| Summary: | With the declining Arctic sea ice extent, it is increasingly demanded for high resolution sea ice concentration (SIC) monitoring. This paper introduces a new thermal infrared ice concentration algorithm (TIRIA), which directly uses thermal infrared (TIR) brightness temperatures thus does not rely on surface temperature products as traditional algorithms. Factors such as sea water salinity and observation angle and their impacts on ice/water brightness temperature are accounted. TIRIA and a traditional algorithm, namely MPA, were applied to FY-3D MERSI-II TIR imagery. Results were inter-compared with passive microwave (PM) SICs and validated with near-infrared (NIR) SICs. Overall speaking, both TIRIA and MPA tend to underestimate SIC in ice-covered area, while overestimate in open water area. Compared to MPA, the overestimation of SIC is largely mitigated in TIRIA. This consequently leads to an overall better performance of TIRIA. Although the bias tends towards slightly negative value, the higher correlation meanwhile smaller MAE and RMSE indicates good potentials of the new algorithm. TIRIA can be theoretically applied to any TIR data. However, factors such as cloud mask and data accuracy play important roles in the TIR SIC retrieval thus should be taken with caution.
|
|---|