Abstract. In October 2024, southeastern Spain was severely impacted by the DANA (Depresión Aislada en Niveles Altos), a high-intensity weather event that triggered extensive flooding, disrupted infrastructure, and caused major damage across both urban and rural landscapes. Rapid and accurate mapping of flood extent is critical to support emergency management efforts and inform mitigation strategies. This study focuses on a highly vulnerable area near Valencia, where the goal is to map with maximum precision the flooded zones located within the defined study boundaries. For this purpose, the effectiveness of five spectral indices commonly used for flood detection is evaluated: the Normalized Difference Water Index (NDWI), the Modified NDWI (MNDWI), the Automated Water Extraction Index (AWEI), the Normalized Difference Flood Index (NDFI), and the recently introduced Flood Mud Index (FMI), which is specifically designed to detect sediment-rich waters. The indices are calculated using Landsat-8 imagery acquired shortly after the DANA event. A supervised classification is then performed using the Maximum Likelihood Classification algorithm to separate flooded and non-flooded covered areas for each index. The analysis revealed significant differences in the capacity of each index to delineate flood extent, especially in turbid waters, where traditional indices tend to underperform. Among all tested methods, the FMI consistently produced the most accurate and spatially coherent results. The FMI achieved an Overall Accuracy of 0.9764 and identified a flooded area of 4081.490 hectares. These outcomes emphasize the importance of selecting suitable spectral tools based on floodwater characteristics and demonstrate how remote sensing methods can play a key role in supporting urban disaster response and recovery following extreme events such as DANA.
Alcaras et al. (2026) studied this question.