The Dynamic Change Of Water-energy Balance And Its Driving Force In The Loess Plateau Under The Changing Environment

In the context of global warming,the Loess Plateau region has experienced significant climate change.Meanwhile,human activities in this region have been continuously enhanced,leading to dramatic changes in underlying surface conditions.Under the influence of climate and underlying surface conditions,the process of hydrological cycle in this area is strongly disturbed.Among them,evapotranspiration,as the link between water cycle and energy cycle of a basin,its changes are affected accordingly,which lead to the dynamic changes of watershed water-energy balance.Therefore,assessment of the impact of climate and underlying surface changes on the water-energy balance dynamics can not only provide a scientific basis for more accurate simulation of evapotranspiration in the Loess Plateau,but also guide the rational development,utilization and effective management of water resources,and contribute to the ecological environment protection and sustainable development of the eco-economic system in the Loess Plateau region.Taking the Wei River Basin(WRB)—a typical loess plateau region as the study area,the spatial and temporal characteristics of vegetation coverage and its response to water and energy condition are analyzed under the Budyko framework based on the meteorological and hydrological time series and remote sensing data.Afterwards,the stationarity of the vegetation and water-energy condition dependent structures is identified.In addition,the impacts of climate condition,underlying surface,human activity,large-scale climate circulation patterns and solar activity on the water-energy balance dynamics over multiple time scales are evaluated.Moreover,the correlation among climatic factors,underlying surface,human activities,large-scale climate circulation patterns and solar activity is explored,and considering their synergistic effect on the Budyko parameter,a semi-empirical formula with certain physical meaning is established.Based on the semi-empirical formula of the Budyko parameter,attribution of the evolution of drought condition in the WRB is quantitatively analysed.The primary findings obtained in this paper are as follows:(1)Based on the Kriging method and the modified Mann-Kendall(MMK)trend test,the evolution characteristics of the hydroclimatic factors of the WRB are revealed.Results show that there is a warming-drying tendency of the climate in the WRB under the background of global warming.The spatial distribution of the average annual rainfall is uneven,increasing from the northwest to the southeast.Annual precipitation shows a non-signficant decreasing trend in the three sub-catchments and the whole basin,and the decreasing trend is obvious in the 1990s.The runoff of the WRB is mainly concentrated in the flood season,showing a significant decreasing trend during the whole study period with a more obvious decreasing trend after the 1990s.Annual average temperature of the WRB shows a significant upward trend in the past decades,and the most obvious increasing trend appears after 1990s.The spatial distribution of annual average temperature is uneven,gradually increasing from the northwest to the southeast.Although the potential evapotranspiration of the WRB shows a non-significant trend,the aridity index exhibits an upward trend,indicating that the WRB is continuously drying.(2)The lag time of the response of vegetation cover changes in different seasons to water and energy conditions in the WRB was determined based on the partial correlation analysis Generally,the vegetation cover shows a gradual greenness trend in the Jing River Basin(JRB)and Beiluo River Basin(BLRB)since 1982,and the greenness trend was most significant in autumn.Vegetation changes are more sensitive to precipitation than temperature,and the lag time of vegetation to precipitation and temperature was less than 3 months.There is also strong teleconnections between vegetation and large-scale climate circulation as well as solar activity,and the lag time of vegetation growth to large-scale climate circulation and solar activity is generally longer than that of precipitation and temperature.In order to avoid the multicollinearity of precipitation/temperature and large-scale climate circulation/solar activity,the ridge regression model is applied to calculate the contribution of climate change on vegetation growth.Result shows that climate change can explain more than 50%of vegetation growth in spring and autumn.(3)Using the copula functions,the joint distribution of water-energy condition and vegetation cover is established at seasonal scale,and the variation characteristics of the vegetation-precipitation/temperature dependence structure are identified by the Copula-based Likelihood-ratio method.Results show that there is a slight change in the relationship between spring vegetation and precipitation/temperature,and autumn vegetation and precipitation in the JRB,while the relationships between summer vegetation and precipitation/temperature and autumn vegetation and precipitation remain stable.The Arctic oscillation(AO),Pacific decadal oscillation(PDO),ENSO event and sunspot shows significant teleconnection with spring and autumn vegetation-precipitation relationship,while the spring and autumn vegetation-temperature relationship is controlled by AO and PDO.The summer vegetation-precipitation relationship is not only affected by potential evapotranspiration and PDO,but also affected by human activities such as irrigation.(4)The annual values of the Budyko parameter is fitted by the least squares method under moving windows.Moreover,the Empirical Mode Decomposition(EMD)method is used to extract the detailed fluctuation information under different time scales,and then the Pearson correlation coefficient method is adopted to determine the the driving force of Budyko parameter fluctuation under different time scales.Results show that the interannual variation of the Budyko parameter of the main stream of the WRB shows a significant upward trend,while that of the JRB and BLRB shows non-significant downward trend.By adopting the EMD method,detailed information of parameter series can be extracted effectively,and components of 4a,12a,20a and beyond 20a and a residual trend component are finally obtained.Among them,the 4a fluctuation is mainly affected by partial climate factors and underlying surface,while the 12a fluctuation is mainly affected by baseflow,PDO and sunspot.The fluctuation of 20a and beyond 20a is influenced by comprehensive factors.(5)The correlation among climatic factors,underlying surface,human activities,large-scale climate circulation patterns and solar activity was explored by the Pearson correlation analysis.In addition,the synergy mechanism of influencing factors was revealed by the multivariate adaptive regression splines(MARS)method.Based on the most critical factors selected,a semi-empirical formula for the annual Budyko parameter was developed.Results show that there is a significant correlation between climate,underlying surface,human activities,large-scale climate circulation as well as solar activity in the WRB.The effective irrigated area and vegetation coverage have a synergistic response to the Budyko parameter in the upper reaches of the WRB,while for the middle and lower reaches of the WRB,the effective irrigated area and seasonality index of precipitation show a synergistic effect on the Budyko parameter.In addition to climate factors(precipitation and potential evapotranspiration),the controlling factors of Budyko parameter in the upper reaches of the WRB are vegetation coverage and effective irrigated area,while that in the middle and lower reaches of the WRB are vegetation coverage,effective irrigated area,seasonality index of precipitation and baseflow,and the effective irrigated areas is the key factors affecting the Budyko parameter in the JRB and BLRB.(6)Based on the Budyko hypothesis and the semi-empirical equation of the Budyko parameter,the actual evapotranspiration is estimated,and the residual water-energy ratio(WER)index was developed to characterize the dry and wet conditions of the WRB.The sensitivity of the dry and wet conditions to climate and underlying surface conditions is also evaluated based on the sensitivity coefficient method.Besides,the contribution of climate and underlying surface changes to the change of dry and wet conditions in the WRB is assessed by numerical experiments.Results show that the inter-annual variation of the WER in the WRB exhibits a decreasing trend.The dry and wet condition in the upstream of the WRB is most sensitive to vegetation coverage and least sensitive to effective irrigated area.The dry and wet condition in the middle and lower reaches of the WRB is more sensitive to precipitation and potential evapotranspiration than to underlying surface conditions.The response of dry and wet condition to irrigation in JRB is obviously stronger than that to climate change,while the beiluo river is just the opposite.The change of effective irrigated area is the leading factors for the change of the dry and wet conditions in the mainstream area of the WRB,while precipitation and effective irrigated area variations dominate the dry and wet conditions in the JRB,and precipitation controls the change of the dry and wet condition in the BLRB.