PLEASE DO NOT REMOVE FROM FILES FORESTRY ISSUES IN URBAN AMERICA Proceedings, 1974 National Convention Society of American Foresters New York City - September 22-26 Copyright 1975 Society of American Foresters IMPACT OF FOREST FERTILIZATION ON WATER QUALITY IN THE DOUGLAS-FIR REGION -- A SUMMARY OF MONITORING STUDIES Duane G. Moore Pacific Northwest Forest and Range Experiment Station U. S. Forest Service Corvallis, Oregon 97331 Forest fertilization on a commercial scale began in the Pacific Northwest in 1965 with the aerial fertilization of a 607-hectare unit in western Oregon. Under pressures of increasing demands for wood fiber and predictions of timber supply deficits in the near future, this management practice has received increased emphasis and has grown rapidly. Between 1965 and the end of 1969, approximately 36,400 hectares were fertilized in western Oregon and Washington. By 1974 this figure increased to over 285,000 hectares. One company alone fertilized 76,900 hectares in 1973 (Hansen 1974), and the annual rate of fertilization for the region is approaching 120,000 hectares. Research on forest nutrition was initiated in the Douglas-fir region in 1949 (Gessel 1969). Laboratory studies first indicated and field trials later confirmed that low nitrogen availability is the most common growth-limiting nutrient deficiency in this area. In fact, major commercial coniferous timber types in the region have responded only to nitrogen. Urea fertilizer formulated as a granule contains 46 percent nitrogen and is physically well adapted to aerial application. Other nitrogen fertilizers are available and have been used to a limited extent, but logistical problems of transporting large tonnages into mountainous terrain will continue to restrict their use. Coincident with the development of forest fertilization as a management tool, there has been a growing and equally important concern over the possibility of detrimental effects on water quality. Although the total forest acreage fertilized to date is relatively small and widely scattered, application of 168 or 224 kilograms of nitrogen per hectare to all or a large part of a forested watershed may increase the amount of this nutrient entering surface streams. Recognizing this potential hazard to water quality, Federal and State agencies and private industry have conducted a number of monitoring studies throughout the Douglas-fir region over the past 6 years. Figure 1 shows the location of 22 water quality monitoring studies that have been conducted in western Oregon and Washington. In addition, two fertilized watersheds were monitored in eastern Washington (Klock 1971), three studies were conducted in northern Idaho (Loewenstein et al. 1973), and two fertilized watersheds were monitored in the spruce-hemlock region of southeast Alaska (Meehan et al. 1974). These monitoring studies cover a wide range of soils, climate, and vegetative species. Soil parent materials range from recent alluvium, glacial deposits, and volcanic tuffs, breccias, pumice, and ash to older sedimentary, metamorphic, and basaltic rock. Soils range from deep, well developed profiles to those that are young, shallow, and skeletal (largely undifferentiated and with more than 35 percent coarse fragments). Climate varies considerably among study sites but is generally characterized by high winter rainfall, moderate temperatures, and dry summers. Douglas-fir, ranging in age from 1P years to mature old growth, is the dominant overstory species in the fertilized stands. Density of both overstory and understory vegetation varies greatly. Most of the stand types in the Douglas-fir region that are 209 Jirnmycomelately Creek r" Crest of the Cascade Range Trapper Creek I Tahuya River Mill Creek Waddell Creek Thrash Creek Skookumchuck River Elochoman River Newaukum River Jackson Creek Fairchilds Creek Pat Creek Turner Creek Roaring Creek Canyon Creek t I Trout Creek I I Dollar Creek I Crabtree Creek Ne/son Creek I 1 Row River Coyote Creek Spencer Creek Figure I. Location of forest fertilization - water quality monitoring studies in western Oregon and Washington. 210 likely to be fertilized are represented. Other variables represented by the monitoring studies include stream size, total area treated, proportion of a watershed fertilized, and rate of application. Affected surface waters range from the small headwater streams of individual forested watersheds to much larger third and fourth order streams with drainage basins of several square kilometers. Total area fertilized in eaeh of the monitoring studies varied from less than 40 to 3,100 hectares. Entire small watersheds were fertilized in five studies, but in most projects the treated area represented only a small percentage of the drainage basin above the stream sampling station. Depending on the organization conducting the fertilization project, the usual rate of application is 370 or 493 kilograms urea per hectare (168 or 224 kg N per hectare). Two watersheds in eastern Washington were fertilized at 56 kilograms nitrogen per hectare in an experimental erosion control seeding program following a severe burn by wildfire (Klock 1971). The data obtained from each monitoring study are directly influenced by time and rate of application, age and density of stand, soil conditions, climate, stream characteristics, and intensity and duration of that particular sampling program. In addition, the monitoring studies have been conducted by several different agencies and companies, and the samples analyzed in different laboratories. Therefore, the results obtained are peculiar to the specific conditions and characteristics of a given project and cannot be extrapolated individually to other field situations. However, we now have enough individual studies to provide a broad base for drawing some conclusions. Rather than attempt to present all of the data, results from selected monitoring studies will be used to illustrate the kind of data obtained. Pertinent information from all of the studies will then be summarized, and conclusions will be drawn from an interpretation of the combined results. Field studies were initiated at one site in the fall of 1969 to determine how much fertilizer nitrogen, if any, entered streamflow as a result of The site selected included four small watersheds fertilization with urea. at the head of Coyote Creek in the South Umpqua Experimental Forest in southwestern Oregon. Background hydrological data for 6 years were already available, and it was possible to include fertilization as an additional treatment without confounding other studies already in progress. These watersheds are typical of the mixed conifer area of southwest Oregon. Each is T9uipped with a 120-degree, V-notch weir gaging station and a Fischer-Porter—I digital recorder. A standard precipitation gage has been maintained at Watershed 2 since 1960. In August and September of 1969, automatic proportional stream samplers were installed on each watershed, and sampling started on October 1, 1969. Detailed chemical analysis of streamwater samples from all four watersheds has been carried out since that time. In March of 1970, Watershed 2 was fertilized by aerial application of urea at the rate of 224 kilograms nitrogen per hectare. The boundaries and gaging station were marked with yellow plastic bunting, and fertilizer traps were placed along transects across the watershed at three elevations and along 1/ Use of brand names does not imply endorsement by the U.S. Department of Agriculture. 211 the boundaries to estimate uniformity of application. Urea was applied by helicopter using a large, forest-granule formulation to avoid dust drift to adjacent watersheds. The drainage channel in Watershed 2 is small, and only very little surface water is exposed. Therefore, the total watershed was fertilized, and no attempt was made to leave an untreated buffer zone. Water quality was monitored by collecting grab samples at the stream gaging stations on treated Watershed 2 and on control Watershed 4 during and after fertilization. Following the close interval sampling during the first 6 weeks after treatment, monitoring continued through use of the automatic proportional samplers. Pretreatment levels of urea-, ammonia-, and nitrate-nitrogen were 0.006, 0.005, and 0.002 parts per million nitrogen, respectively. Fertilization of this 68-hectare forested watershed did temporarily increase these nitrogen concentrations but did not result in toxic levels of any component in streams draining the treated area. Concentrations of nitrogen in selected samples collected over the first 15 weeks following treatment are shown in table 1. Urea concentrations increased slowly and reached a maximum of 1.39 parts per million urea-N, 48 hours after application started. Ammonia-N increased to levels only slightly above background concentrations and never reached 0.10 part per million. Nitrate-N reached a peak concentration of 0.168 part per million in 72 hours, was 0.140 part per million at the end of 2 weeks, and had returned to near pretreatment levels after 9 weeks. Nitrite-N was never detected except as a trace. Table 1. Concentrations of fertilizer nitrogen in selected water samples collected at Watershed 2, South Umpqua Experimental Forvit, following application of 224 kilograms urea-N per hectare.-1 Date Time NH3 -N Urea -N 2.7 NO -N 3 Total .002 .040 .069 .067 .107 .150 .168 .117 .092 .140 .030 .021 .022 .004 .002 .006 .010 .493 .318 .272 1.544 .792 .685 .228 .185.183 .040 .031 .035 .004 .002 .006 ppm 3/25 3/26 II II 3/27 II II 3/28 4/1 4/8 4/15 4/22 5/6 5/27 6/17 7/8 1/ 21 0800 0815 1230• 2025 0805 1640 2005 0805 _, - .001 .016 .012 .034 .048 .036 .029 .036 .016 .015 .010 .010 .013 0 0 0 .007 .437 .232 .171 1.389 .606 .488 .075 .007 .028 0 0 0 0 0 0 From Norris and Moore (1971). Includes both ionized (NH + ) and un-ionized (NH ) ammonia-nitrogen. 4 3 All losses of applied nitrogen as urea occurred during the first 3 weeks. 212 Losses in the form of ammonia-N, even though small, continued for 6 weeks. During the first 9 weeks after application, loss of applied nitrogen amounted to only 1.81 kilograms for the entire watershed (table 2). Table 2. Nitrogen lost from treated Watershed 2 and untreated Watershed 4, South Umpqua Experimental Forest, during the firiC 9 weeks after application of 224 kilograms urea-N per hectare.-1 Unit Urea - N NH3 - N NO -N 3 Total Kilograms N Watershed 2 Watershed 4 Net loss Percent of total loss 1/ 0.65 0.02 0.63 34.75 1.01 0.05 0.96 53.00 0.28 0.06 0.22 12.25 1.94 0.13 1.81 ,100.00 From Moore (1971). Low streamflow due to limited precipitation throughout the summer and fall months resulted in essentially no loss of applied nitrogen during the next 24 weeks. Early storm activity in November brought the soil moisture level back to maximum storage capacity, and in December the nitrate-N concentration in samples for the fertilized watershed reached a second peak of 0.177 part per million (fig. 2). Both streamflow and nitrate-N levels remained high through December and January, resulting in the loss of an additional 23.8 kilograms ap!-. plied nitrogen. This is 92 percent of the total amount of fertilizer nitrogen which was lost during the first year. Total loss of applied nitrogen from the fertilized watershed (68 hectares) during the fitt year amounted to 25.85 kilograms or 0.38 kilogram of'nitrogen per hectare (table 3). Over the same period the total amount of soluble inorganic nitrogen lost from the control watershed (49 hectares) was 2.15 kilograms, or 0.04 kilogram nitrogen per hectare. Data for soluble organic nitrogen, total phosphorus, silica, and exchangeable -cation content.of the stream samples, including sodium, potassium, calcium, magnesium, iron, manganese, and aluminum, indicate that there was no apparent effect of nitrogen fertilization on loss of native soil nitrogen or other plant nutrients. Movement may have occurred in the soil profile, but there was no measurable change in streamwater quality. Table 3. Nitrogen lost from treated Watershed 2 and untreated Watershed 4, South Umpqua Experimental Forest, during the fir year after application of 224 kilograms urea-N per hectare.- I Unit Urea-N NH -N 3 NO -N 3 Total Kilograms N Watershed 2 Watershed 4 Net loss Percent of total loss 1/ 0.65 0.02 0.63 2.44 0.28 0.06 0.22 0.86 From Moore (1971). 213 27.09 2.07 25.02 96.70 28.03 2.15 25.88 100.00 0. 2 - A. 1. 5 0. 1 - 1. 0 mg/I O. 5 0.0 „ „ MAR MAY JUL SEP NOV JAN MAR 0. 0 25 29 2 6 10 14 MARCH APRIL 1970 1970 1971 TOW AFTER APPLICATION Figure 2. Fertilization of a 68-hectare watershed with 224 kilograms urea-nitrogen per hectare in March 1970. A. Immediate effect on water quality; B. Effect on nitrate-nitrogen concentration in streamflow for 1 year following fertilization (Fredriksen et al. 1974). Initial losses of applied nitrogen were largely due to direct application of urea fertilizer to the drainage channel. These losses were measured first as an increase in urea-nitrogen and then as a small increase in ammonia-nitrogen, the latter as a result of hydrolysis of urea applied to open water. Nitrate-nitrogen entering the stream shortly after application is probably leached from the soil immediately adjacent to the stream channel- Approximately half of the applied nitrogen which was lost during the first 9 weeks after application was due to direct application, and half entered the stream as nitrate-nitrogen. However, all of the applied nitrogen lost Auring this 9-week period amounted to only 7 percent of the total loss which,c6curred.. over the first year. 7. trt i • • High streamflow coupled with the second peak in nitrate-nitrogen levels during the winter storm period accounted for 92 percent of the tot 1pss. In February and March 1971, streamflow remained high, but most ok..0e.mmbile nitrogen had already been lost, and nitrate-nitrogen concentratipngs4a4,re-.. turned to near normal. Monitoring of streamwater samples from t0e,pr4teft, and control watersheds continued through the second and third years following fertilization, but there has been no further loss of applied nitrogen. . Similar data have been obtained in each of the monitoring stuAjgs dpn7:ducted throughout the Douglas-fir region. However, most data arg:pqc,aq complete as in the Coyote Creek study. Peak concentrations of urea-,..ammonia and nitrate-nitrogen measured at each site during the monitoring,peqo4 lowing fertilization are given in table 4. The monitoring studies are , identified by the name of the stream sampled, and information on time .04,,re of application and size of area fertilized are shown. The length of. 0p.monitorand ing period varied from a few weeks following treatment to . 6 or 7 in a few studies monitoring continued at least a full year. Most sampling continued until the forms of nitrogen being measured had decreased,Fp near pretreatment levels. Increases in the concentration of urea-N ranged from very low to a high of 44.4 parts per million. These increases are due almost entirely to direct application to surface water, and the peak concentration reached is. proportional to the amount of open surface water in the treated unit. Buffer strips were * not left along the main streams in a few of the earlier fertilization projects, and peak concentrations reached 15 to 20 parts per million urea-N. In other projects it is not feasible to leave buffer strips because the stream channel is very small, and it is not possible for . the helicopter pilot to avoid direct application. The Coyote Creek site is an example of the latter situation, but the amount of urea entering the stream was small because the amount of exposed stream channel is small. The high peak concentrations of urea-N measured in Dollar Creek and Fairchilds Creek are due to greatly expanded stream drainage systems caused by spring runoff of snowmelt. A buffer strip was left along Fairchilds Creek, but the expanded system of small feeder streams still resulted in direct application of fertilizer to a much larger area of surface water than would normally occur. Forested watersheds of western Oregon and Washington are drained by a relatively dense network of small tributary streams. Many of these small tributaries are only a foot or 2 wide and cannot be identified from the air. Thus, even when buffer strips of 30 to 90 meters are left along main streams and tributaries, some direct application of fertilizer to water surfaces still occurs. Peak concentrations of urea-N in the range of 0.5 to 5.0 parts per million can be expected under these conditions. The high concentration .215 ■ Table 4. Study site Peak concentrations of nitrogen in stream samples following forest,/ fertilization with urea in Alaska, Idaho, Oregon, and Washington.Rate of application Date of application Kg N/ha Burns Creek Canyon Creek Coyote Creek Crabtree Creek Dollar Creek Elochoman River Fairchilds Creek Falls Creek Jackson Creek Jimmycomelately Creek McCree Creek Mica Creek Mill Creek Nelson Creek Newaukum River Pat Creek Quartz Creek Roaring Creek Row River Skookumchuck River Spencer Creek Tahuya River Thrash Creek Three Lakes Trapper Creek Trout Creek Turner Creek Waddell Creek Wishbone Creek 562. / 224 224 224 224 224 224 213 168 224 56 224 224 224 168 224 224 224 168 168 224 224 6/ 224213 224 224 224 224 224 Treated area Urea-N 2/ NH3-N- NO -N 3 Hectares Oct. 1970 Oct. 1969 Mar. 1970 May 1969 Apr. 1971 Nov. 1969 Apr. 1972 May 1970 May 1969 Apr. 1970 Oct. 1970 Sept.1972 Dec. 1969 Apr. 1970 Sept.1971 Apr. 1972 May 1972 Mar. 1972 Oct. 1972 Sept.1969 Nov. 1972 Oct. 1972 May 1974 May 1970 Apr. 1970 Mar. 1968 Mar. 1972 Dec. 1969 May 1972 562 1346 68 230 34 297 192 263 95 49 513 47 228 38 2463 243 51 267 2630 191 3108 1620 121 69 64 648 352 600 46 ppm 0 15.20 1.39 24.00 44.40 19.10 23.40 n.d. 0.09 0.71 0.62 0.30 0.68 8.60 0.26 3.26 1.75 0.76 0.13 2.63 0.37 27.20 n.d. 0.70 14.00 4.36 2.48 0.30 0 4/ n.d.0.048 0.080 0.490 n.d. 0.280 1.28 0.044 0.040 0 0 0.12 0.32 0.008 0.079 trace 0.040 0.022 0.026 0.123 1.40 0.06 0.13 0.010 0.700 0.046 0.340 0 0 068 ' 0.80 0.177 0.25 0.13 4.00 0.828 1.67 0.116 0.042 0.210 0.28 1.32 2.10 0.438 0.388 0.70 0.210 0.044 0.0855/ 0.0051.83 1.88 2.36 0.121 0.160 0.243 0.99 0.28 ■•• 1/ Data included in the table have been provided by the following companies and State and Federal agencies: Crown Zellerbach Corporation, Camas, Washington; Weyerhaeuser Company, Centralia, Washington; Willamette Industries, Inc. Albany, Oregon; Forest, Wildlife and Range'Experiment Station, University of Idaho, Moscow, Idaho; Washington Department of Natural Resources, Olympia, WashingtOn; Bureau of Land Management, POrtland, Oregon; -Pacific Northwest Water Laboratory, Environmental Protection Agency, Corvallis, Oregon; and Pacific Northwest Forest and Range Experiment Station, Juneau, Alaska; Corvallis, Oregon; and Wenatchee, Washington. See figure 1 for site locations in Oregon and Washington. 2/ 3/ 4/ Includes both ionized (NH 4* ) and un-ionized (NH ) ammonia-nitrogen. 3 Applied as ammonium sulfate. n.d. = no data available. 5/ In this study there was no increase in NO -N. Concentration given is back3 ground level. 6/ Applied as ammonium nitrate. 216 of urea-N in Tahuya River resulted when the pilot could not identify the stream channel and fertilized the required buffer strip along the stream. The peak concentrations of urea-N, whether high or low, do not persist for more than a few hours. Concentrations characteristically reach a peak each day of application and then drop rapidly toward background. Within 3 to 5 days after fertilizer application is completed levels of urea-N in the stream have returned to pretreatment concentrations. Ammonia-nitrogen levels also increased as a result of direct application of urea fertilizer to open water. Urea is readily hydrolyzed to ammonia-N in the stream system. Urea applied to forest floor and soil surfaces will not reach the stream since it hydrolyzes rapidly to ammonium carbonate and is then held on cation exchange sites in the . soil and forest floor like any other ammonium salt. Ammonia-N concentrations in the stream are rapidly reduced through uptake by aquatic organisms and by adsorption on.stream sediments. Levels in the streams sampled did increase following fertilization, but ex. ceeded 0.10 part per million NH -N in only 10 of the projects monitored. The 3 peak concentration of ammonia-N exceeded 1.00 part per million in two studies, and in each case direct application of urea to the stream was noted. Lower ammonia-N levels occurring coincident with other high urea-N concentrations are attributed to lower temperatures at time of application. Peak concentrations did not persist for more than a few hours, although levels of ammonia-N slightly above background could be detected in some studies for up to 3 and 4 weeks. Peak concentrations of nitrate-N entering streams following forest fertilization ranged from no increase in Spencer Creek to a maximum of 4.00 parts per million in a tributary stream of the Elochoman River study. This high concentration of nitrate-N can be attributed at least in part to runoff from ad jacent agricultural and dairy lands caused by continued heavy rains -following fertilization. Because of the late fall date of application (Nov. 30 and Dec. 1, 1969), both urea hydrolysis and nitrification of ammonium would be markedly inhibited by ,cold, wet soils. In contrast, the lack of any increase in nitrate-N in Spencer Creek is attributed - to the Combined effects of a 90-meter buffer strip along the stream channel, the virtual lack of small feeder and tributary streams, and very low rainfall. The Spencer Creek site is located east of the Cascade Range. The concentration of nitrate-N in stream samples usually reaches a peak 2 to 4 days after spring application of fertilizer. Concentrations then in decrease but may remain above background for 6 to 8 weeks. Losses of applied nitrogen are very small because the maximum concentrations reached are generally below 1 part per million NO 3 -N and streamflow is rapidly decreasing with the onset of the dry summer season. About half of the applied nitrogen entering the stream during the first 30 days is from direct application and is measured as urea- and ammonia-N. The other half enters as nitrate. In early fertilization projects where buffer strips were either inadequate or not used, estimated total loss was between 2 and 3 percent of the applied nitrogen. However, in later projects where direct application to open surface water is minimized by buffer strips along the main streams and tributaries, measured losses are less than 0.5 percent. In monitoring studies where sampling has continued through the first winter following fertilization, additional peaks in the concentration of nitrate-N have been measured. These peaks usually coincide with more intense winter storms, and the concentration drops sharply between storms. Maximum 217 concentrations measured are still low and tend to decrease with each successive storm. Patterns of nitrate-N loss to streams following early fall application of fertilizer (Sept.-Oct.) are similar to those following spring application. However, peak concentrations measured during winter storms may not be as high because of less time during warm weather for conversion of applied nitrogen to the nitrate form. The initial peak in nitrate-N concentration also occurs following fertilizer application in November and December. Subsequent peaks during winter storms are similar to those in streams draining untreated areas. However, additional losses as nitrate-N may occur during the following winter. Maximum concentrations of urea-, ammonia-, nitrite-, and nitrate-N measured in streams following forest fertilization have not approached levels considered unacceptable in public water supplies (NAS-NAE 1972). Samples were collected at the intake for domestic water supplies in two of the studies included in table 4. Recommended standards for nitrate-N in drinking water set the maximum permissible level at 10 parts per million NO -N(highest peak level recorded was 4.00 parts per million). Nitrite-N has not been detected at levels above a trace. Un-ionized ammonia is known to be harmful to fish, and the recommended permissible level is 0.02 part per +million (NAS-NAE 1972). At the maximum level of total inorganic ammonia-N (NH, plus NH 3 ) measured, the concentration of un-ionized ammonia is less than 0.002 part per million (Trussell 1972). There is no recommended standard for urea since this form of nitrogen is non-toxic at concentrations well above 1,000 parts per million. Total amounts of applied nitrogen entering streams draining treated areas are relatively small and should not have any measurable impact on eutrophication in downstream impoundments. Samples collected at points 3 to 5 miles below fertilized units show that the peak concentrations are rapidly diluted with downstream movement. Both the amount of applied nitrogen entering streams and the maximum concentrations measured can be kept at minimum levels by ensuring that adequate buffer strips are left along main streams and larger tributaries. In addition, fertilizer should not be applied to forested watersheds when the stream drainage system is greatly expanded by spring snowmelt or heavy storm activity. Following these few guidelines will ensure that forest fertilization is an environmentally acceptable practice. LITERATURE CITED Fredriksen, R. L., D. G. Moore, and L. A. Norris. 1974. The impact of timber harvest, fertilization, and herbicide treatment on streamwater quality in western Oregon and Washington. In: Bernier, B., and C. H. Wingett (eds.), Soils and Forest Management, Proc. Fourth North American Forest Soils Conference. Quebec, Canada. (In press.) Gessel, Stanley P. 1969. Introduction to forest fertilization in North America. For. Ind. 96(10):26-28. Hansen, Robert A. 1974. Operational forest fertilization: results, problems, and land-use planning implications. In: Proc. 1973 National Convention, Society of American Foresters, pp. 221-225. Washington, D.C. Klock, G. O. 1971. Streamflow nitrogen loss following forest erosion 218 control fertilization. USDA For. Serv. Res. Note PNW-169, 9 pp., illus. Pac. Northwest For. & Range Exp. Stn., Portland, Oreg. Loewenstein, H., F. H. Pitkin, and D. C. Scanlin. 1973. Nitrogen compounds in streams as affected by aerial fertilization of northern Idaho forests. Univ. of Idaho, College of Forestry, Wildlife and Range Sciences, Station Paper No. 12, 11 pp., illus. Moscow, Idaho. Meehan, William R., Frederick B. Lotspeich, and Ernst W. Mueller. 1974. Effects of forest fertilization on two southeast Alaska streams. Environ. Qual., vol. 3. (In press.) J. Moore, Duane G. 1971. Fertilization and water quality. In: Proc. 1971 Annual Meeting Western Reforestation Coordinating Committee, pp. 2831. Western Forestry and Conservation Assoc., Portland, Oreg. National Academy of Sciences-National Academy of Engineering. 1973. Water quality criteria, 1972. Ecological Research Series EPA-R3-73-033, 594 pp., illus. U. S. Govt. Printing Off., Washington, D.C. Norris, Logan A., and Duane G. Moore. 1971. The entry and fate of forest chemicals in streams. In: Krygier, J. T., and J. D. Hall (eds.), Forest Land Uses and Stream Environment Symp. Proc., pp. 138-158. Oreg. State Univ., Corvallis, Oreg. Trussell, R. P. 1972. The percent un-ionized ammonia in aqueous ammonia solutions at different pH levels and temperatures. J. Fish. Res. Bd. Canada 29:1505-1507. 219