C. Lang way, Jr., Physical stratigraphy C Finished H. Shoji, and co-workers Structure Physical properties Mechanical properties I C, I I Finished Finished In progress C.C. Langway, Jr., K. Osada, and co-workers Chemical stratigraphy Volcanic Layers Ionic concentration levels Organic chemistry Embedded solids C, I I C, I Finished Finished Finished In progress In progress University of Copenhagen H. Clausen, C. Hammer, C.C. Langway, Jr., and co-workers Acidity Oxygen isotopes $-activity levels Dust University of Colorado W. Dansgaard, S. Johnsen, and J. White Deuterium University of Bern H. Oeschger I I Finished Finished Finished In progress C. I In progress Carbon dioxide, methane, and nitrous oxide I In progress A. Neftal Hydrogen peroxide, I Finished B. Stauffer Beryllium-10 I In progress U. Stigenthaler, C.C. Langway, Jr., and co-workers Carbon- 13/carbon- 12 I In progress University of Rhode Island M. Bender and co-workers Oxygen, argon, and nitrogen isotopes in gas phase I In progress Massachusetts Institute of Technology I. Olmez, C.C. Langway, Jr., and co-workers Time-series trends of trace elements (THIS WORK) A new method for air content measurement ' C denotes continuous; I denotes intermittent C, I Finished I Finished 1.4. Scope of Research The Antarctic continent covers approximately 10% of the Earth's land surface with almost 99% of the continent permanently covered by snow and ice. Consequently, the predominant sources for the natural aerosol particles found in the ice are not of local origin but the result of long-range transport of aerosol and particulate matter from distant continental and oceanic areas of the globe. Transport is accomplished via upper and middle-troposphere trajectories with active gas to particle conversion reactions occurring during particle transport (Dick, 1991). The anthropogenic impact in Antarctica is minimal due to its distance from population centers. The long-range transport and relatively slow vertical mixing in the Antarctic atmosphere results in a homogeneous and size-segregated aerosol input to the Antarctic region (Cunningham and Zoller, 1981). Snow accumulates at high-elevation polar regions in a layered structure without melting due to the low temperatures throughout the year. In the higher elevation regions of Antarctica snow accumulation exceeds ablational processes and it is possible to observe seasonal variations in augered ice cores. One of the most powerful parameters used in dating polar ice cores is the oxygen isotope ratio but seasonal cycles are much more difficult to determine in Antarctica than Greenland. This method was first introduced into glaciology by Dansgaard (1954). The seasonal change of isotope ratios in precipitation is strongly related to the temperature during the condensation. The ratio of oxygen-18 and oxygen-16 isotopes in precipitation is larger during the summer season than the winter season due to higher temperatures during the summer precipitation (vapor pressure of H2160 is approximately 1 % higher than that of H2180 (Jouzel and Merlivat, 1984). The isotopic composition is measured by mass spectrometry and traditionally is expressed by the 8 function (Hammer, 1989) as, 8180 = [ (Rs-RsMow)/RsMow ]x1000 %c (1.1) where Rs and RsMow are the isotopic ratios of oxygen-18 and oxygen-16 in the sample and in standard mean ocean water, respectively. In a similar way hydrogen isotopes can also be used for dating purposes. In older ice, the diffusion process distorts the seasonality of the signal (Johnsen, 1977) and presents a difficulty in Antarctic cores. The temperature dependence of the oxygen and hydrogen isotopes can also be used to obtain past atmospheric temperatures. However, this relationship is not as straightforward due to the fact that the amount of original water vapor remaining in the air mass is determined by the complicated air mass mixing effects during transportation to the ice sheets (Charles et. al., 1994). These mixing effects include the location of the original moisture source, the trajectories of the moisture, and the mixing between sources with different isotope distillation levels. Another useful measuring parameter with a seasonal relationship in ice cores is the ECM technique (Electric Conductivity Measurement) (Hammer, 1983). This technique relies on the determination of acidity levels via direct electric conductivity measurements on frozen ice cores. The acidity contained in the deposited ice layers is mainly the result of precipitation on to the ice sheets of gas-derived atmospheric acids, mainly H250 4 and HNO 3. The increasing sulfate concentration levels in spring and summer snow deposits are thought to be the result of: photochemical formation of excess sulfate from its organic gaseous precursors (Legrand and Delmas, 1984), sea water, or volcanic sources. The origin of nitrate in polar precipitation is still controversial (Delmas and Legrand, 1989) but lightening was suggested as a possible source (Kley, 1983). The ECM method is fast and easy (with competent ice cores), with high resolution as long as the precipitation is acidic enough with a modest annual deposition rate and the seasonal signal is not distorted by possible conflicting seasonal trends of different acids or by volcanic origin acidic species (Hammer, 1989). Another method for ice core dating that shows regular seasonal dependence is microparticle concentration measurements. Since during the austral summer season the crustal aerosol production is higher and aerosol removal by precipitation is lower, there is an increasing aerosol transport into Antarctica mainly from the continental areas of the Southern Hemisphere (Tuncel et al., 1989). The most commonly applied technique for the measurements is the use of a Coulter microparticle counter. This technique has the advantage of in situ applicability but time consuming and subject to complicated interpretation of the results. Applications of the method have been made on both Greenland and Antarctic ice cores (Thompson, 1977; Hammer et. al., 1985). Some chemical species also show seasonal dependence, and thus can be used for stratigraphic dating purposes. S042-, N0 3-, and H20 2 are some examples. The degree of seasonality for any one of these species might vary with the site and time interval covered (Hammer, 1989). The seasonality of these species has been demonstrated both in Greenland and Antarctica (Langway et. al., 1977; Herron, 1982; Neftel et. al., 1984; Langway et. al., 1994) and have been shown to be an invaluable chemistry index marker for historically recorded volcanic eruptions (e.g., Tambora, Krakatau, Katmai). Any one of these parameters might show a strong or a weak seasonal signal for stratigraphic dating purposes, depending upon the location and period of deposition. Therefore, the best overall method for accurate dating is to use all of these methods on an ice core and cross-correlate the results. For example, in an exceptional and unique study, the estimated standard deviation for continuous dating of the 0 to 980 meters depth interval (1979 A.D - 625 B.C.) in Dye 3, Greenland deep ice core has been reported as ±5 years at the bottom by cross-correlating the results of oxygen isotopes, acidity, and microparticle measurement (Hammer et. al., 1986). In addition to these detailed and reliable stratigraphic dating methods, some other less accurate, but sometimes crucial, methods are also used for specific purposes. Although no in depth information about these methods will be given here, it might be useful to mention the basic principles. One of these methods that is always used for checking the accuracy of stratigraphic dating is called the "Reference Horizons" method and relies on the signatures of known global or regional scale events, such as volcanic eruptions and atmospheric nuclear bomb tests, on the polar ice sheets in a specific year. Two other independent, non-stratigraphic methods use naturally produced radioactive isotopes and ice flow phenomena as ice core dating parameters. These methods are especially useful for dating deep ice cores that have either thin annual layers due to plastic deformation, and, hence, whose seasonal signals are not clear enough for stratigraphic dating, or that have lost stratigraphy due to bed-rock scouring of their deep layers. In addition to their usefulness under these specific conditions, these methods might also be preferable where continuous stratigraphic analysis is not done or the ice core is not continuous. In the first method natural radioactive isotopes enclosed in ice as part of the water substance (3H), in form of aerosols (e.g., 32Si, 1'Be), or in the air bubbles (e.g., 39Ar, 14C, 81Kr) are used for dating (Stauffer, 1989). The latter of these methods combines ice flow modeling with the air content-sample depth relationship to date the deep ice layers. This method relies on the fact that the amount of air scavenged in the ice layers is proportional to the density of air (a function of pressure and temperature), and therefore, the elevation at the time of pore close-off, which is the complete isolation of air bubbles from the surrounding atmosphere (Langway, 1958; 1970; Shoji and Langway, Jr., 1989). Thus, the original elevation of a specific sample can be found by air content measurement (assuming the temperature history is known), and the total distance traveled by that sample through ice flow is calculated from the elevation difference and the slope of the ice sheet. Although the relationship between the air content and elevation might change due to the possible long-term modifications of atmospheric circulation patterns over the ice sheets, so the atmospheric pressure-elevation relationship (Martinerie et al., 1992), the scales of these changes are difficult to predict and ignored in the calculations. In addition, possible variation of pore volume per unit ice mass at the time of pore close-off also affects the relationship between the air content and elevation. It was shown that, this variation is mainly due to the temperature dependence of this parameter, and should be treated if the temperature data (from oxygen and/or deuterium isotopes analysis) is available for more accurate results (Raynaud and Lebel, 1979). Flow velocities of specific ice layers can be calculated from mathematical models, which employ physics laws and topographical information. In addition, these models can be verified by experimental borehole measurements on the ice flow trajectory (e.g., Russell-Head and Budd, 1979), and to certain extent by the satellite measurements on surface layer movement (e.g., Goldstein et al., 1993; Fahnestock et al., 1993). Finally, the sample age is calculated from the travel distance and velocity profile. 1.5. Basic Theory of the Instrumental Neutron Activation Analysis The Instrumental Neutron Activation Analysis (INAA) is a non-destructive multielement measurement technique that relies on the neutron capture and subsequent gammaray emissions of atomic nuclei. The following is a brief review of the basic theory. The historical development, applications, and in depth theory can be found in the extensive literature (e.g., Olmez, 1989; Knoll, 1989; Alfassi, 1990). When target nuclei in a sample are bombarded (irradiated) by low energy thermal neutrons, some of the nuclei are transformed into a metastable state by capturing the incoming neutrons. The main source of the thermal neutrons used for the irradiations of samples are nuclear research reactors that are specifically designed for high neutron flux output. The half-life of a nucleus in this excited state is very short (< 101 2 s) and it decays to a lower state by emitting "prompt" gamma-rays that can also be used for trace elements analysis by another technique (prompt gamma activation analysis). Following this (n, y) reaction, the radioactive nucleus will often further decay by electron emission, electron capture, or positron emission, and the resulting "daughter" nucleus will emit "delayed" gamma-rays. These are the gamma-rays used in instrumental neutron activation analysis. The case of an (n, y) reaction followed by positron decay is illustrated schematically in Figure 1.2. (X+n)* Prompt y's (PGAA) .. +n . Delayed y's (INAA) "'*f AX X A+lI A+ 1 zX Figure 1.2. Illustration of (n,y) reaction followed by Z f3 X decay. A calculation of the amount of activity induced in a sample at the end of the irradiation can be made using the following formula: Ag = 4oN(1-e - kt i ) (1.2) where, Ao : Activity at the end of the irradiation 0 : Flux [neutrons/cm 2-s] a : Neutron capture cross-section, in barns [barn= 10-24 cm2] X: Decay constant [In 2/t112] tl/2 : half-life N : Number of target nuclei in sample [nuclei/cm 3] ti: irradiation time In general there will be some delay between the end of irradiation and beginning of gamma-ray spectrum collection, and this time is called the "cooling time" (t,). Activity at the end of the cooling time will be (A=Ao e- Xtc ), and it is related to the gamma-ray spectrum peak area (interference and background corrected) by the following equation: A = Ao e- tC = P (1- td)(1-e - (1.3) s) where, A : Activity at the end of the cooling time P : Measured net gamma-ray spectrum peak area ts: Spectrum collection time td : Fractional "dead-time" Here, the "dead-time" correction is made to account for the time that analyzer is not active due to signal processing limitations. To calculate the absolute elemental concentrations from this equation, the efficiency of the detector should be known. There are however some drawbacks in using this direct calculation method. One is the dependence of detector efficiency on the counting geometry, and the other is the possibility of neutron flux change during the irradiation. To eliminate these uncertainties, a comparative method can be employed as in the case of this work. In this method, a standard material with an accurately known composition is irradiated along with the sample. The unknown weight of an element of interest in a sample can be found from the following relationship: Wsample - Asample Ws tandard As Astandard where, Wsample : Weight of element in sample Wstandard : Weight of element in standard Asample : Activity of sample at the end of irradiation Astandard : Activity of standard at the end of irradiation (1.4) 2. EXPERIMENTAL 2.1. Sample Selection and Preparation 2.1.1. Selection A total of 97 individual ice core samples were selected and measured by instrumental neutron activation analysis (INAA) to accomplish the main part of this study (Table 2.1). The samples cover both pre-industrial (< 1815 A.D.) and industrial timeperiods. Continuous sampling was done for the first 15 meters depth, which extends over the time-period between 1926 AD and 1989 AD. There are sixty-nine samples covering this time period. In addition, selective interval sampling was done between the 15 meter and 42 meter depths, which covered the overall time period from 1711 A.D. to 1926 A.D., to obtain background information. Twenty-eight samples, each covering approximately a one year time-interval for a 6 to 12 year continuous record, were obtained over this timeperiod. An additional 20 samples were carefully selected to measure specific episodic events, identified by other investigators working on this core (Langway and Osada, personal communication), and from a previous study (Keskin et al., 1992). Among these, a group of 12 samples (Table 2.2) were selected from the depth interval between 88.15 to 89.51 meter depths to locate the peak position of a previously observed (Keskin et al., 1992) rare earth elements enrichment around the 89 meter depth. Another group of 8 samples (Table 2.3) was selected from certain years that either directly coincide with historically known volcanic eruptions or show high sulfur concentrations (Langway and Osada, personal communication) as an indicator of volcanic eruptions. Additional 3 pairs of samples were selected from the time-period of 1711 AD to 1989 AD for quality control purposes (Table 2.4). Table 2-1. Ice core samples selected for the study of trace elements time-series trends. Sample No Sample ID 1IC1 1IC2 1IC3 2IC1 2IC2 2IC3 3IC1 3IC2 3IC3 3IC4 4IC1 4IC2 4IC3 4IC4 5IC1 5IC2 5IC3 5IC4 5IC5 6IC1 6IC2 6IC3 6IC4 6IC5 7IC1 7IC2 7IC3 7IC4 7IC5 8IC1 8IC2 8IC3 8IC4 8IC5 9IC1 9IC2 9IC3 9IC4 9IC5 10IC1 10IC2 10IC3 10IC4 10IC5 11IC1 11IC2 11IC3 11IC4 11IC5 Approximate Depth Interval [m] Date [A.D.] 1989 0.00-0.33 0.33-0.66 0.66-1.00 1.00-1.33 1.33-1.66 1.66-2.00 2.00-2.25 2.25-2.50 2.50-2.75 2.75-3.00 3.00-3.25 3.25-3.50 3.50-3.75 3.75-4.00 4.00-4.20 4.20-4.40 4.40-4.60 4.60-4.80 4.80-5.00 5.00-5.20 5.20-5.40 5.40-5.60 5.60-5.80 5.80-6.00 6.00-6.20 6.20-6.40 6.40-6.60 6.60-6.80 6.80-7.00 7.00-7.20 7.20-7.40 7.40-7.60 7.60-7.80 7.80-8.00 8.00-8.20 8.20-8.40 8.40-8.60 8.60-8.80 8.80-9.00 9.00-9.20 9.20-9.40 9.40-9.60 9.60-9.80 9.80-10.00 10.00-10.20 10.20-10.40 10.40-10.60 10.60-10.80 10.80-11.00 1989 1989 1988 1988 1987 1986 1986-85 1985-84 1984 1984 1984-83 1982 1982 1981 1980 1980-79 1979-78 1978 1977 1976 1976-75 1975 1974 1973 1972 1972-71 1970 1969 1969 1968 1967 1966 1966-65 1964 1963 1962-61 1960 1959 1958 1958-57 1956 1955 1954 1954-53 1952 1951 1950 1950-49 Sample No Sample ID 12IC1 12IC2 12IC3 12IC4 12IC5 13IC1 13IC2 131C3 13IC4 13IC5 14IC1 14IC2 14IC3 14IC4 14IC5 15IC1 15IC2 15IC3 15IC4 15IC5 16IC1 17IC1 18IC1 19IC1 20IC1 21IC1 22IC1 23IC1 241C1 25IC1 26IC1 27IC1 28IC1 29IC1 30IC1 31IC1 32IC1 33IC1 34IC1 35IC1 36IC1 37IC1 38IC1 39IC1 40IC1 41IC1 42IC1 431C1 Depth Approximate Interval [m] Date [A.D.] 11.00-11.20 1948 11.20-11.40 1947-46 11.40-11.60 1946-45 11.60-11.80 1945-44 11.80-12.00 1944-43 12.00-12.20 1942-41 12.20-12.40 1941 12.40-12.60 1941-40 12.60-12.80 1940-39 12.80-13.00 1939-38 13.00-13.20 1938-37 13.20-13.40 1937-36 13.40-13.60 1936-35 13.60-13.80 1935-34 13.80-14.00 1934-33 14.00-14.20 1933-32 14.20-14.40 1931-30 14.40-14.60 1930-29 14.60-14.80 1929-28 14.80-15.00 1927-26 15.00-15.17 1926-25 16.00-16.17 1919-18 17.00-17.17 1913-12 18.00-18.17 1907-06 19.00-19.17 1901-00 20.00-20.17 1895-94 21.00-21.17 1889-88 22.00-22.16 1881-80 23.00-23.16 1870-69 24.00-24.16 1858-57 25.00-25.16 1846-45 26.00-26.15 1837-36 27.00-27.15 1830-29 28.00-28.15 1822-21 29.00-29.15 1815-14 30.00-30.15 1807-06 31.00-31.15 1799-98 32.00-32.14 1791-90 33.00-33.14 1782-81 34.00-34.14 1773-72 35.00-35.14 1766-65 36.00-36.14 1759-58 37.00-37.14 1752-51 38.00-38.14 1744-43 39.00-39.14 1736-35 40.00-40.14 1728-27 41.00-41.14 1720-19 42.00-42.14 1711-10 Table 2-2. Ice core samples selected to locate a previously observed Rare Earth Elements enrichment peak around 89 meters depth. Sample No Sample ID Depth Interval [m] Approximate Date [A.D.] 1 2 3 4 5 6 7 8 9 10 11 12 89LC1 89LC2 89LC3 89LC4 89LC5 89LC6 89LC7 89LC8 90LC1 90LC2 90LC3 90LC4 88.15-88.26 88.26-88.37 88.37-88.48 88.48-88.59 88.59-88.70 88.70-88.81 88.81-88.92 88.92-89.02 89.02-89.14 89.14-89.26 89.26-89.38 89.38-89.51 1341 1341 1340 1339 1338 1338 1337 1336 1335 1334 1333 1332 Table 2-3. Ice core samples selected from volcanic eruption horizons. Sample No Sample ID Depth Interval [m] Approximate Date [A.D.] 1 2 3 4 5 6 7 8 22V1 22V2 22V3 22V4 29V1 29V2 51V1 51V2 21.38-21.53 21.53-21.70 21.70-21.85 21.85-22.00 28.15-28.28 28.28-28.40 50.02-50.17 50.17-50.32 1887-86 1885-84 1884-83 1883-82 1821-20 1820-19 1647-46 1646-45 Table 2-4. Ice core samples selected for quality control purpose. Sample No Sample ID Depth Interval [m] Approximate Date [A.D.] 1 2 3 4 5 6 17CP1 17CP2 23CP1 23CP2 28CP1 28CP2 16.75-17.00 16.75-17.00 22.75-23.00 22.75-23.00 27.17-27.42 27.17-27.42 1915-14 1915-14 1869-68 1869-68 1829-28 1829-28 2.1.2. Preparation After the field drilling operation in 1989, the NBY-89 ice cores were transported and stored at the Ice Core Laboratory (ICL) at SUNY, Buffalo. Each 1 meter ice core section was placed in a polyethylene bag and kept in a protective capped tubing. These tubes were stored on shelves in a cold room at -28 C'. Prior to the chemistry preparation sampling process, all the necessary tools for core surface cleaning, such as pure nickel plates and polyethylene bags, were pre-cleaned at the Environmental Research & Radiochemistry (ER&R) Division Laboratory at Massachusetts Institute of Technology Nuclear Reactor Laboratory (MIT-NRL). A 99.5 % pure-Ni metal plate was used to surface clean the outer layer of each sample to remove possible surface contaminants induced during the field core drilling and handling. The plates (10x12x0.079 cm) were thoroughly cleaned with methanol, rinsed with deionized (DI) water, and stored in acid cleaned polyethylene bags prior to use. Nickel is one of the elements to which instrumental neutron activation analysis (INAA) is not very sensitive, and its hardness is proper to shave solid ice surfaces. The polyethylene zip-lock ice storage bags were soaked inside and out with 25 % nitric acid by volume for 36 hours. They were rinsed four times with DI water and dried in a clean bench after turning them inside out again. Laboratory grade Mallinckrodtfm brand nitric acid was used for cleaning processes. The DI water for dilution and rinsing bags was produced by a ComingMT LD-5 system with Bamstead TM MEGA-PURE® high purity cartridges (resistivity > 2MG). After preparation, the necessary equipment was transported to ICL at SUNY, Buffalo. The core sampling process was completed in three separate trips. After the removal of each ice core tube from the inner cold room (-28 Co) to the sample preparation room (-23 Co), the core was removed from its container and samples from pre-determined depths were cut with a band saw. Each sample weighed 150 to 200 grams before surface cleaning (portions of the core segments from the same depth remain untouched and are available for possible use in other research projects). After the cutting step, each sample was placed horizontally in a clamp on a clean bench and its surface shaved with the clean nickel plate. Shaving was done by layers all around the sample surface, removing an approximately 6 to 10 mm thick layer. The clean inner part was enclosed in an acid cleaned polyethylene bag. Holding this bag and the contained ice sample firmly, the sample was separated from the clamped section by creating a weak point with the nickel plate and bending it to break off. This way, the surface of the shaved ice samples was not touched by any material except the acid cleaned bags. The remaining section of each sample within the clamp was discarded. A simple schematic of this stage is given in Figure 2.1. Weak point < > Clamp Bending lll noiteid - ....... > Shaved ice Polyethylene < block bai Figure 2-1. Schematic illustration of ice sample surface cleaning setup. After closing the zip-lock, the sample was placed into another zip-lock bag (a label card was placed in this second bag) and stored in a freezer in the cold room. The net sample weights after the surface cleaning process were 60 to 100 grams each. During the sample cleaning process polyethylene gloves, protective clothing, and masks were used to prevent any possible contamination to the samples. 2.1.3. Pre-concentration by Freeze-drying The majority of the elemental concentrations in Antarctic ice samples are at nanogram and sub-nanogram levels (Batifol et al., 1989; Gbrlach and Boutron, 1992). These low concentration levels require the use of over 50 gram of ice for each sample. This is necessary to substantially exceed the INAA detection limits of most of the elements for accurate concentration measurements. However, there are limitations on the volume of a sample that can be irradiated and counted without some artifacts. The INAA technique is best suited for solid samples in small volumes. Small volumes prevent the neutron self-shielding effect in the inner region of sample matrix, and solid samples simplify and speed up the handling phase. A second limitation is the difficulty involved during gamma-spectrum collection of radioactive samples. Sample counting geometry is an important factor for detector collection efficiency, which in turn directly affects the measurement errors due to counting statistics. In addition, sample self-shielding of gamma-rays in large volumes and contamination problems related to liquid samples are unwanted features for large volume liquid sample analysis. Two volume reduction techniques are applicable to ice samples: freeze-drying ice samples in the solid state, and evaporation of melted ice samples in liquid state. Each technique has advantages and disadvantages. Evaporation of melted ice is substantially faster than freeze-drying. However, one problem of evaporation is the possibility of loss of some elements by volatilization (Tanizaki, 1990). Although for a very slow, low temperature, non-boiling evaporation process this loss may be prevented (G6rlach and Boutron, 1990), the prevention of loss of volatile species and contamination from the surrounding environment throughout the long evaporation time is a difficult task. For freeze-drying, the major drawbacks are the slow turnaround time and possible loss of elements with high vapor pressure. However, slow separation of water molecules from only the frozen surface layer makes it less likely that some elements will be carried out of the sample container. In addition, the closed, under-vacuum surrounding prevents external contamination. It has been shown previously (Harrison, 1977) that as a result of the freeze-drying process of water samples with known amounts of radioactive elemental tracers, all but two (Hg and I) out of 22 elements studied showed retention yields of 96 % to 100 %. Considering that there are only a few elements that can be lost significantly due to their high vapor pressure (Hg, I, and Cl), freeze-drying was used here as the sample pre-concentration technique. The basic principle of this technique is separation of ice and the residue it contains by sublimating the ice under vacuum. To prevent the sublimating water vapor from going into the vacuum line and to produce humidity free chamber, a cooled coil (- 55 oC) is used to condense it. A simple schematic diagram of the freeze- drying system is Figure 2.2. The low elemental concentration levels in the ice samples require the contribution from the sample containers of the freeze-dried samples to be minimal. Although it would be possible to freeze-dry each 50 to 80 grams of ice without transferring it from its original polyethylene bag, the contribution of each element from the container material itself would be high compared to the net concentration levels expected in ice residue. To overcome this problem, small (approximately 100 milligram), thin-wall polyethylene bags were used as sample containers during freeze-drying (bag selection and characterization steps are explained in detail in section 2.2.3). It was possible to freeze-dry 4 to 6 grams of ice at a time using a multi-step freeze-drying process. ....> Vacuum chamber 'QUUU' : ,> Condenser coil .> Vacuum pump Ice sample Dlug Figure 2-2. Schematic illustration of the freeze-drying system. The polyethylene sample container bags and transparent polypropylene vials (0 35 mmx60 mm), used as rigid holders for container bags, were acid cleaned in the same way ice storage bags were cleaned as explained previously. They were dried in a clean bench on clean analytical grade washed filter papers (S&S M T , # 589). After they were dried, the holders and bags inside them were placed into clean, closed, glass vacuum desiccators. Before closing the cover of each desiccator, clean analytical grade filter papers were placed on top of the holder vials to prevent any possible contamination during closing and opening of the desiccator cover. Since the freeze-drying capacity of the system (LABCONCO , Model 75200) is limited to a certain amount of ice mass, 30 individual sample containers could be processed in one or more vacuum desiccators at a time, which meant approximately 150 grams of total ice mass. This limitation is due to the need for holding a high enough vacuum level inside the vacuum chamber and, thus, each desiccator, to prevent the melting of ice, which essentially stayed at room temperature. Melting would cause the suction of partially frozen samples out of the container bags, which would result in the loss of elements from the containers. For our research sample preparation and 15 consecutive stage freeze-drying cycles (six group of 30 samples, including empty control bags) were completed in over a one and a half year period. 2.2. Instrumental Neutron Activation Analysis 2.2.1. Irradiations At the end of the fifteen consecutive freeze-drying steps, each desiccator was taken onto a clean-bench, and each sample bag was heat-sealed to prevent any loss of residue during the sample irradiations. After heat-sealing, each bag was folded three times on an analytical grade washed filter paper, in a way to minimize the sample size and shape variations to assure a constant geometry during counting. After the folding, each bag was placed into another acid-cleaned polyethylene bag, heat-sealed, and labeled. The same steps were repeated for empty bags, which were used as blank control samples. During these sample preparation steps, standard clean-room procedures were followed. Since comparative INAA was being used as the measurement technique, NIST certified standard reference materials were used as standard samples in the analysis. The standard and control samples were "Coal Fly Ash - 1633a" and "Orchard Leaves - 1571", respectively. For short (10 minutes) irradiations one 5 to 6 mg standard sample and one 8 to 9 mg control sample was used for each detector. To prevent excessive Na-24 isotope production, and therefore high background noise, these standard and control sample weights were reduced for long (24 hours) irradiations, and two 2-3 mg standard samples and one 5-6 mg control sample was used for each detector. The same clean room procedures were followed during the preparation of the reference material samples. In the short irradiation process, elements with radioactive isotopes with half-lives between a few minutes to a few hours were determined. For this group of elements, 10 minute irradiation duration was determined as an optimum time in the 8x10 12 n/cm 2-sec thermal neutron flux. Irradiations were done by inserting four samples inside a sample carrier (rabbit), and sending the container to the reactor irradiation location from the radioactive sample handling room by an automatic pneumatic transfer system. In general, there are two considerations for the number of samples to be irradiated at one time for short periods. The first is the number of high resolution gamma-ray detectors available for use, and the second is the time necessary to prepare the samples for counting at the end of irradiations, which should be kept minimal not to allow very short-lived elements to decay before counting. After 10 minutes irradiation, the samples were allowed to decay for one minute. The purpose of this was to reduce the working dose by letting some very shortlived (on the order of seconds) elements (mainly fluorine) to decay, since they can not be determined at trace concentrations anyway, due to the comparatively long sample packaging and transfer time required before the gamma-ray spectrum collections. After receiving each irradiation container from the reactor through the pneumatic system, the dose equivalent value at 30 cm was measured by a Geiger counter in the radioactive sample handling laboratory. If this value was over five mrem/hr, it has been cooled until the level fell below this value, and then samples were removed with long tweezers onto a radioactive sample preparation hood, which is shielded for worker radiation protection. Each outer sample bag was cut open with scissors and the inner sample bag was placed into a clean, non-radioactive bag using separate clean TeflonTM tweezers. After finishing the sample packaging step, the four samples were quickly transferred to the sample counting room in a lead container. The same steps were followed for standard and control samples. After an average decay period of one week, the previously irradiated samples (10 minutes for short-lived element determination) were placed into an irradiation container for longer irradiation. Irradiation duration of 24 hours was chosen as optimum in the 8x 1012 n/cm 2-sec thermal neutron flux. A maximum of eight samples per detector (total of 32) can be irradiated at a time to allow for the counting of samples before a considerable level of decay. In addition, a total of eight standard samples and four control samples were placed into each container for four detectors. At the end of each irradiation, samples were allowed to decay two to three days to reduce the activity from some highly abundant and comparatively short-lived (of the order of hours) elements (mainly sodium) and prevent detector saturation and high background values during the gamma-ray spectrum collection. After this decay period, samples were packaged as explained above and transferred to the counting laboratory in lead containers. 2.2.2. Gamma-Ray Spectroscopy The gamma rays emitted from the radioisotopes were counted using high resolution HPGe detectors coupled with an 8192-channel pulse-height analyzer; spectra were analyzed using computer-directed programs. As explained in the previous section, irradiation durations, cooling and spectrum collection times required optimizing depending upon the sample matrix and elements of interest. Gamma-ray spectrums of each sample were collected in four stages, two after the short irradiation and two after the long irradiation. At each stage, certain elements with similar half-lives were determined. The counting duration was chosen not to be longer than three half-lives of the shortest-lived element within a group of elements that will be determined in a particular stage. After each short-period irradiation (10 minutes), four samples were transferred to the Counting Laboratory. After the transfer, each sample was placed on top of a detector at 7 cm distance. Sample-detector distance was chosen so that at least ninety percent of the time detector would be collecting gamma-rays emitted by the sample. In other words, detector "dead time" would not exceed ten percent of the real counting time. The occurrence of "dead time" is due to the coupled detector analog-digital conversion electronic circuitry requirement of a certain time period to process an incident gamma-ray from a radioactive isotope. During this signal processing step of an incident gamma-ray, the detector is blocked electronically from other incident gamma-rays. The selection of under ten percent "dead time" was made to minimize the spectral distortions. The first stage (shorts-1) of counting was done for the net counting period of six minutes and forty seconds. Following this, the second stage (Shorts-2) of counting was done right after finishing the first stage for the net counting period of thirty minutes with the same counting geometry. In these counting stages elements given in Table 2.5 were scanned for concentration determination. Figure 2.3 shows a representative gamma-ray spectrum for short lived (shorts-1) elements. Gamma-ray spectrum collections following each long-period irradiation (24 hours) were done in two stages. Since activity of the samples were low enough not to cause high "dead time" values, each sample was placed right on top of detector to improve counting efficiency. The first stage (Longs-1) of countings were done two to three days after the irradiations with the counting periods of around six hours. The second stage (Longs-2) countings were done two to three weeks after the first ones to let the medium half-lived isotopes to decay extensively, so decrease the background and interferences for the second stage elements. The counting periods were around twelve hours with the same sample position as the previous stage. The elements scanned for concentration determination during these counting stages are also given in Table 2.5. After these collection stages, each spectrum was checked on the computer screen for possible interferences on some of the peaks that were used by the computer program 8Z-IV - .. 8S-1J - ZS-A OO i l I LWW vo acu U.. m ID 4a LLOUf 9S-UN I-r!I for the determination of element concentrations. Interactive peak fitting was performed on each low intensity gamma-ray peak and portions of the spectra, wherever the detector resolution would not be enough to separate peak doublets. After these final checks, elemental concentrations were calculated by the computer program and printed out as hard copies. Table 2-5. Elements scanned in ice core samples by INAA. Element Half-life Key Energy Scanning Element Half-life Key Energy Scanning [keV] Stage [kevy] Stage Al 2.25 m 1779.0 Short-i Te 1.35 d 149.8 Long-1 Cu 5.10 m 1039.3 " W 0.996 d 479.6 Mg 9.45 m 843.8 " U 2.36 d 106.4 Ti 5.76 m 320.1 " Yb 4.19 d 396.3 V 3.76 m 1434.1 " Ce 0.089 y 145.4 Long-2 Ba 1.396 h 165.9 Short-2 Cs 2.065 y 795.9 Cl* 0.62 h 1642.4 " Cr 0.076 y 320.1 " Dy 2.33 h 94.7 " Co 5.271 y 1173.2 Ga 14.1 h 834.1 " Eu 13.48 y 1408.0 In 0.903 h 417.0 " Gd 0.662 y 97.4 I 0.417 h 442.9 " Hf 0.116 y 482.1 Ir 0.202 y 316.5 2.578 h 846.8 " Mn Na** 14.96 h 1368.6 " Fe 0.122 y 1099.2 208.4 388.4 " Lu 0.018 y Sr 0.291 h Sb 2.70 d 564.1 Long-1 Nd 0.03 y 91.1 It Rb 0.051 y 1076.7 As 1.096 d 559.1 "I Br 1.471 d 554.3 "I Sc 0.23 y 889.3 Cd 2.228 d 336.3 "I Se 0.328 y 264.7 Au 2.694 d 411.8 "I Ta 0.314 y 1221.4 0.198 y 298.6 487.0 " Tb La 1.678 d Mo 2.748 d 140.5 " Th 0.066 y 311.9 i 0.668 y 1115.5 " Zn 0.515 d 1524.6 K " Zr 0.175 y 724.2 Sm 1.929 d 103.2 (*) Determined by direct liquid sample analysis. (**) Determined by both freeze-dried and liquid sample analysis. 2.2.3. Blank Correction The ice residue which accumulated in the polyethylene bags following the freezedrying process was extremely small, and could not be removed from the bags. Therefore, bag blank contributions to the ice elemental concentration values had to be corrected. Before starting the sample freeze-drying process, a preliminary study was done to determine the elemental concentrations in different batches of polyethylene bags. As a result, a single batch of bags was chosen because of lower elemental total mass per bag. The final concentration results and uncertainties for the bag blanks were obtained from the 24 procedural blank bags placed into the desiccators together with the samples during the sample pre-concentration stages, and they are summarized in Table 2.6. The total uncertainties (CT) were calculated by including both the measured variations in the procedural blank bags (aB) and the analytical uncertainties (OA) by the following equation: ( T = ( +Y A (2.1) Contribution of blank correction to the overall measurement uncertainty of the samples is discussed in section 3.1. Table 2-6. Container bag-blank concentrations. Element Concentration Uncertainty Number Element Concentration Uncertainty Number [ng/g-bag] of Bags [ng/-bag] of Na 13986 879 24 Sb 2.3 1.0 17 Mg 5218 2891 24 Te 0.46 0.38 4 Al 14386 1398 24 I 23 9 8 C1 1536 555 24 Cs 3.2 1.6 10 K 3499 743 19 Ba 280 131 12 Sc 2.8 0.2 19 La 6.3 1.1 19 Ti 1843 1180 15 Ce 18 6 19 V 121 8 24 Nd 76 76 19 Cr 105 37 19 Sm 1.13 0.16 19 Mn 49 12 24 Eu 0.68 0.38 10 Fe 7019 1905 18 Gd 1.4 0.7 12 Co 9.1 4.0 19 Tb 0.49 0.54 11 Cu 381 269 7 Dy 1.8 0.9 18 Zn 117 65 19 Yb 0.66 0.33 16 Ga 94 66 6 Lu 0.14 0.21 11 As 3.1 0.6 17 Hf 0.86 0.35 10 Se 13 9 17 Ta 8.9 5.0 5 Br 79 32 17 W 1.8 0.7 16 Rb 65 51 3 Ir 0.067 0.046 4 Sr 218 168 9 Au 0.077 0.045 13 Zr 273 226 2 Hg 1.5 0.8 19 Mo 15 4 17 Th 2.3 0.8 18 Cd 2.5 1.4 14 U 3.7 1.3 16 In 0.81 0.36 10 2.3. Air Content Determination Before starting the experiments with the actual ice core samples, the total experimental uncertainty of this method had to be determined. One contribution to the uncertainty comes from the adsorption of air from the working environment on the sample surface during the preparation and analysis. This contribution was determined by creating air-free ice, and subjecting it to the identical sampling and analysis stages to those of real samples. A special experimental apparatus was designed and constructed (Figure 2.4) to produce air-free ice. Vacuum Liquid N2 ooling coil Figure 2.4. Illustration of air-free ice production system. A polycarbonate desiccator was modified to contain a cooling coil and a purging nozzle. The cooling coil (0 8 cm x 15 cm) and purging nozzle were made of a copper tubing (0 8 mm) and aluminum tubing (0 2 mm), respectively. The purging nozzle had a spiral shaped tip (0 5.5 cm) with small holes in it. Approximately 250 ml of water in a cylindrical, polyethylene container (0 6 cm x 14 cm) was placed in the middle of the cooling coil. The purging nozzle was placed in the water container, and the desiccator cover was closed. The time chart of the experiment is given in Table 2.7. During the first stage of the experiment a portion of air was released from the water by the application of a vacuum. Following this, the water was purged with CO 2 gas to enhance the release of the remaining air. The purging gas was free of Ar since the amount of air in the ice was determined by measuring the Ar (argon) content by gamma-ray spectroscopy. The gas used for purging was 99.995 % pure CO 2 with a maximum Ar (argon) content of 5 ppmv, which is negligible in comparison of the argon in air of 9340 ppmv. Table 2.7. Air-free ice production time chart. Time [m] Action Temperature [°C] (inside the desiccator) 0 Vacuum started 25 60 CO 2 flow started 20 19 65 Liquid N2 flow started 118 CO 2 flow stopped -18 -18 119 Vacuum stopped 120 Liquid N2 flow stopped -18 Since comparative INAA was used to measure the air content in ice, standards were prepared from atmospheric air, whose Ar content is well known. For accurate volume determinations two polyethylene vials with caps were filled with water, as seen in the Figure 2.5. Net water masses were converted to volumes by dividing them by the density of water, 0.998 g/ml, at 20 'C. After these steps the vials were emptied and dried. They were closed by new caps, and left in the indoor air environment overnight to allow the interior and exterior air pressures to equilibrate. Later their caps were heat sealed to ensure air tightness. Water outlet hole Water injector Cap Polyethylene vial Figure 2.5. Air standard vial volume determination. In addition to the artificially produced, air-free ice samples, a total of nine specimens from the NBY-89 ice core were sampled for air content analyses to demonstrate the applicability of the method. Five were chosen specifically to determine the precision of the method (Figure 2.6a), and the remaining four to understand the seasonal impact on the air content over a continuous year (Figure 2.6b) of snow accumulation. 1/11 rlll -- IVI.ZU m 101.22 m S I S5 S2_S1 S41 iS4-" 85 IUI.3U m 101.52 m S6 .s . ?57 101.54 m 101.24 m iS 101.56 m I-S 101.58 m EI----~-~ (a) (b) Figure 2.6. Ice sample selection for air content measurement. All samples, including the air-free blanks, were cut in the same way as discussed in section 3.1 using a special pure nickel blade, the average sample weight was about 3 g. Immediately after the cutting step, the samples were placed into marked polyethylene bags, and the bags heat-sealed. The samples were stored in a freezer at -28 'C until the time of irradiation. Three ice samples were irradiated at a time in a sample carrier for two minutes in 8x10 1 2 n/cm 2-sec neutron flux. Since the specific irradiation location used is at approximately room temperature, the duration of irradiation did not cause observable surface melting. Once samples were received in the radioactive sample handling laboratory, they were taken out of the carrier, and each bag was cut open. The samples were allowed melt on their external surfaces to flush possible surface air contamination prior to the irradiation by keeping them at room temperature for a few minutes. Following this, each sample was placed into an unirradiated polyethylene bag, and the bag heatsealed. Each sample then was placed into a second labeled bag for safer handling and better Ar gas containment in case of melting during gamma-ray counting for the Ar-41 radioactive isotope. After these steps, samples were transferred to the counting laboratory and counted for 15 min at 7 cm distance from the detector surface. At the end of the counting time, partial melting had occurred in each sample bag. However, the air released by the melting of the samples should have still stayed inside the container bags. The heat sealed air standard vials also were irradiated in the same way and counted for 5 min on each detector. After letting the samples decay (t1 / 2=1.82 h for Ar-41 isotope) for twelve hours, they were weighed to calculate the net weights by subtracting the previously measured container bag weights from the total measured sample weights. Finally the air content of each sample was determined by comparing its Ar-41 gamma-ray peak to that of the standard sample. 3. RESULTS AND DISCUSSION 3.1. Data Quality Assessment In the past, many trace element measurements on ice cores suffered from unknown contamination problems occurring during core drilling, sample handling and preparation stages (Wolff and Peel, 1985; Gbrlach and Boutron, 1992). Although the measurement of elements in nanogram and sub-nanogram concentration levels is already a major challenge, to achieve this goal without contaminating the sample is a more difficult task. Recognizing the problems in earlier studies in this field, special attention was given from the beginning of this research to contamination prevention and quality control. A special difficulty inherent to the study of ice cores is related to the availability of an adequate sample. The NBY-89 ice core has only a 10 cm diameter, and any one ice specimen cannot be replaced. The first 50 meters (1647 A.D.-1989 A.D.) of the NBY-89 core was first used heavily by other researchers and the amount available for this investigation was terribly limited. Since the pores in the core are not totally closed to the atmosphere over the upper 40 to 50 meters (firn structure), it would not be representative enough to use samples from deeper levels, whose pores are closed to the atmosphere (ice structure), for quality control purposes, although those samples were more available for use. Due to this limitation, only three pairs of samples (see Table 2.4) were obtained from the first 50 meters depth of the core for duplicate analysis and cross-profiling. Cross-profiling can also be considered a duplicate analysis of a sample, if there is not a contamination problem. The preparation of these three pairs of samples was somewhat more complicated than that of the rest of the samples. This is due to the need of separating a surface cleaned sample into two parts (interior and exterior) without contaminating it. Since the sample weights were limited, it would not be possible to repeat the surface cleaning by shaving outer layers from each piece after the cutting step. Therefore, a clean way to divide the samples was necessary. One way of doing this could be melting and collecting the outer layers in another container on a clean-bench. However, this was not possible for these samples, since they had a permeable firn structure, which would allow the melted water to diffuse into the inner section of a sample. To overcome this problem, a high purity nickel blade was manufactured and mounted on a drill for cutting the samples on a clean-bench. A simple schematic of this system is given in Figure 3.1. Clean polyethylene Nickel blade Ice sample Figure 3-1. Sketch of sampling system for cross-profiling sample preparation. During the preparation of these samples, one section of a pair (17CP2) was dropped and contaminated accidentally. Therefore, only the remaining two pairs were analyzed for their elemental concentration profiles. The results for the inner and outer sections of each pair are given in Table 3.1. As can be seen in Figure 3.2, the inner and outer sections of each pair showed a good agreement within the experimental errors for most of the elements. For a few elements, the results show a considerable difference between the two sections. However, this difference does not follow a pattern of contamination, which would always show higher concentrations for outer (CP2) layers. Therefore, this disagreement is more likely a result of a combination of factors, namely the cutting process to separate inner and outer parts, sample inhomogeneity, and analytical uncertainty. Although a pure nickel blade and clean room procedures were used to cut the samples, some contamination may have occurred for these elements. Another possibility is a real concentration difference due to the difference of each snow flake deposited. However, this argument is somewhat unlikely, since there are thousands of snow flakes in each section that should create a homogenized deposition of aerosols. Since this extra cutting step was not involved for other samples used in this work, it is likely that elemental concentration results on those samples will have less analytical uncertainty compared to the control samples. Considering that for elements, such as aluminum and sodium, that are readily available in the sampling and preparation environments, the results show satisfactory agreement between the inner and outer sections, any serious and systematic contamination problem influencing the results can be ruled out. Table 3-1. Concentration results for cross-profiling samples [ng/g-ice]. 23CP1 (inner) Na Al Cl K Sc Ti V Cr Mn Fe Co Zn As Br Mo Cd Sb La 44±5 10±4 91+11 23CP2 (outer) 51±5 9±+3 88±10 0.060±-0.003 0.051+0.003 10.0±4.2 0.15+0.04 14.4±3.8 0.15±+0.03 0.37±+0.06 3.25±0.05 12.9±3.6 0.65±+0.09 0.64±0.03 11.5±4.8 0.009±0.010 4.7±0.4 0.004+0.002 0.08±0.09 0.007+0.012 0.002+0.004 0.0731+0.005 0.001+0.003 0.014+0.008 10.6±0.7 0.005±+0.002 0.08-0.07 0.022-0.011 0.009-0.005 0.067+0.004 0.006_+0.002 . 28CP1 (inner) 42+5 12+4 87+11 1.2+3.5 0.058+0.004 5.7±3.0 0.01+0.02 0.39-0.08 0.81±0.03 7.0±4.3 0.010-0.009 6.7±0.5 0.001-0.002 0.12±+0.08 0.011-0.011 0.003+0.005 0.026+0.003 0.006+0.003 28CP2 (outer) 48+5 5±3 108+13 5.9±4.6 0.057+0.003 3.1+2.6 0.01-0.02 1.01-0.07 0.87±0.03 15.1±4.0 0.017_-0.008 11.0±0.8 0.006--0.002 0.19±0.09 0.007-0.009 0.014_+0.007 0.037+0.003 0.002+0.002 Aluminum Scandium ^ ^• ' 0.07 18 0.06- o14 - ~ :55 ::::::~:~;Z·5 0.05- . i~i~i~j 55555' 55 0.04- 0 68 S20- I115 ...... ' 23CP2 28CP1 ·5·- t.:~:~:~::::~s ~ ·~5· ·;·;·:·.·.zz:: 0.03S0.02- 55:~ .v: ... ........ ··Y I·-1·123CP1 ~ 5i~5·~ i·5·Zi··:~: ·ns;s·~l ~i 5555555 :-:s~ ·5·Si·I ~s~ 0. 28CP2 23CP2 23CP1 28CP1 ~ s~ 55 :::::::::~::: :~:~::: :·:5·:·:· ·: cs :5;:;~;:·· ,s~: -~ 28CP2 Sample ID Sample ID Manganese Iron 20 18 16 ,'9 3 0 0 02 12 0, 108 0 23CP1 23CP2 28CP1 5 -,f 0. Mle 0 uV1 e0 16 14 0 -- 23CP1 28CP2 23CP2 28CP1 Sample ID Sample ID Titanium Lanthanum 28CP2 -- 20 000 tl &lt• 16 "b14 12 10 u U 0.008 - S0.007 - 8 o 0.006 o 0.005 0.004 6• 0.003 G 0.002 0.000 2 IJ 23CP1 23CP2 28CP1 Sample ID 28CP2 iT - - - - - 71 . r55- .· ..... - 0.000 23CP1 23CP2 28CP1 28CP2 Sample ID Figure 3-2. Elemental concentrations measured in cross-profiling samples (CP1 and CP2 denote inner and outer samples, respectively. Molybdenum Cadmium II^^^' ""4 0.024 0.022 - 8 T 0.030 - 0.020S0.018- M 0.025 L =S0.014 0.016- - .0 0.012 - ....... .... 8 0.010- 0.005 - U 0.002- 0.000 - I _________ 23CP1 0.000 - 23CP2 28CP1 ,· · ~-~ - 0.015 0.010- So0.0080.0060.004- 22222 ~f~5~-' 0.020 - iiji~ii~ t~ pr 23CP1 28CP2 ~i -~s ,, ______·____ 23CP2 ~f; ~ "·"~""` """""' 28CP1 28CP2 l Sample ID Sample ID Sodium Chlorine 6so T ggg--g -50 0 * T 120 - ~ 100 - ~::::~i~ 40 ~ 030-2 ~t~-~ 20 ~ss~ 10 ~ ff ~ts~ ". . "'" T X:K :·: Sso .... ..... .. .. .. ......... .... 60 ~i ~· F~ss~s~ ~~s~ ss~ ·z~f;ss t~ s-.~s~ ~ .sss·`·- 0 '----- 23CP1 23CP2 28CP1 28CP2 23CP1 Sample ID 23CP2 28CP1 28CP2 Sample ID 0.0 0 Arsenic Antimony 0.009 0.008 8 7 0.016 T - ro 0.007 - c 0.006 o 0.005 V - •4 3 0o. i - o 0.0X o0 - U 1 0.X - 0.0(1 - 0 0.01 I - 23CP1 23CP2 $81 28CP1 Sample ID Figure 3-2. (cont.) 28CP2 0.004 S0.003 S0.002 0.000 0.000 23CP1 23CP2 28CP1 Sample ID 28CP2 1 Bromine Chromium 0.30 0.25 - 0.20 . - S015 . ... .. .. . 0=1 ... .... - C! ........ . ...... ..... .. 0.10 S0.05 ....... ... U 0 0 0.00 23CP1 23CP2 28CP1 28CP2 23CP1 23CP2 28CP1 Sample ID Sample ID Vanadium Zinc 28CP2 0.20 018 0.16 b 0.14 0.12 S010 0.08 = 0.06 0.04 O 002 0.00 - - - - - S12- IT T U T.... . . ... ....... ... .. .. .. .. . . ... .. . ......... .... :~~ .... .. .. .. .. . S10- S 0 - E 8- 0642- - :e.'x~;s -:···j ~~ i ~:·1~ · :LSOZ :·: X~~`l ·:~~t ::· ··~~j ~ ~ ·~:oll· 0- - 23CP1 23CP2 28CP1 Sample ID Figure 3-2. ~;O · · (cont.) 28CP2 23CP1 23CP2 28CP1 Sample ID .-. .-.. . . 28CP2 After this assessment, the procedures for blank correction and experimental uncertainty calculation steps will be described. As was discussed in section 2.2.3, sample container blanks were analyzed and their average elemental concentrations were calculated. Define [Ck]sample and [Ckblank as the concentrations of element k in a sample (includes ice and its container) and blank container, Mice and Mblank as the ice mass and container mass. One can then calculate the net concentration in ice as follows: [Ck]ice = { ( [CkIsample X (Mice + Mblank) ) ( [Cklblank X Mblank-) }/ Mice (3.1) Since, Msample = (Mice + Mblan) = (3.2) ce the final result can be given as: [Ck]ice = [Ck]sample - { ( [Ck]blank X Mblank) / Mice } (3.3) Error propagation calculations were performed in each step to obtain the final concentrations. Uncertainties due to sample weighing, container blank measurement, and gamma-ray spectrometry counting statistics were all included. Although it would be desirable to include the uncertainty due to the preparation and handling steps of samples, assessment of this uncertainty was not possible for the reasons mentioned above. One alternative to assess this uncertainty could be the use of artificially created duplicate samples. However, it would be very difficult to simulate the real samples which had firn structure (pores open to the atmosphere) and low elemental concentrations. Median uncertainties of the final concentrations for the 18 elements with significant observation frequency (> 70 %) are given in Table 3.2. For most of the elements, the major contribution to the total uncertainty on the final concentrations came from the blank correction step for sample containers. Plots of cumulative percent data with respect to overall relative uncertainty are given in Figure A. 1.in Appendix A. Table 3-2. Median uncertainties on the concentrations of elements with significant observation frequency. Number of Median % Element Number of Median % Element samples Uncertainty samples Uncertainty 97 15 Mn 97 11 Na 95 37 Fe 72 123 Mg 97 13 Zn 97 24 Al 97 30 As 97 13 C1 95 48 Br 73 57 K 66 128 90 Mo 52 Sc 95 8 Sb 73 112 Ti 89 La 77 89 V 78 84 56 Sm 97 9 Cr 3.2. Time-Series Trends The final concentrations for the main group of 97 samples analyzed and the statistical summary information regarding this data set are given respectively in Table A.1 and Table A.2 in Appendix A. Since there was more than one sample in some years, and some samples coincided with two or more years, the depth-date calibration data (Table A.3 in Appendix A) provided by Osada (1994, personal communication) was used to calculate the concentrations for a specific year of interest. Note that the provided dating calibration data is more precise for the period between 1941 and 1989 compared to the period between 1711 and 1941. Annual concentrations were calculated by weighted averaging of the samples associated with a specific year, and the percent length of each sample portion that coincided with that year was used as the weight coefficient for that sample. These final annual average concentrations are given in Table A.4 in Appendix A. In these calculations missing values were replaced by the most probable values obtained by log-normal distribution curve fitting to the original data, and the results of these calculations are given in Table A.5 in Appendix A. The physical and mathematical bases of this step are explained in section 3.3.2. One parameter that is helpful in evaluating whether or not a source category might be an important contributor for an element of interest is called enrichment factor (EF) and defined as (Gordon et al., 1971): IX] EFx= [Y]I sample sa(3.4) {[Y] reference where, [X] : Concentration of element of interest [Y] : Concentration of element used as marker for a source category This parameter compares the ratio of these elements in a sample medium to that of a reference medium such as soil, sea water, or plume from a volcano. In a physical sense, the EF parameter shows whether a source category might be a candidate for the observed concentrations of an element or not. When this parameter is less than 5, it is assumed that the element of interest is not enriched significantly with respect to the reference medium. This implies that medium might be a candidate as a source category for that element. In an ideal situation, the EF of an element that originates only from a specific source of interest (e.g., soil dust) should be 1 at the receptor site. However, due to the inhomogeneities on the source compositions and possible physical and chemical fractionation effects during the transport to the receptor site, this parameter deviates from 1 in most cases. That is why EF values as high as 5 may not be claimed with certainty to be real in most cases. Crustal and marine enrichment factors for some elements measured in this work are given in Table B. 1 and Table B.2 in Appendix B respectively. Aluminum and sodium were used as the marker elements for crustal and marine EF calculations. The crustal contribution to the total sodium concentrations was subtracted to find the marine component, and these corrected sea-salt sodium concentrations (Na.) were used in the marine EF calculations. Taylor's (1972) standard crustal and Goldberg's (1963) standard marine reference compositions were used in the calculations. One can safely assume that the elements measured were deposited almost entirely as components of aerosols, although some volatile and highly soluble elements might also have been deposited in individual form. In Antarctica, aerosols can be classified into four general categories with respect to their origin, as crustal, marine, volcanic, and anthropogenic (Dick, 1991; Tuncel and Zoller, 1995). Although there might be some extraterrestrial contribution, it is thought to be very minor in comparison (Tuncel et al., 1989). A few certain elements are almost entirely associated with a certain aerosol source category. 3.2.1. The Marine Aerosols The two major elemental components of marine aerosol used as marker elements are sodium and chlorine, mostly associated with the sea-salt component. Since there might be considerable loss of chlorine due to chemical fractionation during transport (Legrand and Delmas, 1988), and significant contribution of chlorine from volcanic emissions from time to time (Kyle and Meeker, 1990), Sodium was chosen as the marker element for marine aerosols. The time-series trend for sea-salt sodium (Na.) concentrations is given in Figure 3.3. In this plot, and the ones following, the abscissa is not given to scale, so the data covering the time period from 1926 A.D. to 1989 A.D., which has continuous annual data, is clear. The data is comparatively sparse prior to that period. As can be seen, a significant and sharp increase of concentration occurs between 1969 and 1989. Concentrations fluctuate from year to year throughout the time periods both before and after this increase, but do not show a monotone trend within a time period. The ratio of average concentrations in the post and pre 1969 period is 3. This is a sudden and significant increase in the last two decades. Although not as significant, there seems to be another similar era between 1870 and 1895. However, due to the sparse data, it is difficult to draw any conclusions for this earlier period. As mentioned earlier, due to the high vapor pressure of chlorine, the possibility of volatilization loss during the freeze-drying steps exists, so this element was not determined from the freeze-dried samples but directly from small (1-2 g) melted samples. Therefore, it is more meaningful to compare the chlorine concentrations with the sodium concentrations measured within the same small samples. In further discussions these specific measurement results for chlorine and sodium will be designated as Cliq and Naiq, respectively. Figure 3.4 shows the concentration time-series trends of ClIq and Nalq. As 0 0 0 (" '- O 0 C0 0 0- - C0 Dr [93!-6/Bu] uoqleilueouoo t O0 C4 °..• 2 Q et 0 U U S r0 DA *M C [33!-2/2u] uo!jlua55o3 55 seen, these elements are well correlated with each other, as expected because of their common origin. They also exhibit a sharp increase within the last two decades. Two possible causes of this increase are a more efficient aerosol mixing over the Antarctic ice sheets and/or an enhanced source strength in the southern oceans. It is helpful to consider the production and transport mechanisms of marine aerosols to be able to comment further on this behavior. The marine aerosols, made up of mostly sea-salt particles, are released from the sea surface into the air mainly by two well-known release mechanisms, which are the surface films in breaking waves on the ocean and jet drops from rising air bubbles (Shaw, 1989). Air can be brought into the water by raindrops, snowflakes, supersaturation (produced by temperature rises), and whitecaps, with the last being the most effective (Blanchard and Woodcock, 1980). Sea-salt aerosols were shown to be as small as 0.03 gIm (M6szaros and Vissy, 1974) and as large as several hundred microns (Junge and Jaenicke, 1971) in the ocean surface air, although only 1-2 % are small enough (r < 0.5 pm) to arrive on the Antarctic ice sheets (Junge, 1963). It was also shown that wind is an important factor in the source strength of sea-salt aerosols (Fairall et al., 1983). An increasing number of aerosol concentrations in the Aitken mode (r < 0.1 pm), associated with higher sea surface temperatures and stronger surface winds, were observed throughout the tropical trade wind region (Lal and Kapoor, 1992). Increasing storminess is one of the reasons for the enhanced marine aerosol concentrations observed in central parts of Antarctica during the winter seasons (Cunningham and Zoller, 1981; Parungo et al., 1981; Tuncel et al., 1989). The general source region for the warm, marine aerosol enriched air seems to be the Ross-Amundsen-Bellingshausen seas (Bodhaine, et al., 1986). The seasonal behavior of the sea-ice cover in these regions was suggested as a defining factor for the observed discreet sea-salt episodes that accompany warming during wintersummer transitions (Hogan et al., 1990). The transport of marine aerosol from the oceans to the Antarctic ice sheets takes place through advection and turbulent diffusion in the atmosphere. However, this transport is somewhat complicated by the special meteorological features around Antarctica. The especially intense and highly zonal circumpolar vortex is an important barrier for the aerosol transport to the interior of the ice sheets (van Loon, 1972). The circulating cyclonic storm systems that form over the oceans in the vicinity of the polar front (Wilson, 1968) and migrate clockwise around the continent due to the circumpolar vortex can rarely penetrate this barrier (Shaw, 1979). The cyclonic wind system and surface temperature inversion are weaker during the summer season, which results in more favorable conditions for aerosol transport into the ice sheets compared to the winter season (Cunningham and Zoller, 1981; Hogan et al., 1984; Tuncel et al., 1989). Therefore, lower marine aerosol concentrations observed during the summer months should be the result of calm weather conditions that results in the less efficient production and intrusion of aerosols into the ice sheets. An upper troposphere-lower stratosphere mixing was demonstrated, especially during the spring season, both in lower latitude regions and Antarctica (Danielsen, 1968; Ellsaesser, 1983; Wagenbach et al., 1988; Murphey et al., 1991), which might indicate a stratospheric transport route for lowerlatitude water vapor and aerosols into the Antarctic atmosphere. However, it was shown that not this transport mechanism but an enhanced source of aerosols in the maritime polar air source region was the main reason for the observed spring time marine aerosol concentration increase at the South Pole (Hogan et al., 1990). In light of these details about the production and transport of marine aerosols, production is more likely to be the defining factor for the enhanced deposition within the last two decades, although the possibility of more efficient aerosol mixing over the Antarctic ice sheets cannot be ruled out. Since Antarctica is surrounded by oceans, the source strength for marine aerosol is stronger and transport range is shorter than those of other aerosols. Therefore, it is expected that significant changes in the oceans and their surroundings will affect the marine aerosol concentration in Antarctica faster and stronger than the other aerosol categories. Changes in the sea-ice extent, sea surface temperature, ocean currents, ocean microbial activity, wind strength, precipitation patterns, and air temperature are among those that can affect marine aerosol concentrations. In reality, all of these parameters are components of the ocean-atmosphere system and they strongly affect one another. For newly produced sea-salt aerosol, the chlorine to sodium ratio reflects the ratio within the sea-water which is 1.8 (Loureiro et al., 1992). However, several studies showed chlorine depletion in the aged marine aerosols collected in the Antarctic atmosphere (Maenhaut et al., 1979; Bodhaine et al., 1987; Legrand and Delmas, 1988; Guerzoni et al., 1992;), as well as in the tropical marine atmosphere (Kritz and Rancher, 1980). Measurements of the marine aerosols collected over the tropical and equatorial Pacific showed a 40 % loss of chlorine from the submicron particles, while no loss was observed for larger particles (Raemdonck et al., 1986). The overall reaction responsible for this depletion is thought to be (Keene et al., 1986); NaCl (p) + H => Na' (p) + HCI (g) where "p" and "g" denote the particulate and gas phases, respectively. As a source of hydrogen ions, atmospheric acids H2SO4 and HNO 3 have been proposed (Legrand and Delmas, 1988). Figure 3.5 shows the Cliq to Nalq concentration ratios obtained in this work. The overall average ratio is 2.03-+0.04. Since these measurements were done in small samples, which cannot represent annual average concentrations, these values should be taken as informational values due to the possible seasonal differences in this ratio. However, one interesting feature on this plot is the apparent drop of chlorine-to-sodium ratio after 1940. Pre- and post-1940 ratios are 2.24±0.08 and 1.87±0.04, respectively. Two reasons could explain this behavior; one is the possibility of increasing crustal contribution to the total sodium budget during the post-1940 period, and the second one is the possibility of higher chlorine contribution from volcanic activities to the total chlorine budget during the pre-1940 period. As will be discussed in sections 3.2.3 and 3.2.4, both of these possibilities seem likely to contribute to this trend. The average marine Enrichment Factor (EFm) results calculated from Table B.2 in Appendix B indicate that oceans are the main source for the elements sodium (EFm,=1.1_0.03), magnesium (EFm=1.3_0.2), chlorine (EFm=3.8±0.4), and bromine (Efm=1.2+0.1). (EFm=1.3±0.03), potassium a, S-j O0 U'c: o or= E 0 C) "I3 *Mi ! f• 1,i [bB N / [bIi ] 3.2.2. Possible El Nifio Events-Marine Aerosols Connection When one looks at the major changes in the ocean-atmosphere system within the last two decades, one sees that the occurrence of the so-called El Niflo Southern Oscillation (ENSO) events in the Pacific Ocean increased considerably compared to the previous five decades (Fig 3.6). Although it is not the scope nor the intention of this work to get into much detail about this natural phenomenon, a short summary might be useful before getting into further discussion. The term El Nifio (Spanish for the Christ Child) was originally used for a weak warm coastal ocean current which annually flows southwards along the coasts of Ecuador and Peru in the first three months of the year (Burroughs, 1992). In some years, abnormally high and persisting temperatures disrupt the ecological balance much more seriously and in scientific usage the term El Niflo has come to be associated with these extreme inter-annual events. The Southern Oscillation, named by Sir Gilbert Walker in his papers in the 1920's and 1930's, originally referred to the oscillating atmospheric pressure difference between the Pacific and Indian Oceans and its connection with temperature and rainfall patterns (Burroughs, 1992). Essentially, this pressure difference is the driving force for the easterly trade winds from the eastern to western Pacific. Later in the 1960's, Jacob Bjerknes was the first scientist to see a connection between unusually warm seasurface temperatures and the accompanying heavy rainfall in the eastern equatorial Pacific and weak easterlies (Wallace and Vogel, 1994). This discovery led to the recognition that the El Niflo and Southern Oscillation are parts of the same phenomenon which is now called ENSO. During normal years, the easterly winds that blow along the equator and the southeasterly winds that blow along Peru's and Ecuador's coasts drag the warm surface water along with them towards the western Pacific region, and cold deep water replaces this warm surface water. This regular movement of ocean currents ceases during El Nifio years, when the weakening easterly trade winds are no longer able to drag the warm surface water from the eastern to western Pacific. As a result, the sea-surface temperature distribution and the precipitation pattern become drastically different than in normal years. 6861 861 861 PýAl 3861 8961 C961 4861 1861 D861 6L61 RL61 'L6 T ýL61 EL61 9L61 EL61 ZL61 IL6I 9L61 6961 R96I £961 9961 P961 £96I Z961 1961 0961 1£61 £561 0r' 9561 R561 - £661 0 RSC61 1-61 a) V) C) r•.*.::::::*.•.*..•.*.••.: ::::..•::•:::•.:.::•:: 0 ... ::*•*::.•*.*:.*::::::: .::::.••:::*:+ • :•::•;; <• :.:.::*.:.:.• £I'61 0,6 1 < I'6•61 € .**::::::::::**•*..•..::::::..*:*•..*::::• £661 9661 u C ý'6•61 661 £•61 zt6i It6I 4 61 £61 0. *::::•.:*::::::5::*:5*:• r 5:.:.::••:*::::•::**::••:::*•:.:::•:•: L::::: o ýS61 S'61 6f16 1 ,I-,,T ,0 1061 cE61 EE61 E6 C1 6Z61 6Z6 1 PZ61 6161 E161 4061 1061 E681 6881 1881 0)l 00 a- 0a- 0 r-.-c 11C 0'4 O 03 D£8I MTL 4181 D081 5L8 858 56 REST11, 5191 E66L I 16LT 9££8 0/) l)..::.*: 0.:. :*:. . .: .: .*.... :.. . .: pf8L 6SL1I PUT1 I'I'L1 ELI-I C 8ZL1 tIL1 11L1 leIOS OU!N 13 L During these years, sea-surface temperatures and precipitation rates increase in the central and eastern Pacific regions, while decreasing in the western Pacific region. Although the exact reason(s) for the initiation of ENSO event is(are) not known, its effects on the climate are quite visible today. Changing temperature, precipitation, and wind patterns especially in Australia, South America, and part of Africa are drastic. After this short review one can comment on the possible connection to the observed marine aerosol concentration increase in Antarctica (this work). The observed changes on the sea-ice extent, sea-surface temperature distribution, precipitation patterns, and storminess over the Pacific Ocean during the ENSO events are likely to affect the marine aerosol production. It was already mentioned in a review article (Carleton, 1992) that the sea-ice-circulation anomalies might have some connection to the ENSO, and the extent of the sea-ice has an effect on the migration of moisture sources important for the sea-salt aerosol accumulation on the Antarctic ice sheets. It was also mentioned by the same author that signatures recorded in ice cores hold the best promise to obtain longerterm data. Also the Pacific sector of West Antarctica, from which the NBY-89 ice core used in this work was obtained, appears to hold promise for detecting the ENSO signal, although some problems might arise from the dominance of the annual snowfall regime by relatively few cyclonic events. Therefore, the enhanced marine aerosol concentrations observed in this work might have some relationship to the increasing ENSO activity within the last two decades. One way of checking this hypothesis is to search for possible periodicities in the Na, concentration time-series data and to compare these to the previously identified periodicities on the ENSO related parameters, such as sea level pressure (SLP), sea surface temperature (SST), and precipitation patterns in the Pacific region. For this purpose a computer program developed by Khalil and Moraes (1995) was used. This program relies on the Linear Least Squares Spectral Analysis (LLSSA) method which is a simple method of time-series analysis based upon linear least squares curve fitting. The main idea of this method is to find the best fit sinusoid to the data for each member of a set of periods and then assign a weight factor to each fit by calculating the power (square of the amplitude of the fit) for that period (Khalil and Moraes, 1995). The major advantages of this method are the possibility of unevenly spaced data analysis for any length period and easy determination of statistical significance for each period. Detailed information about this method can be found in literature (e.g., Bloomfield, 1976; Moraes and Khalil, 1993; Khalil and Moraes, 1995). The power spectrum of the Na. time-series data (1711 AD-1989 AD) for periods between 2.1 and 15 years are given in Figure 3.7. In spite of its highest power value, the 2.0 year period was eliminated due to its very high uncertainty. Figure 3.8 shows the measured Na. time-series data (1926 AD-1989 AD) together with the curve fitting using the most significant period (4.2 years). In this fit the general trend seems quite close to the real data except the absolute concentration values. This difference comes from the use of a single period for the fitting. When the top eight periods were fitted (2 year period excluded) to the same data, the two curves showed much better agreement as expected (Figure 3.9). The power spectrum calculation results for two different time intervals are summarized in Table 3.3. Table 3-3. Periodicities and their power values in Na,, data set. Time Period Period power 1 [year] (AD) 2.0 802+±1806 1711-1989 3.8 149+129 4.2 165±137 9.1 136±123 11.7 139+127 31.4 367+_195 34.6 319±186 46.5 310±180 55.8 525±221 98.1 686-243 160.0 735±283 1926-1989 2.0 1292±12690 3.8 191+191 4.2 207_+199 5.7 149+170 7.5 122+155 9.1 124±156 11.7 107+145 17.1 119±155 607_±322 34.6 Ih %&j vv C11 L' 6 o -o L aE 8 "6 H o. 88"S CO E . c,© E- o r3 oM I 0 I C) o 0 Co 0 i I 0 0 Cj 0 JEMOd I 0 I 0 CD c I 0 a I 0 0 - -- I -- 066 1 6861 8861 L861 9861 £861 V861 £861 861 9 . .4.......... 4.44j 1861 0861 6Z 61 8L61 LL6I 9L 61 SLZ6 EL61 VL61 CL61 1L61 .R" ~,,......4·~ 4"Ss': .'rr OL61 6961 8961 L961 9961 £961 ZR96 £961 1966 1961 0961 6£61 8061 L£61 .. I~~=··.,.... c© cr •2 rr;. -- C 9Z S Z 66 11 C• 9S6[ 9£ 61 ££61 9Z61 6961 6•6[ l£·61 ZS6Z CE61 0£ ?61 9961 S96 06PG 6Z 61I 81F61 Lf761 0) V) 86E6 ZL;; IYfr 9I61 9i61 9E61 9C61 SE61 BE61 R Z ZZ66 I1 OV61 4 - 6t61 .4. - I I I I I I [azT-f/2u] uoIrlwnuauo D I i I 6C61 8E61 L961 I- cl 066 1 66 4 ............ 8*61 4 a01LILIIC 2.. LPL961 ~1( r Z961 99 6961 P961~8L c E9616 _ 8 6Zt +r :: ... . .... .. ... 19 61 0961 •6L61 9161 LL6I.y ~~~rc:_~p I~c~ ~~l~l~s 4.a............I 9L61 6 i R. Z6 1 t,EZ6,Z61 1Z61 . -1 • >.,~ I Z 61 C 9161 VL61 £161 *~ 66 3 ..J wow - OZ961 L961 9961 ,-r, .-- ~ -. -" - j)L 1ý1 () C,L v C0 * * c7 1961a 1i 0961 6S61< Cd ,i= ' L£;61"" VS61 r Zc61 9£61~' 1S61 0£61 6t;61 9i~61 Ot~-l S S61 61 zS I S6 1 £V~61 t, S611 £V61 Z¶761 E 61 V - ci: i -' - v ~ -Tor:~ uCZ 4 Y1 f 6 E61 . ;f L£61 961 " v 9£61 S E6 1 c ~6 061 91 S6~61 • % "• 61 Z f " 19161 /I [ 00 \C N I O U61 I s It N 0 ~ ; In an independent study, periodicities on the variation of the Southern Oscillation (SO) were identified from accumulated Darwin (Australia) mean sea level pressure (MSLP) anomalies (Zhang and Casey, 1992). This analysis, covering the interval between 1882 and 1989, showed significant periods between 2 and 6 years. It was also reported that (Burroughs, 1992) the spectral analysis of the annual SO figures in Australia and New Zealand showed a series of peaks, the most marked of which are at 3, 3.75, around 6, around 9, and 10 to 12 years. As can be seen, these periods are quite similar to the periods found in this work. Therefore, the hypothesis about the possible relationship between the ENSO related events and sea-salt sodium concentrations deposited on the Antarctic ice sheets has some validity. However, a word of caution is necessary because of the possible problems with the time-series analysis of environmental data as summarized by Burroughs (1992), and more detailed and sophisticated statistical analysis studies are necessary before reaching any final conclusion. 3.2.3. Crustal Aerosols and Their Possible Role in Polar Atmosphere Crustal aerosols are customarily represented by the marker elements aluminum, scandium, and, to a certain extent, iron. In this work aluminum was used as the marker element for crustal aerosols. The time-series trend of aluminum concentrations is given in Figure 3.10. As seen, there is a significant upward trend after 1980, while the overall concentration trend is reasonably constant prior to that time with the exception of sporadic episodes in certain years. The ratio of average concentrations in the post- and pre-1980 periods is 3.0±0.1. This is a significant increase in a considerably short time period. Although not as significant, there is also some enhancement between 1969 and 1980. The ratio of the average concentration between 1969 and 1980 to that of the pre1969 period is 1.4±-0.1. This increase is also evident for scandium which is another mainly crustal origin element (Figure 3.11). Although the aluminum and scandium concentration trends are similar in general, they differ at certain years mainly because of the higher measurement uncertainties for scandium caused by near-detection limit concentrations in the samples. The number of missing values for certain years might bias scandium. Figure 3.12 shows the crustal enrichment factor time-series trend for scandium. Even though the EF for scandium is expected to be close to one, it is only 0.22+0.03 on the average. This behavior was also observed in some previous atmospheric aerosol sampling programs in Antarctica (Tuncel et al., 1989). Although it seems that the post-1948 era has higher scandium EFs compared to that of the pre-1948 era, high uncertainties on the data make it difficult to be sure about the reality of this behavior. The average crustal Enrichment Factor (EFt) results calculated from the Table B. 1 in Appendix B indicate that crustal contribution is the main source for most of the elements measured in this work (Table 3.4). However, this conclusion does not rule out the possibility of some other episodic contributions from specific sources, such as volcanoes. Among these potentially crustal origin elements, magnesium and potassium also come from the oceans at the same degree as shown in the previous section. Table 3-4. Average EFe results for potentially crustal origin elements. Element Average EFe Element Average EFe Mg K Sc Ti V Mn 1.9±0.3 1.9±+0.2 0.2±0.03 2.9_0.5 1.5±0.2 1.2±0.05 Fe Co La Ce Sm 1.4±0.1 1.0±0.2 2.1±0.2 2.0±0.3 1.6±0.1 Since 99 % of Antarctica is covered with ice, and the continent is far from other land areas in the Southern Hemisphere, the major mechanism for crustal aerosol input into the Antarctic ice sheets is expected to be long range atmospheric transport from distant source regions. Although some small arid areas in Antarctica, such as the Transantarctic Mountains, also were proposed recently as potential crustal aerosol source areas for the ice sheets (Guerzoni et al., 1992), these frozen soil areas are not considered as important sources for the central parts of the continent (Cunningham and Zoller, 1981; Dick, 1991). Therefore, increasing crustal aerosol input into the ice sheets should be either due to a, C3C r•r 0 0 0 *M ·-!E om 0= F.v, S. [0 0 [9o!-6/6uI uo!lailue3uoo 0oo 7961 ESTC,6 59•6K tar< 87- Ii 696•• L96 ct Col C- 6n796 1-£6 c•-, cn b v: at61 E *M r4688- ET i-'b C *M 6ti?975 'J•8•U8 S•$Y ''81 6617 •ia7 6ii [6• 39_7 95= o cC c t: *. om <) I- 0) 0 E f.> En [L=iv] 3i3 increasing source strength at some distant land areas and/or increasing transportation toward the Antarctic from those areas. This view is supported by the observed seasonal crustal aerosol concentration trends, which show summer maximum and winter minimum (Cunningham and Zoller, 1981; Wagenbach et al., 1988; Tuncel et al., 1989). Since scavenging is more effective during the winter season due to increasing precipitation, aerosols are less likely to survive the long-range transport over the oceans. In addition, the cyclonic wind system and surface temperature inversion are weaker during the summer season, which result in more favorable conditions for aerosol transport into the ice sheets compared to the winter season (Cunningham and Zoller, 1981; Hogan et al., 1984; Tuncel et al., 1989). Deserts, which cover approximately one-third of the global land surface, are the most important crustal aerosol sources (Jaenicke, 1993). Due to extreme physical and chemical weathering processes, some parts of deserts (dry valleys) produce very high numbers of small size aerosols that can participate in long-range transport (d'Almeida and Schiitz, 1983). Wind patterns over the deserts affect both aerosol production and transport into the atmosphere. The prime candidates for providing crustal aerosols in Antarctic ice sheets are the Kalahari (south Africa) and Atacoma (south America) deserts and the deserts of central and western Australia (Shaw, 1989). Results from a modeling effort using the Atmospheric General Circulation Model (AGCM) showed that dust contribution to West Antarctica, where Byrd Station is located, is dominated by Australia and South America during the austral summer season and by Australia during the austral winter season for the present day conditions (Gaudichet et al., 1992). Since the increasing crustal aerosol concentration trend observed after 1980 is quite monotone, it seems more likely that this reflects changes in production rather than transportation. It is somewhat unlikely that the transportation efficiency would be increasing continuously over the last decade. It also seems unlikely that the local aerosol source strength would increase in this way. Therefore, the remaining possibility of increasing aerosol source strength in the distant Southern Hemisphere land areas seems likely to be the reason. After the major 1982-83 ENSO event, the changing weather patterns in the Pacific created large scale droughts in parts of Australia, Indonesia, South America, and Africa, as well as disastrous forest fires in Indonesia and Australia (Wallace and Vogel, 1994). Considering the frequent ENSO events over the last two decades, it is highly likely that enhanced aerosol production could be caused by increasing erodable soil areas, mainly due to the desertification of inner land areas and sea level drop in the Western Pacific region during ENSO events. In addition to the changes in aerosol source strength, changing weather patterns (Wallace and Vogel, 1994) also might have affected the transport of these aerosols. However, it is not known with certainty whether or not these weather pattern changes affected the aerosol transport over the Antarctic ice sheet. It was the general assumption until recently that crustal aerosols were produced larger than several tenths of a micrometer, and most of the measurements in deserts and arid land areas have been designed for those sizes (Jaenicke, 1993). One of the reasons for this assumption was the impracticality of smaller size particle production by physical weathering processes (Shaw, 1989). In addition, the domination of particle mass distribution by larger particles (Gomes and Gillette, 1992) attracted more interest for their measurement, so that approximate mass deposition fluxes could be obtained. However, it was suggested later that dissolution of water-soluble minerals might reduce the size of particles (Schroeder, 1985), and particles of radii as small as 0.01 pm with frequent number distribution peaks below 0.1 pm were measured (Junge, 1977; d'Almeida and Schiitz, 1983). It was reported that background aerosol originating from the continents dominates the aerosol load in the middle and upper troposphere almost globally, and it has the size range of 0.03 pm to 20 gm (Junge, 1977). On the other hand, marine aerosol is removed efficiently over the oceans due to the high humidity and cloudiness together with its high solubility (Junge, 1977). As mentioned before, a certain amount of mixing occurs between the upper troposphere and lower stratosphere. Therefore, it is possible that some of the crustal aerosol in the upper troposphere might mix into the stratosphere during its transport to the Antarctic. The residence times of aerosols in the upper troposphere and stratosphere can be up to a month or a year, respectively (Barrie, 1985), and are long enough for longrange transport of these aerosols into the Antarctic atmosphere. This possible stratospheric residence of crustal aerosols brings up the question of whether or not there might be some impact on the heterogeneous chemistry of the stratosphere due to the increasing crustal aerosol concentrations. One important and famous example of the heterogeneous chemistry is the spring-time stratospheric ozone depletion mechanism in Antarctica. Following this example, the annual average ozone concentrations homogenized between the 600 S and 900 S latitudes from the Total Ozone Mapping Spectrometer (TOMS) ozone data set provided by NASA Goddard Ozone Processing Team were retrieved for the period between 1979 and 1989. Figure 3.13 shows the timeseries trends of ozone and aluminum concentrations between 1979 and 1989. The reason to start with the year 1979 is the lack of satellite measurement data for globally averaged values prior to this date. However, land-based measurements for the Antarctic spring time ozone concentrations did not show a trend prior to late 1970's (Bojkov, 1986). Interestingly, a strong negative correlation (-0.9) exists between these two parameters and the linear regression lines fitted to the data have very close slopes with opposite signs (4.4 for ozone and +4.8 for aluminum). Although the short time period (11 years) prevents drawing a strong conclusion from this correlation, it is worth thinking about this result in light of the recent findings about the importance of heterogeneous chemical processes, mainly due to the formation of polar stratospheric clouds (PSC), on the formation of the Antarctic ozone hole. General information about the ozone depletion mechanisms is summarized in Appendix E. Aerosol concentrations in the stratosphere are much lower than those of the troposphere. The most abundant aerosol type in the stratosphere is sulfuric acid solution droplets, which are formed by the condensation of sulfuric acid gas that is produced from highly available sulfur dioxide gas through photochemical reactions (Hamill and Toon, 1991). These sulfate aerosols, with diameters around 0.14 gm, contain approximately 25 % water and 75 % sulfuric acid and are believed to be in a supercooled liquid state. In addition, particulate material of natural and anthropogenic origin can also penetrate to the stratosphere from the troposphere and outer space. Among the terrestrial aerosol sources, strong volcanic explosions are best known for their high gaseous and particulate matter injection capability into the stratosphere. In a similar way, particulate material originating from the ablation of meteorites entering the earth's atmosphere is assumed to be the major CA E E 0661 0 6861 S L961 E S861 t861 ct 0 ES61 E-C•,, Z6 1f861 L 0 - 0 O%61 c, oC o 6L61 SL61 n; n 0 CN CN4 In 0 0 0 V-" [a!o-B/Bu] pup [Ilufl uosqoG] 0 o !s extraterrestrial source contribution to the stratosphere. However, information about the stratospheric particulate material composition is limited due to the complexity of sampling at those altitudes, together with the very low concentrations and small sizes of particles. If the increasing crustal aerosol deposition trend observed in these Antarctic ice samples is an indication of increasing production and transport from the mid-latitude regions through the upper and middle tropospheric layers as discussed earlier, it might be possible for some of the smaller size aerosols to penetrate into the stratospheric layers and participate either in the nucleation processes to form the polar stratospheric cloud particles (PSC) (see Appendix E) or directly in the heterogeneous reaction mechanisms. This participation might be in different ways, such as nucleating sulfuric acid and water vapor as sulfuric acid solution droplets or frozen particles, nucleating water ice particles, nucleating nitric acid trihydrate (NAT) particles (see Appendix E), or catalyzing heterogeneous reactions directly on the particle surfaces. It was already suggested that heterogeneous reaction mechanisms at mid-latitude stratospheric regions are likely to be responsible for the observed ozone depletion trends after large scale volcanic eruptions. For example, column NO 2 depletion was observed at northern mid-latitudes during the winter of 1982-83, following the eruption of El Chich6n in Mexico (Hofmann and Solomon, 1989). Similarly, low stratospheric ozone levels were observed in the mid-latitudes in the months following the eruption of Mount Pinatubo in Philippines (Grant et al., 1992). One example of volcanic impact on the polar regions is the observed ozone reduction simultaneously with higher aerosol concentrations in the lower stratosphere in Antarctica after the August 1991 eruption of Mount Hudson in Chile (Hofmann et al, 1992). However, outside the polar caps, the temperature in the stratosphere seldom drops below -63 'C, which is considerably warmer than the PSC formation temperatures by known mechanisms (Garcia, 1994). Therefore, some other mechanisms must exist for the observed ozone depletion events due to heterogeneous processes. An attractive possibility is PSC formation at higher temperatures due to interaction with other solid particulate matter. It is already known from the tropospheric cloud formation mechanisms that cloud droplets can stay in a supercooled liquid state at much lower temperatures than the freezing point unless they are activated as ice nuclei (Barrie, 1985). This activation mechanism is known to be selective since only some of the cloud droplets become ice nuclei and can form snowflakes through condensation growth (Barrie, 1985). It is also known that clay minerals covered with soluble sulfur particles are the most efficient ice nuclei starters (Kumai, 1976). They cause ice nuclei formation either by acting as nucleation particles or by causing ice shattering when they come in contact with supercooled droplets (Beard, 1992). In addition to increasing heterogeneous reactions, these aerosols can cause increasing denitrification and dehydration of the stratosphere through sedimentation of large cloud droplets or frozen particles due to selective condensation growth process, thus causing further ozone depletion due to the reasons discussed in Appendix E. Another impact of the increasing crustal aerosol production and the subsequent deposition on the oceans might be on the organic sulfur production in oceans due to increasing marine biological activity. Iron is known to be essential for the growth of microorganisms in the oceans, and aeolian mineral dust is the principal source of marine iron in many areas (Idso, 1992). Recently it was suggested that iron and sulfur cycles in both the atmosphere and the ocean may be closely coupled (Zhuang et al., 1992). Although the relationship between marine biological activity and the sulfur production rate is not fully understood (Charlson et al., 1987), it is safe to assume that the net sulfur emission to the atmosphere will be affected. For example, it was pointed out that sulfate and methanesulfonic acid (MSA), the two oxidation products of dimethylsulphide (DMS), both decreased in concentration during the transition from the last ice age to the present interglacial, possibly because of the decrease in atmospheric dust loads and subsequent reduction of iron delivery to the oceans (Idso, 1992). Furthermore, these authors pointed out the strong dependence of cloud-condensation nuclei (CCN) formation over the oceans on the dimethylsulphide production by the marine planktonic algae, and its importance on the cloud albedo and climate. 3.2.4. The Volcanic Aerosols In this section, after a short review about current knowledge, the overall characteristics of the anomalous elemental enrichment episodes and their possible relationship to volcanic emissions will be discussed. This will be followed by a discussion of possible volcanic signatures in a group of specifically chosen samples from certain dates with suspected or known volcanic activity (see section 2.1.1). Finally, a discussion about the possibility of volcanic impact on a previously observed rare earth element (REE) enrichment episode (Keskin et al., 1992) (see section 2.1.1) will be presented. The impact of volcanic emissions on the total atmospheric aerosol budget has long been recognized and became clearly evident after the recent volcanic eruptions of El Chich6n (Mexico, 1982) and Mount Pinatubo (Philippines, 1991), whose effects were extensively examined (e.g., Kotra et al., 1983; Woods et al., 1985; Wallace and Gerlach, 1994). Depending on the type and strength, a volcanic eruption event can affect atmospheric composition on a local, regional, or global scale (Buat-Menard, 1990). Large-scale eruptions inject volcanic gases (mainly sulfur dioxide and halogenated acids) and dust (generally silicate ash) into the stratosphere in which they often remain suspended for several weeks to several years (Cadle et al., 1976). A major increase in atmospheric dust content has a serious effect on the climate by changing the atmospheric albedo (Rampino and Self, 1982). It is now believed that even small-scale eruptions disturb the global budgets of sulfur and some volatile heavy metals and metalloids to a considerable extent (Buat-M6nard, 1990). The determination of volcanic contributions is also needed to accurately assess the anthropogenic contributions because the enrichment mechanism of volatile elements on volcanic aerosols is similar to that occurring on anthropogenic aerosols produced by high-temperature industrial activities (Buat-Menard, 1990). As explained earlier, polar ice sheets contain and preserve atmospheric substances in their strata by wet and dry deposition mechanisms. In recent years, many volcanic eruption horizons identified in deposited ice layers were recognized by their high sulfate concentrations (converted from sulfur dioxide) (see Moore et al., 1991; Dai et al., 1991; Delmas, et al., 1992; Langway et al., 1994; Langway et al., 1995). However, possible nonlinearities between the scale of an eruption and the amount of sulfur dioxide gas released was recently pointed out (Bluth et al., 1993; Williams, 1995). Although trace element compositions of volcanic plumes and ashes were determined for some volcanoes (Phelan et al., 1982; Zoller et al., 1983; Phelan-Kotra, 1983; Olmez, et al., 1986; Kyle and Meeker, 1990; Dunbar and Kyle, 1990), understanding of the trace element contribution to the ice sheets from volcanic activity is still limited. Sampling results by these authors showed that the elements fluorine, sulfur, chlorine, chromium, cobalt, copper, zinc, arsenic, bromine, molybdenum, indium, antimony, cesium, tungsten, gold, mercury, cadmium, iridium, and selenium are generally highly enriched in plume and ash samples. However, it has been established that the composition of volcanic plumes can vary with time even for the same volcano (Buat-M6nard, 1990), and this complicates the determination of a definite signature for a given volcano. In the previous sections marine and crustal enrichment factors (EF) were examined for the elements measured in this work. The results indicated that most of the elements are mainly of crustal origin, while some others are mainly of marine origin. The remaining group of elements that has high marine and crustal EF's should have additional sources. These elements are chromium, zinc, arsenic, molybdenum, cadmium, and antimony. However, as can be seen from Table A.1 in Appendix A, molybdenum, and cadmium have low observation frequencies and high measurement uncertainties. Among the rest of the enriched elements, antimony and arsenic show the distinct trend of decreasing crustal enrichment factors, especially after the 1950's, in addition to their overall correlation (Figures 3.14 and 3.15). This behavior rules out anthropogenic contribution as the main source for these elements. The most likely natural source category seems to be volcanic emissions, since it was already mentioned that these elements are highly enriched in volcanic emission plumes. Also, no observed signature exists to indicate any significant extraterrestrial contribution to the observed trends. Although the annual number of large-scale volcanic eruptions seems constant throughout the last two centuries (Simkin, 1994), it is likely that the early records are not complete and some of the eruptions at remote locations, such as Antarctica, were missed. Therefore, more active volcanism (at least locally) could be (1*0 SE"1. L*6L SL8$IC U96ý 61)6ý U•6ý L96ý ;:96ý S96I. $SE4 06~L 6s SS6ý LS6L SE6& 0o.i. LZ6ý LM4· ZZOl UL & 6SL& .Lo*4 *ZLL I4LC -4O O [I=IvT] I80 80 •. S96 96e 9C, ESESIC 17S6l S 1567 TIE VIE7 1 60-. TSF- LUT: GSTV LC8K 66L T6L- 7SLs7i 0 L- [x)!-2/2u] uo! jlua3UO3 responsible for the observed EF's of antimony and arsenic (called volcanic marker elements (VME) in the rest of the discussions) in the pre-1950 era. It should be mentioned that there are other elements (e.g., selenium, indium) highly enriched in volcanic emission plumes that can be used as marker elements. However, some of these elements had near or below detection limit concentrations for the analytical technique used in this work (INAA), and the number of measurement results with low enough uncertainty levels were limited. Remember that in section 3.2.1 the Cl/Na ratio was also found higher in the pre-1940 samples as a sign of possible larger volcanic emission contribution to the annual chlorine concentration budget. If the aluminum concentration distribution is assumed uniform in an annual deposition sample (which is not realistic), an approximate correction to the Nalq concentrations can be made so that Cliqto-(Nalq)X ratio can be calculated. This calculation gives 2.7±0.1 and 2.4+0.4 for pre- and post-1940 periods, respectively, which do not seem very different. However, as explained before, Cl, and Naiq concentrations were measured in very small samples which cannot fully reflect the annual trends, and these results should be taken as informational values. In spite of this uncertainty, ratios much higher than that of sea water (1.8) are likely to indicate extra chlorine contribution from volcanic emissions. Significant enrichment episodes of VME's were observed in six specific years in this work. In addition, one of these episodes (1948) was followed by enrichments of rare earth elements in the subsequent two years. The crustal Enrichment Factors (EF,) of these enriched elements at the time of the observed signal are summarized in Table 3.5. The first VME enrichment episode occurred at 1969 together with an enrichment of Br. However, the marine Enrichment Factor (EFm) of Br is 0.9+0.2, which indicates a marine source contribution. The combination of these enrichments might indicate a specific type of volcanism (submarine) of possible local origin. Among the recorded volcanic eruptions (Simkin et al., 1981) one possible candidate seems to be the 1969 eruption of Deception Island Volcano (62.93 S and 060.57 W) in Antarctica. This was recorded as a "subglacial" and "explosive" eruption that indicates a sudden release of water along with other gaseous and particulate material (Simkin et al., 1981). The volcanic explosivity index (VEI) (Newhall and Self, 1982) for this eruption was three Table 3-5. The crustal Enrichment Factors of the enriched elements in suspected volcanic eruption horizons. Date of signal [A.D.] 1949 1948 1940 1870 1822 1791 7950 46650 98150 31100 18900 23550 +1280 ±28480 ±36150 ±9000 ±3800 ±9800 60 300 400 150 100 95 ±15 +190 ±150 ±50 ±25 ±40 EF 1969 Sb As La Ce Sm 1950 20 91 ±2 ±15 19 88 ±2 ±14 13 63 ±2 ±11 which is a moderate to large scale emission with potential stratospheric injection. The observed increase of Br and Na, concentrations, while their ratio is similar to that of the sea water, might be due to the interaction of ejected hot material and the surrounding ocean that might enhance marine aerosol production. Even though this volcano is quite close to Byrd Station, no significant sulfate signal that coincides with this year was observed in the ice core used in this work (NBY-89) (Langway et al., 1994). However, a further examination of the data provided by Langway and Osada (personal communication) showed that there is approximately a three fold increase on the excess sulfate (non sea-salt origin) for a short period during the 1969-70 transition that might have originated from the Deception Island eruption. The observed delay in the occurrence of the sulfate peak might be due to the time necessary for the sulfur dioxide-sulfuric acid conversion and deposition processes. Another observed signal that might be related to this eruption comes from an ice core drilled in East Antarctica (G 15) in 1984 (Moore et al., 1991). The continuous conductivity measurements on this 100 m long ice core revealed a very sharp peak that coincided with the 1969-70 horizon. These authors suggested that the characteristic of this peak points out a local non-stratospheric eruption, most likely the Deception Island Volcano in the Antarctic Peninsula. This conclusion is in good agreement with this paper's conclusion, and the signal in 1969 is almost certainly come from the Deception Island Volcano. The reason that this eruption was not observed in ice cores from the Antarctic Peninsula, much closer to the eruption, was suggested to be related to the wind patterns that transported the eruption plume to Antarctic ice sheets (Moore et al., 1991). It was also mentioned that (Boutron, 1980) no other eruption occurred at Deception Island between 1842 and 1967, except for five successive major eruptions between 1912±5 and 1917±3, as revealed by the study of the successive pyroclastic deposits in local ice (Orheim, 1972). Since only three samples were analyzed between 1907 and 1920 in this work (1907, 1913, 1919), and none of them showed any significant signal, it is not possible from this work to understand whether the elemental compositions of different eruptions are similar to each other or not. One interesting result that is worth mention is the lack of any elemental enrichment signal related to the prominent and well-recorded acidity signal of the Mount Agung (08.342 S and 115.508 E) volcanic eruption (VEI=3) in 1963 (e.g., Moore et al., 1991; Langway, Jr. et al., 1994). One explanation might be the distance of the eruption from the central Antarctic ice sheets and the different transport properties of the gaseous and particulate substances that eventually cause the acidity or elemental concentration enhancements in the ice sheets. While it is more likely for gaseous sulfur dioxide to be elevated up to the stratosphere and transported into the Antarctic atmosphere, it may not be easy for particulate species to escape removal in the troposphere for a moderate to large scale eruption. The second VME enrichment episode was observed in 1948 followed by very high REE enrichments in 1949 and 1950. In addition, the zinc enrichment was moderately high during this three year period while the Br enrichment was moderately high in 1948. Similar to the previous signature, the Br-to-Nas ratio (0.007+0.003) is close to that of sea water (0.0062), which again might indicate a specific type of volcanism of local origin. Among the recorded volcanic eruptions (Simkin et al., 1981), one possible candidate is the 1947 eruption of the Mount Erebus Volcano (77.58 S and 167.17 E) in Antarctica. This was recorded as an "explosive" eruption with lava flow (Simkin et al., 1981). The volcanic explosivity index (VEI) for this eruption was two which is a moderate scale eruption with no stratospheric injection. Similarly the observed increase of Br and Na, while the Br-to- Na. ratio stayed close to that of sea water, might originate from the interaction of ejected hot material and the surrounding ocean as explained before. The highest enrichments of VME's and REE's were observed in 1948 and 1949, respectively (Figures 3.14 and 3.16). This lag time in the peak enrichments might result from the different physical properties of these elemental groups. While the VME's are quite volatile, this is not the case for REE's. Therefore, the release of volatile elements might have begun earlier with non-explosive fumarolic activity, while that of the refractory elements might have intensified only when the explosive eruptions started. However, one cannot rule out the coincidence of two different events that follow each other. There is no regular observational information regarding the volcanic activities in Antarctica prior to 1956 (Boutron, 1980) which makes it difficult to reach a final conclusion. It was previously suggested that concentration ratios of certain elements, such as REE's, in volcanic emission plumes might be used to characterize certain groups of volcanoes (Palais and Mosher, 1989). However, the results of a study to characterize the plume composition of the Mount Erebus Volcano (Kyle et al., 1990) indicated that the elemental concentration ratios were not constant even on a day-to-day basis. Therefore, it is not possible to make any concrete identification on the basis of the elemental enrichments that coincided with this episode. Although sulfate and conductivity measurements did not indicate any significant signal for these years in the NBY-89 ice core (Langway et al., 1994) and others from Antarctica (e.g., Moore, 1991; Delmas et al., 1992), Mount Erebus Volcano cannot be ruled out as a candidate for this episode. The third VME enrichment episode was observed in 1940. This event is the strongest and longest (1937-44) one in the continuous annual sampling period (19261989) of this work. In addition, chromium, bromine, sodium, and chlorine were also enriched significantly. As mentioned in section 3.2.2, 1939-40 was the time of one of the strongest El Nifio events in this century, and the observed sodium, chlorine, and bromine U.- , -.- . , ... , .- •-• - c; -;c .oc 8 %S. 4i.. .... U.. .. . " -- 7E - •-- - 655 ,6-65 7S6 ~c6T E ,•.--... -'U -U- - :: - LO; •Ei ... .......- =, : :••_ C -C [i=ivl 3,43 I =*-'- ,- 99j C 'C enrichments probably resulted from this event. One strong candidate for the observed VME's peak is the 1937-43 eruptions of the Rabaul Volcano in Papua New Guinea (04.271 S and 152.203 E). All were "explosive" eruptions with decreasing VEI's between 4 and 2 (Simkin et al., 1981). Although the distance from the Antarctic ice sheets is very far, the strong scale and duration of eruptions with highly likely stratospheric impact could allow aerosols from the eruptions to participate in long-range transport. In addition, the coincidence of this event and 1939-40 El Nifio event might have increased the transport efficiency of aerosols due to the changing atmospheric currents. As seen in Figure 3.14, the 1940 enrichment peak is very sharp, while the others show more gradual behavior. Moderate scale (VEI=3) 1938-40 eruptions of the Krakatau Volcano in Indonesia (06.102 S and 105.423 E) might have also contributed to this signal. It is somewhat surprising for this scale eruption that no significant sulfate signal was recorded in ice cores obtained from Antarctica (e.g., Moore, 1991; Delmas et al., 1992, Langway et al., 1994) that coincide with these years. However, it was already pointed out that (Williams, 1995) sulfur dioxide released from the 1994 eruption of the Rabaul Volcano was surprisingly low for its class (only 0.5 % of the 1991 Pinatubo eruption), which might also be the case for 1937-43 eruptions. The remaining VME episodes observed in 1870, 1822, and 1791 occur in the time period (1711-1926) that has sparse data. Therefore, it is not possible to be sure whether these signals show the year of peak deposition. The signal in 1870 might be part of the signal from the listed 1871±40 eruption of the Deception Island Volcano dated by tephrochronology (Simkin et al., 1981). However, no more information is available about this eruption. No large scale (VEI 2 4) eruptions within the Southern Hemisphere are listed around this date (Newhall and Self, 1982), and further comment on this signal is difficult. The signal in 1822, combined with the results of the analysis (Table C.2 in Appendix C) of two other specifically chosen samples from 1819-20 and 1820-21 (see section 2.1.1), indicates a possible volcanic event between 1820 and 1822. In a study to characterize the morphological properties of microparticles recovered from a South Pole ice core's suspected volcanic activity horizons (Palais et al., 1989), results showed good agreement between microparticles from an approximately dated 1816-21 layer and tephra known to be from Deception Island Volcano (Orheim, 1972). Therefore, it is likely that this signal originated from another eruption of Deception Island Volcano. However, the observed zinc enrichment is much higher than that of the 1969 signal. Although no significant sulfate signal is observed in Antarctic ice cores for this time, another potential source listed in literature (Simkin et al., 1981) is the very large scale (VEI=5) 1822 eruption of the Galunggung Volcano in Java (07.25 S and 108.05 E). The 1791 signal does not coincide with any recorded large scale volcanic activity listed in the literature (Newhall and Self, 1982). It probably originated from a local volcano. As mentioned at the beginning of this section, eight separate samples were selected from certain dates with suspected volcanic activities (see section 2.1.1). However, because of the prior use of some of the samples that coincided with the exact dates of the eruptions, samples closest to those dates were selected. The elemental concentrations measured in these samples are given in Table C. 1 in Appendix C. Four of these samples (1882-87) coincided with the eruptions of Krakatau Volcano in Indonesia (06.102 S and 105.423 E) and Tarawera Volcano in New Zealand (38.229 S and 176.508 E) in 1883-84 (VEI=6) and 1886 (VEI=5), respectively. The Krakatau eruption is universally observed in both Antarctic and Greenland ice core sulfate and conductivity records (e.g., Delmas et al., 1992; Crowley et al., 1993; Langway et al., 1994), while the Tarawera eruption signal is resolved in some of the Antarctic ice core records (e.g., Moore et al., 1991). The sulfate peak caused by the Krakatau eruption was observed in the 1884 ice layer (21.6521.89 m) of the NBY-89 ice core (Langway et al., 1994). Among the four samples analyzed in this work, the sample from 1885 (21.53-21.70 m) shows the highest enrichments for the elements arsenic, bromine, chromium, and zinc, while the sample from 1886 (21.38-21.53 m) shows the highest enrichments for the elements antimony and cadmium. Among these elements, zinc, bromine, and cadmium have the highest enrichments observed among the more than one hundred samples analyzed in this work. This indicates the impact of a very special event, possibly the Krakatau eruption. The difference between the observance of a sulfate peak and elemental enrichment peaks is related to the physical and chemical properties of aerosols and individual elements, as they define the transport and deposition efficiencies. One interesting point is that the VME's do not show unusually high enrichments as would be expected from eruptions of these scales. Another group of two samples (1819-21) was chosen from the years following the eruption of Tambora Volcano in Lesser Sunda Island (08.25 S and 118.00 E) in 1815 (VEI=7, highest listed in the last five centuries). This eruption is also universally observed in both Antarctic and Greenland ice core sulfate and conductivity records (e.g., Moore et al., 1991; Delmas et al., 1992; Crowley et al., 1993; Langway et al., 1994). The sulfate peak caused by this eruption was observed in the 1816 ice layer (28.50-28.90 m) of the NBY-89 ice core (Langway et al., 1994). Among these two samples the 1819/20 sample did not indicate any enrichment signal, possibly because of the four years passed after the Tambora eruption. However, the second sample from the 1820/21 layer showed significantly higher enrichments than the first one. Possibly this signal is related to the higher 1822 signal discussed previously. The remaining two samples (1645-47) were chosen to coincide (as close as possible) with an unknown sulfate signal observed in 1648 in the NBY-89 ice core (Langway et al., 1994). These samples showed very high zinc but no significant VME enrichments. Therefore, it is difficult to claim a volcanic event from the elemental concentrations obtained in this work. As mentioned at the beginning of this section, in a preliminary study (Keskin et al, 1992) a specific REE enrichment signal was identified in a layer between 89.02 and 89.17 meter depths. Twelve more samples were obtained between the depths 88.15 and 89.51 meters continuously to locate the peak position of this signal (concentrations and EF's are given in Table C.3 and C.4, respectively, in Appendix C). The results indicated a moderate enrichment of REE's and VME's in 1341 A.D. and 1340 A.D. respectively. This signal is quite similar to that of the 1948 A.D. signal discussed earlier, except that the scale of the enrichments are an order of magnitude smaller for both groups of elements. Therefore, the local Mount Erebus Volcano might be a candidate for this signal. 3.2.5. The Anthropogenic Aerosols As mentioned in the first chapter, one of the goals of this work was the understanding of whether or not a detectable anthropogenic contribution to the elemental concentrations levels found in Antarctica exists. The 278 year-long ice core record (1711 A.D.-1989 A.D.) obtained in this work is long enough to determine the average preindustrial and modem era elemental concentration levels for comparison. However, one complication to this characterization comes from the contribution of volcanic emissions, whose enrichment mechanisms of volatile elements on released aerosols are similar to that occurring on anthropogenic aerosols produced by high-temperature industrial activities (Buat-M6nard, 1990). Therefore, an accurate determination of volcanic contribution to the total concentrations observed is an important and necessary task. Unfortunately, this task is not easy as volcanic contributions are sporadic in nature, different in strength, and inhomogeneous in aerosol composition. Another complication is the lack of information about the stability of weather patterns throughout the last three centuries as they affect the transport of aerosols from continental areas to the Antarctic ice sheets. In a simple way, the time-series trends of the excess elemental concentrations that are assumed to come from volcanic and anthropogenic sources can be obtained by subtracting the crustal and marine contributions from the measured concentrations. For this purpose, crustal and marine contributions were calculated by using the standard concentration ratio of the element of interest to aluminum in crust (Taylor, 1972) and seasalt sodium (Na.) in sea-water (Goldberg, 1963). The resulting concentrations should show the contribution from volcanic and anthropogenic emissions in an approximate manner. Among the elements examined, arsenic, zinc, and chromium, which are known to be enriched in emissions both from volcanic and anthropogenic sources seem to deserve further discussion. As seen in Figure 3.17 antimony concentrations show a decreasing trend after the 1950's, which was assumed in Section 3.2.4 to indicate the decreasing impact of volcanic emissions. However, the excess (crustal and marine contribution corrected) concentrations of arsenic do not show a significant change (Figure 3.18). Since arsenic is also one of the marker elements used to characterize volcanic emissions, it is expected that it would also show a decreasing trend. Therefore, in a qualitative sense this behavior indicates an increasing anthropogenic contribution after the 1950's, which is possibly the beginning of large scale industrial activities in the Southern Hemisphere. Increasing concentration trends are somewhat more apparent for zinc and chromium in recent years (Figures 3.19 and 3.20). Using the same argument, these trends are likely the result of increasing anthropogenic emissions. To get an approximate quantitative sense about these trends, the average concentrations were calculated in three time periods (1711-1901, 1907-1950, 1951-1989), and percent changes relative to the 1711-1901 time period were calculated (Table 3.6). In these calculations, the years with unusual episodic behavior were eliminated. Table 3-6. Percent average concentration changes for Sb, As, Zn, and Cr relative to the 1711-1901 time period. Element 1907-1950 1951-1989 Years ignored Sb +31 -47 1822, 1870, 1940 As +80 +45 1822, 1870, 1940, 1944, 1945 Zn -32 +28 1799, 1822, 1949, 1950, 1951, 1985 Cr +28 +155 1939, 1940, 1970, 1979 In a recent study, lead, cadmium, copper, and zinc concentrations were measured (Gbrlach and Boutron, 1992), and no significant increase was observed between 1940 and 1980 with the possible exception of lead for which a significant increase seems to have taken place after the mid 1960's in Antarctic snow samples from Adelie Land, East Antarctica. However, since the rate of change is likely to be less within this time interval compared to that of the time interval examined in this work, it might have been more difficult to observe any possible change. Otherwise, some increase in the concentrations of anthropogenic elements is expected even if on a very small scale. Among the anthropogenic source regions of the Southern Hemisphere, the coast of Peru was identified as a substantial source of copper, zinc, arsenic, and lead concentrations that are a6ý ca v; C C cjp 0 o ri - - [eo!-6/,u] uO!JBiJuoGOUo C) -0 0- CO 0o o QQ 6 E ad o DD * II e• 0 0 0 0 o [oo!-5/6u] 0 0 uo!leJ1uO3UOO 0 oa 0 0 0V 0O \0w(14 [eo!-616u] uo!LeJlu93uoo 0 C3 a, 0~0 -5 0 0 1E ol,I E ·s 0 0 Cz~ 0: om 0s 0 [ 1 u u o [eo!-6/6u] uo!le1lueouoo r-.q thought to be related to the large emissions of the extensive mining and smelting industries in that region (Raemdonck et al., 1986). In addition, human activities in manned scientific research stations in Antarctica are suspected to be an additional source for some elements in recent years (Boutron and Wolff, 1989). An attempt has been made to calculate the approximate natural background elemental concentration levels within the pre-industrial (1711-1901) era. These values were obtained from the frequency distributions of the elemental concentrations within this era. The most probable values of the distribution curves were chosen as the background values due to the reasons explained in section 3.3 and the results are given in Table 3.7. Table 3-7. Calculated natural background concentrations. Element Concentration Element Concentration [ng/g-ice] 3.3. Na Mg Al Cl K Sc Ti V 23 2.4 4.4 50 2.2 0.0001 1.0 0.01 Cr 0.23 [ng/g-ice] Mn Fe Zn As Br Sb La Ce 0.05 3.6 0.9 0.001 0.009 0.019 0.001 0.001 Source Characterization 3.3.1. Source-Receptor Relationship One can define aerosol production, transport, and deposition mechanisms as part of an integral system. A simplified overall picture of this system can be given, as in Figure 3.21. Here, the receptor site is the location of sample collection, and the sources are continental regions, oceans, volcanoes, and possible others. The transfer functions contain information about factors that affect the transport of aerosols from production sites to the deposition site. Some of these factors are transport distance, wind trajectory, transport altitude, and chemical / physical processes. For example, as discussed in section 3.2.1, marine aerosols released from the oceans may not have the same composition when they arrive at the deposition site after a long transport because of the loss of chlorine due to some chemical reactions during the transport. Another example is the size segregation of aerosols during long-range transport that might change the original aerosol composition, since some elements are preferentially contained in smaller or larger aerosols. However, knowledge about these transfer functions is far from complete due to the unpredictability of atmospheric processes or a lack of understanding about some mechanisms. In addition, knowledge about these sources, their strengths and compositions is also limited, since the system covers almost all of the Southern Hemisphere and, in some cases, the whole globe as possible source regions. Therefore, the only other alternative for identifying sources with a significant contribution to the aerosol composition in Antarctica is to use multivariate statistical analysis techniques, such as Principal Component Factor Analysis. Transfer function 1 Transfer function 2 Figure 3-21. Schematic of source-receptor relationship. 3.3.2. Principal Component Factor Analysis Principal Component Factor Analysis (PCFA) is a multivariate statistical analysis technique for the interpretation of large data sets. Today this and similar techniques are increasingly being used in environmental sciences (e.g.,Okamoto et al., 1990; Artaxo et al., 1992; Olmez et al., 1994), as well as other branches of science. Part of this increase is due to the availability of fast measurement techniques for obtaining large data sets and the advancement of computational power within the last few decades. The PCFA is performed to identify possible correlations among variables of a data set, and its success is closely dependent upon the number and accuracy of observations of the variables. In this work, the variables are the elements, and the observations are the annual average concentration measurements of them. The mathematical background of this technique is given in Appendix D. The PCFA was applied to 97 individual concentration measurement results for 14 elements. Prior to the analysis, the summary statistics for each element were obtained from Statgraphics TM version 6.1 statistical analysis package, and these results are given in Table A.2 in Appendix A. The first consideration was the treatment of missing values for some elements with concentrations close to the analytical detection limits. Since the success of the PCFA technique depends on the completeness and accuracy of a data set, this step was necessary before starting the analysis. For this purpose, the number of missing values for each element was determined from the summary statistics, and only elements with less than 25 % missing data were accepted for the analysis. In addition, elements with high uncertainties were excluded from the analysis. The first step in the treatment of missing values was the determination of the frequency distribution characteristics of the measured elemental concentrations. Results showed that the frequency distribution function for each element is close to logarithmic. This behavior is expected since most of the time concentration values should be close to background values due to dilution effect, and higher values should be observed only due to the specific contributions from various sources for limited durations (Ott, 1990). The next step was to find the most probable concentration values for each element to replace the missing values. They were obtained from the frequency distribution curves and are given in Table A.5 in Appendix A. After the treatment of missing values, the raw data matrix containing elemental concentration values for 97 samples, was entered into the factor analysis program within the statistical analysis package Statgraphics TM version 6.1. The necessary options were chosen to standardize the data matrix, not to change the diagonal elements of the correlation matrix with the estimated communalities (principal components analysis), and to rotate the factor matrix using the varimax method. The mathematical details of these options are explained in Appendix D. After the first run, factor scores were plotted for each factor, and samples with factor scores above six (standardized factor score value) were excluded in the next run. Note that the factor score of a sample for a specific factor represents the relative contribution of that factor (source) to the sample. Since the standardized factor scores were determined by subtracting the average value from each raw score, and dividing the difference by the standard deviation of the scores, a sample with a factor score greater than six would be affected six standard deviations more than the average sample. However, since this kind of impact should result from a very specific and rare event, exclusion of these extremes is preferred to prevent their dominance on the overall analysis. Otherwise, this could affect the factor loadings and cause bias for the overall correlations of the elements in a group. After this sample exclusion step (only sample # 11IC5 was excluded), the new data matrix was entered into the program, and the new results were analyzed in the same fashion. Since the new set of factor scores were below the predefined limit six (Figures 3.22-3.25), the factor loading results were accepted as the final values. Note that the standardized factor scores have the average value of zero, and the negative scores show less than average relative contribution of the associated factor to the related samples. The final factor analysis results are summarized in Tables 3.7 and 3.8. Table 3.7 contains the communalities for each element, eigenvalues for each factor (source), percent variance explained by each factor, and the cumulative percent variance. As can be seen, four factors have eigenvalues above unity and they explain 63.1% of the variance in the data set. As discussed in Appendix D, factors with eigenvalues below unity (average value) are generally assumed to result from errors in the data set. Using the first four factors, the varimax rotated (see Appendix D) factor loadings matrix in Table 3.8 was obtained. Factor loadings with values above 0.3 were assumed significant and retained. The elements in a factor with loadings above this value were considered to be coming from a common source. Although no clear rule for determination of this cut point exists, values between 0.25 and 0.50 were used in the literature for different applications (Olmez et al., 1994). Table 3-8. Factor analysis results for related parameters. Element Na Al Cl K Sc Cr Mn Fe Zn As Br Sb La Sm Communality 0.68 0.58 0.69 0.31 0.63 0.52 0.41 0.45 0.39 0.30 0.24 0.37 0.59 0.54 Factor No Eigenvalue 1 2 3 4 5 6 7 8 9 10 11 12 13 14 4.23 1.95 1.53 1.12 0.92 0.81 0.73 0.68 0.57 0.51 0.33 0.23 0.21 0.17 %variance %cumulative of explained explained variance 30.2 13.9 11.0 8.0 6.6 5.8 5.2 4.9 4.1 3.7 2.4 1.7 1.5 1.2 30.2 44.1 55.1 63.1 69.6 75.4 80.6 85.5 89.6 93.2 95.6 97.3 98.8 100.0 The elements aluminum, potassium, scandium, zinc, lanthanum, and samarium are grouped together in the first factor, which explains the most variance of the data set. The other factors follow the first one in decreasing order of importance for explaining the variance. These elements are known to exist in crustal aerosols, and this factor implies a crustal source for the observed concentrations of these elements. As discussed in section 3.2.3, the arid regions of Australia and South America are the likely candidates for the crustal aerosol input into the Antarctic continent. 100 -- REL 0 == - 5cman r- _ 5E ima liiLDla c mar T7 , tm .5 :......,.. • EIt ~i·I~i·l·l·JI·l~ ~I·I·TI~ [-11~·'·~·~··~~~.1~ 1 rx~t~t~ 0 o CC 4-h -31 ". ,2 ;z toi! 5-' C1 IC 0 t• LI LD C.) I-' O [_ SD Oiu•£• C £~ £2I Ii2 £322 _ _ ••:-.--:-:.:.:-_:- L... . . . _ •giT t....•.:::::::::..::.......°....::•.....•<.••..:•:::..g;;;-':: L JL LO I + c ............. CV4 101 T ... ... ... . 0 . I C%% I :_____ I ~.~·1T~ F·:~~:~· 1;?·~~I- 7t - ~-~ _ 31 _ ~ :22K FT~ ~rrr I~il B 1 2 -- _ I•<.:'•'-.'• ".:,:+:-:";-":<,'<.:°:-...<-:-:• -:• .:...:- - -:';';'7:': ":"":":-:-:-1"7;'.-: :';-<=• •OZ' Z31c_ t31S · -3I - ioi _[Ol I LOC) SI I (Y I I I i t i ( 0. 102 L CY) F F' c "r·. 1L/:···_· -0i F-7 iiii T ex ZZZW· L- -C- -i I=I~7 ~~· ~cl _r -L·-, 31 v 7--;.I T L L iii ~L·-,· L-i L-- -- cL;---vc_ L· L $1' c~c~ T3I -7~L·-, .3__ ;?irt~t~T~ c~-C iv .J, ;;~. PL'I ,~·~,~·n3I= j3I~ ~· Ts ~v TZIP t~;;t -t 71iL'-t 13-, 1-7 rV_ i-· iu^ t i , 10 LO C4 m 103 ___ IZ i IToT E4ki i~T~ sSS55~fS5~~~ _521 Ii B E s-amE DI: CIlI Ljlna mmI~ TZsIes ................... :7..... rnr--- P...::.........•..::..'::::..:•.....:...::M......:...........'.......:.......:...7:`.:...•-.... .... T ma:^ EMEES;~ EMBE * - 2 I 9i -m IED a=ses EES22El ICl tIS ' rr' III I 0 LO( I I c) I ""r ...... .. ...... . I CN - 104 o II I C'4 The second factor shows sodium, chlorine, potassium, and chromium grouped together. The first three elements clearly indicate a marine aerosol impact on the observed elemental concentrations. Considering that Antarctica is surrounded by oceans, it is certain that these oceans are the sources for these elements. Although chromium is not enriched in the bulk of ocean water, its enrichment in the ocean surface microlayer and attachment to sea-salt aerosols is a possibility. Table 3-9. Factor loadings (after varimax rotation) and estimated communalities. Element Estimated Factor No Factor No Factor No Factor No communality 1 2 3 4 Na Al Cl K Sc Cr Mn Fe Zn As Br Sb La Sm 0.72 0.69 0.84 0.56 0.70 0.59 0.57 0.59 0.48 0.63 0.37 0.69 0.78 0.62 0.77 - 0.52 - 0.91 0.62 - 0.37 0.74 - 0.64 - - - - - 0.53 - - - - - - - 0.86 0.76 0.64 - - 0.31 0.40 0.74 0.70 - 0.49 - 0.42 0.79 0.36 0.82 - - - The third factor has significant loadings of manganese, iron, aluminum, and moderate loadings of bromine, chromium, and scandium. Although most of these elements are known to be mainly crustal in origin, the coexistence of bromine, which is known to exist mainly in sea water is interesting. In recent years, an almost a near-total depletion of the surface level ozone has been observed in the Arctic spring, coincident with high atmospheric concentrations of inorganic bromine (Fan and Jacob, 1992). It was previously suggested that the ozone depletion could be due to a catalytic cycle involving the radicals Br and BrO (Barrie et al., 1988). This gas phase catalytic cycle is given as, 105 Br + 03 * BrO + 0 2 BrO + BrO -- 2Br + 02 (i) However, bromine is quickly converted to HBr, HOBr, BrNO3, and other brominated organic compounds (BOC) by gas phase reactions with species such as HO 2 and HCHO. Therefore, the ozone destruction cycle would quickly be terminated (Fan and Jacob, 1992). Recently, it was proposed that the HBr and BOC's could be scavenged by the ambient aerosols and ice crystals, and Br 2 could be released back through heterogeneous reactions on these particles' surfaces (McConnell et al., 1992). Therefore, this cycling would maintain sufficiently high levels of Br atoms and BrO radicals in the ambient air for ozone destruction. These authors suggested that Br might be formed by the photolysis of Br 2(g), which in turn can be formed by the conversion of sea-salt derived Br- through photoactivated reactions on aerosols and snow-packed surfaces. In the past, a possible mechanism for this conversion was suggested as (Duce et al., 1965), 4Br- + 02 + 4H++ hv -+ 2Br 2 + 2H 20 (ii) Although the detailed sequence of the reaction (2) is not known, one of the suggested mechanisms was given as (Carey and Langford, 1975), Mn+Br Br - 400nm M (n-1)+++ Br" 2 Br' -- Br 2 2M (n- l ) + 102+ H+ -- 2M n + + OH- 2 (iii where, M is a transition metal such as Fe(III), Mn(IV), or Cu(II). In light of this information, the third factor that was mentioned before suggests a similar transition metal catalyzed bromine production mechanism, and possible subsequent ozone destruction in the Antarctic surface level ambient air. Since the aerosols and ice crystals are eventually deposited onto the surface snow layers, these possibly crustal origin 106 aerosols could also cause the deposition of some bromine together with some transition metals. Therefore, the elements manganese and iron, which are grouped with bromine, seem to be good candidates for the catalytic reaction (3). Since copper was not measured in this work, it is not possible to comment on it. The possible crustal origin of these aerosols might explain the existence of aluminum and scandium in this group. These conclusions are partly supported by another study on South Pole aerosols (Tuncel and Zoller, 1995) as it was found that particles reaching the South Pole during summer are enriched in bromine, while no enrichment appears in the winter. This would require the availability of a gaseous bromine reservoir during summer, which possibly could result from the proposed reaction (3). In addition, these authors observed bromine in a factor that also included manganese and iron. However, one difference comes from the observation of moderate bromine loading in another factor with sodium and chlorine (marine factor) in the South Pole aerosols study while its additional occurrence is in the volcanic factor in this work. The reason for this difference is not clear at this time. The fourth factor contains the elements antimony, arsenic, bromine, and zinc. As mentioned in section 3.2.4, these elements are highly enriched in volcanic emission plumes and are probably of volcanic origin. A similar factor for the elements antimony, arsenic, and zinc was found in the South Pole aerosol study and was also suggested to be volcanic origin (Tuncel and Zoller, 1995). 3.4. Air/Snow Partitioning 3.4.1. Current State of Knowledge To this point, a clear distinction has not been made regarding the relationship between the elemental concentrations measured in snow deposits and the ambient atmospheric aerosol concentrations. To understand this relationship better, the details of the aerosol air/snow partitioning processes should be examined. The importance of making an accurate determination of the ambient aerosol concentrations stems from the 107 fact that aerosols have an important place in the radiation balance of the earth/atmosphere system (Pyle, et. al., 1988). Aerosols affect the radiation balance both through direct interaction with incoming solar, and outgoing infrared radiation, and by indirect effects due to their role in cloud formation processes. A direct effect of ambient aerosols on surface radiation flux is caused by scattering, absorption, and emission processes. However, the major effect of aerosols in the atmosphere is on cloud optical properties through the cloud condensation nucleation processes (CCN), which in turn affect the cloud/radiation interaction through changes in the drop-size spectrum, cloud lifetime, cloud albedo, and in-cloud absorption (Ghan et al., 1990; Preining, 1991). The cloud/radiation interaction was specified (Grassl, 1988) as one of the most important processes in the climate system. It has been shown that the magnitude of the global cloud albedo effect could be of the same order as the greenhouse effect on increased carbondioxide (Twomey et al., 1984). Therefore, an accurate radiation budget calculation depends strongly on the determination of aerosol characteristics such as number density, size distribution, and chemical composition. Radiation balance and general circulation climate models should be tested for past climatic changes to rely on their future predictions. Because Antarctica is the largest sink for energy in the natural global system, it plays a very important role in determining the climate balance over a large part of the earth's surface (Simmonds, 1990). Any serious feedback mechanism, such as cloud albedo forcing, must be treated accurately for realistic predictions. It is therefore important to reconstruct the aerosol composition and number densities in past atmospheres to be able to use them in these models. A short review about the characteristics of polar atmosphere aerosols and deposition mechanisms might be useful to understand the air/snow fractionation processes. A more comprehensive topical discussion can be found in the literature (e.g., Junge, 1977; Shaw, 1980; Davidson, 1989). Aerosol size characterization studies showed that a bimodal size distribution exists over the Antarctic ice sheets. Using multiwavelength optical measurements (Shaw, 1980), the small and large modal diameters were found to be < 0.1 pLm and 0.28 ± 0.04 ptm, 108 respectively. Results of the same work showed that the particle number concentration at inland sites is dominated (96 %) by the small mode while particle mass is dominated (99.8 %) by the large mode. In another study (Dick, 1990), these bimodal sizes were found to be larger for the Antarctic peninsula region which is close to the oceans and affected by the large sea-salt aerosols. However, the South Pole measurements are more representative of inland regions like Byrd Station. Sulfur was found to be the main component (70-80%) of the Antarctic ice sheet aerosol mass (Cunningham and Zoller, 1981; Wagenbach et al., 1988). It is concentrated mainly in the Aitken particles (- 0.01 gtm diameter) and the accumulation mode particles (-0.1 to 1 Jtm diameters), which are formed by the coagulation of Aitken particles due to Brownian motion. It is believed that these Aitken particles are formed by gas-to-particle conversion of sulfur containing trace gases that are released from oceans and industrial activities (Shaw, 1989). The lifetime of these particles is only a few days because of coagulation, hence, these particles are thought to be produced in situ over the ice sheet. Marine and crustal aerosols are found mainly in the large mode; their properties were discussed previously in sections 3.2.1 and 3.2.3, respectively. The crustal aerosols play an especially important role in the ice phase nucleation process in the Antarctic atmosphere in spite of their small mass fraction (- 5%). It has been shown that the most common nuclei for snowflakes are clay minerals originating from crustal source regions (Kumai, 1976), and also that crustal aerosol is removed more efficiently than other aerosols (Cunningham and Zoller, 1981; Boutron, 1982). These ice nuclei generally satisfy the requirements of insolubility in water, are large, have chemical bonding and crystallographic structures similar to those of ice, and have certain topographic surface features (Hobbs, 1993). The aerosol deposition mechanisms are classified into two main groups as the dry deposition and wet deposition mechanisms. Dry deposition of aerosols onto the snow surface includes gravitational settling, impaction, and diffusion mechanisms which proceed regardless of the existence of precipitation. Among all processes, gravitational settling is the most effective for large particles (> 20 jtm diameter) while diffusion is the most effective process for small particles (< 0.1 itm) (Seinfeld, 1986); impaction affects the 109 particles with diameters larger than 0.1 ýtm (Davidson, 1989). Wet deposition mechanisms are separated into two main groups as the in-cloud and below-cloud scavenging mechanisms (Junge, 1977). In-cloud scavenging processes include cloud droplet nucleation, ice crystal nucleation, interstitial aerosol uptake, and riming of cloud droplets. The below-cloud scavenging processes include sweeping of aerosols by falling snow crystals by impaction and diffusion. Among the various aerosol deposition mechanisms, the dry deposition and belowcloud scavenging mechanisms are comparatively better understood and mathematical models for these mechanisms have been developed (e.g., Junge, 1977; Ibrahim et al., 1983; Noll and Fang, 1989; Dick, 1990; Frank et al., 1990; Miller and Wang, 1991). Recently an increasing number of modeling efforts have been made for in-cloud processes (e.g., Flossmann, 1991; B6hm, 1992; Tlisov et al., 1992; Qin, 1992; Schulist and Freeman, 1992), but success has been limited. Significantly, it has been observed and reported (Shaw, 1980) that the in-cloud nucleation mechanism is responsible for more than 90% of the total aerosol mass deposited onto the Antarctic ice sheets, and dominates the removal of all particles from the atmosphere with diameters from about 0.2 to 10 .tm (Junge, 1977). Another model study (Dick, 1990) revealed that wet deposition processes are responsible for removing more than 95% of the total atmospheric aerosols having diameters between 0.3 and 2 km. In summary, the apparent domination of aerosol deposition by in-cloud nucleation mechanisms, and the lack of satisfactory understanding of these mechanisms so far prevented the development of successful theoretical air/snow partitioning models. However, it has already been suggested that when in-cloud nucleation processes overwhelmingly dominate the removal of the significant mass of aerosols and aerosols are well mixed and homogenized over the inland ice sheets, the elements with common origin (e.g., crustal, marine) are expected to have similar scavenging ratios, and the chemistry of ice layers should reasonably reflect the composition of the atmosphere at the time of deposition (Junge, 1977). In other words, if crustal aerosols are taken as an example, since they favor nucleation more than other kinds of aerosols (Kumai, 1976), the elements 110 having a predominant crustal origin should show similar scavenging ratios. This expectation was supported by an early work that in the existing clean air conditions in central parts of the Antarctic ice sheet, the chemical composition of precipitation (Boutron and Lorius, 1979) and aerosols (Maenhout et al., 1979) were reasonably interrelated (Davidson et al., 1981) in contrast to what is observed in the sub-polar regions, in locales such as Alaska (Rahn and McCaffrey, 1979). The word "reasonably" is always necessary since there are inherent uncertainties related to these processes. As an example, alteration of particle composition after initial production by slow evaporation of volatile components or by the condensation of trace gases onto the particles during transport or fallout can occur (Shaw, 1979). The degree of these impacts depends upon the particle surface area and, thus, the particle size. At the same time, the nucleation mechanism operates more efficiently on the larger particles due to the lower Gibbs free energy, and thus higher thermodynamic stability of the solid-liquidvapor interface, and consequently preferential selection and removal of the larger aerosols (Fletcher, 1958). A slightly different crustal aerosol size distribution in the snow and ambient air might result in a difference in the scavenging ratios because of the mentioned alterations. Other factors that complicate the air/snow partitioning include variety of nucleation modes and the rimming of supercooled cloud droplets onto the ice particles. Cloud droplet and ice nucleation mechanisms are the two major nucleation processes which are effective in clouds (Barrie, 1985). Cloud droplet nucleation occurs more readily on soluble components such as sulfate and marine aerosols, and does not lead to aerosol removal by itself whereas ice nucleation occurs more readily on insoluble crustal aerosols and causes aerosol removal by precipitation. Additional mechanisms include freezing of supercooled water by contact-freezing and immersion-freezing nuclei, direct formation from the vapor on deposition nuclei, and freezing of condensing droplets by condensationfreezing nuclei (Beard, 1992). Under certain conditions large snow crystals (> 200 jim diameter) scavenge supercooled cloud droplets (2 to 50 ýjm in diameter) by inertial impaction and interception as they fall past the small drops (Barrie, 1985). These cloud 111 droplets freeze upon contact with the snow crystals, a process called riming. Riming is more likely to happen in warm clouds which permit a large number of cloud droplets to form. Since the substances in cloud droplets are deposited onto the snow crystal surfaces, concentration levels for rimed snow are higher than those of the unrimed snow crystals (Davidson et al., 1987). Contributions from the riming process to the elemental concentrations measured in deposited snows should be less in central parts of an ice sheet (Hammer, 1985) due to the absence of adequate moisture for supercooled cloud water droplets to coexist in the cold clouds of inland ice sheets. 3.4.2. Scavenging Ratios Calculation of the elemental scavenging ratios was made using the following dimensionless equation: W =CsPa Ca (3.5) where Cs, Ca, and pa designate elemental concentration in snow [ng/g-ice], elemental concentration in air [ng/m 3-air], and air density [g-air/m 3-air] (1440 at -28 Coand 1 atm), respectively. Since ambient elemental concentration data representative of the annual average at Byrd Station, Antarctica is not available, data from the South Pole Station was used (Tuncel et al., 1989; Tuncel and Zoller, 1995). The distance between these locations is about 1000 km and a uniform aerosol distribution over the central ice sheets, as discussed in chapter 1, is the basis for the use of South Pole ambient air elemental concentrations. The South Pole data covers the period between 1979 and 1983. Annual concentrations were obtained by averaging the summer and winter concentration measurements. The scavenging ratios calculated from these ambient air measurements and the ice core measurements (this work) are given in Table 3.9 for some of the elements with the best available measurement accuracy. However, uncertainty levels are still high; the major contribution to the final uncertainties came from the ambient aerosol 112 measurements, due to the low aerosol mass loadings in clean air conditions in central Antarctica. Note that the aerosol samples were collected at surface level, and it is known that the near-surface removal processes can deplete aerosol mass loadings in the turbulent boundary layer (few hundred meters) (Shaw, 1979). This could create a systematic error in the calculated scavenging ratios. However, the relative difference between the scavenging efficiencies of various aerosol categories should not be affected significantly. Table 3-10. Air-to-snow scavenging ratios for some elements. 1979 1980 1981 1982 1983 Al 25750±21800 42000±24800 47900±24300 72150±41000 37500±27250 Fe 20000±15150 33800±19850 22100±13350 42000±27750 44800±31000 La 6250±8050 12150-9450 17900±14200 7050±5050 13900±9700 Na 8250±6200 8400±3000 7200±7850 5200±3300 3350±3850 Zn 5400±3600 3350±2700 11600±7450 15300±14600 55250±52200 Sb 1950±2400 16000±12400 10750±12300 11700±17000 31300±26300 As 350±250 700±350 2350±1900 1450±900 2750±2400 As shown in Figure 3.26, the scavenging ratios for aluminum and iron (mainly crustal origin elements) are the highest listed and in agreement within the data uncertainties, confirming the more efficient removal of crustal aerosols as discussed in the previous section. Although the earth's crust is the major source for the rare earth elements (REE) in the ambient aerosol, the scavenging ratio of lanthanum was found to be smaller (with high uncertainty) than that of other crustal origin elements. This seems to result from experimental problems since a significant contribution of REE from other sources, whose aerosols are not scavenged as efficiently, is unlikely in the Antarctic atmosphere. The element sodium (mainly marine origin) has lower scavenging ratios, which is thought to be related in part to its high solubility in cloud water as discussed in the previous section. The more volatile elements arsenic and antimony have the lowest scavenging ratios in general, even though uncertainties are too high to be sure. This might 113 be due to their probable association with very small sized aerosols, which are not efficient for nucleation removal. In summary, the results showed that scavenging ratios can be used to reconstruct the past ambient elemental concentrations in the atmosphere in an approximate way, and in turn, these elemental concentrations can be used to find the number concentrations of crustal, marine, and smaller size aerosols in the Antarctic atmosphere, assuming the aerosol size spectrum has not changed considerably throughout this reconstruction period. However, there are some drawbacks to this empirical approach. First of all the overall uncertainty levels are far from satisfactory and should be reduced by more detailed and longer duration aerosol sampling and analysis programs. Second, regional differences are not treated in this approach, which could be quite important, especially for locations near the ocean. Therefore, individual scavenging ratios should be obtained for regions within Antarctica that have different geographic and atmospheric properties. Finally, the possible seasonal differences on the scavenging ratios might be identified and treated for further refining of this approach. The impact of seasonal differences was shown in Greenland (Davidson and Honrath, 1987). It might also be important for the warmer sites of Antarctica due to the riming process as mentioned in the previous section. 114 _ C . o c ilL:I -o < CD cON 0) co CVJ ed ed L4 co 0) 0) 0 9-c (D CV 0 0 0 0(D -000 00000000 0 0 0ý00 Pý D,Uý4 CI), C 115 0 0 Na 0) CO - 3.5. Air Content Analysis The results of the air content measurements in ice samples, including artificial ice blanks, are given in Table 3.10. Here, samples and blanks were designated by the letters S and B, respectively, and uncertainties were given as 1 values from the spectrum analysis. As explained in section 2.3, artificially produced air-free ice samples were used to determine the experimental uncertainty due to sample preparation and analysis steps. Table 3-11. Air content measurement results for samples and blanks. Sample ID Depth Interval [m] Weight [g] Air Content [ml/kg] B1 3.4 4.6 ± 1.1 B2 3.4 3.5 ± 1.0 B3 2.3 3.7 ± 1.1 B4 2.6 2.5 + 1.1 B5 2.5 5.7 ± 1.1 B6 3.0 5.6 ± 1.0 S1 101.20 - 101.22 3.6 46 ± 1.3 S2 101.20 - 101.22 3.6 49 ± 1.4 S3 101.20 - 101.22 3.1 43 ± 1.2 S4 101.22 - 101.24 3.8 37 ± 1.1 S5 101.22 - 101.24 2.2 39 ± 1.3 S6 101.50 - 101.52 2.0 39 + 1.3 S7 101.52 - 101.54 2.1 48 ± 1.4 S8 101.54 - 101.56 1.6 37 ± 1.3 S9 101.56 - 101.58 3.6 50 + 1.4 Although the blank samples were assumed to be air-free, in reality some amount of air would have been left dissolved in the water prior to freezing. However, since it is difficult to know the exact amount of the remaining air in the blank samples, the amount of air measured (average 4.3 ± 1.3 ml/kg) was assumed (conservatively) to be the result of external contributions due to the experimental stages. Considering that the average air 116 content in a Greenland ice core has been measured to be greater than 100 ± 5 ml/kg between 100 and 1300 m depths (Herron, 1982b), the maximum uncertainty due to experimental handling would be under five percent. The air contents measured in real samples from the NBY-89 ice core showed good agreement within each of the two same-depth layers. Samples from the first layer (S1, S2, S3) and second layer (S4, S5) gave average air content values of 46 ± 3 ml/kg and 38 ± 1 ml/kg, respectively. The analytical uncertainties of these measurements are quite low due to the very sharp and interference free Ar gamma-ray peaks observed for these samples. Samples from the adjacent depth levels within an annual accumulation layer (S6, S7, S8, S9) showed larger deviation (44 ± 7 ml/kg) compared to the samples from the same-depth layers. This difference is probably the result of seasonal impact on the air content. It has been shown that seasonal variations in air content can be as high as 25 % in high accumulation areas, and this variation has been attributed to the sealing effect on the summer layers by the denser winter layers (Martinerie et al., 1992). As a result, summer layers have higher porosity than winter layers. In addition, wind and sunlight can form thin, dense layers on the snow surface, which can also cause the isolation of the previous layers within an annual accumulation layer. However, for deep ice cores with annual layer thicknesses of a few centimeters, these seasonal variations become less important. It has been shown that the air composition in the bubbles of temperate glaciers can change slightly if some melting occurs (Horibe et al., 1985). This type of melting can happen especially in deep layers due to the high pressure applied by upper layers. These authors observed a slight increase (around 1 %)in the N2/Ar ratio in the air extracted from bubbles in deep ice samples from the Camp Century, Greenland, ice core. They suggested that the higher solubility of argon in melted water and the subsequent push of this melted water into the inter-granular layers between the ice (due to thawing of ice and bubble pressure) could create this difference. Since the argon content of the ice samples was used to find the air content of them in this work, this phenomenon could potentially have an effect on the results. However, this difference of 1 %is negligible compared to the total measurement uncertainties in this work. In addition, this method measures total argon 117 content (dissolved and undissolved), and unless the melted water migrated distances more than a few percent of the core diameter, this phenomenon would not effect the results. In summary, this new method for direct air content measurement in ice samples might play a useful role in future ice core research, and additional work to develop the method further is justified. 118 4. SUMMARY AND SUGGESTIONS FOR FUTURE WORK 4.1. Summary The polar ice sheets contain a wide range of paleodata entrapped in stratigraphic layers. They contain records of changing climatic conditions, deposits of atmospheric constituents, and are sinks for atmospheric trace elements and other chemical species which are transported by wind systems from various global sources. In this study, snow strata that cover 278 years (1711 AD to 1989 AD) were investigated to observe whether or not there is a significant difference in the Antarctic ice sheet elemental concentrations due to anthropogenic impact. In addition, within this timeperiod there have been significant natural events, such as volcanic emissions and variations in global air-mass movements, which could affect the elemental composition of the atmosphere, and which could be investigated using trace element patterns. Even though there have been previous studies to determine the trace element concentrations in polar ice sheets, most have involved the measurements of a relatively small number of trace element over discontinuous time-intervals. They were also limited by the specific earlier analytical techniques used, the physical quality of the firn or ice core samples used, and an evident lack of a close interdisciplinary and multiparameter approach to an integrated study program. Of great importance, most of the previous data available prior to this investigation are not continuous over long time-periods. However, since changes in chemical composition trends in the atmosphere are extremely slow, detecting these changes requires long records. The results of this study will hopefully fill an existing gap in our knowledge. An auxiliary goal was to develop a new method for measuring the air content in very small ice samples from deep ice cores as it would be useful to study the relationship between both the flow patterns and air content changes of glacier ice with depth (as well as for dating purposes). 119 The Instrumental Neutron Activation Analysis (INAA) technique was used to measure the elemental concentrations in ice core samples after preconcentrated them by freeze-drying. INAA is a non-destructive, multi-element technique which is easy to apply to solid samples. A significant increase in marine elemental concentrations have been observed within the time period between 1969 and 1989. This increase coincided with the increasing occurrence of the so-called El Niflo Southern Oscillation (ENSO) events in the Pacific Ocean. The subsequent analysis showed similar periodicities for the sea-salt sodium concentrations observed in Antarctica and the sea-level pressure measurements around Australia which indicates a possible connection between these events. Similarly, a significant upward trend was observed for crustal aerosol concentrations after 1980. The observed ENSO related droughts in the southern hemisphere are suspected to cause this trend in recent years. At the same time, the annual average ozone concentrations homogenized between the 600 S and 90' S latitudes shows a strong negative correlation to the measured aluminum (a crustal aerosol marker) concentrations between 1979 and 1989. Recent findings about the importance of heterogeneous chemical processes and the formation of polar stratospheric clouds (PSC), on the Antarctic ozone hole gives some incentive for further study, knowing that crustal aerosols are efficient for cloud nucleation and ice-phase initiation processes. However, this short time period prevents drawing a strong conclusion from this correlation. The impact of volcanic emissions on the total atmospheric aerosol budget has long been recognized and became evident after the recent volcanic eruptions of El Chich6n (Mexico, 1982) and Mount Pinatubo (Philippines, 1991). In this work antimony and arsenic were identified as useful marker elements for volcanic emission signals in Antarctica. Although these elements are already known to be enriched within volcanic emission plumes, the establishment of the coincidence of their enrichments with the known 120 or suspected volcanic events was essential for their use as marker elements in ice core research. The most interesting volcanic episode was observed in 1940, and this event is both the strongest and longest in duration (1937-44) in the continuous annual sampling period (1926-1989) of this work. In addition to As and Sb, Cr, Br, Na, and Cl were also significantly enriched during this period. One strong candidate for this VME's signal is the 1937-43 eruptions of the Rabaul Volcano in Papua New Guinea. The lack of a high sulfate signal during those years showed that the volcanic marker elements approach can be very useful for the identification of historical eruptions from volcanoes with low sulfur emissions as this one was. Comparison of the pre-industrial and modem era elemental concentration measurements indicated some increase for the excess concentrations (i.e. crustal and marine contributions subtracted) of the elements As, Zn, and Cr in recent years. Since these elements are known to be enriched in the emissions both from volcanic and anthropogenic sources, the role of anthropogenic emissions in these trends is not obvious. However, decreasing antimony concentrations observed after the 1950's is thought to indicate decreasing impact from volcanic emissions because this element has been well identified as a volcanic marker element in Antarctica. Therefore, it appears that the observed increase of As, Zn, and Cr in recent years is partly of anthropogenic origin. Principal Component Factor Analysis (PCFA), a multivariate statistical analysis technique, was applied to identify possible correlations among the measured elemental concentrations. The results indicated three aerosol source categories for the measured elemental concentrations in Antarctica. These are crustal, marine, and volcanic sources as identified by their marker elements. In addition, one factor indicated a possible transition metal catalyzed release mechanism for gaseous bromine in the Antarctic atmosphere. This may be the cause of surface level ozone destruction as observed in Greenland. The accurate determination of the ambient aerosol concentrations is quite important because of their role in the radiation balance of the Earth/Atmosphere system To understand the relationship between the elemental concentrations measured in snow deposits and the ambient atmospheric aerosol concentrations, a simple empirical parameter (scavenging ratio) was utilized. The results, in spite of the high uncertainties, indicated generally more efficient scavenging for crustal aerosols followed by marine aerosols and volatile elements. Air content measurements of several samples from the NBY-89 ice core by INAA technique showed good agreement within each of two same-depth layers. Samples from the adjacent depth levels within an annual accumulation layer showed larger deviation possibly because of seasonal impacts on the air content. This new method seems promising for the air content measurements in very small samples from deep ice cores as the thin annual accumulation layers in those cores make the availability of large samples required for conventional measurements almost impossible. The availability of this method can ease the dating of deep ice cores, a task necessary for the understanding of the past environmental history. 4.2. Suggestions for Future Work The following considerations during the sampling, preparation, and analysis stages are thought to be useful to reduce the uncertainties in trace elements measurements: Cutting a sample from a depth interval that exactly coincides with a specific year, assuming that the ice core is dated precisely prior to the sampling. This will eliminate the extra uncertainty involved during the calculation of annual concentrations as mentioned in section 3.2. 122 * Separating a longitudinal thin slice from each annual sample for use in direct chlorine measurements (by melting and using an aliquot of it) assures the correct representation of a full year. * Obtaining samples for cross-profiling that are large enough to eliminate the need for a secondary cutting step as mentioned in section 3.1 will reduce the risk of contamination. * A special material for sample containers with minimal trace elements content, such as Teflon MT , will reduce the uncertainties due to the blank correction step as mentioned in section 3.1. The following subjects are suggested for future research using polar ice core trace elements analysis: * Continuous seasonal measurements of total sodium and chlorine for the last three centuries to understand the possible ENSO connection in more detail. In addition, utilizing more sophisticated time-series analysis techniques for examining this data set. * Analysis of more annual accumulation samples from the depth intervals that contain well identified large and moderate scale volcanic eruption horizons to further justify the use of volcanic marker elements approach for historical volcanic eruption identification. * Collecting aerosols with longer durations to reduce the uncertainties in ambient elemental concentrations so that uncertainties in scavenging ratio calculations can be reduced. 123 Improvement of the developed INAA air-content analysis method by finding additional methods for keeping the ice samples cooled during the irradiation and gamma-ray spectroscopy. 124 5. REFERENCES Abbatt, J. P. D., and M. J. Molina, 1993, "Status of stratospheric ozone depletion". Annu. Rev. Energy Environ. v. 18, pp. 1-29. Alfassi, Z. B.,edt. 1990 "Activation Analysis.", v. 1, 2, CRC Press, Inc. Artaxo, P., M. L. C. Rabello, W. Maenhaut, and R. Van Grieken, 1992, "Trace elements and individual particle analysis of atmospheric aerosols from the Antarctic peninsula.", Tellus, v. 44B, pp. 318334. Barrie, L. A., 1985, "Atmospheric particles: Their physical and chemical characteristics, and deposition processes relevant to the chemical composition of glaciers", Annals of Glaciology, v. 7, pp. 100108. Barrie, L. A., 1986, "Arctic air pollution: An overview of current knowledge.", Atmos. Env., v. 20, p. 643. Barrie, L. A., J. W. Bottenheim, R. C. Schnell, P. J. Crutzen, and R. A. Rasmussen, 1988, "Ozone destruction and photochemical reactions at polar sunrise in the lower Arctic atmosphere.", Nature, v. 334, pp. 138-141. Batifol, F., C. Boutron, and M. de Angelis, 1989, "Changes in copper, zinc, and cadmium concentration in Antarctic ice during the past 40,000 years.", Nature, v. 337, pp. 544-546. Beard, K. V., 1992, "Ice initiation in warm-base convective clouds: An assessment of microphysical mechanisms.", Atmos. Res., v. 28, pp. 125-152. Blanchard, D. C. and A. H. Woodcock, 1980, "The production, concentration, and vertical distribution of the sea salt aerosol", Ann. NY Acad. Sci., v. 338, pp. 330-347. Bloomfield, P., 1976, "FourierAnalysis of Time Series: An introduction.", John Wiley & Sons. Bluth, G. J. S., C. C. Schnetzler, A. J. Krueger, and L. S. Walter, 1993, Nature, v. 366, p. 327. Bodhaine, B. A., J. J. Deluisi, J. M. Harris, P. Houmere, and S. Baumen, 1987, "PIXE analysis of South Pole aerosol", Nucl. Inst. Meth. Phys. Res., v. B22, pp. 241-247. Bodhaine. B. A., J. J. DeLuisi, J. M. Harris, P. Houmere, and S. Bauman, 1986, "Aerosol measurements at the South Pole", Tellus, v. 38B, pp. 223-235. BOhm, J. P., 1992, "A general hydrodynamic theory for mixed-phase microphysics. Part III : Riming and aggregation.", Atmos. Res., v. 28, pp. 103-123. Bojkov, R. D., 1986, "Spring-ozone change in Antarctica and the role of the polar vortex", Adv. Space Res., v. 6, pp. 89-98. Boutron, C. F. and E. W. Wolff, 1989, "Heavy metal and sulfur emissions to the atmosphere from human activities in Antarctica.", Atmos. Env., v. 23, pp. 1669-1675. Boutron, C. F., and C. C. Patterson, 1983, "The occurrence of lead in Antarctic recent snow, firn deposited over the last two centuries and prehistoric ice.", Geochim. Cosmochim. Acta, v. 47, pp. 1355-1368. Boutron, C. F., and C. C. Patterson, 1987, "Relative levels of natural and anthropogenic lead in recent Antarctic snow.", J. Geophy. Res., v. 92, pp. 8454-8464. Boutron, C., 1980, "Respective influence of global pollution and volcanic eruptions on the past variationsof the trace metals content of Antarctic snows since 1880's.", J. of Geoph. Res., v. 85, pp. 7426-7432. Boutron, C., 1982, "Atmospheric trace metals in the snow layers deposited at the South Pole from 1928 to 1977.", Atmos. Env., v. 16, p. 2451. Boutron, C., and C. C. Lorius, 1979, "Trace Metals in Antarctic Snows Since 1914.", Nature, v. 277, p. 551. Buat-M6nard, P., 1990, "Global source strength and long-range atmospheric transport of trace elements emitted by volcanic activity.", In: The Long-Range Atmospheric Transport of Natural and ContaminantSubstances, ed. A. H. Knap, Kluwer Academic Publishers, pp. 163-175. Budd, W.F., 1991, "Antarctica and global change.", Climatic Change, v. 18, pp. 271-299. 125 Burroughs, W. J., 1992, "Weather Cycles Real or Imaginary?", pp. 48-50, 108-117. Cambridge Univ. Press. Cadle, R. D., C. S. Kiang, and J. -F. Louis, 1976, "The global scale dispersion of the eruption clouds from major volcanic eruptions.", J. Geoph. Res., v. 81, pp. 3125-3132. Carey, J. H., and C. H. Langford, 1975, Canadian Journal of Chem., v. 53, pp. 2436-2440. Carleton, A. M., 1992, "Synoptic interactions between Antarctica and lower latitudes.", Aust. Met. Mag., v.40, pp. 129-147. Charles, C. D., D. Rind, J. Jouzel, R. D. Koster, and R. G.Fairbanks, 1994, "Glacial-interglacial changes in moisture sources for Greenland: Influences on the ice core record of climate.", Science, v. 263, pp. 508-511. J. E. Lovelock, M. O. Andreae, and S. G. Warren, 1987, "Oceanic phytoplankton, Charlson, R. J., atmospheric sulphur, cloud albedo and climate.", Nature, v. 326, pp. 655-661. Comrey, A. L., 1973, "A First Course in FactorAnalysis.", Academic Press. New York. Crowley, T. J., T. A. Criste, and N. R. Smith, 1993, "Reassessment of Crete (Greenland) ice core acidity/volcanism link to climate change.", Geoph. Res. Lett., v. 20, pp. 209-212. Cunningham, W. C., W. H. Zoller, 1981, "The chemical composition of remote area aerosols", J. of Aerosol Science, v. 12, pp. 367-384. d'Almeida, G. A., and L. Schtitz, 1983, "Number, mass, and volume distributions of mineral aerosol and soils of the Sahara.", J. Clim. Appl. Meteor., v. 25, pp. 233-243. Dai, J., E.Mosley-Thompson, and L. G. Thompson, 1991, "Ice core evidence for an explosive tropical volcanic eruption 6 years preceding Tambora.", J. Geoph. Res., v. 96, pp. 17361-17366. Danielsen, E. F., 1968, "Stratospheric-tropospheric exchange based on radioactivity, ozone, and potential vorticity", J. of Atmospheric Science, v. 25, pp. 502-518. Dansgaard, W., 1954, "Oxygen-18 abundance in fresh water.", Nature, v. 174, p. 2 3 4 . Davidson, C. I., 1989, "Mechanisms of wet and dry deposition of atmospheric contaminants to snow surfaces.", In: The EnvironmentalRecord in Glaciers and Ice Sheets. eds. H. Oeschger and C.C. Langway, Jr., John Wiley & Sons. Davidson, C. I., C. Liyang, T. C. Grim, M. A. Nasta, and M. P. Qamoos, 1981, "Wet and Dry Deposition of Trace Elements Onto the Greenland Ice Sheet.", Atmospheric Environment, v. 15, pp. 14291437. Davidson, C. I., R. E. Honrath, J. B. Kadane, R. S. Tsay, P. A. Mayewski. W. B. Lyons, and N. Z. Heidam, 1987, "The scavenging of atmospheric sulfate by Arctic snow.", Atmos. Env., v. 21, pp. 871-882. De Angelis, M., N. I. Barkov, and V. N. Petrov, 1987 "Aerosol concentrations over the last climatic cycle (160 kyr) from an Antarctic ice core.", Nature, v. 325, pp. 318-321. Delmas, R. J., and M. Legrand, 1989, "Long-term changes in the concentrations of major chemical compounds (soluble and insoluble) along deep ice cores.", In: The Environmental Record in Glaciers and Ice Sheets, eds. H. Oeschger and C. C. Langway, Jr., John Wiley & Sons, pp. 319341. Delmas, R. J., S. Kirchner, J. M. Palais, and J.-R. Petit, 1992, "1000 years of explosive volcanism recorded at the South Pole.", Tellus, v. 44B, pp. 335-350. Deshler, T., D. J. Hofmann, J. V. Hereford, and C. B. Sutter, 1990, "Vertical profiles of ozone and aerosol at McMurdo Station, Antarctica, in the spring of 1989", Antarctic Journal, 1990 Review. Dick, A. L., 1990, "A simple model for air/snow fractionation of aerosol components over the Antarctic peninsula.", J. Atmos. Chem., v. 11, pp. 179-196. Dick, A. L., 1991, "Concentrations and sources of metals in the Antarctic Peninsula aerosol", Geochimica et Cosmochimica Acta, v. 55, pp. 1827-1836. Duce, R. A., J. W. Winchester, and T. W. Van Nahl, 1965, "Iodine, bromine, and chlorine in Hawaian marine atmosphere", J. Geophysical Research, v. 70, pp. 1775-1799. Dunbar, N. W. and P. R. Kyle, 1990, "Volatile contents of melt inclusions in anorthoclase phenocrysts from Mount Erebus: Implications for magmatic crystallization.", Antarctic Journal, 1990 Review, pp.7-9. 126 Ellsaesser, H. W., 1983, "Stratospheric water vapor", J. of Geophysical Research, v. 88, pp. 3897-3906 Fahnestock, M., R. Bindschadler, R. Kwok, and K. Jezek, 1993, "Greenland ice sheet surface properties and ice dynamics from ERS-1 SAR imagery",Science, v. 262, pp. 1530-1534. Fairall, C. W., K. L. Davidson, and G. E. Schacher, 1983, "An analysis of the surface production of seasalt aerosols", Tellus, v. 35B, pp. 31-39. Fan, S-M., and D. J. Jacob, 1992, "Surface ozone depletion in Arctic spring sustained by bromine reactions on aerosols.", Nature, v. 359, pp. 522-524. Farman, J. C., B. G. Gardiner, and J. D. Shanklin, 1985, "Large losses of total ozone in Antarctica reveal seasonal CIOx/NOx interaction", Nature, v. 315, pp. 207-210. Fletcher, N. H., 1958, "Size effect in heterogeneous nucleation.", J. of Chem. Phys., v. 29, pp. 572-576. Flossmann, A. I., 1991, "The scavenging of two different types of marine aerosol particles calculated using a two-dimensional detailed cloud model.", Tellus, v. 43B, pp. 301-321. Fogg, G. E., 1992, "A History ofAntarctic Science" Cambridge University Press. Frank, G., J. Tschiersch, and H. Behrens, 1990, "Wet deposition of tracer-marked aerosol.", J. Aerosol Sci., v. 21, suppl. 1, pp. S213-S216. Garcia, R., 1994, "Causes of ozone depletion.", Physics World, April, pp. 49-55. Gaudichet, A., M. De Angelis, S. Joussaume, J. R. Petit, Y. S. Korotkevitch, and V. N. Petrov, 1992, "Comments on the origin of dust in East Antarctica for present and ice age conditions", J. of Atmospheric Chemistry, v. 14, pp. 129-142. Ghan, S. J., K. E. Taylor, and J. E. Penner, 1990, "Model Test of CCN-Cloud Albedo climate Forcing.", Geophysical Research Letters, v. 17, pp. 607-610. Goldberg, E. D., 1963, "The oceans as a chemical system.", In: The Sea. v. 2, edt. M. N. Hill, pp. 23, Wiley-Interscience. Goldstein, R. M., H. Engelhardt, B. Kamb, and R. M. Frolich, 1993, "Satellite radar interferometry for monitoring ice sheet motion: Application to an Antarctic ice stream", Science, v. 262, pp. 15251530. Golub, G. H., and C. F. Van Loan, 1989, "Matrix Computations.", The Johns Hopkins Unv. Press, pp. 331-340. Gomes, L., and D. A. Gillette, 1992, "Chemical and mineral composition by size of dust deposited during dust storms in SW Tadzhikistan", In: Precipitation Scavenging and Atmosphere-Surface Exchange, v.2, pp. 921-932. Gordon, G. E., W. H. Zoller, E. S. Gladney, and A. G. Jones, 1971, "In: Proceedings of the ANS Topical Meeting on Nuclear Methods in EnvironmentalResearch.", Univ. of Mo., Columbia, p. 30. Girlach, U., and C. F. Boutron, 1990, "Preconcentration of lead, cadmium, copper and zinc in water at the pg g-1 level by non-boiling evaporation.", Analytica Chimica Acta, v. 236, pp. 391-398. Gorlach, U., and C. F. Boutron, 1992, "Variations in heavy metals concentrations in Antarctic snows from 1940 to 1980.", J. Atmos. Chem., v. 14, pp. 205-222. Grant, W. B., J. Fishman, E. V. Browell, V. G. Brackett, D. Nganga, et al., 1992, "Observations of reduced ozone concentrations in the tropical stratosphere after the eruption of Mt. Pinatubo", Geophys. Res. Lett., v. 19, pp. 1109-1112. Grassl, H., 1988, "What Are the Radiative and Climatic Consequences of the Changing Concentration of Atmospheric Aerosol Particles?.", In: The ChangingAtmosphere. eds. F. S. Rowland and I. S. A. Isaksen, pp. 187-199. Guerzoni, S, R. Lenaz, G. Quarantotto, and M. Taviani, 1992, "Geochemistry of airborne particles from the lower troposphere of Terra Nova Bay, Antarctica", Tellus, v. 44B, pp. 304-310. Hamill, P., and 0. B. Toon, 1991, "Polar stratospheric clouds and the ozone hole", Physics Today, pp. 3442. Hammer, C. U., 1983, "Initial direct current in the build-up of space charges and the acidity of ice cores.", J. Physical Chem., v. 87, p. 4099-4103. Hammer, C. U., 1985, "The influence on atmospheric composition of volcanic eruptions as derived from ice-core analysis.", Annals of Glaciology, v. 7, pp. 125-129. 127 Hammer, C. U., 1989, "Dating by physical and chemical seasonal variations and reference horizons.", In: The Environmental Record in Glaciersand Ice Sheets, eds. H. Oeschger and C. C. Langway, Jr., John Wiley & Sons, p. 99-121. Hammer, C. U., H. B. Clausen, and C. C. Langway, Jr., 1985, "The Byrd Ice Core: continuous acidity measurements and solid electrical conductivity measurements.", Annals of Glaciology, v. 7, p. 214. Hammer, C. U., H. B. Clausen, and H. Tauber, 1986, "Ice core dating of the Pleistocene/Halocene boundary applied to a calibration of the 14C time scale.", Radiocarbon, v. 28, pp. 284-291. Handler, P., 1989, "The effect of volcanic aerosols on global climate.", J. Volcanol. Geotherm. Process., v. 37, pp. 233-249. Hanson, D., and K. Mauersberger, 1988a, "Laboratory studies of the nitric acid trihydrate: Implications for the south polar stratosphere", Geophy. Res. Lett., v. 15, pp. 855-858. Hanson, D., and K. Mauersberger, 1988b, "Vapor pressures of HNO3/H 20 solutions at low temperatures", J. Phys. Chem., v. 92, pp. 6167-6170. Harman, H. H., 1976, "Modern FactorAnalysis", Unv. of Chicago Press, Chicago. Harrison, S. H., 1977, "Trace Elements in Water: A Development and Application of Sampling and Sample Handling Procedures.",Ph. D. Thesis, University of Maryland. Harriss, R. C., R. J. Charlson, R. J. Delmas, U. Feister, C. C. Langway, Jr., J. E.Lovelock, G. I. Pearman, R. G. Prinn, W. Seiler, B. R. Stauffer, and R. F. Weiss, 1988, "How has the atmosphere already changed?", In: The ChangingAtmosphere, eds. F. S. Rowland and I. S. A. Isaksen, John Wiley & Sons, pp. 201-216. Henry, R. C., 1991, "MultivariateReceptor Models", In Receptor modeling for air quality management, ed. P. K. Hopke, pp. 117-147. Henry, R. C., C. W. Lewis, P. K. Hopke, and H. J. Williamson, 1984, "Review of receptor model fundamentals", Atmospheric Env., v. 18, pp. 1507-1515. Hermann, D., and I. Olmez, 1990, "Source attribution studies in Eastern Massachusetts", Presentation at the 83rd Annual Meeting of Air & Waste Management Association, June 24-29, Pittsburgh, PA. Herron, M. M., 1982, "Glaciochemical dating techniques." In: Nuclear and Chemical Dating Techniques, Interpretingthe Environmental Record, ed. L. A. Currie. ACS Symposium Series, 176, pp. 303318. Herron, S. L., 1982b, "PhysicalPropertiesof the Deep Ice Core from Camp Century, Greenland"', Ph. D. Thesis, State University of New York at Buffalo. Hobbs, P. V., 1993, "Aerosol-cloud interactions.", In: Aerosol-Cloud-Climate Interactions, ed. P. V. Hobbs, Academic Press, Inc. Hofmann. D. J., and S. Solomon, 1989, "Ozone destruction through heterogeneous chemistry following the eruption of El Chichon", J. Geophys. Res., v. 94, pp. 5029-5041. Hofmann, D. J., J. W. Harder, S. R. Rolf, and J. M. Rosen, 1987, "Balloonborne observations of the temporal development and vertical structure of the Antarctic ozone hole in 1986", Nature, v. 326, pp. 59-62. Hofmann, D. J., S. J. Oltmans, J. M. Harris, S. Solomon, T. Deshler, and B. J. Johnson, 1992, "Observation and possible causes of new ozone depletion in Antarctica in 1991", Nature, v. 359, pp. 283-287. Hogan, A. W., W. G. Egan, J. A. Samson, S. C. Barnard, D. M. Riley, and B. B. Murphy, 1990, "Seasonal variation of some constituents of Antarctic tropospheric air", Geophy. Res. Lett., v. 17, pp. 2365-2368. Hogan, A., K. Kebschull, R. Townsend, B. Murphey, J. Samson, and S. Barnard, 1984, "Particle concentrations at the South Pole, on meteorological and climatological time scales; is the difference important?", Geophy. Res. Lett., v. 1, pp. 850-853. Horibe, Y., K. Shigehara, and C. C. Langway, Jr., 1985, "Chemical and isotopic composition of air inclusions in a Greenland ice core", Earth and Planetary Science Letters, v. 73, pp. 207-210. Huang, X, I. Olmez, and N. K. Aras, 1994, "Emissions of trace elements from motor vehicles: Potential marker elements and source composition profile.", Atmos. Environ., v. 28, pp. 1385-1391. 128 Ibrahim, M., L. A. Barrie, and F. Fanaki, 1983, "An experimental and theoretical investigation of the dry deposition of particles to snow, pine trees, and artificial collectors.", Atmospheric Environment, v. 17, pp. 781-788. Idso, S. B., 1992, "The DMS-cloud albedo feedback effect: Greatly underestimated?", Climatic Change, v. 21, pp. 429-433. Ivey, J.P., and D.M. Davies, 1987, "Ion Chromatographic Determination of Selected Ions in Antarctic Ice.", Analytica Chimica Acta, v. 194, p. 281. Jaenicke, R., 1993, "Tropospheric aerosols", In: Aerosol-Cloud-Climate Interactions, ed. P. V. Hobbs, Academic Press pp. 1-31. Johnsen, S. J., 1977, "Stable isotope homogenization of polar firn and ice." In: Isotopes and Impurities in Snow and Ice, Proc. IU66 Symposium, Grenoble, 1975. IAHS-AISH Publication 118, p. 210219. Jouzel, J., and L. Merlivat, 1984, "Deuterium and oxygen 18 in precipitation: modeling of the isotopic effect during snow formation.", J. Geophys. Res., v. 89, pp. 11749-11757. Junge, C. and R. Jaenicke, 1971, J. Aerosol Sci. v. 2, pp. 305-314. Junge, C. E., 1963, "Air Chemistry and Radioactivity", International Geophysical Series, v. 4, Academic Press, NY. Junge, C. E., 1977, "Processes responsible for the trace content in precipitation", In: Isotopes and Impurities in Snow and Ice, IAHS, Publication 118, pp. 63-77. Kaiser, H. F., 1958, "The varimax criterion for analytic rotation in factor analysis.", Psychometrica, v. 23, pp. 187-200. Keene, W. C., A. A. P. Pszenny, J. N. Galloway, and M. E. Hawley, 1986, "Seasalt corrections and interpretation of constituent ratios in marine precipitation", J. of Geophy. Res., v. 91, pp. 66476658. Keskin, S.S., X. Huang, I. Olmez, and C.C. Langway, Jr., "Trace Elements in a Dated Ice Core From Antarctica", (Abstract), American Nuclear Society, National Meeting. June 7-12. 1992, Boston. MA. Khalil, M. A. K. and F. P. Moraes, 1995, "Linear Least Squares method for time series analysis with an application to a methane time series.", J. Air&Waste Manag. Assoc.. v. 45, pp. 62-63. Kley, D., 1983, "The budget of tropospheric nitrate in remote oceanic regions." Paper presented at the 5th CACGP International Conference, Oxford, Aug, 28- Sep. 4. Knoll, G. F., 1989, "RadiationDetection and Measurement", John Wiley and Sons. Kotra, J. P., D. L. Finnegan, W. H. Zoller, M. A. Hart, and J. L. Moyers, 1983, "El Chich6n: Composition of plume gases and particles.", Science, v. 222, pp. 1018-1021. Kritz, M. A. and J. Rancher, 1980, "Circulation of Na, Cl, and Br in the tropical marine atmosphere", J. of Geophy. Res., v. 85, pp. 1633-1639. Kumai, M., 1976, "Identification of nuclei and concentrations of chemical species in snow crystals sampled at the South Pole.", J. Atmos. Sci., v. 33, pp. 833-841. Kyle, P. R., K. Meeker, and D. Finnegan, 1990, "Emission rates of sulfur dioxide, trace gases and metals from Mount Erebus, Antarctica.", Geoph. Res. Lett., v. 17, pp. 2125-2128. Lal, M., and R. K. Kapoor, 1992, "Submicron aerosols over Indian Ocean: some meteorological characteristics", Atmospheric Research, v. 27, pp. 291-303. Langway, C. C., Jr., 1958, ""Bubble pressures in Greenland glacier ice, Symposium of Chamonix, Physics of the movement of ice (Proceedings of the Chamonix Symposium, 1958), IAHS-AISH Pub. no. 47, pp. 336-349. Langway, C. C. Jr., 1970, "Stratigraphic analysis of a deep ice core from Greenland.", Geological Society of America, Special Paper 125. Langway, C. C., Jr., 1992, "A shallow ice core drilling project at Byrd Station, Antarctica.", Ant. Jour. U.S., 1991 Review v. 26, p.60-61. Langway, C. C., Jr., G. A. Klouda, M. M. Herron, and J. H. Cragin, 1977, "Seasonalvariations of chemical constituents in annual layers of Greenland deep ice deposits.", In: Isotopes and Impurities in Snow and Ice, Proceedings of IUGG Symposium, Grenoble, 1975. IAHS-AISH Publ. 118: 302-306. 129 Langway, C. C., Jr., H. B. Clausen, and C. U. Hammer, 1995b, "The stable isotope, acidity, and P-activity levels in a 1360-year-old ice core from Byrd Station, Antarctica.", Manuscript in preparation. Langway, C. C., Jr., H. Shoji, A. Mitani, and H. B. Clausen, 1993, "Transformation process observations of polar firn to ice.", Annals of Glaciology, v. 18, p. 199-202. Langway, C. C., Jr., K. Osada, H. B. Clausen, C. U. Hammer, and H. Shoji, 1995a, "A ten century comparison of prominent bi-polar volcanic events in ice cores.", Manuscript in preparation. Langway, C. C., Jr., K. Osada, H. B. Clausen, C. U. Hammer, H. Shoji, and A. Mitani, 1994, "New chemical stratigraphy over the last millennium for Byrd Station, Antarctica.", Tellus, v. 46B, pp. 40-51. Langway, C. C., Jr., M. M. Herron, and J. H. Cragin, 1974, "Chemical profile of the Ross Ice Sheet at Little America V, Antarctica.", J. Glaciology, v.13, pp. 431-435. Legrand, M. R. and R. J. Delmas, 1988, "Formation of HC1 in the Antarctic atmosphere", J. of Geophysical Research, v. 93, pp. 7153-7168. Legrand, M. R., C. Lorius, N. I. Barkov, and V. N. Petrov, 1988b, "Vostok (Antarctica) ice core: Atmospheric chemistry changes over the last climatic cycle (160,000 years).", Atmos. Env., v. 22, pp. 317-331. Legrand, M., and R. J. Delmas, 1984, "The ionic balance of Antarctic snow: a 10 yr detailed record.", Atmos. Envir., v. 18, pp. 1867-1874. Lindqvist, 0., 1985, "Atmospheric mercury-a review.", Tellus, v. 37B, pp. 136-159. Loureiro, A. L. M., M. B. A. Vasconcellos, and E. B. Pereira, 1992, "Trace element determination in aerosols from the Antarctic Peninsula by neutron activation analysis", J. Radioanaly. and Nuc. Chem., v. 159, pp. 21-28. Lowenthal, D. H., and K. A. Rahn, 1985, "Regional sources of pollution aerosols at Barrow, Alaska, during winter 1979-80 as deduced from elemental tracers.", Atmos. Environ., v. 19, p. 2011. Lowenthal, D. H., and K. A. Rahn, 1987, "Application of the factor-analysis receptor model to simulated urban- and regional-scale data sets", Atmospheric Env., v. 21, pp. 2005-2013. Maenhaut, W., W. H. Zoller, R. A. Duce, and G. L. Hoffman, 1979, "Concentration and size distribution of particulate trace elements in the South Polar atmosphere", J. of Geophy. Res., v. 84, pp. 24212431. Malinowski, E. R. and D. G. Howery, 1980, "FactorAnalysis in Chemistry.", John Wiley, New York. Mardia, K.V., J. T. Kent, and J. M. Bibby, 1992, "Multivariate Analysis", Academic Press, Inc., San Diego, California. Martinerie, P., D. Raynaud, D. M. Etheridge, J.-M. Barnola, and D. Mazaudier, 1992, "Physical and climatic parameters which influence the air content in polar ice", Earth and Planetary Science Letters, v. 112, pp. 1-13. McConnell, J. C., G. S. Henderson, L. Barrie, J. Bottenheim, H. Niki, C. H. Langford, and E. M. J. Templeton, 1992, "Photochemical bromine production implicated in Arctic boundary-layer ozone depletion", Nature, v. 355, pp. 150-152. McCormick, and M. P. and C. R. Trepte, 1986, "SAM II measurements of Antarctic PSCs and aerosols", Geophy. Res. Lett., v. 13, pp. 1276-1279. McElroy, M. B., R. J. Salawitch, S. C. Wofsy, and J. A. Logan, 1986, "Reductions of Antarctic ozone due to synergistic interactions of chlorine and bromine", Nature, v. 321, pp. 759-762. Meanhaut, W., W. H. Zoller, R. A. Duce, and G. L. Hoffman, 1979, "Concentration and size distribution of Particulate Trace Elements in the South Pole Atmosphere.", J. Geophys. Res., v. 84, p. 2421. Mdsz~ros, A. and K. Vissy, 1974, J. Aerosol Sci. v. 5, pp. 101-109. Michelangeli, D. V., M. Allen, and Y. L. Yung, 1989, "El Chichon volcanic aerosols: Impact of radiative, thermal, and chemical perturbations.", J. Geophy. Res., v. 94, pp. 18429-18443. Miller, N. L., and P. K. Wang, 1991, "A theoretical determination of the collection rates of aerosol particles by falling ice crystal plates and columns.", Atmos. Env., v. 25A, pp. 2593-2606. Molina, L. T., and M. J. Molina, 1987, "Production of C1202 from the self-reaction of the CIO radical" J. of Physical Chemistry, v. 91, pp. 433-436. 130 Molina, M. J., 1991, "Heterogeneous chemistry on polar stratospheric clouds", Atmos. Env., v. 25A, pp. 2535-2537. Molina, M. J., and F. S. Rowland, 1974, "Stratospheric sink for chlorofluoromethanes: Chlorine atom catalyzed destruction of ozone", Nature, v. 249, pp. 810-812. T.-L. Tso, L. T. Molina, and F.C.-Y. Wang, 1987, "Antarctic stratospheric chemistry of Molina, M. J., chlorine nitrate, hydrogen chloride and ice: Release of active chlorine", Science, v. 238, pp. 1253-1257. Moore, J. C., 1991, H. Narita, and N. Maeno, 1991, "A continuous 770-year record of volcanic activity from East Antarctica.", J. Geoph. Res., v. 96, pp. 17353-17359. Moraes, F. P. and M. A. K. Khalil, 1993, "LLSSA: Acomputer programfor spectral analysis of unevenly spaced data.", Technical Report, Oregon Graduate Institute. Murphey, B. B., T. Hare, A. W. Hogan, K. Lieser, J. Toman, and T. Woodgates, 1991, "Vernal atmospheric mixing in the Antarctic", J. of Applied Meteorology, v. 30, pp. 494-507. Neftel, A., P. Jacob, and D. Klockow, 1984, "Measurements of hydrogen peroxide in polar ice samples.", Nature, v. 311, pp. 43-45. Newhall, C. G.and S. Self, 1982, "The volcanic explosivity index (VEI): An estimate of explosive magnitude for historical volcanism.", J. of Geoph. Res., v.87, pp. 1231-1238. Noll, K. E., and K. Y. P. Fang, 1989, "Development of a dry deposition model for atmospheric coarse particles.", Atmos. Env., v. 23, pp. 585-594. Okamoto, S., M. Hayashi, M. Nakajima, Y. Kainuma, and K. Shiozawa, 1990, "A factor analysis-multiple regression model for source apportionment of suspended particulate matter.", Atmos. Env., v. 24A, pp. 2089-2097. Olmez, I., 1989, "Instrumental Neutron Activation Analysis of atmospheric particulate matter.", In: Methods ofAir Sampling and Analysis, ed. J. P. Lodge, Jr., Lewis Publishers, Inc. Olmez, I., A.E. Sheffield, G. E.Gordon, J. E.Houck, L. C. Pritchett, J. A. Cooper, T. G. Dzubay, and R. L. Bennett, 1988, "Compositions of particles from selected sources in Philadelphia for receptor modelling applications.", J. Air Pollut. Cont. Assoc., v. 33, p. 1392. Olmez, I., and G. E.Gordon, 1985, "Rare earths: Atmospheric signatures for oil-fired power plants and refineries.", Science, v. 229, p. 966. Olmez, I., D. L Finnegan, and W. H. Zoller, 1986, "Iridium emissions from Kilauea Volcano.", J. Geophys. Res., v. 91, pp. 653-663. Olmez, I., E.L. Fireman, and C.C. Langway, Jr., 1993, "Trace elements in basal ice at Dye-3.", Atmos. Env., v. 27A, pp. 2921-2926. Olmez, I., J. W. Beal, and J. F. Villaume, 1994, "A new approach to understanding multiple-source groundwater contamination: Factor Analysis and Chemical Mass Balances.", Wat. Res., v. 28, pp. 1095-1101. Orheim, 0., 1972, "Volcanic activity on Deception Island, South Shetland Islands.", In: Anarctic Geology and Geophysics, ed. R. J. Adie, pp. 117-120. Universitetsforlaget, Oslo, 1972. Ott, W. R., 1990, "A Physical explanation of the lognormality of pollutant concentrations.", J. Air Waste Manag., v.40, pp. 1378-1383. Palais, J. M. and B. W. Mosher, 1989, "Elemental tracers of volcanic emissions in Antarctic aerosol and snow samples.", Antarctic Journal, 1989 Review, pp. 217-218. Palais, J. M. and M. Legrand, 1985, "Soluble impurities in the Byrd Station ice core, Antarctica: Their origin and sources.", J.Geophy. Res., v. 90, pp. 1143. Palais, J. M., S. Kirchner, and R. Delmas, 1989, "Identification and correlation of volcanic eruption horizons in a 1000-year ice-core record from the South Pole.", Antarctic Journal, 1989 Review, pp. 101-104. Parungo, F., B. Bodhaine, and J. Bortniak, 1981, "Seasonal variation in Antarctic aerosol", J. of Aerosol Science, v. 12, pp. 491-504. Peel, D. A., and E.W. Wolff, 1982, "Recent variations in heavy metal concentrations in firn and air from the Antarctic peninsula.", Ann. Glacio., v. 3, pp. 255-259. 131 Peel, D.A., 1989, "Trace metals and organic compounds in ice cores in the environmental record in glaciers and ice sheets.", In: The Environmental Record in Glaciers and Ice Sheets, eds. H. Oeschger and C. C. Langway, Jr., John Wiley and Sons, p.207. Petit, J. R., M. Briat, and A. Royer, 1981, "Ice age aerosol content from East Antarctic ice core samples and past wind strength.", Nature, v. 293, pp. 391. Phelan Kotra, J., D. L. Finnegan, W. H. Zoller, M. A. Hart, and J. L. Moyers, 1983, "El Chich6n: Composition of plume gases and particles.", Science. v. 222, pp. 1018-1021. Phelan, J. M., D. L. Finnegan, D. S. Ballantine, W. H. Zoller, M. A. Hart, and J. L. Moyers, 1982, "Airborne aerosol measurements in the quiescent plume of Mount St. Helens: September, 1980.", Geoph. Res. Lett., v. 9, pp. 1093-1096. Powers, C. F., and D. L. Rambo, 1981, "The occurrence of acid precipitation on the West Coast of the United States.", Environ. Monitoring Assess., v. 1, pp. 93-105. Prather, M. J., 1992, "More rapid polar ozone depletion through the reaction of HOCI with HCI on polar stratospheric clouds", Nature, v. 355, pp. 534-537. Preining, 0., 1991, "Aerosol and climate-An overview.", Atmos. Env., v. 25A, pp. 2443-2444. Pyle, J. A., C. Bruhl, R. J. Cicerone, H. Grassl, B. Jahne, M. A. K. Khalil, I. Levin, Y. Makide, H. Oeschger, 1988, "Group Report: Trace Substances, Radiation Balance, and the Climate of the Earth.", In: The ChangingAtmosphere. eds. F. S. Rowland and I. S. A. Isaksen, pp. 259-269. Qin, Y., 1992, "Numerical simulation of pollutants removal by precipitation.", J. Atmos. Chem., v. 14, pp. 143-151. Raemdonck, H. W., W. Maenhaut, M. O. Andreae, 1986, "Chemistry of marine aerosol over the tropical and equatorial Pacific, J. Geophys. Res., v. 91, pp. 8623-8636. Rahn, K. A., and R. J. McCaffrey, 1979, "Compositional Difference Between Arctic Aerosol and Snow.", Nature, v. 280, pp. 479-480. Ramanathan, V., 1988, "The radiative and climatic consequences of the changing atmospheric composition of trace gases.", In: The Changing Atmosphere, eds. F. S. Rowland and I. S. A. Isaksen, John Wiley & Sons, pp. 159-199. Rampino, M. R., and S. Self, 1982, "Historic eruptions of Tambora (1815), Krakatau (1883), and Agung (1963): Their stratospheric aerosols and climatic impact.", Ouart. Res.. v. 18. pp. 127-143. Raynaud, D., and B. Lebel, 1979, "Total gas content and surface elevation of polar ice sheets", Nature, v. 281, pp. 289-291. Rowland, F. S., 1991, "Stratospheric ozone depletion", Annu. Rev. Phys. Chem.. v. 42, pp. 731-768. Rummel, R. J., 1970, "Applied FactorAnalysis", Northwestern Unv. Press, Evanston, Illinois. Russell-Head, D. S., and W. F. Budd, 1979, "Ice-sheet flow properties derived from bore-hole shear measurements combined with ice-core studies", J. of Glaciology, v. 24, pp. 117-130. Saigne, C., S. Kirchner, and M. Legrand, 1987, "Ion-Chromatographic Measurements of Ammonium, Fluoride, Acetate, Formate, and Methanesulphonate Ions at Very Low Levels in Antarctic Ice.", Analytica Chimica Acta, v. 203, p. 11. Schneider, S. H., 1994, "Detecting climatic change signals: Are there any 'fingerprints'?", Science, v. 263, pp. 341-347. Schroeder, J. H., 1985, "Eolian dust in the coastal desert of the Sudan: Aggregates cemented by evaporates.", J. of African Earth Science, v. 3, pp. 371-380. Schulist, S. 0., and G. N. Freeman, 1992, "A hybrid bin-moment model for aerosol nucleation.", Particulate Sci. and Tech., v. 10, pp. 65-72. Seinfeld, J. H., 1986, "Atmospheric Chemistry and Physics of Air Pollution.", John Wiley & Sons, Inc. Shaw, G. E., 1979, "Considerations on the origin and properties of the Antarctic aerosol", Rew. of Geophysics and Space Physics, v. 17, pp. 1983-1998. Shaw, G. E., 1980, "Optical, Chemical, and Physical Properties of Aerosols Over the Antarctic Ice Sheet.", Atmos. Env., v. 14, pp. 911-921. Shaw, G. E., 1989, "Aerosol transport from sources to ice sheets.", In: The Environmental Record in Glaciersand Ice Sheets, eds. H. Oeschger and C.C. Langway, Jr., John Wiley & Sons. 132 Shoji, H., and C. C. Langway, Jr., 1989, "Physical property reference horizons", In: The Environmental Record in Glaciers and Ice Sheets, eds. H. Oeschger and C. C. Langway, Jr.. John Wiley & Sons, p. 99-121. Simkin, T., 1994, "Distant effect of volcanism-How big and how often?", Science. v. 264. pp. 913-914. Simkin, T., L. Siebert, L. McClelland, D. Bridge, C. Newhall, and J. H. Latter, 1981, "Volcanoes of the World.", Hutchinson Ross Pub. Co. Simmonds, I., 1990, "Review: Improvements in General Circulation Model performance in simulating Antarctic climate.", Antarctic Science, v. 2, pp. 287-300. Solomon, S., 1990, "Progress towards a quantitative understanding of Antarctic ozone depletion ", Nature, v. 347, pp. 347-354. Solomon, S., R. R. Garcia, F. S. Rowland, and D. J.Wuebbles, 1986, "On the depletion of Antarctic ozone", Nature, v. 321, pp. 755-758. Stauffer, B. R., 1989, "Dating of ice by radioactive isotopes.", In: The Environmental Record in Glaciers and Ice Sheets, eds. H. Oeschger and C. C. Langway, Jr., John Wiley & Sons, p. 123-139. Suttie, E. D., and E.W. Wolff, 1992, "Seasonal input of heavy metals to Antarctic snow.", Tellus, v. 44B, pp. 351-357. Tanizaki, Y., 1990, "Neutron activation analysis of water samples.", In: Activation Analysis, v. 2,Ed. Z. B. Alfassi, CRC Press, Inc., pp. 377-405. Taylor, S.R., 1972, "Abundance of chemical elements in the continental crust: A new table", Geochimica et Cosmochimica Acta, v. 28, p. 1273. Thompson, L. G., 1977, "Variations in microparticle concentration, size distribution, elemental composition found in Camp Century, Greenland, and Byrd Station, Antarctica, deep cores." In: Isotopes and Impurities in Snow and Ice, Proceedings of IUGG Symposium, Grenoble, 1975. IAHS-AISH Publ. 118: 351-364. Tlisov, M. I., V. N. Filatkin, and I. I. Pilip, 1992, "Heat and mass transfer of drops and growth and melting of freely floating hailstones.", Atmos. Res., v. 28, pp. 21-39. Tolbert, M. A., M. J. Rossi, and D. M. Golden, 1988, "Antarctic ozone depletion chemistry: reactions of N20 5 with H20 and HCI on ice surfaces", Science, v. 240, pp. 1018-1021. Tolbert, M. A., M. J.Rossi, R. Malhotra, and D. M. Golden, 1987, "Reaction of chlorine nitrate with hydrogen chloride and water at Antarctic stratospheric temperatures", Science, v. 238, pp. 12581260. Toon, O. B., R. P. Turco, J. J.Jordan, J. Goodman, and G. Ferry, 1989, "'Physical processes in polar stratospheric ice clouds", J. of Geophysical Research, v. 94, pp. 11359-11380. Tuncel, G., and W. H. Zoller, 1995, "Temporal variations and sources of elements in the South Pole atmosphere: 2. Enriched elements and halogens.", Manuscript in preparation. Tuncel, G., N. K. Aras, and W. H. Zoller, 1989, "Temporal variations and sources of elements in the South Pole atmosphere: 1. Nonenriched and moderately enriched elements", J. of Geophysical Research, v. 94, no. D10, pp. 13025-13038. Twomey, S., M. Piepgrass, and T. L. Wolfe, 1984, "An assessment of the impact of pollution on global cloud albedo.", Tellus, v. 36B, pp. 356-366. Ueda, H. T., and D. E. Garfield, 1969, "Core drilling thorough the Antarctic Ice Sheet." CRREL Technical Report 231, Hanover, New Hampshire, U.S. Army Cold Regions Research and Engineering Laboratory, 20 p. Van Loon, H., 1972, "Wind in the Southern Hemisphere", In: Meteorology of the Southern Hemisphere, Ed. C. Newton, Meteor. Monogr., No. 35, American Meteor. Soc., pp. 549-552. Wagenbach, D., U. Girlach, K. Moser, and K. O. Miinnich, 1988, "Coastal Antarctic aerosol: the seasonal pattern of its chemical composition and radionuclide content.", Tellus, v. 40B, pp. 426436. Wallace, J. M. and S. Vogel, 1994, "El Nifio and Climate Prediction", Reports to The Nation, Spring 1994, No. 3. Wallace, P. J., and T. M. Gerlach, 1994, "Magmatic vapor source for sulfur dioxide released during volcanic eruptions: Evidence from Mount Pinatubo.", Science, v. 265, pp. 497-499. 133 Walton, D. W. H., 1987, "Antarctic Science" Cambridge University Press. Williams, S. N., 1995, "Erupting neighbors at last.", Science, v. 267, pp. 340-341. Wilson, C., 1968, "Climatology of the cold regions: Southern hemisphere", Rep. 1-A3C, Cold Regions Research and Engineering Laboratory, Hanover, NH. Wolff, E. W., and D. A. Peel, 1985, "Closer to a true value for heavy metal concentrations in recent Antarctic snow by improved contamination control.", Ann. Glacio., v. 7, pp. 61-69. Woods, D. C., R. L. Chuan, and W. I. Rose, 1985, "Halite particles injected into the stratosphere by the 1982 El Chich6n eruption.", Science, v.230, pp. 170-172. Zhang, X.-G. and T. M. Casey, 1992, "Long-term variations in the Southern Oscillation and relationships with Australian rainfall.", Aust. Met. Mag., v. 40, pp. 211-225. Zhuang, G., Z. Yi, R. A. Duce, and P. R. Brown, 1992, "Link between iron and sulphur cycles suggested by detection of Fe(II) in remote marine aerosols", Nature, v. 355, p. 537. Zoller, W. H., J. R. Parrington, and J. M. Phelan Kotra, 1983, "Iridium enrichment in airborne particles from Kilauea volcano: January 1983.", Science, v. 222, pp. 1118-1121. 134 APPENDIX A Elemental concentrations and related statistical information. 135 l-'-''T.··' I"·' r·'-- 0 0 '1 * 3 0 o .e :j 51 S · E· 9PG)IN o0i N r • J W IU od seO )ue•Jed 1o0 WueoJed T: t 0 N In .c. 3 O I., A 3 ,p. i, CN a' . . . . . . .. ' j: o o D o C4 c 0 Q N4 BIN lUDOJd _I( U D~ r K . . . . . . . . . . - - . . . . - . .. . . . . . 21 o }.. 0 0 0 S 310 IUSOOOd 136 0 fag uCa4d 0 0 0 0 r· •'',, · . ·'T- ·-- -- (0 o i-.. CD - . . CN o u 0U o c E 0 0 f0 IUODJu uo BIBCI cc setB •,ueole d 0- 0oJd C140 via(0O~ 0 ~''''`''''''''''' u, I 'i. o 1. E I o . - 1 ~~. 1 ,· o .; -h' o C~c; l"c~ o s. E. U 0 0 N i = o 0 a- o o 0 0 co e•O •,ueo~ed i w o · ' ' I '-·i-··1 · ·- I · · · -r i ~ 0 0 (0 0 00 0 1 (0 .? 0o C NN 1 eOa Wu0~u6d 0 0 0 _~YLLY 0 0 0 w N 1, 0 03 "5~,~ N ' 0 0 o ' ' ' ' ' ' ' ' ' ' " ' No d erz luOlOad 8W00 4UGOJ~d 137 U· 0 w0 (0 0, 0 ' ' ' c ~ ~ ~ ' __ 0 0 0 0 '·- 0 --- 0 , o 0 r -- · eo 0o o j t 0 o 0 N -3- 0 014 0 o 0 0 0 0 w -it N 0 i.,..~.,..l,~..l..~ o ,~,,~,~, u, a *ea luGOJOd N me - 0100 tuGDled 0 0r ~ I" ' I ' ` r--- 3--. . o - .E o i -"-Y,. i ~ L. ii 0 0 0 0 o 0 o\ 0 8100] ,u03Jd· eBta ueOoJed o o 0 F rr 0 o o E0e t o CN o i., 0 U ~ c.. ~.. i-`'--- 0, r N cc v E 00 0 N 0 N i_ ~. .~...I..,I...1...~ 0 NU o 0u o 0 O e0Q ,ueoJOed 0 0 0 No 018Q Iu831ad 138 ~0 _ i~~_ :0 I, ---5 0 o . I N i·-, ,.i1 8v =o ~ o tuOOJOG s • 0 = 0 0 . .0 . . 0 0 .. . (0 %k% o2 C U I`,. 0 0 o 0 o €. im Cu WCu d N B~waluO~Jad e00 `r-··l-··r···r c, r ~ -E. cn 'S i "1 i, CU 0 0 i·i 0 (0 Sco40 viea V lu U BIBO IUOOJ~d .0 N, 0 0ud 04 IUODJad 139 t) .E 0d ~ -I- q I 0 tlý - Cý C 1.44 N- C 00 r- cN - M - N " - t N "V44 m -4 t,- . C-C .O \O 't W) I r-ý Vý C0 000 0014 14 N ,IRTi4"- tr- C\ O\ r- ' t C r- C m C,N-VN a~: O 01, come ~ \ O C 61-41-4 Qn \O \O O c, d r;r; O U U U O ~ ci 0D - N C ro C ~( /u 16 ýd 06 v' tr' Ci"d z0 ,t r- w r- 4 0o v, I- C4e eOv 0 v ceCl, C ti r- O 0 a ( -N . c, t 00c, tr0 e00 0r - - r,Cw 4 - ON 00 ir r- cl: t 00 ki, Z , - O m m o mfCc 'Rm rm ",c"Nwr r tt-. v, q cCNCltcoc tcfý 000 ficc:, Cfi v0 Cfi Nv, 0 9 NNu' N' Ucjri c U U'j C PUUU N m 9m m9m C,4 C,4999 e1 00 tI -. - cq cC\ c c O' 06 C, d,; ar00-r r00 VIt r v ( N --" C- C1 N- 00 ý-%r,14c 4w: ',m F4 = , cr C\ ýO F C ýt,(,c 00 r-\ ý- Z - Fr-!ý I r-. N -f 10c 00w-O CC"cNzN\ rC,,O \. 1fC t wtczr NV; m i 5C S ý r- CN o0000oc "4"0 Cc·: CN00 odd 'ITcIrl, "T-- oor eq r r, oc-ý\6NC6c,tr, l-qt 4 t,4 6elf, ý6 0 trl 00 N T4 N mr 7-4 q"""4 c N I r -" ONm7, I 00 N -NN--- 0Copmeý4CýMM -4 -4" 0 4 00 WVl C-i I c tom>;I C", 0 C5 C06 C, N NNNC7, o V-, z I c-4 ,t V' t ý wO\Nc-4 q w r, tr,, -4tt, t e~,c: N M 'It tr, -" r-4 N m o 't UUUP ULjL;UQ g rW, C-4CC C UiiP C-)UUUl , ..qC,4 M DC I- . r W,W,WO O \O O W,, W-) S\ CC-04ON ON00ON rIt5 0"0"-"0'* \C wN =N m "t trýNO N C14CIA N ("I 00 eq N Cm mCr C4, 140 4-:, ý cc C, 6 --: d 00 0 ~ lrý C "I ' t 4- I c' Cl l 00 N C 1 ZZZ>\\t 0 cN \ O c~te r C~ClC~ClCl~ C t C t ~Cl l l z 1/OL/\I/ m ccc Co l CtC r-ý t-ý o0 x c ýczz "CNR = r tx ;t-O0 tproZ N C C', l 6Z l C , e C C - ClCl~ C-6 Lf: x c: 1 - 0 00 c C tr1 l m ,- zCt +1 "Ctrl ,N cr;c,~, 0 c· ~CfN ~~~~~-bCq CV' -ý -o `c Cq "t cr- 1-f-; C,'C'Nb-X3A rqt~ m ~~ ~~~~~ b C', c ~~~~~~~(emU + ~~ z 00 0000 00 Cý "SD22 d.0 ~ 0 ~ ~ - ~Clt~C~ ~1- c o\~ -~cl 8R S,\ RES ERS 0 22222nsass= ===aeacXOmedies .hCl zcOOcococoN 00 r-r00c (N 00 00It oczwcN o c-CýCýbc Cý\C o ,----; -- O q S 00 1 q C- q r-4 ý-0- ".(0"P" "" 0-4Q oc 00 u Cr\ CNCN m M I W · = 1· · 1-· 4zrr m m 1,"! tIt Y I lo t t 141 y ~ I L- r 11 l-ILlLl r rc L,1C- ýc 1C ým -4 Nci-w 0 cC ci C CN r r~~cWo- C cl- Ci Nc ci : ~t t-c r-c I-rl + COC + ci cýN Cý IO N m m m + oc rVr °I o o f C z r- t, z lr~ r-C \6 C4 . 6 7,o6 OC 1 4w mcN C r m N . ' c r-- NN m m m c\ r NN m m4N o + z mN c C\ Cýr: C N 66 4fv4 ml m N N r- mN CZ f-c - r N N -t C4 C t cf cr- L - m tr, N cmCm z Nm tr C CC 0e Cj, C m C ~~~ " c, 0 , '- xC CCC C4 t t m c x mmC4 m m , r- •, cmc. mc c. M rc -MC4X-c ~~- r-oc c, 4NNNNNN oooo•o•.... 8 -4 m , I C l - tZX- - oV 44 ·s a p 1 m m N t- - CiC6Cl!C- 000= c 00 E Z krv; Lv-lr tl-:lrcqV·l4x tl-V·f mc m tt y, r- OC C;DCO CN m r-- c, m CA mc mc mc qOrN 00 00 p, cu E ep ol ', M r- tc~~;v tr, I~t cle, - -Itt - fc1v z -- • clell mM Lrl t fl?"x L c ~M~M~~~~~~OCc~~oC~ C -4 i4 dq 0 0 rc" 0 c " .rT ~3bN~~~-~~N~bS-N~~NNN rw-•" " Lr 7t rIA, ev o mx ýT C oc mC -Nc,,, ý CNzmzz ci -~ .. ~0 c NN .. ~ o LfMM Tt -T It d'mtt emmyqq~r oo.. oco ~0-c -0 t r r oc. 0I - I oo m m 0C m-0 rCr -c4-1 CqIC'4cUI C4 C14C'4 r,4 c M m m CCI m mm m cr m t d ~tt: o, p r- w C = - t C t C r- oc - E c~J 142 - C r oc - i C r- I C\ 00 tf , - N Cm C\ Cl; N r O -. 'q d -- - 00 r0 C~C4 C (, ?CC ( · 0 C, N C 0 rC - C cmC ON , N 00 O\ eI N 0 0 00 r- d; wd r- 00 00 ,C rý N N - N ddso 00 dr-r-rcooo - 666 00 c z dcdddd ~"0 dir 00-~ 000 0 0 o66 W'ý- g6 ddd C - N o C66o 88888 N 0 660 ~C~~ ~ C~ I~CI C 00000000-00000000000000000000000 00= =0 C c C5 6d0 0 0 d + ýOCE-llq \6 er; cc; rcc:, 00oýr-C 00 T " 00 C\Ccor,vO" tr, MN 0 mCý O N NC" lN q:4 w q: ",QN " ,t6 oq 4 4M V . 4 00000~00*00~0*0*00000 odddddoooooo6 do cc ~ += C\ C 0 0r -- +, 0 Col N C-1C4 N 4 0 r- - 00 0 N Cf CC; 11t- V--( 0 -.m4 Vo c 00 co c 0 00- 00o0or- 0. r- r- r- r6 66ococCoCCooo66c ~ oocociddddco c 0 66o6c66co 00 er \ N0 cq t 4r-·I~~ Y 000C0*0*000000 6 T-4 00C Cf 0c 41 *N 4 ~~~~·r~~~ ~·ICc~ \-L VocM 6 .O cI oco 6 C5 C5 N tr; tt N tV,o r--.4 -.4 - d 0 -- f N ---4 02 6dddC~ dddd ol, 00 r~- r- r- i- N- 000000 cocco6C6CoooCoC o (, ' dc' c c Nc ON 00 r-- tr coadccoco d jg S-r--OM-0 -0 mm -- G -- 0 ý r- Ce( Co-- 4---Ir-IC 0 m -4w0-N -- 6N N NNNNNNN C666666 N 6q N C-4 C4J NN-NNNNNNN mccoo . . \. N. m. t . \C. . o . .~~"o 6G66666 66 OOO=====0O coococo NNN-N cc oo c o 00l.r-000r 66 C6 N 664 NN 0 0 ~ O r-oc w oc 0 OCQCOCCOO 0000 a=00 00 0* 0 m~-C4m m ' Cý i 66 r--N dC O c6c jdC5 6CO 5C ooooooooo666 rll qOQ oq O"t \q406Nr 0 c NC \q Cciwc0tNNNNN N Nc NNNNNNNNE 666 - 0 , C 00 6 ,..CCCC \ =6, 6C 6C cc 666666666dd -00 00 NNNNN +l 44 NO Wr-NNY-ZNO-NNQ -- C N ri 00 W~ C4Ir,Cs: Q-C-i 41: C-cr:1 clr~ tNN--N all ci Q Soo 800 888 COD 000 888880 menm W-4 66666dddddd - -- (*9 m~(UCc C14m m u u u u u N m IRt u u -u --- 4P" NNNygyONONOygor-r-r-r-ocuococ m qt I.-, u C.ý uU "." ) m wt W,, C.ýCd u ".4Nf V kr r ,4 m u C.C. 9C 0 NC-P u -" + -- co - --c -c---o--0 66666666666666C6 eac- - e - 6666ooo '- `7 C 0 000 -Ne - 0 0 ~e e.e -4 ON 06 006 0 Ne CCe 00000 0 o 00 00 w 00 w - r- ýO '& cj cj cl cli cjC Cj 00wr- '. rcliC 000*9i tr, m; ýc qO - ef-, r- tL ON ; cl - Cf c c'. ý rl r- r- zO r- r- '4- c&ficr0' ci j e4c; 00 mrC I r, cc "t Wý oc ýo t ýo r- W-i (2't ..... 66i dddddd ( C0 E NON V 0 0C0C0 6666666o66666 0 CC 00 I.o d ddddd 4 N - 0 C00 N dC4 C Cd 00 C 60e 00D 6o6 dU C C0 ooooooooo, U 6 coo c00 4dC4 -. 4 -4 6c00C 6c o666, O OCOC 00COC 0 o..o..o.oo .... o.o.o.. (3 Cr'- MCC00 tr r- 0 r- C. c dd (d ·- (C d ' 0 M. d r 00 -0 6666666666ooc 666 t, 1oo, V, c Ccc C C C dd o6 cocococcocococococococcocococod cm r- kill0 .-CON V) 0 C14 - N F.- = mc - cN (= o o-= = o o 0= c r".C·V;eOQ00cr,; 66666i(V~-1 666CCc oooo6666666666666666666 Z 00 cw c 00 Iid 00 ci N 6Ni 6 - U CN 00 ON 00 GN-C "-. --00000 co 00 00 66 666 ~N ON Os00 C-0C00 666 c c\ CU CON 000 F. 666 C-------CCO r0001 - F- 0" q CN o co6666c6 cT-4 C cd c c do c ''qt 0-04000 CIC c ON Or- -fCr ~O '.N.F- jC S- ONC0~ 'C~ F. ~ 0c ;CjCjC 00 00 ---4 -e1 jC _-;c oýcl 00ON\C\1 J C4 -oc r-co C y0C) 666 000 N m 14, co-"- W)-, W r t- 00 C e -- 145 -t W- m 0-1tm- C\-0 -l- -- eq qelf co dd -zO Cc 66d 99 cc c I + 000 id ccd ccot (C~ UCc C J TtC ccc(U~Ucc~ ~Cc cccccccc666666666666666666C cc CI:C~ CC CCcc ( c cc c~Uc 6666 9Z9O9OCOC99OCCOZOCCZGCcZZ cccc666666666666c cc6 ' ddcc 6 666:de c c~-·C99 OCc S00000X- 10 CccVCCCCcCC6 000000000000 CCNOCOOOOOOCCOO 00000r-0000C0000 cI=c cc cc' d:~C5 c N66666666666666mr-clN 06N tr r r r-, C cONct1 d cc66 N'cdc ccc~ c c OC6cc= c 00-~v F c c N p U c~cc -0 o -C NCm ONNO-· 000000 c· 66 v0cc-c0 66 c -- 'r, r-c 3Ccc C~CCc 00 ~ t- c 60 000c0000c0c01=c0 C C: C NN ~30Cl-C~~ CCCcccccccccccccc 00 -- 000000 · 00000 r~~ 0 NN 000 y 66666666666666'ctde:c:de'c'c~ +I - c 6666666666 'oCX- ------- N66666 OC OOC~-· - - mC:~~ - -- O +C O\I, •• 0 0cl 666 r -e• It C • =' C\ 00 "r, . . . . . . R \C W"q r" Z , ccrC, 8 8cc 8c cC, 0c00000000c000cC CC) C)C v; 0-'1 . . . . . 08c0c8 00000000-000. r- 00 d 6:C c C) (= c C 0 C) 0 = = O C C C (= = dC66666c6ccocccocccccococc6666co c;o-c-c 0-1 0-4 P" PP-0 P-4 0--" P" Oý 1-" Cý M Oý It 0-4 P" Pý ýT tr) v f 146 i c·C· Cd·~ v;r-r- cc8 CcccccccV C CCcc6OCC cc, - Cl k , = C) cc C ý ý ýc r- r- -C r- 00 oc Oc 000 00C 000 o66 8c8 000 666 \ 0 000 C\ 0,0 en o\ c a 0 a a d0 0 0 6 o dd0d 0006 8 Oc C 0. cC0 00 880888m88 a q 0 C 000 66666ddd +p U~ 0 c cq e t MwiN c0008-00cooococ sf~ Ir 1:o 0tr, w- 0 occ 66 cNOcoco qq -0 000dd 6 066a 0 0-0 4N4N m NNN C4N C14 N 888 N NN 0000C0000 666ddd Ccom CO5 d66dddddo6dd mocoSA cqq cocoococoj oc q59 r~- Ncq 0000~000.0.0.00.0.0.000-000000 666666ooooocooooooo66666666666 666 6dd cocccoC o00occ4· 000000 0 6 coo d6 66666 66 66666 r-o-oir--oco a.dCo -m k -- 6 ddC: -5 6 I:66 6 6 d6 •-- €' 0e 00 a O c0 CC dO 0 co t 0C00 4 6 666 -OC -000 0 \ 000 6 ee, 0 0 6 + ~~0000- 00-00..0.0..0-Q00-000- ddddd 6666ooooo6666666666666666666 N NC00 0000 -- 003~ 66666666666666666666666666660~; 6666 hi C8888888880088800808C 8008 )80 oooocoo66coco6cococc6co66c6oooooo 0C0 Q' 0 000 - -- C- 6 ci- 6 --- 6 6 6 d 0, -0 Cý0 Co- C) ý 00 00 CN CN Cý CN C% 147 t1TI Iti - cc, C C)C; c 00 dee' d 006 000 60006 C5 lddV) t I,~ 666 666 VV 6dd 66 0o 8 6 coo 4 ~1 + + N C~~ ~0C4~ m6 6 oneo66e 6 6 066 we _m d d dd d ddddiddd d CAe-A VW)OC) 0 oi 666 8e 0000 8 SS= o666 C14 clCl 000 0m000 888 o8" coo 60 60 cd5 C1 0 a a'd ddd No c z z c ;z C.C.4 o C\ c- moC-4N N os O 05.0ooc° .. ~ m9S ~0 d 0 0 0 0 dddd m C. o.w.C\ N r.IT m 66666 COO 6 0 occoco6ococococ6coccccccoco 0 \a 00 0 -( ddddddddddd 0 666666660666666 800 800 8800 cd 00 000 CMNC1 c cq 88888889888 6 OC 000.c 6o6 6 66666666666 C4 11 C14C C -4C* 88 d0d C-\c ~r r~C\ r o6c6 6m6 6 6 6 N ecs e> d d (sr C14 mt ci - d dococcoceed 4 C\ C e-N d m W"00 e.-6m6c6C6c d d d + cc c0c 00%0000000000 66 0 0c 6 66666666666 66 dd 0 0 0 0 0 C 0000000000000' 00 00 0 88 00 00 r 00 0 \r000 80000 dd O80. 666 ddddo o0 dd Crr- r0*0 0000d0d r- = ecMcM 0c0 0c0c00 0c U-, Cn' ý -00 000 6UOoo ES 666 ia + dd0)0 W 00 W0", 00)00 \ cV 0000 w '40 C 666666666666666666666660o t ,0 dda' 0: 0e -- do6666od 0 r Ld ddddddd CoCo r0 d od - O -ýc m, dd000aCM 0-oced00 CCNM ti5 V-om csoc0C\r- r- 00( 66666666666660000oo066666ddd elCf MCM '0 SC~C E8 88888088888888888888e"888888 ddddddddddddddddddd rn C goc 0000 - dddededad coococoocoocodddaccoodddddd 66 666 'e ----- r; 0N.~ - - -' - \, ,- NW 2,Nr f, 148 4., , , r w,• , C" , - Table A-2. Statistical summary information for the data in Table A. 1. Variable: Na Na (*) Mg Al Sample size Average Median Mode Geometric mean Variance Standard deviation Standard error Minimum Maximum Range Lower quartile Upper quartile Interquartile range Skewness Standardized skewness Kurtosis Standardized kurtosis Coeff. of variation Sum 97 52.96 34.97 34.84 40.15 1762.83 41.99 4.26 7.78 177.71 169.92 23.70 74.11 50.41 1.39 5.57 1.20 2.41 79.28 5137 97 48.35 36.10 35.12 38.38 1319.56 36.33 3.69 6.95 217.47 210.52 23.78 61.70 37.92 1.87 7.52 4.58 9.21 75.13 4690 72 5.88 4.83 4.56 4.56 16.58 4.07 0.48 0.26 21.89 21.63 3.15 8.11 4.96 1.58 5.49 3.78 6.54 69.29 423 97 15.70 11.03 10.70 11.43 240.93 15.52 1.58 2.26 94.68 92.42 7.21 16.26 9.06 2.84 11.42 9.63 19.37 98.88 1523 Variable: Sample size Average Median Mode Geometric mean Variance Standard deviation Standard error Minimum Maximum Range Lower quartile Upper quartile Interquartile range Skewness Standardized skewness Kurtosis Standardized kurtosis Coeff. of variation Sum Cl (*) 97 87.89 68.25 68.20 75.23 2952.25 54.33 5.52 21.89 301.06 279.17 50.31 105.01 54.70 1.59 6.39 2.45 4.92 61.82 8525 K 73 6.20 3.67 3.64 3.92 32.51 5.70 0.67 0.05 29.04 28.99 2.21 10.19 7.98 1.56 5.45 2.70 4.71 92.02 452 Sc 90 0.0009 0.0007 0.0007 0.0005 0.000001 0.0011 0.0001 0.000005 0.0072 0.0072 0.0002 0.0011 0.0008 3.59 13.91 15.87 30.74 126.64 0.08 Ti 73 2.62 2.21 2.15 1.80 5.05 2.25 0.26 0.12 11.52 11.40 0.97 3.28 2.31 1.77 6.18 3.86 6.74 85.63 192 149 Table A-2. (cont.) Variable: Sample size Average Median Mode Geometric mean Variance Standard deviation Standard error Minimum Maximum Range Lower quartile Upper quartile Interquartile range Skewness Standardized skewness Kurtosis Standardized kurtosis Coeff. of variation Sum Variable: Sample size Average Median Mode Geometric mean Variance Standard deviation Standard error Minimum Maximum Range Lower quartile Upper quartile Interquartile range Skewness Standardized skewness Kurtosis Standardized kurtosis Coeff. of variation Sum V Cr Mn Fe 89 97 97 95.00 0.03 0.02 0.02 0.02 0.0003 0.02 0.002 0.004 0.09 0.09 0.01 0.04 0.02 1.33 5.12 2.76 5.32 59.55 2.4 0.91 0.64 0.59 0.65 0.65 0.81 0.08 0.08 4.11 4.03 0.35 1.23 0.88 1.81 7.27 3.40 6.84 88.70 88 0.22 0.13 0.13 0.14 0.13 0.36 0.04 0.03 2.63 2.59 0.08 0.23 0.15 5.20 20.89 30.19 60.70 161.23 21.4 11.44 8.30 8.16 8.54 210.56 14.51 1.49 1.74 122.92 121.18 5.78 11.92 6.15 5.73 22.80 39.62 78.82 126.82 1087 Co Zn As Br 56 97 97 95 0.0055 0.0025 0.0023 0.0027 0.0001 0.0097 0.0013 0.0002 0.0569 0.0567 0.0013 0.0061 0.0048 4.24 2.07 1.57 1.56 1.55 4.12 2.03 0.21 0.34 12.87 12.53 0.89 2.27 1.38 3.26 0.011 0.008 0.008 0.008 0.0002 0.014 0.001 0.001 0.122 0.121 0.005 0.012 0.007 5.40 0.21 0.16 0.16 0.13 0.04 0.20 0.02 0.003 1.06 1.06 0.06 0.30 0.24 1.85 12.96 13.11 21.70 7.35 19.51 29.81 174.64 0.31 12.90 25.93 97.97 201 38.44 77.28 127.03 1 4.30 8.55 95.70 20 150 Table A-2. (cont.) Variable: Sample size Average Median Mode Geometric mean Variance Standard deviation Standard error Minimum Maximum Range Lower quartile Upper quartile Interquartile range Skewness Standardized skewness Kurtosis Standardized kurtosis Coeff. of variation Sum Variable: Sample size Average Median Mode Geometric mean Variance Standard deviation Standard error Minimum Maximum Range Lower quartile Upper quartile Interquartile range Skewness Standardized skewness Kurtosis Standardized kurtosis Coeff. of variation Sum Mo 66 0.012 0.008 0.007 0.007 0.0002 0.016 0.002 0.0004 0.095 0.095 0.003 0.015 0.012 3.21 10.66 13.37 22.17 127.29 0.80 Cd 45 0.012 0.006 0.006 0.006 0.0004 0.019 0.003 0.0005 0.121 0.120 0.003 0.013 0.010 4.78 13.08 27.12 37.14 161.56 0.52 Ce Sm 48 84 0.03 0.01 0.01 0.01 0.02 0.15 0.02 0.000005 1.08 1.08 0.003 0.02 0.01 6.88 0.0015 0.0005 0.0005 0.0005 0.0001 0.0084 0.0009 0.00002 0.0775 0.0775 0.0004 0.0006 0.0003 9.11 19.46 47.55 67.24 34.10 83.35 155.94 470.13 542.81 1.6 0.13 (*) indicates direct measurements on 1-2 ml liquid samples. Sb 95 0.14 0.07 0.07 0.07 0.04 0.20 0.02 0.003 1.56 1.56 0.03 0.16 0.13 4.41 17.56 26.84 53.39 147.37 13 La 89 0.011 0.002 0.002 0.003 0.003 0.058 0.006 0.00003 0.553 0.553 0.001 0.004 0.003 9.31 35.86 87.43 168.37 553.18 0.94 Table A-3. Depth-age relationship for NBY-89 ice core (Osada, 1994). Depth Interval [m] 0.00-1.00 1.00-1.64 1.64-2.00 2.00-2.45 2.45-2.68 2.68-3.30 3.30-3.52 3.52-3.97 3.97-4.18 4.18-4.54 4.54-4.74 4.74-5.00 5.00-5.17 5.17-5.46 5.46-5.80 5.80-6.00 6.00-6.20 6.20-6.48 6.48-6.59 6.59-6.79 6.79-7.19 7.19-7.42 7.42-7.60 7.60-7.87 7.87-8.00 8.00-8.19 8.19-8.37 8.37-8.52 8.52-8.61 8.61-8.77 8.77-9.05 9.05-9.25 9.25-9.35 9.35-9.60 9.60-9.80 9.80-10.10 10.10-10.19 10.19-10.40 10.40-10.59 10.59-10.87 Date of Deposition [A.D.] 1989 1988 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 1972 1971 1970 1969 1968 1967 1966 1965 1964 1963 1962 1961 1960 1959 1958 1957 1956 1955 1954 1953 1952 1951 1950 Depth Interval [m] 10.87-11.00 11.00-11.19 11.19-11.30 11.30-11.50 11.50-11.71 11.71-11.86 11.86-12.00 12.00-12.12 12.12-12.50 12.50-13.00 13.00-14.00 14.00-15.00 15.00-16.00 16.00-17.00 17.00-18.00 18.00-19.00 19.00-20.00 20.00-21.00 21.00-22.00 22.00-23.00 23.00-24.00 24.00-25.00 25.00-26.00 26.00-27.00 27.00-28.00 28.00-29.00 29.00-30.00 30.00-31.00 31.00-32.00 32.00-33.00 33.00-34.00 34.00-35.00 35.00-36.00 36.00-37.00 37.00-38.00 38.00-39.00 39.00-40.00 40.00-41.00 41.00-42.00 42.00-43.00 152 Date of Deposition [A.D.] 1949 1948 1947 1946 1945 1944 1943 1942 1941 1940-1939 1938-1934 1933-1927 1926-1920 1919-1914 1913-1908 1907-1902 1901-1896 1895-1890 1889-1882 1881-1871 1870-1859 1858-1847 1846-1838 1837-1831 1830-1823 1822-1816 1815-1808 1807-1800 1799-1792 1791-1783 1782-1774 1773-1767 1766-1760 1759-1753 1752-1745 1744-1737 1736-1729 1728-1721 1720-1712 1711-1705 N w-: C,:-!-C,:-O jj m ý r r- , =- 4 i-- CC-i i eD~* cN~cU~ . 0 oý4 + - e =\ t, cj It , '- ~ ti-, ýc 9 ~r-I~r c,· as~~~ V)~rV-4 ~C ~3~ n - a, w N t r-c- d: r.4r , m,• + c zN 0 (ON C14 0 CNo +1 w ~ ~G 06 6 4 N m 000 + - tr, ; r-ý N C I WEl ý6 r- m-- * N C C\ m 00 N m y 16 16 Ci qt ( \6 N c W w, 0 't - - -(WtN-4 clý Ch m w C\ C,4 Ln O N N r- ~ a~ r: -t Cf~·r C":, Ni Ci kr, ,:- I--C r- z , N 48 - Im tr \c l hi m=z cN C, CV,~-IC C5 C:- O m = C w o Ci ýc d 06 m N d r- ef" W:. r-: ql 3· z C'ý, tjN·, c , Cý Z I 14*z z en - C\ (2, "ý - N r 4 C ·I-,a N a ,T"4 z N in -c- Cý .4 ýq W- ý%, "4 -. = t' 0C c ý NO V-4 - r'4 tt-", C0\ ,- 0 om -ý NcCNC C -4 -. qtmN Nc, -4 -- - - r-m r-c-- = mr- 0C -4 C\ ONv -4 "- C\ - \ " - CNC, C\ 4 4 .4 cf -ItmN ý -. C -4 - "C 4+ NC4 C\ z= ýC LI 0000 Z C " e NSCEN C 'r C', r r C'4 - -. -*-, - (Ni 4 tr) 4 r-ý r-- -- rfý \c C,4 - \DO~~NZZZ~~NCN os C-4 rl rol L, r N N0N,0 t it, + cCsr-'r- , c. -, m• Z "t -, +) N( N N E. N N e mC, c00C + N N m N N N N N q 1 N2 C N m EN N m 00 m N- DECC O' V0 C:N 0 N L, f ,*c oC N - C C xC C zf C-4~~ 0 00 4 Zr Z t 0Z -N C 'Itre, ~c~ ~ ( iCCr~ 4~ C, C "tt m C,oc r-~ z CCc · t0 NC4 CNC NN - NN HNN N N C N m N · N NflNl C C, N NN c, Nq N N C,r- r- mN N N r• rN rý 0 r, ý' \,c~ tr, N Cj Nt c- N N -00 rI r-, -N C N •,m C NNN N m t =E oc ,4Lt 00Nr, 00tN , VC. N mP N m r~lrN C4 tm N N N N N (-N m . . .N T c-oc r- z v-, m C\ Cl\ C4 C\ C'~N~, I" 154 4 I"'" - C - - - Nc, C - - N - - - - CN - + Cl* m N-t m e C m e -- - - - - N tt -l - - tC r0 cj N dd z tr, c,z r-r- z rz z cý, cý z wr- wz w0'-' w cU - O CO mro S, kf0it C re m, m, m + C C r-N ClC~ Cl + r 09 tW em >-- C > <a Cell NC mN C, C O W-. cl N ClC j:n rC oo6-, C, o r , Ci V7, k r, l c, N ,C C i~ o , -I- Cl Cl r-0 o ý t tr NaNc z0C-C - a e e CU·Cj wwCC wj s e c Mggace 4- W r ZC rZIt 0c NClm C, ýVO a - - rr, Oa mClN Ce, NC Cl CN C-4C C1 CN Cl O 0 r' r 9 01 lCf-C = Z C4 t o lc -ý q r cr N4 z v,cc C-l C Cl Cl Cl e - C cl Cl Cc, 3~~~NN- M CNl It Cl o- - vN ttNNN m e C C- Cl4 Cl CC\ ON C\NONOC\ 3-- C- W Cl N -N N-NN-e Cl -No - m~ ac-m a m N m - - r M N 0N Cc> = 00 Z 0C W Om W > NONww 000 - r = Cl Q-, r> C tm C C\ x00000000000000 - - -- 155 Cl t> c c~ e - lC14 m 00> wN -- r Ct 't a cm-- ClN m N~ Cl -man mcNNNNm ýc ýCl C- 00 w Cc tr , Cl Cl r0>>>>>>>>> -- ClMlta cocoC a----------------ItcNN N N NmmNc c cMM N Cc N N N C', ZCZZCZZCZCZZ Zc ~ T~· C;Chj + c r r 33Cl Cl Cl oIO OCO 6 0 C: CCCdddCCCCCCCCCCCCCCCCCCCCC Cl, C CC % O CC C^O~i C O~CCCCCCC Cl Cl C= - C-· CC ccOcO O =ro cococOacoocCO~CCCoCCCCCCCCCC Cl -· "C OE rzoz CC ccoc CvCCCCC 666666666666666666666666666 6 cl cl~ Cc Clj C, Cl .C, Cl C ~ OC)OC Cl rC rCl Cl - - - - j r Cl c r . C, ( , Cl Cl -" N N , Cl . Cr . ,.o, + CCCCCCCCCCCCCCC CCCCCCC 6 6 6 -f~ r- C 6 6 Nr Cý 6 CCOOC CuC 6 6 -w- 6 SCr CIC4 C4Crl CCCCOCCCCCCCC C666 6 z~ N, 6 6 6 N 6 r- r Oj C C--ClNl + oCl coac -C~C Cl e-. rC NC- - coc Nede cNC oc Cl Cl c Cl N Cl Cl ~ N t- 0,, oocoackrccZ CCCCC CC - 66-6--cCCC-Cc 0vC z 0C- CCCCCCCCCCCCCC C\ C14 C4 66666666666666 JccocLcoctr, Lr , CC, N 00ý Cl-CNl-CocCCCC co---4 lc1cc N NCcl Ot~CClNN~N0 6 x- c c c= -ýc c' cl-' 1 + wi r-4 c -. 9,C ; -ý 6-~~~ 666 d C= c C6 6 6 CN c:de CC dC-C- 0 c:c8 C ,-0oN=C- wCo\Q\ It mCC lý r·- r. rVTl clý Ccli Cýw Cý r-4 60·Fc r6 C\--;C-4 -ci ; Cý-; TId2\ cr-5cC4 -- =-·3 S C-=- w 00 lt 00 00 00 \ I" 00 1- C--\1.0 14 -* w00 00 00 oN \ C-----. C- -- rO --Cý "" rN rr4C\-( ----- CI\ 156 Nrr-rz -- zSzzC Cý ---- v-Cý S x - N 6CCUCC cl Cc C CN C\ W 0 - 6C 6 Czz0 ý - coctedd66666666666666ddadedage + a Ico-N Nve -- -a--- M OCýc 8888888 co oco 8CcC dd CCC 8 cooco ccooc 0 ed --- r-X t- X co dddd x 5C ccd cococ cc 00ON-Cl C c· Cv c=-c o Z - ~CCzzzz~6 C~ z z C c' cceZZZZdd +) =ae M ( =M cli mcl Cl MCZ Cl x x CN C N' C Cl Cl C tr 0 cN -4 Cl CN ClNl C rN C ,ý;r, .'I = = ý 06 r-: zýd W, k C \dc d m ci =-I-I rCf\dNC·t~ZCl- z x Cl ClCr ,c x Cl cf:C ý r Cý x C 0000 v Cl ClN CN c4 ~ /3 Z\ococ.d~c~dCN- +, 0r SEEMESESSESSESEE + COC +l C,,zN 0000 OC oc \Cl N BBEBBBBBBBEUS c:-CCCNl---- oC zC\ r - C -r C\ 3 C o 00 r- zCi dm oNclCoarc C - Lr -00 tc, Lrekr, 000 x cN rN tr N- 00 ON t ~ ~- ~ ----------------~ ~ I cx e nter cococo -Cl Zrcf - c oýC=CoCo 6 6 C5 cwc m Cl ZC m ee N Cl--Cl---Cl occccoccococccceccccecco - Cf00 X + 00,ý Cl~c , Cll Cll Zr COC N o m 6c 6 6 \-Z0- "C Cl C 00 Lr oCl Cf -r- -C--5Z cC C O N ttNC Z 6cOCO O 1-4 c c No-~IN' rZ ocococ 00 X , t N cl00NO 6r-C0 -4 c da- 606 jN -; c 00 Cl 00 '-; o l 6 'N -; 4o N x - -;ddCj e C=c Cr, t \C ON\ 'f, N- ClC c l0fZCl w a e Cl4 0000oN o cococ VZ Ir ao Lr e-Cl-ed C O Czcz o d N I, -,O - ccd&a NZX ir;N - m · dd d co-moo-emon-ovoco-a-me-oc4Mddd S U 41 -( 4" *1"-4 -1"--4------- - EEEEEEEEEEEEEEEbmibbbBEEEEEEEE 157 ,A------- ------ ------ -------- ---------- -- 4---- ocococococococo-cocacc(Nci-ci- 0---CAj oo-c-coa---- Iý CCCCCCCCCC8CC""~CCCCCCCCCCCCC CCC CCCCCCCCCOCC 0 C O S CCCCCCCCCC DCCOCCCCC CCCC CCCC d +) N 00 N 00 N trv~, oo N N v-,: N N N N N r Ct C r- Lr, m 6C I.- & 6 C & 6 Cc ME rq NNNNNN cIccco ~~ ~ ~ ~ ~~ ~ c ~ 6coc 66N6zccmzItt 0 yc·( N 0ý \g j ;V ~~f N ":NC C C ~ ~ ZNN CiP CI -\ \ m r - t , cl\ -( C - C- - X -rý kN \, cocC -e0-P 00 .- ~Cý N-,r,ý C 3- C N Z CtCc N \ - = x w 158 00 Nr- - lý - C, r N - Cf - - - C ?~C c0C00NNNNNNNNNNNN -- 00 I-. 6XZE N ci 00r 00tr ZN ct- s e s e cce d cc Z6C6CC6C6ZC6Zc N C CCZN00000Ci~CZNZN00Nbcic 00CN citNZ-i0c~ ii-0 ONZNZN0000000000000Mm -------------------- CCCC - r ic -~ -i c e(s f( ý C - ccc ci V ' i Z I-~c OCCCCCCCCCCCCCCCCCoCCCCCCCCCCC + C N NN CN- c L BE N 6 6C 6 C6 C C C C C C C CCCCCCCCCCCCCCCCCCCCC 6c6cco6Cc cccCcccCc . . . . . . . . . .c . .e coc cc, 0 m\c \c oc r- r- rý CCCCCCCCCCCCCCCC CCC5CCC &CCCCCCCCCCCCCCCCCCC~CCZCCCCCC NNNNNNN m V CP, N- - -- C C - - - , r ~X0 r-0 00 cV c' ýc 0o -- r - (N (N e, C~ C, N m A CIN mN m mm C14C1 m C C C ('N 0000000000 -C*N N( 000000000000 'C14 m r, m (N CC + 00000000000000=0000000 C, 000000000000000000000000000000 0 ., COE -0 c 8 qt z00 000000 ococo 0000000000000000000000 66666666666666666666666666666 dco d 'dddddooco dddddddddddd gE; gegg a oo 000 1 gg g semo~o -mo 0c m .0 t dd a) 8 C0 a0 0 0c) '3 dddddddddd 66666666666666666666666666666 N-ROGYNNCCO SC\ 6aMoNCCC. T-ý cf) N er) ~ N o0o 0CO0 C - 0 ~ aa 0000000000000 " ddddddddddddddddjdd q N N N C0 0 0o tCI 0 0 c00000CooO 0 00 - 0 NNCCOO C, N- N( cq en N eq Cc, (yC.4NC. 0 00 o CO ac I00 ( ON NO N 00CNc C000 N 66666666oqooc6ooocooocooo666o rn U qoq0000qq cc cooa 000 .. ...........oco. 66666666000066666666666600006 00000000000000000000000000000 C") Mr--00 000 aNO 000000 ON N NO erC ON coooo666oooo6666666666666666oc na~mom-m--ocumemmw No d'ddddddddddd " eq 00 m 8088088880 N ddddwmdwdd'rO a Ir(U , (:r Cc, 0000O000 0400000000 SON ON ON -0 C:ý0 80888o 000 0 0? W, r- a e c- a'ddd (N 0 NO (N N C - C0c 0 N W r- (2" ciC5 C5 me--oco t- a'dddd ( N. r (:,N cN ON cc cc ( d'dddddd O i 00 r00000 -00 NO C,-,0N NO C NOr0 cN00 0 r- O'-e(N-e 00 "1-In0C C CC cc t- m 00OC0C -ti · ('- cqN (N 00O 00 0 ON ON O0 0 00\Vj00N ON ON C--- 0 0 C- 0 14 o 0 0 - C c 00C Nd c.4 C CC t l%1= C O 0 ~~ V -- ONNONONONONOa 0S1 -rONON ON ONON O N ON O N N O: CONN\O O C C C - 0 C.) a) eP + = v~fu-· at c8d o Cc,~N C c c0 rC56cC5 (5 6C5C5C5C5C56 mrN p r- \ c, 00 -c0\0 1-w 0oc00 Cc 00 oe 0 00 cC, cý (7~ (2N (2 (7N \0A 0\ CON CN ON 0\ C\ CN a', ON 159 -, = = 0 't0 cf d0 ~Z~,,~6go~ C C0 0 c q oo r CIS C CN CN CN \ CN C W -- ~~ X-C VI- t- 091 4L. - %A',A ýA 33333333333333333333333 333333 O-4009040-O ~3333333333"33 '_Ipl ýp~~~L +! ~~90000-00-0-O ~ 8Sg -0 0-i co O 3 +5 000033333003000000300000000000 3 a3 3-3-a 33300333333033333333333333333333 0· 033003003330033003000030000003 c ococo oc ON Z S S^ 00 -.j \C t x c co C1-1 x cC o +j Qk~k~k)4 kQ~ -333 tQ3 WtQ 033333333003303003333333030303 88Z Z ZZ888 33388333Z33883333r333 S3 3 N) tj 300 t 3 3 3 t00 0 3X 3 = 3 3 tJ= 3 3 3 3 3 3 3 3~N~~~ 33Z5333 ~ 3 30 tc000000 3 0333333333333333333 bjbjj"8888jj8888 3333c33 N 303030333 xam 3 00 3 3 30 333 W t".) +E + N -3 ~ 33333333 3~3 88898989888888 N~~N~~ Moom-4 33333 ~ex 030033333333303330333003333333 b b S S S S S S 333333333-33333333333333333333 = --------3 0 3 0 33 00 J) 3 ~> 33033333333 N>> C wA S~ bbjR C'=----------s-=----X X C1 I \.C 3 000 C333 O 33 3 333 uM cu 33 + 33 +! 3333303333333333333333333333333 + 333333333333333333003333333330 333333 3 Z 3 3 3 3 3a 333 --333 -33----3-3 33 - jbbjjjj j0 S Sbjb 33 8 jb~b 0 3 - ---e -- 300333300033033300300003000303 3b 3 o0 330330000303330003000 33333333333333333333303333333 00033333333333333333300033333 'OC88 '0 ZZZZZZZZZZZ0ZZ~8 0000 3000 33 4. Z3Z3ZOZZ3 3330333 33 33 33 333 333 j 3\3Z3 -.C\ 11 CNVý 33 I =j55~~~~WN~~N -A3 33 33 -JI-A3 33 33 -Q-33333 _ 3 + 191 I --j *+ -4 -.1 -4 C) C\ t 00N) "C CN ·2W t f.-'-, 00 00 00 00000 --j 000 000 00 x I30 c P -O \C I': W \4DC,% --j ýz •. o o. Eii- N =-- ~b + 0co NlN-NNiiiia iiiiiii o0 0 0 + PPPPPPPPPppppppppococpopppppppc 0t 0~ 0. = .0 ppppoooo N- 0.N. N1 -CN PPPPPPppppppoococoppooc 00--coo0 - -JI N \I- _In N ON --- %A -JI'~' 0o0ooococoooooooooc V -01tjV 0C JC Jt. WJ O00000300000pppeacococococooco tZZZZ ZZ %A t- 00 tQ - N) Q Cý t0 000O J O 00 W~ O w N)' 0 0 0 00-0-000000ooooooooooo Eg o ON900N00000000000-000N000 ooco +; 010 ,...,,. 0000 000000000 t")t-j \CtQt-j J~ OPPooooooooooooooacozooz g tQ ONJ - N-- J t-. N OOZ000303p30 N- 11 NN N,- '-,I tJ N 0 N NJ NQ N- tN N- N 3300000-00 Z +, 0 O000000000000ooopppppppppppp "-ACý -.5 b~~o ooo 0 r0.0 X: oooo 8~~ 888 poopppoooppoooooooooooooo \'C tj Fjooo tj s E t-j b 93 0 0 0 oo 0 b 0 3 RatSat ")gg 0 0 N 0t") 0 g oc00 +Jr tto 0 pppPppppp ppppooococoocooooocoo 0 4000 coogapo 0 0 0,. 0000 N N 0N iii - 0 0 t) O 00 0 0 cocj~~ocococooo ~ a0 ON0 ----- - NN 0000000 00000000000 cococo3coc~oococ 0 NN N NN N. N~ N- ococc N) N- N N- 00 N N N N iiiiiiiiiiiiiiiiiiiii 0 0 00000 00 00 0 -n 0n 0 N0i P p 000000000000000000 000000o3o0oom00Nococ000 PP P 000000000000 oco cocooc ~ ~ 00 - 0 NN -N 0 0 N CN 0 0 0 0 oc 0 V0 0 0 0 0 Ln -A000 n\%"0 0 0 0 0 oc 0 0 0 0 0 00000 r I Z91 sdiduins p!nbi uI Z-1 uo slumuansum n lna.p sommpuI (,) 9 1000*0 mS LO*'O uwA J3 6ZO00"O P3 9100"0 TLL qS 1710"0 A L'-j 1717OO0"O !I 9100"0 L'I 90 oIq (W)ID 9100"0 S'V 03 atd ý9000"0 uz Z6"0 t"ý AdW luoW013 E'9 I7hO L0; I Adl •[a-a/,u] suopounj uonnqilsip uoiJu33uo0uo iuotuowio uJoJj pollnolo3 soni~LA olqjqqo.id WSOl Wv LN 1UOtLUoa *S-V;alqq, £91 "SJOloIu Iuowqa:uuug OUU1M pup [elsruD 8 XIaIN•ddV tCLo m ' Z 3u 4:ý- -300 Wt91 s"C E 00) QC. \ -J \C )--00iZ A. t I ce ý,c ýc "C ýwo 0w w wi C w -I w - t'.,J 4L- -;ý z: +, Q z3& 0 bj~~~~~~~jjn j iPi u tJ t, - ý *_nt\ E3 -21 z ý 0 Q ýPNý7 -,I~~ . -- ý- . ý .- ý t') t') ,,) 7N"0 111ý- . .--I00 3J~ + 300 b~~jbaj~j jjbboo CON t"J '-#J '-A VI 00 -0.)~ J ·- ; :,iwt~j ýQ :IJ -nj5:- bbjb~ ýý5ýtli - --.] = t- -- j -- j ~, vt - i\:J~ Cý o;, ---:ý--= = ý-boýýý, C t~i ýc N W0 W (ON14J -A t, CNVIN t-J- ', N)tJ w 00 -A~ - X= - Z3 Z z 00=~3~3 3 z zz -j = tl + -0 tj C3\ CQ~\3 *: Ln 0P~3~300 z Z '3Z P Z - Z -- Z Z Z a Z z 11-i 11 1. I.j 41 L Vt C,~\ -Q .0 .44 .. 00 \.0. Zý tI' ,k 11. 991 .41,.4.4 .4, -#1~ .4. - 4 .41 4.4 00 ýc , Z 00 .4 "A ýA Q %A . -- jw \C -,j .- 00 VI ~ 11 C\ 11 -- i II--I \C rI. *00 ---1 -_ ,J ~~ 00 '-- 0 ~ ~ ON 0-JZON~ 00~00.4ON oc-,O~4 t xJ J VIZ.'-A-. - Q 00 00 C, -,1 \Cw ' t) x -Zý :, CN 00 ' ,-A.c... , 'j). " c -, -C\ . . . = , -j . Ph- t) - + KK 4- -c .4- -e S t tQ \c iboj 3 t.) ,o- =c 0=. t) a--- ON w QONww, ON ot ON '.5 00O 4 00 CS, 1-1 1-0 VJI ZOC 4- -- 1C N t" * 4 -- t + t ~-=J00 '0 J t-Ji ON 3 4n ;o ;- -9~ 3 1 --s Kc'h tn be, -4 -. -~ 00P ~' -0 j:4 j - K 4 ~ O tQ ~p :sJh o 00~t ~- zQ :4 ,~~E- + - ON 4n~ -h -~ 00%O\ Z tN) tN) J 'J N) \ CN \ _j L -_j~~~ _ -1 ~ - ~~ %o ýc ý 00 N) - ----------- =- -- Ji -1 N) N w 991 --0 w P 'J3 wN ~h) 3 I- 0 o - ----- - p's)~icr c ~n u 3~ ppu~~ ~3~~P-- - Z - X \ - ~ - X \C 1C\ = - -J ' - - C ) - - -- N -c \* O \-L ý- N) Nt t-,) -- - C;r p, cp CP 41~~h~~3~~~~P X ~ ~ 9 ~ ~~~a~jj'~w~~ N~ N E30~3c-·0~3h + u ~~P3 u ~=03 p~g~~P~b~jo~o L~;JI~ Z ,+_ X -~3h-P=~~3N~P~~---=~P ,,h·'.p;JI"Lizb'l P~3~~3~~N3N3N~3~~~3~3~P~~P~3~ j~~~j~~j~~g~~t~L~~~~~t~~~~j~~~ + ~~N~P3~-N~P3~~-o~c-N=~~~~~~3N~ ~~~~~~jj~~j~~~~~~;~~~Lww~~t~tj , Q- " .= ýc llý= t-J Cýg X o8c~~ '-A -ACIJ #nW4W= -.j t I ý, - - -tJ .4 -. - ) W CN X t-j _In o\D Lh ,yo"~ O-~1 P330 C C~ E -~~C~ -s ' * • . - 14.)1'1ý .4.~ 00 --, Z t-j N) ý= =Z00 k t-N) 1 b = CN j , Q Ic z '= j ==N)=-==== N) )j =i · Z".j:_ c C C Ncw -3 CNý Aýcý ) -N)j j N: , Zc ib ý ý L let 'n 'C t-.) Z 0i--.I\ NN.) ., P N \. j ýc1CII'c c10\ ---j w L so~ 00 E 00C N O a =ý = usemw-s 0L \C A--j l \C X t-j C w - ,L 0N tO.jCý )XCO b - -0 -0 L91 C \C ýc ýc ýlc ýc 11C \C \00 30c\ 3c 0c \C \C ii -A C -.j X \C = - N)W 4;, '- CN-J j ýc tJ 000-- jj rj w b w Bs- C zQ~Y=~3~N-w3~UP~~ + - 4c r0Z \O N +t Z0-0000 Skee~~~b~~~ -000 oc I.-I\C oc obbhbkbbek OC%, ,"0, .0--1-0 t) 3MA 3 A>>MANACM ·-- MMA ANAA -000000000000000000000000 000 Q00 0 0000 __ __ _V-N toj =Z' t-J- 0 4 ;o to 0c0 JON eme>>-mommunew xcI.,.) Eo c 3333 4. -1 4 . JI 904 c .· C .0i- A 3333 O' +, + 0( ZAMAMOMM AL + ' = = - 'XQ ý *ý-- -- h)-ON 30000\O - i zel~lh)N30-1-4 jl00 3O 0 .144P 1-#-3~Q_ ~ -, ;- -4 - A e030033300333C03030330 h>cn : P~~~: P~c~~~P~~~NO~~O~ ~n ~; j;'p ~Ef ', ~ 'N iN L j~~~ t b +r NO 01) \O- CINx P N P 'O -' x 0 o N __ ' 4 IM - c Q 00 0t + t'.) ~ ~ =ic, :_ -X \Ol Z ON~ 4c 'j" -4 -N Ne~~j -- 1 VI 0Z 4\ -~ 11C - ~ NfN ~ ~ - 1 = - ".O wý.'-Aw ý --4 00 \C = eCceecceecemse 0 E0b 90 00 L 00 L.:j, 000\--j00 CA> WWCACD--MMONM-Mach ".o.) ý-Asemsemseme -- O \C 00000-0994f~ .- 4- t w Ln-ACneece -ý 00 C Z0 .4 Z3333 3'IJ1 ~ W g c'J +l t,) \.D OC \-C C) Cý '.C -,j %A W ~ ---'A)w '-4n ~4, 00 n 3 1-- ~ *.Z N j - ý30j 3 -_j h W - + 0000 CN t- C\j 'Jc 00-t-000000-0000 b býj o c b- "-io - j11-~..A.4 - .- :. ,.w4 Z~- Z. -Q 00 "C Fj - VI =,-ý. - . .C\ "A i '-Ao .- bo a + C *.4~~~ I - - - C t 0 4-. -' -t k.--4'-Pb akebbhbM' U\ --coo w -- kbhkb Omm -a 'I-~- "wAýC-P-A -) - - kk-M'o -a - -0 'e case-<><h -A " - woo b k>k-k - ob s O' 00 w -coo u cook Ps + 3333--MM0 0330~=~usNUm 'h 13 0 300h 33'J > > C 1 /rP'J mA 3 t- 3m zzzzzzzzzzzzzzzzz-=333zzz0z - ~ L~ C 0 ~ ) ~ 4~hh~· ..J\b N)J~ -J00. t- t,)W Q~r3' c \ .4t -JN) - = 3J:_j=*oý:0-r9. bN" :.-'- ýý b3-co LA CN csocý Mt4,) U t + VI ;n'P'Pw' 13t> w cw %A L· - t ';'\ _ A .# e:J )h N)t-- Z ~ = " N) l O ='- PxN 'J . .o.-. . c\Z0 1#-oo4.c N) N) 00 , . .o. .c. --INi OZZ-ZZ3ZZ'-~ZN) .4P IJ "= .4 -r N~ i-J 00 .. 4 .0. JONw-c N)N) -j . . ~O 0D N) c00 . . \ =\ a- + 0 'L1 \ .0'-J I --,] 00 00 CN 00 \,: --j --am-com a~Y as web-~ 0 ~ -- ,) 4- 4 4 '.0-*J O 0 -- 111 00 1-0 .4- 1-.14-1 "5 N) + 1N C C\ - \t S004 3 ' - 00WCsA 000 691 rj Co re (p 00000000000000 \ \W \ 00 c, O '/ 'JON W - 00 - O 00 0 N -- w 0 N\ -- j 0 tJ 00 % W00 - ;t -n m ý= . ' h w a e PP eIý4jPP ;3\ u P e k PPo-u Pe · -j ".00 h t- '-I fQ Q 0 0 I-A -I 0N ON Q- h swee 1-11 o t0, i ,-ý.ý1 c W 00 tJ ON Q0Ck)J =- -= iJ t--e) QZW~cQ ON bc ý 0 - = C W - O n' VI · ' = O ', .4'~ I-A -- n C i -3 c Z 00- = W N - e h 00 3 -~ cQ - = Z = 'J ( 0 .c - jL + N) c - ];rb~ 3j s a3 1.0 0 -00 Z Z Z -0 = +1 \O ' J-acI'4\Z--c - -3 - a - '.A oc 0\ ZZs-a b ~~~LLLjLL~L~~3~~ Q t ,,) ý in ' jj c>E'<13' :-j zQ 00 - -ý :-a"o= tlý - - . tj'-n'4 lark ---] CN*- tllj 00 1-11 1-11 colS L" t'Q~w~ _I Z 3& c zk 1_ = = Zm . Z 'Ji- -o 0= > = = Z tQ 0 O N Zmw-p c- 0 Z = - =- = XXZ - + ouZ = = - t- - 4,ý, -1ý --j --j-1 ý-Cý 0ý j Cý X . - 4 a . - co 0 bo K o 0 -o Z . c K \ O w 0'ac 0 ,• o bi X ',A X l-A--j O.4--JZ0. ON3~030000000,- , .c',• mewo- • n ° ON-i oc o wmQ • ', M + - . 0 . o '• C: , + ZcZc.4-ZON- =o = =J \ N ý - -A .. :-, -j ,C -. I ýcO . C \c . -A - = -'i\ = C= = O 4,.. o N\S- S -4--j CO\C.4 ON 00'00ON . + - ll t3iocCb,• I,.. IA -I t.E ,) J .00 ON 00 ON -,00 "A 0. n110i -0C tl) IIJQ = t- . t-j 0\u~3\\,J\ ocn-cl t- ~ -21 ~ ~ J OLI o 0 0000 00 00 00 00 Oc x3 \c + 9p 4. Q x0 ~ o Z ~-·~-C\' ~IJ) w 00 00-j-- 00 t-Q 0- w 00000\ 00' c 0 0 w~c~ -. J"C0 00P 0 0C00 oc~ '. oc - c~ o00 -_ -A I t-j 00 '..Q JJ 0 x -4-J~ 'Jcl .-c-)JZ N pp~Y 3 'JI t3X0 i 3 -_ 00 '_4 -_j~h 00 ~ 30 "P~~~~~ h) r--· W cl ·- N W \3 N~~~~~~P 'JI r n r ---j4 36 ~'9 'J3 hc~~~p~O S 00 ,- - I· II . . 24-% 2 0 0 . . . . A " o- 3 ~Q tQ- k~0 Z 5 0 1 N 0 00 0 0 00 W 0 0 0 000WC P 00 0 -4 - 0 '~'.4 ; - 0 \j 4 -3 \ J' ') \ \ý PPPPPPPP ;-'4J Ih--os-Moo 00 a 13ceJ mewoma 00oo Masoes- -M02222222200000 M-Cos m-t-cool N Y NC osA3 OC J .) 0k)00q0c ,-A 00 -C -A t, w _.,I CJ ,04 00 \c OC j 00 \c 4 003 --.J c 111C N) -- 0~0 0 - n\=" ý= - -1 00 tJ O ýc Pý- h>h -- J 00 - 0 --J W -- j ýc CN =-j C 0 QC 00 4 00 =-ý 0 0 0t 00 -0'00k oc t-J o ~w 0 -j 004 oc oc 00 \C C +j Q~~~ +t -. 00 ý h)00 -- j 0 00 t-j 00 -A t-j \ E 3C w c ý -A N N'l c] kl2 +f ~00 ~3 I -J -J t~Q - - ,,) 0 ,-J \0C 0 + ,~ ' 0A \C 340 A-0 , :- I-A 110 m~x '-,n 0 =W-I - k00 Cý -w> zzzzzzPPC-zz '-A CN -PPP C P-0 wee-a PPP PPP Xe--mwoo OC *M P P.--PPA ace smazome-on- 0 333 j"co CNOO~~P PP-.~-tnP) PP e -m-W ammaxwe w>- wo =~ u-o a33ewc3cw 3 o emammammmoujp~n: -- wcoovo cr--a -00a-oo 00-eve + 2 -00 00 00 ec)c 0 e10 0 r- It O 0 't 00 \ - C5d 6 5 _; 6 C5 C5 C5 6 4 6d -e66666oo66666666--6n666666ooo 00 c) -"r 0 + qt nZ m ddd 0 w 0 WV \W 0c w 0 0 (= 6 C5 C5 6 r- d 0 0 Coo mc " CW 0 n I/· 0 N = d Cý w r- N i 6 6 l ; r- = Cfi 6 \6 d 6e 0 1:- N r tt. V6 4 -r C r R C4r fJ - (14 n 00 f dd C; Vý C\ r--: C\N c000 : Cl d6 0 C. o - trmm It 6666o-o-moeo666666666ono(= W) - 6W) W)6C) C14 C'4r V" w r'4T Cc N Wm l:CD V, 0n, -; C m Vi ItT M (ON ýc m 6ýc CI\O w 6IRT 6 (=6M "4 64 C 6dCN-so _-; -4, 6 r" 6m 6C,4 -\ 610 6C) r6 r- dd~ t 6 zoco zm ro C ON0 d 5 C5 d e 0C)0Cr0C00=WC CD)000-:T 606-0 6 6 6 66 6 6 + O- Lw?C14 -Xc \ CM C M 00 C14 - Cl N C'4 C\6 6 C 6 N tt N t 0= N d N Z R*r0 mr 6 ý- tO C5 C5 C5c,4 c 6NcC5 r- _: mc6m C5-O6cOroc 6t cN - c\ow) oc 6 6d oo m c6 6mO o 6c 6c\C5t C5 6,4 6 c 6 o op C5 Cl Ce' .:\ re oe'cj CoN C mm 0 m m Q C14 C-1( ;q C r dr;en q dq d ci -ýci-; l o' 00CN c rC) 00 cf,, 00 00 C\ tV rC\ c00 m 00 N"(C C14 00 -4 m 00~ 00 .., C\ \0 CN W) C\A N ~t C, c, U ýc .. _r ~ C\-,T~ 4 - - - \ ( O0 CN rC) 00 ý CN 0\ 0 CN N c -r 00 (=r rr CN Cl\ "t " -4 c 04 C) W W RT I z IT Cfc ) "It 00 W" , C\ m mV-ý r 00 r, r 14 r o v, + '03 00 Ctr N C C trt~vr, C14r, I. C14Wrý 00 C'4 rý 00 00r-r) NOON 00 -. 0 'T 00tW 00 N Cr ý Mt T T N -t U 0Uc-w t O 0e, Cl4N;mmN0 -" C14 trl tr Cl 000 tO\DC\ -\OC C14ClN Ct, 00 00 + C14C\ \c Cc, 0l m N=w z tr-00 o, \z mz " 4 ý0\-4 r--4 =) N C) tc, ~ C\ r- \ cr\ m rQ o) 1-4 CJ~C000- \ C\ C\ ,-C, (7, C\ \ , c, r- f C w, C1 004 00"00 c Vd cO tr" C\ w 0 CUN C. C4 - 0 - CN V m -"F-1 1- c 172 o oc 00C00 clc~l tro o C14 Cf% m - r, cr0 c-4 C ,c cC L-(,C N C c·c m v Cý W \C r- Cý N V-ýr- cN - N tý \o v v O\ 0000000000000000000000 w w w w w w-w tO C\~~~cc \ vC14 ,t ,, wa D A c,- C\e c00 m;; r-,,;-r (I, C14 -- , r-, tý- (r4 00- M\ 8 rC., M tt V C-4 rc r- C\ F\ 00 C,4 00 N -( =~vl V \0 00 m 't-M enC14 ('f) r-4 -"l o r- 00 F\ f- F-- F or, 0 * VN -ýN tW VO *, ] _j -_j -j t -*** --4 ELT -_ -_j -- ] 4-A 0 x x x - x -_j x x x W4I 0"0.- ~~~- - -_j X\= b- - - - - ).- - ý- m- - - - = x x -XCo tl tJ ~ cb P-" - - - - - -~~~ 2 fo S PPPPP. ...... ... tJ ý-- ý-- t'.) ý ý ý ý tJ ~~~~N~~~3~~~u0 ~ ~ L 'N :j,'h, t E3 j &.AJ OCt bcrt 0"3 P- L c~c~ L PP- - - L i jo 3- -o-3 - ZZZZPZZZZPZZZ= p - - - - - 3 - 3 - : -- 4 :=- 7c zzzz0zzzzzzzzzz-zz * ccACnp 'j, P3 -4t t OM QC~ 0 33'J -J p :3, Q -p 'p - * ý-- = .- Lo3 p -= = - ° .. - w- - - = = - ,o cooL + o t'LT wa -c to-o w I-,.%A -.1 00 to ~~~~LL~~~~~~~y~~~~~C~~~L~~~~~ ~- ---- zzzzzz33zz ~' tEJ E~ =- ý ý =\ = -~ LQ -. - =O -- tQ - k)ýQ -'- ~ - ~ ~- - ---- -~y~ i.========o-o===== - E i - i' ..)6E-ttQ E'ýZ o --*Q -) ' - Q-- ý = = t--- 0- 0- cococococcoaccocococococcococo :,- -- ý -- =ýý- - - - >000= ý7N \,C C\ -n p..-.4 t'~j ý-t- 00po t,0 t,,) CN Cý x ý C\ V, "A.)JN) tN) ,, oc tjý- C, ý-b :-eh ý-b p C 'u, x b ý ým - týoA -tCN S tQ!-,.A j '.-n E 'O -,j + - SA 00 00--jC 00 tQ C N)i00 00 41 i3 Z-, - t" 'e Jb --- Q-i- - QtQ.~ Qt )~ N) 0 - N)bbB'e .k . o--wm N) k * N) + ~~~~~LL~~~~~U~~~~y~ ~~~cl~ Id ~NP3~ ~ ~ h ~ -~ ~ ~ ·30 \3 9LI ~ 00 30 00 00 00 00 00 00 00 00 30 \O ~O 3 ~--r h> w 'J, P Ch ~ Oo ~So O ~NO~h~0~9~~~~~~~~ ~ O ~ ~t, u L \3 ~ \O E3 h~h> ~~YCi~~~~~C~~Cy~~C~~~C~~YL~ c~ , ~ =,,, ~ t· ~ ~-.'N L P~oos=~=~~o~~ooo F~, h) i ~ '3 f·-, j L ~ , i-~ 'P L- i~ ~ ~- t~ wo p, Z + =003353~30003 ~~ttt~~~tb~~~ ;~n "'.0 CN".0 C\ C\ o N.)- JI--- + o -P-~NCPY 33 a 3 - m 3 od- 3 a - 3 3 -A tJ 0z A-w-Zj kbUekb tj103-M3c-33ewootj oc - MM333b'm 'wk 00 +p t,,h) tJ -4, '-A-4- 4;- A- b A.) -a- - , z z z z tj t,.) oM-ý] z~hU z z z I on aL J L. z zz L A L oo s- A o jý z z 3 z 9LI GGGGZGGGGGZGGGGZGGGGGGGG GG y 00-b00ch 000M0 0x 0 --jm --j00 tQ,AJAm~S~Woo~~~~mm~ .nCN--p X " w --. tcý bbb - cý O w t,.) wjc ý,c tý 41, So.J C-, ex' ~n'Jh)SpS~~c~Or03~WtQ00~~ bNinw~ 3intin~ E ~,is 3, oweb -,nt-.4 F 0, ct (O 3F cokk 00 w "C f 00 --e -J .W -a tjs m ocC) e , , 'S - N ~00 N-· ) 00 'stP Ph~t~ a S00u om -ma 0- 1 N _ ,,.1 "j') --j 00 OC , ýc 3, -nt-r, j '25h ~\3 n tQ 00 +t r 00h tw EJ --j 3 w --.1 tj ON 11.0i 00 " oc --I --4rh-,j~-~W P ) 3 c 11 + L~0h o -- j c 00 5N W * .Q0 .-- ý," 00 -. 0 P ON '-n "Aj ',..Qt J X 00 C t 00 0-- 0 P.- t") t,> tJ00 o .4 %Ax ý.A\ýc--j t-j w "A tOAC x-74OCZj_j 0 3 P4. " %At. 00tjw Z r, w 4. cc~ + 4. , w~ 00 CN - oc0J J 000 LLI t-o - + t-j 0\ 00 \C X x 00 wC ,+_ ~ ~ QO Jt~ 00 J- 00 ~ 00 00 _ ~00\ 3 11-_ c00 ) %A IJý.N%C t- 110c 00 tl-W3j~ t-j =Chc .4 - Z P P 3w P· )" ýc - ýti PW~~W~~r~~~upP~ O Cr, Bt· ~3 17; Rct 'J, so ccn N w c~ wc-· \3 h) t~ 3 \3 Cn g tnh;, tr~ ý= Cý x t\ ýp 10 ýt ,+ ~ 00 rh)h)N '"' 00 w \3 h> h) h) 3 N~3W00 P w h3 w ~~P~p~~ + 00 40o w -4 er• 0 w sw -O 0 -5 a', • s os P W O- ON 00 3XdPd W - 'J, 0\ \0 --- 0 x 00 t0 di P - P3 W 8LI '0 c CQ Z ý, t-J ýC' 'SI 'J, iQ . J C\ 0 e5 = N~PO~W 00x00 \ c \,c "M \c \,c Cý --I -ý 5 'JI ="@8855888~~ ~~~UWN ~- h t-. t. J > + IC wa- t ,- 00:0j 00 at. 0 WN) w 00 t, 00 --jJI \,C) w '-A%A tQ - x CNh, - - --\ - - - 00 -_j " ".0'j +: - W N)%A W -.J \ý " x --] = N..\c 00\Xl CNo -_j= 11.01) C\ ý"..o--Q c~~~~O -_jtQ= -. --_3 w w w cl\w ) 3c11C w -Q t jI-A c0~~3~t~~~~~ 00t-J -- aý 4:ý- 4.. .Jl)-- _4n I ,.j == ,-A -,j CN - ".) t;ý- j o.,".0iC) t00 t w,3\\ + 00-3 0 wCN=-- - Z '-n Z _ -j Q00--J-j wtjtiZ ,C\ --.] -j t") 000 ,. " %A 00 0 PE3 ~3 C3 pN \O 00 0ý ,C0ý" ,\, "0 2 "') = A _j , =-, - = I-, .-. ". F j"Za-j 1.0"C o I-, ý,= == Ic -;ý 00 00 h) --j h~tQ t-i -ý, . jC Wt, -lýý C00 AW t .)0 - - w . ,.ANc ;ý O -A r~-- 3-ý4 \ýc 00 ~ 3r,30 00 0-.100~ ~~muman + 00 ht 00 a~~~~ 6LI \ CN ONZ - Cý CýCNCýO 0-- Cn t,, ".0 J'J)-;6ýA j x \ý - -- -_j -_J--j -_j - x x- x-),.o O-C-4 x '# x xý--j0 x 0x x C 0f wJ00 h)nO 00, h)C L, o c w'36= P t CNwi C %A X Xh c- ? 3 3 00 ht -I o\ ih c t-J ýc 00 0ý Qs ,.A wN o t'.) O 41 00 I 00 '-00 / .A.4J\ -. +o C\ 00In 0 0 W L A,\ '1 z +. c-a-a--coccocoocococ-co coom :,Q c jo11C i,'P .-A 00 "- - L iýCN ij Eo -j i. 1-nMAMMULLU3e + oco occo L·j Nj\b'p' =;~: w; coo- IcIc-Ao ti ·;3;· P 00 00 f>'h =oc-Ah Fr'NL·t ;\t i o;. 0\ .41 t--j t--j~~un oP I 00 0 N = C ii .A-I 00 0 c \C 3 O 00 ~0 cS 3ir II --Iww --10 000 oc 0 C0 oC .# _#,I~k I p~i oc Ch30 -,a \W%A Lt ~ o0 ".5 \3 SO tL ~ ~ - -N 5 Cc I3 oc oc 0- oc t-j 00 --1xp +. 00 00t-j tPp 00 '-A - '. 00 x0- a e- 30 CT ~c)~i ~n ~Jp~a3P\D ',n \3 P P3 3~5: 0~3~L) NhN~y~h, 3\ '•'• .-- t ~3W OO',n'~n ~~~C~~ tn 081 PPk. J .46 J; ý. -J 351 P O\ •w ' ',C -A 4;, 00 \C-0\C -- j W---w w.. w• I'= , - ýC PO E3~: - > $M i• • _rjn ~ C '-ft n OI 0 h• w C\ ý. t 4ý • Y • W 0 • 00 Nw 0 1 - 00 00 'JI·2 ·1]-jtt-j '-n W~N~-N't, w oc ' 3\IC ;ýI.t, ,) ',.Ah +z + 0 - Oc 00 ---Ioc o- c 00 cc n00 coo CN o 0 a OC 00 00 -ýj -_j 11 + r--o oc j --jP'S O o~P~ w +1 t-J-a'J3 C ON t-J - 00 --j _In00 Ec 4, oc '_4nPO cý N -p 2z t j--- jecoNm;ý A)11 Nc-mINC C)13C) --l 00 L_j n ~~Wo CJ A.) o z\ 0 \\I t 3 -_j--P ch)N L nlJ h J ~, i ~co \.LC ho A w 00 00 00 --1 00 00-0.00-_ WOCW00 - ". z, ~w . 00 '..A \Z w ' ý, 3 00 aC ~) C) cnc-· 00 0 N,c· 0 p,-~ C\I- - h 0e +D C ~-·c-. , (j.,u 0 j E - *WJ c~c-·j e 00 - h -J jj ON00 %A 00 30 00C J w "1 oc - oa -- I- s jw a + a\ 0 \ý o 00 -- l 0 Cý=w ý,-00 -A ") -. 1 " ýc o 00 C + CrC C tJJ~LC +1 "-A VI CN 00 ~ 0 0 N 0000 t)) -j 0 t~ ) o CN1141 "4. OC-j=" O ý1 0".'.ý _j_)O c 0- 0 F + 0- ** ' * N -t") CNONC CVI VN j =0 \C0Cý-1=0\ -- -4--.4-t- 00 cn cn 00 ---j -i4ý, -I00.4C-, I --= -J0000 \,OC) - N) 'J1 c 00C 00 A. 4P rC C\cC c 00N) tj" = O t- ýA -- cNcc A w N) -.1 00 (AJ I 00 00 \, \ ;.. ) 5-. A ,j00 t 00 0- w F50,F) " 00- .4~ t-.4 Ji + 0. \C 11C 00 ý 00 li' rj 5' +J ON · JI~S~C.4 0 -CINý-W -A 00\I -j = -4 O 0 N ý-- W -- 6 0 ON 1D 00 CA\C00 crONW\O 0C -41 'J ONQ\OU~i.4 -4'J ~O-.4\ ON -4 OON 8 +a O n· 00 00 11.i JI 1.1 l*j 11*jIIJ 1.0 4ý- 46.0. 'I C -- 0 C= j tQ --j '4j-.A . t,, - ,A z) .4 -- oc u-Ij .4. -j tC i 00N ýa t 4h .ý6 .4 .. ý6ýACý-1 0\Z=- Ot3 --I C= VI= 00 I0 0 0 - Z--] \.S' o '00 C -400" W 'o 4;---.]'-AA C \3 l-- t t3 N_ J 'JJ~ ,-A' - Ic '-A .,) .4 00 W\I JC) #.)t oc-llW c~N -j"Ji110 3 ~~~~~~ 00n4 ON ~Z C tj ti oc oc \4-1~h ON .,j t).4 ,Iz -A -j -j\1 L rF'-4n 30 '0.4 -Q .'.) _ ,.k j Q ~J 4 00 )- I.=.4 00 1-1 0 00 00 x ~ O 00 00OO "00 .00QAn- JI0 '- , OO 'A N'AO -_ -_A Q 3 Ln01 3 3 ;f-h 0 --_j 4 0 0 "/ 0 00 -- 1 x 00 00 xN x\00c ýc x 00 11C 3 -4C t: rr cD r 0) -- Ct 7 -j IJJ 1.1-) -OC-00 t- t,- - \C 0 0.. t N) -A ;ý-.--00 z."4--- - N oc 0--,1 C t Z00.j4Z ' N) 0.4.~ 0 cE 00 c 00 0 E81 "uozuJoq x1mtuoum 3•}'d LprnL suozloq uoldnio 3.UM131OA uIoj pou!mqo soldu•s al ioj solqm nGu a XIuINcddV ul rr N N N N \C N N N N C' N N N Nh N M 53 LEI~ t,81 U~tANNNNNN NV~~ N N N3 VI V! CJ a"j z N~(N~NC Ul Ul -r N N h '\ hWW t~i ca 4 CI~ ~5~Let ca 1 o rq Pa B 0fl 4 ir3 o, b3 zYe (D · +S 0·--030000 In~i~~~Npp OPOoo 91 c-·100\ 33333633 +. ~8989BBkh0~3 ý 4ýk-ak Cý t) w= N st-, E3 0 coaccoco C. 33333033 'j, \3 t~ W ~~w 3G X ~ c-· N 30 bc~ j\ 00\b z*i 09,- - z 30 00000=00 8 88zaszzzz V zz 0z zoo ~Q ~8, 0 .ýj tb 0000jc "04 " *r! cE 8: , a\Z '300 N c h) O C 01 4- ,0 ,+. CnPP~3\h>~P\S N C co -~t 14-. .4;- bo ~Z "Ji O S) -0 000 Wc- h,'j, 0333330c~ AN 01 2-i LnCn InE~t~P j, ;n~Ptnj\~f·slnb h,'j, w 9h 01 ct h, bZnb'pb 00 03030 3 j ~psss8~ LnLnWO\ + h) 0 0 k - 3 tJ - ) 00000000 00 3t330 0000000 00000000 o03000 -c 00 3 0000000)~~Ji -4- J 'J 'J - I + Cl-- -ClN-~ Cl ~-r- l CCl-C Cl l C d655556d666b6~CCCCCCCCC c ~ dddd ON C C- -0-0- ONC C 0-t- -CC 1 CCeCCCCCC c888ccccc 00000dddd C Cd C cl C-cl CC~CCC d:dCc -C -c*6C CCCCCCCCCC OCCCCCC:CC 66666ddC Clc c: NNCl Cli cli c'4 ci C-ý clý r ClC ON0~0~O d ' rC 0000' 00l N czzc Cl C Cl C~l -, -z· CC CC CCd CCC c C N -:b'CO c Cl C .Cl r -V-C C C Cl C "N C C rZx C - tTC 00~ 2. · 0000000-0- CCCCCCCCCCCCOCCC C C0l00CONONCC-CON0t00000C0l---CC0N-0 Co-'cV', C zr C C C C ýC C rC~ VC; tr C C\ Cc Ce Vý C V,\C t C r-C CNC\tU3 C C; \C C; \C C\ \C tC0 tr -C3 C\l C\1 V1 ocCCCC6ZcccccczcCcCccZcc6ZZZCz6 \C C 't, C C- V-, C\ C +, - Ct m t C, C C 666666-666CCCC6ddd6CdIt V - C- r or- C r- m cCe C'c OC M re -C Cl - r, C Cl r- C 0 O (Cl r- c: - r-. r\CC' t C'C + C c d C-ON0000000 0r C - XO -C al-------------Cls CCCCCCCCCCCCCCCCCCCCC CCCCOCCCCCCCCCCZZZCC t-Cl--Cl - 0C C m--Cl 00CCl0000COr~-'dNW ClCl---ClCl-.ddddd \C d CCCCCCCCCCCCCCCCCCCCCC O C O -cC0 , I ÷ C CCCCCCCCCC C00 O 0Cl0 00 Cl O 0 ir, - CCCCC6CC CcC - v, CO ~- 6c'6c6c ON N O) - C gygooecooc c C\ C CP/T Cl\ 000000Oanya CNN -IC C N N ONeC C - CICCcN C4 N -4 ý 144 C- CC4 1CCr CC u c, p ON _ L0 __ ZI7 'Z 89"I 06- L'I 61t7 17 __ 98 Lt,89 2222 qs I-1 33 mS C'* 0'I 8X'Y tI 9E*Z 88010 88'0 9L'0 60'0 9"0 0"0 06L 0£ tL 96I 1IZI 690E VO V 8'0 ~ 91r-gS91 ZAIS IAIS LV-9tV91 I6ZI OL'O L'' zO1 L'O E' 0'1 91 C'0 L6 L*' 8Z91 ZA6Z 0-618I 66 Lt 8'7 Pg'0 67 9E 9'1 L6"0 6E*0 FO 0N 9991 P3 9'0 L'O )'0 aS 11 6L01 8ELI 061 6909 69gi 98L oNI .r1 £*' 0V6 s 1 I 1 g7 1V-0Z81 •8-V88I L8-9881 IA6Z ZAZZ IAZZ uw['IV] aleI 9t7St)91 HI alm!xoJddV U uZ aldmws ZAIS fAZZ 17I8-E881 1 ZOI 00 Or' t'AZZ 8-881 I 9E8E SL't' IZ-OZ81 9; I L6991 8E' ZA6Z 0O-6181 E I ý9Nt IAIS LP-991 o'0 99'0 A tl 9z8 Z61 i'o 6'0 8 6"0 91'0 69"1 81 ~I__ 99 7AZZ £8-M88 it, 8V .9 0.'0 Z'g t9'0 £AZZ V8-£88I Z, EE IVot 9"£ LL 8"1 L17' O'0 16'0 07 60W I UNJ a3 •suozuoq uO!1dnm5 8LE•P7I 6"• £6Et7 x 13 Zl!UaltfOA uLoUJ PlOu.qo socdULts 81 IA6Z L8-9881 IAZZ ['*I'v] awa• ••um!xoiddv aIdmtueS *Z-3 alq• N IV tH I ZAZZ 98-t881 9E Oql JOJ SJl13RA lU3fi lIa!JImq -?snlJ D _ _ 8I'00 O£*O'O gZZOOO 6Z00O0 Z10'0 t791'0 00*0 6S0'0 1700'0 Z910'0 £gO0"O OL10' -C800'0 Z00"O IZO0"O ZZO0 L£I00' 8010"0 £E10'O L01'0O OI10*0 8000"0 L1100 EZO000 9ZO1'O ý910'O 1 00o O10*IO'O ZZOOO ZZO0"0 00Z*O Z00O* E00* 900'O L100"0 1100"0 L10O'O I00*O 6900"0 0t00"0 9£000 ZO0"O 900"0 Z00"0 800"0 £00"0 1170"0 V00"0 800"0 £00"0 690'0 Z"O*O 00" LOO*O EO£"06E0'0 981 ~ _ o00"0£00"0 900"0 I00"0 LO'O _~ £0"0 91"0 9Z'O LU" 1'O 01'0 01'0 80"0 60'0 80'0 800 LO'O LO'O 1'O 60'0 800"0 LOOO 900"0 700"0 8V'0 £00 80"0 gV'O LO'O 80,0 900"0 LOOO 900'0 1700"0 900'O 900"0 900"0 100,0 LOO'O 170ff0 ~__ __ _ EZ£00O L900"0 '100*0 9ZOO'O S1O0'O SZO0"O £Z000 t£OO0 91OO0 V100"0 100"0 6000"0 V100"0 100'O £EZOOO Z0 £OO*O 9100 91O0O0 9zO00 t:0'O 9100"0 L£O'O 9000"0 VZOOO _ _I t,3106 Z£I £23106 £££1EEI ££;I Z3106 £ IM 13106 9££M 83168 L££M L3168 93'68 S3168 t,3168 £3"168 Z3168 13168 8££I 8££1 6£M1 Of£I IV£I If£I alwi] -1+ 0ci -i+ 1i -/+ S.V 9£'0 9£'0 9EVO ý "E VIIfO' 1£00"0 PIV2 EZI0*0 99EOO1 SL'[ 8010"0 LZOO'O 8010*0 9910,0 L£'O 8E'0 ZS'1 UZ -1+ 0800'0 00Z'O I"V Z600"O ZS00*0 117 EL'I ýVI IZ'O 61"0 Z* 8'ý g8I' -1+ 03 qs -"+ P3 0V·t 20*O ZIV0 LI'0 70'0 80'0 t o'o 60"0 80,0 LO'O ZO'O 9*0 z0Oo 07 i0o*O 60'0 20"0 9Z'O 0"6 iO*O 910 Z'S9 o'o 0z.0 ZO'o O-C'O 2 hO z FO 00O 8Z0 -+ I -"+ -1+ ag LO'O LO*0 LO'O LO00 LO'O LO'O -i+ tT'O 0*'O ££00 ZO'00 ý£'O -1+ 3 -1 u a1j(l aIdumuS sY -££CI t3106 VZO'O 1100 VZOO 0[OO0 EHII C33106 zzToO 90ff0 t9'0 L£'O L9"0 68"0 LZO'O 80"0 EZO' £01 LO0 LO'O £EI E Z3106 H££I 13106 9££1 83'168 L££1 L3'168 8££I 93'168 8££1 •'168 6££1 t,'3168 OtPI £'168 IVI Z23'168 L•0'0 8000 UNI A 1tVE aI•( 13168 aIdtuS ZE'I t3106 O Z'8 '*I g LE FE 'I EZ 9E MI £3106 0*E 6"9 O'L O'E• 9'1 'E 11 91 100'0 1000 LI t£ ZY'106 E 0*E 1"8 6"7 9 L 9'I I00,0 100"0 9 9 M££T13106 •O L'L Z'L t 8 C1 67 ' L'I 83'168 £'6 t 9£ 9££1 Z'E 9' '1I 9"I 6 E 9'1 £ Z L££ L168 19 9', ££ ['61 6 '£ 9 100"0 1000 "I 9 Lt 8££I 93168 17"L £'9 Z Z'E 9'6 0£ 6Z 8£I 53'168 '6 1"6 V 9'E 9"0 £1I I L'81 V 8£ 6££1 t')168 VE £'11 L'6 81 8"6 z1 100"0 100"0 oW i'6 8'8 Z'Z1 V ££ OV£I £M168 V700 £0g'O V', 9" L 6 V'£ ZY168 V VZ I£1 9 0£'E 01I 9 IV£I 13168 81 V7 9£ L17 £'Z1 6 £E Z'£ ['OI~l (I 'N lrU ;ilduxeS -r+ -1+ -+ IV N -1+ 13 Ss -/+ -/+ •[a-•0/Fu] uozuoq XI&puouL agNt L moij p ur.qo sodures tql roj suoplua3uo3 MuIflgJ ":-3 alqeL LSI \000Q000 S~~~ff\D P hhhci rOO~Ch ti -- j x00 N) .kj")ja.,N) --4 00 -Pý-Ic '-" - t-- - ~ ~ t'Q .) - 00 IC C O0000 00 c '../1 0 - .•al I oc -0 o =Q 00 %A " - li . 00 I,n 0 '-A "-)• - t•. t'Q) , - o- -- -, . -AZ tQn 'a > c 4 ' '> - 0 881 •sisXrIuVo j:o3o IqiuauodUwoD ld~.u.d 3OLjo X•ooapl 3XsIe[ U XIUNIddV 681 . luoiraoo ug jo uotlavloouoo 0l4 'aojo0jtlL "uolooas :sr uzlutM oq ugo folduizs uin snolAoid 0q4 u. pomjop uopounj aojsuug• 0o1 01 1U•Aitnbo Ajllntuosso st q.qm 'Jopmj lqgiom r pu qlfuoans oainos Jo stuol u. pozujop oq uo oaominos e tuoij uonnq9nuoo Ot.L, sooainos Ur tuojj suoi nqupuoo Jo uoluquloo mouu1 u si [3] xLurtu oq041 uql!AM uauo o4o ",•oAlroodsoi '(suotcAaosqo) soldums jo aoqtunu pug '(sojqrýMA) Suouaoo jo iaoqtunu 'olus fuqiduts o04 it uornuoauooo 1uomuoIo poansroui oql olouop ui pus 'u '. saloI1o oatp 0oqm [wU = [U] Iro.J ZIo uWro [r ZZOj UWZO s d l,....................• :smolloJ se xulu u Ui mvp uowJuaoouoo ii = v S polnsruo oq0l luosoidoai ol q!ssod si 1I *(Z661 "T 1o RIp.nI '.0861 'JOmloH oanliWoul p4oqs4qmso roHM o41 u1 punoJ put .r4smloutIEW '9L61 'utuLMtH '•'o) oq uro sdois [~t•ua i poaMaoAw oq tIM onbluqom 'v;4oq oqiu l Jo luotaudoIoAop qfnojoql siql .oj qoat0ddu oisrq 0oq q4noqlv "tuoi l jo sluoawonsvotu uo0WlaJuouoo OWeIAe Inuug o0l war suotlvAjosqo 0oq put 'sIauouolo 0q4 OJL sgjq0t..rA 0q4 '3IJOM stI51 uI "solqvtrlA oq4 jo suolwAjosqo jo Aotunooe pug aoquwnu oql uodn luopuodop •losolo st ssooons sl put 'los llep u jo soThqeUeA Oq4l FUOug suortlOaioo olqtssod Xj-nuopi ol posn si VqDd "soproop moj Ol pu SloS vlep 0AJI JO UOtU103UApe sulI 0oq1 u.l!M aomod Ieuorwndtuo jo sjJ JO jo sonbiuqom luotuoans1uf l tuinmmqo .ioj 1O1Jd "oouo0os jo s. osvoaouI sitt jo l o0 0np t qufmA oqlA Sv '(t,66I "I' 10 ZgUlO !Z66I "It 10 ox1v '.0661 '-t 10 01OLurlO soqougthq a0oqo s IIao UITis.oaJOL. "'o) s03uIs0 I1Unmuuo0., u0 t. posn ixnoq 'L861 'uqvN put ItlquOumo aioj onbtuq:om air sonbtuqom n•ituPs pug sit4 ApoL "slos -erp ognI jo uoilaJdaimui oql I1:1sWMS OJL.AnIlntu SI•.sR•U sA y loruV luouoduoJ Idon.ud r si (VHqd) sisX S.•-!luV 1o8 juauoloj I .1PUd a jO 13AJ51 !seff 061 jo OutraUA tuntuax'ur jwtqi os xuwtmu elvp v uoij sluouoduoo pd.:tu, d oql inixg oi st VDd oqi jo osodind o.isq oq. si 'slopotu ioldona "mlp oql jo (VJd) sts&CeuV luouodtuoD rdim..d OlJ.UApIlnw e lsotqu ol uotuuoo s! q.rqmw 'dals Isnaj o •posn am slul..suol onbiun ou svq pur p•AuoJqi sq uoptnbo stp ur x•wtu uMourr suonlnlvj wuouiuoriuo L o l:irsXqd 1ruoit.pp- otuos ssolun uounlos mj•ms iou si [7] putm [] sa•wuw jo uon•umuuolop oql '[j] Xuo zi za.uS "aroqpaioui. zq U Pt '(1661 '"IUOH) ddv oanos uo iordun opiq stq XIzauog uumo joJan suqp 'JOAOMOH (r') 'irqi os 'sio ~Ix] + [] [S] = [ID] inuotwA.Isqo tuopui jo uoilnqjiiuo jv ptrq iq.u oql oi poppr zq piuoqs [A] xmntwu mo. 'swuwtu uoiunqjajuo qi0o0lonp Z-' "b oql jo op.s u. 'MilFex 'a in lirql giON "AJnaOF odsoa ;ai3nos pue uonisoduwoo aanos paollr am [7] put [SI 'OaIH [1] [s] = [I] (Z') .10 : s! • a: u; ur* l W11 . ?y 1 v.I F us Zus [Usl• [[] j,1= Z1 aul I Zdalt I7 s 11 l tlu s .,.................... UZs a jZos Zs i sn . . . . . . . . . . . :sr [j •a•tu uot0lualuouoo o1 ssoadxo utm ouo 'aojojoaoqL oiis uoissino oamnos 3i3o.ds r wluuomol ur jo uo.lTa1uuo0oo 0q SM•OqS 1i zoous 'uorloTwJ~iuopi ol I AIlnumnb j•otskqd Injosn - St S aoiotuw-id ot, o3Jmos f 'SOoJflOS "[OApoodso.I o qA014m • amanos 0q1 m1ouop I pu' 's'1 snau0o 0l ms fos 'lqgaiM jo jaqtunu put 'tqiuao. fy1 .YsS +........ +f F. !s + Z1.Z s +f 1t.is = o (Q"I) (9l) 161 114D "11] = [1-H :snMolJoJ ST XLIUJU a0U'UhAO3 Oqll JO uotflljsuoO3 OR1 ol •AM •nluIts r u.pman1Jsuoa st [ý' XLIWtu uo•Ia ooo i 'dals suji ioajj "l1un auo aoq I•M os vip aotp Jo aotUmmA Iloi aol4 ol luatuala qt0a Jo uorlnq!.lUOO oq4 'aOojoajoaU UPa Atq SUOI.1.IUZOUOO L jMuatuIao "auo Jo UOITiAOp pMpURIs Vput onaz jo QnVlj paz!pJUpUMS oar0JAu 0aql.teA IMtl JO uolT.Aap prputm1s ail Aq pnsoi ow ~uTpIAip put ,oa alqj'.UVA aql ti lmod mltp qmea uio.ij aoqgýMA aql jo unatu aol •uri.rmqns Aq (uo.vzpplvpuvs sp uMout osIr) uaul agqt inoqt to 'MOo alqO.UnA irq1 I Slulod lvp a0q [I jo samnbs a4 jo tuns 4oqJo looi oainbs aoql q M•o v ut iUod VJVp qORa BuIp!ATp zjqt.MeA ui.uo iq oaql inoqv auop aq Jaoqtoa uo uoyvzvlmUaou aoqL •xuPtrU1 MUp MEJ aql Ut sajqnuPA jo UoiI•ZlUUou aql SaAIOAU. daois isnj aoql 'suoTpuo03 asOti1 AJsnS IOU Op VIVp Oq4 J1 'AioA!nTUJaMV "xUWtu a0q jo osodsuma a0z4 smouap ,£a0JM j[D] [L] = [,] (g') :sAolIoJ s'pa013MlSUO3 s9t [J XulnuV oqlJI a3URUVAO03 uTaq1 ioj suolIpuoQ 'AI:OZaOIp pasn 0 oqi js1srs mlp MrJ "Ij AroTadsaj 'soo.Ultu (aldturs) utunloo pu (luatuaolo) moi a0J [1 Pur [Al] aaoq1M [10 [1] = "[j] (P'1) :st UaoIum,aq osi, utro N[j] X.WUI mP W paz[mttuou a0l 'Zu' "b'0O Aolum Aq "sluotuainsr•u oq04 jo sopflugtu a0q o0ItUOIliodoad are svoauo oinosq uaoqA JO Slaun lUa.Ijjjp q•.aM sOjqIt.MA apnlou " mlp oql uaoqM daois kmss00u r sT uotzltmuou 'aaojOJaLO '(0861 'AJOH puP .,smoupl@ l) siuiod virp a0ji jo lOArJ U!Spq I•nqimsns :SI3VJ OMI LuoJJ jo amJflu JUr.mAU. 0leos a0q4 a04 put '(Z661 '"Iu 10 vtPJmN) anbniqoa a0l4 sauooo iTssaoaou stiUJ MtAJ oql Jaoq4aqAM "A1XA aotuOS Ut PaZI1•Uuou oq plnoqs Jo st sr pasn aq uro los ri-p oqt,optnu aq plnoqs uo.sTOap t 's.sAltuu a0q ~SutJms aOiojaq '.OAoMOH '(166T 'AluaH) sjOnj1 Jo iaqwnu tunununu r Aq paq.idxa oq ue:3 mvp PlIvamalu. Z6I '(6861 'u'o"I UBA pue qnloD) uaoaail uoyisodwuooap viijAads ojl ol Sutpaoo0V :sMolioj sr pouildxo oq uro xwnltu uolnloaoo oqt jo soals uoptsoduoo:op o *suoitluuoJJsuiJ1 ouo-ol-ouo opun uum L .UEMAUT. OaR sIJOILUTISO pooqIo0)I e.xLtu 041 put 'uonounJ ouo-ol-ouo r Jo soumw Xq [] o0l po0ilai a0J[o ] jo slne02AU•0. pur sluouodtuoo Ijudiupd olp wtpl ioj oqT Aq pooluamng oslu st uonnlos slp jo ssouonbrun OtL "uoloos sql u! zxow Jz1IlnJ pguLU~x0 Nq l 3lom J stu.tL '(661 "771 10PJNlW) 7 (suoilwAosqo UZATS otq Jo poo•Iaq I 04l oziu.rxtLu 4r0AT sa0outtId otp Jo sTnlA) saJotu•w d uo!ilIndod gmuipuodsoaoo ot1 Jjo sWoIupso poo4q)I1 umuiru 041 .i s0nflAU02t puE (s.o10oAUo02o) sluouodtuoo Idiouud otdums oZq nip Imuou oj 'ploutisi oan [e] jo sonIVAUO.To Oql uoaM lwtql ioJ Zip 0 o np sti st4tL "sJo33jjo aoquinu tunusp v Aq poutildxo oq IFi•M A los L'Ci "bg sp [=] •u Unlp . pMtnl;,JJUI Jo 03UBTrJA umufxUtu oql 01 ql soJnsuo *(SoM sp.i u sumoIauo Jo aoqtunu 0q1 'u) xulJtum "URi ninp MAthi ql Jo uotISUOLUIp .130ltus Oql o0 1U01VAimbra tunxuw rxq l4AM '[''] JO jri oq0l s. b jo onlA OtLL ([Ufx ...... Jx) AI0Allodsai fx'fx] cx f Iat04 put' 'SJO13OAUt10 Oq4'(by< (u'Cz) ioqtunu '((oods ioi:)j 041 Jo Aqrtuoisuoarp) ..... < y< ty) b SXY....... b LL S0on[AU310 •01 O1OUOP b pu "X'yX• oaqM +=XZXZY+ LL~I- I-Ixx-L~[-] :s- SJ1030Au0tO pur sonlVAU03T0 JO sUtJO Ul Ou•u•M Oq u73 11 IM41 os SJ0IOOAU0•10 JO loS r oiUt [-•] osodtuooop o0 si dols lxzu 04L "(jaOM 141 ui soldurus) suoInmoasqo Jo aqtunu zql s0louzp uI OaOH (t,861 '"I' 10 IAMOaUH) xl.ntun VVp POZIuJOU 04ql Jo S10133A MOJ ulrp 0ql uoaaomq S0Jg1U Oql JO SOUIsoO Oql o1JU0VAfmbo xL1lntu uoI•lawIuoO (I-w) Aq papIAip oq oqlaJinj PIl.OS XWur 3u X.JiLU vvep Mri PozipipPUm1S 0q4 urtJ 041 jO SuLuOIO 0ql lnsoJ 041 jo Uou•aj LUOJ pouitlqo si xUntu pOztUIJOu 041 JO SJO130A MOJt RPp ol41 uoAtmoq sOIUt xulvu XUl. uoyllnW0o)o O04 Jo sluOLUOLO OtL ".1 xptu ol 4 qoIo '(3ioM stp Ti sr) uoitlOJaoo oq0 JI "x1JPu. v•Wp O4l jo sOUIsoO Oq401 1U0gIAitnbo am 0T4 Jo uoltIzlzuuou Oql sOlouOp AN aOoqaM U (t t'a) (o01) E61 u! slinsOj 6"i "bg2uisn pur L[j] Aq II'IcIba uAld•innl I [r3] [a]= [•a] Q I *C[)~ ,[a1 [a] = 1 ,1[ 1 3j.b] -J[] [a] = [y] veql punoj si M!'6"U "bgtuisn puE 8"1 "bg oiui 9"1 b' ~trnninisqns ,q (6) j[l] = -[q] a 'qapao oIqm •siInsoJ 'idaodoad •XiIiuo2oqlao oql uoaj "uoipos silp ui int I pzumldxz oq UO aqipooi. poonld arm sonIVnuo020 OSOU T muiseo.oa p v in ['yj xmntruo muotu odxo oqi jo ainwlu jo uoTIOuT1Sip oIs!iWs oqjt q pooniuamnn isouq[ st sonlEnu0aU1 OL "aoJq pOaNOIAW oq iou I•m put oanniviu oqi ul poqsqtlso IloaM OaJn WIq1 spoq•iu OAIVaOIi snOIrnA Aq poumlqo oq uro [j xmntiu Amnitu oq IBM SJO133AU0o21 OSOq 'pOziUtuou s91 funr.su LL "qmju.iouoquo pur 'IjUuo0oqJo sT [j] OOUIS "!J jo SJO1OOAU00 OttaI o0 I-nbo 2rtoq [[j xrntuw Autnputns zqp Jo SMoJ oqi tl!m 'uonnflos OAtq IIM 8"Cj b~ '[JuupiM aiopio umo.i v ui poorld pu t ounsip ar sonFAU02!O onbru nL oql nU JI -XTLIuW ltuogoqtPo ur ponuo st [I] pun 'uonttuUojsU.n iEuo~o2 o uo pollo si uo.raodo siujL "uorwado uotnzqruomrp sup Aq p-u3rqo ar [] jo sonluAAu o 4oq'sonlInAUZI; si oq oslr '.~ [JO SoLUouomlo zq wuS "oures oqj4av sonljnAU02! lay;tI put 'sooatultl JU!WI PSIpolloav [ (gcz) pu[y[-L ;)LL "x. lnut -euogrp r s [y]J aloqm [•]==.[J][-][a] vqi os [j] A1iClfuris v Aq poziruogrip oq tum 'xuWltu oflnotuuvs xIlutu n sqo qmljtM '[] x.Jlt uo lUlTJJo3 qtp SJoi13oj JO Joqunnu uwnwlunx oqiSOA1r t,61 u Jco13turJd 1.iwu. Jddn uouwwtuns oql w3qi o1oN ,oAil3aodsoai 'sluiod impp xulvw uuwnlo pu- x.nuLwu AxOJ o [=f (9['(I) •fgf n fi'm• (9vO:) = olvoltnpi. puL A osoqm IV0 U sL Uoi.UAm oq u-e xlun- ewmp pozl•uuou oql Uiiuiod vwp L ',XpiiTu!S [=f (sj'(1) A-lfnfs X = 2 -' U su U3IiiAi oq U~3 1ii uoql '3ads JO13)J op '[3] xuwai itp M cu oql tUTJ013OA P Su pojaopsuo3 Si 'E'- "b3 •q pou'jop iuiod irpp ~jj "ooads lOiouj srpl urds •qi '"'12 ...... ' 'Izf 'tI .fln Irtwouoq1jo u ougjp ol olqlssod oste si 'SJO13A I 31l ut aoufisop uunlo, pu 000uiSOp MOJ I-qL 14ITMA pM)O.oss "xuir. l~ Ip poziTtuou wnlup 31q1 Jo fnlLA 31l4 S;A1i J0130A U oliuo iuiod - Jo uo101ao[od oUtLL "aodsjloinj s.UP u. muiod v Xq plu3saiJda. aq u-a (aldaus) 'u•isap uufnloo qtia pozi•tLu ou atp jo (iua;u)w3) pup 'do•idA e sv paaoplsuoo aq umo xJtIui m1p u2uisap AOJ 4tea3 11vql os poujzp Si oands ioiJ3J V :(0861 ',JaOmH puu .qsmouT.•u) smojIoj sp luiodAOMiA Jo130A PWoIj pOu1jedxa aJ1oq aq uro uoilisoduwoaop xultu Jo 1doauoo atJ "xluw urip pazrtuou oql asodtuoaop ol 'XlOla1oodsaJ 'sa3lIuJ uufnloo pur A•O• pql oj suoiss;adxo pyT[A O.M tI-'1 put ýI'-j "sbg ',[]= [A• (•I1) -'L[] [A] = [3] = [0] u) (vi pu. SaATZ I"1 puT iTC" "sbH jo uosurduoo3 V (z-I(1) [31] •1[i] = N[3] oq (61 () 96I iuql punoj s! 11 '6-'CI pup La "-usbg uuPdwo,- uux 1=1 . ux [uUX ............ + [ux] !UO Jux [ux UIx [U x JJw [Jr] 7 ZJx w= /=# .. +z~3[U'X ±ucx [ x] 1 kg,;=[/] u' [[x [Jx :sAxolloj S s Igj bH u! siuiod mup oql tuoij pour1oj uuo xLPwu uoilplauoo oql '[ UX ........ 'JX'X = [d] = [AA] su E 'I "b3 ~U!UhUJo- (s8u) U U su u.ajuim oq uro uoiwnbo sul 'sJoOOA iiun oqq jo Xnladoid ,iili•uuouoqlio oqi ~umsn pWe LI' "bg olu 91'-a "bg 5u!ulnsqns •q 'axojaoiLu _" . IV2NO :K= qv (LVU) 7l sr umaum oq upo xulrtu supJo luotulo up '9"g-C ubxauL Xw uon1laiuo3 oqi Jo soldturs Jo Oqwunu j.ai uopiulfop oqi uno.iq "amouird y oq Xq XlsnoiAoid poirugisop su• OTl p~U sitp •q poamOle (sooanos) '(xuvwu uoiaO.UO•3Qq jo jpuml . Oqi Xq pao!ur.) onbriaqo 961 OATSSo33lns qS o Os 'joJJno inpsaoJ 1pro uo poqiow saonbs isroI oRi Xq poialnou X{OA.pnaosuoo oq um3 SJOlOAU091. 'daois tlo (zz'(I) (Y) :NO ui JOJ.on oqi jo uoluollddu sui azru..luu oJi No [=. - No = (j) ,~J.18m a SU Uollum oq uto iaaoulwtd ioj.ao I[npsaoi stp '(sluouodwoo Ipediouud) sjOli•J y lsjxj ;tl Xýq auop si uouonpoaidai oqi jI "lddlIunoo poanpoidoa Jitsn sii pur lmrod erip imuoauodxo u-r uooAoq auoao;jjlp oR1 su poutjop oq uto a oalou•mud JoJao Ipnpisao V :(0861 '"aWmOH puP .qsAmo.I1N) SAHOflHoJ sr oaornj pouTwuxo oq mou uP L'a "bgJo uJ.uoj Otl cuI xu~nu uoIlwJauO0 Oql osoduoaop ci uosFai otL 01 gLupJooo-3 JlouJ L uT SOJOOS "3OuoaoJai pus oJiz jo LnrmA J3S s LL "auo Jo UO1.WIap pgwpurs Joioqaql aJiudwuo oi sdpi; )qlas aOSh.IOAUUu• OaI4 [.AX soaocs o01IOJ ol 'aojOaqj '(EL61 'i6"OJuoD) sao.Is Oql Jo UOT.1TAOp p.wpumls oqi •q OauOaJJIp Oql .UTP.AIppU Pao3S A\mi pIo a Ifoj onlJeA OagJoAr I.o! Sui.nilqns •q sawoos lmoij oziparpuris Oi o.•plrld jaouo2 oslp si 1I "odurms otl Oi jolo•j lIM1 Jo UOcunquliuOc otl 'spiom aoqlo uI OTl lI1l IoN OAzLIIOip1J i uosaidoa jlioJ gToads r ioj odtums v jo aocs JOlouj ssaldures art uo sluuomuoioddr oainos qi1 lUosoidad -7 JO13OA McOJ Mp Otl uo fJO1OOAU;%.• OT4! Jo otmiiodui, saicos Joinj oq1 jo ains-aui USI O13aOAUOBO qoo uL tl'x inmonmrd ooairti (tI'(I "b!) uoulnbo oamrs oqt tuoij uoos oq uga !. ',qxelIu.S "•" jo ooumiaoduIn OA1WJlOJ Otpl 0 JolW3Ipu. ur si Ai flJnlAUaia 0qp '(vi'( "bg -os) fX JO13aAUo2!i Oqocuo . xu3isop uwunlo oqi jo uoilool3od Oql sI (JaloOs POllO osle) ?f. olmouwawd Oql Oui.S -TX Ji3o~3Au0.0 cqtl ([') lATM paPQ.OOssV OnlJAU03.0 0t4 SI 0JRA O.. [ufx...... (fx'fx'Ufc fX x]= put [=t. (0oQ') = 'y wL L61 u.•lqo uro ouo g'Q "b3 ui uoiluoJp oqi Su!sn put 'LZ'I puu 'tZ'Cj "sb3 wuotI (LZ'Cl) Iro r[ ]IX ui salnsaJIx Aq 9Z'T "bI jo uoiwmoidintuaid (9•'G) uDX + ........ + ZDZX + IX = [ND] St SJO3OhA ZSgtl Jo sturum u xurntu 1?Wp pozit.Iuuou oqjl onuM ol olqlssod osjr s. I ".OiooAuo!ai lsJJ Dql TIX pu '[9] XLllU Oql JO o0130a Mo.1 IS.•.J (sz:a) a sT I'....... [STt 'crg z] ,]= Tx 19 oioqm = rDE[NA] s'e uuoj xutLu ui uoluM oq uro jZ'(I bg 'u oi i uoij SOUVA i .•muemird oqi pu- "fx = f!m wqi Ouupqui-onia pupe OZ0'u "bgSuisl [=. [=1. rQ X!M= 3 (t~z) JINO1 U, U, UT salnsaoi CZ' 0- (Ev~a) "bg ut ZZ-' "bgRuTs1 l m :uo!lnbo BuiMolloJ oqi ADqo oi poioodxo air siolonjoo amoi t o iroodsai 4lM (i)a snaloturad onao oqil Jo soajnbs ;n4 jo SOAnUauOp oql 'IaOOOAUOStO lrditouud isiU oqil urlqo o0 ',j uipioooV e ioj sunoom "lsm1up 4oql U uoteuA oq4l Jo LunmtuJ lJo30AUW2t0 (a'CI) 86I 7x 6X = 6x ,3], [] siag ouo 'wE'I pup OE'(j "sbg tuoiI f=[J']Lx (*o) u, sllnsai fX Aq I 'C "b"jo uourouIdintuiJd + 'X ........ u Uy~~ + EIOX + z _ZX (wu) (~a),_;9 'D[x [ANI]= '•] st uOllum zq uro qoiinA 'siona Imnplsai lsnj oqj jo posodwuoo xi~rtu lonao uLxu up s!t[3 ] aioqM (oWu1) Zx = ZD I[g] ui sllnsaJ sTUL "oJaz o0 SOA.lA.UOp Zql Jo inns owl Surlls puPlPulsuoo ([Y'a Su.dooll ollwq Jo uo nTuZIu.u 'oagojoaoLIJ oi poqloui sowunbs isvol oqj g.usIddr Aq pAoarq: si Omuox (y)':a (6Z'a) _(r)a = zg (Z) 1.1v j! NO- = (Z) did Aq UOA! S J•aUO mnp~sa puooos oqL 'A•~M ours oqlJ uI •IIp Oql jo ueuLIA Oql jo IsouI .oj siunomv pup (os poirugisop AIsnoTAoid) OnlEAUO!O Is0a91I Oq4 ql!A pMW!OOSS- ST JO130AUOSTO ~gdomuud isnj oqtl Wq smoqS llnsoaJ su.L ([x[ =I [] (gzci) 661 "O3UtI•I•A unnurxeum oql umldxo I. (siuuodwuoo) SJO1AUS!O •dioui.Id jo azqunu mnunun ui oqi imp os XuM I•rundo uur bj muosaidai 'L'CI ds mvnp atp jo soxy mauipooo oqt u! posn 'SJOIOAUOS!I I•tUOUOt1uO XUtmnui jo los onbiun qoqi M04s oqisuoIVInolo SOtsoLL [=/ Y+fX r+[y = [+fQ [•,] (Lwa) i~'_x 'x~l -[_=l] [] (sE•a) puts :suogpnbo pozi.tumo BuTollofoj oql utsnfAq ,Xvm oures Jo 30UMMIA uUiMturaZi Oql jo ISOAIL' pUOOOS OqL wsoTu U.poutm toq Toj qo oiv SJOP1A0jAUS•TB ~UTUIuJJ .•l• lj p ;qi slunooo• pue (os pOwuS!szp AsnoiAoid) OnflAU0g!O LJUM pOIW1OSSE Si JO1OAUO!TO edimuud puo:os ozp 'aoloaotjLL (9ga) ZX ZY = Zx '[u] ui sllnsa EE'(I "b oiu! 'Ci "b3 2uT~uosuI 1x'x [Y--]= '[•] (Eua) su~tlqo ouo '[•I tijol z xultwu Inplsz' is.~j oqi su uontnbo sup jo apjs ptnq a l .ugmujzp'putm 'uouiiM q uwo rx byJNY - D][ ]= g[]]•[g] pue 'I 'Cl '9Q 'gE'C "sbg tuoj_ uoiTnbo oqi 'D IID t= Iy dtqsuo!ilai oqi1 W1q4 lOS VJVp OtJl JO Oou Ajo sJO Tun 00Z s•l iSaIlorj u ut sluotUolU J lioj sulpeol .ol: j painbs oqi jo tuns oql 'axoJo.otj '(0661 'zOtu1O pPUP utMUOH) .xo1o3j o3q jo 3.rLeA uotuioo oqii Aq poun. dxo oq uro Imapiuouolo Oql jo oou.reA Jo uonotij Ozi szl Ojpu. .iopi:j mJ.omds oqi jo qooe .ioj iuotuolo uu jo Sinpeol iomoj oqi Jo onnbs otlL "swuououla poinsvotu uo oanos qow~ jo wuruodtn oAptla.i zqi owTotpu sfltOi .olimj '•.o•M Siqt uo J.oPAUOAIO tdidou.d qolg jo oouriiodt•. UI "SJOIOA MW MVp poz.ILuou OqJ Jo 143WO Xqi pu, 'su•.poovo .owif ponv o f•.oxr~ sluoogjJooo v On•z oinsf OAT1IO.J 101 J oso MAOJ MVl p ,*f Oq40.oqm Ot4L ("([J]N•[]) JO (!A) SJO•11AU3913 I •q pauJuoj ST 01300A [=.1 f (wr'I) '6ECII "b woI 'X = Nj"_ "[I[A•]=[N][I] 'v,'u "b3 oi uosureduioo q 'liqi iOON ",qoAlloodsoJ dJOos wuouodtuoo pur guipvol luouoduoo 'SODueWu t4l1 poyo •pg[oI] pUre [y] oajqm [N][x] = [b] (otvu) pg[r]=[ ] u.urjoo pue [L, 1y]I=I[xI = oql put 'iuooqjl.o oa *"enbo ame (NbI[D]£•¶]) w oj q [f] put [I] swoult wqi pue (N•[•N[D]) jo sfnl•[ua0! oloN *sOnfrAUOZ0 2ulpuodsauoo oqi jo slooi oanbs 0q4 jo dn oprn xumwu IeUOge!p tnxu re s! [ny] pun '(N,]•[]x[m]) JO sJO1a30AUZS!O LUXLU Oq1 s1 JO sJo•1130AUotO uxu oqI s1 [1] aJoqLn '•[D]N[D]) '( (6'C) l[r][fn]-[I] = N[j] (1661 'AluOH) se ,lonbiun uollux oq uto (xaleutu Atr o) N[D] xulltu 0q4 'tuaoiooql (GAS) "SJolDOAU0310 put sfnlreAU02TO 04q Jo uoflsoduWoap anlvit •lnguis ql o0 2uipmo03V suIol mupossoidxo oq osie uvo xplatu mvp pozqnteuuou 4q1 jo sJol130A MW lelwp 0q4 Ioz •umdooopTM opASI s oi onp posn frm uouaSPa o anlUvAuagda agflad3vPIps.lT-os~4 'qioM Si!Lp IO qa uI "(0861 'aJAMOH put 'rsMou-iA[) sil 'uoissaidwuoo joiovf pogm• si ssgood uoisnlxa AaT4o O1 posn snal spo0l4tu1 snoIV3 puS "ords joieoj oql uaoij popnloxz oq plnoqs put saoao ImUou.ocldxo utoIj ilnsax sioI3-J oqt Jo isai oqL "mIp tlp o1 paIzla Oa sonlEAU0T80 ISSJl•qqlq.IM SJOI;J , 431P Ajjuo 'Wlmp oLi o0oinqUiuo) soonJos 4 jj "SJIoPOA MoJ m1p jo aoqtunu oi oi iou 'soainos jo jaqtunu oqi o0 Imnbo zq pinom (sJoI3)AUasTO) sioioj jo iaqtunu 0oq 'sioio Imuou. *dxz jo oaij aomo mvp Ituruo oqp jI -'O - f( j 041 u! su oaods JOj3;A 1u poziJnsTA Oq uO smUPeroI J01)J pue saioos JoIoej jo id0ouoo oqj "(z661 'IV10 !PJIMW '9L61 'ugttuL-H S"o) a1nflo0!i0l ui punoj zq uro poptwu .IssvXur O0'v io1irJ p.irpumis 4ql inoqe uoiltuuoju! po.ieog3I "uoivenb3 oisuq 0q4 se posn svM "bg puin 'paiouai srm xwui oloirJ iou ' CI onbiun oqp 'pIoM sItI1 I. poJajoid srx '•'j "bg "b3 ooumiS "soq:poiddv snoTfiA Aq poirurn•so ami taot put 'uoud r umout, lou zip sotltunwuuoD "(oIqvuPA i•tq ioj it i:oouiod r sorltpui ouo) s.ioiolj oql jo 11vq pou•idxo s1 Blql OIq.UPA V JO 0OtULIVA Oq1 Jo uo!IOVJJ l41 501OpUl AI4tmlunu00oo 0q4 'ooJoJoaigqL •sJOj3VJ HE al.AO 0lq.uPA W1q4 Jo sduiproI oior3j pairnbs 4oqJo uins oqi SI (IspoM s!li ut lu ooo) oIqyuVA VJo AirIuntuLuoo zqiO'spiOm aoqio uI "sJoIo3J uomumoo rA solqrMA .Olo Oql 4111M pOa.qs ST WII4 gjq0trTJA Vjo ZoULTrA oq4 smuosaidoai tr!tvunouiwoo IOaL.E.d O.,L xz4Vwmu uoyviladdoo patnpai o41 PolTr si xLtlu uIf5 0lnsa oqL "s0i¶eunttuoO portuinso oqi tpl (I = !ie) x.netu uopolaIoo 041 JO SiuoLuolo luofrip 0t1 U!F0eldaI Aq ouop si uommopjpou siutL "solqJtrA oqp uooA oq disuogwIoa 041 ur.idxo o0 s5o0o1J uotumoo oqp qilm powuropod ona suopluodo oq0jo isao oqi itlq os 'sioiaoj onbtun oqt jo I3joJ oql opnlfxo ol poTIpowu si xutui uoivolauoo oql 'sislApu 01 UT posn si Zt-C' "b3 j1 "uououoo parq1s iou put oqletPEA 413o 01 ojtoiods zae .iopVJ uxUn onbiun v sT [,] OaOH "'y wveq suu.m oqp su-uuoo lwql X.LuI "bg se uuoj loumtuorui ourms oqp srq qo0qM [-] + [p 1 jff,] = [31] (z7"u) SE X.LxlU mv1p poz.IuuJou oql s1uosoadoJ Iopouw sisX-ert zolov1j pirpuris oql mtpl oloN "Jo10lJ Wrq4 Aq pouwidxo oaJ oql ul punoj oq uro oajoiao!I "(9L6T 'utuurH '.L6T'IotuumI "'19o) inop inoqp uoiuluuoju, po.[mlop pu, 'zuo InJgu.unotu A •o!sqd aIou o oi o ds JorJ ut.O.uo Oql UOAUOO O1 posn oq upo spotpoux uo!lploa sno trOA "u.umu I3TsAqd OAlq weqj sJo0 j OA!S ilM UaOIql Jo Oinos AIJuo 'prnmoloi q Au sox- jo Jos r qotrq qinoatw sisixo suomllsod jo iaqtunu oaut•.t up qtnotwllV "wlzp o iols uonnlos lonoo v LuoJ os[1 soxv pol•o.J oJL "sotru U=.•a o1 fulpioXv tUOtl MOloJ o1 sT sOous!sop ol jo r IO:sAqd ouos oiro!pu. soaodoaid o:srq qltisoxv oouaiojai ozq ug!j ol ArMaouO "Surotu oi soloinj utunloo put MOJ oqj joadxo oi sisvq ou si awoq put 'sooiwvJ otuos jo sunti ut los virp oqi umIdxo ol podoloAop stM qooaiddu Iwortuotquw Aoand v 'Xou F.ufl '(006 31 SJOPOAUM5! ztl pUE SoU!I paqsup aqi uzOmA•q safItu ztl 11) sautl paqsvp Aq UMoqs WaE (sIoMAu03! Oqp uo •D 10a3oA MOI mvp ;qi jo d!l qtl Jo suo!3ofod) t1l uO 'd aaugtsap uumnioo pqJo suoprafoid) swooas pUr (SJO133AU30!) sioltj oaa.pl qpAm ;3ds JOICA aldum•S "1-(I aan!ni s•gup!o .1oPJ I .o13IJ pu~ (S.•OyaAU2!a 'SJI0o3A UUInlo3 inp 3A!J ZX ZMD "Iopotu oql t idol oq plnoqs pup OsIou upq ur jups aows ojo ouo SaOlOOAUO ol Jo sotlWI jrll polsoa 'aioJaqL OJOIO IUmroja sonIVAUOSIO qIAM (1661I 'AIUOH) luouodtuoo ediouud io 'jOPOAUO!O polWOossr OSIou-o1-rugis oql WIt'nuosso on, xLflvuu uonoioJuoo v jo sonjUAUO2o1 1J pOooi z ns uooq os" svqII-p U Ur 'Alun Jo onIVA OaT1OAV up1~ .Oi A n oolq sonIV tl. sjorvJ STAU0!O t uoWnI UOIVo.o a0q oouTS "popnlPxo sonIVAU2O saq xSm4 oq pInoqs OaiNOA, Oql moolaq sonIVAUO!O qliM sio•PJ IUp Wrql soips uoJuO!uo) situ atl uo 101o0A AOJ VIVp ap iioj sixt ,IO •pmaldtuoo uooq svq uolmoi oqJ Sx Jo 2uipPoI oxl si put 'lonIVAU131.O,, !T1Y '0aiOH "IoATrloPdsoJ '(szlqvM.A) f Olp S.I0I:3;A MOJ MVp puP (siolopj) S.o010AUa0i! JO iaqunu aql arotpu, utpup 4 sjaizumt~d oiql aolqA 7=u i )I- :I=A Uy U sT pataUp sT pup x.wLtu 2uip•.o oqi jo uTunIoo qovo utn.w. sduipoli pounbs zaTl Jo sooU.UgA oq4i jo tuns X[AIZAIJ~I. pouuojppd sduiproI oqi jo uolrunj optrpmnb s qnb -IL 5Iyst uorpunj rtlnw·p r Jo uontzituuixU oql Xq poazitp~ sl uOIVol-I xVtu1rA "ulOt uaoaMuq OaInos uoutuoo v ino muod uo alovlj v ui smuiproi q•nq ql!mM sojqý.u A 'lnsaJ i svV "ioloJ ouo qiyAM pomo!ossr si qotqM jo qtao 'sios macmdos olui i.ds oq I.M solqt.UA ql WqlIp Os S u.pe~o AI jMaOJ ojoz-jaou AuvLu put IOPOAUO!o0 SoXlv•oO s0ropl oismq oJtL "sx) ql.~Im oaosoad suoplioi IuogoqIo IUi.uo oq uooAaoq Salim oq 3l 'aosniXI) onbuqo3m uolurmoi [uo•oqlo uouimuoo "posn s•~M '(896I 'ruopvia ' XVwlv4 ').IOAm sItp UI t7OZ • asonap uonpodop ouozo Jjo.uq V oia.qdsolr.iis snoauoo•al oqa jo A OaiAO 3 XIUNIddV :Molpq UOAI2 O.e sLusIutpuo ", OZ-OT mnq.luo3 (9861 "9l m '(E661 'muloI :u.sruq3oul OI/01D uolloi osztL -uounqpuoo Ivtus r a u osl OaV (9861 '*IV 1 uotuoIos) u.sMq3lOau uI suoT.1JOuaw OID/ZOH OI gu.lAOATIT suoR0ap lioAioi93j) usTr•npaOu OIID/O5 O1iiqnot 01q 'uoPIppE . ONl ?UIOAX " pup uuqqv) uonrlnnsop ouozo Jo % 9L AIMtmu.xo a ddr .oj olqIsuodsoi ai (La61 'eu.'lON putl suoiitean irw llUioiN) LuSTUl O 013/0OD D/OuID OT 5UTAOAU. poAo!Ioq Si ii ,epol -(VL61 'PtMIo~tI pu tU.ION) sD3:D OT jo sisAlootqd oNtu.poonpoad spunodtuoo ouuo.oqo ;N Npi • ouozo ouqodsoiers Aonsop plnoo 'soseoIal oiuaoodooNup jo 1insai r 'oaOqdsotuIe zp u.(DqD)suoqIooaongoolqot 2uls•taIu! Wqit poisouns suIM ii 'iq•q ol otuo puoai uonoidop ;uozo sun urqN iaoio opeoop V "(L861 '"IE10 uuVulJOH .9861 'AoCot) oljoaid ouozo poqiuaodun ue jo Vod oii q•l• luop!0uioo 'sIartuorol, OZ put 01 UoaAiaoq Jnoo p!p ou.oop ouozo toq jo isot iueq poiodai swm I "dn svoa) q XOJJOA Jelod oi uoqmlsopPmPll plt Pul siaLOuIoli. JZ aou sopninllt t uoij JI. qX!u ouozo Jo xnl• U" oNrPIM iaoqo10 o olm snloq fa3AOa0J putl 'aoqoioo-pau Xq uonlnuzauooo oqN jo 9%og Sst3ffo S ol SNsoqotoJ uoilnpoi u Illol lllip poaoqs osir sluouoaJnsluo U .n•! osotLL sopniplr aoquoidos m sunwo uon;ild;p Iruospos umuxrlntxu otL -a "(0661 "Il 0IJlqsO) Jao•oI o sp)ooood u;ql pue saouioolrq ZZ pul OZ uoouzq opniplv u tesurls uonoldop Ouozo W141 pOMoqs E3nfaelUV aOAO siumuuoansTOtu UOOllq "OlOq ouozo zi jo aonloruls IlenEds pue ieaodwuo oqi se IoM se 'olqlsuodsoa sosso3ood oaaZM sluotuoansvowu 9861 "'loVu10uullE) pontuq3 zip puvisaopun o Opetu ole)S SjoiraoqeI put rll s m oASUoiXo 'faoAOoSip siTq U.pa0IAO3STp raNw "(86I srtIJ SUM 3T113JlUIUV JOAO uopoldop ouozo our.i 2uudS "(E66I 'tu!iOlA pue utqqV '!166T 'putlmo-a "2"o) aintaimIl ONj U!pOaMAoa SIZA!SUOvax Oaaloouozo opoqdsovalls jo susnmrqootu uoo3npoid DtLL -(E661 'eV.lolA pUt nuqqV) oaoqdsoivWs oqtu.sopsaoi ouozo m!aoqdsouft Itloi a3w jo % 06 A10euo•.poddV aljsnufaUq uof.a1lap auozo apaqisocaris snoau5o3a5aq aq4 Jo Al.•ai jaLq V -sU2 OsaO 'uolIppP uI 90Z "(0661 'uowoloS) ouo1- Xnsiutaiw asrd-sn2 aond Xq pouit{dxo Oq ioUUm puan uooljdop ouozo oplsmap pajosqo opl veqi pamoqs aoqdso!is Oql uit smo, ds 0AIJOU.X Jao[mu Jo sllnsaoi luuoainsotu oqi ~tusn suoilrlnlcmo 'JOAOAOH OE <-'-OZ W:°N zO + zOH <- 'O + HO ZO + OlD <- lO + ID HO + 1 <-- At + IDOH 'O + IDOH <- OD + zOH :tusiurtqOu OID/ZOH ZO+0-1I<--10 + is kO + OID <- £O + ID ZO + ID + ja <- 010 + 0•[ :usnr•STUlaur zOý -- Oz OID/Oia :lN ID0013 <-- OID + 01O ( ZO + OID -- lO + [D) Z O + IDO001O <-001D + ID -- A't + [D001D LOZ '(i) Z~ONOD - (2) ZON + (9) 013O :uo!iorai SuimloOJ oql qtnoiil saoIids zONOI3 JToAJasZ0o1 UOTSJOAUO0 l•' p•. Jaj 03D SSOl U! sllnsai qorM 'ONH S1tid plIos o0 zON smq 'uo.ppg uI ostqd sS otp uuojsura stusiunqoou uo oroa voiwtuoq snoouoS~oi i '(s) OZH + (S) 3ID- (s) D13H + (S) D30H :(Z661 'Jaql'wJd) uovooai u•wuolloj orl uvo tusirutoou uouiouai snoouo•oaoq r qrnoilp sooods D130H zD3 Iuo~nPPV "uow3ioossipoloqd q4nol0j oaI qoiIM 'solnalooou IDOH Ppu suo0t1oaJ osoiq 'oaiojoioqaLL atpAq oauld oql uioij uolionpoid suor 13 ooaJ ol pou0ojsu-.n ,ll~nmuoAo I3 sr qons 'SUJOJ oATrl1 oiu "IgoAnoodsoa 'sosvqd pqos pup sv ONOID013 ol0 lJoUL uojsu,.n uipm s pu S oajoqm (s) £ONH + (S)IDOH <- (s) OzH + (S) ON13O pute (s) £ONH + (2) ID <-"(s) 13H + (S) zON0O3 :SMOlloJ sV uOAIT OaJ suo0T3VaJ Io0A.OSOaJ Zsta s3Sd Aq pop!Aoid soojjans uo oorId oi qlnoj0Jt Solnaolowu o!tuoqo4 'sossoooad 2,qo 'L861 "(L86i "II ao i qotYtIM OZH pue 13H ql.M I3•ov solnolotu ZON0O3 'oAo,•OH "*I 10 nUIO0I ) snomuo0o0ao1q "suoo.ta mglod oq ul ouozo il!AX m 1AiJp iorao iou op pup sJoAosoaJ zuuolq:4 ajou A1irluosso oai soioods OsOU. "(1661I u IlSOwu 'tu.qol) zONOI3 uoaSOpixq Jo tuoJ a0 'MaV1IU OUPOIT put' '13H 'OPUOlqo oaJqdsoldls opm!TWI q4.rq oq4l qoaJ uoTnsoduoU p 3D3 tuoiJ sonpoJd •olo q 0uozo oPnJ'luV oql gu41.. ldxz ut stusuerqootu Sol aoiam (DSd) Spnolo ouzqdsomlns irlod 0q4 o0 onp sossgood 3n.tuoqo snoouo2o0aoiq ipqi •oIlo otuo3ooq sN 1. 's.goX lu0a3a uI "(1661 'uooJL put 1truIH) uostzs 2uuds 0q4 uT samooo 11 XqAM pUr voiUrmuV ol politu. si uonoldop Ouozo a0JAos 0q4 Aqm ur!ldx0ol lqv iou oaOa stsusrqmLtu uoop•ga os~tqd oiqd~rJoio pldri jo ilnsoaJ APuonboaj ssoI OUlL 80t s' uuoj 'spnolo oi q ol iqgnoqi osl 'sSd III odkl poaJOsqo "(1661 'UOoQL put 'IImH 'q8861 : I odAil 3usixo oqp uo oo! a~-io slmSas 'Ja0J0qsaonrlAN pug uosugH) jo uoimisuopuo3 0q1 Aq pouuoj an, Amo Itp pOAMoIq s 1I '(1661 'uFlqoN) oauiJ ozis tu(i001-01 o0rui D0 98- inoqr tAoloq uLoj sDSd II odXL '(V8861 'jaoJgqsJaonlA put uosUgH) slulsJIO ' E(OzH) •EONH '(LVN) aBJp(q14 moloq sainwiodurm 1ituLoj meqi sJSd 0aJ q:WtqM turiI X[qno. OAL1q pug D. 8L- inoqe p 1oldoap p~og opTrltu Al•.wu Jo uou!isodwuoo t qlwt 'I Ocl( si ouo uouituOO IsouI OtuL pui (9861 'rndoij pug yotuuoJ3l) saloumt "(6861 'I- lo uooj) III pue 'II'Isodf4 sE poJiss•po J DSd luoaOJJIp 0zJUJL porIuop! OJoM soJflonfls otpI noqe ur.tuai UIps sonssI POalosoajun "t1utl Jo susumqTu1m uorlUtuuoj Autui 'qoaosai uoroldop ouozo ouoqdsouJis jo seoA Aie oaol poaeduoo '-pol s:Sd Inoqg uMout s!zaotu qoLnuiqfno0tI •oa•( op!xo ouuoplqo o!iikw oq qgnoil ooId soae, uoipnaisop ouozo otioanj os 'uoneuaouIp0s 0a1ou UOISJaUO3 zitp odroso slp O) 0.D puI 0nslTu OULOlqO IJOt 0141 01o qinoiql •lwuoueuaod 11 Jo Ouos OAOuIa.1 flqIssod put p!o:e :)aL oi op!xo!p uonojiiu oqi IJOAUOO Aorq 'puooZS "smoods SuiAoxisop zuozo uretu 0q1 s qoiqtA 'IDjo uononpold "1i30jja 01ql oz0•ito A0[1 '1sJj o snoouaool0zq jaOI qynoaqi SurzlXe orol•aluV JOAO uotloldop ouozo 3uoqdsole•s 0ql uo soloJ iumiodtul oAi Auld sDSd 'Ituwums tuI uuoIWuiuipos O41 uIOJJ ILAOU0Jo0Ial!u jo suolEoU!pu JSd Iqfnoql 0iazqdsolirls Oai ZaXoql 'uo01ppr ul "(T661 'UOlL pre IFUeIMH) o motql uooa4oq o3ugleq o:iugukp oq 01 fnp (uonrotyxouop) sotoods XON jo uolonpoa 0ql Sq pamOlIOJ '1U0WUOU.AUO 0ql U! 9OZN JO uo3onpoaJ 0q uT.lnsoJ suon13avo os0ULL (s) EONHZ <- (s) OZH + (9) £OzN 'pue (I) ID+ (2) ZON <-- Aq + (2) IJZON (s) 'ONH + (3)lDON *- (s) IDH + (2) £OZN :(8861 "lIe jo uaoqloL) toloq UQATS oa ';OzN 'opjxoluod uo0oaiuTp qifno4aq omue•lq XON ill suwsnttqotu uotlova0 snoou03oJi1oq IruoIn1ppv q1 loaojje "(1661 'euiloN) soplo onp•lm 013 qfnoJlq uotlonalsop ouozo oaowu OD surou qowqm 60Z "(1 661 'uooL put) IHUMH) slosoJor pltop ounlJns uozoJj zql uo Uuoj sDSd I odCl IuqW oql u! plot o.njins st p 4linoql s. ii 'lJaoqdsolmns oulnru no• qonu s soum 0i si ijoql qxnoqllf 'sozn.ird osoJo• oiJlns punowiijopq Jo oqtunu ot oi Wnbo Jo ureq ssol zq oqsmoddr solo3ld plt ,3uiiu jo .nqunu oqj owuS "aoqdsodon oql tuoj pauodsuri sljo•md io saojuoaotuoJoitw st qans 'uljmvtu olnn.nrd .nqlo jo sz: Jins toquo ssoooJd uonrzilnu snomuoo•rnoq v Xq io (D. rn ol t ssoojd uonrolnu snoouogotuoq v ,q .qi!o porimouf 8L- Moloq) saJm•rnodium oq o liqnoqi air sioido.p uopnlos plop a.ImJIns osatLL "smiJpAq tuuoj pup ozooij ,Ap ssolun tusiuvqo3u osU•wia ouu.olq: Alr u q uoion.nsop ouozo tuloutuaod snTp 'solnoolou IDH 2uTdd•an i uoijjo iou oe saoldoJp uoQnlos OH-t'OSzH plmbq 'lunpunqg lqilrq am otp qinoqll (1661 'tuqolW) itql pouzod.• osjj svm 1i '(1661 '•uTIoN) ojoitdsolils Oq ul OAZI ainssaid I-eiud iuaunn q i i odA IJDH OIpPOtuto3m o0 JiugIFJa ayq JTqiajjo osnmooq suognorw jIuawuq3 snooufonioo q oq mrotuoad oi ,jqo 0Jouw 3.- sDSd I odAi oqi wqt iqrnoqi si iI "(E661 'IuiloW pup uvqqv) sUoA0 fu.qooo