Your web browser (Internet Explorer) is out of date. Some things will not look right and things might not work properly. Please download an up-to-date and free browser from here.

We got it wrong!

Top Debate — Not too often you hear forecasters going public with their mistakes – but on day one of setting up WeatherWatch.co.nz we introduced a policy of explaining why we – and sometimes other forecasters – get things wrong from time to time.

Clearly no forecaster is 100% perfect and often when we’re conflicted about a prediction we use words like “possibly” and “maybe” or we give a risk factor – like “low” or “high” or “60%”.

Forecasting accuracy is going up overall, and we like to think WeatherWatch.co.nz has seriously raised the bar in Auckland.

But when we make a black and white prediction and then get it wrong – we believe it’s our duty to explain why.  It A) helps you understand the tricky and chaotic side to forecasting, and B) it holds us accountable. 

So what did we get wrong today?  Well – our predicted high for Auckland.

Often highs can be 1 or 2 degrees out…sometimes 3.  But anything beyond that we consider a bad forecast.

Today we predicted a high of 26 in Auckland…it only reached 22.  Quite a big difference on a wet, gloomy, day.

So what went wrong?

It’s actually very simple – the front was slower than we thought.  In fact, you’ll find many forecasts that end up being wrong simply come down to timing. 

For example, if we predict rain tonight, and it falls at 3am instead of 10pm, most people don’t care.

If we predict rain tomorrow afternoon and it doesn’t arrive until late tomorrow night – we take a lot of heat.

In that example the rain still arrived – but timing is everything – especially when it’s during daylight hours.

Today we thought the clouds would break, allowing the sun to rapidly heat the ground (it’s also why we had a low to moderate risk for thunder today).  But the thick cloud and slow movement of the front stopped thunder development in Auckland and meant that the heating never occured.

By 4pm the rain band had started to clear Auckland and you could feel the sun through the clouds for the first time – but it was too little, too late.  The cut off time for highs at MetService is 4pm. It’s more than likely some places started to see the mercury rise after the high of 22 had been reached – but not officially.

So there you have it – an explanation on why our high for Auckland today was off the mark. 


What do you think?  Are there unfair expectations on weather forecasts to be 100% accurate?  Or do you think all forecasters should be held to a higher standard of accountability?  And is WeatherWatch.co.nz honest in your view?  Love to hear your thoughts – positive, negative or constructive – but please, keep it polite.

Cheers

The WeatherWatch.co.nz team

Comments

Guest on 25/02/2012 12:54am

You mention getting the timing right! – Always thought previously you guys had this more accurate? but, along with some other forecasters, when forecasting rain coming from an East to North East quarter, the rain seems to come *much* later, and not last as long as forecast.[more so when the forecast is longer range -. few days before].
Most recent example this week Wednesday [Auckland] was initially [long range] forecasted to have rain, sometimes heavy all day whereas the rain didn’t turn up till late Thursday. Time to adjust those computer models huh ,,, 🙂

Guest on 23/02/2012 10:31pm

I remember looking at metservice earlier this week, and they said it would be 26c on the 23rd of Feb in Auckland
Weatherunderground, which has some good weatherstations, but isn’t as accurate, had a forecast of 28c for the 23rd… obviously this happened across the board, not just for yourselves 🙂

Andy in Timaru on 23/02/2012 4:48pm

Howdy. To me it doesn’t really matter if you’re spot on with the forecast – it’s harder to pick than a broken nose anyway sometimes! Keep up the good work. You only need to have a look at your ratings over the past 12-18mths to see that people DO take you seriously. Well done WW. Love your site – even if it doesn’t have a live streaming rain radar and all the trimmings. Can’t wait to see what it looks like in 5-10yrs time. Surely you may be able to get a few bob out of TRN hehe. Cheers, Andy

Peter of Dunedin on 23/02/2012 10:51am

Medium and longterm forecasting by you and NZ Met. Service was astray for the south of the South Island on this low. It tracked further south than forecast and we received much needed rain. In the period since the rain commenced in the early hours of Thursday morning to the present time when rain has virtually ceased, I have recorded 98mm of rain. For once I am not complaining as the ground which was bone dry below a semi-damp surface, is now relatively moist. I would say most farmers will be happy as it will give a boost to grass growth for winter feed.

Sue on 23/02/2012 9:00am

Some philosophical thoughts in response to your questions. First I would like to say that the definition of the word forecast is to "estimate" or "calculate" – not to define precisely, be exact or otherwise. I am a social scientist and endeavour to be humanitarian in all that I do and say and I do not believe at all that any of the sciences be they social, environmental have any way whatsoever of being completely accurate – nor do those humans that choose to play in those arenas. We are all dealing with the unknown (even though we think we know some things), with knowledge acquired through observation, experiment and theory (often smattered with a bit of confusing technology) complicated by the impact of content and context amongst many other factors that may arise completely out of the blue. To those that blame weather forecasters, seismologists or anyone whomsoever for not being 100% accurate – I say this. We dance on a planet that is full of atoms, nanoparticles, wind currents, electrical charges, forces and so on right down to human emotion, cultural and societal factors, that choose to interplay in such a myriad of combinations and with so variant impacts and results that it is impossible I believe to be 100% accurate in a predictive sense. As a human being it is not possible to be 100% accurate. We wouldnt want to be – what a boring, predictable world it would be. And most certainly we should not ridicule, judge, put down, criticise, sneer or otherwise the efforts of another human being or group of human beings for that matter when they are simply trying to make sense of a rather complicated universe and share what knowledge, observations and theories they have with those around them with good intention. As for the honesty of WW? Well. One thing tells me that WW is honest. And Phil that is you – you are the face of WW and come across as a very honest, caring, clever, eager to assist person that could only gather those of the same around you. Finally – I often listen to this song – its a bit of a mantra for my life. Kinda takes the pressure off you when you listen to the words – we are not perfect – we are only human and its pretty counterproductive to try and do anything else but our best nor to judge ourselves or our fellow man too harshly. http://www.youtube.com/watch?v=iJew4fxHl1U Oh and ….. a saying of mine follows. Phil – when someone says to you "what you are reporting is not normal or totally correct for a weather forecaster"… You say, loud and proud – "For the record Normal is but a cycle on the washing machine. And Hey – cut me some slack – Im just having a bash at writing a page in the universal book of right/wrong/total correctness – somehow I think Im going to have trouble finding us all a publisher brave enough to print it though."

View more comments

Related Articles