Not all forecasts are equal
AS the rain pelted down across much of the South over the holiday break, few could honestly call it a stellar start to summer. While timely weather warnings were certainly provided for much of the heavy rainfall by MetService, WeatherWatch and Niwa, they came against a background of expectation of a “summer scorcher”, based on longerterm seasonal predictions by Niwa.
When it comes down to it, this prospect of a hot, dry summer was probably more about how southerners subconsciously deciphered the message rather than any such overt phrasing from Niwa.
This disparity between fact and fiction holds true for the reality of most weather predictions, whether shortterm, for tomorrow, or longerterm for the next few months. There is a gulf between the way many members of the public interpret the forecast and what the forecast actually says.
The classic case — and one which must be incredibly frustrating for forecasters — is the prediction of “a few afternoon showers”. Despite the hitandmiss nature of showers, this often turns into something like, “it was supposed to rain all afternoon, but it never happened”.
Radio and television newsreaders and announcers, and others in the media, are also often guilty of generalising a forecast of a week with showers as simply, “it’s going to rain all week”.
As easy as it is to malign the forecasters, those whose lives and livelihoods depend on a detailed understanding of what the atmosphere is up to are always careful not to simply check out their local forecast without considering context. That context comes from looking at the “fine print” — the myriad maps, satellite and radar images, videos, warnings and news stories posted by the weather agencies several times every day.
The science of weather forecasting has advanced hugely in the past few decades, spurred on by evermore powerful computers able to crunch billions of figures in the blink of an eye and produce a range of models mapping where the anticyclones and depressions and fronts are up to two or more weeks ahead.
These models are generally very accurate in the first week and still give good indications of likely weather in the second week. Only 30 years ago, it was difficult to get a forecast more than four days ahead; more than 40 years ago, the day after tomorrow was about as good as it got.
It is hard to imagine a world now without weather forecasts. But we need to learn how to discriminate between the strong guidance of likely conditions in the next few hours or days and seasonal predictions.
A good example of this is the La Nina phenomenon, which is steering the weather conditions this summer. In a La Nina — the opposite of an El Nino — the eastern Pacific Ocean is colder than usual. This generally leads to an increase in mild or even subtropical northeasterly airstreams across New Zealand.
Unfortunately, as any Dunedinite will tell you, northeasterlies can be pretty cold, damp and cloudy as they blow onshore, which puts paid to any “summer scorcher”, even if the wind is originally from the subtropics.
Another difficulty with seasonal forecasting is that, because the data becomes more nebulous the further out you look, it is even more difficult to pinpoint where the most severe weather may occur to make the predictions any more useful than just a general guide.
New Zealand forecasters have always struggled to predict the weather in the South. Its proximity to the cold Southern Ocean and the fastmoving cold fronts of the Roaring Forties, to tempestuous Foveaux Strait, the temperature and rainfall extremes of Central Otago and its rough topography make predicting wind directions and weather a difficult task.
Southerners can now have more faith in their forecasts, particularly with the MetService’s new Otago radar plugged in and constantly watching for rain, snow and thunderstorms. But it is worth remembering to treat the vaguer seasonal prognostications with more than a few tonnes of salt.