Thursday, August 9, 2018

Why Do Meteorologists Get The Forecast Wrong?

On Wednesday, August 8, 2018 the weather forecast for Texoma called for chances of rain, up to 70% in some areas.  But, many folks in the area would quickly point out on Thursday, August 9, 2018 that there was no rain the day and night before or that it was very little, especially in Wichita Falls.

So this creates the question, "why do we get the forecast wrong?"

Although many many meteorologists/forecasters will claim their forecast was 100% correct (even locally); the fact remains that many if not a majority of the folks in the area would say, "no, you were wrong."

So, what's going on here?  Who's right? And, why?  Why is it so hard to forecast a simple area of thunderstorms?

There is no quick answer but I can do my best to explain.

A meteorologist/forecaster has a vast array of data at his/her fingertips to produce a weather forecast.  There are computer models of all sorts and kinds for all areas of the United States and the world.

The computer models take in data from the atmosphere all the way from the ground to tens of thousands of feet above.  This information is then put into each of the many computer programs to analyze the data.

Each computer model will compare the current information from the atmosphere and times in the past when it was similar.  For instance if the computer looks at the atmosphere over Texoma and sees that last time these conditions were present it did rain, it would produce a forecast that produces rain now.  The computer sees the situations as similar with a likelihood of a similar outcome.  But, this is just one part of producing a forecast by the computer models.

In addition to comparing what happened in a similar situation in the past, the computer models have written into them independent and proprietary software that forecasts the weather.  Many of these elements of specializing each computer model are not known outside of the developers and companies simply because the computer model formula is a secret and is not shared with competitors.

Another aspect of computer model forecasting is the fact that sometimes the computer model itself does not 'initialize' correctly.  For example, if a computer model takes in the atmospheric data and then produces a forecast for an area such as Texoma and Wichita Falls the actual forecast may begin incorrectly.

Here is an example. If a computer model looks at the data and says the temperature in Wichita Falls at the beginning of the forecast period should be 80° but the actual recorded temperature was 70° then we can determine that this computer model is already incorrect from the beginning.  Being 10° off in the very first time frame of the forecast is a major problem.  When this happens often forecasters will then throw out that model for this forecast period because it didn't "initialize" correctly.

Another problem with some computer models is they have a built in (either by accident or design) bias towards one element or another.  Let me give you an example.

Three of the most used forecast models are the GFS, NAM and EURO.  For this area of the nation, western north Texas and southwest Oklahoma,the NAM model has a bias in summer to produce forecast temperatures that are a bit too warm.  For example if the NAM says it will get to 105° in Wichita Falls today then most forecasters understand that this model has a bias and the high will most likely be around 99°.  As mentioned above this bias is only evident in the summer months.  Therefore, in winter and surrounding seasons the model does a bit better.  A good meteorologist/forecast must take this into consideration.

These are just a few of the reasons (not excuses, at least not for me personally . . . others have ego that get in the way, LOL, they think they are always right no matter what) why the forecast can end up wrong.  And, sometimes we get it very wrong.

In reality I've just scratched the surface on why a forecast can go wrong.

If you found this interesting and want to know more about this let me know.  Become a follower/member of my blog and then you can leave a message on there and let me know what else you would like to know about meteorologists/forecasters getting the forecast wrong.

Bryan W. Rupp