How to Interpret
21 June 2007
- NIWA’s outlooks indicate the likelihood of climate conditions being at, above, or below average for the season as a whole. They are not ‘weather forecasts’. It is not possible to forecast precise weather conditions three months ahead of time.
- The outlooks are the result of the expert judgment of NIWA’s climate scientists. They take into account observations of atmospheric and ocean conditions, and output from global and local climate models. The presence of El Niño or La Niña conditions and the sea surface temperatures around New Zealand can be useful indicators of likely overall climate conditions for a season.
- The outlooks state the probability for above average conditions, average conditions, and below average conditions for rainfall, temperature, soil moisture, and streamflows. For example, for winter (June–July–August) 2007, for all the North Island, NIWA assigned the following probabilities for temperature:
- Above average: 60%
- Average: 30%
- Below average: 10%
- This three–way probability means that a random choice would only be correct 33% (or one–third) of the time. It would be like randomly throwing a dart at a board divided into three equal parts, or throwing a dice with three numbers on it. An analogy with coin tossing (a two–way probability) is not correct.
- A 50% ‘hit rate’ is substantially better than guess-work, and comparable with the skill level of the best overseas climate outlooks. See, for example, analysis of global outlooks issued by the International Research Institute for Climate and Society based in the U.S. (http://portal.iri.columbia.edu/portal/server.pt) published in the Bulletin of the American Meteorological Society (Goddard, L., A. G. Barnston, and S. J. Mason, 2003: Evaluation of the IRI’s "net assessment" seasonal climate forecasts 1997–2001. Bull. Amer. Meteor. Soc., 84, 1761–1781).
- Each month NIWA publishes an analysis of how well its outlooks perform. This is available on-line and is sent to about 3,500 recipients of NIWA’s newsletters, including many farmers. See The Climate Update: www.niwascience.co.nz/ncc
- All outlooks are for the three months as a whole. There will inevitably be wet and dry days, hot and cold days, within a season.
- Probabilistic forecasts are most useful to businesses which have to make multiple, climate-influenced decisions. They are not generally a good basis for individuals to plan a one-off event. For example, unless past climate statistics suggest it is an almost dead cert that it will rain or that it will not rain in October, a probabilistic outlook may not be especially useful in determining whether to hold an outdoor wedding that month. However, it would be useful to an insurance company in setting the premium to cover such outdoor events. Farmers also make many risk management decisions that can be influenced by the relative probabilities of various climate conditions, such as what crop mix to use or how much acreage to plant.