Predicting weather not an exact science

I think I might have a go at running for parliament... why? Because I'm learning the skill of giving politicians answers. Yesterday while doing one of my routine radio interviews I was asked for my prediction for the weather in Dunedin for the All Blacks game. I stammered and tried to answer but couldn't give a firm prediction. I was, in fact, grilled by these guys - and these were just breakfast announcers on Classic Hits - not exactly asking the tough questions that the Newstalk ZB or Herald journalists ask. They were simply wanting to know the forecast for kick off and I simply couldn't give them a definite answer. I gave them several different possibilities but no firm answer. That's how I operate - "I'll be honest with ya, it's too far out...to give you the forecast for a 2 hour window with this particular weather event is too risky". Some believe if you give a forecast with confidence, any forecast, you'll sound trustworthy. But I don't want to be judged on sounding trustworthy... I want to be trustworthy.

The forecast data I had yesterday morning was conflicting...there was the weather map that showed a huge low in the Tasman crossing the country on Saturday...there was the weather data we had that showed just a 10% chance of showers with a westerly breeze....and I noticed MetService was forecasting showers and a southerly change. Add all of that on to the fact that the low in the Tasman Sea had at least 7 fronts attached to it, yet was on the wrong side of the South Island to really bring showers to Dunedin and I was stumped. Ronnie, a guy from my work, won't let me finish a sentence if I use the words "might, perhaps, chance, risk, could, should, possible, may and slight".

Ron: "Phil, will I get wet riding my bike home tonight?"
Phil: "Looks like it should-"
Ron: "NO!" (hand comes up like a stop sign)
Phil: "I mean, a few showers are possible-"
Ron: "NO!"
Phil: "Just let me finish"
Ron: "No, I just want to know will there be a shower or not?"
Phil: "Well this low is very large and-"
Ron: "NOPE! Showers - yes or no?"
Phil: "There will be showers but the risk for Auckland-"
Ron: "Yes...or NO?"
Phil: big long awkward pause...."Yes"
Ron: "THANK YOU".
Phil: "...but you won't get wet"
Ron: Smacks hand over face "Ohhh you almost had it Phil, you almost had it".

I get what he means. A lot of the public probably don't care too much about the reasons behind the forecast...will it be warm, will it be wet, can I hang out my washing etc. The problem is, weather still isn't an exact science despite all the millions that goes into research and training. Currently forecasters have to rely on predictions of the chaotic weather patterns and sometimes gut instinct (which is painful when it's at odds with the weather data believe me).

And the weather really is chaotic. A contact of mine at MetService in Wellington once told me "you can see the same weather event happen 9 times in a row, every time it does the exact same thing. Then on the 10th time it behaves completely differently". That's the chaos of weather.

Weather enthusiasts in the weather forums here in New Zealand know this all too well. Watching big thunderstorms moving in from the Tasman...they can track towards somewhere like Auckland for several hours then just (and I mean just) before they reach the west coast they completely fall apart. Completely. Not even drizzle. It's frustrating and sometimes when we've put out a weather forecast for Auckland warning of heavy showers and then nothing eventuates all we want to do is display the rain radars online to prove they existed in the first place. "See! They were there. We weren't making it up!". But as Ron would say "So what? You gave me a forecast that was wrong...I don't care what happened out at sea...I don't live out at sea".

And that is why I'm a fan of percentages. At face value it sounds a lot like "might, perhaps, chance" etc, but it's really more scientific than that. The important thing is watching the trend. If today is 50%, tomorrow is 70% and the following day is 10% and yet all days say "Showers" you can clearly see which day is most likely to be wetter. It's sometimes a better way of saying "Slight risk" or "A few". We try to add percentages to our forecasts but on their own it sounds dodgy. For example if I said "50% chance of showers" it sounds like I'm saying "Flip a coin". 50% doesn't represent a 50/50 option...it represents a moderate confidence.

Our current forecast for Dunedin? Yes...well let me firstly say my confidence levels are still only moderate...but at this stage we think cloudy periods increasing and mostly dry with west to south west winds. Temperature around 8 degrees for kick off. But with several fronts around New Zealand...it might pay to check www.weatherwatch.co.nz on Saturday for our latest forecast!