W. Caribbean Disturbance and the Difficulties of Pre-Genesis TC Forecasting

The Western Caribbean has come under quite the microscope in recent days with models continuing to come into agreement on something forming from the Central American Gyre (CAG, see: Ryan's previous discussion). The most recent National Hurricane Center (NHC) tropical weather outlook currently forecasts a 40% chance of tropical cyclone (TC) development in the next seven days.

Forecasting a TC that hasn't formed yet

It is somewhat of a misnomer to classify this as a "disturbance". For one, this is not your classic Cape Verde system that forms from a long-tracked African Easterly Wave. For another, as of now there simply isn't a disturbance for us track. Looking at satellite imagery for the area surrounding the Yucatan Peninsula - the exact area we will be looking for future TC genesis - you'd be hard pressed to find anything of note.

This brings us to the first point of contention with forecasting this system, the time range we are talking about. Current model guidance does not even begin to show TC genesis until sometime early-mid next week. Potential land impacts? We are talking potentially next weekend. The other difficulty with this system is forecasting how the gyre will evolve. Will it be able to spin up a discrete low level center (LLC) that is 1) far enough from land to not cause disruptions in the vortex development and 2) not too broad that it takes too long to get going?

TCs spawning from the CAG can be particularly tricky to forecast. Recall that the CAG is essentially a very large and broad low pressure that forms over portions of the Western Caribbean and Eastern Pacific. Where and when an area of vorticity may become concentrated and thus help form a TC can be a challenge to forecast. We have seen this in some of the model runs, with sort of a sloppy evolution for much of it's early life as it struggles to consolidate vorticity. Some model runs have even shown development in the EPAC from this same gyre - something that lead to more uncertainty regarding any potential NATL system.

Climatology

When forecasting long-range and complex setups such as this, it can be prudent to revert to some of the basics. Climatology, or in essence a historical look at TC behavior at a particular time/location, can give us a good understanding of how TCs might behave. Below is a composite of all TCs since 1949 that have occurred between September 21-30. We are interested in the area surrounding the Yucatan Peninsula in far Western Caribbean. So what does the climatology say? It appears that most TC that pass through or near this area end up curving N/NE once entering the Gulf. This makes sense, as more frequent, deeper, and farther south troughs are to be expected as we creep towards late fall/early winter. These troughs can act as a net to "catch" TCs and slingshot them N/NE before a ridge can build back in to force the TC westward.

Current Model Guidance

Model guidance has been all over the place in whether or not this system even forms, where it will go, when it should form, how intense it will be, etc. Let's see how the last six runs of the deterministic GFS have turned out for one week from now:

Not good! Wildly different outcomes in both track and intensity. I will sound like a broken record, but I implore you to stay away from these individual model runs, as they will lead you astray.

Let's instead take a look at the GEFS (left) and the ECENS (right) ensemble solutions:

What do we see? For one, both the GFS and EURO ensembles do agree on some kind of TC development in this area through the next ~9 days - although there is a very large spread in eventual tracks/intensities. The EURO ensembles seems to be favoring a slower moving, potentially weaker, and more western solution - evident by the large clustering of members in the Bay of Campeche (BOC). The GFS ensembles, on the other hand, prefer a quicker, stronger, and more eastern solution, with less members making it to the BOC and more staying off the eastern coast of the Yucatan. While not as clear of a signal, it appears that in both models the stronger solutions end up heading generally more northward, likely feeling more of an influence from a nearby trough, and weaker solutions having more of a westward component. This westward component would push the system into the Yucatan -> causing even more weakening -> causing more westward movement. This may be what the EURO ensembles are trying to show.

Which solution, if any, will verify? It's anyone's guess at this point. Best course of action is to look for any kind of convergence in solutions across different models or take some kind of blend between the two of them.

Concluding Thoughts

With a complex gyre interaction and wildly varying model solutions for a TC that isn't forecast to form until next week, it is near impossible to pin-point what this may look like, if at all it even forms. Attempting to predict fine details such as TC intensity, track, steering flows, etc. would be nothing more than speculation - and it would be highly irresponsible. For now, stay away from deterministic model runs until (if) the storm forms an LLC to track. Focus on trends in model guidance, ensemble support, convergence in solutions across multiple different models, and climatology. No matter what happens with this system, it is a great opportunity to review your hurricane plans.

-Ian Mutschler