Vulnerability of the North Water ecosystem to climate change

Abstract

High Arctic ecosystems and Indigenous livelihoods are tightly linked and exposed to climate change, yet assessing their sensitivity requires a long-term perspective. Here, we assess the vulnerability of the North Water polynya, a unique seaice ecosystem that sustains the world's northernmost Inuit communities and several keystone Arctic species. We reconstruct mid-to-late Holocene changes in sea ice, marine primary production, and little auk colony dynamics through multi-proxy analysis of marine and lake sediment cores. Our results suggest a productive ecosystem by 4400-4200 cal yrs b2k coincident with the arrival of the first humans in Greenland. Climate forcing during the late Holocene, leading to periods of polynya instability and marine productivity decline, is strikingly coeval with the human abandonment of Greenland from c. 2200-1200 cal yrs b2k. Our long-term perspective highlights the future decline of the North Water ecosystem, due to climate warming and changing sea-ice conditions, as an important climate change risk.

Fig. 1. Location and configuration of the North Water polynya.

a Marine and lacustrine core sites, main ocean currents, prehistoric human migration routes, and present-day distribution of little auk colonies. The extent of the polynya is defined as in ref. . b Example of late-June polynya configuration with a stable ice arch c. Example of late-June configuration in the absence of the Kane Basin ice arch. Satellite images from NASA EOSDIS. Several Inuit communities rely directly on the polynya resources today, including Qaanaaq, Siorapaluk, Qeqertat and Savissivik in Northwest Greenland, and Grise Fiord in Nunavut. Background map figures were created using Ocean Data View.

Fig. 2. Marine sediment core.

a Computerised tomography scan image of the Casq1 core. b CT number. Denser areas appear whiter in the CT scan image. c Total organic carbon (TOC) percentage weight. d TOC-normalised concentrations of the sea ice biomarker IP_25. e TOC-normalised concentrations of HBI III (Triene). f Sedimentation rates. g Modelled median age-depth relationships constructed in BACON for CASQ1 and CASQ1 BC (insert). Error bars and dashed lines represent 95% confidence intervals (for dates see Supplementary Table 1). The grey bar indicates the stratigraphic interval covered only by the box-core record.

Fig. 3. Lake sediment core.

a Core photograph showing laminations and shift from organic-poor to organic-rich sediments at c 150 cm core-depth. b Percentage (weight) of organic material loss on ignition (LOI). c Carbon to Nitrogen ratio (C:N). d Log of Cl counts based on X-ray fluorescence (XRF) scanning. e Log Cl:Ti XRF data. f Sedimentation rates. g Modelled median age-depth relationship constructed in BACON. Error bars and dashed lines represent modelled 95% confidence intervals (for dates see Supplementary Table 1).

Fig. 4. Holocene changes at the North Water as evidenced by marine and lake multi-proxy records.

The blocks shaded in grey indicate periods of polynya instability. The main phases of human prehistory in Northwest Greenland are represented by bars. Stars on the individual proxy time -series indicate points of significant change based on generalised additive model (GAM) statistics (see ‘Methods’ and Supplementary Figs. 7–15), and dashed vertical lines denote significant changes in more than one proxy at the same time. a Marine primary production as indicated by the fluxes of Chaetoceros resting spores. b Marine primary production as indicated by fluxes of diatoms (excluding Chaetoceros resting spores). c Fluxes of the sea ice biomarker IP₂₅. d Fluxes of the ice-marginal zone indicator HBI III (z triene). e Principal component 1 for the marine record proxies (Supplementary Fig. 4 and Supplementary Table 2). f Principal component 1 for the lake record proxies (Supplementary Figs. 5, 6 and Supplementary Table 2). g Principal Components 1 and 2 for the lake diatom assemblages (Supplementary Fig. 3). h Changes in Cadmium (Cd) to Titanium (Ti) ratios tracing the level of Cd-enrichment of the lake sediments by bird guano. The dark line shows a LOESS smoothed curve. i δ¹⁵. j A combination of total cholesterol flux, and the ratio of cholesterol to cholesterol plus β-sitosterol (greyscale) indicates relative bird colony size. RWP Roman Warm Period, DA Dark Ages cold period, MCA Medieval Climate Anomaly, LIA Little Ice Age, LD Late Dorset.

Fig. 5. Evolution of the North Water ecosystem and cultural transitions in Greenland.

a A stable and highly productive polynya is inferred from our records after 4400–4200 cal yrs b2k, coincident with the arrival of the first humans in Greenland and the first appearance and expansion of little auks in the area. b From 2700 to 800 cal yrs b2k, the polynya is unstable and reduced in extent, particularly after 2200 cal yrs b2k. This period spans a void in the human settlement of Greenland from c. 2200–1200 yrs b2k and absence/low abundance of little auks. c From c. 800 cal yrs b2k, a stable but low productive polynya is inferred and little auk colonies recover. During this time, there is a replacement of Late Dorset groups by the Thule Culture, the direct ancestors of modern Inuit. d Predicted disappearance of the polynya following the current trajectory of Arctic warming and sea-ice decline. BC Baffin Current, WGC West Greenland Current. Changes in WGC influence in the polynya region are based on ref. . Triangles represent drift ice. Shades of green represent the interpreted late-spring relative extent and productivity of the polynya (darker green corresponds to a more productive polynya and vice-versa). Background map figures were created using Ocean Data View.

Fig. 6. Climate variability in the Arctic and North Atlantic region during the late Holocene from selected high-resolution records.

a Air temperature anomalies from the Agassiz ice core record from. b Arctic Oscillation (AO) reconstruction from a Kara Sea record as presented in ref. . c North Atlantic Oscillation (NAO) reconstruction from a southwest Greenland lake record as presented in ref. . d Atlantic multidecadal variability (AMV) as reconstructed from a lake record from Ellesmere Island. RWP Roman Warm Period, DA Dark Ages Cold Period, MCA Medieval Climate Anomaly, LIA Little Ice Age, MoW Modern Warming. Grey bars indicate periods of polynya instability as inferred by our lake and marine records. The darker grey bar overlaps with the period of human abandonment of Greenland from c. 2200–1200 cal yrs b2k.

Fig. 7. Trends in atmospheric and sea ice conditions in the North Water region since the mid-twentieth century.

a Anomaly map showing the difference in May–September sea ice concentration following the extreme AO + anomaly year of 1989 compared to the extreme AO-anomaly year of 1996 and location of the weather station at Pituffik (Thule Air Base). Extreme AO + conditions resulted in 5% sea-ice cover reduction in the ice arch area and an increase in drift ice in the polynya area compared to AO- conditions. b Air temperature anomaly for May–September at Pituffik. c Winter AO index. d. Average sea ice concentrations during May–September for the two regions of interest based on satellite observations (see Methods for details); e, f Fluxes of the sea ice biomarkers IP₂₅ and HBI III, respectively, in the box-core marine sediment record. The grey bar denotes a regime shift in sea ice (declining %) and air temperatures (positive anomalies) after 1998.