The Role of Downward Infrared Radiation in the Recent Arctic Winter Warming Trend Steven Feldstein1, Tingting Gong2, Sukyoung Lee1 1Department of Meteorology and Atmospheric Science, Pennsylvania State University, University Park, PA 2Institute of Oceanography, Chinese Academy of Sciences, Qingdao, China 21st Conference on Atmospheric and Oceanic Fluid Dynamics, Portland, Oregon, June 26-30, 2017

Question: What process accounts for the inter-decadal Arctic Amplification trend? Data: European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-Interim) data

ΔC = ΔId + ΔIu + ΔFsh + ΔC Surface Energy Budget Analysis Trend (Δ) of the Surface Energy Budget terms (Lesins et al. 2012) ΔC = ΔId + ΔIu + ΔFsh + ΔC Downward IR Storage (very small) Upward IR surface turbulence heat fluxes conduction through ice Expressing the upward infrared radiation (IR) as -εσTs4, the energy balance equation can be written as ΔTs = (ΔId + ΔFsh + ΔC - ΔG)/(4εσTs3) Neglect ΔC and ΔG. Ts and SAT are highly correlated (r=0.97; Chen et al. 2003) over most of the Arctic.

Trend Calculation Δf = (δf/δxi)∆xi; f = SAT; ∆ = trend; (δf/δxi) =rσ(SAT)/σ(xi) = regression coefficient x1 = downward IR x2 = surface heat flux x3 = horizontal temperature advection; x4 = adiabatic warming/cooling For each winter, DJF mean values of f and xi for are subtracted in the regression coefficient calculation Regression coefficient expresses the intraseasonal relationship between f and xi Reference ENSO, our SOM papers, and Tim Palmer’s papers

Surface Energy Budget SAT trend SAT (Downward IR) SAT (surf heat flux) SAT (horiz temp adv) SAT (adiabatic warm/cool) Cooling Period 1980-1999 Mention surface heat flux warming (from above) large over Greenland Barents and Kara Seas. Results suggest that ice albedo feedback small. Warming Period 1991-2010 SAT trend driven primarily by trend in downward IR, not by the surface heat flux trend

Energy Budget (regression coefficients and trends) Downward IR Surface Heat Flux Horizontal Temp Advection Adiabatic Warming/Cooling Regression Coefficients Trends

What processes drives the downward IR trend? Downward IR trend index obtained by projecting daily downward IR (poleward of 70N) onto downward IR trend pattern Discuss increase in frequency of IR index. IR(x,t) = IRindex(t)IRtrend(x) + residual IRindex(t) = (∑i,jIR(x,t) IRtrend(x) cos θ)/(∑i,j IRtrend(x)2cos θ) e-folding time scale = 10 days

What processes drives the downward IR trend? Obtained from regression against the IR index Moisture flux & conv Total column water Downward IR εsT4 εsT4 (adv) Cooling Period 1980-1999 Warming Period 1991-2010 Downward IR trend is driven largely by a trend in the intrusions of warm moist air into the Arctic

Moisture flux (convergence) & 250-hPa streamfunction (1991-2010 trends) LVq Total moisture flux trend LV’TqC Moisture flux trend due to circulation trend LVCq’T Moisture flux trend due to specific specific humidity trend L(V’q’)T Moisture flux trend due to transients Moisture transport from midlatitudes 250-hPa Streamfunction

Daily Evolution Lagged pattern correlations between regressed fields and trend patterns Upward IR SAT Downward IR Mention time sequence of processes and time scale Moisture flux convergence Total Column Water Surface heat flux

Daily evolution (lagged regressions) Downward IR Upward IR Net IR Heat flux Surface air temperature Downward heat flux Note that at positive lags upward IR > downward IR and surf heat flux changes from downward to upward. Consistent with sea ice melting. Upward heat flux

250-hPa wave activity fluxes & streamfunction: Wave activity transits the Arctic Lag - 4 Lag - 2 Lag 0 Lag + 2 Lag + 4 Lag + 6 Lag + 8 Lag + 10

CONCLUSIONS Arctic Amplification: Most of the inter-decadal Arctic surface warming trend arises from a trend in downward IR, not a trend in the surface heat flux (ice albedo feedback). The downward IR increase is due to moisture flux intrusions associated with poleward propagating Rossby waves, which bring warm moist air in the Arctic. Moisture intrusions appear to be increasing in their frequency and amplitude. Ice-albedo feedback during the previous summer does not appear to account for Arctic Amplification.