Synergistic impacts of wintertime regional snow anomalies in the Northern Hemisphere on the summer rainfall pattern in China

Changes in winter snow cover in the Northern Hemisphere (NH) could have a profound impact on mid-latitude weather. Previous studies have focused on the role of regional, e.g. Eurasian or Tibetan, snow cover in summer precipitation anomaly, without considering the synergistic impacts of hemispheric wintertime snow. In this study, we find that the dominant pattern of the NH winter overall snow cover anomaly with a synergistic impact, has a stronger cross-seasonal association with the China’s summer rainfall pattern than regional snow cover anomaly. We summarize three synergistic impact paths of regional snow cover. One is extratropical path, that is the westerlies are affected by less snow in Europe through the snow-soil moisture-atmospheric feedback, and the influence is strengthened by less snow in Mongolia through enhanced temperature anomalies. The second is subtropical path, that is the meridional thermal difference anomaly caused by more snow anomaly on the Tibetan Plateau is strengthened by less Mongolian snow and then impacts the behavior of the upper-tropospheric westerly jet. Third, concurrently, more North American snow enhances the above two synergistic influence paths via the Circumglobal Teleconnection pattern. These three paths can be simultaneously reflected in the associated circulations of the first mode of NH snow cover. Their synergistic impacts eventually influence the meridional East Asia-Pacific pattern circulation anomalies in summer, leading to increased precipitation in the Yangtze River Basin. The cross-seasonal influences of synergistic effects of multiple regional snow anomalies can be identified by CMIP6 multi-model ensembles, particularly the impact of European snow cover.


Introduction
Snow cover significantly impacts atmospheric circulation, exhibiting more seasonal variability than sea surface temperature (Arduini et al 2019).Research has emphasized its role in regional climate change, especially in winter-to-summer effects (Coumou et al 2018).Snow cover's impact on global climate and hydrology draws attention in studies, with a focus on some regions like Eurasia, the Tibetan Plateau (TP), and North America (NA) (Henderson et al 2018).
Eurasian snow cover significantly affects summer precipitation in China (Wu et al 2009).This cross-season correlation is independent of the El Niño-Southern Oscillation (ENSO) and has remained stable over the past 40 years (Wu et al 2009).Snowmelt, with delayed hydrological effects, enhances soil moisture and heat absorption, creating a 'land bridge' that affects summer atmospheric circulation (Koster et al 2011).This leads to persistent soil moisture anomalies, triggering Rossby waves in Eurasia (Zhang et al 2020), which influence East Asian atmospheric circulation during summer (Zuo et al 2015, Shen et al 2020).
Snow cover on the TP may influence summer rainfall in China.Research confirms that increased TP snow cover reduces surface heat flux, leading to lower tropospheric temperatures nearby (Liu et al 2020).This reduction in temperature weakens the East Asian summer monsoon intensity (Ding et al 2008).The correlation is dependent on the radiative cooling of snow and 'wet soil' acting as cold sources (You et al 2020).
Recent studies have emphasized the significance of snow cover in northern China and Mongolia (Shen et al 2020).Persistent surface cooling in the region extends into summer owing to the memory effect of snow and its interaction with the monsoon.This results in unusual shifts in summertime temperature gradients and upper-level westerly airflow (Koster et al 2010, Chiang et al 2017).Enhanced westerly winds over the northern TP combined with plateau topography induced convergence over North China and compensatory divergence over the southern basin, causing anticyclonic circulation and reduced rainfall (Xiao et al 2016).
Furthermore, increased NA snow-cover induces atmospheric cooling through heightened albedo and reduced upward sensible heat flux (Henderson et al 2018).This sustained cooling prompts low-level westerly anomalies in the tropical central-eastern Pacific by adjusting the geopotential and temperature gradients.These anomalies promote convergence and upward motion in the tropical eastern Pacific (Cao et al 2018, Wang et al 2020).Then, affecting China's summer circulation via teleconnection patterns, such as Circumglobal Teleconnection (CGT), East Asia-Pacific, and East Asia-Pacific-NA atmospheric teleconnection (Vavrus et al 2017, Zhou et al 2020).
Previous researchers have identified the influence of regional snow cover on summer circulation and China's precipitation.However, the synergistic impacts of Northern Hemisphere (NH) snow anomalies remain unclear.To address this shortcoming, this study focuses on overall NH snow cover and explores the relationship(s) between winter snow and summer precipitation over China.Specifically, the synergistic impacts of multiple snow cover on atmospheric circulation and the process of cross-seasonal transmission of snow signals.

Calculation of wave activity flux (WAF)
The energy propagation characteristics of quasi-stationary waves were described by the three-dimensional WAF defined by Takaya and Nakamura (2001).In spherical coordinates, the expression is as follows: where U = (U, V) is the basic flow field, |U| is the climatological magnitude of the winds, α is the mean radius of Earth, and λ, φ , p represent the longitude, latitude, height, and pressure scaled by 1000 hPa, respectively.ψ ′ denotes the perturbed geostrophic stream function.

Association between NH wintertime snow cover and Chinese summer precipitation
Empirical orthogonal function (EOF) analysis is applied to examine the NH winter snow cover and The correlation coefficient between the two PC1s was 0.40 (p < 0.01; figure 1(c)).The two largest PC1 values occurred in 1998 and 2020, coinciding with major flooding in the YRB and corresponding to the two large PC1 values for winter snow cover.After removing the influence of ENSO, the correlation coefficient between the two is 0.32 (p < 0.05).The map of the correlation coefficient between China's summer precipitation and winter NH snow cover PC1 (figure 1(e)) also shows a pattern similar to that of the summer precipitation EOF1 (figure 1(b)), with a spatial correlation coefficient of 0.70.
To further evaluate the impact of statistical synergies, we selected four regions (red boxes in . This comparison of hemispheric and regional results can help elaborate on the common and different influences of snow cover in different regions on the summer circulation in China.Moreover, such comparisons can help us identify which effect is more important for summer precipitation patterns: the synergistic effect (i.e. the main mode) of NH snow cover or the regional snow effect.The supplementary material contains a comprehensive discussion and analysis of the results (figure S1, tables S2 and S3), including correlations, removal of influences, and the correlation fields.Overall, we confirm that synergies exert a greater influence than single regions and thus are more likely to cause summertime flooding in the YRB.This underscores the importance of evaluating hemispheric snow cover in relation to regional precipitation.

Dynamic analysis 3.2.1. Extratropical path of snow synergistic impacts
After removing European (Mongolian) snow cover, the correlation coefficient between NH winter snow cover and Chinese summer precipitation PC1 decreased to 0.13 (0.22) (table S3).Therefore, statistically speaking, European snow cover appears to be more significant.In Europe, following winters with relatively less snow cover, springtime soil moisture was generally low (figures 2(a)-(c)).Due to the persistence of soil moisture anomalies from winter to summer (highlighted by red boxes in figures 2(d)-(f)), the latent heat flux during European summer decreases.Simultaneously, both the sensible heat flux and the net surface-upward energy flux increase (figure S2).This leads to elevated surface air temperatures and the formation of low-pressure anomalies (figures 2(g)-(i)).Moreover, a low-pressure circulation, induced by thermal forces over Europe and featuring easterly anomalies to the north, disrupts the typical trajectory of westerly airflows (e.g.Haarsma   As depicted in figures 2(d)-(f), low soil moisture at around 45 • N tends to persist in summer, reflecting the positive feedback between winter snow cover, springtime soil moisture, and summer atmospheric circulation.This phenomenon is referred to as the snow-soil moisture-atmospheric circulation feedback.It involves a process in which diminished spring snow cover postpones the drying effect on soils until mid-summer.This delay results in higher temperatures due to suppressed evaporative cooling, impacting both regional and hemispheric atmospheric circulation (Coumou et al 2018).
Figure 3(a) shows that European winter snow cover influences summertime Rossby-wave activity.This activity leads to downstream energy flow from Europe, splitting into two branches over Central Asia.One branch continues eastward towards eastern China, as depicted in figures 3(a) and 4(a), which can strengthen the southern YRB high-pressure system.However, it is worth noting that the low-pressure anomalies in the northern part of YRB are not very pronounced.On the other hand, the other branch propagates northeastward and strengthens the highpressure ridge over northern East Asia.Moreover, Mongolian winter snow cover anomalies not only reinforce the high-pressure anomaly over northern East Asia but also enhance the low-pressure anomaly over the northern YRB.Therefore, it is the synergy between dominant European snow cover and the enhancing effect of Mongolian snow cover.The synergy leads to the formation of a meridional '+ _ +' wave train structure.This structure strengthens the East Asia and Pacific wave train, a dominant feature of the East Asian Summer monsoon period (Nitta 1987, Lau et al 2000).Additionally, it enhances the northsouth propagation of wave energy, making the East Asian precipitation pattern more sensitive to meridional changes.Similar circulation and planetary wave characteristics are observed in the correlation fields of NH snow-cover PC1 and summer precipitation PC1 (figures 3(c) and (d)).These characteristics align with conditions conducive to summertime flooding events in the YRB (Ding et al 2008, Wu et al 2009).

Subtropical path of snow synergistic impacts
Based on the characteristics of Rossby-wave activity, we constructed vertical profiles for the four areas depicted in figure 4 to investigate the synergistic impacts of snow cover in Europe, Mongolia, and TP.An anomalous upward and eastward propagation of WAF occurs from Europe and TP to the YRB, strengthening the cut-off low-pressure system over the YRB (figures 4(a), (b) and 5(a)).This phenomenon can also be induced by the accumulation of snow in Mongolia (figure 3(b)).However, solely relying on winter snowfall from the TP is insufficient to significantly trigger the East Asia and Pacific pattern (figure 5(a)).Figure 4(c) shows the anomalous eastward propagation of WAF that starts in Mongolia.The propagation of these WAFs collaboratively reinforces the high-pressure ridge in Northeast Asia (figure 5(b)).The correlation figures of NH winter snow-cover PC1 (figures 3(c) and 4(d)) reveal a synergistic interaction with Chinese summer precipitation PC1 (figures 3(d) and 4(e)).
When the TP experiences relatively more winter snow cover, it hinders warming in the subsequent spring and summer seasons (Liu et al 2020, Sun et al 2021).This leads to the development of an anomalous meridional temperature gradient over the region.Consequently, a strong westerly airflow occurs over the TP, accompanied by anticyclonic circulation to the south and cyclonic circulation to the north.Additionally, a reduction in winter snow cover in Europe leads to a thermal gradient between the northern TP and Mongolia.This reinforces the subtropical westerly jet and causing it to shift southward (figures 5(a) and (b)).This configuration promotes upward motion, leading to rainfall over the middle and lower reaches of the YRB (Yang et al 2002, Zhao et al 2007).
The positive winter snow-cover anomaly in the TP strengthens the meridional temperature gradient and zonal wind anomaly (figure S3).Southerly winds transport a significant amount of water vapor from the Bay of Bengal and South China Sea, leading to increased summer precipitation in the middle and lower reaches of the YRB.

The teleconnection effect of snow in North America
Owing to snow-hydrological effects, a significant wetness signal develops at 40 • N latitude in NA during spring and summer (figure 6, upper panels).Additionally, a cold temperature signal at 2 m persists into the summer (figure 6, lower panels).Crucially, we note that (1) the significant soil moisture signal emerges before the cold temperature signal, suggesting that soil moisture at about 40 • N may be a potential driver, and (2) the late-winter soil moisture signal shows seasonal transmissions.This relationship drives changes in the meridional thermal gradient and fosters easterly (westerly) anomalies north (south) of 40 • N, resulting in an anomalous summertime cyclonic circulation in the lower troposphere over NA (figure 7(a)).According to figure 7(c), anomalous cyclonic eddies near 45 • N strengthen the westerly jet.They propel the eastward propagation of the North Atlantic jet and enhance its connection to the Asian-African jet.A cyclonic eddy forms downstream and excites Rossby waves that reach the Caspian Sea, Mongolia, and China.Similar to the CGT, this wave-wave interaction not only strengthens the cyclonic anomaly over northern China but also leads to the intensification of the high-altitude westerly jet over the YRB.This shift southward of the jet's position could lead to increased precipitation in the YRB (Wu et al 2016).

CMIP6 simulations
We used 20 CMIP6 models to verify the reanalysis described above.First, we compared reanalysis data with the simulated model data and found that ( 1 S6, r = 0.42).
The figures S6 and S7 illustrate the simulation of cross-seasonal relationships between snow cover and soil moisture, as well as the reproduction of circulation patterns in different regions.While the influence of snow cover on atmospheric circulation needs further refinement, CMIP6 models generally capture the combined impacts of various regional snow anomalies.

Discussion and conclusions
Extreme precipitation is increasing in China owing to climate change (IPCC 2021).In this study, snow cover in various NH regions has three synergistic impact paths.These paths serve as a more informative indicator of summer precipitation compared to the influence of snow cover within a single locality.
Firstly, the extratropical path involves westerlies affected by reduced European snow, driven by the snow-soil moisture-atmospheric circulation feedback mechanism (Coumou et al 2018, Henderson et al 2018, Yao et al 2022), further intensified by decreased Mongolian snow and enhanced temperature anomalies.Secondly, the subtropical path sees a meridional thermal difference anomaly.This anomaly is a result of increased snow on the TP and strengthened by reduced snow in Mongolia.The anomaly, in turn, influences the upper-tropospheric westerly jet.Thirdly, increased NA snow reinforced previous paths via CGT patterns.These paths are reflected in the EOF1 of winter NH snow cover and related circulations.This leads to meridional East Asia-Pacific pattern circulation anomalies in summer and results in heightened YRB precipitation.Additionally, snow depth in topographic depressions and treelines has a lasting presence until late spring and occasionally into summer (Mackiewicz 2012).This extended presence plays a more crucial role in influencing soil moisture and the snow-soil moisture-atmospheric circulation feedback mechanisms, especially in late summer, compared to shallow snow cover during winter.The significance of the extended snow depth is attributed to its close association with the seasonal dynamics of the regional climatic system (Lievens et al 2019, Yang et al 2023).Furthermore, while the MME of CMIP6 models capture the cross-seasonal relationship and role of European snow, they tend to overestimate NA's influence on NA and underestimate that of Mongolia and the TP.
While explaining ⩽20% variance, EOF1 of winter snow cover shows mid-to-high latitude patterns, which is important for understanding synergistic impacts and identifying extreme events.The temperature difference between the sea and land caused by the snow is also a significant factor contributing to the excess summer precipitation in the YRB (Ding et al 2008).Moreover, future research should focus on snow depth or snow-atmosphere coupling for better seasonal predictions.Understanding the sensitivity of extratropical variability to climate change, especially the snow-(N)AO/NAM linkage, is crucial for improving predictions on weekly to monthly timescales, and is sensitive to anthropogenic effects.

Figure 2 .
Figure 2. (Left panels) Correlations between the standardized time series of winter snow cover over Europe (multiplied by -1; with effects from distant regions eliminated; same as below) and (a) spring soil moisture, (d) monthly mean soil moisture averaged over 0 • -40 • E (red box in figure (a)), and (g) summertime 2 m temperature (T2m; shading) and 925 hPa geopotential height (P925; contours).Middle and right-hand panels are similar to the left-hand panels, but for the correlations with the NH wintertime snow PC1 (middle) and summer precipitation PC1 (right), respectively.

Figure 4 .
Figure 4. Vertical profiles of the correlation field between geopotential height (shading) and WAF (vectors) with the standardized time series of winter snow in (a) Europe and (b) the TP, (d) the NH winter snow-cover PC1, and (e) the Chinese summer precipitation PC1 at points A (30 • E, 45 • N), B (80 • E, 30 • N), and C (120 • E, 30 • N) depicted in figure 3. Panel (c) is the correlation field with the standardized time series of Mongolian winter snow cover at points B (80 • E, 30 • N) and D (120 • E, 60 • N).The green line at 105 • -125 • E represents the YRB.

Figure 5 .
Figure 5. Vertical cross-sections of the correlation field, averaged for 80 • -100 • E, between summer temperature (shading) and the u-component of wind (contours), composite of vertical velocity and meridional wind (vectors), the standardized time series of winter snow cover (a) on the TP and (b) in Mongolia, (c) the NH winter snow-cover PC1, and (d) the Chinese summer precipitation PC1.The solid yellow line denotes the climatological summer 30 m s −1 zonal wind, which corresponds to the subtropical westerly jet.

Figure 6 .
Figure 6.(Upper panels) Correlation coefficients between monthly averaged soil moisture at 70 • -120 • W and (a) the NA standardized winter snow time series, (b) the NH winter snow-cover PC1, and (c) the Chinese summer precipitation PC1.(Lower panels) (d)-(f) As for the upper panels (a)-(c) but for the correlation coefficient with T2m.

Figure 7 .
Figure 7.The composite field of (a) the summertime detrended 500 hPa geopotential height field (shading, units: gpm) and WAF (arrows, units: m 2 s −2 ) for the high-value years of (a) the NH winter snow-cover PC1, (b) the NH winter snow-cover PC1 after the removal of NA, and (c) the difference between (a) and (b) (dif.).The green and yellow contours represent the 25 m s −1 wind speed at 200 hPa in 2020 and 2018, which are typical high year and low year of NH winter snow-cover PC1, respectively.The purple plus signs indicate the main anticyclonic centers of the CGT pattern.
) most models can reasonably capture the EOF1 of NH winter snow cover and Chinese summer precipitation (see figures S4 and S5), and (2) the simulation performance of multi-model ensembles (MME; figure 8) is considerably better than that of individual climate models (figures S4 and S5).The spatial correlation coefficients of EOF1 patterns between the MME and reanalysis data were >0.70 (figures 8(a) and (b)).Second, we evaluated the MME's ability to simulate the cross-seasonal relationship between NH winter snow cover and Chinese summer precipitation.Results show that the correlation coefficient

Figure 8 .
Figure 8. Spatial distribution of first modes of EOF analysis for the CMIP6 MME of (a) NH winter snow cover and (b) Chinese summer precipitation.The ACC between the simulated EOF1 and reanalysis EOF1 are given at upper right.(c) The corresponding series of PC1s for NH wintertime snow cover (brown line) and Chinese summer precipitation (blue line).