Previous PageTable Of Contents

Climate change risk assessment associated with wheat production systems in Southern NSW

Qunying Luo1 and Helen Fairweather1

1NSW Department of Primary Industries, PO Box 100, Beecroft NSW 2119. Email qunying.luo@dpi.nsw.gov.au

Abstract

Statistically downscaled daily outputs of ten general circulation models under A2 emission scenario for two future periods centred on 2055 and 2090 were coupled with the Agricultural Production System Simulator (APSIM)-Wheat to demonstrate a methodology for assessing climate change risks in a wheat production system. Indices that best describe the climate risk to the production system were analysed with the APSIM modelling output providing the time periods for this analyses. Data inputs in the APSIM model were two wheat cultivars (Sunbrook and Hartog) and daily time step historical/baseline and climate change data from a location in southern NSW. It was found that there was minimal change in temperature related indices across the models. Decrease in GSR was projected for the two future periods. As time progresses, rainfall decreased more. Soil water deficit decreased for the period centred on 2055 while it was increased for the period of 2081-2100 for both cultivars. It is concluded that both opportunities and risks exist in future wheat cropping at the study location. Statistical tests show that there are significant differences in the median growing season rainfall and soil water deficit for both cultivars across the four climate scenarios. This implies that adaptation options should be oriented toward to soil water conservation practices and improving water use efficiency.

Key words

climate indices, APSIM-Wheat, wheat production

Introduction

Frost, heat and soil moisture stress at reproductive stages such as flowering and early grain fill are common climate risks in Australian wheat production systems. Frost injury to cereals in the reproductive stages is a significant economic problem in eastern Australia regularly causing losses in excess of $100M (Fuller et al. 2007). High temperatures around anthesis can substantially reduce grain yield (Wheeler et al. 2000). Half-way through anthesis (when half of the ears in a population have flowered), temperatures above 27oC can result in a high number of sterile grain (Wheeler et al., 1996). A maximum temperature of 35.4oC was identified as a threshold for wheat crops at grain filling period (Porter and Gawith 1999). This value is the mean of five literature sources under controlled experimental conditions. In southeast Australian cereal zones, the effect of high temperature is usually confounded with the effect of limited water supply later in the crop growth cycle. At this time of limited soil water supplies, high temperature intensifies drought affects via increased transpiration (Cawood, 1996). In addition to these reproductive stage climate risks, growing season rainfall is a very important climate index for determining wheat crop yield under rainfed conditions.

There is a growing recognition that extreme low and high temperatures and drought will be more frequent under probable future climates. These conditions will have significant implications for the Australian wheat industry. This paper presents a preliminary analysis of these conditions to describe a method for assessing the changed risk to wheat production.

Downscaled daily climate change projections are available and provide the potential to quantify the probability of exceeding extreme temperature thresholds or the occurrence of other climate risks, with both changes in mean climate and changes in climate variability able to be considered. Our study aims to quantify a number of important climate indices in the wheat production system of New South Wales (NSW) under current and future climate change conditions, based on recently released statistically downscaled daily outputs from a range of general circulation models (GCMs).

Methods

Study site

A location in southern NSW was chosen for which temperature and rainfall data are available in the high quality climate dataset. Wheat is rotated with rice in practice at this location, however only wheat was modelled for this exercise.

Climate data

Two types of climate data were used in this analysis: historical climate data and statistically downscaled climate data. Historical daily climate data for the period of 1980-1999 were obtained from ftp://ftp.bom.gov.au/anon/home/ncc/www/change/. Missing data in the high quality dataset were replaced by the corresponding climate data from SILO CLIMARC database for this analysis. The rationale of using the historical climate data is to compare these results with those derived from statistically downscaled climate data for the same period (1980-1999). Statistically downscaled (analogue approach) climate data were obtained from the Bureau of Meteorology and derived from ten general circulation models (GCMs) for three periods: 1980-1999, 2046-2065, 2081-2100, under the special report on emission scenarios A2 scenario (Timball et al., 2008). Scenario A2 assumes a very heterogeneous world. The underlying theme is self-reliance and preservation of local identities.

Climate indices

The Agricultural Production System Simulator (APSIM)-Wheat model was used to quantify climate change risks using two cultivars - Sunbrook (SB) and Hartog (HT). Sunbrook is a late maturity cultivar and sowing date was set on the 13th of May each year. Hartog is an earlier maturity cultivar with sowing date set as 10th of June. Five climate indices were considered in this study, viz. the number of frost days and hot days at flowering time, the number of hot days at early grain filling time, growing season rainfall and soil water deficient at flowering time. Soil water deficit (0-1) indicates the degree to which crops are stressed. “1” means that crops are fully stressed. Scripts were written in the APSIM package to quantify these climate indices. The number of frost days was calculated as the days on which the minimum temperature is less than 2.2oC. The number of hot days associated with heat stress was calculated as the days on which the maximum temperature is equal or greater than 27oC for flowering time and 35oC for early grain fill.

Statistical analysis

Simple statistics such as average and/or maximum value were calculated by and across climate models. Averages were compared for the number of hot days, growing season rainfall and soil water deficits. A comparison of the maximum value was more appropriate for the number of frost days at anthesis due their infrequent occurrence in the periods considered.

Results

Changes in flowering time

The average flowering time decreased by six days for the period of 2046-2065 and nine days for the period of 2081-2100 compared with baseline for the two cultivars considered (Table 1). There were also small decreases projected for the number of days from flowering to maturity (data not shown). The climate indices analyses should be considered in conjunction with the changes in phenological stages.

Table 1 Flowering time under baseline and climate change (day of the year)

Cultivars

Historical Baseline

Modelled Baseline

2055

2090

Sunbrook

291

291

285

282

Hartog

289

289

283

280

Temperature related indices

Number of frost days at flowering time A maximum number of two frosts at flowering time were found in the historical record for the periods considered. Only one model produced this same result for the baseline period (data not shown). The response for the future periods varied across models, but none projected more than that which occurred historically.

Number of hot days at flowering time Historically only one hot day at flowering time has occurred on average for both cultivars. The maximum average number of hot days projected by all the models was two with most indicating this increase for both cultivars by 2090 (Table 2). A decrease or increase in the number of hot days depends on the balance between the direct effect of temperature increase and the impact of increased rate of crop growth and development under a warmer environment.

Number of hot days at early grain filling period Again there was only an average of one hot day at early grain filling from the historical data and very little change projected, on average, for the future periods (Table 2).

Table 2 Temperature related indices

Indices

Scenarios

CCM

CNRM

CSIRO

GFDL1

GFDL2

GISSR

IPSL

MIROC

MPI

MRI

Average no. of hot days at anthesis

Baseline

2

2

1

1

1

1

2

2

2

2

2055

2

2

1

2

2

1

2

2

1

2

2090

2

2

1

2

2

2

2

2

2

2

Historical

1

 

 

 

 

 

 

 

 

 

Baseline

1

2

1

1

1

1

1

1

2

2

2055

2

2

1

1

2

2

1

2

1

2

2090

2

2

1

2

2

2

2

2

2

1

Historical

1

 

 

 

 

 

 

 

 

 

Average no. of hot days at early gran filing

Baseline

2

1

1

1

1

1

1

2

1

2

2055

1

1

2

2

1

1

1

2

1

1

2090

1

1

1

1

1

1

1

1

1

2

Historical

1

 

 

 

 

 

 

 

 

 

Baseline

2

1

1

1

1

1

1

2

1

1

2055

1

1

2

1

1

1

1

2

1

1

2090

1

1

1

1

1

1

1

1

1

2

Historical

1

 

 

 

 

 

 

 

 

 

Clear background is for Sunbrook while shading is for Hartog. The same applies to Table 3.

Rainfall related indices

Growing season rainfall Seven models projected a decrease in GSR in 2055 for both cultivars. Nine models showed a decrease in growing season rainfall in 2090 across the two cultivars (Table 3). The average annual growing season (sowing-maturity) rainfall decreased (5% for SB, 6.6% for HT) for the period centred on 2055 and further decreased (15% for SB and 16% for HT) for the period centred on 2090 when compared with that of modelled baseline.

Soil water deficit at flowering time More than half of the models projected an increase in soil water deficit in 2055 for both cultivars. Most of the models projected an increase in soil water deficit for 2090 across the two cultivars (Table 3). The average annual soil water deficit decreased (2.4% for SB and 3.1% for HT) for the period of 2046-2065, and increased (11.1% for SB and 7.6% for HT) for the period of 2081-2100.

Table 3 Rainfall related indices

Indices

Scenarios

CCM

CNRM

CSIRO

GFDL1

GFDL2

GISSR

IPSL

MIROC

MPI

MRI

Average

Average growing season rainfall (mm)

Baseline

228

188

193

208

215

189

177

189

213

195

200

2055

221

147

166

191

203

206

186

209

183

177

189

2090

201

117

183

181

173

191

141

185

149

177

170

Historical

235

                 

235

Baseline

191

159

168

175

185

158

151

160

173

157

168

2055

179

115

148

159

177

162

157

178

146

144

157

2090

174

102

153

152

143

163

122

148

109

143

141

Historical

199

 

 

 

 

 

 

 

 

 

199

Average Soil water deficit

Baseline

0.48

0.52

0.57

0.55

0.43

0.50

0.58

0.56

0.63

0.52

0.53

2055

0.50

0.53

0.47

0.56

0.48

0.47

0.55

0.58

0.44

0.62

0.52

2090

0.54

0.60

0.59

0.64

0.60

0.60

0.57

0.53

0.64

0.63

0.59

Historical

0.44

                 

0.44

Baseline

0.47

0.54

0.56

0.52

0.44

0.50

0.58

0.56

0.61

0.52

0.53

2055

0.48

0.57

0.44

0.53

0.48

0.46

0.53

0.55

0.47

0.63

0.51

2090

0.50

0.60

0.56

0.60

0.59

0.57

0.55

0.50

0.65

0.59

0.57

Historical

0.41

                 

0.41

Statistical analysis

Significant differences in the median of growing season rainfall and soil water deficit among the four climate scenarios and the two cultivars were found (data not shown).

Conclusion and Discussion

It was found that there was minimal change in temperature related indices across the models. However this analysis needs to be furthered by considering the changed phenological stages (ie. probability of occurrence of frosts and hot days). Decreases in growing season rainfall were projected for the two future periods by most models. As time progresses, rainfall decreased more. Soil water deficit decreased for the period of 2046-2065 and increased for the period of 2081-2100 for both cultivars. This result reinforces the need to consider adaptation options oriented toward to soil water conservation practices and improving water use efficiency.

Statistical downscaled daily outputs (20 years) of ten GCMs were directly used for the quantification of climate indices considered in this study. The advantage of this procedure is that both changes in climate variability and in the mean climate can be considered and easily applied in climate change risk assessment. The disadvantages of this approach lie in that (a) directly using the downscaled outputs of GCMs may introduce model bias (difference exists in quantified climate indices between modelled baseline and historical baseline), especially for climate variability as weather variables are reproduced poorly by GCM on a daily scale, and (b) 20 years time series may be not long enough for risk assessment.

The APSIM-Wheat model used in this study was not calibrated nor validated against experimental phenological data as part of this exercise and therefore the results shown in this study only provide an indication of the relevant differences between the climate change projections and baseline conditions. Further sensitivity and statistical analysis is required before these results can be used with confidence to infer the impacts of climate change on wheat production.

References

Cawood R (1996) Principals of Sustainable Agriculture: Climate, Temperature and Crop Production in South-eastern Australia. Victorian Institute for Dryland Agriculture.

Fuller MP, Fuller AM, Kaniouras S, Christophers J and Fredericks T (2007) The freezing characteristics of wheat at ear emergence. European Journal of Agronomy 26, 435-441.

Porter JR and Gawith M (1999) Temperatures and the growth and development of wheat: a review. European Journal of Agronomy 10, 23-36.

Timball B, Murphy B, Fernandez E and Zhihong L (2008) Development of the analogue downscaling technique for rainfall, temperature, dew point and pan evaporation, Final Report for South Eastern Australian Climate Initiative, Project 1.3.1.

Wheeler TR, Batts GR, Ellis RH, Hadley P and Morison JIL (1996) Growth and yield of winter wheat (Triticum aestivum) crops in response to CO2 and temperature. J. Agric. Sci. 127, 37–48.

Wheeler TR, Craufurd PQ, Ellis RH, Porter JR and Vara Prasad PV (2000) Temperature variability and the yield of annual crops. Agriculture, Ecosystems & Environment 82, 159-167.

Previous PageTop Of Page