|Khedouri et al. ||Estimate subsurface temperature from satellite altimetry.||Gulf Stream region (analyzed depth: upper 700 m)||The least square linear between XBT measurement and GOES satellite data and inferring subsurface thermal structure from sea surface topography.||There are high correlations between subsurface temperature measured by XBT and sea level height variability measured by GOES, and this relationship can be used to infer subsurface thermal structure.|| |
|Fiedler ||Detect potential surface manifestations of subsurface thermal structure.||Northern to Southern California Oceanic coastal (analyzed depth: upper 500 m)||Correlation analysis between surface temperature and vertical structure parameters (thermocline strength and mixed layer depth) were applied by linear regression. ||Temperature correlation profiles (between surface temperature and subsurface temperature) at different regions and seasons. ||Improvement of the precision of vertical structure parameters by considering the sea surface winds from satellite scatterometers or ocean color from new color sensors.|
|Chu et al. ||Determine the subsurface thermal structure from satellite SST observation.||South of China sea (analyzed depth: upper 600 m)||Parametric model (relationship between SST and subsurface parameters such as MLD, TBD, and TTG in multi time scales profiles).||The parametric model is more accurate than the simple method of estimating subsurface temperature. ||After testing the validity of the multi time scale hypothesis between the mentioned parameters, had better globally apply the multi time scale inverse method.|
|Fischer ||Estimating the vertical thermal structure.||Tropical Pacific (analyzed depth: upper 300 m)||Using the SSTA and SSHA data to compute regression matrices that were obtained from a forced integration of the Modular Ocean Model (MOM).||He obtained significantly better results by the multivariate projection method than by the univariate method.||For other regions, other surface observations such as salinity and pressure gradients seem to be necessary.|
|Willis et al. ||Estimation of steric height, heat storage, subsurface temperature, and sea-surface temperature variability.||Southwestern Pacific enclosing the Tasman Sea (analyzed depth: upper 800 m)||Combining altimetry height (AH) and SST with in situ data to produce improved estimates of steric height (SH), heat content, and temperature variability by linear regression.||Nine-year time series of heat storage and temperature variability are calculated. The RMSE of estimated approximately was 4.6 W/m2 in heat storage, 0.10 °C in subsurface temperature and 0.11 °C in surface temperature.|| |
|Ali et al. ||Determination the subsurface thermal structure from surface parameters.||Central Arabian Sea (analyzed depth: upper 300 m)||Estimation of the subsurface temperature from SST, sea surface height, wind stress, net radiation, and net heat flux by neural network approach.||There were proper relationships between the estimated temperature profiles and in situ observations, meanwhile, 50% and 90% of the estimations were a ±0.5 °C and ±1 °C error, respectively.|| |
|Guinehut et al. ||Estimation of the ocean 3-D temperature fields by combining Argo and remote-sensing data||North Atlantic Ocean (analyzed depth: upper 500 m)||Multiple linear regression was used to derive a synthetic 200-m temperature from simulated altimeter and SST data, then they are combined with individual Argo 200-m simulated temperature to correct the high-resolution synthetic T using an optimal interpolation method.||Obtaining a large reduction of RMSE by combining both data types of the large-scale and low-frequency temperature fields at 200-m depth (by a factor of four in large mesoscale variability regions) as compared to the results obtained using only in situ profiles.||(1) Providing a better estimation of the 3-D thermohaline structure of the ocean by this method.|
(2) Using the dynamic height instead of sea surface height for reducing the errors due to the regression method.
(3) Extending the analyses to other depths and also to the salinity fields.
|Swart et al. ||Estimating GEM and AGEM sections of temperature, salinity, and density.||South-western tip of South Africa and the Antarctic continent (analyzed depth: upper 2500 m) ||Using the CTD, satellite altimetry data (SLA, ADT), Argo float data of temperature and salinity and XBT data. 2-D GEM produced, and then dynamic topography data derived from satellite altimetry are combined with the gravest empirical mode (GEM) to obtain a 16-year time series of temperature and salinity fields. ||RMSE of assimilation for temperature, salinity, and density is 0.15 °C, 0.02, and 0.02 kg m−3, respectively. ||The accuracy of the AGEM to reproduce subsurface thermohaline conditions serves as a catalyst to further studies that utilize time series analysis.|
|Meijers et al. ||Estimating the four-dimensional structure of the ocean using satellite altimetry.||Southern Ocean (analyzed depth: upper 2000 m)||By using the altimetry SSH values and Argo floats, a GEM projection of temperature and salinity fields is presented and is used to correlate with the AVISO data to produce gridded, full depth, time-evolving temperature, salinity, and velocity fields.||Strongly correlated with in situ measurements and satellite data for estimating meaningful subsurface properties. The combination of altimetry with the GEM fields allows the resolution of the subsurface structure of the filamentary fronts and eddy features.||Authors confronted the smooth frontal and eddy features due to the limitation of the spatial and temporal resolution of the satellite altimetry, so they suggested the satGEM requires that each dynamic height be associated with just one T–S profile at each longitude.|
|Guinehut et al. ||Combining the main components of the global ocean observing system (in situ and satellite altimeter) using statistical methods and deriving high-resolution 3-D temperature and salinity fields.||Atlantic, Pacific, and Indian Oceans (analyzed depth: upper 1500 m)||Deriving synthetic temperature fields from altimeter and sea surface temperature observations and salinity fields from altimeter observations, through multiple/simple linear regression methods. Then combining the synthetic fields with in situ temperature and salinity profiles using an optimal interpolation method.||Up to 50% of the variance of the temperature fields can be reconstructed from altimeter and sea surface temperature observations and a statistical method. Because of reconstructing only about 20 to 30% of the signal altimeter observations for salinity, making the in situ observing system essential for its estimates.||In this study, only thermohaline fields are presented, but the Global Ocean Observation-based products also include 3-D geostrophic velocity estimates that are calculated using the thermal wind equation combined with absolute surface altimeter geostrophic currents and can be used towards integrating climate-relevant global ocean Datasets.|
|Wu et al.  ||Estimation of subsurface temperature anomaly by remote sensing data.||North Atlantic Ocean (analyzed depth: upper 2000 m) ||Training self-organizing map (SOM) neural network method using anomalies of SST, SSH, and SSS data from Argo gridded monthly anomaly datasets in order to estimate a STA.||Estimating STA time series for 1993–2004 from remote sensing SST. Obtaining good agreement between the STA estimations from the SOM algorithm and in situ measurements taken from the surface down to 700-m depth. SOM algorithm can only predict the STA that lies in the range of labeling data.|| |
|Su et al. ||Determining the subsurface temperature anomaly by satellite measurements (AMSR-E, MIRAS, AVISO altimetry).||Indian Ocean (analyzed depth: upper 1000 m)||Estimation STA from a suite of satellite remote sensing measurements including SSTA, SSHA, and SSSA by the support vector machine.||Detecting the SSTA, SSHA, and SSSA parameters had improved and describe the STA estimation accuracy.||Providing a useful technique for studying subsurface and deeper ocean thermal which has played an important role in global warming.|