Litcius/Paper detail

Sea Ice Thickness Estimation From TechDemoSat-1 and Soil Moisture Ocean Salinity Data Using Machine Learning Methods

Qingyun Yan, Weimin Huang

2020Global Oceans 2020: Singapore – U.S. Gulf Coast16 citationsDOI

Abstract

In this paper, two machine learning methods, specifically, convolutional neural network (CNN) and support vector regression (SVR), are employed for retrieving sea ice thickness (SIT) from TechDemoSat-1 (TDS-1) and Soil Moisture Ocean Salinity (SMOS) data. The input for both methods consists of scattering coefficient (σ°), the incidence angle (θ), sea ice salinity (S) and sea ice temperature (T). The first two variables are derived from the TDS-1 data, and the latter two are from the SMOS data. Evaluation of the proposed methods is based on measurements in 2017 and 2018 of thin sea ice with thickness less than 1 m. Comparisons showed good consistency between the derived and reference SIT, with correlation coefficients of 0.95 and 0.90 and root mean square differences of 5.49 cm and 7.97 cm for SVR and CNN, respectively. This demonstrates the capability of these machine learning-based methods and the utility of TDS-1 data for SIT retrieval.

Topics & Concepts

SalinitySea iceWater contentSupport vector machineCorrelation coefficientConsistency (knowledge bases)Convolutional neural networkCoefficient of determinationSeawaterEnvironmental scienceRemote sensingSea ice concentrationMean squared errorSea ice thicknessGeologyArtificial intelligenceComputer scienceMachine learningArctic ice packMathematicsClimatologyStatisticsOceanographyGeotechnical engineeringArctic and Antarctic ice dynamicsClimate change and permafrostCryospheric studies and observations
Sea Ice Thickness Estimation From TechDemoSat-1 and Soil Moisture Ocean Salinity Data Using Machine Learning Methods | Litcius