Efeitos de mudança de resolução espacial em representações topográficas e redes de drenagem derivadas de modelo digital de terreno LiDAR de alta resolução espacial

Abstract

The use of digital terrain models (DTM) has great importance for environmental studies, for example, hydrological studies. In recent decades, due to technological advances, several DTM products have become free, among them, products with high spatial resolution, especially derived from LiDAR technology. The high spatial resolution presents greater detailing of the topographic characteristics, which justifies the acquisition of such models. However, working with high spatial resolution models imply on higher computational cost and require high-performance equipment, which may make the use of these models unfeasible. One of the alternatives to enable the use of this type of model is to coarsen the spatial resolution. However, this procedure causes loss of information. The objective of this work is to evaluate the effects caused by the change of spatial resolution in topographic representations and drainage networks extracted from high spatial resolution digital terrain models derived from LiDAR technology. For this, three resampling techniques were applied: mean aggregation, bilinear interpolation and nearest neighbour, to coarsen a DTM at 1-meter spatial resolution in multiple resolutions (2, 5, 10, 30 and 100 m). The origin of DTM data is the Pernambuco Tridimensional – PE3D program. The sub-basin of the Sirinhaém river basin with approximately 550 km² was taken as case study area. The evaluation of the results was carried out using eight metrics: quantification of operational cost, comparison between cross-sectional topographic profiles (PTT), relative variation of hypsometric curves, visual inspection of drainage networks, average distance between drainage networks tracings, relative variation of river length, percentage within the buffer, watershed delimitation area and computational cost quantification (processing time and space size in disk). The results indicate that there was no significant difference between the resampling techniques. On the other hand, among spatial resolution there were considerable differences, varying according to the applied metric. In the metrics that evaluated topographic representations, a minor difference was identified in the PTT for a resolution of 100 meters. The assessment of drainage networks, showed that there were displacements of confluences and a decrease in the length of rivers. There was no significance difference between the delimitation of the catchment area. Furthermore, the computational cost reduced significantly with resolution coarsening. In general, the spatial resolution of 2 meters presented the best results, being the most indicated to use in case of resample need. The others spatial resolution in order of the best performance were 10, 5, 30 and 100 meters.