A "Post Mortem's" Role in Improving Forecasting?

[Darren Addy]

What role does a "Post Mortem" examination have in understanding storm-related meteorology and forecasting? It would seem that, since hindsight is 20/20, that looking BACK one could learn a lot that might make one's forecasts better. "Storms triggered HERE because an outflow boundary from the previous day's storms lifted unstable air at the time of day that the cap was breaking from daytime heating." (to give a succinct hypothetical example)

Does anyone do this? If so, would you share what you are looking for (and perhaps your favorite data sources for the Post Mortem?)

I can see why this might be difficult to do. First, we are generally more interested in looking into the future than looking back (particularly during a busy chase season, there may not be enough time to analyze PAST day's data). Secondly, many people are content to simply be in the general area and then pick a target as it develops in real time. Understanding WHY a particular cell became a monster may be less important (for some) than simply being on it.

Are post mortem's a dumb idea or part of good science?

Darren Addy
Kearney, NE

 

[rdale]

They're great - that's why every severe weather conference you go to will be filled with quick and dirty explainers about storm days (or busts.) Many times even looking back knowing exactly what's going to happen still doesn't help though, but it certainly can make you a better forecaster by trying...

- Rob

 

[Glen Romine]

As you would expect, a part of getting better at forecasting tornadic events is learning what to look for - and the way you do that is looking at events where tornadoes occurred, and dissecting the elements that came together to create it. While no two events are ever exactly alike - often very similar events can yield similar results. What is less popular is figuring out why days when everything was expected to go nuts failed to materialize, but understanding these are equally important. Ideally, you would make a forecast in the morning, outlining your thought process in making that forecast, and then at the end of the day see if things evolved the way you expected them to. Sometimes you will get the "right" answer but for the wrong reasons, and that is critical to note in further refining forecasting skills.

Glen

 

[Joel Wright]

I think with just about everything, it's always a good idea to go back and see why something happened. The best way to predict the future is to study the past.

 

[Jeff Snyder]

I like to use the term "comparative forecasting" for the use of pattern recognition as one of the main components in one's forecast. While much of pattern recognition is gained through experience, you can go through and look at past events to see if the outcome of those events can shed any light on the possibilities of the current event/setup. Below are just two webpages that help this, though they are also very interesting just go glance through...

SPC Severe Thunderstorm Events: http://www.spc.noaa.gov/exper/archive/events/

Bangladeshtornadoes.org US Case Study page:
http://bangladeshtornadoes.org/UScases.html