Elementa: Science of the Anthropocene (Dec 2015)

Temporal and spatial variability in sea-ice carbon:nitrogen ratios on Canadian Arctic shelves

  • Andrea Niemi,
  • Christine Michel

DOI
https://doi.org/10.12952/journal.elementa.000078

Abstract

Read online

Abstract To enhance the accuracy of carbon cycling models as applied to sea ice in the changing Arctic, we analyzed a large data set of particulate organic carbon (POC) and nitrogen (PON) measurements in first-year bottom sea ice (n = 257) from two Arctic shelves, the Canadian Arctic Archipelago and Beaufort Sea shelf, including dark winter and spring seasonal measurements. Wide ranges of sea-ice POC:PON ratios were observed during both the dark winter (12–46 mol:mol) and spring (3–24 mol:mol) periods. Sea-ice POC:PON ratios and chlorophyll a concentrations were significantly higher in the Archipelago versus the Beaufort Sea shelf (p < 0.01), yet there was a highly significant relationship between sea-ice POC and PON during spring for both shelves (r2 = 0.94). POC:PON ratios were not consistent over the range of measured POC and PON concentrations, justifying the use of a power function model to best describe the relationship between POC and PON. Distinct relationships between POC:PON ratios and chlorophyll-based biomass were observed for the dark winter and the spring: dark winter sea-ice POC:PON ratios decreased with increasing sea-ice biomass whereas spring POC:PON ratios increased with increasing sea-ice biomass. The transition from the dark period to the spring growth period in first-year sea ice represented a distinct stoichiometric shift in POC:PON ratios. Our results demonstrate that the Redfield ratio has limited applicability over the four-order of magnitude range of biomass concentrations observed in first-year sea ice on Arctic shelves. This study emphasizes the need for variable POC:PON stoichiometry in sea-ice biogeochemical models and budget estimates, in particular at high biomass concentrations and when considering seasonality outside of the spring period in first year ice. The use of a power function model for POC:PON relationships in sea ice is also recommended to better constrain carbon estimates in biogeochemical sea-ice models.

Keywords