1
Adaptation to Climate Change of Wheat Growing in South
Australia: Analysis of Management and Breeding Strategies
*
Qunying Luo
1
, William Bellotti
2
, Martin Williams
1
and Enli Wang
3
1
Department of Geographical and Environmental Studies, University of Adelaide,
South Australia, 5005, Australia
2
School of Agriculture & Wine, University of Adelaide,
South Australia, 5371, Australia
3
CSIRO Land and Water, GPO Box 1666, Canberra, ACT2601, Australia
Abstract
Evaluation of adaptive management options is very crucial for successfully dealing with negative
climate change impacts. Research objectives of this study were (1) to determine the proper N
application rate for current practice, (2) to select a range of synthetic wheat (Triticum aestivum L.)
cultivars to expand the existing wheat cultivar pool for adaptation purpose, (3) to quantify the potential
impacts of climate change on wheat grain yield and (4) to evaluate the effectiveness of three common
management options such as early sowing, changing N application rate and use of different wheat
cultivars derived in (2) and given in the APSIM-Wheat model package in dealing with the projected
negative impacts for Keith, South Australia. The APSIM-Wheat model was used to achieve these
objectives. It was found that 75kg ha-1 N application at sowing for current situation is appropriate for
the study location. This provided a non-limiting N supply condition for climate change impact and
adaptation evaluation. Negative impacts of climate change on wheat grain yield were projected under
both high (-15%) and low (-10%) plant available water capacity conditions. Neither changes in N
application level nor in wheat cultivar alone nor their synergistic effects could offset the negative
climate change impact. It was found that early sowing is an effective adaptation strategy when initial
*
Corresponding author current organisation and contacts
NSW Department of Primary Industries
Postal address: PO BOX 100, Beecroft, NSW, 2119 Australia
Email: qunying.luo@2003.adelaide.edu.au (perpetual email), qunying.luo@dpi.nsw.gov.au
Phone: 61 2 9872 0117 Fax: 61 2 9871 6941
2
soil water was reset at 25 mm at sowing but this may be hard to realise especially since a drier
environment is projected.
Key words: wheat grain yield, climate change, impact assessment, adaptation evaluation, early
sowing, cultivars choices, N application level
1. Introduction
Impact and adaptation are key components of climate change risk assessment. While
the former issue has been extensively studied the latter still needs to be
comprehensively investigated. Compared with the large number of impact assessment
studies, adaptation evaluation is seldom adequately assessed, even though a few
studies have considered these two issues together by using process-oriented crop
models (Wang et al, 1992; Qureshi and Iglesias, 1994; Seino, 1995; Brklacich and
Stewart, 1995; Baethgen and Magrin; 1995; Delécolle et al., 1995; Bayasgalan et al.,
1996; Rosenzweig and Iglesias, 1998; Howden et al., 1999; Reyenga et al., 1999a,
Torriani et al., 2007). However, most climate change risk assessment studies so far
end with impact assessment (some examples are Aggarwal and Sinha, 1993; Barry
and Geng, 1995; Tubiello et al., 1995; Pilifosova et al., 1996; Karim et al., 1996;
Reyenga et al., 1999b; Luo et al., 2003; 2005a, b; Van Ittersum et al., 2003). To some
extent, the role of impact assessment is to set the scene for adaptation evaluation.
Without addressing adaptation, climate change risk assessment is incomplete. The
ultimate purpose of climate change risk assessment is to identify adaptation strategies
and evaluate their effectiveness in counteracting the negative climate change impacts
for the sustainable development of a specific sector/region. Several factors have
impeded the balanced development of adaptation studies compared with impact
assessment. One concerns the considerable uncertainties in regional climate change
3
risk assessment. The other concerns the difficulty in quantifying certain management
options.
A few studies dealt with adaptation issues in Australia. Wang et al. (1992) assessed
the interactive impacts of increase in CO
2
concentration and in temperature on wheat
yields in Victoria. They suggested that doubling of pCO
2
to 700ppm would increase
yields by 28% to 43%, but that simultaneous increases in temperature of 3
o
C would
decrease yields by 25% to 60% using current cultivars or cause a substantial increase
in yield if a late-maturing variety from Queensland was used. Howden et al. (1999)
quantified the potential impacts of climate change on wheat production and explored
the benefit of early sowing at nine wheat production areas in Australia for the period
centred on 2070 based on the CSIRO (1996) climate change scenarios, with
atmospheric CO
2
set at 700ppm. Reyenga et al. (1999a) assessed the possible impacts
of climate change and increased atmospheric pCO
2
on wheat production in southeast
Queensland by applying the same source of climate change scenarios as Howden et al.
(1999). Management options such as nitrogen application and cultivar maturities were
evaluated in dealing with climate change risk. It was found that N application
enhanced wheat yield across all scenarios considered and that late maturity and early
maturity varieties generally have lower wheat yields than standard varieties.
Similar studies were conducted in Europe. Torriani et al. (2007) quantified the
potential impacts of changes in mean climate and in climate variability on crop yields
in Switzerland. Increasing growing degree days (equivalent to the use of later maturity
cultivar) and later sowing were evaluated in adapting to negative climate change
impacts. In contrast to above mentioned process-oriented modelling approach,
Reidsma et al. (2007) addressed adaptive capacity issue in Europe by adopting a
statistical modelling approach.
4
This study aims to quantify the potential impacts of climate change on wheat grain
yield and to evaluate the effectiveness of a range of management options in dealing
with climate change risks in South Australia (SA) by coupling the outputs of a higher
spatial and temporal resolution climate model with a wheat model. To achieve this
aim, two ancillary studies were carried out before the core study. One is a sensitivity
study of N application rate at sowing. The purpose of this ancillary study is to
determine an appropriate N application rate to avoid haying-off and to achieve a non-
limiting N supply condition for climate change impact and adaptation studies. The
other is the identification of synthetic cultivars through changes in vernalisation and
photoperiod coefficients used by the wheat model to expand the cultivar pool for
adaptation evaluation in addition to existing wheat cultivars included in the APSIM-
Wheat package.
2. Methodologies
2.1 Study site
This study focused on Keith, which is located in the southeast of South Australia and
is one of the major wheat production areas in this state. This location receives mid-
high annual rainfall (468mm) with average growing season (May-Oct. inclusive)
rainfall of 315mm under a Mediterranean climate.
2.2 Method
The Agricultural Production System sIMulator (APSIM)-Wheat model (version 4.1)
was used in the two ancillary studies and the core study (climate change impact
assessment and adaptation evaluation). The APSIM-Wheat module has been
described in detail elsewhere (Keating et al., 2003; Luo, 2003). The performance of
5
APSIM-Wheat in the Australian environment (Keating et al., 2003) and in the South
Australian environment (Luo, 2003; Yunusa et al., 2004) has been evaluated. The
physiological effects of increased atmospheric CO
2
on wheat production were
included in the simulations. Modifications have been made to the Wheat module
through changes to radiation use efficiency (RUE), transpiration efficiency (TE) and
to critical nitrogen concentration (CRC) based on experimental data (Reyenga et al.,
1999a; Luo, 2003).
2.3 Climate and soil data
Climate data
Historical daily climate data (solar radiation, maximum temperature, minimum
temperature and rainfall) for the period of 1906-2005 for Keith were gathered from
SILO patched point dataset (PPD) at http://www.nrw.qld.gov.au/silo/ppd. This period
of historical climate data was directly used by the APSIM-Wheat model in the two
ancillary studies. Historical climate data for the period of 1958-2005 were used by a
stochastic weather generator (LARS-WG) to produce 100-year baseline climate and
climate changes scenarios for the quantification of climate change impacts and
adaptive options. The rationale for this procedure is to produce climate change
scenarios with both changes in mean climate and in climate variability considered,
which is an important issue in the field of climate change impact assessment.
Semenov et al. (1998), Semenov (2007, 2008) and Qian et al. (2004) applied and
evaluated the performance of LARS-WG across a wide range of environments in the
world. Figure 1 details the usage of historical climate data in this study.
Figure 1
