ClimaTea Journal Club: "On the nature of the sea ice albedo feedback in simple models"

This week GS Chris Horvat is going to lead the discussion and explore how modern GCMs simulate sea ice, and two ways modelers have been led into traps while modeling it. First, Moon and Wettlaufer (2014) (attached) explore modeling of growth and decay in variable ice-cover regions in GCMs, showing that the when simple sea ice models were extended to these regions (which constitute a majority of the Arctic and Antarctic regions), the extension: (a) "violates the basic rules of calculus" (b) "neglects a leading order latent heat flux" and therefore (c) cannot capture the influence of the ice-albedo feedback in 15 out of 29 of the AOMIP models (!) Second, time allowing, we will discuss a similar pathology in the treatment of ice thermodynamics by examining model sensitivity to floe size. A small manila folder of sensitivity studies has indicated that how we deal with sea ice lateral thermodynamics is not relevant to global sea ice cover. Inspired by Moon and Wettlaufer I will show that these studies (a) violate some basic rules of modeling (b) neglect a leading order mass flux and (c) may not capture the influence of the ice-albedo feedback in, well, all 29 of the AOMIP models (!).