| Summary: | With the backdrop of continuous global change, it is beneficial to create consistent long-term records of sea ice area on regional scales for ice disaster prevention and risk mitigation. In this study, a piecewise multiple nonlinear regression model was developed to reconstruct long-term daily sea ice area dataset in the Bohai Sea from 1958 to 2015 by linking the related meteorological data and the satellite-derived ice area. The validation analysis show that related meteorological status corresponding to physical process had stable skill of predictive ability, which was able to account for 81% of the observational variance under consideration of sea ice state, freezing and melting phases. The reconstructed daily sea ice area dataset was further used to study the interannual and seasonal variability of sea ice area. The annual maximum ice area (AMIA) and the annual average ice area (AAIA) in the Bohai Sea exhibited a decreasing trend with fluctuation of -0.33 +/- 0.18% yr(-1) and -0.51 +/- 0.16% yr(-1) over the period of 1958-2015, respectively. The most obvious change of the Bohai Sea ice area occurred in time scale of similar to 30 years. The whole study period could be divided into slight increasing stage (1958-1980), significant decreasing stage (1980-1995), and moderate increasing stage (1995-2015). In most years, the annual changes of sea ice area showed an unimodal variation and the freezing period (similar to 65 days) was longer than the melting phase (similar to 40 days) due to the relatively higher freezing rate. In addition, high correlations between AAIA and Arctic Oscillation (AO) index (r=-0.60, p Peer reviewed
|
|---|