Developments in Landsat Land Cover Classification Methods: A Review

Class-aware feature attention-based semantic segmentation on hyperspectral images

This research explores an innovative approach to segment hyperspectral images. Aclass-aware feature-based attention approach is combined with an enhanced attention-based network, FAttNet is proposed to segment the hyperspectral images semantically. It is introduced to address challenges associated with inaccurate edge segmentation, diverse forms of target inconsistency, and suboptimal predictive efficacy encountered in traditional segmentation networks when applied to semantic segmentation tasks in hyperspectral images. First, the class-aware feature attention procedure is used to improve the extraction and processing of distinct types of semantic information. Subsequently, the spatial attention pyramid is employed in a parallel fashion to improve the correlation between spaces and extract context information from images at different scales. Finally, the segmentation results are refined using the encoder-decoder structure. It enhances precision in delineating distinct land cover patterns. The findings from the experiments demonstrate that FAttNet exhibits superior performance compared to established semantic segmentation networks commonly used. Specifically, on the GaoFen image dataset, FAttNet achieves a higher mean intersection over union (MIoU) of 77.03% and a segmentation accuracy of 87.26% surpassing the performance of the existing network.

Mapping the Green Urban: A Comprehensive Review of Materials and Learning Methods for Green Infrastructure Mapping

Green infrastructure (GI) plays a crucial role in sustainable urban development, but effective mapping and analysis of such features requires a detailed understanding of the materials and state-of-the-art methods. This review presents the current landscape of green infrastructure mapping, focusing on the various sensors and image data, as well as the application of machine learning and deep learning techniques for classification or segmentation tasks. After finding articles with relevant keywords, the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyzes) method was used as a general workflow, but some parts were automated (e.g., screening) by using natural language processing and large language models. In total, this review analyzed 55 papers that included keywords related to GI mapping and provided materials and learning methods (i.e., machine or deep learning) essential for effective green infrastructure mapping. A shift towards deep learning methods can be observed in the mapping of GIs as 33 articles use various deep learning methods, while 22 articles use machine learning methods. In addition, this article presents a novel methodology for automated verification methods, demonstrating their potential effectiveness and highlighting areas for improvement.

Changing pattern of urban landscape and its impact on thermal environment of Lahore; Implications for climate change and sustainable development

Rapid urbanization in Lahore has dramatically transformed land use and land cover (LULC), significantly impacting the city's thermal environment and intensifying climate change and sustainable development challenges. This study aims to examine the changes in the urban landscape of Lahore and their impact on the Urban thermal environment between 1990 and 2020. The previous studies conducted on Lahore lack the application of Geospatial artificial intelligence (GeoAI) to quantify land use and land cover, which is successfully covered in this study. This study analyzes how urban sprawl has driven LULC shifts and assesses their direct impact on Land Surface Temperature (LST) using Geographic Information System (GIS) and remote sensing techniques. Landsat imagery, processed using Google Earth Engine (GEE), was employed for LULC classification and LST calculation, ensuring high accuracy through multi-level change detection and a thorough accuracy assessment. Pearson's correlation was also calculated in this study to assess the impact of decreased green cover on LST. The findings highlight a substantial decrease in green cover, from 1,292.8 km2 in 1990 to 754 km2 in 2020, alongside a marked increase in built-up areas, expanding from 262 km2 to over 550 km2. Additionally, barren land showed significant growth, while water bodies diminished. The spatiotemporal analysis of LST indicates a considerable rise in high-temperature zones, specifically the industrial zones, with areas exceeding 40 °C expanding from 2 km2 to 1,075 km2 over the study period. A strong positive correlation between increased urbanization and rising LST, particularly in areas within a 10 to 40 km radius of the Central Business District (CBD), is evident. The overall accuracy of LULC classification surpassed 94%, with the kappa coefficient above 92%, ensuring the robustness of the results. Future research should focus on evaluating the long-term socioeconomic impacts of urban sprawl and LST increment while developing heat mitigation strategies. Recommendations include adopting sustainable urban planning practices prioritizing green infrastructure, energy-efficient building designs, and policies promoting environmental preservation. This study offers valuable insights for policymakers and underscores the urgency of balancing urban growth with strategies that mitigate thermal stress, combat climate change, and foster sustainable development in Lahore.

Road urban planning sustainability based on remote sensing and satellite dataset: A review

Infrastructural development and urbanization effects have been investigated over the past decades with novel approaches and adaptation strategies. Road network expansions are more useful for the socio-economic development from urban to rural areas where 75 % of the passenger, and goods transportation sectors are influenced by the road. Road infrastructure and urbanization are perpendicular to each other, and this research investigation indicates that the novel approaches and adaptation strategies for road infrastructure and urbanization effects. This study evaluated the trend in the road network and urbanization-related literature from 2010 to 2022 with some measurable keywords. Around 370 pieces of research literature are analysis and around 85 research evaluations for the road network and urbanization-related Land use and land cover (LULC) studies while numerous road network analysis approaches and LULC-related investigations are evaluated in this research. Three major parts road network analysis-related approaches, LULC, and urbanization-related approaches related to road network expansion and urbanization, were investigated. In this work, many research publications' approaches to LULC simulation, kernel density, shortage distance, and picture classification are discussed and assessed. The survey is more valuable for urban planners, future disaster management teams, and administrators to implement the shortage distance analysis, reduction of road accidents, and urbanization effects on the environment.

Analysis of land use land cover change dynamics in Habru District, Amhara Region, Ethiopia

This study examines spatiotemporal changes in Land Use Land Cover (LULC) patterns within Habru District, Amhara Region, Ethiopia, between 1985 and 2021. Employing Landsat imagery and a supervised classification approach, we mapped six LULC classes: water body, settlement area, cultivated land, bare land, shrub land, and forest land. Ground control points (GCP) were obtained through field observations to ensured image accuracy. Complementing the remote sensing analysis, focus group discussions (FGDs) and key informant interviews were conducted to understand local perspectives on LULC changes. The analysis revealed significant LULC transformations over the study period. Between 1985 and 2021, there was a significant increase in cultivated land and settlement areas, which expanded from 12.27 % to 0.23 %-42.12 % and 2.58 % of the total area, respectively. Conversely, shrub and forest lands saw a marked reduction during the same period, declining from 71.52 % to 12.88 %-45.35 % and 7.06 % of the total area, respectively. Water body area showed a minor increase, while bare land exhibited negligible change. Agricultural land and settlement area expansion, population growth, land tenure insecurity, economic challenge, and climatic change were identified as key drivers of observed LULC changes. Our findings underscore the urgency of implementing multifaceted interventions to address these drivers and promote sustainable resource use practices. Such actions are critical for safeguarding Habru District's natural resources and ensuring the long-term viability of ecosystem services.

Fusion of spectral and topographic features for land use mapping using a machine learning framework for a regional scale application

This study investigated the dynamics of land use and land cover (LULC) modelling, mapping, and assessment in the Kegalle District of Sri Lanka, where policy decision-making is crucial in agricultural development where LULC temporal datasets are not readily available. Employing remotely sensed datasets and machine learning algorithms, the work presented here aims to compare the accuracy of three classification approaches in mapping LULC categories across the time in the study area primarily using the Google Earth Engine (GEE). Three classifiers namely random forest (RF), support vector machines (SVM), and classification and regression trees (CART) were used in LULC modelling, mapping, and change analysis. Different combinations of input features were investigated to improve classification performance. Developed models were optimised using the grid search cross-validation (CV) hyperparameter optimisation approach. It was revealed that the RF classifier constantly outstrips SVM and CART in terms of accuracy measures, highlighting its reliability in classifying the LULC. Land cover changes were examined for two periods, from 2001 to 2013 and 2013 to 2022, implying major alterations such as the conversion of rubber and coconut areas to built-up areas and barren lands. For suitable classification with higher accuracy, the study suggests utilising high spatial resolution satellite data, advanced feature selection approaches, and a combination of several spatial and spatial-temporal data sources. The study demonstrated practical applications of derived temporal LULC datasets for land management practices in agricultural development activities in developing nations.

Impacts of climate change on water-related mosquito-borne diseases in temperate regions: A systematic review of literature and meta-analysis

Mosquito-borne diseases are a known tropical phenomenon. This review was conducted to assess the mecha-nisms through which climate change impacts mosquito-borne diseases in temperate regions. Articles were searched from PubMed, Scopus, Web of Science, and Embase databases. Identification criteria were scope (climate change and mosquito-borne diseases), region (temperate), article type (peer-reviewed), publication language (English), and publication years (since 2015). The WWH (who, what, how) framework was applied to develop the research question and thematic analyses identified the mechanisms through which climate change affects mosquito-borne diseases. While temperature ranges for disease transmission vary per mosquito species, all are viable for temperate regions, particularly given projected temperature increases. Zika, chikungunya, and dengue transmission occurs between 18-34 °C (peak at 26-29 °C). West Nile virus establishment occurs at monthly average temperatures between 14-34.3 °C (peak at 23.7-25 °C). Malaria establishment occurs when the consecutive average daily temperatures are above 16 °C until the sum is above 210 °C. The identified mechanisms through which climate change affects the transmission of mosquito-borne diseases in temperate regions include: changes in the development of vectors and pathogens; changes in mosquito habitats; extended transmission seasons; changes in geographic spread; changes in abundance and behaviors of hosts; reduced abundance of mosquito predators; interruptions to control operations; and influence on other non-climate factors. Process and stochastic approaches as well as dynamic and spatial models exist to predict mosquito population dynamics, disease transmission, and climate favorability. Future projections based on the observed relations between climate factors and mosquito-borne diseases suggest that mosquito-borne disease expansion is likely to occur in temperate regions due to climate change. While West Nile virus is already established in some temperate regions, Zika, dengue, chikungunya, and malaria are also likely to become established over time. Moving forward, more research is required to model future risks by incorporating climate, environmental, sociodemographic, and mosquito-related factors under changing climates.

Mapping decadal land cover dynamics in Sicily's coastal regions

Coastal zones are dynamic interfaces shaped by the interplay of Land Cover (LC) and Land Use (LU), influenced by both natural processes and anthropogenic activities. Grasping the historical shifts in land is essential for safeguarding coastal benefits such as defense mechanisms, biodiversity conservation, and recreational spaces, alongside enhancing their management. LC and LU products offer a valuable option for monitoring urban development, vegetation coverage, and dry-beach areas. Herein, we present the first study of the spatiotemporal evolution of LC specifically tailored for coastal zones, using the coast of Sicily as an illustration. We used classified satellite imagery from Landsat and Sentinel missions as input for a semantic segmentation model based on deep neural networks. We trained the model with an extensive dataset of coastal images. Our classification and analysis of coastal LC dynamics from 1988 to 2022 provide insights at a high spatiotemporal resolution. We identified key factors driving urban transformation, underscoring the impact of urban expansion on vegetated areas, and explored its correlation with economic and demographic growth. This study includes a multiscale analysis of coastal changes, encompassing long-term trends and seasonal fluctuations across Sicilian beaches. Our findings can contribute to preserve coastal areas by informing policymaking aimed at sustainable management.

Past and future land use change dynamics: assessing the impact of urban development on agricultural land in the Pantura Jabar region, Indonesia

The conversion of large-scale agricultural land into urban areas poses a significant challenge to achieving national and global food security targets, as outlined in Sustainable Development Goal number 2, which aims to eradicate hunger. Indonesia has experienced a significant decline in rice field areas, with a reduction of approximately 650 thousand hectares within a year (2017-2018), the largest being in Java. Hence, this study aims to examine the impact of urban expansion on agricultural land in the north coast region of West Java Province from 2013 to 2020 and develop a predictive model for 2030 to support sustainable land use planning. The primary methods employed were random forest (RF) analysis using Google Earth Engine, intensity analysis, multilayer perceptron-neural network (MLP-NN), Markov chains-cellular automata (Markov-CA), and stakeholder interviews. The model also evaluated the influence of "distance to tollgates" as a previously unexplored driving factor in existing land use modeling studies. Landsat image classification results using the RF algorithm showed 87-88% accuracy. Cropland has historically been and is projected to remain the primary target for the expansion of built-up areas. Spatial planning irregularities were found in the growth of these areas that adversely affected farmers' socioeconomic and environmental conditions. Evaluation of land use models using MLP-NN and Markov-CA demonstrated an accuracy rate of 86.29-86.23%. The distance to tollgates factor significantly impacts the models, albeit less than population density. The 2030 intervention scenario, which implements a firm policy for sustainable agricultural land use, offers the potential to maintain the predicted cropland loss compared to business as usual.

Remote sensing-based seasonal surface urban heat island analysis in the mining and industrial environment

Cooling spaces have an optimistic influence on surface urban heat islands (SUHI). Blue spaces benefit from balancing the changing climate and heat variations. Because of the rapid deforestation and SUHI increase, the climate is gradually changing in Paschim Bardhhaman, West Bengal state, India. Paschim Bardhhaman has two sectors: specifically, Durgapur is the main industrial centre and Asansol has coal mines. This investigation aims to categorize spatiotemporal variations and seasonal differences in cooling spaces and their influence on SUHI, land use and land cover (LULC), and thermal differences using Landsat datasets for the years 1992, 2004, 2012, and 2022 in summer and winter. The coal mining and industrial range decreased from 10,391.92 (1992) to 3591.1 ha (2022), respectively. Open pit mining distresses fresh water by heavy water uses in ore processing, and mining water was applied to excerpt minerals. Among the two sub-divisions, the blue space amount was higher in Asansol because mining actions were higher in Asansol than in Durgapur. The open vegetation volume has reduced from 46,441.03 (1992) to 25,827.55 ha (2022) and dense vegetation has erased from 7368.02 (1992) to 15,608.56 ha (2022). Dense vegetation improved because of heavy precipitation in those regions. Mostly, Raghunathpur, Saraswatiganja, Bhagabanpur, Bistupur, Paschim Gangaram, Garkilla Kherobari, and Gourbazar have dense vegetation. The outcomes similarly demonstrate that the total built-up part has increased by 8412.82 ha in between 30 years. The built-up zone changes near the southeast and western Paschim Bardhhaman district. Those region needs appropriate attention and planning to survive soon.

Long-Term Effects of Ecological Restoration Projects on Ecosystem Services and Their Spatial Interactions: A Case Study of Hainan Tropical Forest Park in China

Ecological restoration projects aim to comprehensively intervene in damaged or deteriorating ecosystems, restore them, improve the provision of ecosystem services, and achieve harmonious coexistence between humans and nature. Implementing ecological restoration projects leads to continuous changes in land use/land cover. Studying the long-term changes in land use/land cover and their impacts on ecosystem services, as well as the trade-off and synergy between these services, helps evaluate the long-term effectiveness of ecological restoration projects in restoring ecosystems. Therefore, this study analyzes the land use/land cover, and ecosystem services of the Hainan Tropical Forest Park in China to address this. Since 2000, the area has undergone multiple ecological restoration projects, divided roughly into two stages: 2003-2013 and 2013-2021. The InVEST model is used to quantify three essential ecosystem services in mountainous regions (water yield, carbon storage, and soil conservation), and redundancy analysis identifies the primary driving factors influencing their changes. We conducted spatial autocorrelation analysis to examine the interplay among ecosystem services under long-term land use/land cover change. The results indicate a decrease in the total supply of water yield (-5.14%) and carbon storage (-3.21%) in the first phase. However, the second phase shows an improvement in ecosystem services, with an increase in the total supply of water yield (11.45%), carbon storage (27.58%), and soil conservation (21.95%). The redundancy analysis results reveal that land use/land cover are the primary driving factors influencing the changes in ecosystem services. Furthermore, there is a shift in the trade-off and synergy between ecosystem services at different stages, with significant differences in spatial distribution. The findings of this study provide more spatially targeted suggestions for the restoration and management of tropical montane rainforests in the future.

Environmental impact assessment of transportation and land alteration using Earth observational datasets: Comparative study between cities in Asia and Europe

Developments in the transportation field are emerging because of the growing worldwide demand and upgrading requirements. This study measured the transportation development, shortage distance, and decadal land transformation of Kuala Lumpur and Madrid using various remote sensing and GIS approaches. The kernel density estimation (KDE) tool was applied for road and railway density analysis, and hotspot information increased the knowledge about assessable areas. Landsat datasets were used (1991-2021) for land transformation and related analyses. The built-up land increased by 1327.27 and 404.09 km2 in Kuala Lumpur and Madrid, respectively. In the last thirty years, the temperature increased 6.45 °C in Kuala Lumpur and 4.15 °C in Madrid owing to urban expansion and road construction. Chamberi, Retiro, Moratalaz, Salama, Wangsa Maju, Titiwangsa, Bukit Bintang, and Seputeh have very high road densities. KDE measurements showed that the road densities in Kuala Lumpur (4498.34) and Madrid (9099.15) were high in the central parts of the city, and the railway densities were 348.872 and 2197.87, respectively. The observed P values were 0.99 and 0.96 for traffic signals and 0.98 and 0.99 for bus stops, respectively. The information provided by this study can support local planners, administrators, scientists, and researchers in understanding the global transportation issues that require implementation strategies for ensuring sustainable livelihoods.

Monitoring land cover changes and droughts using statistical analysis and multi-sensor remote sensing data

Climate change is of paramount importance for ecosystems, especially in arid and semi-arid areas. The major aim of the current study is to monitor vegetation and land use changes, and run drought assessment using field and satellite data. The main precipitation proportions in the studied region are influenced by the Westerlies, meaning that any variations in these precipitation systems significantly impact the region. The utilized data entailed MODIS images for 16- and 8-day intervals between 2000 and 2013, TM and OLI sensor images recorded in 1985 and 2013, precipitation network data of TRMM satellite between 2000 and 2013, and synoptic data 32-year period. The Mann-Kendall (MK) test was used to monitor temporal changes in meteorological station data in annual and seasonal scales. The results indicated that there was a downward trend in 50% of the meteorological stations in the annual scale. This falling trend was statistically significant at the level of 95%. At the end, drought was assessed using PCI, APCI, VSWI, and NVSWI. The results showed that vegetation, forest, pasture, and agriculture areas recorded the strongest correlations with initial precipitation at the beginning of the study. Based on interactions among various factors influencing vegetation indices, reduction in green vegetation, especially the area of oak forests in the studied period, is around 95,744 hectares, which is attributed to lower precipitation rate. Increasing of agricultural land and water zones during the studied years is the result of human management and depends on how surface and underground water resources are exploited.

Introducing ARTMO's Machine-Learning Classification Algorithms Toolbox: Application to Plant-Type Detection in a Semi-Steppe Iranian Landscape

Accurate plant-type (PT) detection forms an important basis for sustainable land management maintaining biodiversity and ecosystem services. In this sense, Sentinel-2 satellite images of the Copernicus program offer spatial, spectral, temporal, and radiometric characteristics with great potential for mapping and monitoring PTs. In addition, the selection of a best-performing algorithm needs to be considered for obtaining PT classification as accurate as possible. To date, no freely downloadable toolbox exists that brings the diversity of the latest supervised machine-learning classification algorithms (MLCAs) together into a single intuitive user-friendly graphical user interface (GUI). To fill this gap and to facilitate and automate the usage of MLCAs, here we present a novel GUI software package that allows systematically training, validating, and applying pixel-based MLCA models to remote sensing imagery. The so-called MLCA toolbox has been integrated within ARTMO's software framework developed in Matlab which implements most of the state-of-the-art methods in the machine learning community. To demonstrate its utility, we chose a heterogeneous case study scene, a landscape in Southwest Iran to map PTs. In this area, four main PTs were identified, consisting of shrub land, grass land, semi-shrub land, and shrub land-grass land vegetation. Having developed 21 MLCAs using the same training and validation, datasets led to varying accuracy results. Gaussian process classifier (GPC) was validated as the top-performing classifier, with an overall accuracy (OA) of 90%. GPC follows a Laplace approximation to the Gaussian likelihood under the supervised classification framework, emerging as a very competitive alternative to common MLCAs. Random forests resulted in the second-best performance with an OA of 86%. Two other types of ensemble-learning algorithms, i.e., tree-ensemble learning (bagging) and decision tree (with error-correcting output codes), yielded an OA of 83% and 82%, respectively. Following, thirteen classifiers reported OA between 70% and 80%, and the remaining four classifiers reported an OA below 70%. We conclude that GPC substantially outperformed all classifiers, and thus, provides enormous potential for the classification of a diversity of land-cover types. In addition, its probabilistic formulation provides valuable band ranking information, as well as associated predictive variance at a pixel level. Nevertheless, as these are supervised (data-driven) classifiers, performances depend on the entered training data, meaning that an assessment of all MLCAs is crucial for any application. Our analysis demonstrated the efficacy of ARTMO's MLCA toolbox for an automated evaluation of the classifiers and subsequent thematic mapping.

Integrated model for land-use transformation analysis based on multi-layer perception neural network and agent-based model

The efficacy of land-use changes on aquatic ecosystems has been extensively studied in recent decades. Water resource management needs to understand the relationship between land-use change patterns and water quality, especially in urban areas. Hence, recognizing spatial-temporal changes in land use is required for sustainable development and proper water resource management. This research has developed an integrated model based on agent-based model (ABM) and multi-layer perceptron (MLP) neural network technique to predict the future land-use transformation tested on the North Ahvaz watershed, Iran. Random forest-supervised classification technique was applied to derive the land-use maps using Landsat 1989, 2004, and 2019 images in the Google Earth Engine (GEE) platform. The overall accuracy of classified land-use images was 0.82, 0.81, and 0.84, respectively, with the kappa coefficient of 0.74, 0.72, and 0.78. Land-use change analysis and generating transition potential maps were carried out in land change modeler (LCM) through MLP based on seven driving factors. Then, the land-use map for 2019 (for validation) and 2040 was simulated using the transition potential map and an agent-based approach. The ABM scenario was farmers' and urban landowners' decisions to convert undeveloped and unprotected lands to residential lands. The results showed that residential areas and pasture lands would grow by 67.96 km² and 64.63 km², and agricultural and barren lands would degrade about 84.19 km² and 47.98 km² during 2019-2040, respectively. Predicting land-use change through the integrated MLP-ABM model may be used to evaluate the effects of land-use change coming out of human decision-making.

Loss of Relict Oak Forests along Coastal Louisiana: A Multiyear Analysis Using Google Earth Engine

Coastal forests along the southeastern Gulf of Mexico are known to be diminishing at an alarming rate. The live-oak dominant chenier forests of southeast Louisiana are amongst those exhibiting the steepest declines. The remnant stands have experienced numerous hurricanes and intense storm events in recent years, calling into question the current status and immediate future of this imperiled natural resource. Despite their noted ecological and physiographic importance, there is a lack within national geographic data repositories of accurate representations of forest loss and wetland extent for this region. Supervised machine learning algorithms in the Google Earth Engine were used to classify and process high-resolution National Agricultural Image Product (NAIP) datasets to create accurate (>90%) tree cover maps of the Louisiana Chenier Plains in Cameron and Vermilion Parishes. Data from three different years (2003, 2007, and 2019) were used to map 2302 km2 along the southwestern coast of Louisiana. According to the analyses, there was a 35.73% loss of forest cover in this region between 2003 and 2019. A majority of the land-use change was from tree cover to saltmarsh, with losses in pastoral land also documented. We found variable rates of loss with respect to elevation. Forest cover losses corresponded strongly to rises in mean sea level. These findings deliver a baseline understanding of the rate of forest loss in this region, highlighting the reduction and potentially the eventual extirpation of this imperiled ecosystem.

Recent Recovery of the World’s Rarest Primate Is Not Directly Linked to Increasing Habitat Quality

Due to habitat loss and hunting, the Hainan gibbon (Nomascus hainanus), the world’s rarest primate, was reduced to only two social groups and seven known individuals in 1978. Following the establishment of Bawangling National Natural Reserve (BNNR), gibbon forest habitat increased within this landscape from 56 km2 in 1980 to 300 km2, and the species had increased to five groups and 35 individuals by 2021. It is important to assess whether the large increase in habitat area was responsible for gibbon population increase, or whether gibbon recovery was associated with other factors. Here we use a 21-year longitudinal dataset of Hainan gibbon population change and habitat change, combined with vegetation survey plot data for 2021, to establish an accurate distribution baseline for natural tropical broadleaf forest across the BNNR landscape from 400 to 1300 m (the elevational range of gibbons at BNNR) and within the home range for each of the five Hainan gibbon social groups. We then utilized Landsat time-series images and analysis to compute non-linear causal relationships between forest dynamics and gibbon population growth from 2000 to 2021, both across BNNR and within each gibbon group home range. Metrics of forest dynamics include change in total forest area and forest fragmentation, and metrics of gibbon population dynamics include variation in total number of individuals for the entire population and within each social group, and variation in total number of groups. Our results demonstrate that overall gibbon population growth shows a positive relationship with improved habitat quality, with a one-year time lag of population response. However, changes in numbers of individuals within social groups do not show a similar relationship with improving habitat quality, suggesting that increasing forest cover and connectivity within the BNNR landscape are not direct determinants of Hainan gibbon recovery and that other environmental and/or anthropogenic factors are likely to be involved.

Comparing CNNs and Random Forests for Landsat Image Segmentation Trained on a Large Proxy Land Cover Dataset

Land cover mapping from satellite images has progressed from visual and statistical approaches to Random Forests (RFs) and, more recently, advanced image recognition techniques such as convolutional neural networks (CNNs). CNNs have a conceptual benefit over RFs in recognising spatial feature context, but potentially at the cost of reduced spatial detail. We tested the use of CNNs for improved land cover mapping based on Landsat data, compared with RFs, for a study area of approximately 500 km × 500 km in southeastern Australia. Landsat 8 geomedian composite surface reflectances were available for 2018. Label data were a simple nine-member land cover classification derived from reference land use mapping (Catchment Scale Land Use of Australia—CLUM), and further enhanced by using custom forest extent mapping (Forests of Australia). Experiments were undertaken testing U-Net CNN for segmentation of Landsat 8 geomedian imagery to determine the optimal combination of input Landsat 8 bands. The results were compared with those from a simple autoencoder as well as an RF model. Segmentation test results for the best performing U-Net CNN models produced an overall accuracy of 79% and weighted-mean F1 score of 77% (9 band input) or 76% (6 band input) for a simple nine-member land cover classification, compared with 73% and 68% (6 band input), respectively, for the best RF model. We conclude that U-Net CNN models can generate annual land cover maps with good accuracy from proxy training data, and can also be used for quality control or improvement of existing land cover products.

Land Change Science and the STEPLand Framework: An Assessment of Its Progress

This contribution assesses a new term that is proposed to be established within Land Change Science: Spatio-TEmporal Patterns of Land (‘STEPLand’). It refers to a specific workflow for analyzing land-use/land cover (LUC) patterns, identifying and modeling driving forces of LUC changes, assessing socio-environmental consequences, and contributing to defining future scenarios of land transformations. In this article, we define this framework based on a comprehensive meta-analysis of 250 selected articles published in international scientific journals from 2000 to 2019. The empirical results demonstrate that STEPLand is a consolidated protocol applied globally, and the large diversity of journals, disciplines, and countries involved shows that it is becoming ubiquitous. In this paper, the main characteristics of STEPLand are provided and discussed, demonstrating that the operational procedure can facilitate the interaction among researchers from different fields, and communication between researchers and policy makers.

Machine Learning Approaches for Developing Land Cover Mapping

In remote sensing data processing, cover classification on decimeter-level data is a well-studied but tough subject that has been well-documented. The majority of currently existent works make use of orthographic photographs or orthophotos and digital surface models that go with them (DSMs). Urban land cover classification plays a significant role in the field of remote sensing to enhance the quality of different applications including environment protection, sustainable development, and resource management and planning. Novelty of the research done in this area is focused on extracting features from high-resolution satellite images to be used in the classification process. However, it is well known in machine learning literature that some of the extracted features are irrelevant to the classification process with a negative or no effect on its accuracy. In this work, a genetic algorithm-based feature selection approach is used to enhance the performance of urban land cover classification. Neural networks (NNs) and random forest (RF) classifiers were used to evaluate the proposed approach on a recent urban land cover dataset of nine different classes. Experimental results show that the proposed approach achieved better performance with RF classifier using only 27% of the features. The random forest tree has achieved highest accuracy 84.27%; it is concluded that the RF algorithm is an appropriate algorithm for classifying cover land.

Seeded Classification of Satellite Image Time Series with Lower-Bounded Dynamic Time Warping

Satellite Image Time Series (SITS) record the continuous temporal behavior of land cover types and thus provide a new perspective for finer-grained land cover classification compared with the usual spectral and spatial information contained in a static image. In addition, SITS data is becoming more accessible in recent years due to newly launched satellites and accumulated historical data. However, the lack of labeled training samples limits the exploration of SITS data, especially with sophisticated methods. Even with a straightforward classifier, such as k-nearest neighbor, the accuracy and efficiency of the SITS similarity measure is also a pending problem. In this paper, we propose SKNN-LB-DTW, a seeded SITS classification method based on lower-bounded Dynamic Time Warping (DTW). The word “seeded” indicates that only a few labeled samples are required, and this is not only because of the lack of labeled samples but also because of our aim to explore the rich information contained in SITS, rather than letting training samples dominate the classification results. We use a combination of cascading lower bounds and early abandoning of DTW as an accurate yet efficient similarity measure for large scale tasks. The experimental results on two real SITS datasets demonstrate the utility of the proposed SKNN-LB-DTW, which could become an effective solution for SITS classification when the amount of unlabeled SITS data far exceeds the labeled data.

Limiting the Collection of Ground Truth Data for Land Use and Land Cover Maps with Machine Learning Algorithms

Land use and land cover (LULC) classification maps help understand the state and trends of agricultural production and provide insights for applications in environmental monitoring. One of the major downfalls of the LULC technique is inherently linked to its need for ground truth data to cross-validate maps. This paper aimed at evaluating the efficiency of machine learning (ML) in limiting the use of ground truth data for LULC maps. This was accomplished by (1) extracting reliable LULC information from Sentinel-2 and Landsat-8 s images, (2) generating remote sensing indices used to train ML algorithms, and (3) comparing the results with ground truth data. The remote sensing indices that were tested include the difference vegetation index (DVI), the normalized difference vegetation index (NDVI), the normalized built-up index (NDBI), the urban index (UI), and the normalized bare land index (NBLI). Extracted vegetation indices were evaluated on three ML algorithms, namely, random forest (RF), k-nearest neighbour (K-NN), and k dimensional-tree (KD-Tree). The accuracy of these algorithms was assessed with standard statistical measures and ground truth data randomly collected in Prince Edward Island, Canada. Results showed that high kappa coefficient values were achieved by K-NN (82% and 74%), KD-Tree (80% and 78%), and RF (83% and 73%) for Sentinel-2A and Landsat-8 imagery, respectively. RF was a better classifier than K-NN and KD-Tree and had the highest overall accuracy with Sentinel-2A satellite images (92%). This approach provides the basis for limiting the collection of ground truth data and thus reduces the labour cost, time, and resources needed to collect ground truth data for LULC maps.

Classification of airborne multispectral imagery to quantify common vole impacts on an agricultural field

BACKGROUND

The common vole (Microtus arvalis) is a very destructive agricultural pest. Particularly in Europe, its monitoring is essential not only for adequate management and outbreak forecasting, but also for accurately determining the vole's impact on affected fields. In this study, several alternatives for estimating the damage to alfalfa fields by voles through unmanned vehicle systems (UASs) and multispectral cameras are presented. Currently, both the farmers and agencies involved in the integrated pest management (IPM) programs of voles do not have sufficiently precise methods for accurate assessments of the real impact to crops.

RESULTS

Overall, the four multispectral classification methods presented showed similar performances. However, the normalized difference vegetation index (NDVI)-based segmentation exhibited the most accurate and reliable appraisal of the affected areas. Nevertheless, it must be noted that the simplest method, which was based on an automatic classification, provided results similar to those obtained by more complex methods. In addition, a significant direct relationship was found between the number of active burrows and damage to the alfalfa canopy.

CONCLUSION

Unmanned vehicle systems, combined with multispectral imagery classification, are an effective and easily transferable methodology for the assessment and monitoring of common vole damage to agricultural plots. This combination of methods facilitates decision-making processes for IPM control strategies against this pest. © 2022 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Urban Land Use and Land Cover Change Analysis Using Random Forest Classification of Landsat Time Series

Efficient implementation of remote sensing image classification can facilitate the extraction of spatiotemporal information for land use and land cover (LULC) classification. Mapping LULC change can pave the way to investigate the impacts of different socioeconomic and environmental factors on the Earth’s surface. This study presents an algorithm that uses Landsat time-series data to analyze LULC change. We applied the Random Forest (RF) classifier, a robust classification method, in the Google Earth Engine (GEE) using imagery from Landsat 5, 7, and 8 as inputs for the 1985 to 2019 period. We also explored the performance of the pan-sharpening algorithm on Landsat bands besides the impact of different image compositions to produce a high-quality LULC map. We used a statistical pan-sharpening algorithm to increase multispectral Landsat bands’ (Landsat 7–9) spatial resolution from 30 m to 15 m. In addition, we checked the impact of different image compositions based on several spectral indices and other auxiliary data such as digital elevation model (DEM) and land surface temperature (LST) on final classification accuracy based on several spectral indices and other auxiliary data on final classification accuracy. We compared the classification result of our proposed method and the Copernicus Global Land Cover Layers (CGLCL) map to verify the algorithm. The results show that: (1) Using pan-sharpened top-of-atmosphere (TOA) Landsat products can produce more accurate results for classification instead of using surface reflectance (SR) alone; (2) LST and DEM are essential features in classification, and using them can increase final accuracy; (3) the proposed algorithm produced higher accuracy (94.438% overall accuracy (OA), 0.93 for Kappa, and 0.93 for F1-score) than CGLCL map (84.4% OA, 0.79 for Kappa, and 0.50 for F1-score) in 2019; (4) the total agreement between the classification results and the test data exceeds 90% (93.37–97.6%), 0.9 (0.91–0.96), and 0.85 (0.86–0.95) for OA, Kappa values, and F1-score, respectively, which is acceptable in both overall and Kappa accuracy. Moreover, we provide a code repository that allows classifying Landsat 4, 5, 7, and 8 within GEE. This method can be quickly and easily applied to other regions of interest for LULC mapping.

Multi-Temporal Landsat-8 Images for Retrieval and Broad Scale Mapping of Soil Copper Concentration Using Empirical Models

Mapping soil heavy metal concentration using machine learning models based on readily available satellite remote sensing images is highly desirable. Accurate mapping relies on appropriate data, feature extraction, and model selection. To this end, a data processing pipeline for soil copper (Cu) concentration estimation has been designed. First, instead of using single Landsat scenes, the utilization of multiple Landsat scenes of the same location over time is considered. Second, to generate a preferred feature set as input to a regression model, a number of feature extraction methods are motivated and compared. Third, to find a preferred regression model, a variety of approaches are implemented and compared for accuracy. In this research, 11 Landsat-8 images from 2013 to 2017 of Gulin County, Sichuan China, and 138 soil samples with lab-measured Cu concentrations collected from the area in 2015 are used. A variety a metrics under cross-validation are used for comparison. The results indicate that multi-temporal images increase accuracy compared to single Landsat images. The preferred feature extraction varies based on the regression model used; however, the best results are obtained using support vector regression and the original data. The final soil Cu map generated using the recommended data processing pipeline shows a consistent spatial pattern with a ground-truth land cover classification map. These results indicate that machine learning has the ability to perform large-scale soil heavy metal mapping from widely available satellite remote sensing images.

A Block Shuffle Network with Superpixel Optimization for Landsat Image Semantic Segmentation

In recent years, with the development of deep learning in remotely sensed big data, semantic segmentation has been widely used in large-scale landcover classification. Landsat imagery has the advantages of wide coverage, easy acquisition, and good quality. However, there are two significant challenges for the semantic segmentation of mid-resolution remote sensing images: the insufficient feature extraction capability of deep convolutional neural network (DCNN); low edge contour accuracy. In this paper, we propose a block shuffle module to enhance the feature extraction capability of DCNN, a differentiable superpixel branch to optimize the feature of small objects and the accuracy of edge contours, and a self-boosting method to fuse semantic information and edge contour information to further optimize the fine-grained edge contour. We label three sets of Landsat landcover classification datasets, and achieved an overall accuracy of 86.3%, 83.2%, and 73.4% on the three datasets, respectively. Compared with other mainstream semantic segmentation networks, our proposed block shuffle network achieves state-of-the-art performance, and has good generalization ability.

Evaluating the Impact of the Influx of Syrian Refugees on Land Use/Land Cover Change in Irbid District, Northwestern Jordan

The refugee movement creates urban and environmental pressures at their destination locations. This pressure often presents in the form of Land Use/Land Cover (LULC) change. This study seeks to understand the impact of the Syrian refugees’ influence on changing the urban and agricultural land dynamics in Irbid district in northwestern Jordan from 1985 to 2021, including the period of the civil war in Syria, using Landsat Thematic Mapper (TM) images for the years 1985 and 2004, and the Landsat-8 Operational Land Imager (OLI) for the years 2013 and 2021. The Google Earth Engine (GEE) platform was used to conduct all image processing and perform calculations and classification analysis using the Random Forest (RF) approach. The study of the classified images compared LULC before and during the Syrian crisis using images from 1985, 2004, 2013, and 2021. The results show that the urban area increased. In parallel, agricultural land increased. During the Syrian refugee crisis, agriculture became a significant livelihood activity for Syrian refugees. In summary, the movement of the refugees to Irbid district caused an increased demand for land and housing, which accelerated the building and construction process.

Groundwater potential mapping of Tawi River basin of Jammu District, India, using geospatial techniques

The demand for water is increasing around the world due to population growth, urbanization, industrialization, etc., which is making groundwater a vital natural resource for meeting the growing demand for water. According to the central groundwater report, Jammu district has adequate groundwater potential (GWP) and comes under the safe category. However, the GWP has not been fully utilized, thereby leading to a water shortage in the district. Therefore, this study has been designed to examine the GWP zones in the Tawi River basin of Jammu district using geospatial techniques. For this, several GWP conditioning parameters, such as elevation, slope, geology, geomorphology, rainfall, soil, land use/land cover, topographic wetness index (TWI), drainage density, lineament density, roughness, and curvature, were utilized. The analytical hierarchy process (AHP) technique was used to evaluate the weights of the selected criteria after a pair-wise comparison of each criterion with the rest of the criteria. The result showed that the high GWP zone accounts for 21.98% of the area, the moderate zone covers an area of 40.54%, the low GWP area accounts for about 34.91%, and only 2.57% of the area lies under the very low GWP zone. The validation of GWP zones using 25 monitoring wells showed an accuracy of 80% in GWP modeling. The findings of this study may be utilized in meeting the rising demand for water in the region.

Can a Hierarchical Classification of Sentinel-2 Data Improve Land Cover Mapping?

Monitoring of land cover plays an important role in effective environmental management, assessment of natural resources, environmental protection, urban planning and sustainable development. Increasing demand for accurate and repeatable information on land cover and land cover changes causes rapid development of the advanced, machine learning algorithms dedicated to land cover mapping using satellite images. Free and open access to Sentinel-2 data, characterized with high spatial and temporal resolution, increased the potential to map and to monitor land surface with high accuracy and frequency. Despite a considerable number of approaches towards land cover classification based on satellite data, there is still a challenge to clearly separate complex land cover classes, for example grasslands, arable land and wetlands. The aim of this study is to examine, whether a hierarchal classification of Sentinel-2 data can improve the accuracy of land cover mapping and delineation of complex land cover classes. The study is conducted in the Lodz Province, in central Poland. The pixel-based land cover classification is carried out using the machine learning Random Forest (RF) algorithm, based on a time series of Sentinel-2 imagery acquired in 2020. The following nine land cover classes are mapped: sealed surfaces, woodland broadleaved, woodland coniferous, shrubs, permanent herbaceous (grassy cover), periodically herbaceous (i.e., arable land), mosses, non-vegetated (bare soil) and water bodies. The land cover classification is conducted following two approaches: (1) flat, where all land cover classes are classified together, and (2) hierarchical, where the stratification is applied to first separate the most stable land cover classes and then classifying the most problematic once. The national databases served as the source of the reference sampling plots for the classification process. The process of selection and verification of the reference sampling plots is performed automatically. To assess the stability of the classification models the classification processes are performed iteratively. The results of this study confirmed that the hierarchical approach gave more accurate results compared to the commonly used flat approach. The median of the overall accuracy (OA) of the hierarchical classification was higher by 3–9 percentage points compared to the flat one. Of interest, the OA of the hierarchical classification reached 0.93–0.99, whereas the flat approach reached 0.90. Individual classes are also better classified in the hierarchical approach.

Land Cover Changes in Selected Areas Next to Lagoons Located on the Southern Coast of the Baltic Sea, 1984–2021

The aim of the study is the evaluation of land cover changes in selected areas next to three lagoons (the Curonian Lagoon, the Vistula Lagoon and the Szczecin Lagoon) located on the southern coast of the Baltic Sea (in Lithuania, Russia, Poland and Germany) from 1984 to 2021. The changes are evaluated using multispectral (visible light—RGB and near infrared—NIR) satellite images from the Landsat 5 and Sentinel-2 sensors. Due to their high importance for ecosystem services, two main land cover types are evaluated, i.e., forest area and inland water reservoirs. The classification of the images is performed using a random forest algorithm. Areas of water bodies and forests are evaluated for the years 1984 and 2021. During period 1984–2021, positive changes in land cover are observed in all three regions included in the study. In almost all parts, with the exception of the Polish part of the area located next to the Szczecin Lagoon, of these regions, an increase in forest area is observed. The increase ranges from 0.1% (Poland, area next to the Vistula Lagoon) to 1.2% (Germany, area next to the Szczecin Lagoon). The area of inland water reservoirs has not changed significantly in the long term. Despite the global warming, no reduction in the area of these water reservoirs is observed, even new seminatural reservoirs have been created in some parts of the study area.

Mapping native and non-native vegetation in the Brazilian Cerrado using freely available satellite products

Native vegetation across the Brazilian Cerrado is highly heterogeneous and biodiverse and provides important ecosystem services, including carbon and water balance regulation, however, land-use changes have been extensive. Conservation and restoration of native vegetation is essential and could be facilitated by detailed landcover maps. Here, across a large case study region in Goiás State, Brazil (1.1 Mha), we produced physiognomy level maps of native vegetation (n = 8) and other landcover types (n = 5). Seven different classification schemes using different combinations of input satellite imagery were used, with a Random Forest classifier and 2-stage approach implemented within Google Earth Engine. Overall classification accuracies ranged from 88.6-92.6% for native and non-native vegetation at the formation level (stage-1), and 70.7-77.9% for native vegetation at the physiognomy level (stage-2), across the seven different classifications schemes. The differences in classification accuracy resulting from varying the input imagery combination and quality control procedures used were small. However, a combination of seasonal Sentinel-1 (C-band synthetic aperture radar) and Sentinel-2 (surface reflectance) imagery resulted in the most accurate classification at a spatial resolution of 20 m. Classification accuracies when using Landsat-8 imagery were marginally lower, but still reasonable. Quality control procedures that account for vegetation burning when selecting vegetation reference data may also improve classification accuracy for some native vegetation types. Detailed landcover maps, produced using freely available satellite imagery and upscalable techniques, will be important tools for understanding vegetation functioning at the landscape scale and for implementing restoration projects.

Assessment of bank erosion and its impact on land use and land cover dynamics of Mahananda River basin (Upper) in the Sub-Himalayan North Bengal, India

Bank erosion is the predominant character of River Mahananda in the Sub-Himalayan North Bengal. The present study aims to identify the bank erosion mechanism as well as the impact of river bank erosion on land use and land cover (LULC) dynamics of the study area. Survey of India (SOI) topographical map 78 B/5 (1975) and satellite imageries for the temporal year of 1991 and 2019 from USGS have been used for the study. For the assessment of bank erosion process Bank erosion hazard index (BEHI) model has been adopted here. The channel migration has been delineated by the superimposition of temporal bank lines extracted from the temporal satellite imageries. LULC analysis has been carried out through the supervised classification technique using remote sensing and GIS tools. Form the assessment of BEHI it can be visualized that the scores have been ranging from 30.75 to 44.30 which indicates high to very high vulnerable areas under fluvial erosion. The channel migration for the temporal period from 1991 to 2019 is ranging from 7.72 to 411.16 m along the studied reach which reflects the high erosion effectiveness. From LULC classes it has been assessed that settled or built-up areas have been increased and the water body is gradually decreased overall in the study area. The study resulted that the river bank erosion has its direct impact on land use of the studied area. In the study vulnerable sites to fluvial erosion have been delineated and unplanned land use can be managed through sustainable way.

An Object-Based Markov Random Field with Partition-Global Alternately Updated for Semantic Segmentation of High Spatial Resolution Remote Sensing Image

The Markov random field (MRF) method is widely used in remote sensing image semantic segmentation because of its excellent spatial (relationship description) ability. However, there are some targets that are relatively small and sparsely distributed in the entire image, which makes it easy to misclassify these pixels into different classes. To solve this problem, this paper proposes an object-based Markov random field method with partition-global alternately updated (OMRF-PGAU). First, four partition images are constructed based on the original image, they overlap with each other and can be reconstructed into the original image; the number of categories and region granularity for these partition images are set. Then, the MRF model is built on the partition images and the original image, their segmentations are alternately updated. The update path adopts a circular path, and the correlation assumption is adopted to establish the connection between the label fields of partition images and the original image. Finally, the relationship between each label field is constantly updated, and the final segmentation result is output after the segmentation has converged. Experiments on texture images and different remote sensing image datasets show that the proposed OMRF-PGAU algorithm has a better segmentation performance than other selected state-of-the-art MRF-based methods.

Mapping Land Use Land Cover Changes and Their Determinants in the Context of a Massive Free Labour Mobilisation Campaign: Evidence from South Wollo, Ethiopia

Northern Ethiopia is characterised by fragile mountain eco-systems that are highly susceptible to land degradation, impacting food security and livelihoods. This study appraises Land Use Land Cover Changes (LULCC) and their determinants from 2000 to 2020 in Dessie Zuria and Kutaber Woredas. It explores the LULCC and the key anthropogenic drivers of the change over the past 20 years through a mix of satellite imagery and a survey of local understandings. Six land use types (agriculture, forest, area closure, grazing, settlement and bare land) were mapped from satellite imagery that was acquired from Landsat 7 for the years 2000, 2005, and 2010, and Landsat 8 and OLI multispectral imageries for the years 2015 and 2020 with a spectral resolution of 30-m obtained from USGS. The results showed that agricultural land increased from 29.68% in 2000 to 35.77% in 2020.Furthermore, settlement and grazing lands enlarged from 5.95% and 6.04%, respectively, to 8.31% and 6.35% during the same period, while bare land increased from 9.89% to 10.92% in 2020. On the contrary, forest and area closure decreased from 18.45% and 29.99% to 17.8% and 17.38%, respectively. Meanwhile, population growth, unrestricted grazing, losing a sense of ownership of protected area closures and forests, lack of cooperation, using the free labour mobilisation schemes for government-induced agendas, weak enforcement of laws and bylaws, and engaging farmers for extended days on the campaign were prominent determinants of the changes. This research has implications for development actors across land management and food security towards implementing sustainable land management in the area and beyond.

Mapping the spatial distribution of the dengue vector Aedes aegypti and predicting its abundance in northeastern Thailand using machine-learning approach

BACKGROUND

Mapping the spatial distribution of the dengue vector Aedes (Ae.) aegypti and accurately predicting its abundance are crucial for designing effective vector control strategies and early warning tools for dengue epidemic prevention. Socio-ecological and landscape factors influence Ae. aegypti abundance. Therefore, we aimed to map the spatial distribution of female adult Ae. aegypti and predict its abundance in northeastern Thailand based on socioeconomic, climate change, and dengue knowledge, attitude and practices (KAP) and/or landscape factors using machine learning (ML)-based system.

METHOD

A total of 1066 females adult Ae. aegypti were collected from four villages in northeastern Thailand during January-December 2019. Information on household socioeconomics, KAP regarding climate change and dengue, and satellite-based landscape data were also acquired. Geographic information systems (GIS) were used to map the household-based spatial distribution of female adult Ae. aegypti abundance (high/low). Five popular supervised learning models, logistic regression (LR), support vector machine (SVM), k-nearest neighbor (kNN), artificial neural network (ANN), and random forest (RF), were used to predict females adult Ae. aegypti abundance (high/low). The predictive accuracy of each modeling technique was calculated and evaluated. Important variables for predicting female adult Ae. aegypti abundance were also identified using the best-fitted model.

RESULTS

Urban areas had higher abundance of female adult Ae. aegypti compared to rural areas. Overall, study respondents in both urban and rural areas had inadequate KAP regarding climate change and dengue. The average landscape factors per household in urban areas were rice crop (47.4%), natural tree cover (17.8%), built-up area (13.2%), permanent wetlands (21.2%), and rubber plantation (0%), and the corresponding figures for rural areas were 12.1, 2.0, 38.7, 40.1 and 0.1% respectively. Among all assessed models, RF showed the best prediction performance (socioeconomics: area under curve, AUC = 0.93, classification accuracy, CA = 0.86, F1 score = 0.85; KAP: AUC = 0.95, CA = 0.92, F1 = 0.90; landscape: AUC = 0.96, CA = 0.89, F1 = 0.87) for female adult Ae. aegypti abundance. The combined influences of all factors further improved the predictive accuracy in RF model (socioeconomics + KAP + landscape: AUC = 0.99, CA = 0.96 and F1 = 0.95). Dengue prevention practices were shown to be the most important predictor in the RF model for female adult Ae. aegypti abundance in northeastern Thailand.

CONCLUSION

The RF model is more suitable for the prediction of Ae. aegypti abundance in northeastern Thailand. Our study exemplifies that the application of GIS and machine learning systems has significant potential for understanding the spatial distribution of dengue vectors and predicting its abundance. The study findings might help optimize vector control strategies, future mosquito suppression, prediction and control strategies of epidemic arboviral diseases (dengue, chikungunya, and Zika). Such strategies can be incorporated into One Health approaches applying transdisciplinary approaches considering human-vector and agro-environmental interrelationships.

Assessment of Land Cover Dynamics and Drivers of Urban Expansion Using Geospatial and Logistic Regression Approach in Wa Municipality, Ghana

The current trends of land use dynamics have revealed a significant transformation of settlement spaces. In the Wa Municipality of Ghana, the changes in land use and land cover are inspired by a plethora of driving forces. In this study, we assessed the geo-physical drivers of settlement expansion under land use dynamics in the Wa Municipality of Ghana. The study employed geospatial and remote sensing tools to map and analyse the spatio-temporal dynamics of the landscape, using Landsat satellite imageries: thematic mapper (TM), enhanced thematic mapper (ETM) and operational land imager (OLI) from 1990 to 2020. The study employed a binomial logistic regression model to statistically assess the geo-physical drivers of settlement expansion. Random forest (RF)–supervised classification based on spatio-temporal analyses generated relatively higher classification accuracies, with overall accuracy ranging from 89.33% to 93.3%. Urban expansion for the last three decades was prominent, as the period from 1990 to 2001 gained 11.44 km2 landmass of settlement, while there was 11.30 km2 gained from 2001 to 2010, and 29.44 km2 gained from 2010 to 2020. Out of the independent variables assessed, the distance to existing settlements, distance to river, and distance to primary, tertiary and unclassified roads were responsible for urban expansion.

Evaluation of Riparian Tree Cover and Shading in the Chauga River Watershed Using LiDAR and Deep Learning Land Cover Classification

River systems face negative impacts from development and removal of riparian vegetation that provide critical shading in the face of climate change. This study used supervised deep learning to accurately classify the land cover, including shading, of the Chauga River watershed, located in Oconee County, South Carolina, for 2011 and 2019. The study examined the land cover differences along the Chauga River and its tributaries, inside and outside the Sumter National Forest. LiDAR data were incorporated in solar radiation calculations for the Chauga River inside and outside the National Forest. The deep learning classifications produced land cover maps with high overall accuracy (97.09% for 2011; 97.58% for 2019). The most significant difference in land cover was in tree cover in the 50 m buffer of the tributaries inside the National Forest compared to the tributaries outside the National Forest (2011: 95.39% vs. 81.84%, 2019: 92.86% vs. 82.06%). The solar radiation calculations also confirmed a difference between the area inside and outside the National Forest, with the mean temperature being greater outside the protected area (outside: 455.845 WH/m2; inside: 416,770 WH/m2). This study suggests that anthropogenic influence in the Chauga River watershed is greater in the areas outside the Sumter National Forest, which could cause damage to the river ecosystem if left unchecked in the future as development pressures increase. This study demonstrates the accurate application of deep learning for high-resolution classification of river shading combined with the use of LiDAR data to estimate solar radiation reaching the Chauga River. Techniques to monitor riparian zones and shading at high spatial resolutions are critical for the mitigation of the negative impacts of warming climates on aquatic ecosystems.

Determine the Land-Use Land-Cover Changes, Urban Expansion and Their Driving Factors for Sustainable Development in Gazipur Bangladesh

At present, urbanization is a very common phenomenon around the world, especially in developing countries, and has a significant impact on the land-use/land-cover of specific areas, producing some unwanted effects. Bangladesh is a tightly inhabited country whose urban population is increasing every day due to the expansion of infrastructure and industry. This study explores the land-use/land-cover change detection and urban dynamics of Gazipur district, Bangladesh, a newly developed industrial hub and city corporation, by using satellite imagery covering every 10-year interval over the period from 1990 to 2020. Supervised classification with a maximum likelihood classifier was used to gather spatial and temporal information from Landsat 5 (TM), 7 (ETM+) and 8 (OLI/TIRS) images. The Geographical Information System (GIS) methodology was also employed to detect changes over time. The kappa coefficient ranged between 0.75 and 0.90. The agricultural land was observed to be shrinking very rapidly, with an area of 716 km2 in 2020. Urbanization increased rapidly in this area, and the urban area grew by more than 500% during the study period. The urbanized area expanded along major roads such as the Dhaka–Mymensingh Highway and Dhaka bypass road. The urbanized area was, moreover, concentrated near the boundary line of Dhaka, the capital city of Bangladesh. Urban expansion was found to be influenced by demographic-, economic-, location- and accessibility-related factors. Therefore, similarly to many countries, concrete urban and development policies should be formulated to preserve the environment and, thereby, achieve sustainable development goal (SDG) 11 (sustainable cities and communities).

Spatio-Temporal Mixed Pixel Analysis of Savanna Ecosystems: A Review

Reliable estimates of savanna vegetation constituents (i.e., woody and herbaceous vegetation) are essential as they are both responders and drivers of global change. The savanna is a highly heterogenous biome with high variability in land cover types while also being very dynamic at both temporal and spatial scales. To understand the spatial-temporal dynamics of savannas, using Earth Observation (EO) data for mixed-pixel analysis is crucial. Mixed pixel analysis provides detailed land cover data at a sub-pixel level which are essential for conservation purposes, understanding food supply for herbivores, quantifying environmental change, such as bush encroachment, and fuel availability essential for understanding fire dynamics, and for accurate estimation of savanna biomass. This review paper consulted 197 studies employing mixed-pixel analysis in savanna ecosystems. The review indicates that studies have so far attempted to resolve the savanna mixed-pixel issues by using mainly coarse resolution data, such as Terra-Aqua MODIS and AVHRR and medium resolution Landsat, to provide fractional cover data. Hence, there is a lack of spatio-temporal mixed-pixel analysis for savannas at high spatial resolutions. Methods used for mixed-pixel analysis include parametric and non-parametric methods which range from pixel-unmixing models, such as linear spectral mixture analysis (SMA), time series decomposition, empirical methods to link the green vegetation parameters with Vegetation Indices (VIs), and machine learning methods, such as regression trees (RT) and random forests (RF). Most studies were undertaken at local and regional scale, highlighting a research gap for savanna mixed pixel studies at national, continental, and global level. Parametric methods for modeling spatio-temporal mixed pixel analysis were preferred for coarse to medium resolution remote sensing data, while non-parametric methods were preferred for very high to high spatial resolution data. The review indicates a gap for long time series spatio-temporal mixed-pixel analysis of savannas using high resolution data at various scales. There is potential to harmonize the available low resolution EO data with new high-resolution sensors to provide long time series of the savanna mixed pixel, which, according to this review, is missing.

Estimation of Urban Land-Use Efficiency for Sustainable Development by Integrating over 30-Year Landsat Imagery with Population Data: A Case Study of Ha Long, Vietnam

Humans are moving into urban areas at an accelerated pace. An increasing urban population fuels urban expansion and reduces nearby agricultural lands and natural environments such as forests, swamps, other water-pervious areas. Unsustainable development creates a disproportion between the growth of urban areas and the growth in urban population. The UN SDG indicator 11.3.1 specifically addresses the issue of the measurement of land-use efficiency. While the metric and methodology to estimate the indicator are straightforward, it faces problems of data unavailability and inconsistency. Vietnam has a record of tremendous economic growth that has translated into more urban settlements of size. Consequently, rural population movement into urban areas has led to many urban sustainable planning and development challenges. In the absence of previous work on estimating land-use efficiency in Vietnamese cities, this study makes the first attempt to examine land-use efficiency in Ha Long, one of the country’s fast-growing cities in recent decades. We mapped land use from high-resolution Landsat imagery (30 m) spanning multi-decadal observations from 1986 to 2020. An advanced machine learning approach, the Support Vector Machine algorithm, was applied to estimate the built-up area, which, by integration with census data, is essential for calculating SDG indicator 11.3.1. This study shows that the land-use efficiency metric was positive but small at the beginning of the considered period but increased in 2000–2020. These results suggest that before 2000, the urban land consumption rate in Ha Long was lower than the population growth rate, implying denser urban land use. The situation changed to the opposite when the urban land consumption rate exceeded the population growth rate in the past two decades. The study’s approach is applicable to regional and district levels to provide comparative analyses between cities or parts of a region or districts of the city. These analyses are valuable tools for assessing the impact of local urban and municipal planning policies on urban development.

Integrating Remote Sensing and a Markov-FLUS Model to Simulate Future Land Use Changes in Hokkaido, Japan

As the second largest island in Japan, Hokkaido provides precious land resources for the Japanese people. Meanwhile, as the food base of Japan, the gradual decrease of the agricultural population and more intensive agricultural practices on Hokkaido have led its arable land use to change year by year, which has also caused changes to the whole land use pattern of the entire island of Hokkaido. To realize the sustainable use of land resources in Hokkaido, past and future changes in land use patterns must be investigated, and target-based land use planning suggestions should be given on this basis. This study uses remote sensing and GIS technology to analyze the temporal and spatial changes of land use in Hokkaido during the past two decades. The types of land use include cultivated land, forest, waterbody, construction, grassland, and others, by using the satellite images of the Landsat images in 2000, 2010, and 2019 to achieve this goal to make classification. In addition, this study used the coupled Markov-FLUS model to simulate and analyze the land use changes in three different scenarios in Hokkaido in the next 20 years. Scenario-based situational analysis shows that the cultivated land in Hokkaido will drop by about 25% in 2040 under the natural development scenario (ND), while the cultivated land area in Hokkaido will remain basically unchanged in cultivated land protection scenario (CP). In forest protection scenario (FP), the area of forest in Hokkaido will increase by 1580.8 km2. It is believed that the findings reveal that the forest land in Hokkaido has been well protected in the past and will be protected well in the next 20 years. However, in land use planning for future, Hokkaido government and enterprises should pay more attention to the protection of cultivated land.

Two Decades Progress on the Application of Remote Sensing for Monitoring Tropical and Sub-Tropical Natural Forests: A Review

Forest covers about a third of terrestrial land surface, with tropical and subtropical zones being a major part. Remote sensing applications constitute a significant approach to monitoring forests. Thus, this paper reviews the progress made by remote sensing data applications to tropical and sub-tropical natural forest monitoring over the last two decades (2000–2020). The review focuses on the thematic areas of aboveground biomass and carbon estimations, tree species identification, tree species diversity, and forest cover and change mapping. A systematic search of articles was performed on Web of Science, Science Direct, and Google Scholar by applying a Boolean operator and using keywords related to the thematic areas. We identified 50 peer-reviewed articles that studied tropical and subtropical natural forests using remote sensing data. Asian and South American natural forests are the most highly researched natural forests, while African natural forests are the least studied. Medium spatial resolution imagery was extensively utilized for forest cover and change mapping as well as aboveground biomass and carbon estimation. In the latest studies, high spatial resolution imagery and machine learning algorithms, such as Random Forest and Support Vector Machine, were jointly utilized for tree species identification. In this review, we noted the promising potential of the emerging high spatial resolution satellite imagery for the monitoring of natural forests. We recommend more research to identify approaches to overcome the challenges of remote sensing applications to these thematic areas so that further and sustainable progress can be made to effectively monitor and manage sustainable forest benefits.

Natural and anthropogenic forcings lead to contrasting vegetation response in long-term vs. short-term timeframes

Understanding vegetation response to natural and anthropogenic forcings is vital for managing watersheds as natural ecosystems. We used a novel integrated framework to separate the impacts of natural factors (e.g. drought, precipitation and temperature) from those of anthropogenic factors (e.g. human activity) on vegetation cover change at the watershed scale. We also integrated several datasets including satellite remote sensing and in-situ measurements for a twenty-year time period (2000-2019). Our results show that despite no significant trend being observed in temperature and precipitation, vegetation indices expressed an increasing trend at both the control and treated watersheds. The vegetation cover was not significantly affected by the natural factors whereas the watershed management practice (as a human activity) had significant impacts on vegetation change in the long-term. Further, the vegetation cover long-term response to watershed management practice was mainly linear. We also found that the vegetation indices values in the 2011-2019 period (as the treated period in treated watershed) were significantly higher than those in the 2000-2010 period. In the short-term, however, the drought condition and decreased precipitation (as natural factors) explained the majority of the change in vegetation cover. For example, the majority of the breakpoints occurred in 2008, and it was related to a widespread extreme drought in the area. The watershed management practice as a human activity along with extreme climatic events could explain a large part of the vegetation changes observed in the treated and control watersheds.

Soil Erosion Assessment and Prediction in Urban Landscapes: A New G2 Model Approach

Soil erosion is a global problem that negatively affects the quality of the environment, the availability of natural resources, as well as the safety of inhabitants. Soil erosion threatens the functioning of urban areas, which was the reason for choosing the territory of the Master Plan of Belgrade (Serbia) as the research area. The calculation of soil erosion loss was analyzed using the G2 erosion model. The model belongs to a group of empirical models and is based on the synthesis of the equation from the Revised Universal Soil Loss Equation (RUSLE) and the Erosion Potential Method (EPM). The estimation of soil degradation was analyzed in two time periods (2001 and 2019), which represent the time boundaries of the management of the Master Plan of Belgrade. The novel approach used in this research is based on using the land cover inventory as a dynamic indicator of the urbanization process. Land cover was identified using remote sensing, machine learning techniques, and the random forest algorithm applied to multispectral satellite images of the Landsat mission in combination with spectral indices. Climatic parameters were analyzed on the basis of data from meteorological stations (first scenario, i.e., 2001), as well as on simulations of changes based on climate scenario RCP8.5 (representative concentration pathways) concerning the current condition of the land cover (second scenario). A comparative analysis of the two time periods identified a slight reduction in total soil loss. For the first period, the average soil loss value is 4.11 t·ha−1·y−1. The analysis of the second period revealed an average value of 3.63 t·ha−1·y−1. However, the increase in non-porous surfaces has led to a change in the focus of soil degradation. Increased average soil loss as one of the catalysts of torrential flood frequencies registered on natural and semi-natural areas were 43.29% and 16.14%, respectively. These results are a significant contribution to the study of soil erosion in urban conditions under the impact of climate change.

Improving Land Cover Classification Using Genetic Programming for Feature Construction

Genetic programming (GP) is a powerful machine learning (ML) algorithm that can produce readable white-box models. Although successfully used for solving an array of problems in different scientific areas, GP is still not well known in the field of remote sensing. The M3GP algorithm, a variant of the standard GP algorithm, performs feature construction by evolving hyperfeatures from the original ones. In this work, we use the M3GP algorithm on several sets of satellite images over different countries to create hyperfeatures from satellite bands to improve the classification of land cover types. We add the evolved hyperfeatures to the reference datasets and observe a significant improvement of the performance of three state-of-the-art ML algorithms (decision trees, random forests, and XGBoost) on multiclass classifications and no significant effect on the binary classifications. We show that adding the M3GP hyperfeatures to the reference datasets brings better results than adding the well-known spectral indices NDVI, NDWI, and NBR. We also compare the performance of the M3GP hyperfeatures in the binary classification problems with those created by other feature construction methods such as FFX and EFS.

Spatial Analysis of Urbanization Patterns in Four Rapidly Growing South Asian Cities Using Sentinel-2 Data

The rapid and dominant urbanization in Asian cities has fueled concerns regarding the local and global efforts toward urban sustainability. Specifically, South Asian cities have been a topical issue concerning ecological and environmental threats due to their unplanned and haphazard urban development. However, comparative urbanization studies in South Asian cities remain uncommon. Therefore, in this study, we sought to comparatively examine the land use and land cover (LULC) dynamics and to detect the urbanization patterns of four rapidly developing South Asian lowland cities: Mumbai (India), Colombo (Sri Lanka), Karachi (Pakistan), and Dhaka (Bangladesh). Sentinel-2 (10 m) data and various geospatial approaches, including urban–rural gradient and grid-based methods, statistics, and urban landscape metric techniques, were used to facilitate the analysis. The study revealed that Mumbai, Karachi, and Dhaka had larger built-up landscapes compared to Colombo. Mumbai had the highest percentage of green spaces, followed by Colombo. Dhaka and Karachi had relatively small percentages of green spaces. Colombo and Dhaka had more croplands, which consistently increased along the urban–rural gradient compared to Mumbai and Karachi. Karachi showed that the only major land use was built-up, while most of the areas were left as open lands. On the other hand, Colombo’s urban setup was more fragmented than the other three cities. Mumbai and Karachi had larger patches of urban footprints compared to Colombo and Dhaka. Thus, this study provides vital information on the past land utilization priorities in the four cities, and comparatively proffers guidance on certain critical areas of focus for local, regional, and global future sustainable urban planning.

Long-Term Changes of Aquatic Invasive Plants and Implications for Future Distribution: A Case Study Using a Tank Cascade System in Sri Lanka

Climate variability can influence the dynamics of aquatic invasive alien plants (AIAPs) that exert tremendous pressure on aquatic systems, leading to loss of biodiversity, agricultural wealth, and ecosystem services. However, the magnitude of these impacts remains poorly known. The current study aims to analyse the long-term changes in the spatio-temporal distribution of AIAPs under the influence of climate variability in a heavily infested tank cascade system (TCS) in Sri Lanka. The changes in coverage of various features in the TCS were analysed using the supervised maximum likelihood classification of ten Landsat images over a 27-year period, from 1992 to 2019 using ENVI remote sensing software. The non-parametric Mann–Kendall trend test and Sen’s slope estimate were used to analyse the trend of annual rainfall and temperature. We observed a positive trend of temperature that was statistically significant (p value < 0.05) and a positive trend of rainfall that was not statistically significant (p values > 0.05) over the time period. Our results showed fluctuations in the distribution of AIAPs in the short term; however, the coverage of AIAPs showed an increasing trend in the study area over the longer term. Thus, this study suggests that the AIAPs are likely to increase under climate variability in the study area.