Friday, June 16, 2017

Tropics Trying to Come Alive - 6-16-2017

DISCLAIMER: The views expressed in the blog post are strictly my own and do not necessarily express the views of the National Hurricane Center, National Weather Service, or other official sources unless explicitly stated. The ideas expressed here should not be the sole source of your information. Always consult official sources before making any decisions.


Tropical Weather Outlook from the National Hurricane Center showing our two areas of interest. Valid as of 7 PM CDT Friday (June 16, 2017).
We are in Hurricane Season, yes. But having 2 systems that bear watching in the tropics is a bit unusual for June, especially Invest 92L in the central Atlantic. Climatologically speaking, Invest 92L is in an area uncommon for development this time of year.

The climatologically-favored areas of tropical cyclone formation during June. Image from the National Hurricane Center. Source: http://www.nhc.noaa.gov/climo

Invest 92L

Infrared satellite imagery of Invest 92L. Source: http://www.ssd.noaa.gov/PS/TROP/floaters.html
Let's start with the system that has a higher chance of development in the short term--Invest 92L. Since this is one of the first times we're talking about "invests" this season, remember that an "invest" is an area of investigation. These are called 90L to 99L ("L" for Atlantic), and the numbers are reused through the season.

92L isn't looking too shabby, especially for June. This tropical wave came off the coast of Africa a couple days ago, and has maintained its organization fairly well. Here's the National Hurricane Center's outlook for 92L:

Satellite images indicate that an area of disturbed weather
associated with a tropical wave is located about 1800 miles east of
the southern Windward Islands. This disturbance has become
better organized today, and additional development is possible
during the next day or two before conditions become less favorable
for tropical cyclone formation. This system is expected to continue
moving toward the west at 15-20 mph over the tropical Atlantic
during the next several days.
* Formation chance through 48 hours...medium...40 percent.
* Formation chance through 5 days...medium...60 percent.
Conditions are expected to become less favorable for development in a few days as it approaches the Windward Islands. But it's still several days away from land--plenty of time to watch this. Just so you can see them, here are the spaghetti plots:

Spaghetti models for Invest 92L (as of 7 PM CDT June 16, 2017). Source: http://www.tropicaltidbits.com/
 It's important to note a couple things about spaghetti models:

  • These are just one run of these models, and only give a general idea of it's potential movement. Thus they can change from run to run.
  • These don't tell you about the possible intensity of whatever system.
  • These don't tell you about any of the impacts, only about the center of the system.
Generally, the system is expected to move toward the southern Windward Islands, and maybe brush the northern coasts of Guyana and Venezuela, bringing some squally weather for those areas.

The Mess in the Caribbean

GOES-16 Infrared satellite imagery showing the Gulf of Mexico and the NW Caribbean Sea. Source: http://weather.cod.edu/satrad/exper/. NOTE: GOES-16 data is undergoing testing as is considered experimental and non-operational.
Yes, I'm using "mess" as a technical term. This view doesn't even give the full view of the Caribbean Sea, but I used it because the GOES-16 imagery is meteorological awesomeness. Very high resolution and rapid data. But, let's give you a complete view of the Caribbean.

Infrared satellite imagery of the Caribbean Sea. Source: http://www.nhc.noaa.gov/satellite.php
Here's the official word from the National Hurricane Season (as of 7 PM CDT, June 16, 2017):
Surface pressures are falling in the western Caribbean Sea and
the areal coverage of the disturbed weather has continued to
increase. Conditions appear to be favorable for gradual development
of this system while it moves slowly northwestward across the
Yucatan Peninsula this weekend and into the south central Gulf of
Mexico early next week. Regardless of development, heavy rains are
likely to spread over portions of Central America, the Yucatan
Peninsula and western Cuba during the next several days.
* Formation chance through 48 hours...low...20 percent.
* Formation chance through 5 days...medium...60 percent.
There's nothing organized about it at this time, and that's causing a lot of problems for forecasting. How? Allow me to explain.

Without anything defined in the Caribbean, the models can't lock on to anything, but they will try to make some sort of sense of what's going on with what data they have. Unfortunately, there isn't a whole lot of data in this area, so the model uses its own forecasts. This can lead to the model consistently performing poorly. Also, as of this writing, this area has not been designated an "invest" (again, due to the lack of an organized low pressure center). This means that we do not have spaghetti models to show you. Let's break down some models.

GFS

GFS model valid Monday morning. Model image source: tropicaltidbits.com. Annotations made by me.
Okay, let's break down the image. The different color shadings represent the flow at 500 millibars (about 17,000 feet). I tried to illustrate the flow using the black and white arrows. Our surface low is the red and white "L."

The GFS predicts that there will be a ridge from Texas westward (part of what will bring record heat to the Desert Southwest) and another ridge from the East Coast into the Atlantic (ridges = high pressure = "H"). However, there is a bit of a break, or "weakness" in the ridge. This may help draw the low pressure system further north.

GFS model valid Wednesday morning. Model image source: tropicaltidbits.com. Annotations made by me.
By Wednesday morning, a trough dips down, according to the GFS, and starts to pick up the low pressure system and starts lifting it northeastward while it weakens further. The low at this point is relatively weak, but the main threat would not be wind, but would be heavy rain according to the GFS.

GFS model precipitation through Thursday afternoon. Model image source: tropicaltidbits.com.
Wherever this system goes, the eastern and northern sides will see quite a bit of rain. The GFS paints well over 6" of rain over the Big Bend of Florida through Thursday afternoon.

But, if you read my posts, you know that I am quick to point out to never use one run of one model. So, let's take a look at what the past runs have done with this system.

GFS model trend using the last 10 runs, all valid for Wednesday morning. Model image source: tropicaltidbits.com
The last 10 runs of the GFS haven't been exactly consistent, ranging from the low being in the southwestern Gulf to northwest Florida in just 10 runs. However, it is somewhat consistent in strength, keeping it weak. But, again, we don't have an organized system to pinpoint yet, so we don't know if this model (or any of its runs) will be correct.

Let's take a look at another reliable model...

ECMWF (European)

ECMWF model valid Monday morning. Model image source: tropicaltidbits.com
The ECMWF, another reliable model, shows that the ridging is a bit stronger than the GFS. The result? Any northward motion gets blocked since there's little to no weakness. It gets shoved to the west. So what happens further down the line?

ECMWF model valid Wednesday morning. Model image source: tropicaltidbits.com
It moves toward Mexico instead of the northern Gulf Coast. The mid to upper level flow blocks any northward progression.

However, a lot of moisture will still be drawn up from the Caribbean, and very heavy rain would still happen to the east and north of the center. I can't show the precipitation data from the ECMWF due to proprietary issues. But how consistent has the ECMWF been?

ECMWF model trend of last 6 runs, all valid Wednesday morning. Model image source: tropicaltidbits.com
Actually, the ECMWF has been relatively consistent in placement and strength. But does that mean that it's more accurate? Not necessarily, and only time will tell what actually happens.

One last model (or set of models) to show you.

GFS Ensembles (GEFS)

GFS Ensembles showing different low positions from Sunday afternoon through Wednesday afternoon. Model image source: tropicaltidbits.com
You've actually seen this model before, sort of. This is the GFS model run several times, each time being tweaked slightly. This results in different outcomes. Each red number you see is a low pressure center that each run produced. The more clustered they are, the more confident the model is in that solution. The more spread out they are, the more uncertain the forecast. The GFS Ensembles are actually fairly consistent in the central Gulf of Mexico, but it quickly become uncertain, and it has to do with the ridge and/or weakness we've seen in the other discussions.

Summary

So, you've read this far, what are the takeaways?
  • Invest 92L is still really far away from land, and won't bother anyone for several days. But it does have a 60% chance of formation during the next 5 days.
  • The mess in the Caribbean is just that right now, a mess.
  • There is much uncertainty in the forecast with the mess in the Caribbean. Part of this is due to nothing being organized right now. Another factor is its large size. It takes longer for larger systems to become organized.
  • The models are very split as to what will actually happen.
So, what should you do?
  • Don't panic! There's still time to watch what will happen.
  • Keep up with forecasts this weekend.
  • Consult official sources when making any decisions.
  • Now is a good time to review your emergency plans and make sure your kit is stocked. You can never be over-prepared!

Sunday, June 11, 2017

The Dangers of Storms that Haven't Even Formed

Yeah, maybe the title was a little "clickbaity" (if I may make up a word), but it certainly got your attention. Every Hurricane Season, we seem to go through the same thing: someone posts a picture of the GFS out 384 hours (16 days) that shows a hurricane hitting somewhere popular. Then the same page posts the lone image with a caption reading something like:
"Major Hurricane will Strike this U.S. City! Catastrophic Destruction is CERTAIN!"
And every meteorologist collectively rolls their eyes. Well, let's talk about some ways that you can be sure you don't fall prey to these social media storms.

Don't Believe Everything You Read

Social media is a very useful tool for dispersing information, but that doesn't mean that it's always trustworthy. Sometimes you end up with posts like this:

Hyped-up hurricane "forecast." Image source: http://media.nola.com/hurricane_impact/photo/bogus-hurricane-path-d2cf288128cb5eb0.png
 Notice a few things on this single image:
  • Have you ever heard of WeatherAlertCentral.com? Research before believing. Confirm that what they post is accurate and corresponds with official sources (National Weather Service, National Hurricane Center, your favorite TV meteorologist).
  • The title automatically goes to the worst case scenario, and it's in ALL CAPS! Hyped titles like this should cast some doubt about its credibility.
  • The map. Notice how nearly every major Gulf Coast city is in the "path" of this "hurricane," including the entire state of Florida. That should also be another red flag. Again, pay attention to official sources.
Pages like this are posting content like this simply to get more likes and shares. Look at how many people shared the post: 77,159. Yikes! This spread of misinformation makes social media potentially dangerous.

Errors Grow Quickly

Let's get a little geeky here. Every computer model is based on math and physics that attempt to describe the atmosphere. There are also some things that have to be calculated using workaround methods (parameterization). Small differences in the math/physics from the real atmosphere can lead to enormous errors rather quickly. Not only that, but the model tries to make a guess at what the atmosphere looks like using its own data and some real observations. While the models do adjust somewhat for the observations, it can also reject actual observations if it's too different from what the model thinks. If there is no data available (a big problem over oceans), then there is (next to) nothing that can correct the model. Thus, the model keeps any errors, and it can maintain those errors for a long time. Remember, all of these errors quickly multiply the further out you run the model.

For example, let's take a simple math problem:


We know that 2+2 is equal to 4, but what if we made a slight mistake and said it was equal to 5. An error of 1 doesn't seem so bad, but those errors can grow very quickly. If we use the error and multiply it by itself once, we're now off by 9. Keep multiplying, and you see the error grow substantially. Once we multiplied the wrong answer by itself for the fifth time, we are now off by over 2,100. Yikes! While the errors likely don't accumulate that fast in the computer models, this gives you the idea that errors can grow quickly. This is why you shouldn't trust the models too far out. They are just guidance, not gospel!

Also, don't just look at one run of a model or just one model. "If you live by one model, you will die by that model." No one model is perfect all the time. Also, things change from run to run as we get new data put in. You want to look for some sort of consistency among the models and the runs of the models.

Stay Calm, but Be Prepared!

If you do see something from an official source that could be threatening to you, stay calm. I know its easier said than done, but one thing that can help you remain calm is being prepared.
  • Have your emergency kit fully stocked at all times (including food, water, batteries, etc.)
  • Know your evacuation zone and route.
  • Keep up with forecasts and the latest information from official sources.
  • Don't read the hyped posts. Those may cause you to panic more.
Remember, we are in Hurricane Season right now, so we do have to keep our eyes out for anything that may form.

TL;DR
  • Know your sources, and don't believe everything you read.
  • Do some research before believing or sharing something.
  • Don't trust models that show something more than a week out. Errors can grow rapidly.
  • Consistency in the models is better, but models can be consistently wrong.
  • No one model is always accurate.
  • It only takes one storm. You should always be prepared.