[Eoas-seminar] MET Dissertation Defense - Daniel Halperin - 6/1/15 - 3:30 - LOV 353

eoas-seminar at lists.fsu.edu eoas-seminar at lists.fsu.edu
Tue May 26 09:59:29 EDT 2015

Meteorology Seminar

Daniel Halperin

PhD Meteorology Candidate

Major Professors:
Dr. Henry Fuelberg and Dr. Robert Hart

Investigating the Potentially Untapped Predictability of Tropical Cyclone Genesis in Operational Global Models

Monday June 1st

3:30 PM

Werner A. Baum Seminar Room (353 Love Building)
(Please join us for refreshments served outside room 353 Love @ 3:00 PM)


There is an operational need for accurate tropical cyclone (TC) genesis forecasts. Global numerical models are an important genesis guidance tool, but each model has strengths and weaknesses. Further, the interpretation of when genesis occurs in a model forecast field can be subjective. Thus, this study seeks to create an automated, objective, statistical-dynamical TC genesis guidance tool for the North Atlantic and eastern North Pacific basins based on output from the Environment Canada Global Environmental Multiscale Model, the National Centers for Environmental Prediction (NCEP) Global Forecast System, and the United Kingdom Meteorological Office global model. Another goal is to determine how well important genesis processes in global models agree with those theoretically proposed. This research also attempts to identify the characteristics of successful and failed genesis forecasts.

First, historical global model forecasts of TC genesis over the past decade (2004-2013) are verified. Using this archive of genesis forecasts, univariable logistic regression equations are created to reveal statistical relationships between relevant variables and genesis probability. These statistical relationships are compared to theoretical relationships between atmospheric variables and TC genesis. Results show several expected and counterintuitive statistical relationships, with some disagreement among the models. Multiple logistic regression equations then are developed to provide probabilistic genesis forecasts. Separate equations are developed for each global model, basin, and forecast window. A consensus regression equation also is developed. These equations are tested operationally during the 2014 hurricane season. Verification using this independent dataset indicates generally well-calibrated guidance. Homogeneous comparisons of the consensus regression equation and National Hurricane Center Tropical Weather Outlook genesis probabilities are presented.

Case studies and composite analyses are conducted to gain further insight. Case studies from the following categories are selected: (1) African Easterly Wave genesis over the Main Development Region; (2) genesis from stalled frontal boundaries; (3) genesis via tropical transition; and (4) genesis over the Gulf of Mexico. Hit, miss, and false alarm events are compared. Storm centered composite analyses also are constructed to examine differences in the environments between hit and false alarm forecasts. Separate composites are made for the eastern Main Development Region (where the GFS false alarm rate is greatest) and the remainder of the North Atlantic basin. Statistically significant differences between hit and false alarm cases are found for all variables analyzed with various areal extents.

Results from the case studies and composite analyses will help guide the selection of new predictors to test for inclusion into the multiple logistic regression equations. The case study of Sean (2011) suggests that changes to the TC identification algorithm are needed to better detect subtropical to tropical transition. Real-time guidance products based on the logistic regression equations are being evaluated by hurricane specialists at the National Hurricane Center. It is possible that the products will be selected for operational implementation pending further testing and evaluation during 2015.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.fsu.edu/pipermail/eoas-seminar/attachments/20150526/9f65f9cb/attachment.html>

More information about the Eoas-seminar mailing list