Despite their critical importance for policy and research, high-resolution poverty data remain limited across much of Africa. Machine learning (ML) with earth observation (EO) imagery has recently emerged as a way to supplement these data by predicting (i.e., estimating) poverty where it has not been directly measured. Yet to be used reliably, decision-makers and analysts need assurances that they will not be misled by the errors in these predictions. To meet this need, we develop an uncertainty-aware EO-ML method for poverty mapping based on simultaneous quantile regression and a novel form of conformal prediction. Using a spatiotemporal transformer trained on sequences of Landsat and nighttime-light images, we produce prediction intervals for neighborhood-level International Wealth Index estimates across Africa which are statistically guaranteed to achieve their desired coverage rates. While our method's point-prediction performance matches the state of the art, its prediction intervals are wider than might be expected given its high R^2 of 0.75. However, other models of similar accuracy likely suffer from comparable uncertainty, pointing to an inherent limitation: even with its remarkably high explanatory power, EO-ML cannot naively be relied upon for policy-making, such as when designing poverty-targeting programs. To handle this challenge, we develop a procedure to efficiently allocate aid using both ground-truth surveys and model predictions while provably ensuring the risk of excluding eligible neighborhoods remains below a prespecified level. In simulations, this approach delivers substantially more aid per eligible recipient than other strategies, thereby demonstrating that EO-ML can indeed be a reliable supplement to traditional data sources—as long as methods

Accurate and consistent mapping of urban and rural areas is crucial for sustainable development, spatial planning, and policy design. It is particularly important in simulating the complex interactions between human activities and natural resources. Existing global urban-rural datasets such as such as GHSL-SMOD, GHS Degree of Urbanisation, and GRUMP are often spatially coarse, methodologically inconsistent, and poorly adapted to heterogeneous regions such as Africa, which limits their usefulness for policy and research. Their coarse grids and rule-based classification methods obscure small or informal settlements, and produce inconsistencies between countries. In this study, we develop a DeepLabV3-based deep learning framework that integrates multi-source data, including Landsat-8 imagery, VIIRS nighttime lights, ESRI Land Use Land Cover (LULC), and GHS-SMOD, to produce a 10 m resolution urban-rural map across the African continent from 2016 to 2022. The use of Landsat data also highlights the potential to extend this mapping approach historically, reaching back to the 1990s. The model employs semantic segmentation to capture fine-scale settlement morphology, and its outputs are validated using the Demographic and Health Surveys (DHS) dataset, which provides independent, survey-based urban-rural labels. The model achieves an overall accuracy of 65% and a Kappa coefficient of 0.47 at the continental scale, outperforming existing global products such as SMOD. The resulting High-Resolution Urban-Rural (HUR) dataset provides an open and reproducible framework for mapping human settlements, supporting UN Sustainable Development Goal (SDG) 11—Sustainable Cities and Communities—and enabling more context-aware analyses of Africa’s rapidly evolving settlement systems, which indirectly support other SDGs, such as SDG 1 (No Poverty), by distinguishing human settlement types. We release a continent-wide urban-rural dataset covering the period from 2016 to 2022, offering a new source for high-resolution settlement mapping in Africa.oxy_aqreply_end

This study examined wetland trends in the St. Lawrence Seaway (similar to 500,000 km(2)) in Canada over the past four decades. To this end, historical Landsat data within the Google Earth Engine (GEE) big geo data platform were processed. Reference samples were scrutinized using the Continuous Change Detection and Classification (CCDC) algorithm to identify spectrally unchanged samples. These spectrally unchanged samples were subsequently employed as training data within an object-based Random Forest (RF) model to generate wetland maps from 1984 to 2021. Subsequently, a change analysis was conducted to calculate the loss and gain of different wetland types. Overall, it was observed that approximately 45% (184,434 km(2)) and 55% (220,778 km(2)) of the entire study area are covered by wetland and non-wetland categories, respectively. It was also observed that 2.46% (12,495 km(2)) of the study area was changed during 40 years. Overall, there was a decline in the Bog and Fen classes, while the Marsh, Swamp, Forest, Grassland/Shrubland, Cropland, and Barren classes had an increase. Finally, the wetland gain and loss were 6,793 km(2) and 5,701 km(2), respectively. This study demonstrated that the use of Landsat data, along with advanced machine learning and GEE, could provide valuable assistance for wetland classification and change studies.

Ensuring accountability in emissions reductions is critical as many nations struggle to meet their Nationally Determined Contributions (NDCs) under the Paris Agreement. Traditional greenhouse gas (GHG) monitoring approaches are hindered by the long atmospheric lifetimes of key gases like carbon dioxide (CO2) and methane (CH4). In contrast, nitrogen dioxide (NO2) is a short-lived pollutant with high spatial and temporal variability, making it an effective proxy for tracking anthropogenic emissions. This study introduces a novel predictive framework integrating satellite-derived NO2 data with socioeconomic indicators, specifically the International Wealth Index (IWI). Using machine learning techniques, we establish IWI as a reliable predictor of NO2 column densities, demonstrating that socioeconomic development patterns significantly influence emission trends. Our convolutional neural network (CNN) model achieves an average predictive accuracy (R2 = 0.56) across the African continent, allowing for anticipatory analysis of emission hotspots. Forecasts for 2030 reveal substantial disparities between projected NO2 levels and NDC commitments, highlighting regions at risk of non-compliance. These findings emphasize the potential of integrating socioeconomic and environmental data to improve emissions monitoring, inform policy decisions, and enhance accountability in global climate action.

Human settlement areas significantly impact the environment, leading to changes in both natural and built environments. Comprehensive information on human settlements, particularly in urban areas, is crucial for effective sustainable development planning. However, urban land use investigations are often limited to two-dimensional building footprint maps, neglecting the three-dimensional aspect of building structures. This paper addresses this issue to contribute to Sustainable Development Goal 11, which focuses on making human settlements inclusive, safe, and sustainable. In this study, Sentinel-1 data are used as the primary source to estimate building heights. One challenge addressed is the issue of multiple backscattering in Sentinel-1’s signal, particularly in densely populated areas with high-rise buildings. To mitigate this, firstly, Sentinel-1 data from different directions, orbit paths, and polarizations are utilized. Combining ascending and descending orbits significantly improves estimation accuracy, and incorporating a higher number of paths provides additional information. However, Sentinel-1 data alone are not sufficiently rich at a global scale across different orbits and polarizations. Secondly, to enhance the accuracy further, Sentinel-1 data are corrected using nighttime light data as additional information, which shows promising results in addressing multiple backscattering issues. Finally, a deep learning model is trained to generate building height maps using these features, achieving a mean absolute error of around 2 m and a mean square error of approximately 13. The generalizability of this method is demonstrated in several cities with diverse built-up structures, including London, Berlin, and others. Finally, a building height map of Iran is generated and evaluated against surveyed buildings, showcasing its large-scale mapping capability.

Reducing global poverty is a key objective of the Sustainable Development Goals (SDGs). Achieving this requires high-frequency, granular data to capture neighborhood-level changes, particularly in data scarce regions such as low- and middle-income countries. To fill in the data gaps, recent computer vision methods combining machine learning (ML) with earth observation (EO) data to improve poverty estimation. However, while much progress have been made, they often omit intra-annual variations, which are crucial for estimating poverty in agriculturally dependent countries. We explored the impact of integrating intra-annual NDVI information with annual multi-spectral data on model accuracy. To evaluate our method, we created a simulated dataset using Landsat imagery and nighttime light data to evaluate EO-ML methods that use intra-annual EO data. Additionally, we evaluated our method against the Demographic and Health Survey (DHS) dataset across Africa. Our results indicate that integrating specific NDVI-derived features with multi-spectral data provides valuable insights for poverty analysis, emphasizing the importance of retaining intra-annual information.

Reducing global poverty, particularly in low- and middle-income countries, is a critical objective of the sustainable development goals (SDGs). To track progress toward these goals, high-frequency, granular geo-temporal data that capture changes at the neighborhood level is essential for researchers and policymakers. Recent advancements in methodology have combined machine learning (ML) and Earth observations (EOs) for poverty estimation, thereby addressing significant data gaps. However, a notable limitation of these EO-ML methods is their frequent deployment without a robust mechanism to quantify predictive uncertainty. Understanding this uncertainty is crucial for making informed decisions, effectively managing risks, and instilling confidence in users and stakeholders regarding the model's predictions. Although deep learning (DL) offers methods to quantify predictive uncertainties, their reliability is often constrained, failing to accurately reflect the underlying variations in predictions. Our proposed method aims to enhance confidence in predictive uncertainty without sacrificing accuracy. It begins by integrating an external model to explicitly capture data variability. Subsequently, we employ two orthogonal metrics-accuracy and uncertainty-to evaluate the influence of training data, especially in scenarios involving satellite imagery (e.g., selecting a subset of source domain countries for prediction in the target country). By applying these metrics, we formulate criteria to assess the importance of choosing specific countries from the source domain as training data. Our analysis highlights the effectiveness of this methodology in situations where the target country offers high-dimensional data, like satellite images, but faces a shortage of adequate training samples for DL models.

Using High Resolution (HR) and Very High Resolution (VHR) Remote Sensing (RS) images for post-disaster building damage assessment provides more information than low-resolution images. Consequently, a damage map is expected to be at building level, which requires both rooftop and facade information. Oblique imagery is therefore becoming increasingly popular for post-disaster analysis. However, oblique images are rarely available in the pre-disaster acquisition, and thus without any prior information it is a challenging task to assess the facade damage just from a post-disaster scene. As a solution, we aggregated information at the cluster level using pre-disaster neighborhood buildings' feature analysis, and thus the post-disaster building-level damage assessment is supported by the cluster-level information. Therefore, the proposed method, Intra-Cluster-Classification (ICC), uses hierarchical steps of unsupervised and supervised methods to detect damaged and undamaged areas within each cluster of buildings. The procedure is imple-mented on Google Earth Engine platform, and the results are evaluated using Hurricane Michael (2018) images. At the building-level, damage information is shown as a fractional number be-tween 0 and 1, with the higher number indicating more destruction. R-squared (R2) value is 0.9688 between actual and predicted damage scores. In addition, the Overall Accuracy (OA) and the Kappa coefficient (K) in the 4-class RS-scale are 83.2% and 0.7438, respectively. Furthermore, in 3-class RS-scale, the OA and K of our results are 91.08%, and 0.8582, respectively.

The Great Lakes (GL) wetlands support a variety of rare and endangered animal and plant species. Thus, wetlands in this region should be mapped and monitored using advanced and reliable techniques. In this study, a wetland map of the GL was produced using Sentinel-1/2 datasets within the Google Earth Engine (GEE) cloud computing platform. To this end, an object-based supervised machine learning (ML) classification workflow is proposed. The proposed method contains two main classification steps. In the first step, several non-wetland classes (e.g., Barren, Cropland, and Open Water), which are more distinguishable using radar and optical Remote Sensing (RS) observations, were identified and masked using a trained Random Forest (RF) model. In the second step, wetland classes, including Fen, Bog, Swamp, and Marsh, along with two non-wetland classes of Forest and Grassland/Shrubland were identified. Using the proposed method, the GL were classified with an overall accuracy of 93.6% and a Kappa coefficient of 0.90. Additionally, the results showed that the proposed method was able to classify the wetland classes with an overall accuracy of 87% and a Kappa coefficient of 0.91. Non-wetland classes were also identified more accurately than wetlands (overall accuracy = 96.62% and Kappa coefficient = 0.95).

The 11th Sustainable Development Goal (SDG) is focused on sustainable cities and communities and is closely related to other SDGs such as Good health and well-being (the third SDG) and climate action (the 13th SDG). However, the lack of data has made it difficult to evaluate the success of reaching these goals. To address this, a method is proposed in this paper to generate temporal building height maps and extract features from the 3D structure of urban areas to examine their relationship with environmental variables, acquired from remote sensing satellites. Therefore, no survey data is required from the study area. Building height map is generated by processing Sentinel-1, Sentinel-2, and Nighttime light data by a UNet-based deep model. The results showed significant improvements in Mean Square Errors compared to available building maps in Berlin and London. In the second step, several features were extracted from the 3D structure of urban areas, and their relationship with environmental variables such as atmosphere contents from Sentinel-5 data and Urban Heat Island (UHI) from MODIS was examined via shallow regression models. The spatial study shows high correlation between each environmental variable and height map features in a neighborhood, with R2 scores of 0.78, 0.94, 0.92, 0.7, and 0.88 for CO2, CO, NO2, SO2, and UHI, respectively. It is found the environmental parameters are shaped by the collective building heights within a specific neighborhood, rather than hinging on the individual building heights at the sampling site. Furthermore, spatial resolution plays a significant role. In the case of the MODIS-based heat island map, a 3 km neighborhood yields a high R2-score, whereas when utilizing Sentinel-5 data, it is advisable to employ a larger neighborhood. Furthermore, the temporal study shows even higher R2 scores than the spatial domain, indicating the temporal reliability of the proposed method. The findings of this study can be used by governors and decision makers for sustainable urban development.

Soil moisture content (SMC) plays a critical role in soil science via its influences on agriculture, water resources management, and climate conditions. There is broad interest in finding relationships between groundwater recharge, soil characteristics, and plant properties for the quantification of SMC. The objective of this study was to assess the potential of optical satellite imagery for estimating the SMC over cropland areas. For this purpose, we collected 394 soil samples as targets in Gonbad-e Kavus in the Golestan province in the north of Iran, where a variety of crop types are cultivated. As input data, we first computed several spectral indices from Sentinel 2 (S2) and Landsat 8 (L8) images, such as the Normalized Difference Water Index (NDWI), Modified Normalized Difference Water Index (MNDWI), and Normalized Difference Salinity Index (NDSI), and then analyzed their relationships with surveyed SMC using four machine learning regression algorithms: random forests (RFs), XGBoost, extra tree decision (EDT), and support vector machine (SVM). Results revealed a high and rather similar correlation between the spectral indices and measured SMC values for both S2 and L8 data. The EDT regression algorithm yielded the highest accuracy, with an R-2 = 0.82, MAE = 3.74, and RMSE = 1.08 for S2 and R-2 = 0.88, RMSE = 2.42, and MAE = 1.08 for L8 images. Results also revealed that MNDWI, NDWI, and NDSI responded most sensitively to SMC estimation.

To combat poor health and living conditions, policymakers in Africa require temporally and geographically granular data measuring economic well-being. Machine learning (ML) offers a promising alternative to expensive and time-consuming survey measurements by training models to predict economic conditions from freely available satellite imagery. However, previous efforts have failed to utilize the temporal information available in earth observation (EO) data, which may capture developments important to standards of living. In this work, we develop an EO-ML method for inferring neighborhood-level material-asset wealth using multi-temporal imagery and recurrent convolutional neural networks. Our model outperforms state-of-the-art models in several aspects of generalization, explaining 72% of the variance in wealth across held-out countries and 75% held-out time spans. Using our geographically and temporally aware models, we created spatio-temporal material-asset data maps covering the entire continent of Africa from 1990 to 2019, making our data product the largest dataset of its kind. We showcase these results by analyzing which neighborhoods are likely to escape poverty by the year 2030, which is the deadline for when the Sustainable Development Goals (SDG) are evaluated.

Shadow detection provides worthwhile information for remote sensing applications, e.g. building height estimation. Shadow areas are formed in the opposite side of the sunlight radiation to tall objects, and thus, solar illumination angle is required to find probable shadow areas. In recent years, Very High Resolution (VHR) imagery provides more detailed data from objects including shadow areas. In this regard, the motivation of this paper is to propose a reliable feature, Shadow Low Gradient Direction (SLGD), to automatically determine shadow and solar illumination direction in VHR data. The proposed feature is based on inherent spatial feature of fine-resolution shadow areas. Therefore, it can facilitate shadow-based operations, especially when the solar illumination information is not available in remote sensing metadata. Shadow intensity is supposed to be dependent on two factors, including the surface material and sunlight illumination, which is analyzed by directional gradient values in low gradient magnitude areas. This feature considers the sunlight illumination and ignores the material differences. The method is fully implemented on the Google Earth Engine cloud computing platform, and is evaluated on VHR data with 0.3m resolution. Finally, SLGD performance is evaluated in determining shadow direction and compared in refining shadow maps.

Wetlands provide many benefits, such as water storage, flood control, transformation and retention of chemicals, and habitat for many species of plants and animals. The ongoing degradation of wetlands in the Great Lakes basin has been caused by a number of factors, including climate change, urbanization, and agriculture. Mapping and monitoring wetlands across such large spatial and temporal scales have proved challenging; however, recent advancements in the accessibility and processing efficiency of remotely sensed imagery have facilitated these applications. In this study, the historical Landsat archive was first employed in Google Earth Engine (GEE) to classify wetlands (i.e., Bog, Fen, Swamp, Marsh) and non-wetlands (i.e., Open Water, Barren, Forest, Grassland/Shrubland, Cropland) throughout the entire Great Lakes basin over the past four decades. To this end, an object-based supervised Random Forest (RF) model was developed. All of the produced wetland maps had overall accuracies exceeding 84%, indicating the high capability of the developed classification model for wetland mapping. Changes in wetlands were subsequently assessed for 17 time intervals. It was observed that approximately 16% of the study area has changed since 1984, with the highest increase occurring in the Cropland class and the highest decrease occurring in the Forest and Marsh classes. Forest mostly transitioned to Fen, but was also observed to transition to Cropland, Marsh, and Swamp. A considerable amount of the Marsh class was also converted into Cropland.

Global environmental changes have increased the frequency of natural disasters and the demand for rapid postdisaster mapping. In this regard, remote sensing (RS) is a leading technology because it provides consistent near-real-time images. In this chapter, we studied different disasters, Joplin MO Tornado (2011), Hurricane Harvey (2017), and Hurricane Michael (2018), using satellite sensors such as Landsat 5 and Sentinel 2 and airborne imagery acquired within the National Agriculture Imagery Program (NAIP) and by the National Oceanic and Atmospheric Administration (NOAA). We compared different RS methods, such as pixel- and object-based classification techniques and spectral/spatial feature analysis to compare the potential of vertical and oblique images to produce regional- and building-level damage maps. We illustrated several large-scale and zoomed scenes for visual interpretation and the corresponding assessment analysis. Finally, the further development of RS technology and its effect on the development of the algorithm are discussed.

Land Use/Land Cover (LULC) maps can be effectively produced by cost-effective and frequent satellite observations. Powerful cloud computing platforms are emerging as a growing trend in the high utilization of freely accessible remotely sensed data for LULC mapping over large-scale regions using big geodata. This study proposes a workflow to generate a 10 m LULC map of Europe with nine classes, ELULC-10, using European Sentinel-1/-2 and Landsat-8 images, as well as the LUCAS reference samples. More than 200 K and 300 K of in situ surveys and images, respectively, were employed as inputs in the Google Earth Engine (GEE) cloud computing platform to perform classification by an object-based segmentation algorithm and an Artificial Neural Network (ANN). A novel ANN-based data preparation was also presented to remove noisy reference samples from the LUCAS dataset. Additionally, the map was improved using several rule-based post-processing steps. The overall accuracy and kappa coefficient of 2021 ELULC-10 were 95.38% and 0.94, respectively. A detailed report of the classification accuracies was also provided, demonstrating an accurate classification of different classes, such as Woodland and Cropland. Furthermore, rule-based post processing improved LULC class identifications when compared with current studies. The workflow could also supply seasonal, yearly, and change maps considering the proposed integration of complex machine learning algorithms and large satellite and survey data.

Remote sensing (RS) oblique imagery provides valuable information of buildings' facades. Facade detection using spectral-, spatial-, and texture-only features does not accurately separate facade and nonfacade regions in a single-view oblique image. Therefore, a facade index and a unimodal thresholding method were proposed to characterize and detect facade regions. This new index, named the probability-spatial facade index (PSFI), first highlighted facade areas. Then, the facade map was created through a histogram-based thresholding. In unimodal histograms, the thresholding procedure was according to the roots of the second derivative of a fourth-degree polynomial model that is fitted to the PSFI's histogram. All the steps of the proposed method were implemented in the Google Earth Engine cloud computing platform and could automatically handle very high resolution oblique imagery (in terms of both angle and direction) without any limitations. Furthermore, various high- and low-rise buildings could be effectively processed without any assumptions about the structure of facades. The evaluation showed the high performance of the proposed method in different test areas, in which the average overall accuracies in distinguishing facade and nonfacade regions and in separating facade and rooftop regions were 97% and 86%, respectively. (C) 2021 Society of Photo-Optical Instrumentation Engineers (SPIE)

In this study, wetland trends in Alberta were investigated in the past four decades using Landsat satellite imagery to produce updated information about wetland changes and to prevent further degradation of these valuable natural resources. All the processing steps and analyses were conducted in Google earth engine (GEE) to produce 16 wetland maps from 1984 to 2020. A comprehensive change analysis showed 1 ) approximately 18% of the province was subjected to change; 2 ) in terms of wetland classes, there was a decreasing trend for the Shallow Water and Swamp classes and an increasing trend for the Fen and Marsh classes; 3 ) in terms of nonwetland classes, there was a considerable decreasing trend for the Forest class and increasing trend for the Grassland/Shrubland class; 4 ) wetland loss was approximately 22 000 km 2 , which was mainly due to the conversion of wetlands to Forest and Grassland/Shrubland; 5) wetland gain was approximately 24 000 km 2 , which was mainly due to the conversion from the Forest class to wetlands, especially the Swamp and Fen classes; 6) the highest class transition was from Cropland to Grassland/Shrubland and vice versa (29 000 km 2 ), from Forest to different wetland classes (18 000 km 2 ), and from Fen to Forest (6000 km 2 ). In summary, the results of this study provided the first comprehensive information on wetland trends in Alberta over the past 37 years and will assist policymakers to adjust the required/established policies to mitigate the potential wetland changes due to anthropogenic activities and climate-related events.

The first Canadian wetland inventory (CWI) map, which was based on Landsat data, was produced in 2019 using the Google Earth Engine (GEE) big data processing platform. The proposed GEE-based method to create the preliminary CWI map proved to be a cost, time, and computationally efficient approach. Although the initial effort to produce the CWI map was valuable with a 71% overall accuracy (OA), there were several inevitable limitations (e.g., low-quality samples for the training and validation of the map). Therefore, it was important to comprehensively investigate those limitations and develop effective solutions to improve the accuracy of the Landsat-based CWI (L-CWI) map. Over the past year, the L-CWI map was shared with several governmental, academic, environmental nonprofit, and industrial organizations. Subsequently, valuable feedback was received on the accuracy of this product by comparing it with various in situ data, photo-interpreted reference samples, land cover/land use maps, and high-resolution aerial images. It was generally observed that the accuracy of the L-CWI map was lower relative to the other available products. For example, the average OA in four Canadian provinces using in situ data was 60%. Moreover, including reliable in situ data, using an object-based classification method, and adding more optical and synthetic aperture radar datasets were identified as the main practical solutions to improve the CWI map in the future. Finally, limitations and solutions discussed in this study are applicable to any large-scale wetland mapping using remote sensing methods, especially to CWI generation using optical satellite data in GEE.

Accurate information about the location, extent, and type of Land Cover (LC) is essential for various applications. The only recent available country-wide LC map of Iran was generated in 2016 by the Iranian Space Agency (ISA) using Moderate Resolution Imaging Spectroradiometer (MODIS) images with a considerably low accuracy. Therefore, the production of an up-to-date and accurate Iran-wide LC map using the most recent remote sensing, machine learning, and big data processing algorithms is required. Moreover, it is important to develop an efficient method for automatic LC generation for various time periods without the need to collect additional ground truth data from this immense country. Therefore, this study was conducted to fulfill two objectives. First, an improved Iranian LC map with 13 LC classes and a spatial resolution of 10 m was produced using multi-temporal synergy of Sentinel-1 and Sentinel-2 satellite datasets applied to an object-based Random forest (RF) algorithm. For this purpose, 2,869 Sentinel-1 and 11,994 Sentinel-2 scenes acquired in 2017 were processed and classified within the Google Earth Engine (GEE) cloud computing platform allowing big geospatial data analysis. The Overall Accuracy (OA) and Kappa Coefficient (KC) of the final Iran-wide LC map for 2017 was 95.6% and 0.95, respectively, indicating the considerable potential of the proposed big data processing method. Second, an efficient automatic method was developed based on Sentinel-2 images to migrate ground truth samples from a reference year to automatically generate an LC map for any target year. The OA and KC for the LC map produced for the target year 2019 were 91.35% and 0.91, respectively, demonstrating the efficiency of the proposed method for automatic LC mapping. Based on the obtained accuracies, this method can potentially be applied to other regions of interest for LC mapping without the need for ground truth data from the target year.