Influence of Atmospheric Rivers on Glaciers

Abstract

Atmospheric rivers (ARs) are long, narrow, and transient corridors of robust horizontal water vapor transport commonly associated with a low-level jet stream ahead of the cold front of an extratropical cyclone. These weather features are essential for Earth’s hydrological cycle, transporting water vapor poleward, delivering precipitation for local climates, and having societal repercussions, such as intense storms and flood risk. The polar regions have experienced increasing AR activity in recent years. ARs usually transport substantial amounts of moisture and heat poleward that can potentially affect glaciers and sea ice. Many studies have demonstrated that ARs cause surface melting of glaciers in Antarctica and Greenland. Predicting and understanding the characteristics of ARs under global warming is a challenging task because there is not a consensus among scientists on a quantitative definition of ARs and the tracking methods. Understanding how ARs affect the surface mass balance of glaciers is crucial to increase our knowledge of how a warming atmosphere associated with warm ocean water will impact glaciated areas. In this work, we review recent advances in AR, including the methods used to identify them, their impacts on glaciers, their relationship with large-scale ocean-atmosphere dynamics, and variabilities under future climate.

1. Introduction

Over the last decade, interest in atmospheric river (AR) science has significantly increased because ARs are increasingly recognized globally as an essential weather phenomenon associated with extreme precipitation [1, 2, 3, 4, 5]. ARs potentially affect the global water vapor distribution and the spatiotemporal structures of the energy and water cycles of the planet [3]. Recent studies suggested that AR significantly impacts the energy and water budgets of the cryosphere, mountain glaciers [6], and polar regions [7, 8, 9, 10, 11, 12, 13, 14]. ARs can produce significant snow accumulation over the ice sheet [7, 15, 16, 17], melt events with consequences for ice shelf stability [18, 19, 20, 21], or calvin events [22]. The American Meteorological Society Glossary defines an AR as “a long, narrow, and transient corridor of strong horizontal water vapor transport typically associated with a low-level jet stream ahead of the cold front of an extratropical cyclone” [1]. ARs account for over 90% of the polar water vapor transport in the midlatitude and, therefore, have important implications for extreme precipitation when they make landfall, affect extreme precipitation events, regional water supply, as well as flooding hazards along western edges of continents (North America, South America, and Western Europe [23]) due to the interaction with the topography. This definition provides a qualitative description of ARs and opens the door to different quantitative definitions that can result in differences in AR climatologies. Understanding ARs in a warming climate is challenging because of the numerous algorithms that have been developed to identify and track the characteristics of ARs. Moreover, the AR community has not agreed on identification methodologies.

Here, we review the impact of AR on sea ice, mountain glaciers, and glaciers in the polar regions. We report on recent advances in AR, including the identification methods, their main climatological characteristics, impacts on glaciers, relationship with large-scale ocean-atmosphere dynamics, and some variabilities under future climate. For a review of the main characteristics of ARs that are responsible for the transport of large amounts of water across the midlatitudes toward higher latitudes, readers are referee to [24]. The climate change impacts of ARs, including physical processes such as the moistening of the atmosphere due to warming, shifting extratropical storm tracks, and the impacts of ARs on the hydrological cycle and hydrologic extremes and crucial AR research directions, extending from the urgency for higher resolution modeling, better observations (especially of regions globally where they are lacking) are reviewed in Payne et al. [4].

Section 2 gives an overview of ARs, their characteristics, their formations, and implications of ARs on sea ice and glaciers in higher latitudes regions and their roles in extreme events and in providing beneficial water supply. The methods used to identify AR are presented in Section 3. Section 4 presents observations and modeling of AR with a focus on high latitudes, while the effect of ARs on sea ice, glaciers, and ice sheets is shown in Section 5. Section 6 addresses AR under a warming climate, and in Section 7, we conclude with suggestions for future research.

2. Overview of atmospheric river and implications

ARs are large-scale weather features that transport significant amounts of moisture that are crucial for large-scale and local hydrological climate across the globe [25]. The impact of ARs on the western coasts of continents in midlatitude locations has been widely studied. In the polar region, emerging research has found that the ARs’ related moisture transport can interact with land and sea ice and lead to precipitation events that affect the local cryosphere [7, 8, 9, 10, 11, 12, 13, 15, 16, 17]. ARs traveling to Antarctica or Greenland from the midlatitudes are usually accompanied by warm air intrusion that can likely affect the sea ice and the stability of ice sheets. A study by Dare et al. [26] displayed a relationship between ARs and the precipitation associated with warm conveyor belt ascent. They argued that precipitation and water vapor transport increase with enhanced moisture in the feeder airstream. ARs can lead to snowfall anomaly, surface melt over glacier, and significantly influence the surface energy balance.

The effect of climate change on ARs or their characteristics under a warming climate is poorly understood. A study by Corringham et al. [27] suggested that ARs are responsible for the majority of the economic losses related to flooding in the Western United States and are anticipated to intensify in a warming climate. They argued that flood damages caused by extreme AR events may triple from 1 billion to over 3 billion a year toward the end of the century if mitigation actions are not implemented to diminish anthropogenic greenhouse effects. Although the impacts of ARs are increasingly recognized globally, at the global scale, essential questions remain unanswered, such as basic observations with ARs core, the development, and evolution of an AR from its initial setup to dissipation, how an AR interacts with large-scale dynamics, and quantifying the amount of surface melt ARs can trigger over the Arctic and Antarctic [4]. Understanding ARs responses under a warmer climate from a global perspective is of great interest and imperative for predicting weather systems, hydrological extremes, and the preparation for potential associated hazards [28, 29]. An increased trend in AR frequency is demonstrated in climate projection models [28, 30], particularly ARs that make landfall along the California Coast [4, 31]. Moreover, under future warming, it is expected that ARs will become longer, wider, and wetter [32]. For example, a recent study by Michaelis et al. [33] suggested climate change enhanced the precipitation associated with the February 2017 AR over Northern California. However, the physical drivers of AR genesis, development, and dissipation are still poorly understood.

An extratropical cyclone propagates originating from relatively warmer latitudes and propagating across the Southern Ocean is usually characterized by a continuous and strong moisture convergence in the cyclone’s warm conveyor belt circulation or along the trailing cold front that feeds the ARs [34, 35, 36]. The polar regions have experienced increasing AR activity in recent years [8, 36, 37]. Recently, studies have suggested that sea ice variation in Antarctica and the Arctic can be affected by the poleward transport of moisture and heat from midlatitudes [9, 10, 12, 13, 38]. Recent studies have shown that ARs contribute to the West Antarctic surface melting [11, 21, 39]. Over the West Antarctica Ice Shelf (WAIS), the blocking high and low pressure located along the coast of West Antarctica helps to dive the warm, and moist air to the WAIS. Previous studies have shown that both blocking high- and low-pressure systems are influenced by the El Niño-Southern Oscillation and the Southern Annular Mode and other modes of natural variability that affect the Antarctic continent [25, 40]. Moreover, AR precipitation promotes the rapid increase in surface height over West Antarctica during the 2019 austral winter [15] and abnormal snow accumulation in East Antarctica [7].

In August 2012 and July 2020, Li et al. [10] show that ARs associated with large cyclones triggered rapid sea ice melt through modulating turbulent heat fluxes and winds. They also find a significant negative correlation between atmospheric moisture content and the rate of changes in sea ice concentration over almost the entirety of the Arctic Ocean. The associated warmer air temperature induced by landfalling ARs in the Northeast Greenland ice stream triggers meltwater ponds and rivers that can modify the landscape of the ice stream [41]. The meltwater ponds and rivers on the ice stream absorb more sunlight than the surrounding glacier, and since Northeast Greenland is known to be an area of fast-flowing ice stream draining a large portion of the Greenland Ice Sheet into the ocean, increasing ARs may accelerate the outlet glaciers of this region [41].

ARs also affect the surface hydrological processes of many mountainous regions on the planet. They trigger avalanches and mudslides during winter when they land in mountainous areas [42, 43]. In the Western United States, Hatchett et al. [43] linked the highest percentage of casualties during avalanches to AR conditions. ARs regularly stroke British Columbia (Canada) in the fall and winter, which causes avalanches and heavy precipitation that significantly affect transportation systems [44]. Several studies have demonstrated that ARs triggered extreme precipitation events in mountainous regions with a variety of flooding hazards and ground failures including landslides, and riverbank erosion, for example, [30, 45, 46, 47, 48, 49, 50], with a significant impact on the population living in those areas. In the central Hindu-Kush, Karakoram, and Himalaya [46] reported that ARs contribute a significant fraction of the non-monsoon (October–May) precipitation.

2.1 Atmospheric rivers and warm air intrusion

Moist air reaches the Antarctic continent through a limited number of atmospheric processes. The primary one is the advection of warm, marine air. Warm, moist air is a key player in understanding the surface mass balance of the Antarctic [7, 51, 52, 53, 54]. Warm air intrusion (WAI) events are more frequent in austral winter in Antarctica than in other seasons. WAI events are crucial in cloud formation and precipitation events on the ASE [55]. Although WAI is frequent in summer, they can trigger surface melting and potentially affect ice sheet stability. ARs and WAI have been documented in the Amundsen Sea Embayment region [14, 40]. Refs. [56, 57] reported heat anomalies from the Atlantic Ocean reached the Greenland Ice Sheet and caused surface melt.

Studying AR and WAI events in the polar region is relevant because surface melted snow or ice can increase ice loss through runoff [58] and also by modifying ice flow dynamics and thermomechanical properties [59, 60]. Moreover, meltwater plays a prominent role in ice shelf hydrofracturing and ice-cliff collapse at deep grounding lines [61]. AR and WAI events are believed to be influenced by regional and large-scale atmospheric variability, including the Southern Annular Mode (SAM) and the El Niño-Southern Oscillation (ENSO), during the summer melt season [62, 63], and it is essential to understand that connection if any.

Figure 1a-d present water vapor imagery obtained from the UW-AMRC repository (ftp://amrc.ssec.wisc.edu/archive/2013/and/2014) during austral summer AR events from 18 to 25 February 2013 (FWAI2013), a fall AR event from 7 to 13 March 2013 (MWAI2013), and early austral winter in the Amundsen Sea Embayment, West Antarctica [14].

Figure 1.

Composite water vapor imagery data at 5 km spatial resolution on (a) 0000 UTC 20 February 2013, (b) 0000 UTC 10 March 2013, (c) 0012 UTC 21 February 2013, and (d) 0003 UTC 01 June 2014. The data are from the University of Wisconsin Antarctic Meteorological Research Center (UW-AMRC) repository.

3. Theory

ARs are characterized by intense moisture transport. AR intensity is determined by the column-integrated water-vapor transport (IVT),

$$\mathrm{IVT}=\sqrt{{\left(\frac{1}{g}{\int }_{p_t}^{p_s}\mathit{\text{qudp}}\right)}^2+{\left(\frac{1}{g}{\int }_{p_t}^{p_s}\mathit{\text{qvdp}}\right)}^2},$$E1

where the gravitational acceleration g is $\mathrm{m}{\mathrm{s}}^{-2}$, the specific humidity q is in kg ${\mathrm{kg}}^{-1}$, p is the atmospheric pressure (hPa), $p_s$ surface pressure, and $p_t$ an upper-atmospheric reference pressure (typically set between 500 and 200 hPa, and between 1000 and 300 hPa in the polar regions), and u and v represent the zonal and meridional component of the wind. For the high-latitude regions, vertically integrated water vapor (IWV) fields.

$$\mathrm{IWV}=\frac{1}{g}{\int }_{p_t}^{p_s}\mathit{qdp},$$E2

are combined with the IVT. Changes in IVT can be partitioned into the change in atmospheric moisture contain and the change in atmospheric motion (through wind vector). The Clausius-Clapeyron equation, which states an exponential-like increase of the water vapor content of saturated air $\left(q^{\mathrm{sat}}\right)$ with temperature (T), can be employed to investigate the change in the thermodynamic component of IVT. The Clausius-Clapeyron relationship is defined as

$$\frac{{\mathit{dq}}^{\mathrm{sat}}}{\mathit{dt}}=\alpha \left(T\right)q^{\mathrm{sat}},$$E3

where $\alpha \left(T\right)$ denotes the Clausius-Clapeyron scaling factor, described as

$$\alpha \left(T\right)=\frac{L_v}{R_v{T}^2}$$E4

where latent heat of fusion $L_v$ and $R_w$ is the gas constant of water vapor [4]. From the Clausius-Clapeyron equation, one can deduce that the fraction of change in the humidity per degree of surface warming is related to the magnitude of warming at a particular height relative to the surface.

$$\frac{1}{\mathrm{\Delta }{T}_s}\frac{\mathrm{\Delta }{q}^{\mathrm{sat}}}{q^{\mathrm{sat}}}=\alpha \left(T\right)\frac{\mathrm{\Delta }T}{\mathrm{\Delta }{T}_S}$$E5

The response of IVT at the surface is well understood, while above the surface, $\alpha \left(T\right)$ is not constant. An increase in the specific humidity enhances a release of latent heat when saturated air ascends. The lapse rate then decreases with warming and, therefore $\mathrm{\Delta }T$ increases with height [4].

The response of atmospheric circulation to warming is much less certain [64] and the dynamical responses of IVT that are crucial to the impact of ARs on land are usually their location of landfall and intensity.

3.1 Atmospheric river identification and tracking

Advances in research on ARs have demonstrated their significance for the global water cycle, revealing uncertainties in their mapping, tracking, categorizing, and forecasting landfalling events that make them one of the challenging branches of atmospheric sciences. Numerous methods to identify and track ARs have been developed, usually to address specific research questions. Some criteria used include geometry, threshold values of critical variables, and time dependence, to name a few. These different methods produce differences in AR climatologies and, therefore, differences in the impacts attributable to ARs. More than 20 AR detection methods were assessed in the atmospheric river tracking method intercomparison project [25, 65, 66].

The majority of AR tracking methods often choose a thresholding and are based on the analysis of vertically integrated water vapor transport (IVT) or vertically integrated water vapor (IWV) [25, 66]. The magnitude of the thresholding variable can be either absolute $\left(\mathrm{IVT}\ge 250\mathrm{kg}{\mathrm{m}}^{-1}{\mathrm{s}}^{-1}\right)$ or relative ($\mathrm{IVT}\ge 85\mathrm{th}$ percentile of local IVT). Probabilistic IVT forecasts have also been used to determine AR location and intensity [67]. Recent advances in AR tracking methods include machine learning techniques that do not require threshold values have also been employed [68], and a near-global AR detection algorithm that incorporates 3-D wind information from satellite observations [69].

The heat and moisture transport associated with ARs can significantly affect the cryosphere in the polar regions. Indeed, even a slight increase in surface temperature above the freezing point can trigger surface melting over the ice. The impacts of ARs on the Antarctica ice sheet include surface melt in West Antarctica [14, 21], foehn events near the northern tip of the Antarctic Peninsula [39], and snowfall accumulation in East Antarctica [7]. Over the Greenland Ice Sheet, the combined effect of downwelling long-wave radiation and warm air intrusion can negatively impact the ice mass [8, 56, 57].

Djoumna and Holland [14] employed the algorithm developed by Wille et al. [21] for Antarctica to analyze AR events in the Amundsen Sea Embayment (ASE). They combined an IWV-based and IVT-base algorithm between the latitude band of ${35}^{°}\mathrm{N}$ and ${80}^{°}$S, to locate regions where IVT exceeds IVT’s 98th percentile. The IVT threshold of 250 kg ${\mathrm{m}}^{-1}{\mathrm{s}}^{-1}$ was employed to identify AR lateral boundaries as recommended in the literature [7, 39, 70].

Ralph et al. [70] developed a scale to categorize AR events using IVT values that vary from weak AR $\left(\mathrm{IVT}\ge 250‐500\mathrm{kg}{\mathrm{m}}^{-1}{\mathrm{s}}^{-1}\right)$ to extreme AR ($\mathrm{IVT}\ge 1250$ kg ${\mathrm{m}}^{-1}{\mathrm{s}}^{-1}$). Based on a scale that categorizes AR events based on the maximum instantaneous IVT associated with a period of AR conditions and the duration of those conditions at a point developed by Ralph et al. [70] for the northeastern Pacific Ocean and the Western United States, the MWAI2013 and FWAI2013 AR events in the Amundsen Sea Embayment by Djoumna and Holland [14] were moderate $\left(\mathrm{IVT}\ge 500‐750\mathrm{kg}{\mathrm{m}}^{-1}{\mathrm{s}}^{-1}\right)$ in Figure 2. They were weaker than the 25–30 May 2016 event reported by Wille et al. [71].

Figure 2.

Panels (a–c) and (g) show the integrated water vapor (IWV, kg ${\mathrm{m}}^{-2}$) and the 800 hPa wind vectors obtained from ERA-Interim during two atmospheric rivers that stroke the Amundsen Sea Embayment during 19–22 February 2013 and 10 March 2013. The vertically integrated vapor transport (IVT, kg ${\mathrm{m}}^{-1}{\mathrm{s}}^{-1}$) fields and 500 hPa geopotential heights (m, blue contour lines) are shown in panels (d–f) and (h). The figures are from [14].

4. Observations and modeling of AR in the high-latitude regions

Many studies have shown that ARs cause extreme rain events and flooding hazards and can supply water in midlatitude regions. However, the use of numerical weather prediction (NWP) systems to accurately forecast ARs is still under development. Most of the research on the topics has been conducted only during the last two decades. Gaps in the global observing system, for example, across remote ocean areas, and over glaciated areas, hindered efforts to improve forecasts of landfalling ARs. Observation campaigns across AR over glaciers and sea ice are nearly nonexistent [72] used radiosonde measurements collected in austral summer of 2018–2019 (November 2018 to February 2019) at the Dronning Maud Land coast (East Antarctica) to detect AR signatures in the vertical profile of wind speed and direction, air temperature, relative humidity, and specific humidity. Most observation campaigns targeting AR have been conducted in midlatitudes and subtropical regions. Several observation campaigns using the drop-sondes have been conducted in the midlatitudes and subtropical regions; these observations include the vertical profile of water vapor, wind, and pressure and were obtained from 304 aircraft drop-sondes across 21 ARs [73]. Recently, the AR reconnaissance in the northeast Pacific [74] aimed to improve the science and forecasts of landfalling ARs to help better inform decision-makers on water management and flooding hazard in the Western United States. To achieve this goal, NWP models will be combined with observations collected within ARs’ core from drop-sondes (deployed by aircraft), and also on the ocean surface (drifting ocean buoys), and airborne radio occultation [75, 76].

In recent years, many studies have highlighted the significance of ARs in shaping the global water vapor distribution, water and energy budgets, and hydrology extremes. Since ARs can affect the Earth’s climate through their effect on the poleward transport of water vapor, it is crucial to accurately represent their associated dynamics, thermodynamics, and hydrodynamics in climate models [31, 77]. Climate models to date represent the observed statistics of ARs relatively well, while significant regional biases still exist [77]. An accurate simulation of ARs is crucial to pursue robust projections of AR changes under a warming climate. However, climate models are complex and often bear some degree of uncertainty, and incorporating the fundamental dynamical processes of ARs in a climate model is still under development, and this hindered our understanding of the AR response to a warming climate [31]. Zhang et al. [31] elaborated an idealized atmospheric general circulation model in which an Earth-like global circulation was combined with a hydrological cycle model. The model used passive tracers, simplified cloud microphysics, and precipitation to model water vapor and clouds. They found a good representation of observed dynamical structures for individual ARs, statistical characteristics of ARs, and spatial distributions of AR climatology.

For an AR forecast to be of interest to the following forecast users’ properties at seasonal time scales, late medium range and early extended range and the short and early medium range (1–7 days) [76] should be considered depending on the region of interest.

• Good knowledge of the ARs frequency and the amount of total precipitation induced by ARs within a season across a given region is valuable information that can help decision-makers to manage the consequences of ARs. In the polar region, we know that ARs effects are likely critical during the summer melting season.

• In the late medium range and early extended-range forecast (2–3 weeks), the approximate time when an area may experience extreme precipitation and flooding is desired.

• From a forecast user’s perspective, predicting AR in the short and early medium range (1–7 days) is highly important and challenging from a modeling perspective. It is essential to accurately predict the intensity, location, and exact time an AR will make a landfall.

Two newly available AR forecast products are the Extreme Forecast Index (EFI) for water vapor flux [78, 79, 80, 81] and the AR Scale [70]. The EFI compares the probability distribution of the European Centre for Medium-Range Weather Forecasts (ECMWF). The climatology of ARs in the polar regions and mountain glaciers is still being studied. There is still a long way to go before developing AR forecasting tools for high-latitude regions.

5. Effect of ARs on sea ice, glaciers, and ice shelves

The importance of ARs has been demonstrated over the Antarctic ice sheet [7, 14, 15, 18, 19, 21, 22, 82, 83, 84], the Antarctic sea ice [12], and the Arctic continent [8, 10, 36, 37, 41]. Extreme warmer events usually accompany ARs traveling to Antarctica or Greenland from the midlatitudes. Most of these weather events are formed at lower latitudes, over the Southern Ocean and Atlantic Ocean. They transport heat and moisture, propagate eastward and poleward, dissipate, and lose their baroclinic energy near the continent [36, 85, 86]. ARs generate positive anomalies in temperature, moisture, and winds in the coastal regions bordering the Antarctic and Greenland. Several studies have shown ARs impact both the Antarctic ice sheet and the Greenland Ice Sheet through high snowfall events, surface melt events, and sea ice decays [7, 8, 19, 37, 41, 82, 83, 84].

ARs occur in Antarctica on low frequency (around 12 events per year in West Antarctica). They account for the most significant percentage of precipitation observed over Antarctica and have important consequences for ice shelf stability [17, 18, 21]. For example, Wille et al. [82] showed that the exceptional heat wave with widespread $30–{40}^{°}\mathrm{C}$ temperature anomalies across the ice sheet that occurred over East Antarctica between 15 and 19 March 2022 was triggered by an intense atmospheric river advecting subtropical/midlatitude heat and moisture deep into the Antarctic interior. The March 2022 East Antarctica “Heat” wave caused extensive precipitation and surface melt along coastal areas. They also recorded extensive high snowfall accumulations within the interior of the East Antarctic region that resulted in a primarily positive surface mass balance [83]. Extremes warm events can trigger instantaneous knockout effects on the cryosphere because of positive feedbacks (ice/snow albedo) susceptible to enhance and escalate the warming trend [84]. ARs and the associated warm air intrusion events cause about 3 days of surface melting over the Pine Island Glacier in February and March 2013 [14]. Liang et al. [12] suggested that low-frequency ARs can lead to prominent sea ice reduction over marginal ice zone primarily through thermodynamic processes over the Antarctic sea ice. They also argued that ARs could amplify sea ice melting during cold weather periods in contrast to narrow effects during summertime.

In the Arctic, Li et al. [10] found that individual AR events associated with large cyclones initiate a rapid sea ice decrease through turbulent heat fluxes and winds.

6. AR under future warming

The characteristic of ARs, their number, frequency, and intensity are likely to change in future climates, and it is essential to understand and predict how ARs will change under a warming climate. An increase in atmospheric moisture in a warmer climate will lead to an increase in the magnitude of ARs, with significant consequences for flooding, ice sheet stability, and sea ice cover. There is now a growing number of studies highlighting the importance of investigating the contribution of a globally warming atmosphere and changes in the inflow of relatively warm water when examining future changes in the polar region, especially the surface mass balance over glaciated areas and air-sea interactions over the ice cover region.

AR occurrences for different regions around Antarctica and Greenland have been associated with various teleconnections and modes of natural variability, such as the Southern Annular Mode (SAM, [17]), the Pacific South American Mode 2, the Pacific Decadal Oscillation [87] Arctic Oscillation, North Atlantic Oscillation, and Pacific/ North America [88]. We hypothesize that the intensity and frequency of AR events may vary with the SAM and the ENSO. It has been suggested that ENSO events can lead to more blocking in the Amundsen Sea and more robust westerly flow on the continental shelf, while a positive SAM is linked to weaker easterly flow north of the ASE [89]. Recently, [60] reported that the Amundsen Sea blocking activity and a negative SAM correlate with ENSO conditions in the tropical Pacific Ocean during the peak summer warming (December–January) in West Antarctica. Moreover, Shields et al. [25] argued that the Indian Ocean Dipole teleconnections in phase with ENSO produce a stronger AR precipitation response in Antarctica compared to other modes of natural variability that affect the Antarctic continent.

The AR Recon program focus on the North Atlantic and represents a recent initiative to improve our understanding to achieve better forecasts of ARs and their impacts [76]. The program targets to develop advanced numerical methods and observation campaigns, assimilate the observed data, and theoretical physical studies to investigate ARs responses to climate change in Europe and the United States of America. A similar program will be needed for an in-depth study of polar ARs. However, the polar regions are characterized by a range of extreme and continuously harsh environmental conditions, and observation campaigns within polar ARs are challenging.

7. Summary points and future issues

The critical role of ARs in various extreme precipitation and flooding events in Western North America, the Western Pacific, Europe, New Zealand, and South America has been demonstrated in multiple studies over the last two decades. In the polar regions, the effects of ARs have been uncovered in the last decade. ARs are less frequent in the polar regions than in the midlatitudes. However, ARs are shown to have a high impact on the Antarctic Ice Sheet mass balance, as they have been linked to surface melting on the West Antarctic Ice Sheet, record high temperatures, extreme snowfall events on the Antarctic Peninsular and the East Antarctic ice sheet, and sea ice decline in marginal ice zones. In the Arctic, ARs have been shown to affect warming in the troposphere during summertime, and the surface radiative balance with potential consequences for Arctic sea ice [9, 10]. A good understanding of the dynamics that drive ARs is vital for predicting how weather extremes in the Arctic and Antarctic and their impacts will change in response to climate change in those regions.

In this work, we have reviewed the impact of AR on sea ice, mountain glaciers, and glaciers in the polar regions. We have also reviewed recent advances in AR, including the identification methods, their main climatological characteristics, their impacts on glaciers, their relationship with large-scale ocean-atmosphere dynamics, and some variabilities under future climate. Significant progress has been made toward a better understanding of ARs, including addressing uncertainty in the tracking and identification algorithm through the ARTMIP project, a quantitative definition of ARs, observations campaign (the AR Recon program), and modeling efforts for better forecasts of ARs (the EFI for water vapor flux and the AR Scale). However, several questions remain about the mechanisms driving atmospheric rivers and their life cycles (genesis, development, and dissipation), observations of their development, their interaction with large-scale dynamics, their role in ephemeral, extent melt events over the Arctic and Antarctica, forecast ARs accurately using NWP systems to provide warnings and awareness, and a better understanding of how ARs interact with large-scale circulation in a warming climate.

As high latitude is sensitive to warmer temperatures, it is imperative to understand how extreme weather or climate conditions will influence moisture intrusions on the Antarctic continent and Arctic and mountain glaciers. Improving our understanding of the exact mechanisms (thermodynamic or dynamical changes) that underlay the long-term changes in moisture transport associated with ARs remains uncertain in the pair region. Moreover, fields campaign to collect observations within ARs’ core in the polar region will be necessary. A better prediction of the influence of anthropogenic changes in the climate of Antarctica and Greenland will require an accurate understanding of the processes and impacts of extreme temperature events associated with atmospheric rivers in these regions.

Conflict of interest

The authors declare no competing interests.