Deriving Karst Groundwater Storage Changes Of Guizhou Province Based On GRACE Data

Special hydrogeological condition and lack of basic data make it difficult to apply traditional evaluation methods of groundwater resources in karst area. GRACE provides a new method to study groundwater storage in regional scale, which is of vital significance in karst area short of groundwater monitoring data and hydrogeological parameter data.The continental water mass anomalies from January, 2003 to June, 2014 in Guizhou Province are calculated, and combined with GLDAS and the equation of water equilibrium, the thesis calculates time series of the groundwater storage changes in the study region, which suggests a slow rising trend with a growth rate of about 6mm/a and clear seasonal variations, and through analyzing, seasonal fluctuation of the groundwater storage is mainly caused by the difference of precipitation and evapotranspiration in the study area. To test the reliability of GRACE, the thesis compares the groundwater storage changes derived from GRACE data with the groundwater monitoring data in the study area, and the results show that GRACE is of high accuracy and reliability, and is an important supplement to traditional evaluation methods of groundwater resources and has vital significance in karst area short of monitoring data. Furthermore, based on time series derived from GRACE data, NAR model is established to forecast the trend of the groundwater storage changes in the study area, and checked by groundwater monitoring data, NAR model is of high feasibility and accuracy, and is an important analysis tool of time series. At the same time, the thesis builds NARX model with the difference of precipitation and evapotranspiration in the study area as external input data, however, the training error of the model is relatively large and has a low generalizability. Compared between NAR model and NARX model, the accuracy of ANN model depends on whether the training sample is adequate and if related factors is taken into account or not.In summary, the thesis tries evaluating karst groundwater resources without using monitoring data and hydrogeological parameter, builds prediction model of groundwater storage change in the study region based on the results derived fromGRACE data, and has made preliminary progress. GRACE will get more extensive application in hydrologicalgeology.