Avaliação da qualidade em redes de drenagem processadas a partir de modelos digitais de elevação

Abstract

In many environmental studies it is important to know the quality of drainage network generated by digital elevation model (DEM) processing, as studies that works with drainage networks derived from different procedures obtained from different sources or test new algorithms of obtaining flow directions. In these and in other cases it is necessary evaluate drainage network quality, judging how close it is to the automatically generated drainage system of quality drainage system known, taken as reference. Therefore, the are cited in the literature different evaluation techniques for assessing the quality of drainage network. In most cases, these techniques are utilized with little or no discretion. The present research aims at the assessment of techniques used for the analysis of drainage networks, conducting a survey of the main techniques and metrics cited in the literature identifying the main limitations and potential for the application of these techniques in high and low drainage networks resolution. The Uruguay river basin. Was taken as a case study. It was generated drainage networks obtained from DEM processing in high-resolution spatial and low spatial resolution by manual tracing, processed from DEM for resampling and upscaling of flow directions. The evaluation of drainage networks was evaluated in qualitative and quantitative form from various criteria and metrics, taking as reference, in the most cases, the network vector of LANDSAT imagens, but in some cases, the comparison or evaluation was done taking the low resolution done manually as the reference network. It was verified that the visual inspection of drainage systems was essential for evaluation of drainage networks, as allowed critically evaluate the performance of different techniques for assessing the quality of drainage networks. It can be seen that the techniques are based on the comparison of drainage network characteristics failed to quantitatively express the actual performance of the traces of the drainage network and the errors are larger for the rivers belonging to the largest drainage basins area. The techniques to quantify the area formed between the drainage systems without having to take into consideration the length of the drain. The spatial comparison techniques to quantify areas out of agreement with the watershed high resolution and average distance between the boundaries of river basins showed adequate results for assessing the quality of drainage networks. The quantification of the percentage of drainage networks inside the buffer could identify the standard of quality seen in visual inspection. It is expected that this study has provided results that work as a guide for the development of future work using DEM to obtain drainage network and as intermediate step the quality of drainage network processed.