Long-range weather forecasting using the moon was introduced to New Zealand by Clement Wragge in 1910. One of the more colourful characters in early Australian meteorology, Wragge was among the first forecasters to name cyclones. He started with the Greek alphabet, moved on to women’s names, then settled for the names of politicians he disliked.
Wragge resigned as Queensland state meteorologist following intense criticism of his failure to break a drought by firing six hail cannons repeatedly at the clouds. He left Australia after missing out on the job of commonwealth meteorologist but published a nine-year seasonal forecast before he left. In Auckland he set up an observatory and forecasting service and lectured widely.
After Wragge’s death, his son Kismet continued moon-based long-range forecasting from the Waikato. In turn, he was followed by Harry Alcock, who was in the umbrella trade and claimed special knowledge of the weather from observing the rise and fall of demand for his product. Today, Ken Ring is following in these characters’ footsteps.
Ken Ring’s Predict Weather Almanac and Isobaric Maps 2005 includes a continuous sequence of weather maps from 18 years and 10 days earlier, and uses them as a guide to forecast the weather. How successful is this method?
New Zealand is situated in the mid-latitude westerly wind belt. Choose a map at random from the past and there is a good chance it will show a westerly situation. Likewise, choose a day at random in the future and there is a good chance there will be a westerly flow over the country. This being so, it is likely that past and future weathers will match, especially in eastern areas, which are largely sheltered from frontal rain-bands.
Napier and Blenheim, for example, average only 92 and 78 rain days per year respectively. This means you could forecast dry weather every day for a year in Blenheim and expect a success rate of about 80 per cent—higher if you ignored the days when rain fell only during the night. If you allowed your forecast to be a day out, the percentage of correct forecasts would probably be in the 90s.
In western areas, the number of rain days is higher. New Plymouth has 137 and Hokitika 171. Even so, if you forecast rain in Hokitika every third day for a year and allowed a day either way, your chances of forecast success would be very high.
A better way to judge the usefulness of forecasts based on old weather maps is to look at situations that bring severe weather such as prolonged heavy rain and gale-force winds. The most common severe-weather situation in New Zealand is a strong north-west flow ahead of an active front, which causes heavy rain in the west of the South Island and gale-force winds in many areas, especially over coastal waters. On such occasions, the weather map shows five or more isobars crossing New Zealand from the north-west.
A search through the weather maps that show what actually happened in 2005 reveals 22 occurrences of this nature, none of which is matched by a similar event on the weather maps from 18 years and 10 days before. Allowing a day either way catches only three events. The older maps show 17 such situations, none of which re-materialised 18 years and 10 days later.
Heavy rain in eastern areas occurs when a low moves east across New Zealand and strong easterly winds blow from sea to land on the pole-ward side of the low. On the maps for 2005 there are 18 situations in which a low with two circular isobars cross New Zealand. Just one of these matches a similar situation 18 years and 10 days before, while the maps from that time show 13 other occasions when a two-isobar low crossed New Zealand, none of which was reincarnated in 2005.
If there was much predictive skill in using the maps from 18 years and 10 days ago, you might expect the method to pick up the long-lasting highs that sometimes influence New Zealand’s weather for many days in a row. One such affected the South Island from January 25 to February 4, 2005, bringing dry weather, apart from brief drizzle in Canterbury and Southland and isolated showers in Buller on February 3 and 4. During this time Ring’s weather maps showed a sequence of seven fronts crossing the South Island, and he forecast rain in the west on most days and heavy rain on four days.
Considering the year as a whole, Ring forecast heavy rain in Hokitika on 53 occasions during 2005. In fact, heavy rain fell on only five of those days—a success rate of less than 10 per cent. Allowing a day either way captures only another four days, making for a success rate of less than 20 per cent. Allow two days either way, which amounts to a five-day window, and the success rate rises to 34 per cent.
In the eastern locations of Napier, Blenheim and Christchurch, Ring forecast a total of 25 days with heavy rain. Only one of these forecasts was correct—a success rate of 4 per cent. Allowing a day either way only catches two more events—a success rate of 12 per cent. Allowing two days either way catches only one more event, raising the success rate to 16 per cent.
If long-range forecasting were as easy as following the same sequence of weather as occurred 18 years and 10 days ago, the veracity of the method would have been clearly established a long time ago and everyone would be using it to great financial advantage. But the method doesn’t work very well, as the dismal success rate for forecasting severe-weather events shows.
Success this year has been no less elusive. The map Ring used to forecast the weather for January 29, 2006, was from January 19, 1988. On that day a depression formed over the Tasman Sea from the remains of cyclones Anne and Agi. This deepened explosively as it crossed New Zealand on January 19, bringing hurricane-force winds to the east coast of the South Island and heavy rain to Buller and Otago. On January 29, 2006, by comparison, a large high covered the South Island, bringing dry weather, and the winds along the Canterbury coast barely reached 15 kn.
During March 7–10, 1988, a deep depression formed as the remains of cyclone Bola drifted westward past the top of the North Island, bringing severe flooding to Gisborne, northern Hawke’s Bay and Northland, and widespread destructive winds to many parts of the North Island. Ring used the maps for this period to forecast the weather for March 18–21, 2006. Nothing resembling Bola happened on those days. A week later the remains of tropical cyclone Wati moved south over the Tasman, then weakened considerably before crossing Northland. It brought some heavy rain and strong winds to northeastern parts of the North Island, but nothing to compare with the widespread destruction wreaked by Bola.
When interviewed by the Gisborne Herald on March 14 about the non-arrival of a Bola-like disaster, Ring allowed that there was still time and encouraged people to watch for seagulls moving inland, sheep moving downhill and horses sniffing the air, as animals can predict the weather as well.
While some people may look for clues to tomorrow’s weather in the aerobatics of a fantail or a wink from a farmer’s pig, others see more sinister forces at work. For example, when Hurricane Katrina was trashing New Orleans last year, Scott Stevens, a TV weather-forecaster in Idaho, claimed to detect an insidious human influence that went way beyond the effects of climate change driven by excess atmospheric carbon dioxide.
His conspiracy theory involved a secret weapon allegedly developed by the Soviet Union in the 1970s that exploited apparent flaws in classical thermodynamics. According to Stevens, this weapon had fallen into the hands of the Japanese mafia, who had directed Katrina over New Orleans as retribution for the wartime bombing of Hiroshima and Nagasaki.
Soon afterwards, Stevens quit his TV job to pursue opportunities gained from his newfound internet notoriety. “There’s a chess game going on in the sky,” he told an Idaho newspaper.