Coupling Simulation And Uncertainty Analysis Of Surface Water And Groundwater In Hun River Basin

Surface water and groundwater often have a close hydraulic connection,the change of surface water will affect groundwater,and vice versa.Both surface water numerical simulation and groundwater numerical simulation can't fully describe the movement and distribution of water flow in the basin,and can't effectively solve the problem of joint dispatching and rational allocation of surface water and groundwater.A watershed-scale fully coupling simulation model of surface water and groundwater based on the physical mechanism of hydrological cycle was established.It can not only describe the migration and distribution of water flow,but also realize the integrated analysis and unified planning of water resources.The global climate change will affect the spatial and temporal distribution of atmospheric precipitation in local areas.The atmospheric precipitation is a important link in the process of hydrologic cycle,and its changes directly affect the surface water runoff and groundwater recharge in the watershed.The extreme precipitation events will also lead to frequent hydrological disasters such as floods and droughts.Therefore,it is necessary to implement the influence of global climate change on local atmospheric precipitation change.On this basis,combined with the coupling simulation model of surface water and groundwater,the influence of future precipitation change on surface water and groundwater in the basin is predicted and analyzed.Above researches can provide scientific basis for rational development of water resources in the basin,which has important theoretical and practical significance.Under the influence of random changes of surface hydraulic parameters and hydrogeological parameters,the output results of the coupling simulation model of surface water and groundwater are uncertain.Without consideration,it is difficult to evaluate the reliability of model outputs.Therefore,it is necessary to carry out the uncertainty analysis research of the coupling simulation model of surface water and groundwater,and analyze the influence of random changes of model parameters on the output results of the model.In the process of uncertainty analysis,it is necessary to invoke the coupling simulation model of surface water and groundwater several times,which will lead to huge computational load and lengthy computational time.An effective way to solve this problem is to establish a surrogate model for the coupling simulation model of surface water and groundwater.Compared with the simulation model itself,the surrogate model can greatly reduce the computational load and maintain better computational accuracy.In this paper,the kernel extreme learning machine method and linear weighted sum method are used to establish the surrogate model of surface water and groundwater coupling simulation model.The hun river basin is selected as the study area.Through coupling simulation model of surface water and ground water,sensitivity analysis,surrogate model,general circulation model,statistical downscaling model and uncertainty analysis and other theories and methods.The coupling simulation and prediction of surface water and groundwater flow in the study area are carried out,and the value ranges and probability distribution of sensitive parameter in the coupling simulation model of surface water and groundwater is identified.On this basis,the uncertainty analysis of the coupling model of surface water and groundwater is carried out,and the influence of the random changes of the sensitive parameter of the model on the prediction results of the coupling model is analyzed.Firstly,through field investigation,dynamic monitoring and collection and arrangement of relevant data of surface water and groundwater,the transport characteristics of surface water and groundwater in the study area are analyzed and determined.The hydrology and hydrogeological conditions of the study area are simplified,and the conceptual model of coupling simulation of surface water and groundwater in the study area is established,and then the numerical model of coupling simulation of surface water and groundwater in the study area is established.HydroGeoSphere software was used to solve the coupling model in parallel.The coupling model is calibrated and verified by using the synchronous observation data of field surface water and groundwater.Then,local sensitivity analysis method is used to screen out the parameters that have the greatest influence on the output of the model in the coupling simulation model of surface water and groundwater in the study area,which are taken as sensitive parameters and regarded as random variables.The Latin hypercube sampling method was applied to conduct sampling in the feasible region of random variables.The sampling results as model input were substituted into the surface water and ground water coupling simulation model in the study area,and the corresponding model output was obtained by running the coupling model,thus forming the input-output data set.According to the input-output data set,the surrogate model of the coupling simulation model of surface water and groundwater in the study area is established by using the kernel limit learning machine method,kriging method and support vector regression method,and the ensemble surrogate model including the surrogate model of the kernel limit learning machine is established by using the linear weighted sum method.The genetic algorithm is used to optimize the weight of each single surrogate model in the ensemble surrogate model.By comparing and analyzing the approximation degree of each single surrogate model and ensemble surrogate model to coupling model,the surrogate model with the best precision is selected.Finally,the relative errors of the measured and simulated values of surface water runoff and the absolute errors of the measured and simulated values of groundwater level are taken as likelihood functions.The critical value of likelihood functions is set and the value range and probability distribution of sensitive parameter in the coupled simulation model of surface water and groundwater in the study area are identified by the generalized likelihood uncertainty estimation method.On this basis,the monte carlo method is used to analyze the uncertainty of the coupling model of surface water and ground water,and to analyze the influence of the random change of the sensitive parameters of the model on the prediction results of the coupling model.Based on the above research,the following main conclusions are drawn:(1)According to the collected data of surface water and groundwater in the study area,after model calibration and verification.The simulated and measured values of surface water and groundwater have good fitting accuracy,the surface water and groundwater coupling simulation model is set up in line with the actual situation in the study area,can be used to predict some scene changes the response of surface water and groundwater.(2)The statistical downscaling model is established by using the statistical downscaling method,and the fitting effect is good after calibration and verification,which is in line with the actual situation of the study area.The predicted results show that the future precipitation in the study area will increase,but not obviously.In the later stage of the prediction periods,the annual precipitation changes havea a larger range,it prone to extreme precipitation events.(3)The influence of future precipitation change on surface water runoff and groundwater table was predicted and analyzed by using general circulation model,statistical downscaling model and coupling simulation model of surface water and groundwater.The results show that both surface water runoff and groundwater table will increase in the future.The increasing trend of groundwater table in mountainous area is obviously higher than that in plain area.(4)The results of precision analysis of different surrogate models show that the precision of ensemble surrogate model is significantly higher than that of the other three single surrogate models,which indicates that ensemble surrogate model has better approximation precision to the simulation model.For different practical problems,the selected ensemble surrogate model will be different.(5)The generalized likelihood uncertainty estimation method can identify the value ranges and probability distribution of sensitive parameters in the surface water and groundwater coupling simulation model in the study area.Through the parameter uncertainty analysis of the coupling model,the statistical characteristics of the prediction results of the coupling simulation model of surface water and ground water in the study area and the confidence intervals at different confidence levels can provide a scientific basis for the reliability analysis and the risk analysis of the development and utilization scheme.