Causality Analysis Between Precipitation And Soil Moisture And Precipitation Prediction

The main reason of droughts and floods is the uneven distribution of water resources in time and space which was caused by the changes in the climate system and water cycle.Therefore,identifying the relationship between the climate system changes and the water cycle plays a key role in effectively preventing drought and flood disasters.In addition,because precipitation is a complicated process,it has several characteristics such as randomness,variability,uneven regional distribution and large interannual variability in spatiotemporal distribution.Therefore,it is very difficult to accurately forecast rainfall.As an important parameter,soil moisture has become the entry point in many precipitation researches.However,there is still much controversy about the causal relationship between soil moisture and precipitation.This thesis is devoted to exploring the causal relationship between soil moisture and precipitation,then apply the relationship to improve the accuracy of precipitation forcasts.For this reason,we employ New causality method,which is based on the traditional Granger causality method,to establish a causal relationship model between soil moisture and precipitation,and use historical records to rebuild the process of climate change,and demonstrate the relationship between the two causal influences.Particularly,we also use the BP neural network model optimized by genetic algorithm to predict the monthly precipitation and analyze its influence of it on soil moisture.The main work and achievements are as follows:1)Using the Granger causality and the new causality method to study the causal influences on precipitation and soil moisture at same depth.The result shows that there is causal effect between the two causal variables,and the causal effect of the former on the latter is greater than the causal influence of the latter on the former.The results also show that the causal effect is more pronounced in dry areas;conversely,causal effect in wet areas is relatively weak.2)Using Granger causality and new causality method to study causal effects on precipitation and soil moisture at different depths.The results show that the causality influence on the two are more obvious in the shallow soil,while the causal effect of the former on the latter becomes weaker with the increase of soil depth.At the sametime,the causal influence on the latter to the former will also gradually weaken.3)By using the optimized BP neural network,we investigate the influence of soil moisture on precipitation prediction based on genetic Algorithm.The experimental results show that the soil moisture has a significant influence on the prediction of monthly precipitation,then,we further verifies the validity of the new causality method to analyze the causal relationship between soil moisture and precipitation.