|
|
Murdoch graduate Dr David Stephens' decade of research has taken crop yield forecasting into a new dimension  with potential major economic benefits for Australia's wheat industry.
 According to Professor Tom Lyons, Executive Dean of Murdoch's Division of Science, Dr Stephens' models are a considerable improvement on current forecasts.
"Getting a better handle on the amount of wheat produced in a year is absolutely crucial to economic planning," he said.
"It's a bit like the futures market if you know the way a particular commodity is going to go six months in advance you can do extremely well."
Professor Lyons supervised Dr Stephens' honours and doctoral research which began in 1986 and covered shire, state and national wheat yield variability.
Now, Dr Stephens works within Agriculture Western Australia where he is a research officer in the Spatial Resource Information Group.
This has allowed him to further develop and apply his findings in practical problems.
Dr Stephens developed three wheat forecasting models.
"In the first, I took district rainfall from around the country and weighted it according to the importance of the particular month in the crop growing cycle," said Dr Stephens.
This has been applied in ABARE (Australian Bureau of Agricultural Resource Economics) to forecast state and national grain production.
"The second, more complicated model formed part of my original honours thesis at Murdoch in 1986/87. In it, daily rainfall figures collected from individual rainfall stations in each shire are used in the model to calculate a stress index which is correlated with shire yields.
"The third model also used by the Canadian Wheat Board Model to forecast prairie wheat production included daily rainfall and temperature data."
Dr Stephens said satellite imagery in Western Australia also picked up crop yield variations very well, as such a high proportion of the state's surface was under wheat.
"Underlying crop yield variation is the influence of rainfall on yields," he said.
"The total rainfall received in a growing season is important, but so is the distribution in time. I found autumn and spring rainfall contributed to yields in a positive way right across the wheat belt, but winter rainfall had an increasingly negative effect from Queensland to Western Australia. This is due mainly to poor soils in Western Australia where the amount of soil moisture we can store is very limited compared to the eastern states."
|
|
As a result, in wetter years in Western Australia, the highest wheat yields occur in the central and eastern wheat fields while the south-west corner has lower yields.
"It is the reverse situation in a dry year when low yields are experienced in the north-east and high yields in the south-west," said Dr Stephens.
Dr Stephens also looked at the broad scale weather systems in the Australian region and how they caused crop variation each year.
"Most major dips in grain production in Australia are also related to broad-scale atmospheric pressure changes in the Indian and Pacific Oceans. In autumn the weather systems can swing into a wet or drought mode which can persist."
Dr Stephens said the Southern Oscillation Index the normalised difference in pressure between Darwin and Tahiti drove most of our rainfall and crop production in Australia.
"In our operational yield forecasting we make assumptions about future rainfall based on this index," he said.
"If the yield model correctly accounts for existing soil moisture and the seasonal rainfall predictions are correct, we can accurately forecast yields months in advance of harvest."
Fortunately, Western Australian production varied less than in the eastern states, said Dr Stephens.
This was good for the Australian Wheat Board which markets the wheat and needs to guarantee a certain production.
Also as part of his thesis, Dr Stephens carried out a survey of 650 wheat farmers across the country, asking them when they had sown their wheat and how much rain they needed to sow.
"Based on this I looked at trends in the sowing dates," he said.
"I found that farmers particularly in Queensland and WA had recently brought their sowing forward by two or three weeks."
This was due to quicker and bigger machinery and better herbicides, which could remove weeds and reduce the number of passes needed over the ground, said Dr Stephens.
Newer, semi-dwarf varieties of wheat added to the efficiency and, if planted early, did particularly well, especially with the addition of nitrogen fertiliser.
"This has contributed to quite a marked jump in Western Australian wheat yields probably as much as half a tonne per hectare which is almost a 50 percent increase."
At Agriculture Western Australia, Dr Stephens is looking at the technological increase in yields spatially at a shire level across the state.
"By removing the climatic effect, I have shown that the highest yield increases have been in the north-west of the Western Australian wheatbelt, where there has been a big swing towards the planting of lupins in a lupin/wheat rotation," he said.
Professor Lyons said Dr Stephens' models were a considerable improvement on ABARE (Australian Bureau of Agricultural Resource Economics) forecasts, particularly in Queensland.
"I think even ABARE themselves would recognise that," said Professor Lyons.
"The problem ABARE faces is getting a good forecast of the national yield as early as possible. With a good forecast the Wheat Board can plan the price at which they want to sell the wheat."
Professor Lyons said other factors also included storage problems and organising wheat transport.
"When I spoke to the former head of the Canadian Wheat Board, he said the Canadian government was able to recover nearly all of the money they had invested in improving long range forecasting in just a few years."
See also
Technology Transfer Centre to boost State's agriculture
Keeping our noodles number 1
|