This study evaluates the effectiveness of combining remote sensing techniques with the Random Forest algorithm for estimating the Periodic Annual Increment (PAI) in a dry tropical forest located within the Caatinga biome in northeastern Brazil. The analysis integrates forest inventory data collected from permanent plots monitored between 2011 and 2019 with Landsat satellite imagery processed through the Google Earth Engine platform. By incorporating surface reflectance and vegetation indices, the approach significantly improved the accuracy of productivity estimates while reducing the costs and efforts associated with traditional field-based methods. The Random Forest model achieved a strong performance (R² = 0.8867; RMSE = 0.87), and its predictions were further refined using post-processing correction factors. These results demonstrate the potential of data-driven modeling to support forest monitoring and sustainable management practices, especially in ecosystems vulnerable to the impacts of climate change.

The uncontrolled expansion of human activities may lead to a reduction in vegetation cover, an increase in erosion processes and soil sealing. The aim of this study is to examine forest cover using the Weights of evidence method based on land use maps between the years of 1996 and 2011 in the micro-region of the Campanha Ocidental located in the state of Rio Grande do Sul (Brazil). The spatial database was constructed in SPRING software (version 5.2.1) based on LANDSAT 5 images, which were georeferenced and classified. Geophysical and socioeconomic variables were included in the database for further analysis in Dinamica EGO software (version 2.2.8). In order to parameterize the probabilistic model for the analysis of the dynamics of forest cover change, we calculated the percentage of class change through transition matrices; calculated the intervals for discretization of continuous variables; calculated the weights of evidence (W+); analyzed the correlation between the weights of the input variables for all transitions; simulated future scenarios and; validated the simulated final map based on the historical map. This model was adequate for understanding the variables that most contribute to forest cover change in the region. The results showed that the emergence of new forest areas was influenced by hypsometry, distance to sandy lands and per capita income, while deforestation by rural population and distance to the sandy lands.

This study evaluates the effectiveness of remote sensing combined with the Random Forest algorithm compared to traditional forest inventory methods for estimating the Annual Periodic Increment (API) in a dry tropical forest within the Caatinga biome. Data were collected from permanent plots monitored between 2011 and 2019 and integrated with Landsat imagery processed in Google Earth Engine. The results show that both surface reflectance and vegetation indices significantly improves the accuracy of API estimates while reducing fieldwork costs and efforts. The Random Forest model demonstrated robustness (R²= 0.8867, RMSE=0.87), further enhanced with post-processing correction factors, offering a sustainable approach for forest management in ecosystems vulnerable to climate change.

O inventário florestal convencional é realizado para obter os dados dendrométricos de uma floresta, e o sensoriamento remoto vem sendo bastante utilizado como uma nova tecnologia associada ao inventário florestal, uma vez que ajuda a diminuir os erros de amostragem e promove a redução de custos. A utilização de aeronaves remotamente pilotadas permite obter dados avançados que podem gerar resultados de altimetria de determinada área. Com isso, o objetivo deste estudo foi apresentar método que apoie o inventário florestal convencional na obtenção da detecção de árvores individuais, delineamento de copas de árvores individuais e estimativa de parâmetros de árvores individuais com foco nas Florestas Tropicais secas do semiárido Pernambucano. O estudo foi realizado no município de Floresta-PE, onde foi realizado o inventário florestal de 40 parcelas, coletando os dados de altura e CAP de cada árvore. As imagens foram obtidas utilizando um ARP de modelo Phantom 4 PRO com sensor RGB embarcado, adquirindo 346 imagens, que posteriormente criaram a nuvem de pontos. Como resultado do processamento, foram obtidas a ortoimagem, MDT, MDS e MDA. O método proposto se mostrou efetivo no apoio ao inventário florestal, sendo capaz de detectar árvores individuais e, por conseguinte, delinear as copas, mensurar alturas e diâmetros das espécies florestais que ocorrem em Floresta Tropical Seca.

O presente estudo tem por objetivo analisar a influência do relevo na ocupação das florestas. Para tanto, realizou-se o cruzamento do mapa de uso e cobertura da terra oriundo da classificação Maxver, com mapas de declividade e altitude.