Abstract. The synergistic integration of spatial layers (data fusion) within Geographic Information Systems (GIS) is pivotal for enhancing environmental decision-making and spatial planning. However, this process is fraught with inherent complexities, including the necessity for explicitly defined decision objectives, the judicious selection of conceptually and statistically relevant data layers, ensuring robust data quality and balance, devising optimal reclassification strategies, and rigorously validating model outputs through appropriate metrics. Addressing these multifaceted issues, this paper introduces a structured framework complemented by an operational checklist. This methodology is designed to minimize subjective biases, bolster methodological transparency, and significantly enhance the reproducibility of spatial analysis results. The proposed framework is versatile, accommodating a spectrum of data environments: from data-scarce contexts where knowledge-based or statistical approaches, such as the Analytic Hierarchy Process (AHP) or Dempster-Shafer Theory, are crucial for compensating for limited data, to data-rich settings that can leverage advanced machine learning (ML) techniques. Furthermore, the framework explicitly acknowledges inherent limitations. These include the restricted transferability of models to disparate geographical regions, the necessity for localized parameter tuning, and the fundamental understanding that computational modeling, while powerful, serves as a decision support tool rather than a replacement for indispensable field-based assessments. By delineating between addressable methodological challenges and unavoidable contextual constraints, this framework provides a practical and comprehensive guide for advanced GIS-based spatial analysis across critical domains such as natural resource management, environmental assessment, and urban planning.
Azizi et al. (Fri,) studied this question.