The accumulated remote sensing data of altimeters and scatterometers have provided new opportunities for ocean state forecasting and have improved our knowledge of ocean–atmosphere exchanges. Studies on multivariate, multistep, spatiotemporal sequence forecasts of sea level anomalies (SLA) for different modalities, however, remain problematic. In this paper, we present a novel hybrid and multivariate deep neural network, named HMnet3, which can be used for SLA forecasting in the South China Sea (SCS). First, a spatiotemporal sequence forecasting network is trained by an improved convolutional long short-term memory (ConvLSTM) network using a channelwise attention mechanism and multivariate data from 1993 to 2015. Then a time series forecasting network is trained by an improved long short-term memory (LSTM) network, which is realized by ensemble empirical mode decomposition (EEMD). Finally, the two networks are combined by a successive correction method to produce SLA forecasts for lead times of up to 15 days, with a special focus on the open sea and coastal regions of the SCS. During the testing period of 2016–18, the performance of HMnet3 with sea surface temperature anomaly (SSTA), wind speed anomaly (SPDA), and SLA data is much better than those of state-of-the-art dynamic and statistical (ConvLSTM, persistence, and climatology) forecast models. Stricter testbeds for trial simulation experiments with real-time datasets are investigated, where the eddy classification metrics of HMnet3 are favorable for all properties, especially for those of small-scale eddies.

This study uses a variational method combined with satellite observations to reconstruct three‐dimensional temperature and salinity profiles for the Northern Indian Ocean. Sensitivity experiments show that sea surface temperature dominantly improve the temperature reconstruction of upper 100 m; sea surface salinity determines salinity estimation in the upper 100 m; sea surface height anomaly dominates the reconstruction of thermocline. The reconstructed temperature fields can be greatly improved in the thermocline by removing barotropic signal from the altimeter sea surface height data through a linear regression method. Ocean reanalysis and in situ temperature and salinity data are used to evaluate the results of reconstruction. Comparing with Simple Ocean Data Assimilation (SODA) in 2016, the spectral correlation between the reconstruction and the SODA density anomalies show that the reconstruction fields can retrieve mesoscale and large‐scale signals better. Moreover, the reconstruction salinity is much more accurate than SODA salinity in the upper ocean over the Bay of Bengal (BoB). Compared with CTD section observations, the reconstruction fields can capture the mesoscale eddy structure in the Arabian Sea and BoB well, respectively. The long time series of reconstruction along Argo trajectory shows that the reconstruction fields can better reproduce the observed intraseasonal oscillations of thermocline/halocline in the BoB. Compared with the World Ocean Atlas 2013 climatology, the reconstruction fields can better characterize upper ocean water mass variability.

Using sea surface data, including sea surface temperature (SST), sea surface height (SSH), and sea surface salinity (SSS), from two reanalysis data sets as input, we explore the potential for retrieving subsurface three‐dimensional (3‐D) structures in the Southwest Indian Ocean (SWIO) based on the “interior+surface quasi‐geostrophy” (isQG) method. We find that the isQG method is particularly successful south of 35°S corresponding to greater SSH anomaly (SSHA) and to higher ratio of the barotropic mode (BT) and the first baroclinic mode (BC1). Meanwhile, higher‐resolution surface data may provide more coherent retrieved results with original reanalysis data in the inactive eddy regions north of 35°S, where coarser data set cannot resolve mesoscale signals. Based on a long time series reconstruction by the Simple Ocean Data Assimilation (SODA) surface data from 2011 to 2015 in the SWIO, our spectrum correlation analysis shows that the isQG solution achieves a better skill near the surface layer and under the thermocline depth with horizontal scale larger than 50 km, but it degrades from ~30‐ to 200‐m depth south of 35°S. In contrast, the spectral correlation north of 35°S is relatively low, with correlation coefficient beyond (below) 0.5 for signals with a wavelength longer (shorter) than 200 km near surface layer and from 250‐ to 500‐m depth. Mechanisms about the regional‐dependent isQG skill are also discussed. This study should have important implications for both operational oceanography and ocean dynamics research in the SWIO, including the greater Agulhas system.

Accurately inverting global and regional subsurface temperature (ST) by multisource satellite observations is a challenging but hot topic. This study proposes a new method to invert daily ST from the sea surface information in China's marginal seas based on generative adversarial network (GAN) model. The proposed GAN‐based model can project the STs from sea surface information (SLA, SSTA, SST) with a high resolution of 1/12°. A traditional regression‐based model, Modular Ocean Data Assimilation System (MODAS), is set up same experiments for comparison. The results show that the averaged root mean square error results are less than 1.45°C in the upper 200 m and the highest averaged R² of 0.97 at the 70 m level, which is better than that of MODAS. Errors analysis and typical oceanographic phenomena analysis results show the superiority of the proposed GAN‐based model in this study. This study can provide high‐precision daily ST data from sea surface information, which can be expanded to further studies on the interior ocean variation characteristics.

This study explores the seasonally lagged impact of wintertime sea surface temperature (SST) in the Atlantic main development region (MDR) on the subsequent Atlantic hurricane season. It is found that wintertime SST anomalies in the MDR can persist into the summer, explaining 42% of the variance in the subsequent hurricane season's SST during 1951–2010. An anomalously warm wintertime in the MDR is usually followed by an anomalously active hurricane season. Analysis shows an important constraint on the seasonal evolution of the MDR SST by the water vapor feedback process, in addition to the well‐known wind‐evaporation‐SST and cloud‐SST feedback mechanisms over the tropical North Atlantic. The water vapor feedback influences the seasonal evolution of MDR SST by modulating seasonal variations of downward longwave radiation. This wintertime thermal control of hurricane activity has significant implications for seasonal predictions and long‐term projections of hurricane activity over the North Atlantic.