Australian Journal of Experimental Agriculture, 1996,36, 1037-48
Simulation of a legume ley farming system
in northern Australia using the Agricultural
Production Systems Simulator
P. S. c a r b e r r y A , R. L. ~ c ~ o w n ~ , R. C. ~ u c h o w ~ , J. I? ~ i m e s ~ , M. E. prober tA, P. L. ~ o u l t o n ~
and N. P. DalglieshA
A Agricultural Production Systems Research Unit, PO Box 102, Toowoomba, Qld 4350, Australia.
CSIRO Division of Tropical Crops and Pastures, 306 Carmody Road, St Lucia, Qld 4067, Australia.
Summary. An innovative ley farming system,
involving cereal crops grown in rotation with pasture
legumes, has been tentatively adopted by farmers in the
semi-arid tropics of northern Australia. Yet, after more
than a decade of experimental research, the long-term
potential of this system remains uncertain. The
approach used to address this question has been to use
the APSIM (Agricultural Production Systems
Simulator) model in conjunction with historical~climate
records to simulate system performance. Thus, the
objectives of this paper were to describe APSIM, to test
its performance against data from cropping systems
experiments, and to use it in assessing the long-term
consequences of alternative farming practice in this
region of northern Australia.
APSIM is able to simulate the soil carbon, water and
nitrogen balances arising from interactions between
different crops and pastures grown in rotation. In this
paper, APSIM was configured to simulate either
conventional rotations of sorghum or maize crops, or
crops grown in rotation with Stylosanthes hahata
(Verano) ley pastures. In the latter case, APSIM
simulates the establishment of a Verano pasture sward,
its growth and death, and its effects on subsequent
cereal crops. The crop is either kept free of weeds or,
alternatively, an understorey of volunteer legume can
establish to form an intercrop where the crop and
pasture compete for resources. Simulation of crop and
pasture residues can encompass either their retention on
the soil surface and decomposition over time, or their
complete removal from the system (as hay).
In comparisons of simulations against limited
experimental data, APSIM was able to reproduce the
measured yields from sorghum, maize and Verano
grown either as sole crops, as intercrops, or in rotations
of several years. Likewise, a simulation analysis using
APSIM of several cropping options for Katherine,
Northern Territory, resulted in the preferred outcome
reflecting current farming practices in the region. This
preferred option, a combination of legume hay and
sorghum grain production, was shown to be superior in
terms of both gross margin returns and long-term soil
fertility status.
It was concluded that APSIM now provides a useful
tool with which farming systems in northern Australia
can be further explored.
Additional keywords: legume ley, farming system, rotation, intercrop, simulation model, APSIM.
Introduction
The establishment of successful primary industries in
the semi-arid tropics of northern Australia has been the
goal of Governments , research institutions and
entrepreneurial farmers for at least the past 5 decades
(Chapman et al. 1996). Despite considerable expenditure
on agricultural research and development over this
period, until recently, the region's achievements have
been modest . However, in recent years, primary
producers in the Northern Territory have been
successfully producing dryland crops and livestock
(Price et al. 1996). They have achieved this success
using innovative systems based on conservation tillage
and ley farming that have emerged from research
undertaken in the 'Top End' over the past 15 years.
A major focus of the research to enhance the
feasibility of dryland farming systems has been on the
use of conservation t i l lage technology and the
integration of livestock with pasture legumes as leys
(McCown et al. 1985; Chapman et al. 1996). In these
systems, grain crops are grown during the wet season
and are sown directly into a mulch of chemically-killed

1038 P. S. Carberry et al.
pasture which provides both protection from high soil
surface temperature and a source of mineralisable
nitrogen (N). The legume pasture re-establishes from
hard seed during and after the grain crop. Cattle graze
native grass or improved pastures in the wet season and
crop residues and legume pasture during the dry season.
Research on conservation tillage and ley farming
systems in northern Australia over the past 15 years has
ranged from large farming systems trials to intensive
physiological studies on crop development and growth.
The purpose of this paper is to describe an approach
where much of this information has been integrated into
a dynamic simulation capacity of this farming system.
This capacity, in the form of the APSIM model
(McCown et al. 1995), provides the opportunity for
exploring not only today's issues of agronomic
management but also the longer-term prospects for
sustaining agricultural production in this climatic zone of
Australia.
The specific objectives of this paper are to:
(i) describe APSIM as configured for a ley farming
system, (ii) test the model 's performance against
cropping systems experimental data, and
(iii) demonstrate the utility of the model through an
assessment of the long- term consequences of alternative
farming practices in northern Australia.
APSIM cropping systems model
General description
The APSIM (Agricultural Production Systems
Simulator) model has been developed as a flexible
software system for simulating agricultural production
systems. A detailed description of APSIM, including its
capabilities, design features, structure, user interface and
the derivation of its main biological and environmental
modules is provided by McCown et al. (1995).
In providing the capability to simulate system
performance, APSIM has the soil as its central unit. It is
through the soil that the cumulative effects of climate,
crops, pastures and agricultural management affect
productivity (McCown et al. 1995). This concept was
achieved by separating system processes into distinct
modules which can all interact with a common set of soil
process modules. This interaction between modules
occurs only via a communications 'engine' and a
standard interface. Thus, APSIM contains an extensive
library of modules describing different soil, biological
and managerial processes, the combination of which, at
the time of APSIM configuration, defines the system
simulation capability. Selection of module combination
is user-defined and so may differ between users and
applications.
The combination of APSIM modules used to simulate
a ley pasture system in northern Australia is described in
the following sections. APSIM V1.O engine and protocol
were used to interface each module.
Sorghum, maize and Stylosanthes hamata modules
Many of the biological modules within APSIM have
been derived from existing stand-alone models that
simulate crop, pasture or soil processes in Australia and
elsewhere (McCown et al. 1995). This is the case with
the sorghum, maize and Stylosanthes modules used in
this study. The 2 crop modules resulted from redesigning
and re-engineering the AUSIM (Australian SIMulator)
models described by Carberry and Abrecht (1991),
which themselves were derivatives of the CERES-Maize
model (Jones and Kiniry 1986). The Stylosanthes
hamata (cv. Verano) module resulted f rom
re-engineering an existing model described by
Carberry et al. (1992, 1993).
Whereas the re-engineered sorghum module, APSIM-
CSSAT VO.l, largely remains true to the process
descriptions used in its parent model, the new generation
maize module, APSIM-Maize V1.O, was developed from
a generic APSIM crop template (McCown et al. 1995).
The resulting difference in process description is chiefly
in how soil water is extracted by the crop, with APSIM-
Maize V1.O using the approach where the rate of soil
water extraction is defined for each soil layer (Passioura
1983; Meinke et al. 1993). Likewise, APSIM-Stylo
V1.O, being based on the crop template, resembles its
parent in all but the water extraction routines.
Soil watel; nitrogen, temperature and residue
breakdown modules
The soil water (APSIM-SoilWat V1.0), soil N
(APSIM-SoilN V1.0), and surface residue breakdown
(APSIM-Residue V1.O) modules used in this study have
been described in some detail by Probert et al. (1996).
Their applicability to the soils of northern Australia is
underlined by the fact that the validation datasets used
by Probert et al. (1996) to test APSIM included data
from Katherine, Northern Territory.
The soil surface temperature (APSIM-SoilT VO.l)
module in APSIM consists of a 'patched-in' version of
the TCIMAX model of Ross et al. (1985). This module
predicts daily maximum soil surface temperature for
different levels of surface cover. The original model was
validated for soils at Katherine by Carberry and Abrecht
(1991) and this validation has been confirmed for
APSIM-SoilT VO.l (data not shown).
Arbitrator (intercropping) module
APSIM permits any number of biological modules to
be 'grown' together as mixtures or intercrops, each of
which compete for light, water and N resources. APSIM
allocates resources using rules defined in the arbitrator
(APSIM-Arbit V1.O) module. Carberry et al. (1996)
have described this capability in detail.
Managel; fertilisel; irrigation and report modules
In this study, APSIM was configured with APSIM
V1.O utility modules that permitted implementation of