| Affected by climate change and intensified human activities,the hydrological cycle process in the Loess Plateau region of northern Shaanxi has undergone significant changes,and the response of runoff to environmental changes is particularly strong.In recent years,the continuous reduction of runoff in northern Shaanxi has attracted widespread attention.Base flow,as an important component of runoff,is the main source of runoff during the dry season,especially for the northern Shaanxi basin located in arid and semi-arid areas.It has a significant impact on flood control,water supply,and ecological protection projects within the basin.Therefore,it is particularly important to conduct research on the evolution laws of runoff and base flow in the northern Shaanxi basin under changing environments,quantitatively analyze the impact of climate change and human activities on both,and explore their future change characteristics.This study focuses on the Jialu River,Tuwei River,and Gushanchuan basins in the northern Shaanxi region.Multiple numerical simulation methods were used to segment the basin base flow,and the most suitable base flow segmentation method was selected for each basin to analyze the evolution law of the base flow;Secondly,a two-stage precipitation allocation model based on the principle of water balance and the elasticity coefficient method based on Budyko theory were used to quantitatively analyze the impact of climate change and human activities on base flow and runoff,revealing the driving factors of typical watershed runoff and base flow changes;Finally,with its main influencing factors as input parameters,an LSTM-BP coupling model was established to predict the future status of runoff and base flow in three watersheds,in order to provide reference for ecological environment construction and scientific management of water resources in the Loess Plateau region of northern Shaanxi.The main research results obtained are as follows:(1)Evaluated the applicability of various basic flow segmentation methods in typical watersheds in northern Shaanxi.Ten basic flow segmentation methods,including digital filtering,smooth minimum value,and time step method,were used to segment the daily runoff series of three watersheds to obtain the daily basic flow series.The rationality and reliability of each basic flow segmentation method were discussed in three aspccts:the basic flow index,the basic flow hydrograph,and the effectiveness of the annual basic flow.The Eckhardt fitering method was analyzed to obtain higher basic flow accuracy,The basic flow hydrograph is more in line with the actual situation of the watershed,so it is recommended to use the Eckhardt filtering method for basic flow estimation in typical watersheds in northern Shaanxi.(2)Explored the annual distribution characteristics and interannual evolution trends of base flow in typical watersheds.Analyzed the monthly and seasonal distribution patterns of base flow within a year,and explored the characteristics of base flow distribution within a year from three aspects:non-uniformity,concentration,and amplitude of change.There were significant differences in base flow distribution among the three watersheds within a year,with extremely uneven distribution in the Gushan River Basin.During the flood season,base flow accounted for 68.66%of the annual base flow,while the Tuwei River Basin had relatively uniform distribution within a year,with a stable base flow index of around 0.50 in each month,The unevenness of annual distribution in the Jialu River Basin is between the two.Using statistical methods to analyze the interannual evolution trend of base flow in each basin,it was found that the annual base flow in all three basins has significantly decreased and has a further decreasing trend in the future.During the period from 1957 to 2019,there was one variation in the annual base flow in the Jialu River,Tuwei River,and Gushanchuan basins,and the variation year and runoff in each basin remained consistent,with 1971,1979,and 1979,respectively.Wavelet analysis showed that:,There is a complete main cycle for the annual base flow of all three watersheds.(3)Analyzed the characteristics of hydrological and meteorological elements and underlying surface conditions in typical watersheds.The temperature,precipitation,and potential evapotranspiration of the three watersheds all show an upward trend on an annual scale,while runoff decreases in a stepped manner.The significant reduction in runoff and base flow in each watershed should be the result of the dual effects of climate change and human activities.The changes in underlying surface conditions in each watershed were analyzed from three aspects:land use mode,vegetation coverage,and silt dam construction.From 2000 to 2020,the land use mode in each watershed was mainly farmland and grassland,and vegetation coverage showed an overall trend of improvement during this period.The number of silt dam construction in the three watersheds increased sharply in the 1970s,which is consistent with the years of sudden changes in runoff and base flow in each watershed.It can be considered that,The construction of silt dams has had a significant impact on the sudden changes in watershed runoff and base flow.(4)Revealed the main driving factors of typical watershed runoff and base flow changes.A two-stage precipitation allocation model based on the principle of water balance was used to simulate the surface runoff and base flow generated by a single precipitation.Based on this,the contribution rate calculation method was used to quantitatively analyze the impact of climate change and human activities on surface runoff and base flow.The contribution rate of climate change to surface runoff in ench watershed calculated from the simulation results was greater than that to base flow,The contribution rate of human activities to surface runoff in each watershed is above 53%,and the contribution rate to base flow is above 74%,indicating that human activities play a leading role in reducing surface runoff and base flow in the three watersheds.Using the elasticity coefficient method based on the Budyko assumption to analyze the driving factors of runoff,it was found that changes in underlying surface conditions are the main cause of reduced runoff in the three watersheds,followed by precipitation and the smallest impact of potential evapotranspiration.(5)A LSTM-BP coupling model was constructed to predict the future status of typical watershed runoff and base flow.Using precipitation,temperature,potential evapotranspiration,and accumulated storage capacity of silt dams as input parameters for the LSTM-BP model,typical watershed runoff and base flow from 2020 to 2030 were predicted on a monthly scale.The accuracy evaluation indicators RMSE,NSE,and R2 were used to evaluate the simulation accuracy of the model,confirming the rationality and reliability of the LSTM-BP coupled model for future prediction of typical watershed runoff and base flow.Due to differences in the annual distribution of runoff and base flow among the three watersheds,the simulation accuracy of the model varies slightly.From the prediction results,the trend of monthly base flow prediction values in each watershed remains consistent with the trend of monthly runoff prediction values,both of which comply with the annual and monthly distribution law of runoff and base flow. |
PDF Full Text Request