Review of application of geostatistical techniques to groundwater salinization problems

Groundwater salinization is a complex and dynamic process often related to multiple causes, such as seawater intrusion, soil salinization associated with water irrigation, and geogenic factors such as evaporate dissolution. Consequently, a reliable assessment of salinization risks depends heavily on the ability of statistical methods to accurately capture the spatial variability and interrelation among salinization indicators. Geostatistical methods are often used to identify and map salinization-affected regions, investigate how salinization indicators influence groundwater mechanisms, and eventually design optimal groundwater management policies.

In the context of the MEDSAL Project (www.medsal.net), this study reviews the recent key applications of geostatistical methods to address problems relevant to groundwater salinization. The basic principles of geostatistics are briefly described, and several studies are discussed that employ geostatistical and multivariate tools for identifying salinization sources, clarifying the relationship among salinization indicators and groundwater processes, and facilitating uncertainty propagation in physically-based models of the groundwater systems affected by salinization.

The literature review identifies most used methods and offers several recommendations in terms of future directions and challenges on the role of geostatistics for improved mapping of the spatial and/or spatiotemporal distribution of geochemical data related to salinization. These recommendations include the integration of geostatistics and machine learning methods for improved understanding and modeling of groundwater salinization processes, as well as the application of modern geostatistical simulation algorithms, accounting for diverse information sources, for exploring parameter uncertainty in spatially distributed hydrogeochemical models.