Role of atmospheric rivers in shaping long term Arctic moisture variability

The Arctic has experienced warming at a rate more than twice the global average, a phenomenon known as Arctic Amplification (AA). This warming, governed by the Clausius-Clapeyron (CC) relationship, leads to atmospheric moistening, resulting in increased specific humidity, cloud cover, and precipitation. This increased moisture has significantly altered Arctic hydrological and cryospheric variability. While the CC relationship explains part of the Arctic moistening trend, especially during summer, large-scale circulation variability also plays a significant role. This variability influences moisture distribution and transport through its effects on weather systems, including atmospheric rivers (ARs). ARs, characterized by intense plumes of water vapor, contribute over 90% of the poleward water vapor transport into the Arctic, with their highest occurrence during June-August. The frequency of ARs in the Arctic has increased in recent decades, particularly over western Greenland, contributing to significant Greenland Ice Sheet (GrIS) melt. The precise causes of these increases remain uncertain, with suggestions attributing them to thermodynamic effects of global warming, aerosol forcing, and atmospheric internal variability. A key unanswered question is how global warming and low-frequency circulation variability, both individually and interactively, have impacted Arctic moisture through AR activity. This study addresses this gap by utilizing a nudging approach in the Community Earth System Model version 1 (CESM1) to investigate AR responses to observed wind trends and compare them with responses to anthropogenic forcing. The focus is on the summer months (JJA) because this period exhibits the most significant and positive Pan-Arctic moistening trend.

Existing literature highlights the significant role of Arctic Amplification (AA) and its associated warming in driving changes in the Arctic's hydrological cycle. Studies have emphasized the Clausius-Clapeyron (CC) relationship, linking increased temperatures to higher specific humidity and increased precipitation. However, the observed trends in Arctic moisture content and associated impacts like accelerated sea ice and GrIS melt have not been fully explained by the CC relationship and anthropogenic forcing alone. Previous research has implicated increased evaporation from local oceans and surrounding continents, enhanced sublimation of ice and snow, and intensified moisture transport from lower latitudes as contributors to this phenomenon. Climate models forced by historical anthropogenic emissions generally reproduce a warmer and more humid Arctic, but discrepancies between model outputs and observations remain. These discrepancies have been attributed to limitations in capturing the observed low-frequency variability of tropical-extratropical teleconnections, among other factors. Studies suggest that internal climate variability also plays a crucial role in modulating observed moisture trends and that both anthropogenic forcing and internal variability should be factored in when examining the mechanisms behind recent moisture trends in the Arctic. Previous works have explored the role of atmospheric rivers (ARs) in Arctic moisture transport, but a comprehensive understanding of their interaction with large-scale circulation changes is lacking. Specifically, how global warming and low-frequency circulation variability independently and together contribute to Arctic moistening through AR activity needs further investigation, especially in light of significant melt events linked to AR activity.

This study utilizes a multi-faceted approach combining observational data analysis with numerical modeling experiments. Observational data from ERA5 reanalysis data are used to analyze changes in Arctic atmospheric circulation (Z200), specific humidity, temperature, cloud cover, and radiation from 1979 to 2019. The results from ERA5 are consistent with data from ERA-Interim. Both monthly and 6-hourly data are used, with higher temporal resolution data employed for analyzing AR variability. Atmospheric rivers (ARs) are identified using an integrated water vapor transport (IVT)-based detection algorithm, based on ARTMIP recommendations, that identifies areas exceeding the 85th percentile of climatological monthly IVT. Further criteria are used to define AR length and length-to-width ratio. The CMIP6 multi-model ensemble mean of 34 climate models and the CESM2 large ensemble (40 realizations) are used to study the response of the atmosphere to historical anthropogenic forcing and internal variability. CESM1 is used for nudging experiments where the observed wind trends (from ERA5) are added to the model (within the Arctic region north of 60°N) to examine how ARs respond to large-scale circulation changes. Three sets of simulations were performed in CESM1: a control run (CTL) with constant CO2 levels of 2000, a wind nudging run (WIN) where 3-D wind anomalies were added to the model each summer, and a combined run (WIN+CO2) that includes both wind nudging and anthropogenic forcing up to 2020 levels. Each run is integrated for 40 years and the 40-year means are compared to analyze stable responses. A fingerprint analysis was performed using CESM2-LEN to further evaluate the contribution of internal variability by comparing members of the ensemble with differing AR trends, especially in key regions like western Greenland. Maximum Covariance Analysis (MCA) is extensively used to identify coupled patterns between AR frequency and atmospheric variables on both interannual and shorter timescales. A statistical method was developed to separate the contributions of ARs from other factors to the long-term trends of specific humidity. The methodology involves removing specific humidity values at grid points where ARs occur when calculating long-term trends unrelated to ARs.

The study reveals a significant positive trend in summertime AR frequency in several Arctic regions from 1979 to 2019, with notable increases over northern Canada, western Greenland, eastern Siberia, and parts of northern Europe. This increase is strongly correlated with trends in large-scale JJA circulation, specifically the upper-level geopotential heights (Z200). However, CMIP6 and CESM2 simulations of anthropogenic forcing do not fully capture these observed changes, suggesting a significant role for internal variability. MCA analysis identifies a dominant coupled pattern between detrended JJA Z200 and AR frequency, exhibiting a zonal wave number 2 structure. The spatial pattern of this mode closely resembles the long-term trends of both variables, especially over Greenland. The wind anomalies associated with this MCA mode reveal that strong southerly wind anomalies enhance AR frequency, while northerly or easterly winds reduce it. Analysis of daily AR frequency and specific humidity anomalies in 2012, particularly focusing on western Greenland, shows strong co-variability, with ARs preceding significant increases in specific humidity. This pattern is consistent across summers from 1979 to 2019 in western Greenland, northern Europe, and eastern Siberia. Composite analyses show that specific humidity increases approximately one day before ARs, peaks one day after, and then quickly returns to pre-AR levels. This indicates a close day-to-day relationship between AR activity and specific humidity changes across much of the Arctic. Quantifying the impact of ARs on long-term moisture changes reveals that the trend in AR activity may contribute to 36% of the overall increase in Pan-Arctic atmospheric moisture since 1979. The impact is even more significant in regions like western Greenland, northern Europe, and eastern Siberia, where ARs account for over 50% of the moisture increase. CESM1 nudging experiments, using observed wind trends, show a pattern of AR change consistent with observations. The increase in AR frequency is particularly noticeable over northwestern Greenland and northern Eurasia, while decreases are observed in southern and eastern Greenland and the Bering Strait, mimicking observed patterns. These changes are linked to southerly and northerly wind anomalies related to large-scale circulation. The fingerprint analysis using CESM2-LEN reinforces the role of large-scale circulation in modulating AR activity, especially over western Greenland and eastern Siberia, as only a subset of the ensemble captures the observed trends of both AR frequency and large-scale circulation.

The study's findings demonstrate that large-scale atmospheric circulation plays a significant role in shaping long-term Arctic moisture variability through its influence on AR activity. The observed increase in summertime AR frequency is not fully captured by models focusing solely on anthropogenic forcing, highlighting the importance of internal variability and large-scale circulation trends. The statistically significant correlation between AR frequency and large-scale circulation, revealed through MCA, suggests that changes in the background flow directly influence AR pathways and intensity. The model experiments, especially the nudging simulations in CESM1, confirm that changes in large-scale winds can effectively modulate AR activity. These findings underscore that while anthropogenic forcing undoubtedly contributes to Arctic warming and increased atmospheric moisture, low-frequency changes in circulation are essential in determining the spatial patterns and magnitudes of these changes, particularly through the regulation of ARs. The study's estimates of AR contributions to long-term Arctic moistening quantify this impact, particularly in specific regions.

This study provides compelling evidence for the significant role of large-scale circulation in modulating the frequency and impact of atmospheric rivers on Arctic moisture variability. The observed increase in summertime AR activity, particularly in specific regions, is not solely attributable to anthropogenic forcing but is substantially influenced by low-frequency circulation changes. The findings highlight the need to incorporate internal variability and large-scale atmospheric dynamics into climate models to accurately project future Arctic moisture changes. Future research should investigate how these findings will persist in a warming world with a potentially weakened and poleward-shifted jet stream, and explore the role of ARs in high-latitude moisture feedbacks and their influence on AA and accelerating melt events.

While the study provides a comprehensive analysis, there are some limitations to consider. The nudging approach in CESM1, while effective in simulating observed wind trends, might not fully represent the complex interactions within the coupled climate system. The statistical method for separating AR contributions to moisture trends relies on assumptions about the independence of AR events, which may not always hold. Further research is needed to fully understand the complex interaction between anthropogenic forcing and internal climate variability in influencing AR activity, particularly in the eastern Siberia region where the large-scale circulation trends appear to have a weaker impact according to the nudging analysis.