In many respects the global carbon cycle has constituted a paradigm for GAIM studies: (a) it is an interdisciplinary field crosscutting several IGBP core project activities, (b) it is strongly coupled to the physical climate system, and (c) the main processes controlling the storage and cycling of carbon are sufficiently well known to make it amenable to global scale modeling. A comprehensive model of the carbon cycle requires components of the carbon systems in the atmosphere, the ocean and on land. A fundamental question concerns the accurate quantification of the fate of the anthropogenic CO2 in the global carbon cycle - where is it being stored and for how long? The global carbon cycle is also strongly coupled to the physical climate system, however, the different feedback mechanisms between the two systems are still only poorly understood.
The Carbon Cycle Model Linkage Project (CCMLP), is an attempt to investigate this complex interplay by means of global modeling studies involving various combinations of different comprehensive model components of the various carbon cycle subsystems. The main emphasis during the first phase of CCMLP lies on the terrestrial carbon cycle. Various global simulation experiments are being conducted within CCMLP, which explore the different global and regional responses to a series of external forcing factors, such as the historical increase in the atmospheric CO2 concentration since 1800, changes in land-use and impacts induced by the climate variations. As an additional feature also the behavior of the carbon isotopes (13C, 14C) is being investigated in various model simulations. In order to facilitate the model intercomparison, each of the experiments is being conducted with a strict simulation protocol, specifying in detail the model settings (e.g. grid resolution, simulation period, etc.), the input forcing data (e.g. atmospheric CO2 concentration, climate data) and the output data for the subsequent analysis.
One of the fundamental objectives of the CCMLP concerns the quantification of the climatic feedback effects on the carbon cycle. For this an investigation is being conducted, which explores the response of the carbon cycle models on the ENSO and decadal time scales by forcing them with the observed historical climate. The results can then be intercompared with variations in the atmospheric CO2 concentration and isotope ratios as recorded at the global monitoring station networks. The observed growth rate of atmospheric CO2 shows significant variations on interannual time scales which are correlated to climate fluctuations (Fig. 6.1). Presumably a substantial fraction of these fluctuations is caused by variations of the CO2 exchange fluxes between the atmosphere and the terrestrial biosphere. This hypothesis is explored by means of simulations with 4 comprehensive, global prognostic terrestrial biogeochemical models (TBMs) driven by observed monthly climate datasets (temperature and precipitation) over the time period 1900-1994. The TBMs include the HRBM model of the University of Giessen, the FBM model of the University of Frankfurt, the TEM model of the University of New Hampshire and the Marine Biological Laboratory, Woods Hole and the SILVAN model of the Max-Planck-Institute for Meteorology in Hamburg.
Each of the models was subjected to the following experiment protocol:
In these experiments atmospheric CO2 concentration was held fixed, and the models employed potential vegetation only.
The TBM simulations are able to explain a substantial fraction of the observed anomalous CO2 source flux, even though the latter presumably also includes variations caused by the oceans. This is illustrated in Figure 6.2 which shows the globally integrated net terrestrial CO2 source flux from the combined experiment (D4), together with the anomalous source flux derived from the atmospheric CO2 record (from Figure 6.1). In particular the variations associated with strong El Ni–o-Southern Oscillation (ENSO) events (e.g. 1972-3, 1982-3 and 1987), during which the land biota acts as a net CO2 source, are reproduced in the simulations, although with an absolute magnitude varying by a factor of two among the models. The dominant process (photosynthesis, autotrophic or heterotrophic respiration, feedback effects due to changes in soil water, etc.) that induces the fluctuations of the net CO2 flux, however, is found to be different from model to model.
Figure 6.1: Observed interannual variations in the global carbon cycle. Upper panel: Fluctuations of the atmospheric growth rate of CO2 determined from the average of the seasonally adjusted records of the Mauna Loa and South Pole stations. The dashed line is the growth rate that would result from an atmospheric balance taking into account the documented CO2 inputs from fossil fuel and changes in land use together with the uptake rates computed by an ocean and a terrestrial model. Lower panel: Anomalous, presumably climate driven, CO2 source implied by the difference between the solid and the dashed line shown in the upper panel.
Figure 6.2: Globally integrated CO2 source flux simulated by four TBMs of the CCMLP (red line), together with the anomalous source flux shown in Figure 6.1 (blue dashed line).
As an example, Figure 6.3 shows the anomalous net surface-air CO2 flux predicted by the four models in the ENSO year 1987, which demonstrates that also the predicted regional patterns are quite different among the models.
None of the models exhibits in sufficient magnitude the anomalous net CO2 uptake that was observed after the Pinatubo volcanic eruption in 1992 and 1993. Hence, while the models short term sensitivity to climate variations in tropical regions appears realistic, this might not be the case in midlatitudes where the climate anomaly in 1992-93 was most predominant.
A series of additional, short term simulations are currently being performed over the 1979-1994 period, using as climate drivers the newly available meteorological reanalyses from the ECMWF weather forecast model. These simulations allow an assessment of the sensitivity of the results with respect to the different climate driver data. Further experiments are also scheduled to explore the concurrent effects of land-use changes and of the rising CO2 concentration. It is also expected, that additional models will participate in the simulations in a future phase of the project.
Figure 6.3: Net CO2 source flux predicted by the four TBMs for the ENSO year 1987. The uppermost panels show the annual average of the anomalous climate fields that were used as forcing fields in the model simulations.
