Arctic hydrology is experiencing rapid changes including earlier snow melt, permafrost degradation, increasing active layer depth, and reduced river ice, all of which are expected to lead to changes in stream flow regimes. Recently, long-term (>60 years) climate reanalysis and river discharge observation data have become available. We utilized these data to assess long-term changes in discharge and their hydroclimatic drivers. River discharge during the cold season (October-April) increased by 10% per decade. The most widespread discharge increase occurred in April (15% per decade), the month of ice break-up for the majority of basins. In October, when river ice formation generally begins, average monthly discharge increased by 7% per decade. Long-term air temperature increases in October and April increased the number of days above freezing (+1.1 d per decade) resulting in increased snow ablation (20% per decade) and decreased snow water equivalent (-12% per decade). Compared to the historical period (1960-1989), mean April and October air temperature in the recent period (1990-2019) have greater correlation with monthly discharge from 0.33 to 0.68 and 0.0-0.48, respectively. This indicates that the recent increases in air temperature are directly related to these discharge changes. Ubiquitous increases in cold and shoulder-season discharge demonstrate the scale at which hydrologic and biogeochemical fluxes are being altered in the Arctic.

Winter is an important period for ecological processes in northern regions; however, compared to other seasons, the impacts of winter climate on ecosystems are poorly understood. In this review we evaluate the influence of winter climate on carbon dynamics based on the current state of knowledge and highlight emerging topics and future research challenges. Studies that have addressed this topic include plot-scale snow cover manipulation experiments that alter soil temperatures, empirical investigations along natural climatic gradients, laboratory temperature incubation experiments aimed at isolating influential factors in controlled environments, and time series of climate and carbon data that evaluate long-term natural variation and trends. Combined, these studies have demonstrated how winter climate can influence carbon in complex ways that in some cases are consistent across studies and in other cases are difficult to predict. Despite advances in our understanding, there is a great need for studies that further explore: (i) carry-over effects from one season to another, (ii) ecosystem processes in the fall-winter and winter-spring shoulder seasons, (iii) the impacts of extreme events, (iv) novel experimental approaches, and (v) improvements to models to include ecological effects of winter climate. We also call for the establishment of an international winter climate change research network that enhances collaboration and coordination among studies, which could provide a more thorough understanding of how the snow-covered period influences carbon cycling, thereby improving our ability to predict future responses to climate change.

The atmospheric methane (CH4) concentration, a potent greenhouse gas, is on the rise once again, making it critical to understand the controls on CH4 emissions. In Arctic tundra ecosystems, a substantial part of the CH4 budget originates from the cold season, particularly during the zero curtain (ZC), when soil remains unfrozen around 0 degrees C. Due to the sparse data available at this time, the controls on cold season CH4 emissions are poorly understood. This study investigates the relationship between the fall ZC and CH4 emissions using long-term soil temperature measurements and CH4 fluxes from four eddy covariance (EC) towers in northern Alaska. To identify the large-scale implication of the EC results, we investigated the temporal change of terrestrial CH4 enhancements from the National Oceanic and Atmospheric Administration monitoring station in Utqiagvik, AK, from 2001 to 2017 and their association with the ZC. We found that the ZC is extending later into winter (2.6 0.5 days/year from 2001 to 2017) and that terrestrial fall CH4 enhancements are correlated with later soil freezing (0.79 0.18-ppb CH4 day(-1) unfrozen soil). ZC conditions were associated with consistently higher CH4 fluxes than after soil freezing across all EC towers during the measuring period (2013-2017). Unfrozen soil persisted after air temperature was well below 0 degrees C suggesting that air temperature has poor predictive power on CH4 fluxes relative to soil temperature. These results imply that later soil freezing can increase CH4 loss and that soil temperature should be used to model CH4 emissions during the fall. Plain Language Summary Methane (CH4) is a powerful greenhouse gas, capturing more heat per molecule than carbon dioxide (CO2). Although CH4 is less concentrated in the atmosphere, it is the second most important greenhouse gas with respect to climate change after CO2. Arctic tundra ecosystems are potentially major sources of CH4, given large soil carbon storage and generally wet conditions, favorable to CH4 production. This study investigates if the persistence of unfrozen soils is associated with higher CH4 emissions from the Arctic. We combined long-term soil temperature measurements, terrestrial CH4 enhancements from the National Oceanic and Atmospheric Administration monitoring station in Utqiagvik, AK, and CH4 emissions from Arctic tundra ecosystems across four stations in the North Slope of Alaska. Our results show that from 2001 to 2017 the soil is freezing later and that later soil freezing is associated with higher fall CH4 enhancements. Given that unfrozen soils are related to higher CH4 emissions, a later soil freezing could contribute to the observed increase in the regional atmospheric CH4 enhancement. Unfrozen soil layers persisted after the air temperature was well below 0 degrees C, suggesting that air temperature does not properly predict the sensitivity of CH4 emissions to climate warming.