Investigating the quantity variation trend of ground water table using geostatistics and GIS (Case study: Sirjan Plain)

The purpose of this research was to determine the temporal and spatial variation of groundwater levels of Sirjan Plain using geostatistical methods in the statistical period (1997-2007). The water level data of 30 wells were collected. The quality، accuracy and normality of data were controlled and the water table was interpolated by different interpolation methods including inverse distance weighted (IDW)، local polynomial interpolation (LPI)، radial basis function (RBF) completely regularized spline، spline with tension، multiquadric، inverse multiquadric، thin plate spline، ordinary kriging (OK)، simple kriging (SK)، disjunctive kriging (DK). The best interpolation method was determined based on mean error (ME)، root mean square error (RMSE) and coefficient of determination. The zoning map was also traced during investigation. The results showed that the Gaussian variogram model with 0. 7 coefficient of determination was the best fitted model for the spatial data and local polynomial interpolation method and ordinary kriging method with root mean square error of 43. 6 and 23. 85 were the best method of interpolation respectively. The results also showed that the ordinary kriging with Gaussian variogram was statistically more accurate in comparison to other methods for estimating the depth of the groundwater table in the two periods. Zoning with ordinary kriging showed that groundwater level has dropped in most parts of the plains during the study period. The groundwater level drop ranged to 40 meters with an average of 15 meters. The water level has increased in the western parts of the plain which causing a shift of water from the western salty water aquifer into sweet water aquifer; and decreasing of Sirjan Plain Aquifer quality accordingly.