On Warm and Moist Air Intrusions into Winter Arctic

. In this study, warm and moist air intrusions (WaMAI) over the Arctic Ocean sectors of Barents, Kara, 8 Laptev, East Siberian, Chukchi and Beaufort Seas in recent 40 winters (from 1979 to 2018) are identified from 9 ERA5 reanalysis using both Eulerian and Lagrangian views. The analysis shows that WaMAIs, fuelled by Arctic 10 blockings, causes a relative surface warming and hence a sea ice reduction by exerting positive anomalies of net 11 thermal irradiances and turbulent fluxes to the surface. Over Arctic Ocean sectors with land-locked sea ice in 12 winter, such as Laptev, East Siberian, Chukchi and Beaufort Seas, total surface energy budget is dominated by 13 net thermal irradiance. From a Lagrangian perspective, total water path (TWP) increases linearly with the 14 downstream distance from the sea ice edge over the completely ice-covered sectors, inducing almost linearly 15 increasing net thermal irradiance and total surface energy-budget. However, over the Barents Sea, with an open 16 ocean to the south, total net surface energy-budget is dominated by the surface turbulent flux. With the energy in 17 the warm-and-moist air continuously transported to the surface, net surface turbulent flux gradually decreases 18 with distance, especially within the first 2 degrees north of the ice edge, inducing a decreasing but still positive 19 total surface energy budget. The boundary-layer energy-budget patterns over the Barents Sea can be categorized 20 into three classes: radiation-dominated, turbulence-dominated and turbulence-dominated with cold dome, 21 comprising about 52%, 40% and 8% of all WaMAIs, respectively. Statistically, turbulence-dominated cases with 22 or without cold dome occur along with one order of magnitude larger large-scale subsidence than the radiation- dominated cases. For the turbulence-dominated category, larger turbulent fluxes are exerted to the surface, 24 probably because of stronger wind shear. In radiation-dominated WaMAIs, stratocumulus develops more strongly 25 and triggers intensive cloud-top radiative cooling and related buoyant mixing that extends from cloud top to the 26 surface, inducing a thicker well-mixed layer under the cloud. With the existence of cold dome, fewer liquid water 27 clouds were formed and less or even negative turbulent fluxes could reach the surface.

terms are also interpolated along the trajectories as previously discussed (also see You et al. 2020You et al. , 2021.
Note that while the surface energy budget depends on the surface fluxes, the atmospheric energy budget depends 119 in the vertical gradient of fluxes.

122
EMIs were identified in the Arctic ocean basins of Barents, Kara, Laptev, East Siberian, Chukchi and Beaufort.  are also found in all other ocean basins.

132
As warm and moist air is advected into the Arctic over the Barents Sea, it interacts with the cool ice 149

155
The composites of large-scale pattern discussed above are extracted from the stronger EMI events to

164
In this section, we will explore the transformation of temperature inversion, cloud formation and surface energy-

169
Over the completely ice-covered sea sectors such as the Laptev, East Siberia, Chukchi and Beaufort Seas, strong 170 temperature inversion develops with cloud formation below, as the warm-and-moist air propagates over the sea 171 ice. A detailed analysis of this boundary-layer structure follows in Section 3.3. In this case, hsh is above ht, and 172 both higher than ℎ ( Figure 9a). From the ice edge and onward up to 10 degrees north of the ice edge, hsh, ht and 173 ℎ increase almost linearly, by 30-40 m degree -1 (Figure 9a). TCW and PRCP also increase northward, although 174 more slowly for the first two degrees, in total by 6 g m -2 degree -1 and 0.4 mm day -1 degree -1 , respectively, implying 175 that stratocumulus develop continuously along the trajectories (Figure 9b, c). The increasing TWP is mainly due 176 to the increase in IWP since LWP is almost constant along the trajectories (Figure 9b). The increase of ℎ is comparable to that of summer WaMAIs, while the increase in TWP is about half of that of summer WaMAIs (You et al., 2021), since less moisture is available for cloud development in winter ( Figure 4c).

179
The gradual increase of ℎ , a manifestation of increased boundary-layer mixing, leads to a reduction in

200
Fsh anomaly decreases fast by nearly 50% over the first two degrees from the sea ice edge (Figure 10c).

201
From 2 to 10 degrees north of the sea ice edge, the decrease is more moderate at a rate of 4 W m -2 degree -1 ( Figure   202 10c) but still faster than that over the completely frozen ocean sectors. However, Fsh anomaly even at ten degrees  with upstream open ocean (e.g. Barents Sea) form at the altitude of ~1 km, above the warm-and-moist air-masses.

232
The boundary-layer energy-budget here is categorized into three categories (RAD, TBL, TCD  Figure 15. for winter WaMAIs over completely frozen ocean sectors; see Figure 15a.

253
Clouds are relatively thin and radiative cooling near the cloud top is therefore weak (Figure 11f) and

254
only in a few cases the magnitude of radiative cooling is comparable to the turbulent cooling. Generally, in this 255 category, turbulent heating is larger than radiative heating as well as latent heating, and hence boundary-layer 256 warming is dominated by turbulence, but since turbulence only redistribute heat inside the PBL, as a whole it is 257 gradually cooled as the warm air progresses northward. Over the Barents Sea, the maximum air temperature (Figure 12a, 13a, 14a) and specific humidity (Figure 12d, 260 13d, 14d) over open ocean south of the ice edge are always located right above the sea surface as a result of the 261 strong air-sea interaction and are also typically larger than those over ocean sectors with land-locked sea ice. As 262 this air-mass, considerably affected by air-sea interaction, is advected over the sea ice, different stories take place.

263
Around 8% of all WaMAIs over the Barents Sea belong to category RAD (Table 2). In this category, the 264 total temperature tendencies are forced by radiative processes. For this category, the large-scale subsidence is an 265 order of magnitude smaller than that in category TBL (Table 3, CONV) and LWP is three times larger than that 266 in category TCD (Table 3, (Table 3), and increases almost linearly along the trajectory (Figure 16d1) due to the enhancing TWP ( Figure   281 16c1). ℎ and ℎ are generally smaller than those of category TBL since stronger mixing weakens vertical 282 gradients in the PBL and hence suppresses the surface turbulent heat flux (Table 3). The decreasing rates of ℎ 283 and ℎ from 0 to 2 degrees north of the sea ice edge are larger than for categories TBL and TCD as a result of  warm and cold air-mass by cooling (heating) warmer (colder) air (Figure 13h), simultaneously inducing a 13b). In this category, the well-mixed layer is substantially thinner than in category RAD, since the turbulent 298 mixing here is mainly forced by surface friction, weaker and less effective than the buoyant mixing in category 299 RAD (Figure 12b). Turbulence is mainly forced by wind shear and buoyancy, but buoyancy is negative here in 300 the initially very stable near-surface layer. Therefore, wind shear mostly fuels the turbulent mixing. In category 301 TBL, turbulent mixing is stronger than in category RAD, but the surface fluxes are still stronger, due to the 302 stronger gradients; ℎ and ℎ are 77% and 42% larger than those in category RAD. Also see Figure 15c.

304
40% of WaMAIs over the Barents Sea belong to this category. For this category, the boundary-layer energy-

305
budget is generally similar to that in category TBL. The main difference is that there is always a layer of cold air 306 (cold dome) laying below the warm-and-moist air-mass especially in the central Arctic (Figure 14c). This cold 307 dome enlarges the vertical temperature gradient and hence intensifies turbulent heat near the surface (Figure 14h).

308
As the warm-and-moist air-mass is advected over the cold dome, it is gradually lifted up by the cold dome and 309 consequently, ht and hsp are increasing at a faster rate than in category TBL (Figure 16b3). With faster lifting ht 310 and hsp, ℎ and ℎ would be reduced more rapidly or even become negative in the high Arctic (Figure 16a3).

311
Unlike in category RAD, where TWP is dominated by LWP, and category TBL, where TWP is contributed almost 312 equally by LWP and IWP, in this category TWP is gradually more dominated by IWP; the IWP-to-TWP ratio 313 increases linearly from ~50% to ~100% ( Figure 16c3); also see Figure 15d. Arctic, supplying moisture for cloud formation, exerting a positive total energy-budget anomaly on the surface.

321
Statistically, as warm-and-moist air is advected over ocean sectors with land-locked ice cover, such as 322 the Laptev, East Siberian, Chukchi and Beaufort Seas, the longwave irradiance anomaly increases linearly by 2.5 323 W m -2 degree -1 , while the total column cloud liquid water increases linearly by 6 g m -2 degree -1 . We have also 324 analysed the boundary-layer vertical structure along these trajectories, as well as the associated surface energy-325 budget pattern of over these sectors, and find one main category, elevated lifting temperature inversion (INV),
radiation-dominated (category RAD), turbulence-dominated (category TBL) and turbulent-dominated with cold 329 dome (category TCD), comprising 8%, 52% and 40%, respectively, of all WaMAIs. Unlike over the sectors with 330 land-locked sea ice, air-masses over the ice-free Barents Sea is warmed by the sea surface (local process) before 331 advected over the sea ice (remote process), consequently resulting in more intensive surface warming.
( Figure 15b). However, this strong radiative cooling induces intensive buoyant mixing extending from the cloud 335 top till the surface, supresses the surface turbulent mixing and decreases the lifting rate of the height to the 336 maximum temperature (ht) and to the maximum specific humidity (hsp). Therefore, surface turbulent fluxes in 337 category RAD and the lifting rate of ht and hsp are apparently smaller than those in category TBL (Figure 15c).

338
With cold dome, less liquid cloud water could be formed and fewer or even negative turbulent fluxes could access 339 to the surface, in comparison with category TBL (Figure 15d). In category TCD, turbulent fluxes decrease faster 340 along the trajectory since warm-and-moist air is lifted to higher altitude above the cold dome (Figure 15d).

341
Under the background of global warming, the rate of local process has been accelerated by 9% per year (Kim et al., 2019), while the meridional heat and moisture transports (remote processes) over the Barents Sea are 343 also enhanced in recent decades (Nygå rd et al., 2020). This implies that WaMAI may play a more significant role 344 in the future Arctic warming. Therefore, the potential mechanism which enhances the occurrence and intensity of

345
WaMAI deserves more attentions from atmospheric scientists.

346
Data availability

347
All data used can be found on the ERA5 data repository at DOI: www.ecmwf.int/en/forecasts/datasets/reanalysis-348 datasets/era5.