Since the ECMWF started freed up more of their NWP climate data I’ve been making good use of their long-range 42 day forecast charts for temperature anomalies. I now have my doubts about just how just accurate and useful these long-range temperature predictions are. The main reason I say this is because they are continually being updated, which is obviously good, but they are also continually changing sometimes quite radically from previous runs, which doesn’t give me a lot of confidence in how realistic they are going to be. It’s rather like the three hourly forecast in the BBC and Met Office smart app that is continually synchronised to fit in with the current weather and latest weather radar. Perhaps what this is highlighting are the inexactitudes climate models have beyond day five.
Compare the latest forecast (above) with the one from last week (below) to see what I mean. This weeks version of the forecast for example now goes for a much colder second week in December as you can see. The second week is closer to average and now not mild. Temperatures across the Christmas period are not dissimilar and look near average but looking ahead the first full week of January is forecast to be milder than average.