Combined impact of summer NAO and northern Russian shortwave cloud radiative effect on Eurasian atmospheric circulation

Based on ERA-Interim and CERES_SYN1deg Ed4.1 datasets, the combined influence of summer North Atlantic Oscillation (SNAO) and positive shortwave cloud radiative effect (SWCRE) events in northern Russia on Eurasian atmospheric circulation is investigated at the intraseasonal scale. The impact of the SNAO on the position of the North Atlantic storm track is modified combined with the Ural anticyclone anomaly contributed by positive northern Russian SWCRE anomalies, which could affect the summer stationary wave pattern. During positive northern Russian SWCRE events under SNAO+, the upstream wave train enhanced by the southward Ural anticyclone anomaly is easily trapped by the northward South Asian jets, thus propagating to low latitudes and causing extreme heat events in East Asia. Under SNAO-, the wave train propagates in the British–Baikal Corridor pattern along polar front jet towards the Far East, slowing down the dramatic melting of sea ice in the Laptev and East Siberian seas. Summer positive SWCRE events in northern Russian act as a bridge by promoting the emergence of the Ural anticyclone anomaly, influencing extreme weather in East Asia and Arctic sea ice variability.


Introduction
Tiny modifications in the cloud radiative effect (CRE) field act as an influential climate feedback mechanism. In general, the longwave CRE (LWCRE) is completely offset by the comparably high and negative shortwave CRE (SWCRE). Thus, clouds as a whole provide a net cooling effect on Earth (Zelinka et al 2017, Järvinen et al 2018, Guo et al 2019. Since solar heating reaches its peak in summer, a wealth of research has focused on exploring the summer SWCRE. Numerous studies highlight that summer cloud cover, in particular its shortwave radiation effect, appears likely to be an important local factor contributing to heatwaves and droughts in Eurasia (Dai et al 1999, Myers et al 2018, Tang et al 2020. Zhang et al (2020b) suggested that summer SWCRE in East Asia dominates the LWCRE at the surface and top of the atmosphere. Moreover, the monsoon march heavily affects the seasonal evolution of SWCRE in the Asian monsoon region (Zhao et al 2007, Guo and Zhou 2015, Li et al 2019, Zhang et al 2020a. Huang et al (2020) indicated that the rapidly varying atmospheric circulation in the Asian monsoon region around the onset of the South China Sea summer monsoon can significantly modulate regional SWCRE. Instead, the spatial and temporal evolution of the SWCRE is a response to the modification of the monsoon circulation.
In addition to its local impact, a recent study has shown that SWCRE also exerts remote impact. Summer SWCRE exhibits distinct regional characteristics and the greatest variability in SWCRE occurs over northern Russia (not shown). Liu et al (2022) found that summer positive northern Russian SWCRE anomalies facilitate the evolution of Ural blocking. The decreased cooling effect of cloud radiative forcing results in abnormal warming. The consequent decrease in the meridional temperature gradient causes deceleration of prevailing tropospheric westerly winds, thereby diminishing baroclinicity in the mid and low troposphere. This might dynamically favor the formation of the Ural blocking in summer. It would enhance the wave train and make it propagate to low latitudes causing persistent heatwaves in East Asia. Recently, extreme heat events have occurred frequently, caused considerable damage to human life and socio-economics (Luo and Zhang 2012, Sun et al 2014, Larcom et al 2019, Wu and Francis 2019, McElroy et al 2020, Hong et al 2021. That indicated that extreme events should be explored at the intraseasonal scale to capture the real heatwave signals. In addition, the fact that the wave train originates from the North Atlantic suggests that North Atlantic acts as a vital factor in the remote impact of the SWCRE (Liu et al 2022). The summer North Atlantic Oscillation (SNAO), as the dominant mode of summer North Atlantic atmospheric circulation variability, profoundly influences weather and climate of Asia. Hong et al (2022) proposed that the shapes for SNAO southern pole can partly determine whether the SNAO is related to the Silk Road pattern (SRP) and affects East Asian climate. But previous studies on the relationship between SNAO and East Asian climate have mainly focused on precipitation and summer monsoon (Liu and Yin 2001, Li et al 2008, Wang et al 2018, Du et al 2020, Liu et al 2021, and a few studies have pointed out that SNAO is associated with temperature anomalies on a small regional scale in middle East Asia (Sun et al 2008), more often with high temperatures in Europe (Cassou et al 2005, Drouard et al 2019. And the connection between SNAO and extreme heat waves in East Asia has not been explored. Studying the collaborative impact of SNAO and positive northern Russian SWCRE anomalies on Eurasian atmospheric circulation at the intraseasonal scale provides a deeper understanding of the impact of mid and high latitude variability on extreme weather events in East Asia.

Data and methods
Daily cloud radiative fluxes used were from the CERES of NASA. And the latest CERES_SYN1deg Ed4.1 dataset is available from March 2000 with a spatial resolution of 1 • × 1 • . We employed the net shortwave radiative flux difference between all-sky and clear-sky as SWCRE that is downward positive (Ramanathan et al 1989). Daily atmospheric variables and sea ice concentration (SIC) were accessed by ERA-Interim dataset compiled with a spatial resolution of 2.5 • × 2.5 • . Daily SNAO index is obtained from NOAA/CPC. The study period is from 2000 to 2018.
The first two modes of summer (June-August) daily SWCRE over 40 • -100 • E north of 50 • N are extracted using the empirical orthogonal function method and account for 8.8% and 7.7% of the total variance, respectively. They are well separated from each other based on North et al (1982). And the statistical significance estimates are based on the twotailed student's t-test. The leading mode (figure S1) is dominated by a large region of significant negative SWCRE anomalies except for a small portion of positive anomalies in southwestern Russia. Because of its weak connection with Eurasian atmospheric circulation, we mainly focus on the second mode. The spatial distribution of the second mode (figure 1(a)) is featured by a north-south dipole. Significant positive SWCRE anomalies in northern Russia both in extent and magnitude are larger than negative anomalies to its north, implying a general reduction in low cloud cover and relatively high surface insolation, which together result in abnormal warming. A positive SWCRE event over northern Russia is identified when the time series (PC2) of the second mode exceeds 1 standard deviation and persists for at least three consecutive days. In such an event, the day with the maximum value of PC2 is taken as its peak day. The lifetime of a positive SWCRE event is defined by the number of consecutive days with the PC2 greater than 1 standard deviation. And the interval between events should be longer than 15 d. Totally, 32 positive SWCRE events were recognized during the study period ( figure 1(b)).
The horizontal component of the wave activity flux (WAF) derived by Nakamura (1997, 2001) is employed. The formula is shown as follows: where p, ψ and U = (U, V) denote the pressure scaled by 1000 hPa, the stream function and the background flow, respectively. The primes refer to the deviations from the climatological mean and the subscripts represent the partial derivatives. Eddy kinetic energy (EKE) is calculated by the formula EKE = , where u ′ and v ′ indicate synoptic components of horizontal winds obtained through a 2-8 d Lanczos bandpass filter (Duchon 1979).

Collaborative effects of summer positive SWCRE events in northern Russia and SNAO in different phases
To investigate the evolution of summer positive SWCRE events in northern Russia and their impact on Eurasian atmospheric circulation, a composite analysis of 32 cases (figure 1(b)) is carried out (figure 1(c)). At lag −8 d, SNAO+ pattern occurs (c) Composite of summer daily 500 hPa geopotential height anomalies (contour, gpm) and surface air temperature (SAT) anomalies (shading, K), and WAF (vectors, m 2 s −2 ) at 300 hPa for positive SWCRE events (32 cases) list in (b). Lag 0 denotes the peak day of the positive SWCRE event, and lag −n (n) denotes n days before (after) the peak day. Stippling areas represent anomalies at the 95% confidence level. Vectors larger than 3 m 2 s −2 are plotted in (c). in North Atlantic, and Ural anticyclone anomaly is not present. At lag −6 d, anomalous Ural anticyclone and positive surface air temperature (SAT) anomalies over northern Russia appear, while a branch of wave train is generated in the southern center of SNAO+ towards Europe. The composite field of summer daily SWCRE anomalies (figure S2) shows that significant positive SWCRE anomalies are present in northern Russia at lag −8 d. It confirms the conclusion Liu et al (2022) drew that positive SWCRE anomalies there are approximately two days ahead to promote the anomalous Ural anticyclone formation. From lag −4 to −2 d, as Ural anticyclone anomaly expands, the wave train inspired by subtropical North Atlantic travels through western Europe to northern Russia. From lag −2 d onwards, the wave train is enhanced considerably over northern Russia and begins to split into two north-south branches. During lag 0-4 d, the paths of the two become clear. The northern one travels to the Far East along high latitudes. The southern one travels to East Asia via Lake Balkhash along low latitudes. Significant anomalous anticyclone (cyclone) and positive (negative) SAT anomalies appear in East Asia (North Asia). At lag 6 d, the southern wave train disappears.
Considering the role of SNAO, we divide positive SWCRE events into two categories for composite analysis: under SNAO+ (when SNAO index is positive throughout the lifetime of a positive SWCRE event) and SNAO−, with 12 and 19 cases, respectively (figures S3(a) and (b)). Under SNAO+ (figure 2(a)), the anomalous cyclone appears in Greenland from lag −8 to −6 d, and anomalous anticyclones are observed in North America and central North Atlantic. The wave train passes from North America to Greenland and then splits into two branches. One branch reaches central North Atlantic and then spreads to western Europe. The other reaches northern Russia and immediately travels southwards to the central Eurasian continent. During lag −4 to −2 d, the wave train strengthened over northern Russia mainly travels southward through Lake Balkhash. At lag 0 d, the wave train travels to East Asia, where positive 500 hPa height and SAT anomalies occur. Simultaneously, a small amount of the planetary waves zonally propagate from northern Russia to the north of Lake Baikal, causing negative SAT anomalies.
Under SNAO− (figure 2(b)), from lag −8 to −4 d, anomalous anticyclones are observed surrounding Greenland and Mediterranean, with anomalous cyclones in central North America, central North Atlantic, and western Europe. From lag −2 d onwards, the Ural anticyclone anomaly expands into the surrounding area and the SNAO− southern pole intensifies further and moves into western Europe. After traveling from western Europe to northern Russia, the wave train spreads directly eastwards with no tendency southwards. At lag 0 d, the anomalous cyclones in western Europe, central North America, and central North Atlantic are contiguous. At this time, both the Ural anticyclone anomaly and the eastward propagating wave train are at their strongest. Significant cyclonic and negative SAT anomalies occur in the Far East and the Laptev sea. The overall -+-spatial distribution is thus in the British-Baikal Corridor (BBC) pattern (Xu et al 2019).
Obviously, under SNAO− ( figure 2(b)), insignificant anomalies are observed over East Asia. We calculate the frequency of extreme heat events in East Asia (figure S3(c)). Notably, there are five extreme heat events that peak days lie within the lifetime of positive SWCRE events under SNAO+ while none of extreme heat events occurs under SNAO−. We further performed a lead-lag analysis of summer daily East Asian SAT index (see text S2 for definition) and SWCRE index (PC2) for positive SWCRE events under both SNAO phases (figure S5). Under SNAO+ ( figure S5(a)), there is a tight correlation between SWCRE and SAT index from lag −6, and the correlation coefficient reaches its maximum at lag −2. Whereas under SNAO− ( figure S5(b)), the correlation is insignificant.
Under SNAO−, the wave train propagates towards the Far East and its north. It is necessary to check out Arctic SIC variation. Under SNAO+ ( figure  S4(a)), the significant anomalous cyclone around Greenland brings warm air from North Atlantic to the Barents and Kara seas (BKS), promoting sea ice melting in BKS. It also carries the unmelted ice to the Beaufort sea, causing increased sea ice there. Under SNAO− ( figure S4(b)), anomalous anticyclones in Greenland and northern Russia go along with an anomalous cyclone near the Laptev and East Siberian seas. The cyclonically induced cooling leads to local positive SIC anomalies, with the strongest value at lag 0 d. Arctic sea ice extent (SIE) is sensitive to sea ice variability in this region (Ogi andWallace 2007, Kapsch et al 2013). We carried out a lead-lag analysis of summer daily Arctic SIE (see text S2) and SWCRE index (PC2) under different SNAO phases ( figure S5). Under SNAO+ ( figure S5(a)), SWCRE index and SIE do not correlate well. However, under SNAO− ( figure S5(b)), a significant positive correlation appears between SWCRE index and SIE from lag −8 and their correlation coefficient reaches its maximum (nearly 0.8) at lag −7. It indicates that SNAO− and summer positive SWCRE anomalies in northern Russia can jointly affect Arctic SIC on the Pacific side, and have a mitigating effect on the decreased Arctic SIE (Wang and Overland 2009, Overland and Wang 2013, Francis and Wu 2020.

Discussion of possible mechanisms
In this part, possible physical mechanisms are explored. SNAO+ (−) is accompanied by a northward (southward) shift of the North Atlantic storm track (NAST), which is related to the summer European climate (Dong et al 2013). Under SNAO+ (figure 3(a)), from lag −8 to −6 d, EKE anomalies are spatially located north and south to the core of the climatological mean of EKE, respectively. Significant negative EKE anomalies occur in the south to the core and extend towards western Europe. At lag −4 d, positive EKE anomalies near Iceland become significant. It implies a northward shift of NAST. From lag −2 d onwards, with the strengthening of Ural anticyclone anomaly intensity and extent, widespread negative EKE anomalies appear in northern Russia, even northern Eurasia. At lag 4 d, positive EKE anomalies near Iceland continuously extend to the northeast, reaching as far east as 90 • E, which is more easterly than the extension under normal SNAO+ (Linderholm et al 2011). At the same time, negative EKE anomalies in central North Atlantic extend towards North America. Thus, two zonal belts appear in the high and mid latitudes of North Atlantic and Eurasia, respectively. Up to lag 6 d, the EKE anomalies in North Atlantic weaken.
Under SNAO− ( figure 3(b)), from lag −8 to −6 d, significant negative EKE anomalies appear in North America and southwest Greenland. Significant positive EKE anomalies appear in a zonal belt from east North Atlantic to central Europe. Such an east-west distribution implies an eastward shift of NAST, unlike the southward shift under normal SNAO−. From lag −4 d onwards, positive EKE anomalies in western Europe begin to shrink, while negative EKE anomalies in North America and southwest Greenland deepen and intensify eastward, reaching as far east as 30 • E. This is about 30 • E of the position under normal SNAO− (Linderholm et al 2011). And the spatial Figure 2. Composite of summer daily 500 hPa geopotential height anomalies (contour, gpm) and SAT anomalies (shading, K), and WAF (vectors, m 2 s −2 ) at 300 hPa for positive SWCRE events under (a) SNAO+ (12 cases) and (b) SNAO− (19 cases). Lag 0 denotes the peak day of the positive SWCRE event, and lag −n (n) denotes n days before (after) the peak day. Stippling regions denote anomalies at the 95% confidence level. Vectors larger than 4 m 2 s −2 are plotted.
distribution is similar to the first mode of summer NAST (Wang et al 2020).
The NAST is the region within which cyclones preferentially generate and propagate due to baroclinic instability (Shaw et al 2016). Summer synopticeddy activity related to NAST transition is crucial in shaping the stationary wave regime (Fukutomi  A wealth of research has been conducted on the possible physical mechanisms or dynamical reasons affecting the propagation path of Rossby waves (Chiang et al 2015, Li et al 2015, Zhao et al 2015, Son et al 2019, 2021. In the case of different position of Ural anticyclone anomaly, we consider the explanation for the wave train propagation along different paths from the perspective of jets. The . Composite of summer daily 300 hPa zonal wind (shading, m s −1 ) and absolute vorticity meridional gradient anomalies (contour, 10 −12 m −1 s −1 ) for positive SWCRE events under (a) SNAO+ (12 cases) and (b) SNAO-(19 cases). Lag 0 denotes the peak day of the positive SWCRE event, and lag −n (n) denotes n days before (after) the peak day. Stippling areas denote anomalies at the 95% confidence level. The green contours denote the climatological mean of summer 300 hPa zonal wind, with contour intervals of 5 m s −1 .
implications of the jet on planetary wave propagation and the consequent atmospheric anomalies rely on the meridional path of Rossby waves compared to the geographic location of the jet (Graf and Zanchettin 2012). Positive zonal wind anomalies around north-ern Russia occur in the northern part and negative anomalies in its southern part, as the existence of Ural anticyclone anomaly. Here, we mainly focus on the South Asian jets (SAJ). Under SNAO+ ( figure 4(a)), the SAJ around the Caspian sea shifts northwards during lag −8-lag -6 d. This makes it easier for the wave train to be trapped. Therefore, at lag −8 d, the train wave tends to travel southward ( figure 2(a)). From lag −4 d onwards, the jets from the Caspian sea to Lake Balkhash undergo a more significant northward shift. The zonal wind maximum in the jet core generates a strong meridional gradient of absolute vorticity along the jet (Shaman and Tziperman 2007). According to Rossby wave theory, the maximum in the meridional gradient of absolute vorticity produces barotropic Rossby wave refraction or bending, toward the jet core (Hoskins and Ambrizzi 1993). In addition, the jet is flanked by minima in the meridional gradient of absolute vorticity. These minima act as refractive barriers to most northward and southward Rossby wave propagation. As a result, the wave train is mainly trapped by the jets there and then travels southwards to East Asia ( figure 2(a)). The SAJ is activated as a zonal waveguide for the wave train . Notably, from lag 0 d onwards, the jets near Lake Baikal start to move northwards significantly so that a small fraction of the planetary waves traveling to northern Lake Baikal is also trapped by the jets for southward spread ( figure 2(a)). In addition, from lag 0 to 6 d, the zonal wind in northern China strengthens and weakens significantly in the south, which favors the formation of the anomalous anticyclone and even the extreme heat. Under SNAO− ( figure 4(b)), no significant anomalies and northward shift of the SAJ occurred, which is not conducive to the Rossby wave refraction or bending. It combined with the northerly disturbance makes the wave train unlikely to be trapped by the jet, so the wave train continues to travel eastward along the pattern along polar front jet (PFJ) in the BBC pattern. It also suggests the BBC is not only associated with the preceding winter NAO (Xu et al 2019) but can also be linked to SNAO via the northern Russian SWCRE.

Summary and discussion
We examine the collaborative impact of SNAO in different phases and summer positive SWCRE events in northern Russia on Eurasian atmospheric circulation at the intraseasonal scale. Under SNAO+, the wave train that originates in North America and passes through Greenland to northern Russia is enhanced by the southerly Ural anticyclone anomaly. Then, trapped by the stronger SAJ shifting northwards, it undergoes a shift from northern Caspian sea towards the low latitudes and propagates along the SAJ to East Asia via Lake Balkhash, causing extreme heat. Under SNAO−, the wave train from western Europe to northern Russia is strengthened by the northerly Ural anticyclone anomaly and then travels in the BBC pattern along the PFJ to the Far East, generating significant cyclonic anomalies in the Laptev and East Siberian seas, favoring local sea ice increase. A positive SWCRE event combined with SNAO− is beneficial in mitigating the dramatical reduced SIE. This process is conceptualized schematically in figure 5.
Besides the combined impact of SNAO and northern Russian SWCRE, we also examine their respective roles. We split the SNAO events into SNAO+ (figure S6) and SNAO− events (figure S7) related and unrelated to positive northern Russian SWCRE events, respectively (see text S3 for definition). The composite results of SNAO+ (figure S6(a))/SNAO− events (figure S7(a)) related to positive SWCRE events are generally consistent with figures 2(a) and (b). However, the composite results of SNAO+ (figure S6(b))/SNAO− events (figure S7(b)) unrelated to positive SWCRE events show no significant anomalies over East Asia. We further performed statistical calculations. East Asian extreme heat events (four cases) accounted for 80% of the SNAO+ events (five cases) related to positive SWCRE events. In contrast, only 12.5% of the SNAO+ events (16 cases) unrelated to positive SWCRE events consist of extreme heat events in East Asia (two cases). It suggests that the link between SNAO and East Asian extreme weather requires the amplification and driving effect of the summer northern Russian SWCRE. In turn, the association between the summer northern Russian SWCRE and Eurasian atmospheric circulation is closely related to the SNAO.
In this paper, we point out the tight connection between SNAO and large areas of extreme heat waves in East Asia for the first time. And no one has ever explored the connection between northern Russian SWCRE and East Asia weather specifically at the intraseasonal scale. These are where our work differs from previous studies. In brief, northern Russian SWCRE in summer is a bridge to influence extreme heat in East Asia as well as Arctic SIE variability by promoting the emergence of the Ural anticyclone anomaly. The relationship will be examined in the future via ensemble runs of atmospheric and coupled models, and the source of northern Russian SWCRE variation will be studied further.