5:15 PM - 6:45 PM
[ACG42-P09] Regional characteristics of sea ice dynamics and its interannual variability in the Northern Hemisphere
Keywords:Sea ice, Arctic region, Sea ice deformation, Sea ice rheology
Recently sea ice conditions have been changing rapidly in the northern hemisphere (NH). The sea ice extent has decreased significantly especially in summer, resulting in the decrease in multi-year ice (MYI) fraction and increasing expanse of seasonal ice zones. Associated with the loss of MYI, mean ice thickness in the Arctic Ocean also decreased significantly. Some scientists point out that a regime shift in ice thickness distribution occurred in 2007 (Sumata et al., 2023). These significant changes in ice conditions have caused changes in sea ice dynamics such as increases in ice drift speed, the significant wave height in the open water, and granular ice fraction in the sea ice structure. The motivation of our study comes from a need to examine this change from the perspective of deformation processes. This issue is closely related to the treatment of sea ice rheology. While traditionally the Viscous-Plastic (VP) rheology (Hibler model) has been adopted in many numerical sea ice models, its validity was not necessarily confirmed and needs to be clarified based on the observational data. To address this, we examined the regional characteristics and the interannual variability of sea ice dynamics parameters such as ice drift speed, strain rate magnitude, divergent and shear components for the wintertime sea ice in NH, using AMSR-E&2 derived gridded ice drift datasets with a grid spacing of 60 km for the period 2002 to 2022.
We divided the sea ice extent in the NH into 12 regions based on the NSIDC definition and attempted to characterize the dynamics parameters and their interannual variability in each region. In the analysis we paid particular attention to the aspect ratio (e) of the elliptic yield curve for VP rheology as an indicator representative of the dynamical state. This is because e expresses various physical processes in a simplified way and is related directly to the dynamical processes in the numerical sea ice model (e=2 for Hibler model). According to Rothrock (1975) theory in which sea ice behaves like an isotropic plastic material and all the work done by internal stress be consumed for ridging, if an elliptic yield curve is adopted, e is expected to take a value of 1.73~2. For e>2, it is expected that part of the work done by internal stress was consumed for other processes such as lateral friction between floes. e<1.73 implies that we may need to reconsider the VP rheology. e can be estimated using ice drift data following the method of Stern et al. (1995) and Toyota and Kimura (2018). We estimated optimal e in individual regions for each month (December to April) and examined the geographical properties by mapping and its interannual variations. This analysis is expected to elucidate the dynamical properties related to deformation in each region and their interannual trends. Another merit of this analysis is it gives us an opportunity to examine the validity of the traditional Hibler model and if necessary, to modulate it easily as a tuning parameter in models.
The results are summarized as follows. 1) Ice drift speed was shown to increase by 10-15% per decade in every region of the Arctic Ocean, consistent with past results (e.g. Spreen et al., 2011). 2) In the Barents, Greenland, Hudson, Bering, and Okhotsk seas divergence and shear components contribute to deformation almost evenly, whereas the shear component works preferentially in the Arctic Ocean. 3) The VP rheology in Hibler model appears to work efficiently in most cases, but some regionality was found in optimal e values. Whereas higher values are estimated in the Greenland Sea (~2.3) and the Beaufort Sea (~2.0), relatively low values (~ 1.65) were observed in Hudson Bay (Fig). 4) An increasing trend was observed for optimal e values in the Beaufort Sea as well as for shear component of strain rate. This is consistent with the consideration that energy consumption due to lateral friction is becoming more significant in the Beaufort Sea.
We divided the sea ice extent in the NH into 12 regions based on the NSIDC definition and attempted to characterize the dynamics parameters and their interannual variability in each region. In the analysis we paid particular attention to the aspect ratio (e) of the elliptic yield curve for VP rheology as an indicator representative of the dynamical state. This is because e expresses various physical processes in a simplified way and is related directly to the dynamical processes in the numerical sea ice model (e=2 for Hibler model). According to Rothrock (1975) theory in which sea ice behaves like an isotropic plastic material and all the work done by internal stress be consumed for ridging, if an elliptic yield curve is adopted, e is expected to take a value of 1.73~2. For e>2, it is expected that part of the work done by internal stress was consumed for other processes such as lateral friction between floes. e<1.73 implies that we may need to reconsider the VP rheology. e can be estimated using ice drift data following the method of Stern et al. (1995) and Toyota and Kimura (2018). We estimated optimal e in individual regions for each month (December to April) and examined the geographical properties by mapping and its interannual variations. This analysis is expected to elucidate the dynamical properties related to deformation in each region and their interannual trends. Another merit of this analysis is it gives us an opportunity to examine the validity of the traditional Hibler model and if necessary, to modulate it easily as a tuning parameter in models.
The results are summarized as follows. 1) Ice drift speed was shown to increase by 10-15% per decade in every region of the Arctic Ocean, consistent with past results (e.g. Spreen et al., 2011). 2) In the Barents, Greenland, Hudson, Bering, and Okhotsk seas divergence and shear components contribute to deformation almost evenly, whereas the shear component works preferentially in the Arctic Ocean. 3) The VP rheology in Hibler model appears to work efficiently in most cases, but some regionality was found in optimal e values. Whereas higher values are estimated in the Greenland Sea (~2.3) and the Beaufort Sea (~2.0), relatively low values (~ 1.65) were observed in Hudson Bay (Fig). 4) An increasing trend was observed for optimal e values in the Beaufort Sea as well as for shear component of strain rate. This is consistent with the consideration that energy consumption due to lateral friction is becoming more significant in the Beaufort Sea.