5:15 PM - 6:45 PM
[ACG34-P09] Parameterizations of the air-ice and ice-ocean drag coefficients depending on the roughness of sea ice floes
Keywords:air-ice drag coefficient, ice-ocean drag coefficient, sea ice
Both the ice-ocean drag coefficient Cio and the air-ice drag coefficient Cai are important parameters affecting sea ice motion, sea surface stress, and ice-ocean interactions in coupled climate system. The drags at upper and lower sea ice interfaces determine air-ice and ice-ocean momentum exchanges. Cio and Cai are expected to be large when the roughness of sea ice is large. In most climate models, however, constant drag coefficients regardless of roughness of sea ice are adopted for future climate projection. We can make a simple but serious mistake for understanding a future climate.
Theoretical parameterizations of Cio have been proposed using simple geometric properties (e.g., Lu et al. (2011)). They examined variations of Cio as functions of ideal sea ice parameters such as keel depths and their spatial intervals. However, even such simple parameters, actual observed data are difficult to acquire over the entire ice-covered ocean. Therefore, now we need a new practical parameterization to understand ice-coupled climate system and to predict a future climate.
In the present study, a parameter of the roughness (Δd) as an observable quantity is introduced. Δd is defined as the difference between daily averaged sea ice draft and daily mode one. Δd involves the geometric properties of both the pressure keel depths and their spatial intervals. We succeeded to parameterize both Cio and Cai as a function of Δd using in-situ moored upward looking sonar data and 10 m wind of reanalysis data. The results suggest that the constant coefficients are not adequate in the current situation in the Arctic Ocean, since the spatial and temporal variations of Δd were large. For example, Δd of geometrically rough multi-year ice was large all through the year. On the other hand, Δd of geometrically flat first-year ice at the initial thermodynamic growing stage was small but roughed first-year ice which experienced mechanical growth by the end of winter was large, which was almost same value as that of the multi-year ice. Cai was also increased as well as Cio depending on Δd, however, the rate of change Cai with Δd was much smaller than that of Cio. The ratio Cai/Cio, which affects "wind factor" (WF~sea ice speed/ wind speed), increased as Δd became small. In the present Arctic Ocean after a regime shift around 2007, the fraction of the flat sea ice has increased, then the spatially averaged Cio(WF) in the region where rough multi-year ice have been replaced by flat first-year ice has been considerably decreased (increased). The actual ocean circulation in the present Arctic Ocean would be less activated than in a case using constant Cio. The discrepancy leads to different heat transportation in the ocean, which in turn changes the behavior of sea ice in future projection.
Theoretical parameterizations of Cio have been proposed using simple geometric properties (e.g., Lu et al. (2011)). They examined variations of Cio as functions of ideal sea ice parameters such as keel depths and their spatial intervals. However, even such simple parameters, actual observed data are difficult to acquire over the entire ice-covered ocean. Therefore, now we need a new practical parameterization to understand ice-coupled climate system and to predict a future climate.
In the present study, a parameter of the roughness (Δd) as an observable quantity is introduced. Δd is defined as the difference between daily averaged sea ice draft and daily mode one. Δd involves the geometric properties of both the pressure keel depths and their spatial intervals. We succeeded to parameterize both Cio and Cai as a function of Δd using in-situ moored upward looking sonar data and 10 m wind of reanalysis data. The results suggest that the constant coefficients are not adequate in the current situation in the Arctic Ocean, since the spatial and temporal variations of Δd were large. For example, Δd of geometrically rough multi-year ice was large all through the year. On the other hand, Δd of geometrically flat first-year ice at the initial thermodynamic growing stage was small but roughed first-year ice which experienced mechanical growth by the end of winter was large, which was almost same value as that of the multi-year ice. Cai was also increased as well as Cio depending on Δd, however, the rate of change Cai with Δd was much smaller than that of Cio. The ratio Cai/Cio, which affects "wind factor" (WF~sea ice speed/ wind speed), increased as Δd became small. In the present Arctic Ocean after a regime shift around 2007, the fraction of the flat sea ice has increased, then the spatially averaged Cio(WF) in the region where rough multi-year ice have been replaced by flat first-year ice has been considerably decreased (increased). The actual ocean circulation in the present Arctic Ocean would be less activated than in a case using constant Cio. The discrepancy leads to different heat transportation in the ocean, which in turn changes the behavior of sea ice in future projection.