Freshwater Provisioning Ecosystem Service Quantification And Responses To Environmental Changes

Consideration of ecosystem services has recently emerged as an important tool in developing systems-based watershed management strategies.Freshwater provisioning(FWP)is one of critical ecosystem services that is sensitive to climate change and variability and agricultural management strategies.With the high demand for freshwater and its vital role in sustaining multiple ecosystem services,it is important to quantify and evaluate freshwater provisioning and its responses to environmental changes.In this study,freshwater provisioning was quantified annually from 1995 to 2013 for 13 watersheds in the Upper Mississippi River Basin(UMRB),using an index-based ecosystem service approach.Soil and Water Assessment Tool(SWAT)model for UMRB was setup,then the hydrology and water quality as well as sediment load were calibrated and validated using monthly observed data from US Geological Survey(USGS)gauge stations.Corn and soybean yield were also validated using annual measured yield reported by National Agricultural Statistics Service(NASS)county-level crop data.Based on model outputs,FWP and food provisioning services(FP)were quantified for 14 HUC4 subbasins in the UMRB.Three representative drought indices(standardized precipitation index(SPI),standardized soil water content index(SSWI)and standardized streamflow index(SSI))were also calculated for drought characteristic analysis and for evaluating drought implications on freshwater and food provisioning service in the UMRB.Then five Best Management Practices(BMPs)scenarios(No Tillage-NT,Cover Crops-CC,Filter Strips-F,Grassed Waterways-GW and Combination of all BMPs-CM)were simulated using SWAT.The effectiveness of each BMP in improving FWP was evaluated both for baseline climate(1975-2004)and projected near future climate(2021-2050)from two regional climate models(CCCma-CanESM2 RCA4,ICHEC-ECEARTH RCA4)under two Representative Concentration Pathway scenarios(RCP4.5 and RCP8.5).The main conclusions of this study are listed below:(1)The annual freshwater provision indices(FWPI)for all watersheds were less than one indicating that freshwater provisioning is diminished in the UMRB.The concentrations of sediment(TSS),total nitrogen(TN)and total phosphorus(TP)are the most sensitive factors that impact FWP in the UMRB.A significant linear relationship was observed between precipitation and FWPI.During wet periods freshwater provisioning generally decreased in the study watersheds,primarily because of relatively high concentrations and loads of sediment and nutrients delivered from nonpoint sources.The results indicate a need to improve water quality in intensively managed agricultural watersheds if the watershed management goal is to increase freshwater provisioning.(2)The SWAT model indicated a satisfactory model performance for both hydrology and water quality predictions.The coefficient of determination(r2),Nash-Sutcliffe efficiency(NSE)and Percent Bias(PBIAS)values for all 13 streamflow stations met with the standards(r2>0.6,NSE>0.5,PBIAS<25%)for hydrology evaluation.The monthly streamflow for most stations was fitted well with the observed data throughout the time period.Generally,the simulated TN,TP and TSS loads were well correlated with observed data.The r2 and PBIAS met with the standards(r2>0.6,PBIAS<70%)for water quality evaluation.The crop yields were also predicted well by the model.The PBIAS between simulated and observed corn yield were less than 10%.The PBIAS between simulated and observed soybean yield ranged from-18.5%~14.8%.(3)A relatively high correlation was observed from July to December between annual FWP and two drought indices(SPI12 and SSI12).The correlation has statistical significance with p<0.05.The higher correlation between annual FWP and SSI12 than SPI12 indicates a greater importance of hydrological drought than meteorological drought for FWP.The seasonal FWP was highly affected by the precipitation deficits and/or surplus in summer and autumn.The maximum r values between SPI3 and FWP for most watersheds ranged from 0.7-0.88 with p<0.05.For some agriculturally dominant subbasins,the FWP was significantly and negatively correlated with SPI3 in spring and summer with r ranging from-0.61~-0.51 and p<0.05,indicating a negative implication of precipitation deficits and/or surplus on FWP during this period primarily because of the improved water quality conditions during dry periods and declined water quality conditions during wet periods.(4)A greater importance of agricultural drought effects on FP was evident for most of the subbasins during crop reproductive stages.FP was more affected by the precipitation and soil water deficits during summer and early autumn with relatively less affect observed in winter.For some mainstream subbasins(HUCs 0704,0706-0708,0711-0714),a significant and negative correlation between annual FP and two drought indices(SPI3 and SSWI3)was observed in spring(p<0.05),indicating that the food production was relatively less affected by drought during the planting period.(5)The annual FWP for all BMPs scenarios was improved under both baseline climate and future climate periods projected by two climate models.The efficiency for CM scenario is higher than that for other four BMPs scenarios.The efficiency in improving FWP for NT and CC scenarios were relative higher in eight agriculturally dominant subbasins.The efficiency for FS and GW scenarios were relative higher in five non-agriculturally dominant subbasins.The mean annual FWP of Twin Cities and Lower Des Moines watersheds for FS scenario increased by 7.3% and 27% respectively under baseline climate.The mean annual FWP of Peruque-Piasa,Lower Illinois and Upper Mississippi-Cape Girardeau watersheds for GW scenario increased by 10%,6%,12.6% respectively.While the efficiency in improving FWP for FS and GW scenarios were relative lower than that for NT and CC scenarios in most agriculturally dominant subbasins.(6)The BMP efficiency in improving FWP under future climate projected by two models were improved and higher than that under baseline climate,particularly for 6 watersheds(Twin Cities,Lower Minnesota,Lower Wapsipinicon,Upper Iowa,Lower Des Moines and Upper Mississippi-Cape Girardeau).However,the BMP efficiency for three watersheds(Turkey,Peruque-Piasa and Lower Illinois)under future climate projected by both RCP scenarios of CCCma-CanESM2 RCA4 model were reduced and lower than that under baseline climate.It is primarily because the increased precipitation and temperature under future climate,resulting in increased losses of pollutants from crop fields to streams.The TN,TP and TSS concentrations in these watersheds were increased with relatively larger degrees than streamflow,leading to reduced water quality and freshwater availability.Thus,the BMP efficiency in improving FWP was decreased under future climate in these watersheds.On the basis of studing the relative new theoretical methods about ecosystem services and BMPs in improving water quality in US,this study has conducted to evaluate the freshwater provisioning ecosystem service and analyse its responses to environmental changes in a large-scale river basin.The results would be helpful to improve integrated watershed management,and protect and restore the ecosystem services in our country.