Climate extremes in Svalbard over the last two millennia are linked to atmospheric blocking

The Arctic is warming two to three times faster than the global average, with Svalbard experiencing about 4 °C warming over the last century and an accelerated warming since 1991. Svalbard lies along the pathway of Atlantic heat and moisture, including frequent atmospheric rivers, making it highly sensitive to hydroclimatic changes. Recent episodes (2015–2017) featured exceptional warmth, reduced sea ice, and extreme rainfall leading to mudflows, slushflows, and damaging avalanches in and around Longyearbyen, indicating a shift toward a rain-dominated hazard regime. Understanding short-term variability and circulation dynamics is essential to assess regional extremes. Atmospheric blocking—particularly Scandinavian, Ural, and Greenland blocking—modulates moisture transport and temperature anomalies across the North Atlantic–Arctic sector. While Greenland blocking has contributed to enhanced Greenland Ice Sheet melt and sea-ice loss, Scandinavian and Ural blocking advect warm, moist air toward Svalbard, producing anomalously warm and wet conditions. However, limited instrumental records, lack of long blocking proxies, and model uncertainties hinder a comprehensive understanding. This study integrates instrumental data, lake sediments, tree-ring reconstructions, and a paleoclimate simulation to evaluate how Scandinavian (and Ural) blocking has controlled Svalbard’s temperature and precipitation variability from daily to millennial scales, and to place recent extremes in a long-term context.

Prior work identifies atmospheric blocking as a key driver of Arctic hydroclimate variability, with Greenland blocking linked to enhanced Greenland melt and sea-ice decline, and Ural/Scandinavian blocking associated with warm air advection and regional extremes. Observational and modeling studies have noted increased blocking episodes in recent decades, but their depiction and future changes remain uncertain across climate models. The literature also highlights the roles of atmospheric rivers in Arctic moisture transport and the influence of sea-ice conditions on precipitation extremes around Svalbard. A critical gap remains due to short instrumental records and the absence of direct pre-instrumental proxies for blocking, complicating attribution and long-term assessments. The study builds on tree-ring temperature reconstructions for Scandinavia that capture warm summers associated with anticyclonic conditions and on reconstructions of Nordic Seas sea ice that document large historical fluctuations relevant to Svalbard hydroclimate.

Observations and reanalysis: Seasonal (June–November) precipitation over Svalbard for 1955–2019 was taken from two gridded datasets (UDEL and GPCC). Atmospheric fields (500 hPa geopotential height Z500, 500 hPa winds, surface air temperature) came from ERA5. Sea surface temperature was obtained from HadISST and ERSST. Regression analyses were performed between standardized Svalbard precipitation indices and anomaly fields, reporting changes per one standard deviation change; significance assessed using SciPy linregress. Lake monitoring and traps: In Linnévatnet, sedimentation occurs when ice-free (June–November). Since 2004, intervalometers and sediment traps at mooring C recorded half-hourly sedimentation timing and thickness. Grain-size was measured (Beckman Coulter LS 13 320) on 0.5 cm slices; elemental Ca was measured using an Itrax XRF core scanner. Event-scale analyses linked coarsest grains and highest Ca to heavy precipitation and days with Scandinavian/Ural blocking (identified via Z500 anomalies). Sediment cores and annually resolved proxies: In April 2019, a 41.8 cm gravity core and a 498 cm piston core were recovered at mooring C. Ninety-four overlapping thin sections covered a 5.03 m composite; ~5,000 SEM backscattered electron images (1 µm pixel) were used to identify varve boundaries and extract annual grain-size for the upper 370 cm with regular laminations. Annual Ca was obtained via µ-XRF and averaged annually. Loss-on-ignition (550 °C) provided organics as a proxy for glacial activity. Paleoclimate simulation and blocking index: The MPI-ESM1-2-LR PMIP4 past2k simulation (1–1850 CE) provided SAT, PR, Z500, and lower-tropospheric moisture/winds. A Scandinavian Blocking (SB) index over the Common Era was computed by PCA of Z500 anomalies over 60°W–50°E, 60°N–85°N; the leading mode corresponds to NAO and the second principal component (PC2) captures the SB pattern. Regressions of SAT and PR onto the SB index characterized blocking impacts. Future projections: Two CMIP6 models (IPSL-CM6A-LR, MPI-ESM1-2-HR) were used to estimate 2015–2099 trends in June–November SAT and PR across SSP2-4.5, SSP3-7.0, SSP5-8.5 (raw CMIP6). Changes in extreme rainfall days (>10 mm) were analyzed using ISIMIP3b bias-adjusted, statistically downscaled data under SSP3-7.0, comparing 2015–2034 with 2080–2099. Data integration: Lake Ca and grain-size were interpreted as rainfall-sensitive proxies modulated by glacier size (grain-size) and catchment lithology (Ca from eastern carbonate bedrock). Ca is a robust proxy for medium-to-large rainfall events (>10 mm/day). Ca and tree-ring summer temperature reconstructions for Scandinavia were compared, as were Ca and reconstructed Nordic Seas sea-ice concentration, to evaluate co-variability with blocking-linked warmth and moisture.

- Modern circulation link: Wet months (Jun–Nov) in Svalbard are associated with positive Z500 anomalies centered over northern Scandinavia (Scandinavian blocking), tilted jet, and low pressure over Greenland, enabling moisture advection into Svalbard. These periods also feature higher SSTs in the Nordic Seas and warmer surface air temperatures regionally. The blocking–precipitation linkage holds across seasons, including winter. - Event-scale confirmation: Sediment trap and intervalometer records show the coarsest grain-size and highest Ca deposition coinciding with days of strong Scandinavian blocking (e.g., 11 Sep 2015; 28 May 2016; 15 Oct 2016). Days with the coarsest sediments dominate annual sediment flux and are characterized by blocking over northern Scandinavia or Eurasia and warm Nordic Seas SSTs. - Millennial perspective: Annually resolved Linnévatnet Ca exhibits a long decline from the Medieval period to the Little Ice Age, with minima from the 1600s to mid-1800s, followed by a steady rise beginning ~1850. The 2016 floods stand out as exceptional within the last ~1600 years. Ca co-varies with Scandinavian tree-ring summer temperature reconstructions (r = 0.30 annual; r = 0.42 with 5-year running mean; p < 0.001) and with other Eurasian tree-ring reconstructions over 1500–2000 years. Ca is significantly anti-correlated with reconstructed Nordic Seas sea-ice concentration over the past ~800 years (r = −0.37 annual; r = −0.46 with 5-year running mean; p < 0.001), consistent with fewer rainfall events during periods of extensive sea ice. - Model-based mechanisms: In the MPI-ESM1-2-LR past2k simulation, the SB pattern increases regional temperatures and precipitation and strengthens low-level moisture transport over the Greenland Sea toward Svalbard, consistent with observations and proxies. - Future changes: Across two CMIP6 models and three SSPs, Jun–Nov mean temperature and precipitation in Svalbard increase during 2015–2099, with a rise in the frequency of days with rainfall >10 mm by late century (SSP3-7.0). While model uncertainty remains for blocking projections, a wetter future climate with more extreme rainfall is indicated. - Spatial coherence: The spatial patterns linking Svalbard precipitation to Z500 and SSTs mirror those linking Scandinavian tree-ring temperatures to Z500 and North Atlantic SSTs, highlighting a common blocking-driven mechanism.

Findings demonstrate that atmospheric blocking over Scandinavia and, at times, extending toward the Urals, is the dominant driver of wet and warm extremes in Svalbard on daily to millennial timescales. The sediments from Linnévatnet provide a long-term record of when blocking was active, with suppressed Ca (and inferred reduced rainfall) during the coldest phase of the Little Ice Age (1600s–mid-1800s), coinciding with expanded Nordic Seas sea ice, cooler SSTs, and a larger local glacier (Linnébreen). The co-variability of Ca with Scandinavian tree-ring temperatures, and the anti-correlation with Nordic Seas sea ice, corroborate the mechanism: blocking fosters regional warmth, enhanced moisture transport, and rainfall that deposits coarse, Ca-rich sediments. Proxy-model agreement suggests that atmospheric variability, rather than regional SST variability alone, primarily drives hydroclimate excursions in the Greenland–Eurasian Arctic sector. In recent decades, increased blocking, declining sea ice, and Atlantification have amplified moisture availability, setting the stage for more frequent and intense rainfall events. Projections indicate continued warming and wetting, with potential intensification of Ural blocking in summer, implying heightened flood hazards for Svalbard. The results underscore the central role of blocking in past and future hydroclimatic extremes and provide a framework for evaluating changes under continued Arctic warming.

By integrating instrumental observations, an annually resolved 2,000-year lake sediment record, Scandinavian tree-ring reconstructions, and a paleoclimate model, the study establishes a persistent linkage between Scandinavian (and Ural) atmospheric blocking and warm, wet extremes in Svalbard. Rainfall events during blocking deposit coarse, Ca-rich sediments in Linnévatnet; this mechanism holds at event, seasonal, and centennial–millennial scales. A millennial-scale decline in precipitation culminated during the 1600s–mid-1800s and reversed around the 1850s, with unprecedented recent extremes. With ongoing warming, sea-ice decline, and Atlantification, future blocking episodes are expected to deliver more intense rainfall and flooding, elevating risks to communities and ecosystems. Future work should target improved simulation and detection of blocking in climate models, leverage daily-resolution multi-model ensembles across scenarios, and refine proxy constraints to reduce uncertainties in projecting extreme precipitation linked to blocking.

- Instrumental records are relatively short for characterizing multi-decadal to centennial variability in blocking and extremes. - There is a lack of direct pre-instrumental proxies for atmospheric blocking, necessitating indirect inference via sediments and tree rings. - Current climate models have considerable uncertainties in representing and projecting atmospheric blocking, complicating attribution of future precipitation changes to blocking characteristics. - Proxy differences in seasonal sensitivity (tree rings reflect primarily summer, sediments reflect summer–fall) and modulation by sea-ice presence can lead to discrepancies in inferred precipitation despite blocking. - The analysis of future extremes relies on a limited set of models and, for extremes, a single scenario (SSP3-7.0); broader multi-model, multi-scenario daily analyses are needed to test robustness.