Bohai Sea Ice Parameter Estimation Based on Thermodynamic Ice Model and Earth Observation Data

Juha Karvonen,Marko Mäkynen,Bin Cheng,Markku Similä,Timo Vihma,Lijian Shi

doi:10.3390/rs9030234

Juha Karvonen, Marko Mäkynen + Show 4 more

Open Access

https://doi.org/10.3390/rs9030234

Copy DOI

Abstract

We estimate two essential sea ice parameters—namely, sea ice concentration (SIC) and sea ice thickness (SIT)—for the Bohai Sea using a combination of a thermodynamic sea ice model and Earth observation (EO) data from synthetic aperture radar (SAR) and microwave radiometer. We compare the SIC and SIT estimation results with in-situ measurements conducted in the study area and estimates based on independent EO data from near-infrared/optical instruments. These comparisons suggest that the SAR-based discrimination between sea ice and open-water works well, and areas of thinner and thicker ice can be distinguished. A larger comprehensive training dataset is needed to set up an operational algorithm for the estimation of SIC and SIT.

Highlights

A large portion of the Bohai Sea is typically covered by sea ice during the winter from December to March
The synthetic aperture radar (SAR)-based sea ice concentration (SIC) estimates were compared to the SAR segment-wise SIC computed from open-water/sea ice classification based on the HJ-1B (9 January 2013) optical image
The MODIS-based ice thickness for the Bohai Sea is sensitive to the changes of air temperature and only works reliably in cold air and ice surface conditions—presumably, Ta below −10 o C [41]

Summary

Introduction

A large portion of the Bohai Sea is typically covered by sea ice during the winter from December to March. A large portion of the Bohai Sea is covered by sea ice during normal winters. The background ice thickness (h B in cm) was estimated using the HIGHTSI model and AMSR2 radiometer data, which were both interpolated to 1 km pixel size. The thermodynamic sea ice growth for the Bohai Sea is modeled using HIGHTSI. The local-scale spatial distribution of SIT is often inadequately described by the model This explains, for example, the negligible correlation between the in-situ and HIGHTSI SIT (r = 0.1). Earlier studies [69] demonstrated that the HIGHTSI model is capable of estimating the thermodynamic growth of sea ice with reasonable accuracy. The uncertainty of the HIGHTSI estimate for Vi is unknown

Results

Discussion

Conclusion