| Although the government has taken a series of water quality protection measures in headwater watershed recently,non-point source pollution(NPS)is the major reason resulting in the deterioration of water quality.Non-point source pollution is very difficult to control in watershed water environment management,because the discharge of non-point source pollutants is random,multi-sources and multi-paths.Clarifying “who should be controlled in priority” and “to what degree are the various pollution processes controlled” these two questions is important for effective control of water pollution,especially for the headwater watersheds,where the non-point source pollution is characterized by complex multi-sources.We selected the Hengxi reservoir watershed in Ningbo,a drinking water source area with clear requirements for water quality,as a case study.We collected the data of hydrology,meteorology and Yearbook of Social and Economic Statistics,and monitored the dynamic changes of water quality in the watershed.The contribution rates of different pollution sources to river total nitrogen(TN)pollution were quantified,through the inversion and traceability of TN load in rivers,by adopting the combination of LOADEST model,two-parameter digital filtering baseflow load separation model and one-dimensional water quality equation –exported coefficient coupling model.At the same time,taking agricultural production as the key point of NPS control,we analyzed the net nitrogen input and the proportion of nitroge from different source in farmland by adopting a regional nitrogen inputoutput balance model.Based on field experiments,we further studied the influence of different water and fertilizer management measures on the nitrogen discharge of farmland.Based on the targets of nitrogen emission reduction required in headwater watersheds,we analyzed the optimal scheme of farmland water and fertilizer management of the main cropping systems.We also explored the influences of landscape patterns on river water quality in NPS areas by adopting regular/partial redundant analysis;we proposed to control NPS by further regulating the watershed landscape pattern after optimizing the management of water and fertilizer in farmland.The main results were as follows.(1)We calibrated and validated a LOADEST model to simulate the change of river TN load based on the gydrological and water quality monitoring data in this watershed.The daily TN load of the monitoring section at the outlet of the watershed was simulated,and the annual average TN load of the Hengxi reservoir headwater watershed was 55.46±11.86 t·year-1 during the period of 2015 to 2019;the higher average monthly TN load occurred from June to September,and the monthly TN load was highly correlated with monthly rainfall.Further,we established a two –parameter digital filtering baseflow load separation model to separate the baseflow and surface runoff non-point source TN loads.During this study period,the annual average contribution of baseflow and surface runoff to TN load was 27.23±5.93t·year-1 t·year-1 and 28.24±5.98 t·year-1,and the annual average contribution rates were 49.06±0.91% and 50.94±0.91% respectively.(2)We used a one-dimensional water quality equation-export coefficient coupling model to quantitatively analyze the surface runoff TN load of different land uses.The results of Bayesian method showed that the TN export coefficient of surface runoff for each land use were followed in order: dryland(34.65±0.40 kg·ha-1·yr-1),residential land(26.91±0.38 kg·ha-1·yr-1),garden land(16.89±0.88 kg·ha-1·yr-1),paddy field(12.30±0.41 kg·ha-1·yr-1),forest land(6.33±0.25 kg·ha-1·yr-1).The contribution rates of different land use to TN load through surface runoff were as follows: forest land(41%)> paddy field(22%)> residential land(15%)> dry land(12%)> garden land(10%).Then,we calculated the different nitrogen input components of farmland(including paddy field and dry land)by using the agricultural net anthropogenic nitrogen input model.The results showed that the proportion of each component to the agricultural net anthropogenic nitrogen input were followed in order: chemical fertilizer application(81.90%),atmospheric nitrogen deposition(17.60%),agricultural nitrogen fixation(7.80%),human feces returning to the field(1.20%),animal feces returning to the field(0.20%)and seed nitrogen input(0.10%).(3)The field experiment results showed that the combination mode of “alternate dry-wet irrigation + suitable for soil nutrient management(SSNM)+ slow controlled fertilizer replacement” was the optimal irrigation and fertilizer management mode for TN runoff loss reduction under the rice – potato rotation system in this region.Compared with traditional irrigation and farmer fertilization treatment,the average annual TN runoff loss under the optimal water and fertilizer mode treatment was reduced by 16.54±3.34 kg N·ha-1,with a reduction ratio of 52.34±1.38%;the rice yields only decreased by 0.78% in 2019,and even increased in 2018;the potato yields also decreased by 4.54±5.13% during the study period.If the optimal water and fertilizer management mode of farmland are adopted in this watershed,the TN load of surface runoff can be reduced by 22.28%.(4)Taking the reservoir reaches the water quality standard of class II as the target,the water environment capacity of TN,calculated by Dillon model,was 42.46 t·year-1,49.75 t·year-1,and 60.59 t·year-1 under the annual runoff flow with 90%,75% and50% guarantee rate,respectively.After deducting the TN load from the base flow,the input TN load from the forestland surrounding the reservoir and the annual average input TN load from the surface runoff of river headwater watershed,the excessive TN discharge of surface runoff was 12.61 t·year-1,8.90 t·year-1,3.38t·year-1,respectively;thus,based on the current TN discharge situation,the proportion of TN flow into the river by surface runoff should be reduced by 40.74%,30.58%,15.47%,respectively.Thus,we concluded that under the premise of guaranteeing high crop yield,it is difficult to achieve the water quality target under the 75% guarantee rate only by adopting water and fertilizer management measures in farmland.(5)Maintaining a high level of agricultural production while ensuring the water quality meets the standard is a realistic development goal for many regions.Therefore,it is necessary to integrate multi-approach and multi-method water quality control measures while controlling agricultural NPS pollution to a large extent.This study explored the spatio-temporal effects of landscape pattern on river water quality.Among the topographic metrics,landscape composition metrics and landscape configuration metrics,the spatial scale effect of physiography impact on water quality was the most significant,while the impact of landscape configuration on water quality had the highest seasonal sensitivity.The key landscape metrics at subwatershed scale and buffer scale were the largest patch index of farmland(LPIfar)and the landscape shape index of forest(LSIfor),respectively.Furthermore,we used the non-parametric abrupt change analysis to reveal the thresholds of key landscape metrics causing the abrupt change of TN concentration.The results showed when the LPIfar is less than 7.0% and/or LSIfor is greater than 5.5 in this region,the risk of river water quality deterioration with the increase of TN concentration will increase significantly. |
PDF Full Text Request