What’s new in 2021 from climate science– a quick update for busy advisers

Author: | Date: 08 Feb 2022

Take home messages

  • For most SA cropping regions, the 2021 season had a late start and winter. Forecasts in July for an increased chance of above average rainfall in spring were followed by a very dry August to mid-October and wet November
  • New products released by BoM in 2021 as part of the Forewarned is Forearmed (FWFA) project include forecasts of extreme wet or dry and extreme hot or cold for the coming fortnight, month, and three months.
  • In August 2021, the Intergovernmental Panel on Climate Change (IPCC) issued a report and summary for policy makers expressing increased confidence in warming and the urgency of reducing emissions. This is worth a read.
  • Climate Services for Agriculture is a website developed by BoM and CSIRO. This site is worth bookmarking for future reference.

New forecast products that go beyond exceeding the median

Seasonal climate forecasts from the Bureau of Meteorology (BoM) are expressed as the per cent chance of exceeding median rainfall or temperature. Grain growers have pointed out that there is a big difference between a season that is a few mm wetter or drier than the median and extremely dry or wet seasons. The middle deciles are easy to manage compared to the extremes. Forecasts of the chances of the driest or wettest, coldest or warmest two deciles have been developed as part of the Forewarned is Forearmed project (FWFA), funded by the Australian Government Department of Agriculture, Water and the Environment’s Rural R&D for Profit program, with co-investment from 14 project partners including GRDC and SARDI. A grains industry reference group with GRDC panel members from each of the three regions and GRDC staff have reviewed products and provided feedback.

Three new forecast products were made available on the BoM website (bom.gov.au) late in 2021:

  • Overview
  • Summary
  • Climate Outlooks.

Examples will be given in the presentation. The first product is a series of maps showing the chance of having extreme high (deciles 9 and 10) or low (deciles 1 and 2) rainfall, maximum temperatures or minimum temperature for the weeks, months and seasons ahead. The second product is location-specific bar graphs that indicate the shift in the probabilities compared to usual across the deciles. They are available for rainfall, maximum and minimum temperatures for the weeks, months and seasons ahead. These location specific bar graphs come from the clickable map. The third product is the “climagram”. These are location-specific timeseries graphs showing the forecast of rainfall totals, maximum and minimum temperatures for the coming weeks and months. Two further forecast products of extremes will be released in 2022.

The extra information on climate extremes meets a request from long term users of the forecasts. It is important to note that this extra detail is not adding accuracy to seasonal climate forecasts. There is ongoing work to characterise and improve the skill of the forecast, however they are likely to remain in the category of ‘too good to ignore but not good enough to be sure’.

What happened in 2021 and what can we learn about using seasonal climate forecasts

Most SA cropping regions had a late start to the 2021 cropping season followed by a wet June and July. As can be seen in the regular updates on climate drivers from BoM (Table 1), the development and declaration of a negative Indian Ocean Dipole event coincided with the wet winter. The Indian Ocean Dipole (IOD) is measured as the difference in sea surface temperatures between the eastern and western tropical Indian Ocean. A negative phase (8 weeks below -0.4oC) is typically associated with above average winter-spring rainfall in Australia. The very wet season of 2016 was a strongly negative IOD and the very dry 2019 was the strongest positive IOD on record. In early spring of 2021, although crops were less advanced than ideal, the winter rain in the profile and increased chance of a wet spring led to widespread confidence for nitrogen topdressing in early spring.

Table 1: Climate driver updates over 2021 from BoM.

The wet June and July were followed by extremely dry August and September, a patchy October and wet November. For most of the SA grains belt, the 2021 growing season rainfall (April to October) was dry (decile 1—3) or average (decile 4—7). If spring is defined as September, October and November, the season was average (decile 4—7) or wet (8—10). The November rainfall, consistent with the developing La Niña, presented harvest challenges and, in some cases, opportunities to build soil water for the 2022 crop.

Figure 1 shows a time series of the of IOD and Nino 3.4. The IOD index fell to below the threshold of -0.4oC in April before rising in July and falling again in August. The Nino 3.4 index fell during spring (Aust BoM threshold is -0.8oC, US NOAA use a higher threshold of -0.4oC)

Figure 1. Time series of IOD and Nino 3.4. BoM website

Figure 1. Time series of IOD and Nino 3.4. BoM website 

Growers and advisers correctly interpreted the expectation for increased chances of above average spring rainfall from most forecasts and commentators. So, what happened?

As shown in Figure 1, in 2021, the relatively weak negative IOD fluctuated around the -0.4oC threshold. This contrasts with a year like 2016 where the IOD was clearly negative with monthly values below -1.4oC. IOD and ENSO are major drivers of seasonal rainfall in southern Australia, but there are shorter term drivers such as the Southern Annular Mode and the Madden Julien Oscillation. The excellent Climate Dogs animation from Agriculture Victoria introduced many in agriculture to ENSO, Indy, SAM and Mojo as four dogs along with the subtropical ridge (Ridgy) that can pull in different directions to ‘round up’ weather for southern Australia Climate Kelpie. Factors that may have contributed to the dry spring include the Madden Julian Oscillation (MJO) early in spring staying in a phase associated with drier conditions in southern Australia and a positive Southern Annular Mode. Advisers and growers who followed the forecasts in the 2020 season will recognise the pattern of an outlook for increased chance of a wet spring from a negative IOD that failed to eventuate until a La Niña later in spring. In 2020, the rain came in October rather than November which was timely for most of the SA grain crop. For example, for some early crops in low rainfall areas, October rain in 2020 was more disruptive to harvest than November rain in 2021.

The failure of the IOD to deliver a wet spring in 2021 poses the question ’should growers and advisers follow the trends and forecasts for climate drivers such as SAM and MJO along with ENSO and IOD?’  Individual growers and advisers are the best to judge how much time and attention they want to give to following climate drivers. The most straightforward approach is to use the seasonal climate forecast and let the experts worry about the alphabet soup of climate drivers. It is important to acknowledge that understanding ENSO and IOD has been an important step in building trust and confidence in the basis of seasonal forecasts and useful when interpreting the forecasts. Following climate drivers such as SAM and MJO will come at a greater cost of time because they tend to fluctuate within a given month rather than locking into a positive, neutral or negative phase during the growing season. It is very hard to collate the information and almost impossible to give the appropriate influence that each driver will affect rainfall, be that a mental calculation for busy growers or even by using a spreadsheet. There is a good reason that climate science increasingly relies on dynamic climate models run on super computers, the oceans and atmosphere.

In 2013, the BoM switched from statistical to dynamic seasonal climate forecasts. Statistical forecast systems, used since the 1990s, are based on correlations between patterns of sea surface temperatures and local rainfall in the following 3—6 months. Dynamic forecast systems use the power of super computers to model future changes in the ocean and atmosphere for each day over the 6—9 months ahead. Because the model captures some of the stochastic nature of weather, no two runs of the model will be the same. The range of possible futures is widened by incorporating uncertainty in the starting conditions and uncertainty in the model. The result of the many model runs is 100 possible futures, some drier than the long-term median and some wetter. A forecast of 70% chance of exceeding the median rainfall is a statement that while 70% of the model runs were wetter, 30% were drier.

Gigerenzer (2005) asked participants in US and European cities what they understood from the statement of 30% chance of rain tomorrow. Did it mean, with current weather conditions, rain was expected i) 30% of the time (that is, about 7 hours), ii) 30% of the area or iii) 30% of days? A recent Australian example of a misinterpretation is here.

Many studies have shown that probabilistic forecasts are difficult to understand. The statement ‘it will be a wet spring’ is simpler than ‘70% chance of above median rainfall’. The problem with an interpretation based on headlines of ‘BoM predicts wet spring’ is that this simple categorical forecast will be wrong 30% of the time. A useful rule suggested by Gigerenzer (2005) is that when someone mentions a percentage, take time to ask yourself ’a percentage of what?’ The problem with probabilistic forecasts isn’t the maths. Grain growers understand and use percentages, such as 80% germination rate, 5% of a wheat sample contaminated with weeds or the soil moisture is 70% full. A probabilistic forecast is simply expressing the percentage of futures that will be wetter than the median. Thinking of a seasonal forecast as 30% of the model runs ended up with a drier season at my location and 70% were wetter is one way to better understand the forecast. When this shifts to 80% chance of exceeding the median, there is still 20% of computer runs that are drier. This clarity in thinking becomes even more important when dealing with extremes. A statement that there is 40% chance of decile 9 and 10 is a doubling of the long-term climate odds of being in the top two deciles.

Working with Upper North grower and consultant Barry Mudge and other colleagues, SARDI Climate Applications is working on methods to use these revised probabilities in decision making. For a quick explanation, see Rapid Climate Decision Analysis Tool. For a longer discussion, contact Barry Mudge or Peter Hayman.

An important IPCC report released in 2021

A number of important climate change science and policy responses occurred in 2021. The most media attention was on the November climate conference in Glasgow. This conference was preceded in August by an important report released from Working Group 1 of the Intergovernmental Panel on Climate Change (IPCC). This 30-page summary document is worth a read.

The IPCC aims to provide a rigorous and transparent update on science relevant to climate change. Rigour and transparency are achieved through an extensive review of drafts by scientists and policy specialists from governments. This is the 6th IPCC report and many commentators have observed the stronger language expressing more confidence in warming and the urgency to reduce emissions. The presentation will provide a brief introduction to some of the key graphs.

For a local South Australian focus, the AgExtra annual conference in July 2021 included a plenary session on climate change with Professor Mark Howden from ANU, Fiona Simpson from National Farmers Federation and Professor Richard Eckard from the University of Melbourne. AgEx Plenary - Session 1. Other local presentations worth following up include Dr Amanda Schapel (Rural Solutions, now SARDI) on soil carbon status of SA soils.

In March 2021, climate scientist Dr Kimberly Nicholas published a book ‘Under the sky we make’. She provides a clear summary in five short sentences: ‘It's warming. It's us. We're sure. It's bad. But we can fix it’. In the last two decades, well-meaning growers and advisers have drawn my attention to strong contrarian arguments on the first four points. These include a challenge to whether the warming is really happening, the relative role of humans compared to volcanos and solar cycles, the fact that some scientists have different views and that the benefits of higher carbon dioxide will outweigh the problems. It is noteworthy that some of the same voices that dismissed the problem of human-induced climate change are now questioning the effectiveness of attempts to reduce emissions. Those of us involved in agriculture need to carefully and critically consider the impacts and adaptation options and the role for agriculture in reducing emissions including sequestering carbon. A local perspective on the challenges and opportunities presented by climate change is the PIRSA website Climate change resilience and adaptation - PIRSA.

A new website worth bookmarking – climate services for agriculture

As part of the National Drought Initiative, CSIRO and BoM have been funded to produce the Climate Services for Agriculture tool. This tool provides historical data (1961-2020), seasonal forecasts (1—3 months) as well as future climate projections (2030, 2050, 2070) for a given location. The tool is in prototype form and feedback is welcome. See Climate Services for Agriculture (indraweb.io) User name: csa Password: demo

Acknowledgements

The research undertaken as part of this project is made possible by the significant contributions of growers through the support of the GRDC’, the author would like to thank them for their continued support. Ideas presented in this paper have benefited from discussion with Barry Mudge, Geoff Kuehne, Dale Grey, Liam Ryan, Dane Thomas and Bronya Cooper.

References

Gigerenzer G, Hertwig R, Van Den Broek E, Fasolo B, Katsikopoulos KV (2005) “A 30% chance of rain tomorrow”: how does the public understand probabilistic weather forecasts? Risk Analysis 25(3), 623-629.

Contact details

Peter Hayman
Climate Applications SARDI
0401 996448
Peter.Hayman@sa.gov.au