Agriculture Reference
In-Depth Information
25
years duration) have not been estab-
lished for all environments, soils, land uses
and climates.
The early developments of SOM models,
therefore, focused on sites where long-term
experimental data were easily available. For
example, the CENTURY model was devel-
oped first for grassland systems, against data
from long-term sites in the northern Great
Plains of the USA (Parton
et al
., 1988); the
RothC model was developed for arable,
grassland and forestry systems using data
from the Rothamsted classical experiments
(Jenkinson and Rayner, 1977). Because of
this constraint, following the prevalence of
long-term experiments, more SOM models
have been developed in temperate regions
and on mineral soils than in tropical sys-
tems or on highly organic soils.
If we aim to use SOM models to deter-
mine global changes in soil C stocks, it is im-
portant that models be expanded to include
the wide range of soils and conditions found
around the world. Jenkinson
et al
. (1991)
used RothC to simulate changes in global C
stocks and noted that RothC was not suit-
able for use in anaerobic conditions. Smith
et al
. (2007) also noted that RothC could not
be used to simulate SOM turnover in highly
organic soils. The establishment of SOM-
NET, a network of long-term experiments
and models (Powlson
et al
., 1998), opened
up the way to further developments of SOM
models by providing access to a wider range
of long-term experiments to allow adequate
testing of long-term simulations.
Models have been developed to include
the impact of nutrient limitations on SOM
turnover. For example, RothC was expanded
to include nitrogen (N) limitation on de-
composition if the organic and mineral N
content of the soil was inadequate to main-
tain a stable C:N ratio in the decomposed
material (SUNDIAL, Bradbury
et al
., 1993).
The CENTURY model includes N, phos-
phorus (P) and sulfur (S) limitation in a
similar way, but uses a more complex ap-
proach that assumes immobilization of nu-
trients by decomposition of structural plant
material and allows mineralization of nutri-
ents from metabolic plant residue and the
SOM pools according to different stable C to
nutrient ratios for each organic matter pool
and using a range of values for the stable ra-
tios (Parton
et al
., 1988).
The impact of anaerobic conditions on
decomposition has been described very
simply in the ECOSSE model (developed
from RothC and SUNDIAL by Smith
et al
.,
2010) using a moisture rate modifier for
decomposition that shows a linear decline
in its value above field capacity. Similarly,
anaerobic conditions are described in
CENTURY as a linear decline in the rate of
decomposition above a fitted parameter de-
fining partial anaerobic conditions, up to
another fitted parameter defining the soil
as fully anaerobic (Parton
et al
., 1993;
Chimner
et al
., 2002). In flooded anaerobic
wetlands, particularly those undergoing
cycles of wetting and drying, the effect of
alternative electron acceptors in the soil
(i.e. NO
−
, Mn
4+
, Fe
3+
and SO
2−
ions) be-
comes important and influences strongly
the production of methane during anaer-
obic decomposition (van Bodegom and
Stams, 1999). In some models, this effect
is accounted for by simulating the pattern
of decline of reduction (redox) potential
(e.g. Cao
et al
., 1995). The Wetland-DNDC
model estimates the redox potential of the
soil layers in the saturated zone from ob-
served patterns of variation in soils under
fluctuating water table or continuously
submerged conditions (Zhang
et al
., 2002).
Other models simulate the behaviour of al-
ternative electron acceptors directly, so
avoiding the need to estimate redox poten-
tial (Matthews
et al
., 2000).
In order to simulate highly organic
soils, some authors have developed new
models, specifically designed to simulate
the different processes occurring in organic
soils (e.g. Frolking
et al
., 2001). An alterna-
tive approach modifies existing models for
mineral soils for use in organic soils; this
has the advantage that the adapted model
can be applied seamlessly across the full
range of soil types found in the landscape,
but the models may be less well adapted to
describe processes in organic soils than a
purpose-built model. The ECOSSE model
uses the description of anaerobic conditions
and nutrient limitation, with added rate
Search WWH ::
Custom Search