Future climate projections for the grainbelt

Key messages

  • The grainbelt of Western Australia has experienced a drying and warming trend in April to October rainfall since the mid-1970s (Hope 2006).
  • Global Climate Models, which can be used to project future climate based on different greenhouse gas emissions scenarios, suggest that rainfall and temperatures in 2030 and 2050 will continue this warming and drying trend.
  • It is important that Western Australian agribusinesses have adaptations and processes in place to ensure a reliable and profitable future for the grains industry.

Aims

Climate Services for Agriculture (CSA) is a website developed by CSIRO and the Bureau of Meteorology to ‘help farmers and communities plan for the impacts of climate variability’. The aim of this paper is to review the CSA for the grainbelt of WA and present its projections in a series of maps. This will help in planning for a future climate that is projected to be drier and warmer than current.

Introduction

Projections are developed using Global Climate Models (GCMs) that couple various components of Earth’s systems, including atmospheric processes, land processes, ocean temperatures and circulation, sea-ice, aerosol feedbacks and carbon cycle feedbacks. By using prescribed scenarios of greenhouse gas emissions, it is possible to estimate how quickly the Earth system will warm and model varying responses to this warming by the different Earth system components (e.g., melting sea ice) (DWER 2021).

Many GCMs exist that model future rainfall and temperature for WA. In the late 1960s, National Oceanic and Atmospheric Administration (NOAA) developed the first-of-its-kind general circulation climate model that combined both oceanic and atmospheric processes. Over time GCMs have been refined but have only been available to climate scientists. Current GCMs are better calibrated and are publicly accessible, including the Bureau of Meteorology and CSIRO Climate Services for Agriculture (CSA).

It is important to note that projections are not forecasts. They do not provide a prediction of exactly what will happen by a particular date. Instead, they draw on the best available information to indicate how the climate is likely to change based on different scenarios of human activity (DWER 2021).

Method

The CSA website displays interpolated data on a 5km x 5km grid. The climate change patterns are derived from eight GCMs from the Coupled Model Intercomparison Project (CMIP5) pool of models. CSA allows users to select a commodity. Example commodities are (general, apple, barley, canola, lupins, southern sheep, southern beef, and wheat). When wheat is selected, future rainfall for April to October, November to March, temperatures below 0°C for 1 August to 15 October, above 32°C for August to November, and soil moisture at sowing (1 April) are given. An example of the data is presented in Figure 1.

Historical data sourced from the Bureau of Meteorology’s Australian Gridded Climate Data dataset, is split into two different time periods: 1961–1990 (past average) and 1991–2020 (recent average). Future climate is represented under two emissions scenarios. Medium greenhouse gas emissions or Representative Concentration Pathways, developed by the Intergovernmental Panel on Climate Change (RCP 4.5) assume that greenhouse gas emissions are reduced substantially by end of the century, but not enough to stop continued warming. Under RCP 4.5 adaptation will become more difficult over time. High emissions (RCP 8.5) assume rapid increases in greenhouse gases to continue towards the end of the century. Under RCP 8.5 some systems are unlikely to be able to adapt to the large changes in climate.

Projections are presented as the mean and range for the eight GCMs. In Figure 1, projected mean April to October rainfall for the example location of Merredin is 216mm, with a range of 167mm to 245mm. For the purposes of this paper the range of 167mm is considered as the worst-case scenario, with 216mm the mean and 245mm as the best-case scenario.

Maps were produced using projections from the CSA website for 507 stations for April to October rainfall and August to November heat days, and 371 stations for November to March rainfall and soil moisture (as CSA is still under development, newer versions are displayed differently and currently both Nov-March and soil moisture are only available for ‘wheat’ locations). These locations were chosen for ease of mapping, as they are the same stations used in seasonal DPIRD rainfall, temperature and frost maps. If a location was not in the CSA database (for example Williams), the next closest location was selected.

Figure 1

Mapping

Data was mapped using R with the same methods as reported in Hepworth et al (2019). Future frost was not selected for mapping as CSA uses 0°C for frost. DPIRD uses temperatures below 2°C (as this equates to 0°C on the ground as temperature is recorded approximately 2m off the ground by an automatic weather station).

Results

For Merredin (Table 1) CSA’s gridded historical data closely matches actual historical data (something which earlier GCMS failed to do), giving some confidence in the website. The large variation in the range, for 2050 RCP 8.5 for example, lessens the confidence. Therefore, using the mean projections is important. Under RCP 8.5 by 2050, the mean of modelled projections in the CSA website projects that Merredin Nov–March rainfall will revert to the 1961–1990 average, April–Oct rainfall will continue to decline (by 19%), heat in August to November will increase by six days and soil moisture will remain the same.

Table 1

Selected maps

Figure 2 shows four maps that highlight the projected climate change for the Western Australian grainbelt.

Conclusion

Future climate projections from CSA using the mean of the eight models indicate that the climate for the Western Australian grainbelt in both 2030 and 2050 will continue to dry and warm in both medium and high emission scenarios. There is a large variation in the eight models, as represented by the range, however by using the mean model output, the CSA website gives similar projections to other published models.

Other reports also indicate a drier and warmer climate for the grainbelt in the future. The sixth ICPP report indicates a projected reduction in mean rainfall particularly in the cool season, an increase in aridity, and an increase in meteorological and agricultural and ecological droughts, but with only medium confidence. DWER (2021) reports that using projections from 40 GCMs from CMIP5 indicates the recent trend of a drying climate will continue.

The CSA website has a few problems including maps of future projected soil moisture (Figure 2D) which indicate a wetter start to April than the past and current averages maps (not shown). This does not reflect the current trend of later breaks, resulting in drier autumns since 2000 (Guthrie and Bowran 2021). DWER (2021) reports that drier and hotter conditions will lead to decreases in soil moisture and runoff because of increased water loss from evapotranspiration. This could further exacerbate drought conditions. The past and current average time frames (see Figure 1) are not a perfect match for the grainbelt of Western Australia. For more reliance, it would better to use 1975–1999 and 2000–2020 to reflect the climate shift experienced (Guthrie and Bowran 2021). Also, spring frost should be reported as below 2°C to reflect automatic weather stations records.

Frost events in 2016 cost the WA agricultural industry an estimated $600 million in crop losses (GIWA 2017). Current GCMs do not model frost correctly. Not only does frost depend on minimum temperature, but it is also affected by pressure. CSIRO analysis of climate data from 1960–2011 suggests that increasing frost incidence is due to the southerly displacement and intensification of high-pressure systems (sub-tropical ridges) and to heightened dry atmospheric conditions associated with more frequent El Nino conditions during this period (GRDC).

The CSA website is still at prototype stage, and with more consultation with agribusiness and agronomists, it is hoped that it can be better matched to suit the WA grainbelt. The Department of Water and Environmental Regulation (DWER) is currently developing the Climate Science Initiative, that is working on downscaling GCMs to the grainbelt, which should give a better understanding of future climate predictions. Regardless of future projections, the climate of the grainbelt will remain highly variable and if variability and trends remain similar to those experienced since the mid-1970s, farmers will have to produce with less rainfall and higher temperatures.

Figure 2

Acknowledgments

This work was part funded by the MLA funded SheepLinks: Climate adaptation to ensure a sustainable WA Sheep industry.

References

Climate Services for Agriculture https://www.awe.gov.au/agriculture-land/farm-food-drought/drought/future-drought-fund/climate-services

DWER – Department of Water and Environmental Regulation 2021 Western Australian climate projections

Guthrie M and Bowran D 2021 Recent changes in weather and crop growth drivers for the Western Australian grainbelt, and potential impacts on crop growth. Grains Research Update Perth.

GRDC grownotes – tips and tactics managing frost risk

Grains Industry Association of Western Australia 2017 Crop Report 8th February 2017

Hepworth A, Guthrie M and Evans F 2019 Improved weather and climate maps. Grains Research Updates Perth Western Australia.

Hope, P.K. 2006 Projected future changes in synoptic systems influencing southwest Western Australia. Clim Dyn 26, 765–780.

Contact details

Meredith Guthrie
Department of Primary Industries and Regional Development
1 Nash Street, Perth
Ph: 08 9368 3058