In this study, a novel end-to-end image fusion method , DCAFusion, is proposed. The method is based on Swin Transformer and introduces a dual cross-attention mechanism for the task of fusing infrared and visible, multi-focus and medical images. Although infrared and visible datasets contain source image pairs, they lack corresponding labels and cannot be trained by a unified supervised learning framework. To address this problem, DCAFusion designs image reconstruction blocks that generate reconstructed images as labels to guide model feature learning and provide dynamic information retention. The reconstructed images enable the full reference loss function to intervene in a supervised learning manner and participate in the computation of cross-attention scores through information mapping to more efficiently integrate complementary information between source images. In the comparison experiments of infrared and visible image fusion, DCAFusion's fusion result reaches 13.9420 and 1.0895 in the two metrics, ahead of the second-ranked 13.2547 and 0.9983, respectively. These metrics also maintain the lead in comparison experiments of other fusion tasks, which proves DCAFusion's unique advantages in fusion results.

Internal waves (IWs) are widely present in the global ocean and play a crucial role in ocean dynamics, material transport, and climate change research. However, due to the large-scale variations and complex morphological features of internal waves objects, existing detection methods have limitations in cross scale feature extraction and information fusion. To address the aforementioned issues, this letter proposes a Cascade Hybrid Global Adaptive Feature Fusion YOLO (CHGAFF-YOLO) model for ocean internal waves detection. First, we employ the HGNetV2 structure as the backbone network of YOLO to better capture global information and extract complex features in a lightweight manner. Second, we introduce a Cascade Hybrid Cross-head Attention Module (CHCAM), which integrates in-group cascade attention with cross-head parallel self-attention mechanisms to achieve optimized multi-scale feature extraction and enhance global feature representation capability. Finally, we design a Four-Head Adaptive Feature Fusion Module (FHAFF), which dynamically fuses feature information from different scales by constructing four detection heads, further enhancing cross-scale information interaction. Extensive experimental results show that our method significantly outperforms existing approaches on both the Sentinel-1 SAR and MODIS satellite internal waves remote sensing datasets.

Lagrangian drifter datasets are widely used in oceanographic research but remain limited in eddy identification. This study proposes an eddy identification method using segmented drifter trajectories. The idea is to take advantage of the cycloidal drifter in a swirling flow as a tracer to determine the eddy’s spinning direction. We use the density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm on eddy-trapped drifter trajectories to extract drifter spinning channels, which are used to establish criteria based on trajectory lengths and rotation angles, and then applied to drifters outside altimetrically identified eddies. Despite limitations related to the uneven spatial distribution and applicability of drifter datasets, the new approach captures an additional 25.48% of anticyclonic eddies (AEs) and 27.36% of cyclonic eddies (CEs) with the rest remaining outside eddy (OE), with 89.04% smaller than the local Rossby radius of deformation to be defined as submesoscale eddies, demonstrating its complementary capability to altimeter-based eddy identification. Furthermore, the algorithm is independently applied to drifter data in 2023, and the high-resolution Surface Water and Ocean Topography (SWOT) data products are used to validate the drifter-only identified eddies. The overall consistency between drifter-identified eddies and background fields reaches 94%, demonstrating the reliability and effectiveness of the drifter-based eddy identification method.

Ocean waves can either be the target of interferometric radar altimeter (IRA) observations or may affect observations of mesoscale and submesoscale ocean dynamic phenomena. Due to the low incidence angle of the IRA, in addition to the azimuth shift caused by the motion of the ocean waves, the range shift caused by the height of the ocean wave is also significant. In this article, based on previous research, we propose a new nonlinear integral transform relating ocean wave spectra to phase image spectra of an IRA at a low incidence angle that includes the effect of range shift. This new nonlinear integral transformation can be replaced by a quasilinear transformation when the nonlinear variation is weak or the wavelength of the ocean wave is large. We analyze the effect of swell and wind waves on the phase spectra under different conditions based on the new nonlinear integral transform through simulation. The height of the ocean wave and the incidence angle of the radar affect the range shift, leading to a range nonlinear effect in the phase spectra. The orbital velocity of the ocean wave affects the azimuth shift, leading to an azimuth nonlinear effect in the phase spectra. The wavelength and propagation direction of the ocean wave also induces a distortion in the phase spectra. Overall, the nonlinear effect has a small effect on the swell and a large effect on the wind wave.

The widespread presence of oceanic internal waves (IWs) across continental shelves, straits, and islands has been confirmed using multiple satellite payloads, including optical and synthetic aperture radar (SAR) sensors. However, the efficiency and accuracy of IWs detection are severely limited by the cloud contamination of optical images and the availability of SAR data. In other words, although IWs can be observed by multiple sensors, achieving full-time coverage remains challenging. The SAR altimeter (SRAL), characterized by high spatial resolution and continuous observation capability, holds substantial potential for IWs detection. Consequently, this study proposes a deep-learning-based method, named the IWs detection network (IWD-Net), to detect IWs from SRAL data. The IWD-Net is trained and tested in the Andaman Sea, achieving a detection precision of 96.2%. In addition, the model is subsequently applied to the South China Sea (SCS), where the detection precision of 94.9% reconfirms its robustness and reliability in detecting IWs. Statistical results indicate that the IWs detection efficiency using altimeter data improves by 227% compared to SAR and optical sensors combined. Finally, spatiotemporal analysis reveals that IWs are primarily concentrated in the western Luzon Strait and the Sulu Sea, but the seasonal variations of IWs in these two regions exhibit opposite trends: IWs are more active in summer/autumn within the western Luzon Strait, whereas IWs are more active in late winter/early spring within the Sulu Sea. These findings highlight the potential of altimeter data to fill gaps in IWs observations and enhance our understanding of ocean dynamics.

The echo signal received by ionosonde contains both ordinary wave (O-wave) and extraordinary wave (X-wave), and its separation result directly determines the accuracy of mode discrimination and ionospheric parameter inversion, which is of great significance to ionospheric research. The separation of O-wave and X-wave is exceptionally complex due to environmental noise, instrumental thermal noise, external interference, and the time-varying dispersion properties of the ionosphere itself. In this article, a method for separating O-wave and X-wave of ionosonde based on dual-channel phase difference statistics is proposed, which uses constant false alarm rate (CFAR) detection to extract the vertical ionospheric measurement echo signals and dynamically calculates the compensated phases of O-wave and X-wave separations at each frequency, achieving the robust and effective separation of O-wave and X-wave. The results of the measured data show that: 1) this method dynamically calculates the phase compensated for O-wave and X-wave separation by considering the variation of the compensation phase with frequency and time, through real-time statistics of the phase difference in the echo signal; 2) it compensates for the amplitude difference between the two channels and the signal-to-noise ratio of the separated O-wave and X-wave is improved; 3) it exhibits strong robustness and is suitable for vertical ionospheric signals in various modes; and 4) it demonstrates good performance, with the accuracy of O-wave and X-wave separation reaching 98.64%, which is 21.37% higher than the traditional dual-channel phase compensation methods.

Reconstruction of ocean subsurface temperature anomalies (STA) is of great significance for the study of complex and variable ocean phenomena. Nowadays, STA reconstruction using deep learning is an important method, but the current reconstruction method still has the problem of incomplete feature extraction for multi-scale ocean dynamical processes. To address this problem, this study proposes a Transformer Network based on K-means clustering, referred to here as KMT, to reconstruct the STA in the Western Pacific using Argo measurements and multi-source satellite remote sensing data, and compares it with the Transformer Network inversion. The results show that the reconstruction accuracy of KMT is significantly better than that of Transformer Network, especially in the upper 200 m ocean where the temperature change situation is complicated.The average RMSE of the STA reconstruction accuracy of Transformer Network is 0.91 °C, and the average MAE is 0.73 °C , while the average RMSE of the STA reconstruction accuracy of KMT reaches 0.74 °C , and the average MAE reaches 0.57 °C. This study provides a more efficient and accurate STA reconstruction method for the West Pacific region, which can help to better understand and predict complex dynamical processes at different depths in the ocean.

Ocean subsurface thermal structure prediction is an area of active research field because of its scientific importance attach to ocean dynamic, air-sea interaction, and climate change, but currently, most of the ocean temperature predictions are oriented to the sea surface temperature (SST) due to the lack of observed profile data inside the ocean, especially in some marginal sea areas, such as the South Yellow Sea Cold Water Mass (SYSCWM). In fact, the prediction for ocean subsurface thermal structure is more important than SST in some ocean fields. In this letter, a dynamic coupling vertical multifeature difference time series prediction model based on bi-long short-term memory (DVMFD-Bi-LSTM) is proposed for the subsurface thermal structure prediction in the SYSCWM. Bi-LSTM with the “bi-directional” structure enables information association in temporal dimension. The dynamic coupling vertical mechanism is used to realize the spatial correlation between two adjacent layers of the subsurface ocean, and the difference algorithm is introduced to ensure the accuracy and robustness of the new method. Besides, we construct multifeature datasets to improve data scale and quality and rely on a multistep prediction strategy for multiday prediction. For a more comprehensive evaluation, multiple groups of experiments are set up for comparison, and the RMSE of the new model is reduced to 0.517, and R2 is increased to 0.937, which verifies the good performance of it in both temporal and spatial dimensions.

Systematic attitude errors (the roll, pitch, and yaw errors) are important sources of altimetric error in interferometric radar altimeters (IRA). So far, the effect of systematic attitude errors on the measurement of IRA has been investigated based on the interferometric geometry, in which the impact of the attitude errors on the IRA image quality and the coupling between each attitude error are not considered. In this work, the influence of attitude errors on the measurement accuracy of IRA is reanalyzed according to the IRA imaging principle and the interferometry theory. The theoretical formulas derived in this paper demonstrate that the roll is the most important factor affecting the measurement accuracy of IRA. Although the altimetric error directly introduced by yaw and pitch is smaller, the yaw and pitch would cause the altimetric error to be offset along the azimuth direction, and the offset grows approximately linearly across the range direction in the IRA images. Meanwhile, the attitude jitter would also introduce additional high-frequency altimetric errors, although these high-frequency altimetric errors can be removed through a Gaussian low-pass filter with the cut-off frequency determined by wavelet analysis. To verify the validity of the theoretical formulas and exhibit the impact of diverse attitude errors, a full-link simulation method is proposed, which encompasses the entire process from IRA observation to imaging and interference processing. Furthermore, the theoretical results are compared with the results of the full-link simulation and the experimental data acquired by an airborne IRA.

Subsurface chlorophyll maxima (SCM) is a crucial ecological indicator for marine ecosystems. Previous studies have indicated that this phenomenon is globally widespread. Although the biogeochemical Argo assimilation results have yielded positive results, the sparse data prevents them from being effectively used in oceanographic operations. Considering the dependence of ocean parameter, a deep learning model termed AT-GRU based on gated recurrent units is proposed. By incorporating an attention mechanism, the model can effectively address missing data in the biogeochemical-Argo (BGC-Argo) profiles, achieving the transition of data from sparse to dense (Sparse2Dense) and improving the accuracy of estimating subsurface chlorophyll-a (Chla) concentration. Specifically, the dataset of satellite remote sensing data and the associated BGC-Argo profiles is first established. AT-GRU is employed to reconstruct Chla concentration profiles from 1 to 300 m, utilizing several sources of ocean surface data. Next, an in-depth investigation is conducted to determine the characteristics of SCM. The objective is to enable wider research on SCM by analyzing vertical Chla profiles in four geographical locations. Finally, the general improvement of skill performance metrics, with R-squared reaching 0.84, demonstrates the feasibility of the proposed methodology through extensive experiments. In addition, we apply AT-GRU to global surface satellite data from January 2023 and compare the results with numerical modeling data to further validate the performance. This study presents promising opportunities for leveraging artificial intelligence in subsurface oceanic phenomena with the idea of Sparse2Dense and holds significant implications for the field of marine ecology.

Efficient eddy trajectory prediction driven by multi-information fusion can facilitate the scientific research of oceanography, while the complicated dynamics mechanism makes this issue challenging. Benefiting from ocean observing technology, the eddy trajectory dataset can be qualified for data-intensive research paradigms. In this paper, the dynamics mechanism is used to inspire the design idea of the eddy trajectory prediction neural network (termed EddyTPNet) and is also transformed into prior knowledge to guide the learning process. This study is among the first to implement eddy trajectory prediction with physics informed neural network. First, an in-depth analysis of the kinematic characteristics indicates that the longitude and latitude of the trajectory should be decoupled; Second, the directional dispersion prior knowledge of global eddy propagation is embedded into the decoder of the EddyTPNet to improve the performance; Finally, EddyTPNet predicts global eddy trajectories through pre-training and adapts to complex local regions via model transfer. Extensive experimental results demonstrate that EddyTPNet can reliably forecast the motion of eddies for the next 7 days, ensuring a low daily mean geodetic error. This exploratory study provides valuable insights into solving the prediction problem of ocean phenomena by using knowledge-based time series neural networks.

Recently, fronts (eddies) at the margins of eddies (fronts) have been discovered by observing sea surface temperature (SST) and sea level anomaly (SLA) data. They can both induce strong vertical motions and submesoscale processes, and are important for the vertical exchange of ocean mass and energy as well as ocean ecological processes. However, it raises an important challenge about the global spatiotemporal distribution of eddy-induced fronts and frontal eddies. This letter proposes a deep learning (DL) approach, dubbed eddy-front association detection network (EFADN), that is appropriate for mining eddy-front associations (EFAs) to extract the features of eddy-induced fronts (anticyclonic and cyclonic eddy-induced fronts) and frontal eddies [frontal anticyclonic eddies (AEs) and frontal cyclonic eddies (CEs)] from SLA and SST satellite data during 2006–2015 in the global ocean. The EFADN model integrates encoder-decoder and attention structures. The introduced spatial attention (SA) module in attention structure utilizes large-scale convolutional kernels to extract spatial information, which enlarges the receptive field to enhance the recognition of topological structures between eddies and fronts, improving the ability of EFA detection. The results of comparative experiments demonstrate that EFADN surpasses the state-of-the-art (SOTA) eddy detection model. Ablation studies underscore the crucial importance of all modules within EFADN for achieving accurate detection of EFAs. Moreover, the spatiotemporal distribution characteristics of eddy-induced fronts and frontal eddies are displayed. They are widely dispersed in the western boundary current (WBC) and Antarctic Circumpolar Current (ACC) regions, and they are active in the boreal summer while weak in the austral summer.

Oceanic eddy is the ubiquitous ocean flow phenomenon, which has been the key factor in the transportation of ocean energy and materials. Consequently, oceanographic understanding can be enhanced by the intelligent identification of eddy. State-of-the-art deep learning technologies are gradually improving identification methods. This letter proposes the pyramid split attention (PSA) eddy detection U-Net architecture (PSA-EDUNet) that targets oceanic eddy identification from ocean remote sensing imagery. As for the PSA-EDUNet, its inspiration comes from U-Net, which contains encoder and decoder parts, making the integration of inferior and senior features efficient and ensuring the feature information will not be lost in large quantities through nonlinear connection mode. Meanwhile, the PAS module is introduced to enhance feature extraction. In terms of the fusion data, the sea surface feature is the main criterion of eddy identification, including sea surface temperature (SST) and sea level anomaly (SLA). The experiments are implemented on the Kuroshio Extension (KE) and the South Atlantic regions, the results demonstrate that the proposed method can outperform other methods, especially for eddy edges and small-scale eddies.

Tropospheric path delay (TPD) is one of the primary factors affecting the accuracy of sea level anomaly (SLA) measurements by satellite altimetry. This article focuses on the effects of differential tropospheric error (DTE), resulting from the different TPD between two antennas with different incident angles in the cross-track direction of the wide-swath interferometric altimetry (WSIA). Based on the principle of interferometry, this article presents the mathematical model of the DTE and conducts a simulation experiment for validation, demonstrating a tangent-squared of incidence angles relationship between the DTE and TPD. Combining observational parameters of the SWOT and Guanlan satellites, the DTE increases nonlinearly from the near nadir, reaching approximately 1 and 3 cm at the far end of the swath for SWOT and Guanlan satellites, respectively. In addition, a comparison of relative magnitude between the SLA and DTE is carried out by computing the root mean square (rms) values using the wavenumber spectrum. The ratio of rms values is larger in the typical strong current regions [Kuroshio, Gulf Stream, and Antarctic Circumpolar Current (ACC)], while smaller in the low-latitude tropics regions. Furthermore, the global distribution of the power ratio is analyzed by calculating the ratio of spectral densities between the SLA and DTE at different wavelengths. The ratio of power is largest at mesoscale (250 km) and gradually homogeneous over the global ocean at submesoscale (50 km).

Predicting the trajectory of ocean eddies can promote the understanding of the transport of matter and energy in the ocean. However, accurately and rapidly predicting the trajectory of eddies poses a significant challenge due to their intricate nonlinear motion within a physical environment. Regrettably, existing data-driven methods primarily focus on the migration and combination of models, as well as the fusion processing of diverse observational data on oceanic eddies. These ways often overlook the crucial aspect of modeling the underlying motion mechanism of the eddies. We believe that the expeditious and precise prediction of eddies is closely intertwined with the physical mechanism and historical time series. Consequently, a medium-range eddy trajectory prediction neural network (ETPNet) compliant with the physical constraint is proposed, which embeds the physical regulation, intrinsic relations, and mutual interactions into the network via constraints. Then, a novel variant of the long short-term memory (LSTM) cell is designed to enhance the dynamic interaction and representation ability of the features, constraints, and knowledge. Finally, a geographically informed comprehensive loss function for marine tasks is formulated, namely mean absolute geodetic error (MAGE), which optimizes the network in Euclidean and sphere space. The proposed network is evaluated by predicting the future seven days trajectory of anticyclone eddies in the$15^{\circ }\text{N}$to$40^{\circ }\text{N}$. The extensive experiments and evaluations demonstrate that the proposed network guided by the comprehensive loss function can implement a state-of-the-art performance. The code is available athttps://github.com/AI4Ocean/ETPNet

Wide swath near-nadir interferometric altimetry is a newly developed technology for sea surface height (SSH) measurement. However, the absence of actual measurement data makes this novel SSH mapping technique difficult to verify and apply for wide swath interferometric altimeters. To verify the designed performance of the scheduled wide swath single-pass interferometric altimeter in the "Guanlan Mission", an airborne campaign was carried out off the coast of Rizhao, China on November 16, 2020. An airborne dual-frequency interferometric radar altimeter system (ADIRAS) with a single-pass mode was utilized for SSH measurement as the first flight. Two pioneering and fundamental works have been conducted: an intensive altimetry error analysis according to the ADIRAS parameter settings along the incident direction, an effective SSH reconstruction approach based on a multichannel likelihood (ML) function, and detailed validation procedures through airborne campaigns illustrated in this study. The results indicated that the difference between the wave-induced sea surface elevation (WSSE) variances derived by the ML approach and GNSS buoy was 2 cm2, which was smaller than the results of the single band on Ku (11 cm2) and Ka (6 cm2). Moreover, the estimated Significant Waves Height (SWH) bias of joint bands was 10 cm, which was also superior to that of Ku (39 cm) and Ka (24 cm). Both simulated data and real airborne dual-frequency InSAR data were employed in this study for cross-validation of the proposed method. This approach represents an effective technique for SSH reconstruction of future spaceborne/airborne interferometric altimeters.

Interferometric radar altimeter (IRA) is a new ocean remote sensing sensor. It can be used to retrieve sea surface height (SSH) by means of cross-track interferometry. Compared with the traditional cross-track interferometric synthetic aperture radar (XT-InSAR), IRA works at very small incidence angles for higher altimetry sensitivity. In this case, multiple discontinuous surface scatterers at sea surface would be cut into a same range pixel which leads to severe layover. This layover induced by ocean waves will reduce the correlation between the master-slave images acquired by IRA and increase random interferometric phase noise. At present, how to quantitatively analyze the impact of the ocean wave layover on the decorrelation of IRA images is still a problem that needs in-depth discussion. In this letter, theoretical analysis of the effect of ocean waves on the decorrelation of IRA images has been carried out when the ocean waves layover is considered. And the theoretical results are also compared with the airborne IRA data. It is found that the layover of ocean waves has significant influence on the decorrelation between the master-slave IRA images, especially at very low incidence angles.

The eddy identification is an important part of human cognition of the ocean. Significant achievements have been made by using sea level anomaly (SLA) data observed by the altimeter. However, the abundant eddies, which do not cause sea surface characteristic anomalies, cannot be identified. In this study, the eddy subsurface vertical structure-oriented 3-D neural network is developed to classify the oceanic eddies. This study is among the first that explores the ability of deep learning in eddy identification with vertical structure. First, the purified eddy profiles dataset is constructed based on the fact that the structure derived from vertical profiles is highly correlated with the sea surface topography detected by altimetry. Then, the eddy vertical structure-oriented 3-D neural network based on the residual network (ResNet) is constructed, which can classify the eddies as anticyclonic eddies (AEs), cyclonic eddies (CEs), and noneddies (NEs) effectively. Furthermore, the spatial and temporal features can be combined in the proposed network as external factors. Meanwhile, through 3-D convolutions and 3-D pooling, the proposed network is capable of modeling 3-D eddy data and can be extended to the deeper network structure. Finally, the classification experiments are implemented to validate the performance of the proposed methodology. The most striking result emerging from experiments is that the proposed method can expand the capacity of eddy identification by using vertical profiles as calibrated by altimetry with competitive classification performance. Together these results provide important insights into the application of artificial intelligence in oceanic eddy research.

Although the zonal propagation of mesoscale eddies is intrinsically westward in oceans worldwide, eddies occasionally propagate eastward. Westward propagating eddies (WPEs) and eastward propagating eddies (EPEs) have distinctively different characteristics. In this article, the propagation features of WPEs and EPEs were studied using a combination of current, wind, and topographical data. We found that, except for in the North Atlantic, the energetic areas for WPEs and EPEs did not overlap. Furthermore, the propagation channels of WPEs and EPEs were located at different depths. The magnitude of change in current and wind stress simultaneously affected both the spread direction and velocity of eddies. The average translation speeds of eddies increased as subsurface current velocity increased, regardless of the angle between the direction of eddy propagation and flow. Compared with WPEs, increases in wind stress had little effect on the propagation speeds of EPEs. The directions of EPEs were more sensitive to changes in flow and wind speed. Even small changes in bathymetry slowed the propagation speeds of eddies. This topographical inhibition was biggest when the topographic gradients were around 0.02 and 0.04 for WPEs and EPEs, respectively. The propagation directions of EPEs were also sensitive to the surrounding terrain and tended to move along seamounts and ridges. These findings may refine eddy predictions and aid eddy observations in the future.

The interferometric radar altimeter (IRA) is one of the main payloads of the “Guanlan” ocean science satellite proposed by the National Laboratory for Marine Science and Technology of China. To evaluate the effectiveness and accuracy of the IRA in retrieving the ocean dynamic parameters, such as sea surface height (SSH), ocean wave spectrum, and wind speed, two airborne IRA experiments were carried out at Qingdao Xiaomaidao (XMD) sea area on March 31, 2019, and Rizhao sea area on November 16, 2020. In the present work, wave-induced sea surface elevation (SSE) and its spectrum have been retrieved based on the interferograms acquired by the airborne IRA. To suppress the random phase noise, a mean filtering algorithm has been used in the multilook process of calculating the complex IRA images. The results show that the size of the filter window has a significant effect on the retrieved SSE. If the size of the filter window along THE range direction is too large, the flat earth would cause the spectral density of the retrieved ocean wave to be higher. In addition, the comparisons of the retrieved spectra with the buoy measurements demonstrate that the swell can be well-retrieved by IRA images at low sea-state conditions with significant wave height (SWH) less than 0.7 m. However, for wind wave, because of the effect of the velocity bunching along the azimuth direction, the wind wave spectrum can be extracted only when it propagates approximately along the ground-range direction of the IRA images.

The spatio-temporal residual network (ST-ResNet) leverages the power of deep learning (DL) for predicting the volume of citywide spatio-temporal flows. However, this model, neglects the dynamic dependency of the input flows in the temporal dimension, which affects what spatio-temporal features may be captured in the result. This study introduces a long short-term memory (LSTM) neural network into the ST-ResNet to form a hybrid integrated-DL model to predict the volumes of citywide spatio-temporal flows (called HIDLST). The new model can dynamically learn the temporal dependency among flows via the feedback connection in the LSTM to improve accurate captures of spatio-temporal features in the flows. We test the HIDLST model by predicting the volumes of citywide taxi flows in Beijing, China. We tune the hyperparameters of the HIDLST model to optimize the prediction accuracy. A comparative study shows that the proposed model consistently outperforms ST-ResNet and several other typical DL-based models on prediction accuracy. Furthermore, we discuss the distribution of prediction errors and the contributions of the different spatio-temporal patterns.

The scanning microwave radiometer (RM) onboard the Haiyang-2A (HY-2A) satellite has low-frequency channels with the capability of observing sea surface temperature (SST) from space. To improve the accuracy of the HY-2A RM SST data, the intercalibration of RM brightness temperature (BT) and SST retrieval were carried out. Based on the simulated BTs using the microwave radiative transfer model (RTM), the double-difference approach was applied to perform the intercalibration of RM BTs for the 10.7, 18.7, 23.8, and 37.0 GHz channels, with the Global Precipitation Measurement (GPM) Microwave Imager (GMI) as the reference. The RM 6.6-GHz BTs were corrected with the RTM modeled BT on account of a lack of the similar channel in GMI. The comparison of RM original BTs with GMI and modeled BTs showed large biases and the strong dependence on latitude. We obtained latitude-dependent coefficients using robust linear regression for RM BT correction and applied recalibrated RM BTs for the SST retrieval. The validation of the RM retrieved SST showed the bias of -0.12 °C and the robust standard deviation (RSD) of 1.10 °C compared with buoy SST in the region between 70°S and 70°N. In the tropical and subtropical regions, the bias was -0.12 °C and the RSD was 0.93 °C. In addition, the relationships between the SST difference and the sea surface and atmospheric parameters were investigated. Both statistics of validation results and error analysis indicated significant improvement of RM SST accuracy.

This work presents EddyNet, a deep learning based architecture for automated eddy detection and classification from Sea Surface Height (SSH) maps provided by the Copernicus Marine and Environment Monitoring Service (CMEMS). EddyNet consists of a convolutional encoder-decoder followed by a pixel-wise classification layer. The output is a map with the same size of the input where pixels have the following labels {`0': Non eddy, `1': anticyclonic eddy, `2': cyclonic eddy}. Keras Python code, the training datasets and EddyNet weights files are open-source and freely available on https://github.com/redouanelg/EddyNet.

The reconstruction of high-resolution gridded altimetry maps from irregularly sampled along-track data remains a key challenge in ocean remote sensing science. Operational products use optimal Interpolation (OI) techniques, which may not deal with nonlinear dynamics at short space-time scales. Here, we investigate an analog data assimilation scheme to improve the reconstruction of fine-scale structures. The analog data assimilation combines an ensemble Kalman model and a dataset of exemplars issued from high-resolution numerical simulations to perform an exemplar-based spatio-temporal interpolation of along-track data. As a case-study, we consider a region in the South China Sea and demonstrate the proposition analog data assimilation outperforms the classical OI by about ≃ 20% in terms of mean square reconstruction error.

Haiyang-2 (HY-2) is the first marine dynamic environmental satellite of China, which was launched on August 16, 2011. The scanning microwave radiometer (RM) onboard HY-2 has low-frequency channels with the capability of observing sea surface temperature (SST) from space. In this paper, the Level 2A (L2A) SST products of HY-2 RM are evaluated. The global HY-2 RM L2A SST products are compared with the buoy SST measurements, WindSat SST, and National Oceanic and Atmospheric Administration Optimum Interpolation (OI) weekly SST products for the period from January 2012 to December 2014. The collocations of HY-2 RM, WindSat, and buoy SST data are generated with the spatial window of 0.25° and the temporal window of 0.5 h. The biases are -0.45 °C (RM minus buoy) and -0.41 °C (RM minus WindSat) and the corresponding standard deviations are 1.73 °C and 1.72 °C. The comparisons of the weekly averaged HY-2 RM and OI SST show that the biases of each week difference are from -1.06 °C to 0.48 °C with the mean value of -0.30 °C. The standard deviations of the SST difference are from 0.83 °C to 1.47 °C with the mean value of 1.05 °C. The relationships between SST difference and the sea surface and atmospheric parameters, such as wind speed, wind direction, SST, and water vapor are investigated.

A network virtual reality (VR) engine for interactive analysis and smart city immersive visualization which integrates several latest information technologies, i.e., Web virtual reality (WebVR), Internet of Things (IoT), and three-dimensional (3-D) geographical information system (3-DGIS) with peer-to-peer (P2P) network, has been proposed in this study. It can be used in big spatial data (e.g., remote sensing data) organization and presentation while achieving the online sharing by hash-based P2P networks. In the P2P network, a map, which contains the real geographic space users and the virtual scene user avatar with prep the virtual network nodes, has been made. Moreover, it can also support the following integrated functions of VRGIS: 3-D spatial analysis function, 3-D visualization for spatial information process, and Web engine for 3-D globe and smart big city. To evaluate WebVRGIS system, a user study had been conducted by comparing it with virtual community system. The performance evaluation results had confirmed the higher usability and user satisfaction of the proposed WebVRGIS system.

Abstract We present a method that improves the spatio‐temporal resolution of original data with fluid scalar data. The basis of the method, velocity estimation, is constructed to be inverse and optimized problem. We reduce the calculation cost through convex optimization and make the velocity field more accurate by coupling with Navier–Stokes equations. The spatial resolution receives significant enhancement by applying advection on original data with higher‐resolution velocity field data generated by our method. The temporal resolution is improved by generating intermediate velocity fields through the solution of Navier–Stokes equations. In this paper, we demonstrate that the accuracy of our velocity estimation method is clearly better than that of optical flow methods and the enhanced data show an attractive performance in fluid visualization. Copyright © 2016 John Wiley & Sons, Ltd.

Daily 5-km sea surface temperature (SST) data from the Visible and Infrared Scanning Radiometer (VIRR) onboard the Fengyun-3C (FY-3C) satellite are evaluated for latitudes greater than 60° N against in situ and daily 4-km SST data from the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard the Terra satellite. Approximately five months of data are used for the evaluation, during August to December 2014. The in situ and 4-km MODIS SST data are averaged and resampled at a spatial resolution of 5 km, which is similar to that of the VIRR SST data. The bias and standard deviation of the SST difference between VIRR and buoy data are -0.12 °C and 0.93 °C, respectively. A three-way error analysis is conducted between VIRR, MODIS, and buoy SSTs to obtain standard deviations of error for each kind of observation. The separated standard deviations of the error are found to be 0.91 °C for the VIRR SST, 0.20 °C for the in situ SST, and 0.51 °C for the MODIS SST. The error sources are discussed. The results indicate that the FY-3C/VIRR SST product needs to be improved.

ABSTRACT A novel hybrid level‐of‐detail (LOD) algorithm is introduced. We combine point‐based, line‐based, and splat‐based rendering to synthesize large‐scale urban city images. We first extract lines and points from the input and provide their simplification encoded in a data structure that allows for a quick and automatic LOD selection. A screen‐space projected area is used as the LOD selector. The algorithm selects lines for long‐distance views providing high contrast and fidelity of the building silhouettes. For medium‐distance views, points are added, and splats are used for close‐up views. Our implementation shows a 10 × speedup as compared with the ground truth models and is about four times faster than geometric LOD. The quality of the results is indistinguishable from the original as confirmed by a user study and two algorithmic metrics. Copyright © 2014 John Wiley & Sons, Ltd.

Since the launch of the Soil Moisture and Ocean Salinity (SMOS), different sea surface salinity (SSS) level 3 (L3) research products have been developed by several institutions, including the Barcelona Expert Centre (BEC) in Spain and the Centre Aval de Traitement des Données SMOS (CATDS) in France. This letter assesses the performances of BEC “Ocean Reprocessing Campaign 2012001” L3 data and CATDS Version 2 (V02) L3 data by comparing both of them with Argo upper salinity measurements (within 5-m depth). Twenty-month data from May 2010 to December 2011 are selected for the study. Through computing differences and root-mean-square errors of the differences between BEC and Argo data as well as between CATDS V02 and Argo data at each 1° ×1° grid point, it is observed that CATDS V02 data are of better quality along coasts and in high latitudes than BEC data owing to different-SSS-bias mitigation such as applying a correction with respect to the climatology in 5° ×5° boxes and using the same ocean target transformation for the whole processing. In open ocean, however, both SMOS data sets perform very well. Several regions are picked out to conduct the time series comparisons, and the results are similar.

Abstract This paper introduces a virtual city oriented VR–GIS platform which synthesizes several latest information technologies including virtual reality, 3D geographical information system, remote sensing, and multi‐dimensional visualization. The platform is a seamless integration of VR functions and GIS analysis methods, which can be used to organize and present massive spatial data. It also supplies 3D spatial analysis functions, 3D visualization for spatial process and natural simulation, and serves as an engine platform for digital city. Copyright © 2009 John Wiley & Sons, Ltd.

The scientific term "amphidrome" is usually associated with tides in oceanography. The dozen tidal amphidromes observed in the ocean are critical points that determine the fundamental pattern of the global tidal system. Exploration of recently available satellite data with an unprecedented 1-2 decades duration suggests that an amphidrome is not a tide-only phenomenon in the ocean. Analysis of altimeter-derived sea level anomaly (SLA) data and radiometer-derived sea surface temperature (SST) data allows ten amphidromic points to be clearly identified in annual SLA and SST variations. These amphidromes are located in the tropical areas of the Pacific, Atlantic, and Indian oceans. Their existence implies that the annual cycle (in time) of the atmosphere-ocean system is translated into a rotary variation (in space) for many of the geophysical parameters. It can be concluded that annual amphidromes are common, the knowledge of which is of particular interest, given their annually "constant" nature, for the monitoring and understanding of oceanic, climatic, as well as biological variabilities at seasonal to decadal scales, which strongly affect many aspects of the natural and societal activities on the globe.

Using 9.5 years of simultaneous TOPEX and TMR (TOPEX microwave radiometer) data, a joint TOPEX/TMR precipitation index, which is generally consistent and quantitatively comparable with existing precipitation climatologies such as GPCP (global precipitation climatology project), is defined. The new TOPEX/TMR precipitation climatology, on the other hand, confirms the fundamental features of global oceanic rainfall; on the other hand, reveals a number of interesting characteristics which are previously unknown or poorly defined.