Water Environment Simulation And Prediction Research In Xinlicheng Reservoir Watershed Based On Semi-distributed Hydrological Model HSPF

With the rapid development of social economy, the influence of human activitieson water environment is growing, and it will cause water environment deteriorationwhen the degree of damage is greater than water self-purification ability. China isfacing three water problems, namely, drought, floods and water environmentdeterioration. The most direct and main manifestations are the shortage of waterresources and water quality deterioration.“Pollution-induced water shortage” hasbecome an important factor restricting our country national economy development ina sustainable, rapid and healthy manner. Water pollution accident have not onlycaused serious destruction of ecological environment, economic losses and societyinstability, but also directly influenced people’s health. Water environment simulationis the main technical method for carrying out water environment evaluation and waterquality prediction and early warning, and making water environment planning andwater pollution control projects. The Xinlicheng Reservoir watershed was regarded asthe study area, and the water environment situation was analyzed, andsemi-distributed Hydrological Model HSPF was applied for water environmentsimulation, and the non-point source pollution load was quantified, and the spatialconfiguration of best management practice was discussed by using model method, andwater environmental capacity of Xinlicheng Reservoir was calculated by simulatedresults of pollution load, which provided decision basis for water environmentprotection and management.(1) Water environment situation analysis in Xinlicheng Reservoir watershed,including analysis of limnology characteristics, water environment quality assessment,eutrophication assessment and attribution analysis of water environment problems.Limnology characteristics analysis mainly includes the relational features of waterquality variables, temporal and spatial variation characteristics of water quality, hydraulic retention time and water temperature stratification structure. Significantanalyses of reservoir water quality were applied for zones, seasons and years by SPSSstatistical test method. Significant differences (P<0.05) were observed betweenseasons for SD, T, Chl-a, FC, DO, pH, NH3-N and TP. Significant differences (P<0.05)were detected among years for Chl-a, DO, pH, NH3-N, TN, TP and CODMn. Therewere no significant differences among zones for all water quality variables expect SD.Correlations of water quality variables were analyzed by Pearson correlationcoefficients. SD was negatively correlated with Chl-a (P<0.05), T, FC and pH(P<0.01), and positively correlated with DO (P<0.01). T was positively correlatedwith Chl-a, FC and pH (P<0.01), and negatively correlated with DO (P<0.01). Chl-awas positively correlated with FC (P<0.01), and negatively correlated with DO(P<0.01). FC was positively correlated with pH (P<0.01), and negatively correlatedwith DO (P<0.01). DO was negatively correlated with pH (P<0.05). NH3-N waspositively correlated with TN (P<0.05) and TP (P<0.01). Factor analysis revealed thatreservoir water quality according to the composition could be interpreted as two maincomponents, biological factor and trophic factor, and according to score sorting waterquality during dry season was better than that during wet season. According toreservoir flow features, the average hydraulic retention time was about518d. Thereservoir water temperature showed that stable stratification structure was absent, andonly epilimnion existed. The water temperature changed with the seasons and theinfluence range of temperature was from surface to bottom. Based on the improvedNemerow comprehensive pollution index method, the water quality evalution resultsshowed that reservoir water quality level belonged to class Ⅱ or Ⅲ, regarded asbetter. Along the longitudinal axis from the upstream to the dam, water quality tendedto be better, which indicated the reservoir had self-purification ability. Based on thecomprehensive nutrition state index menthod, the eutrophication assessment resultsshowed that comprehensive nutrition state indexes during wet season were higher thanthat during dry season, and that of upstream fluvial zone were higher than that ofdownstream lacustrine zone. The water body most of time was mesotrophic. The mainfactors affecting reservoir water environment includes rainfall-runoff, rapid economicand social development and municipal facilities shortage, land use pattern, multiplesource pollution and bank-line hardening, and sediment release.(2) Water environment simulation in Xinlicheng Reservoir watershed based on HSPF model. Based on the watershed spatial and attribute database, HSPF model forXinlicheng Reservoir watershed was established to simulate the processes ofhydrology, sediment, water temperature and water quality and to be calibrated andvalidated, whose simulation effects were evaluated by relative error (Re), correlationcoefficient (R) and Nash-Sutcliffe coefficient (Ens). Based on PEST automaticcalibration and manual adjustment, hydrological simulation was calibrated andvalidated for annual flow, seasonal flow, monthly flow and daily flow respectively.The relative errors of total annual flow during calibration period and validation periodwere both within±10%, respectively-0.41%and-9.31%, which indicated simulationeffects were very good. The relative errors of seasonal runoff volume for winter andsummer were among-10.86%~18.89%during calibration period, and those wereamong-19.36%~-5.90%during validation period, which indicated simulation effectswere reasonable or better. The correlation coefficient and Nash coefficient of monthlyrunoff volume were respectively0.984and0.95during calibration period, and thosewere respectively0.936and0.87during validation period, which indicated simulationeffects were good or better. The correlation coefficient and Nash coefficient of dailyflow were respectively0.932and0.87during calibration period, and those wererespectively0.927and0.86during validation period, which indicated simulationeffects were very good. The correlation coefficient and Nash coefficient for sedimentsimulation were respectively0.89and0.73during calibration period, and those wererespectively0.982and0.80during validation period, indicating simulation effectswere good. The correlation coefficient and Nash coefficient for water temperaturesimulation were respectively0.983and0.96during calibration period, and those wererespectively0.919and0.84during validation period, indicating simulation effectswere good or better. The correlation coefficient and Nash coefficient for NH4-N wererespectively0.961and0.85during calibration period, and those were respectively0.877and0.66during validation period, indicating simulation effects were good andgeneral respectively. The correlation coefficient and Nash coefficient for NO3-N wererespectively0.958and0.78during calibration period, and those were respectively0.927and0.82during validation period, indicating simulation effects were both good.The correlation coefficient and Nash coefficient for TN were respectively0.892and0.61during calibration period, and those were respectively0.865and0.64duringvalidation period, indicating simulation effects were both general. The correlationcoefficient and Nash coefficient for TP were respectively0.979and0.92during calibration period, and those were respectively0.932and0.79during validationperiod, indicating simulation effects were good or better. Overall, the simulationeffects for HSPF model were good and met the requirements of simulation accuracy.(3) Non-point source pollution load analysis in Xinlicheng Reservoir watershedand scenario simulation. Based on calibrated and validated HSPF model, inflownon-point source pollution loads of2010~2013were estimated, the temporal andspatial variation characteristics of non-point source pollution load were analyzed, theinfluences of land use changes on non-point source pollution load were explored, andthe implement and spatial configuration of best management practices (BMPs) wereresearched through the scenario simulation. Without regard to point source pollutionload, the inflow annual average loads of sediment, TN and TP were respectively26225t,200t and24.39t. Non-point source load variations were consistent withrainfall, and its peaks generally appeared during wet season (from June to September).The ratios of sediment load during wet season to annual load were71%~95%,similarly49%~55%for TN and75%~94%for TP. The largest non-point sourcepollution load was located in the top upstream subbasin, where should be the focus ofnon-point source pollution management. Non-point source pollution load is closelyrelated to human activities. Agricultural land and urban or built-up land have beenland use types which are the biggest contributors to non-point source pollution.Influences of BMPs (filter strips, constructed wetland, grassed swale and bioretention)on non-point source pollution load were simulated by BMP module. The resultsrevealed that comprehensive reduction effects of filter strips on non-point source loadwere best, constructed wetland came second, and the interception effects of grassedswale and bioretention on sediment particle were better. The subbasin unit withinHSPF model could be considered as a BMP management unit, and the same ordifferent BMPs could be selected in the upstream and downstream units according tothe characteristics of land use types, and the spatial optimum allocation of BMPscould be simulated by setting the percentage of BMP implement area.(4) Water environmental capacity calculation in Xinlicheng Reservoir. Based onHSPF model, this research selected2013as status quo year, and forecasted TN and TPpollution load and water environmental capacity during2020with differentprecipitation frequency (P=20%as wet year, P=50%as normal year, P=75%as dryyear). TN and TP inflow loads in status quo year were respectively227.30t and28.74t, and their water environmental capacities were respectively2526.45t and525.33t. Their infow loads in2020with the frequency of P=20%were respectively268.57tand37.36t, and their water environmental capacities were respectively2210.09t and441.17t. Their infow loads in2020with the frequency of P=50%were respectively244.65t and21.95t, and their water environmental capacities were respectively2175.89t and459.33t. Their infow loads in2020with the frequency of P=75%wererespectively218.74t and9.29t, and their water environmental capacities wererespectively2177.96t and451.81t. The water environmental capacities in presentsituation and forecast level years changes dynamically, and water environmentalcapacities of each pollution index are greater than the pollution loads in the sameperiod. Dilution capacity is less than self-purification capacity, which indicates theself-purification become the dominant position in the reservoir water purificationprocess. The water environmental capacities during the flood season are greater thanthat during the dry season. Different values of comprehensive pollutant degradationcoefficient affects the magnitude of self-purification capacity. The waterenvironmental capacity of Xinlicheng Reservoir is related with precipitationfrequency, water process, pollution load, comprehensive pollutant degradationcoefficient, etc. Efforts to improve the water environmental capacity includeecological protection measures, environmental control measures and ecologicalrestoration measures, which comprehensive measures should be taken.From the above, in this paper the Xinlicheng Reservoir was selected as thereseach object, and semi-distributed hydrological model HSPF was applied to waterenvironment research for the river basin. It was the first time to systematicallysimulate the process of hydrology, sediment, water temperature and water quality, andafter calibration and verification the simulation effects were good. It was the first timeto apply HSPF-BMP module to watershed best management practices in XinlichengReservoir, and spatial optimum allocation of BMP was discussed through scenariosimulation. It was the first time to use HSPF pollution load results for calculatingwater environmental capacities in Xinlicheng Reservoir. This study provided thereferenced example in watershed simulation and management with HSPF model.