Inverse estimation of nonpoint source export coefficients for total nitrogen and total phosphorous in the Kako river basin

Abstract

Export coefficients of nutrients from nonpoint sources are significant parameters in managing nutrient loading into coastal waters. Estimating the export coefficients based solely on field measurements requires water sampling at multiple sites with different land use characteristics in a watershed, which leads to high sampling costs. In this study, we propose a method that combines simulation tools and observed data at a single point to inversely estimate the export coefficients for total nitrogen (TN), and total phosphorous (TP) from different land uses in one of the most important watersheds of western Japan. Combining the results from a physically based distributed runoff model with nutrients’ observed concentrations at a single point near the estuary, multiple linear regression (MLR) models for predicting nutrient loads in the mainstream were developed. Geographical and observed meteorological data were used to simulate the total direct runoff discharged from each land use type during rainfall events, adding time series and climate deviation detection into the MLR. Observed data from 24 months between April 2010 and March 2012 were split into two sets to construct two MLRs for each nutrient and obtain two sets of annual export coefficients. The applicability of the obtained export coefficients was tested in numerical simulations for the studied area using a physically distributed hydro-chemical model. Verification and validation at different time resolutions of the export coefficient sets were performed and found to be highly satisfactory for estimating loads in the mainstream for both nutrients. The presented method implementation reduces the amount of observed data, jointly avoiding data transformation and excessive data filtering issues. Additionally, it has shown to be effective in considering the influence of the typical extreme climatic events in East Asia (typhoons, rainy season, etc.) on the estimation of the export coefficients.