Advanced decision support methods for solving diffuse water pollution problems

Abstract: Dealing with water diffuse pollution is a major problem for watershed managers. This problem raises many complicated questions, which are important to answer in order to reach water environment protection goals. This study suggested some possible answers for the country of Lithuania. Among them were the identification of critical source areas, the identification of sensitive areas and the application of multi-objective spatial optimization. Those decision support methods were not only suggested, but also examined through literature review and their application was demonstrated practically on the Graisupis river catchment, which is located in the middle of Lithuania. For this purpose, the SWAT (Soil and Water Assessment Tool) model was prepared and successfully calibrated and validated for water flows and nitrate load simulations. The model was calibrated for 7 years (2000-2006) and validated for 3 years period (2007-2009). The model was run for 10 years period (2000-2009) in order to obtain results for decision support methods. Critical source areas were defined as those areas, which have nitrate loads to surface water bodies higher by two standard deviations from average in the catchment. Sensitivity (nutrient leaching potential) of areas was assigned based on the response of modeled physical nature to the addition of nitrogen fertilizers. The SWAT model was also used for the simulation of effects of best environment practices. The results were imported into the genetic algorithm, which was used for the purposes of multi-objective spatial optimization. Model results indicated average nitrate loading of 15.9 kg nitrate nitrogen per hectare in the catchment. The identification of critical source areas located 12.4% of the Graisupis river catchment as risk areas. The sensitive areas identification assigned medium or low sensitivity to 99.5% of the catchment. Only 0.4% of the catchment territory was identified as high or very high sensitivity. Multi-objective spatial optimization increased the costeffectiveness of diffuse pollution abatement 24 times (up to 50 times with lesser implementation scale), if compared to the random selection of best environmental practices. Optimization with equal weights for environmental and economic objectives resulted in 16.9 LTL for reduction of 1 kg nitrate nitrogen to surface water bodies, while providing 62% reduction of total loads to surface water bodies. This scenario required 24% of additional catchment territory to be converted to grasslands and consideration of filter strips for 34% of the catchment territory. Optimization for obtaining Pareto optimum between environmental and economic objectives provided the most cost effective solution of 9.7 LTL for reduction of 1 kg nitrate nitrogen, while providing 25% reduction of total loads to surface water bodies. This scenario required the application of cover crops on 2.6%, new grasslands on 1.6% and consideration of filter strips on 11% of the Graisupis river catchment area. Optimization for obtaining Pareto optimum between environmental and economic objectives also provided quantifiable relationship between economic and environmental objectives in the form of regression equation.