To test the influence of these perturbations in the carbonate factory on atmospheric CO2 and climate, we carried out a suite of simulations with the GEOCLIM model [Donnadieu et al., 2006, 2009; Godderis et al., 2008]. The numerical model GEOCLIM couples a 3D general circulation model (FOAM) to a model of the biogeochemical cycles of carbon, alkalinity, oxygen and phosphate (COMBINE) (see Donnadieu et al.  for a full description). The COMBINE model has been upgraded since its original version [Godderis and Joachimski, 2004]. The geometry of the ocean has been changed, and the model now includes 9 oceanic boxes divided into, 2 high-latitude oceans (each including a photic zone and a deep ocean reservoirs), a low- to middle-latitude ocean (with a photic zone, thermocline and deep oceanic reservoirs) and an epicontinental sea (with a photic zone and a deep epicontinental reservoirs), and one box for the atmosphere. Exchange water fluxes between boxes are tuned to fit the present-day vertical distribution of the oxygen, DIC and alkalinity in the low-to-middle-latitude ocean. The atmospheric component of FOAM is a parallelized version of NCAR's Community Climate Model 2 (CCM2) with the upgraded radiative and hydrologic physics incorporated in CCM3 v. 3.2. The atmosphere runs at R15 spectral resolution (4.5° × 7.5°) with 18 levels. We use FOAM in mixed-layer mode, i.e., the atmospheric model is linked to a 50 m mixed-layer ocean, which parameterizes heat transport through diffusion, mainly for computation time considerations. A full coupling between COMBINE and FOAM cannot be achieved owing to excessive computation times. Hence we adopt an indirect coupling that employs lookup tables from a catalog of simulations. For a given paleogeography we run a suite of FOAM experiments (30 years for each to reach the steady state) in which the only factor that varies is atmospheric CO2. Atmospheric CO2 is tested over a range from 4200 to 200 ppm, which covers all plausible atmospheric CO2 content for the Mesozoic time period. We linearly interpolate between experiments to obtain climatic variables (temperature and runoff) for any CO2 value within our CO2 range. These climatic parameters allow the calculation of the weathering rates within the 1920 grid elements using weathering laws linking climatic factors (temperature and runoff) to CO2 consumption through silicate weathering. We want to emphasize here that phosphorus flux from the continents are also calculated through the climatic dependency of the continental weathering. Fixing the CO2 degassing to a given constant value, the numerical feedback loop between FOAM and COMBINE is run until a steady state PCO2 is reached. Each continental configuration is thus finally characterized by a steady state atmospheric PCO2. The MLJT land-ocean distribution used here is derived from a synthesis of paleomagnetic data, hot spot tracks and geologic constraints [Besse and Courtillot, 2002; Dercourt et al., 1993]. Surface types are set to average model surface characteristics (i.e., deciduous forest) and the Earth's orbit around the Sun is circular (eccentricity = 0) and the Earth's obliquity is 23.5° (this setting leads to an equal annual insolation for both hemispheres). Solar luminosity is assumed to evolve through time according to the stellar evolution models [Gough, 1981]. As described by Donnadieu et al. , the long-term steady state atmospheric CO2 level characterizing the Middle Jurassic paleogeography is 700 ppm (see Figure 1 also). In other words, 700 ppmv of CO2 are required in the Middle Jurassic for the calculated silicate weathering to balance the prescribed solid Earth degassing. Note that because there is no consensus about the Earth degassing rate for the last 200 million years, we choose to keep it constant at its present-day value, assuming the present-day equality between CO2 consumption by continental silicate rock weathering (6.8 × 1012 moles CO2/yr today [Gaillardet et al., 1999]) and CO2 degassing. In contrast to previous investigations of the past climatic and geochemical evolution of the Earth system with the GEOCLIM model, GEOCLIM is run dynamically in this study. This means that the long-term steady state of the carbon cycle, requiring the close balance between silicate weathering and solid Earth degassing, is not prescribed. Over the course of the simulation, and depending on the dynamics of the alkalinity, and organic and inorganic carbon cycles, significant departures between silicate weathering and solid Earth degassing are calculated [Walker et al., 1981].
Figure 1. Model results simulating the effects of the collapse and the recovery of the carbonate platforms during the Middle Late Jurassic transition. (a) Fdown function as described in the main text. (b) Changes in atmospheric CO2 content. Note that the first collapse is more diluted in times than the second one based on geological data. Nevertheless, the effect remains the same when increasing the time for the collapse (see dashed curve). (c) Changes in the burial flux coming from tropical and subtropical platforms. Note the remaining flux of the tropical shallow carbonate platform during the crisis (i.e., 20% here). We have also plotted the mean saturation state. (d) Distribution of the carbonate species. (e) Changes in sedimentary calcium carbonate burial fluxes. The black curve represents the neritic platform carbonate burial flux, the red curve represents the pelagic shelfal carbonate burial, and the green curve represents “abiotic” carbonate production flux (see text). Note that the neritic/global carbonate production ratio is 65% in this run. Platform collapse lasts for 1 million years; 20% of the preperturbation platform burial flux is conserved during the crisis in order to account for the likely possibility that carbonate platforms still exist in specific environments.
Download figure to PowerPoint
 We use the GRISLI ice sheet model (ISM) to reconstruct ice sheet topography during the MLJT. GRISLI simulates the dynamics of grounded ice as well as ice shelves and ice stream regions. Inland ice deforms according to the stress balance using the shallow-ice approximation. Ice shelves and dragging ice shelves (ice streams) are described following MacAyeal . This 3-D model has been developed and validated over Antarctica by Ritz et al. . Ritz et al.  provide a comprehensive description of the model. Three data sets are used from the FOAM simulations as the input to the GRISLI model: surface temperature from the hottest month, mean annual temperature and mean annual precipitation (i.e., snow and rain). The climatological variables are interpolated to the fine grid topography of the ISM using constant lapse-rate corrections (for the temperature) and an exponential law (for the precipitation). The first two of these data sets are used to compute ablation at the ice sheet surface, and the last two to compute accumulation. Therefore, the feedback of the ice on the GCM-simulated climate is not taken into account but the lapse-rate corrections still capture much of the interactions. GEOCLIM provides time evolution of atmospheric CO2 and climate during the Middle Late Jurassic Transition. These scenarios are then used to force the GRISLI model (Figure 2).
Figure 2. Sensitivity of CO2 and induced sea level fall. (a) Atmospheric CO2 simulated in GEOCLIM zoomed on the 3.5–5 Ma period corresponding to the Middle Late Jurassic Transition. Black, red, and green solid lines represent results for a percentage of total carbonate burial originating from neritic platform like producers being 40, 65, and 85%, respectively. Twenty percent of the preperturbation carbonate platform burial flux is conserved during the crisis for these runs. Top and bottom dashed lines of the same color represent runs in which 40 and 0% of the preperturbation carbonate platform burial flux is conserved during the crisis. Blue lines come from a run in which settings are the same as those from the red lines except the duration of the crisis that is diminished to last 100 kyr. (b) Sea level fall simulated using the GRIZZLI model forced by GEOCLIM output. Color codes remain the same as those described in Figure 2a. Blue areas denote CO2 ranges for which the sea level fall is found to be around 12, 34, and 58 m.
Download figure to PowerPoint