Agricultural land use and its relationship to nutrient levels in four eastern Iowa river systems.

摘要

Concentrated animal feeding operations (CAFO) and fertilizer application to row crops may contribute to poor water quality in surface waters. To test this hypothesis, we evaluated nutrient concentrations and fluxes in four Eastern Iowa watersheds sampled between 1996 and 2004. We found that these watersheds contribute nearly 10% of annual nitrate flux entering the Gulf of Mexico, while representing only 1.5% of the contributing drainage basin. Mass budget analysis shows stream flow to be a major loss of nitrogen (18% of total N output), second only to crop harvest (80%). The major watershed inputs of nitrogen include applied fertilizer for corn (52% of total N input) and nitrogen fixation by soybeans (25%). Despite the relatively small input from animal manure (∼7.5%), the results of spatial analysis indicate that row crop and CAFO densities are significantly and independently correlated to higher nitrate concentration in streams. Pearson Correlation Coefficients of 0.59 and 0.89 were found between nitrate concentration and row crop and CAFO density, respectively. Multiple linear regression analysis produced a correlation for nitrate concentration with an R2 value of 85%. High spatial density of row crops and CAFOs are linked to the highest river nitrate concentrations (up to 15 mg/l normalized over five years).; We constructed a network model with the capability to predict nitrate concentration and flux for the Cedar, Iowa, Skunk and Wapsipinicon Rivers in eastern Iowa. This model accounts for nitrate flux from agricultural activities, point source emissions and from upstream river sections. We found evidence suggesting that denitrification plays a small but measurable role in estimating river nitrate concentrations. Using our water quality database and network model, we evaluated several mitigation strategies to reduce nitrate concentration below the USEPA limit of 10 mg/l. Eliminating CAFOs from some sub-watersheds, as well as redistributing CAFOs within watersheds, showed promising results. Reducing row crop acreage was not universally successful in reducing nitrate below the 10 mg/l target and reductions in point source emissions had little effect on riverine nitrate concentrations. Our analysis of riparian buffer zones around streams was inconclusive in that statistical correlations and confidence levels.