Modelling Soil Carbon - Soil Carbon: Science, Management and Policy for Multiple Benefits - page 203

well established. They use mathematical re-
lationships to describe the transformation
of plant and animal material into organic
matter under varying climatic and geo-
graphical circumstances, which in turn af-
fect conditions in the soil (texture, moisture
content and temperature). Application of
SOC models is wide and varied. They can
be used at the small scale to give farmers an
indication of how much organic C to expect
following different land management prac-
tices. They can be used to estimate soil C
stocks across areas of similar land-use cat-
egories or land management activities where
measured data are limited. In addition, they
are also used in national greenhouse gas
(GHG) inventories and to make projections
about future soil C stocks under different
land use and climate scenarios in climate
change mitigation projects. In this chapter,
we explore the history of SOC models and
how they can be applied at different scales.
We then consider the uncertainties associ-
ated with their use and research gaps in our
current understanding.

C = fA c / k + (Co - fA c / k )e - kt (17.1)

Plant material decomposes rapidly in the
first few months of addition to the soil, and
more slowly from then on. This leads to the
deduction that at least two different frac-
tions of organic matter (and therefore or-
ganic C) exist, a labile fraction with a fast
turnover time and a more resistant, stable
fraction with a slow turnover time (Sauer-
beck and Gonzalez, 1977). In addition, the C
to nitrogen ratio in plant material varies,
making it unlikely that all plant material has
the same decomposition rate. Sugars and
proteins break down first, followed by cellu-
lose and finally lignin (van Keulen, 2001).
These two facts spurred the development of
models with multiple compartments, each
with different decomposition rates.
Many multicompartmental models were
developed to describe the decay of varying
types of plant material in the first few days
to months of addition to the soil (Hunt,
1977; Smith, 1979; McGill et al ., 1981).
Models of the longer-term turnover of or-
ganic C (10s to 100s of years) were fewer, as
long-term data sets were needed to develop
and parameterize models, and few of these
existed at the time (a problem which per-
sists today). The best-known early multi-
compartmental model is RothC, which was
first devised in the 1970s (Jenkinson and
Rayner, 1977) and is still in wide use
today, be it in a modified form (Coleman and
Jenkinson, 1996). RothC was developed us-
ing the long-term fertilizer experiments at
Rothamsted Agricultural Research Station
(UK), which provided unique data on the
turnover of C in soils over a period of nearly
150 years (Jenkinson and Rayner, 1977). A
challenge encountered in the development
of RothC, universal to all models of the
longer-term turnover of SOC, was finding data
sets to test the model which were independ-
ent of those that had been used to develop
it. The long-term experiments at Rothamsted
provided a means of doing this. In essence,
RothC has two compartments for incoming
plant C, decomposable material (DPM) and re-
sistant material (RPM). These then decay at
different rates, according to first-order pro-
cesses, to two more compartments; biomass

Brief History of SOC Models

In 1941, Jenny developed a model for the
turnover of nitrogen in soil (Jenny, 1941).
This was then elaborated by Henin and Du-
puis (1945), to describe the decomposition
of organic C. They realized that during the
initial decomposition of plant material, ni-
trogen was retained in the soil, whereas C
was lost rapidly as CO 2 , and therefore a
modification of Jenny's model was needed
(Jenkinson, 1990). By the end of the 1950s,
Henin and Dupius's model had been fitted
to varying data sets and modified further
(Bartholomew and Kirkham, 1960; Nye and
Greenland, 1960). In the model, the soil
contains a pool of humus ( fA c ), and yearly
inputs of C from plant material are assumed
to decompose, according to first-order kin-
etics, to add to this pool. The model can be
described by Eqn 17.1, where C is the or-
ganic C content in the soil, t is the time in
years and k is the fraction of this C decom-
posing each year (Jenkinson et al ., 1994).