Wednesday, October 7, 2015

MWN Lightning Round: Joaquin forecasts, 1000-year rain, and the weekend forecast

It's time for another edition of the MWN Lightning Round! Today we tackle a couple topics related to recent events on the east coast and tropics, as well as take another quick look at the weekend forecast, which has changed a bit in the past couple of days.

Hurricane Joaquin track forecasts

Early last week, eyes turned towards the warm waters of the Caribbean as Joaquin spun up and quickly became a major hurricane as it churned into the Bahamas where above average warmth and low wind shear aided rapid development. There's been a lot of talk in the aftermath (when you can break through the conversation about SC flooding, which is definitely worthy of plenty of it!) about the performance of the American (GFS) model vs. the European (ECMWF) model.

Once again, the ECMWF handily beat its American counterpart on the forecast track of Joaquin, indicating that the nearly 10-year long streak of no major hurricane landfalls on U.S. shores would remain intact. While the European model nearly pegged the eventual track of the storm as early as Monday night, the GFS didn't release the east coast from its sights until Friday morning (after the storm had reached the Bahamas) and took another 4 model runs (24 hours) to correctly show that there was a higher threat to Bermuda than the Canadian Maritimes. Credit goes to the National Hurricane Center for keeping the official forecast track off the east coast and not issuing tropical watches on U.S. soil in deference to the European model, despite the consensus of the rest of the models showing an eastern U.S. landfall.

The European model forecast for Hurricane Joaquin as of Monday night, September 28th, correctly showing a trip into the central Bahamas, followed by a turn to the northeast, close approach to Bermuda, and eastward turn south of 40° north. Graphic courtesy WeatherBell Analytics. Used with permission.
The American (GFS) model forecast for Tuesday night, September 29th, showing landfall near Norfolk, VA and significant direct impact to the Mid-Atlantic region. This forecast was made 24 hours after the above track produced by the European model. The GFS didn't eliminate direct impact to the U.S. as a possibility until late in the week. Graphic courtesy WeatherBell Analytics.
Having followed these models regularly for years, my primary comment here is this: the American model generally does a very good job in the short to medium-range (out to a week) and compares favorably with the European model in general. However, when the spotlight is on (Hurricanes Sandy and Joaquin immediately come to mind), it has fallen short. What must not be forgotten is that the massive northeast blizzard that was predicted to bury NYC last winter (and ended up being a "regular snowstorm" for a well-prepared northern metropolis) was actually best forecast by the American-made GFS, whereas the European led forecasters astray, at least in the area that gets the most media attention, the Big Apple.

What is a 1,000-year rain event?

Since we're talking about Joaquin, this is a good opportunity to discuss another misunderstood, and unfortunately poorly-explained, topic (at least by the mainstream media) that we've all heard in reference to the historic flooding this past weekend. Portions of South Carolina received a 1,000-year rain event. What does that mean, besides that it is historic, catastrophic, and devastating all at the same time?

Let's start with what it is NOT. First, it was not a 1,000-year FLOOD event. That is different from a 1,000-year RAIN event. A 1,000-year rain event does NOT mean that: a) it's been 1,000 years since the last rain of this magnitude, despite what Stephen Colbert or the Governor of South Carolina says, b) it'll be 1,000 years before this much rain falls again, or c) it only rains this much every 1,000 years.

A 1,000-year rain event actually means that there is a 1/1,000 chance that a certain amount of rain will fall in the given period, or that there is a 0.1% chance of it occurring. Could it occur again next year? Sure! A good way to describe this (and credit to the CoCoRaHS blog for this great example) is to imagine 1,000 ping pong balls in a box. All but one are white, the other is red. With your eyes closed, you pull a random ball out of the box. You have a 1/1,000 (0.1%) chance of grabbing that red ball. You then put that ball back in the box, mix them up, and grab a ball again. Once again, you have a 0.1% chance of pulling the red ball. The fact that you picked it out the first time does NOT affect your chances the subsequent time. Therefore, a 1,000 year rain event may not happen for 2,000 years, or you may see it much more often than every 1,000 years!

The last comment on these "recurrence interval" events, whether 1,000-year, 500-year, or 100-year event, is that the amounts of rain vary significantly from one place to another. For example, for Memphis, a 1,000 year rain event is 10.64" in 24 hours, while in Phoenix, it is 4.82" in 24 hours. Also, average recurrence intervals are calculated not just on 24 hour rainfall amounts, but on several different timeframes. The recurrence intervals for Memphis are shown below. As an example, a 100-year rain event is 8.02" in 24 hours or 3.20" in 1 hour.

Average recurrence intervals for precipitation events for Memphis, TN. Data courtesy City of Memphis/Shelby County Storm Water Management Manual.

Weekend forecast

In a post earlier this week, we discussed the warm weather mid-week (which we're experiencing... mid 80s can leave anytime thank you) and the passage of a cold front on Friday. That post mentioned minimal rain chances with the front, but computer models are picking up on more moisture along and behind the front and we've added a 30% chance of showers and maybe a few thunderstorms to the Friday afternoon forecast, as well as a 40-50% chance of evening showers.

GFS Wednesday morning forecast model of total rainfall for Friday and Friday night. Graphic courtesy WeatherBell Analytics.
The good news: we could really use some rain after a dry month, so a quarter inch of liquid will be welcome for some, plus it'll be gone for the Columbus Day weekend, so expect cooler and drier conditions Saturday and Sunday with clouds departing by mid-day Saturday. The bad news: you might need a poncho or umbrella at any Friday night events outdoors, including football games and the Levitt Shell concert. If forecast trends hold, we could see a steady rain during the evening with north breezes making it feel a fair amount cooler than the daytime hours when highs will reach the lower to mid 80s.

Keep abreast of the latest forecast conditions from MWN with our mobile apps, perfect for checking radar and our human-powered forecast while out and about. Links to download can be found below.

Erik Proseus
MWN Meteorologist

Follow MWN on Facebook, Twitter, and Google+
Visit on the web or on your mobile phone.
Download our iPhone or Android apps, featuring StormWatch+ severe weather alerts!

No comments: