Following on from the chatter on some of the threads regards to temp forecast, has anyone noticed that perhaps the accuracy of 3 day forecasts has dropped somewhat over the past say 1-2 years?
Obvs this isnt targeted at TWO - I wouldnt be that rude. This is a general sweeping statement from the BBC widget to GFS etc. through to the forecast you find on all manner of weather sites, inc Google.
So whilst perhaps we have seen a improvement in the modeling of days T3-10. The accuracy when we get to within a few days, now often flip flops so much that right up until the weather is actually happening, can the forecast change, and quite considerably too.
Take Sunday for example. One day it was forecast rainy, the next was dry, the next was rainy , the next was dry. Turned out to be rainy, and an outdoor event we had planned ended up being a bit damp (but not spoilt 🙂 )
Ive seen countless situations where a forecast will start off as rain - all day, to be downgraded in the hour the rain was meant to start, to perhaps an hour of rain, to no rain at all.
Ive also seen swings the complete opposite way.
Its almost laughable when we consider how close to "Nowcast" the forecast has become, with Radar data deciding what the forecast will be, rather than any scientific knowledge being applied.
My point to all of this is:
- Are we in an era of TOO much information, meaning we are checking every minute,, rather than perhaps 3 or 4 times a day (Pre internet regular weather forecast slots linked with News forecasts), so perhaps previously didnt notice such forecast swings.
- Have the models perhaps dropped their accuracy
- Are we flooded by too many forecasting models, some better than others, which perhaps steer forecasters decisions the wrong way
- Are the forecasts provided too automated and rely too much on Modeling, calculations of which have no real intelligence to them.
Genuinely interested to see what peoples views are to this.