Search

The recent paper by Morillas et al. [Morillas, L. et al. Using radiometric surface temperature for surface energy flux estimation in Mediterranean drylands from a two-source perspective, Remote Sens. Environ. 136, 234–246, 2013] evaluates the two-source model (TSM) of Norman et al. (1995) with revisions by Kustas and Norman (1999) over a semiarid tussock grassland site in southeastern Spain. The TSM - in its current incarnation, the two-source energy balance model (TSEB) - was applied to this landscape using ground-based infrared radiometer sensors to estimate both the composite surface radiometric temperature and component soil and canopy temperatures. Morillas et al. (2013) found the TSEB model substantially underestimated the sensible H (and overestimated the latent heat LE ) fluxes. Using the same data set from Morillas et al. (2013), we were able to confirm their results. We also found energy transport and exchange behavior derived from primarily the observations themselves to differ significantly from a number of prior studies using land surface temperature for estimating heat fluxes with one-source modeling approaches in semi-arid landscapes. However, revisions to key vegetation inputs to TSEB and the soil resistance formulation resulted in a significant reduction in the bias and root mean square error (RMSE) between model output of H and LE and the measurements compared to the prior results from Morillas et al. (2013). These included more representative ground-based vegetation greenness and local leaf area index values as well as modifications to the coefficients of the soil resistance formulation to account for the very rough (rocky) soil surface conditions with a clumped canopy. This indicates that both limitations in remote estimates of biophysical indicators of the canopy at the site and the lack of adjustment in soil resistance formulation to account for site specific characteristics, contributed to the earlier findings of Morillas et al. (2013). This suggests further studies need to be conducted to reduce the uncertainties in the vegetation and land surface temperature input data in order to more accurately assess the effects of the transport exchange processes of this Mediterranean landscape on TSEB formulations. [ABSTRACT FROM AUTHOR]

Mapping of highly dynamic changes in land use and land cover (LULC) can be hindered by various cloudy conditions with optical satellite images. These conditions result in discontinuities in high-temporal-density LULC mapping. In this paper, we developed an integrated time series mapping method to enhance the LULC mapping accuracy and frequency in cloud-prone areas by incorporating spectral-indices-fused deep models and time series reconstruction techniques. The proposed method first reconstructed cloud-contaminated pixels through time series filtering, during which the cloud masks initialized by a deep model were refined and updated during the reconstruction process. Then, the reconstructed time series images were fed into a spectral-indices-fused deep model trained on samples collected worldwide for classification. Finally, post-classification processing, including spatio-temporal majority filtering and time series refinement considering land–water interactions, was conducted to enhance the LULC mapping accuracy and consistency. We applied the proposed method to the cloud- and rain-prone Pearl River Delta (i.e., Guangdong–Hong Kong–Macao Greater Bay Area, GBA) and used time series Sentinel-2 images as the experimental data. The proposed method enabled seamless LULC mapping at a temporal frequency of 2–5 days, and the production of 10 m resolution annual LULC products in the GBA. The assessment yielded a mean overall accuracy of 87.01% for annual mapping in the four consecutive years of 2019–2022 and outperformed existing mainstream LULC products, including ESA WorldCover (83.98%), Esri Land Cover (85.26%), and Google Dynamic World (85.06%). Our assessment also reveals significant variations in LULC mapping accuracies with different cloud masks, thus underscoring their critical role in time series LULC mapping. The proposed method has the potential to generate seamless and near real-time maps for other regions in the world by using deep models trained on datasets collected globally. This method can provide high-quality LULC data sets at different time intervals for various land and water dynamics in cloud- and rain-prone regions. Notwithstanding the difficulties of obtaining high-quality LULC maps in cloud-prone areas, this paper provides a novel approach for the mapping of LULC dynamics and the provision of reliable annual LULC products. • Novel LULC mapping via indices-fused deep models and time series reconstruction. • High-temporal-density 10 m seamless LULC mapping in cloudy and rainy areas (PRD). • Overall mapping accuracy yielded up to 87.01%, outperforming existing LULC products. • Potential to generate seamless near real-time maps and high-quality LULC data sets. [ABSTRACT FROM AUTHOR]

This paper is the second part of companion papers describing the development of GRASP approach for aerosol and surface retrieval from Sentinel-5P/TROPOMI. Here we focus on the S5P/TROPOMI GRASP aerosol and surface products global validation and systematic intercomparison with other products from independent instruments and algorithms. Specifically, we have validated the S5P/TROPOMI GRASP, Suomi-NPP/VIIRS DB and MODIS/TERRA DT + DB aerosol products with the ground-based AERONET referenced measurements using the same methodology and intercompare the validation results. In addition, the global pixel-to-pixel intercomparisons of the aerosol products (AOD, fine/coarse mode AOD and SSA) are performed over different surfaces, i.e., ocean and land surface with different NDVIs. Besides, we compared the S5P/TROPOMI GRASP, MODIS MCD43 surface BRDF/albedo as well as OMI, GOME-2 and SCIAMACHY Lambertian-Equivalent Reflectivity (LER) albedo climatology developed by Royal Netherlands Meteorological Institute (KNMI) with the surface reference dataset generated based on the synergetic retrieval of AERONET and S5P/TROPOMI measurements. Finally, the intercomparisons of the surface BRDF and albedo datasets were performed globally at the UV, VIS, NIR and SWIR parts of the spectrum. Overall, generally good agreement was observed between independent aerosol and surface datasets with a high percentage of pixels satisfying the Optimal and Target requirements. We would emphasize two advantages for TROPOMI/GRASP aerosol and surface products: (i) it provides spectral AOD together with detailed aerosol properties, such as fine/coarse mode AOD, spectral AAOD and SSA at UV, VIS, NIR and SWIR wavelengths, which are important for constraining aerosol environmental and climate effects; (ii) the TROPOMI/GRASP aerosol and surface products are globally retrieved simultaneously in a fully consistent manner. • S5P/TROPOMI GRASP aerosol and surface products are validated and intercompared. • A remarkable agreement was found between TROPOMI/GRASP and VIIRS/DB spectral AOD. • TROPOMI/GRASP BRDF product show good agreement with MODIS MCD43 product globally. • TROPOMI/GRASP provides AOD, AODF, SSA, and BRDF, albedos at UV, VIS, NIR and SWIR. [ABSTRACT FROM AUTHOR]

With the upcoming spaceborne synthetic aperture radar (SAR) missions (BIOMASS, LuTan-1, NISAR, and TanDEM-L), it will become possible to extract vegetation height at a global scale by utilizing spaceborne low-frequency (L- and P-band) polarimetric synthetic aperture radar interferometry (PolInSAR) data. However, in the context of single-baseline parameter retrieval by the random volume over ground (RVoG) model, three main error sources that affect the inversion accuracy should be carefully considered, i.e., the ground scattering contribution, the spatial baseline configuration, and the temporal decorrelation (the main part of non-volume decorrelation). To make the estimation more reliable, several kinds of multibaseline PolInSAR inversion methods have been proposed over the past few years and have achieved improved inversion performances. Dual-baseline inversion effectively avoids the ambiguity of the ground contribution, whereas the performance is highly dependent on the appropriate combination of two spatial baselines as well as the mitigation of the non-volume decorrelation. In this study, we conducted in-depth research into the effect of these influencing factors on dual-baseline inversion, aiming to provide effective guidance on dual-baseline combination selection among multibaseline data. Accordingly, a novel multibaseline inversion scheme (MBLFPI) suitable for low-frequency PolInSAR data is proposed in this paper. The significant advantage of the new method is that the three aforementioned error sources can be taken into account simultaneously, without relying on external data. The proposed scheme was validated using L- and P-band SAR data acquired by the DLR's E-SAR/F-SAR and ONERA's SETHI systems, as well as corresponding light detection and ranging (LiDAR) data collected during the BioSAR-2008, AfriSAR-2016, and TropiSAR-2009 campaigns. A series of experiments was performed to evaluate the applicability and generalizability of the proposed method. The results showed that this innovative scheme produced forest height maps with a root-mean-square error (RMSE) of 2.37 m (R2 = 0.88) and 3.13–4.43 m (R2 = 0.35–0.94) in the L-band and P-band scenarios (boreal and tropical forest), respectively, indicating a significant improvement over the three conventional multibaseline methods and pure dual-baseline inversion. The comprehensive analysis provided in this paper should assist with and provide strong support for SAR system and mission design, and the proposed scheme could be considered a promising way for future spaceborne missions to invert vegetation parameters at a global scale. • A generalizable MB method is proposed for global forest mapping by BIOMASS and LT-1. • Three general issues in low-frequency PolInSAR inversion are simultaneously solved. • B1 is more sensitive than B2 in dual-baseline PolInSAR geometric inversion method. • An error model of dual-baseline forest height inversion is constructed. • Findings help for system design of spaceborne low-frequency PolInSAR missions. [ABSTRACT FROM AUTHOR]

Remote sensing of urban environments has unveiled a significant shift from single-city investigations to the inclusion of multiple cities. Originated from the ideas of the Remote Sensing of Environment special issue entitled "Remote Sensing of the Urban Environment: Beyond the Single City," this paper offers a comprehensive examination of the state of the science in multi-city remote sensing, and aims at fostering the rapid advancement of this emerging field to address global sustainability challenges and support knowledge development needed for a new discipline – urban sustainability science (USS). Through a synthesized review of eight key research fields within urban remote sensing [i.e., land use and land cover (LULC) and change, urban vertical structure, urban heat islands, hazards, energy use and emissions, air quality, carbon budgets, and green space], the paper provides insights into the underlying rationale for conducting multi-city studies, the criteria employed in the selection of cities, the societal applications, as well as the opportunities and future directions for expanding the scope of assessments in multi-city remote sensing. • We provide a comprehensive review of multi-city remote sensing. • We systematically analyze eight key research fields in urban remote sensing. • We provide insights into the underlying rationale behind conducting multi-city studies. • We summarize the criteria employed for city selection and the societal applications. • We further discusses opportunities and future directions in multi-city remote sensing. [ABSTRACT FROM AUTHOR]

Dust plays a key role in many Earth system processes and is ubiquitous in the Martian atmosphere. Various intensive field campaigns, laboratory analyses, space-based remote sensing missions, and global modeling efforts aim to characterize dust optical properties. This is a bountiful time for dust scientists, and yet the interpretation of retrievals and comparison to models remains complicated by various conflicting assumptions that are part of each algorithm. For example, the conversion of satellite radiance measurements into products like aerosol optical depth for model evaluation depends upon aerosol properties like particle size and shape that are often prescribed and not part of the retrieval. Conversely, the model calculation of aerosol optical depth often uses different assumptions. The goal of this review is to first document algorithmic assumptions by various satellite retrieval products and models, and identify where there is consistency and where there are differences. In general, the differences documented in this paper reflect uncertainties resulting from incomplete observational characterization of dust aerosols and limitations in our understanding. Second, we note what observations might reduce uncertainties in our knowledge and bring greater consistency to retrievals and models, allowing for a more rigorous and harmonious comparison. The lack of comprehensive and realistic shape models for dust is an outstanding issue, such that closure between forward modeling from particle refractive index, shape, and size and observed optical properties cannot be achieved. Limitations in the computational methods that must be applied to model scattering from complex shapes also makes accurate optical modeling for dust challenging. Field observations indicate the persistence of coarse and giant dust particles at higher altitudes and farther downwind from their source than previously expected. Remote sensing retrieval algorithms based on observations at visible wavelengths have limited sensitivity to these particles and generally do not consider them, although a recent product based on longwave radiances is encouraging. Current measurements of the refractive index of bulk dust and fundamental dust minerology components such as hematite vary widely, inhibiting attempts to represent the variability in dust optical properties and forcing, as expected from different major dust source regions on Earth that have varying mineralogical composition. Some remote sensing retrieval algorithms allow for limited refractive index variability in their inversion solutions through mixing with other fine mode aerosol models, or optimizing the single scattering albedo, but Earth system models surveyed for this paper assume a globally uniform, size-invariant refractive index. Although no Martian dust samples have yet been returned to Earth, remote sensing observations indicate that Martian dust is globally homogenous in composition, and a single spectral refractive index assumption has been widely adopted to represent Martian dust. The lack of comprehensive, statistically representative measurements of dust particle microphysical properties (size distribution, morphology, complex index of refraction spectra, internal structure heterogeneity), and the resulting optical properties, limits our ability to verify the fidelity of these assumptions. A chain of measurements is needed, ranging from characterizing individual dust mineralogy components (e.g., pure hematite and goethite) to in situ sampling of complex atmospheric aerosol mixtures. Such results could be applied to both remote sensing retrievals that characterize the optical properties of the total aerosol burden in the atmosphere from total radiance measurements, and to global models that represent the total aerosol burden in the atmosphere by building it up from the balance of individual aerosol sources and sinks. • Algorithmic assumptions by various satellite retrieval products and models • No closure between forward modeled and observed dust optical properties. • Retrieval algorithms using visible channels have low sensitivity to large particles. • Field observations indicate the persistence of coarse and giant dust particles. • Retrieval algorithms and global models make different refractive index assumptions. [ABSTRACT FROM AUTHOR]

This paper provides a review and summary status of the research underway by the NASA Terra Aqua Suomi-NPP Land Discipline Team to provide continuity of global land data products from the NASA Moderate resolution Imaging Spectroradiometer (MODIS) to the Visible Infrared Imaging Radiometer Suite (VIIRS). The two MODIS instruments on the NASA Earth Observing System Terra (morning overpass) and Aqua (afternoon overpass) platforms have provided more than twenty years of data. The peer-reviewed land products generated from MODIS are now being transitioned to production using VIIRS inputs, with the intention of providing dynamic continuity for the Aqua observations. As part of that process, the products from the two instruments are undergoing intercomparison and evaluation. These results are provided where available and show promising levels of agreement and accuracy in all cases. The paper also offers options for establishing continuity of Terra MODIS data products. • Most MODIS data products can achieve continuity with VIIRS data products. • Continuity can access improved data products and new supplementary data. • There are options for MODIS continuity data sources other than VIIRS. [ABSTRACT FROM AUTHOR]

Aerial moving target detection is an important technology to ensure the safety of civil flight and airspace. Different from ground-based active radar and space-based visible imager, on-orbit thermal infrared sensor (TIS), which records the temperature difference between background and targets, is proved to be a stable means of monitoring the aircraft throughout the day. Limited by low spatial resolution and dataset scarcity, there are few existing the research of space-based aircraft detection based on TIS. In this paper, a new multi-model joint method to detect on-orbit flying aircrafts with only 1 × 1 to 2 × 2 pixels is proposed and validated. And the first three-band thermal infrared flying aircrafts dataset (TIFAD) with 30 m spacial resolution is produced from SDGSAT-1 thermal infrared images, including the analysis of its infrared feature. Since the point target cannot by distinguished by the shape, all samples are selected by an improved aircraft contrails detection algorithm based on progressive probabilistic Hough transform, from more than 325GB SDGSAT-1 TIS remote sensing images (RSIs). And training by TIFAD, state-of-the-art YOLOv8 algorithm is applied to more than 17,767 medium spatial resolution thermal infrared images, and the potential aircrafts (with contrails or non-contrails) in RSIs are effectively detected. Moreover, the giga floating point operations (GFLOPs) of detecting flying targets from RSIs with 300 km width range is only 8.9. This paper is believed to provide a technical support of detecting aircraft targets around the globe throughout the day, and supplement on-orbit flying targets dataset of TIS images. • Contrails cloud is joint to detect on-orbit flying aircrafts. • Thermal infrared (TI) is proved to be a stable way to monitor aircraft all the day. • The flying aircrafts data is labeled from SDGSAT-1 TI images. • The potential aircrafts with 1 × 1–2 × 2 pixels are effectively detected. [ABSTRACT FROM AUTHOR]

Hyperspectral images are increasingly being used in classification and identification. Data users prefer hyperspectral imagery with high spatial resolution, finer spectral resolution, and high signal-to-noise ratio (SNR). However, tradeoffs exist in these core parameters in imagery acquired by different hyperspectral sensor systems. Data users may find it difficult to utilize all the advantages of hyperspectral imagery. How to select hyperspectral data with optimal parameter configuration has been one of the essential issues for data users, which also affects the back-end applications. With advancements in computer science, various artificial intelligence algorithms from conventional machine learning to deep learning have been utilized for hyperspectral images classification and identification. Few researchers study the mechanism between the core parameters of hyperspectral imaging spectrometers and advanced artificial intelligence algorithms, which affects the application efficiency and accuracy. In this paper, we delved into the evolution of machine learning and deep learning models applied to imagery acquired by different hyperspectral sensor systems having different SNR, spectral, and spatial resolutions. Additionally, we also considered the tradeoffs among the core parameters of hyperspectral imagers. We used two conventional machine learning models, including the classification and regression tree (CART) and random forest (RF), two deep learning methods based on convolution neural network architectures—3D convolutional neural network (3D-CNN) and hamida, and two deep learning methods based on vision transformers architectures—transformer models vision transformer (VIT) and robust vision transformer (RVT), to compare the characteristics of different algorithms. In addition, five hyperspectral datasets with different species categories and scene distributions and aggregated datasets with different spatial resolutions, spectral resolutions, and SNRs were used to validate our study. The experimental results indicate that: (1) The overall accuracy (OA) using CART, RF, 3D-CNN, and VIT models decreased with coarser spectral resolution, but almost remained unchanged using the RVT classifier. The number of class and classification species affect the results. (2) The influence of spatial resolution on classification accuracy is related to the scene complexity, target size, and classification purpose. The coarser spatial resolution can achieve higher OA than the original spatial resolution for the uniform scene distribution. For the datasets with small objects and intersections of different species, OA first increased, plateaued, and then decreased with coarser spatial resolution. (3) The SNR has an obvious impact on OA for the CART and RF classifiers, and the impact decreased for deep learning models, especially for the VIT and RVT models, which were almost unaffected by SNR. Additionally, slight variations in experimental results were observed for datasets with different scene distributions and categories. Furthermore, we conducted a detailed analysis of the role of traditional machine learning and deep learning models in the experimental outcomes. The study can provide insights into understanding the relationship between the core parameters of hyperspectral imager and the artificial intelligence algorithms used for hyperspectral classification. It serves to bridge the knowledge gap between the front-end hyperspectral imager, mid-end model, and back-end applications, and further promote the development of hyperspectral imaging technology. • Optimize the processing chain for the design and application of hyperspectral sensor. • Analysis the role of hyperspectral imager parameters on classification accuracy. • Bridge the gap between front-end hyperspectral sensor design and back-end application. • Improve the classification accuracy by suitable sensor configuration. [ABSTRACT FROM AUTHOR]

Cloud geometrical thickness (CGT) retrieved from satellite observations is vital in the investigation of cloud microphysics and radiative balance. However, there are few algorithms designed for retrieving CGT over land by using passive remote sensing observations. This is due to the inherent difficulties in correctly retrieving atmospheric profiles because of inadequate knowledge of land surface reflection. A passive algorithm for retrieving CGT in single-layer liquid clouds over land and ocean has been developed in this paper. The retrieval method utilizes OCO-2 (Orbiting Carbon Observatory-2) hyper-spectral oxygen A-band (O2A) observations and employs an O2A white-sky albedo estimation method to correct for the disturbances from land surface reflected radiation under cloudy conditions. The retrieval is achieved in near real-time through the use of a fast radiative transfer approximation scheme and is as accurate as using a full radiative transfer model. The new algorithm retrieves CGTs with a correlation coefficient of 0.619 as compared to results from satellite-based lidar-radar products that have been adjusted for outliers. The retrieval of the cloud geometrical thickness exhibits a mean bias of 5 hPa (46 m) and a root-mean-square error of 40hPa (393 m). The validation uncertainties arise from inaccurate inputs, including land surface albedo estimates, as well as uncertainties and errors in the reference CGT. In addition, there are non-negligible biases in retrieval results due to cloud vertical inhomogeneity and three-dimensional radiative transfer effects at cloud edges. • An algorithm is created for the thickness of global single-layer liquid clouds. • A new estimation for land albedo facilitates the retrieval of terrestrial clouds. • A leaner retrieval model utilizing multiple channels reduces errors and saves time. • It is a competitive remote sensing approach to replace spaceborne radar in future. [ABSTRACT FROM AUTHOR]

Operational canopy height mapping at high resolution remains a challenging task at country-level. Most of the existing state-of-the-art inversion methods propose physically-based schemes which are specifically tuned for local scales. Only few approaches in the literature have attempted to produce country or global scale estimates, mostly by means of data-driven approaches and multi-spectral data sources. In this paper, we propose a robust deep learning approach that exploits single-pass interferometric TanDEM-X data to generate accurate forest height estimates from a single interferometric bistatic acquisition. The model development is driven by considerations on both the final performance and the trustworthiness of the model for large-scale deployment in the context of tropical forests. We train and test our model over the five tropical sites of the AfriSAR 2016 campaign, situated in the West Central state of Gabon, performing spatial cross-validation experiments to test its generalization capability. We define a specific training dataset and input predictors to develop a robust model for country-scale inference, by finding an optimal trade-off between the model performance and the large-scale reliability. The proposed model achieves an overall estimation bias of 0.12 m, a mean absolute error of 3.90 m, a root mean squared error of 5.08 m and a coefficient of determination of 0.77. Finally, we generate a time-tagged country-scale canopy height map of Gabon at 25 m resolution, discussing the potential and challenges of these kinds of products for their application in different scenarios and for the monitoring of forest changes. • Country-scale mapping of canopy height over Gabon, Africa. • Bistatic TanDEM-X InSAR features essential for bio-physical parameter estimation. • State-of-the-art canopy height estimation performance with deep learning. • Generation of calibrated estimates independent of SAR and InSAR geometries. • Novel reliability estimation approach of the country-scale product. [ABSTRACT FROM AUTHOR]

Global urban mapping is vital for understanding various environmental challenges and supporting Sustainable Development Goal 11. Although deep learning models present a potential unified solution, their effectiveness is intrinsically tied to the quality and diversity of the training data, which often present limitations in existing research. To overcome these limitations, this paper introduced a data engine tailored to generate high-quality and diverse training samples at the global scale. This semi-automatic procedure operated in two stages. The initial stage focused on the generation of globally-distributed accurate samples by harmonizing existing open-source datasets. The subsequent stage broadened the sample coverage to the global scale by leveraging published global data products and OpenStreetMap data, ensuring the sample's diversity. Using the dataset generated by the data engine, we trained a Global Urban Mapper (GUM), achieving superior global testing results, outperforming the second-best product (i.e., GISA-10) by 2.89% in Overall Accuracy (OA) and 5.92% in mean Intersection over Union (mIoU). The advancements can primarily be ascribed to the superior quality and heterogeneity of the data generated by the proposed data engine, providing a precise and diverse set of samples for the deep learning model to assimilate. The proposed data engine, built exclusively on open-source data, offers promising prospects for global mapping tasks beyond urban land cover. We will release GUM and the associated preprocessing code in https://github.com/LauraChow77/GlobalUrbanMapper , which will empower users to map specific areas of interest worldwide, thereby facilitating timely urban assessment and monitoring. • A data engine to generate an accurate and diverse dataset for global urban mapping • Open-source datasets alleviated the minimum mapping unit issue in labeling samples • Existing products and OpenStreetMap data utilized to expand and diversify samples • The Global Urban Mapper enables timely urban assessment and monitoring globally [ABSTRACT FROM AUTHOR]

This paper presents a scheme for retrieving wind direction without 180° ambiguity over the sea solely based on Synthetic Aperture Radar (SAR) images. The dataset utilized for training, validating, and testing wind direction estimation deep neural network consists of 19,210 spatiotemporal match-ups of Sentinel-1 sub-images from Google Earth Engine (GEE) platform and wind direction records from National Data Buoy Center (NDBC) stations, from 2014 to 2023. Evaluation metrics, namely Line Angle Error (LAE) and Wind Direction Ambiguity accuracy (WDA accuracy), are proposed to facilitate modelling and assessing the performance of model in various conditions, whether open or not, and under different marine atmospheric boundary layer (MABL) stratification states. Post-processing method is devised to derive final robust wind direction field. Stratification can be quantified by a bulk Richardson number, Ri. With a wide spatiotemporal distribution, 58 full-frame Sentinel-1 images that satisfy −0.016 < Ri < +0.001 constitute an independent dataset for validating the entire wind direction retrieval process. Statistical analysis was conducted to compare the SAR-derived wind directions with those from COSMO-REA6, ASCAT and ERA5. It reports bias <1.4° and centered root mean square difference (CRMSD) < 17.3°. Wind streaks (WS) and shallow convective cells (MC) are common coherent structures in these images, both of which contain wind direction information. The scheme has a certain ability to counteract the contamination from rain cells, internal waves and atmospheric gravity waves, and also demonstrates great potential for high-resolution wind direction retrieval of tropical cyclones (TC). Additionally, we introduced Gradient-weighted Class Activation Mapping (Grad-CAM) to emphasize the significance of land-shadow features to wind direction inference. • The deep neural network can remove 180° wind direction ambiguity to some extent. • Surface layer stratification affects the retrieval of wind direction. • A scheme for retrieving unambiguous wind direction from SAR imagery is presented. • The results have been compared with wind data from COSMO-REA6, ASCAT and ERA5. [ABSTRACT FROM AUTHOR]

Understanding the sensitivity of remote sensing (RS) observation to land surface parameters (e.g., reflectance and temperature) is very important for estimating the accuracy of RS products and optimizing inversion algorithms. The most precise method for quantifying this sensitivity is physical modelling of derivative propagation in simulated 3D landscapes. However, to our knowledge, present land surface radiative transfer models (RTM) do not simulate derivative propagation. This paper proposes an original "differentiable radiative transfer modelling" that simulates the derivative propagation in natural and urban landscapes, for reflectance. It is integrated in the framework of DART RTM. We validated it both analytically and with a finite difference method applied to a 3D landscape. This new modelling extends the efficiency of 3D RTMs for sensitivity studies. It is implemented in the DART version freely available for research and education (https://dart.omp.eu). • First derivative (Jacobian) propagation modelling in land surface RT models. • Validation of this derivative modelling by the finite difference method. • Discussion of the major factors that influence the derivative. • Outlook of potential applications of the derivative modelling. [ABSTRACT FROM AUTHOR]

The use of remote sensing technique to monitor atmospheric water vapor is significant for weather and climate studies. However, the general methods of retrieving precipitable water vapor (PWV) with high precision and high resolution using remote sensing satellite has hardly been investigated, which becomes the focus of this paper. A general remote sensing PWV retrieval (GRPR) method that uses level-1 near-infrared (NIR) data sympathized by a global navigation satellite system (GNSS) is proposed. In this method, the atmospheric transmittance coefficients are determined by combining the radiative transmission model and the central wavelength interpolation method instead of directly using the traditional empirical values. Next, the PWV derived from remote sensing NIR data is obtained using the adaptive seasonal exponent model rather than the traditional transmittance–water vapor lookup table method. Furthermore, the accuracy of remote sensing NIR PWV is further calibrated by establishing the seasonal relationship between PWV residual and elevation. The corresponding NIR data from the Fengyun-3A (FY3A) satellite over the period of 2013–2015 in China are selected to validate the proposed method. Statistical results show the good performance of GRPR method for internal and external accuracies with root mean square (RMS) improvement rates of 76.8% and 72.4%, respectively, compared with the FY3A level-2 PWV products. In addition, the proposed method has good robustness and is almost unaffected by the PWV magnitude at different seasons. Proposed GRPR method is also applied for PWV retrieval at different time scales, further showing its superiority of retrieving PWV with high precision and high resolution. These results indicate the good application prospect of the GRPR method proposed in this study for generating remote sensing PWV products. • The optimal atmospheric transmittance coefficients are determined in this study. • An adaptive seasonal exponent model is developed to retrieve remote sensing PWV. • The accuracy of remote sensing NIR PWV is further calibrated using DEM data. • GRPR method is applied for remote sensing PWV retrieval at different time scales. [ABSTRACT FROM AUTHOR]

The uncertainty of surface soil moisture (SM) retrievals from satellite brightness temperature (TB) observations depends primarily on the choice of radiative transfer model (RTM) parameters, prior SM information and TB inputs. This paper studies the sensitivity of several established and experimental SM retrieval products from the Soil Moisture Ocean Salinity (SMOS) mission to these choices at 11 reference sites, located in 7 watersheds across the United States (US). Different RTM parameter sets cause large biases between retrievals. Whereas typical RTM parameter sets are calibrated for SM retrievals, it is shown that a parameter set carefully optimized for TB forward modeling can also be used for retrieving SM. It is also shown that the inclusion of dynamic prior SM estimates in a Bayesian retrieval scheme can strongly improve SM retrievals, regardless of the choice of RTM parameters. The second part of this paper evaluates the ensemble uncertainty metrics for SM retrievals obtained by propagating a wide range of RTM parameters through the RTM, and the relationship with time series metrics obtained by comparing SM retrievals with in situ data. As expected for bounded variables, the total spread in the ensemble SM retrievals is smallest for wet and dry SM values and highest for intermediate SM values. After removal of the strong long-term SM bias associated with the RTM parameter values for individual ensemble members, the remaining anomaly ensemble SM spread shows higher values when SM deviates further from its long-term mean SM. This reveals higher-order biases (e.g. differences in variances) in the retrieval error, which should be considered when characterizing retrieval error. The time-average anomaly ensemble SM spread of 0.037 m3/m3 approximates the actual time series unbiased root-mean-square-difference of 0.042 m3/m3 between ensemble mean retrievals and in situ data across the reference sites. • Wetter SM retrievals for higher roughness and vegetation parameter values • Good SM retrievals possible with RTM parameters calibrated for TB forward modeling • Improved SM retrievals with inclusion of model-based prior SM estimates • Similar SM mean - variance relationship sampled in time, space, or across an ensemble • Higher-order biases in SM retrieval error after long-term mean removal [ABSTRACT FROM AUTHOR]

Numerous algorithms have been developed to retrieve chlorophyll- a (Chla) concentrations (mg m−3) from Earth observation (EO) data collected over optically complex waters. Retrieval accuracy is highly variable and often unsatisfactory where Chla co-occurs with other optically active constituents. Furthermore, the applicability and limitations of retrieval algorithms across different optical complex systems in space and time are often not considered. In the first instance, this paper provides an extensive performance assessment for 48 Chla retrieval algorithms of varying architectural design. The algorithms are tested in their original parametrisations and are then retuned using in-situ remote sensing reflectance (R rs (λ), sr−1) data (n = 2807) collected from 185 global inland and coastal aquatic systems encompassing 13 different optical water types (OWTs). The paper then demonstrates retrieval performance across the full dataset of observations and within individual OWTs to determine the most effective model(s) of those tested for retrieving Chla in waters with varying optical properties. The results revealed significant variability in retrieval performance when comparing model outputs to in-situ measured Chla for the full in-situ dataset in its entirety and within the 13 distinct OWTs. Importantly, retuning an algorithm to optimise its parameterisation for each individual OWT (i.e. one algorithm, multiple parameterisations) is found to improve the retrieval of Chla overall compared to simply calibrating the same algorithm using the complete in-situ dataset (i.e. one algorithm, one parameterisation). This resulted in a 25% improvement in retrieval accuracy based on relative percentage difference errors for the best performing Chla algorithm. Improved performance is further achieved by allowing model type and specific parameterisation to vary across OWTs (i.e. multiple algorithms, multiple parameterisations). This adaptive framework for the dynamic selection of in-water algorithms is shown to provide overall improvement in Chla retrieval across a continuum of bio-geo-optical conditions. The final dynamic ensemble algorithm produces estimates of (log 10 -transformed) Chla with a correlation coefficient of 0.89 and a mean absolute error of 0.18 mg m−3. The OWT framework presented in this study demonstrates a unified approach by bringing together an ensemble of algorithms for the monitoring of inland waters at a global scale from space. • Comprehensive evaluation of algorithm performance for remote sensing applications. • Algorithm recalibration based on an extensive dataset collected from 185 waters. • Recommendation for an optical water type framework to improve overall accuracy. [ABSTRACT FROM AUTHOR]

Scientific contributions from remote sensing over the last fifty years have significantly advanced our understanding of urban areas. Key contributions of urban remote sensing include but are not limited to characterization of urban areas, urban land cover changes and thermal remote sensing of urban climates. Today, the proliferation of new sensors, long time series of the satellite record, joint analysis of Earth observation data with ancillary data sets, widespread availability of high-performance computing facilities, and slow but increasing use of remote sensing data and methods in non-remote sensing fields together offer new opportunities to generate scientific knowledge for an urbanizing planet. Simultaneously, the scale and pace of contemporary urbanization require new information about urban areas from both the science and policy communities. This paper examines remote sensing contributions to the scientific understanding of urban areas over the last 50 years until today. Based on this assessment and current needs by user communities, we identify four strategic directions for future urban remote sensing research: high temporal frequency analysis, characterization of urban heterogeneity, characterization of urban form and structure in two and three dimensions, and linking remote sensing with emerging urban data. Advances in these four areas are likely to generate significant new insights that will be useful to both science and policy. • This paper reviews remote sensing contributions to the knowledge of urban areas. • Four strategic directions for future urban remote sensing research are proposed. • Advances in these four areas provide new insights that are useful to science and policy. [ABSTRACT FROM AUTHOR]

Abstract Accuracy is an integral component in consumption of land cover and change information. Accuracy characterization in change categorization is more complicated than that in single-date classifications, especially when pursued at a local or per-pixel basis. In this paper, reference sample data consisting of reference class labels at sample locations are referred to as (model) training and validation sample data, respectively, when they are used for building and testing predictive models of local accuracy. They are preferably collocated to verify land cover and change types at the same locations across multiple single-date classifications. With collocated training data, methods devised for local accuracy mapping were usually implemented as direct extensions to those designed for accuracy predictions in single-date land cover maps, given that both response and explanatory variables are defined properly based on change maps. However, alternative methods are required when collocated training sample data are unavailable or sparse but non-collocated data collected at sample units not aligned across time exist or can be made available more conveniently and flexibly. This is typically the case for change-classification accuracy analyses, whereby prior sampling design and sample data may not remain informative in time but can be made adaptive by restructuring and augmentation to better represent strata of transitions as well as persistence. Such adaptability leads to great cost-saving in sampling than undertaking collocated sampling anew each time performing or revising accuracy mapping for a specific change analysis over a particular period. With a coherent review of related research, this paper proposes a geostatistical method for predictive mapping of local accuracies in any land-cover change maps created properly through adaptive use of complexly configured training sample data. This method works on a functional relation between accuracies in change-categorization and single-date classifications whereby temporal correlation (cross-correlation) between bi-temporal classification correctness is accommodated. In this cross-correlation-adjusted product method (named PXCOV), quantification of single-date classification accuracies is based on logistic regression. Cross-correlation is estimated via cross-validation cokriging through exploiting a useful relation between traditional and pseudo cross-variograms in the absence of collocated sample data or calculated more easily on the basis of collocated sample data if available. Studies were undertaken to test this method and compare it with direct logistic-regression-kriging (named LogRK direct), its simpler version, direct logistic regression without kriging (named LogR direct), and the baseline product method (named Product, being a simpler option of PXCOV without accounting for cross-correlation), using GlobeLand30 2000 and 2010 land cover at five sites in China. For each site, eleven training samples of equal sizes but differing configurations and one independent validation sample were collected. Logistic regression was performed on map class occurrence pattern indices quantified in size-optimized moving windows for mapping local accuracies in single-date classifications (as necessary intermediate steps in Product and PXCOV) and bi-temporal change-categorization (LogR direct and LogRK direct). In LogRK direct, geostatistical modeling and kriging were based on standardized residuals after logistic regression (based on change maps directly). In PXCOV, cross-correlation between single-date classification correctness was estimated and accounted for properly, with their non-stationarity in change vs. no-change strata of sample pixels accommodated, after estimating aforementioned single-date local accuracies. Statistical testing of relative performances of alternative methods was undertaken based on validation samples for individual study sites. It was confirmed that the proposed method PXCOV is more accurate than LogR direct and LogRK direct when there are relatively few collocated sample pixels (i.e., the typical scenarios justifying use of PXCOV). PXCOV is generally more accurate than Product and should be pursued when enough collocated data are available for estimating cross-correlation, although Product is sometimes adequate for practical purposes. LogR direct is recommended over LogRK direct because collocated sample data are often sparse due to low sampling intensities as in large-area applications. Therefore, as a method complementing LogR direct and similar methods, PXCOV facilitates adaptive use of training sample data of complex configurations for increased accuracy and efficiency in predictive mapping of local change-categorization accuracies. Highlights • Non-collocated reference data exist and should be reused for local accuracy mapping. • Sample data (collocated or not) are adaptively used by PXCOV for accuracy mapping. • PXCOV is applicable for mapping accuracies in any change maps created properly. • PXCOV is significantly more accurate than a simpler method Product in general. • PXCOV is more accurate than LogR direct with largely non-collocated sample data. [ABSTRACT FROM AUTHOR]

The penetration ability of Light Detection and Ranging (LiDAR) pulses into vegetation cover makes it a valuable tool in forest inventory. Extraction of a Digital Terrain Model (DTM) using LiDAR data is a challenging topic, especially in steep and complex terrains with forest canopy. This paper presents some approaches for ground filtering and interpolating the point cloud to generate DTM in forested terrains. Interpolating the points in dense forests that have high attitude variation is very difficult and make the popular interpolation methods unsatisfactory. This paper proposed a modified Invasive Weed Optimization (IWO) method for finding the optimized coefficients of the polynomial interpolation method. This method had a good performance with 0.210 mm RMSE value in the forested terrains. The interpolated point cloud was used as the input of the proposed ground filtering method for detecting the non-ground pixels. The proposed ground filtering method was structured with two main sections including, iterative geodesic morphology and scan labeling. In the iterative geodesic morphology, some geometric and structural parameters were introduced to investigate the quality of extracted points in each iteration. The scan labeling searched the data pixel by pixel in four directions and labeled the pixels with the high slope value in all directions. The non-ground pixels were obtained by integrating the result of iterative geodesic and scan labeling. Assessment of the ground filtering results using the International Society for Photogrammetry and Remote Sensing (ISPRS) showed 3.92% total error compared to the other reported algorithms. This demonstrated the ability of the proposed approach in recognition of the non-ground pixels. The extracted pixels were removed and the DTM was generated by filling the gaps using the proposed IWO and polynomial interpolation. Some forested regions with various characteristics such as sparse and dense trees on the hills and steep slopes were utilized to evaluate the accuracy of the generated DTM. The computed RMSE in the test areas was 0.463 m, on average, which was acceptable for the complex and forested terrains. [ABSTRACT FROM AUTHOR]

Satellite altimetry has been widely used for inland water monitoring and modeling in the present decade. In this paper, the availability, data quality, and quantity of a number of spaceborne radar altimetry missions (including Topex/Poseidon (TP), ERS-2, Envisat, Jason-1/−2/−3, CryoSat-2, Saral, and Sentinel-3) are investigated. This is, to the best of our knowledge, the first paper to investigate the influences of onboard trackers, satellite flying directions, and land contamination on Level-1 waveforms and Level-2 measurements over large mountainous lakes. The analyses were conducted over eight Tibetan lakes, which are of varying sizes from ca. 93 to 1046 km2 and are sampled by multiple altimeters. From the results presented in this paper, it is demonstrated that the previous generation of altimeters, such as TP, Jason-1, ERS-2, and Envisat, shows a higher percentage of data unavailability and lower valid observation rates. Investigations of Level-1 waveforms show onboard tracker failing in locking on the lake surface, resulting in data loss or invalid measurements. In addition, the onboard tracking systems are sometimes unstable, resulting in larger along-track standard deviations (SD) (SD > 15 cm). In contrast, the recent generation of altimeters, i.e., Jason-3 and Sentinel-3 with open-loop tracking mode, significantly increases the data availability (> 90% passes and > 60% individual observations) and delivers higher precision measurements (SD around 10 cm). Moreover, the waveform leading edge is very consistent, indicating more stable onboard tracking. Higher quality Jason-3 and Sentinel-3 data allow monitoring of sub- and quasi-monthly lake variations, which are not revealed by e.g., TP and ERS-2 data due to poorer signal-to-noise ratios. Exceptionally, Saral Ka-band altimeter, operated in closed-loop tracking mode, is less affected by flying directions and land contamination (topography and landscapes), delivering high quality measurements (SD around 7 cm for 40 Hz). This study provides a comprehensive comparison of multiple altimeters in monitoring lakes in mountainous areas, and demonstrates the substantial improvements achieved in the past decades. However, care has to be taken when merging data from old generation altimetry missions over mountainous lakes. The lessons learned from these Tibetan lakes can be taken across to other mountain lakes. • Comprehensive assessment of multiple altimeters performance over Tibetan lakes. • Investigation of influences of onboard trackers on Level-1 waveforms and Level-2 measurements. • Onboard tracking failure is largely responsible for erroneous measurements contained in the GDR. • Distinctive improvements (∼ 100% data availability & < 10 cm precision) achieved during 1992–2021. [ABSTRACT FROM AUTHOR]

The distributed scatterer (DS) interferometry (DSI) technology is a powerful geodetic tool for measuring the deformation over abundant land covers. As a state-of-the-art DSI technology, the SqueeSAR approach considered the statistical behavior of the decorrelated DS and proposed the phase triangulation algorithm (PTA) as the pioneering implementation of the phase linking (PL) theory to estimate the equivalent single-reference (ESR) phases from all interferometric combinations. Subsequently, many advanced estimators of the systematic phase series have been introduced. Interestingly, this paper has demonstrated that the utilization of weakly coherent interferometric pairs is not conducive to the theoretical accuracy improvement of the PL theory in case of the fast decorrelation scenario. Moreover, the reduction of the small baseline scales will worsen the positive definite degree of the time series coherence matrix, which makes the ill-posed solution problem more serious. Therefore, this paper proposes a novel adaptive PL (AdpPL) method integrating the following two innovations for mapping surface deformation: 1) in terms of the interferometric pair selection optimization (IPSO), this paper performs the homogeneity measure-based IPSO (HoMeIPSO) method to remove the lowly coherent candidate subset progressively and select the moderate interferometric pairs adaptively; 2) in terms of the robust parameter estimation, an adaptive regularization (AdpReg) approach suggests selecting an optimal damping factor according to the minimum fitting error of all interferogram observables. Experimental results demonstrate that the proposed method can restore the systematic phase series more clearly and detect the deformation area with spatially highly dynamic characteristics successfully. • Small baseline configuration can improve the theoretical accuracy of the PL theory. • Small baseline configuration further worsens the ill-posed problem for the MLE. • A coherence magnitude-based statistical comparison can construct the adaptive network. • An optimal regularization factor is selected according to the minimum fitting error. • AdpPL restores the fringe clearly and reduces the deformation monitoring blind area. [ABSTRACT FROM AUTHOR]

The launch of Soil Moisture Active Passive (SMAP) satellite in 2015 has resulted in significant achievements in global soil moisture mapping. Nonetheless, spatiotemporal discontinuities in the soil moisture products have arisen due to the limitations of its orbit scanning gap and retrieval algorithms. To address these issues, this paper presents a physics-constrained gap-filling method, PhyFill for short. The PhyFill method employs a partial convolutional neural network technique to explore spatial domain features of the original SMAP soil moisture data. Then, it incorporates variations in soil moisture induced by precipitation events and dry-down events as penalty terms in the loss function, thereby accounting for monotonicity and boundary constraints in the physical processes governing the dynamic fluctuations of soil moisture. The PhyFill model was applied to SMAP soil moisture data, resulting in continuous spatially daily soil moisture data on a global scale. Three validation strategies are employed: visual inspection through global pattern, simulated missing-region validation, and soil moisture validation with in situ measurements. The results indicated that the reconstructed soil moisture achieved a higher spatial coverage with satisfactory spatial continuity with neighbouring pixels. The simulated validation of the missing regions revealed that the averaged unbiased root mean square difference (ubRMSD) and correlation coefficient (R) were 0.01 m3/m3 and 0.99, respectively versus the gap filled SMAP product. The core validation sites demonstrated that the reconstructed soil moisture data has a consistent ubRMSD compared with the original SMAP soil moisture data (0.04 m3/m3 vs. 0.04 m3/m3). The PhyFill method can generate globally continuous, high accurate soil moisture estimates, providing remarkable support for advanced hydrological applications, e.g. , global soil moisture dry-down events and patterns. • A physics-constrained deep learning model to fill SMAP soil moisture gaps is developed. • Globally continuous daily SMAP soil moisture data from 2015 to 2022 is generated. • The reconstructed soil moisture data exhibits consistently high accuracy compared to the original soil moisture. • The reconstructed soil moisture can better capture dry-down events. [ABSTRACT FROM AUTHOR]

Atmospheric aerosols have strong impact on climate, environment, and health. To account correctly for such impact, extended aerosol characterization, including spectral Aerosol Optical Depth (AOD), Angstrom Exponent (AE), spectral Single Scattering Albedo (SSA) etc., are required to be derived globally from space-borne observations. Together with the aerosol, the Earth's surfaces are an important component of climate system, reflecting and absorbing solar and atmospheric radiation and being sources of emission of different natural aerosol, for example, sea-salt, mineral dust or organic aerosol. Since the beginning of space-borne atmospheric observation it was recognized that the most comprehensive extended aerosol and surface characterization can be achieved from multi-angular polarimetric measurements. In this paper, we show that the extended aerosol characterization can be obtained even from single viewing radiance only satellite measurements when several crucial conditions are fulfilled both for the observations and retrieval algorithm: (i) wide spectral range (for example, covering UV, VIR, NIR and SWIR spectra) providing rich spectral information about aerosol and surface; (ii) wide swath of measurements ensuring frequent overpassing over the same ground pixel to account for the different temporal variability of aerosol and surface properties as well as for angular dependence of surface reflectance; (iii) applied algorithm should be able to treat multi-temporal and multi-spatial observation accounting for spatial and temporal dependencies of aerosol and surface characteristics. The new possibilities of retrieval of aerosol and surface properties are investigated by applying the advanced GRASP (Generalized Retrieval of Atmosphere and Surface Properties) algorithm to the space-borne S5P/TROPOMI observations. The most optimal for these purposes forward and inversion approaches are discussed, and the quality of retrieved aerosol and surface properties is evaluated. The possibilities of TROPOMI/GRASP approach for aerosol extended characterization are demonstrated on several aerosol events including dust storms, biomass burning and anthropogenic aerosol pollution outbreaks in worldwide locations. • Extended aerosol/surface characterization can be derived from single angle measurements. • The conditions for the extended characterization are analyzed • The quality of TROPOMI/GRASP extended product is discussed. • The new possibilities of the TROPOMI/GRASP extended retrieval are demonstrated. [ABSTRACT FROM AUTHOR]

Downward shortwave radiation (DSR) is the main energy source for the Earth's system and a dominant component of the radiation budget. Although many algorithms have been proposed for estimating DSR, most models only consider either the radiative effects of terrain or atmosphere, and no research has considered the conditions of clouds on the mountainside. In this paper, a high-resolution Mountain Radiative Transfer model with Clouds (MRTC) is proposed to characterize the cloud-terrain coupling effects on DSR under different surface positions relative to the clouds. To operate the MRTC, a look-up table is employed to obtain initial radiative parameters. The comparison with Monte-Carlo based radiative transfer indicates that MRTC can be applied to conditions of clouds on the mountainside. And the reliability of MRTC under conditions when clouds are above mountains is also validated against in-situ measurements, with an RMSE of 149.9 W / m 2 , a bias of 28.3 W / m 2 , and an R 2 of 0.74. By using more accurate cloud cover data extracted from All-Sky images, the accuracy of MRTC has improved, with the RMSE decreasing to 115.8 W / m 2 , and the R 2 increasing to 0.8. This suggests that the uncertainty in satellite cloud products significantly contributes to errors in DSR estimations. MRTC is expected to more accurately represent cloud-terrain coupling effects on DSR, and improve the accuracy of DSR estimation over rugged terrain under cloudy skies. • A high-resolution mountain radiative transfer model with clouds (MRTC) is proposed. • The coupling effects of clouds and terrain on DSR are fully considered. • The first time the conditions of clouds on the mountainside are handled in DSR estimation. • Validation against a 3D radiative transfer model and in-situ measurements shows good agreement. • Uncertainty in cloud products and the complex 3D cloud structures significantly affect the accuracy of estimation. [ABSTRACT FROM AUTHOR]

Drought is a complex and extremely destructive natural disaster that seriously threatens the sustainable development of human society and ecosystems. The integrated drought index with remote sensing offers an efficient approach for drought monitoring and assessment, and has become one of the main trends in drought monitoring research. The existing integrated drought indices are generally used to monitor drought conditions at the national scale, with a spatial resolution of 500 m - 1 km. However, this resolution is still too coarse for farmland, unable to capture the spatial heterogeneity of drought stress at a finer field scale (such as 30 m × 30 m), and cannot meet the needs of field-scale drought monitoring such as refined agricultural water resources management and drought loss assessment. Therefore, this paper proposes an integrated framework for monitoring vegetation drought at the field scale, which integrates spatiotemporal fusion, deep learning, remote sensing, in-situ stations, and biophysical information. First, the framework established a classification system of drought factors based on the disaster system theory, including vegetation condition, drought-pregnant environment and drought-inducing factors. Second, the dense time series of 30 m monthly vegetation conditions (greenness, moisture, temperature) from 2001 to 2014 were generated based on spatiotemporal fusion. The monthly anomalies were calculated and then integrated based on 3D Euclidean distance method to generate the 30 m Vegetation Condition Anomaly Index (VCAI). Then, the drought-pregnant environment and drought-inducing factors were integrated through deep learning to generate the 30 m Environmental Drought-Inducing Index (EDII). Finally, the joint cumulative distribution was used to couple VCAI and EDII, and the 30 m High spatial resolution Drought Response Index (HiDRI) was generated. Results showed that the comparative analysis of HiDRI with in-situ Standardized Precipitation Evapotranspiration Index (SPEI), meteorological reanalysis data and remote sensing soil moisture data performed well. The overall Pearson correlation coefficient between HiDRI and in-situ SPEI was 0.601 (p < 0.01). Meanwhile, HiDRI can effectively map 30 m × 30 m field-scale drought spatial patterns in different climate zones, including heterogeneous spatial distribution and detailed texture features. In addition, the generated results of spatiotemporal fusion and deep learning have been effectively verified. • A new integrated framework for monitoring field-scale vegetation drought. • Deep learning-based 30 m environmental drought information integration. • Spatiotemporal fusion-based 30 m vegetation conditions generation. • Effective coupling by joint cumulative distribution. • A 30 m High spatial resolution Drought Response Index (HiDRI) was generated. [ABSTRACT FROM AUTHOR]

This paper presents the Nighttime Ocean Aerosol Optical Depth (AOD) Retrieval Algorithm (NOARA), an innovative method for deriving the oceanic AOD from the reflected moonlight observation of Day-Night Band (DNB) aboard the Visible Infrared Imaging Radiometer Suite (VIIRS) over ocean surfaces. Building upon the recent development of nighttime AOD retrievals for rural/wild areas over land, the NOARA introduces several advancements tailored for oceanic application: (a) the new dust and sea salt optical property models based on the AERONET observation; (b) the Cox-Munk bidirectional reflectance distribution function (BRDF) model that better represents the ocean surface reflectance; (c) the new internal cloud and dust masking schemes for the ocean areas. Case studies show that the nighttime AODs are in good consistency with multiple AOD measurements such as from AERONET and CALIOP. The algorithm fills the observational gap for a comprehensive understanding of the diurnal cycle of the dust transport over the ocean and has the potential to provide vital information to better constrain the chemical transport model. • First nighttime oceanic AOD retrieval algorithm from VIIRS lunar observation. • AOD retrievals agree well with the nighttime AERONET and CALIOP observations. • Day/night AOD together enable more coherent monitoring of the Sahara dust transport. • Assimilation of daytime AOD is insufficient to constrain dust transport in CTM. • Nighttime oceanic AOD provides insights on the evolution of diurnal dust transport. [ABSTRACT FROM AUTHOR]

Monitoring grazing activities on grassland is crucial for ensuring sustainable grassland development and for protecting it from grazing-led degradation. The Leaf Area Index (LAI), which measures leaf coverage over a surface area, is commonly used as a proxy for grassland condition. However, current studies focus on the year-round or seasonal aggregated LAI change rather than the change that can be attributed explicitly to grazing, which is the important indicator for quantifying grassland grazing. This paper presents a new exponential growth function under grazing with an estimation algorithm, the purpose of which is to extract grazing-led LAI changes for every 8 days' satellite observations. All the analyses are based on the Moderate Resolution Imaging Spectroradiometer (MODIS) MOD15A2H products. An improved MODIS LAI and an expected LAI are produced separately, considering both current and previous grazing-led LAI changes. The differences between expected LAI and improved LAI are then converted to the equivalent carbon mass of grazed material. This grazed carbon mass is then aggregated within the growing season, and compared with the expected carbon mass consumed by livestock (calculated from statistics yearbooks). In addition, Net Primary Productivity (NPP) is produced using the improved LAI, simulated by a Light Use Efficiency with Vegetation Photosynthesis Model (LUE-VPM). This is compared with the NPP produced by LUE-VPM based on original MODIS LAI, MODIS NPP products (MOD17A2H) and grassland monitoring stations' in situ measured data. Results show that the NPP calculated from the improved LAI is statistically the same as in situ converted NPP with a p-value equalling 0.998 (the RMSE between the two is 97.77 gC/m 2 ). Conversely, the p-value between converted in situ measured carbon mass and the MODIS NPP product is 0.011 (the RMSE between the two is 133.98 gC/m 2 ), indicating they are statistically different. The results detailed in this paper provide precise and almost real-time grassland grazing monitoring information for policy makers managing grassland. [ABSTRACT FROM AUTHOR]

This paper provides an overview of existing literature on vessel/ship detection and classification from optical satellite imagery. Although SAR (Synthetic Aperture Radar) is still the leading technology for maritime monitoring, the number of studies based on optical satellite data is quickly growing. Altogether we analysed 119 papers on optical vessel detection and classification for the period from 1978 to March 2017. We start by introducing all the existing sensor systems for vessel detection, but subsequently focus only on optical imaging satellites. The article demonstrates the temporal development of optical satellite characteristics and connects this to the number and frequency of publications on vessel detection. After presenting the methods used for optical imagery-based vessel detection and classification in detail, along with the achieved detection accuracies, we also report possibilities for fusing optical data with other data sources. The studied papers show that the most common factors greatly influencing the vessel detection accuracy are the following: different weather conditions affecting sea surface characteristics, the quantity of clouds and haze, solar angle, and imaging sensor characteristics. All these factors bring great variations in the selection of the most suitable method; some still continue to pose unsolved challenges. For higher relevance and wider usage, we suggest that the algorithms for detection and classification should support a variety of targets and meteorological conditions, and ideally also a variety of optical satellite sensors. At least, they should be tested on many images under different conditions. This is not usually the case in the existent literature. We also observed that many authors omit an appropriate performance quantification, which is critical for a practical assessment and a numerical comparison of the presented algorithms. Overall it can be seen that vessel monitoring from spaceborne optical images is a popular research topic and has a great operational potential in the near future due to the large amount of satellite data, much of it free and open. [ABSTRACT FROM AUTHOR]

In this paper we present the activities performed at the Microwaves and Radar Institute of the German Aerospace Center (DLR) to derive global forest/non-forest classification mosaics from interferometric synthetic aperture radar (InSAR) data acquired by the TanDEM-X mission. The data have been collected between 2011 and 2016 in bistatic stripmap single polarization (HH) mode, with the main goal of generating a consistent, timely, and highly accurate 3D representation of the global terrain’s surface (digital elevation model, DEM). The global data set of quicklook images, which represent a spatially averaged version of the original full resolution data at a ground independent pixel spacing of 50 m × 50 m, was used as input, in order to limit the computational burden. For classification purposes, several observables, systematically provided by the TanDEM-X system, can be exploited, such as the calibrated amplitude, the digital elevation model (DEM), and the interferometric coherence. Among the several factors contributing to a coherence degradation in InSAR data, the so-called volume correlation factor quantifies the coherence loss due to volume scattering phenomena, which typically occur in presence of vegetation. This quantity is directly derived from the interferometric coherence and used as main indicator for the identification of vegetated areas. For this purpose, a fuzzy multi-clustering classification approach, which takes into account the geometry and acquisition configuration, is applied to each acquired scene separately. A certain variability of the interferometric coherence at X band was observed among different forest types, mainly due to changes in forest structure, density, and tree height, which led to an adjustment of the algorithm settings depending on the considered type of forest. The identification of additional information layers, such as urban settlements or water areas, is also discussed, and the procedure for mosaicking of overlapping acquisitions (two at global scale, up to ten over mountainous terrain, forests, and desert regions) to improve the classification accuracy is detailed. The resulting global forest/non-forest map was validated using external reference information as well as with other existing classification maps and an overall agreement was observed that often exceeds 90 % . Finally, examples for high-resolution (at 12 m × 12 m) forest maps and potentials for deforestation monitoring over selected regions are presented as well, demonstrating the unique capabilities offered by the TanDEM-X bistatic system for a broad range of geoinformation services and scientific applications. The global TanDEM-X forest/non-forest map presented in this paper will be made available to the scientific community for free download and usage. [ABSTRACT FROM AUTHOR]

The FLuorescence EXplorer (FLEX) satellite mission, selected as ESA's 8th Earth Explorer, has been designed for the measurement of sun-induced fluorescence ( F ) spectra emitted by plants. This will be accomplished through a multi-sensor approach by placing it in a common orbit in tandem with the Sentinel-3 (S3) mission, which will have two optical sensors on board, OLCI (Ocean and Land Colour Instrument) and SLSTR (Sea and Land Surface Temperature Radiometer) to complement FLEX. These S3 instruments will be used in combination with the imaging spectrometers on board FLEX to provide data useful for atmospheric correction of FLEX data. However, a fully synergetic approach, i.e. by exploiting the spectral and directional information from all tandem mission instruments together, is an attractive alternative which is explored in this paper. By employing all combined top-of-atmosphere (TOA) spectral radiance data, one can (i) characterize the relevant optical properties of the atmosphere, (ii) retrieve biophysical canopy properties including the associated reflectance anisotropy, and (iii) retrieve a more accurate and consistent canopy F . Regarding retrieval methods, Fraunhofer Line Depth (FLD) and Spectral Fitting (SF) are well-known techniques applied to hyperspectral data. Both methods depend on a high spectral resolution and assume a Lambertian (isotropic) canopy reflectance. However, most vegetation canopies are non-Lambertian. This implies that, in particular when ignoring the anisotropic surface reflection, substantial retrieval errors can occur due to the interaction between atmospheric absorption bands and surface reflectance anisotropy. In this paper, a novel method based on spectral radiative transfer (RT) modeling is proposed, in which coupled RT models are used to simulate TOA radiance spectra. These are then matched with ‘measured’ spectra in order to retrieve surface fluorescence, along with a suite of biophysical parameters, by model inversion through optimization. By applying coupled RT models of the soil-leaf-canopy and the surface-atmosphere systems, TOA radiance spectra can be simulated for all optical sensors of this tandem mission. In this way, complex effects due to surface reflectance anisotropy and the spectral sampling by the various instruments, which are difficult to compensate for in the end products, are properly taken into account by their incorporation in the forward modeling. Next, by model inversion of TOA radiance data via optimization, the most accurate F retrievals can be achieved in a consistent manner, along with important canopy level biophysical parameters that may help interpret the F spectrum, such as chlorophyll content and leaf area index (LAI). The potential of this approach has been explored in a numerical experiment, and the results are presented in this paper. We find that, with the assumed well-characterized and plausible FLEX/S3 instrument performances, the simultaneous retrieval of biophysical canopy parameters and F spectra would be possible with a remarkable accuracy, provided the correct atmospheric characterization is available. [ABSTRACT FROM AUTHOR]

This paper presents measurements of the azimuth and zenith angle-dependence of the targets that are used for supervised vicarious calibration (SVC) methodology of hyperspectral sensors. Surface radiance measure at-sensor from airborne instruments, such as the AisaDUAL and HyMap sensors, show an apparent noise and shift that limit the information content and even impair the collected data (after removing the atmospheric attenuation). The drift from actual radiance is contributed by various sources; e.g., atmospheric gases and aerosols, and sensor stability. Yet one of the major contributors to it during nadir scanning is the directional effect of target reflectance at variable illumination geometry, also known as Bidirectional Reflectance Distribution Function (BRDF). The SVC method, which is based on ground dark nets of varying densities, was able to correct for the above factors and recalibrate the sensor's radiance during data acquisition. The main assumption of the SVC method is that the net targets are a Lambertian surface. In this paper, the BRDF measurements of the SVC net targets were carried out using a goniometer and controlled laboratory conditions. These measurements revealed anisotropic properties of the nets targets, which were further introduced to the radiometric calibration process are significant, especially with the high density nets (> 50% black net cover). The paper presents a look-up-table of the BRDF correction coefficient factors for each net target (identified and separated by its density and mixing level once placed over a bright background) at various zenith (from − 10° to − 30°) and azimuth (from 0° to 360° at intervals of 10°) angles. The coefficients were calculated for the near infrared spectral region (at 981 nm), and should be applied to the full-reflected spectral range across visible, near and shortwave-infrared wavelengths (400–2500 nm). The acquired results in this empirical study showed that the SVC calibration nets targets are characterized by incoherent BRDF effect. The darker and denser calibration targets absorb the down-welling radiance, and thus, reduce the amount of intermediate multiple scattering, normally accrued in all other (less dense) examined targets. The impact of target density on the BRDF correction coefficient and its dependence with illumination zenith and azimuth angles is studied and presented. The correction coefficient factors obtained in the laboratory were implemented in two real airborne hyperspectral campaigns of two different sensors: AisaDUAL and HyMAP over the same calibration site. The proposed BRDF SVC net corrections enables fine-tuning of the vicarious calibration procedure, while increasing the quality of the radiometric data. Thus, it is recommended to use the BRDF coefficients obtained in this paper for any airborne hyperspectral sensor that uses the SVC method in order to achieve better calibration accuracy. [ABSTRACT FROM AUTHOR]

Ocean eddies play a crucial role in the global energy cycle and significantly impact the transport of heat, salt, and nutrients in the global ocean. Spaceborne synthetic aperture radar (SAR), with its high spatial resolution (O (20 m)) and wide coverage, is an important tool for studying ocean eddies. In this paper, we propose a deep-learning network, named EOLO, to automatically detect ocean Eddies observed in C-band spaceborne SAR imagery, based on the You-Only-Look-Once (YOLO) deep learning algorithm. A Sentinel-1 (S1) SAR data-based ocean eddy dataset (named EddyDataset) is established to train the EOLO network. To effectively improve the performance of the EOLO network, we introduced a channel attention mechanism, adopted the up-sampling operator with the larger receptive field, and improved feature fusion method, anchor box size, and loss function. With the support of a high-performance detection network, the geographic information extraction module based on the affine geographic transformation model and a data preprocessing module were added, making EOLO an application-level framework. Our experiment results on EddyDataset demonstrate that EOLO achieves a high quality of eddy detection with 91.5% average precision. We further applied EOLO to entire scenes of S1 images to detect eddies in the Red Sea, the Baltic Sea, and the Western Mediterranean Sea, achieving 96.6%, 98.8% and 98.9% precision, respectively. Moreover, EOLO was used to extract size and location information of ocean eddies based on 6135 S1 scenes acquired in 2021 over the Western Mediterranean Sea. Their spatial characteristics were derived and compared with the eddies extracted from radar altimeter data, presenting interesting discrepancies between sub-mesoscale and mesoscale eddies in the ocean. • A deep-learning method is proposed to automatically detect ocean eddies by spaceborne SAR. • Applying the method to detect eddies in various seas shows a high accuracy of up to 95%. • In addition to detection, geographic information on eddies is extracted. • Discrepancies of sub-mesoscale and mesoscale eddies are revealed based on the SAR and RA observations. [ABSTRACT FROM AUTHOR]

Land surface temperature (LST) is an important factor in studies of surface energy fluxes between the Earth's surface and atmosphere. The Landsat LST product has been applied widely due to its fine spatial resolution and high data quality. Frequent cloud coverage, however, results in different degrees of gaps in the Landsat LST images, restricting greatly their application in practical cases requiring continuous LST time-series. Gap filling is an important technique used to reconstruct the missing data in the Landsat LST images. Using only temporally neighboring data from the same sensor, however, current gap filling methods struggle to deal with large spatial gaps and large temporal changes in LST. To solve these challenging issues, in this paper, multi-scale data (i.e., the 2 km, all-sky, hourly GOES LST product) were proposed for gap filling of Landsat LST images, and a spatial-temporal random forest (STRF) method was developed. The GOES LST product can provide effective information on LST in the gaps of the target Landsat LST data, with great potential to decreasing the uncertainty in gap filling when large LST change exists or the gap size is large. Through experiments in six regions of the USA, the proposed STRF method was shown to be more accurate than three typical gap filling and spatio-temporal fusion methods, and particularly advantageous for cases involving large temporal changes or large spatial gaps in LST. Specifically, STRF produces an average root mean square error (RMSE) of 1.6444 K and average correlation coefficient (CC) of 0.8653 for the six regions, which are 0.1133 K smaller and 0.0176 larger than the most competitive benchmark method (i.e., the spatial random forest (SRF) method). Conceptually, by spatial-temporal fusion of multi-scale data in gap filling, STRF provides a new paradigm for both spatial-temporal fusion and gap filling. Furthermore, the proposed STRF has the potential to be applied to various gap filling tasks in other situations. • 2 km, all-sky, hourly GOES LST data were used to assist gap filling of Landsat LST. • A spatial-temporal random forest (STRF) method was proposed. • STRF provides a new paradigm for both spatial-temporal fusion and gap filling. • The method is advantageous for large temporal changes and spatial gaps in LST. [ABSTRACT FROM AUTHOR]

Previous studies have indicated and shown the feasibility of retrieving snow density from ground based passive microwave measurements at the L band (1 – 2 GHz) from theoretical and experimental viewpoints. This paper expands the previous studies by presenting a case study of the retrieval problem with space borne brightness temperature measurements from the SMOS satellite over Sodankylä, Finland. To successfully retrieve snow density, also ground and forest parameters were included to the retrieval process. The retrieved variables were validated against in-situ ground, snow, and forest measurements made around the Sodankylä area over 12 winters from 2010 to 2022. A Bayesian framework was used to account for and quantify the uncertainties associated with the retrieval process. The results show good agreement between the retrieved ground and forest variables with the respective reference values. The snow density retrievals were seen to suffer from multiple sources of geophysical noise. However, the monthly average bulk snow density was successfully retrieved under stable midwinter conditions. The best agreement with the in-situ measurements was found for February, with a bias of -0.4 kg/m3, an unbiased root mean square difference of 12.2 kg/m3, and a correlation coefficient of 0.75. • Ground, snow, and forest parameters were retrieved over 12 winters. • The results were validated against comprehensive in-situ measurements. • The study site represents a typical heterogeneous boreal forest site. • Retrieved vegetation optical depth follows the modelled dynamics and value range. • Retrieved snow density agrees with measurements during the cold mid-winter period. [ABSTRACT FROM AUTHOR]

The monitoring of forest ecosystems is significantly affected by the lack of consistent historical data of low-severity (forest partially disturbed) or gradual disturbance (e.g. eastern spruce budworm epidemic). The goal of this paper is to explore the use of a subset of Landsat time series and deep learning models to identify both the type and the year of disturbances, including low-severity and gradual disturbances, in the boreal forest of eastern Canada at the pixel level. Remote sensing data such as the spectral information from Landsat time series are the best available option for large scale observations of disturbances that go back decades. Traditional modeling approaches, like LandTrendr, require substantial handcrafted pre-processing to remove noise and to extract temporal features from the image sequences before using them as input to a classical machine-learning model. Deep-learning models can autonomously discern which features are relevant within the coarse temporal and spectral information from the Landsat annual dense time series. We evaluated the performance of TempCNN and Transformer model in detecting and classifying the type and the year of the forest disturbance using Landsat time series subsequences. Our findings resulted in the generation of four disturbance maps outlining the forest history from 1986 to 2021 within the eastern Canadian boreal forest. Our experimental outcomes demonstrate several significant benefits of employing deep learning models. Firstly, using noisy Landsat time series they achieve comparable accuracy for classifying fire and total harvesting than existing publicly available disturbance maps. Secondly, the use of shorter time series subsequence with deep learning models enables to map adequately different overlapping disturbances occurring in the complete time series. Finally, they increase the number of distinguishable disturbance classes by adding partial harvesting, gradual disturbances, and forest recovery from older events, making them useful approaches for obtaining the first remote sensing-based map for areas affected by the eastern spruce budworm. • TempCNN, Transformer tested on Landsat time series for disturbance classification. • The type and the year of disturbance events were classified with good accuracy. • Deep learning effective for identifying partial and progressive disturbances. • Landsat subsequences extract overlapping disturbances. [ABSTRACT FROM AUTHOR]

Accurate aerosol optical depth (AOD) with high temporal resolution is required for dynamic monitoring changes of aerosol properties. The Himawari-8 Advanced Himawari imager (AHI) data with 10-min temporal resolution has been widely used for AOD retrieval but usually involves radiative transfer models and assumptions on surface reflection and aerosol properties that are hard to satisfy. This paper introduces a data-driven algorithm without any priori assumptions on surface or aerosol properties called Time series retrieval of Multi-wavelength AOD by adapting Transformer (TMAT). TMAT directly retrieves AOD at eight-wavelengths every 10-min throughout the day, using AHI top of atmosphere (TOA) reflectances in 6 wavebands and ancillary data. TMAT explores diurnal temporal change patterns of AOD and surface reflection using time series data. TMAT was trained and evaluated using samples derived by matching a year (2019) of AHI full-disk data and AERONET/SONET observations with 12,768 diurnal time series, each of which includes up to 81 10-min 6-band AHI TOA observations and 81 (time) × 8 (wavelength) AOD values per day. There are up to 273,413 (<12,768 × 81) TOA observations in 6 wavebands and up to 1,420,716 (<12,768 × 81 × 8) AOD values in all the time series samples, with filled values (−9999) to represent no data due to cloud contamination and incomplete wavelength configurations of AOD stations. The masking mechanism in Transformer is leveraged to exclude the contribution of the AHI filled values in time series AOD retrieval and exclude the contribution of AOD filled values in training. TMAT was trained with a physical constraint that the retrieved AOD decreases toward longer wavelengths. TMAT AOD was evaluated by separating training and validation samples using site-specific leave-one-station-out validation. TMAT AOD shows extremely high accuracy at all wavelengths, with RMSE = 0.101, R = 0.949 and 81.69% of retrievals within ± 0.05 + 15 % ∗ AOD sites for the AOD at 550 nm. We also validated the accuracy of the NASA multiangle implementation of atmospheric correction (MAIAC) AHI AOD over these sites with RMSE = 0.186, R = 0.832 and 52.10% of retrievals within ± 0.05 + 15 % ∗ AOD sites . The Ångström exponent (AE) derived from the multi-wavelength TMAT AOD also shows good agreement with AERONET/SONET AE, with a correlation coefficient of 0.620 and an RMSE of 0.291. The TMAT effectiveness to capture diurnal temporal change patterns of surface reflection and AOD was evidenced by the significant reduction of AOD accuracy when using a non-time series deep learning method. The effectiveness of multi-wavelength joint retrieval was evidenced by the significant reduction of AE accuracy when using single wavelength retrieval models. Evidence shows the training and evaluation split methods have a large impact on evaluation accuracy and this study further used ground network independent training and evaluation split to confirm the leave-one-station-out validation accuracy. Despite this, future study will focus on more spatially independent validation and more comprehensive comparison between the MAIAC and TMAT results. • Time series and multi-wavelength retrieval of AOD with deep learning for Himawari-8. • Trained with >1.4 million AOD samples from ground-based network measurement. • Accuracy of AOD is high at all wavelengths with RMSE = 0.101 and R = 0.949 at 550 nm. • Improvement over NASA MAIAC AOD at 550 nm with RMSE = 0.186 and R = 0.832. • Derived AE is accurate with RMSE = 0.620 and R = 0.291 due to multi-wavelength retrieval. [ABSTRACT FROM AUTHOR]

Geothermal systems can be used to produce low-emission energy throughout the day and night, regardless of the weather conditions. These features make geothermal systems a sustainable and reliable energy source, which can be exploited on a much larger scale than it is now. Remote sensing techniques can support detecting areas potentially suitable for geothermal energy production, thereby reducing the costs of preliminary exploration. The Ecosystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) can provide nighttime thermal imagery, which can be used for geothermal anomaly detection. This paper presents a method for automated detection of geothermal anomalies using nighttime ECOSTRESS data of the study area in Olkaria, Kenya. The proposed detection method is a kernel-based one, and includes adaptions of kernel size for the cases of large geothermal anomalies. The accuracy of the method is verified with reference data acquired during field work. A producer's accuracy of 82% is achieved, which is on average 56% points better than in randomised anomaly maps. The possible sources of errors in detection are heat capacity of surfaces, terrain features and vegetation masking the thermal signatures. The high producer's accuracy proves potential for application in global mapping of geothermal anomalies. [ABSTRACT FROM AUTHOR]

Surface waves are the most widely distributed texture in synthetic aperture radar (SAR) images acquired over the oceans, and long ocean waves are generally imaged in SAR images with a spatial resolution of 5 m as wave-looking structures. If these ocean-wave textures cause interference when detecting or extracting small-scale ocean features on the ocean surface, it is necessary to suppress them. However, owing to the complexity and unknown nature of ocean-wave distribution, existing filtering methods based on image spatial domain or transform domain and deep-learning-based methods cannot effectively suppress ocean-wave components. To solve this problem, this paper proposed a depth counteraction method to take advantage of the time-varying characteristics of ocean waves and other ocean phenomena to counteract ocean-wave components in deeper feature spaces. The proposed method first used sub-aperture segmentation to obtain multi temporal SAR images, determining the depth features of these images using encoder networks. Next, counteraction techniques were used in this depth-feature space to suppress the ocean-wave components, after which the SAR image could be reconstructed without ocean-wave components using the decoder network. Additionally, to further exploit the information in the images, parameter-sharing encoder networks were employed. Compared with seven other state-of-the-art SAR filtering methods, the proposed method can improve the average peak signal-to-noise ratio (PSNR) to 13 dB on simulated data and reduce the average target–to–clutter ratio (TCR) to 0.16 on real world SAR data. These indicators show that the proposed method has the potential to make useful contributions to detection or extraction of small-scale ocean features such as oil detection. • The ocean-wave distribution is more complex than speckle noise distributions. • Traditional filtering methods cannot effectively suppress ocean-wave components. • Sub-aperture segmentation can extract time-varying characteristic of ocean waves. • Simple neural network structure has powerful representation ability. • The counteraction method is more effective than fitted wave distribution method. [ABSTRACT FROM AUTHOR]

Synthetic Aperture Radar (SAR) imagery is the primary data type used for sea ice mapping due to its spatiotemporal coverage and the ability to detect sea ice independent of cloud and lighting conditions. Automatic sea ice detection using SAR imagery remains problematic due to the presence of ambiguous signal and noise within the image. Conversely, ice and water are easily distinguishable using multispectral imagery (MSI), but in the polar regions the ocean's surface is often occluded by cloud or the sun may not appear above the horizon for many months. To address some of these limitations, this paper proposes a new tool trained using concurrent multispectral Visible and SAR imagery for sea Ice Detection (ViSual_IceD). ViSual_IceD is a convolution neural network (CNN) that builds on the classic U-Net architecture by containing two parallel encoder stages, enabling the fusion and concatenation of MSI and SAR imagery containing different spatial resolutions. The performance of ViSual_IceD is compared with U-Net models trained using concatenated MSI and SAR imagery as well as models trained exclusively on MSI or SAR imagery. ViSual_IceD outperforms the other networks, with a F1 score 1.30% points higher than the next best network, and results indicate that ViSual_IceD is selective in the image type it uses during image segmentation. Outputs from ViSual_IceD are compared to sea ice concentration products derived from the AMSR2 Passive Microwave (PMW) sensor. Results highlight how ViSual_IceD is a useful tool to use in conjunction with PMW data, particularly in coastal regions. As the spatial-temporal coverage of MSI and SAR imagery continues to increase, ViSual_IceD provides a new opportunity for robust, accurate sea ice coverage detection in polar regions. • ViSual_IceD is a convolutional neural network (CNN) trained for sea ice detection. • The CNN inputs multispectral and radar images outperforming single image tools. • The network selects the most informative image type during segmentation. • Network outputs are more accurate than coastal AMSR2 sea ice concentration data. [ABSTRACT FROM AUTHOR]

Surface emissivity (ε) is used to characterize surfaces and to determine surface temperature from thermal radiation data. While in many applications it is treated as a constant, it is known to change with surface water content. Several ASTER/SEVERI based studies have speculated that diurnal changes in ε over deserts are linked to diurnal soil water content cycles resulting from water vapor adsorption during the night and subsequent evaporation during the day. This paper aims, for the first time, to validate the relationship between diurnal changes in surface ε and the changes in soil water content due to water vapor adsorption and evaporation under natural conditions. Measurements were conducted with a 6-band infrared radiometer, designed to validate ASTER bands 10–14, with study-specific recalibration for improved accuracy of ε. The evaluation included two different approaches to determine ε : using a single reference band (1B) and using the temperature/emissivity separation algorithm (TES). water content. While the TES has proven itself in many applications, it was found that for the soils studied (sand and loess) the use of 1B approach gave more consistent results for ε changes with soil water content than TES. Emissivity could be a powerful tool to characterize little studied soil water content changes in arid regions, but will require better characterization of surface properties to quantify the relationship between ε and soil water content for various soil types. Additional challenges to upscale this method include properly accounting for air irradiance and spatial heterogeneity. Meeting these main challenges will lead the way to detect small changes in soil water content under dry conditions at larger scales. Whether these are a result of water vapor adsorption or other processes, detecting such small changes in soil water content will provide new insights into desert hydrology. • On-ground validation of an infrared radiometer with ASTER bands. • Calibration and measuring protocol for minute changes in emissivity. • Observed emissivity changes of 0.2–0.6 due to water vapor adsorption. • Main challenges: air radiance, surface heterogeneity, and soil spectral properties. • Potential for new insights into desert hydrology. [ABSTRACT FROM AUTHOR]

Knowledge of spatio-temporal patterns of global land surface evapotranspiration (ET) is essential for understanding the exchanges of energy, water and carbon between the earth surface and atmosphere and responses to human activities, climate changes, and extreme weather events. This paper comprehensively reviewed accuracies and spatio-temporal patterns of 25 state-of-the-art global datasets of land surface ET from 2001 to 2018, where the global and/or local distributions of annual mean, interannual variability, and annual trend for both ET and its components [canopy transpiration (Ec), soil evaporation (Es), and interception evaporation (Ei)] were investigated. Overall, 23 out of the 25 ET products exhibited good consistency with monthly eddy covariance observations at 83 globally distributed sites, with the root mean square error (RMSE) ranging from ∼19 mm/month to ∼29 mm/month and the coefficient of determination (R2) ranging from 0.45 to 0.76. Comparatively, almost all ET products were in good agreement with the water balance-based observations at 137 large basins worldwide, with the R2 of >0.60 and the RMSE ranging from ∼200 mm/yr to ∼260 mm/yr. In general, remote sensing-based, machine learning-based, and hybrid-based ET products generally agreed better with site observations than reanalysis-based products. Global annual ET ranged from 530 mm/yr to 700 mm/yr with the ensemble mean of 628 ± 44 mm/yr (excluding areas of the urban, barren, wetlands, permanent snow and ice, and water bodies). Most products captured consistent global spatial distributions of annual mean, interannual variability, and annual trends of land surface ET and its components. Both mean annual ET and its components generally decreased with increasing latitudes though with an increasing trend in time. Interannual variability also decreased with increasing latitudes and with an increasing trend in time, but only in the Northern Hemisphere; these signals were not clear in the Southern Hemisphere. Most of the 25 products presented unimodal shapes of frequency curves, and displayed a sequenced multi-year mean of seasonal ET of JJA > MAM > SON > DJF. The Ec exhibited the largest mean annual value, in most areas followed successively by the Es and Ei. The ET products displayed a relatively high consistency in the spatial distribution of the Ec/ET with higher estimates of Ec/ET in the Southern Hemisphere than in the Northern Hemisphere. This up-to-date review could benefit the understanding of the states and advances of global ET datasets, afford scientific instruction for the development and improvement of ET models and products, and serve as a basis for the choice of ET datasets for specific use. • Spatio-temporal patterns of 25 global datasets of land surface ET are reviewed. • The 25 global ET products are validated against EC- and WB-based observations. • Global and local patterns for both ET and its components are investigated. [ABSTRACT FROM AUTHOR]

Recent studies have shown that radar temporal coherence over tropical and boreal forests undergoes a diurnal cycle as a result of a combined effect of the wind-induced motion of scatterers and of the change and displacement of water within the plant in response to the transpiration process. Within this context, the objective of this paper is to investigate, for the first time, the diurnal cycle of temporal coherence over wheat crops in relation to its development and physiological functioning throughout the agricultural season. A ground-based experiment was installed in Morocco, targeting a wheat field during the 2020 agricultural season. The radar system, essentially based on a Vector Network Analyzer (VNA) connected to 6C-band antennas installed at the top of a 20 m tower, has enabled quad-polarimetric acquisitions every 15 min. In parallel, evapotranspiration, soil moisture and meteorological variables are automatically measured in addition to above-ground biomass and vegetation water content collected during field campaigns. The results show that the temporal coherence with a 15 min baseline follows a marked diurnal cycle characterized by variable amplitude according to the phenological stage, with high values during the night, a significant morning drop to reach the lowest values in the late afternoon followed by an increase to recover the high nighttime values. The rate of the drop at dawn is shown to be related to the increase of evapotranspiration (r = 0.80 at VV polarization) when the wheat is covering the soil and the transpiration dominate the evapotranspiration process. This supports the assumption of a physiological effect related to water movement entailing a decorrelation. By contrast, the daily minimum of temporal coherence occurring in the late afternoon correlates well to the daily maximum of wind (r = 0.7). Interestingly enough, the amplitude of the diurnal cycle exhibit a marked seasonal evolution characterized by an increase of 85% from tillering to maturity in relation to the wheat development. At the early start of the season when the soil is almost bare, irrigation events impact slightly the diurnal cycle of temporal coherence. Likewise, it is shown that the presence of dew in the early morning has led to a decrease of the decorrelation rate. Temporal coherence dynamic has also been investigated for longer baselines up to 22 days. Results indicate a stronger decorrelation than what has been observed on tropical and boreal forests by previous studies with values below 0.4 for baselines above 2 days. Taken together, the results of this work demonstrate the unique potential of sub-daily C-band data for monitoring crop water status by future geostationary radar missions such as Hydroterra. • Diurnal cycles of 15-min C-band temporal coherence over wheat crops are analyzed. • The morning drop of the temporal coherence is related to the transpiration process. • Temporal coherence exhibits a marked seasonal dynamic according to phenological stages. • Irrigation events and morning dew have an impact on sub-daily temporal coherence. • Early morning hours showed potential for monitoring the vegetation physiological functioning. [ABSTRACT FROM AUTHOR]

Floods are a great threat to Pakistan with increasing concern. As the consequences of increased extreme weather related to climate change, Pakistan experiences severe floods almost every year. This study aims to explore and analysis the actual inundated situation, magnitude, the possible causes of the 2022 devastating floods, and future trends. We presented an enhanced nationwide flood mapping method and compared with other pixel-based image processing techniques including active contours and change detection. These algorithms were applied to Sentinel-1 Ground Range Detected (GRD) Synthetic Aperture Radar (SAR) imagery (10 m spatial resolution) with various land types and inundation scenarios in Pakistan, and were evaluated using other reference flood products. Accuracy evaluation analysis demonstrated that our algorithm has high robustness and accuracy, with the overall accuracy (OA) higher than 0.83 and critical success index (CSI) up to 0.91, and is suitable for automated flood monitoring in near real time. Nearly one-third of the lands were flooded in 2022, and more than half were inundated croplands. Punjab and Sindh provinces were the most severely affected regions, with the proportions of inundated area in 2022 (21.26% and 20.55%) nearly twice of that in 2010 (11.40% and 12.70%), indicating an intensified flooding trend. Analysis of possible influential factors showed that the intense and cumulative rainfall during the monsoon season (June to August) was the major cause of the 2022 flood event. Although the snow melted rapidly in June (the average change in snow depth is ∼10 mm), the overall ablation contributed insignificant amount to the flood water. The glacial lake outburst floods (GLOFs) induced by abnormal April–May heatwave provide water flowed into the tributaries of the Indus River, but are difficult to spread for thousands of kilometers from mountains to the plain downstream. The combination of the intrinsic arid climate and extreme floods exacerbate the already severe situation. • This paper presents an automated method based on tile thresholding and fuzzy logic. • The inundation in Punjab and Sindh in 2022 are nearly twice of that in 2010. • The cumulative rainfall is the major cause of the 2022 flood event. • The snow melting and GLOFs are likely small relative to the flood water gains. • These observations imply an intensified flooding and water cycle in Pakistan. [ABSTRACT FROM AUTHOR]

The marine shipping industry is among the strong emitters of nitrogen oxides (NO x) – a substance harmful to ecology and human health. Monitoring of emissions from shipping is a significant societal task. Currently, the only technical possibility to observe NO 2 emission from seagoing ships on a global scale is using TROPOMI data. A range of studies reported that NO 2 plumes from some individual ships can be visually distinguished on selected TROPOMI images. However, all these studies applied subjectively established pre-determined thresholds to the minimum speed and length of the ship — variables that to a large extent define the emission potential of a ship. In this study, we investigate the sensitivity limits for ship plume detection as a function of their speed and length using TROPOMI data. For this, we train a classification model to distinguish TROPOMI image patches with a ship, from the image patches, where there were no ships. This way, we exploit ground truth ship location data to potentially exceed human visual distinguishability. To test for regional differences, we study four regions: the Mediterranean Sea, Biscay Bay, Arabian Sea, and Bengal Bay. For the Mediterranean and the Arabian Sea, we estimate the sensitivity limit to lie around a minimum speed of 10 knots and a minimum length of 150 m. For the Biscay Bay — around 8 knots and 100 m. We further show that when focusing the analysis on the biggest emitters (junctions of several ships in the area), the detectability can be improved up to above 0.8 ROC-AUC. Finally, we show that increasing the size of the dataset, beyond the dataset used in this study, yields further improvements in the detectability of smaller/slower ships. The rate of improvement in both experiments is dependent on the region studied. This paper is the first comprehensive study on the real-world sensitivity of the TROPOMI instrument to distinguish the NO 2 emission produced by seagoing ships. • We established sensitivity limits for NO 2 ship plume detection with TROPOMI. • Obtained sensitivity limits: Med/Arabian Sea — 10 kt, 150 m; Biscay Bay — 8 kt, 100 m. • Detectability of ship plumes from short/slow ships can be improved with more data. • Detectability of NO 2 plumes differs between the regions. [ABSTRACT FROM AUTHOR]