Climate and Land Use Change Impacts on N-Loads in Iowa Rivers and Remediation of Tile Water with an Anion-Exchange Resin

Abstract

This research was conducted to (1) better understand the underlying reasons for a continuous increase in nitrate loads in the Gulf of Mexico, and (2) if an industrial anion resin can be used at a field scale to reduce N losses from tile-drained watersheds to the rivers. The first objective was accomplished through statistical analyses of climate and land use change impacts on streamflow, baseflow, flow weighted nitrate-N concentrations (FWNC) and nitrate-N-loads in three major rivers of Iowa. The rivers included the Des Moines River, the Iowa River, and the Raccoon River. The results from this analysis showed that natural log of annual streamflow, baseflow, and N-loads were primarily controlled by the precipitation in the corresponding watersheds. For streamflow and baseflow, this precipitation corresponded to the current years as well as previous year precipitation. Previous year precipitation reflected the lack or excess presence of stored water in the soil and its consequences in terms of increased or decreased overland flow, infiltration, and percolation processes. For N loads, the precipitation effect was limited to one-year precipitation for the Des Moines and the Iowa Rivers and two-year precipitation for the Raccoon River. There were individual years when streamflow, baseflow, and N loads were impacted by up to three previous years’ precipitation. Effect of land use change, in terms of increased soybean area, had no effect on annual streamflow, annual baseflow, annual flow-weighted N concentrations or annual N-loads in all three rivers. Additional regression analysis of FWNC and N-loads from 1987-2001 showed no effect of N fertilizer use as an explanatory variable for any of the three watersheds. Statistical analysis of the combined annual data from all three rivers showed that there was a unique relationship between the natural log of streamflow, the baseflow, and the N-yield (N-loads/watershed area) versus the precipitation. The precipitation effects were both in terms of current year precipitation and the previous year precipitation. The coefficient of determination (R2) of Ln(streamflow), Ln(baseflow) and Ln(N load) with precipitation for the combined data were 0.74, 0.70 and 0.54, respectively. Limited scatter in the N-yield data at a given annual precipitation level over three rivers suggested that variation in annual precipitation has much bigger impact on N losses than the differences in cultural or cropping practices between the three river watersheds over the study period. Considering that there has been a 10-15% increase in precipitation in the Upper Midwestern United States in recent years, the combined N Yield relationship with precipitation would suggest that the recent increases in N-loads or increased hypoxic area in the Gulf of Mexico are likely due to increased precipitation. Statistical analysis of N-loads over a shorter period of time (1987-2001) also showed that changes in fertilizer use had no effect on river N-loads. Regression analysis of monthly streamflow, baseflow, N-loads and FWNC concentration showed that natural log of streamflow, baseflow, and N-loads were generally linearly related to precipitation in a given month and a few prior months. In some cases earlier in the season, these variables were also related to previous year’s precipitation, an indication that some of the past water stored in the soil both above and below the drain tile is interacting with current months precipitation and affecting the streamflow and baseflow. In most cases, there was no effect of soybean area on natural log of monthly streamflow, baseflow, or N-loads. A field test on the use of anion exchange resin to remediate tile water for nitrate showed that nitrate adsorption by the resin is instantaneous. The efficiency of the resin to retain nitrate varied 7-46%. This efficiency generally decreased with time due to the presence of sulfate, bicarbonates, and organic anions in tile water, which competed with nitrate ions for adsorption to the resin. In some instances, nitrate concentration in the percolating water was higher than the tile water most likely due to the expulsion of adsorbed nitrate ions on the resin by sulfate ion in the tile water. The results also showed that potassium chloride (KCl) is an effective resin-regenerating agent and provides a means to recycle wastewater as KNO3 fertilizer back on land. Although the use of anion exchange resin is an attractive alternative to passive technologies like bioreactors, saturated buffers, control drainage, etc. for remediating nitrate in tile water, it also presents some challenges in its use under field conditions. These challenges include the fouling up of the resin by sediment, sulfate, bicarbonate, and organic anions in tile water; costs associated with buying of resin and regenerating salt (KCl versus NaCl); need for a large volume of clean water for cleaning of resin; and the difficulty of treating large volume of tile water in-situ. However, the feasibility study shows that small-scale units similar to home water softener can be developed for individual homes in rural area where groundwater may be high in NO3-N concentration and NO3-N remediation is needed.