Decadal trends of ocean and land carbon fluxes from a regional joint ocean‐atmosphere inversion. Issue 12 (31st December 2015)

• Record Type: Journal Article
• Authors: Steinkamp, K.; Gruber, N.
• Is Part Of: Global biogeochemical cycles. Volume 29:Issue 12(2015:Dec.)
• Page Start: 2108
• Page End: 2126
• Publication Date: 2015-12-31
• Subjects: atmospheric inversion -- carbon sinks and sources -- decadal trends -- carbon cycle; Biogeochemical cycles -- Periodicals; Electronic journals; 577.1405
• DOI: 10.1002/2014GB004907
• Languages: English
• ISSNs: 0886-6236

Abstract

Abstract: From 1980 until 2010, the combined CO2 sink strengths of ocean and land increased by nearly 50% (−0.55 Pg C yr −1 decade −1 ), but the spatial distribution of this trend is not well known. We address this by performing a joint cyclostationary ocean‐atmosphere inversion for the three decades 1980–1989, 1990–1999, and 2000–2008, using only carbon data from the ocean and atmosphere as constraints, i.e., without applying any prior information about the land fluxes. We find that in the inversion, most of the 30 year sink trend stems from the ocean (−0.44 Pg C yr −1 decade −1 ). The contribution of the terrestrial biosphere is commensurably smaller but has more decadal variability. First, the land sink strength intensified in the 1990s by 0.4 (±0.3) Pg C yr −1 compared to the 1980s but then weakened slightly by 0.2 (±0.4) Pg C yr −1 in the 2000s. The different land regions contributed very variedly to these global trends. While the northern extratropical land acted as an increasing carbon sink throughout the examined period primarily driven by boreal regions, the tropical land is estimated to have acted as an increasing source of CO2, with source magnitude and trend dominated by enhanced release in tropical America during the Amazon mean wet season. This pattern is largely unchanged if the oceanic inversion constraint, which is based on a stationary ocean circulation, is replaced by an estimate based on simulation results from an ocean biogeochemical general circulation model that includes year‐to‐year variability in the air‐sea CO2 fluxes and also has a trend (−0.07 Pg C yr −1 decade −1 ) that is at the very low end of current estimates. However, the land/ocean partitioning of the trend contribution is adjusted accordingly. Oceanic carbon data has a major impact on carbon exchange for all tropical regions and southern Africa but also for observationally better constrained regions in North America and temperate Asia. The European trend exhibits a strong sensitivity to the choice of the atmospheric CO2 network. Key Points: Land sink increased in the 1990s and then decreased in the 2000s Boreal land is the main contributor to trend in northern land sink CO2 release from the Amazon basin drives increased tropical source