Can you trust the weather forecast? - The Irish Times

vicwatson

Irish Times article today -

http://www.irishtimes.com/newspaper/weekend/2012/0714/1224320029733.html

Hammer Archer

One of the things that I like to do when people complain to me about weather forecasts not being 100% correct is hand them my notes I used while studying Dynamical Meteorology and Numerical Weather Prediction (full of quasi-geostrophic approximations and 4D-Variational data assimilation etc.) and ask them to use these equations/assimilation techniques to predict tomorrows weather. Then explain that these equations that they see are actually extremely crude approximations of the atmosphere and can't really be used in operational forecasting as they are too "simple".

Interesting read. I think Gerry makes a good point on probability forecasting. I think it should be used here as it's used in other countries. But then again, I'm not sure if the majority of the population would necessarily understand these types of forecasts.
I hope this article will help to educate at least some of the population to the fact that weather forecasting, especially in Ireland, is pretty feckin difficult.

The trouble is that a look at 24-36 hours previous satelite imagery would produce a fairly accurate forecast for Ireland over the last few weeks. The crappy rain we've had has at times been incredibly localised so reading "up to 20mm" of rain and getting 2mm one day and "5mm" and getting an inch the next day doesn't instill confidence in forecasts.

M.T. Cranium

It is tough, but I think most regulars on this forum realize and understand that weak precip systems are subject to quasi-random fluctuation processes, sometimes it is a lot easier to nail precip forecasts when you're dealing with stronger systems that have a history of regular motion (whether linear or curved). I do a lot of rainfall forecasting in UK, Ireland and North America and find that the results are best during winter and spring, also later autumn storm systems that have those regular and relatively long-lived histories.

We all tend to use the same guidance and so really this is more a question of state of the art technology than individual forecaster skill. If no guidance shows the outcome, any correct forecast made would have to rely on either private knowledge (outside the mainstream and available guidance) or compensating errors. Let's say you stumbled in and made a quick forecast, then later realized that the precip guidance you consulted was for 12h later than you thought it was -- then the system slows down unexpectedly -- and your forecast verifies ... that's compensating error.

I'm not saying this is a big part of my work history but I've known such things to happen. I once caught an experienced forecaster giving wind forecasts off a reversed flow paradigm around low pressure. That could have been ugly. I was also watching a public weather network broadcast one night around fifteen years ago on TV and realized that the announcer was giving a heads up for a violent thunderstorm off an "anomalous propagation" radar image (ground clutter not clouds). I saved them a lot of future embarrassment by sending a message saying check that out, no satellite imagery of storm etc.

So they took that out of their presentation at a very early hour and probably I was the only person who actually saw the severe storm forecast. :cool:

By the way, there will be between 0 and 323 mms of rain on Tuesday. I hope you get the amount you're expecting.