Abstract
Abstract. We describe the design and evaluation of a large ensemble of coupled climate–carbon cycle simulations with the Earth system model of intermediate complexity GENIE. This ensemble has been designed for application to a range of carbon cycle questions, including the causes of late-Quaternary fluctuations in atmospheric CO2. Here we evaluate the ensemble by applying it to a transient experiment over the recent industrial era (1858 to 2008 AD). We employ singular vector decomposition and principal component emulation to investigate the spatial modes of ensemble variability of oceanic dissolved inorganic carbon (DIC) δ13C, considering both the spun-up pre-industrial state and the transient change. These analyses allow us to separate the natural (pre-industrial) and anthropogenic controls on the δ13CDIC distribution. We apply the same dimensionally-reduced emulation techniques to consider the drivers of the spatial uncertainty in anthropogenic DIC. We show that the sources of uncertainty related to the uptake of anthropogenic δ13CDIC and DIC are quite distinct. Uncertainty in anthropogenic δ13C uptake is controlled by air–sea gas exchange, which explains 63% of modelled variance. This mode of variability is largely absent from the ensemble variability in CO2 uptake, which is rather driven by uncertainties in thermocline ventilation rates. Although the need to account for air–sea gas exchange is well known, these results suggest that, to leading order, uncertainties in the ocean uptake of anthropogenic 13C and CO2 are governed by very different processes. This illustrates the difficulties in reconstructing one from the other, and furthermore highlights the need for careful targeting of both δ13CDIC and DIC observations to better constrain the ocean sink of anthropogenic CO2.
Highlights
We employ singular vector decomposition and principal component emulation to investigate the spatial modes of ensemble variability of oceanic dissolved inorganic carbon (DIC) δ13C, considbon emissions, especiallyEgaivretnhreSceyntsotbesemrvations that the oceanic carbon instance, while csionukpilsedwcelaikmeantien–Sgc(acLrbieeoQnnucceyrceeleestmaol.d, 2el0s0c7o)n. sFios-r tently predict a weakened efficiency of the ocean–terrestrial ering both the spun-up pre-industrial state and the transient carbon sink under future-emissions scenarios, they do so change
We show that the sources of uncertainty related to the uptake of anthropogenic δ13CDIC and DIC are quite distinct
By varying sea ice diffusivity (SID) we attempt to represent uncertainty introduced by brine rejection on Antarctic Bottom Water (AABW)
Summary
We utilise a coupled carbon cycle–climate configuration of GENIE (version 2.7.7). The physical model comprises the 3-D frictional geostrophic ocean model GOLDSTEIN (at 36 × 36 × 16 resolution) coupled to a 2-D energy moisture balance model of the atmosphere (EMBM) and a thermodynamic/dynamic sea ice model (Edwards and Marsh, 2005; Marsh et al, 2011). Atmospheric heat diffusivity (m2 s−1) Atmospheric moisture diffusivity (m2 s−1) Atlantic–Pacific moisture flux scaling Clear skies OLR reduction (Wm−2) Isopycnal diffusivity (m2 s−1) Reference diapycnal diffusivity (m2 s−1) Power law for diapycnal diffusivity depth profile Ocean inverse drag coefficient (days) Wind scale factor Sea ice diffusivity (m2 s−1) Fractional vegetation dependence on vegetation carbon density (m2 kgC−1) Base rate of photosynthesis (kgC m−2 yr−1) Vegetation respiration activation energy (J mol−1) Leaf litter rate (yr−1) Soil respiration temperature dependence (K) PO4 half-saturation concentration (mol kg−1) Initial proportion of POC export as recalcitrant fraction e-folding remineralisation depth of non-recalcitrant POC (m) Rain ratio scalar Thermodynamic calcification rate power Initial proportion of CaCO3 export as recalcitrant fraction e-folding remineralisation depth of non-recalcitrant CaCO3 (m) Iron solubility Air–sea gas exchange parameter. We refer to the resulting uncertainty as parametric uncertainty, but are cognisant of the fact that the broad ranges applied are in some cases deliberately larger than the “true” parametric uncertainty associated with the pre-industrial state
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
