Search Results (253)

Reliable remote sensing-based Land Use and Land Cover (LULC) information is crucial for assessing Earth’s surface activities. Brazil’s agricultural dynamics, including year-round cropping, multiple cropping, and regional climate variability, make LULC monitoring a highly challenging task. The country has thirteen remote sensing-based LULC products specifically tailored for this purpose. However, the differences and the results of these products have not yet been synthesized to provide coherent guidance in assessing their spatio-temporal agricultural dynamics and identifying promising approaches and issues that affect LULC analysis. This review represents the first comprehensive assessment of the advantages, challenges, and limitations, highlighting the main issues when dealing with contrasting LULC maps. These challenges include incompatibility, a lack of updates, non-systematic classification ontologies, and insufficient data to monitor Brazilian LULC information. The consequences include impacts on intercropping estimation, diminished representation or misrepresentation of croplands; temporal discontinuity; an insufficient number of classes for subannual cropping evaluation; and reduced compatibility, comparability, and spectral separability. The study provides insights into the use of these products as primary input data for remote sensing-based applications. Moreover, it provides prospects for enhancing existing mapping efforts or developing new national-level initiatives to represent the spatio-temporal variation of Brazilian agriculture. Full article

Semantic image synthesis (SIS) in remote sensing aims to generate high-fidelity satellite imagery from land use/land cover (LULC) labels, supporting applications such as map updating, data augmentation, and environmental monitoring. However, the existing methods typically focus on pixel-level semantic-to-image translation, neglecting the spatial and semantic relationships among land cover objects, which hinders accurate scene structure modeling. To address this challenge, we propose GMDiT, an enhanced conditional diffusion model that extends the masked DiT architecture with graph-prior modeling. By jointly incorporating relational graph structures and semantic labels, GMDiT explicitly captures the object-level spatial and semantic dependencies, thereby improving the contextual coherence and structural fidelity of the synthesized images. Specifically, to effectively capture inter-object dependencies, we first encode the semantics of each node using CLIP and then employ a simple yet effective graph transformer to model the spatial interactions among nodes. Additionally, we design a scene similarity sampling strategy for the reverse diffusion process, improving contextual alignment while maintaining generative diversity. Experiments on the OpenEarthMap dataset show that GMDiT achieves superior performance in terms of FID and other metrics, demonstrating its effectiveness and robustness in the generation of structured remote sensing images. Full article

Fire refugia are unburned and low severity patches within wildfires that contribute heterogeneity that is important to retaining biodiversity and regenerating forest following fire. With increasingly intense and frequent wildfires in the Pacific Northwest, fire refugia are important for re-establishing populations sensitive to fire and maintaining resilience to future disturbances. Mapping fire refugia and delayed canopy loss is useful for understanding patterns in their distribution. The increasing abundance of satellite data and advanced analysis platforms offer the potential to map fire refugia in high detail. This study uses the Bayesian Updating of Land Cover (BULC-D) algorithm to map fire refugia and delayed canopy loss three years after fire. The algorithm compiles Normalized Burn Ratio data from Sentinel-2 and Landsat 8 and 9 and uses Bayes’ Theorem to map land cover changes. Four wildfires that occurred across Washington State in 2020 were mapped. Additionally, to consider the longevity of ‘durable’ fire refugia, the fire perimeters were analyzed to map delayed canopy loss in the years 2021–2023. The results showed that large losses in fire refugia can occur in the 1–3 years after fire due to delayed effects, but with some patches enduring. Full article

Large-scale high-resolution (HR) land cover mapping is essential in monitoring the Earth’s surface and addressing critical challenges facing humanity. While weakly supervised methods help to mitigate the scarcity of HR annotations across wide geographic areas, existing approaches struggle with feature extraction instability. To address this issue, this study proposes AECA-FBMamba, an efficient weakly supervised framework that enhances model perception by stabilizing feature transitions during encoding. Specifically, this work introduces the Adaptive Environment Channel Alignment (AECA) module at the input stage, processing independently grouped color channels to enhance robust channel-wise feature extraction. Additionally, we incorporate the Feature Bridging Mamba (FBMamba) module, which enables smooth receptive field reduction, effectively addressing feature alignment issues when integrating local contexts into global representations. The proposed AECA-FBMamba achieved a 65.27% mIoU on the Chesapeake Bay dataset and a 56.96% mIoU on the Poland dataset. Experiments conducted on these two large-scale datasets demonstrate the method’s effectiveness in automatically updating high-resolution (HR) land cover maps using low-resolution (LR) historical annotations. This framework advances weakly supervised learning in remote sensing and offers solutions for large-scale land cover mapping applications. Full article

The Xinjiang Uygur Autonomous Region, characterized by its complex and fragile ecosystems, has faced ongoing ecological degradation in recent years, challenging national ecological security and sustainable development. To promote the sustainable development of regional ecological and landscape conservation, this study investigates Fractional Vegetation Cover (FVC) dynamics in Xinjiang. Existing studies often lack recent data and exhibit limitations in the selection of driving factors. To mitigate the issues, this study utilized Google Earth Engine (GEE) and cloud-free MOD13A2.061 data to systematically generate comprehensive FVC products for Xinjiang from 2000 to 2024. Additionally, a comprehensive and quantitative analysis of up to 15 potential driving factors was conducted, providing an updated and more robust understanding of vegetation dynamics in the region. This study integrated advanced methodologies, including spatiotemporal statistical analysis, optimized spatial scaling, trend analysis, and Geographical Detector (GeoDetector). Notably, we propose a novel approach combining a Theil–Sen Median trend analysis with a Hurst index to predict future vegetation trends, which to some extent enhances the persuasiveness of the Hurst index alone. The following are the key experimental results: (1) Over the 25-year study period, Xinjiang’s vegetation cover exhibited a pronounced north–south gradient, with significantly higher FVC in the northern regions compared to the southern regions. (2) A time series analysis revealed an overall fluctuating upward trend in the FVC, accompanied by increasing volatility and decreasing stability over time. (3) Identification of 15 km as the optimal spatial scale for FVC analysis through spatial statistical analysis using Moran’s I and the coefficient of variation. (4) Land use type, vegetation type, and soil type emerged as critical factors, with each contributing over 20% to the explanatory power of FVC variations. (5) To elucidate spatial heterogeneity mechanisms, this study conducted ecological subzone-based analyses of vegetation dynamics and drivers. Full article

Fires drive global ecosystem change, impacting carbon dynamics, atmospheric composition, biodiversity, and human well-being. Biomass burning, a major outcome of fires, significantly contributes to greenhouse gas and aerosol emissions. Among these, fine particulate matter (PM_2.5) is particularly concerning due to its adverse effects on air quality and health, and its substantial yet uncertain role in Earth’s energy balance. Variability in emission factors (EFs) remains a key source of uncertainty in emission estimates. This study evaluates PM_2.5 emission sensitivity to EFs variability in Brazil’s Amazon and Cerrado biomes over 2002–2023 using the 3BEM_FRP model implemented in the PREP-CHEM-SRC tool. We updated the EFs with values and uncertainty ranges from Andreae (2019), which reflect a more comprehensive literature review than earlier datasets. The results reveal that the annual average PM_2.5 emissions varied by up to 162% in the Amazon (1213 Gg yr⁻¹ to 3172 Gg yr⁻¹) and 184% in the Cerrado (601 Gg yr⁻¹ to 1709 Gg yr⁻¹). The Average peak emissions at the grid-cell level reached 5688 Mg yr⁻¹ in the “Arc of Deforestation” region under the High-end EF scenario. Notably, the PM_2.5 emissions from Amazon forest areas increased over time despite shrinking forest cover, indicating that Amazonian forests are becoming more vulnerable to fire. In the Cerrado, savannas are the primary land cover contributing to the total PM_2.5 emissions, accounting for 64% to 80%. These findings underscore the importance of accurate, region-specific EFs for improving emission models and reducing uncertainties. Full article

Multi-temporal synthetic aperture radar interferometry (MT-InSAR) has evolved from a niche research technique into a powerful global monitoring tool. With the launch of nationwide and continent-wide ground motion services (GMSs), freely available deformation data can now be analyzed on a large scale. However, their applicability for monitoring critical infrastructure, such as dams, has not yet been thoroughly assessed, and several challenges have hindered the integration of MT-InSAR into existing monitoring frameworks. These challenges include technical limitations, difficulties in interpreting deformation results, and the rigidity of existing safety protocols, which often restrict the adoption of remote sensing techniques for operational dam monitoring. This study evaluates the effectiveness of persistent scatterer (PS) data from the German ground motion service (Bodenbewegungsdienst Deutschland, BBD) in complementing time-consuming in situ techniques. By analyzing a gravity dam in Germany, BBD time series were compared with in situ pendulum data. We propose a two-stage assessment procedure: First, we evaluate the dam’s suitability for PS analysis using the CR-Index to identify areas with good radar visibility. Second, we assess the interpretability of BBD data for radial deformations by introducing a novel index that quantifies the radial sensitivity of individual PS points on the dam. This index is universally applicable and can be transferred to other types of infrastructure. The results revealed a fair correlation between PS deformations and pendulum data for many PS points (up to $R^2$ = 0.7). A priori feasibility assessments are essential, as factors such as topography, land cover, and dam type influence the applicability of the PS technique. The dam’s orientation relative to the look direction of the sensor emerged as a key criterion for interpreting radial deformations. For angle differences (${\Delta }_{\mathrm{RAD}}$) of up to 20° between the true north radial angle of a PS point and the satellite’s look direction, the line-of-sight (LOS) sensitivity accounts for approximately 50 to 70% of the true radial deformation, depending on the satellite’s incidence angle. This criterion is best fulfilled by dams aligned in a north–south direction. For the dam investigated in this study, the LOS sensitivity to radial deformations was low due to its east–west orientation, resulting in significantly higher errors (6 mm $\le RMSE\le 43$ mm) compared to in situ pendulum data. Eliminating PS points with an unfavorable alignment with the sensor should be considered before interpreting radial deformations. For implementation into operational monitoring programs, greater effort must be spent on near-real-time updates of BBD datasets. Full article

Land cover classification (LCC) using satellite images is one of the rapidly expanding fields in mapping, highlighting the need for updating existing computational classification methods. Advances in technology and the increasing variety of applications have introduced challenges, such as more complex classes and a demand for greater detail. In recent years, deep learning and Convolutional Neural Networks (CNNs) have significantly enhanced the segmentation of satellite images. Since the training of CNNs requires sophisticated and expensive hardware and significant time, using pre-trained networks has become widespread in the segmentation of satellite image. This study proposes a hybrid synergistic semantic segmentation method based on the Deeplab v3+ network and a clustering-based post-processing scheme. The proposed method accurately classifies various land cover (LC) types in multispectral satellite images, including Pastures, Other Built-Up Areas, Water Bodies, Urban Areas, Grasslands, Forest, Farmland, and Others. The post-processing scheme includes a spectral bag-of-words model and K-medoids clustering to refine the Deeplab v3+ outputs and correct possible errors. The simulation results indicate that combining the post-processing scheme with deep learning improves the Matthews correlation coefficient (MCC) by approximately 5.7% compared to the baseline method. Additionally, the proposed approach is robust to data imbalance cases and can dynamically update its codewords over different seasons. Finally, the proposed synergistic semantic segmentation method was compared with several state-of-the-art segmentation methods in satellite images of Italy’s Lake Garda (Lago di Garda) region. The results showed that the proposed method outperformed the best existing techniques by at least 6% in terms of MCC. Full article

Accurate identification of dust emission sources is crucial for simulating dust aerosols in atmospheric chemical models. Therefore, a dynamically updated dust source function (DSF) was developed within the dust emission scheme of the Weather Research and Forecasting model coupled with chemistry (WRF-Chem) to simulate an East Asian dust storm event from 13 to 16 March 2021. Utilizing satellite-derived input of vegetation cover, snow cover, soil texture, and land use, the DSF was updated to better identify dust source areas over bare soils and sparsely vegetated regions in western China and central-western Mongolia. With the updated DSF, simulated dust emissions increase significantly over western China and Mongolia. The dust aerosol simulations demonstrate substantial improvements in near-surface PM₁₀ concentrations, a better agreement with remotely sensed dust aerosol optical depth (DOD), and a more accurate representation of the vertical distribution of dust extinction coefficients compared to observations. This study highlights the importance of integrating real-time data to accurately characterize dust emission sources, thereby improving atmospheric environment simulations. Full article

Isoprene emissions can affect the oxidizing capacity of the atmosphere and are likely to increase with an increase in the world’s biomass. The emission of isoprene is strongest in tropical forested regions, suggesting a major portion of tropospheric chemistry occurs in the tropics. As well as deforestation and reforestation having a direct impact on the world’s climate through land cover, there is also an indirect environmental impact (e.g., global warming, air pollution) through the resulting change in isoprene emissions. Previously, incomplete understanding of isoprene oxidation chemistry caused a model-measurement breakdown for concentrations of HO_x radicals observed over certain low-NO_x regions, such as the pristine Amazon rainforest. Over the last decade, however, understanding of isoprene oxidation chemistry has been vastly improved. Numerous research studies have provided evidence for the involvement of 1,6-H and 1,5-H shift reactions in the isoprene oxidation mechanism, which increases the level of HO_x recycling that occurs. As well as helping to reduce the model-measurement breakdown observed, the updated isoprene oxidation mechanism affects the tropospheric burdens of other species, including carbon monoxide (CO), methane (CH₄), ozone (O₃), organic peroxides (ROOH), secondary organic aerosol (SOA), and organic nitrates (RONO₂). There are still gaps in the understanding of the impacts and oxidation chemistry of isoprene emissions, which this literature review identifies and discusses. In the future, there is still much scope for further research, including modeling future reforestation scenarios with isoprene emissions and their impacts on both global and regional scales. Full article

Kazakhstan’s pasture, as a spatially extended agricultural resource for sustainable animal husbandry, requires effective monitoring with connected rational uses. Ranking number nine globally in terms of land size, Kazakhstan, with an area of about three million square km, requires proper assessment technologies for climate change and anthropogenic impact to track the pasture lands’ degradation. Remote sensing (RS)-based adaptive approaches for assessing pasture load, combined with field cross-checking of pastures, have been applied to evaluate the quality of vegetation cover, economic potential, service function, regenerative capacity, pasture productivity, and changes in plant species composition for five pilot regions in Kazakhstan. The current stages of these efforts are presented in this project report. The pasture lands in five regions, including Pavlodar (8,340,064 ha), North Kazakhstan (2,871,248 ha), Akmola (5,783,503 ha), Kostanay (11,762,318 ha), Karaganda (19,709,128 ha), and Ulytau (18,260,865 ha), were evaluated. Combined RS data were processed and the Normalized Difference Vegetation Index (NDVI), Leaf Area Index (LAI), Fraction of Vegetation Cover (FCover), Fraction of Absorbed Photosynthetically Active Radiation (FAPAR), Canopy Chlorophyll Content (CCC), and Canopy Water Content (CWC) indices were determined, in relation to the herbage of pastures and their growth and development, for field biophysical analysis. The highest values of LAI, FCOVER, and FARAR were recorded in the Akmola region, with index values of 18.5, 126.42, and 53.9, and the North Kazakhstan region, with index values of 17.89, 143.45, and 57.91, respectively. The massive 2024 spring floods, which occurred in the Akmola, North Kazakhstan, Kostanay, and Karaganda regions, caused many problems, particularly to civil constructions and buildings; however, these same floods had a very positive impact on pasture areas as they increased soil moisture. Further detailed investigations are ongoing to update the flood zones, wetlands, and swamp areas. The mapping of proper flood zones is required in Kazakhstan for pasture activities, rather than civil building construction. The related sustainable permissible grazing husbandry pasture loads are required to develop also. Recommendations for these preparation efforts are in the works. Full article

Land use and land cover are the main anthropogenic factors that lead to the rapid and aggressive degradation of land and interfere with the functioning of ecosystems, especially through the expansion of urbanization and the reduction in forested areas. The purpose of this article is to identify sources of official data and to build an updated dataset upon which analysis algorithms can be applied. Basic analyses were performed on the dataset to study the evolution of land use and land cover in Romania during the period 1990–2022 and to analyze whether they align with and support national and European strategies for economic development. The study is based on the historical evolution of land use and land cover over the past 30 years, which has allowed us to identify current trends and to analyze potential future positive and negative effects, and facilitates the adoption of documented corrective measures. The methodology applied is based on the collection and analysis of historical data obtained mainly from the TEMPO database managed by the National Institute of Statistics and from the annual reports of the National Environmental Protection Agency, analyzing these data, identifying current trends, and issuing specific recommendations. Monitoring the dynamics of land use during the mentioned period allowed us to identify the most important developments with positive effects (a slight increase in forested areas and minimal variations in arable land) but mostly negative effects (a drastic reduction in areas occupied by vineyards and orchards, a decrease in areas occupied by water, and a concerning increase in unproductive land). The results highlight the importance of the immediate implementation of the national general cadastre system, as well as the importance of implementing appropriate policies, legislative provisions, and land management practices aimed at environmental protection and supporting sustainable development and biodiversity in general. Full article

The dynamics of Land Use/Land Cover changes are crucial to environmental sustainability, socio-economic development, and spatial planning. These changes stem from complex interactions between human activities, natural processes, and policies. In recent decades, LULC transformations have been linked to global challenges such as biodiversity loss, climate change, and resource degradation. Key drivers include urban sprawl, agricultural expansion and abandonment, and deforestation, emphasizing the need for effective frameworks to monitor and assess their impacts. This study investigates Land Use/Land Cover (LULC) changes in Tuscany (Italy) over the period from 2007 to 2019. To achieve this, statistical analyses were conducted to quantify variations in LULC across different classes and administrative territories represented by provincial local authorities. Specifically, data spanning five temporal intervals (2007, 2010, 2013, 2016, and 2019) enabled a comprehensive comparative analysis of spatial persistence in LULC patterns. Changes were assessed using a statistical approach based on Odds Ratios (OR). Additionally, Generalized Linear Models (GLMs) at the provincial level were employed to facilitate one-to-many provincial comparisons and to evaluate the statistical significance of observed LULC changes. The analysis revealed that certain classes exhibit a greater susceptibility to changes compared to others. Specifically, the classes categorized under ’Artificial Surfaces’ (LC_100) were, on average, 6.7 times more likely to undergo changes than those classified as ’Agricultural Areas’ (LC_200) and 11 times more likely than those under ’Forest and Semi-natural Areas’ (LC_300). Over time, the areas classified as artificial territories have exhibited a progressively decreasing probability of change. Notably, during the first update period (2007–2010), these areas were 3.5 times more susceptible to change compared to the most recent update period (2016–2019). An additional significant finding emerged from the statistical comparison of LULC changes across administrative regions governed by different authorities (Provinces). These findings underscore the potential of using administrative indicators and morphological parameters to analyze LULC change trends. The proposed approach provides a robust framework for interpreting territorial resilience and informing spatial planning strategies effectively. Full article

National Forest Inventories (NFIs) provide valuable land cover (LC) information but often lack spatial continuity and an adequate update frequency. Satellite-based remote sensing offers a viable alternative, employing machine learning to extract thematic data. State-of-the-art methods such as convolutional neural networks rely on fully pixel-level annotated images, which are difficult to obtain. Although reference LC datasets have been widely used to derive annotations, NFIs consist of point-based data, providing only sparse annotations. Weakly supervised and self-supervised learning approaches help address this issue by reducing dependence on fully annotated images and leveraging unlabeled data. However, their potential for large-scale LC mapping needs further investigation. This study explored the use of NFI data with deep learning and weakly supervised and self-supervised methods. Using Sentinel-2 images and the Portuguese NFI, which covers other LC types beyond forest, as sparse labels, we performed weakly supervised semantic segmentation with a convolutional neural network to create an updated and spatially continuous national LC map. Additionally, we investigated the potential of self-supervised learning by pretraining a masked autoencoder on 65,000 Sentinel-2 image chips and then fine-tuning the model with NFI-derived sparse labels. The weakly supervised baseline achieved a validation accuracy of 69.60%, surpassing Random Forest (67.90%). The self-supervised model achieved 71.29%, performing on par with the baseline using half the training data. The results demonstrated that integrating both learning approaches enabled successful countrywide LC mapping with limited training data. Full article

Erosion is a significant environmental challenge in Serbia, shaped by natural and human factors. Pronounced relief, fragile geological substrate, a developed hydrographic network, and a climate characterized by an uneven distribution of precipitation throughout the year make this area prone to activating erosion processes and flash floods whenever there is a significant disruption in ecological balance, whether due to the removal of vegetation cover or inadequate land use. Researchers have recorded approximately 11,500 torrents in Serbia, most of which were activated during the 19th century, a period of significant social and political change, as well as intensive deforestation and the irrational exploitation of natural resources. By the mid-19th century, the effects of land degradation were impossible to ignore. As the adequate assessment of soil erosion intensity is the initial step in developing a prevention and protection strategy and the type and scope of anti-erosion works and measures, this article presents the path that the anti-erosion field in Serbia has taken from the initial observations of erosion processes through the first attempts to create the Barren Land Cadastre and Torrent Cadastre to the creation of the Erosion Potential Method (EPM) and its modification by Dr. Lazarević that resulted in the creation of the first Erosion Map of SR Serbia in 1971 (published in 1983). In 2020, a new Erosion Map of Serbia was created with the application of Geographic Information System (GIS) technologies and based on the original method by Professor Slobodan Gavrilović—the EPM—without the modifications introduced by Lazarević. We compared the 1983 and 2020 erosion maps in a GIS environment, where the change in soil erosion categories was analyzed using a confusion matrix. The updated erosion maps mirror the shift in methodology from a traditional approach (Lazarević’s modification) to the modern GIS-based method (Gavrilović’s original EPM) and reflect technological improvements and changes in land use, conservation practices, and environmental awareness. Full article