干旱区生态环境知识资源中心

Designing an optimal groundwater level monitoring network (GLMN) is a major challenge and one of the primary goals of groundwater management. This paper proposes a methodology that incorporates comparative analysis of IDW, spline, ordinary kriging (OKrig), and empirical Bayesian kriging (EBK) interpolation methods with "1-fold cross-validation" technique for the regionalization and redesigning of the existing GLMN in an arid hardrock-alluvium Al-Buraimi region, Oman-UAE border. The performance indicators (weighted RMSE = 10.84 m and weighted R-2 = 0.79) show the superiority of the EBK interpolation method over other methods and reveal reasonably accurate results when applied in areas with sparse and scarce observation wells. A new GLMN is proposed with regard to the results of the EBK method in which the idea of a "secondary observation network" is presented to reduce time and cost needed for groundwater level measurements. The proposed GLMN consists of 14 new observation wells added to 39 existing wells in order to provide better quality data compared to the current GLMN. The procedure used herein provides a practical and straightforward method to regionalize groundwater level variations and redesign the GLMN in regions with complex hardrock-alluvium geological setting.