This study validated wind speed (WS) and significant wave height (SWH) retrievals from the Sentinel-3A/3B and Jason-3 altimeters for the period of data beginning 31 October 2019 (to 18 September 2019 for Jason-3) using moored buoy data and satellite Meteorological Operational Satellite Program (MetOp-A/B) Advanced Scatterometer (ASCAT) data. The spatial and temporal scales of the collocated data were 25 km and 30 min, respectively. The statistical metrics of root mean square error (RMSE), bias, correlation coefficient (R), and scatter index (SI) were used to validate the WS and SWH accuracy. Validation of WS against moored buoy data indicated errors of 1.19 m/s, 1.13 m/s and 1.29 m/s for Sentinel-3A, Sentinel-3B and Jason-3, respectively. The accuracy of Sentinel-3A/3B WS is better than that of Jason-3. All three altimeters underestimated WS slightly in comparison with the buoy data. Errors in WS at different speeds or SWHs increased slightly as WS or SWH increased. Over time, the accuracy of the Jason-3 altimeter-derived WS improved, whereas that of Sentinel-3A showed no temporal dependence. The WSs of the three altimeters were compared with ASCAT wind data for validation purposes over the global ocean without in situ measurements. On average, the WSs of the three altimeters were lower in comparison with the ASCAT data. The accuracy of the three altimeters was found to be consistent and stable at low/medium speeds but it decreased when the WS exceeded 15 m/s. Validations of SWH against buoy wave data indicated that the accuracy of Jason-3 SWH was better than that of Sentinel-3A/3B. However, the accuracy of all three altimeters decreased when the SWH exceeded 4 m. The accuracy of Sentinel-3A and Jason-3 SWH was temporally stable, whereas that of Sentinel-3B SWH improved over time. Analyses of SWH accuracy as a function of wave period showed that the Jason-3 altimeter was better than the Sentinel-3A/3B altimeters for long-period ocean waves. Generally, the accuracy of WS and SWH data derived by the Sentinel-3A/3B and Jason-3 altimeters satisfies their mission requirements. Overall, the accuracy of WS (SWH) derived by Sentinel-3A/3B (Jason-3) is better than that retrieved by Jason-3 (Sentinel-3A/3B).
The validation of significant wave height (SWH) data measured by the Sentinel-3A/3B SAR Altimeter (SRAL) is essential for the application of the data in ocean wave monitoring, forecasting and wave climate studies. Sentinel-3A/3B SWH data are validated by comparisons with U. S. National Data Buoy Center (NDBC) buoys, using a spatial scale of 25 km and a temporal scale of 30 min, and with Jason-3 data at their crossovers, using a time difference of less than 30 min. The comparisons with NDBC buoy data show that the root-mean-square error (RMSE) of Sentinel-3A SWH is 0.30 m, and that of Sentinel-3B is no more than 0.31 m. The pseudo-Low-Resolution Mode (PLRM) SWH is slightly better than that of the Synthetic Aperture Radar (SAR) mode. The statistical analysis of Sentinel-3A/3B SWH in the bin of 0.5 m wave height shows that the accuracy of Sentinel-3A/3B SWH data decreases with increasing wave height. The analysis of the monthly biases and RMSEs of Sentinel-3A SWH shows that Sentinel-3A SWH are stable and have a slight upward trend with time. The comparisons with Jason-3 data show that SWH of Sentinel-3A and Jason-3 are consistent in the global ocean. Finally, the piecewise calibration functions are given for the calibration of Sentinel-3A/3B SWH. The results of the study show that Sentinel-3A/3B SWH data have high accuracy and remain stable.
The HY-2B satellite was successfully launched on 25 October 2018. One of the main payloads of the HY-2B was a radar altimeter. In the present study, the quality of the HY-2B along-track sea surface heights (SSH), significant wave heights (SWH), and sea surface wind speeds (SSWS) were assessed, including their precision and accuracy. In order to achieve this goal, the mono-mission metrics of the HY-2B were analyzed and compared with those of the Jason-2 and Jason-3 over the same periods of time. The results of both direct comparison and cross comparison methods were presented in this study. The comparison results indicated that the quality of the HY-2B satellite’s geophysical data records (GDRs) data was excellent, with 95% of the sea surfaces effectively observed between 82 degrees north and south latitudes. In addition, the standard deviation of the sea level anomalies (SLA) at the single mission crossovers was 4.6 cm to 5.8 cm, and at the dual-crossovers with Jason-3, the standard deviation was determined to be 5.1 cm to 5.8 cm. The accuracy levels of the significant wave heights and products of the HY-2B satellite radar altimeter were observed to be greater than 0.3 m and 1.4 m/s (STD), respectively. Therefore, it was concluded in this study that the data quality and system performance of the HY-2B satellite were excellent and stable, and could be widely utilized in such fields as global sea-level change monitoring, wave numerical assimilation predictions etc.
Significant wave height (SWH) forecasting is a key process for offshore and costal engineering. However, accurate prediction of the SWH is quite challenging due to the randomness and fluctuation features of waves. This paper employs a novel deep learning method, the gated recurrent unit network (GRU), to forecast SWH with lead times of 3, 6, 12 and 24 h. The data sets used in this study include the wind speed of the past 3 h and the current SWH as inputs, which were obtained from six buoy stations in the Taiwan Strait and its adjacent waters. The GRU results are compared with those of back propagation neural network (BP), extreme learning machine (ELM), support vector machine (SVM), and random forest (RF). Although the error indices of the six stations are different, the general performance of GRU is satisfactory, with a faster forecasting speed, smaller volatility and better adaptability. Using buoy station 46714D as an example, the root mean square error (RMSE) predicted by GRU reaches 0.234, 0.299, 0.371, and 0.479 with lead times of 3, 6, 12, and 24 h, respectively.
Automatic modulation recognition (AMR) of underwater acoustic (UWA) signals is incredibly challenging due to the complexity of UWA channels and the severity of ocean noise. In the presence of noise interference, single-modal features fail to fully represent the characteristics of different modulated signals. While the in-phase/quadrature (I/Q) and time-frequency maps can adequately represent the signal features in the time, frequency, and time-frequency domains, the direct integration of the two modalities is ineffective because of the variations in shape, information granularity, and noise manifestation. To address the low recognition rate caused by the above issues, we propose a two-stream transformer (TSTR) based network for AMR of UWA signals. First, the input pre-processing layer obtains the I/Q and time-frequency features from the received signals. Then, the feature capture layer extracts high-dimensional signal features in the time, frequency, and time-frequency domains. Finally, the classification layer estimates the modulation of the signals. A multi-head self-attention module with adaptive soft thresholding is used in the feature capture layer to provide noise reduction and redundant feature rejection while retaining context information. Moreover, multi-scale ghost convolution is employed to address the inability of the transformer to efficiently extract spatial characteristics from the signals. Results are presented using real UWA channels from the Watermark dataset for two different seas which show that the TSTR improves recognition by 1.2% and 5.9% over the best existing model. Further, it has better generalization capabilities and the model has a small number of parameters so the time complexity is low.