Skip to main content
ARS Home » Midwest Area » Columbus, Ohio » Soil Drainage Research » Research » Publications at this Location » Publication #197913

Title: OPTIMAL MANAGEMENT OF NON-POINT SOURCE POLLUTION FROM AGRICULTURE: AN APPLICATION OF DYNAMIC PROGRAMMING

Author
item SURENDRANNAIR, SUJITHKUMAR - THE OHIO STATE UNIV.
item SOHNGEN, BRENT - THE OHIO STATE UNIV.
item Fausey, Norman
item King, Kevin

Submitted to: Soil and Water Conservation Society
Publication Type: Abstract Only
Publication Acceptance Date: 2/9/2007
Publication Date: 7/23/2007
Citation: Surendrannair, S., Sohngen, B., Fausey, N.R., King, K.W. 2007. Optimal management of non-point source pollution from agriculture: an application of dynamic programming [abstract]. Soil and Water Conservation Society. p. 118.

Interpretive Summary:

Technical Abstract: Agricultural non-point source pollution is a major source of water quality impairment. When considering responses to non-point source pollution, several policy options have been considered historically, including reducing inputs (e.g. fertilizers) altering technologies on the landscape (e.g. conservation tillage). More recently, some authors have suggested that it may be possible to alter technologies within streams to increase the assimilative capacity of the environment (Ward et all., 2004). As this new technology emerges, and scientists advocate for its use, it is important to assess the benefits and costs relative to other policy options. This paper assesses the efficiency of three methods to reducing nutrient concentrations in streams. A dynamic programming framework is used to assess optimal management of nutrients within a continuous corn cropping system in the a small watershed in Ohio, the Upper Big Walnut watershed north of Columbus, Ohio. The watershed is one of several watersheds across the country being intensively studied as part of the US Department of Agriculture Conservation Effects Assessment Program. The model is a deterministic, infinite horizon, and full-information dynamic program. The model solves the optimal planning problem for maximizing the net benefits of reducing nitrogen concentrations in the stream. The actions in the model are to reduce nitrogen inputs, to improve technology associated with applying nitrogen, and to improve the technology in the stream so as to increase assimilative capacity. Stock variables in the model are the stock of nitrogen in the field, and the stock of nitrogen in a downstream reservoir. Parameters in the model are derived or estimated from several sources. The nitrogen stock building in the agricultural field and the nitrogen-loading factor were derived from the soil nitrogen balance equations. A quadratic nitrogen production function was used to relate the nitrogen application to corn production (Jeong and Sohngen, 2005). A quadratic social cost function was utilized to capture the negative externality of the nutrient loading to the Hoover reservoir. The social cost function was calibrated using results from a recent conjoint analysis-on benefits of water quality improvements in the study area (Tennity, 2005) and benefit transfer equations capturing the benefits of recreation in the streams and downstream reservoir. The results of the analysis suggest substantially lower rates of application than currently observed would be efficient when environmental costs are incorporated. The possibility of introducing in-stream technologies to enhance the assimilative capacity of streams has only a small effect on nutrient loads, and consequently, these technologies do not appear to be as efficient as activities on the landscape. Two important additional areas are currently being explored and the results for these will be presented. First, the effect of enhanced subsurface drainage on nutrient levels and the actions in the model will be assessed. In this region, subsurface drainage is an important activity that could influence the results. Second, many of the parameters are uncertain in the model. Methods to conduct Monte Carlo analysis to assess the stability of the results across different parameter distributions is being developed and will be presented.