Flood delineation in a large and complex alluvial valley, lower Pabasin, Mexico

Determining the extent of flooding is an important role of the hydrological research community and provides a vital service to planners and engineers. For large river systems located within distant settings it is practical to utilize a remote sensing approach. This study combines a remote sensing and geomorphic approach to delineate the extent of a large hurricane generated flood event in the lower Panuco basin (98,227 km2), the seventh largest river system draining into the Gulf of Mexico. The lower Panuco basin is located within the coastal plain of eastern Mexico and has a complex alluvial valley. Data sources included a Landsat 5TM and Landsat 7ETM+ scene, and topographic and particle size data from fieldwork and laboratory analysis. The Landsat 5TM image was acquired after the peak of a large flood event in 1993, whereas the Landsat 7ETM+ scene was acquired during the dry season in 2000. The increasing number of days between flood crest and the date of flood image acquisition along the river valley provided the opportunity to examine several methods of flood delineation and to consider differences in floodplain geomorphology. Backswamp environments were easily delineated in flooded reaches within the Panuco and Tamuin valleys, whereas in the Moctezuma valley more sophisticated methods were required because of the greater time between image acquisition and flood peak, and the complex floodplain topography. This included Principal Component (PC) analysis and image classification. Within the floodplain, residual Holocene terraces complicated flood mapping. Classification of both images allowed consideration of the influence of permanent standing water. Although the flooded areas were greater in the lower reaches of the study area, because this portion of the valley contained large floodplain lakes, the amount of inundation was actually lower. Remote sensing offers the ability to examine large alluvial valleys in distant settings but does not imply that geomorphic criteria should be excluded. Indeed, because of heterogeneous floodplain topography this study illustrates the importance of including field based geomorphic analysis so that the complexity of distinct floodplain environments are considered. The findings from this study are significant because most remote sensing data obtained for the purpose of flood mapping will not coincide with the flood crest. Thus, this study provides an appropriate method for mapping flood inundation in large and complex