Delineation of main relief subdomains of central Amazonia for regional geomorphometric mapping with SRTM data

Abstract

The geomorphometric characterization of central Amazonia reliefs is specially challenging because of its large dimension and rather uniform low topography of sedimentary units. In such conditions, usual techniques applied for general geomorphometry are ineffective and thus disabling direct bottom-up approaches for delimitations. Specific geomorphometry is a more promising approach for terrain analysis in this area, since procedures can be adapted to analyze the particular features of each geomorphological context. Therefore, delimitation of subdomains is a requirement for specific geomorphometry, to address proper approaches in the different strata. This paper presents a first level delimitation of major subdomains of Central Amazonia, carried out through combined techniques of geoprocessing guided by visual assessment. The three targeted subdomains, floodplains, plateau and highland, were mapped at the 1:2,000,000 scale using SRTM and derivative data as inputs. The relief of highland, as influenced by Precambrian to Cretaceous basement rocks, was firstly delimited using mostly height and relative relief combinations. The partition of the remaining lowland area into plateaus (Neogene) and floodplains (Quaternary sediments) was based on residual relief (subtraction of regional relief from full detail elevation data), with a regional relief modeled as the 9th order trend surface of elevations. Thresholds of negative residues were determined to create extents of visually assessed boundaries between plateau and floodplain areas. The continuous visual supervision of procedures and concerning interventions were based on interpretation of output displays of DEM and derived data prepared with image processing techniques. Delineation of subdomains was implemented through steps of supervised classifications and map algebra to operationally enable the task over the full area extension and corresponding data volume. Observation of data combined with knowledge and conceptual inputs were decisive to provide useful and unambiguous criteria to choose specific variables and analyses along this development.