Abstract:Sea surface temperature (SST) and wind speed (SSWS) are two significant parameters in the coupled ocean-atmosphere system. Nowadays, spaceborne microwave radiometers are the main approaches to measuring global SST and SSWS with high coverage and high accuracy. For conventional processing of passive microwave data, radiometer calibration is conducted to convert the radiometer raw output voltage data into the top-of-atmosphere (TOA) brightness temperatures (TBs). After calibration, sea surface parameters are retrieved from TBs using statistically or physically based retrieval methods. Considering the complex procedures of instrument calibration and retrieval, a novel estimation method based on deep learning is proposed to obtain SST and SSWS directly from radiometer voltages. A comprehensive matchup dataset, including the data measured by the scanning microwave radiometer (SMR) onboard the Chinese HY-2B satellite, the European reanalysis 5 (ERA5) products, and the WindSat measurements, is used to develop and test the deep learning models. Validation against ERA5 and WindSat products indicates that the deep learning models perform well. To compare with traditional methods, we also retrieve SST and SSWS from TB data of SMR using a statistical regression retrieval algorithm. The comparison suggests that the deep learning models provide results close to and sometimes better than the regression algorithm. Furthermore, the feature importance of deep learning models and the dependence of their performances on sea state are analyzed. In this article, it is demonstrated that the deep learning method is a reliable and feasible tool for SST and SSWS estimation from the radiometer voltages with high accuracy, which can simplify the data processing procedure and improve processing efficiency.

Passive ocean acoustic thermometry with machine learning

Sparsity-based Approach for Ocean Acoustic Tomography Using Learned Dictionaries

Study of Acoustic Tomography Based on Data Assimilation for Shallow Water Environments

Ocean Acoustic Tomography As a Data Assimilation Problem

Enhancing Remote Sensing of Water Temperature Using Acoustic Tomography and 3-D-RPS Algorithm

Short-Range Water Temperature Profiling in a Lake with Coastal Acoustic Tomography.

An Observation Instrument for Artificial Upwelling Based on Coastal Acoustic Tomography

Coastal Ocean Current Tomography Using a Spatio-Temporal Kalman Filter

Acoustic Data Assimilation: Concepts and Examples

Underwater Sound Speed Field Forecasting Based on the Least Square Support Vector Machine

Machine learning approach to prediction of real-time ocean sound speed profile

A Multi-Spatial Scale Ocean Sound Speed Prediction Method Based on Deep Learning

Sound Speed Inversion Based on Multi-Source Ocean Remote Sensing Observations and Machine Learning

A physics-informed machine learning approach for predicting acoustic convergence zone features from limited mesoscale eddy data

Ocean Surface Parameters Estimation From Microwave Radiometer Voltages Using Deep Learning

Using Ocean Ambient Sound to Measure Local Integrated Deep Ocean Temperature

Estimation of Sound Speed Profiles Using a Random Forest Model with Satellite Surface Observations

Learning-based estimation of in-situ wind speed from underwater acoustics

A Sound Ray Correction Method Based on Historical Data of Marine Acoustic Environment

An Ensemble-Based Machine Learning Model for Estimation of Subsurface Thermal Structure in the South China Sea