Evolution Characteristics And Attribution Analysis Of Runoff In A Changing Environment

The evolution of surface runoff is affected by human activities and climate change,and its dynamic characteristics change significantly.Identifying and evaluating the changing environment of runoff evolution characteristics and driving mechanism,decoupling human activities and climate change effects rate in the process of change of land surface runoff are the hotspots and full of difficulties in the field of hydrological science currently,these are also practical problems for the relevant management departments.The Dawen River is the largest tributary of the lower Yellow River in Shandong Province,also,it is the main battlefield of ecological protection and high-quality development in the lower reaches of the Yellow River,and it is an important water resource for the storage lake(Dong ping Lake)of the South-North Water Diversion Project.However,the frequen t occurrence of floods and droughts in the Dawen River Basin,accompanied with complex human activities,had brought great difficulty and uncertainty to the water resources management in the Dawen River Basin.This study uses the Dawen River Basin as a case study,then,the evolution characteristics of water cycle elements in the basin,the attribution analysis of runoff changes,and the runoff response lawdriven by different conditions in the future are analyzed to explore the runoff evolution law and its attribution decoupling.The main research results are as follows:(1)Based on the integration of multiple statistical methods,this study revealed the tendency mutation point and space distribution characteristics of precipitation,extreme precipitation,temperature,extreme temperature and runoff,and evaluated the land use change in the basin,the reservoir operation,and water use dynamic characteristics,the cut-off points for the base period and the change period of the hydrological phase of the basin are determined.The results revealed that:the extreme precipitation index and runoff had an insignificant downward trend,while the extreme temperature index had a significant upward trend.there were significant periodic oscillations of 0.71.9 years and 0.7-2.5 years for extreme precipitation and temperature,respectively.The annual variation period of runoff was 0.5 years,and the interannual variation period was 1.55-2.26 years.The extreme precipitation and temperature had abrupt changes in the late 1970s and the middle and late 1980s respectively,and the runoff had abrupt changes in 1976-1977.The extreme precipitation and temperature indicators changed from west to east in space,and the direction of the interannual tendency rate of extreme precipitation indicators was different among the meteorological stations in the basin,while the extreme temperature indicators basically had the same direction of change.Runoff decreases gradually from upstream to downstream.The main change of land use is that the gray space occupied the green space,the reservoir storage water accounts for 60-70%of the total water storage of water conservancy projects,the water withdrawal shows an increasing trend,and the surface water source gradually replaced the underground water source.The changes of these driving factors comprehensively led to the frequent occurrence of extreme precipitation events,the intensification of drought,and the significant decrease of runoff in the Dawen River Basin.Considering the change point of each element of the water cycle and the characteristics of the basin,The cut-off point of the base period and the change period of the hydrological phase of the basin was determined to be 1976.(2)Establishing the SWAT model with multi-period "spatiotemporal dynamic parameters".The hydrological phase of the basin(1961-2019)was divided into the base period(1961-1976)and the change period(1977-2019),and the change period was divided into four time periods in 10-year steps.Parameter sensitivity analysis,calibration and verification are carried out based on SWATCUP combined with "two-way" parameter calibration method.The model decision coefficient(R2)and Nash Sutcliffe coefficient(NSE)for all models were above 0.71,and the highest was 0.98,indicating the model had high accuracy.The established multi-period model was then used to set up a variety of scenarios to simulate the runoff under certain conditions,in order to separat the runoff evolution under the background of climate change,human activities and future climate scenarios.(3)This study constructed a comprehensive framework for basin runoff attribution analysis considering runoff driving factors and basin characteristic factors,evaluated the applicability of various runoff attribution methods,and decoupled the contribution ratio of climate change and human activities to basin runoff,reveals the Dawen river basin runoff for human activity is the main driving factor,contribution rate of about 70-82%.In this study,correlation analysis was used to screen the relatively independent runoff driving factors and watershed characteristic factors,and the random forest method was used to rank the relative importance of the selected factors.According to the results,the appropriate attribution analysis method was selected,and the comprehensive framework of watershed runoff attribution analysis was constructed.The results revealed that:the influence of human activities on runoff and the river basin characteristics factor relative importance was higher,the double mass curve of precipitation-runoff and elastic coefficient method does not consider the influence of these factors and SWAT model can be integrated modeling the factor’s contribution on the runoff.Therefore,the SWAT model is the most suitable for runoff attribution analysis in the Dawen River basin.The results of decoupling of runoff by SWAT model revealed the main driving factors of runoff change during the whole period of change(1977-2019)and four sub-periods of change.During the whole change period,the main driving factor of runoff change in the sub-basins controlled by Laiwu station,Beiwang Station,Dawenkou Station and Daicunba station in the upper and lower reaches of the Dawen River basin was human activities,with contribution rates of 70%,79%,76%and 82%,respectively.Land use change and reservoir operation lead to the increase of runoff.Water use activities have resulted in a reduction in runoff.Section change period 1977-1990,climate change was the main driving factor for runoff change in the upper reaches of the basin(the Laiwu hydrological Station),with a contribution rate of 55%human activities were the main driving factor of runoff change in the middle and lower reaches of the basin(the Beiwang hydrological Station,the Dawenkou hydrological Station and the Daicunba hydrological Station),with a contribution rate of 62%,55%and 58%,respectively.The main driving factor of runoff change in the basin was human activities at the period of 1991 to 2000,2001 to 2010 and 2011 to 2019,and the contribution rates of runoff change in the sub-basins controlled by the Laiwu hydrological station,the Beiwang hydrological station,the Dawenkou hydrological station and the Daicunba hydrological station in these three periods were 77%,81%,75%and 84%,respectively.104%,95%,146%,145%and 77%,93%,88%,86%.(4)This study considered water use and integrated deep learning model-hydrological model(DeepAR-SWAT)as runoff prediction coupled simulation model,which decoupled the runoff evolution characteristics of the Dawen River Basin under different future climate and land use scenarios by considering water use.and reveals that the runoff will show an upward trend in the future period(2021-2060).Three scenarios of shared socioeconomic path(SSP)and representative concentration emission path(RCP)of five climate models and land use scenarios in 2050 under three scenarios of natural development state,cultivated land protection and ecological protection simulated by the GeoSOS-FLUS model are set.Three climate scenarios SSP126,SSP245,SSP585 and three land use(LUCC)scenarios are combined to set nine scenarios of SSP+LUCC.The results revealed that under the scenarios of SSP126,SSP245 and SSP585,the future precipitation has an insignificant increasing trend,and the temperature has a significant increasing trend,and the gray space land in the land use type occupies the green space,the demand for water is increasing,and the surface water source gradually replaces the underground water source.The simulating results of the DeepAR-SWAT coupling model for the future runoff under different scenarios show that the annual distribution of the nine scenarios(the wet season accounts for 60-73%of the annual runoff)are more uniform than that of the historical period(the wet season accounts for 87%of the annual runoff).The future runoff exhibited an upward tendency under the nine scenarios,among which the annual average and wet season runoff showed a significant upward tendency under the SSP126+LUCC scenario,while the runoff at all scales(annual average,wet season and dry season)showed an insignificant upward tendency under the other scenarios.Compared to the historical period,the runoff increased in all scenarios except the SSP126+LUCC natural development scenario(runoff decreased).The comparative analysis of all scenarios found that the runoff under the SSP126+LUCC scenario,especially the SSP126+LUCC natural development scenario,had the most change during the comparative historical period.When formulating relevant water resources management policies to face with the changing environment in the future,special attention should be paid to the runoff changes under the SSP126 scenario,the green space should be controlled to turn into gray space,and the water intake should be controlled by improving the water saving level and adjusting the industrial structure.