Source DocumentPrevious PageTable Of ContentsNext Page

Drought experienced by Australian wheat: current and future trends

Karine Chenu1 and Scott Chapman2

1 The University of Queensland, Queensland Alliance for Agriculture and Food Innovation (QAAFI), 203 Tor Street, Toowoomba,
CSIRO Plant Industry, Queensland Bioscience Precinct, 306 Carmody Road, St. Lucia, 4067, Australia


Drought frequently limits Australian wheat production and the expected future increase in extreme temperatures and rainfall variability will further challenge productivity. A modelling approach captured plant x environment x management interactions to simulate drought patterns experienced by wheat crops for representative locations and managements across the wheatbelt. Simulations were performed for 123 years of ‘current’ climate (1889-2011) and four future climatic scenarios for 2030, accounting for predicted shifts in sowing dates and soil water content at sowing.

Across the wheatbelt, four main drought-environment types have been identified for the ‘current’ climates, ranging from stress-free/light-stress to severe stress with terminal drought. The frequency of occurrence of these environment types greatly varied across seasons and locations, and these variations tended to accentuate in future climate. Frequency of the most-severe (terminal) drought type was predicted to increase by 10 to 77% on average across the wheatbelt for the studied future climatic scenarios, with some high spatial variability between regions. While the Wet & Low emission scenario had no substantial impact on drought frequency in most regions, the Dry & High emission scenario more than doubled severe-drought frequency in several regions and substantially impacted the others (>25% increase). Further study is needed to assess alternative options (genotype and management) to reduce future drought impact, in particular for the Dry & High emission scenario.


Water deficit, environment characterisation, mega-environment, climate change, modelling, APSIM.


While growth in population and urban/industrial water demands are increasingly limiting water supply for agricultural production, improving crop yield remains a key strategy globally to meet projected demand (Borlaug and Dowswell, 2005). Recent trends in increased extreme temperature and rainfall events are forecast to amplify with climate change (e.g. IPCC, 2007; Battisti and Naylor, 2009; Coumou and Rahmstorf, 2012).

The Australian wheatbelt is characterised by large variation in inter-annual rainfalls (e.g. Williams et al., 2002; Potgieter et al., 2002) and by soils ranging from shallow sands to deep clays. Extending the modelling approach applied by Chenu et al. (2011), this paper aims to analyse the drought patterns that wheat crops experience in Australia in current and future climatic scenarios.

Materials and Methods

The APSIM crop model (e.g. Keating et al., 2003) was used to simulate crop drought pattern for the quick/medium maturity variety ‘Hartog’ across the Australian wheatbelt over 122 years. To represent the Australian wheat cropping system, the major production areas (West, South, South-East, East) were divided into 22 regions (Fig. 1) and 60 locations, each representing between 130 000 and 230 000 hectares of planted wheat (averaged data from 1975-2000, 2005 and 2006; source: Australian Bureau of Statistics). The simulations used weather records for 1889-2010 (SILO Patched Point Dataset; Jeffrey et al., 2001; and for four future climatic scenarios for 2030. Future climates had been generated for a wet and dry scenarios (Global Climate Models (GCM) ECHAM5 and GFDL-21, respectively) and a high and low CO2 emission scenarios (A1FI and A2, respectively; IPCC, 2007). They had been calculated by QCCCE based on historical data (baseline from 1889 to 2010), using Consistent Climate Scenarios projections (Version 1.1).

For each climatic scenario and each location, an assessment of the sowing window and soil water content at sowing was performed. The sowing windows were determined as periods that resulted in flowering time occurring during low-risk periods of extreme temperatures (less than 10% chance of frost (Tmin < 0C) and less than 30% chance of heat (Tmax > 35C); Zheng et al., 2012). Five representative sowing dates and soil water contents at sowing were calculated for each site and each climatic scenario, based on preliminary simulations beginning 1-Nov with a fallow (Fig. 2). Those initial conditions were used to simulate the drought pattern for the current and each of the future climatic scenarios.

The daily drought pattern was calculated for each crop based on a water-deficit index (“water supply/demand ratio”) which indicates the degree to which the soil water extractable by the roots (“water supply”) is able to match the potential transpiration (“water demand”). For each environment (site x year x sowing date x initial soil water x climatic scenario), this daily index was centred around flowering and averaged over 100Cd from emergence to 450Cd after flowering. In a previous study (Chenu et al., submitted), four main drought-environment types (Fig. 2) were identified across the wheatbelt (using a similar methodology applied for sowing windows defined based on local practices). Simulations of the present study were classified based on which environment type they were the most similar to, i.e. based on the minimum sum of squared differences for the considered water-deficit pattern compared to the water-deficit pattern of the environment types. Analyses were done with R (R Dev. Core Team, 2011).

Fig. 1 The 22 regions (coloured and named in each box) and 60 sites (dots) used in the simulations to represent four cropping areas (‘West’, ‘South’, ‘South-East’ and ‘East’) of the Australian wheatbelt. Figure adapted from Chenu et al. (submitted).

Fig. 2 Simulated water-stress index for four environment types identified across the Australian wheatbelt for the period 1889-2011. The stress index corresponds to the ratio of soil water supply to crop water demand and is represented as a function of cumulative thermal time relative to flowering, from crop emergence to 450Cd after flowering. Figure adapted from Chenu et al. (submitted).

Results and discussion

Four main drought-environment types were identified by Chenu et al. (submitted) across the wheatbelt (Fig. 2): stress-free or light-stressed environments (ET1), mild water shortages during grain filling that were relieved by maturity (ET2), more severe water stresses that occurred during the vegetative stage and relieved during mid-grain filling (ET3), and water deficits from early stage onwards, with severe stresses throughout the grain filling period (ET4). The frequency of these environment types varied spatially (Fig. 4) and over time (data not shown), with a tendency for higher frequency of severe stresses during the last decade.

By 2030, the average temperature during the crop cycle is expected to increase compared to the baseline by 0.3C (Wet & Low emission scenario) to 0.8C (Dry & High emission) across the wheatbelt, with some substantial variations across regions (Fig. 3). Cumulative rainfall during the crop cycle is forecast to decrease by 14 mm (Wet & Low emission scenario) to 52 mm (Dry & High emission) on average, which corresponds to a 7 to 25% reduction in within-season rainfall, respectively. The change in climatic conditions led to a shift in sowing and flowering windows (Fig. 3; Wang et al., 1992; Madgwick et al., 2011; Zheng et al., 2012), with e.g. the median sowing date occurring 4 days earlier on average across the wheatbelt for the Dry & High emission scenario. While a high spatial variability was observed in the shift in sowing and flowering windows (Fig. 3), this shift is expected to further increase over time with the effect of climate change (data not shown). Zheng et al. (2012) forecast early sowing to be shifted up to 1 to 2 month(s) earlier by 2050 in

locations like Merredin, for a medium and short-season variety, respectively. Soil water content at sowing was also forecast to decrease, with a maximum impact in WA zones 4 and 5 for the Dry & High emission scenario (>35% reduction in median soil water content; Fig. 3).

Fig. 3 Effect of four climatic scenarios for 2030, compared to the baseline data (1889-2010) on (i-ii) change on average temperature and cumulative rainfall during the cropping season, (iii) median sowing date and (iv) the median level of initial soil water at sowing, and (v) simulated yield. Data averaged for multiple locations in each of 22 regions of the Australian wheatbelt (Fig. 1).

Fig. 4 Effect of four climatic scenarios for 2030, compared to the baseline data (1889-2010) on the frequency of four main drought environment types identified for current climate (presented in Fig. 2). Data from simulations in multiple locations for each of 22 regions of the Australian wheatbelt (Fig. 1).

In future climates, the predicted change in initial and seasonal conditions indicate an increase in occurrence of severe ET4 stresses (Fig. 2 and 4), with no substantial change or reduced occurrence of the other environment types. Across the wheatbelt, ET4 frequency increased by 10, 21, 33 and 77 % on average for the Wet & Low emission, Wet & High emission, Dry & Low emission and Dry & High emission scenarios, respectively. Hence, the ET4 frequency for these scenarios was predicted to reach an average of 25, 28, 30 and 41% of occurrence, respectively, across the Australian production area. Across regions, the rise in ET4 frequency varied from a 25% (South-eastern NSW region) to a 2-fold (Mid North (SA) and WA zones 1-2-4-5) increase for the most pessimistic scenario (Dry & High emission).

The simulations predicted a substantial decrease in yield, especially for the Dry & High emission scenario. Note that the effect of CO2 was not integrated here, nor were heat-shock impacts considered. Improving current crop models to simulate effects of extreme climates on crop production is urgently required as part of a strategy to adapt to future climate (e.g. Howden et al., 2007; Moriondo et al., 2011).


Impact of future climates on drought occurrence greatly varied depending on the scenario considered. Although the optimistic Wet & Low emission scenario results in only a 10% increase in most-severe stresses by 2030, a much greater proportion of these stresses was forecast for all regions under the Dry & High emission scenario (77% increase on average; 41% of occurrence). Recent climatic observations show that we are currently tracking toward a high CO2 emission scenario (A1FI; Le Quere et al., 2009). Both scenarios studied with a high emission forecast (i.e. wet and dry options) had a substantial impact on severe drought occurrence (28 and 77% increase on average, respectively).

While sowing dates have been adapted to the future temperature trends in this study, extending the approach to a broad range of genotypes and managements would allow the assessment of different options to best adapt to future climatic scenarios, and in particular the likely high emission scenarios. In terms of crop improvement, it is becoming urgent to begin the adaptation of varieties to the future, as breeding cycle takes 5 to 20 years (Chapman et al., 2012).


We thank Dr. Steve Crimp for advice on selecting future climate scenarios, QCCCE ( for providing scenarios, and Dr. Bangyou Zheng for the future sowing windows. This research was partly funded by the Australian Government Department of Agriculture, Fisheries and Forestry (DAFF GMS-0335), who also funded the research to generate the consistent climate scenarios.


Battisti DS, Naylor RL. 2009. Science 323: 240-244.

Borlaug NE, Dowswell CR 2005. In: R. Tuberosa, R.L. Phillips and M. Gale (Eds.), Proceedings of the International Congress “In the Wake of the Double Helix: From the Green Revolution to the Gene Revolution”, 27–31 May 2003, Bologna, Italy. Avenue media, Bologna, pp. 3–24.

Chapman SC, Chakraborty S, Dreccer MF, Howden SM. 2012. Crop and Pasture Science 63: 251-268.

Chenu K, Deihimfard R, Chapman SC. Large-scale characterization of drought pattern: A continent-wide modelling approach applied to the Australian wheatbelt. Submitted.

Chenu K, Cooper M, Hammer GL, Mathews KL, Dreccer MF, Chapman SC. 2011. Journal of Experimental Botany 62: 1743-1755.

Coumou D, Rahmstorf S. 2012. Nature Climate Change. Advance online publication.

Howden SM, Soussana JF, Tubiello FN, Chhetri N, Dunlop M, Meinke H. 2007. Proceedings of the National Academy of Sciences 104: 19691–19696.

IPCC. 2007. Core writing team, Pachauri RH, and Reisinger A, eds. Geneva, Switzerland, IPCC.

Jeffrey SJ, Carter JO, Moodie KB, Beswick AR. 2001. Environmental Modelling and Software 16: 309-330.

Keating BA, Carberry PS, Hammer GL, Probert ME, Robertson MJ, Holzworth D, Huth NI, Hargreaves JNG, Meinke H, Hochman Z, McLean G, Verburg K, Snow V, Dimes JP, Silburn M, Wang E, Brown S, Bristow KL, Asseng S, Chapman S, McCown RL, Freebairn DM, Smith CJ. 2003. European Journal of Agronomy 18: 267-288.

Le Quere C, Raupach MR, Canadell JG, Marland G, Bopp L, Ciais P, Conway TJ, Doney SC, Feely RA, Foster P, Friedlingstein P, Gurney K, Houghton RA, House JI, Huntingford C et al. 2009. Trends in the sources and sinks of carbon dioxide. Nature Geoscience 2: 831-836.

Madgwick JW, West JS, White RP, Semenov MA, Townsend JA, Turner JA, Fitt BDL. 2011. European Journal of Plant Pathology 130: 117–131.

Moriondo M, Giannakopoulos C, Bindi M. 2011. Climatic Change 104: 679-701.

Potgieter AB, Hammer GL, Butler D. 2002. Australian Journal of Agricultural Research 53: 77-89.

R Development Core Team. 2011. R foundation for statistical computing, Vienna, Austria. ISBN 3-900051-07-0, url

Wang YP, Handoko J, Rimmington GM. 1992. Climate Research 2: 131–149.

Williams JH, Hook RA, Hamblin A. 2002. Canberra: CSIRO Land and Water.

Zheng B, Chenu K, Dreccer MF, Chapman SC. 2012. Global Change Biology. 18: 2899–2914.

Previous PageTop Of PageNext Page