The objective of this work was to evaluate multitemporal data, obtained by moderate resolution imaging spectroradiometer (MODIS) sensor, for the study of spatial-temporal dynamics in two subregions of the Pantanal biome. One hundred and thirty nine enhanced vegetation index (EVI) images, from MOD13 vegetation index product, altimetry data from shuttle radar topography mission (SRTM) and tropical rainfall measuring mission (TRMM) precipitation data were used. In order to reduce data dimensionality, MODIS EVI images were sampled based on contour lines spacing of 10 m. The autocorrelation and cluster analysis were used for spatial and temporal evaluation of the samples; and the principal components analysis was applied to all dataset for spatial and temporal analysis. Results showed a spatial and temporal dependence between spectral response and precipitation. The cluster analysis indicated two spatial groups, suggesting the need for the analysis of the entire study area. The principal components analysis allowed to distinguish four behaviors: the areas permanently flooded; nonflooded areas composed by vegetation; flooded areas with higher spectral vegetation response; and riparian vegetation areas.
remote sensing; time series; cluster analysis
