While advances in tropical cyclone (TC) track forecasting have been substantial over the past few decades, and modest advances in intensity forecasting have occurred more recently, the quality of TC rainfall forecasts has not undergone the same rigorous verification. This is despite the 27% of total TC-related deaths being due to rainfall-induced flooding and that rainfall-related deaths occur more frequently than those due to any another weather-related hazard. A continual effort is needed to understand and better-forecast TC rainfall. This dissertation research seeks to contribute to this endeavor. A climatological dataset is created using 6-h Stage IV rainfall accumulations combined with Best Track 6-h locations for all TCs within 300 km of the U.S. Gulf and Atlantic coastlines during years 2004 - 2013. Stage IV data are used due to their higher spatiotemporal resolution, their extension to high latitudes, and because they have been found to be the superior option when compared to other TC rainfall data sources. The 6-h Stage IV rainfall accumulations are composited by shear magnitude and storm intensity in earth-, motion-, and shear-relative reference frames. Additionally, a full composite comprised of all storms is created. This compositing is done for TCs impacting the U.S. Gulf and Atlantic coastlines. Seven geographical regions are created within this domain to further composite the rainfall. The geographical regions are determined based on 2004 - 2013 Best Track (HURDAT2) landfall locations. Results show that some Stage IV rain rate characteristics, especially those in specific regions, are different when compared to prior findings based on satellite-derived rain rates. Results from the Stage IV-derived climatological datasets then are used together with track forecasts from the Global Forecast System (GFS) during years 2014 - 2016 to create 72-h TC rainfall forecasts. Separate forecasts are created for each 6-h TC position forecast based on shear magnitude, storm intensity, and the all-storms composites in earth-, motion-, and shear-relative reference frames. This yielded 1,290 verifiable forecasts during the 3-yr period. These statistical rainfall forecasts along with forecasts from the GFS and an R-CLIPER created from Stage IV data are verified using the Fractions Skill Score (FSS) metric. Results show that the statistical method based on shear magnitude in a shear-relative reference frame that used regional rainfall composites is the best performing of the methods. Additionally, FSSs from the statistical model are shown to be larger than those from R-CLIPER. The preliminary results from the statistical model show that this method is a viable candidate to supplement R-CLIPER as a statistical baseline TC rainfall forecast method. GFS analysis and forecast environmental parameters are composited based on the skill (FSS) of each forecast. Three categories are created: Top (FSS > 0.6), Bottom (FSS < 0.3), and Middle (0.3 < FSS < 0.6). This methodology is based on the desire to provide "guidance on guidance," i.e., suggesting to a forecaster whether the TC's environment is conducive to a skillful or not-skillful GFS rainfall forecast, and to help determine possible factors to increase the FSS of the statistical model. Results show that some aspects of the mean sea level pressure, 1000 - 500 hPa thickness anomalies, eddy flux convergence, and upper-level winds and divergence differ between skillful and non-skillful TC rainfall forecasts. / A Dissertation submitted to the Department of Earth, Ocean, and Atmospheric Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / 2019 / July 9, 2019. / forecast, precipitation, rainfall, tropical cyclone / Includes bibliographical references.
| Identifer | oai:union.ndltd.org:fsu.edu/oai:fsu.digital.flvc.org:fsu_752397 |
| Contributors | Hall, Tristan J. (author), Florida State University (degree granting institution), College of Arts and Sciences (degree granting college), Department of Earth, Ocean, and Atmospheric Science (degree granting department) |
| Publisher | Florida State University |
| Source Sets | Florida State University |
| Language | English, English |
| Detected Language | English |
| Type | Text, text, doctoral thesis |
| Format | 1 online resource (144 pages), computer, application/pdf |
Page generated in 0.0015 seconds