Assessing the Effectiveness of Agricultural BMPs Using the Copula Approach

A variety of best management practices (BMPs), suitable for the abatement and control of non-point source pollution have been developed and recommended for application in the state of Louisiana. Owing to comparatively poor water quality levels, the Mermentau and Vermilion-Teche River basins in southwestern Louisiana have been prioritized for BMP implementation. Water bodies in these river basins have been found impaired to various degrees for their designated uses of fishing, swimming, and/or wildlife propagation. Evaluation of the improvement in water quality is essential in order to ascertain the effectiveness of these BMPs at the watershed scale. Consideration of inherent uncertainty is required while evaluating improvements in the water quality of various water bodies. Concentrations and loads of these pollutants show considerable dependence on flow characteristics. A better assessment of the variability of water quality can be based on the bivariate joint distribution of pollutant and flow variables. Conventional formulations for such joint distributions have certain restrictions with respect to the type of marginal distributions and the dependence structure. The new copula-based approach overcomes such restrictions and allows modeling of joint behavior on the basis of marginals in their untransformed forms. This paper presents an evaluation of the effectiveness of BMPs that have been implemented in 11 watersheds of Mermentau and Vermilion-Teche River basins. Statistical evaluation of improvements in turbidity and total suspended solids has been made for one of the watersheds on the basis of copula-based conditional distribution concept for purposes of illustration. As concentration of these pollutants has significant dependence on flow magnitudes, the effect of flow is incorporated by employing these conditional distributions. The results are compared with those obtained using other conventional parametric and non-parametric methods. The turbidity and total suspended solids levels are worst during spring months and this coincides with the sediment laden releases from the rice fields. Various hypothesis tests identify the watersheds for which the BMPs have been effective and also those for which improvements could not be established. Consideration of inherent uncertainties in various processes allows an objective assessment of the improvements that have been brought about by the implementation of various BMPs in Mermentau and Vermilion-Teche River basins.