Geostatistical analysis of the spatial distribution of environmental data: a survey on methods and applications

Abstract

The aim of this study was to examine the methods, tools, and platforms used for the analysis of spatial data, as well as to assess their potential for solving environmental challenges. The study considers geostatistical methods, including kriging, variogram analysis, semivariance, the Thiessen polygon method, inverse distance weighting (IDW) interpolation, and regression models such as linear regression, multiple regression, and geographically weighted regression (GWR).
A literature search turned up 52 peer-reviewed and indexed articles on methods including regression models, variogram analysis, and kriging. The selection criteria included: (1) relevance to geostatistical analysis of environmental data, (2) methodological rigor, (3) publication in high-impact peerreviewed journals, and (4) citation frequency indicating scientific significance. These studies highlight the effectiveness of geostatistical methods, geospatial platforms, and Python in environmental monitoring and predictive modeling. For classification tasks, logistic regression and decision trees were examined. The study results demonstrate that the application of modern geostatistical methods allows for the identification of spatial distribution patterns of environmental data and improves prediction accuracy. In particular, it was found that the spatial autocorrelation index effectively determines areas with high levels of similarity in environmental parameters, while local indicators of spatial association (LISA) help identify regional clusters with high pollution intensity or other anomalous characteristics. It was demonstrated that the use of spatial modelling platforms, such as Geographic Information System (GIS) software like ArcGIS and Quantum GIS (QGIS), along with the Python programming language and spatial data analysis libraries such as GeoPandas and the Python Spatial Analysis Library (PySAL), significantly enhances the effectiveness of environmental phenomenon analysis. The integration of satellite image data with geostatistical methods was found to contribute to the creation of more accurate maps for forecasting environmental risks. The proposed approaches demonstrate significant potential for environmental monitoring and natural resource management, enhancing the understanding of spatial patterns and serving as a basis for further research in this field.