Comparing Three Regression Models for Reconstructing Groundwater Level Data (A Case Study)

The base for hydrology studies is accurate data. However, the gaps and shortage of sufficient data exist n the most hydrology data such as  underground water data as the most important and cheapest water source,  lack of  data  take places due to various reasons such as Inability to measure and faille to register statistics. Missing data or incorrect statistics, Therefore, estimating the missing data is necessary which depending on the conditions of each station may demand a specific method to yield the best solution. In this article regression methods were applied in restoring underground water contour of piezometer stations of Lorestan province. In this regards, after deliberate deletion of about 15% the monthly observation data for four consecutive years in 22 piezometer stations in Alashtar of Lorestan province, their values are estimated and assessed them through RMSE and percentage of relative deviation of mean module. Finally, the obtained results are show that the simple linear regression method outperforms other methods.