Colin Edwards: Rainfall accuracy not an exact science

This year January was 49 per cent above and February 30 per cent above their average monthly rainfall amounts.

So what happened to the predicted drought in Whangarei this summer?

Many people have blamed the forecasters for yet another blunder.

But what are the facts? Who forecast what? To understand what happened it is necessary to know how Niwa predicts weather. They define average predicted rainfall as the long-term average plus or minus 20 per cent. For the first three months of this year Niwa's rainfall prediction was a 20 per cent chance of above-average rainfall, 35 per cent chance of average and 45 per cent below average.

Thus Niwa did not say there would be a drought, merely that below average rainfall was the most likely. If below average means a drought, then they actually forecast less chance of a drought (45 per cent) than the combined average and higher than average rainfall (55 per cent).

In fact this year January was 49 per cent above and February 30 per cent above average, so certainly not a drought.

Rainfall predictions for Northland are arguably more difficult than anywhere else in the country.

This is because summer rainfall is often the result of a cyclone or depression coming down from the tropics. If and when this occurs there is a big dump of rain. Whether this happens, and how often, is almost impossible to forecast several months ahead.

Recently a reader claimed that the time of cabbage tree flowering was a more reliable forecast of drought. Certainly the growth and flowering of trees is weather dependent, but only dependent on past weather. Even the cleverest tree cannot tell the future.

Ken Ring claims to forecast weather using the moon. Many fans believe him but it seems they do not compare what actually occurred with his predictions.

I did an analysis over an extended period and found no evidence to support his claims. His most dramatic failure was a forecast of two weeks dry weather but in the middle of this period we had one of the worst storms ever to hit Northland. An east coast road was closed for several months due to a huge slip and guests were unable to get to a wedding.

This year there has been an El Nino weather pattern which often results in dry weather on the east coast.

Again, a drought may be more likely but that does not mean it will occur.

Long-term weather prediction is like forecasting the result of a general election. A week before, the prediction may be pretty accurate, but several months ahead it is anyone's guess.

So what can Northlanders do to predict summer rainfall? The answer is not very much. You just never know what will happen.