Intercalibration at Sea of Hydrographic and Chemical Observations Aboard the USSR Ship PROFESSOR VIESE and the US Research Vessel 1PSON during FDRAKE 76 G C 8 5 6 1 0735 AnISOS Technical Report a.. no.77-7 cop.2 'Mar 1 ne Louis I. Gordon George School of Oceanography Oregon State University. Corvallis, OR 97331 C. Anderson Scripps Institution of Oceanography La Jolla, CA 92037 Worth D. Nowlin, Jr. Department of Oceanography Texas A&M University College Station, TX 77843 Sponsored by the Office for the International Decade of Ocean Exploration of the National SCience Foundation August 1977 OSU Reference No. 77-7 Results of an Intercalibration at Sea of Hydrographic and Chemical Observations and Standards Aboard the USSR Ship PROFESSOR VIESE and the US Research Vessel THOMAS G. THOMPSON during FDRAKE 76 An ISOS Technical Report Louis I. Gordon School of Oceanography Oregon State University Corvallis, OR 97331 George C. Anderson Scripps Institution of Oceanography La Jolla, CA 92037 Worth D. Nowlin Department of Oceanography Texas A&M University College Station, TX 77843 Sponsored by the Office for August 1977 the International Decade of Ocean Exploration of the National Science Foundation OSU Reference No. 77-7 ABSTRACT On 15 February 1976 the USSR Research Ship PROFESSOR VIESE and the US R/V T.G. THOMPSON rendezvoused in the southern Drake Passage. The ships simultaneously occupied hydrographic stations as close together as feasible. Calibration standards for salinity, dissolved oxygen, and dissolved reactive phosphate and silicate were then exchanged by the two scientific parties. The calibration standards were found to be in very good agreement, the differences being insignificant with respect to measurement precision. The results of the hydrographic station intercomparison showed aver- age differences with regard to sign between the two ships' results for temperature, salinity, dissolved oxygen, reactive phosphate, and reactive silicate to be realtively small but the root-mean-square differences to be quite large. This indicates that general comparisons of the two data sets are possible, whereas detailed considerations are not reasonable because of serious measurement errors aboard VIESE. INTRODUCTION During February of 1976 the R/V T.G. THOMPSON, operated by the University of Washington, and the R/V PROFESSOR VIESE were both working in the Drake Passage. The THOMPSON was being used in the field operation FDRAKE 76 as part of International Southern Ocean Studies (ISOS), while the VIESE, operated by the Arctic and Antarctic Research Institute of the USSR, was participating in the POLEX-South program. Prior agreement had been made for these two vessels to rendezvous in the Drake Passage in order to exchange standard materials, and to simultaneously occupy hydrographic stations at positions as close together as safety permitted. On 15 February the vessels rendezvoused at 60024'S, 61°19'W. COMPARISON OF THE STANDARD MATERIALS During the reciprocal visits by Soviet and American scientists to the two research vessels, standard materials were exchanged for subsequent laboratory comparisons on both ships. Standards exchanged were for salinity, dissolved oxygen, phosphate, and silicate. In addition, the Soviets supplied pH standards and the Americans supplied nitrate standards. We now present the results of the comparisons of the salinity, oxygen, phosphate, and silicate standards done aboard THOMPSON. a. Salinity One case (25 vials) of Soviet standard seawater was received from the PROFESSOR VIESE. This water was labeled as batch C81 a nominal chlorinity of 19.379%;. dated 6/V 1974 with The salinity of six Soviet samples was determined using a University of Washington (1960) conductive salinometer 2 Table 1. Comparison of USSR and SIO dissolved oxygen standards.* Nominal USSR Concentration Observed USSR Concentration (N) (N) N-2/SIO-1 0.010154 0.010152 0.02 N-2/SIO-2 0.010154 0.010164 -0.10 N-14/SIO-3 0.010073 0.010044 0.29 Standard Pair Mean Diff. + 1 SD Difference (%) + 0.07 + 0.20 * The standard solutions made up from the USSR starting materials were measured as unknown samples using the SIO materials as standards. Three or Nominal more replicate titrations were made of all standards and blanks. concentrations given here are based upon label weights. The USSR vial labels indicated that the contents were for preparation of a .01000 N solution, but the listed weight of KIO on the two vials were 1.54 and 0.73% greater than that recommended fo11 an .01000 N solution (Carpenter, 1965). 3 Oceanography. (no. 14), owned and maintained by the Scripps Institution of samples were The salinometer was standardized before and after the Soviet batch P67, dated 10/8 processed using Copenhagen standard seawater: Based on this standardization, 1974, and nominal chlorinity of 19.3735°/°°. lower than the inthe values obtained for five of the Soviet standards were Soviet sample dicated salinity by approximately .002°/°°, while the sixth was lower than the indicated salinity by .003°/ °. The accuracy of this method of salinity determination is about this same value. The precision of the salinometer is between .0005 and .001°/°°. b. Dissolved oxygen (KI03) Two sealed glass ampoules of dried preweighed potassium iodate powder were received from the PROFESSOR VIESE. These were compared with Institution three preweighed standards, also KIO35 prepared at the Scripps of Oceanography (SIO). Standard solutions (nominally O.O1000N) were pre- outlined pared from both sets of dry materials according to the procedure by Anderson (1971). The same volumetric glassware and distilled water were used for preparing all standard solutions. Blanks determined using either the same set of standards according to the method in Anderson (1971) were within experimental error. Ten cc. aliquots of both standards were titrated in replicates of three to eight titrations and the averages compared. The results are shown in Table 1. We made the comparison on the basis the of the exact weights of KIO3 given on the VIESE ampoule labels, using THOMPSON solutions as working standards. We consider the agreement of 0.1% to be very good, and probably not significantly different from zero. If with real, the difference could have arisen from (1) our lack of experience possithe sealed glass ampoules used for the VIESE standards and resulting 4 Comparison of USSR and OSU nutrient standards.* Table 2. Experiment 19 Feb. 1976 7 Mar. 1976 Mean: * Silicate Phosphate Date of Nominal Concn. Observed Concn. Diff. ( mol/1) ( mol/1) (%) 2.409 2.42 " 2.39 2.409 2.405 Observed Concn. Diff. ( mol/1) ( mol/1) (%) 85.4 83.2 Nominal Concn. 86.2 0.15 85.4 84.7 0.8 The standard solutions made up from the USSR starting materials were measured as unknowns using the OSU materials as standards. Five or more replicate comparisons of the solutions were made on the AutoAnalyzer on each date. Blanks made up in the same matrix as the standards were subtracted from both sets of absorbances before calculations. 5 ble loss of some material during transfer to the volumetric flask; (2) a suggestion from previous data on standard SIO-1 that it might have been slightly high in concentration. c. Nutrients Standard materials for phosphate and silicate brought aboard THOMPSON from the PROFESSOR VIESE were preweighed dry potassium dihydrogen phosphate (KH2PO4) powder in two glass vials and solutions labelled "500 mg Si as The Oregon State University SiO2 3per liter", in two plastic bottles. (OSU) nutrient standards in use on FDRAKE 76 were preweighed powdered materials stored in screw capped glass vials. The phosphate standard was also KH2PO4 that had previously been compared with US National Bureau of Standards NH4H2PO4, Standard Reference Material No. 149. The silicate standard was sodium fluosilicate, Na2SiF6 that had been compared with S102 of stated 99.89% purity obtained from a vendor who supplies high purity material to the semiconductor industry. Both comparisons had been carried out in natural seawater matrices. In the present work two comparisons were made, in identical natural seawater matrices, each starting with a different sample of both US and USSR standard on two different dates about two weeks apart. Standard solutions were prepared as described by Nowlin et al. (1977). Five or more replicate comparisons were made on the ISOS precision AutoAnalyzer system on each date. The results are given in Table 2 with the OSU materials arbitrarily used as standards and the measured USSR concentrations compared with their nominal values. The two sets of standards are indistinguishable for phosphate and probably so for silicate. 6 THE HYDROGRAPHIC STATIONS Two hydrographic casts with STD attached to the wire were.made from the THOMPSON (TGT). The STD used was a Plessey model 9040 with both analog and The recorded sonic depth, corrected using Matthews digital data recording. (1939) tables and the depth of the ship's transducer, was 3733 m, and the second from 1 m to 590 m. two casts.. A total of 33 Nansen bottles were used on the Depths were determined from paired protected reversing thermometers on all bottles below 75 in. and unprotected For each bottle in situ temperature, salinity, dissolved oxygen, and reactive inorganic phosphate, silicate, nitrate and nitrite were determined. With the exceptions of nitrate and nitrite measurements, the bottle data from both casts are listed in Table A of the Appendix and plotted in Figure 1. The quality of these data is representative of that obtained on THOMPSON during FDRAKE 76 and that obtained on the R/V MELVILLE during FDRAKE 75. The profiles are quite smooth. PROFESSOR VIESE (PV) made two hydrographic casts, the first from 661 to 3575 m, and the second from the surface to 636 in. were used on the two casts. A total of 30 bottles Temperature, chlorinity, dissolved oxygen, in situ pH, reactive inorganic phosphate, and reactive silicate were measured. The measured data, excepting pH, are listed in Table B of the Appendix and plotted in Figure 2. The nutrient and oxygen profiles are quite irregular and contain many one-point maxima and minima indicating that large measurement errors are present. Table 3 summarizes the average differences for the vaiables measured by both vessels, the rms deviations between the data sets, and the numbers of differences employed in making the computations. The differences were TEMP (DEG C) 0 1 1 2 1000 2000 0 W 0 3000 40001 1 FIG ' 1 -- : TGT STN. H21 SALINITY (PPT) 34.0 35.0 34..5 i 4000 FIG 1-B: TGT Sm. H21 OXYGEN (ML/L) 8.0 7.0 6.0 5.0 r-. 2000 F- W 0 3000 4000 1 1 --, -L FIG 1-C: TGT Sl'N. H21 0.0 2 000 PHOSPHATE (UM/L) 1.0 2.0 3.0 4.0 5.0 i - 3000 4000 FIG 1-D: TGT STN. H21 SILICATE (UM/L) 50 100 1000 2800 3ee0 L FIG 1-E: TGT STN. H21 150 TEMP (DEC C) 0 1 2 FI C 2-A: 8U 8T N. 28 ? 7 SALINITY (PPT) 33.5 0 34.0 34.5 35.0 0 3000L FTC 2-B: PU STN. 2817 OXYGEN CMLiL) 5.0 6.0 7.0 1000 2000 F--- 0_ LU 0 3000 !- 4000 FIG 2-C: PU STN. 2817 8.0 r-- 0.0 0 PHOSPHATE (UMiL ) 1.0 2.0 3.0 4.0 1000 2000 n LU Cl 3000 4000 FIG 2-D: FU SIR 2817 5.0 16 SILICATE CUM/L) 50 100 3000 4000 F I ( 2-E: PU STN. 2817 150 17 Table 3. x rms dev. n Statistical comparison of the THOMPSON and VIESE intercalibration station results. AT oS (°C) (°/°°) -0.003 0.12 30 0.0000 0.04 31 002 (ml/1) AP04 nSi (umol/1) (umol/1) -0.02 -0.15 -0.94 0.19 0.47 7.6 30 31 31 x denotes the averages,with regard to sign, of oT's, AS's, etc. (TGT-PV). n denotes the numbers of differences used in the computations for each variable. Individual differences were computed as explained in the text. 18 computed as the difference (TGT-PV) between the TGT observations and the PV observations linearly interpolated at the depths of the TGT observations. In general, the rms deviations are one or more orders of magnitude greater than either the signed average differences or the calibration discrepancies discussed earlier in this report. Although the numbers of paired observations are not large and the conclusion, therefore, not definite, histograms of the differences indicate that they are not randomly distributed in the cases of most of the measurements. SUMIARY AND CONCLUSIONS The compatability of two data sets depends upon the agreement of analytical standards, the measurement techniques, and procedures employed by shipboard analysts. From this work it appears that the USSR and USA stand- ard materials used to calibrate the measurements in the respective national Antarctic programs are in very good agreement. The salinity standards agree to within 0.002°/°° and are indistinguishable as sets. The dissolved oxygen standards agreed to within 0.1% while the nutrient standards agreed to within 0.2% for phosphate and .08% for silicate. discrepancies are within experimental error. The oxygen and nutrient Thus it appears that the the fundamental bases for accuracy and intercomparability of data sets, analytical standards employed, are in good agreement. Since the signed average differences between TGT and PV measurements (Table 3) are small, it is clear that the basic measurement techniques and analytical procedures are comparable. The rms deviations between individual pairs of observations are large, however. This is true for each of the is not variables commonly measured except salinity, where the disagreement exceptionally large in view of the fact that the PV measurements were reported as chlorinity to .01°/°0 only. As can be seen clearly from comparing 19 the property versus depth plots of Figures 1 and 2, the large deviations result from scatter of the PV data about the TGT data values, especially in the cases of phosphate and silicate. Since the signed average differences between TGT and PV data are much smaller than the rms deviations (In fact, they are of the same order of magnitude as the difference between standard materials.), it seems likely that the differences between the two data sets are due to measurement error. Further, because the TGT vertical profiles are smooth while the PV profiles are quite irregular with many one-point maxima and minima we conclude that the errors are present in the PV data set. 20 REFERENCES Anderson, G.C. 1971. "Marine Technicians Handbook. Oxygen Analysis." Scripps Institution of Oceanography. Reference No. 71-8. Sea Grant Publication No. 9. La Jolla, California. 29 p. The Chesapeake Bay Institute technique for the 1965. Carpenter, J.H. 141-143. Winkler dissolved oxygen method. Limnol. and Oceanog., 10: pure water and Matthews, D.J. 1939. Tables of the velocity of sound in Second Edition. sound-ranging. sea water for use in echo-sounding and 52 p. 282: Hydrographic Department, Admiralty, H.D. 1975 Nowlin, W.D. Jr., T. Whitworth III, L.I. Gordon, and G.C. Anderson. Aboard R/V MELVILLE during '.'Oceanographic Station Data Collected Oceanography, Texas A&M FDRAKE 75." Reference 77-2-D, Department of University. Department of Oceanography Technical University of Washington. 1960. Report No. 66, U.W. Ref. 60-18. 21 APPENDIX Table A. Corrected, observed hydrographic and chemical data, Station 21, FDRAKE 76, R/V THOMPSON. Table B. Corrected, observed hydrographic and chemical data, Station 2817, Expedition 24-I, PROFESSOR VIESE. 22 Corrected, observed hydrographic and chemical data, Station 21, FDRAKE 76, R/V THOMPSON. Table A. Depth (m) 96 106 120 144 -0.61 0.25 02 (mQ/ 33.837 34.100 34.130 34.167 34.276 7.91 5.01 7.29 7.05 6.74 5.90 PO 4 Si (PM/9-) (uM/Q) 1.58 2.12 2.19 2.23 14.5 44.6 47.3 50.3 2.31 58.1 2.39 2.42 2.43 2.42 2.42 67.9 72.5 75.5 78.2 80.6 82.4 83.4 84.4 86.0 87.4 291 1.867 1.848 34.399 34.481 34.509 34.558 34.594 340 389 439 489 539 1.905 1.89 1.906 1.898 1.89 34.623 34.638 34.656 34.674 34.690 4.15 4.16 4.18 4.24 4.19 2.40 2.37 2.33 590 667 A 763 A 1.875 1.85 859 A 955 A 1.72 1.63 34.702 34.712 34.722 34.728 34.732 4.25 4.23 4.35 4.36 4.44 2.26 2.28 2.23 2.24 2.21 88.2 88.4 89.6 92.7 95.8 34.731 4.51 34.727 34.724 34.716 34.712 4.58 4.66 4.79 2.23 2.25 2.25 2.27 2.28 103.0 108.5 113.8 119.0 123.6 34.707 34.706 34.703 34.705 34.703 4.82 4.82 4.87 4.89 4.90 2.28 2.29 2.28 2.28 2.29 128.0 131.7 133.0 133.8 134.8 34.697 34.698 34.697 4.93 4.95 4.92 2.31 133.7 133.5 133.7 1.12 1.60 169 193 217 242 1.61 1.81 1191 1432 1669 1910 2153 A A A A A 1.429 2390 2630 2873 3115 3358 A A A A A 0.65 0.60 0.53 0.48 0.48 3552 A 3650 A 3699 A B) (°/°°) 2.03 -0.95 -0.82 1 A) S T (°C) 1.251 1.082 0.85 B ---- 0.436 0.439 0.435 B 4.50 4.41 4.21 4.22 4.71 2.31 2.27 2.32 2.30 Cast I. The temperature, listed to Both protected thermometers malfunctioned. inferred from the pressure the nearest five (.05) hundredths, was thermometer, wire depth and L-Z curve. 23 Corrected, observed hydrographic and chemical Expedition 24-I, PROFESSOR VIESE. Table B. 0 PO44 data, Station 2817, Si Depth T Cl (m) (°C) (°/°°) (m11) (uM/1) (uM/1) 26 52 79 1.99 2.00 1.85 1.59 -1.09 18.725 18.710 18.723 18.724 18.832 ---7.84 7.84 7.82 7.65 2.23 2.23 2.35 2.24 2.30 14.27 13.94 13.30 14.16 105 -0.71 18.888 19.032 19.094 6.84 5.78 5.04 4.59 4.29 2.23 2.26 2.43 2.42 2.43 47.39 4.14 ---4.15 4.26 2.43 ---2.19 2.43 2.45 79.81 1.63 1.55 1.65 1.62 3.13 97.27 91.97 0 11 131 157 208 261 0.27 0.98 1.53 1.77 19.117 19.127 314 366 419 472 524 1.81 1.84 1.88 1.90 1.89 19.140 19.155 19.170 19.177 19.182 636 1.82 1.82 1.76 1.68 1.63 19.197 19.204 19.225 19.226 19.224 4.27 4.25 4.32 4.35 1.39 1.18 1.02 0.85 0.67 19.222 19.215 19.223 19.215 19.208 4.56 0.53 0.50 0.46 0.45 0.40 19.197 19.196 19.194 19.188 19.191 4.96 4.94 4.89 4.86 5.04 661 756 848 935 1207 1481 1755 2029 2288 2546 2804 3061 3318 3575 4.21 4.51 4.61 4.70 4.85 5.51 2.24 2.40 4.12 2.24 2.26 3.25 3.74 2.35 2.21 2.28 40.28 65.91 54.41 85.43 91.35 ----- 62.36 93.53 90.10 98.21 101.9 99.77 96.96 112.5 101.6 126.5 126.5 135.9 104.1 143.7 136.8 139.3