Determination of valuable piezometric wells in groundwater level prediction by considering spatiotemporal information

The precision of groundwater level estimation is a function of the density, location and accuracy of the wells. As the number of piezometric wells increase, the precision of surface estimation is improved, but digging and maintaining piezometric wells is always costly and time consuming. The main objective of this research is to minimize the number of piezometric wells and to identify valuable wells for precise prediction of groundwater level. The main innovation of this research is the inclusion of spatial and temporal information in optimizing the number of wells. This research was conducted in three main phases: 1) Obtaining information of 75 piezometric wells of Tabriz plain and exploratory spatial data analysis (ESDA), 2) Implementation of the ant colony optimization algorithm in combination with the simple kriging method and the identification of wells in each year, 3) Calculation of the final value of wells based on number of times each well was present in the calculations for 2014-2017. The results show that of the 75 wells per month, an average of 18 wells adds limited information to the calculations. Of these 75 wells, only two wells in none of the 36-month calculations were considered as valuable wells, and 15 wells have a relative value less than 11 percent. Therefore, the number of piezometric wells can be reduced 20% (15 wells), without a sensible data loss in the estimation of groundwater levels in the area. The results show that using temporal information of wells and proposed method, valuable wells can be identified and, by saving time and cost, to some extent reached the precision of the initial level.