Study On Non-point Pollution In Tao River Basin’s Agricultal Field Based On AnnAGNPS Model

With point source pollution including industrial waste water and city sewage has been effectively controlled, agricultural non-point source pollution has become the primary source of pollution and the first major pollution source of water environment. In our country, water pollution control presents that the point source adjustable space is becoming smaller, however, non-point source adjustable space has an increasing trend. Therefore, more and more researchers focus on occurrence mechanism and control measures of agriculture non-point source pollution. As one of the headwaters of Gan River which is the biggest tributary of Poyang lake basin, Tao River basin’s rural population accounts for about81.2%, so agricultural non-point source pollution is the most important source to water pollutions. Therefore, it has highly theoretical and practical significance, also the demonstration effect to study on the occurrence mechanism and loss characteristics of agricultural non-point source pollution in Tao River basin.In this paper, means of field investigation, laboratory analysis and querying historical data and existing literature outcomes were applied in building basic database in Tao River basin. Method of qualitative analysis and quantitative calculation combined together to carry out the study of the Tao River basin agricultural non-point source pollution. The research steps are as follows:1. Nitrogen balance change model was established at two scales in Tao River basin and study area by means of Nitrogen balance income and expenditure, and was used for quantitative analysis of the regional nitrogen input, output, and net increase. It provided not only important Parameters and boundary conditions for the followed up establishment and application of the AnnAGNPS model to quantitative simulation of occurred mechanism and loss characteristics agricultural non-point source nitrogen, but also a scientific basis for the interception of nitrogen. The results showed that average annual nitrogen input in Tao River basin and study area was32992t/a and6957t/a from2006to2008, among which, chemical fertilizer input nitrogen accounted for44.94%and44.05%, respectively, was the main source of nitrogen input; Crop nitrogen fixation accounted for49.89%and50.78%of the total nitrogen output, was the main source of output, respectively; Nitrogen input increment was much higher than nitrogen output increment, net nitrogen increase was14612t/a and3167t/a with annual growth in Tao River Catchment (9.78%and4.28%) and in study area (8.27%、2.83%), showing a significantly upward trend. The studies of net nitrogen and its whereabouts showed that the share ratio of net nitrogen increment in Tao River Catchment and study area were water body (64.18%and62.12%), soil (30.86%and32.96%) and living creature (Organism)(4.96%and4.93%), respectively. So it shows clear that, there were two existence types of residual nitrogen:First, it flowed into the river through surface and groundwater runoff and polluted the water body; Second, it was left in the soil, causing soil pollution. Agricultural non-point source nitrogen is the main watershed nitrogen sources, so studies on the topic are becoming significant and urgent.2. Based on3S technology, AnnAGNPS model was established to simulate agricultural non-point source pollution in study area and its applicability was calibrated and validated. The calibration and validation processes and results showed:1) on different time scales of year, month and day, the simulation accuracy of the model was decreasing Sequential. With time scales of year or month, model showed higher simulation accuracy than that of day. Maybe, it dues to that, the actual occurrence of heavy rain is a relatively short-term process, usually lasting less than one day (24hours). But when daily simulation has to be carried out, because of limitations of meteorological data and experimental conditions, it could lead to weaken rainfall intensity, prolong rainfall time and impact accuracy of the model;2) AnnAGNPS model simulation accuracy order of surface runoff, sediment erosion and total nitrogen output from high to low were as follows:surface runoff> sediment erosion> total nitrogen output. The reason for this phenomenon might be the migration of sediment and total nitrogen in runoff for the carrier, the error in turn accumulate and be enlarged;3) Compared with domestic and foreign counterparts, it can be drawn that main factors which affected the accuracy of the model were basin-scale, climatic characteristics and topography of the watershed;4) Calibration and validation of year and month scales meet with model accuracy requirements, the relative error of day scale was less than50%. Therefore, the AnnAGNPS model, which established in this paper was suitable for the simulation and prediction of long-term non-point source pollution, but wasn’t suitable for risk assessment and risk management of non-point sources.3. Based on rainfall from2001to2008, the AnnAGNPS model above was applied in quantitative prediction, spatio-temporal analysis and evaluation of best management practices. The results showed that:1) From2001to2008, the average annual runoff depth was579.74mm, sediment loss was152.66X103t/a and non-point source nitrogen was2781.73t/a; monthly runoff, sediment loss and non-point source nitrogen in flood season (April to August) accounted for64.30%,85.64%and86.34%of the whole year, respectively, reflecting that, runoff, sediment loss and non-point source nitrogen output mainly caused by high intensity rainfall;2) The high similarity of inter annual variability curve of rainfall, runoff, sediment and total nitrogen and the correlation coefficient (rainfall-runoff, rainfall-sediment, rainfall-nitrogen, runoff-sediment and runoff-nitrogen) were between0.7825and0.8421, showed a close relationship between the four parameters;3) variability curve of rainfall, runoff, sediment and total nitrogen in flood season also had similarly trend, the correlation coefficient were between0.6022and0.9855, showing a close relationship between the four and reflecting the correlation coefficient had bigger fluctuation in flood season than that of the whole year. The correlation coefficient of rainfall-runoff was0.9855, showed that flood season rainfall runoff had a very good linear relationship, while the R2of runoff-nitrogen was0.6022, reflecting non-point source nitrogen output were highly uncertain in flood season. Therefore It shows that, in all year round, the crop growth stage, fertilization, management mode and the nitrogen intercept ability could be different, so the non-point source pollution levels of nitrogen can also be different correspondingly, even under the same rainfall situation;4) In different hydrological years, runoff difference was small, but the sediment and non-point source nitrogen output difference were bigger. From2001to2008, runoff, sediment and non-point source nitrogen in wet year were1.4times,3.5times and1.8times as that of dry year, respectively. It can be seen that the sediment erosion was more sensitive to rainfall than runoff and nitrogen, at the same time showed indirectly that, high intensity rainfall was the main factor of sediment erosion;5) By identification of critical source areas of nitrogen in study area, nitrogen non-point source pollution critical areas was confirmed as Level5.(The highest Level);6) Balanced fertilization was most efficient to cut down the non-point source nitrogen pollution, vegetation buffer had obvious effect on runoff, sediment erosion and nitrogen loss.