ABSTRACT: Relief characterization using a digital terrain model (DTM) is widely applied in erosion, soil and vegetation modeling. However, factors, such as acquisition technology and the spatial resolution of the digital model, affect modeling results. The aim of this study was to characterize noises in a DTM of the entire state of Santa Catarina recently made available through the state's Sustainable Economic Development Secretary and to evaluate different methods of interpolation and smoothing of the original 1 m resolution to a new digital model with a spatial resolution of 15 m. Using the SAGA GIS program, spurious data that appeared as peaks and sinks were removed from the digital model. Of five processing procedures, the following three were used for smoothing: a Gaussian filter, a Lee filter, and mesh denoising. The remaining two were for interpolation: nearest-neighbor, and ordinary kriging. Altimetric reference data were collected in the study area of 11,597 km2 with two dual-frequency GPS RTK receivers located in the Celso Ramos community in the municipality of Frei Rogério (SC). The root mean square error showed that the documented values of the aerial survey report were consistent with the findings of this study. However, the GPS RTK data showed a difference of 1.07 m compared with the original DTM with 1 m spatial resolution. There were pixels with peaks and sinks in the digital model of the Santa Catarina State.