Análise da qualidade ambiental em remanescentes de Cerrado por meio de avaliação in loco e sensores remotos multiescalares

Abstract

Environmental quality monitoring in remnants of native vegetation is an important step for biodiversity conservation in ecosystems such as the Cerrado, which are threatened by intensive land use in a context of ongoing climate change. This assessment typically faces challenges related to spatial coverage and precision: traditional field-based methods cover small areas at high cost, whereas the predominant use of Remote Sensing data at medium–high and/or moderate spatial resolution is often not accompanied by field protocols that ensure the reliability and suitability of environmental and cartographic products for monitoring objectives. In this context, the general objective of this thesis is to develop and evaluate a set of methodological procedures that integrate advanced remote sensing techniques, based on airborne and orbital platforms, with in situ environmental quality assessments. More specifically, it aims to understand how remote sensing products, including vegetation indices derived from RGB aerial images, capture interactions among vegetation, water quality and overall ecosystem health, using a multisensor and multiscale approach. Different methodological tools were employed, including literature review; acquisition and processing of Planet and Sentinel-2 satellite images; fieldwork using an Unmanned Aerial Vehicle (UAV/drone); application of the Rapid Assessment Protocol (RAP) to characterize environmental quality in vegetation fragments; measurements of physicochemical water parameters with a multiparameter probe; organization of a geographic database; statistical analyses (with emphasis on Principal Component Analysis and regression models); and the production of thematic maps. The results indicate significant correlations between vegetation indices obtained from remote sensing and biophysical–chemical data, with a predominance of environmental quality classes classified as “High” and “Moderate” in more than 60% of the sampled sites, according to the criteria adopted. A consistent correlation was also observed between Principal Component Analysis (PCA) and physicochemical water parameters, highlighting the role of variables such as pH, total dissolved solids (TDS) and electrical conductivity. The integration of the PAR with remote sensing data from UAV, Planet and Sentinel-2 enabled the assessment of environmental quality considering multiple landscape aspects, such as channel geomorphology and land use/land cover. The spectral indices GLI, PRI, RGR, TGI and VARI proved effective in discriminating environmental quality classes, evidencing the potential of remote sensing techniques integrated with in situ evaluations for detailed monitoring of areas impacted by human activities. The proposed method showed adequate capacity to identify both conserved areas and stretches requiring restoration, a capacity enhanced by the integration of multisensor and multiscale data. Taken together, the results demonstrate the potential of these techniques for monitoring and evaluating environmental assets of the Cerrado, such as water, soil, climate and biodiversity, at multiple spatial and temporal scales, providing support for public policies and conservation strategies in this critical and threatened biome.