Cover Classification Research Articles

Continuity of Top-of-Atmosphere, Surface, and Nadir BRDF-Adjusted Reflectance and NDVI between Landsat-8 and Landsat-9 OLI over China Landscape

The successful launch of Landsat-9 marks a significant achievement in preserving the data legacy and ensuring the continuity of Landsat’s calibrated Earth observations. This study comprehensively assesses the continuity of reflectance and the Normalized Difference Vegetation Index (NDVI) between Landsat-8 and Landsat-9 Operational Land Imagers (OLIs) over diverse Chinese landscapes. It reveals that sensor discrepancies minimally impact reflectance and NDVI consistency. Although Landsat-9’s top-of-atmosphere (TOA) reflectance is slightly lower than that of Landsat-8, small root-mean-square errors (RMSEs) ranging from 0.0102 to 0.0248 for VNIR and SWIR bands (and larger RMSE for NDVI at 0.0422) fall within acceptable ranges for Earth observation applications. Applying atmospheric corrections markedly enhances reflectance uniformity and brings regression slopes closer to unity. Further, Bidirectional Reflectance Distribution Function (BRDF) adjustments improve comparability, ensuring measurement reliability, and the NDVI maintains robust consistency across various reflectance types, time series, and land cover classes. These findings affirm Landsat-9’s success in achieving data continuity within the Landsat program, allowing interchangeable use of Landsat-8 and Landsat-9 OLI data for diverse Earth observation purposes. Future research may explore specific sensor correlations across different vegetation types and seasons while integrating data from complementary platforms, such as Sentinel-2, to enhance the understanding of data continuity factors.

Mapping the Land Suitability Rating of Arabica Coffee Crops: A Geographical Indication Factor-Based Approach

Aims: To map land suitability ratings and organoleptic characteristics of arabica coffee using physical and human dimension attributes constructed from geographical indication components.
 Study Design: The study was designed in an integrated manner, using a quantitative approach integrated with a qualitative-descriptive approach and supported by field surveys.
 Place and Duration of Study: Conducted in 3 sub-districts within the scope of Sinjai Regency, South Sulawesi Province, Indonesia, at altitudes > 800 m.asl., between February 2021 and August 2022.
 Methodology: The method for determining the observation location point is on-screen digitization based on the land cover class seen with satellite imagery. Ground truth implementation, 16 location points were determined with their respective geographical coordinate positions. To analyze soil properties and characteristics, 41 samples were obtained: An organoleptic test of coffee beans, as many as 3 samples representing the sub-district area.
 Results: Mapping using physical dimensions in the West Sinjai area resulted in a moderate land suitability rating with an area of 6,976.46 ha; the Sinjai Borong area obtained a less suitable rating of 7,530.32 ha and a moderate rating of 1,457.17 ha, the Central Sinjai area obtained a less suitable rating of 822.16 ha, and a moderate rating of 1,586.47 ha. The compilation of physical and human dimensions in the West Sinjai area shows a moderate rating of 1,057.97 ha and a suitable rating of 5,918.49 ha, Sinjai Borong produces a moderate rating of 8,821.48 ha and a suitable rating of 166.01 ha, Central Sinjai obtained land suitability with a moderate rating of 986.55 ha and a suitable rating of 1,422.08 ha. West Sinjai and Central Sinjai have an organoleptic score of 88.50, with a more specific floral aroma than Sinjai Borong.
 Conclusion: Mapping the land suitability ranking of arabica coffee plants using physical and human dimensions provides clear and logical results.

One metric or many? Refining the analytical framework of landscape resistance estimation in individual-based landscape genetic analyses.

One of the allures of landscape genetics is the ability to leverage pairwise genetic distance metrics to infer how landscape features promote or constrain gene flow (i.e. landscape resistance surfaces). Critically, properly parameterized landscape resistance surfaces are foundational to applied conservation and management decisions. As such, there has been considerable effort expended assessing methods and metrics to estimate landscape resistance from genetic data (Balkenhol et al., Ecography, 32, 2009, 818; Peterman et al., Landsc. Ecol., 34, 2019, 2197; Shirk et al., Mol. Ecol. Resour., 17, 2017, 1308; Shirk et al., Mol. Ecol. Resour., 18, 2018, 55). Nonetheless, a primary challenge to assessing the effects of landscapes on gene flow is in the estimation of landscape resistance values, and this problem becomes increasingly challenging as more landscape features or land cover classes are considered. It quickly becomes infeasible to adequately assess the potential parameter space through manual or systematic assignment of resistance values. The development of ResistanceGA (Peterman, Methods Ecol. Evol., 9, 2018, 1638) provided a framework for using genetic algorithms to optimize landscape resistance values and identify the best statistical relationship between pairwise effective distances and genetic distances. ResistanceGA has seen extensive use in both population- and individual-based landscape genetic analyses. However, there has been relatively limited assessment of ResistanceGA's ability to identify the landscape features affecting gene flow (but see Peterman et al., Landsc. Ecol., 34, 2019, 2197; Winiarski et al., Mol. Ecol. Resour., 20, 2020, 1583) or the sensitivity of ResistanceGA results to the choice of genetic distance metric used. In the current issue of Molecular Ecology Resources, Beninde et al. (2023) aim to address these knowledge gaps by examining the impact of individual-based genetic distance measures on landscape genetic inference.

Semi-Supervised Learning Method for the Augmentation of an Incomplete Image-Based Inventory of Earthquake-Induced Soil Liquefaction Surface Effects

Soil liquefaction often occurs as a secondary hazard during earthquakes and can lead to significant structural and infrastructure damage. Liquefaction is most often documented through field reconnaissance and recorded as point locations. Complete liquefaction inventories across the impacted area are rare but valuable for developing empirical liquefaction prediction models. Remote sensing analysis can be used to rapidly produce the full spatial extent of liquefaction ejecta after an event to inform and supplement field investigations. Visually labeling liquefaction ejecta from remotely sensed imagery is time-consuming and prone to human error and inconsistency. This study uses a partially labeled liquefaction inventory created from visual annotations by experts and proposes a pixel-based approach to detecting unlabeled liquefaction using advanced machine learning and image processing techniques, and to generating an augmented inventory of liquefaction ejecta with high spatial completeness. The proposed methodology is applied to aerial imagery taken from the 2011 Christchurch earthquake and considers the available partial liquefaction labels as high-certainty liquefaction features. This study consists of two specific comparative analyses. (1) To tackle the limited availability of labeled data and their spatial incompleteness, a semi-supervised self-training classification via Linear Discriminant Analysis is presented, and the performance of the semi-supervised learning approach is compared with supervised learning classification. (2) A post-event aerial image with RGB (red-green-blue) channels is used to extract color transformation bands, statistical indices, texture components, and dimensionality reduction outputs, and performances of the classification model with different combinations of selected features from these four groups are compared. Building footprints are also used as the only non-imagery geospatial information to improve classification accuracy by masking out building roofs from the classification process. To prepare the multi-class labeled data, regions of interest (ROIs) were drawn to collect samples of seven land cover and land use classes. The labeled samples of liquefaction were also clustered into two groups (dark and light) using the Fuzzy C-Means clustering algorithm to split the liquefaction pixels into two classes. A comparison of the generated maps with fully and manually labeled liquefaction data showed that the proposed semi-supervised method performs best when selected high-ranked features of the two groups of statistical indices (gradient weight and sum of the band squares) and dimensionality reduction outputs (first and second principal components) are used. It also outperforms supervised learning and can better augment the liquefaction labels across the image in terms of spatial completeness.

Evaluation and comparison of MODIS aerosol optical depth retrieval algorithms over Brazil

Brazil experiences significant aerosol loads throughout the year, particularly during the biomass-burning season in the Amazon. Thus, given the importance of aerosols for climate and health, this research aimed to validate and compare Aerosol Optical Depth (AOD) products over Brazil. This evaluation considers three algorithms that retrieve AOD by using data from Moderate Resolution Imaging Spectroradiometer (MODIS) sensor: Dark Target (DT) at 3 and 10 km resolution, Deep Blue (DB), and Multi-Angle Implementation of Atmospheric Correction (MAIAC). To validate the satellite data, 17 sunphotometers from the AErosol RObotic NETwork (AERONET) were utilized. The results show a high correlation (R > 0.9) between the MODIS-AOD products and ground-based data. However, MODIS-AOD products tend to overestimate or underestimate AOD values, depending on the specific AOD value and algorithm evaluated. Additionally, it was observed that the performance of the algorithms is influenced by factors such as land cover type, view geometry, and the spatiotemporal distribution of aerosols. In particular, challenges were encountered when retrieving robust AOD data for Savanna and Urban cover classes. In conclusion, the results indicate that MAIAC and DB algorithms demonstrate greater stability in retrieving AOD values. Nevertheless, caution should be exercised when applying these products to map aerosols on specific surfaces, such as urban areas.

Elucidating factors driving post-fire vegetation recovery in the Mediterranean forests using Landsat spectral metrics

Wildfires represent one of the primary disturbance agents in the Mediterranean, significantly affecting the ecological integrity of forests. Therefore, understanding the spatial patterns of post-fire vegetation recovery is crucial to improving forest restoration planning and assessing the regeneration capacity of different forest stands that have been impacted by wildfires. In this study, we analysed post-fire vegetation recovery rates within the context of fire severity, pre-fire vegetation, and post-fire climate conditions, for different Mediterranean forest classes, namely, Mediterranean pine, holm, deciduous oak forests, sclerophyllous vegetation, and thermophilous shrublands. Basilicata, in Italy, was chosen as a study area, as it represents a wide range of forests.The Relative Recovery Indicator (RRI) was derived from Normalized Burn Ratio (NBR) patterns extracted from 30-meter Landsat time series for different wildfires that occurred during the 2004–2016 within Google Earth Engine (GEE) environment. A Linear Mixed Model (LMM) was used to test the effect of the different variables on the RRI. Results showed a general decrease in recovery rate within five-years post-fire for each forest cover class, which is mainly related to pre- and post-fire conditions. Pre-fire vegetation conditions significantly influenced post-fire vegetation recovery, especially in sclerophyllous and deciduous oak forests. Post-fire climate conditions (e.g., temperature) were also important predictors of vegetation recovery explaining the variation in post-fire RRI patterns. The proposed method could provide new insights into the restoration and management of forest ecosystems in the Mediterranean.

Relationship between the hydrological regime and the land use and cover of the Rio das Mortes Basin

Interactions between land use and cover and a basin hydrological regime are complex and dynamics. The identification of trends and changes in rainfall and flow regimes in a region is relevant in water resources planning and management so that environmental, economic and social adversities are avoided. Thus, this study aims to identify possible trends in the historical series of rainfall and fluviometric stations in Rio das Mortes basin, Minas Gerais, Brazil, as well as to relate the modifications in the flow rate behaviour due to changes in land use and cover in the basin between the years of 1985 and 2017. For that, hydrological data were used from the Ibituruna-MG river station and 12 rainfall stations and orbital images from Landsat-5 and Sentinel satellites. Flow rate and precipitation behavior were performed using the nonparametric Mann-Kendall test and the Pettitt’s test, while the land use and cover classification was performed in an eCognition Developer® 64 environment. Significant trends were observed in the average of total annual rainfall series and decreasing trends in the average of monthly flow rate series of lowest runoff over the years 1925 to 2014. There were changes in land use and cover in the studied region, however, these were not significant enough to modify the hydrological regime of Rio das Mortes basin during the analysed period.

Object-Based Image Analysis (OBIA) and Machine Learning (ML) Applied to Tropical Forest Mapping Using Sentinel-2

The purpose of this research was to distinguish and estimate natural forest areas at Paraná, Brazil. Forest plantations (Silviculture) and natural forests have high annual vegetative vigor, as well as agricultural areas in the periods of agricultural harvests, which can bring classification errors between these classes of Land Use and Land Cover (LULC), these classes have similar spectral signatures, but have a distinct texture that can be separated in the supervised classification process, with the joining of object and pixel-to-pixel classification method approaches. Thus, image segmentation techniques through Object-Based Image Analysis (OBIA) and Machine Learning (ML) made forest mapping possible over a large territorial extension. The Google Earth Engine (GEE) platform was used to calculate the vegetation indices (VIs) and Spectral Mixture Analysis (SMA) fraction spectral from Sentinel-2 images, and the creation of homogeneous spectrally shaped regions under supervised classification of phytoecological regions and mesoregions. The overall precision obtained in the mappings resulted in 0.94 Kappa Index (KI) and 96% of Overall Accuracy (OA), which indicates a high performance in large-scale forest mapping. The proposed dataset, source codes and trained models are available on Github (https://github.com/Cechim/simepar-brazil/), creating opportunities for further ad vances in the field.

Development of an Ontology-Based Technique for Labeling Land Cover Classes with Minimum Utilization of SAR Features

Development of an Ontology-Based Technique for Labeling Land Cover Classes with Minimum Utilization of SAR Features

Land Cover Classification Using Sentinel 2A Image in Kolaka Subdistrict, Kolaka Regency, Southeast Sulawesi

Land cover in Kolaka District continues to change. Mapping and identifying land cover types using the Maximum Likelihood method is more accurate than other methods. This research aims to analyze the capabilities of Sentinel 2A Imagery and the Maximum Likelihood classification method for mapping and identifying land cover types in Kolaka District. This research was carried out from July to September 2023 and was carried out in 4 stages, namely the first stage of image pre-processing by carrying out the layer stacking process. The second stage is image analysis and classification. The third stage is carrying out a Ground Check, and the fourth stage is validation and accuracy testing. The value of the accuracy test results with Overall Accuracy (OA) is 88.75% which is in the good category. The results of the land cover classification obtained 8 land cover classes, namely secondary dry land forest covering an area of 3974.20 Ha or 31.84%, plantation land cover covering an area of 3,886.87 Ha or 31.14%, dense bushes covering an area of 1,641.42 Ha or 13.15%, mixed dry land agricultural land cover covering an area of 1,415.62 Ha or 11.34%, residential land cover of 744.26 or 5.96%, paddy field cover of 613.53 Ha or 4.92%, open land cover of 148.66 Ha or 1.19% and water body land cover of 56.22 Ha or 0.45% of the total area of Kolaka District.

Land Use and Land Cover Change Modulates Hydrological Flows and Water Supply to Gaborone Dam Catchment, Botswana

Identifying the mechanism through which changes in land use and land cover (LULC) modulate hydrological flows is vital for water resource planning and management. To examine the impact of LULC change on the hydrology of the Gaborone Dam catchment within the upper Limpopo basin, where Notwane river is the major river within the catchment, three LULC maps for the years 1997, 2008, and 2017 were established based on a mosaic of Landsat 5 for 1997 and 2008 and Landsat 8 for 2017. The 10 m-resolution Version 200 ESA World Land Cover Map for 2021 is used as a ground truth to train the random forest (RF) classifier to identify land cover classes from Landsat 8 imageries of 2021 using the Google Earth Engine (GEE) Python API. The overall accuracy/kappa coefficient of the RF classifier is 0.99/0.99 for the training and 0.73/0.68 for the validation data sets, which indicate excellent and substantial agreements with the ground truth, respectively. With this confidence in the LULC classification, the impact of LULC change on the hydrological flow within the catchment was estimated by employing the Soil and Water Assessment Tool (SWAT) and indicator of hydrological alteration (IHA). The SWAT model calibration and validation were first performed, and the ability of the model to capture the observed stream flow was found to be good. The LULC maps from Landsat images during the 1997–2017 period show a decrease in forests and shrubland in contrast to an increase in pasture land. The expansion of pasture and cropland and the reduction in forests and shrubland led to a decline in the amount of evapotranspiration and groundwater recharge. Furthermore, the LULC change also caused a reduction in low flow during dry periods and an increase in high flow during the rainy season. The findings clearly demonstrate that LULC changes can affect the water table by altering soil water recharge capacity. The study highlighted the importance of LULC for catchment water resource management through land use planning to regulate the water level in the Gaborone Dam against the impact of climate change and growing water demands by the city of Gaborone due to population growth.

Monitoring vegetation dynamics using multi-temporal Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) images of Tamil Nadu

Vegetation indices serve as an essential tool in monitoring variations in vegetation. The vegetation indices used often, viz., normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) were computed from MODIS vegetation index products. The present study aimed to monitor vegetation's seasonal dynamics by using time series NDVI and EVI indices in Tamil Nadu from 2011 to 2021. Two products characterize the global range of vegetation states and processes more effectively. The data sources were processed and the values of NDVI and EVI were extracted using ArcGIS software. There was a significant difference in vegetation intensity and status of vegetation over time, with NDVI having a larger value than EVI, indicating that biomass intensity varies over time in Tamil Nadu. Among the land cover classes, the deciduous forest showed the highest mean values for NDVI (0.83) and EVI (0.38), followed by cropland mean values of NDVI (0.71) and EVI (0.31) and the lowest NDVI (0.68) and EVI (0.29) was recorded in the scrubland. The study demonstrated that vegetation indices extracted from MODIS offered valuable information on vegetation status and condition at a short temporal time period.

Assessing the Accuracy of Multi-Temporal GlobeLand30 Products in China Using a Spatiotemporal Stratified Sampling Method

The new type of multi-temporal global land use data with multiple classes is able to provide information on both the different land covers and their temporal changes; furthermore, it is able to contribute to many applications, such as those involving global climate and Earth ecosystem analyses. However, the current accuracy assessment methods have two limitations regarding multi-temporal land cover data that have multiple classes. First, multi-temporal land cover uses data from multiple phases, which is time-consuming and inefficient if evaluated one by one. Secondly, the conversion between different land cover classes increases the complexity of the sample stratification, and the assessments with different types of land cover suffer from inefficient sample stratification. In this paper, we propose a spatiotemporal stratified sampling method for stratifying the multi-temporal GlobeLand30 products for China. The changed and unchanged types of each class of data in the three periods are used to obtain a reasonable stratification. Then, the strata labels are simplified by using binary coding, i.e., a 1 or 0 representing a specified class or a nonspecified class, to improve the efficiency of the stratification. Additionally, the stratified sample size is determined by the combination of proportional allocation and empirical evaluation. The experimental results show that spatiotemporal stratified sampling is beneficial for increasing the sample size of the “change” strata for multi-temporal data and can evaluate not only the accuracy and area of the data in a single data but also the accuracy and area of the data in a multi-period change type and an unchanged type. This work also provides a good reference for the assessment of multi-temporal data with multiple classes.

Land Use and Land Cover Change Detection Using the Random Forest Approach: The Case of The Upper Blue Nile River Basin, Ethiopia.

Monitoring land use change dynamics is critical for tackling food security, climate change, and biodiversity loss on a global scale. This study is designed to classify land use and land cover in the upper Blue Nile River Basin (BNRB) using a random forest (RF) algorithm. The Landsat images for Landsat 45, Landsat 7, and Landsat 8 are used for classification purposes. The study area is classified into seven land use/land cover classes: cultivated lands, bare lands, built-ups, forests, grazing lands, shrublands, and waterbodies. The accuracy of classified images is 83%, 85%, and 91% using the Kappa index of agreements. From 1983 to 2022 periods, cultivated lands and built-up areas increased by 47541 and 1777 km2, respectively, at the expense of grazing lands, shrublands, and forests. Furthermore, the area of water bodies has increased by 662 km2 due to the construction of small and large-scale irrigation and hydroelectric power generation dams. The main factors that determine agricultural land expansion are related to population growth. Therefore, land use and land cover change detection using a random forest is an important technique for multispectral satellite data classification to understand the optimal use of natural resources, conservation practices, and decision-making for sustainable development.

Park conservation or degradation? iCLUE modelling of land use change projections in the upper Manafwa watershed on Mount Elgon, Uganda

Land use changes result from interactions between different factors over time. Land use models support natural resource conservation by helping stakeholders to understand and predict these changes. Through such model-based projections, governments and conservationists can identify hotspots of encroachment and formulate informed decisions for sustainable conservation. Land use on Mount Elgon has undergone rapid changes in the conservation area and community land since the 1970s. Park encroachment has threatened natural cover due to population pressure and increasing demand for agricultural land. To support conservation efforts, this study conducted near-future potential land use change projections using historical land use change and iCLUE model simulations. This study established hotspots of land use conversion, locations of park encroachment and the extent of land use intensification in the upper Manafwa watershed of Mount Elgon. The study applied remote sensing, a geographic information system, and a land use change model (iCLUE), while considering nine land use classes to project land use change for 2030, 2040 and 2050 using a “business-as-usual” scenario. Stakeholder verification of the model output was conducted in a dedicated workshop. The results showed that in 2050, agriculture (44.24%) and low-stocked tropical high forest (22.56%) will be the largest land cover classes with bushland and shrubs as the smallest land cover classes (∼0.5%). Most of the land use conversion in the park is projected to occur along the boundary of the conservation area by agriculture and built-up areas. This study will guide stakeholders and decision makers in planning future sustainable management strategies to conserve and enhance remaining forest tracts in Mount Elgon.

TTNet: A Temporal-Transform Network for Semantic Change Detection Based on Bi-Temporal Remote Sensing Images

Semantic change detection (SCD) holds a critical place in remote sensing image interpretation, as it aims to locate changing regions and identify their associated land cover classes. Presently, post-classification techniques stand as the predominant strategy for SCD due to their simplicity and efficacy. However, these methods often overlook the intricate relationships between alterations in land cover. In this paper, we argue that comprehending the interplay of changes within land cover maps holds the key to enhancing SCD’s performance. With this insight, a Temporal-Transform Module (TTM) is designed to capture change relationships across temporal dimensions. TTM selectively aggregates features across all temporal images, enhancing the unique features of each temporal image at distinct pixels. Moreover, we build a Temporal-Transform Network (TTNet) for SCD, comprising two semantic segmentation branches and a binary change detection branch. TTM is embedded into the decoder of each semantic segmentation branch, thus enabling TTNet to obtain better land cover classification results. Experimental results on the SECOND dataset show that TTNet achieves enhanced performance when compared to other benchmark methods in the SCD task. In particular, TTNet elevates mIoU accuracy by a minimum of 1.5% in the SCD task and 3.1% in the semantic segmentation task.

Investigations of Spatial and Temporal Land Use/Land Cover Changes in Trabzon Province (1990-2018) Using CORINE Maps and Landscape Metrics

In this study; it is aimed to analyze the temporal and spatial changes in land use/land cover of Trabzon province with the population increase in the 28-year period between 1990-2018 by analyzing it with CORINE maps and landscape metrics. Within the scope of the study, temporal variation analysis was carried out using CORINE maps of the years 1990, 2000, 2006, 2012 and 2018. In Trabzon, according to the CORINE classification system, a total of 4 main groups have been identified at the first level, and at the third level, 19 land use/land cover classes have been identified for 1990, 20 for 2000, 24 for 2006, and 26 for 2012 and 2018. It was observed that the most differentiation occurred in the agricultural areas class in the 28-year period. In addition, the metrics of class area (CA), number of patches (NumP), mean patch size (MPS), edge density (ED), total edge (TE) and mean shape index (MSI) were calculated through ARCGIS's Patch Analyst plugin in the study. The changes that have occurred have been examined. Accordingly, there has been a decrease in the size of forest and seminatural areas and agricultural areas with the increase in the size of the artificial surfaces due to the increase in human activities and population over the past 28 years. In addition, due to the continuous increase in the number of patches in the agricultural areas over time and the decrease in the average size of the patches, the fragmentation of the agricultural areas has been more common. The least change was in the water bodies. In line with the findings obtained, sustainable planning and land use suggestions for land use were presented.

Impacts of Human Activities on Urban Sprawl and Land Surface Temperature in Rural Areas, a Case Study of El-Reyad District, Kafrelsheikh Governorate, Egypt

Anthropogenic activities affect the surrounding environment dynamically in different ways. In the arid and hyper arid, agriculture is concentrated in rural communities, which are cooling surfaces that help mitigate surface temperature increases. Recently, rural communities are suffering from increasing urban sprawl. The current work focuses on evaluating the changes in land cover and their impacts on land surface temperature (LST) during (1988–2022) and predicting the changes until 2056 in El-Reyad District, Kafrelsheikh Governorate, Egypt. For achieving this purpose, Landsat images (TM, ETM+, and OLI) were used. The support vector machine (SVM) was applied using Google Earth Engine (GEE) to monitor changes in land use/cover and LST. The prediction of land use until 2056 was achieved using the CA-Markov simulation model. The results showed six land cover classes: agricultural lands, bare lands, urban areas, natural vegetation, Lake Burullus, and fish farms. The results showed the effects of human activity on the conversion of agricultural land to other activities, as agricultural lands have decreased by about 3950.8 acres, while urban areas have expanded by 6283.2 acres, from 1988 to 2022. Fish farms have increased from 3855.6 to 17,612 acres from 1988 to 2022. While the area of bare land decreased from 28.3% to 0.7% of the total area, it was converted to urban, agricultural, and fish farms. The spatiotemporal change in land cover affected the balance of LST in the study area, although the average temperature increased from 32.4 ± 0.5 to 33.6 ± 0.2 °C. In addition, it is expected to reach 36 ± 0.5 °C in 2056, and there are some areas with decreased LST where it is converted from bare areas into fish farms and agricultural uses. The prediction results show that the agricultural area will decrease by −11.38%, the urban area will increase by 4.6%, and the fish farms area will increase by 6.1%. Thus, preserving green spaces and reducing urban sprawl in rural communities are very important objectives.

Urban Land Use and Land Cover Mapping

The Brazilian urbanization process produced a complex urban space, with a variety of urban land use and cover as a result. The study of these forms through a classification system is essential, but most current systems don’t capture this complexity, condensing it. Urban forms are difficult to distinguish and classify, resulting in the need for a system with a high degree of detail, for a more accurate urban planning. The objective of this work is to propose a classification system for urban land use and cover, which can demonstrate the multiplicities through remote sensing, using data from CBERS 04A satellite. The methodology surveys the visual elements of remote sensing images, through visual interpretation, relating them to each proposed use and cover class. With this, a classification system was developed that covers the urban space in 17 classes, being an effective way to raise information about the different urban forms.

Evaluation of Machine Learning Algorithms in the Classification of Multispectral Images from the Sentinel-2A/2B Orbital Sensor for Mapping the Environmental Dynamics of Ria Formosa (Algarve, Portugal)

With the growing availability of remote sensing orbital spatial data, the applications of machine learning (ML) algorithms have been leveraging the field of process automation in image classification. The present work aimed to evaluate the precision and accuracy of ML algorithms in the classification of Sentinel 2A/2B images from an area of high environmental dynamics, such as Ria Formosa (Algarve, Portugal). The images were submitted to classification by groups of ML algorithms such as the Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbors (KNN), and Decision Tree (DT). The Orfeo Toolbox (OTB) open-source programming package made the algorithms available. Ten samples were collected for each of the 14 land use and cover classes in the Ria Formosa area, totaling 140 samples. Of these, 70% were for training and 30% for validating the classification. The evaluation metrics used were the class discrimination measures: Recall (R), the Global Kappa Index (k), and the General Accuracy Index (OA). The results showed that the KNN and DT algorithms demonstrated a greater discrimination capacity for most classes. SVM and RF significantly improved class discrimination when using larger samples for training. Merging the classified images significantly improved the classification accuracy, ranging from 71% to 81%. This evaluation made it possible to define sets of ML algorithms sensitive to change detection for mapping and monitoring dynamic environments.