|
Activity Number:
|
313
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #304304 |
|
Title:
|
Comparison of Methods to Understand the Link Between Climate Variability and Hurricane Counts
|
|
Author(s):
|
Roshanak Nateghi*+ and Seth D. Guikema and Steven Quiring
|
|
Companies:
|
Johns Hopkins University and Johns Hopkins University and Texas A&M University
|
|
Address:
|
3400 North Charles, Balitmore, MD, 21218,
|
|
Keywords:
|
count regression ; hurricane activity ; data mining ; climate variability
|
|
Abstract:
|
This paper compares different methods for estimating the relationship between climate variables and hurricane activity in the U.S. These methods include data mining and count regression analysis. The analysis is done to better understand the factors that drive the frequency of hurricanes that make landfall in the U.S. Data mining techniques are first employed to assess the impacts of eighteen climate variables (with their quarterly and annual averages) on hurricane counts. The variables that data mining suggests are most important are then used in regression to determine the extent of their contribution to the hurricane activity. This approach is compared with more standard variable selection methods.
|
