Transparency and reproducibility in the selection of relevant studies are essential, involving the identification, screening, eligibility, and inclusion of appropriate articles, in reviewing the extent to which remote sensing has emerged as a non-invasive solution in overcoming the limitations of conventional methods, especially in mapping karst as a system with complex geomorphological and ecological characteristics that are difficult to observe directly, especially soil-landscape conditions. Hence, this study aims to review the extent of geospatial technology application, identifies knowledge gaps, and outlines future research directions by establishing a conceptual foundation for data-based mapping in karst environments. This study was conducted using a systematic literature review with the PRISMA framework, which produced 56 articles (16 landscape, 22 landform, 3 soil class, and 15 soil properties articles) from 2010 to 2023 and analyzed to evaluate driving factors and indicators, spatial patterns, environmental covariates, analysis techniques, and types of sensors used. In karst regions, the landform and soil landscape approach has been the method of choice for mapping unit characteristics at land units. The results indicate that DEM-based geomorphometric analysis remains the dominant approach, primarily using SRTM, ASTER, and LiDAR for landform detection (e.g., dolines and sinkholes), while multispectral imagery such as Landsat and Sentinel-2 is widely applied for landscape dynamics through spectral indices (e.g., NDVI and carbonate-related indices). Recent studies demonstrate a methodological shift toward object-based image analysis, sub-pixel techniques, and machine learning approaches (e.g., Random Forest and Support Vector Machine), improving classification accuracy and predictive performance. Spatially, research is concentrated in Asia and Europe, reflecting disparities in data availability and research infrastructure. Overall, integrating high-resolution topographic data and multispectral imagery consistently yields better performance than single-sensor approaches in capturing karst heterogeneity. However, limitations remain in detecting micro-scale features, methodological inconsistency, and limited multi-sensor and temporal integration. Future research should prioritize standardized, multi-scale, and integrative frameworks incorporating high-resolution data, artificial intelligence, and multi-sensor fusion to enhance mapping accuracy and support sustainable karst management.
Karim et al. (Thu,) studied this question.