Stochastic Simulation Of Groundwater System Based On HydroGeoSphere Technology In The Western Jilin Province

The method that groundwater flow and pollutant movement in porous media are simulated by stochastic model is rapid developing in foreign countries in recent 20 years. Because some countries pay attention to geological disposal of radioactive nuclear waste, some scholars studied the model of bedrock fractured groundwater flow and solute transport and made an important breakthrough. However, the stochastic theory of groundwater movement used to solve practical problems is also few particularly the successful solution. Therefore, stochastic simulation technology must be more widely appreciated.With the development of numerical simulation, using numerical model to evaluate and predict groundwater resources has long been from theoretical research phase to practical application. The results of evaluation and prediction have also become the main basis of the relevant management for managers. In this case, the model quality is particularly important. Random model not only can take into account uncertainty of parameters in the simulation model, but also directly reflect the sensitive factors and conditions for parameters changing in the model. It is beneficial to know the model deeply, and then guide the simulation work in depth.This paper selected Western Jilin Province as the study area. The geological, hydrogeological conditions, groundwater recharge, runoff, discharge conditions and dynamic characteristics are analysed. The boundary conditions and groundwater sources and sinks of study area are generalized. The hydrogeological conceptual model of study area is established. The corresponding mathematical model of groundwater is established using by three-dimensional Richards equation. The model is solved with HydroGeoSphere software that based on the finite element method. Through the identification and validation of model to prove the model and hydrogeological parameters which had been adjusted can represent the actual groundwater flow system of the study area. The sources and sinks of future 15 years are predicted by using time series methods in the study area. The forecast results are brought to the simulation model and calculated the distribution of groundwater level from 2001 to 2015. Sensitivity analysis method is used to calculate influence degree of hydrogeological parameters for the simulation model and obtain parameters and parameters partition which is the larger sensitivity. The permeability coefficient, specific yield and storage coefficient have a large degree of influence for the model. Those are located in the fold belt of Da Hinggan Mountains and the table plateau that between Songhua River and Lalin River. Therefore, the parameters accuracy should be improved when complementary hydrogeological survey work of Western Jilin Province is carried out in the future. The impact of reliability of prediction and evaluation of groundwater resources due to the uncertainty of parameters will be reduced in this way.Many experiments and researches have been done since 1944. The probability distribution characteristics of hydrogeological parameters are obtained by Law and many hydrogeologists during 60 years. Based on the known distribution probability shape of parameters, parameters are randomized by Monte-Carlo method and brought into the established simulation model. Finally the stochastic model of groundwater in Western Jilin Province and random characteristics of groundwater level distribution which take into account parameter uncertainty are obtained. The distribution of standard deviation of groundwater level shows that the dispersion degree of calculation results of the simulation model is little. The reliability of the numerical simulation model is high. In a probability the range of groundwater level of any node can be obtained by stochastic model in the study area. That is a basis of risk analysis and project budget which are environmental geological problems caused by groundwater exploitation (such as groundwater pollution, springs cutoff, etc.). Compared with the deterministic model, it is more scientific and reasonable that using the groundwater level mean value of random model to evaluate and predict the amount of groundwater resources.