Congress Resources: Papers, posters and presentations

< Return to abstract list

Identification Of Areas With Flooding Potential In South America

Congress: 2015
Author(s): Sergio Rosim, João Ricardo de Freitas Oliveira, Alexandre Copertino Jardim
National Institute for Space Research - INPE1

Keyword(s): Sub-theme 11: Key vulnerabilities and security risks,
Abstract Introduction The American continent has 39.6% of the world's fresh water, and of this total, 61.3% is in South America. The existence of this huge amount of water is directly related to various situations of floods, both in cities and in rural areas, fast and slow floodings. Those floods, which can cause major social and economic damage, must be studied and whenever possible prevented. This work employed a procedure, called Above the Nearest Drainage (HAND) [1], for determining the areas with greatest potential for flooding. The whole South American geographic region was used, considering drainage networks with different densities. These networks were defined by the TerraHidro system [2] which is a distributed hydrological model system that is being developed at in Image Processing Division of the National Institute for Space Research, located in the city of Sao Jose dos Campos, Brazil. HAND determines the variation of relief in relation to the nearest drainage segments of each point, considering a great regular structure representing the relief from their study geographic region. A comparison between HAND flood results and real flooded extracted from Landsat 8 image was done to show the accuracy of the HAND prediction in a real situation. Methods and Materials This work was carried out using TerraHidro system and HAND process, also a TerraHidro process. TerraHidro is a distributed hydrological system created to develop water resource applications. It uses regular grid Digital Elevation Model (DEM) as the surface and elevation structure for drainage extraction. TerraHidro is a plugin of the geographic visualizer TerraView that loads and stores data in a geographical library called TerraLib [3], an open source geographical library implemented in C++ language that has also been developed at the Image Processing Division. This approach has allowed TerraHidro project team of designers and programmers to keep focused on the development of TerraHidro system functionality. TerraLib is an open-source GIS software library. TerraLib supports coding of geographical applications using spatial databases, and stores data in different DBMS including MySQL, PostgreSQL and other databases. The TerraHidro funcionalities used here are: first, the definition of local flow, called of Local Drain Directions (LDD) extraction [4]. For each DEM grid cell the LDD was defined considering the steepest downstream regarding the 8 neighbors grid cell. At the end of the task, a new grid was created with the same number of columns and rows of DEM and same resolution. Each grid cell received a code indicating the water flow from this cell; second the creation of the grid called contribution grid area. The user wants to work only with representative drainages regarding his application, not with drainages of all LDD grid. Each cell of the contribution area grid receives a value that is the amount of the areas of all cells that participate in the path arriving at that cell; third, the definition of a particular drainage network, using a thresold value. The value of each cell of the contribution area grid is compared with the thresold value. If the value of contribution area grid is iqual or greater than the threshold value the cell is selected as a drianage network cell. At the end of this process a new grid is created, defining the draiange network. TerraHidro uses the HAND procedure to identify these areas. It calculates, for every DEM cell, the altimetry difference between this cell and the nearest cell belonging to the drainage network, following the local drain directions. As the HAND terrain descriptor is sensitive to drainage changes in the regions of sudden terrain variations, it was used as an attempt to determine critical drainage areas. The materials used to develop this work were SRTM of 90 meters of horizontal resolution as surface elevation data set. TerraHidro and HAND used this data set to extract their information. A Landsat 8 image acquired on July/05/2014, showing a flooded region was compared with HAND result of the same area. Results and Discussion Results of HAND information extraction based on TerraHidro drainage network definition were shown as thematic maps containing altimetry tracks. Each altimetry track informs the greater or lesser proximity of its altimetry in relation to the nearest drainage altimetry. Different scenarios were created with the use of altimetry tracks with different relief intervals. The result verification was done by using a Landsat 8 image, bands 4, 5, 6, orbit point 224/080, of São Borja city region, in the south of Brazil. A color composition using these three bands was generated and the HAND result was overlaid on the image to check the degree of coincidence between the two flooding areas. The result discussion concerns the relationship between selected drainage and HAND results. Each drainage gives a particular potential flood region context. For quick floods it is better to have dense drainages to grasp the water behavior in small regions. For slow floods the drainage can be less dense. In this case, river the larger floods must be considered. Conclusion TerraHidro and HAND are used to define respective drainage network and potential flooded areas for the South America region. These areas serve as priority areas for experts to carry out studies on flooding. A comparison between HAND result and a real flooded area shown by a satellite image revealed coincidence between both flood areas. This confirms that the methodology shown in this paper is useful to determine critical areas for flooding. [1] Rennó, C. D., Nobre, A. D., Cuartas, L. A., Soares, J. V., Hodnett, M. G., Tomasella, J., Waterloo, M. J., "HAND, a new terrain descriptor using SRTM-DEM: Mapping terra-firme rainforestenvironments in Amazonia," Remote Sensing of Environment 112 3469–3481, (2008). [2] Rosim S., de Freitas Oliveira J. R., Jardim A. C., Namikawa L. M., Rennó C. D. " TerraHidro: a distributed hydrology modelling system with high quality drainage extraction". GEOProcessing 2013 : The Fifth International Conference on Advanced Geographic Information Systems, Applications, and Services, February 24 - March 1, Nice, França. 2013. [3] Camara, G, Souza, R.C.M, Freitas, U.M, Garrido, J. "SPRING: Integrating remote sensingand GIS by object-oriented data modelling. Computers & Graphics, 20: (3) 395-403, May-Jun 1996 [4]Burrough, P. A., McDonnell, R. A. "Principles of Geographical Information Systems". New York: Oxford University Press. 1998
© 2011 IWRA - International Water Resources Association office@iwra.org - http://www.iwra.org - Admin