| Agricultural production plays a leading role in the hilly regions of southeast China,thus excessive fertilization or other irrational agricultural management practices in these areas increased nitrogen and phosphorus loss. Therefore, it is very urgent and essential home and abroad to improve water environment by recognizing the formation and evolution of non-point source pollution(NPS) clearly and deeply. As the formation and transportation of NPS is very complex and full of regional characteristics, the distributed computer model based on the process of NPS has been an effective means to simulate, evaluate and quantitatively monitor NPS. Agricultural Policy and Environmental eXtender, APEX can quantitatively study the NPS output features in higher spatial resolution because of its discrete method of watershed-to-subarea. As ZhongTianshe Watershed belongs to the typical small agricultural watershed and is the drinking water source of Liyang city and the conserving water of its downstream reservoirs, the influence on water quality brought by frequent agricultural activities within ZhongTianshe Watershed has raised more and more focus. So distributed computer model has become a useful and meaningful tool to quantitatively monitor and quickly forecast NPS.With the support of Chinese National Natural Science Foundation Project(grand number: 41271382): Study on Distributed Modeling of Non-point Source Pollution Processes in Plain River Network Area, and based on APEX and the Water Environment Experiment Station of Nanjing Normal University, this paper has simulated and quantitatively analyzed the NPS output features in finer spatial scale by adopting both model simulation and experimental observation, and discussed the applicability of APEX model for simulating NPS in typical small agricultural watersheds in Taihu Lake,and analyzed the temporal and spatial output characteristics of NPS load and related quantitative calculation methods in particular landuse type. The chief work and results of this research include the following aspects:1.After field investigation and selection of sample point, the collection and analysis of samples in streamway and interior drainage area has been carried out regularly, and eventually assisted to complete the study of nitrogen and phosphorus output features in overland with a particular landuse cover and different subbasin.2.To eliminate the unreasonable homogenization brought by complete automatic discretization only using DEM, this paper .combined the computer automation and artificial modification to carry out the spatial discretization. Based on the in-depth analysis of APEX modeling theory, including its distributed spatial discretization, and quantitative calculation method of NPS, and integrating landuse, soil and topography of the study area, this paper manually accomplished stream network extraction and subarea delineation. These manually divided subareas descripted and simulated the emergence and development of NPS more objectively and accurately. After using basic data to extract model parameters, APEX input files were established. Finally, with the observing data, the surface runoff, sediment yield and nutrients transportation from January 2008 to December 2016 were simulated and evaluated.3、The accuracy assessment of simulated results of surface runoff, sediment yield and nutrients show that, APEX supplied satisfactory modeling values for surface runoff,sediment yield, which means APEX could reflect the streamflow and sediment loss of study area ideally. APEX performed slightly poorer when simulating the total nitrogen(TN) and total phosphorus(TP) load, the quantity disparity between simulated and observed values was relatively large, but the changing trend between both was similar, which indicates that APEX could represent TN and TP load to a certain extent.And we checked the measured data and surveyed the study area to find out the possible reasons why the few simulated results diverged from observed ones. Iin general, the APEX model shows good capability and applicability for simulating NPS in ZhongTianshe Watershed. The manually divided subareas adopted in this study were capable of analyzing the loading features of nitrogen and phosphorus at higher resolution especially for different overlands with a particular landuse cover.4、The temporal and spatial characteristics of total nitrogen and total phosphorus in ZhongTianshe River Basin were studied respectively in the whole watershed,different subbasins, and different overlands with a particular landuse cover. The analysis of TN and TP load characteristics in the study area included following aspects:(1) For the whole ZhongTianshe Watershed, TN load amount(kg) was much higher than TP output(kg), which means nitrogen load accounts for a substantial part of nutrient loss. In terms of temporal change, the annual output of nitrogen and phosphorus nutrients had been increasing year by year, and nutrients load amount was highest in summer. In terms of spatial distribution, the spatial features of TN were fairly close to those of TP. Nutrients load mainly concentrated in the middle and lower region of ZhongTianshe Watershed, and nutrients output of the fields like cropland, orchard, tea plantation and residential communities which spread along the main river channel were enormous. Some tea plantation,orchard and construction land which scattered in the upstream area or the regions far away from main river channel were also the major source of NPS.(2) For the upper, middle and lower region of ZhongTianshe Watershed, the vertical comparison in the same month and horizontal comparison in different months of TN and TP concentration(mg/L) were carried out. In the same month, the simulated or measured TN and TP concentration in the upstream area were slightly lower than those in the downstream area. In different months, the peak concentration of TN in upstream or midstream or downstream area appeared in June,September and December respectively, while the peak concentration of TP in upstream or midstream or downstream area appeared in June.(3) Based on the measured data of overland flow caused by four rainfalls which occurred sequentially on June 23rd 2016, July 4th 2016, September 15th 2016 and September 29th 2016, the TN and TP output features in different overlands corresponding to a particular landuse cover were studied. The typical overlands selected by this research included 6 types: pine forest, chestnut orchard, bamboo forest,shrubbery,tea and cropland. Peak concentration of TN and TP in six types of overland mentioned above all appeared on June 23rd, and concentration of TN and TP continuously decreased in the following three days. However, average concentration of TN in cropland increased again on September 11th as a result of massive fertilizer applied to the crops in the end of August. Due to the regular fertilization managements not conducted within the observing period or the overland belonging to primary woodland, the average concentration of TN and TP in the other five types of overland kept decreasing in the following three days. |