Remote Sensing Applications-Society and Environment最新文献

Early detection of wildfires is essential for mitigating their impact on forests and surrounding areas. In this study, we propose a wireless sensor node system that combines multiple low-cost sensors with an artificial intelligence-based detection method for early wildfire detection. The system architecture includes temperature, humidity, and smoke sensors, as well as a wireless communication module. Four machine learning classifiers, including decision trees, random forests, support vector machines, and k-nearest neighbors, were evaluated for their effectiveness in predicting wildfire detection using a dataset collected in a forest area. The results showed that the random forest algorithm with optimum hyperparameters had the highest accuracy in classifying fire and non-fire samples (77.95% correctly classified). The proposed system provides an effective and cost-efficient solution for early wildfire detection in large forest areas.

Digital land use data before the age of satellites is scarce. Here, we build a machine learning model, using Extreme Gradient Boosting, that can automatically detect land use classes from an orthophoto map of Sweden (economic maps, 1:10 000 and 1:20 000) constructed from 1942 to 1988. Overall, the machine learning model demonstrated robust performance, with Cohen's Kappa and Matthews Correlation Coefficient of 0.86. The F1 values of the individual classes were 0.98, 0.95, 0.84, and 0.87 for graphics, arable land, forest, and open land, respectively. While the model can be used to detect land use changes in arable land, higher uncertainties associated with forest and open land necessitate further investigation at regional scales or exploration of improved mapping techniques. The code is publicly available to enable easy adaptation for classifying other historical maps.

Mangrove forests are found across the Gulf Cooperation Council (GCC) region despite challenging environmental extremes, including highly variable temperatures and hypersalinity. Understanding the biophysical and anthropogenic factors that influence mangrove forest growth is key to locate suitable areas for regeneration and afforestation activities. The main objectives of this study were to develop a mangrove forest regeneration age map that represents the age of all the existing secondary mangroves in the past 37 years (1986–2023). Long-term Landsat satellite imagery, the random forest classification algorithm, and logistic regression analyses were used to identify the existing secondary mangroves and determine the underlying drivers that contribute to the successful afforestation of mangroves in the region. Our results showed that only around 8.5% of secondary mangrove forests in the GCC region were older than 30 years, with mangroves younger than 5 years being the most abundant age class (41.3%). Saudi Arabia and Oman have the highest percentages of young mangroves, while relatively older secondary mangrove forests were most common in Bahrain, Qatar, and UAE. The current trends in overall mangrove area show that the UAE and Saudi Arabia have the largest total mangrove area among the GCC countries, followed by Qatar, Oman, Bahrain, and Kuwait. The results of the stepwise logistic regression show that the main drivers that influence mangrove regeneration are lower elevation, lower slope, higher available soil moisture, lower average temperatures, higher precipitation, greater proximity to freshwater sources, lower population density and greater distance from agricultural and urban areas. Our results aim to offer support to decision-making in selecting optimal areas for new planting initiatives in the region.

The risk of wildfires is increasing globally and models are critical to reducing this risk. Such models require information on fuel load, a crucial factor of fire behaviour, which is generally determined using a combination of fuel age and fuel accumulation models. Traditionally, estimating fuel load relies on manually compiled fire history data (MCFH). In this paper, we introduce an approach to estimate fuel load using readily available earth observation (EO) data, MODIS MCD64A1. The approach is applied to a wildfire-prone region in Southwestern Australia from 2001 to 2021. Results suggest that MODIS produces more accurate and reliable estimates of fuel load compared with MCFH. It is effective in maintaining spatially and temporally complete records of fires, as it reports 11,019 more hectares of burned areas associated with wildfires over the study period. MODIS performs better in capturing wildfires than prescribed burns, as the spatial overlapping ratio is higher for wildfires (0.63) than prescribed burns (0.42). The high agreement between the two datasets for fuel load estimation (weighted kappa of 0.91) results from grassland covering the majority of the landscape. However, the agreement is reduced for other vegetation types — 0.24 for pine, 0.36 for mallee heath, 0.39 for shrubland, and 0.58 for forest. MODIS has lower effectiveness in detecting small and under-canopy fires such as prescribed burns, suggesting the value in combining EO and manually compiled data to obtain improved estimates of fuel load. Due to the scope of objectives, the integration of EO and MCFH has not been fully explored in this study, which will be included in our future research. This study highlights the potential of earth observation data in assessing wildfire risk as the data are easily accessible and reliable, as well as efficient and cost-effective, and they provide the opportunity to develop mitigation strategies at regional scales.

The identification of susceptible dust sources (SDSs) based on the analysis of effective factors (i.e. dust drivers) is considered to be one of the primary and cost-effective solutions to deal with this phenomenon. Accordingly, this study aimed to identify SDSs and delineate their drivers using remote sensing data and machine learning (ML) algorithms in a hotspot area in the Lower Mesopotamian floodplain in southern Iraq. To model SDSs, a total of 15 environmental features based on remote sensing data such as topographic, climatic, land use/cover, and soil properties were considered as dust drivers and fed into the four well-known ML algorithms, including linear discriminant analysis (LDA), logistic model tree (LMT), extreme gradient boosting (XGB)-Linear, and XGB-Tree-based. Dust emission hotspots were identified by visual interpretation of sub-daily MODIS-Terra/Aqua true color composite imagery (2000–2021) to train (70%) and validate (30%) ML algorithms. Considering the variability of the spatial-temporal patterns of SDSs as a result of changes in dust drivers, the modeling process was carried out in four periods, including 2000–2004, 2005–2007, 2008–2012, and 2013–2021. Our results show that dust events in the study area occur most frequently in April, June, July, and August. Overall, all ML algorithms performed well and provided reliable results for identifying SDSs. However, the XGB-Linear provided the most reliable results with an average area under curve (AUC) of 0.79 for the study periods. Precipitation was determined as the most important dust driver. The SDS maps produced can be used as a basis for the development of rehabilitation plans in the study area to mitigate the adverse effects of dust storms.

Machine learning, a subset of artificial intelligence, has emerged as a powerful tool for generating new knowledge from observations. In the realm of geographic information systems (GIS), machine learning techniques have become essential for spatial analysis tasks. Satellite image classification methods offer valuable decision-making support, particularly in land-use planning and identifying asbestos cement roofs, which pose significant health risks. In Colombia, where asbestos has been used for decades, the detection and management of installed asbestos is critical. This study evaluates the effectiveness of the RoofClassify plugin, a machine learning-based GIS tool, in detecting asbestos cement roofs in Sibaté, Colombia. By employing high-resolution satellite imagery, the study assesses the plugin's accuracy and performance. Results indicate that RoofClassify demonstrates promising capabilities in detecting asbestos cement roofs, achieving an overall accuracy score of 69.73%. This shows potential for identifying areas with the presence of asbestos and informing decision-makers. However, false positives remain a challenge, necessitating further on-site verification. The study underscores the importance of cautious interpretation of classification results and the need for tailored approaches to address specific contextual factors. Overall, RoofClassify presents a valuable tool for identifying asbestos cement roofs, aiding in asbestos management strategies.

The purpose of this article is to present the scope and the dynamics of the environmental changes unfolded in the vicinity of Mtendeli refugee camp. It presents a new method, which combines geospatial analysis of high-resolution Earth observation data (Sentinel-1&2) with ground-based observations and input from local experts. Time series classifications of annual land use/land cover in the surroundings of the camp is developed from remote data. Subsequently main transitions and trends are quantitatively achieved. This is a first study which, not only treats the land transition process in a comprehensive manner, but also tracks the changes and their main drivers on an annual scale over the lifetime of the camp (2016–2021) and the post-closure situation in 2022. Most importantly, thanks to the involvement of social studies, it unfolds the socio-economical drivers of those changes. Drawing upon a random forest algorithm and available databases, we achieve overall classification accuracies of 83.5% (2020) and 82.0% (2022). Our findings indicate an ongoing expansion of cropland between 2016 and 2021, to the detriment of natural vegetation classes. The impact of environmental restoration programs implemented in the former camp area is visible by 2022. The proposed method can be used to identify areas of environmental risk and thus support decisions linked with sustainable development and land management.

Although climate change is impacting various aspects of our environment, it is important to note that the overall risk to trees remains low, especially in urban areas like Hong Kong where the benefits of trees to society are significant. The trees planted in an urban setting are isolated and have several limiting factors including, excessive run-off, urban pollution, physical damage and limited root growth, which sometimes lead for tree failure incidents. The conventional on-site tree health assessment method is time consuming thus, requiring a remote sensing based method to effectively and routinely monitor the health status of urban trees. In this study several types of remote sensing datasets have been exploited to assess the health status of more than 700 Old and Valuable Trees (OVTs) and Stone Wall Trees (SWTs) around Hong Kong. These datasets include the data from Terrestrial LiDAR (Light Detection and Ranging) Surveys (TLS), Handheld Laser Scanner (HLS), Airborne LiDAR Surveys (ALS) and airborne multispectral data. For validation purpose, the in situ tree parameters data was also obtained from the Tree Management Office (TMO) of the Greening, Landscape & Tree Management Section (GLTMS) under the Development Bureau of the Hong Kong SAR Government. The results have indicated that over the period of four years (2017–2020) there has been a decline in the health of some target trees which can be attributed to the increased infestation rate in trees and severe weather conditions. The usage of LiDAR data has supported the fact that different tree structural forms can effectively be extracted and can help making informed decisions on the precise health conditions of urban trees.

Mapping hydrothermal alteration zones associated with porphyry copper deposits (PCDs) is crucial for identifying new exploration targets on a regional scale. Hydrothermal alteration indicator layers play a fundamental role in recognizing potential areas for PCDs, highlighting the need for precise delineation of these zones and their integration with geochemical and geological data to reduce uncertainty in mapping porphyry copper prospectivity. This study focuses on the Pariz district within the Urmia-Dokhtar Metallogenic Belt (UDMB) in southern Iran, a region known for its significant porphyry copper mineralization. First, logical operator algorithms (LOA) were applied to ASTER remote sensing data to map and distinguish argillic and phyllic alteration zones associated with PCDs. Subsequently, propylitic alteration zones associated with chlorite-epidote and propylitic alteration associated with calcite were also delineated, as were silica-rich hydrothermal alteration zones. Five evidence layers corresponding to these geologic features were generated and weighted with logistic functions, independent of expert judgment and without consideration of the spatial distribution of known mineral occurrences (KMOs). In addition, two layers of information were developed, including multivariate geochemical signatures and proximity to intrusive rocks. The geochemical analysis identified two significant factors associated with porphyry copper mineralization: Factor-I (Zn, Pb, Cu, Sn, B) and Factor-II (Mo, Cu). These factors contributed to a multivariate geochemical signature in addition to the alteration layers derived from remote sensing. Evaluation using prediction-area (P-A) plots and Normalized density index (ND) confirmed the effectiveness of all seven layers for mineral prospectivity mapping (MPM). Geometric average (GA), data-driven index overlay (IO), and deep autoencoder neural network (DEA) integrated these layers, with IO showing superior performance in identifying high potential zones, as indicated by higher prediction rates compared to other methods. Therefore, IO proves to be the most efficient approach for mapping the regional porphyry copper minerals in the Pariz district of the UDMB.