Meltwater discharge from tidewater glaciers impacts the adjacent marine environment. Due to the global warming, tidewater glaciers are retreating and will eventually terminate on land. Yet, the mechanisms through which meltwater runoff and subglacial discharge from tidewater glaciers influence marine primary production remain poorly understood, as data in close proximity to glacier fronts are scarce. Here, we show that subglacial meltwater discharge and bedrock characteristics of the catchments control the phytoplankton growth environment inside the fjord, based on data collected in close proximity to tidewater glacier fronts in Kongsfjorden, Svalbard from 26 to 31 July 2017. In the southern part of the inner fjord, glacial meltwater from subglacial discharge was rich in fine sediments derived from erosion of Devonian Old Red Sandstone and carbonate rock deposits, limiting light availability for phytoplankton (0.6 mg m −3 Chl a on average, range 0.2-1.9 mg m −3). In contrast, coarser sediments derived from gneiss and granite bedrock and lower subglacial discharge rates were associated with more favourable light conditions facilitating a local phytoplankton bloom in the northern part of the inner fjord with mean Chl a concentration of 2.8 mg m −3 (range 1.3-7.4 mg m −3). In the northern part, glacier meltwater was a direct source of silicic acid through weathering of the silica-rich gneiss and granite bedrock. Upwelling of the subglacial freshwater discharge plume at the Kronebreen glacier front in the southern part entrained large volumes of ambient, nutrient-rich bottom waters which led to elevated surface concentrations of ammonium, nitrate, and partly silicic acid. Total dissolved inorganic nitrogen transported to the surface with the upwelling of the subglacial discharge plume has a significant potential to enhance summer primary production in Kongsfjorden, with ammonium released from the seafloor being of particular importance. The transition from tidewater to land-terminating glaciers may, thus, reduce the input of nutrients to the surface layer with negative consequences for summer productivity.

A ntarctic krill Euphausia superba, a key species in Southern Ocean food webs 1 , plays a central role in ecosystem processes and community dynamics of apex predators, and is the target of a commercial fishery 2 . Krill spawn in late spring and their larvae develop during summer, autumn and under the ice in winter to emerge as juveniles in the following spring. The newly spawned eggs sink from the surface to up to 1,000 m depth where they hatch and the developing larvae actively swim upwards to feed in the upper water column 1 . Larval Antarctic krill have low lipid reserves, which are insufficient to support long periods of starvation. Therefore, winter, when primary production is minimal, is assumed to be a critical bottleneck for larval krill development and hence recruitment to the adult population 3,4 . Present hypotheses suggest that high algal biomass in winter sea ice enhances larval krill winter-feeding conditions and growth [5][6][7][8] . This implies that larvae have access to this high algal biomass within the sea ice. However, recent observations 9,10 indicate that the linkage between sea ice and krill recruitment success is not as direct as has been suggested. The timing of ice-edge advance and annual ice-season duration is highly variable, and does not necessarily show a clear link to krill recruitment in the following year ( Supplementary Figs. 1 and 2). Along the Antarctic Peninsula, adult krill have a five to six year population cycle with oscillations in biomass exceeding an order of magnitude 9 . According to bioenergetics models, part of the variability is due to interannual variation in reproductive output 11 as well as autumn blooms that may govern the possible overwinter survival rate of larvae 11,12 . In the Bransfield Strait, three krill winter surveys have shown that krill abundance is an order of magnitude higher than in summer, regardless of concurrent sea-ice conditions 10 A dominant Antarctic ecological paradigm suggests that winter sea ice is generally the main feeding ground for krill larvae. Observations from our winter cruise to the southwest Atlantic sector of the Southern Ocean contradict this view and present the first evidence that the pack-ice zone is a food-poor habitat for larval development. In contrast, the more open marginal ice zone provides a more favourable food environment for high larval krill growth rates. We found that complex under-ice habitats are, however, vital for larval krill when water column productivity is limited by light, by providing structures that offer protection from predators and to collect organic material released from the ice. The larvae feed on this sparse ice-associated food during the day. After sunset, they migrate into the water below the ice (upper 20 m) and drift away from the ice areas where they have previously fed. Model analyses indicate that this behaviour increases both food uptake in a patchy food environment and the likelihood of overwinter transport to areas where feeding conditions are more favourable in spring.

Pigmented microalgae inhabiting snow and ice environments lower the albedo of glacier and ice-sheet surfaces, significantly enhancing surface melt. Our ability to accurately predict their role in glacier and ice-sheet surface mass balance is limited by the current lack of empirical data to constrain their representation in predictive models. Here we present new empirical optical properties for snow and ice algae and incorporate them in a radiative transfer model to investigate their impact on snow and ice surface albedo. We found ice algal cells to be more efficient absorbers than snow algal cells, but their blooms had comparable impact on surface albedo due to the different photic conditions of their habitats. We then used the model to reconstruct the effect of ice algae on bare ice albedo spectra collected at our field site in southern Greenland, where blooms dropped the albedo locally by between 3 and 43%, equivalent to 1–10 L m $^{-2}$ d $^{-1}$ of melted ice. Using the newly parametrized model, future studies could investigate biological albedo reduction and algal quantification from remote hyperspectral and multispectral imagery.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.