The Mechanism Of Soil And Nutrient Losses And Their Modeling In Purple Hilly Area

Mechanism and simulation of soil and nutrient losses have been an interdisciplinary focus with evolvement of soil erosion and expansion of non-point source pollution in recent years. Purple soil is a lithologic soil with high productivity but characterized by severe erosion. It is mostly distributed in upper stream region of Yangtze River. The region is an important food bases in China, however, unreasonable reclamation and land use methods have accelerated soil erosion process and losses of soil nutrient. This had not only caused the irreversible loss of soil resources and thus limited the increase of land productivity, but also threaten downstream water environmental quality and water conservancy. Three small agro-watersheds with different catchments area, physical characteristic and land use type were employed to investigate mechanism of soil and nutrient losses in purple soil region in the middle of Sichuan province combining closely with Euro-Chi cooperation projects. Three-year natural rainfall events, corresponding runoff process, soil, and nutrient loss process together with change of planting pattern were sequentially observed. The mechanisms of runoff generation were discovered from watersheds scale. The characteristics of soil and nutrient losses in response to rainfall type, spatial scale variability, land use change and planting pattern were set out. Based on the response and the key influential factors, the forecasting model was put forward. The objectives were to serve for reducing, controlling soil, and nutrient loss and evaluating the status of non-point source pollution in agro-watersheds of this region. The main conclusions and innovations were following. 1. The mechanisms of runoff generation in purple hilly area could be characterized by saturation-excess with the components of surface runoff, subsurface runoff and base flow. The process of runoff behaved a compound type of single-peak-shape hydrograph of surface runoff with higher flow rate and shorter duration, which response the high intense rainfall process, and the following multi-peak-shape hydrograph of subsurface runoff and base flow with lower flow rate and longer duration. The peak period during the processes of subsurface runoff and base flow was 24 hour. 2. The reasonable depth for soil sampling during forecasting the loss of soil nutrient by surface runoff was demonstrated through the measurement and analysis of nutrient content among 0~2.5 cm, 0~5 cm and 0~10 cm soil layers. The depth of 2.5 cm below was proved to be reasonable for taking soil sample to forecast the loss of soil nutrient by surface runoff in woodland-dominated watershed, while 0~5 cm was feasible for other watersheds but to forecast soil phosphorus loss, which should be 2.5 cm below. 3. Sediment concentration behaved three types as indention and descending trend, logarithm-decreasing trend and peak-shape undulation trend in purple soil watersheds. The sediment concentration was highest (9.58±10.90 g.l-1) under the erosion of IM-type rainfalls (the duration is discontinuous, having obvious segmentation phenomenon during rainfall process). While the lowest sediment concentration (1.13±0.58 g.l-1) was found during the erosion of PP-type rainfalls (the high-intensity-rainfall duration appeared in the prophase of rainfall process). The peak flow rate was proved to be the main influential factors of soil loss in woodland-dominated watersheds. While the main factors are rainfall erosivity factor in other watersheds. Based on the flow rate, power function was put forward to forecast sediment concentration in surface runoff during erosion of a rainfall event. 4. Nitrate nitrogen was the main form of N loss, which accounting above 80 percent of N loss by surface runoff. The peak flow rate was proved to be the main influential factors of nitrate nitrogen losses in woodland-dominated watersheds, while the main factors in other watersheds are total runoff amount. Based on the flow rate, Logarithm-decreasing function was put forward to forecast nitrate nitrogen concentration in surface runoff during erosion of a rainfall event. The loss of soil ammonium nitrogen was strongly affected by the change of plantingpattern and vegetation cover degree. The concentration of ammonium nitrogen in surface runoff in interplant pattern was much higher than that of in monoculture pattern. With the increase of vegetation cover degree, the loss of soil ammonium nitrogen increased in woodland-dominated watershed but decreased in farmland-dominated watershed. The average concentration of nitrogen in surface runoff in farmland-dominated watershed and the average concentration of water-solute phosphorus in watersheds with various landscape structures were remarkable high, which exceeded 10 to 24 times than nitrogen critical value of eutrophication and 2 to 19 times than phosphorus critical value of eutrophication, respectively. 5. Relationships of the concentrations of ammonium nitrogen and water-solute phosphorus in surface runoff and flow rate both took on power function decreasing in woodland-dominated watersheds, while they showed logarithm decreasing and power function increasing relationship with sediment concentration in farmland-dominated watersheds, respectively. During the erosion of two rainfalls with a short interval period (less than 5 days), the concentration of sediment and water solute phosphorus in surface runoff during erosion of the latter rainfall were lower than that of the former. In farmland-dominated watershed, the concentration of nitrate nitrogen in surface runoff of the latter rainfall was significantly higher than that of the former. No significant difference was found for the loss of ammonium nitrogen in the two rainfalls.