THE CLIMATE KELPIE BLOG: Finding the window of opportunity with new forecasting products

Posted by BCG on 7th December 2020

When you’re dealing with an invasive pest, finding a rain-free window of opportunity for control can be difficult enough, but what if you needed a two-week window with enough lead time to organise a spray program for over 13,000 different properties in the western suburbs of Perth? 

This was the challenge facing the Department of Primary Industries and Regional Development’s (DPIRD) Queensland fruit fly (Qfly) eradication program. 

Not only did they need two weeks to undertake the initial fortnight of twice weekly bait sprays, there was also a requirement for at least two weeks’ advance notice to coordinate hundreds of staff, prepare the insecticide baits and liaise with property owners.

To help the team identify a start date, DPIRD climate expert Dr Ian Foster turned to new Bureau of Meteorology experimental forecasts that were able to provide more informative weekly forecasts with double the lead time of existing products. 

The Queensland fruit fly eradication program is targeting over 13,000 properties in the western suburbs of Perth.

No small undertaking

Qfly is an exotic pest in Western Australia that threatens the ability to retain access to important horticultural markets. If Qfly were to become established in WA, it would cost the state $38 million per year, impacting on production and overseas market access and affecting commodity exports such as avocados to Japan. It would also be more difficult for home gardeners to grow fruit and fruiting vegetables in their yards.

If eradication is not successful, WA could lose access to national and international markets, such as the export of avocados to Japan.

The Qfly eradication program is a huge logistical undertaking that needed to target thousands of trees, ideally starting in September as daily temperatures increased but before the Qfly started to breed.

To be most effective the bait needs to be applied during a period of relatively dry conditions, which are not typical of September. Spraying in October would increase the probability of dry conditions, but would also increase the likelihood of breeding, inevitably extending the length of the outbreak. This meant that the operations team needed weekly forecasts a month in advance so they could identify a September spray window far enough ahead to allow time to prepare the baits.

“The rainfall forecasts currently available on the Bureau website provide weekly forecasts for the coming two weeks,” said Dr Foster.

“And the longer-term forecasts are aggregated into multiple weeks, which do not supply enough detail to decide which two-week period might provide the best opportunity to spray.”

Forecasting the distance

Instead Dr Foster was able to access a new experimental forecast being trialed as part of the Forewarned is Forearmed (FWFA) project. The experimental climagrams show rainfall probabilities for specific weeks, tailored for a specific location (Figure 1).

“These forecasts provide us with much more information. For example, we can look at individual forecasts for each week, up to four weeks in advance. This allowed us to select a period when rainfall was likely to be lower than average, with enough lead time to prepare the response.

“One of the real advantages of these forecasts is that they are location-specific rather than regional maps,” says Dr Foster. “This means that we can compare forecast rainfall with the historic median and range for that period at that location.”

Current forecasts are presented as regional maps, which means that they cannot provide this level of detail.

Figure 1 Climagram for Perth showing the weekly rainfall forecasts (box and whisker plots in green) were lower than the historical climatology (in grey) for much of the forecast period.

Climagram forecasts are presented as box and whisker plots, which show the most likely rainfall range (the box) and the potential extremes (the whisker). Each week the team reviewed the updated climagram to get a better picture of the coming month.

For example, the climagram for Perth in late August showed weekly rainfall forecasts as box and whisker plots (in green). The box shows the 25th and 75th percentiles with the average as a black bar, while the whisker show the 5th and 95th percentiles. They are overlain with historical climatology (in grey).

This example indicates that predicted rainfall was more likely to be drier than normal during September with the exception of week three.

“If we see that a period of a couple of weeks is likely to have below-average rainfall then we have the confidence to go ahead with the spray program. There might always be an unexpected rain event, but the risk is lower. Apart from a couple of rain days, the team got their wish.”

“We still used the weekly forecasts available on the Bureau website to direct daily planning, but the climagrams provided the longer-term view needed for strategic planning.”

In October, the Qfly eradication program reached a significant milestone with response personnel conducting their 100,000th Qfly inspection.

“Without these forecasts we could not have hoped to be this well advanced in the baiting program. They enabled us to identify an optimal window and start the baiting program with more preparation and confidence than we could have if we relied on the existing less-detailed forecasts.”

Product for agriculture

These new forecasts are being developed to improve the delivery of short, medium and long-term forecasts for agricultural businesses.

They are being developed by the Bureau and its research partners as part of FWFA project in consultation with industry representatives from the grains, dairy, red meat, rice, sugar, cotton, viticulture and pork industries.

The forecasts are intended to provide better forecast tools to help with planning weather-dependent farm activities, such as spray programs. They will also help land managers to identify and prepare for extreme climate events such as periods of heavy rainfall or hot conditions.

“It’s all about the risk assessment,” says Dr Foster.

“There is always the possibility of an unexpected event, but the risk is low enough to make it worth committing our resources.”


Dr Ian Foster, 0428 944 478,

Forewarned is Forearmed project information

Visit DPIRD’s Queensland fruit fly updates for more information on the eradication program, a map of the Quarantine Area and the Quarantine Area Notice.

The Forewarned is Forearmed project is supported by funding from the Australian Government Department of Agriculture 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 and Livestock Australia, Dairy Australia, Wine Australia, Sugar Research Australia, Grains Research and Development Corporation, Agrifutures Australia, Cotton Research and Development Corporation and Australian Pork.

Leave a Reply