Soil erosion poses a significant global environmental challenge, causing land degradation, deforestation, river siltation, and reduced agricultural productivity. Although the Revised Universal Soil Loss Equation (RUSLE) has been widely applied in Brazil, its use in the tropical river basins of the Amazon remains limited. This study aimed to apply a GIS-integrated RUSLE model and compare its soil loss estimates with multiple linear regression (MLR) models based on terrain attributes, aiming to identify priority areas and key geomorphometric drivers of soil erosion in a tropical Amazonian river basin. A digital elevation model based on Shuttle Radar Topography Mission (SRTM) data, land use and land cover (LULC) maps, and rainfall and soil data were applied to the GIS-integrated RUSLE model; we then defined six risk classes—slight (0–2.5 t ha⁻¹ yr⁻¹), slight–moderate (2.5–5), moderate (5–10), moderate–high (10–15), high (15–25), and very high (>25)—and identified priority zones as those in the top two risk classes. The Caeté River Basin (CRB) was classified into six erosion risk categories: low (81.14%), low to moderate (2.97%), moderate (11.88%), moderate to high (0.93%), high (0.03%), and very high (3.05%). The CRB predominantly exhibited a low erosion risk, with higher erosion rates linked to intense rainfall, gentle slopes covered by Arenosols, and human activities. The average annual soil loss was estimated at 2.0 t ha⁻¹ yr⁻¹, with a total loss of 1005.44 t ha⁻¹ yr⁻¹. Additionally, geomorphological and multiple linear regression (MLR) analyses identified seven key variables influencing soil erosion: the convergence index, closed depressions, the topographic wetness index, the channel network distance, and the local curvature, upslope curvature, and local downslope curvature. These variables collectively explained 26% of the variability in soil loss (R² = 0.26), highlighting the significant role of terrain characteristics in erosion processes. These findings indicate that soil erosion control efforts should focus primarily on areas with Arenosols and regions experiencing increased anthropogenic activity, where the erosion risks are higher. The identification of priority erosion areas enables the development of targeted conservation strategies, particularly for Arenosols and regions under anthropogenic pressure, where the soil losses exceed the tolerance threshold of 10.48 t ha⁻¹ yr⁻¹. These findings directly support the formulation of local environmental policies aimed at mitigating soil degradation by stabilizing vulnerable soils, regulating high-impact land uses, and promoting sustainable practices in critical zones. The GIS-RUSLE framework is supported by consistent rainfall data, as verified by a double mass curve analysis (R² ranging from 0.64 to 0.77), and offers a replicable methodology for soil conservation planning in tropical basins with similar erosion drivers. This approach offers a science-based foundation to guide soil conservation planning in tropical basins. While effective in identifying erosion-prone areas, it should be complemented in future studies by dynamic models and temporal analyses to better capture the complex erosion processes and land use change impacts in the Amazon.

Predictive Digital Soil Mapping (PDSM) in Eastern Amazonia faces challenges due to its environmental complexity, difficult access, and scarce legacy data. While legacy soil maps contain valuable tacit knowledge, updating them requires methods that can handle uncertainty. This study evaluates the integration of old soil maps with machine learning to update soil information in Tracuateua, Pará, with a specific focus on the performance of ensemble learning and the explicit incorporation of uncertainty metrics in soil mapping units under hydromorphic influence, which, in addition to being difficult to access, are influenced by complex pedogenetic processes. We combined 270 sampling points, equivalent to the total pixels that captured the variability of soil mapping units, with environmental covariates and historical data. Several algorithms were tested, including an ensemble approach, to predict mapping units and quantify uncertainty through entropy and confusion indices. The ensemble model demonstrated improved stability and reduced classification uncertainty compared to single models, particularly in challenging hydromorphic environments. Although accuracy gains were modest, the models captured soil–environment relationships, with climate as: Annual Mean Temperature 22,000 years ago (Tmean_22k), relief: Channel Network Base Level (CNBL and altitude) and organism variables: Land Surface Temperature (LST) emerging as the main predictors. Spatialized uncertainty estimates, expressed through entropy and the confusion index, provide a practical decision-support tool for guiding field surveys and identifying areas of low mapping reliability. By explicitly transferring the pedologist’s mental model—encoded as tacit knowledge in legacy soil maps—into ensemble learning, this study presents a robust and transferable framework for updating soil maps in data-scarce tropical regions, balancing predictive performance, spatial consistency, and uncertainty-aware interpretation.

Context: Pedological mapping in the Amazon is essential but highly challenging, particularly in hydromorphic environments where soil properties and boundaries are difficult to delineate. Improving mapping accuracy is critical for sustainable land-use planning and ecosystem management, yet conventional approaches are often time-consuming and costly. Digital soil mapping (DSM) offers a promising alternative by integrating machine learning algorithms with geospatial data to predict soil mapping units. Aims: This study evaluates DSM approaches in the Eastern Amazon using legacy soil data, environmental covariates, and multiple machine learning algorithms, with emphasis on classification accuracy and spatial uncertainty. Methods: We compared the performance of eight machine learning algorithms for predicting soil mapping units using predictors derived from a digital elevation model, Landsat-9 imagery, and legacy soil maps. A total of 244 spatial sampling units (30 m × 30 m pixels), corresponding to the centroids of soil mapping units and supported by field observations across nine soil classes, were used, with 70% allocated for training and 30% for validation. Recursive Feature Elimination was applied to identify the most relevant predictors. Model performance was assessed using overall accuracy, Kappa index, user’s accuracy, and producer’s accuracy. Key results: The random forest implementation algorithm (ranger) achieved the best performance, with an overall accuracy of 0.74 and a Kappa index of 0.71. Other algorithms showed moderate agreement with the conventional pedological map (accuracy = 0.56; Kappa = 0.49), whereas the artificial neural network showed low performance (accuracy = 0.26; Kappa = 0.14). Conclusions: Results indicate that the density and spatial configuration of soil mapping units influence the agreement between conventional and digital maps, particularly in hydromorphic environments where classification uncertainty is higher. Implications: These findings highlight the potential of decision tree-based algorithms for DSM in complex tropical landscapes, while emphasizing the need for further research to address limitations related to data availability, spatial resolution, and the quality of reference soil maps.

Land use and land cover (LULC) changes are major drivers of environmental transformation in sensitive regions, such as the Marajó Archipelago in the Brazilian Amazon. This study assessed accumulated anthropization of LULC in the Immediate Geographic Region of Soure–Salvaterra, Eastern Amazon, between the reference years 1985 and 2023, using MapBiomas data and spatial statistical techniques. Bivariate Local Moran’s I (LISA) was applied to evaluate intertemporal spatial associations between areas classified as natural in 1985 and anthropized in 2023. In this approach, High–Low indicates natural areas associated with low anthropization in 2023, whereas High–High indicates areas where natural cover in 1985 was spatially associated with higher anthropization in 2023. The results indicated a strong predominance of High–Low, with values above 94% in all municipalities and up to 99.86% in Santa Cruz do Arari. In contrast, High–High had localized concentrations in Salvaterra (3.53%), Cachoeira do Arari (1.28%), and Soure (1.16%), especially in coastal zones and inland sectors. Low–Low, associated with lower anthropogenic pressure or possible signs of natural regeneration, was extremely low (≤0.0004%). These findings indicate that LISA is useful for identifying local LULC patterns and supporting environmental assessment and territorial planning in tropical regions.