Modeling Extreme Warm‐Air Advection in the Arctic: The Role of Microphysical Treatment of Cloud Droplet Concentration

Abstract

As the Arctic climate is changing fast, with increasing areas of open water in summer, there is a growing interest in the processes related to the marginal ice zones. Recent studies have indicated that such a critical process may be the advection of warm and moist air from the south. In this study, the performance of the Weather Research and Forecasting (WRF) model is evaluated during an extreme warm advection episode over melting sea ice that occurred near the Arctic ice edge in summer 2014. The model gives a reasonably good representation of the atmospheric conditions and the Arctic boundary layer, characterized by very strong surface inversions and the frequent presence of low‐level jets. However, the representation of the highly variable cloud conditions, from optically thick to optically thin, dissipating clouds, is sensitive to the choice of cloud droplet treatment in WRF. Simulations with relatively high cloud droplet number concentrations (Ndrop ≥ 100 cm−3) are more successful in representing the optically thick cloud state, whereas to reproduce optically thin and tenuous clouds Ndrop should be <50 cm−3. The WRF‐Chem model, with a realistic treatment of the cloud‐aerosol interactions, allows for large variations in Ndrop and hence can reproduce the cloud water properties reasonably well for most of the simulation time. This contributes to an improved representation of the cloud longwave radiative effect, compared to the simulations where a less adaptive treatment of Ndrop is applied.

Modeling Extreme Warm‐Air Advection in the Arctic During Summer: The Effect of Mid‐Latitude Pollution Inflow on Cloud Properties

“…The transition Elev‐Cl period exhibits the largest N drop concentrations (>200 cm −3 , Figure 3 and mean value 63 cm −3 , Table 1) and the highest LWC (Figures 10a1 and 10b1), in agreement with the LWP observations (Sotiropoulou et al., 2019). The largest accumulation in cloud number concentrations are also found during this period (mean value 72 cm −3 , Figure 7d1), while Aitken number concentrations in cloud are <3 cm −3 (mean value <1 cm −3 , Figure 7b1).…”

“…All four domains have a high vertical resolution with 40 terrain‐following sigma levels, where the model top is set to 10 hPa. The lowest 1 km includes 16 model levels with the lowest model level at about 14 m. The simulation period spans from July 25, 2014 00:00 UTC to August 9, 2014 00:00 UTC, split in 48 h segments; a 24 h spin‐up is commonly used in WRF Arctic atmospheric/meteorological studies (e.g., Hines & Bromwich, 2017; Sotiropoulou et al., 2019). Although the days of interest span from July 29 00:00 UTC to August 9 00:00 UTC, the simulations started 3 days earlier, on July 25, so that the chemical and aerosol pollution can be advected to the innermost domain through large‐scale transport.…”

“…An extensive discussion about model set up and evaluation of the model results with ACSE observations can be found in Sotiropoulou et al. (2019). Table gives a brief description of the configurations and physical options (parameterizations) used in simulations including sea‐ice.…”

“…A comparison of WRF‐Chem and ACSE observations from the extreme advection period is presented in Sotiropoulou et al. (2019). They showed that the WRF‐Chem can successfully reproduce the enhanced heat (Figures 5a and 5b) and moisture transport (Figures 5c and 5d) observed during this episode and realistically reproduce the variable cloud conditions.…”

“…This particular episode resulted in anomalously high‐air temperatures over the East Siberian Sea (∼6°C higher than the 1981–2010 average temperature) and accelerated sea‐ice melting (Tjernström et al., 2015). The episode is simulated with the Weather Research and Forecasting model, fully coupled with chemistry (WRF‐Chem), which explicitly solves cloud‐aerosol interactions; this model has been shown to successfully reproduce the thermodynamic aspects of this episode and to realistically represent cloud characteristics (Sotiropoulou et al., 2019). Through a number of sensitivity simulations we aim to: (1) investigate the effect of the mid‐latitude pollution transported in the study area (anthropogenic, BB, and their synergy) on the vertical distribution of aerosol number concentration and chemical composition, (2) investigate how fog and low‐level clouds interact with aerosols throughout this episode, (3) quantify the contribution of each emission inflow in cloud formation, radiation budget and CRF during such warming episodes.…”