Performance of Carbon Flux Models Across the ABoVE Domain Using Eddy Covariance Measurements

Abstract:

The Arctic region is experiencing unprecedented environmental changes, such as thawing permafrost and increased wildfires that are releasing more and more carbon dioxide and methane into the atmosphere. Several of these land-atmosphere exchange processes can be quantified with eddy covariance tower measurements and ecosystem models; however, there is often a mismatch between the flux estimates derived from these two techniques. Mismatch between models and measurements creates uncertainty in whether ecosystems and regions act as carbon sources or sinks and how they will evolve in the future. This study evaluates net ecosystem exchange (NEE) measurements from a handful of eddy covariance tower sites representing five ecoregions across the ABoVE domain and compares them with flux estimates from models, including process-based bottom-up, empirical and top-down models. We found that while model outputs were able to identify trends of carbon uptake in the summer and carbon release in the winter, the strength of uptake and release was inconsistent across models and had significant regional variations. We highlight these comparisons and report on standard deviation, correlation coefficient and root mean square error (RMSE) using a Taylor diagram. Our findings suggest that top-down models were closest to the measurements at the majority of sites whereas the empirical class of models also performed well. For example, at a site in the Southern Arctic ecoregion, top-down models showed an R value of 0.96 and an RMSE of 0.008. Ongoing analyses attempt to identify several causes of mismatch by focusing on model performance during the growing season and by comparing results to previous studies at similar geographic locations. We believe that the findings from this study have the potential to inform and improve ecosystem models, especially across the ABoVE domain.