@Article{ValerianoRoss:2017:ReLoGe,
               author = "Valeriano, M{\'a}rcio de Morisson and Rossetti, Dilce de 
                         F{\'a}tima",
          affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto 
                         Nacional de Pesquisas Espaciais (INPE)}",
                title = "Regionalization of local geomorphometric derivations for 
                         geological mapping in the sedimentary domain of central 
                         Amaz{\^o}nia",
              journal = "Computers and Geosciences",
                 year = "2017",
               volume = "100",
                pages = "46--56",
                month = "Mar.",
             keywords = "Amaz{\^o}nia, Geological mapping, Geomorphometry, Sedimentology, 
                         SRTM.",
             abstract = "This paper reports procedures to prepare locally derived 
                         geomorphometric data for geological mapping at regional scale in 
                         central Amaz{\^o}nia. The size of the study area, approximately 
                         1.5 million km2, and the prevailing flat topography of the 
                         targeted environment were the constraints motivating the aims, at 
                         spatial and numerical synthesis of the detailed geomorphometric 
                         information derived from SRTM DEM. The developed approach 
                         consisted in assigning single (average) values to terrain patches, 
                         to represent the regional distribution of pixel-based 
                         geomorphometric information (slope, profile curvature and relative 
                         relief). In analogy to the nature of sedimentary packs, patches 
                         were established as contiguous elevation strata, constructed 
                         through a procedure combining segmentation, filterings and range 
                         compressions. For slope only, pre-processing of locally derived 
                         data with median filtering effectively avoided the typical 
                         flattening of the regionalized results due to input distribution 
                         characteristics. Profile curvature was transformed into absolute 
                         values and thus a different meaning from the original (pixel) 
                         variable was considered in the interpretation, also avoiding the 
                         compensation of original values (positive and negative) tending to 
                         zero value when averaged through a regionally flat extension. 
                         Examinations near major river valleys showed patched elevation to 
                         depict alluvial terraces. In the interfluves and floodplains, 
                         contrasting patterns in the averaged variables among patches of 
                         similar elevations allowed the recognition of important relief 
                         features. In addition to the reduction of the distribution ranges, 
                         the correlation between regionalized geomorphometric variables was 
                         higher than observed in the originally local data, due to the 
                         thematic synthesis following regionalization. Depth of dissection, 
                         claimed to be related to the relative age of sedimentary units, 
                         was the main factor to explain the overall variations of the 
                         geomorphometric results. The developed regionalization process 
                         improved the potential of local geomorphometric data for updating 
                         and revision of geological maps and for guiding future surveys in 
                         the sedimentary domain of Amaz{\^o}nia.",
                  doi = "10.1016/j.cageo.2016.12.002",
                  url = "http://dx.doi.org/10.1016/j.cageo.2016.12.002",
                 issn = "0098-3004",
             language = "en",
           targetfile = "valeriano_regionalization.pdf",
        urlaccessdate = "27 abr. 2024"
}