Modeling spatial-temporal variation of particulate phosphorus at regional scales

Abstract: Phosphorus is the dominant nutrient causing the eutrophication in the Three Gorges Reservoir Area of China where there is abundant rainfall. The related studies report that about 85% of phosphorus exists in particulate phosphorus (PP), and there are rapid and dramatic spatial and temporal variations in PP load particularly during the period of intensive rainfall. To estimate the spatial-temporal variation of PP load, it is important to simulate soil erosion and sediment-associated transport. For mountainous watershed, hillslope sediment delivery ratio (HSDR) coupled with empirical soil erosion model is helpful to improve the capability of simulating sediment delivery and PP pollutant load. Considering the runoff characteristics from upslope and downslope during the transportation course of eroded soil, and the time varying characteristics of runoff caused by climate conditions such as rainfall, a runoff connectivity index (RC) by modifying flux connectivity index (IC) was proposed to define the variable HSDR for the watersheds with inhomogeneous climatic conditions. And by combining the HSDR with the RUSLE and the enrichment ratio of phosphorus in soil, the spatial-temporal distribution model of PP load was developed. The model was applied in Xiaojiang watershed of the Three Gorges Reservoir Area, and the spatial-temporal variation of sediment yield was simulated and analyzed. The results showed that simulated values of sediment yield were consistent with the observed data, the range of relative error was within -10.77%-13.81%, and the Nash-Sutcliffe coefficient and the relative root mean square error were 0.93 and 0.106, respectively, while the Nash-Sutcliffe coefficient and the relative root mean square error were 0.81 and 0.124 respectively using the HSDR from IC. In addition, it could be seen that the model performed well with a determination coefficient of 0.98. For PP load, the range of annual value of relative error was within -7.77%-14.73% and the relative error was comparatively small, and hence the simulation results were still satisfactory. From the distribution of PP load, it could be seen that the most serious regions of PP load caused by the accumulation of erosion sediment along rivers were Puli river, Dong river downstream, Nan river coast and Pengxi river, where the land use types were mainly dry land, paddy field and other fields. For the Taoxi river basin, upstream of the Dong river and Nan river downstream where were mainly grassland and forestland, the PP load was relatively low. According to the analysis of the proportion of different land use types, dry land was the largest, with 59.97%, the second was grassland with 19.75%, paddy field, forest land and water area were relatively small, with 8.12%, 6.31% and 5.62% respectively, and unutilized land and residential land accounted for 0.12% and 0.11% respectively. These were consistent with reported results. Thus, the calculation precision of the developed model in this paper has been greatly improved and the model is feasible. The model can be used as a major tool to assess sediment yield risks and PP load at mountainous watersheds.