Record-breaking Barents Sea Ice Loss Favors to the Unprecedented Summertime Extreme Heatwave in 2021 over Western North America by Enhancing Rossby Wave Ridge

A record-setting extreme heatwave occurred over western North America (WNA) in the summer of 2021, which was associated with an extreme atmospheric Rossby wave ridge (ARR) over WNA and a minimum record event of the pre-winter Barents Sea ice concentration. We identify the temporal-spatial remote effect and investigate how the sea ice loss in the Barents Sea modulates the intensity of the ARR and extreme heatwaves by analyzing the reanalysis data. Our results suggest that atmospheric wave activity �ux associated with Barents Sea ice loss (BSIL) in pre-winter transferred wave energy towards the circumglobal teleconnection (CGT)-like wave trains, increasing the wave amplitude and enhancing the ARR over WNA. Meanwhile, the weakening of the jet streams drives atmospheric subsidence and increases the shortwave radiation, leading to anticyclonic anomalies over WNA and the enhancement of the ARR. Both the CGT-like wave trains and the weakened jet streams contribute to the synergistic effect of the enhancing ARR and more frequent extreme heatwaves. And the BSIL serves as a pioneer predictor for the predictions of extreme heatwaves 6 months in advance. The above analysis can improve the prediction skills of extreme heatwaves over WNA, possibly predicting more precisely.


Introduction
The summertime extreme heatwaves appear to be on the rise in the Northern Hemisphere as a result of climate change.The latest studies show that the mean spatial extent, maximum intensity duration, and frequency of the heatwaves have signi cantly increased under global warming (Perkins-Kirkpatrick and Lewis 2020; Rogers et al. 2022;McKinnon and Simpson 2022).The World Meteorological Organization reported that the temperatures reached a peak and broke the previous records in numerous cities across the United States and Canada in June 2021.With a temperature of 49.6°C (24°C higher than average), Lytton, British Columbia, Canada became the highest temperature ever recorded above 45 °N latitude worldwide according to Environment and Climate Change Canada.This extreme heatwave, which burned western North America (WNA) and Canada and caused huge economic losses and casualties, is around a once-in-a-millennium event (Ebi et  Due to the enormous impact this extreme heatwave puts on human life, how to trigger such a devastating event and the occurrence mechanism have drawn more attention in recent years.In late June of 2021, a large tropospheric blocking ridge was established after the unstable polar vortex center split and allowed an omega block to form at the surface and persistent anticyclonic anomalies over northwestern states of the U.S. and Canada (Overland 2021; Zhou et al. 2022).This strong anomalous high-pressure system over the WNA can explain half of the magnitude of the anomalous temperature (Zhang et al. 2023).
Persistence of the blocking allowed increased radiation, low soil moisture and adiabatic subsidence, which jointly contributed to this heatwave over WNA (Neal et al. 2022;Schumacher et al. 2022; Thompson et al. 2022).The anomalous atmospheric river across the North Paci c and a Rossby wave train excited by weakened monsoon precipitation in East Asia both resulted in this extraordinary heatwave event (Lin et al. 2022;Mo et al. 2022;Qian et al. 2022;Terray 2023).Wang et al. (2023) demonstrate that greenhouse gases are the primary cause for the trend of daily maximum temperatures in WNA from the past to the future.The equatorward shifts and weakening of the zonal jet streams occur more frequently after global warming, which will lead to higher temperatures in WNA (Overland 2021;Philip et al. 2021).In the mid-latitudes, ampli ed quasi-stationary circumglobal Rossby waves associated with anomalous jet streams are more likely to cause extreme heatwaves in the Northern Hemisphere The atmospheric dynamics in midlatitudes are acknowledged as too chaotic to predict, but the atmospheric circulation variations, such as anomalous atmospheric planetary waves, are helpful to make heatwaves more predictable beyond the scope of the typical weather forecast (Teng et al. 2013).Extreme weather and climate predictions are of immense societal and economical bene t.Unfortunately, even though most models were able to predict the above-normal temperature in the summer of 2021, the magnitude of the heatwave was severely underestimated (Emerton et al. 2022).Therefore, selecting the predictor related to speci c atmospheric circulation variations has been a foremost challenge in predicting future changes in heatwave events.The extreme heatwave in the summer of 2021 over WNA has been reasonably addressed by weather-scale processes.Previous studies also focus on the analysis of the synoptic and subseasonal timescale predictors, which make extreme heatwaves unexpected in climate prediction.
Given all the evidence above, two questions need to be fully addressed in this paper.One is what causes such an extreme heatwave event that happened in the early summer of 2021.The other is whether there is any credible predictor to improve the accuracy of the heatwaves forecasts in case to reduce the hazard to human life.To answer these two questions, we have examined the historical extreme heatwaves and the related general circulation conditions over WNA.We also brie y discuss the sea ice loss in the previous winter favors the positive geopotential height (GPH) anomalies via the propagation of Rossby waves and the weakening zonal jet streams.The results show that the Barents Sea ice loss (BSIL) can forecast the extreme heatwaves 6 months in advance over WNA, which makes it a good predictor.The rest of the paper is organized as follows.Section 2 describes the datasets and analysis methods.In section 3, we illustrate the extreme heatwave event in the summer of 2021, and its relationship with the atmospheric Rossby wave ridge (ARR) in the WNA.The impact of BSIL along with the weakening jet streams on the ARR, both contribute to heatwaves over WNA are shown in section 4. A summary and discussion are given in section 5.

Data
One set of data used in this study is the National Center for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis , with a horizontal of 2.5° × 2.5° (Kanamitsu et al. 2002).The daily and monthly atmospheric dataset includes surface air temperature at 2 meters, GPH and winds elds.The monthly sea ice concentration (SIC) dataset employed in our research is the Nimbus Unless otherwise noted, the summertime discussed in this paper is from June to July (JJ) during 1979-2021.The anomalies are calculated by using this period of time as the climatology.

Index de nitions
We de ne the index of the atmospheric Rossby wave ridge over WNA (I ARR ) to evaluate the variabilities of this high-pressure ridge.The normalized I ARR is calculated by area (30°N -60 °N, 120°W -90 °W) averaging of 500-hPa GPH anomalies (GPHa) in early summer.The area is selected by zonal deviation of 500-hPa GPH in climatology, which re ects the most active area of the ARR over WNA.
Likewise, the sea ice loss index (I SIL ) is calculated by averaging the opposite values of SIC anomalies in which region (78°N -83 °N, 12°E -78 °E) demonstrates the highest correlation coe cients with I ARR .The I SIL shows the opposite SIC variabilities, which is intuitionally interpreted as the variations of sea ice loss.

Analysis methods
The composite analysis is used in this study to demonstrate the spatial distribution of the anomalies of each variable to the normalized index.The composite patterns are obtained by averaging linear-detrended anomalies of those periods when the indices are below (above) one standard deviation as the positive (negative) phases.And we employ a student's t-test to assess the statistical signi cance of the differences between positive and negative composites.The probability density functions (PDFs) of the air temperature maximums (T max ) are calculated to measure the intensities of extreme heatwave events (Stefanon et al. 2012).We have calculated the T max of each day in early summer from 1979 to 2021 (43 years) as the cardinality.If there is a value that exceeds the upper 95th centile of the local PDF, we consider that there has been an extreme heatwave event that occurred that day.The magnitude above a threshold is the abnormal degree of the heatwave.
And the abnormal degree turns into 0 if the value is not able to exceed the threshold.The sum of the abnormal degree of the heatwave in a month is used to comprehensively assess the frequency and magnitude of the heatwaves, which can be considered as the integrated intensity of the extreme heat events.We extract the abnormal degrees of June and July per year and turn that into a time series.Then we normalize the index and select the period when the index is above (below) one standard deviation to composite (none) extreme heatwave events.

Atmospheric wave activity ux
The atmospheric wave activity ux is employed in our research to illustrate the atmospheric wave activity associated with BSIL.To diagnose the propagation anomalies of atmospheric Rossby waves, the formula for calculating wave-activity ux derived by Takaya and Nakamura (2001) is as follows: where p is the normalized atmospheric pressure (pressure per 1000 hPa), U (= (U,V) T ) indicate the average climatic wind elds, |U| is the speed of the climatological mean winds, Ψ' is the perturbation stream function, α is the earth's radius, N 2 is the buoyancy frequency squared, and φ and λ denote latitude and longitude, respectively.
Here the perturbation geostrophic stream function (Ψ') is also employed to illustrate the in uence of SIL on the Rossby wave trains.The seasonally averaged 500-hPa geostrophic stream function anomalies are calculated relative to the climatological mean from 1979 to 2021.The formula for calculating geostrophic stream function (Ψ) is as follows: Where Φ is 500-hPa geopotential height and f (= 2Ωsinφ) is the Coriolis parameter.Then we use the stream function anomalies regressing onto the I SIL to reveal the perturbation stream function induced by SIL.

Connection between heatwaves and the ARR over WNA
3.1 The unprecedented extreme heatwave event over WNA in the summer of 2021 In the early summer of 2021, a record-setting heatwave occurred in WNA.As reanalysis data shown in Fig. 1a, the six hours mean temperatures are all over 37°C from June 14th to 18th, 2021 with a maximum of 39.8°C (17.1°C higher than climatology) in California, America (the triangle mark in Fig. 1b).The World Meteorological Organization has de ned a weather process in which the daily maximum temperature exceeds 32°C for over 3 days as a heatwave event.There were 6 times heatwaves (42 days above 32°C) in early summer (JJ), and the average climatology is below 25°C during this time, which made the heatwaves more intense than the late summer.Such extreme heatwaves in the early summer of 2021 affected numerous unprepared people and caused hundreds of deaths (Philip et al. 2021).Therefore, we primarily discuss the heatwaves and the related factors during early summer (JJ) through this article.
The horizontal structure of the GPHa and surface winds anomalies during the early summer of 2021 are shown in Fig. 1b.Note that the robust high-pressure anomalies collaborated with strong anticyclone anomalies have been taken over WNA at that time, which strengthens the intensity of the high-pressure ridge.The maximum GPHa reaches 73.3 gpm.Such conditions trigger the summer blocking high, which usually contributes to the persistence of heatwave events.There are southeasterly anomalies near California compared to the climatological westerly to southwesterly, which leads to less cold and wet advection transport to this area.
Φ f The anomalies along with the climatology of the temperature and GPH over North America in the summer of 2021 are demonstrated in Fig. 2a and 2b Hovmöller diagrams respectively.There are usually two warm centers located at 95 °W and 115 °W in July, and the temperature of the eastern center is higher than the western one.But in the early summer of 2021, the warm anomalies mainly distribute in the west, with a large area of cool anomalies over the southern U.S. The dipole structure is the principal mode of heatwave variability over North America (Yu et al. 2023).And the eastern one shows below-normal anomalies in late June and early July.Such a phenomenon is highly correlated with the variations of GPH.As shown in Fig. 2b, the quasi-stationary Rossby wave ridge centered at 100 °W moving westward (110 °W) during 12-16 June, and one week later on 21-30 June, a blocking high is maintained and centered at 120°W.The conditions of GPHa are in favor of warm extremes and causing the heatwaves during 14-20 June and 26 June to 14 July, which centers moving from 100 °W to 115 °W and quasistationary centered at 120 °W.These heatwaves are one month in advance of the hottest period in the WNA, which brings a great disaster to unprepared people.
Historically, the summertime warm center in Fig. 2c is usually at 100 °W and the synchronous highpressure ridge in Fig. 2d is around 110 °W.When the ridge goes higher, for instance, in 1987, 2005, 2016, etc., it shifts westward, which causes the warm center hotter and further west.In the summer of 2021, the GPH is much higher than the historical records and signi cantly shifting westward from climatology, which leads to the record-setting extreme heatwave that occurred in WNA.As Hovmöller shows, the positive GPH anomalies are highly related to the heatwave events in summer.The temperature response to high-pressure anomalies is consistent with previous studies on heatwaves that occurred in Europe (Black et al. 2004;Zschenderlein et al. 2019).Therefore, we need to measure the variations of the summertime ridge which is highly related to the heatwaves over WNA next.The correlation coe cient between I ARR and the TA index is 0.89 on the 0.01 signi cance level, which con rms the relative role of ARR to the heatwave events in early summer over WNA.
In case to further con rm the level of the heatwave in uenced by the intensity of ARR, we employ composite analysis and PDFs to reveal the correspondence between temperature and pressure variations (see Section 2.3 for details).As displayed in Fig. 4a, the integrated temperature anomalies which exceed the upper 95th centile of the local PDFs are located in the Midwest of North America when ARR demonstrates above-normal anomalies.And the maximum values are just in the area where we select to measure the variations of ARR.On the contrary, there is no heatwave event relevant to the weakening of ARR in central and WNA (Fig. 4b).Only the Alaska region has a slight connection with it.Therefore, we have con rmed that the extreme heatwaves are consistent with the enhancement of the ARR over WNA.The anticyclonic winds anomalies and atmospheric subsidence make the temperature rise through adiabatic heating and shortwave radiation.The heat budget results for the heatwaves con rm the contribution of the adiabatic heating process within the anticyclone dominating the heatwaves over North America (Qian et al. 2022).As we nd that extreme heatwaves are more likely to take place in WNA when the ARR is enhanced.Next, we need to focus on the mechanism of the ARR variations and nd the factors or even better predictors to cause I ARR changes.
4 Record-breaking Barents Sea Ice Loss Favors to the enhanced Rossby wave ridge over WNA which causes more frequent quasi-barotropic ridges and more devastating heatwaves over WNA.However, the SIC anomalies in the Barents Sea in the pre-winter related to summertime ARR over WNA and lasted for more than half a year, which might be a good predictor for the heatwaves and explain such an extreme heatwave event in 2021.The SIC in the Chukchi Sea seems correlated with ARR in winter, but the correlation coe cients are not as signi cant as that in the Barents Sea.Also the signless SIC in Chukchi Sea in spring means the SIC in summer may not contain the signal from the previous winter, which cannot make SIC in Chukchi Sea a predictor.
The wintertime sea ice melting in the Barents Sea can last till early summer by different positive feedbacks.The positive ice-albedo feedback will amplify the initial warming and sea ice retreat (Curry et al. 1995).For instance, the Barents Sea ice feedback usually reaches its peak in winter and spring (Koenigk et al. 2009).Less sea ice in winter will postpone the freezing processes and make the thinner of ice (Zhou et al. 2022).The thinner sea ice would shrink faster to intensify the ice-albedo feedback (Parkinson and Washington 1979).The reduced SIC and thickness at the end of winter could lead to more melting in summer.So the SIC anomalies in winter will last for more than six months after accelerated melting in spring, which explains what makes sea ice loss in pre-winter the leading signal of the ARR in summer.
Due to the thermal inertia of the sea ice, we nd that the I ARR is highly related to the variabilities of SIC in the Barents Sea from previous winter to summer.The signal preserves in the sea ice for long memory, which makes it a good predictor.Thus, we select the Barents Sea region (78°N -83 °N, 12°E -78 °E; green box in Fig. 5) and calculate I SIL (See Section 2.2 for details).Note that the I SIL is the negative areaaveraging value of SIC, which is used to measure the melting of sea ice in the Barents Sea.Normalized I SIL in pre-winter, spring and early summer are shown in Fig. 6 with the rising trends.The correlation coe cient between I SIL in early summer and pre-winter is 0.75 on the 0.01 signi cance level.As the similar variations of I SIL in three seasons, we choose the I SIL in pre-winter as the predictor to study its relative roles with ARR over WNA.Unless otherwise noted, the I SIL refers to the variations of BSIL in prewinter (DJF).According to the extreme heatwave in 2021, we nd I SIL in the pre-winter reaching its maximum.The correlation between I ARR and the pre-wintertime I SIL is 0.62 (p < 0.01).The results indicate that the SIC will serve as a predictor of ARR which eventually leads to extreme heatwave events.Previous studies focused on the synchronous factors causing heatwaves but ignored the possibility of predictors.
By using I SIL , we could make the heatwaves taking place in WNA more predictable at least six months in advance.In this case, we should look into the mechanism of the relative role between BSIL and ARR more deeply.

Way of BSIL on the enhanced ARR via circumglobal teleconnection
To better understand the relative role of BSIL on the ARR, we regress the summertime 500-hPa GPHa onto the I SIL (Fig. 7a).The regression pattern shows the spatial feature of the circumglobal teleconnection (CGT) -like pattern in the midlatitudes.The CGT pattern identi es the leading mode of summertime GPHa in the northern hemisphere (Ding and Wang 2005).The results indicate that BSIL triggers CGT-like wave trains that contribute to the enhancing ARR and lead to increasing extreme heatwaves over WNA.The regressed 500-hPa GPHa onto the summertime I ARR shows a similar pattern to Fig. 7a (not shown).
Therefore, the BSIL will sustainedly affect the summertime general circulation since pre-winter.
To further understand the role of SIL on the ARR and heatwaves, we construct the composite patterns by I SIL > 0.8 (Fig. 7b) and I SIL < -0.8 (Fig. 7c).We nd the asymmetry patterns when comparing these two composites.During the period of BSIL, the robust CGT-like pattern (similar pattern as in Fig. 7a) is triggered and the positive anomalies center in WNA that is related to ARR is signi cantly enhanced, which explains the consistent extreme heatwaves with the ARR (Fig. 4a).However, the ARR center is not shown decrease robustly when sea ice of the Barents increases.This is because the increasing sea ice still has a profound in uence on the atmospheric circulation by a westerly jet waveguide forming an opposite phase shift CGT-like pattern in which center locations and intensities are different.Note that generally the ARR is weakened and the center near Alaska is enhanced, which explains the heatwave events only occurred in Alaska region but not in WNA (Fig. 4b).
Next, we use 500-hPa atmospheric T-N wave activity ux (see Section 2.4 for details) to demonstrate the contributions of the sea ice in the Barents Sea to the general circulation.We draw the atmospheric wave activity patterns in June (Fig. 8a) and July (Fig. 8b), which are relative to BSIL.The pattern in June is clearly shown that there are Rossby wave disturbances in the Barents Sea that transmit to the Siberian, the source center of the CGT-like pattern, which increases the amplitude of each center through the Rossby wave trains and eventually enhances the ARR.Besides, as shown in Fig. 8b, the CGT-like wave trains shift northeastward in July, which explains why the eastern center of temperature maximum is higher than the western one in July (Fig. 2a and 2c).The anomalous energy of the Arctic melting sea ice disperses downstream and modulates the atmospheric circulation through wave-current interactions, which resulted in the enhancement of ARR.In June 2021, there were two unexpected extreme heatwaves occurred.And we compute the T-N wave activity ux patterns of these two events during 12-16th (Fig. 8c) and 21-30th (Fig. 8d) June 2021, respectively.The perturbation wave source in the Barents Sea emanates wave activity ux southeastward to the positive source center of a CGT-like pattern and then enhances the ARR during the rst process in 2021.The second heatwave process shows that the ampli cation of the ARR amplitude in the late June of 2021 was mainly due to a relay Rossby wave source originating from the North Paci c, one of the negative centers of the CGT wave trains.

Role of the weakened jet streams induced by BSIL on the enhancement of ARR
Weakened high-level jet streams usually increase the amplitude of ARR through the energy cascade of wave-ow interaction.A recent study nds that the weakening of the zonal winds in the Eurasian continent is signi cantly in uenced by the Arctic ampli cation effect through Arctic-midlatitude temperature contrasts modulation (Wang et al. 2017).The melting of the sea ice in the Arctic will reduce the meridional temperature gradient.The polar-equatorial temperature gradient in the Northern Hemisphere has signi cantly decreased in the past, which may explain the linear uptrend of ARR.In our study, the 200-hPa zonal winds associated with BSIL turn weaker and the westerly jet core breaks over the WNA in June (Fig. 9a) and July (Fig. 9b).It indicates that the negative zonal winds anomalies are in favor of the development and persistence of the high-pressure system.The jet stream belt in 30-40 °N is signi cantly weakened in July, which leads to the stagnation of the weather systems and eventually enhances the ARR (Fig. 9b).The weakened jet stream in this region indicates that the transmission of energy is from the mean ow to the synoptic scale perturbations through wave-mean ow interaction.The synoptic-scale energy accumulates in the convergence of the wave activity ux over North America, which donates the enhancement of the ARR (Fig. 8).
In addition, we draw the zonal wind patterns during two extreme heatwave processes (Fig. 9c and 9d).
The rst heatwave process shows the weakened jet stream in the WNA.Combined with the enhanced center of CGT-like pattern associated with the BSIL makes the intensity of ARR ampli ed in 2021 (Fig. 8c).
During the second event, the zonal winds are signi cantly weakened over the western coast of North America, which is consistent with the convergences of wave ux in WNA in the late June of 2021 (Fig. 8d), suggesting the enhancement of the ARR and leading to the extreme heatwave.
It is noted that a north-south dipole anomaly zonal wind pattern with a westerly anomaly in the north and easterly anomaly in the south always appears in the gap region of the jet stream core over the WNA (Fig. 9a, c, d).This dipole zonal wind anomaly pattern favors a high-level anticyclone geopotential height anomaly consistent with sinking air ow and ARR over the WNA (Fig. 7a).In a word, weakened high-level westerly jet induced by BSIL plays an important role in enhancing ARR by atmospheric jet stream dynamics.

Summary and discussion
The research elucidates the relative roles of BSIL in the variations of the Rossby wave ridge over WNA in early summer, which eventually causes heatwave events.According to observational reanalysis, a recordsetting extreme heatwave occurred over WNA in the summer of 2021.Meanwhile, we noticed that the intensity of the ARR in WNA also reached the extreme.The composite analysis showed that extreme heatwaves are more likely to take place in WNA when the ARR is enhanced.The ARR leads to extreme heatwaves by increasing atmospheric subsidence and shortwave radiation.We also found that the Barents Sea ice also set a minimum record in SIC just in the pre-winter ahead of the 2021 summertime heatwaves.For the thermal inertia of the sea ice, we have demonstrated the BSIL index I ARR closely associated with variabilities of Barents Sea ice from the previous winter to synchronous summer, which makes the Barents Sea ice a performance predictor 6 months in advance.The impacts of BSIL on the ARR and summertime extreme heatwaves are driven by two pathways: (a) atmospheric wave activity ux associated with BSIL in pre-winter will transfer wave energy towards the CGT-like wave trains, which increases the amplitude of pattern and enhances the ARR over WNA; (b) the weakening of the jet streams drive atmospheric subsidence and increases the shortwave radiation, which leads to the anticyclonic anomalies over WNA and the enhancement of the ARR.Both the CGT-like wave trains and the weakened jet streams contribute to the enhancing ARR and more frequent extreme heatwaves.
Our study identi es a temporal-spatial remote effect between the variations of Barents Sea ice in prewinter and the Rossby wave ridge over WNA in summer.The synergistic effect of the BSIL induced CGTlike wave trains and the weakened jet streams that cause the enhancing ARR and extreme heatwaves are summarized schematically as in Fig. 10.According to the physical mechanism, the extreme heatwaves over WNA will possibly predict more precisely in advance.
Heatwave events vary in their forcing mechanisms and connections to large-scale climate modes.We also have explored the in uence of other factors such as the variabilities of the North Paci c SST on the heatwave events over North America, but there are no robust results or signi cant signals that have been revealed.The researches on the coupled ice-atmospheric processes are necessary to carry a deep understanding of the diverse roles of extreme heatwaves under a warming climate.
Schematic diagram for the mechanisms of Barents Sea ice loss in uencing the atmospheric Rossby wave ridge and jet stream over western North America in the summertime, which eventually leads to heatwaves.
al. 2021; Philip et al. 2021; Schramm et al. 2021; Henderson et al. 2022).Furthermore, along with the enhanced North Paci c sea surface temperature (SST) anomalies and atmospheric circulation variations under global warming, there will be more intense, more frequent and longer lasting heatwaves over WNA in the second half of the 21st century according to the future scenario runs of coupled climate models (Meehl and Tebaldi 2004; Chen et al. 2018; Chen et al. 2023).According to statistics, the extreme events show signi cant rising trends through time in both observational reanalysis datasets and projections of climate models (Bartusek et al. 2022; Thompson et al. 2022; Dong et al. 2023; Heeter et al. 2023; Lucarini et al. 2023).

(
Coumou et al. 2014; Lin and Yuan 2022).Human activities rise the maximum temperature of extreme heatwaves (King et al. 2016; Philip et al. 2021; Bercos-Hickey et al. 2022).And these extreme events tend to be more frequent in the summertime during La Niña years (Luo and Lau 2020).Meanwhile, the establishment of blocking contributes to enhancing radiative forcing, subsiding atmosphere and increasing diabatic heating, which all eventually lead to surface temperature rising (Zschenderlein et al. 2019).The formation and maintenance of the blockings are signi cantly in uenced by synoptic-scale eddies, upstream latent heating and jet streams (Hoskins et al. 1983; Luo et al. 2001; Nakamura and Huang 2018; Kim and Lee 2022).However, the Eastern Paci c-North American region tends to experience blocking less frequently (Tibald et al. 1997).Therefore, there must be some other factors to explain the more frequent heatwaves in recent decades over North America.The Arctic sea ice and snow cover have undergone unprecedented changes.Recent studies reveal that near-surface Arctic warming along with sea ice melting can modify the atmospheric circulation and weather-related factors, such as jet streams, meridional temperature gradient and storm tracks in the midlatitudes, which will contribute to amplifying the intensity and frequency of extreme weather and climate events (Francis and Vavrus 2012; Cohen et al. 2014; Barnes and Screen 2015).Meanwhile, the combined loss of sea ice and snow cover modulate interannual and subseasonal atmospheric variabilities in the midlatitudes and lead the extreme events (Honda et al. 2009; Deser et al. 2010; Tang et al. 2013; Vihma 2014; Gu et al. 2018; Chripko et al. 2021).The maintenance of the sea ice loss in the Arctic relies on local atmospheric and oceanic processes and self-supporting feedback spontaneously, such as sea ice thickness and concentration feedback, albedo feedback, snowfall feedback, cloud and water vapor feedback, and temperature feedback (Curry et al. 1995; Serreze et al. 2009; Screen and Simmonds 2010; Döscher et al. 2014; Zhou et al. 2022).Furthermore, sea ice variations make a signi cant contribution to the long-term prediction of weather and climate (Wu et al. 2009; Xing and Huang 2019; Zhu et al. 2019).
-7 SMMR and DMSP SSM/I-SSMIS Passive Microwave Data Version 2 distributed by NASA National Snow and Ice Data Center Distributed Active Archive Center (NSIDC DAAC) with the resolution of 25 kilometers, which is used to analyze the Rossby wave ridge over North America associated with the Arctic sea-ice loss (DiGirolamo et al. 2022).
The regression analysis is to demonstrate atmospheric variabilities associated with indices of the atmospheric activity.The regression patterns are extracted by regressing the detrended anomalies of the atmospheric elds (relative to the climatological mean for 1979-2021) onto the normalized indices.

3. 2
Frequent extreme heatwaves induced by ARR over WNAThe horizontal structures of the climatological winds and GPH in early summer for 1979-2021 are shown in Fig.3a.To portray the variations of the ARR over WNA, we calculate the zonal deviations of the latitudes and select the appropriate area (30°N -60 °N, 120°W -90 °W) to compute the ARR index I ARR (see Section 2.2 for details).The normalized I ARR and synchronous air temperature (TA) index calculated by averaging the temperature anomalies in the same region with the ARR index are displayed in Fig.3b.There are robust and similar uptrends along with interannual variabilities in both indices.Note that global warming and anthropogenic forcing result in warming and thermal expansion of the lower atmosphere, which leads to linear uptrends(Christidis and Stott 2015).The I ARR in 2021 reached its maximum since 1979.The ARR features a maximum of ~ 5980 gpm in June 2021, which is unprecedented for the WNA since 1979(Overland 2021).The TA index in 2021 also rises to the highest in 43 years of reanalysis data.

4. 1
Barents Sea ice loss in previous winter relevant to the ARR over WNA in early summer In the early winter of 2021, the Barents-Laptev sea ice reached a minimum for the period 1981-2020 and reduced the meridional gradient (Yang and Fan 2022).In this section, we further reveal the potential roles of the Arctic sea ice anomalies in the changes of ARR intensities over WNA.The correlation patterns between I ARR and SIC of the Arctic Ocean in previous winter, spring and synchronous summer are shown in Fig. 5.The patterns demonstrate the signi cant negative correlation between I ARR and SIC from previous winter to summer in the Barents Sea area where displays the minimum correlation coe cients down to -0.60 Wang et al. (2022) indicate that the sea ice loss in the summertime Chukchi Sea is a factor

Figures
Figures

Figure 4 The
Figure 4

Figure 5 The
Figure 5