Floods are natural hazards that have shaped the landscape of our environment over thousands of years. Recently, the effect of floods has been intensified by the sustained population growth, changes in land use, and climate change with more intense rainfall patterns. The impact of these effects depends on the extent of the flood and temporal nature of their occurrence, as well as on the vulnerability and associated risk of the exposed communities and elements (infrastructure, ecosystems, etc). Most flood hazards result from insufficient drainage systems, breaching or topping of levees or dams, and intense rainfall events (Kulkarni, et al., 2014).
One of the most devastating effects of flood events is the loss of life, along with the damage to properties. Other effects include economic loss, disturbance of ecological resources, food shortages and starvation (Haq et al., 2012). The amount of warning time that is provided to the population at risk of flooding has been described as key factor to prevent the loss of lives (Graham, 1999). Therefore, adequate prediction capability is essential to mitigate flooding damages and establish flood-warning systems.
Floods have a complex spatial and temporal dynamic that motivates their study. The spatial dimension of floods ranges from the local to the regional scale, while the temporal dimension varies from slow floods (days) to flash floods (minutes). The spatial extent, variability and magnitude of floods has been addressed using Geographic Information Systems (GIS) (Crossetto et al., 2000), while the temporal dynamic has been tackled using hydrologic and hydraulic modeling (Correia, et al., 1998).
The coupling of GIS and flood modeling is a developing field. There are three approaches in the literature describing the role of GIS in flood modeling: The first approach is called Loose Coupling, which consists in using GIS for pre-processing and post-processing of model data inputs and outputs. A second approach called Tight Coupling allows model inputs and outputs be directly addressed by GIS through a common database; and a more challenging and modern approach is called Embedded Coupling, in which the model and GIS are completely integrated through programming (Jankowski, 1995, Al-Saban et al., 2003, De Roo et al., 2000).
The use of GIS in pre-processing has been widely acknowledged for specific applications such as watershed delineation, terrain processing, and cross section discretization, which allow obtaining a computer representation of the spatial and terrain characteristics of the study area for further computer modeling (Paiva, et al., 2011, Haq et al., 2012, Sarhadi et al., 2012). Post-processing applications include inundation maps that display the depth and spatial extent of the floods, assessment of population and infrastructure impacts, spatial-uncertainty and sensitivity analysis, among others (Crossetto et al., 2000). These applications of GIS can either be used for computer modeling, or...