Estimation of Nitrate Concentrations in Groundwater Using a Whole Farm Nitrogen Budget D. A. J. Barry, D. Goorahoo, and M. J. Goss* ABSTRACT Contamination of groundwater under agricultural land by NO~ is influenced by the kind of farming system. Onepossible methodof selecting farming systems that result in less NO 3 leaching is to cal- culate whole farm N budgets, that are simplified by assuming soiI-N remains constant fromone cycle of a rotation to the next. This method was applied to two model crop rotations using average crop yield data for two regions of Ontario, and to a cash.crop farmand a dairy farm using information on purchases, sales, andcrop yields, for these farms. The model rotations were corn (Zea mays L.)-soybean [Glycine max (L.) Merr.]-wheat (Triticum aestivum L.) and corn-soybean-wheat- hay (mixture of timothy, Phleum pratense L. and alfalfa, Medicago sativa L.)-hay-hay. Atmospheric deposition (18.4 kg N ha -t yr) was obtained by literature review. Symbiotic N2fixation by legume crops with different yields was estimated from regression equations. A net surplus in the N balance was converted to maximum mean NO3-N concentration in groundwater by assuming a groundwater recharge rate of 160 mm yr -~, and no denitrification. Predicted NO3-N con- centrations in leachate for the model corn-soybean-wheat rotation were greater for southwestern Ontario (22.4 mg -a) t han western Ontario (8.5 mgL-~), probably because more fe rtilizer wa s re c- ommended in the southwest. Including hay in the model rotation in- creased the amount of N leached by a factor of two in western Ontario, but only by 9% in the southwest. Predicted NO3-N concentration in groundwater for the cash crop farm was 6.7 mg L -~, compared with an average measured value of 9.5 mg L- ~ in the tile drainage water. For the dairy farmthe predicted value was 58 mgL- ~ and a measured value was not available. The simplified N balance method provided useful estimates of potential NO 3 leaching losses even though it relied on some major assumptions. A major uncertainty was atmospheric deposition of ammonia volatilized from on-farm sources. Denitrifi- cation could be as much as 62 kg N ha -~ yr -~ under continuous pro- duction of grain corn, based on differences between N present after harvest and amountof N leached. A GRICULTURAL LAND USEcan influence the amount of NO 3 lost to groundwater. Losses from arable land are most affected by the presence and type of crop, and are generally greater than from ungrazed grassland (Juergens-Gschwind, 1989). An ability to predict the magnitude of NO 3 leaching losses from different farming practices is necessary for making recommendations to minimize these losses. Onemethod for predicting NO 3 losses to groundwater is by calculation of N budgets for individual farms or farming systems. The N budget could be formulated so that a positive balance would indicate the amountof N potentially available for leaching. This amountand av- erage groundwater recharge could then be used to predict maximum mean NO3-N concentrations in groundwater recharge for the farm. The N budget can be simplified by assuming that soil organic matter content, and con- sequently soil N content, remain constant on a yearly Centre For Land and Water Stewardship, Univ. of Guelph, Guelph, ON, Canada N1G 2W1. Funding provided by Ontario Ministry of Agriculture and Food. Received 29 Sept. 1992. *Corresponding author. Published in J. Environ. Qual. 22:767-775 (1993). basis for monoculture systems or on a rotation basis for crop rotation systems. The N budget would therefore be calculated for one cycle of the farming system, that is, the length of a crop rotation. It wouldindicate the long term potential of a given farming system to cause NO 3 contamination of groundwater. This simplified N balance approach for predicting the long-term effect of farming practices on groundwater quality has been described in detail by Fried et al. (1976). They stated that any continued agricultural practice will result in the soil N content reaching a steady state level. The transfer of N to groundwater should then equal the difference between N inputs and N outputs. The concept was tested by Tanji et al. (1977) with monoculture corn at two sites in California. The budget included N inputs from irrigation, precipitation, N 2 fixation and fertilizer, and N outputs of denitrification and harvested crop. Pre- dicted NO3 concentrations in leachate were compared with measured NO 3 concentrations at various depths in the soil profile. It was concluded that NO3 losses at one site were approaching steady state values, but predicted NO 3 concentrations for the other site were less than mea- sured values. Macduff and White (1984) tested the ap- proach of Fried et al. (1976) using data for arable and grassland soils in England. They concluded that there was net mineralization of organic matter in their soils so the simple N-budget approach was inadequate. How- ever, they did not take account of N inputs from the urine and dung from the cattle grazing the grassland dur- ing spring and summer.Lund (1982) suggested that even where the assumption of steady state was invalid, results can still be indicative of the effects of certain soil and crop management practices. In Ontario, Hill (1986) used the simplified N budget method for a field that had been cropped to potato (Solanum tuberosum L.) since 1960. Amounts of N returned in crop residues and rye (Secale cereale L.) cover crops equalled only 2 to 3% of the soil organic N content and therefore was approximately con- sistent with the assumption of a steady state organic mat- ter level (Hill, 1986). Most cropped soils in Ontario have been under culti- vation for a long enough period that they are probably at or very near steady state in terms of organic matter (Groenevelt et al., 1984), and hence of soil N. The sim- plified N budget approach might therefore be useful for predicting NO3-N contamination of groundwater in areas of Ontario where agriculture is the predominant land use. The current maximum acceptable concentration of NO3- N in water used for drinking is 10 mgL-1 (Ministry of the Environment, 1992). Nitrate-N concentrations > 10 mg L-1 were detected in 21 of 180 farm wells sampled in 1986and 1987 in southern Ontario (Frank et al., 1991). Most of the wells were < 10 m deep and on sandy loam soils in corn-soybean growing areas of southwestern On- tario. A survey of 49 wells in a small Ontario farming community near Lindsay found that NO3-N concentra- tions exceeding 10 mgL- 1 occurred in 25% of the wells and were associated with well-drained fertilized soils 767 Published October, 1993