A simulation-based interval two-stage stochastic model for agricultural non-point source pollution control through land retirement.

This study presents a simulation-based interval two-stage stochastic programming (SITSP) model for agricultural non-point source (NPS) pollution control through land retirement under uncertain conditions. The modeling framework was established by the development of an interval two-stage stochastic program, with its random parameters being provided by the statistical analysis of the simulation outcomes of a distributed water quality approach. The developed model can deal with the tradeoff between agricultural revenue and "off-site" water quality concern under random effluent discharge for a land retirement scheme through minimizing the expected value of long-term total economic and environmental cost. In addition, the uncertainties presented as interval numbers in the agriculture-water system can be effectively quantified with the interval programming. By subdividing the whole agricultural watershed into different zones, the most pollution-related sensitive cropland can be identified and an optimal land retirement scheme can be obtained through the modeling approach. The developed method was applied to the Swift Current Creek watershed in Canada for soil erosion control through land retirement. The Hydrological Simulation Program-FORTRAN (HSPF) was used to simulate the sediment information for this case study. Obtained results indicate that the total economic and environmental cost of the entire agriculture-water system can be limited within an interval value for the optimal land retirement schemes. Meanwhile, a best and worst land retirement scheme was obtained for the study watershed under various uncertainties.