THE CLIMATE KELPIE BLOG: Getting ahead of weather extremes with better forecasting products

Posted by BCG on 25th February 2019

Agriculture is highly dependent on weather and when extreme events are on their way, Australia’s farming community turns to the Bureau of Meteorology for information.

Getting this information far enough ahead can make a big difference in the ability of farmers to reduce risk. With this in mind, the Bureau and a number of research partners are working with agricultural industries to develop new forecasting tools that will improve the lead time in predicting extreme events.

Reducing risk

Extreme events can be expensive and at worst can lead to loss of life.

A sudden cold snap can kill freshly shorn sheep or new-born lambs. A prolonged dry spell can wipe millions off the value of a grain harvest. Extreme heatwaves stress livestock, drop milk production and destroy horticultural crops.

In many cases being able to predict these conditions can enable farmers to prepare and potentially reduce the impact on their business. For instance, livestock can be moved to sheltered paddocks ahead of cold and wet conditions or to higher ground if extreme rainfall is predicted. In hot conditions, timely irrigation can limit the impact on crops and sprinkler systems can help cool livestock.

Forewarned is Forearmed

The Forewarned is Forearmed (FWFA) project aims to identify the type of extreme events that impact on agriculture and improve the delivery of forecasts to land managers. The project is working with representatives from the grains, dairy, red-meat, rice, sugar, cotton, viticulture and pork industries.

Through a series of consultations with industry representatives, the Bureau of Meteorology and its research partners have identified the types of extreme events that impact on farm business profitability. Their task is to test whether the Bureau’s models can successfully predict these extreme events ahead of time and develop new tools to deliver these forecasts.

The project is focusing on three key areas, extremes of heat, cold and rainfall, starting initially with heat.

Extreme heat

The years 2009 and 2014 commenced with extreme heat events and in 2009 these contributed to the devastating Black Saturday bushfires. Record high maximum temperatures across much of eastern Australia in January 2019 have proved challenging for livestock and have damaged fruit and vegetable crops.

While current temperature forecasts project seven-days into the future, the Bureau are testing a seasonal climate model extending to time frames ranging from two weeks to three months.

One prototype with broad application would allow users to identify extreme heat events up to four weeks ahead. The ‘plume forecast’ displays possible temperature scenarios for the coming month presenting them as a median, 25th to 75th percentile and 10th to 90th percentile forecasts (Figure 1). The chart also shows the historic observed 90th percentile – just 10 per cent of days are hotter than this threshold for that time of year. When the forecast plume exceeds the 90th percentile for observed temperatures, the forecast is highlighted in orange or red depending on the likelihood of an extreme heat event.

Figure 1: Prototype plume forecast of daily maximum temperatures for Camden, New South Wales. The 28-day forecast shows the median forecast (50th percentile, white line), the 25th to 75th percentile forecast (dark grey) and 10th to 90th percentile forecast (light grey). Forecasts that exceed the historic observed 90th percentile maximum temperature threshold temperatures (black line) are highlighted in orange or red depending on the likelihood.

This type of forecast can be useful across a broad range of agricultural industries from those who need to mitigate the impact of heat on livestock, manage irrigation supplies or simply to plan activities around avoiding the hottest part of the day.

Other extreme heat prototypes developed by the Bureau include maps of heat waves and numbers of hot days; as well as charts of temperature projections and the temperature-humidity index for specific stations.

Seasonal rainfall

Another prototype takes seasonal planning a step further. For instance, the current Bureau seasonal forecast for rainfall provides the probability of rainfall being above or below median; that is two categories, drier or wetter than the long-term median. The prototype provides a more detailed forecast showing the probability of rainfall across five decile ranges (Figure 2).

Figure 2. A Bureau of Meteorology prototype is testing the value of a more detailed forecast dividing probabilities into decile 1-2, 3-4, 5-6, 7-8 and 9-10. Shown here is an example forecast issued in November 2018 for the Jan-Feb-Mar rainfall season for Mullewa, Western Australia. The forecast suggests a shift in the odds towards an increased chance for the rainfall season to be in decile 1-2 and a reduced chance of having a decile 9-10 season. This is seen by comparing the forecast probability with the usual (climatological) probability of 20% for each category. The chance of being in the lower two deciles has increased from 20% to 35% and the chance of having a really wet season is halved.

These forecasts can offer a better understanding of the potential for dry conditions. On a seasonal time-scale they could give grain and cotton growers time to adjust their cropping program if there was a particularly dry start forecast for the season. For instance, in 2010 Western Australia had a very dry start to the season despite most of the rest of Australia enjoying a relatively wet year. Better information could reduce financial losses by enabling growers to select more drought tolerant crops, reduce fertiliser inputs and elect to fallow more paddocks.

The prototypes are currently being tested with industry reference groups and will be modified in response to user feedback. New prototypes will also be added. The first products are expected to be released on the Bureau’s website in early 2021.

The Forewarned is Forearmed project is supported by funding from the Australian Government Department of Agriculture and Water Resources as part of its Rural R&D for Profit program. Project research partners: Bureau of Meteorology, South Australian Research and Development Institute, University of Melbourne, University of Southern Queensland, Birchip Cropping Group, Agriculture Victoria, Monash University and Queensland Department of Agriculture and Fisheries. Rural R&D Corporation partners: Meat & Livestock Australia, Dairy Australia, Wine Australia, Sugar Research Australia, Grains Research & Development Corporation, Agrifutures Australia, Cotton Research & Development Corporation and Australian Pork.