Are We Throwing Dollars Down The Drain?
The implementation of the US EPA’s Environmental Quality Incentives Program (EQIP) in the Upper Big Walnut Creek Watershed began in 1999 with the goal of reducing atrazine loads in Hoover Reservoir, a 22-billion-gallon drinking water source for Columbus, Ohio. The program, which provides incentives to farmers for selected agricultural best management practices (BMPs), has produced beneficial water quality trends in the reservoir and has decreased the
City’s treatment costs. However, BMP efficiencies and the overall cost-effectiveness of the program have not been assessed.
This project was completed in two phases. The goal of Phase I was the development of a methodology for evaluating the effectiveness of watershed-scale non-point source pollution abatement systems. Phase II of the project, and the focus of this discussion, had the primary goal of developing a framework for optimizing the effectiveness of watershed-scale pollution abatement systems using atrazine loading to Hoover Reservoir as a case study. Although the project focused on atrazine, the framework needed to be relevant and reproducible for other contaminants and watersheds. To achieve these goals the following primary questions had to be answered:
• How can we evaluate BMP performance?
• What are the most efficient BMPs in the project watershed?
• How does the effectiveness of the BMPs measure up to the incentives provided,
and how does this compare to water treatment costs?
• What proportion of the farmed watershed must implement BMPs to meet the
desired atrazine load reduction in the reservoir?
Phase I of the project proved the success of the EQIP program by comparison of pre- and postprogram water quality data. This also opened the door for the goal of any new technology – how can we make something that is functional and necessary even better? Success of the watershed optimization effort hinged on the ability to numerically reproduce the implementation of potential agricultural BMPs. After a GIS-based watershed characterization was completed, the
process of selecting an appropriate modeling package began. The NRCS’s NAPRA (National Agricultural Pesticide Risk Assessment) model, which is built around the GLEAMS (Groundwater Loading Effects of Agricultural Management Systems) model, was used to simulate the outcome of potential BMPs on each of the mapped soil types in the project watershed. The model generates estimations of contaminant losses at the edge of field and bottom of root zone for runoff, sediment loss, and leaching pathways. Modeled BMPs included various tillage types, crop rotations, and pesticide application methods.
The various soil and BMP combinations were processed through the NAPRA model using 50 years of historical climate data as a driving force for atrazine losses. The climate data set provided a reasonable distribution of atrazine loss data to develop probability distributions for each soil type and BMP combination. We compared modeled BMP efficiencies and estimated atrazine losses for farmed soils, which indicated the outcome of BMP implementation is widely varied by soil type. The risk of atrazine losses were characterized in terms of variable climate and soil conditions with a probabilistic simulation approach. To model the distribution of BMPs that minimizes the risk of excessive atrazine concentrations in the reservoir innovative postprocessing tools were developed to up-scale soil-level model output to farm field and watershed scales.
Many farm fields were composed of multiple soil types, and farmers do not typically change BMPs within a field. By up-scaling the soil level NAPRA output to the farm field scale, we were able to show the optimal BMP for each field in the watershed. Monte Carlo modeling methods were used to up-scale the soil level NAPRA output to the watershed scale based on the use of the most efficient and most cost effective (based on per acre incentive costs) BMPs. The results of Phase I of the project included a detailed watershed characterization, reservoir water quality analyses, and preliminary soils-level NAPRA modeling output. This information established the baseline conditions for the optimization phase (Phase II) of the project. The optimization results define the most efficient and cost-effective BMPs for each of the farmed soil types in the watershed, which indicate that atrazine losses from farmed soils can be reduced by up to 80% by implementing appropriate BMPs.
When atrazine concentrations exceed 2 ug/L, powdered activated carbon (PAC) is introduced to the water treatment process and additional treatment costs are incurred. Using 2 ug/L as a threshold, the total number of farmed acres that must implement BMPs to produce a reservoir load reduction that prevents or minimizes additional water treatment costs was estimated. This methodology can be used to established guidelines for future program enrollment and also show watershed areas where BMP incentives are most cost-effective. This framework can also be used for estimating nutrient loading and TMDL studies.