Arctic fjords are hotspots of marine carbon burial, with diatoms playing an essential role in the biological carbon pump. Under the background of global warming, the proportion of diatoms in total phytoplankton communities has been declining in many high-latitude fjords due to increased turbidity and oligotrophication resulting from glacier melting. However, due to the habitat heterogeneity among Svalbard fjords, diatom responses to glacier melting are also expected to be complex, which will further lead to changes in the biological carbon pumping and carbon sequestration. To address the complexity, three short sediment cores were collected from three contrasting fjords in Svalbard (Krossfjorden, Kongsfjorden, Gronfjorden), recording the history of fjord changes in recent decades during significant glacier melting. The amino acid molecular indicators in cores K4 and KF1 suggested similar organic matter degradation states between these two sites. In contrast to the turbid Kongsfjorden and Gronfjorden, preserved fucoxanthin in Krossfjorden indicated a continuous increase in diatoms since the mid-1980s, corresponding to a 59 % increase in biological carbon pumping, as quantified by the delta C-13 of sedimentary organic carbon. The increasing biological carbon pumping in Krossfjorden is further attributed to its hard rock types in the glacier basin, compared to Kongsfjorden and Gronfjorden, which are instead covered by soft rocks, as confirmed by a one-dimensional model. Given the distribution of rock types among basins in Svalbard, we extrapolate our findings and propose that approximately one-fifth of Svalbard's fjords, especially those with hard rock basins and persistent marine-terminated glaciers, still have the potential for an increase in diatom fractions and efficient biological carbon pumping. Our findings reveal the complexity of fjord phytoplankton responses and biological carbon pumping to increasing glacier melting, and underscore the necessity of modifying Arctic marine carbon feedback to climate change based on results from fjords underlain by hard rocks.

Rising atmospheric CO2 is intensifying climate change but it is also driving global and particularly polar greening. However, most blue carbon sinks (that held by marine organisms) are shrinking, which is important as these are hotspots of genuine carbon sequestration. Polar blue carbon increases with losses of marine ice over high latitude continental shelf areas. Marine ice (sea ice, ice shelf and glacier retreat) losses generate a valuable negative feedback on climate change. Blue carbon change with sea ice and ice shelf losses has been estimated, but not how blue carbon responds to glacier retreat along fjords. We derive a testable estimate of glacier retreat driven blue carbon gains by investigating three fjords in the West Antarctic Peninsula (WAP). We started by multiplying ~40 year mean glacier retreat rates by the number of retreating WAP fjords and their time of exposure. We multiplied this area by regional zoobenthic carbon means from existing datasets to suggest that WAP fjords generate 3,130 tonnes of new zoobenthic carbon per year (t zC/year) and sequester >780 t zC/year. We tested this by capture and analysis of 204 high resolution seabed images along emerging WAP fjords. Biota within these images were identified to density per 13 functional groups. Mean stored carbon per individual was assigned from literature values to give a stored zoobenthic Carbon per area, which was multiplied up by area of fjord exposed over time, which increased the estimate to 4,536 t zC/year. The purpose of this study was to establish a testable estimate of blue carbon change caused by glacier retreat along Antarctic fjords and thus to establish its relative importance compared to polar and other carbon sinks.