Influence of Natural Factors on the Quality of Midwestern Streams and Rivers

By Stephen D. Porter, Mitchell A. Harris, and Stephen J. Kalkhoff

U.S. GEOLOGICAL SURVEY

Water-Resources Investigations Report 00–4288

Denver, Colorado 2001 U.S. DEPARTMENT OF THE INTERIOR GALE A. NORTON, Secretary

U.S. GEOLOGICAL SURVEY Charles G. Groat, Director

The use of firm, trade, and brand names in this report is for identification purposes only and does not constitute endorsement by the U.S. Geological Survey.

For additional information write to: Copies of this report can be purchased from:

Regional Hydrologist U.S. Geological Survey U.S. Geological Survey Information Services Box 25046, Mail Stop 406 Box 25286 Denver Federal Center Federal Center Denver, CO 80225–0046 Denver, CO 80225 Influence of Natural Factors on the Quality of Midwestern Streams and Rivers

—Stephen D. Porter, Mitchell A. Harris, and Stephen J. Kalkhoff

Streams flowing through cropland in the Midwestern Corn Belt differ considerably in their chemical and ecological characteristics, even though agricultural land use is highly intensive throughout the entire region. These differences likely are attributable to differences in riparian vegetation, soil properties, and hydrology. This conclusion is based on results from a study of the upper Contents Midwest region conducted during seasonally low-flow conditions in Introduction...... 1 August 1997 by the U.S. Geological Natural Factors Influence Chemical Indicators Survey (USGS) National Water of Water Quality ...... 4 Quality Assessment (NAWQA) Natural Factors Influence Biological Indicators Program. This report summarizes of Water Quality ...... 8 significant results from the study and presents some implications for the Implications for Water-Quality Monitoring and design and interpretation of water- Assessment ...... 12 quality monitoring and assessment References Cited ...... 13 studies based on these results.

Introduction The U.S. Geological Survey percentage of trees in buffer zones (USGS) National Water-Quality along riparian segments. Census data The Midwestern Corn Belt is Assessment (NAWQA) Program from the U.S. Department of Agri- one of the most productive agricul- conducted a water-quality study in culture (USDA) for 1997 were used tural regions in the world. Nearly the upper Midwest region during in this study to quantify cropland, 80 percent of the Nation’s corn and seasonally low-flow conditions in livestock, and chemical application soybeans is grown in the region. August 1997. Chemical and biologi- in each basin. A complete descrip- More than 6 million metric tons cal samples were collected from 70 tion of the study design, methods, of nitrogen fertilizer and over streams and rivers in the Minnesota, and data can be found in Sorenson 100,000 metric tons of pesticides are Wapsipinicon, Cedar, Iowa, Skunk, and others (1999). This report sum- applied to cropland in the Midwest and lower Illinois River basins of marizes significant results from the annually (Goolsby and others, southern Minnesota, eastern Iowa, study and presents some implica- 1993). The 1997 growing season and western Illinois (fig. 1). Sam- tions for the design and interpreta- produced the largest soybean crop pling locations were chosen to tion of water-quality monitoring and and second-largest corn crop represent a range of soil-drainage assessment studies based on these recorded in the Midwest. conditions in stream basins and the results. Area of map

95° 90°

3.5

50° 3.0 WESTERN LAKE

Minnesota 2.5 River Basin BASE (STATSGO) SCORE

MINNESOTA STATE SOIL GEOGRAPHIC DATA 2.0

Lake Western Till Plains Dissected Southern extent of Till Plains Wisconsinan PHYSIOGRAPHIC SECTION Minneapolis glacial advance M in ne so ta Ri ver M is s i s s i p p i WISCONSIN R

45° i v

Eastern Iowa W a Basins p s i p in ic on R iv Chicago er ILLINOIS Cedar River IOWA Io wa River

DISSECTED TILL PLAINS S kunk R.

EXPLANATION

Soil hydrologic group classification r of stream basins e v Ri is Moderately well drained soils o n i l 40° l Poorly drained soils I Very poorly drained soils TILL MISSOURI PLAINS 0 100 MILES

0 100 KILOMETERS Lower Illinois River Basin

Figure 1. Soil drainage in basins for the 70 streams and rivers sampled during August 1997 varied from very poor in the Western Lake physiographic section to moderately well drained in the Dissected Till Plains section.

2 best management practice. These “riparian buffer zones” are thought to intercept runoff of sediment and chemicals from fields, promote bank stability, and provide shading and habitat for aquatic life (Osborne and Kovacic, 1993). For this study, riparian vegetation density, referred to hereinafter as “tree density,” was determined at two spatial scales: segment and reach. At the segment scale, tree density was classified based on the percentage of trees in a 100-meter-wide zone on each side of the stream for a distance of 2 to 3 miles upstream from the sampling location (Sorenson and others, 1999). Tree density was considered Differences in landscape terrain depth to water table, and related “low” if 35 percent or less of the and soil characteristics among drain- characteristics. Stream basins with area was forested and “high” if more age basins in the upper Midwest STATSGO scores of 2.6 or less than 35 percent was forested. At region are associated with historical were characterized as having “mod- the reach scale (refer to Fitzpatrick patterns of glacial advance and erately well drained” soils; basins and others, 1998), tree density retreat. The spatial distribution of with scores exceeding 2.6, “poorly was the percentage of the land sur- soils corresponds generally with drained;” and basins with scores face covered by trees in replicate physiographic sections (Fenneman exceeding 3.0, “very poorly 20-square-meter vegetation plots and Johnson, 1946) in the upper drained” (fig. 1). along transects at the sampling loca- Midwest region (fig. 1). For example, soils generally are moderately Resource agencies such as the tion (Sorenson and others, 1999). well drained in the Dissected Till USDA encourage maintenance of Tree density was considered low if Plains section of eastern Iowa, and strips of trees or grass between the percentage was less than 5 and the proportion of stream water that is agricultural fields and streams as a high if 5 percent or more. derived from ground-water inflow is substantially greater than in adjacent sections (Winter and others, 1998). In contrast, soils with poor drainage, high runoff potential, and relatively higher organic nitrogen content are found in the Western Lake section of southern Minnesota and northern Iowa. Agricultural drainage systems (tiles and ditches) are prevalent throughout the study area. The USDA State Soil Geographic data base (STATSGO) was used to calcu- late Soil Hydrologic Group scores for each basin (Sorenson and others, 1999). Soil Hydrologic Groups are classified on the basis of drainage, runoff potential, permeability,

3 Natural Factors Influence 10

Chemical Indicators of Physiographic Section Water Quality Till Plains Dissected Till Plains Nutrient concentrations in water Western Lake samples from Midwestern streams 1 and rivers were not correlated with measures of agricultural intensity, such as the percentage of cropland, rates of fertilizer application, or the 0.1 number of livestock in the basin. About one-half of the variance in dissolved nutrient concentrations among sites can be explained by runoff relations with rainfall. Figure 2 0.01 shows that concentrations of dissolved nitrogen increased with the amount of streamflow relative to

basin drainage area (unit basin BASIN WATER YIELD, IN CUBIC FEET PER SECOND SQUARE MILE 0.001 0.04 0.1 1 10 20 water yield). Differences in unit DISSOLVED NITRITE PLUS NITRATE, IN MILLIGRAMS PER LITER basin yield among streams in the Till Plains, Dissected Till Plains, Figure 2. Dissolved nitrogen concentrations increased with unit water yield and Western Lake physiographic from stream basins in August 1997, which was associated with regional differences in recent runoff. sections (fig. 2) are associated with differences in the amount of rainfall during the month preceding the study (fig. 3). In streams with low tree density, average dissolved orthophosphorus concentrations

20 + Standard Error Surface runoff to stream Mean – Standard Error 15

Tile drain discharge to stream

10 IN INCHES

5 Ground water to stream ESTIMATED RAINFALL DURING JULY 1997, 0 Till Dissected Western Plains Till Plains Lake PHYSIOGRAPHIC SECTION Figure 3. Average rainfall in stream basins differed among physiographic sections in the upper Midwest region.

4 were smaller in basins with poorly drained soils than those with moderately well drained soils (fig. 4). Moderately well drained soils Poorly drained soils Although the application of nutrient 0.20 fertilizers to cropland is considerable + Standard Error in all stream basins, nutrient concen- Mean – Standard Error trations vary among streams because 0.15 of differences in natural factors such as riparian tree density, soil properties, and unit basin water yield. Average concentrations of sus- 0.10 pended sediment were more than 40 percent smaller in streams with high tree density in riparian seg- 0.05 ments, but only in basins with poorly

drained soils (fig. 5). Riparian zones AVERAGE DISSOLVED ORTHOPHOSPHATE CONCENTRATION, IN MILLIGRAMS PER LITER with high tree density appeared to 0 be effective in reducing runoff High Low High Low of agricultural soils into streams, TREE DENSITY particularly in the Western Lake Figure 4. Average dissolved orthophosphorus concentrations section where soil drainage is very were smallest in streams with low tree density in basins with poorly poor (fig. 1) and antecedent rainfall drained soils. was high (fig. 3). Beneficial effects of buffers on nutrient and sediment concentrations in streams were 300 less pronounced in sections where + Standard Error rainfall was relatively lower, soils Mean were moderately well drained, and – Standard Error ground-water discharge contributed 200 substantially to total streamflow.

100 AVERAGE TOTAL SUSPENDED SEDIMENT CONCENTRATION, IN MILLIGRAMS PER LITER 0 High Low TREE DENSITY IN BASINS WITH POORLY DRAINED SOILS Figure 5. Average concentrations of suspended sediment were smaller in streams with high riparian-tree density in buffer segments, but only in basins with poorly drained soils.

5 0.8 80 Concentrations of herbicides in + Standard Error Midwestern streams were propor- Mean – Standard Error tional to herbicide use (fig. 6; 0.6 60 Sorenson and others, 1999); however, relations between parent and degradation compounds differed

0.4 40 with respect to riparian tree density and soil conditions. Average concentrations of triazine herbicides (for

MICROGRAMS PER LITER example, atrazine and cyanazine) 0.2 20 PER SQUARE KILOMETER were larger in streams with high tree density in riparian segments (fig. 7). AVERAGE ATRAZINE CONCENTRATION, IN ESTIMATED ATRAZINE USE, IN KILOGRAMS 0 0 However, the ratio of triazine parent Till Plains Dissected Western Lake to degradation compounds was larg- Till Plains PHYSIOGRAPHIC SECTION est in streams with low tree density Figure 6. Atrazine application rates varied among physiographic sections; stream concentrations were proportional to herbicide use.

Moderately well drained soils Poorly drained soils 1.5 + Standard Error Mean – Standard Error

1.0

0.5 IN MICROGRAMS PER LITER AVERAGE CONCENTRATION OF TRIAZINE PARENT COMPOUNDS, 0 HighLow High Low TREE DENSITY

Moderately well drained soils Poorly drained soils 3 + Standard Error Mean – Standard Error

1 COMPOUNDS TO TRIAZINE RATIO OF TRIAZINE PARENT DEGRADATION COMPOUNDS

0 HighLow High Low TREE DENSITY Figure 7. Average concentrations of triazine herbicides were larger in streams with high tree density in riparian segments. The ratio of parent to degradation compounds was largest in poorly drained basins with low tree density.

6 in basins with poorly drained soils Moderately well drained soils Poorly drained soils 0.3 (fig. 7), indicating slower rates of + Standard Error triazine herbicide degradation in Mean those streams. Atrazine and its – Standard Error degradation compounds (notably 0.2 hydroxyatrazine) are retained in poorly drained soils, such as those in the Western Lake section, because 0.1 of adsorption to clay or organic mat-

ter in soils (Lerch and others, 1998). IN MICROGRAMS PER LITER AVERAGE CONCENTRATION The use of atrazine on corn is lower in the Western Lake section (fig. 6), OF ACETANILIDE PARENT COMPOUNDS, 0 possibly because of potential carry- HighLow High Low TREE DENSITY over of atrazine in soils that could Moderately well drained soils Poorly drained soils adversely affect the growth of soy- 0.10 beans grown during the following + Standard Error Mean season. 0.08 – Standard Error Average concentrations of acetanilide herbicides (for example, 0.06 acetochlor and metolachlor) were also larger in streams with high 0.04 riparian tree density (fig. 8) and, consequently, greater shading. 0.02 DEGRADATION COMPOUNDS

Acetanilide parent-to-degradation COMPOUNDS TO ACETANILIDE RATIO OF ACETANILIDE PARENT compound ratios were larger in 0 streams with high riparian tree den- HighLow High Low sity, indicating that rates of degrada- TREE DENSITY tion are slower in shaded streams, Figure 8. Average concentrations of acetanilide herbicides, as well regardless of soil-drainage condi- as parent-to-degradation compound ratios, were larger in streams tions. Previous research has sug- with high tree density in riparian segments. gested that photolysis (degradation by sunlight) may be an important process by which herbicide parent compounds are broken down into compounds that are presumed to be less toxic (Kolpin and Kalkhoff, 1993; Torrents and others, 1997). Rates of acetanilide herbicide degradation may be inhibited by canopy shading in streams with high riparian tree density. However, average acetanilide parent-to-degradation compound ratios (less than 0.04; fig. 8) were an order of magnitude less than those for triazine herbicides (greater than 0.4; fig. 7), possibly indicating that acetanilide herbicides degrade more rapidly than triazine herbicides.

7 Moderately well drained soils Poorly drained soils ples, as well as maximum rates of 75 max + Standard Error stream productivity (P ) and respi-

Mean ration (Rmax) (Porter, 2000; Soren-

a – Standard Error son and others, 1999). Concentrations of dissolved nitrogen 50 in Midwestern streams decreased with increases in seston CHLa and Pmax (fig. 10), and mean dissolved orthophosphate concentrations were

25 smallest in poorly shaded streams in basins with poorly drained soils (fig. 4) where seston CHLa (fig. 9)

AVERAGE SESTON CHLOROPHYLL- and Pmax were higher. Thus, shading

CONCENTRATION, IN MICROGRAMS PER LITER from tree cover in riparian segments 0 High Low High Low can influence nutrient concentrations TREE DENSITY in streams indirectly by reducing the growth of algal seston and rates of Figure 9. The relative abundance of algal seston was greater in streams with a low density of trees in the riparian segment. stream metabolism. Periphyton (algae attached to submerged tree branches and logs) abundance was largest in streams Natural Factors Influence concentrations, was largest in draining the Dissected Till Plains section (fig. 11), where water clarity Biological Indicators of streams with low tree density in Water Quality and riparian tree density generally basins with poorly drained soils were high (Porter, 2000), soils (fig. 9). In addition, seston CHLa were moderately well drained, Algal seston (phytoplankton; concentrations were positively corre- and ground-water contributions algae suspended in the water) lated with total organic nitrogen and to streamflow were appreciable can be an indicator of the amount carbon concentrations in water sam- (Winter and others, 1998). Streams of total nutrients and organic enrichment in Midwestern streams during low-flow conditions, such as those during August 1997. Organic enrichment from excessive algal production can increase rates of micro- bial respiration in streams, reducing dissolved oxygen (DO) to concentrations that are detrimen- tal for sensitive invertebrate and fish species. Rates of dissolved nutrient uptake by algae and other aquatic plants are closely associated with rates of primary production (Borchardt, 1996). Streams with low tree density in riparian segments contained relatively large growths of algal seston at levels considered indicative of eutrophica- tion (Porter, 2000). The relative abundance of algal seston, as mea- sured by chlorophyll-a (CHLa)

8 with above-average periphyton CHLa (but low seston CHLa) were productive; however, rates of stream respiration were relatively low (Porter, 2000). The composition of periphyton communities (fig. 12) varied from a dominance of diatoms, a desirable food source for invertebrates and fish, to blue-green algae, which are not consumed. The percentage of blue-green algae in the periphyton community increased with the concentration of total

100 5 Physiographic Section IN MILLIGRAMS PER LITER Till Plains

DISSOLVED NITRITE PLUS NITRATE, Dissected Till Plains 0 Western Lake 0 0.5 1.0 1.5 80 MAXIMUM PRODUCTIVITY, IN GRAMS O2 PER CUBIC METER PER HOUR Figure 10. Concentrations of dissolved nitrogen decreased with increasing stream productivity. 60

75 + Standard Error Mean 40 a – Standard Error PERCENTAGE OF DIATOMS

25 CONCENTRATION, IN

PERIPHYTON CHLOROPHYLL- 0 MILLIGRAMS PER SQUARE METER 0 20 40 60 80 100 PERCENTAGE OF BLUE-GREEN ALGAE 0 Till Dissected Western Plains Till Plains Lake Figure 12. Periphyton community structure varies from a predominance of PHYSIOGRAPHIC SECTION diatoms to blue-green algae in Midwestern streams and rivers. Figure 11. Average periphyton abundance was largest in streams of the Dissected Till Plains section.

9 triazine-herbicide compounds and was larger in streams with large amounts of algal seston (Porter, 2000). As expected, average periphyton biomass was smaller in streams with a large abundance of benthic invertebrates that consume algae (scrapers) (fig. 13).

40 + Standard Error Mean – Standard Error 30

20 PER SQUARE METER 10

100 PERIPHYTON ASH-FREE DRY MASS, IN GRAMS 0 Less 5% or than 5% greater Physiographic Section ABUNDANCE OF INVERTEBRATE Till Plains SCRAPERS Dissected Till Plains 80 Western Lake Figure 13. Average periphyton biomass was smaller in streams with a large abundance of invertebrates that consume algae. 60

Benthic invertebrate communities were influenced by stream velocity, minimum DO concentra- 40 tions, tree density in the sampling reach, and the abundance of periphyton. The composition of invertebrate communities (fig. 14) varied from a dominance of EPT 20 taxa (mayflies, stoneflies, and cad- disflies), which are relatively sensitive to organic enrichment, to a PERCENTAGE OF LESS-TOLERANT MAYFLY, CADDIS FLY, AND STONE FLY (EPT) TAXA 0 group of taxa (midges, true flies, 0 20 40 60 80 100 beetles, and worms) considered to be PERCENTAGE OF TOLERANT MIDGE, TRUE FLY, BEETLES AND WORM TAXA tolerant of organic enrichment and Figure 14. Benthic invertebrate community composition and tolerance to resultant low DO concentrations organic enrichment varies among Midwestern streams and rivers.

10 (Klemm and others, 1990; Johnson 10 0.5 and others, 1993). Sensitive stream + Standard Error invertebrates, such as EPT taxa, Mean – Standard Error require flowing waters with high 8 0.4 oxygen content to survive. Inverte- brate indicators of water-quality deg- 6 0.3 radation in many Till Plains streams may have been responding more to stresses associated with low-flow 4 0.2 PER CUBIC METER HOUR conditions than any substantial dif- 2

ferences in agricultural intensity IN MILLIGRAMS PER LITER across the upper Midwest. Below 2 0.1 average basin water yield and stream GRAMS OF O MAXIMUM RATE OF STREAM RESPIRATION, IN

velocity (fig. 15) in the Till Plains MINIMUM DISSOLVED OXYGEN CONCENTRATION, 0 0 basins probably were a result of Till Plains Dissected Western Lake Till Plains PHYSIOGRAPHIC SECTION 45 Figure 16. Minimum dissolved oxygen concentrations were associated with + Standard Error higher rates of stream metabolism in the Till Plains physiographic section. Mean – Standard Error

30 the Western Lake section (Sorenson food resources. In a related study, and others, 1999; Mitchell Harris, Stauffer and others (2000) also written commun., 2000), despite found that fish communities indi- favorable velocity and DO condi- cated better overall conditions in 15 tions. Extended (and variable) high agricultural streams with wooded streamflow conditions in the section riparian zones than those with more IN CENTIMETERS PER SECOND during spring and early summer open canopy.

AVERAGE VELOCITY IN STREAM REACH, 1997 (USGS streamflow data;

0 Porter, 2000), associated with above- Till Dissected Western 60 Plains Till Plains Lake average rainfall (fig. 3), may have been unfavorable for species that are + Standard Error PHYSIOGRAPHIC SECTION Mean unable to withstand hydrologic dis- – Standard Error Figure 15. Average velocity was turbances. 50 slowest in Till Plains streams, probably because of less rainfall and In general, biological indicators water yield in those basins prior to represented better water quality in 40 the date of sampling. Midwestern streams with riparian zones characterized by high tree density, particularly when velocity 30 relatively low rainfall (fig. 3) prior to and DO conditions were favorable. August 1997, the month of sample Overall, the number of invertebrate 20 collection. Minimum DO concentra- taxa was over 20 percent larger NUMBER OF INVERTEBRATE TAXA tions in these streams were lower in streams with high tree density than in the other physiographic sec- along the sampling reach (fig. 17). 10 High Low tions, and rates of stream metabo- Submerged woody debris in these (5% or greater) (less than 5%) lism (notably Rmax) were higher streams enhances the complexity TREE DENSITY IN THE SAMPLING REACH (fig. 16), indicative of organic of aquatic habitats, and seasonal leaf Figure 17. More invertebrate taxa were enrichment. Relatively fewer inver- fall into the streams, in addition to found in streams with high tree density tebrate taxa were found in streams in the periphyton, provides diverse than in streams with low density.

11 Implications for Water-Quality the basin; however, basin soil condi- algal and invertebrate indicators Monitoring and Assessment tions and riparian tree density along of degraded water quality increased stream segments appear to influence with the intensity of organic enrich- The assessment and manage- the fate of herbicide parent com- ment, as indicated by large amounts ment of water resources could bene- pounds and the rate of herbicide- of algal seston and dominance by fit from an improved understanding degradation processes. The effec- blue-green algae, high rates of of the influence of natural landscape tiveness of riparian zones as a stream productivity and respiration, and hydrologic factors on chemical buffer for intercepting runoff of and commensurately low DO con- and biological indicators of water centrations that occur during early agricultural contaminants may quality. For example, nutrient con- morning hours (Porter, 2000; be influenced by soil properties, centrations in Midwestern streams Sorenson and others, 1999). Thus, drainage characteristics, and the during late summer were related temporal and spatial variability amount of rainfall in stream basins more to antecedent runoff and algal- of natural factors are likely to influ- nutrient processes than rates of fer- prior to water-quality assessment. ence the chemical and biological tilizer or manure application. Herbi- Biological indicators of water qual- classification of water quality in cide concentrations in agricultural ity were influenced by physical streams and rivers, as well as the streams during the same time period factors such as streamflow and effectiveness of land-use manage- were proportional to herbicide use in riparian conditions; however, ment practices.

12 References Cited Freshwater biomonitoring and U.S. Environmental Protection benthic macroinvertebrates: Agency, Office of Water, Borchardt, M.A., 1996, Nutrients, in New York, Chapman and Hall, Office of Science and Technol- Stevenson, R.J., Bothwell, M.L., p. 41–158. ogy, EPA–822–B–00–002, and Lowe, R.L., eds., Algal Klemm, D.J., Lewis, P.A., Fulk, F., and p. A–25—A–42. ecology—Freshwater benthic Lazorchak, J.M., 1990, Macroin- Sorenson, S.K., Porter, S.D., Akers, ecosystems: San Diego, Calif., vertebrate field and laboratory K.K.B., Harris, M.A., Kalkhoff, Academic Press, Inc., methods for evaluating the biologi- S.J., Lee, K.E., Roberts, L.R., and p. 183–227. cal integrity of surface waters: Cin- Terrio, P.J., 1999, Water quality Fenneman, N.M., and Johnson, D.W., cinnati, Ohio, U.S. Environmental and habitat conditions in upper 1946, Physical divisions of the Protection Agency, Environmental Midwest streams relative to ripar- United States: U.S. Geological Monitoring Systems Laboratory, ian vegetation and soil characteris- Survey, accessed April 30, 2001, EPA/600/4–90/030, p. 207–244. tics, August 1997—Study design, http://water.usgs.gov/GIS/ Kolpin, D.W., and Kalkhoff, S.J., 1993, methods, and data: U.S. Geological metadata/usgswrd/physio.html. Atrazine degradation in a small Survey Open-File Report 99–202, 53 p. Fitzpatrick, F.A., Waite, I.R., stream in Iowa: Environmental Stauffer, J.C., Goldstein, R.M., and D’Arconte, P.J., Meador, M.R., Science and Technology, v. 27, Newman, R.M., 2000, Relationship Maupin, M.A., and Gurtz, M.E., p. 134–139. of wooded riparian zones and run- 1998, Revised methods for charac- Lerch, R.N., Blanchard.P.E., and off potential to fish community terizing stream habitat in the Thurman, E.M., 1998, Contribution composition in agricultural National Water-Quality Assess- of hydroxylated atrazine degrada- streams: Canadian Journal of Fish- ment Program: U.S. Geological tion products to the total atrazine eries and Aquatic Sciences, v. 57, Survey Water-Resources Investiga- load in Midwestern streams: Envi- p. 307–316. tions Report 98–4052, p. 21–26. ronmental Science and Technology, Torrents, A., Anderson, B.G., v. 32, p. 40–48. Goolsby, D.A., Boyer, L.L., and Bilboulian, S., Johnson W.E., Mallard, G.E., eds., 1993, Selected Osborne, L.L., and Kovacic, D.A., 1993, and Hapeman, C.J., 1997, Atrazine papers on agricultural chemicals Riparian vegetated buffer strips in photolysis—Mechanistic investiga- in water resources of the mid- water-quality restoration and tions of direct and nitrate-mediated continental United States: U.S. stream management: Freshwater hydroxy radical processes and the Geological Survey Open-File Biology, v. 29, p. 243–258. influence of dissolved organic car- Report 93–418, 89 p. Porter, S.D., 2000, Upper Midwest bon from the Chesapeake Bay: Johnson, R.K., Wiederholm, T., and river systems—Algal and nutrient Environmental Science and Tech- Rosenberg, D.M., 1993, Fresh- conditions in streams and rivers nology, v. 31, p. 1475–1482. water biomonitoring using individ- in the upper Midwest region Winter, T.C., Harvey, J.W., Franke, ual organisms, populations, and during seasonal low-flow condi- O.L., and Alley, W.M., 1998, species assemblages of benthic tions, in Nutrient Criteria Techni- Ground water and surface water— macroinvertebrates, in Rosenberg, cal Guidance Manual, Rivers A single resource: U.S. Geological D.M., and Resh, V.H., eds., and Streams: Washington, D.C., Survey Circular 1139, 79 p.