[A-CG24] Science in the Arctic Region

[ACG24-11] Comparison of CO₂ fluxes estimated by top-down and bottom-up methods -- a case study at Yakutsk, Siberia --

*Kumiko TAKATA^1,2,3, Prabir Patra³, Dmitry Belikov^2,1, Ayumi Kotani⁴, Junko Mori^3,1, GTMIP Group, Kazuhito Ichii³, Tazu Saeki³, Yosuke Niwa⁵, Kazuyuki Saito³, Takeshi Ohta⁴, Shamil Maksyutov², Toshinobu Machida², Alexander Ganshin^6,7, Ruslan Zhuravlev^6,7, Atsuko Sugimoto⁸, Shuji Aoki⁹ (1.Arctic Environment Research Center, National Institute of Polar Research, 2.CGER, National Institute for Environmental Studies, 3.Japan Agency for Marine-Earth Science and Technology, 4.Nagoya Univ., 5.Meteoroligical Research Institute, JMA, 6.Tomsk State University, 7.Central Aerological Observatory, 8.Hokkaido University, 9.Tohoku University)

Keywords:CO2 balance, boreal forest region, topdown and bottomup hethod, tower observation

Carbon balance of the forested ecosystem is widely recognized as an important component in climate change research, but remains uncertain at the same time. Attempts have been made, recently, to understand the origin of the uncertainty by comparing estimates of carbon budgets with bottom-up and top-down methods. In the Green Network of Excellence (GRENE) Arctic Climate Change Research Project (hereafter as, GRENE Arctic project), terrestrial and atmospheric observations are conducted in the Arctic regions, where observational data were not available otherwise, e.g., in Siberia. At the same time, CO₂ fluxes are estimated using process-based terrestrial ecosystem,models and atmospheric CO₂ inversion models as a part of the GRENE Arctic project.
In the terrestrial sub-program, observation on energy-water-carbon balances are conducted in the Circum-Arctic, and the fluxes are estimated by a suite of terrestrial ecosystem models at four super-sites in the GRENE-TEA model intercomparison project (GTMIP) (Miyazaki et al., 2015). In the greenhouse gas sub-program, atmospheric CO2 concentration is measured at high accuracy using aircrafts and at surface stations and top-down/inverse modeling is performed for estimating regional CO₂ fluxes. We have compared the CO₂ fluxes estimated from tower observation at Yakutsk, Siberia with the CO2 flux estimates by the land-surface models for Yakutsk and CO₂ surface fluxes estimated by inverse models around the Yakutsk region (area ~500 x 500 km²). The Net Ecosystem Production (NEP) or Net Biome Production (NBP) are considered for this analysis at monthly time intervals over the period of 1980 - 2012 (from 2004 - 2011 for flux observation).
We find that the seasonal cycle of CO₂ flux consists of a large drawdown in June-August from the atmosphere, and weaker emissions or absorptions in other months. This result agrees well among the models and observation. As for the long-term changes, the model variation is smaller in summer (June-August) than for the annual values. That is because respiration takes a dominant part of CO₂ flux in winter, that would have large uncertainty both for the observation and the model estimation. Thus the large uncertainty in CO₂-flux estimates in winter would affect the large fluctuation for the annual values. The year-to-year variations in summer by some models agree, at least in part, with the observation, but the reasons for the agreement/disagreement should carefully be investigated. At first, the difference in the horizontal scale represented by each method should be considered. Besides, different treatments of forest fire are identified as one of the possible causes for model-to-model differences. The extreme climate, such as very humid or hot-and-dry summer, resulted in year-to-year changes in NEP/NBP with the tower observation, but some models do not agree to those changes. Making thorough examination of each case is required to identify the causal process of the disagreements and to reduce the uncertainty in CO₂ balance.

References
Miyazaki, S., K. Saito et al., GRENE-TEA model intercomparison project (GTMIP): overview and experimental protocol for Stage 1, Geosci. Model. Dev., 8, 1-16, doi:10.5194/gmd-8-1-2015