Simulation Of Regional-scale Water Movement In Saturated And Unsaturated Zone Based On Ensemble Kalman Filter

With the rapid development of economy,many problems,such as water shortage,water pollution and soil salinization,have become significant factors that constrain the sustainable development of agriculture in China.The interaction between surface water and groundwater is increasingly common,and the accurate simulation of water movement in variably saturated zone is the basis of studying the pollution of groundwater and secondary salinization.As far as I am concerned,many saturated-unsaturated coupling models have been developed.However,there are many difficulties in the numerical simulation of regional saturation-unsaturated saturation due to the non-linearity of the model,the variability of the boundary and the source and sink items,and the spatial variability of the parameters.Data assimilation can fuse a lot of observation information from different sources,then extract useful information,and improve the simulation of state variables,finally achieving model parameter estimation.The Kalman filter is a kind of assimilation method which is widely used in groundwater and soil water simulation.However,there are few researches on the application of the Kalman filter in the saturated-unsaturated water flow coupling simulation.The main purpose of this paper is to explore the role of groundwater level and soil moisture content observation data in regional saturated-unsaturated water assimilation system.The following research is carried out:The Yonglian Experimental Zone of Hetao Irrigation District in Inner Mongolia is the research area,(MODFLOW-HYDRUS),the soil hydraulic parameters and hydrogeological parameters of the model are determined and verified by the combination of trial and error methods and PEST,moreover,the sensitivity analysis of model parameters is made at last.At the same time,a regional saturated-unsaturated water flow assimilation system based on the ensemble Kalman filter algorithm is constructed.Two separate assimilation projects of the assimilation water level and assimilation water content are carried out.The main conclusions of this paper are as follows:(1)MODFLOW-HYDRUS coupling model has good applicability to the simulation of saturated-unsaturated water movement area;(2)Sensitivity analysis of model parameters shows that:when the unsaturated soil profile is set to homogeneous soil,the saturated water content,intake value and pore distribution index have a great influence on the simulation results,while the residual moisture content and saturated hydraulic conductivity have an influence on the simulation results;when the aquifer is regarded as homogeneous,the permeability coefficient has little influence on the simulation results,while the water supply of aquifer has an influence on the simulation results.(3)The distribution of soil water parameters and hydrogeological parameters of each unsaturated soil profile can be obtained by inversion of the water level and water content of the aquifers,and the distribution of soil water parameters obtained by inversion of assimilation water content is closer to the distribution of the actual soil profile than the inversion of the water level.Since there are many parameters of the inversion,the soil water parameters are retrieved can cause heterogeneity,and it is difficult for the assimilated water level to effectively invert the distribution of unsaturated hydraulic parameters(4)In the regional saturated-unsaturated water assimilation system,the assimilated water level can significantly improve the simulation of groundwater.However,the soil moisture content of the profile after the assimilation water level is worse than that of the assimilated result,which may be due to the distribution of hydraulic parameters that are got through the inversion of the water level shows a large difference from the real soil profile,leading to the deviation of the soil moisture content from the actual value.While improving the soil water content to some extent,the assimilation water content does not improve the simulation of groundwater.