Abstract:
Abstract: With the advantages of fewer model parameters, fewer difficulties in obtaining necessary data, and more combinations with geographic information system (GIS), remote sensing (RS) technologies, the export coefficient model (ECM) is an applicable model for estimating non-point source (NPS) pollution loads. Based on the traditional ECM, the model can be improved by considering the impacts of precipitation and topographical factors on NPS pollution. The simulation accuracy of improved ECM (IECM) on NPS pollution is higher and it has been successfully used in China, such as the upper reaches of the Yangtze River basin, the Miyun Reservoir basin, and the Baoxiang River basin in Yunnan Province. Beijing-Tianjin-Hebei region, namely Beijing-Tianjin-Hebei, located in North China, is an economically developed region with a dense population and fast urbanization. There is a serious lack of water resources in this region and the water quality is bad. Seventeen drinking-water reservoirs in Beijing-Tianjin-Hebei region were selected in this study, and the meteorological, topographical, land use, social and economic data of their watersheds in the year of 2010 were collected. The IECM was applied to estimate the loads and intensities of total nitrogen (TN) and total phosphorus (TP) of NPS in the 17 reservoir watersheds, and the sources of TN and TP were also analyzed. The results were as follows: (1) TN and TP load had the same spatial pattern, namely the TN load was large, and the TP load was large too. The average ratio of TN to TP load was 6.73. (2) TN and TP load intensities had obvious spatial heterogeneities. The TN load intensity was 1.04-24.91 t/km2·a, ranging from the minimum value in Yunzhou Reservoir watershed to the maximum value in Taolinkou Reservoir watershed. The TP load intensity was 0.11-3.89 t/km2·a, with the minimum value in Yunzhou Reservoir watershed and the maximum value in Yanghe Reservoir watershed. The TN and TP load intensities were both relatively high in Taolinkou, Yanghe, Daheiting, Qiuzhuang Reservoir watershed, indicating that the reservoirs located in the northeast Beijing-Tianjin-Hebei region (Tangshan City and Qinhuangdao City) had higher risk of nitrogen and phosphorus pollution. (3) Based on the source analysis results, a conclusion could be drawn that the order of the mean contribution to TN load was land use > rural life > livestock, while the order of the mean contribution to TP load was rural life > land use > livestock. Rural life was an important source of both TN and TP load in this region. (4) We used the water eutrophication index (EI) of drinking-water reservoirs which was published by Haihe River Water Conservancy Commission in 2010 to test the modeling results. EI was significantly correlated to TN and TP load intensities. Person correlation analysis was employed and the correlation coefficient between TN load intensity and EI and that between TP load intensity and EI were 0.562 (P<0.05) and 0.558 (P<0.05), respectively. Model test results showed that the IECM has a high application value, and the TN and TP load of NPS can be well estimated by IECM at large watersheds which lack large numbers of measured data of water quality.