Skip to the bottom of the diary if you're looking for the latest info on Tropical Storm Debby.
Please note that this is not meant as an attack on the National Hurricane Center. They are great forecasters who do an excellent job. However, they (and the models) got the forecast horribly wrong yesterday, and I've had a few people ask me why. I feel people deserve at least something of an explanation as to why the NHC said yesterday that it would hit Texas, but now it's hitting 800 miles away in Florida.
For the last few days, most people who watch the news (or live in the south) have been aware of Tropical Storm Debby. It wasn't exactly a surprise to meteorologists and weather watchers. Conditions became favorable last week for deep thunderstorm activity to develop in the Caribbean, and the area of storms very slowly drifter northward into the Gulf of Mexico.
At the same time those storms were getting their act together, a large ridge of high pressure built in over the eastern United States. The ridge essentially left the door open to the Gulf of Mexico, allowing the system free reign of the area. Big, practically weatherless areas like the Gulf this week provide very little steering current for a storm to latch onto and move in one definitive direction. Accordingly, the system that would become Debby just sat there and drenched parts of Mexico, Cuba, and Florida for a few days.
As upper-level shear let up and the system was able to get its act together a little more, the National Hurricane Center (and other organizations that run weather models) flipped the switch on the storm Friday morning and started running the weather models on it to see where it would go. One of the best ways to view hurricane track models is a "spaghetti model plot." This plot takes all the available model runs and plots their forecast tracks on the same map. The resulting chart shows all the model forecast lines overlapping each other, resembling a plate of spaghetti. The spaghetti model plot for Invest 96L -- what the low pressure area was designated before becoming Debby- was less than comforting:
The first spaghetti model plot of Invest 96L, which would later become Debby.
There were two main schools of thought in the models on Friday: 1) a high pressure area would form over the south and steer Debby into Texas; and 2) the high wouldn't form (or be very weak) and Debby would continue to drift into Florida.
The models didn't improve much in their agreement by Friday afternoon. The 18z run (about 200PM EDT) of the models showed everyone from Brownsville, TX to Tampa, FL in play:
A few hours after that model run came out, the National Hurricane Center issued their first official advisory on newly-formed Tropical Storm Debby. The NHC acknowledged that the models were split between going west into Texas and going northeast into Florida, but opted to go the westerly route because "more than half of the GFS ensemble members are in agreement with westward movement." The NAM, HWRF, CMC, and NGP models showed Debby going into Texas in their runs on Friday as well.
This was their first forecast track:
Just a few hours after they issued the above forecast, the models already started backing off and going a more northerly/easterly route. The black line in the following spaghetti model plot from around 8PM EDT Friday indicates the NHC's official forecast:
The forecast models continued to trend towards the north and east, making it harder for the NHC to justify their forecast sending Debby into Texas. By 5AM Saturday, the NHC had adjusted their track northward a little bit, but still kept Debby raking the coast and going towards Texas.
The 11AM Saturday advisory saw the NHC finally moving with the model trend by shifting the track of Debby to hit the mouth of the Mississippi River in southeastern Louisiana, probably damn near giving everyone in New Orleans a heart attack. However, they still remained to the west of the model trends.
The Saturday afternoon models came in and pretty much rang the death knell for the NHC's forecast, with a majority of the models agreeing pretty well on a track right into northern Florida:
The Hurricane Center responded by adjusting the storm's track up to the Florida Panhandle, the one spot some of the models (including the GFS) consistently showed throughout the time period, by 500PM EDT Saturday.
This forecast and ultimate path of Debby could change after the posting of this diary, and I'll update accordingly at the 11PM EDT update.
So...why did the National Hurricane Center get their forecast off by about 800 miles? I can offer several explanations.
- Meteorologists notice patterns in the models, and experts like the ones who work at the National Hurricane Center get pretty good at spotting which models perform the best under certain scenarios. The best I can tell, the forecasters at the NHC were pretty confident that one set of models were right and the others were wrong, even as they were trending east.
- The models did a terrible job with this storm, just because of the ambiguity of the scenario. One little change in the atmospheric setup meant a few hundred mile shift in the track of Debby.
- Since the 1990s, the National Hurricane Center has been very skeptical of making large shifts in their forecast tracks over short periods of time. They like to gradually shift the track in one direction or another. They didn't have that luxury this time. Unfortunately, the model they followed was wrong, so they were about 800 miles off the actual trajectory of the storm. They had to stay within the model guidance while still making a forecast that sounded reasonable. That's why they somewhat split the difference and aimed it at Louisiana, then rejigged it to go towards the Florida Panhandle.
The reason they don't want to make large jumps in the forecast is because it would erode the public's confidence in their forecasts if it keeps making sudden changes. Of course, getting it wrong also erodes confidence. It's a fine line.
- In the above reason, I said "since the 1990s." For a while, the Hurricane Center made their forecast tracks in a somewhat political manner. They would purposely forecast storms to hit major cities even if it wasn't going to hit them, just to make sure that the public and media were paying attention to the system. A large internal controversy ensued and the NHC had to drop the practice of political forecasting.
Sometimes even the best get the forecast wrong. It happens to every meteorologist and weather forecast at one time or another, and to some degree. Debby was and still is a hard storm for which to forecast. It's a good case study in why needing to be prepared is key. Things can change on a dime.
Discussion about Debby's impacts and future
As of right now, Tropical Storm Debby seems to be on track to hit Florida with extremely heavy rain, strong winds, and tornadoes.
Heavy Rain/Flooding: Extremely heavy rain and resulting flooding will continue to be the big story with Tropical Storm Debby as it very slowly makes its way towards the coast. The storm is lopsided, with almost all rain and thunderstorm activity confined to the north and east side of the storm. Unfortunately (or fortunately) for Floridians, the rain is pretty much covering the whole state of Florida, and that's not expected to change anytime soon. The HPC is predicting an incredible 20+ INCHES of rain across the bend in the Florida Panhandle south of Tallahassee. Flooding is going to be the biggest concern this week. If you live in a low-lying area, you need to make plans to evacuate with little to no warning. Do not drive through a flooded roadway, nor wade through/swim in flood waters. You could get sick or killed from the toxins and venomous animals/bugs in the water, let alone the fact that you could get swept away and drown.
HPC's 5 day total accumulated rainfall forecast, in inches.
Tornadoes: Tornadoes are a concern with any landfalling tropical system, and Debby is no exception. Numerous tornadoes have already touched down across Florida, including one that killed a woman in her Venus, FL home. Keep a close eye on your local National Weather Service for the possibility of tornado warnings. The warning time for tornadoes in tropical systems is greatly diminished, so you may have little to no warning at all. The tornadoes won't be exceptionally strong -- EF0 to EF2 intensity -- but even a small, weak tornado can be deadly if it hits an inhabited area.
Rip Currents: Numerous people have already been killed due to rip currents all across the northern and eastern Gulf Coast. The rip current risk is very strong. Don't go in the water. If you go in the water, don't expect someone to be there to save your stupid ass. But, if you do find yourself caught in a rip current, swim parallel to the shore until you don't feel the current pulling on you.
Wind: While the wind isn't the biggest threat in Debby, it still has 60 MPH winds (as of 8PM EDT Saturday). That's enough to knock around loose objects outside and do some real tree damage, especially since the ground is so wet with the deluge of rain with the system. Make sure loose objects are tied down and secured, and make sure you avoid tall trees and high-profile vehicles until the winds die down.
Surge/Tide: I'll defer to the 8PM advisory from the National Hurricane Center:
STORM SURGE...THE COMBINATION OF A STORM SURGE AND THE TIDE WILLThis is a slow moving tropical storm. Things can and do change in a short period of time. Keep an eye on it.
CAUSE NORMALLY DRY AREAS NEAR THE COAST TO BE FLOODED BY RISING
WATERS. THE WATER COULD REACH THE FOLLOWING DEPTHS ABOVE GROUND AT
THE TIMES OF HIGH TIDE OVER THE NEXT FEW DAYS...
APALACHEE BAY TO WACCASASSA BAY...4 TO 6 FT
FLORIDA WEST COAST SOUTH OF WACCASASSA BAY...2 TO 4 FT
COASTAL MISSISSIPPI EASTWARD TO APALACHEE BAY...2 TO 4 FT
SOUTHEASTERN LOUISIANA...1 TO 3 FT
THE DEEPEST WATER WILL OCCUR ALONG THE IMMEDIATE COAST IN AREAS OF
ONSHORE FLOW. SURGE-RELATED FLOODING DEPENDS ON THE RELATIVE
TIMING OF THE SURGE AND THE TIDAL CYCLE...AND CAN VARY GREATLY OVER
SHORT DISTANCES. FOR INFORMATION SPECIFIC TO YOUR AREA...PLEASE
SEE PRODUCTS ISSUED BY YOUR LOCAL NATIONAL WEATHER SERVICE OFFICE.
National Weather Service Tampa FL
National Weather Service Tallahassee FL
National Weather Service Mobile AL
National Weather Service New Orleans LA
National Weather Service Lake Charles LA
National Weather Service Houston TX
National Weather Service Brownsville TX
Update 930PM EDT: I forgot to add that the models screwed this one up too, so I've added that into the list of reasons above.