Modelling Soil Carbon - Soil Carbon: Science, Management and Policy for Multiple Benefits - page 202

17 Modelling Soil Carbon

Eleanor Milne* and Jo Smith

Abstract
Models that describe the dynamics of soil organic carbon (SOC) can be useful tools when estimating
the impacts of land cover, land management and climate change on ecosystems. The development of
SOC models started with single-compartment models that assumed a constant decomposition rate. As
understanding of SOC dynamics improved, these were replaced by models with different compart-
ments with varying decomposition rate constants. Models that deal with the decomposition of SOC as
a continuum have been developed, but they require complex mathematics and are therefore less popu-
lar. Compartmentalized soil carbon models are at the core of complex models such as CENTURY and
DNDC, which describe nutrient turnover in the entire ecosystem both above and below ground. The
majority of such models have been developed using data from temperate ecosystems as studies on SOC
stock change in temperate areas outnumber those from tropical areas. Application to tropical and sub-
tropical areas therefore requires substantial parameterization and testing, and the availability of appro-
priate data sets remains a challenge.
The majority of SOC models were originally developed to model dynamics at the plot scale.
Empirical methods to estimate SOC stock changes at the landscape to regional scale have been used
successfully but are limited in their ability to predict future change. Linking dynamic SOC models to
large-scale data sets using geographic information systems offers a way of estimating SOC change across
large areas. However, account needs to be taken of the uncertainties associated with using multiple data
sets with different scales. Uncertainties arise due to errors in process descriptions and imprecision in
input data, and it is important that these are quantified against independent experimental measure-
ments of the appropriate output values. As the scale of simulation increases, the precision of input data
tends to decrease. It also becomes more difficult to quantify accurately the uncertainty in the simula-
tions. These issues need to be addressed if the results of simulations are to be applied in a rigorous way.

Introduction

properties of soils (Chapter 1, this volume).
As such, gaining an understanding of how
SOC levels will change under different envir-
onmental and anthropogenic circumstances
is important. Models that simulate the turn-
over of organic carbon (C) in the soil are

Soil organic matter (SOM, usually meas-
ured and expressed as content of soil or-
ganic carbon (SOC)) is a major determinant
of the physical, chemical and biological