466 Historical Rainfall Data Analysis of Storm Characteristics in Tropical Cyclones and Noncyclonic Storms

Tuesday, 9 January 2018
Exhibit Hall 3 (ACC) (Austin, Texas)
Baxter E. Vieux, Vieux & Associates, Inc., Norman, OK; and J. E. Vieux

Analysis of historical rainfall data from rain gauges and weather radar reveals important storm characteristics of both tropical cyclones and non-cyclonic storms. Rainfall climatology was analyzed for the Bush International Airport (IAH) NOAA cooperative observer station, ID 41-4300. This site records 37 years of continuous record from 1970-2016 with only 0.39% missing data. An event definition of no rain for at least 6 hours was applied to segment the continuous hourly time series into discrete storm events. There were 6,453 events greater than 0.01 inches, but 1814 events greater than 0.25 inches. The average duration of a storm was 5.63 hours, averaged 0.49 inches in depth, and the mean inter-event time was 3.48 days. Seasonally more intense storms with higher depths occurred in the warm season (Jun-Sep) when tropical cyclones occur as opposed to the cool season (Oct-May). There was no trend observed in the number of storms, which averages 41 events per year, and ranges between 23 to 59. The average number per decade ranges from 35.9 to 44.3 with the lowest during the most recent partial decade (2010-2016), averaging only 35.9 per year. All storms with a duration from 4-48 hours were classified by the quarter of the storm in which the greatest accumulation occurred. The first quarter of the duration (1st Quartile) is predominant with 60% regardless of specific durations, 6, 12, and 24 hr. Centered events comprise 29.1%, while 10.6% were 4th Quartile. The ‘Tax Day’ storm on April 18, 2016, was not a tropical cyclone, but was one of the most extreme events on record. Its hyetograph is nearly coincident with the median 3rd quartile storm hyetograph. Rainfall frequency was also analyzed at the IAH gauge. Annual maxima depths for 1, 2, 3, 6, 12, and 24 hr durations were selected for determination of exceedance probabilities. Several probability distributions were considered, with the Generalized Pareto (GPA) exhibiting the best fit to the data. Storm motion was analyzed from archival weather radar data representing storm events during 1995-2016. Out of 744 events analyzed, 485 events passed quality control. Storm motion is measured in terms of direction (azimuth) and speed of storm cells passing over the target area. Storm motion exhibited strong seasonality. Rainfall does not typically fall uniformly, especially over large watersheds where storm total depth decreases as the area covered increases. Depth Area Reduction Factors (DARFs) were computed from NEXRAD for 248 storms during the period of record for KHGX. The 50th, 80th, and 90th percentile DARFs were computed and compared for extreme storm events in 2015 and 2016. Notably only the Tax Day Flood on April 18, 2016 compared with the 90th percentile as seen in the figure below. A functional relationship between the median or 50th percentile DARF and area was determined. Watershed hydrologic response varies greatly considering directions and speed commensurate with flow direction. Physics-based distributed modeling of a watershed near to IAH will be used to demonstrate sensitivity of watershed response to storm characteristics including movement.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner