I will not swear, I will not swear. Om.This morning's medium-range forecast from the MSC for #yegwx is wrong. Worse than that, it is fiction and contradicts the very NWP model was supposedly used to generate it. An analysis in "n" parts. #abstorm #cdnpoli

So what the heck is going on here? Really, one could write a book about this saga. The fictitious forecast represents the confluence of multiple poor decisions & outdated technology used in the MSC. Thanks to @SpotWx for generating these products. 2/n #abstorm #cdnpoli

So a long time ago, the medium-term forecasts were unreliable. There have, however, been significant improvements in the NWP models over recent decades & we can now run the NWP models in ensemble mode to quantify the expected range of forecast values and uncertainties. 3/n

A common technique used in generating forecasts from NWP models is to post process the model data. All NWP models have their weaknesses and biases, and the post processing is meant to mitigate those, especially at longer lead times. 4/n

So back in the day someone came up with the idea of using the temperatures at 850 mb (~1.5 km above sea level) to estimate the near surface temperatures that we experience and which we see in the MSC forecasts. The idea being that temperatures at this level are more reliable. 5/n

Okay great. But there is a catch. This method works fairly well during the warm season, but is absolutely horrendous during the winter months when we have strong low-level inversions and also over complex terrain. So what happens then is that the forecast is degraded, a lot. 6/n