Last Thursday, I wrote about the potential problems with mentioning possible tropical storms the minute they show up on a computer forecast model. I promised to look at a specific date and time on the GFS model every morning until that date – 06Z Thursday, August 17, 2017. The following day, I noticed several meteorologists mentioning a potential tropical storm the week of August 13-19 on social media.
By Monday, my point was proven. The model that I was watching no longer showed any hint of a storm anywhere near the Atlantic Coast on August 17. I had to look days in advance to find a shadow of what the forecast storm had been.
This morning, the GFS had dropped it altogether. Ironically, a few of the other models have picked up on a potential, weak storm that could travel up the coast at the beginning of the week. In fact, I took a few screenshots of 0Z Monday, August 14.
Comparisons between what I have seen today and last week:
Last week, the GFS had the storm slamming the coast of Georgia. Today, the ECMWF and CMC models have it staying well out to sea.
Last week, the GFS showed a strong low pressure system with a pretty high amount of moisture available to it. Today, the ECMWF and CMC models show a weak system with about the same amount of available moisture that the east coast of South Carolina is experiencing today – typical summertime moisture for that coast.
Last week, the GFS showed winds around the center of circulation around 75 miles per hour. Today, the ECMWF and CMC models show winds around 20 miles per hour.
It is still possible that the system develops into something worth talking about in the next few days. The National Hurricane Center is keeping an eye on it. My point here is that a great deal can change in just a week or two when it comes to long-range weather forecasting. The answer to my question of “How far out is too far out?” may not be cut and dry, but in this case, we can safely say that two weeks is still too far out for any real accuracy, at least with the technology we have today.
Note: 0Z is midnight Zulu time, which translates to 8:00 PM Sunday evening in Eastern Time.
For those who love weather maps, here are some of the screenshots I have taken for this little experiment: