Carbon cycle models

The terrestrial carbon cycle model, TRIFFID, is a dynamic global vegetation model (DGVM). A full model description can be found in Cox [2001].

The net flux of carbon into the terrestrial biosphere is determined by the small differences between the large uptake of carbon due to growth and the large release due to respiration. In the long term, if the biosphere was in equilibrium, these fluxes would balance and the total carbon stored in the biosphere would not change. However, in the short term they do not balance: daily, seasonal and inter-annual changes exist. On top of that, changes to the long-term carbon storage due to anthropogenic activities, or natural changes in climate, affect this balance too. In order to capture this behaviour of the terrestrial carbon cycle, TRIFFID has two key components:

1. five types of vegetation (known as Plant Functional Types: PFTs) compete for coverage of the land. They are: Broadleaf tree, Needleleaf tree, C3 grass, C4 grass and Shrub. The absence of vegetation is assumed to be bare soil. (C3 and C4 are different types of grass, which respond differently to changes in climate and carbon dioxide concentration). This competition means that areas of the model can change their vegetation cover - for example, forests can come and go over many hundreds of years depending on climate fluctuations. This is shown schematically in the figure below.
   

   In addition to the exchanges of carbon, the vegetation cover affects surface properties of the climate model, such as roughness, stomatal conductance and albedo. Changes in these properties form another mechanism by which the terrestrial carbon cycle can feedback onto climate behaviour.

2. within each PFT, the growth (gross primary productivity: GPP) and respiration are determined. They depend on climatological conditions such as temperature and soil moisture. The difference between these two terms is called the net primary productivity (NPP) and is the amount of carbon taken up by the vegetation.

   This carbon is then used by the plant to contribute towards root, wood or leaf mass, or to extend its areal coverage. Carbon is then lost from the plant through turnover of the root/wood/leaf or disturbance of its area. This carbon then enters the soil, and eventually is broken down by microbes and released back to the atmosphere as soil respiration. The difference between NPP and soil respiration is the net ecosystem productivity (NEP) which is the total amount gained or lost by the biosphere. This is shown schematically in the figure below.

   

The total amounts of carbon held in the vegetation and soil after a long run of the model under pre-industrial conditions are 495 GtC and 1170 GtC respectively. IPCC estimates of real-world values are 600 and 1600 respectively, but there are large uncertainties in these estimates (there are no direct, global measurements) and the modelled values are within the possible range of values. Typically in the model, GPP is about 120-125 GtC/yr, plant respiration is about 62 GtC/yr and soil respiration is about 60-61 GtC/yr. For a single year, NEP can be as large as 3-4 GtC (either uptake or release), but in the longterm it is very close to zero. This indicates that the model is in a state of balance.

As for vegetation cover, the figures below (panels a-f) show the fraction of each model point that is occupied by each PFT. The model simulates the major biomes of the world. E.g. the Amazon rainforest and other tropical forests, the northern evergreen forests of Canada, Siberia and Scandinavia and the deserts of Africa, Asia and Australia.

The last figure above (panel g) shows how these fractions agree with the observed vegetation cover from the IGBP dataset [ref Loveland and Belward, 1997]. The blue, pink and black colours denote where there is good agreement between the model and observations - these cover most of the world. There are only a few, small, areas of red, yellow and orange which denote poorer agreement.

The overall results from the model show that it is good enough to use in our coupled climate-carbon cycle experiments. Plans to improve the processes represented by the model include treatment of nitrogen, and forest fires.