Ulva prolifera (U. prolifera) green tides have become a recurring ecological issue in the Yellow Sea, posing serious threats to coastal ecosystems and regional economies. Traditional optical satellite index-threshold methods often require scene-specific threshold adjustment under complex marine environments, which may constrain their automation in large-scale applications. In this study, we present Green Tide Detection Network (GTD-Net), a knowledge-guided deep learning model for detecting and quantifying U. prolifera biomass using MODIS imagery. A key innovation of this study is the construction of a highly heterogeneous, multidimensional training dataset that captures variations in observation conditions, environmental backgrounds, and algae distribution patterns. By translating domain knowledge into learnable sample constraints, the dataset effectively guides model learning and improves robustness under heterogeneous observation conditions. Architecturally, GTD-Net augments the classical U-Net with multi-scale Inception modules, Residual Blocks, and the Convolutional Block Attention Module (CBAM), together with a hybrid weighted loss function to improve the delineation of boundaries and the detection of small or sparse targets. Ablation experiments showed that all components contributed positively to model performance, with the Inception and CBAM modules playing particularly important roles in multi-scale green tide feature extraction and background suppression. Comparative experiments showed that GTD-Net consistently outperforms traditional index-based methods and classical deep learning models, achieving an F1-score of 0.837 under complex observation conditions. When applied to MODIS time series data from 2007 to 2024, GTD-Net enables long-term, high-precision biomass estimation. For example, in 2019, its annual estimates were approximately 14% higher than those from the index-based method. These findings highlight the effectiveness and applicability of GTD-Net for large-scale monitoring of U. prolifera green tides.

Efficient segmentation of oiled pixels in optical remotely sensed images is the precondition of optical identification and classification of different spilled oils, which remains one of the keys to optical remote sensing of oil spills. Optical remotely sensed images of oil spills are inherently multidimensional and embedded with a complex knowledge framework. This complexity often hinders the effectiveness of mechanistic algorithms across varied scenarios. Although optical remote-sensing theory for oil spills has advanced, the scarcity of curated datasets and the difficulty of collecting them limit their usefulness for training deep learning models. This study introduces a data expansion strategy that utilizes the Segment Anything Model (SAM), effectively bridging the gap between traditional mechanism algorithms and emergent self-adaptive deep learning models. Optical dimension reduction is achieved through standardized preprocessing processes that address the decipherable properties of the input image. After preprocessing, SAM can swiftly and accurately segment spilled oil in images. The unified AI-based workflow significantly accelerates labeled-dataset creation and has proven effective for both rapid emergency intelligence during spill incidents and the rapid mapping and classification of oil footprints across China’s coastal waters. Our results show that coupling a remote sensing mechanism with a foundation model enables near-real-time, large-scale monitoring of complex surface slicks and offers guidance for the next generation of detection and quantification algorithms.

Beach litter poses persistent ecological and socioeconomic risks to coastal zones, yet reliably detecting extremely small objects (ESOs) in uncrewed aerial vehicle (UAV) imagery remains challenging: these targets occupy only a very small image area on the order of 10-5 of a typical RGB frame and are easily confused with textured sand backgrounds. We propose SRGO-YOLO, a super-resolution and granularity-optimized detector for UAV-based beach-litter monitoring. The network integrates a scale-adaptive spatial recalibration with channel attention module, a lightweight C3K2-MGLite neck, and a spatial-optimization-aware head with an additional P2 tiny-object branch to enhance shallow high-resolution features, perform mixed-granularity multiscale fusion, and refine predictions for low-contrast ESOs. To evaluate SRGO-YOLO, we construct the beach-litter-UAV dataset by screening two public UAV garbage datasets and combining them with a newly collected Zhuhai coastal beach litter dataset, in which more than 80% of annotated instances have normalized sizes below 0.02. Experiments show that SRGO-YOLO achieves 81.4% mAP(50) and 48.2% mAP(50:95), surpassing YOLOv11s by 21.8 and 13.5 percentage points, respectively, while maintaining 127 FPS on an RTX 3090 GPU. On a Jetson AGX Orin 64GB platform with TensorRT FP16 deployment, SRGO-YOLO further sustains 31.21 FPS in a 10-min continuous inference test with an average end-to-end latency of 32.04 ms. Per-class results further show consistent gains across all five categories: compared with YOLOv11s, SRGO-YOLO improves mAP(50) by 28.8%, 31.2%, 5.4%, 29.5%, and 14.5% for plastic, flexible pouch, metal, glass, and food container, respectively, with particularly strong benefits for visually ambiguous small items such as flexible pouch and food container. Qualitative comparisons further indicate that SRGO-YOLO recovers more complete instances under partial occlusion, tight adjacency, and overlapping litter patches. These results indicate that SRGO-YOLO not only improves ESO detection accuracy but also shows strong potential for real-time UAV-based beach-litter detection, supporting timely shoreline litter mapping and targeted coastal cleanup planning.

The Japan Sea is a semi-enclosed marginal sea characterized by complex hydrodynamics and prominent seasonal variability in oceanic fronts and currents. While physical dynamics are known to influence marine ecosystems, the specific spatiotemporal dependence of chlorophyll a (Chl a) on frontal activities and other environmental factors in this region remains quantitatively under-explored. This study utilized satellite-derived sea surface temperature (SST) and Chl-a data to detect fronts using the gradient method and investigated their relationships using empirical orthogonal function (EOF) and correlation analyses. The results revealed distinct seasonal patterns, with the subpolar frontal zone exhibiting a phase reversal in frontal activity compared to the coastal shelf zone. A significant positive correlation between frontal probability (FP) and Chl a was identified in the central and southern basins (35°N–40°N), indicating that frontal dynamics are a primary driver of phytoplankton growth in this region, likely through nutrient supply via vertical circulation. In contrast, wind forcing and surface currents were identified as the dominant factors regulating Chl-a variability in the northern and coastal regions. Furthermore, these regional dynamics were found to be significantly modulated by the North Pacific index on interannual scales. These findings provide a comprehensive quantitative assessment of the physical-biological coupling in the Japan Sea, highlighting the spatially varying roles of fronts and atmospheric forcing in sustaining marine productivity.

Seagrass meadows are vital coastal ecosystems that provide essential ecological functions but are highly vulnerable to human disturbances and environmental change. Effective monitoring is therefore critical for understanding their long-term dynamics and supporting conservation efforts. Satellite remote sensing provides an efficient approach for large-scale and long-term observation of seagrass ecosystems; however, accurate monitoring remains challenging due to tidal variability, seasonal growth dynamics, and complex underwater environments. This study focuses on the Caofeidian (CFD) seagrass meadow, the largest continuous seagrass habitat in China. Using Sentinel-2 Multispectral Instrument (MSI) imagery, we quantified its spatiotemporal dynamics from 2016 to 2024. A standardized workflow integrating pixel-level classification, image composition, and trend analysis was developed to characterize long-term variation patterns and identify areas associated with degradation, stability, and recovery. The results reveal a persistent net decline in seagrass extent over the past decade, despite localized recovery observed during the period of restoration activities after 2019. Five variation classes were identified: Significant Degradation (SD), Slight Degradation (SID), Stable (ST), Slight Recovery (SIR), and Significant Recovery (SR), accounting for 5.2 %, 15.5 %, 55.8 %, 22.9 %, and 0.6 % of the meadow area, respectively. Severe losses in several localized areas showed strong spatial and temporal associations with sustained human activities, including land reclamation, clam harvesting, and waterway development, whereas recovery patterns were mainly observed in regions with targeted restoration backgrounds. This study provides the first region-wide assessment of long-term seagrass spatiotemporal dynamics in CFD and demonstrates the potential of multi-temporal satellite observations for monitoring coastal seagrass ecosystems and supporting adaptive conservation management.

High resolution satellite optical sensors provide detailed sea surface textures and wave patterns, offering significant potential for retrieving sea surface wave parameters. The Coastal Zone Imager (CZI) onboard China’s ocean color satellite Haiyang-1E (HY-1E) provides 20-m multispectral and 5-m panchromatic imagery with a 3-d revisit cycle. Its distinctive optical design, which separates transmission and reflection channels onto different focal planes, introduces specific along-track channel time lags. In this study, the viewing geometry and inter-channel time lags of the CZI sensor were systematically evaluated. A dedicated workflow for wave parameter retrieval was developed and its feasibility was validated using 4 in-situ measurements and 69 reanalysis matchups. There is a statistically linear relationship between CZI-derived and reanalysis wave direction (RMSE=20.50°, Bias=0.24°) and period (RMSE=1.68 s, Bias=0.10 s). The specific requirements and limitations of CZI for resolving the 180° directional ambiguity were analyzed in detail. Furthermore, the proposed method was applied to generate the original coastal regional wave products from HY-1E. The uncertainty in comparison with ERA5 products and the potential of wind speed estimation were additionally discussed. Overall, these findings demonstrate that the HY-1E CZI sensor can effectively retrieve sea surface wave parameters, providing a valuable data source for monitoring coastal and marine dynamical environments.

High-resolution optical remote sensing can capture fine-scale sea surface texture and dynamics, and thus offers a valuable means for estimating surface winds and wave fields. Wind streaks, as direct indicators of wind forcing and roll vortices, have rarely been investigated in optical imagery. In this study, a diagnostic detection method for wind streaks based on cross-spectral analysis is developed using Sentinel-2 Multispectral Instrument (MSI) imagery. The algorithm is adapted from the previously published Angular Spectral Analysis (ASA) and combined with the inter-band time lag to determine the unambiguous wind direction. The cross-spectral framework further distinguishes wind waves from swells, enabling the estimation of wind wave direction, wavelength, and period. We compiled 965 MSI-buoy matchups (2016–2024), from which 157 wind streak scenes were detected and validated. The retrieved wind directions agree well with buoy observations with RMSE = 7.6°. Numerical simulations and synoptic-scale MSI images quantify a sunglint-driven brightness reversal and indicate a critical viewing angle range of θₘ ≈ 19°–24°, consistent with the observed absence of detectable streaks within this range. Wind streaks occur predominantly under moderate winds (6–12 m s−1) and near-neutral stability, highlighting the potential of optical imagery in diagnosing boundary-layer regimes. The method is applicable to other time-lagged push-broom sensors (e.g., HY-1E CZI2 at 20 m and Landsat OLI at 30 m) and, under favorable conditions, can be combined with whitecap coverage to retrieve the surface wind vector.

The uneven distribution of floating macroalgae often leads to missed detections of small, sparsely aggregated patches in medium-low resolution satellite images. This study evaluates the effectiveness of quasi-synchronous MODIS Terra/Aqua and HY-1 C/D CZI imagery from 2019 to 2022 for green tide detection in the Yellow Sea, revealing previously overlooked regional differences in detection capability between high-resolution and medium-low-resolution sensors. The results indicate significant discrepancies in biomass estimation across different spatial resolutions, with MODIS persistently underestimating algae biomass and failing to detect small algae patches. South of the Yellow Sea turbidity front, MODIS exhibits a higher missed detection rate, but the biomass estimation deviation between CZI and MODIS is smaller due to lower biomass per area (BPA). In contrast, north of the front, undetected small patches exhibit higher BPA, leading to greater biomass estimation deviation. The small algae patches missed by MODIS display distinct spatial distribution patterns, particularly forming narrow, long bands in the radial sand ridge area south of the front, likely influenced by topography, wind, and currents. These findings highlight the importance of high-resolution studies on small algae patches in different regions of the Yellow Sea for early detection and effective green tide management.

The extent of the differences between atmospheric and lake heatwaves remains unclear. Here we analysed daily surface water and air temperature data from 265 lakes worldwide (2000–2022) to compare heatwave trends, spatial distributions, and key differences. We find that lake heatwaves are more severe than atmospheric heatwaves, with longer accumulated heatwave days (29.7 days vs. 18.0 days), a shorter reoccurrence period (86.9 days vs. 121.4 days), and greater accumulated heat. Additionally, the frequency and total heatwave days have increased faster for lake heatwaves than for atmospheric heatwaves. When both types co-occur, heatwave severity intensifies. From a long-term perspective, reduced wind speed is the key driver of the differences between lake heatwaves and atmospheric heatwaves. Spatially, lake location is the primary determinant, followed by lake area and depth. Under a fixed-baseline high-emission scenario, by 2100, the difference is expected to diminish as air temperatures rise faster than water temperatures. Over the past two decades, lake heatwaves have become more intense than atmospheric ones due to declining wind speeds, which stabilise water column stratification and increase surface heating, according to analysis of daily surface water and air temperature from 265 lakes during 2000–2022.

A spaceborne optical technique for marine floating debris is developed to detect, discriminate, and quantify such debris, especially that with weak optical signals. The technique uses only the top-of-atmosphere (TOA) signal based on the difference radiative transfer (DRT). DRT unveils diverse optical signals by referencing those within the neighborhood. Using DRT of either simulated signals or Sentinel-2 Multispectral Instrument (MSI) data, target types can be confirmed between the two and pinpointed on a normalized type line. The line, mostly, indicates normalized values of <0.2 for waters, 0.2-0.6 for debris, and >0.8 for algae. The classification limit for MSI is a sub-pixel fraction of 3%; above which, the boundary between debris and algae is distinct, being separated by >three standard deviations. This automated methodology unleashed TOA imagery on data cloud platforms such as Google Earth Engine (GEE) and promoted monitoring after coastal disasters, such as debris dumping and algae blooms.

The distribution of Arctic sea ice is an important direct indicator of climate change, and spaceborne optical remote sensing represents one primary technique for sea ice monitoring due to its high spatiotemporal resolution and wide swath coverage. However, this process is often impeded by heavy cloud cover, which shares similar visual and spectral features with sea ice. To address these limitations, this study proposes a novel methodological framework for discriminating between sea ice and different cloud types (cirrus and cumulus) via the ultraviolet-visible-infrared observations from China's Haiyang-1C/D (HY-1C/D) satellites, and the ultraviolet (UV) data from the onboard Ultraviolet Imager (UVI) are used to study sea ice and clouds over the Chukchi Sea for the first time. The spectral properties are characterized by the top-of-atmosphere (TOA) reflectance (rho TOA) in both UV and visible and near-infrared (VNIR) wavelengths. This indicates that the 355 nm UV band has the optimal sensitivity to the presence of sea ice and clouds, with cirrus clouds composed of high-altitude ice crystals exhibiting extremely high UV reflectivity. A hybrid threshold is subsequently determined to separate sea ice and cloud pixels. In comparison to the MODIS MOD29 sea ice product, which masks cloud pixels with brightness temperature (BT) differences, this algorithm can effectively reduce the misclassification resulting from surface temperature inversions in polar regions. The ice/cloud identification results have been further applied to sea ice concentration (SIC) estimation, and extensive trials of this UV-based ice/cloud detection approach in the Arctic Passages demonstrates its potential applicability.

Oceanic whitecaps, as indicators of increased air-sea exchange, can be effectively captured by Landsat-8 Operational Land Imager (OLI) images with a 30-m spatial resolution. This study used a new adaptive iterative triangle algorithm, which was applied to 400 OLI images synchronized with National Data Buoy Center-measured wind speeds, covering the different offshore areas of the United States over the last ten years to automatically extract whitecap-affected pixels. By integrating radiative transfer equations, we calculated the whitecap coverage (W), determined that a 4 km window size is optimal for calculating W, and established a whitecap coverage-sea surface wind speed model. A sliding window method was used to achieve high-resolution (100 m) large-area sea surface wind speed estimations. More importantly, this study provides new insights into the impact of different wind and wave conditions on W and regional variations in the whitecap coverage-sea surface wind speed model from an optical satellite perspective, which can effectively quantify the wind-wave breaking and distinguish the regional variations of them in different offshore seas. By comparing with 10-m resolution data from the Sentinel-2 Multi-Spectral Instrument, we further clarified the impact of spatial resolution on the selection of methods for calculating W, revealing scale effects in optical remote sensing of whitecap detection and proposing corresponding W calculation strategies. These results demonstrate the potential of high-resolution optical remote sensing for monitoring whitecaps, estimating sea surface winds and studying air-sea interactions.

The determination of thin-ice thickness (TIT) in the Arctic can not only contribute to the understanding of climate change, but also to the ship navigation in the Arctic Passages. Currently, the acquisition of high-resolution distribution of TIT is primarily facilitated by thermal measurements derived from a one-dimensional (1-D) thermodynamic ice model. This study obtained spaceborne records over the Bering Sea and the Chukchi Sea for the last five years (2020-2024). The performance of this model is significantly influenced by the temperature of air (T-a), ice surface (T-s), and seawater freezing point (T-w). The TIT retrieval results demonstrate a high accuracy when compared with operational sea ice thickness products derived by microwave measurements (i.e. SMOS). The mean absolute difference between the two is similar to 0.04 m, and the TIT range of 0-0.2 m exhibits the most reliable accuracy. However, the sensitivity of this model is not applicable during the melt season with rising temperature and inhomogeneous sea ice cover. This limitation can be overcome through the development of an optical model based on the relationship between TIT and reflectivity characteristics. Additionally, notable discrepancies have been revealed between the spatial patterns of TIT retrieval results using optical and thermal measurements, particularly in regions with excessive sensitivity for the thermodynamic model when the temperature difference between the air and ice surface is minimal. The updated schemes for TIT retrieval can effectively compensate for the absence of high-resolution ice thickness products during the melt season, thereby further facilitating the long-term quantification of TIT.

High-resolution optical satellite sensors, such as the multispectral instrument (MSI) and the operational land imager (OLI), show excellent performance in capturing fine-scale sea surface wave motion and spatial patterns. Their interband time lag offers the potential to resolve wave directional ambiguity and parameter estimation. Here, we investigate the spatial resolution and channel time lag of optical sensors, based on cross-spectral analysis of multichannel imagery, to clarify their requirement in ocean wave monitoring. The analysis, validated by matched buoy data and optical imagery, demonstrates that the minimum detectable time lags for 180 degrees directional ambiguity removal are 0.47, 0.52, and 0.74 s for MSI 10-m, 20-m, and OLI 30-m resolution data, respectively. All three resolutions effectively detect wave system wavelengths ranging from 60 to 300 m, with minimum detection limits of 20, 40, and 60 m, respectively. Additionally, the optimal statistics window size for consistent wave detection is about 8 km. These findings not only highlight the strengths of current satellite sensors, but also provide References for future high-resolution optical sensor design in ocean wave monitoring.

High spatial-resolution Sentinel-2 Multi-Spectral Instrument (MSI) images often show image features due to surface waves, swells, fronts, internal waves, eddies, and currents under different sunglint and skyglint reflections. Here, MSI images are used to estimate wave direction and wavelength through a simple but practical method using Angular spectrum analysis (ASA). The method does not require the complex transfer function between the image domain and the physical domain, but is based on the MSI-derived Fast Fourier Transform (FFT) spectrum that reveals the wave direction, whose 180(o) ambiguity is then removed by the inter-band time lag between the B04 and B08 bands (0.74 s). The wavelength of different sea waves (wind wave and swell) can also be estimated by quantifying the MSI FFT images. There is a statistically significant linear relationship between MSI-derived and buoy-measured wave direction and wavelength (N = 144) with low bias. Further analyses according to wave age criterion show a better statistical relationship for wind waves (N = 62) than swells (N = 82). Comparison with the hourly wind products of the fifth generation European Centre for Medium-Range Weather Forecasts (ECMWF) atmospheric reanalysis (ERA5) shows RMS differences of <1.4 m/s and 14.4(o) for wind speed and direction, respectively. These results demonstrate that, under certain observing conditions, high spatial-resolution optical remote sensing images can provide relatively accurate estimates of surface wave directions and wavelengths as well as wind speeds, while operational applications still require further work.

The Bay of Campeche in the southern Gulf of Mexico is abundant with oil reservoirs and is also full of oil platforms and oil seeps. Accidental oil spills occasionally occur, as was the case with the most recent one linked to an explosion on the Nohoch Alfa platform on 7 July 2023, which was captured by multi-sensor satellite imagery (including radar, optical, and thermal data). In addition, we found another continuous and larger extended oil spill from a nearby platform. This latter spill had already occurred before the reported explosion, as evidenced by the satellite imagery captured on 6 July 2023. The spill area on July 6 was about 367 km(2), and large portions (similar to 16.5%) of the spilled oil were emulsified. From July 6 to July 24, about 1250 km(2) of the ocean was polluted due to the cumulative impact. Continuous oil spill from deep waters was also found from the thermal imagery. These findings suggest that multi-sensor satellite imagery, combined with appropriate algorithms, can be used to monitor and assess oil spill events, which also provides information that is otherwise impossible to obtain from other sources.

Estimation of the smoke emissions released by wildfires has become a significant concern in the context of atmospheric environment and global climate change, in which satellite-borne optical remote sensing plays an important role. Previous studies have utilized the satellite-based aerosol optical depth (AOD) algorithm to estimate the wildfire-related total particulate matter emissions (TPM). However, this "top-down" approach is often impeded by false or missed detection due to extensive cloud cover and smokes with high concentrations, resulting in an underestimation of TPM. To address these limitations, this study proposes a new ultraviolet (UV)based methodology for the identification and quantification of wildfire smokes from satellite platforms. The spectral characteristics of smoke plumes and different cloud types, including cirrus and cumulus clouds, are characterized by China's Haiyang-1C/D (HY-1C/D) satellites within both UV and visible wavelengths. Specifically, the 355 nm UV band displays a strong absorption pattern associated with smoke plumes. Based on this diagnostic spectral feature, an ultraviolet smoke index (UVSI) is developed, which can mitigate the impact of high background heterogeneity. Moreover, in comparison with the MODIS AOD algorithm, the UVSI index can not only improve the precision of smoke detection, especially for smokes with high concentrations, but also demonstrate a strong correlation with AOD at the numerical level. Accordingly, this study proposes the use of UVSI in place of AOD for TPM estimation. Combined with the fire radiative energy (FRE) measured by GOES-R satellite, the smoke emission coefficient (Ce) can be determined, which is 34.37 g MJ-1 for the September 2020 burning season in North America. The above results would provide a new reference for operational monitoring and assessment of wildfire smokes.

Based on the global demand for ocean color climate change,this paper reviews the development process of ocean color remote sensors in the past 50 years,sorts out the deployment and scientific objectives of ocean color remote sensors in earth observation missions in various countries,expounds the advantages and disadvantages of the current imaging systems of each remote sensor,and analyzes the historical evolution and technical constraints of remote sensor channel selection.According to the spatiotemporal coverage requirements of global observation,the design objectives of the current payload resolution and width are clarified,and the technical means of on-orbit calibration of ocean color remote sensors are listed and the importance of each means is analyzed.According to the list of remote sensors that have been launched and planned to be launched by countries in the past and next decade,the focus of countries' attention to global changes and coastal disasters is given.Combined with the development history of China's ocean color observation satellites,this paper clarifies the direction of technological breakthroughs in various countries to cope with ocean color climate change,and gives suggestions on the technical route of China's ocean color remote sensing and the idea of China's breakthrough in the field of innovation..

This paper presents a glint correction algorithm for high spatial resolution optical remote sensing imagery captured by the ER-2 Airborne Visual Infrared Imaging Spectrometer (AVIRIS). The algorithm employs linear and differential techniques to mitigate sun glint and sky glint effects, encompassing statistical glint reflections resulting from variations in imaging angles within strips and inter-strip variations due to Fresnel reflectance disparities. It aims to diminish Fresnel reflectance diversity on water surfaces and mitigate the distortions induced by glint reflectance during spectral and ocean color inversion. A comparative analysis of spectral and ocean color information in AVIRIS images before and after correction reveals enhanced accuracy following the glint correction. By systematically addressing multiple glint reflections and their ramifications, this method offers a valuable framework for correcting water surface glint in diverse high spatial resolution optical imagery.

Based on centimeter-level ultra-high-resolution UAV images, the drifting velocity of floating macroalgae are quantified. In this study, a Large Scale Particle Image Velocimetry (LSPIV) method is used for analyzing the drift of floating macroalgae in high-resolution Red-Green-Blue (RGB) images and videos collected from an unmanned aerial vehicles (UAVs) of hovering mode. The method employs Maximum Cross-Correlation (MCC) and Iterative Multigrid Approach (IMA) to achieve high spatial resolution and wide range of velocity gradient. Utilizing floating macroalgae as natural tracers, we enhance tracer signals using the Red-Green band virtual baseline Floating green Algae Height (RGFAH) index. Subsequently, LSPIV and a deep learning U-Net model are employed to acquire high spatiotemporal resolution information regarding the distribution and drift velocity of floating macroalgae. We then establish comprehensive instantaneous and time-averaged flow fields of floating macroalgae. Utilizing input data derived from alterations in water surface grayscale due to natural tracers such as water surface bubbles, suspended sediment, waves, and sun glint, we analyze the periodic variations in the sea surface velocity and wave intensity correlation based on statistical and Fourier analysis in instantaneous and time-averaged flow fields.