CLIMATE forecasters are making steady progress in terms of accuracy, however there remains a critical gap in the late summer and early autumn, when growers are making their planting decisions.
This is the summary of the accuracy of medium to long term forecasts from Agriculture Victoria climate agronomist Dale Grey.
“You look at the results from the models as the have evolved, from the Bureau of Meterology’s POAMA 1 forecasting model, used from 2002 to 2006 and compared them to the ACCESS-S system now used and it is very different.
“The models have moved from statistical to dynamic and can now account for what is actually happening in the oceans and atmosphere.”
He said improvements were being made not just in terms of overall climate forecasting but with localization.
“The ACCESS-S system will give you readings on a 60km grid, compared to 250km grids or larger with the older models.”
Whereas the old systems were statistically based, the newer dynamic-based models, predict weather based on real-time information from the world’s oceans and atmosphere.
“It is a more accurate system, but it requires an enormous amount of computer power to operate,” Mr Grey said.
“The more computer power becomes available, theoretically the better the forecasts will get as they can handle more processes.”
He said while farmers often focused on the times that the forecast was wrong, overall skill was improving right across the board.
“You only have to look at the short term forecast and where we are with that.
“The accuracy we get five days out now is what a decade ago you could get for a two-day outlook.”
Mr Grey said in terms of international climate forecasting, the impact of the Pacific Ocean was well researched, but other drivers of climate in Australia were less well understood.
“There has been a lot of global investment into what happens in the Pacific and that has led to what we know about El Niño and the Southern Oscillation Index (SOI) but other factors for Australia such as the Indian Ocean Dipole and the Southern Annular Mode are less well understood.”
“Not many people are that interested in what happens in the Indian Ocean, it is mainly of interest to Australia, so we haven’t seen the level of investment and research there that has occurred in the Pacific.”
He said growers really looking to get a handle on likely weather outcomes for the season ahead needed to look at a range of models.
“Different models have different attributes, for instance, POAMA-2 has very good skill in central Queensland, but only in spring and summer, while others have strengths in other areas.”
Mr Grey said an interesting trend was emerging with consensus forecasts utilizing all available models in terms of weather events becoming more concentrated than forecasted.
“Overall, the consensus of models have got it right, but have said a year will be only slightly wetter or drier when at the end it has been significantly wetter or drier in years like 2015 and last year.”
“There is an interesting correlation that when half the models move away from average a significant result, whether wetter or drier has occurred.”
“We can probably put that down to something happening that some, but not all models are picking up on.”