Modelling Soil Organic Matter Dynamics – Global Challenges - Sustainable Management of Soil Organic Matter - page 40

based explicitly on 14 C carbon dating, and default estimation procedures
have been published by Falloon et al . (1998a). The inclusion of a small
(~5-10% of total C), totally inert, and very old SOM fraction in the Roth-C
model was based on the need to reconcile the radiocarbon age of the soil
with rates of SOM turnover inferred from bomb-C measurements. In
analysing 14 C dynamics for a tropical and temperate soil, Trumbore (1993)
concluded that the temperate soil was best represented by a relatively even
distribution of organic matter in active, slow and passive pools while the
tropical soil was best represented as comprised of mostly actively cycling
carbon with a small, very recalcitrant pool.
Seeking a closer correspondence between theoretical and measurable
fractions will continue to be a major thrust of SOM research. It should be
realized that both theoretical and measurement-based depictions of SOM
components are, in essence, 'models', i.e. abstractions and simplifications
of reality. As new analytical procedures are developed that give functionally
meaningful results, with consistent and repeatable patterns, they will
doubtless influence the definition of modelled pools. Similarly, the
expanded use of isotopic tracers and the availability of a richer set of field
experiments will allow more robust and constrained testing of conceptually
based models.
Regional Applications
Climate change issues, specifically questions about: (i) the role of soil
carbon as a potential part of the 'missing' terrestrial carbon sink; (ii) climate
change impact studies; (iii) the need to quantify the emissions/sinks of CO 2
from soils as part of national greenhouse gas inventories; and (iv) interest in
the potential of soil C sequestration as a greenhouse gas mitigation option
have helped spur the development and application of SOM modelling
systems at regional and global levels (e.g. Burke et al ., 1991; Donigian et al .,
1994; King et al ., 1997; Falloon et al ., 1998b; Paustian et al ., 2000; Schimel
et al ., 2000). The approaches used are similar in most cases and rely on the
linkage of simulation models with geographically distributed databases,
typically maintained in geographic information systems (GIS) which
contain model-driving variables and initial conditions (i.e. inputs) and help
to manage and display model output (Fig. 2.1). Typically, model-driving
variables (i.e. climate, soils, vegetation and land use) define a unit area
of land, or polygon, having a unique combination of driving variables to
which the model is applied. Approaches vary in the spatial and temporal
resolution employed, the degree to which sub-polygon distributions of soil,
vegetation and land use are represented, the source of the input data and
the model being used. Such approaches have enabled information and
understanding of SOM, originally derived at the field level, to be scaled up