Geomorphometric pattern recognition of SRTM data applied to the tectonic interpretation of the Amazonian landscape

The Amazon landscape spatial variability and anisotropic trends in the Uatumã and Urubu River regions are evaluated using geomorphometric techniques such as fractal dimension, drainage network density and semivariogram. These procedures were applied to the digital elevation model (DEM) of the Shuttle Radar Topography Mission (SRTM). This evaluation facilitated the definition of geomorphometric domains with different degrees of roughness (fractal dimension) and elevation (semivariogram). These areas are consistent with known qualitative relief types. Furthermore, known geological structures in the subsurface and surface apparently influence the spatial variability of these geomorphometric variables. This is the Silves area case, where the hilly topography exhibits several annular and radial rivers, denoting subsurface control due to faults and folds that were mapped by seismic surveys. Another possible influence example is the spatial coincidence between structures mapped by magnetic data with low dissection zones delineated as a result of drainage network density analysis. In addition to the spatial distribution, the anisotropic trends of these geomorphometric variables were analyzed and compared with geological and geophysical information. Results indicate that the predominant directions for topographic semivariance anisotropy are NNE–SSW and NE–SW for the interfluvial regions, as well as NW–SE for the alluvial plains. The highest agglomeration direction in the drainage network, as shown by its anisotropy, coincides with the studied region’s current maximum horizontal stress direction, except in the floodplains. The direction with the most pronounced roughness, NNE–SSW, coincides with the direction of waterfalls and rapids. This study demonstrates that spatial variability knowledge and anisotropic trends of geomorphometric parameters is useful to understand the geology and geomorphology of the Central Amazon region.