HR: 10:55h
AN: H22A-03    [Abstracts]
TI: Obtaining drainage directions in flooded areas: a new burning approach
AU: * Getirana, A C
EM: augusto@hidro.ufrj.br
AF: Water Resources Division - COPPE/UFRJ, Ilha do Fundão - Caixa Postal 68506, Rio de Janeiro, RJ 21941-972, Brazil
AU: * Getirana, A C
EM: augusto@hidro.ufrj.br
AF: Laboratoire des Mecanismes et Transferts en Geologie, LMTG - UMR 5563 UR 154 CNRS Universite Paul-Sabatier IRD, Toulouse, 31400, France
AU: Bonnet, M
EM: bonnet@lmtg.obs-mip.fr
AF: Laboratoire des Mecanismes et Transferts en Geologie, LMTG - UMR 5563 UR 154 CNRS Universite Paul-Sabatier IRD, Toulouse, 31400, France
AU: Rotunno Filho, O C
EM: otto@coc.ufrj.br
AF: Water Resources Division - COPPE/UFRJ, Ilha do Fundão - Caixa Postal 68506, Rio de Janeiro, RJ 21941-972, Brazil
AB: Drainage direction extraction from a Digital Elevation Model (DEM) is becoming a current practice in hydrological studies. Nevertheless, these models are not free of errors and problems may appear along the process. A particular case of error occurs in flat areas caused by inundated areas. In these areas, it is difficult or impossible to identify the exact drainage direction without another source of information. In these cases, it is recommended the execution of corrections in the DEM by using methods such as stream burning (SB). But as it is demonstrated in this paper, there are situations where the use of SB methods is not enough to achieve good drainage directions maps and basin delineations. This paper aims to present a method to obtain drainage directions and consequently more accurate watershed information in flooded areas close to rivers. In this new method the procedures are similar to the ones adopted in the standard SB method. The difference between them is the combination of a classified radar image to the vectors of rivers. In this cense, the DEM is burned twice: in the main rivers with the vectors and in the inundated areas with the classified image. A potential equation is used to vary the deepness burned in each pixel representing the inundated area in the classified image according to the distance of the vectors. In the example, it is used a classified composition map of two JERS-1 images. Different combinations of classes were evaluated and the outputs were compared to previous studies of the same area and validated with the interpretation of the original JERS-1 images. The results demonstrated that the new method offers a better accuracy in obtaining drainage directions and watershed information than the traditional SB methods when dealing with this sort of situation.
DE: 1819 Geographic Information Systems (GIS)
DE: 1821 Floods
DE: 1855 Remote sensing (1640)
DE: 1879 Watershed
SC: Hydrology [H]
MN: 2007 Joint Assembly