11:00 〜 11:15
[ACG43-08] 海氷漂流の風への応答 - 太平洋側北極海での観測
キーワード：海氷、ドリフター観測、波浪
Ocean surface processes are a central component in the climate system, because they connect the atmosphere and the ocean, which have very different timescales, via the exchange of heat, momentum, and materials. In the Arctic sea, the ocean surface processes are strongly modified by the presence of sea ice, which acts as a lid on the sea surface. To correctly predict the impact of the sea ice reduction on the Arctic climate, an understanding of sea ice dynamics is crucial. It is known that the sea ice motion is dominated by wind, but the analysis of sea ice drift is often based on a simple model assuming a constant wind factor α and turning angle θ, such that w=αW10exp(iθ), where w and W10 are complex sea ice drift and wind velocities. However, this model does not account for several dynamical processes such as transient response to wind and nonlinear relation between wind and drift. In this study, we report analyses of the drifter observation in the Pacific-side Arctic Sea to characterize the response of the drifter motion to the wind.
We used the hourly location data of nine Spotter wave buoys produced by Sofar Ocean, deployed in 2019/2020/2021 Arctic cruises of R/V Mirai. The buoys are solar-powered, so the data is concentrated in summer and autumn seasons. The number of hourly records is 12694 in total, which is approximately 529 days. To complement the buoy data, we used hourly 10 m wind speed data from ERA5 reanalysis and daily sea ice concentration (SIC) data from JAXA AMSR2 product. These data are interpolated to the buoy locations to obtain time series.
To characterize the drifters’ response to wind change, wavelet analysis is conducted. The continuous wavelet transform is applied to the drift velocity components, and the velocity magnitude spectrum is calculated as the magnitude of the complex coefficient vector. The resulting spectrum is conditionally averaged in time, depending on the SIC value. The velocity magnitude spectrum shows a peak near the 12-hour period, which represents the near-inertial motion. Both the near-inertial and longer-period motions are much weaker when SIC is higher than 0.8. This indicates that the energy that sea ice and oceanic motions gain is significantly reduced over a fully ice-covered sea compared to partially ice-covered sea and open-water conditions.
Similarly to the velocity magnitude spectrum, the conditionally-averaged coherence between the drift speed and wind speed, which approaches unity when wind and drift are aligned, is calculated. In all SIC classes, the coherence is greater than 0.5 for periods longer than 1 day, suggesting that the drifter motion of such temporal scales is dominated by wind.
Next, we characterize the mean wind-induced drift. Vector rotation is applied to drift velocity to obtain along-wind and across-wind components, which are then normalized with the wind speed and conditionally averaged for each ice condition and wind speed classes. In the open water and partially ice-covered conditions, normalized drift velocity takes a roughly constant value, with magnitude 3% of wind speed and turned 30° to the right of the wind direction. In the close-ice condition (SIC>0.8), however, the normalized drift velocity changes with wind speed, i.e., it shows nonlinear behavior. Its along-wind component grows larger as wind speed increases. This behavior is consistent with the theory by Thorndike and Colony [1982], explained with the shift of momentum balance (Coriolis and wind stress to water and wind stresses), which suggests the dominance of vertical processes.
We used the hourly location data of nine Spotter wave buoys produced by Sofar Ocean, deployed in 2019/2020/2021 Arctic cruises of R/V Mirai. The buoys are solar-powered, so the data is concentrated in summer and autumn seasons. The number of hourly records is 12694 in total, which is approximately 529 days. To complement the buoy data, we used hourly 10 m wind speed data from ERA5 reanalysis and daily sea ice concentration (SIC) data from JAXA AMSR2 product. These data are interpolated to the buoy locations to obtain time series.
To characterize the drifters’ response to wind change, wavelet analysis is conducted. The continuous wavelet transform is applied to the drift velocity components, and the velocity magnitude spectrum is calculated as the magnitude of the complex coefficient vector. The resulting spectrum is conditionally averaged in time, depending on the SIC value. The velocity magnitude spectrum shows a peak near the 12-hour period, which represents the near-inertial motion. Both the near-inertial and longer-period motions are much weaker when SIC is higher than 0.8. This indicates that the energy that sea ice and oceanic motions gain is significantly reduced over a fully ice-covered sea compared to partially ice-covered sea and open-water conditions.
Similarly to the velocity magnitude spectrum, the conditionally-averaged coherence between the drift speed and wind speed, which approaches unity when wind and drift are aligned, is calculated. In all SIC classes, the coherence is greater than 0.5 for periods longer than 1 day, suggesting that the drifter motion of such temporal scales is dominated by wind.
Next, we characterize the mean wind-induced drift. Vector rotation is applied to drift velocity to obtain along-wind and across-wind components, which are then normalized with the wind speed and conditionally averaged for each ice condition and wind speed classes. In the open water and partially ice-covered conditions, normalized drift velocity takes a roughly constant value, with magnitude 3% of wind speed and turned 30° to the right of the wind direction. In the close-ice condition (SIC>0.8), however, the normalized drift velocity changes with wind speed, i.e., it shows nonlinear behavior. Its along-wind component grows larger as wind speed increases. This behavior is consistent with the theory by Thorndike and Colony [1982], explained with the shift of momentum balance (Coriolis and wind stress to water and wind stresses), which suggests the dominance of vertical processes.