Application of Artificial Intelligence in the Evaluation of Positional Accuracy and Statistical Validation of Digital Elevation Models for Hydrological Studies

Abstract

This study uses artificial intelligence to comprehensively evaluate digital elevation models (DEMs), specifically SRTM, AlosPalsar, and ASTER, in the Moquegua region of Peru. Three recognized standards were used to evaluate the positional accuracy of DEMs: EMAS, NMAS, and NSSDA. The DEMs were also assessed through correlation, the coefficient of determination (R2) and the Bland-Altman Graph, which allowed us to understand and visualize the relationship and agreement between the elevations extracted from the DEMs and the altimetric control network of the national chart of Peru at a scale of 1:25000. The correlation and R2 revealed a solid relationship and a high degree of explanation for the variability of the elevations observed by the MDEs. The Bland-Altman plots confirmed the agreement between the elevations predicted by the MDEs and those observed at the points of the altimetric control network. This study highlights the importance and value of combining artificial intelligence techniques with statistical validation methods and positional accuracy standards to ensure the accuracy and reliability of EDMs in hydrological applications, thus providing a robust and verifiable framework for future research in this domain.