In this study, we investigate the feasibility of using historical remote sensing data to predict the future three-dimensional subsurface ocean temperature structure. We also compare the performance differences between predictive models and real-time reconstruction models. Specifically, we propose a multi-scale residual spatiotemporal window ocean (MSWO) model based on a spatiotemporal attention mechanism, to predict changes in the subsurface ocean temperature structure over the next six months using satellite remote sensing data from the past 24 months. Our results indicate that predictions made using historical remote sensing data closely approximate those made using historical in situ data. This finding suggests that satellite remote sensing data can be used to predict future ocean structures without relying on valuable in situ measurements. Compared to future predictive models, real-time three-dimensional structure reconstruction models can learn more accurate inversion features from real-time satellite remote sensing data. This work provides a new perspective for the application of artificial intelligence in oceanography for ocean structure reconstruction.

Quantifying the diffuse attenuation coefficient of the photosynthetically available radiation (K_par) can improve our knowledge of euphotic depth (Z_eu) and biomass heating effects in the upper layers of oceans. An algorithm to semianalytically derive K_par from remote sensing reflectance (R_rs) is developed for the global open oceans. This algorithm includes the following two portions: (1) a neural network model for deriving the diffuse attention coefficients (K_d) that considers the residual error in satellite R_rs, and (2) a three band depth‐dependent K_par algorithm (TDKA) for describing the spectrally selective attenuation mechanism of underwater solar radiation in the open oceans. This algorithm is evaluated with both in situ PAR profile data and satellite images, and the results show that it can produce acceptable PAR profile estimations while clearly removing the impacts of satellite residual errors on K_par estimations. Furthermore, the performance of the TDKA algorithm is evaluated by its applicability in Z_eu derivation and mean temperature within a mixed layer depth (T_ML) simulation, and the results show that it can significantly decrease the uncertainty in both compared with the classical chlorophyll‐a concentration‐based K_par algorithm. Finally, the TDKA algorithm is applied in simulating biomass heating effects in the Sargasso Sea near Bermuda, with new K_par data it is found that the biomass heating effects can lead to a 3.4°C maximum positive difference in temperature in the upper layers but could result in a 0.67°C maximum negative difference in temperature in the deep layers.

This letter describes the new ability to measure wind-generated gravity waves using the airborne $Ka$ -band interferometric altimeter (AirKaIA). Although the original definition of wave parameters is derived from wave-induced sea surface elevation (WSSE), it is difficult to directly measure the WSSE, so most remote sensing of waves use other types of signals to obtain wave parameters. Here, we present the measurement of 2-D WSSE from AirKaIA. A new wave retrieval method based on the 2-D WSSE has been developed, tested with simulation data, and applied to AirKaIA data. The results indicate that the retrieved dominant wave directions, dominant wavelengths, and significant wave heights from AirKaIA are in agreement with in situ measurements while highlighting the need for further method tests using more observations. The study of ocean signals from interferometric altimeter is an emerging research topic. AirKaIA’s measurements of wind-generated gravity waves with 2-D WSSE have implications for the assessment of future interferometric altimeter missions.

Satellite remote sensing enables the extensive, long-term observation of oceanic changes. To achieve transparent ocean observation, artificial intelligence is innovatively used to reconstruct the three-dimensional ocean structure from remote sensing data with spatiotemporal attention. However, high-resolution satellite imagery imposes high computational demands. This study proposes Spatiotemporal Window Ocean (SWO), which is an efficient network that uses a special calculation strategy and spatiotemporal window attention to reduce complexity. The model integrates multiple satellite data sources, including sea surface temperature (SST), absolute dynamic topography (ADT), and sea surface salinity (SSS). Our experimental results demonstrate that SWO achieves lower computational costs and superior performance compared with recent commonly used spatiotemporal sequence models. Specifically, SWO requires a training time equivalent to 1/3 of SimVP, 1/4 of PredRNN, and 1/5 of SA-ConvLSTM while achieving a root mean square error (RMSE) index that is 13.8%, 20.3%, and 13.2% better, respectively. The computational advantages of SWO offer an important technical means for the high-resolution reconstruction of ocean phenomena in the future.