Abstract Understanding the spatial pattern of a particular geographic phenomenon such as deforestation is a key issue to establish monitoring programs to prevent the depletion of natural resources. Thus, the goal of this study was to assess the spatial pattern of deforested areas in the Pardo and Jequitinhonha River basins using Ripley’s K function. First, we mapped all deforested areas in these basins using Landsat multispectral imagery from 2007 to 2015. Then, we used the Ripley’s K function to test for spatial interactions between deforestation events. Our results showed that deforestations predominantly occur in a clustering spatial pattern in these basins. Spatial statistical analyses as Ripley’s K function may provide a baseline for deforestation monitoring, as well as allowing us to understand the spatial pattern of deforestation in different natural ecosystems, especially in countries like Brazil, where the territorial dimension presents a great difficulty for the effectiveness of deforestation monitoring.

Resumo O monitoramento dos recursos naturais do nosso planeta é essencial para a obtenção de informações que possam subsidiar estratégias de conservação e utilização sustentável desses recursos. Tais estratégias se tornam ainda mais importantes em biomas ameaçados, como é o caso do Cerrado brasileiro, uma das savanas mais ricas e ameaçadas do mundo. Comumente o monitoramento das florestas é realizado por meio de inventários florestais, atividade onerosa, cara e extremamente difícil de ser realizada periodicamente em florestas extensas, como o Cerrado. Nesse sentido, o presente estudo teve como objetivo atualizar o mapeamento do estoque de volume e carbono em fragmentos de Cerrado, diminuindo a quantidade de levantamentos de campo por meio da aplicação de geotecnologias. Utilizaram-se dados de levantamento de campo de 61 fragmentos de Cerrado, sendo que em 25 deles foram realizadas 2 medições em um intervalo de 5 anos. Através dos dados de campo e valores de reflectância de imagens Landsat, modelos lineares múltiplos foram ajustados e validados. Posteriormente, esses modelos foram aplicados aos fragmentos que foram mensurados apenas uma vez. Dessa forma, foram obtidos os valores de volume e carbono para todos os 61 fragmentos, no ano da segunda medição, gerando mapas de produtividade atualizados.

Abstract Monitoring natural resources of our planet is essential to gather information to support strategies for conservation and a sustainable use of these natural resources. This monitoring is even more important in endangered biomes, as is the case of brazilian savannas, one of the most threatened and richest, in terms of biodiversity, savannas in the world. Usually, forest monitoring is accomplish by measuring the trees on the field, the so-called forest inventories, which are expensive, costly and extremely difficult to be performed periodically in extensive forests, as is the case of the Brazilian savannas. Therefore, this study aimed to update the information about volume and carbon stocks in savanna fragments in order to reduce tree measurements in the field. We used data from 61 savanna remnants, where 25 of them had 2 measurements with 5 years interval (monitoring). Multiple linear models, based on field data and reflectance values of Landsat images, were fitted and validated. Subsequently, the best models were applied to remnants that had only one measured, and then volume and carbon estimates were obtained for all 61 remnants in the second year of measurement. Additionally, maps updating the productivity of these remnants were generated.

ABSTRACT Volumetric equations is one of the main tools for quantifying forest stand production, and is the basis for sustainable management of forest plantations. This study aimed to assess the quality of the volumetric estimation of Eucalyptus spp. trees using a mixed-effects model, artificial neural network (ANN) and support-vector machine (SVM). The database was derived from a forest stand located in the municipalities of Bom Jardim de Minas, Lima Duarte and Arantina in Minas Gerais state, Brazil. The volume of 818 trees was accurately estimated using Smalian’s Formula. The Schumacher and Hall model was fitted by fixed-effects regression and by including multilevel random effects. The mixed model was fitted by adopting 14 different structures for the variance and covariance matrix. The best structure was selected based on the Akaike Information Criterion, Maximum Likelihood Ratio Test and Vuong’s Closeness Test. The SVM and ANN training process considered diameter at breast height and total tree height to be the independent variables. The techniques performed satisfactorily in modeling, with homogeneous distributions and low dispersion of residuals. The quality analysis criteria indicated the superior performance of the mixed model with a Huynh-Feldt structure of the variance and covariance matrix, which showed a decrease in mean relative error from 13.52% to 2.80%, whereas machine learning techniques had error values of 6.77% (SVM) and 5.81% (ANN). This study confirms that although fixed-effects models are widely used in the Brazilian forest sector, there are more effective methods for modeling dendrometric variables.

ABSTRACT According to previous studies involving biometric variables modeling using remote sensing (RS), data did not consider the effects of anthropogenic disturbance as relevant factor. The main objective of this study was to model aboveground biomass (AGB) and total wood volume (TWV) of Brazilian Savanna biome using vegetation indices (VI) from LANDSAT 5 TM. Multiple linear regression (MLR) and random forest (RF) algorithm were applied across 641 field plots of cerrado sensu stricto of the state of Minas Gerais, Brazil, comparing two models: non-stratified, and stratified according to plot’s anthropization degree. AGB and TWV obtained from non-anthropized plots presented linear relation with VIs (R2 = 0.82 and 0.74, respectively) and, on the other hand, presented nonlinear relation when plots were affected by anthropogenic disturbances or were not stratified. This finding helps improving estimates by stratifying plots into their anthropization degree, mainly in the Brazilian Savanna biome undergoing anthropogenic disturbances.

ABSTRACT Eremanthus erythropappus is a commercially-important tree which has a long history of exploitation in the Brazilian State of Minas Gerais. The knowledge on the potential geographical distribution of E. erythropappus is therefore critical for the species sustainability. Thus, the aim of this study was to estimate and map current and future ecological niche for E. erythropappus in Minas Gerais. We used the Random Forests algorithm to model the ecological niche for current and future climates scenarios. Our predictions indicate Espinhaço, Mantiqueira, and Canastra mountain ranges as core areas of distribution and forecast drastic reductions in potential areas under all climate scenarios. Based on our results, we highlight that the continual harvesting of naturally-occurring E. erythropappus populations will not be sufficient to supply the market demand. Silviculture practices would likely serve as an economically viable and ecological sustainable alternative to harvesting natural populations.

ABSTRACT This research aims to identify the vulnerabilities of Rio Grande basin in Minas Gerais state as well as to investigate, through technical, legal, social, and governmental aspects, the relevance of SEA as an environmental tool for the conservation and preservation of watersheds. Rio Grande basin evaluation was performed by using this following database: soil classes, land use classification, water quality index, and economic ecological index. It was performed multi-criteria decision analysis based on analytic hierarchy process methodology. 52.5% of the studied area was classified as “good” preservation condition. Alto Grande (GD1), Mortes/Jacaré (GD2) and Furnas (GD3) units did not contain significant percentages of their areas corresponding to the preservation class “very good”. In the studied area, 34.51% was classified as “medium” preservation condition, thus indicating the need for effective measures for the river basin conservation. These findings highlight the potential importance of including the SEA in the decision-making process for plans, policies and programs related to the integrated management of water resources.

ABSTRACT The present study aims (a) to assess the correlations between forest stand characteristics (viz., basal area, dominant height, and volume) and the reflectance values derived from Landsat 5 TM spectral bands as well as from vegetation indices and (b) to understand how Eucalyptus sp. stand age influences these correlations. Sampling data comprised a set of 35 permanent plots from a forest inventory conducted annually between 2006 and 2011. Spectral data were derived from Landsat 5 TM images. The results showed that TM4 and TM5 spectral bands, as well as vegetation indices ND54 and TM5/4, were better correlated with basal area and volume, while the TM2 spectral band was better correlated with dominant height. Eucalyptus sp. stand age directly influenced the correlations between spectral data and forest stand characteristics, and could not be disregarded in the spectral characterization of the forest variables.

ABSTRACT Brazil has many rural properties with unmanaged eucalyptus stands. These plantations are heterogeneous, presenting different tree sizes, advanced ages, and large wood volumes that can be quantified using forest inventories. The prediction error of dendrometric variables, mainly in highly heterogeneous areas, can be associated with inadequate forest inventory procedures, i.e. low intensity of sampling plots. However, a larger number of plots increases the cost of inventorying. Therefore, a promising alternative is forest stratification into homogeneous sub areas. Accordingly, the aim of this study was to analyze the reduction of volume estimate errors by post-stratification procedures. We used the normalized difference vegetation index (NDVI) derived from Landsat 8 and Spot 6 images and geostatistical techniques, such as kriging the volume (V) and diameter at breast height (DBH). The most precise method to estimate the total volume was the stratified random sampling (STS), based on geostatistical interpolation, using the DBH (error lower than 10%).

ABSTRACT The Brazilian Forestry Code established the Permanent Preservation Areas (PPAs) to preserve environmentally significant areas, such as the banks of waterways. Grande River is an important Brazilian river whose watercourse contains several hydroelectric plants, with few PPAs containing original features. Thus, this study analyzed land use in PPAs of a non-dammed stretch of the upper Rio Grande, in southern Minas Gerais. For this analysis, we used an image of the Rapideye sensor and the Maximum Likelihood classification method. The results showed the occurrence of pastures (49.63%), exposed soil (9.13%), others (0.77%), water (0.15%) and ornamental vegetation (0.13%) while the remaining native vegetation represented only 40.19% of PPAs. These numbers show that environmental laws have not been fulfilled in this area and there is strong human intervention in the PPAs studied.

RESUMO Assumindo a existência de uma relação entre a heterogeneidade da paisagem e medidas de dependência espacial obtidas de dados de sensoriamento remoto, o objetivo deste estudo foi avaliar o potencial dos parâmetros do semivariograma derivados de imagens de satélite com diferentes resoluções espaciais, para caracterizar áreas cobertas por floresta e áreas sob ação antrópica. Para isso, o NDVI (Índice de Vegetação da Diferença Normalizada) de cada umas das imagens (SPOT 6, Landsat 8 e MODIS Terra) foi gerado em uma área de floresta tropical Amazônica (1.000 km²), onde foram selecionadas amostras (1 x 1 km) de áreas florestadas e áreas antrópicas. A partir destes dados, foram gerados os semivariogramas e extraídos os parâmetros patamar (σ²-variabilidade espacial total) e alcance (φ-distância dentro da qual as amostras apresentam-se estruturadas espacialmente). A análise revelou que a resolução espacial das imagens influencia os parâmetros σ² e φ, apresentando significativo aumento das áreas de florestas para as áreas sob ação antrópica. A maior variação entre estas classes foi obtida com as imagens Landsat 8, indicando estas imagens, com resolução espacial de 30 metros, a mais apropriada para a obtenção dos parâmetros do semivariograma objetivando a caracterização da heterogeneidade espacial da paisagem. Combinando o sensoriamento remoto e técnicas geostatisticas, demonstrou-se que os parâmetros do semivariograma derivados de imagens NDVI podem ser utilizados como um simples indicador de heterogeneidade da paisagem, gerando mapas que permitem aos pesquisadores delinearem com maior eficácia o regime de amostragem. Outras aplicações combinando estas duas técnicas devem ser investigadas, como por exemplo a detecção de mudanças na cobertura do solo e a classificação de imagens utilizando análises orientada a objetos (OBIA).

ABSTRACT Assuming a relationship between landscape heterogeneity and measures of spatial dependence by using remotely sensed data, the aim of this work was to evaluate the potential of semivariogram parameters, derived from satellite images with different spatial resolutions, to characterize landscape spatial heterogeneity of forested and human modified areas. The NDVI (Normalized Difference Vegetation Index) was generated in an area of Brazilian amazon tropical forest (1,000 km²). We selected samples (1 x 1 km) from forested and human modified areas distributed throughout the study area, to generate the semivariogram and extract the sill (σ²-overall spatial variability of the surface property) and range (φ-the length scale of the spatial structures of objects) parameters. The analysis revealed that image spatial resolution influenced the sill and range parameters. The average sill and range values increase from forested to human modified areas and the greatest between-class variation was found for LANDSAT 8 imagery, indicating that this image spatial resolution is the most appropriate for deriving sill and range parameters with the intention of describing landscape spatial heterogeneity. By combining remote sensing and geostatistical techniques, we have shown that the sill and range parameters of semivariograms derived from NDVI images are a simple indicator of landscape heterogeneity and can be used to provide landscape heterogeneity maps to enable researchers to design appropriate sampling regimes. In the future, more applications combining remote sensing and geostatistical features should be further investigated and developed, such as change detection and image classification using object-based image analysis (OBIA) approaches.

RESUMO Recentemente, variáveis geoestatísticas derivadas de imagens de sensoriamento remoto ganharam espaço dentre os procedimentos de detecção de mudanças, porém, o potencial temporal destas variáveis para o mapeamento das mudanças baseado na análise por objetos ainda é pouco estudado. Neste estudo, o desempenho de um conjunto de índices calculados de semivariogramas derivados de imagens NDVI bitemporais para detectar mudanças na cobertura do solo foi analisado e avaliado. O município de Mariana foi selecionado para teste e validação da metodologia devido ao grande impacto ocasionado pelo desastre. O processo iniciou-se com a aquisição de imagens Landsat 8 antes e após o desastre e o cálculo do NDVI. Os objetos foram criados através da segmentação em multiresolução baseada na imagem pós-desastre. Os semivariogramas experimentais foram gerados dentro de cada objeto e os índices foram extraídos e selecionados através da análise de componentes principais. Os índices selecionados foram utilizados como dados de entrada para o algoritmo support vector machines para a classificação de áreas de mudança e não mudança. Os índices selecionados se mostraram efetivos para a detecção de mudanças, indicando a possibilidade de utilização para a detecção de mudanças baseada em objetos, resultando em um mapa precisos das áreas inundadas afetadas pelo desastre. Esta abordagem pode ser usada em muitos outros contextos para uma avaliação rápida e precisa de tais mudanças na cobertura do solo.

ABSTRACT Object-based change detection is a powerful analysis tool for remote sensing data, but few studies consider the potential of temporal semivariogram indices for mapping land-cover changes using object-based approaches. In this study, we explored and evaluated the performance of semivariogram indices calculated from remote sensing imagery, using the Normalized Differential Vegetation Index (NDVI) to detect changes in spatial features related to land cover caused by a disastrous 2015 dam failure in Brazil’s Mariana district. We calculated the NDVI from Landsat 8 images acquired before and after the disaster, then created objects by multiresolution segmentation analysis based on post-disaster images. Experimental semivariograms were computed within the image objects and semivariogram indices were calculated and selected by principal component analysis. We used the selected indices as input data to a support vector machine algorithm for classifying change and no-change classes. The selected semivariogram indices showed their effectiveness as input data for object-based change detection analysis, producing highly accurate maps of areas affected by post-dam-failure flooding in the region. This approach can be used in many other contexts for rapid and accurate assessment of such land-cover changes.

RESUMO No presente artigo realizou-se uma avaliação da ocupação da sub-bacia Córrego dos Bois, no município de Oliveira (MG) - com base nos resultados foram obtidas as classes de uso e ocupação do solo, delimitadas as áreas de preservação permanente, identificadas a ocorrência de conflito entre o uso do solo e a legislação e elaborados indicadores ambientais. Foi construído um mosaico a partir de três imagens provenientes dos satélites RapidEye. O método de classificação adotado foi a classificação orientada a objetos e para isso utilizou-se o software Ecognition Developer 8.0®. O algoritmo utilizado no processo de classificação foi o vizinho mais próximo, que leva em consideração a proximidade dos objetos segmentados na imagem em um espaço de características. A avaliação da exatidão da classificação foi realizada com base na coleta de amostras de acurácia obtidas aleatoriamente em áreas previamente conhecidas na imagem. Foram obtidas as classes água, café, eucalipto, floresta, infraestrutura, mineração, pastagem, rodovia, solo exposto e várzea, e encontradas somente duas categorias de áreas de preservação permanente (APP), ao longo dos cursos d’água e no entorno das nascentes. Conclui-se que a área total da sub-bacia Córrego dos Bois é de 1.540,18 ha, dos quais 40,70% são ocupados por pastagem, 35,84%, por cafezal e 18,84%, por vegetação nativa. O conflito do uso e da ocupação do solo nas áreas destinadas à preservação corresponde a 25,58% do território das APPs.

ABSTRACT In the present work, we evaluated the occupation of the Córrego dos Bois sub-basin, in the municipality of Oliveira County, Minas Gerais State, in Brazil - based on it, it has been obtained land cover classes, defined areas of permanent preservation, identified the conflict among land use and environmental laws, and calculated environmental indicators. Initially, it was constructed a mosaic of three RapidEye images. The classification method used was the object-based classification with the Ecognition Developer 8.0® software. The classifier algorithm applied was the nearest neighbor, which takes in account the proximity of segmented objects in an image in the feature space. A random sampling, in some previously defined areas, performed the evaluation of the classification accuracy. The land cover classes obtained were water bodies, coffee farming, Eucalyptus forests, native forests, buildings, mining, pasture, road, bare land and floodplain. It was identified only two categories of preservation permanent areas in the study area: adjacent to rivers and springs. It concludes that the total area of the Córrego dos Bois sub-basin is 1,540.18 hectares, where 40.70% are pastures, 35.84% are coffee plantations, and 18.84% are native forests. The corresponding conflict of the land use in the preservation areas is 25.58%.

Resumo: O objetivo deste trabalho foi avaliar a utilização de interpoladores geoestatísticos e imagens de sensoriamento remoto, para a estratificação de povoamentos de Eucalyptus sp. em idade de corte. As estimativas da amostragem casual estratificada foram comparadas às estimativas da amostragem sistemática quanto à variável volume de madeira. A área de estudo correspondeu a um povoamento clonal de Eucalyptus sp. no Município de Lagoa Grande, MG. O inventário florestal foi realizado em 2011, em quatro situações de amostragem e diferentes números de parcelas. A estratificação da área foi feita por meio de krigagem da área basal (variável principal), da reflectância na banda TM5 (variável auxiliar) obtida em uma imagem Landsat 5 TM e com base na cokrigagem dessas variáveis. A área basal e a reflectância na banda TM5 apresentaram estrutura de dependência espacial. A estratificação reduziu o erro de amostragem em até 40%, em comparação ao erro da amostragem sistemática. As imagens de sensoriamento remoto e os interpoladores geoestatísticos krigagem e cokrigagem são eficientes na definição dos estratos em povoamentos de Eucalyptus sp., em decorrência da maior precisão dos estimadores da amostragem casual estratificada, que permitem estimativas volumétricas mais precisas do inventário florestal do que os da amostragem sistemática.

Abstract: The objective of this work was to evaluate the use of geostatistical interpolators and remote sensing images for the stratification of Eucalyptus sp. stands, at harvesting age. Estimates by stratified random sampling and systematic sampling were compared for the timber volume variable. The study area was a clonal plantation of Eucalyptus sp. in the municipality of Lagoa Grande, in the state of Minas Gerais, Brazil. The forest inventory was performed in 2011, considering four sampling situations and different plot numbers. The area stratification was carried out based on kriging of the basal area (target variable), on the reflectance values in the TM5 band (auxiliary variable) extracted from a Landsat 5 TM image, and on the cokriging of these variables. Basal area and reflectance in the TM5 band showed a spatial dependence structure. Stratification reduced the sampling error up to 40%, in comparison with that generated from systematic sampling. Remote sensing images and kriging and cokriging interpolators are efficient to define the strata in Eucalyptus sp. stands, as a consequence of the higher precision of the stratified random sampling estimators, which allow more precise volume estimates of the forest inventory than those of systematic sampling.

Objetivou-se, neste trabalho, avaliar o estado de conservação atual da paisagem na região de Ouro Preto, Ouro Branco e Mariana, MG, correspondendo à áreas situadas no alto das bacias dos rios das Velhas, do Carmo, Gualaxo do Sul e ribeirão da Colônia. A região possui diferentes Unidades de Conservação (UCs), localizadas no domínio fitogeográfico da Mata Atlântica. Utilizaram-se imagens RapidEye AG (jun/2010) com 5 m de resolução espacial. A classificação do uso do solo para obtenção das métricas da paisagem foi dividida em três classes: Floresta Estacional Semidecidual Montana, Vegetação natural 2 e Uso Antrópico. Os resultados apontaram que a área de estudo encontra-se bem preservada, principalmente dentro das UCs, onde se localizam os grandes fragmentos, principais para a conservação; as formações naturais presentes na área de estudos corresponderam a 87,08%; as métricas mostraram que a distância média entre os fragmentos (262) é 72,18 m; a forma e o efeito de bordas nos fragmentos apresentaram-se adversas à preservação da biodiversidade; fortes pressões antrópicas: área urbana, solo exposto, minerações etc., estão presentes em maior intensidade fora da Zona de Amortecimento (ZA) do Parque Estadual do Itacolomi; pontos vulneráveis à conservação encontram-se, em grande parte, próximos às áreas de pastoreio e de mineração, com destaque para as áreas de campo das UCs de Ouro Branco e de pastoreio dentro da APA Estadual Cachoeira das Andorinhas.

This study aimed to evaluate the current state of conservation and the landscape connectivity in the area covering part of the counties of Ouro Preto, Mariana and Ouro Branco, MG. This area is located at the upper course of the Velhas river, Carmo river, Gualaxo do Sul river and Colônia creek. The region belongs to the Mata Atlantica phytogeographic ecosystem and encompasses different Conservation Units (CUs). RapidEye AG (Jun/2010) with 5 m of spatial resolution images were used. The study area was divided into three classes: Seasonal semideciduous montane forest; Natural vegetation 2; Anthropic use. The results showed that the study area is well preserved, particularly inside the CUs where the fragments are bigger and are considered priority sites for conservation. The natural formations of the region corresponded to 87.08% of the study area. The metrics showed that the average distance between the 262 fragments is 72.18 m. The shape of the fragments do not favor the preservation of biodiversity, because they are complex and irregular (fractal). The application of edge effects initially increases the fragmentation and, subsequently reduces the size of individual vegetation spots, smoothing their complex shapes. Strong anthropic pressures are present outside the buffer zone of the State Park of Itacolomi. The weak points are those surrounding pastoral and mining areas, mainly field areas of CUs in Ouro Branco and pasture areas inside the Cachoeira das Andorinhas State APA.

O Índice da Vegetação por Diferença Normalizada (NDVI) é, frequentemente, usado para extrair informações de áreas vegetadas, uma vez que está diretamente relacionado com parâmetros da vegetação, como o percentual de cobertura do solo, a atividade fotossintética da planta e o índice de área foliar. Objetivou-se, neste trabalho, analisar o potencial dos semivariogramas gerados a partir do índice de vegetação NDVI, para a detecção de desmatamentos, analisando seu comportamento (forma) e métricas derivadas (alcance, contribuição e efeito pepita). Semivariogramas foram gerados a partir do índice de vegetação NDVI, derivados de imagens Landsat TM de maio de 2010, junho de 2010 e julho de 2011. A área de estudo é uma região de Cerrado, no norte do estado de Minas Gerais, Brasil, totalizando 1.596 ha. Foram gerados semivariogramas, após a análise exploratória dos dados. Modelos foram ajustados, validados e suas métricas analisadas. Observou-se que a forma dos semivariogramas, alcance e contribuição foram diferentes quando houve desmatamentos na área e similares quando a área não foi alterada. O modelo que melhor se ajustou aos dados foi o gaussiano, porém os três modelos testados mostraram comportamentos similares quanto a eficiência na detecção dos desmatamentos. Os resultados sugerem que estudos futuros explorem o grau de intervenção que os semivariogramas temporais consigam detectar, bem como a influência que a sazonalidade de algumas fisionomias pode ter nas análises.

The Normalized Difference Vegetation Index (NDVI) is often used to extract information from vegetated areas since it is directly related to vegetation parameters such as percent of ground cover, photosynthetic activity of the plant and leaf area index. The aim of this paper was to analyze the potencial of semivariograms generated from NDVI values to detect changes in vegetated areas, analyzing their behavior (shape) and derived metrics (range, sill and nugget). Semivariograms were generated from NDVI values derived from Landsat TM images of May 2010, June 2010 and July 2011. The study area is located in the northern state of Minas Gerais, Brazil, and is covered by Brazilian savannas vegetation, totalizing 1,596 ha. Semivariograms were generated after the exploratory data analysis. Models were fitted, validated and their metrics analyzed. The results showed a very clear trend where the shape of semivariograms, sill and range were different when deforestation occurred and were similar when the area had not been changed. The model that generated best fit was the Gaussian, however, the three models tested showed behavior that makes it possible to detect changes in vegetation. It suggests that further researches should explore the degree to which the semivariogram can be used to quantify this spatial variability as well as to analyze the influence of sazonality for changing detection in vegetated areas.