1. Introduction
The atmospheric transport of energy into the Arctic serves as an important external energy source. It has been shown to greatly alters the Arctic near-surface air temperature, and even enter the upper ocean and alter the marine environment of the Arctic Ocean through the coupling process (Zhang 2005; Burgard and Notz 2017). Graversen et al. (2008) conclude that changes in atmospheric energy transport may be an important cause of the recent Arctic temperature amplification. And the internal variability of Arctic sea ice is primarily caused by atmospheric temperature fluctuations (Olonscheck et al. 2019). The energy transport accomplished by the atmosphere at the Arctic boundary (70°N) is estimated to be comparable to absorbed solar radiation (Oort and Peixóto 1983). Therefore, changes in poleward energy transport could have a significant impact on the interannual variability and long-term trends of the Arctic climate system. The atmospheric energy flux consists of dry static energy (DE) and latent energy (LE) components. The internal, potential and kinetic energy constitute the former, and the latter is closely related to moisture transport.
It has been indicated that the latent energy transport may affect the near-surface air temperatures more than the dry static counterpart, as the LE transport brings not only energy through the condensation process but also water vapor. Moisture in and of itself and the associated cloud water can enhance the local greenhouse effect through cloud feedback (Johansson et al. 2017; You et al. 2021), which increases the downward longwave radiation (Park et al. 2015), thus warming the surface (Graversen and Burtu 2016; Gong et al. 2017) and favoring earlier melt onset during later spring or early summer (Maksimovich and Vihma 2012). This implies more absorbed solar radiation and more severe ice loss due to the ice-albedo feedback (Mortin et al. 2016; Horvath et al. 2021). Kapsch et al. (2013) found that years with a low September sea ice concentration (SIC) are characterized by more persistent periods in spring with enhanced energy flux to the surface in the forms of net longwave radiation plus turbulent fluxes. The poleward atmospheric transport of energy and moisture toward the high Arctic would also affect the precipitation and its phase, thus affecting the snow accumulation, events of rain-on-snow/ice, and freezing/thawing processes on the surface of sea ice, and ultimately regulating the thermodynamic growth and decay of sea ice (Gimeno-Sotelo et al. 2019), as well as the meltwater cycle on a floe scale (Perovich et al. 2021). Model results reveal that there is an increase in LE flux in the future (Hwang et al. 2011; Skific and Francis 2013), which will continue to contribute to Arctic warming (Graversen and Burtu 2016).
The transport of total energy and moisture from lower latitudes toward the Arctic system is dynamically controlled by changes in large-scale atmospheric circulation (Graversen et al. 2011; Vihma et al. 2016) and synoptic activities (Kim et al. 2017; Villamil-Otero et al. 2018; Fearon et al. 2020). Previous studies revealed that cyclones are the main synoptic activities that carry energy and water vapor, which act as a vital component in maintaining the global atmospheric balance of energy, moisture, and momentum (Ulbrich et al. 2009; Varino et al. 2019). Specifically, due to the almost exclusively meridional flux carried by transient eddies, the moisture transport at 70°N is dominated by cyclones that explain 80%–90% of the total northward transport (Jakobson and Vihma 2010; Dufour et al. 2016; Fearon et al. 2020). Clancy et al. (2022) argued that the spatial patterns in the atmosphere and sea ice during Arctic cyclones are heterogeneous, with cyclonic surface winds, a warm, moist atmosphere and decreasing sea ice to the east of cyclones. The characteristics of cyclone activities have changed substantially in recent decades. Zhang et al. (2004) found that Arctic cyclones exhibited pronounced low-frequency variability, noting that the number and intensity of cyclones entering the Arctic have increased (Sepp and Jaagus 2011). Based on reanalysis data and model results, many studies reported an increased number of cyclones whose trend patterns are not spatially coherent (Zahn et al. 2018; Akperov et al. 2019a,b; Valkonen et al. 2021). For instance, Rinke et al. (2017) found an increased frequency of extreme cyclones in the Arctic North Atlantic during winter, which is consistent with observed significant winter warming in the region. Moreover, the consistent poleward and upward shift and intensification of the storm tracks have been identified in model simulations (Yin 2005), accompanied by the northward shift of surface wind stress, total energy, and precipitation. This is consistent with Bader et al. (2011), whose results showed that storm tracks in midlatitudes are forecasted to experience a poleward shift. There are also indications of a shift toward more meridional storm tracks over the Barents and Greenland Seas (Wickström et al. 2020).
To summarize, previous studies have reached a consensus that poleward atmospheric latent energy transport could impact Arctic near-surface air temperatures substantially, thereby affecting sea ice variations. Especially during the Arctic winter when solar radiation is absent, variations in downward longwave radiation (DLR) associated with LE transport are fundamental in modulating surface air temperature (Kim and Kim 2017). And in recent years, the LE flux toward the Arctic may experience changes due to the alteration of large-scale atmospheric circulation patterns and cyclone activities. However, there still exist some cognitive gaps that limit our understanding. For instance, previous studies focus on different atmospheric energy transport components (e.g., total energy, moist static energy, latent energy transport), thereby leading to inconsistent results. And they tend to address the main gateways through which atmospheric transport enters the Arctic in a general sense with different pathways not being separated (Jakobson and Vihma 2010; Dufour et al. 2016; Mewes and Jacobi 2019; Naakka et al. 2019; Nygård et al. 2020). Thus, we lack knowledge about frequencies of separated LE pathways and their long-term variations during winter in recent decades. This is important given that the sea ice changes in winter exhibit significant regional differences. And to what extent are different LE pathways modulated by cyclone activities, which experienced considerable changes in recent years? Moreover, the distinct impact of different atmospheric energy transport patterns on the Arctic sea ice mass balance remains ambiguous. The environment of the Arctic has transformed into a new state with younger (Rigor and Wallace 2004; Tschudi et al. 2016), thinner (Kwok and Rothrock 2009; Bi et al. 2018), and more easily deformed (Lei et al. 2020) sea ice, which is more susceptible to changes in atmospheric forcing. Against this background, the answers to these questions are crucial.
In the present study, we aim to extract the general pathways of atmospheric LE transport variability during winter by applying self-organizing maps (SOM) classification to daily fields over the satellite record (1979–2021). Therefore, the patterns depicted by the SOM nodes include modes of variations at the synoptic time scale and beyond. Winter is defined as January–March when the temperature is the lowest. Our results are tested to be robust if winter is alternatively defined as December–March. The SOM method is a type of artificial neural network, but it has been extensively applied in Arctic climate research. Previous studies have utilized the SOM method to address linkages between atmospheric circulation and Arctic sea ice anomalies (Lynch et al. 2016; Yu et al. 2019), to classify atmospheric moisture transport pathways in the vicinity of the Greenland Ice Sheet (Mattingly et al. 2016), and to investigate patterns of atmospheric heat transports and their relationship with Arctic air temperature (Mewes and Jacobi 2019). The main advantage of SOM is that the training process allows for nonlinear features of the complex climate system to be captured, thus providing a more detailed and realistic view of the true nature of variables of interest (Reusch et al. 2005). Then, these synoptic-scale LE transport patterns are related to cyclone activities. Further, we identify the transport patterns that give the largest contributions to Arctic sea ice changes and how these patterns act on sea ice loss. To achieve the above diverse aims, we organized this paper as follows: The datasets and methodology utilized are introduced in detail in section 2; The spatial and temporal characteristics of primary patterns of LE transport variability, as well as the regulating roles of cyclones, are outlined in section 3; In section 4, how these LE transport patterns influence the Arctic sea ice loss and the underlying mechanisms will be discussed; and a summary and discussions are presented in section 5.
2. Data and methods
a. Data
We use the satellite-derived daily sea ice concentration (SIC) product provided by the National Snow and Ice Data Center (NSIDC) to investigate the SIC variation and its association with different LE transport pathways. SIC fields are derived from brightness temperature measured with the Scanning Multichannel Microwave Radiometer (SMMR), the Special Sensor Microwave Imager (SSM/I), and the Special Sensor Microwave Imager Sounder (SSM/IS) by applying the bootstrap algorithm (Comiso 2017). The latest version (version 3.1) of the dataset provides improved consistency between sensors through the use of a suite of daily varying tie points generated from the Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) observations. The daily SIC fields are available on a polar stereographic projection with a spatial resolution of 25 km.
Large-scale atmospheric indexes, such as Arctic Oscillation (AO) and North Atlantic Oscillation (NAO), are obtained from the Climate Prediction Center (CPC) at the National Oceanic and Atmospheric Administration (NOAA). AO is the leading EOF mode of monthly mean 1000-hPa height anomalies poleward of 20°N latitude. The positive phase of AO is characterized by stronger westerlies at subpolar latitudes and lower SLP over the Arctic. The NAO and AO indices are highly correlated, yet the NAO captures more of the regional (North Atlantic/European) variability in the atmospheric circulation (Thompson and Wallace 1998). The Arctic dipole anomaly (DA), which represents the second-leading mode of the SLP anomaly in the Arctic north of 70°N, is calculated following Wu et al. (2005). In its positive phase, negative SLP anomalies appear over the Kara and Laptev Seas, and a concurrent positive SLP occurs in the Canadian Arctic Archipelago extending southeastward to Greenland. When developing into its negative phase, the opposite holds.
b. Methods
1) The SOM method
Generally, the SOM represents a neural network algorithm using unsupervised learning to reduce the dimensionality of a dataset by organizing it into a two-dimensional array called a map (Kohonen 1998). Each node in the map includes a spatial pattern and a time series of its occurrence frequency. First, the size of the map is determined (e.g., a 4 × 4 matrix), which is sufficient to capture the important features in the input datasets while maintaining a small size for patterns to facilitate interpretation and visualization. The initial step of the SOM method is to create the first-guess array so that each node has a reference vector that has equal dimensions to the input data. In this first-guess map, the eigenvectors from the covariance matrix of datasets with the largest eigenvalue are placed in the corners, and the reference vectors of the remaining nodes are derived by linear interpolation. The map is then trained iteratively by providing the SOM with samples of the original fields, and the similarity (measured by the Euclidean distance) between the sample and each reference vector is calculated. In each iterative process, the “best match” node is identified and adjusted most. And the nodes around it, within a user-defined radius, are also updated but to a lesser extent depending on their distance from the most similar node. The degree of adjustment is defined by a user-defined learning-rate parameter. Since the SOM procedure depends on adjacent nodes, the more similar patterns are placed next to each other, allowing for a more intuitive interpretation of the patterns. The general topology and continuum of the original dataset thereby can be preserved. More detailed descriptions of the SOM method can be found in Kohonen (1998), Reusch et al. (2005), Mattingly et al. (2016).
In this study, we apply SOM to derive the main pathways of daily LE flux anomalies from January through March during 1979–2021 north of 50°N. Note that the daily LE flux fields are interpolated to a 50-km Equal-Area Scalable Earth (EASE) grid to assure the equal weight of each grid in the training of the SOM. For each day, the climatological value is obtained by averaging daily data over 43 years, and an anomaly for that day is obtained by subtracting the climatological value from the daily data. A 4 × 4 SOM map with 16 nodes was chosen to represent synoptic-scale horizontal LE transport during winter. Different size of metrics was tested to determine a suitable number of nodes for the analysis. If the matrix is too small, some characteristics may not be represented and separated well from each other. If it is too big, there exist redundant nodes that feature a transition state between two pathways, and adjacent nodes will be too identical and visualization is unwieldy. Moreover, the results, including primary pathways and their frequencies, relationship with cyclones and impacts on sea ice, are not sensitive to small differences in the matrix size. Thus, we use a 4 × 4 matrix in the present study, given a balance between a pragmatic minimum of nodes and sufficient patterns to characterize essential variabilities. Then the 4 × 4 SOM map was linearly initialized before the iterative training process. Next, each node (denoting a typical pattern of daily LE flux anomalies) in the SOM was trained and self-adjusted based on the similarity between the sample LE flux pattern and each node iteratively. After many iterations, the SOM map tends to approximate the original LE anomalies fields with primary patterns that emerged (nodes in Fig. 1).
SOM results of vertically integrated horizontal LE transport anomalies during winters from 1979 to 2021. The percentage in the top right of each pattern correspond to the relative frequency of occurrence during the analyzed period. The patterns grouped together are separated by polylines.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
SOM results of vertically integrated horizontal LE transport anomalies during winters from 1979 to 2021. The percentage in the top right of each pattern correspond to the relative frequency of occurrence during the analyzed period. The patterns grouped together are separated by polylines.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
SOM results of vertically integrated horizontal LE transport anomalies during winters from 1979 to 2021. The percentage in the top right of each pattern correspond to the relative frequency of occurrence during the analyzed period. The patterns grouped together are separated by polylines.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Once all daily anomalies of LE flux are assigned to a node in the 4 × 4 map, the frequency of their occurrences can be retrieved (percentage labeled in Fig. 1). Then we use composite analysis to inspect the relationships between different LE pathways and cyclone activities, as well as to assess their contributions to Arctic sea ice loss in winter during 1979–2021. Once anomalies of air temperature, water contents, and surface radiation variables are associated with LE transport regimes of SOM, a Monte Carlo approach is utilized to test the statistical significance of these anomalies. Three thousand random subsets of a given field, which have the same number of samples as the corresponding SOM node, are constructed from the original data. Then we performed a two-tailed test to compare anomalies in the random subsets to the anomalies in related SOM nodes to determine the significant anomalies that are different from zero at a certain level.
2) Cyclone identification and tracking
We use a revised automatic cyclone identification and tracking algorithm developed originally by Serreze et al. (1993) to diagnose the center positions and trajectories of the cyclones from the 6-hourly SLP data (Serreze et al. 1993; Wang et al. 2013). The 6-h SLP fields were first georeferenced to a 50 km EASE-Grid prior to applying the algorithm and regions with surface elevations greater than 1000 m are excluded to avoid larger uncertainty in the SLP over high terrain. There exists a range of ideas and concepts regarding how to identify and track extratropical cyclones due to different definitions and their relatively complicated characteristics. Neu et al. (2013) conducted an intercomparison experiment involving 15 commonly used detection and tracking algorithms for extratropical cyclones, in which they confirmed that the algorithm adopted in our study agrees well with the others in terms of spatial distribution, interannual variability, and long-term trends of the cyclone characteristics. It gives credence to the method utilized in this study.
The cyclone detection and tracking algorithm works as follows: First, iterating all the grids at a time step to determine all the candidates of cyclone centers. A cyclone center should have a minimum pressure value that is at least 0.1 hPa (Wang et al. 2013) lower than the surrounding 7 × 7 array of grid points, with a positive pressure gradient from the center outward. Then, tracking the centers between two consecutive time steps based on the “nearest-neighbor” rule to construct the cyclone trajectories, with further checks including the distance moved in specific directions and pressure tendency. Therefore, a cyclone track consists of a series of cyclone centers identified in sequential time steps at adjacent locations. We use multiple parameters including the maximum travel distance of 800 km, maximum northward, southward, and westward migration of 600 km, and maximum pressure tendency of 20 hPa within two consecutive time steps (6 h) (Serreze 1995; Wang et al. 2006) in this study. Meanwhile, we retrieve the corresponding features for each cyclone, including the duration and intensity. The intensity is referred to as the difference between the SLP of the cyclone center and the climatological monthly mean SLP at corresponding grid points. The density of tracks denotes the number of distinct cyclones occurring in a particular region over a period of time. An integrative parameter called the cyclone activity index (CAI) is used to measure the intensity, number, and duration of a cyclone. The CAI is defined as the sum of the intensity of all cyclone centers in a particular region over a given period (Zhang et al. 2004). Cyclone features are related to primary LE pathways using composite analysis to inspect the regulation of LE transport and to assess which pathways are mostly impacted by cyclone activities.
3. LE transport pathways and the associations with cyclone activities
a. LE transport SOM
In this section, we present and analyze the SOM results of classifying the vertically integrated daily LE flux anomalies north of 50°N in winter from 1979 to 2021 (Fig. 1). Each day in January–March during the 43 years was mapped onto the most identical node in the map based on minimum Euclidean distance. Obviously, similar patterns are typically concentrated together on the map, and in all of the nodes, transitions from one regime to another commonly occur at the neighboring or diagonal positions. In this 4 × 4 map, node 16 represents the most frequent pattern, with the largest (10.1%) frequency of occurrence. It identifies a pathway by which LE is transported toward the Arctic through the Bering Strait and the southern part of the Norwegian Sea, while its almost opposite pattern (node 1) has the third largest frequency (9.8%). Node 13 appears to be the second-frequent pattern (9.9%) and is characterized by a strong poleward LE transport through the Davis Strait and a high southward flux over the Greenland Sea. In contrast, about 360 days (9.3%) witnessed a conspicuous northward LE transport through the Greenland Sea and Fram Strait into the Arctic Ocean during the 43 winters, as represented by node 4 (Fig. 1).
For ease of discussion, the 4 × 4 SOM map of vertically integrated horizontal LE transport during winter has been divided into four categories by grouping similar patterns together (Fig. 1). Putting structures that share related features into one group has been used in previous studies to make sure that the resulting patterns could fit from a meteorological perspective (Higgins and Cassano 2009; Mattingly et al. 2016). In this study, we divide all nodes into four primary pathways based on their similarity measurement as well as the physical characteristics according to the horizontal LE transport from lower latitudes into the Arctic (Fig. 2). The resulting groups are derived by integrating the distinct patterns of each group in Fig. 1 weighted by their relative frequency of occurrence. Note that the climatological LE transport pattern is characterized by a large northward flux predominantly through the North Atlantic (Fig. S2 in the online supplemental material). Therefore, the SOM method was applied to the anomaly fields to better distinguish the general pathways with significant variations in horizontal LE transport. The corresponding patterns of four primary pathways extracted from LE transport data (instead of anomaly data) are delineated in Fig. S1. The northern Europe pathway (Fig. 2a) is featured by a higher southerly LE flux moving along the coast of northern Europe into the Arctic (nodes 1, 2, 5 and 6 in Fig. 1). Four typical patterns on the lower left corner of SOM (nodes 9, 10, 13 and 14), which are characterized by anomalous high LE transport through the northern Labrador Sea, are grouped as the Davis Strait pathway. Besides, the Greenland Sea pathway consists of four nodes (3, 4, 7, and 8) in the upper right of the SOM. The remaining nodes (11, 12, 15, and 16) are defined as the Bering Strait pathway. Manual grouping leads to more straightforward and physically meaningful clusters, which distinguish specific characteristics of winter LE transport patterns. These primary pathways highlight intense LE transport into the high Arctic during winter through northern Europe, the Bering Strait, the Davis Strait, and the Greenland Sea (Fig. S2a), respectively, which are consistent with the main entry channels in the Labrador Sea, the North Atlantic, and the North Pacific along which moisture enters the Arctic (Jakobson and Vihma 2010; Dufour et al. 2016; Naakka et al. 2019).
The four primary pathways by grouping similar nodes in the 4 × 4 SOM map shown in Fig. 1: (a) the northern Europe, (b) the Davis Strait (c) the Greenland Sea, and (d) the Bering Strait pathways.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
The four primary pathways by grouping similar nodes in the 4 × 4 SOM map shown in Fig. 1: (a) the northern Europe, (b) the Davis Strait (c) the Greenland Sea, and (d) the Bering Strait pathways.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
The four primary pathways by grouping similar nodes in the 4 × 4 SOM map shown in Fig. 1: (a) the northern Europe, (b) the Davis Strait (c) the Greenland Sea, and (d) the Bering Strait pathways.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Figure 3 shows the total number of days when each primary LE transport pattern occurred (Fig. 2), in all the winters during 1979–2021. Overall, the four primary pathways occur at a comparable frequency, with a fraction of 25.6% (the northern Europe pathway), 24.2% (the Davis Strait pathway), 24.4% (the Greenland Sea pathway), and 25.7% (the Bering Strait pathway), respectively. Noticeable interannual fluctuations can be observed in each time series of the frequency for these pathways, especially for the Davis Strait and the Bering Strait pathways (Fig. 3), with greater standard deviations (13.1 and 14.5 days). It is shown that, except for the Greenland Sea pathway, the frequencies of these primary pathways are closely related to the large-scale atmospheric circulation indices. A close connection (R = +0.55, P < 0.01) exists between the frequency of occurrence of the northern Europe pathway and the NAO index for the 43 years, indicating that the occurrence of this pathway is modulated by the NAO. In the positive phase of NAO, a strong low pressure system is centered over Iceland while a subtropical high is located over the Azores islands in the central North Atlantic. The anomalously deep Iceland low promotes increased LE through this pathway. For the Davis Strait pathway, we identify that its frequency is highly correlated with the AO index (R = −0.81, P < 0.01). That is to say, the negative phase of AO favors the Davis Strait pathway to occur. This can be a result of the frequently developed or persistent Greenland block pattern and a low over Hudson Bay, which create a large pressure gradient in Baffin Bay. Besides, a clear connection also emerged between the incidence of the Bering Strait pathway and the DA index with a correlation coefficient of +0.59 (P < 0.01). We argue that the meridional winds associated with DA+ facilitate LE transport via the Bering Strait into the Arctic.
Time series of the frequency of occurrence (black solid lines) for each winter from 1979 to 2021 of the primary LE transport pathways shown in Fig. 2 and atmospheric circulation indices (red solid lines). Frequency is defined as the number of days with the occurrence of each pattern of total winter days. Dotted lines indicate the average value during the 43 years and dashed lines denote the time series of the 10-yr running average. The correlation coefficients between the occurrence frequency of the primary LE transport pathways and atmospheric circulation indices shown in each panel are statistically significant at the 99% confidence level.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Time series of the frequency of occurrence (black solid lines) for each winter from 1979 to 2021 of the primary LE transport pathways shown in Fig. 2 and atmospheric circulation indices (red solid lines). Frequency is defined as the number of days with the occurrence of each pattern of total winter days. Dotted lines indicate the average value during the 43 years and dashed lines denote the time series of the 10-yr running average. The correlation coefficients between the occurrence frequency of the primary LE transport pathways and atmospheric circulation indices shown in each panel are statistically significant at the 99% confidence level.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Time series of the frequency of occurrence (black solid lines) for each winter from 1979 to 2021 of the primary LE transport pathways shown in Fig. 2 and atmospheric circulation indices (red solid lines). Frequency is defined as the number of days with the occurrence of each pattern of total winter days. Dotted lines indicate the average value during the 43 years and dashed lines denote the time series of the 10-yr running average. The correlation coefficients between the occurrence frequency of the primary LE transport pathways and atmospheric circulation indices shown in each panel are statistically significant at the 99% confidence level.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Also, lower-frequency variations (3–5 years or decadal signals) can be captured in each time series of the frequency for these pathways. Statistical results indicate that the 10-yr running-averaged frequencies and corresponding averaged atmospheric indices are highly mutually related compared to the original series. In the recent decade (2011–21) with a scenario of a relatively positive DA (+0.11) and NAO (+0.63), the Bering Strait and northern Europe pathways occurred at relatively high frequencies (28.4% and 27.7%) while the other two pathways occurred less frequently compared with the last decade (2001–10) (the long-dashed line shown in Fig. 3 and Table S1). Thus, the large interannual and decadal variations in the frequencies of these primary pathways can be partly attributed to the variations in local atmospheric circulation during winter. Other modes may also make a nonnegligible contribution. For instance, the Atlantic multidecadal oscillation (AMO) and the Pacific decadal oscillation (PDO) were suggested to be major internal drivers of Arctic sea ice variability (Zhang 2015; Screen and Francis 2016). Additionally, more recent studies have revealed that internal sea surface temperature (SST) variability residing in the tropical Pacific can also have a substantial impact on the Arctic climate through teleconnections (Ding et al. 2014; Baxter et al. 2019; Screen and Deser 2019). Enlightened by these studies, we checked the influence of these processes away from the Arctic by constructing the regression maps of SST, and geopotential on 925 and 300 hPa based on the time series of the occurrence for each primary LE transport pathway. There do exist significant associations with the midlatitude Atlantic or Pacific, which may result from their proximity to the Arctic. For instance, the occurrence of the Greenland Sea pathway is related to lower geopotential in the middle Atlantic extending to the middle of Eurasia along with higher geopotential near northern Europe. This is evident throughout the troposphere. Moreover, the Bering Strait (Greenland Sea) pathway is linked with increasing (decreasing) SST at the eastern Pacific, which appears to be associated with ENSO (figures not shown). However, these teleconnections with the tropical Pacific are not statistically significant. The specific mechanism indeed requires more comprehensive analysis and should be explored in greater detail in future studies.
b. Relationships with cyclone activities
In this section, the distinct spatial distributions of wintertime cyclone activities corresponding to different LE pathways during the period 1979–2021 are evaluated (Fig. 4). On this basis, we can assess which LE pathway is mostly impacted by cyclone activities. Considering the cyclones represent synoptic events that are consecutive in time, we first composite the cyclone features of four time steps in all days when each pathway occurred during the winters from 1979 to 2021. Then cyclone features under distinct LE pathways are normalized by the number of occurrence days over 43 winters due to their diverse frequencies and multiplied by the total days of winter (90 days). The anomaly fields are referred to as the difference between the normalized cyclone features and the corresponding climatology during 1979–2021. In addition, the spatial characteristics of cyclone features were calculated in a particular region over a period of time. Therefore, the significance of the anomalies is tested by their magnitudes, other than the Monte Carlo approach.
Anomalies of the cyclone features relative to the 1979–2021 climatology associated with each LE transport pathway during the winter (January–March), including the (a)–(d) density of cyclone tracks [(106 km2)−1] and (e)–(h) CAI [103 hPa (106 km2)−1]. (c),(d) Stippling represents the values where the anomaly exceeds 1.5 standard deviations.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Anomalies of the cyclone features relative to the 1979–2021 climatology associated with each LE transport pathway during the winter (January–March), including the (a)–(d) density of cyclone tracks [(106 km2)−1] and (e)–(h) CAI [103 hPa (106 km2)−1]. (c),(d) Stippling represents the values where the anomaly exceeds 1.5 standard deviations.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Anomalies of the cyclone features relative to the 1979–2021 climatology associated with each LE transport pathway during the winter (January–March), including the (a)–(d) density of cyclone tracks [(106 km2)−1] and (e)–(h) CAI [103 hPa (106 km2)−1]. (c),(d) Stippling represents the values where the anomaly exceeds 1.5 standard deviations.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
When anomalous high LE flux moved along the coast of northern Europe and then deflected northward toward the high Arctic through the Norwegian and Barents Seas, there tended to be more and intense cyclones occurring over the Norwegian Sea (Figs. 4a,e). In these regions, the track density and CAI are higher than the climatological mean by 1–2 standard deviations. More and stronger cyclones detected in the Labrador Sea as well as the Baffin Bay (Figs. 4b,f) are likely to favor a sizeable LE flux through the Davis Strait. Likewise, wintertime days when LE transport patterns are mapped onto the Greenland Sea pathway are associated with numerous cyclones with high intensity extended from the Greenland Sea to the central Arctic (Figs. 4c,g). For the Bering Strait pathway, powerful cyclones occurred more frequently in the North Pacific, along with an increase in cyclone tracks and intensity over the Barents and Kara Seas (Figs. 4d,h). Enhanced cyclone frequency and intensity in these regions may favor poleward LE flux from Siberia (nodes 11 and 15 in Figs. 1 and 2d), although these LE transports are more zonal. However, those over the Barents and Kara Seas mostly contributed to eastward transport (Fig. 2d). As expected, considerable anomalies of cyclone features predominantly emerge in the North Atlantic and the Arctic marginal seas of this sector, which emphasized the dominance of the North Atlantic on cyclones carrying energy from midlatitudes toward the Arctic. It can explain partly why these identified primary pathways, except for the Bering Strait pathway, are close to the North Atlantic.
To clarify the regulation of LE transport into the Arctic by cyclones more insightfully, we analyzed the relationship between the meridional component of LE flux and the cyclone characteristics at the boundary of the Arctic (60°N). Longitudinal distributions of the climatological vertically integrated total (northward) LE across 60°N, as well as the intensity of cyclones nearby, averaged over January–March during 1979–2021, are illustrated in Fig. 5. Note that extratropical cyclones have a poleward migration tendency, thus we also estimated the northward LE transport by ignoring equatorward flux to specifically assess the impact of the cyclones on the northward LE flux across the Arctic boundary. These four primary pathways have prominent meridional LE into the Arctic through the main entry channels [section 2b(1)]. For northern Europe and the Greenland Sea pathways, the spatial distribution of cyclone intensity is in good agreement with the meridional LE flux, especially for the northward transport. On a grid-by-grid basis, a strong correlation exists between the averaged intensity and the poleward LE across the Arctic boundary with R = +0.73 and +0.62 (361 grids, P < 0.01) for the northern Europe and Greenland Sea pathway, respectively. This high consistency suggests a key role of cyclones in carrying water vapor and the associated latent heat poleward. We argued that this could be a result of more frequent and intense cyclones in these regions (Figs. 3 and 4). For instance, compared with those identified in other pathways, cyclones that occurred under the northern Europe pathway are most intense in the region south of Iceland, with a central pressure ∼30 hPa lower than the monthly climatological average (Fig. 4a). In the days with LE transport through the Bering Strait, cyclones that occurred in the North Pacific had higher intensity than others, reaching up to −22 hPa (Fig. 4d). By contrast, the longitudinal pattern of cyclone intensity and the meridional LE flux at 60°N for the Davis Strait are not highly correlated (R = +0.56 and +0.28 for total and poleward LE flux, respectively), although this entry region is characterized by relatively strong cyclones (Figs. 4b and 5b). The relatively weak correlation again highlights that the Davis Strait pathway is mainly controlled by large-scale atmospheric circulation but not the synoptic processes (section 3a). It is noteworthy that for the northern Europe and the Greenland Sea pathways, we observed differing spatial relationships between cyclone intensity and the poleward LE. To be specific, the peaks of longitudinal distributions of LE flux tend to be located east of that of cyclone intensity (Figs. 5a,c). If this spatial offset is taken into account, the correlations for these two pathways exceed +0.80. This can be attributed to the asymmetric patterns in the atmosphere during Arctic cyclones, with cyclonic surface winds, a warm, moist sector to the east of cyclones and the opposite to the west (Clancy et al. 2022). To sum up, the synoptic-scale LE transports toward the high Arctic through the northern Europe and the Greenland Sea pathways seem to be affected significantly by cyclone activities while the Bering Strait and Davis Strait pathways are being modulated to a lesser extent.
Longitudinal distributions of the climatological vertically integrated latent energy flux in the meridional direction across 60°N, as well as the intensity of cyclones at 60°N, averaged over the winter months (January–March) during 1979–2021. Total (northward) LE flux is delineated by the blue (pink) bar while cyclone intensity is plotted by the green line. The number enclosed by the blue (pink) polygon denotes the correlation between the longitudinal intensity and total (northward) LE flux. Note that all these correlations are significant at the 99% confidence level for 361 longitude grids.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Longitudinal distributions of the climatological vertically integrated latent energy flux in the meridional direction across 60°N, as well as the intensity of cyclones at 60°N, averaged over the winter months (January–March) during 1979–2021. Total (northward) LE flux is delineated by the blue (pink) bar while cyclone intensity is plotted by the green line. The number enclosed by the blue (pink) polygon denotes the correlation between the longitudinal intensity and total (northward) LE flux. Note that all these correlations are significant at the 99% confidence level for 361 longitude grids.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Longitudinal distributions of the climatological vertically integrated latent energy flux in the meridional direction across 60°N, as well as the intensity of cyclones at 60°N, averaged over the winter months (January–March) during 1979–2021. Total (northward) LE flux is delineated by the blue (pink) bar while cyclone intensity is plotted by the green line. The number enclosed by the blue (pink) polygon denotes the correlation between the longitudinal intensity and total (northward) LE flux. Note that all these correlations are significant at the 99% confidence level for 361 longitude grids.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Taking all wintertime days together, a more robust connection between the integrated poleward LE transport and the intensity of cyclones nearby arises (Fig. S2), with a correlation coefficient of +0.81 (P < 0.01). This robust connection indicates that about 66% of the variance of the LE transport is explained by changes in cyclone activities, which is not surprising considering the role of cyclones in holding and transporting moisture and energy. A variety of studies have corroborated the vital role that cyclone activities play in modulating poleward atmospheric energy and moisture transport into the Arctic (Jakobson and Vihma 2010; Woods et al. 2013; Dufour et al. 2016; Kim et al. 2017; Villamil-Otero et al. 2018). In particular, Villamil-Otero et al. (2018) found that strong cyclone activity across 60°N, measured by the CAI generally co-occurs with enhanced poleward monthly atmospheric moisture transport in all months. Another case was given by Kim et al. (2017), who argued that a strong Atlantic windstorm into the Arctic in late December 2015 triggered abrupt warming (reaching up to 7–8 K) by bringing enormous moist and warm air masses into the Arctic. Our results augment evidence for the view that the synoptic-scale intrusion of moisture and energy into the Arctic is closely related to the behaviors of extratropical cyclones.
4. Contributions of LE transport to Arctic sea ice loss
a. Regional variability of SIC related to different LE transport pathways
In this section, we continue to examine the contributions of each LE transport pathway to regional SIC loss during winter in the past four decades based on a composite analysis and assess their relative importance. Figure 6 demonstrates the composite map of SIC anomalies in winter during the period 1979–2021. Overall, there tend to exist negative SIC anomalies in regions where poleward LE transport prevailed while the opposite holds in regions with southward LE flux. Regulated by the northern Europe pathway, below-normal SICs were observed in the Greenland Sea and the Barents Sea. These negative SIC anomalies even reach −10% in some regions where the differences are significant at the 99% significance level (tested by a Monte Carlo approach). Meanwhile, the northern Europe pathway appears to be favorable for the sea ice area to expand in the Labrador Sea, an elongated region along the west coast of Greenland (Fig. 6a). The spatial distribution of SIC anomalies averaged over the Greenland Sea pathway days share an analogous pattern with the northern Europe pathway, having broader negative anomalies in the Barents–Kara Seas as a result of larger northward LE flux therein (Fig. 6c). By comparison, the SIC anomalies associated with the Greenland Sea pathway have a broader extent of negative fields expanding further north over the Barents–Kara Seas. Likewise, LE transport to the Arctic through the Davis Strait could induce extensive SIC reduction over Baffin Bay (Fig. 6b). It is estimated that the maximum SIC drop in Baffin Bay reaches up to −14%. In contrast, albeit with the highest frequency (25.71%), the Bering Strait pathway has the weakest ability in diminishing sea ice cover, which merely brings about SIC decline in the Bering Sea (about −3%). In the Atlantic sector, negative SIC anomalies did not spread conspicuously with a localized reduction only detected in the south part of the Barents Sea. This can be attributed to the limited northward LE transport through the North Atlantic Ocean compared with others under this transport scenario (Fig. 2d).
Anomalies of the SIC regulated by various LE transport pathways during the winter months (January–March). Anomalies are relative to the 1979–2021 climatology (%). Green dotted shadings indicate the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Anomalies of the SIC regulated by various LE transport pathways during the winter months (January–March). Anomalies are relative to the 1979–2021 climatology (%). Green dotted shadings indicate the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Anomalies of the SIC regulated by various LE transport pathways during the winter months (January–March). Anomalies are relative to the 1979–2021 climatology (%). Green dotted shadings indicate the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
The abovementioned results of the present study are consistent with Park et al. (2015), who argued that the northward moisture flux into the Arctic during the winter is followed by a decrease in SIC of up to 10% in the Greenland, Barents, and Kara Seas. We list the SIC decline averaged over the LE convergence region for each pathway in Table 1. For these primary pathways, the affected areas with diminished SIC by LE transport are estimated to be 7.99, 10.40, 8.23, and 4.57 × 104 km2, respectively. Utilizing SIC data, the loss of Arctic sea ice only can be identified in the marginal seas because the SIC within the central Arctic Ocean remained at a high level close to 100%. The contribution of atmospheric LE transport toward the Arctic on sea ice cover identified by statistics, as shown in Fig. 6, emphasizes its importance. However, to further quantify the relative contribution of the LE flux to the loss of Arctic sea ice, it is necessary to develop a coupled numerical model of atmosphere, ocean and sea ice. Also note that prominent downward SIC trends are observed in winter, in the Sea of Okhotsk, the Bering Sea, the Baffin Bay, the Greenland Sea, and the Barents–Kara Sea during the period 1979–2021 (figure omitted). If these long-term trends are taken into account, the SIC anomalies related to the decreasing trend for each LE pathway share identical patterns, yielding the affected areas to be 10.53, 16.61, 13.91, and 5.42 × 104 km2. This consistency underscored that the results we revealed are robust.
The regional anomalies in terms of SIC and the surface radiative and turbulent fluxes (including DLR, NLR, LHF, and SHF), as well as the corresponding changes in SIT over the winter months averaged over the LE convergence domain associated with different LE transport pathways.
b. Mechanism of how LE transports impact Arctic sea ice
In winter, pronounced poleward LE transport leads to energy and moisture convergence in the Arctic, which could contribute to the local increases in atmospheric temperature and humidity. The LE transport pathways agree with those of SAT and specific humidity anomalies reasonably well (Figs. 2 and 7). As depicted in Fig. 7b, wintertime transport through the Davis Strait could lead to intense and extensive surface warming (dampening) of the atmosphere in the Baffin Bay and the Labrador Sea. Particularly, the magnitude of the SAT and total specific humidity anomaly even exceeds 5 K and 5 × 10−3 kg kg−1, respectively, in these regions. The second most efficient transport pattern is found to be the Greenland pathway, which prompts a rise of 3–4 K and ∼4 × 10−3 kg kg−1 with a broad area extending from the Greenland Sea to the Barents–Kara Seas (Fig. 7c). Moreover, modulated by the northern Europe LE transport pathway, the regions along northern Europe are characterized by positive anomalies of SAT and water vapor contents of up to 3–4 K and 2.8 × 10−3 kg kg−1 (Fig. 7a), respectively. By comparison, the positive anomalies induced by the intrusion of LE through the Bering Strait are not that remarkable, as shown in Fig. 7d. It is noteworthy that the discernable warming and moistening with relatively small magnitude (1–2 K and 0.8–1 × 10−3 kg kg−1) mostly occur over land except for the Bering Sea and part of the Norwegian Sea. Note that we demonstrate the anomalies of SAT instead of the vertically averaged temperature of the atmospheric column in consideration of the fact that the atmospheric boundary layer plays a more important role in the energy balance over the sea ice surface. Generally, the vertically averaged temperature anomalies revealed no major discrepancy for all the leading pathways. LE transport can increase the temperature and humidity of the local atmosphere, which could enhance the cloud fraction (Johansson et al. 2017). Shupe and Intrieri (2004) argued that the cloud can warm the surface through most of the year except for a short period in the middle of summer. And in all seasons, liquid-containing clouds dominate both longwave and shortwave radiative impacts on the surface. We, therefore, evaluated the total column cloud liquid water and cloud fraction occurring in the different levels of the troposphere. It is noteworthy that the total column cloud liquid water, the amount of liquid water contained within cloud droplets in an atmospheric column, exhibits positive anomalies in the regions where poleward LE flux prevailed for four pathways (figure not shown). This implies that LE transport toward the Arctic in winter may strengthen the local formation of liquid-containing clouds, which be able to induce surface warming. The consistency between spatial patterns of SIC, SAT as well as humidity, supports that anomalously high transport of LE into the Arctic accompanied by the unusually moist and warm atmosphere during winter has a deleterious effect on sea ice mass balance.
Anomalies of (a)–(d) the surface air temperature and (e)–(h) mean specific humidity of the atmosphere column relative to the 1979–2021 climatology regulated by various LE transport pathways during the winter months (January–March). Dots represent the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Anomalies of (a)–(d) the surface air temperature and (e)–(h) mean specific humidity of the atmosphere column relative to the 1979–2021 climatology regulated by various LE transport pathways during the winter months (January–March). Dots represent the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Anomalies of (a)–(d) the surface air temperature and (e)–(h) mean specific humidity of the atmosphere column relative to the 1979–2021 climatology regulated by various LE transport pathways during the winter months (January–March). Dots represent the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
The additional LE flux from the lower latitudes that arrived in the Arctic could exert twofold impacts on surface changes. On one hand, the moisture associated with the LE transport would affect the precipitation in the Arctic. In the LE convergence regions of four primary pathways, the anomalies of total precipitation relative to the 1979–2021 climatology during the winter month are characterized by positive values (Fig. S3). To be specific, LE transport through the entry region in each pathway brings about more precipitation therein. Even though we did not specify the type of precipitation at the surface (e.g., snow or rain), the enhanced snowfall or rainfall both have a deleterious effect on sea ice cover in winter (Fig. S3 and Fig. 6). In winter, snowfall on sea ice enhances thermal insulation and thus reduces sea ice growth. And rainfall is generally related to sea ice melt. Note that enhanced snow accumulation would form a negative freeboard and create snow ice on the upper surface of sea ice, thereby contributing to the mass balance of sea ice. However, this mechanism is not prevalent in the Arctic due to the small snow–sea ice thickness ratio, especially in winter, when seasonal sea ice has grown to a certain thickness.
From the perspective of radiative mechanisms, part of the energy (mainly as longwave radiation in winter) was radiated back to warm the atmosphere while the remainder was radiated to the sea surface and turbulently mixed, slowing down the sea ice growth or even leading to melting in winter. In addition, an increase in humidity and air temperature in the Arctic associated with LE intrusion help to enhance the local greenhouse effect and may strengthen cloud formation (Johansson et al. 2017). These variations together have a critical effect on the surface radiative forcing. Detailed structures of anomalies in surface radiative fluxes in winter connected to the four basic LE transport patterns are depicted in Fig. 8.
Anomalies of (a)–(d) the surface downward longwave radiation (DLR), (e)–(h) net longwave radiation (NLR), as well as (i)–(l) latent heat flux (LHF) and (m)–(p) sensible heat flux (SHF) relative to the 1979–2021 climatology under various LE transport pathways during the winter months (January–March) (W m−2). Note that the surface radiative fluxes are positive downward. Dots represent the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Anomalies of (a)–(d) the surface downward longwave radiation (DLR), (e)–(h) net longwave radiation (NLR), as well as (i)–(l) latent heat flux (LHF) and (m)–(p) sensible heat flux (SHF) relative to the 1979–2021 climatology under various LE transport pathways during the winter months (January–March) (W m−2). Note that the surface radiative fluxes are positive downward. Dots represent the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Anomalies of (a)–(d) the surface downward longwave radiation (DLR), (e)–(h) net longwave radiation (NLR), as well as (i)–(l) latent heat flux (LHF) and (m)–(p) sensible heat flux (SHF) relative to the 1979–2021 climatology under various LE transport pathways during the winter months (January–March) (W m−2). Note that the surface radiative fluxes are positive downward. Dots represent the anomalies that are different from zero at the 99% significance level, tested by a Monte Carlo approach.
Citation: Journal of Climate 36, 19; 10.1175/JCLI-D-22-0789.1
Obviously, an anomalously high LE transport toward the Arctic could trigger a substantial increase in surface longwave radiation (DLR) (Figs. 8a–d). As for the Davis Strait and the Greenland Sea pathways, positive and statistically significant anomalies of DLR spread throughout the convergence zones, with the largest amplitudes observed in the Baffin Bay and Greenland–Barents Seas (∼33 and ∼32 W m−2, Figs. 8b,c, respectively). However, the increase in DLR was less pronounced under the northern Europe pathway, featured by a gain of ∼16 W m−2 in the Barents–Kara Seas. It is not unexpected that the Bering Strait pathway brings about the least positive DLR anomalies. The anomalies of surface net longwave radiation (NLR) exhibit identical patterns with those of the DLR for all pathways, but with a smaller magnitude (Figs. 8e–h). During winter, turbulent heat exchange is limited as a result of a stable atmospheric boundary layer (Serreze et al. 2007). Nevertheless, remarkably excessive sensible and latent turbulent surface flux to the sea surface were noted in the regions with LE convergence during winter, especially for the Davis Strait and the Greenland pathways. These positive anomalies in turbulent fluxes contribute to the energy surplus at the surface as well, and can be attributed to the active interaction between the atmosphere and ice/sea surface through condensation and conductive processes. The modes of variability for latent and sensible heat flux under four distinct LE pathways are almost the same, with the largest positive anomalies identified in the LE convergence regions under the Davis Strait and the Greenland Sea pathways.
5. Discussion and conclusions
In this study, the SOM method was utilized to classify the vertically integrated daily LE transport patterns north of 50°N in winter (January–March) over the observational period (1979–2021). We retrieved a 4 × 4 map to characterize fundamental variabilities in horizontal LE transport toward the Arctic. Then the similar nodes are further grouped together to obtain more straightforward and physically meaningful clusters. A novel point distinguishing this study from previous studies is that we provided a straightforward picture of the horizontal LE transport at the synoptic time scale with separated LE flux through distinct gateways. Four general LE pathways, namely, the northern Europe, the Davis Strait, the Greenland Sea, and the Bering Strait pathways, were identified, with poleward LE toward the Arctic through the entry regions suggested by their names. The frequency of occurrence of these pathways is nearly equivalent, which accounts for 25.63%, 24.21%, 24.44%, and 25.71% of the total days during winter months, respectively. Noticeable interannual fluctuations and decadal variability were observed in each time series of the frequency for these pathways. Except for the Greenland Sea pathway, the interannual variances in the frequencies of occurrence of these primary pathways correlate highly with the large-scale atmospheric circulation indices. The relationship is even more robust in terms of decadal variations, indicating that the large interannual and decadal variations of frequencies of the primary pathways can be attributed, at least partially, to the changes in the energetic wintertime large-scale atmospheric circulation.
Moreover, we present the spatial patterns of cyclone features connected to the primary LE pathways. A clear relationship is distinguished, that is, more frequent and intense extratropical cyclones tend to appear near the entry regions for each pathway. High agreement exists between the wintertime longitudinal distributions of the climatological northward LE across 60°N and the intensity of cyclones nearby during 1979–2021. The correlations between these two quantities during 1979–2021 are no less than +0.54 for all the primary pathways (statistically significant at the 99% confidence level). Stronger connections are observed for the northern Europe and the Greenland Sea pathways, with R = +0.73 and +0.62, respectively, but weaker for the Davis Strait and Bering Strait pathways. Statistical results imply that synoptic-scale LE transport through the northern Europe and the Greenland Sea (Davis Strait and Bering Strait) pathways seem to be more (less) strongly connected with the behaviors of the passing cyclone. It is also noteworthy that the distributions of cyclones under the Greenland Sea and the northern Europe pathways are basically congruent with the western and eastern cyclone clusters demonstrated in Kenigson and Timmermans (2021), which represent two branches along which Nordic seas cyclones migrate. According to climate model simulations, a poleward deflection of the Atlantic storm tracks is generally expected under the background of global warming (Yin 2005; Chang et al. 2012). We also detected positive trends in the number of cyclone tracks and intensity in the North Atlantic during the recent four decades (1979–2021) based on reanalysis data (not shown). Intuitively, more numerous and intense North Atlantic storms reaching the Arctic would lead to an increasing frequency of these two pathways. However, neither frequency of these two pathways during the 43-yr period exhibits significant long-term tendency. How the changes in the Nordic cyclones impact the trends of different synoptic-scale LE transport pathways warrants detailed evaluations in future work.
Previous studies have elucidated that poleward atmospheric heat and moisture transport can affect Arctic sea ice change. Overall, a decline in SIC is detected in entry regions where wintertime poleward LE transport prevailed under four primary LE pathways. For instance, a below-normal sea ice coverage was observed in Baffin Bay when the LE transport via the Davis Strait pathway dominated. The SIC anomalies averaged over the LE convergence regions are estimated as −3.66%, −5.13%, −3.24%, and −2.19% for each LE pathway, respectively. We argued that the LE transports cause a notable sea ice shrinkage, which can be attributed to the locally enhanced greenhouse effect caused by the energy and moisture convergence. Besides, the increased amount of precipitation associated with LE transport also made a contribution. Further analysis showed that a rise in the surface air temperature (5 K) and vertically averaged specific humidity (5 × 10−3 kg kg−1) coincide with the significant sea ice loss. Indeed, the local SAT and humidity increases are accompanied by a regional increase in DLR of ∼33 W m−2. Consistent with our results, Park et al. (2015) showed that the northward flux of moisture into the Arctic during the winter (December–early March) strengthens the DLR by 30–40 W m−2 over 1–2 weeks, which is further confirmed by a moisture intrusion event occurred in 9–11 February 2010 (Doyle et al. 2011). Moreover, additional sensible and latent turbulent surface fluxes in these regions also contribute to the energy budget surplus although these anomalies are less extensive compared with those of DLR. For the Davis Strait and the Greenland Sea pathways, enhanced surface turbulent fluxes to surface even predominate over DLR, reaching up to about 27 W m−2 in the LE convergence regions. Based on a simplified sea ice growth model, there is an approximate 0.27 m (northern Europe pathway), 0.86 m (Davis Strait pathway), 0.56 m (Greenland Sea pathway), and 0.13 m (Bering Strait pathway) reduction of SIT caused by the surface energy budget surplus related to LE transport. The SIT changes combined with SIC anomalies both point to the fact that the LE transports through the Davis Strait and the Greenland Sea pathways could cause a notable sea ice loss particularly effectively, followed by the northern Europe and the Bering Sea pathways.
We address the impact of LE transport on sea ice based on detailed and quantitative analysis through a thermodynamic perspective in this study. One should bear in mind that thermodynamic processes are not applicable alone but mutually coupled with dynamic ones. On one hand, sea ice surface melting induced by thermodynamic processes weakens the ice solidity and promotes breakup. On the other hand, less consolidated sea ice is more susceptible to extreme weather events which bring about strong winds, thereby leading to ice drift and fragmentation. Besides, the reduction of sea ice in the Arctic marginal seas cannot be entirely attributed to the poleward atmospheric LE transport. Other factors, including the drift and breaking of sea ice caused by storms (Brümmer et al. 2008; Kriegsmann and Brümmer 2014; Lei et al. 2020), the poleward transport of ocean heat (Spielhagen et al. 2011; Årthun et al. 2012), especially in the Atlantic sector, the southward sea ice advection from the central Arctic Ocean, especially through the Fram Strait (Kwok et al. 2004; Bi et al. 2016; Smedsrud et al. 2017; Wei et al. 2019), and the preconditions of sea ice of the previous melt season (Markus et al. 2009; Parkinson 2014; Stroeve et al. 2014), also play a crucial role.
In the present study, we focus on vertically integrated horizontal atmospheric LE transport throughout wintertime days. Nevertheless, a disproportionally large fraction of the meridional moisture and energy transport occurs in discrete events, referred to as atmospheric rivers, which could trigger wintertime events with melting temperatures. These episodic, extreme water vapor transports play an important role in modulating Arctic hydroclimate and sea ice cover (Nash et al. 2018; Zhang et al. 2023), and can be captured reasonably by reanalysis. (Guan et al. 2018; Cobb et al. 2021). We have observed younger (Rigor and Wallace 2004; Tschudi et al. 2016) and thinner (Kwok and Rothrock 2009; Bi et al. 2018) trends of sea ice in the Arctic in recent decades. How much do atmospheric rivers contribute to these trends? To answer this question, climate models of higher fidelity combined with more sophisticated analysis methods are needed. Also, in situ and remote sensing observations would complement reanalysis, which provides opportunities for follow-up research. Moreover, the regulation of synoptic activities (cyclones) on atmospheric LE was highlighted in the present study. Nevertheless, it is shown that the frequencies of occurrence of primary pathways correlate highly with the diverse large-scale atmospheric circulations. Therefore, different LE transport patterns occurring at frequencies with large interannual variations may be related to changes in large-scale atmospheric circulations in winter or teleconnections from lower latitudes. In the follow-up research, it is necessary to conduct more in-depth and detailed research to solve these issues.
Acknowledgments.
This work was financially supported by Laoshan Laboratory (LSKJ202202303), the National Natural Science Foundation of China (52192691 and 41976219), the General Project of Natural Science Foundation of Shandong Province (ZR2020MD100), and the Program of Shanghai Academic/Technology Research Leader (22XD1403600).
Data availability statement.
NSIDC sea ice concentration data are obtained from https://nsidc.org/data/NSIDC-0079/versions/3 as cited in Comiso (2017). The ERA5 datasets are available at the website https://cds.climate.copernicus.eu/cdsapp#!/search?type=dataset (Hersbach et al. 2020). Large-scale atmospheric indexes, including AO and NAO were downloaded from https://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/teleconnections.shtml.
REFERENCES
Akperov, M., and Coauthors, 2019a: Trends of intense cyclone activity in the Arctic from reanalyses data and regional climate models (Arctic-CORDEX). IOP Conf. Ser., 231, 012003, https://doi.org/10.1088/1755-1315/231/1/012003.
Akperov, M., and Coauthors, 2019b: Future projections of cyclone activity in the Arctic for the 21st century from regional climate models (Arctic-CORDEX). Global Planet. Change, 182, 103005, https://doi.org/10.1016/j.gloplacha.2019.103005.
Årthun, M., T. Eldevik, L. H. Smedsrud, Ø. Skagseth, and R. B. Ingvaldsen, 2012: Quantifying the influence of Atlantic heat on Barents Sea ice variability and retreat. J. Climate, 25, 4736–4743, https://doi.org/10.1175/JCLI-D-11-00466.1.
Bader, J., M. D. S. Mesquita, K. I. Hodges, N. Keenlyside, S. Østerhus, and M. Miles, 2011: A review on Northern Hemisphere sea-ice, storminess and the North Atlantic Oscillation: Observations and projected changes. Atmos. Res., 101, 809–834, https://doi.org/10.1016/j.atmosres.2011.04.007.
Baxter, I., and Coauthors, 2019: How tropical Pacific surface cooling contributed to accelerated sea ice melt from 2007 to 2012 as ice is thinned by anthropogenic forcing. J. Climate, 32, 8583–8602, https://doi.org/10.1175/JCLI-D-18-0783.1.
Bi, H., K. Sun, X. Zhou, H. Huang, and X. Xu, 2016: Arctic Sea ice area export through the Fram Strait estimated from satellite-based data: 1988–2012. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 9, 3144–3157, https://doi.org/10.1109/JSTARS.2016.2584539.
Bi, H., J. Zhang, Y. Wang, Z. Zhang, Y. Zhang, M. Fu, H. Huang, and X. Xu, 2018: Arctic Sea ice volume changes in terms of age as revealed from satellite observations. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 11, 2223–2237, https://doi.org/10.1109/JSTARS.2018.2823735.
Brümmer, B., D. Schröder, G. Müller, G. Spreen, A. Jahnke-Bornemann, and J. Launiainen, 2008: Impact of a Fram Strait cyclone on ice edge, drift, divergence, and concentration: Possibilities and limits of an observational analysis. J. Geophys. Res., 113, C12003, https://doi.org/10.1029/2007JC004149.
Burgard, C., and D. Notz, 2017: Drivers of Arctic Ocean warming in CMIP5 models. Geophys. Res. Lett., 44, 4263–4271, https://doi.org/10.1002/2016GL072342.
Chang, E. K. M., Y. Guo, and X. Xia, 2012: CMIP5 multimodel ensemble projection of storm track change under global warming. J. Geophys. Res., 117, D23118, https://doi.org/10.1029/2012JD018578.
Clancy, R., C. M. Bitz, E. Blanchard-Wrigglesworth, M. C. McGraw, and S. M. Cavallo, 2022: A cyclone-centered perspective on the drivers of asymmetric patterns in the atmosphere and sea ice during Arctic cyclones. J. Climate, 35, 1–47, https://doi.org/10.1175/JCLI-D-21-0093.1.
Cobb, A., L. Delle Monache, F. Cannon, and F. M. Ralph, 2021: Representation of dropsonde-observed atmospheric river conditions in reanalyses. Geophys. Res. Lett., 48, e2021GL093357, https://doi.org/10.1029/2021GL093357.
Comiso, J. C., 2017: Bootstrap sea ice concentrations from Nimbus-7 SMMR and DMSP SSM/I-SSMIS, version 3. NASA National Snow and Ice Data Center Distributed Active Archive Center, accessed 16 May 2022, https://doi.org/10.5067/7Q8HCCWS4I0R.
Ding, Q., J. M. Wallace, D. S. Battisti, E. J. Steig, A. J. E. Gallant, H.-J. Kim, and L. Geng, 2014: Tropical forcing of the recent rapid Arctic warming in northeastern Canada and Greenland. Nature, 509, 209–212, https://doi.org/10.1038/nature13260.
Doyle, J. G., G. Lesins, C. P. Thackray, C. Perro, G. J. Nott, T. J. Duck, R. Damoah, and J. R. Drummond, 2011: Water vapor intrusions into the high Arctic during winter. Geophys. Res. Lett., 38, L12806, https://doi.org/10.1029/2011GL047493.
Dufour, A., O. Zolina, and S. K. Gulev, 2016: Atmospheric moisture transport to the Arctic: Assessment of reanalyses and analysis of transport components. J. Climate, 29, 5061–5081, https://doi.org/10.1175/JCLI-D-15-0559.1.
Fearon, M. G., J. D. Doyle, D. R. Ryglicki, P. M. Finocchio, and M. Sprenger, 2020: The role of cyclones in moisture transport into the Arctic. Geophys. Res. Lett., 48, e2020GL090353, https://doi.org/10.1029/2020GL090353.
Gimeno-Sotelo, L., R. Nieto, M. Vázquez, and L. Gimeno, 2019: The role of moisture transport for precipitation in the inter-annual and inter-daily fluctuations of the Arctic sea ice extension. Earth Syst. Dyn., 10, 121–133, https://doi.org/10.5194/esd-10-121-2019.
Gong, T., S. Feldstein, and S. Lee, 2017: The role of downward infrared radiation in the recent Arctic winter warming trend. J. Climate, 30, 4937–4949, https://doi.org/10.1175/JCLI-D-16-0180.1.
Graversen, R. G., and M. Burtu, 2016: Arctic amplification enhanced by latent energy transport of atmospheric planetary waves. Quart. J. Roy. Meteor. Soc., 142, 2046–2054, https://doi.org/10.1002/qj.2802.
Graversen, R. G., T. Mauritsen, M. Tjernström, E. Källén, and G. Svensson, 2008: Vertical structure of recent Arctic warming. Nature, 451, 53–56, https://doi.org/10.1038/nature06502.
Graversen, R. G., T. Mauritsen, S. Drijfhout, M. Tjernström, and S. Mårtensson, 2011: Warm winds from the Pacific caused extensive Arctic sea-ice melt in summer 2007. Climate Dyn., 36, 2103–2112, https://doi.org/10.1007/s00382-010-0809-z.
Guan, B., D. E. Waliser, and F. M. Ralph, 2018: An intercomparison between reanalysis and dropsonde observations of the total water vapor transport in individual atmospheric rivers. J. Hydrometeor., 19, 321–337, https://doi.org/10.1175/JHM-D-17-0114.1.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Higgins, M. E., and J. J. Cassano, 2009: Impacts of reduced sea ice on winter Arctic atmospheric circulation, precipitation, and temperature. J. Geophys. Res., 114, D16107, https://doi.org/10.1029/2009JD011884.
Horvath, S., J. Stroeve, B. Rajagopalan, and A. Jahn, 2021: Arctic sea ice melt onset favored by an atmospheric pressure pattern reminiscent of the North American-Eurasian Arctic pattern. Climate Dyn., 57, 1771–1787, https://doi.org/10.1007/s00382-021-05776-y.
Hwang, Y.-T., D. M. W. Frierson, and J. E. Kay, 2011: Coupling between Arctic feedbacks and changes in poleward energy transport. Geophys. Res. Lett., 38, L17704, https://doi.org/10.1029/2011GL048546.
Jakobson, E., and T. Vihma, 2010: Atmospheric moisture budget in the Arctic based on the ERA-40 reanalysis. Int. J. Climatol., 30, 2175–2194, https://doi.org/10.1002/joc.2039.
Johansson, E., A. Devasthale, M. Tjernström, A. M. L. Ekman, and T. L’Ecuyer, 2017: Response of the lower troposphere to moisture intrusions into the Arctic. Geophys. Res. Lett., 44, 2527–2536, https://doi.org/10.1002/2017GL072687.
Kapsch, M.-L., R. G. Graversen, and M. Tjernström, 2013: Springtime atmospheric energy transport and the control of Arctic summer sea-ice extent. Nat. Climate Change, 3, 744–748, https://doi.org/10.1038/nclimate1884.
Kenigson, J. S., and M.-L. Timmermans, 2021: Arctic cyclone activity and the Beaufort high. J. Climate, 34, 4119–4127, https://doi.org/10.1175/JCLI-D-20-0771.1.
Kim, B.-M., and Coauthors, 2017: Major cause of unprecedented Arctic warming in January 2016: Critical role of an Atlantic windstorm. Sci. Rep., 7, 40051, https://doi.org/10.1038/srep40051.
Kim, H.-M., and B.-M. Kim, 2017: Relative contributions of atmospheric energy transport and sea ice loss to the recent warm Arctic winter. J. Climate, 30, 7441–7450, https://doi.org/10.1175/JCLI-D-17-0157.1.
Kohonen, T., 1998: The self-organizing map. Neurocomputing, 21 (1–3), 1–6, https://doi.org/10.1016/S0925-2312(98)00030-7.
Kriegsmann, A., and B. Brümmer, 2014: Cyclone impact on sea ice in the central Arctic Ocean: A statistical study. Cryosphere, 8, 303–317, https://doi.org/10.5194/tc-8-303-2014.
Kwok, R., and D. A. Rothrock, 2009: Decline in Arctic sea ice thickness from submarine and ICESat records: 1958–2008. Geophys. Res. Lett., 36, L15501, https://doi.org/10.1029/2009GL039035.
Kwok, R., G. F. Cunningham, and S. S. Pang, 2004: Fram Strait sea ice outflow. J. Geophys. Res., 109, C01009, https://doi.org/10.1029/2003JC001785.
Lei, R., D. Gui, P. Heil, J. K. Hutchings, and M. Ding, 2020: Comparisons of sea ice motion and deformation, and their responses to ice conditions and cyclonic activity in the western Arctic Ocean between two summers. Cold Reg. Sci. Technol., 170, 102925, https://doi.org/10.1016/j.coldregions.2019.102925.
Lynch, A. H., M. C. Serreze, E. N. Cassano, A. D. Crawford, and J. Stroeve, 2016: Linkages between Arctic summer circulation regimes and regional sea ice anomalies. J. Geophys. Res. Atmos., 121, 7868–7880, https://doi.org/10.1002/2016JD025164.
Maksimovich, E., and T. Vihma, 2012: The effect of surface heat fluxes on interannual variability in the spring onset of snow melt in the central Arctic Ocean. J. Geophys. Res., 117, C07012, https://doi.org/10.1029/2011JC007220.
Markus, T., J. C. Stroeve, and J. Miller, 2009: Recent changes in Arctic sea ice melt onset, freezeup, and melt season length. J. Geophys. Res., 114, C12024, https://doi.org/10.1029/2009JC005436.
Mattingly, K. S., C. A. Ramseyer, J. J. Rosen, T. L. Mote, and R. Muthyala, 2016: Increasing water vapor transport to the Greenland ice sheet revealed using self-organizing maps. Geophys. Res. Lett., 43, 9250–9258, https://doi.org/10.1002/2016GL070424.
Mayer, J., M. Mayer, and L. Haimberger, 2021: Consistency and homogeneity of atmospheric energy, moisture, and mass budgets in ERA5. J. Climate, 34, 3955–3974, https://doi.org/10.1175/JCLI-D-20-0676.1.
Mewes, D., and C. Jacobi, 2019: Heat transport pathways into the Arctic and their connections to surface air temperatures. Atmos. Chem. Phys., 19, 3927–3937, https://doi.org/10.5194/acp-19-3927-2019.
Mortin, J., G. Svensson, R. G. Graversen, M. L. Kapsch, J. C. Stroeve, and L. N. Boisvert, 2016: Melt onset over Arctic sea ice controlled by atmospheric moisture transport. Geophys. Res. Lett., 43, 6636–6642, https://doi.org/10.1002/2016GL069330.
Naakka, T., T. Nygård, T. Vihma, J. Sedlar, and R. Graversen, 2019: Atmospheric moisture transport between mid‐latitudes and the Arctic: Regional, seasonal and vertical distributions. Int. J. Climatol., 39, 2862–2879, https://doi.org/10.1002/joc.5988.
Nash, D., D. Waliser, B. Guan, H. Ye, and F. M. Ralph, 2018: The role of atmospheric rivers in extratropical and polar hydroclimate. J. Geophys. Res. Atmos., 123, 6804–6821, https://doi.org/10.1029/2017JD028130.
Neu, U., and Coauthors, 2013: IMILAST: A community effort to intercompare extratropical cyclone detection and tracking algorithms. Bull. Amer. Meteor. Soc., 94, 529–547, https://doi.org/10.1175/BAMS-D-11-00154.1.
Nygård, T., T. Naakka, and T. Vihma, 2020: Horizontal moisture transport dominates the regional moistening patterns in the Arctic. J. Climate, 33, 6793–6807, https://doi.org/10.1175/JCLI-D-19-0891.1.
Olonscheck, D., T. Mauritsen, and D. Notz, 2019: Arctic sea-ice variability is primarily driven by atmospheric temperature fluctuations. Nat. Geosci., 12, 430–434, https://doi.org/10.1038/s41561-019-0363-1.
Oort, A. H., and J. P. Peixóto, 1983: Global angular momentum and energy balance requirements from observations. Advances in Geophysics, Vol. 25, Academic Press, 355–490, https://doi.org/10.1016/S0065-2687(08)60177-6.
Park, H.-S., S. Lee, S.-W. Son, S. B. Feldstein, and Y. Kosaka, 2015: The impact of poleward moisture and sensible heat flux on Arctic winter sea ice variability. J. Climate, 28, 5030–5040, https://doi.org/10.1175/JCLI-D-15-0074.1.
Parkinson, C. L., 2014: Spatially mapped reductions in the length of the Arctic Sea ice season. Geophys. Res. Lett., 41, 4316–4322, https://doi.org/10.1002/2014GL060434.
Parkinson, C. L., and W. M. Washington, 1979: A large-scale numerical model of sea ice. J. Geophys. Res., 84, 311–337, https://doi.org/10.1029/JC084iC01p00311.
Perovich, D., M. Smith, B. Light, and M. Webster, 2021: Meltwater sources and sinks for multiyear Arctic sea ice in summer. Cryosphere, 15, 4517–4525, https://doi.org/10.5194/tc-15-4517-2021.
Reusch, D. B., R. B. Alley, and B. C. Hewitson, 2005: Relative performance of self-organizing maps and principal component analysis in pattern extraction from synthetic climatological data. Polar Geogr., 29, 188–212, https://doi.org/10.1080/789610199.
Rigor, I. G., and J. M. Wallace, 2004: Variations in the age of Arctic sea‐ice and summer sea‐ice extent. Geophys. Res. Lett., 31, L09401, https://doi.org/10.1029/2004GL019492.
Rinke, A., M. Maturilli, R. M. Graham, H. Matthes, D. Handorf, L. Cohen, S. R. Hudson, and J. C. Moore, 2017: Extreme cyclone events in the Arctic: Wintertime variability and trends. Environ. Res. Lett., 12, 094006, https://doi.org/10.1088/1748-9326/aa7def.
Screen, J. A., and J. A. Francis, 2016: Contribution of sea-ice loss to Arctic amplification is regulated by Pacific Ocean decadal variability. Nat. Climate Change, 6, 856–860, https://doi.org/10.1038/nclimate3011.
Screen, J. A., and C. Deser, 2019: Pacific Ocean variability influences the time of emergence of a seasonally ice-free Arctic Ocean. Geophys. Res. Lett., 46, 2222–2231, https://doi.org/10.1029/2018GL081393.
Sepp, M., and J. Jaagus, 2011: Changes in the activity and tracks of Arctic cyclones. Climatic Change, 105, 577–595, https://doi.org/10.1007/s10584-010-9893-7.
Serreze, M. C., 1995: Climatological aspects of cyclone development and decay in the Arctic. Atmos.–Ocean, 33, 1–23, https://doi.org/10.1080/07055900.1995.9649522.
Serreze, M. C., J. E. Box, R. G. Barry, and J. E. Walsh, 1993: Characteristics of Arctic synoptic activity, 1952–1989. Meteor. Atmos. Phys., 51, 147–164, https://doi.org/10.1007/BF01030491.
Serreze, M. C., A. P. Barrett, A. G. Slater, M. Steele, J. Zhang, and K. E. Trenberth, 2007: The large-scale energy budget of the Arctic. J. Geophys. Res., 112, D11122, https://doi.org/10.1029/2006JD008230.
Shupe, M. D., and J. M. Intrieri, 2004: Cloud radiative forcing of the Arctic surface: The influence of cloud properties, surface albedo, and solar zenith angle. J. Climate, 17, 616–628, https://doi.org/10.1175/1520-0442(2004)017<0616:CRFOTA>2.0.CO;2.
Skific, N., and J. A. Francis, 2013: Drivers of projected change in Arctic moist static energy transport. J. Geophys. Res. Atmos., 118, 2748–2761, https://doi.org/10.1002/jgrd.50292.
Smedsrud, L. H., M. H. Halvorsen, J. C. Stroeve, R. Zhang, and K. Kloster, 2017: Fram Strait sea ice export variability and September Arctic Sea ice extent over the last 80 years. Cryosphere, 11, 65–79, https://doi.org/10.5194/tc-11-65-2017.
Spielhagen, R. F., and Coauthors, 2011: Enhanced modern heat transfer to the Arctic by warm Atlantic water. Science, 331, 450–453, https://doi.org/10.1126/science.1197397.
Stroeve, J. C., T. Markus, L. Boisvert, J. Miller, and A. Barrett, 2014: Changes in Arctic melt season and implications for sea ice loss. Geophys. Res. Lett., 41, 1216–1225, https://doi.org/10.1002/2013GL058951.
Thompson, D. W. J., and J. M. Wallace, 1998: The Arctic oscillation signature in the wintertime geopotential height and temperature fields. Geophys. Res. Lett., 25, 1297–1300, https://doi.org/10.1029/98GL00950.
Tschudi, M. A., J. C. Stroeve, and J. S. Stewart, 2016: Relating the age of Arctic sea ice to its thickness, as measured during NASA’s ICESat and IceBridge campaigns. Remote Sens., 8, 457, https://doi.org/10.3390/rs8060457.
Ulbrich, U., G. C. Leckebusch, and J. G. Pinto, 2009: Extra-tropical cyclones in the present and future climate: A review. Theor. Appl. Climatol., 96, 117–131, https://doi.org/10.1007/s00704-008-0083-8.
Valkonen, E., J. Cassano, and E. Cassano, 2021: Arctic cyclones and their interactions with the declining sea ice: A recent climatology. J. Geophys. Res. Atmos., 126, e2020JD034366, https://doi.org/10.1029/2020JD034366.
Varino, F., P. Arbogast, B. Joly, G. Riviere, M.-L. Fandeur, H. Bovy, and J.-B. Granier, 2019: Northern Hemisphere extratropical winter cyclones variability over the 20th century derived from ERA-20C reanalysis. Climate Dyn., 52, 1027–1048, https://doi.org/10.1007/s00382-018-4176-5.
Vihma, T., and Coauthors, 2016: The atmospheric role in the Arctic water cycle: A review on processes, past and future changes, and their impacts. J. Geophys. Res. Biogeosci., 121, 586–620, https://doi.org/10.1002/2015JG003132.
Villamil-Otero, G. A., J. Zhang, J. He, and X. Zhang, 2018: Role of extratropical cyclones in the recently observed increase in poleward moisture transport into the Arctic Ocean. Adv. Atmos. Sci., 35, 85–94, https://doi.org/10.1007/s00376-017-7116-0.
Wang, X. L., V. R. Swail, and F. W. Zwiers, 2006: Climatology and changes of extratropical cyclone activity: Comparison of ERA-40 with NCEP–NCAR reanalysis for 1958–2001. J. Climate, 19, 3145–3166, https://doi.org/10.1175/JCLI3781.1.
Wang, X. L., Y. Feng, G. P. Compo, V. R. Swail, F. W. Zwiers, R. J. Allan, and P. D. Sardeshmukh, 2013: Trends and low frequency variability of extra-tropical cyclone activity in the ensemble of twentieth century reanalysis. Climate Dyn., 40, 2775–2800, https://doi.org/10.1007/s00382-012-1450-9.
Wei, J., X. Zhang, and Z. Wang, 2019: Reexamination of Fram Strait sea ice export and its role in recently accelerated Arctic sea ice retreat. Climate Dyn., 53, 1823–1841, https://doi.org/10.1007/s00382-019-04741-0.
Wickström, S., M. O. Jonassen, T. Vihma, and P. Uotila, 2020: Trends in cyclones in the high‐latitude North Atlantic during 1979–2016. Quart. J. Roy. Meteor. Soc., 146, 762–779, https://doi.org/10.1002/qj.3707.
Woods, C., R. Caballero, and G. Svensson, 2013: Large-scale circulation associated with moisture intrusions into the Arctic during winter. Geophys. Res. Lett., 40, 4717–4721, https://doi.org/10.1002/grl.50912.
Wu, B., J. Wang, and J. E. Walsh, 2005: Dipole anomaly in the Arctic atmosphere and winter Arctic sea ice motion. Sci. China, 48D, 1529–1536, https://doi.org/10.1360/04yd0174.
Yin, J. H., 2005: A consistent poleward shift of the storm tracks in simulations of 21st century climate. Geophys. Res. Lett., 32, L18701, https://doi.org/10.1029/2005GL023684.
You, C., M. Tjernström, and A. Devasthale, 2021: Warm-air advection over melting sea-ice: A Lagrangian case study. Bound.-Layer Meteor., 179, 99–116, https://doi.org/10.1007/s10546-020-00590-1.
Yu, L., S. Zhong, M. Zhou, D. H. Lenschow, and B. Sun, 2019: Revisiting the linkages between the variability of atmospheric circulations and Arctic melt-season sea ice cover at multiple time scales. J. Climate, 32, 1461–1482, https://doi.org/10.1175/JCLI-D-18-0301.1.
Zahn, M., M. Akperov, A. Rinke, F. Feser, and I. I. Mokhov, 2018: Trends of cyclone characteristics in the Arctic and their patterns from different reanalysis data. J. Geophys. Res. Atmos., 123, 2737–2751, https://doi.org/10.1002/2017JD027439.
Zhang, J., 2005: Warming of the Arctic ice-ocean system is faster than the global average since the 1960s. Geophys. Res. Lett., 32, L19602, https://doi.org/10.1029/2005GL024216.
Zhang, P., G. Chen, M. Ting, L. Ruby Leung, B. Guan, and L. Li, 2023: More frequent atmospheric rivers slow the seasonal recovery of Arctic sea ice. Nat. Climate Change, 13, 266–273, https://doi.org/10.1038/s41558-023-01599-3.
Zhang, R., 2015: Mechanisms for low-frequency variability of summer Arctic sea ice extent. Proc. Natl. Acad. Sci. USA, 112, 4570–4575, https://doi.org/10.1073/pnas.1422296112.
Zhang, X., J. E. Walsh, J. Zhang, U. S. Bhatt, and M. Ikeda, 2004: Climatology and interannual variability of Arctic cyclone activity: 1948–2002. J. Climate, 17, 2300–2317, https://doi.org/10.1175/1520-0442(2004)017<2300:CAIVOA>2.0.CO;2.