Process studies on the interactions between spatio-temporal variations in Arctic longwave surface emission and boundary-layer humidity

Kuo et al, 2018 have demonstrated a reduction in modeled Arctic surface temperature biases in the Community Earth System Model of over 6 K relative to ERA-Interim Reanalysis by incorporating realistic spectral surface emissivity and ensuring consistent surface radiative physics across model components. However, CESM radiative flux diagnostics indicate that the magnitude of the TOA ice-emissivity feedback is only a few percent of the value of the ice-albedo feedback. In order to reconcile this contradiction, we explore a multivariate parameter space of the updated model (CESM.ɛ(ν)) and an original model ensemble (CESM1) to discern the causal mechanisms for CESM.ɛ(ν) warming and whether our inclusion of realistic longwave surface physics has uncovered compensating model errors in CESM.

From these diagnostic exercises, we find that due to higher spectral surface emissivity values, CESM.ɛ(ν) exhibits higher wintertime longwave surface upwelling radiation than CESM1 over sea-ice. CESM1 shows persistent near-surface supersaturation over the Arctic winter, with relative humidity with respect to ice above 130%. While such near-surface ice supersaturation has been measured over Arctic sea-ice, the frequency of such occurrence in CESM1 is inconsistent with observations (Treffesien et al, 2007). Since wintertime surface temperature biases in CESM.ɛ(ν) are only 1 K, its near-surface relative humidity with respect to ice are under 110% and thus in better agreement with the few direct measurements of this quantity that exist from the Surface Heat Budget of the Arctic Ocean mission (SHEBA) (Andreas et al, 2002; Makkonen and Laakso, 2005). Furthermore, the wintertime near-surface in-cloud ice mixing ratio is reduced in CESM.ɛ(ν) compared to CESM1, yet near-surface specific humidity is higher by two-fold, resulting in 25 W/m² higher longwave surface downwelling radiation.

Our model results suggest that measurements of spectral surface emissivity and the near-surface atmospheric state can identify the radiative, thermodynamic, and dynamic drivers of high-latitude surface temperature. Our findings suggest that observations, such as those as part of upcoming MOSAiC campaign, may be useful for systematically constraining models so that they exhibit reduced wintertime surface temperature biases.