HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (11) Citing Articles via Google Scholar Google Scholar Articles by Birr, A. S. Articles by Mulla, D. J. Search for Related Content PubMed PubMed Citation Articles by Birr, A. S. Articles by Mulla, D. J. Agricola Articles by Birr, A. S. Articles by Mulla, D. J. Related Collections Surface Water Quality Water Quality Watershed and Landscape Processes Nutrient Management Water Pollution

TECHNICAL REPORT
Landscape and Watershed Processes

Evaluation of the Phosphorus Index in Watersheds at the Regional Scale

Department of Soil, Water, and Climate, 1991 Upper Buford Circle, Univ. of Minnesota, St. Paul, MN 55108

* Corresponding author (dmulla{at}soils.umn.edu)

Received for publication November 27, 2000.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY REFERENCES

 
Agricultural losses of phosphorus (P) in runoff are a primary cause of eutrophication in many freshwater systems. A modified version of the P Index originally developed jointly by the USDA (Agricultural Research Service [ARS], Cooperative State Research, Education, and Extension Service [CSREES], and Natural Resources Conservation Service [NRCS]) was used to prioritize P loss vulnerability at the regional scale from 60 watersheds located within Minnesota using readily available data related to the transport and sources of P. This modified version of the P Index was created for a regionally based analysis of the index. Validation of the P Index rating was conducted using long-term water quality monitoring data consisting of total P concentrations collected from watersheds and lakes. The modified version of the P Index produced a strong correlation between P Index rating and total P stream monitoring data in watersheds (r² = 0.70) excluding the Red River Basin. An equally strong relationship was observed between P Index rating and lake water quality (r² = 0.68) using the modified P Index. The P Index ratings for the Red River Basin showed good correlation with observed total P stream monitoring data (r² = 0.51); however, the P Index ratings were smaller than in other basins. The results of this study suggest that, with certain limitations, the P Index can be used at the regional scale to prioritize P loss vulnerability using state and national databases. Regional P Index ratings represent the average risk for agricultural land within the entire watershed.

Abbreviations: USGS, United States Geological Survey

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY REFERENCES

 
AQUATIC plant growth in many freshwater systems is limited by phosphorus (P), and P loads to freshwater systems often originate from agricultural soils (Schindler, 1977). A combination of transport and source factors directly influence P movement from agricultural systems to P-sensitive waters (Sharpley et al., 1993). Transport factors include the mechanisms by which P is delivered to surface waters, such as erosion and runoff. Source factors represent the amount of P available for transport, including soil test P and P applied (rate and method) in fertilizer and organic forms. Long-term applications of both organic and fertilizer P in excess of crop uptake needs have contributed to elevated soil test P levels in many regions throughout North America (Fixen, 1998). Reducing P losses from nonpoint sources entails targeting areas with a combination of high P transport and high P source potentials (Sharpley et al., 1999). The USDA developed a P Index that integrates both transport and source factors to identify areas vulnerable to P export (Lemunyon and Gilbert, 1993). The original P Index uses information readily available at the field scale to assess the risk of P losses and to evaluate alternative management practices.

Previous studies indicate that the P Index effectively ranks the potential for P loss. Stevens et al. (1993) applied the P Index to plots within the Dairy–McKay Hydrologic Unit Area (HUA) in western Oregon and 48 farm units comprising a subbasin (1930 ha) within the Granger Drain HUA of eastern Washington. The authors demonstrated that the P Index could be used to prioritize sites based on P loss vulnerability. The study lacked field validation to substantiate that P Index ratings correspond to actual P losses from a site. Sharpley (1995) applied the P Index to 30 P-fertilized and unfertilized fields in the U.S. Southern Plains ranging in size from 1.1 to 125 ha. The study found a strong relationship between P Index rating and P losses monitored over a 16-yr period (r² = 0.70).

Previous studies have focused on the use of the P Index at the field (<1:5000) and local scales (1:12000 to 1:24000). For reference, second-order soil surveys are commonly mapped at a scale of 1:24000. The purpose of this study was to develop a modified version of the P Index to characterize P loss at the regional scale (1:100000 to 1:250000). The P Index values in the modified approach were compared with stream and lake P monitoring.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY REFERENCES

 
Study Area
The study area consisted of 60 watersheds defined by the United States Geological Survey (USGS) as eight-digit hydrologic units ranging in size from 3982 to 540 672 ha. Eight-digit hydrologic units are defined as major watersheds. The study area is comprised of six basins including the Upper Mississippi River, Lower Mississippi River, Minnesota River, St. Croix River, Red River, and Cedar River Basins (Fig. 1) . The Upper Mississippi River Basin has high water erosion potentials attributed to a large proportion of cultivated cropland (Table 1). The southern half of this basin consists primarily of agricultural land, in contrast to the northern half of the basin, which is dominated by forest cover. The Lower Mississippi River Basin is comprised of steep loessial soils draining to deeply dissected river valleys, and has severe water erosion potentials (Minnesota Department of Natural Resources, 1999). Agriculture is the predominant land use in this basin. Topography in the Minnesota River Basin consists of a gently rolling ground moraine and glacial till (Hobbs and Goebel, 1982). Water erosion potentials in the Minnesota River Basin are high because of a combination of relatively steep land, heavy precipitation, and cultivated cropland comprising more than 69% of the basin (USDA, 1992). The St. Croix River Basin is a combination of outwash plains and moraines with rolling to hummocky topography (Hobbs and Goebel, 1982). Total erosion losses are low because the majority of the basin is forested. The Red River Basin is a very flat glacial lake plain with silty, sandy, and clayey deposits (Minnesota Department of Natural Resources, 1999). Water erosion potentials are low due to the level to gently rolling topography; however, wind erosion is significant. The Cedar River Basin has gently rolling topography and high water erosion potentials, with 75% of the basin comprised of cultivated cropland (USDA, 1992).

View larger version (73K): [in this window] [in a new window]   Fig. 1. River basins of Minnesota within the study area.

The erosion potential was calculated for each soil type within a STATSGO map unit using the preceding four variables (A = RKLS). Erosion potential values for STATSGO map units were area-weighted based on the percent of the map unit comprised of each soil type. A 1:250000 scale landuse–landcover coverage developed by the USGS in the late 1970s and early 1980s was used to determine erosion potentials spatially coincident with cropland and pastureland (USEPA, 1994). An erosion potential value for all cropland and pastureland within a watershed was determined using the percent of each STATSGO map unit covering a watershed. The landuse coverage did not differentiate spatially between cropland and pastureland; however, Census of Agriculture data indicate that pastureland represents about 11% of this classification category in Minnesota (National Agricultural Statistics Service, 1999). Differences in potential erosion for the two land uses were accounted for in the determination of the C factor based on the proportion of hay reported for a particular county.

Cropping management factors (C) were adapted from values provided by the USDA (1975) and Wischmeier and Smith (1978) for corn (Zea mays L.), wheat (Triticum aestivum L.), soybean [Glycine max (L.) Merr.], hay, sugar beet (Beta vulgaris L.), potato (Solanum tuberosum L.), oat (Avena sativa L.), and barley (Hordeum vulgare L.). The C factors were calculated for each county based on the area of each harvested crop covering the county. Watershed values for the C factors were weighted based on the proportion of the watershed that was covered by the county. The C factor calculations include crop rotation effects but not the variation in tillage effects. There is no reliable method for estimating the variation in crop residue cover across the watersheds studied. The conservation practice factor (P) was assumed to be 1, because it could not be accurately quantified at the regional scale. The overall erosion potential value for each watershed represents the product of the area-weighted C factor and the variables R, K, and LS for each watershed (A = RKLSCP).

Runoff
Average annual runoff values for each watershed were derived from the average annual discharge monitored from 1951 to 1985 for 327 stations distributed throughout Minnesota (Lorenz et al., 1997). The average annual runoff value is calculated as the average annual discharge divided by the drainage basin area defined for the station. Runoff values corresponding to each station were interpolated with an inverse distance weighted interpolator using a cell resolution of 3000 m, 8 nearest neighbors, and an exponent of 2. An average runoff value was determined for each of the 60 watersheds on the basis of individual watershed boundaries using ARC/INFO software (Environmental Systems Research Institute, 2000a).

Agricultural Land within 91.4 m of a Watercourse
The area of cropland and pastureland within 91.4 m of drainage ditches and perennial streams was determined using hydrography coverages developed by the Minnesota Department of Transportation (1999) and the United States Geological Survey (1999). The 91.4-m distance reflects a setback standard for manure application established by the Minnesota Pollution Control Agency (MPCA) for controlling nonpoint source pollutant loads from agricultural lands. The USGS landuse–landcover coverage (USEPA, 1994) was used to determine the area of cropland and pastureland within the 91.4-m proximity to watercourses for each watershed. The cropland and pastureland area within 91.4 m of drainage ditches and perennial streams was normalized by the total watershed area to account for the variation in watershed sizes. Watersheds with a higher proportion of cropland and pastureland within 91.4 m of drainage ditches and perennial streams are more vulnerable to P transport to surface water bodies from inorganic and organic P amendments than watersheds with a smaller proportion.

Soil Test Phosphorus
Mean soil test P levels for each county represented a 5-yr database consisting of 22 421 Bray-1 extractable P (Brown, 1998) samples analyzed by the University of Minnesota's soil testing laboratory. For reference, agronomic Bray-1 extractable P values that classify Minnesota soils from very low to very high are: 0–5, 6–10, 11–15, 16–20, and >20 mg P/kg (Rehm et al., 1994). The spatial distribution of samples collected within each of the counties is not known. Soil test P varies across field scales; however, the database provides the best available regional representation of soil test P trends for Minnesota. Soil test P levels for each watershed were based on the area of the watershed covered by each county.

Fertilizer Phosphorus and Organic Phosphorus Application Rate
Data for P-fertilizer sales by county were obtained from the Minnesota Department of Agriculture (1997). Fertilizer P values for watersheds were based on a summation of area-weighted county-based values intersecting the watersheds. The total area of fertilized land within each watershed was determined using the same procedure based on reported county values (National Agricultural Statistics Service, 1999). Less than 1% of this area was in pastureland and rangeland. The aggregated fertilizer P value was divided by the aggregated reported fertilized land for each watershed to determine fertilizer P application rates. This procedure assumes that the fertilizer P is distributed equally over the total fertilized area. Counties in which the reported fertilized land represented less than 1% of the reported total cropland area were assumed to have no fertilizer P amendments. These counties represented regions where a large portion of fertilizer sales are used on lawns, gardens, and golf courses.

The P content of livestock manure was calculated based on the total number of cattle, swine, broilers, and turkeys reported within each county (Midwest Planning Service, 1985; Schmitt, 1999; National Agricultural Statistics Service, 1999). The total amount of manure P was derived for each watershed based on the summation of area-weighted county values intersecting the watersheds. The reported total cropland area was also determined using the same procedure (National Agricultural Statistics Service, 1999). The aggregated total P content of manure was normalized by the aggregated total cropland area for each watershed to determine organic P application rates. The distribution of organic P application is assumed to be uniform throughout the total reported cropland area.

Burkart and James (1999) used a similar approach to quantify both nitrogen (N) fertilizer and organic N application rates in the Mississippi River Basin. These workers normalized the reported fertilizer and organic N amounts determined for each watershed by the total area of the watershed, acknowledging that use of this methodology would produce N application rates lower than normalizing N application rates by cropland area.

Modified Phosphorus Index
The original version of the P Index developed by Lemunyon and Gilbert (1993) consists of an 8 by 5 matrix of site characteristics and rating values, respectively. The irrigation erosion site characteristic included in the original P Index was not used because irrigation is not common in Minnesota (National Agricultural Statistics Service, 1999). In the original P Index, the transport factors included soil erosion and runoff. The source factors included soil test P and rate and method of P applied in fertilizer and organic forms. For the modified P Index (Table 2), the proportion of a watershed comprised of cropland and pastureland within 91.4 m of drainage ditches and perennial streams was also included as a transport site characteristic. Recent versions of the P Index have also included a proximity to water component in the P Index (Sharpley et al., 1999). Each site characteristic is assigned a weighting factor based upon the premise that site characteristics have a varying effect on P loss to runoff. Each site characteristic has an associated P loss rating value (very low, low, medium, high, and very high) using a base of 2 to reflect the higher potential for P loss associated with higher rating values. The P Index rating is the summation of the product of the rating value and corresponding weighting value for each site characteristic. The modified P Index was developed in an attempt to more accurately depict regional P loss conditions from agricultural land in Minnesota based on stream and lake monitoring data. Data were not available for site characteristics such as soil test P to characterize other land uses within a watershed. The assumption is that agriculture is the primary source of P within a watershed.

Categories corresponding to the rating values were derived by segregating the distribution of statewide values for each site characteristic into five classes using the quantile classification method available in ArcView software (Environmental Systems Research Institute, 2000b). Quantiles were used because it is a statistically objective method of classifying the distribution of P Index ratings.

The P Index rating values resulting from the application of the modified P Index were validated using two different sets of data. The first set of data consists of a 27-yr record (1968–1994) of total P concentrations collected at the mouth and at interior points in 54 of the 60 watersheds in the study. Because no statewide standard exists for total P concentration in rivers, a tolerance value of 0.25 mg/L total P was selected as it represented the average total P concentration of the Minnesota River (Mulla and Mallawatantri, 1997). These data were statistically compared across a wide range of months and flow regimes. They were not weighted by month or flow regime. The P Index ratings were correlated with the percentage of samples in which total P concentrations exceeded 0.25 mg/L for 37 of the 60 watersheds in the study area. Seventeen of the 54 watersheds with monitoring data derived from main stems of the six major rivers were excluded from the statistical comparison to ensure that both cumulative and point source (urban) effects did not influence the total P observations. Water quality data collected on main stems such as the Mississippi River may reflect P losses from regions outside of the watershed in which the station is located.

The second set of validation data consists of lake water quality parameters maintained by the USEPA's STORET national water quality database. The P Index ratings were statistically compared with median total P concentration of lakes for 20 of the 60 watersheds having more than 14 lakes assessed. A majority of the lakes (66%) were monitored during summer months (June–September) between 1989 and 1998. The remaining data were collected between 1970 and 1988, including nonsummer samples (Heiskary and Wilson, 2000).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY REFERENCES

 
Water Quality Patterns
The highest proportion of stream water quality samples exceeding 0.25 mg/L total P was observed in watersheds of the Minnesota and Lower Mississippi River Basins (Fig. 2a) . Watersheds in the St. Croix River Basin and the northern portion of the Upper Mississippi River Basin had a significantly smaller number of stream water quality samples exceed 0.25 mg/L total P.

View larger version (40K): [in this window] [in a new window]   Fig. 2. Total P monitoring data collected in Minnesota summarized by watershed. (a) Percentage of samples exceeding 0.25 mg/L total P collected from streams. (b) Median total P concentrations (µg/L) collected from lakes.

Site Characteristics
A wide range of soil erosion potential values was calculated (2–103 Mg/ha) for the 60 watersheds studied. The highest values of potential soil erosion were calculated in the southeastern portion of the state and the lowest values were calculated in the northwestern portion of the state (Fig. 3a) . These trends in soil erosion are indicative of both topographic and climatic variations.

View larger version (34K): [in this window] [in a new window]   Fig. 3. Modified P Index transport site characteristics included in the modified P Index for watersheds of Minnesota. (a) Potential soil erosion losses (Mg/ha). (b) Observed runoff values interpolated from 327 United States Geological Survey (USGS) long-term gauging stations (cm/yr). (c) Cropland and pastureland area within 91.4 m of drainage ditches and perennial streams normalized by the watershed area.

Monitored runoff rates increased from the western portion of the study area to the eastern portion of the study area. The highest runoff rates (28 cm) occurred in the northeastern portion of the St. Croix River Basin (Fig. 3b). A wide range of values for the proximity to watercourse site characteristic was observed (0.2–12.1%) throughout the study area with the largest proportions occurring in the Red River and Minnesota River Basins (Fig. 3c). The greatest Bray-1 extractable P values observed in the study area were concentrated in the Upper Mississippi Basin (Fig. 4a) . Based on agronomic P requirements, 82% of the watersheds were in the very high category, at which point further P applications do not generate increases in yield (Rehm et al., 1994). The highest application rates of fertilizer P were concentrated in the Minnesota River Basin (Fig. 4b). Rotations in the Minnesota River Basin consist primarily of corn–soybean crops requiring large inputs of P compared with the small grain crops grown in the Red River Basin. Furthermore, organic P inputs are greater in the Upper and Lower Mississippi River Basins than in the Minnesota River Basin. The highest rates of organic P application occurred in the Upper Mississippi River and Lower Mississippi River Basins (Fig. 4c). Rates of organic P application ranged from 1 to 18 kg P/ha. Elevated organic P application rates and soil test P levels corresponded to areas of intense dairy animal production. Watersheds that had high fertilizer application rates generally did not correspond to watersheds that had high organic P application rates, indicating that farmers generally recognize the nutrient value of applied manure.

View larger version (34K): [in this window] [in a new window]   Fig. 4. Modified P index source factors for watersheds of Minnesota. (a) Bray-1 extractable P levels (mg P/kg). (b) Fertilizer P application rate normalized by reported total fertilized land (kg P/ha). (c) Organic P application rate normalized by reported total cropland (kg P/ha).

View larger version (75K): [in this window] [in a new window]   Fig. 5. Modified P Index ratings for watersheds of Minnesota.

View larger version (18K): [in this window] [in a new window]   Fig. 6. Relationship between water quality data and modified P Index rating for selected watersheds of Minnesota. (a) Relationship between modified P Index rating and percentage of samples exceeding 0.25 mg/L total P for watersheds within the Upper Mississippi River, Lower Mississippi River, Minnesota River, St. Croix River, and Cedar River Basins. (b) Relationship between modified P Index rating and median total P concentrations of lakes. (c) Relationship between modified P Index rating and percentage of samples exceeding 0.25 mg/L total P for watersheds within the Red River Basin.

The relationship between P Index ratings and stream water quality data (Fig. 6c; r² = 0.51) was slightly poorer in the Red River Basin than that observed in the other basins. The mean P Index rating for watersheds in the Red River Basin is significantly lower (P < 0.05) than in the other basins, with mean values of 24.5 for the Red River Basin and 30.9 for the remaining five basins. Watersheds within the Red River Basin generally had low P Index rating values for each of the site characteristics relative to the other watersheds in the study, with the exception of the proximity to watercourse site characteristic. The lower P Index ratings observed in the Red River Basin may be attributed to transport and source factors unique to the Red River that are not accounted for by the current P Index model. Sediment contributions from wind and streambank erosion are problematic in the Red River Basin (Stoner et al., 1993). Source P contributions are low because of a lack of animal production, as reflected by the lower organic P application rates and soil test P values in the Red River Basin relative to the other watersheds.

The asymptotic relationships observed between P Index ratings and monitoring data from watersheds other than the Red River Basin (Fig. 6a) suggest that a value of 32 is an appropriate cutoff for watersheds that are severely impaired with respect to potential P loss in runoff. This suggests that state agencies would prioritize clean-up efforts for watersheds based on the regional scale P Index values, and that individual fields should be identified for the implementation of best management practices (BMPs) if the watershed P Index value exceeds 32. In the Red River Basin, the critical P Index value is approximately 24 to 25, corresponding to a 20% exceedance of the 0.25 mg/L total P level in streams. As stated in previous sections, unique factors related to the transport and source of P in the Red River Basin distinguish it from the rest of the basins in the study area. Consequently, a lower critical P Index value is required in this region to prioritize watersheds for P management strategies. One important caveat here is that the highest rating for both P fertilizer and organic P application method was used in this study. Application methods with less potential for P losses will lower the critical P Index value; however, the relative rankings of the P Index ratings across watersheds would not change.

The Minnesota Pollution Control Agency has adopted the Aquatic Ecoregion framework to establish lake water quality criteria on a regional basis. The majority of the watersheds evaluated in this study for lake water quality are contained within the Northern Lakes and Forests Ecoregion and the North Central Hardwood Forest Ecoregion. The Northern Lakes and Forests Ecoregion extends from the central portion of the Upper Mississippi River Basin to the northeastern portion of the state. The North Central Hardwood Forest Ecoregion extends across the southern portion of the Upper Mississippi River Basin in addition to portions of the Minnesota River, Lower Mississippi River, St. Croix River, and eastern Red River Basins. Total P concentrations for "nonsupport" of swimmable use are 50 µg/L in the Northern Lakes and Forests Ecoregion, and 60 µg/L in the North Central Hardwood Forest Ecoregion. The preceding total P criteria correspond to P Index ratings of 29 and 31, respectively (Fig. 6b). The excellent relationship between total P ecoregion criteria for lakes and the P Index rating suggests that the P Index is a reliable indicator of lake water quality at the watershed scale.

	SUMMARY

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY REFERENCES

 
A modified version of the P Index involving a statistical delineation of the rating values for each of the site characteristics was applied to streams and lakes in Minnesota at the regional scale. The modified P Index was intended to prioritize watersheds for P loss potential based on their comparative differences in transport and sources of P.

This study shows that the modified P Index accurately depicts regional-scale trends in P loss for streams and lakes. The intent of using the P Index at a regional scale is to identify critical transport and source factors that can be controlled at a local scale. The P Index enables managers to identify areas that are a high priority for the implementation of BMPs that minimize P losses to sensitive watercourses.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION SUMMARY REFERENCES

 

• Brown, J.R. 1998. Recommended chemical soil test procedures for the north central region. North Central Regional Res. Publ. no. 221 (Revised). SB 1001. Missouri Agric. Exp. Stn., Columbia.
• Burkart, M.R., and D.E. James. 1999. Agricultural-nitrogen contributions to hypoxia in the Gulf of Mexico. J. Environ. Qual. 28:850–859.[Abstract/Free Full Text]
• Environmental Systems Research Institute. 2000a. Arc/Info Version 8.0.1. ESRI, Redlands, CA.
• Environmental Systems Research Institute. 2000b. ArcView Version 3.1. ESRI, Redlands, CA.
• Fixen, P.E. 1998. Soil test levels in North America. Better Crops 82: 16–18.
• Gibson, C.E. 1997. The dynamics of phosphorus in freshwater and marine environments. p. 119–135. In H. Tunney et al. (ed.) Phosphorus loss from soil to water. CAB Int., New York.
• Hamlett, J.M., D.A. Miller, R.L. Day, G.W. Peterson, G.M. Baumer, and J. Russo. 1992. Statewide GIS-based ranking of watersheds for agricultural pollution prevention. J. Soil Water Conserv. 47:399–404.
• Heiskary, S.A., and C.B. Wilson. 2000. Minnesota lake water quality assessment data: 2000. Minnesota Pollution Control Agency, Environ. Outcomes Division, Environ. Monitoring and Analysis Section, St. Paul.
• Hobbs, H.C., and J.E. Goebel. 1982. Geologic map of Minnesota: Quaternary geology. Vol. S-1. Minnesota Geol. Survey, St. Paul.
• Lemunyon, J.L., and R.G. Gilbert. 1993. The concept and need for a phosphorus assessment tool. J. Prod. Agric. 6:483–486.
• Lorenz, D.L., G.H. Carlson, and C.A. Sanocki. 1997. Techniques for estimating peak flow on small streams in Minnesota. Water-Resour. Investigations Rep. 97-4249. U.S. Geol. Survey, Denver, CO.
• Midwest Planning Service. 1985. Livestock waste facilities handbook. Midwest Planning Serv. Livestock Waste Subcommittee Rep. MWPS-18. 2nd ed. Iowa State Univ., Ames.
• Minnesota Department of Agriculture. 1997. Total fertilizer and nutrients by county. Agron. and Plant Protection Div., St. Paul.
• Minnesota Department of Natural Resources. 1999. Ecological classification system [Online]. Available at http://www.dnr.state.mn.us/ebm/ecs/ (verified 18 July 2001). MDNR, St. Paul.
• Minnesota Department of Transportation. 1999. State of Minnesota base map. Office of Land Management Surveying and Mapping Section, St. Paul.
• Mulla, D.J., and A.P. Mallawatantri. 1997. Minnesota River Basin water quality overview. Minnesota Ext. Serv. FO-7079-E. Univ. of Minnesota College of Agric., St. Paul.
• National Agricultural Statistics Service. 1999. 1997 Census of agriculture: Minnesota state and county data [Online]. Vol. 1, Geographic Area Series Part 23. Available at http://usda.mannlib.cornell.edu/reports/census/ac97amn.pdf (verified 18 July 2001).
• Rehm, G., M. Schmitt, and R. Munter. 1994. Fertilizer recommendations for agronomic crops in Minnesota. Minnesota Ext. Serv. BU-6240-E, Revised 1994. Univ. of Minnesota College of Agric., St. Paul.
• Renard, K.G., G.R. Foster, G.A. Weesies, D.K. McCool, and D.C. Yoder. 1997. Predicting soil erosion by water: A guide to conservation planning with the Revised Universal Soil Loss Equation (RUSLE). Agric. Handb. no. 703. USDA, Washington, DC.
• Schindler, D.W. 1977. Evolution of phosphorus limitation in lakes. Science 195:260–262.[Free Full Text]
• Schmitt, M.A. 1999. Manure management in Minnesota. Minnesota Ext. Serv. FO-3553-C, Revised 1999. Univ. of Minnesota College of Agric., St. Paul.
• Sharpley, A. 1995. Identifying sites vulnerable to phosphorus loss in agricultural runoff. J. Environ. Qual. 24:947–951.[Abstract/Free Full Text]
• Sharpley, A.N., T.C. Daniel, and D.R. Edwards. 1993. Phosphorus movement in the landscape. J. Prod. Agric. 6:492–500.
• Sharpley, A.N., T. Daniel, T. Sims, J. Lemunyon, R. Stevens, and R. Parry. 1999. Agricultural phosphorus and eutrophication. ARS-149. USDA-ARS, University Park, PA.
• Stevens, R.G., T.M. Sobecki, and T.L. Spofford. 1993. Using the phosphorus assessment tool in the field. J. Prod. Agric. 6:487–492.
• Stoner, J.D., D.L. Lorenz, G.J. Wiche, and R.M. Goldstein. 1993. Red River of the North Basin, Minnesota, North Dakota, and South Dakota. Water Resour. Bull. 29:575–615.
• USDA. 1975. Minnesota field office technical guide. Section III. Natural Resour. Conserv. Serv., St. Paul.
• USDA. 1991. State soil geographic data base (STATSGO): Data users guide. Natural Resour. Conserv. Serv. Misc. Publ. no. 1492. Natural Resour. Conserv. Serv., Fort Worth, TX.
• USDA. 1992. 1992 National resources inventory. Natural Resour. Conserv. Serv., Fort Worth, TX.
• USEPA. 1994. 1:250,000 scale quadrangles of landuse/landcover GIRAS spatial data in the conterminous United States [Online]. Available at http://www.epa.gov/ngispgm3/nsdi/projects/giras.htm (verified 18 July 2001). USEPA, Washington, DC.
• United States Geological Survey. 1999. National atlas of the United States: Streams and waterbodies [Online]. Available at http://www-atlas.usgs.gov/hydrom.html (verified 18 July 2001). USGS, Washington, DC.
• Wischmeier, W.H., and D.D. Smith. 1978. Predicting rainfall erosion losses. USDA Sci. and Education Admin. Agric. Handb. no. 537. USDA, Washington, DC.

This article has been cited by other articles:

				G. M. Pierzynski and K. A. Gehl Plant Nutrient Issues for Sustainable Land Application J. Environ. Qual., January 1, 2005; 34(1): 18 - 28. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (11) Citing Articles via Google Scholar Google Scholar Articles by Birr, A. S. Articles by Mulla, D. J. Search for Related Content PubMed PubMed Citation Articles by Birr, A. S. Articles by Mulla, D. J. Agricola Articles by Birr, A. S. Articles by Mulla, D. J. Related Collections Surface Water Quality Water Quality Watershed and Landscape Processes Nutrient Management Water Pollution