Experiment Protocol, Outline for
Manuscript, and Action Items
October 12, 1999
Modified by A. D. McGuire based on
Jena meeting (5 - 7 October 1999)
IMPORTANT NOTE
The following is the Grand Slam
protocol for accounting for land use (croplands only) in historical
simulations.
For the C4MIP exercise, we now have
crops and pasture historical distribution from Ramankutty and Goldewijk.
Therefore, the following protocol should be used as a guide line only.
3.0 Model modifications/LU protocol
The land use data set assigns each
0.5 degree grid cell as either natural vegetation (0) or agriculture (1) for
the historical period 1860 through to 1995. Note that agricultural land is a
subset of the total land appropriated for human use, and only includes
croplands. Other land uses, which
include pastures and important activities such as forestry/logging, are not
considered in this simulation. Here we wish to understand and model the
following:
3.1 Conversion from natural vegetation to cultivation
In general, the methodology of
simulation 3 builds on the approach of Houghton et al., 1983 (Houghton et al.
1983. Ecological Monographs 53:235-262).
3.1.1 Temperate/Boreal forest conversion
Fate
of carbon in vegetation at time of clearing:
Left dead in soils
(from ROOTS) 33 %
1st year
conversion loss (CONVERSION FLUX) 40
%
10 year product
pool (PROD10) 20 %
100 year product
pool (PROD100) 7 %
total
100 %
3.1.2 Tropical Forest conversion
Fate
of carbon in vegetation at time of clearing:
Left dead in soils
(from ROOTS) 33 %
1st year
conversion loss (CONVERSION FLUX) 40
%
10 year product
pool (PROD10) 27 %
100 year product
pool (PROD100) 0 %
total
100 %
3.1.3 Grasslands/Tundra
Fate of carbon in
vegetation at time of clearing:
Left dead in soils
(from ROOTS) 50 %
1st year
conversion loss (CONVERSION FLUX) 50
%
10 year product
pool (PROD10) 0 %
100 year product
pool (PROD100) 0 % total 100 %
3.1.4 Shrublands, Woodlands & Savannas
If a group uses a mixed pixel
approach and disaggregates these biomes into forest and grassland fractions,
then the appropriate conversion estimates given above should be used for each
fraction. Otherwise use the estimates below for the fate of carbon in
vegetation at time of clearing:
Left dead in soils
(from ROOTS) 50 %
1st year
conversion loss (CONVERSION FLUX) 40
%
10 year product
pool (PROD10) 10 %
100 year product
pool (PROD100) 0 %
total 100
%
3.1.5 Additional Considerations
Each model should use the fractions
and decay rates given by Houghton if they are not already explicitly
modelled. For example, if a model
explicitly estimates below ground vegetation carbon (roots), Houghton’s fraction
of vegetation carbon left dead in soils may be ignored. In this case, the
estimate of above ground vegetation carbon is assigned to the product decay
pools using the given ratios (e.g., 40:27 for the case of tropical forest
conversion). Therefore, potential
inconsistencies are avoided and reliance on the book keeping approach is
minimized. Each group should decide how
to assign the vegetation carbon left dead in soils to particular soil
pools. Decisions on the seasonal timing
of inputs into the soil and product pools are left to each modeling group.
3.1.6 Agricultural NPP
The annual agricultural NPP is
divided into above ground and below ground biomass, i.e., harvest vs. residue,
using the ratio 40:60 based on information in Table 2 of Malmstrom et al.
(1997. Global Biogeochemical Cycles 11:367-392). The above ground biomass is
placed in the agricultural productivity pool (AGPROD1). This pool decays to the atmosphere in one
year. The below ground biomass enters
the soil. Each group should decide on
the seasonal timing of these inputs into the soil and AGPROD1.
3.3 Abandonment of cultivated sites
Each model should grow back vegetation biomass from the extant state of
the grid cell at the time of abandonment.
Because, the time required for the ecosystem to fully recover is not
prescribed, time to recovery will be site specific.
3.4 Land Use Fluxes and Pools to be tracked during the simulation
At
the end of each year each group should define four fluxes for each grid cell:
the CONVERSION FLUX and three PRODUCT FLUXES. The product fluxes represent the
sum of the CO2 releases from the AGPROD1, PROD10 and PROD100 pools,
which are also tracked for each grid cell.
Releases from the AGPROD1, PROD10, and PROD100 pools are calculated as a
linear decay of the initial carbon inputs into these pools over 1, 10 and 100
years, respectively. For example, the
annual release from the PROD10 pool represents 10% of the initial carbon
entering the PROD10 pool during the previous 10 years. Therefore, 10 years after conversion to
agriculture the initial carbon entering the PROD10 pool will have been released
to the atmosphere. Please note that the
releases from AGPROD1 (AGPROD1FLUX), PROD10 (PROD10FLUX), and PROD100
(PROD100FLUX) are to be calculated for each grid cell. Each group should decide on the seasonal
timing of the CONVERSION FLUX and CO2 releases from the AGPROD1,
PROD10 and PROD100 pools.