A real-time flood forecasting and simulation system based on GIS and DEM: Analysis of sensitivity to scale factors

The hydrometeorological telemetric networks in real time interrelated with weather forecasting and rainfall information obtained from remote sensing, constitute real forecasting and protection instruments in the event of flash flooding, so typical of semiarid environments.; In this Thesis, spatial analysis approached with functions embedded in a Geographical Information System (GIS) are proposed. The aims are: (a) To combine efficiently information from different sources (telemetric networks and radar-satellite technology). (b) To develop methodology of application of spatially distributed and hybrid hydrologic models, which are topographically based and event-oriented. (c) To extract automatically from Digital Elevation Models (DEM) the relevant parameters of the hydrologic models used.; When extracting the drainage networks from a DEM, various questions arise: what is the most suitable drainage density for the hydrographic network? What degree of affection does the selection of DEM cell size have on the hydrologic results, or are they not sensitive to it? Can any invariable property by defined with the scale which characterizes indexes or parameters based on the drainage network hierarchy?; A clear inter-relationship can be seen between the geomorphological and hydrologic parameters and the DEM resolution. The morphometric parameters are also affected by threshold area variation. It is proposed a methodology to identify a priori the range of DEM resolutions and threshold areas for in which the parameters present a certain stability for modelling based on drainage networks topology.; When working with spatially distributed models, several questions crop up: Are the distributed parameters derived from DEM and the complete hydrologic results affected by cell size? Is it feasible to identify invariable properties with the scale which characterizes the spatial distributions of the parameters?; The terrain slope and the flow path length are affected by the DEM cell-size adopted. The spatially distributed flow velocity presents invariance properties with the DEM scale, founded on the behaviour of the cumulative drainage area distribution curve. Thus, the hydrographs simulated by means of distributed UH models, present slight sensitivity to the DEM cell size. Scale exponents, invariable with DEM resolution, of power distributions which characterize the behaviour of certain distributed parameters, have been identified.