Disaggregating the Carbon Exchange of Degrading Permafrost Peatlands Using Bayesian Deep Learning
Norbert Pirk, Kristoffer Aalstad, Erik Schytt Mannerfelt, François Clayer, Heleen de Wit, Casper T. Christiansen, Inge Althuizen, Hanna Lee, Sebastian Westermann
Abstract
Abstract Extensive regions in the permafrost zone are projected to become climatically unsuitable to sustain permafrost peatlands over the next century, suggesting transformations in these landscapes that can leave large amounts of permafrost carbon vulnerable to post‐thaw decomposition. We present 3 years of eddy covariance measurements of CH 4 and CO 2 fluxes from the degrading permafrost peatland Iškoras in Northern Norway, which we disaggregate into separate fluxes of palsa, pond, and fen areas using information provided by the dynamic flux footprint in a novel ensemble‐based Bayesian deep neural network framework. The 3‐year mean CO 2 ‐equivalent flux is estimated to be 106 gCO 2 m −2 yr −1 for palsas, 1,780 gCO 2 m −2 yr −1 for ponds, and −31 gCO 2 m −2 yr −1 for fens, indicating that possible palsa degradation to thermokarst ponds would strengthen the local greenhouse gas forcing by a factor of about 17, while transformation into fens would slightly reduce the current local greenhouse gas forcing.