Editor’s Choice Articles

Unmanned Aerial Vehicles (UAVs), or drones, have recently emerged as a relatively affordable and accessible method for studying wildlife. Vertical Take-off and Landing (VTOL) UAVs are appropriate for morphometric, behavioural, abundance and demographic studies of marine mammals, providing a stable, nonintrusive and highly manoeuvrable platform. Previous studies using VTOL UAVs have been conducted on various marine mammal species, but specific studies regarding behavioural responses to these devices are limited and scarce. The aim of this study was to evaluate the immediate behavioural responses of common (Delphinus delphis) and bottlenose (Tursiops truncatus) dolphins to a VTOL UAV flown at different altitudes. A multirotor (quadcopter) UAV with an attached GoPro camera was used. Once a dolphin group was located, the UAV was flown at a starting height of 50 m directly above the group, subsequently descending 5 m every 30 s until reaching 5 m. We assessed three behavioural responses to a VTOL UAV at different heights: (i) direction changes, (ii) swimming speed and (iii) diving. Responses by D. delphis (n = 15) and T. truncatus (n = 10) groups were analysed separately. There were no significant responses of T. truncatus to any of the studied variables. For D. delphis, however, there were statistically significant changes in direction when the UAV was flown at a height of 5 m. Our results indicate that UAVs do not induce immediate behavioural responses in common or bottlenose dolphins when flown at heights > 5 m, demonstrating that the use of VTOL UAVs to study dolphins has minimal impact on the animals. However, we advise the use of the precautionary principle when interpreting these results as characteristics of this study site (e.g., high whale-watching activity) may have habituated dolphins to anthropogenic disturbance. Full article

A stack of Sentinel-1 InSAR data in an urban area where flood events recurrently occur, namely Beletweyne town in Somalia, has been analyzed. From this analysis, a novel method to deal with the problem of flood mapping in urban areas has been derived. The approach assumes the availability of a map of persistent scatterers (PSs) inside the urban settlement and is based on the analysis of the temporal trend of the InSAR coherence and the spatial average of the exponential of the InSAR phase in each PS. Both interferometric products are expected to have high and stable values in the PSs; therefore, anomalous decreases may indicate that floodwater is present in an urban area. The stack of Sentinel-1 data has been divided into two subsets. The first one has been used as a calibration set to identify the PSs and determine, for each PS, reference values of the coherence and the spatial average of the exponential of the interferometric phase under standard non-flooded conditions. The other subset has been used for validation purposes. Flood maps produced by UNOSAT, analyzing very-high-resolution optical images of the floods that occurred in Beletweyne in April–May 2018, October–November 2019, and April–May 2020, have been used as reference data. In particular, the map of the April–May 2018 flood has been used for training purposes together with the subset of Sentinel-1 calibration data, whilst the other two maps have been used to validate the products generated by applying the proposed method. The main product is a binary map of flooded PSs that complements the floodwater map of rural/suburban areas produced by applying a well-consolidated algorithm based on intensity data. In addition, a flood severity map that labels the different districts of Beletweyne, as not, partially, or totally flooded has been generated to consolidate the validation. The results have confirmed the effectiveness of the proposed method. Full article

The combined heatwave and drought in 2018 notably affected the state and functioning of European ecosystems. The severity and distribution of this extreme event across ecosystem types and its possible implication on ecosystem water fluxes are still poorly understood. This study estimates spatio-temporal changes in evapotranspiration (ET) during the 2018 drought and heatwave and assesses how these changes are distributed in European ecosystems along climatic gradients. We used the ET eight-day composite product from the MODerate Resolution Imaging Spectroradiometer (MODIS) together with meteorological data from the European Centre for Medium-Range Weather Forecasts (ECMWF ERA5). Our results indicate that ecosystem ET was strongly reduced (up to −50% compared to a 10-year reference period) in areas with extreme anomalies in surface air temperature (T_sa) and precipitation (P) in central, northern, eastern, and western Europe. Northern and Eastern Europe had prolonged anomalies of up to seven months with extreme intensities (relative and absolute) of T_sa, P, and ET. Particularly, agricultural areas, mixed natural vegetation, and non-irrigated agricultural areas were the most affected by the increased temperatures in northern Europe. Our results show contrasting drought impacts on ecosystem ET between the North and South of Europe as well as on ecosystem types. Full article

Ship detection from high-resolution optical satellite images is still an important task that deserves optimal solutions. This paper introduces a novel high-resolution image network-based approach based on the preselection of a region of interest (RoI). This pre-selected network first identifies and extracts a region of interest from input images. In order to efficiently match ship candidates, the principle of our approach is to distinguish suspected areas from the images based on hue, saturation, value (HSV) differences between ships and the background. The whole approach is the basis of an experiment with a large ship dataset, consisting of Google Earth images and HRSC2016 datasets. The experiment shows that the H-YOLO network, which uses the same weight training from a set of remote sensing images, has a 19.01% higher recognition rate and a 16.19% higher accuracy than applying the you only look once (YOLO) network alone. After image preprocessing, the value of the intersection over union (IoU) is also greatly improved. Full article

Climate studies, including trend detection and other time series analyses, necessarily require stable, well-characterized and long-term data records. For satellite-based geophysical retrieval datasets, such data records often involve merging the observational records of multiple similar, though not identical, instruments. The National Aeronautics and Space Administration (NASA) cloud mask (CLDMSK) and cloud-top and optical properties (CLDPROP) products are designed to bridge the observational records of the Moderate-resolution Imaging Spectroradiometer (MODIS) onboard NASA’s Aqua satellite and the Visible Infrared Imaging Radiometer Suite (VIIRS) onboard the joint NASA/National Oceanic and Atmospheric Administration (NOAA) Suomi National Polar-orbiting Partnership (SNPP) satellite and NOAA’s new generation of operational polar-orbiting weather platforms (NOAA-20+). Early implementations of the CLDPROP algorithms on Aqua MODIS and SNPP VIIRS suffered from large intersensor biases in cloud optical properties that were traced back to relative radiometric inconsistency in analogous shortwave channels on both imagers, with VIIRS generally observing brighter top-of-atmosphere spectral reflectance than MODIS (e.g., up to 5% brighter in the 0.67 µm channel). Radiometric adjustment factors for the SNPP and NOAA-20 VIIRS shortwave channels used in the cloud optical property retrievals are derived from an extensive analysis of the overlapping observational records with Aqua MODIS, specifically for homogenous maritime liquid water cloud scenes for which the viewing/solar geometry of MODIS and VIIRS match. Application of these adjustment factors to the VIIRS L1B prior to ingestion into the CLDMSK and CLDPROP algorithms yields improved intersensor agreement, particularly for cloud optical properties. Full article

The uniform application (UA) of agrochemicals results in the over-application of harmful chemicals, increases crop input costs, and deteriorates the environment when compared with variable rate application (VA). A smart variable rate sprayer (SVRS) was designed, developed, and tested using deep learning (DL) for VA application of agrochemicals. Real-time testing of the SVRS took place for detecting and spraying and/or skipping lambsquarters weed and early blight infected and healthy potato plants. About 24,000 images were collected from potato fields in Prince Edward Island and New Brunswick under varying sunny, cloudy, and partly cloudy conditions and processed/trained using YOLOv3 and tiny-YOLOv3 models. Due to faster performance, the tiny-YOLOv3 was chosen to deploy in SVRS. A laboratory experiment was designed under factorial arrangements, where the two spraying techniques (UA and VA) and the three weather conditions (cloudy, partly cloudy, and sunny) were the two independent variables with spray volume consumption as a response variable. The experimental treatments had six repetitions in a 2 × 3 factorial design. Results of the two-way ANOVA showed a significant effect of spraying application techniques on volume consumption of spraying liquid (p-value < 0.05). There was no significant effect of weather conditions and interactions between the two independent variables on volume consumption during weeds and simulated diseased plant detection experiments (p-value > 0.05). The SVRS was able to save 42 and 43% spraying liquid during weeds and simulated diseased plant detection experiments, respectively. Water sensitive papers’ analysis showed the applicability of SVRS for VA with >40% savings of spraying liquid by SVRS when compared with UA. Field applications of this technique would reduce the crop input costs and the environmental risks in conditions (weed and disease) like experimental testing. Full article

Drought is one of the extreme climatic events that has a severe impact on crop production and food supply. Our main goal is to test the suitability of remote sensing-based indices to detect drought impacts on crop production from a global to regional scale. Moderate resolution imaging spectroradiometer (MODIS) based imagery, spanning from 2001 to 2017 was used for this task. This includes the normalized difference vegetation index (NDVI), land surface temperature (LST), and the evaporative stress index (ESI), which is based on the ratio of actual to potential evapotranspiration. These indices were used as indicators of drought-induced vegetation conditions for three main crops: maize, wheat, and soybean. The start and end of the growing season, as observed at 500 m resolution, were used to exclude the time steps that are outside of the growing season. Based on the three indicators, monthly standardized anomalies were estimated, which were used for both analyses of spatiotemporal patterns of drought and the relationship with yield anomalies. Anomalies in the ESI had higher correlations with maize and wheat yield anomalies than other indices, indicating that prolonged periods of low ESI during the growing season are highly correlated with reduced crop yields. All indices could identify past drought events, such as the drought in the USA in 2012, Eastern Africa in 2016–2017, and South Africa in 2015–2016. The results of this study highlight the potential of the use of moderate resolution remote sensing-based indicators combined with phenometrics for drought-induced crop impact monitoring. For several regions, droughts identified using the ESI and LST were more intense than the NDVI-based results. We showed that these indices are relevant for agricultural drought monitoring at both global and regional scales. They can be integrated into drought early warning systems, process-based crop models, as well as can be used for risk assessment and included in advanced decision-support frameworks. Full article

Biological invasions are a major global threat to biodiversity and often affect ecosystem services negatively. They are particularly problematic on oceanic islands where there are many narrow-ranged endemic species, and the biota may be very susceptible to invasion. Quantifying and mapping invasion processes are important steps for management and control but are challenging with the limited resources typically available and particularly difficult to implement on oceanic islands with very steep terrain. Remote sensing may provide an excellent solution in circumstances where the invading species can be reliably detected from imagery. We here develop a method to map the distribution of the alien chestnut (Castanea sativa Mill.) on the island of La Palma (Canary Islands, Spain), using freely available satellite images. On La Palma, the chestnut invasion threatens the iconic laurel forest, which has survived since the Tertiary period in the favourable climatic conditions of mountainous islands in the trade wind zone. We detect chestnut presence by taking advantage of the distinctive phenology of this alien tree, which retains its deciduousness while the native vegetation is evergreen. Using both Landsat 8 and Sentinel-2 (parallel analyses), we obtained images in two seasons (chestnuts leafless and in-leaf, respectively) and performed image regression to detect pixels changing from leafless to in-leaf chestnuts. We then applied supervised classification using Random Forest to map the present-day occurrence of the chestnut. Finally, we performed species distribution modelling to map the habitat suitability for chestnut on La Palma, to estimate which areas are prone to further invasion. Our results indicate that chestnuts occupy 1.2% of the total area of natural ecosystems on La Palma, with a further 12–17% representing suitable habitat that is not yet occupied. This enables targeted control measures with potential to successfully manage the invasion, given the relatively long generation time of the chestnut. Our method also enables research on the spread of the species since the earliest Landsat images. Full article

Lidar remote sensing has proven to be a powerful tool for estimating ground elevation, canopy height, and additional vegetation parameters, which in turn are valuable information for the investigation of ecosystems. Spaceborne lidar systems, like the Global Ecosystem Dynamics Investigation (GEDI), can deliver these height estimates on a near global scale. This paper analyzes the accuracy of the first version of GEDI ground elevation and canopy height estimates in two study areas with temperate forests in the Free State of Thuringia, central Germany. Digital terrain and canopy height models derived from airborne laser scanning data are used as reference heights. The influence of various environmental and acquisition parameters (e.g., canopy cover, terrain slope, beam type) on GEDI height metrics is assessed. The results show a consistently high accuracy of GEDI ground elevation estimates under most conditions, except for areas with steep slopes. GEDI canopy height estimates are less accurate and show a bigger influence of some of the included parameters, specifically slope, vegetation height, and beam sensitivity. A number of relatively high outliers (around 9–13% of the measurements) is present in both ground elevation and canopy height estimates, reducing the estimation precision. Still, it can be concluded that GEDI height metrics show promising results and have potential to be used as a basis for further investigations. Full article

Early Warning Systems (EWS) for near real-time detection of deforestation are a fundamental component of public policies focusing on the reduction in forest biomass loss and associated CO₂ emissions. Most of the operational EWS are based on optical data, which are severely limited by the cloud cover in tropical environments. Synthetic Aperture Radar (SAR) data can help to overcome this observational gap. SAR measurements, however, can be altered by atmospheric effects on and variations in surface moisture. Different techniques of time series (TS) stabilization have been used to mitigate the instability of C-band SAR measurements. Here, we evaluate the performance of two different approaches to SAR TS stabilization, harmonic deseasonalization and spatial stabilization, as well as two deforestation detection techniques, Adaptive Linear Thresholding (ALT) and maximum likelihood classification (MLC). We set up a rigorous, Amazon-wide validation experiment using the Google Earth Engine platform to sample and process Sentinel-1A data of nearly 6000 locations in the whole Brazilian Amazonian basin, generating more than 8M processed samples. Half of those locations correspond to non-degraded forest areas, while the other half pertained to 2019 deforested areas. The detection results showed that the spatial stabilization algorithm improved the results of the MLC approach, reaching 94.36% global accuracy. The ALT detection algorithm performed better, reaching 95.91% global accuracy, regardless of the use of any stabilization method. The results of this experiment are being used to develop an operational EWS in the Brazilian Amazon. Full article

The increasing availability of digital elevation models (DEMs) facilitates the monitoring of glacier mass balances on local and regional scales. Geodetic glacier mass balances are obtained by differentiating DEMs. However, these computations are usually affected by voids in the derived elevation change data sets. Different approaches, using spatial statistics or interpolation techniques, were developed to account for these voids in glacier mass balance estimations. In this study, we apply novel void filling techniques, which are typically used for the reconstruction and retouche of images and photos, for the first time on elevation change maps. We selected 6210 km² of glacier area in southeast Alaska, USA, covered by two void-free DEMs as the study site to test different inpainting methods. Different artificially voided setups were generated using manually defined voids and a correlation mask based on stereoscopic processing of Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) acquisition. Three “novel” (Telea, Navier–Stokes and shearlet) as well as three “classical” (bilinear interpolation, local and global hypsometric methods) void filling approaches for glacier elevation data sets were implemented and evaluated. The hypsometric approaches showed, in general, the worst performance, leading to high average and local offsets. Telea and Navier–Stokes void filling showed an overall stable and reasonable quality. The best results are obtained for shearlet and bilinear void filling, if certain criteria are met. Considering also computational costs and feasibility, we recommend using the bilinear void filling method in glacier volume change analyses. Moreover, we propose and validate a formula to estimate the uncertainties caused by void filling in glacier volume change computations. The formula is transferable to other study sites, where no ground truth data on the void areas exist, and leads to higher accuracy of the error estimates on void-filled areas. In the spirit of reproducible research, we publish a software repository with the implementation of the novel void filling algorithms and the code reproducing the statistical analysis of the data, along with the data sets themselves. Full article

The main objective of the present study is to introduce a novel predictive model that combines evolutionary algorithms and machine learning (ML) models, so as to construct a landslide susceptibility map. Genetic algorithms (GA) are used as a feature selection method, whereas the particle swarm optimization (PSO) method is used to optimize the structural parameters of two ML models, support vector machines (SVM) and artificial neural network (ANN). A well-defined spatial database, which included 335 landslides and twelve landslide-related variables (elevation, slope angle, slope aspect, curvature, plan curvature, profile curvature, topographic wetness index, stream power index, distance to faults, distance to river, lithology, and hydrological cover) are considered for the analysis, in the Achaia Regional Unit located in Northern Peloponnese, Greece. The outcome of the study illustrates that both ML models have an excellent performance, with the SVM model achieving the highest learning accuracy (0.977 area under the receiver operating characteristic curve value (AUC)), followed by the ANN model (0.969). However, the ANN model shows the highest prediction accuracy (0.800 AUC), followed by the SVM (0.750 AUC) model. Overall, the proposed ML models highlights the necessity of feature selection and tuning procedures via evolutionary optimization algorithms and that such approaches could be successfully used for landslide susceptibility mapping as an alternative investigation tool. Full article

Utilising satellite images for planning and development is becoming a common practice as computational power and machine learning capabilities expand. In this paper, we explore the use of satellite image derived building footprint data to classify the residential status of urban buildings in low and middle income countries. A recently developed ensemble machine learning building classification model is applied for the first time to the Democratic Republic of the Congo, and to Nigeria. The model is informed by building footprint and label data of greater completeness and attribute consistency than have previously been available for these countries. A GIS workflow is described that semiautomates the preparation of data for input to the model. The workflow is designed to be particularly useful to those who apply the model to additional countries and use input data from diverse sources. Results show that the ensemble model correctly classifies between 85% and 93% of structures as residential and nonresidential across both countries. The classification outputs are likely to be valuable in the modelling of human population distributions, as well as in a range of related applications such as urban planning, resource allocation, and service delivery. Full article

A project was constructed to integrate remotely sensed data from multiple sensors and platforms to characterize range of ecosystem characteristics in the Peace–Athabasca Delta in Northern Alberta, Canada. The objective of this project was to provide a framework for the processing of multisensor data to extract ecosystem information describing complex deltaic wetland environments. The data used in this study was based on a passive satellite-based earth observation multispectral sensor (Sentinel-2) and airborne discrete light detection and ranging (LiDAR). The data processing strategy adopted here allowed us to employ a data mining approach to grouping of the input variables into ecologically meaningful clusters. Using this approach, we described not only the reflective characteristics of the cover, but also ascribe vertical and horizontal structure, thereby differentiating spectrally similar, but ecologically distinct, ground features. This methodology provides a framework for assessing the impact of ecosystems on radiance, as measured by Earth observing systems, where it forms the basis for sampling and analysis. This final point will be the focus of future work. Full article

Epidemiological studies on the health effects of air pollution usually rely on measurements from fixed ground monitors, which provide limited spatio-temporal coverage. Data from satellites, reanalysis, and chemical transport models offer additional information used to reconstruct pollution concentrations at high spatio-temporal resolutions. This study aims to develop a multi-stage satellite-based machine learning model to estimate daily fine particulate matter (PM_2.5) levels across Great Britain between 2008–2018. This high-resolution model consists of random forest (RF) algorithms applied in four stages. Stage-1 augments monitor-PM_2.5 series using co-located PM₁₀ measures. Stage-2 imputes missing satellite aerosol optical depth observations using atmospheric reanalysis models. Stage-3 integrates the output from previous stages with spatial and spatio-temporal variables to build a prediction model for PM_2.5. Stage-4 applies Stage-3 models to estimate daily PM_2.5 concentrations over a 1 km grid. The RF architecture performed well in all stages, with results from Stage-3 showing an average cross-validated R² of 0.767 and minimal bias. The model performed better over the temporal scale when compared to the spatial component, but both presented good accuracy with an R² of 0.795 and 0.658, respectively. These findings indicate that direct satellite observations must be integrated with other satellite-based products and geospatial variables to derive reliable estimates of air pollution exposure. The high spatio-temporal resolution and the relatively high precision allow these estimates (approximately 950 million points) to be used in epidemiological analyses to assess health risks associated with both short- and long-term exposure to PM_2.5. Full article

Surface melt, driven by atmospheric temperatures and albedo, is a strong contribution of mass loss of the Greenland Ice Sheet. In the past, black carbon, algae and other light-absorbing impurities were suggested to govern albedo in Greenland’s ablation zone. Here we combine optical (MODIS/Sentinel-2) and radar (Sentinel-1) remote sensing data with airborne radar and laser scanner data, and engage firn modelling to identify the governing factors leading to dark glacier surfaces in Northeast Greenland. After the drainage of supraglacial lakes, the former lake ground is a clean surface represented by a high reflectance in Sentinel-2 data and aerial photography. These bright spots move with the ice flow and darken by more than 20% over only two years. In contrast, sites further inland do not exhibit this effect. This finding suggests that local deposition of dust, rather than black carbon or cryoconite formation, is the governing factor of albedo of fast-moving outlet glaciers. This is in agreement with a previous field study in the area which finds the mineralogical composition and grain size of the dust comparable with that of the surrounding soils. Full article

Mangroves are found globally throughout tropical and sub-tropical inter-tidal coastlines. These highly biodiverse and carbon-dense ecosystems have multi-faceted value, providing critical goods and services to millions living in coastal communities and making significant contributions to global climate change mitigation through carbon sequestration and storage. Despite their many values, mangrove loss continues to be widespread in many regions due primarily to anthropogenic activities. Accessible, intuitive tools that enable coastal managers to map and monitor mangrove cover are needed to stem this loss. Remotely sensed data have a proven record for successfully mapping and monitoring mangroves, but conventional methods are limited by imagery availability, computing resources and accessibility. In addition, the variable tidal levels in mangroves presents a unique mapping challenge, particularly over geographically large extents. Here we present a new tool—the Google Earth Engine Mangrove Mapping Methodology (GEEMMM)—an intuitive, accessible and replicable approach which caters to a wide audience of non-specialist coastal managers and decision makers. The GEEMMM was developed based on a thorough review and incorporation of relevant mangrove remote sensing literature and harnesses the power of cloud computing including a simplified image-based tidal calibration approach. We demonstrate the tool for all of coastal Myanmar (Burma)—a global mangrove loss hotspot—including an assessment of multi-date mapping and dynamics outputs and a comparison of GEEMMM results to existing studies. Results—including both quantitative and qualitative accuracy assessments and comparisons to existing studies—indicate that the GEEMMM provides an accessible approach to map and monitor mangrove ecosystems anywhere within their global distribution. Full article

Our paper presents a remote sensing workflow for identifying modern activities that threaten archaeological sites, developed as part of the work of the Endangered Archaeology of the Middle East and North Africa (EAMENA) project. We use open-source Sentinel-2 satellite imagery and the free tool Google Earth Engine to run a per-pixel change detection to make the methods and data as accessible as possible for heritage professionals. We apply this and perform validation at two case studies, the Aswan and Kom-Ombo area in Egypt, and the Jufra oases in Libya, with an overall accuracy of the results ranging from 85–91%. Human activities, such as construction, agriculture, rubbish dumping and natural processes were successfully detected at archaeological sites by the algorithm, allowing these sites to be prioritised for recording. A few instances of change too small to be detected by Sentinel-2 were missed, and false positives were caused by registration errors, shadow and movements of sand. This paper shows that the expansion of agricultural and urban areas particularly threatens the survival of archaeological sites, but our extensive online database of archaeological sites and programme of training courses places us in a unique position to make our methods widely available. Full article

The Niger Delta Region is the largest river delta in Africa and features the fifth largest mangrove forest on Earth. It provides numerous ecosystem services to the local populations and holds a wealth of biodiversity. However, due to the oil and gas reserves and the explosion of human population it is under threat from overexploitation and degradation. There is a pressing need for an accurate assessment of the land cover dynamics in the region. The limited previous efforts have produced controversial results, as the area of western Africa is notorious for the gaps in the Landsat archive and the lack of cloud-free data. Even fewer studies have attempted to map the extent of the degraded mangrove forest system, reporting low accuracies. Here, we map the eight main land cover classes over the NDR using spectral-temporal metrics from all available Landsat data centred around three epochs. We also test the performance of the classification when L-band radar data are added to the Landsat-based metrics. To further our understanding of the land cover change dynamics, we carry out two additional assessments: a change intensity analysis for the entire NDR and, focusing specifically on the mangrove forest, we analyse the fragmentation of both the healthy and the degraded mangrove land cover classes. We achieve high overall classification accuracies in all epochs (~79% for 1988, and 82% for 2000 and 2013) and are able to map the degraded mangroves accurately, for the first time, with user’s accuracies between 77% and 87% and producer’s accuracies consistently above 82%. Our results show that mangrove forests, lowland rainforests, and freshwater forests are reporting net and highly intense losses (mangrove net loss: ~500 km²; woodland net loss: ~1400 km²), while built-up areas have almost doubled in size (from 1990 km² in 1988 to 3730 km² in 2013). The mangrove forests are also consistently more fragmented, with the opposite effect being observed for the degraded mangroves in more recent years. Our study provides a valuable assessment of land cover dynamics in the NDR and the first ever accurate estimates of the extent of the degraded mangrove forest and its fragmentation. Full article

Areas at risk of land subsidence in Jakarta can be identified using a land subsidence susceptibility map. This study evaluates the quality of a susceptibility map made using functional (logistic regression and multilayer perceptron) and meta-ensemble (AdaBoost and LogitBoost) machine learning algorithms based on a land subsidence inventory map generated using the Sentinel-1 synthetic aperture radar (SAR) dataset from 2017 to 2020. The land subsidence locations were assessed using the time-series interferometry synthetic aperture radar (InSAR) method based on the Stanford Method for Persistent Scatterers (StaMPS) algorithm. The mean vertical deformation maps from ascending and descending tracks were compared and showed a good correlation between displacement patterns. Persistent scatterer points with mean vertical deformation value were randomly divided into two datasets: 50% for training the susceptibility model and 50% for validating the model in terms of accuracy and reliability. Additionally, 14 land subsidence conditioning factors correlated with subsidence occurrence were used to generate land subsidence susceptibility maps from the four algorithms. The receiver operating characteristic (ROC) curve analysis showed that the AdaBoost algorithm has higher subsidence susceptibility prediction accuracy (81.1%) than the multilayer perceptron (80%), logistic regression (79.4%), and LogitBoost (79.1%) algorithms. The land subsidence susceptibility map can be used to mitigate disasters caused by land subsidence in Jakarta, and our method can be applied to other study areas. Full article

Geotagged smartphone photos can be employed to build digital terrain models using structure from motion-multiview stereo (SfM-MVS) photogrammetry. Accelerometer, magnetometer, and gyroscope sensors integrated within consumer-grade smartphones can be used to record the orientation of images, which can be combined with location information provided by inbuilt global navigation satellite system (GNSS) sensors to geo-register the SfM-MVS model. The accuracy of these sensors is, however, highly variable. In this work, we use a 200 m-wide natural rocky cliff as a test case to evaluate the impact of consumer-grade smartphone GNSS sensor accuracy on the registration of SfM-MVS models. We built a high-resolution 3D model of the cliff, using an unmanned aerial vehicle (UAV) for image acquisition and ground control points (GCPs) located using a differential GNSS survey for georeferencing. This 3D model provides the benchmark against which terrestrial SfM-MVS photogrammetry models, built using smartphone images and registered using built-in accelerometer/gyroscope and GNSS sensors, are compared. Results show that satisfactory post-processing registrations of the smartphone models can be attained, requiring: (1) wide acquisition areas (scaling with GNSS error) and (2) the progressive removal of misaligned images, via an iterative process of model building and error estimation. Full article

Monitoring drought impacts in forest ecosystems is a complex process because forest ecosystems are composed of different species with heterogeneous structural compositions. Even though forest drought status is a key control on the carbon cycle, very few indices exist to monitor and predict forest drought stress. The Forest Drought Indicator (ForDRI) is a new monitoring tool developed by the National Drought Mitigation Center (NDMC) to identify forest drought stress. ForDRI integrates 12 types of data, including satellite, climate, evaporative demand, ground water, and soil moisture, into a single hybrid index to estimate tree stress. The model uses Principal Component Analysis (PCA) to determine the contribution of each input variable based on its covariance in the historical records (2003–2017). A 15-year time series of 780 ForDRI maps at a weekly interval were produced. The ForDRI values at a 12.5km spatial resolution were compared with normalized weekly Bowen ratio data, a biophysically based indicator of stress, from nine AmeriFlux sites. There were strong and significant correlations between Bowen ratio data and ForDRI at sites that had experienced intense drought. In addition, tree ring annual increment data at eight sites in four eastern U.S. national parks were compared with ForDRI values at the corresponding sites. The correlation between ForDRI and tree ring increments at the selected eight sites during the summer season ranged between 0.46 and 0.75. Generally, the correlation between the ForDRI and normalized Bowen ratio or tree ring increment are reasonably good and indicate the usefulness of the ForDRI model for estimating drought stress and providing decision support on forest drought management. Full article

Fast-growing Eucalyptus spp. forest plantations and their resultant wood products are economically important and may provide a low-cost means to sequester carbon for greenhouse gas reduction. The development of advanced and optimized frameworks for estimating forest plantation attributes from lidar remote sensing data combined with statistical modeling approaches is a step towards forest inventory operationalization and might improve industry efficiency in monitoring and managing forest resources. In this study, we first developed and tested a framework for modeling individual tree attributes in fast-growing Eucalyptus forest plantation using airborne lidar data and linear mixed-effect models (LME) and assessed the gain in accuracy compared to a conventional linear fixed-effects model (LFE). Second, we evaluated the potential of using the tree-level estimates for determining tree attribute uniformity across different stand ages. In the field, tree measurements, such as tree geolocation, species, genotype, age, height (Ht), and diameter at breast height (dbh) were collected through conventional forest inventory practices, and tree-level aboveground carbon (AGC) was estimated using allometric equations. Individual trees were detected and delineated from lidar-derived canopy height models (CHM), and crown-level metrics (e.g., crown volume and crown projected area) were computed from the lidar 3-D point cloud. Field and lidar-derived crown metrics were combined for ht, dbh, and AGC modeling using an LME. We fitted a varying intercept and slope model, setting species, genotype, and stand (alone and nested) as random effects. For comparison, we also modeled the same attributes using a conventional LFE model. The tree attribute estimates derived from the best LME model were used for assessing forest uniformity at the tree level using the Lorenz curves and Gini coefficient (GC). We successfully detected 96.6% of the trees from the lidar-derived CHM. The best LME model for estimating the tree attributes was composed of the stand as a random effect variable, and canopy height, crown volume, and crown projected area as fixed effects. The %RMSE values for tree-level height, dbh, and AGC were 8.9%, 12.1%, and 23.7% for the LFE model and improved to 7.3%, 7.1%, and 13.6%, respectively, for the LME model. Tree attributes uniformity was assessed with the Lorenz curves and tree-level estimations, especially for the older stands. All stands showed a high level of tree uniformity with GC values approximately 0.2. This study demonstrates that accurate detection of individual trees and their associated crown metrics can be used to estimate Ht, dbh, and AGC stocks as well as forest uniformity in fast-growing Eucalyptus plantations forests using lidar data as inputs to LME models. This further underscores the high potential of our proposed approach to monitor standing stock and growth in Eucalyptus—and similar forest plantations for carbon dynamics and forest product planning. Full article

The ability of the Canadian agriculture sector to make better decisions and manage its operations more competitively in the long term is only as good as the information available to inform decision-making. At all levels of Government, a reliable flow of information between scientists, practitioners, policy-makers, and commodity groups is critical for developing and supporting agricultural policies and programs. Given the vastness and complexity of Canada’s agricultural regions, space-based remote sensing is one of the most reliable approaches to get detailed information describing the evolving state of the country’s environment. Agriculture and Agri-Food Canada (AAFC)—the Canadian federal department responsible for agriculture—produces the Annual Space-Based Crop Inventory (ACI) maps for Canada. These maps are valuable operational space-based remote sensing products which cover the agricultural land use and non-agricultural land cover found within Canada’s agricultural extent. Developing and implementing novel methods for improving these products are an ongoing priority of AAFC. Consequently, it is beneficial to implement advanced machine learning and big data processing methods along with open-access satellite imagery to effectively produce accurate ACI maps. In this study, for the first time, the Google Earth Engine (GEE) cloud computing platform was used along with an Artificial Neural Networks (ANN) algorithm and Sentinel-1, -2 images to produce an object-based ACI map for 2018. Furthermore, different limitations of the proposed method were discussed, and several suggestions were provided for future studies. The Overall Accuracy (OA) and Kappa Coefficient (KC) of the final 2018 ACI map using the proposed GEE cloud method were 77% and 0.74, respectively. Moreover, the average Producer Accuracy (PA) and User Accuracy (UA) for the 17 cropland classes were 79% and 77%, respectively. Although these levels of accuracies were slightly lower than those of the AAFC’s ACI map, this study demonstrated that the proposed cloud computing method should be investigated further because it was more efficient in terms of cost, time, computation, and automation. Full article

Identifying the flooding risk hotspot is crucial for aiding a rapid response and prioritizes mitigation efforts over large disaster impacted regions. While climate change is increasing the risk of floods in many vulnerable regions of the world, the commonly used crisis map is inefficient and cannot rapidly determine the spatial variation and intensity of flooding extension across the affected areas. In such cases, the Local Indicators of Spatial Association (LISA) statistic can detect heterogeneity or the flooding hotspot at a local spatial scale beyond routine mapping. This area, however, has not yet been studied in the context of the magnitude of the floods. The present study incorporates the LISA methodology including Moran’s I and Getis–Ord Gi* to identify the spatial and temporal heterogeneity of the occurrence of flooding from super cyclone Amphan across 16 coastal districts of Bangladesh. Using the Synthetic Aperture Radar (SAR) data from Sentinel-1 and a Support Vector Machine (SVM) classification, “water” and “land” were classified for the pre-event (16 May 2020) and post-events (22 May, 28 May, and 7 June 2020) of the area under study. A Modified Normalized Difference Water Index (MNDWI), and visual comparison were used to evaluate the flood maps. A compelling agreement was accomplished between the observed and predicted flood maps, with an overall precision of above 95% for all SAR classified images. As per this study, 2233 km² (8%) of the region is estimated to have been inundated on 22 May. After this point, the intensity and aerial expansion of flood decreased to 1490 km² by 28 May before it increased slightly to 1520 km² (2.1% of the study area) on 7 June. The results from LISA indicated that the main flooding hotspots were located in the central part, particularly in the region off the north-east of the mangrove forest. A total of 238 Unions (smallest administrative units) were identified as high flooding hotspots (p < 0.05) on 22 May, but the number of flooding hotspots dropped to 166 in the second week (28 May) after Amphan subsided before it increased to a further 208 hotspots (p < 0.05) on 7 June due to incessant rainfall and riverbank failure in the south-west part of the study area. As such, an appropriate, timely, and cost-effective strategy would be to assess existing flooding management policies through the identified flooding hotspot regions. This identification would then allow for the creation of an improved policy to help curtail the destructive effects of flooding in the future. Full article

Artificial skyglow, the brightening of the night sky by artificial light at night that is scattered back to Earth within the atmosphere, is detrimental to astronomical observations and has an impact on ecosystems as a form of light pollution. In this work, we investigated the impact of the lockdown caused by the COVID-19 pandemic on the urban skyglow of Berlin, Germany. We compared night sky brightness and correlated color temperature (CCT) measurements obtained with all-sky cameras during the COVID-19 lockdown in March 2020 with data from March 2017. Under normal conditions, we expected an increase in night sky brightness (or skyglow, respectively) and CCT because of the transition to LED. This is supported by a measured CCT shift to slightly higher values and a time series analysis of night-time light satellite data showing an increase in artificial light emission in Berlin. However, contrary to this observation, we measured a decrease in artificial skyglow at zenith by 20% at the city center and by more than 50% at 58 km distance from the center during the lockdown. We assume that the main cause for the reduction of artificial skyglow originates from improved air quality due to less air and road traffic, which is supported by statistical data and satellite image analysis. To our knowledge, this is the first reported impact of COVID-19 on artificial skyglow and we conclude that air pollution should shift more into the focus of light pollution research. Full article

Response efforts in emergency applications such as border protection, humanitarian relief and disaster monitoring have improved with the use of Unmanned Aerial Vehicles (UAVs), which provide a flexibly deployed eye in the sky. These efforts have been further improved with advances in autonomous behaviours such as obstacle avoidance, take-off, landing, hovering and waypoint flight modes. However, most UAVs lack autonomous decision making for navigating in complex environments. This limitation creates a reliance on ground control stations to UAVs and, therefore, on their communication systems. The challenge is even more complex in indoor flight operations, where the strength of the Global Navigation Satellite System (GNSS) signals is absent or weak and compromises aircraft behaviour. This paper proposes a UAV framework for autonomous navigation to address uncertainty and partial observability from imperfect sensor readings in cluttered indoor scenarios. The framework design allocates the computing processes onboard the flight controller and companion computer of the UAV, allowing it to explore dangerous indoor areas without the supervision and physical presence of the human operator. The system is illustrated under a Search and Rescue (SAR) scenario to detect and locate victims inside a simulated office building. The navigation problem is modelled as a Partially Observable Markov Decision Process (POMDP) and solved in real time through the Augmented Belief Trees (ABT) algorithm. Data is collected using Hardware in the Loop (HIL) simulations and real flight tests. Experimental results show the robustness of the proposed framework to detect victims at various levels of location uncertainty. The proposed system ensures personal safety by letting the UAV to explore dangerous environments without the intervention of the human operator. Full article

The short-term prediction of earthquakes is an essential issue connected with human life protection and related social and economic matters. Recent papers have provided some evidence of the link between the lithosphere, lower atmosphere, and ionosphere, even though with marginal statistical evidence. The basic coupling is hypothesized as being via the atmospheric gravity wave (AGW)/acoustic wave (AW) channel. In this paper we analyze a scenario of the low latitude earthquake (Mw = 6.9) which occurred in Indonesia on 5 August 2018, through a multi-instrumental approach, using ground and satellites high quality data. As a result, we derive a new analytical lithospheric–atmospheric–ionospheric–magnetospheric coupling model with the aim to provide quantitative indicators to interpret the observations around 6 h before and at the moment of the earthquake occurrence. Full article

The use of Sentinel-1 (S1) radar for wide-area, near-real-time (NRT) tropical-forest-change monitoring is discussed, with particular attention to forest degradation and deforestation. Since forest change can relate to processes ranging from high-impact, large-scale conversion to low-impact, selective logging, and can occur in sites having variable topographic and environmental properties such as mountain slopes and wetlands, a single approach is insufficient. The system introduced here combines time-series analysis of small objects identified in S1 data, i.e., segments containing linear features and apparent small-scale disturbances. A physical model is introduced for quantifying the size of small (upper-) canopy gaps. Deforestation detection was evaluated for several forest landscapes in the Amazon and Borneo. Using the default system settings, the false alarm rate (FAR) is very low (less than 1%), and the missed detection rate (MDR) varies between 1.9% ± 1.1% and 18.6% ± 1.0% (90% confidence level). For peatland landscapes, short radar detection delays up to several weeks due to high levels of soil moisture may occur, while, in comparison, for optical systems, detection delays up to 10 months were found due to cloud cover. In peat swamp forests, narrow linear canopy gaps (road and canal systems) could be detected with an overall accuracy of 85.5%, including many gaps barely visible on hi-res SPOT-6/7 images, which were used for validation. Compared to optical data, subtle degradation signals are easier to detect and are not quickly lost over time due to fast re-vegetation. Although it is possible to estimate an effective forest-cover loss, for example, due to selective logging, and results are spatiotemporally consistent with Sentinel-2 and TerraSAR-X reference data, quantitative validation without extensive field data and/or large hi-res radar datasets, such as TerraSAR-X, remains a challenge. Full article

Several satellite-based systems have been developed over the years to study and monitor thermal volcanic activity. Most of them use high temporal resolution satellite data, provided by sensors like the Moderate Resolution Imaging Spectroradiometer (MODIS) that if on the one hand guarantee a continuous monitoring of active volcanic areas on the other hand are less suited to map thermal anomalies, and to provide accurate information about their features. The Multispectral Instrument (MSI) and the Operational Land Imager (OLI), respectively, onboard the Sentinel-2 and Landsat-8 satellites, providing Short-Wave Infrared (SWIR) data at 20 m (MSI) and 30 m (OLI) spatial resolution, may make an important contribution in this area. In this work, we present the first Google Earth Engine (GEE) App to investigate, map and monitor volcanic thermal anomalies at global scale, integrating Landsat-8 OLI and Sentinel-2 MSI observations. This open tool, which implements the Normalized Hot spot Indices (NHI) algorithm, enables the analysis of more than 1400 active volcanoes, with very low processing times, thanks to the high GEE computational resources. Performance and limitations of the tool, such as its next upgrades, aiming at increasing the user-friendly experience and extending the temporal range of data analyses, are analyzed and discussed. Full article

Aerial photogrammetry by Unmanned Aerial Vehicles (UAVs) is a widespread method to perform mapping tasks with high-resolution to reconstruct three-dimensional (3D) building and façade models. However, the survey of Ground Control Points (GCPs) represents a time-consuming task, while the use of Real-Time Kinematic (RTK) drones allows for one to collect camera locations with an accuracy of a few centimeters. DJI Phantom 4 RTK (DJI-P4RTK) combines this with the possibility to acquire oblique images in stationary conditions and it currently represents a versatile drone widely used from professional users together with commercial Structure-from-Motion software, such as Agisoft Metashape. In this work, we analyze the architectural application of this drone to the photogrammetric modeling of a building with particular regard to metric survey specifications for cultural heritage for 1:20, 1:50, 1:100, and 1:200 scales. In particular, we designed an accuracy assessment test signalizing 109 points, surveying them with total station and adjusting the measurements through a network approach in order to achieve millimeter-level accuracy. Image datasets with a designed Ground Sample Distance (GSD) of 2 mm were acquired in Network RTK (NRTK) and RTK modes in manual piloting and processed both as single façades (S–F) and as an overall block (4–F). Subsequently, we compared the results of photogrammetric models generated in Agisoft Metashape to the Signalized Point (SP) coordinates. The results highlight the importance of processing an overall photogrammetric block, especially whenever part of camera locations exhibited a poorer accuracy due to multipath effects. No significant differences were found between the results of network real-time kinematic (NRTK) and real-time kinematic (RTK) datasets. Horizontal residuals were generally comparable to GNSS accuracy in NRTK/RTK mode, while vertical residuals were found to be affected by an offset of about 5 cm. We introduced an external GCP or used one SP per façade as GCP, assuming a poorer camera location accuracy at the same time, in order to fix this issue and comply with metric survey specifications for the widest architectural scale range. Finally, both S–F and 4–F projects satisfied the metric survey requirements of a scale of 1:50 in at least one of the approaches tested. Full article

Agriculture provides for the most basic needs of humankind: food and fiber. The introduction of new farming techniques in the past century (e.g., during the Green Revolution) has helped agriculture keep pace with growing demands for food and other agricultural products. However, further increases in food demand, a growing population, and rising income levels are likely to put additional strain on natural resources. With growing recognition of the negative impacts of agriculture on the environment, new techniques and approaches should be able to meet future food demands while maintaining or reducing the environmental footprint of agriculture. Emerging technologies, such as geospatial technologies, Internet of Things (IoT), Big Data analysis, and artificial intelligence (AI), could be utilized to make informed management decisions aimed to increase crop production. Precision agriculture (PA) entails the application of a suite of such technologies to optimize agricultural inputs to increase agricultural production and reduce input losses. Use of remote sensing technologies for PA has increased rapidly during the past few decades. The unprecedented availability of high resolution (spatial, spectral and temporal) satellite images has promoted the use of remote sensing in many PA applications, including crop monitoring, irrigation management, nutrient application, disease and pest management, and yield prediction. In this paper, we provide an overview of remote sensing systems, techniques, and vegetation indices along with their recent (2015–2020) applications in PA. Remote-sensing-based PA technologies such as variable fertilizer rate application technology in Green Seeker and Crop Circle have already been incorporated in commercial agriculture. Use of unmanned aerial vehicles (UAVs) has increased tremendously during the last decade due to their cost-effectiveness and flexibility in obtaining the high-resolution (cm-scale) images needed for PA applications. At the same time, the availability of a large amount of satellite data has prompted researchers to explore advanced data storage and processing techniques such as cloud computing and machine learning. Given the complexity of image processing and the amount of technical knowledge and expertise needed, it is critical to explore and develop a simple yet reliable workflow for the real-time application of remote sensing in PA. Development of accurate yet easy to use, user-friendly systems is likely to result in broader adoption of remote sensing technologies in commercial and non-commercial PA applications. Full article

Spaceborne remote sensing can track ecosystems changes thanks to continuous and systematic coverage at short revisit intervals. Active remote sensing from synthetic aperture radar (SAR) sensors allows day and night imaging as they are not affected by cloud cover and solar illumination and can capture unique information about its targets. However, SAR observations are affected by the coupled effect of viewing geometry and terrain topography. The study aims to assess the impact of global digital elevation models (DEMs) on the normalization of Sentinel-1 backscattered intensity and interferometric coherence. For each DEM, we analyzed the difference between orbit tracks, the difference with results obtained with a high-resolution local DEM, and the impact on land cover classification. Tests were carried out at two sites located in mountainous regions in Romania and Spain using the SRTM (Shuttle Radar Topography Mission, 30 m), AW3D (ALOS (Advanced Land Observation Satellite) World 3D, 30 m), TanDEM-X (12.5, 30, 90 m), and Spain national ALS (aerial laser scanning) based DEM (5 m resolution). The TanDEM-X DEM was the global DEM most suitable for topographic normalization, since it provided the smallest differences between orbital tracks, up to 3.5 dB smaller than with other DEMs for peak landform, and 1.4–1.9 dB for pit and valley landforms. Full article

Vertical profiles of wind, temperature, and moisture are essential to capture the kinematic and thermodynamic structure of the atmospheric boundary layer (ABL). Our goal is to use weather observing unmanned aircraft systems (WxUAS) to perform the vertical profiles by taking measurements while ascending through the ABL and subsequently descending to the Earth’s surface. Before establishing routine profiles using a network of WxUAS stations, the climatologies of the flight locations must be studied. This was done using data from the North American Regional Reanalysis (NARR) model. To begin, NARR data accuracy was verified against radiosondes. While the results showed variability in individual profiles, the detailed statistical analyses of the aggregated data suggested that the NARR model is a viable option for the study. Based on these findings, we used NARR data to determine fractions of successful hypothetical flights of vertical profiles across the state of Oklahoma given thresholds of visibility, cloud base level (CBL) height, and wind speed. CBL height is an important parameter because the WxUAS must stay below clouds for the flight restrictions being considered. For the purpose of this study, a hypothetical WxUAS flight is considered successful if the vehicle is able to reach an altitude corresponding to a pressure level of 600 hPa. Our analysis indicated the CBL height parameter hindered the fractions of successful hypothetical flights the most and the wind speed tolerance limited the fractions of successful hypothetical flights most strongly in the winter months. Northwest Oklahoma had the highest fractions of successful hypothetical flights, and the southeastern corner performs the worst in every season except spring, when the northeastern corner performed the worst. Future work will study the potential effect of topology and additional variables, such as amount of rainfall and temperature, on fractions of successful hypothetical flights by region of the state. Full article

Recently, deep learning has been reported to be an effective method for improving hyperspectral image classification and convolutional neural networks (CNNs) are, in particular, gaining more and more attention in this field. CNNs provide automatic approaches that can learn more abstract features of hyperspectral images from spectral, spatial, or spectral-spatial domains. However, CNN applications are focused on learning features directly from image data—while the intrinsic relations between original features, which may provide more information for classification, are not fully considered. In order to make full use of the relations between hyperspectral features and to explore more objective features for improving classification accuracy, we proposed feature relations map learning (FRML) in this paper. FRML can automatically enhance the separability of different objects in an image, using a segmented feature relations map (SFRM) that reflects the relations between spectral features through a normalized difference index (NDI), and it can then learn new features from SFRM using a CNN-based feature extractor. Finally, based on these features, a classifier was designed for the classification. With FRML, our experimental results from four popular hyperspectral datasets indicate that the proposed method can achieve more representative and objective features to improve classification accuracy, outperforming classifications using the comparative methods. Full article

The nonlinear radiation distortions (NRD) among multimodal remote sensing images bring enormous challenges to image registration. The traditional feature-based registration methods commonly use the image intensity or gradient information to detect and describe the features that are sensitive to NRD. However, the nonlinear mapping of the corresponding features of the multimodal images often results in failure of the feature matching, as well as the image registration. In this paper, a modality-free multimodal remote sensing image registration method (SRIFT) is proposed for the registration of multimodal remote sensing images, which is invariant to scale, radiation, and rotation. In SRIFT, the nonlinear diffusion scale (NDS) space is first established to construct a multi-scale space. A local orientation and scale phase congruency (LOSPC) algorithm are then used so that the features of the images with NRD are mapped to establish a one-to-one correspondence, to obtain sufficiently stable key points. In the feature description stage, a rotation-invariant coordinate (RIC) system is adopted to build a descriptor, without requiring estimation of the main direction. The experiments undertaken in this study included one set of simulated data experiments and nine groups of experiments with different types of real multimodal remote sensing images with rotation and scale differences (including synthetic aperture radar (SAR)/optical, digital surface model (DSM)/optical, light detection and ranging (LiDAR) intensity/optical, near-infrared (NIR)/optical, short-wave infrared (SWIR)/optical, classification/optical, and map/optical image pairs), to test the proposed algorithm from both quantitative and qualitative aspects. The experimental results showed that the proposed method has strong robustness to NRD, being invariant to scale, radiation, and rotation, and the achieved registration precision was better than that of the state-of-the-art methods. Full article

A satellite survey of the cumulative radiant emissions from electric lighting across China reveals a large radiance decline in lighting from December 2019 to February 2020—the peak of the lockdown established to suppress the spread of COVID-19 infections. To illustrate the changes, an analysis was also conducted on a reference set from a year prior to the pandemic. In the reference period, the majority (62%) of China’s population lived in administrative units that became brighter in March 2019 relative to December 2018. The situation reversed in February 2020, when 82% of the population lived in administrative units where lighting dimmed as a result of the pandemic. The dimming has also been demonstrated with difference images for the reference and pandemic image pairs, scattergrams, and a nightly temporal profile. The results indicate that it should be feasible to monitor declines and recovery in economic activity levels using nighttime lighting as a proxy. Full article

The information extraction of polarimetric synthetic aperture radar (PolSAR) images typically requires a great number of training samples; however, the training samples from historical images are less reusable due to the distribution differences. Consequently, there is a significant manual cost to collecting training samples when processing new images. In this paper, to address this problem, we propose a novel active transfer learning method, which combines active learning and the deep forest model to perform transfer learning. The main idea of the proposed method is to gradually improve the performance of the model in target domain tasks with the increase of the levels of the cascade structure. More specifically, in the growing stage, a new active learning strategy is used to iteratively add the most informative target domain samples to the training set, and the augmented features generated by the representation learning capability of the deep forest model are used to improve the cross-domain representational capabilities of the feature space. In the filtering stage, an effective stopping criterion is used to adaptively control the complexity of the model, and two filtering strategies are used to accelerate the convergence of the model. We conducted experiments using three sets of PolSAR images, and the results were compared with those of four existing transfer learning algorithms. Overall, the experimental results fully demonstrated the effectiveness and robustness of the proposed method. Full article

We live in a sphere that has unpredictable and multifaceted landscapes that make the risk arising from several incidences that are omnipresent. Floods and landslides are widespread and recurring hazards occurring at an alarming rate in recent years. The importance of this study is to produce multi-hazard exposure maps for flooding and landslides for the federal State of Salzburg, Austria, using the selected machine learning (ML) approach of support vector machine (SVM) and random forest (RF). Multi-hazard exposure maps were established on thirteen influencing factors for flood and landslides such as elevation, slope, aspect, topographic wetness index (TWI), stream power index (SPI), normalized difference vegetation index (NDVI), geology, lithology, rainfall, land cover, distance to roads, distance to faults, and distance to drainage. We classified the inventory data for flood and landslide into training and validation with the widely used splitting ratio, where 70% of the locations are used for training, and 30% are used for validation. The accuracy assessment of the exposure maps was derived through ROC (receiver operating curve) and R-Index (relative density). RF yielded better results for both flood and landslide exposure with 0.87 for flood and 0.90 for landslides compared to 0.87 for flood and 0.89 for landslides using SVM. However, the multi-hazard exposure map for the State of Salzburg derived through RF and SVM provides the planners and managers to plan better for risk regions affected by both floods and landslides. Full article

Spatiotemporal precipitation data is one of the essential components in modeling hydrological problems. Although the estimation of these data has achieved remarkable accuracy owning to the recent advances in remote-sensing technology, gaps remain between satellite-based precipitation and observed data due to the dependence of precipitation on the spatiotemporal distribution and the specific characteristics of the area. This paper presents an efficient approach based on a combination of the convolutional neural network and the autoencoder architecture, called the convolutional autoencoder (ConvAE) neural network, to correct the pixel-by-pixel bias for satellite-based products. The two daily gridded precipitation datasets with a spatial resolution of 0.25° employed are Asian Precipitation-Highly Resolved Observational Data Integration towards Evaluation (APHRODITE) as the observed data and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR) as the satellite-based data. Furthermore, the Mekong River basin was selected as a case study, because it is one of the largest river basins, spanning six countries, most of which are developing countries. In addition to the ConvAE model, another bias correction method based on the standard deviation method was also introduced. The performance of the bias correction methods was evaluated in terms of the probability distribution, temporal correlation, and spatial correlation of precipitation. Compared with the standard deviation method, the ConvAE model demonstrated superior and stable performance in most comparisons conducted. Additionally, the ConvAE model also exhibited impressive performance in capturing extreme rainfall events, distribution trends, and described spatial relationships between adjacent grid cells well. The findings of this study highlight the potential of the ConvAE model to resolve the precipitation bias correction problem. Thus, the ConvAE model could be applied to other satellite-based products, higher-resolution precipitation data, or other issues related to gridded data. Full article

Remote sensing is a useful tool for monitoring spatio-temporal variations of crop morphological and physiological status and supporting practices in precision farming. In comparison with multispectral imaging, hyperspectral imaging is a more advanced technique that is capable of acquiring a detailed spectral response of target features. Due to limited accessibility outside of the scientific community, hyperspectral images have not been widely used in precision agriculture. In recent years, different mini-sized and low-cost airborne hyperspectral sensors (e.g., Headwall Micro-Hyperspec, Cubert UHD 185-Firefly) have been developed, and advanced spaceborne hyperspectral sensors have also been or will be launched (e.g., PRISMA, DESIS, EnMAP, HyspIRI). Hyperspectral imaging is becoming more widely available to agricultural applications. Meanwhile, the acquisition, processing, and analysis of hyperspectral imagery still remain a challenging research topic (e.g., large data volume, high data dimensionality, and complex information analysis). It is hence beneficial to conduct a thorough and in-depth review of the hyperspectral imaging technology (e.g., different platforms and sensors), methods available for processing and analyzing hyperspectral information, and recent advances of hyperspectral imaging in agricultural applications. Publications over the past 30 years in hyperspectral imaging technology and applications in agriculture were thus reviewed. The imaging platforms and sensors, together with analytic methods used in the literature, were discussed. Performances of hyperspectral imaging for different applications (e.g., crop biophysical and biochemical properties’ mapping, soil characteristics, and crop classification) were also evaluated. This review is intended to assist agricultural researchers and practitioners to better understand the strengths and limitations of hyperspectral imaging to agricultural applications and promote the adoption of this valuable technology. Recommendations for future hyperspectral imaging research for precision agriculture are also presented. Full article

This work presents the latest calibration results for the Copernicus Sentinel-3A and -3B and the Jason-3 radar altimeters as determined by the Permanent Facility for Altimetry Calibration (PFAC) in west Crete, Greece. Radar altimeters are used to provide operational measurements for sea surface height, significant wave height and wind speed over oceans. To maintain Fiducial Reference Measurement (FRM) status, the stability and quality of altimetry products need to be continuously monitored throughout the operational phase of each altimeter. External and independent calibration and validation facilities provide an objective assessment of the altimeter’s performance by comparing satellite observations with ground-truth and in-situ measurements and infrastructures. Three independent methods are employed in the PFAC: Range calibration using a transponder, sea-surface calibration relying upon sea-surface Cal/Val sites, and crossover analysis. Procedures to determine FRM uncertainties for Cal/Val results have been demonstrated for each calibration. Biases for Sentinel-3A Passes No. 14, 278 and 335, Sentinel-3B Passes No. 14, 71 and 335, as well as for Jason-3 Passes No. 18 and No. 109 are given. Diverse calibration results by various techniques, infrastructure and settings are presented. Finally, upgrades to the PFAC in support of the Copernicus Sentinel-6 ‘Michael Freilich’, due to launch in November 2020, are summarized. Full article

Recent advances in deep learning have made it possible to use neural networks for the detection and classification of oceanic mesoscale eddies from satellite altimetry data. Various neural network models have been proposed in recent years to address this challenge, but they have been trained using different types of input data and evaluated using different performance metrics, making a comparison between them impossible. In this article, we examine the most common dataset and metric choices, by analyzing the reasons for the divergences between them and pointing out the most appropriate choice to obtain a fair evaluation in this scenario. Based on this comparative study, we have developed several neural network models to detect and classify oceanic eddies from satellite images, showing that our most advanced models perform better than the models previously proposed in the literature. Full article

The melting of the polar ice caps is considered to be an essential factor for global sea-level rise and has received significant attention. Quantitative research on ice cap mass changes is critical in global climate change. In this study, GRACE JPL RL06 data under the Mascon scheme based on the dynamic method were used. Greenland, which is highly sensitive to climate change, was selected as the study area. Greenland was divided into six sub-research regions, according to its watersheds. The spatial–temporal mass changes were compared to corresponding temperature and precipitation statistics to analyze the relationship between changes in ice sheet mass and climate change. The results show that: (i) From February 2002 to September 2019, the rate of change in the Greenland Ice Sheet mass was about −263 ± 13 Gt yr⁻¹ and the areas with the most substantial ice sheet loss and climate changes were concentrated in the western and southern parts of Greenland. (ii) The mass balance of the Greenland Ice Sheet during the study period was at a loss, and this was closely related to increasing trends in temperature and precipitation. (iii) In the coastal areas of western and southern Greenland, the rate of mass change has accelerated significantly, mainly because of climate change. Full article

A comprehensive understanding of temporal variability of subsurface soil moisture (SM) is paramount in hydrological and agricultural applications such as rainfed farming and irrigation. Since the SMOS (Soil Moisture and Ocean Salinity) mission was launched in 2009, globally available satellite SM retrievals have been used to investigate SM dynamics, based on the fact that useful information about subsurface SM is contained in their time series. SM along the depth profile is influenced by atmospheric forcing and local SM properties. Until now, subsurface SM was estimated by weighting preceding information of remotely sensed surface SM time series according to an optimized depth-specific characteristic time length. However, especially in regions with extreme SM conditions, the response time is supposed to be seasonally variable and depends on related processes occurring at different timescales. Aim of this study was to quantify the response time by means of the time lag between the trend series of satellite and in-situ SM observations using a Dynamic Time Warping (DTW) technique. DTW was applied to the SMOS satellite SM L4 product at 1 km resolution developed by the Barcelona Expert Center (BEC), and in-situ near-surface and root-zone SM of four representative stations at multiple depths, located in the Soil Moisture Measurements Station Network of the University of Salamanca (REMEDHUS) in Western Spain. DTW was customized to control the rate of accumulation and reduction of time lag during wetting and drying conditions and to consider the onset dates of pronounced precipitation events to increase sensitivity to prominent features of the input series. The temporal variability of climate factors in combination with crop growing seasons were used to indicate prevailing SM-related processes. Hereby, a comparison of long-term precipitation recordings and estimations of potential evapotranspiration (PET) allowed us to estimate SM seasons. The spatial heterogeneity of land use was analyzed by means of high-resolution images of Normalized Difference Vegetation Index (NDVI) from Sentinel-2 to provide information about the level of spatial representativeness of SMOS observations to each in-situ station. Results of the spatio-temporal analysis of the study were then evaluated to understand seasonally and spatially changing patterns in time lag. The time lag evolution describes a variable characteristic time length by considering the relevant processes which link SMOS and in-situ SM observation, which is an important step to accurately infer subsurface SM from satellite time series. At a further stage, the approach needs to be applied to different SM networks to understand the seasonal, climate- and site-specific characteristic behaviour of time lag and to decide, whether general conclusions can be drawn. Full article

Land surface temperature (LST) is an important variable in the physics of land–surface processes controlling the heat and water fluxes over the interface between the Earth’s surface and the atmosphere. Space-borne remote sensing provides the only feasible way for acquiring high-precision LST at temporal and spatial domain over the entire globe. Passive microwave (PMW) satellite observations have the capability to penetrate through clouds and can provide data under both clear and cloud conditions. Nonetheless, compared with thermal infrared data, PMW data suffer from lower spatial resolution and LST retrieval accuracy. Various methods for estimating LST from PMW satellite observations were proposed in the past few decades. This paper provides an extensive overview of these methods. We first present the theoretical basis for retrieving LST from PMW observations and then review the existing LST retrieval methods. These methods are mainly categorized into four types, i.e., empirical methods, semi-empirical methods, physically-based methods, and neural network methods. Advantages, limitations, and assumptions associated with each method are discussed. Prospects for future development to improve the performance of LST retrieval methods from PMW satellite observations are also recommended. Full article

Atmospheric desert-dust aerosol, primarily from north Africa, causes negative biases in remotely sensed climate data records of sea surface temperature (SST). Here, large-scale bias adjustments are deduced and applied to the v2 climate data record of SST from the European Space Agency Climate Change Initiative (CCI). Unlike SST from infrared sensors, SST measured in situ is not prone to desert-dust bias. An in-situ-based SST analysis is combined with column dust mass from the Modern-Era Retrospective analysis for Research and Applications, Version 2 to deduce a monthly, large-scale adjustment to CCI analysis SSTs. Having reduced the dust-related biases, a further correction for some periods of anomalous satellite calibration is also derived. The corrections will increase the usability of the v2 CCI SST record for oceanographic and climate applications, such as understanding the role of Arabian Sea SSTs in the Indian monsoon. The corrections will also pave the way for a v3 climate data record with improved error characteristics with respect to atmospheric dust aerosol. Full article

Accurate knowledge of the carbon budget on global and regional scales is critically important to design mitigation strategies aimed at stabilizing the atmospheric carbon dioxide (CO₂) emissions. For a better understanding of CO₂ variation trends over Asia, in this study, the column-averaged CO₂ dry air mole fraction (XCO₂) derived from the National Oceanic and Atmospheric Administration (NOAA) CarbonTracker (CT) was compared with that of Greenhouse Gases Observing Satellite (GOSAT) from September 2009 to August 2019 and with Orbiting Carbon Observatory 2 (OCO-2) from September 2014 until August 2019. Moreover, monthly averaged time-series and seasonal climatology comparisons were also performed separately over the five regions of Asia; i.e., Central Asia, East Asia, South Asia, Southeast Asia, and Western Asia. The results show that XCO₂ from GOSAT is higher than the XCO₂ simulated by CT by an amount of 0.61 ppm, whereas, OCO-2 XCO₂ is lower than CT by 0.31 ppm on average, over Asia. The mean spatial correlations of 0.93 and 0.89 and average Root Mean Square Deviations (RMSDs) of 2.61 and 2.16 ppm were found between the CT and GOSAT, and CT and OCO-2, respectively, implying the existence of a good agreement between the CT and the other two satellites datasets. The spatial distribution of the datasets shows that the larger uncertainties exist over the southwest part of China. Over Asia, NOAA CT shows a good agreement with GOSAT and OCO-2 in terms of spatial distribution, monthly averaged time series, and seasonal climatology with small biases. These results suggest that CO₂ can be used from either of the datasets to understand its role in the carbon budget, climate change, and air quality at regional to global scales. Full article

An accelerating trend of global urbanization accompanying population growth makes frequently updated land use and land cover (LULC) maps critical. LULC maps have been widely created through the classification of remotely sensed imagery. Maps of urban areas have been both dichotomous (urban or non-urban) and entailing of discrete urban types. This study incorporated multispectral built-up indices, designed to enhance satellite imagery, for introducing new urban classification schemes. The indices examined are the new built-up index (NBI), the built-up area extraction index (BAEI), and the normalized difference concrete condition index (NDCCI). Landsat Level-2 data covering the city of Miami, FL, USA was leveraged with geographic data from the Florida Geospatial Data Library and Florida Department of Environmental Protection to develop and validate new methods of supervised and unsupervised classification of urban area. NBI was used to extract discrete urban features through object-oriented image analysis. BAEI was found to possess properties for visualizing and tracking urban development as a low-high gradient. NDCCI was composited with NBI and BAEI as the basis for a robust urban intensity classification scheme superior to that of the United States Geological Survey National Land Cover Database 2016. BAEI, implemented as a shadow index, was incorporated in a novel infill geosimulation of high-rise construction. The findings suggest that the proposed classification schemes are advantageous to the process of creating more detailed cartography in response to the increasing global demand. Full article

Land cover classification with the focus on chlorophyll-rich vegetation detection plays an important role in urban growth monitoring and planning, autonomous navigation, drone mapping, biodiversity conservation, etc. Conventional approaches usually apply the normalized difference vegetation index (NDVI) for vegetation detection. In this paper, we investigate the performance of deep learning and conventional methods for vegetation detection. Two deep learning methods, DeepLabV3+ and our customized convolutional neural network (CNN) were evaluated with respect to their detection performance when training and testing datasets originated from different geographical sites with different image resolutions. A novel object-based vegetation detection approach, which utilizes NDVI, computer vision, and machine learning (ML) techniques, is also proposed. The vegetation detection methods were applied to high-resolution airborne color images which consist of RGB and near-infrared (NIR) bands. RGB color images alone were also used with the two deep learning methods to examine their detection performances without the NIR band. The detection performances of the deep learning methods with respect to the object-based detection approach are discussed and sample images from the datasets are used for demonstrations. Full article