My Take on Erika

By: Lee Grenci , 2:56PM,GMT on August 30,2015

I watched with interest the evolution of forecasts for Tropical Storm Erika as it approached the Caribbean Islands this past week. Fairly or unfairly, the first impression I got was that southern jog in the storm as it approached Hispaniola was a huge surprise (see track below). The associated detachment of the low-level center of circulation from the deep moist convection also seemed to catch some forecasters off guard.



Yet both of these dynamic processes could have been expected. Oh, yes, I agree, hindsight is always 20-20, and I’m not trying to disparage anyone. But I worry that the preoccupation with computer models is pushing real weather analysis into the background.

A paper by Bender, Tuleya, and Kurihara was published in Monthly Weather Review in 1986. It essentially simulated the response of tropical cyclones whose circulation interacted with tall mountains on a number of tropical islands (given my concerns about relying too much on computer models, I think it is ironic that the relevant weather analysis in this case was derived from a computer model). In a nutshell, the results of these simulations were that tall mountains on tropical islands affected the basic wind flow upstream of the islands (Hispaniola was one of the islands in this study). In turn, the altered wind flow ahead of a tropical cyclone alters the movement and the structure of an approaching tropical cyclone, with the storm sometimes taking a southward jog (like Erika did). All the simulations showed a northward deflection in the track of the storm (when southward shifts were observed, the northward deflection occurred after the southward jog, which Erika’s track clearly shows). The altered basic flow also promotes the detachment of the low-level center of circulation from the deep moist convection (as was the case for Erika).

I heard explanations on television that were based on dry air and/or the decline in latent heat energy over land to support the expected weakening of Erika (which were fine), but I did not hear any forecasts related to Hispaniola’s tall mountains possibly altering the basic flow and thus causing subsequent changes in Erika’s track. Nor did I hear any forecasts heralding the detachment of the low-level center of circulation from deep, moist convection.

The moral of my story is that track and intensity forecasts based on complex interactions with tall tropical mountains usually require forecasters to perform a weather analysis that transcends the predictions by the operational computer models. In short, choosing the “model of the day” is sometimes not the best way to forecast complicated scenarios.

For me, combining ensemble forecasts with weather analysis based on past peer-reviewed research was the way to go (in times of large uncertainty, the use of ensemble forecasts is warranted, in my opinion).

Yes, hindsight is 20-20. I don’t pretend to have all the answers, but I can tell you that red flags went up in my mind (no pun intended) when I first realized that the mountains on Hispaniola would come into play.