Cover Classification Research Articles

European Union crop map 2022: Earth observation’s 10-meter dive into Europe’s crop tapestry

To provide the information needed for a detailed monitoring of crop types across the European Union (EU), we present an advanced 10-metre resolution map for the EU and Ukraine with 19 crop types for 2022, updating the 2018 version. Using Earth Observation (EO) and in-situ data from Eurostat’s Land Use and Coverage Area Frame Survey (LUCAS) 2022, the methodology included 134,684 LUCAS Copernicus polygons, Sentinel-1 and Sentinel-2 satellite imagery, land surface temperature and a digital elevation model. Based on this data, two classification layers were developed using a Random Forest machine learning approach: a primary map and a gap-filling map to address cloud-covered gaps. The combined maps, covering 27 EU countries, show an overall accuracy of 79.3% for seven major land cover classes and 70.6% for all 19 crop types. The trained model was used to derive the 2022 map for Ukraine, demonstrating its robustness even in regions without labelled samples for model training.

Comparative analysis of the performance of selected machine learning algorithms depending on the size of the training sample

Abstract The article presents an analysis of the effectiveness of selected machine learning methods: Random Forest (RF), Extreme Gradient Boosting (XGB), and Support Vector Machine (SVM) in the classification of land use and cover in satellite images. Several variants of each algorithm were tested, adopting different parameters typical for each of them. Each variant was classified multiple (20) times, using training samples of different sizes: from 100 pixels to 200,000 pixels. The tests were conducted independently on 3 Sentinel-2 satellite images, identifying 5 basic land cover classes: built-up areas, soil, forest, water, and low vegetation. Typical metrics were used for the accuracy assessment: Cohen’s kappa coefficient, overall accuracy (for whole images), as well as F-1 score, precision, and recall (for individual classes). The results obtained for different images were consistent and clearly indicated an increase in classification accuracy with the increase in the size of the training sample. They also showed that among the tested algorithms, the XGB algorithm is the most sensitive to the size of the training sample, while the least sensitive is SVM, which achieved relatively good results even when using training samples of the smallest sizes. At the same time, it was pointed out that while in the case of RF and XGB algorithms the differences between the tested variants were slight, the effectiveness of SVM was very much dependent on the gamma parameter – with too high values of this parameter, the model showed a tendency to overfit, which did not allow for satisfactory results.

Assessing the impact of land use and land cover change on the Densu Delta wetland using Markov chain modeling and artificial neural networks

This study investigates the dynamics of land use and land cover (LULC) changes in the Densu Delta wetlands, a critical ecosystem in Ghana. Here, satellite images spanning from 1998 to 2023 were used to analyse the spatio-temporal patterns of LULC changes and their implications for water bodies, wetlands, vegetations, bare lands and urban areas in the Densu Delta wetland. Employing advanced techniques such as Markov chain modelling and artificial neural networks (ANNs), the research assesses and predicts LULC alterations. Significantly, the largest loss of LULC is observed in the Densu Delta wetland, where wetlands transition to waterbody cover type (14.02 km²). Model validation for 2023 attests to the accuracy of the model, boasting a correctness percentage of 70% and a kappa value of 0.74. In-depth analyses explore regional variations in the Densu Delta wetlands, revealing distinct patterns in the rates of LULC change before and after 2013. Notably, urbanization emerges as a prominent factor post-2013, with urban areas experiencing remarkable rates of change in the wetland. Transition matrices underscore the intricate interplay of different land cover classes over the years. Simulated LULC predictions for 2033 and 2043 highlight the urban land cover type as having the highest positive change, recording approximately 0.39% for the Densu Delta wetland. The wetland land cover in the Densu Delta wetland exhibit negative changes of about −0.52%. The synthesis of LULC data enhances our understanding of the complex interactions shaping these critical ecosystems. This research offers valuable insights for sustainable environmental conservation, emphasizing the pivotal role of informed urban planning policies. It also unveils potential challenges posed by climate change, advocating for a holistic approach to preserve these vital wetland ecosystems.

Comparative Study of the Coastal Geomorphology of Kajirbhati Sand-Spit Using Google Earth Pro Images (2005, 2011, 2023) and Landsat Satellite Data (2009, 2016, 2023): An Approach Towards the Impact of the Casuarina Plantation

Coastal geomorphology undergoes constant transformation due to a combination of coastal and fluvial processes and human intervention. Remote sensing and GIS tools are invaluable in identifying and studying these changes. Due to the dynamic interaction between land and sea, the coastline requires careful monitoring to detect hotspots and understand the spatial and temporal impacts of climate change on the coastal environment. Remote sensing data offers a comprehensive means of studying coastal landform changes. In a recent study, the focus was to analyze the impact of the Casuarina plantation on the coastal geomorphology of Kajirbhati spit. This involved studying various depositional features such as sand bars, sand spits, dunes, and beaches. The study utilized Google Earth Pro images from 2005, 2011, and 2023 to digitize 13 classes each year, enabling the identification of changing geomorphic classes. These include agriculture, swash zones, dunes on sand spits, berm lines, Casuarina and coconut plantations, estuary, mangroves, mudflats, open land, sea, settlement, and vegetation. The results revealed dramatic changes in land use and land cover classes from 2005 to 2023. In the years from 2011 to 2023, the area of sandspit covered by Casuarina plants expanded rapidly. To analyze the changes, Landsat 5, 8, and 9 data were used to check the intensity of NDVI (Normalized Difference Vegetation Index) in 2009, 2016, and 2023. NDVI is a critical factor in this research. The mean NDVI values for 2009, 2016, and 2023 were -0.002, 0.02, and 0.04 respectively. The changes in higher NDVI values from 2009 to 2023 were -0.10, 0.27, and 0.28, while the lower values were -0.10, -0.10, and -0.07 for the same years. The correlation calculation between the mean NDVI values and the years indicates a strong positive correlation with an R-value of 0.99. Casuarina plants play a crucial role in protecting the coast from strong winds and dynamic sea waves. They help protect mudflats, mangroves, settlements, and estuary land use and cover classes along the coast of Kajirbhati.

Disproportionate contributions of land cover and changes to ecosystem functions in Kazakhstan and Mongolia

Land use and land cover change (LULCC) have profoundly altered land surface properties and ecosystem functions, including carbon and water production. While mapping these changes from local to global scales has become more achievable due to advancements in earth observations and remote sensing, linking land cover changes to ecosystem functions remains challenging, especially at regional scale. Our study attempts to fill this gap by employing a computationally efficient method and two types of widely used high-resolution satellite images. We first investigated the contribution of landscape composition to ecosystem function by examining how land cover and proportion affected gross primary production (GPP) and evapotranspiration (ET) at six macro-landscapes in Mongolia and Kazakhstan. We hypothesized that both ecosystem and landscape GPP and ET are disproportionate to their composition and, therefore, changes in land cover will have asymmetrical influences on landscape functions. We leveraged a computational-friendly linear downscaling approach to align the coarse spatial resolution of MODIS (500 m) with a fine-grain and localized land cover map developed from Landsat (30 m) for six provinces in countries where intensive LULCC occurred in recent decades. By establishing two metrics—function to composition ratio (F/C) and function to changes in composition change (ΔF/ΔC)—we tested our hypothesis and evaluated the impact of land cover change on ecosystem functions within and among the landscapes. Our results show three major themes. (1) The five land cover types have signature downscaled ET and GPP that appears to vary between the two countries as well as within each country. (2) F/C of ET and GPP of forests is statistically greater than 1 (i.e., over-contributing), whereas F/C of grasslands and croplands is close to or slightly less than 1 (i.e., under-contribution). F/C of barrens is clearly lower than 1 but greater than zero. Specifically, a unit of forest generates 1.085 unit of ET and 1.123 unit of GPP, a unit of grassland generates 0.993 unit of ET and GPP, and a unit of cropland produces 0.987 unit of ET and 0.983 unit of GPP. The divergent F/C values among the land cover classes support the hypothesis that landscape function is disproportionate to its composition. (3) ΔET/ΔC and ΔGPP/ΔC of forests and croplands showed negative values, while grasslands and barrens showed positive values, indicating that converting a unit of forest to other land cover leads to a decrease in ET and GPP, while converting units of grassland or barren to other land cover classes will result in increased ET and GPP. This linear downscaling approach for calculating F/C and ΔF/ΔC is labor-saving and cost-effective for rapid assessment on the impact of land use land cover change on ecosystem functions.

Intercomparison of multisource actual evapotranspiration satellite products in Bilate watershed, Ethiopia

Recent advancements in satellite remote sensing have led to increased spatial and temporal resolution of actual evapotranspiration (AET) estimates across scales. Yet, the accuracy of AET products remains unknown for many regions, prompting further investigation to guide selection. This study intercompares five AET products within Ethiopia’s Bilate watershed, focusing on the 2009-2018 period. The products assessed include TerraClimate, Food and Agriculture Organization Water Productivity (FAO WaPOR), Moderate Resolution Imaging Spectroradiometer Operational Simplified Surface Energy Balance (ModisSSEBop), and Synthesis of Global AET. Reference evapotranspiration estimated using ground station climate data served as a basis for comparing the Satellite Products (SP). The intercomparison was conducted using descriptive statistics, scatter plots and Pearson’s Correlation Coefficient to assess correlation, standard deviation, and root mean square error. Additional error statistics were also considered. Findings reveal higher AET values in the highlands compared to the lowlands of the Bilate watershed. A weak correlation (<0.35) exists between ETo and satellite-derived AET, potentially due to the averaging of AET values across diverse land cover classes, contrasting with point-scale reference measurements. The variance among AET products was varied across seasons and elevation ranges. While the annual patterns of AET were consistent across the products, large discrepancies in magnitude (average AET varies from 25 to 83 mm per month in the lower part) were detected. The ModisSSEBop global and continental products showed minimal mismatches, whereas the Synthesis of Global and FAO WaPOR products displayed slight differences. Notably, the FAO WaPOR’s AET estimates showed relatively closer agreement with many products in terms of magnitude and variability of AET. In conclusion, the study highlights significant random and systematic differences between the AET products. The substantial mismatch between the products underscores the necessity for continued research to refine AET product accuracy through improved input dataset and revisiting the algorithms.

Spatial Distribution of Thaw Depth in Palsas Estimated From Optical Unoccupied Aerial Systems Data

ABSTRACTMaximum seasonal thaw depth, referred to as active layer thickness (ALT), is one of the key parameters used to monitor permafrost conditions. ALT maps based on interpolation of point measurements or derived from coarse or moderate spatial resolution satellite data often hide small‐scale spatial variations in thaw depth resulting from differences in surface characteristics and microtopography. To model and predict changes in hydrological and biogeochemical processes in permafrost areas accurately, high‐resolution remote sensing‐based estimations of ALT are needed. Therefore, we applied random forest (RF) regression on a set of topographical and spectral vegetation indices derived from optical unoccupied aerial systems data, Landsat 8 land surface temperature (LST) data, and field measurements to estimate thaw depths in palsas at three mires in north‐west Finland. We also analyzed differences in thaw depths between mires located at different elevations, between dome and plateau‐shaped palsas, and between different vegetation and surface cover classes. The RF models resulted in root mean square errors from 2.4 to 5.7 cm between predicted and observed thaw depths and the R2 values of 0.57–0.96. Height from the surrounding fen surface and LST were the most important variables in thaw depth models, although high‐accuracy results were also achieved without LST. The mean thaw depths did not differ between the sites with lowest and highest elevation, whereas the thaw depths were significantly deeper in dome‐shaped palsas compared to plateaus. The thaw depths were significantly different between vegetation cover classes only on plateau‐shaped palsas. The results indicate the high impact of the topography on the palsa thaw depth, thus highlighting the importance of accurate elevation models in spatial modeling of palsa ALT. The methodology presented in this study can be applied to other permafrost regions where field measurements of ALT are accompanied with high‐resolution topographical and multispectral data.

Bayesian spatio-temporal modeling to assess the effect of land-use changes on the incidence of Cutaneous Leishmaniasis in the Brazilian Amazon

Cutaneous Leishmaniasis (CL) is a vector-borne disease caused by a protozoan of the genus Leishmania and is considered one of the most important neglected tropical diseases. The Brazilian Amazon Forest harbors one of the highest diversity of Leishmania parasites and vectors and is one of the main focuses of the disease in the Americas. Previous studies showed that some types of anthropogenic disturbances have affected the abundance and distribution of CL vectors and hosts; however, few studies have thoroughly investigated the influence of different classes of land cover and land-use changes on the disease transmission risk. Here, we quantify the effect of land use and land-cover changes on the incidence of CL in all municipalities within the Brazilian Amazon Forest, from 2001 to 2017. We used a structured spatiotemporal Bayesian model to assess the effect of forest cover, agriculture, livestock, extractivism, and- deforestation on CL incidence, accounting for confounding variables such as population, climate, socioeconomic, and spatiotemporal random effects. We found that the increased risk of CL was associated with deforestation, especially modulated by a positive interaction between forest cover and livestock. Landscapes with ongoing deforestation for extensive cattle ranching are typically found in municipalities within the Amazon Frontier, where a high relative risk for CL was also identified. These findings provide valuable insights into developing effective public health policies and land-use planning to ensure healthier landscapes for people.

Temporal Dynamics of Global Barren Areas between 2001 and 2022 Derived from MODIS Land Cover Products

Long-term monitoring studies on the transition of different land cover units to barren areas are crucial to gain a better understanding of the potential challenges and threats that land surface ecosystems face. This study utilized the Moderate Resolution Imaging Spectroradiometer (MODIS) land cover products (MCD12C1) to conduct geospatial analysis based on the maximum extent (MaxE) concept, to assess the spatiotemporal changes in barren areas from 2001 to 2022, at global and continental scales. The MaxE area includes all the pixels across the entire period of observations where the barren land cover class was at least once present. The relative expansion or reduction of the barren areas can be directly assessed with MaxE, as any annual change observed in the barren distribution is comparable over the entire dataset. The global barren areas without any land change (UA) during this period were equivalent to 12.8% (18,875,284 km2) of the global land surface area. Interannual land cover changes to barren areas occurred in an additional area of 3,438,959 km2 (2.3% of the global area). Globally, barren areas show a gradual reduction from 2001 (91.1% of MaxE) to 2012 (86.8%), followed by annual fluctuations until 2022 (88.1%). These areas were mainly interchanging between open shrublands and grasslands. A relatively high transition between barren areas and permanent snow and ice is found in Europe and North America. The results show a 3.7% decrease in global barren areas from 2001 to 2022. Areas that are predominantly not barren account for 30.6% of the transitional areas (TAs), meaning that these areas experienced short-term or very recent transitions from other land cover classes to barren. Emerging barren areas hotspots were mainly found in the Mangystau region (Kazakhstan), Tibetan plateau, northern Greenland, and the Atlas Mountains (Morocco, Tunisia).

Support Vector Machine Algorithm for Mapping Land Cover Dynamics in Senegal, West Africa, Using Earth Observation Data

This paper addresses the problem of mapping land cover types in Senegal and recognition of vegetation systems in the Saloum River Delta on the satellite images. Multi-seasonal landscape dynamics were analyzed using Landsat 8-9 OLI/TIRS images from 2015 to 2023. Two image classification methods were compared, and their performance was evaluated in the GRASS GIS software (version 8.4.0, creator: GRASS Development Team, original location: Champaign, Illinois, USA, currently multinational project) by means of unsupervised classification using the k-means clustering algorithm and supervised classification using the Support Vector Machine (SVM) algorithm. The land cover types were identified using machine learning (ML)-based analysis of the spectral reflectance of the multispectral images. The results based on the processed multispectral images indicated a decrease in savannas, an increase in croplands and agricultural lands, a decline in forests, and changes to coastal wetlands, including mangroves with high biodiversity. The practical aim is to describe a novel method of creating land cover maps using RS data for each class and to improve accuracy. We accomplish this by calculating the areas occupied by 10 land cover classes within the target area for six consecutive years. Our results indicate that, in comparing the performance of the algorithms, the SVM classification approach increased the accuracy, with 98% of pixels being stable, which shows qualitative improvements in image classification. This paper contributes to the natural resource management and environmental monitoring of Senegal, West Africa, through advanced cartographic methods applied to remote sensing of Earth observation data.

Land Use and Soil Carbon Sequestration in the Kou Watershed, Burkina Faso

Background: Soil organic carbon is an indicator that must be considered when assessing the sustainability of agroecosystems. Aim: The aim of the study is to assess the impact of different land use patterns on soil organic carbon stock and the contribution of this component to climate change mitigation. Methodology: soil samples were taken from depths of 0-15 cm and 15-30 cm along diagonal transects of study plots, with seventy-two composite samples collected in total. Results: Soils in the riparian zone (23.77%) and irrigated crops (25.73%) were found to be the richest in clay. Only clay content was positively correlated with soil depth (r=0.104). The Kruskal-Wallis test shows that the amount of carbon varies significantly (p-values between 2.510-4 and 7.910-4) across land use/land cover classes (LULC). Wetland (28.91 ± 2.83 t/ha) and woody savannah (28.53 ± 4.24 t/ha) had the highest carbon stock, and shrub savannah had the lowest (12.77 ± 0.82 t/ha). The carbon stock at a depth of 0-15 cm exceeds that found at 15-30 cm. Over the 1,330.80 ha of the Kou River riparian buffer zone, the total atmospheric carbon sequestered in the soil was 119,833.27 metric tons of CO2. Conclusion: The carbon sequestration capacity of the Kou River’s riparian zone is significant in the context of climate change mitigation.

Geoinformation analysis of the vegetation cover of the northern regions of the Irkutsk region and its changes under the impact of fires

The northern regions of the Irkutsk oblast have been prone to fires in recent decades as a result of oil and gas and infrastructure development, as well as changing climatic conditions. Geoinformation mapping and analysis of the spatial structure of the vegetation cover of two key areas located in the area of the Vershina Khandy village and the village of Tokma for two time slices: 2013–2014 and 2018–2019 (before and after active forest fires, mainly in 2016 and 2017) was carried out. Mapping was done using the Random Rorest supervised classification method, Landsat 8 space images. As a result, 10 classes of vegetation cover were identified on the territory. It is shown that in the territories in 2013–2014 coniferous (light coniferous – pine-larch, larch-pine, less often dark coniferous – cedar and spruce) forests prevailed, as well as mixed coniferous-deciduous and deciduous forests and swamps. From 2013 to 2019, more than 20% of Tokma and more than 5% of Khanda key area was covered by fires. Mostly coniferous forests, as well as ernik bogs in the Tokma area were burning.

Star γ Covers and Their Applications to Selection Principles

Abstract By imposing the star operator on γ covers, we establish the notion of st-γ covers. The interrelationships among many recent variants of γ covers have been identified, and various counterexamples have been shown to distinguish them. An investigation has been conducted on the basic topological attributes of these covers. We have also examined the implications between various selection principles and topological games in the context of the class of st-γ coverings.

Investigating the nexus of urban expansion, wetlands, and livelihoods from 1991 to 2021: evidence from Hawassa, Ethiopia

ABSTRACT Urban expansion in Ethiopia, particularly informal construction, has led to complex and dynamic changes in livelihoods and wetlands. This study examined the effects of urban expansion on the Cheleleka Wetland and the livelihood of the peripheral community in Hawassa. A mixed research design was used to quantitatively analyze Landsat imagery (1991, 2001, 2011, 2021) and socioeconomic data. Key informant interviews and focus group discussion data were narrated qualitatively. The built-up area substantially increased from 718.11 ha in 1991 to 4820.71 ha in 2021. However, agricultural and wetland areas decreased from 8807.58 ha and 8177.04 ha in 1991 to 6382.44 ha and 7030.26 ha in 2021, respectively. Water and forest areas fluctuated. Built-up areas had a significant advantage over other land use and cover classes, with agricultural, wetland, and forest areas being converted into built-up areas throughout the study period. As a result, the area of Lake Hawassa expanded. Currently, the Cheleleka Wetland has been negatively impaired by overgrazing, cutting of grass, animal dung dumps, eucalyptus plantations, agricultural expansion, and urban expansion. Urban expansion has exacerbated social and economic divisions, resulting in the concentration of poor-quality neighborhoods in the peripheral community. To protect the wetlands, alternative livelihoods and participatory land-use planning are crucial.

Long-term assessment of fire-induced carbon loss in Southeast Atlantic Forest

Fires threaten tropical forests such as Atlantic Forest in Brazil, compromising the ecosystem service of carbon stock. However, there is a literature gap regarding these studies in these ecosystems. Therefore, we conducted this analysis in different land use and land cover (LULC) classes, considering seasonality and topographic, hydrological, anthropogenic and fire variables correlations, during 2000-2020. The InVEST Carbon model was used, applied to carbon biomass pre-fire and pos-fire, based on field work and linear regression, weighted by pre- and post-fire NBR spectral index. The results, in 21 years, revealed a total loss after fire of 55.7GgC (43%), and of these, 79% is in old-growth Ombrophilous dense. In general, fire negatively impacts the carbon stock of native forests by an average of 38% (ranging from 19.9% to 69.1%, depending on phytophysiognomy and seasonality), Eucalyptus plantations by 87.1%, high-altitude grasslands by 79.5% and pasture in 90.4%. Burn frequency and severity as well as distance from rivers and roads were significantly correlated with carbon loss. A small portion of this biome has shown a high potential for fire-induced carbon loss, indicating a danger for the whole Atlantic Forest conservation and to international agreements commitments.

A machine learning approach for estimating snow depth across the European Alps from Sentinel-1 imagery

Seasonal snow plays a crucial role in society and understanding trends in snow depth and mass is essential for making informed decisions about water resources and adaptation to climate change. However, quantifying snow depth in remote, mountainous areas with complex topography remains a significant challenge. The increasing availability of high-resolution synthetic aperture radar (SAR) observations from active microwave satellites has prompted opportunistic use of the data to retrieve snow depth remotely across large spatial and frequent temporal scales and at a high spatial resolution. Nevertheless, these novel SAR-based snow depth retrieval methods face their own set of limitations, including challenges for shallow snowpacks, high vegetation cover, and wet snow conditions. In response, here we introduce a machine learning approach to enhance SAR-based snow depth estimation over the European Alps. By integrating Sentinel-1 SAR imagery, optical snow cover observations, and topographic, forest cover and snow class information, our machine learning retrieval method more accurately estimates snow depth at independent in-situ measurement sites than current methods. Further, our method provides estimates at 100 m horizontal resolution and is capable of better capturing local-scale topography-driven snow depth variability. Through detailed feature importance analysis, we identify optimal conditions for SAR data utilization, thereby providing insight into future use of C-band SAR for snow depth retrieval.

Determination of the Change of Agricultural Landscapes Based on CORINE Land Cover Agricultural Land Classes Using GIS and Visual Quality Value with the AHP Method: The Case of Tekı̇rdağ Province

Agricultural landscapes contribute positively to the visual texture of the city and change over time. In this context, this study aims to determine the visual landscape quality of agricultural landscapes based on parameters on expert approach the hypothesis; it is important and necessary to determine the visual landscape values based on the change in agricultural landscapes over time. Within the framework of the developed hypothesis, it aims to make an expert-based visual landscape assessment of the agricultural landscapes of Tekirdağ Province, located in the northwestern part of Türkiye, based on nine subclasses created according to the CORINE land cover (CLC) classes, on the basis of 4 main parameters. Using the AHP technique, the priorities of the parameters and agricultural landscape classes and the relationships between basic parameters and visual preferences were determined. In this direction, the questions; What are the priorities of parameters that are effective in determining the visual quality of agricultural landscapes on the basis of expert approach?; What kind of changes have occurred in agricultural landscapes during the years 1990-2000-2006-2012-2018 in 9 subclasses created according to the CLC classes? and; How should the visual landscape quality values of the agricultural landscape subclasses created according to the CLC classes be ranked? The results of the study can be used as a tool in landscape planning and management studies as a factor in strengthening landscape quality.

A novel hybrid methodology integrating pixel- and object-based techniques for mapping land use and land cover from high-resolution satellite data

ABSTRACT The classification of very high-resolution satellite imagery remains a focal point in remote sensing, attracting increased attention across diverse scientific disciplines. Various classification methods, including pixel- and object-based techniques, have been proposed, and their performances and limitations have been discussed in the literature. This paper presents a hybrid method that combines the strengths of pixel- and object-based methods in image classification to minimize errors associated with the segmentation process, particularly under-segmentation errors in object-based image analysis. The core concept behind the method lies in categorizing segmented image objects as either homogeneous or heterogeneous based on their class probability. In this process, the estimated possibilities from the object-based classification model are considered, and segments are designated as homogeneous or heterogeneous using a user-defined threshold. The object-based classification model determines the class labels for homogeneous image objects, while the heterogeneous ones, containing pixels representing different land cover classes, are classified using the pixel-based model. The performance of hybrid classification models, created by varying thresholds, is analysed using high-resolution WorldView-3 and WorldView-2 imagery and compared with pixel- and object-based classification results. For the implementation of image classification methods, Canonical Correlation Forest (CCF), Extreme Gradient Boosting (XGBoost), and Categorical Boosting (CatBoost) were employed. The findings indicated that employing the suggested hybrid strategy with a threshold value selected within a specific range (e.g. between 60% and 80%) and employing a robust classification algorithm that provides class probabilities (e.g. CCF) results in a statistically significant improvement in overall accuracy compared to pixel and object-based methods, with gains of 5% and 4%, respectively. Visual analysis of the produced thematic maps revealed that the proposed method minimizes the salt-and-pepper effect associated with pixel-based classification while mitigating segmentation problems arising from object-based classification.

Variability in land surface temperature concerning escalating urban development using thermal data of Landsat sensor: A case study of Lower Kharun Catchment, Chhattisgarh, India

Over the past few years, there has been a revitalized emphasis on comprehending the shifts in land cover and their implications for a range of environmental factors. This investigation seeks to analyze how changes in land surface temperatures (LST), normalized difference vegetation index (NDVI), normalized difference built-up index (NDBI), and alterations in land cover intersect within the lower Kharun catchment area. The primary dataset utilized in this study is for 2001 and 2021 Landsat 7 and 8, part of the Landsat program managed by the United States Geological Survey (USGS), offer essential Earth observation data using their multispectral and thermal sensors which are designed to detect thermal radiation emitted from the Earth's surface. When these bands are properly processed, they enable accurate temperature measurements. Visual interpretation was conducted on these images, categorizing them into five specific classes of land cover these were vegetation, open land, settlement, waterbodies, and cultivation. Following this, spectral indices like NDVI and NDBI were calculated, and LST was derived using a single-channel algorithm. Subsequently, correlation analysis was utilized to explore the interconnectedness or mutual relationship among the spatial distribution of these parameters. Over the period from 2001 to 2021, the most significant changes in land use were observed in the settlement area and cultivation, which increased by 6.92 and 6.23 sq. km, respectively. Conversely, open land, vegetation, and waterbodies experienced decreases of 7.13, 5.56, and 0.46 sq. km, respectively. The patterns in which LST, NDBI, and NDVI are distributed, exhibited corresponding variations following changes in land cover. The observed alterations in LST, NDBI, and NDVI are believed to be primarily influenced by the expansion of built-up areas. A noticeable association suggests that as built-up areas increase, both NDBI and LST values typically rise.Furthermore, a correlation observed between LST with NDVI was negative, suggesting an inverse relationship between these parameters. On the other hand, the correlation of LST with NDBI observed was positive, indicating that these parameters exhibit a direct relationship. Overall, these findings seem to be complex and highlight the interactions between changing land cover and environmental parameters, underscoring the importance of understanding these relationships for effective land management and environmental monitoring.

What is the actual composition of specific land cover? An evaluation of the accuracy at a national scale – Remote sensing in comparison to topographic land cover

Satellite imagery allows us to capture and collect land cover information for increasingly large areas. This allows us to represent current land cover on maps in a simple and standardized way; however, any land cover determined in this way is subject to some algorithmic uncertainty. This paper aims, for the first time, to indicate the magnitude of this uncertainty through the empirical probability distribution of a given land cover at a given location. By analyzing 3 data sources, i.e. the Corine Land Cover map, the POLSA land cover map and the classic map - the BDOT10k database of topographic objects. Empirical distributions of the occurrence of land cover class data in areas with a given land use on a topographic map were determined. The work was carried out on a large scale, i.e. on the maximum possible sample for Poland, i.e. on the area of the whole country. This makes it possible to introduce and quantify uncertainties. Spatial analyses were carried out using satellite-based methods to determine land cover or using a topographic map. This work and its results will be useful to all users who want to assess the occurrence of a phenomenon in a given area, taking into account the uncertainty of the land cover, and thus obtain more accurate and reliable results. It also provides, for the first time, a methodology for verifying such map correspondences, which can be replicated in work by other researchers, using the confusion matrix and as evaluation metrics the true positive rate (TPR) and weighted accuracy have been adopted. The paper proposes a link between land cover classes in all databases. It was shown that the TPR for BDOT10k was higher than 50% only with CLC Level 1 (72.0%) and POLSA Land Cover (61%), while the TPR for RS classes for each remote sensing data was always higher than 60% with BDOT10k. The class with the highest remote sensing classes was related to water, especially marine (92.0% for POLSA and 85.3% for CLC level 3), arable land (98% for POLSA, lowest for CLC level 3 (80%), and forests (coniferous POLSA – 89%, CLC level 1 and 2–85%), while low values were obtained for wetlands, peatbogs. The authors do not state which land cover approach is better, as each may have multiple uses, but the values presented in this work must raise awareness of uncertainties in land cover and critical implementation in decision-making processes for multiple areas of human activity. The study provides ready-to-use values of the probability of a given land cover class being present on a topographic map, given that remote sensing has classified it as such. These functions can also be used in reverse, to determine the probability of a given land cover class being present in remote sensing, given that a specific class has been identified on a topographic map. The results of the consistency assessment, with the composition structure, can be used by a wide range of users, including public administration, land managers, land architects, public services, academia and individuals.