Over five years from May 2002 to April 2007, our analyst found NIWA forecasters correct in only 48% of their predictions.
The New Zealand Climate Science Coalition
Hon Secretary, Terry Dunleavy MBE, 14A Bayview Road, Hauraki, North Shore City 0622 Phone (09) 486 3859 – Mobile 0274 836688 – Email –
The Coalition is a not-for-profit, membership-based organization dedicated to climate science and has no political affiliation. We take a science and evidence-based approach to climate change issues.
Media release (immediate) 5 June 2007
NIWA weather predictions wrong half the time
When scientists of the National Institute of Water and Atmospheric Research Institute (Niwa) say that New Zealand’s Indian summer is likely to drag on into winter, there’s an equal chance, based on their past performance, that we are headed for a cold snap. This today from Professor Augie Auer, chairman of the science panel of the New Zealand Climate Science Coalition.
“There are two immediately arresting aspects of this morning’s story in the New Zealand Herald headed “Warmer weather looks set to continue.” The first is the prominence given to a supposedly warm month of May, when there was little comparable mention of the fact that the preceding month, April, was colder than normal. The second is the reference to May setting a record for warmth: the actual figures were: 2007, an average national temperature of 12.4oC, up a barely measurable 0.1oC from the previous warm figure for May 1962 of 12.3oC.
“But, more significantly, in the long term, one of our coalition people, Malcolm Taylor, has just completed an analysis of Niwa seasonal forecasts of seasonal weather against what actually occurred. Over the five-year period, from May 2002 to April 2007, he found that Niwa were correct in only approximately 50% of their predictions. In other words, they were as likely to be wrong as they were to be right. “The results were extremely varied. Over the 60 periods studied, Niwa scored a better than 67% accurate prediction on 25 occasions and were worse than 33% accurate on 26 occasions. The overall accuracy was 48%. The surprising outcome is that they were within the 33% - 67% range only nine times. The analysis was based on information available from Niwa’s own website, complete with map diagrams showing ‘outlook’ and ‘outcome’,” said Mr Auer.
“This has serious consequences for our rural community. Farmers in South Canterbury in June last year were hit with a very severe snow dump with minimum warning. With snow already in the deep south, and Metservice forecasts of extremely cold winds over the next couple of days, South Canterbury farmers, and rural dwellers generally in the deep south can be excused for having little faith in seasonal climate forecasters with a demonstrated 50-50 record of accuracy in prediction. We owe our farmers, and all New Zealanders, better than that. “It makes one question whether taxpayers should be funding Niwa forecasters who have no better odds of success than flipping a coin; and whether that funding would not be better placed in the hands of Metservice, the previous long-range forecasting agency,” Mr Auer concluded.
Contacts: Professor Augie Auer, tel 09 480 5300. Email:
Malcolm Taylor: Email:
Caption for image – NIWA-AirT.Oct-Dec06.psd
Outlook is what NIWA predicted for the period October-December 2006; outcome is what actually happened.