Abstract
Few multi‐year observations of CO 2 gas exchange exist over melting first year sea ice (FYI). We report 4 years of CO 2 flux data during the melt season in Dease Strait, Nunavut, Canada. Utilizing a dried, closed‐path eddy covariance system on an island within the strait, we identify two statistically distinct CO 2 flux melt regimes: a pre‐pond period, where melt has commenced but there is no standing water; and a pond period, a surface which encompasses bare ice and melt water, intermittently draining through open features within the icescape. Fluxes transition from small in magnitude during pre‐pond period (mean: −0.50 mmol m −2 day −1 ; range: −2.06 to +2.20 mmol m −2 day −1 ) to larger in magnitude and more variable during pond period (mean: −0.56 mmol m −2 day −1 ; range: −4.74 to +2.99 mmol m −2 day −1 ). As seasonal thermodynamic regimes of FYI can be identified using seasonal shifts in C‐Band synthetic aperture radar HH‐backscatter, we investigated the utility of using backscatter to model CO 2 fluxes during the pond period. Along with wind speed and air temperature, the inclusion of HH‐backscatter, a proxy for liquid water, improved observed versus predicted pond period CO 2 fluxes R 2 from: 0.10 to 0.27. This regression model was applied to Dease Strait, where estimates of uptake during the melt pond period ranged from 1.13 × 10 −3 to 1.77 × 10 −3 Tg C. When considering regions of melt ponds across polar seas, this work highlights that this seasonal period could contribute to non‐negligible portions of the annual Arctic carbon budget.
Funding Information
  • AUTO21 Network of Centres of Excellence
  • Natural Sciences and Engineering Research Council of Canada
  • Polar Knowledge Canada
  • ArcticNet
  • Marine Environmental Observation Prediction and Response Network
  • Social Sciences and Humanities Research Council of Canada
  • Canada Foundation for Innovation

This publication has 65 references indexed in Scilit: