| Summary: | Climate models underestimate Arctic sea ice loss through ocean-atmospheric interactions that are improperly modeled. One reason for this disconnect is the heterogeneity of sea ice in the marginal ice zone (MIZ), causing secondary circulations unable to be captured by climate models within the boundary layer over the MIZ (MIZ-ABL). Large-eddy simulations of the MIZ-ABL were conducted throughout this study to understand how surface-atmosphere fluxes, as well as the dynamics and thermodynamics of the boundary layer, change as the geometric pattern of sea ice changes. A simplified theoretical framework was proposed to non-dimensionalize the dynamics of the MIZ-ABL; this method captured the surface thermodynamics reasonably well, however, they were unable to predict the atmospheric dynamics, suggesting that the individual stabilities over each surface influence the dynamics separately. A suite of large-eddy simulations over idealized surface patterns (with equivalent ice fraction and average floe area) were used to demonstrate that spatial organization plays a crucial role in determining boundary-layer structure. A broader set of surface characterization metrics was then established, minimized, and analyzed (ice fraction, patch density, splitting index, and perimeter-area fractal dimension), detailing the first steps towards further development of methods to quantify the variability of binary surfaces. A method was then proposed to obtain a principal orientation of the surface relative to the mean wind. Real-world sea ice patterns were then simulated, showing that the ice fraction and geostrophic wind direction are not enough to predict bulk surface thermodynamic fluxes. Another simulation suite of real-world satellite-sensed sea ice maps were conducted to understand how each metric affects the MIZ-ABL. Roughness heterogeneity showed minimal contributions to the resulting atmospheric circulations, and a multi-linear regression of such features, with discussion on how future models may be generated, was presented in ...
|
|---|