Estimation Of Baseflow Non-point Source Pollution In A Typical Agricultural Watershed

Non-point source pollution (NPS) has been recognized as the primary threat to the surface water around the world. However, such the threat caused by the NPS will not be limited to the surface water. Baseflow derived from the contaminated groundwater is likely to become a hidden-potential non-point source of the surface water that cannot be ignored in most time. Because of the fact that there still are no direct approaches to measure the baseflow at a large watershed scale or in a long period by now, how to estimate baseflow and its nutrient load more accurately and reliably is considered to be one of the difficult problems needed to be solved urgently in hydrology and water environment researches. To minimize the uncertainty in baseflow separation that caused by the variability of different baseflow recessions under different conditions (i.e.,climate), the Meteorology Corrected Nonlinear Reservoir Algorithm (MNRA) was developed in this study based on the Nonlinear Reservoir Algorithm and a Meteorological Regression Statistical Model (used for recession parameter a estimation). Given that, baseflow nutrient load recession coefficient (?) was proposed in this study, and a Recursive Tracing Source Algorithm(RTSA) was constructed based on the MNRA to estimate baseflow nutrient load in the Changle River watershed. And then,the dynamics of baseflow nutrient load under future climate change scenarios were simulated together with the Statistical Down Scaling Modell (SDSM) and the Soil and Water Assessment Tool (SWAT). The main results and conclusions are as follows:(1) Compared with the linear reservoir algorithm-based digital filters (i.e.,Kalinin and Eckhardt's recursive digital filter), the Meteorology Corrected Nonlinear Reservoir Algorithm (MNRA) could usually be more sensitive for baseflow response to precipitation obtained a higher goodness-of-fit for streamflow recession[Nash-Sutcliffe efficiency (NSE) = 0.98], and effectively eliminated the uncertainty existing in baseflow separation caused by the variations in different baseflow recessions that under different recession conditions (e.g., climate). According to the results of MNRA, the ratio of annual baseflow (188-619 mm) to annual streamflow(284-944 mm) ranged from 57.5 to 74.1%, and averaged at 65.1%.(2) Baseflow nutrient load recession coefficients (r) was proposed in this study to characterize baseflow nutrient load recession. Together with the Nonlinear Reservoir Algorithm, we developed the Recursive Tracing Source Algorithm (RTSA) that can be directly used for baseflow nutrient load separation. By taking full consideration of the effects of meteorological factors on the recessions of baseflow and its nutrient loads(including the changes of nutrient concentrations), the RTSA provided an accurate and reliable baseflow TN, NO3-N, TP and DP load estimation (NSE > 0.7, R2 > 0.7).Compared with the LOADEST model, the baseflow nutrient load estimated by RTSA can fit better to the real baseflow nutrient load recessions, and effectively avoid some significant over- and under- estimation that mainly caused by the significant variations in the flow-loads (or concentrations) relationship for different baseflow nutrient load recessions. Therefore, RTSA can be an effective approach for baseflow nutrient load estimation in the rainy regions.(3) Interannual fluctuations in the non-point source pollutant load largely depended on the annual variation in precipitation and runoff. The estimated annual total TN, NO3-N, TP, and DP loads in streamflow showed significantly positive correlations with the annual precipitation and streamflow. There were significant differences the nutrient loads in the dry year and rainy year. From 2003 to 2012, the mean annual total TN, NO3-N, TP and DP loads in the streamflow were 24.67 kg/ha,18.31 kg/ha, 1.05 kg/ha, and 0.60 kg/ha, respectively; and approximately 57.48%,62.33%, 53.55%, and 64.29% of those nutrient loads in the streamflow were contributed by baseflow, respectively. Obviously, baseflow was the most important pathway for non-point source pollutants export in the Changle River watershed over the 10-year records. In the light of the calculated baseflow concentration index (BCI),baseflow played a major role in diluting the TN, NO3-N, TP and DP in the steam by now in the Changle River.(4) Under the scenario A2 (a very heterogeneous world with high population growth, slow economic development and slow technological change) and B2 (a world with intermediate population and economic growth, emphasising local solutions to economic, social, and environmentalsustainability) of a coupled regional ocean atmospheric model (HadCM3), there was significant interannual variability in the annually precipitation of Changle River watershed from 2013 to 2099, but no significant trend was obtained. However, mainly because of the increase of extreme rainfall events caused by the climate changes, the composition of riverine water was changed obviously, indicating by the decrease of baseflow fraction of streamflow.The results showed that streamflow TN, NO3-N, TP and DP loads and concentrations for in Changle River watershed presented a general upward trend under the scenario A2 and B2, Compared with the baseline period, contributions of baseflow to total TN, NO3-N, TP and DP loads in streamflow declined to some extent because of the decrease of baseflow fraction of streamflow, whereas the those nutrient loads and concentrations in baseflow showed a rising trend on the whole. For the scenario A2, the mean annually baseflow TN, NO3-N, TP and DP loads were 15.29 kg/ha, 13.31 kg/ha, 0.64 kg/ha, and 0.44 kg/ha, respectively; which accounted for approximately 50.36%, 55.01%, 37.63%, and 44.58% of corresponding total loads in streamflow, respectively. For the scenario B2, the mean annually TN, NO3-N, TP and DP loads export from this studied watershed through baseflow were 15.29 kg/ha,13.31 kg/ha, 0.64 kg/ha and 0.44 kg/ha, accounting for approximately 53.62%,59.78%, 43.65% and 47.05% of corresponding total loads in streamflow, respectively.(5) With the ongoing climate change in the future, the enrichment effects of baseflow on TN and NO3-N concentrations in stream would be generally strengthened to some extend; however, the effects of baseflow on TP and DP concentrations in stream would be dominated by the dilution for the most cases, mainly because of the much higher increasing rate of TP and DP concentrations of surface flow relative to that of baseflow. However, such dilution effects of baseflow on TP and DP in stream would not reduce but increase the risk of eutrophication caused by high level of P concentration in the water, because the baseflow TP and DP concentrations would generally increase in actual, due to the climate changes. In other words, extraordinary risks of eutrophication caused by the baseflow nutrients loading should not be neglected.