Docstoc

Atlantic Basin tropical cyclone storm track analysis

Document Sample
Atlantic Basin tropical cyclone storm track analysis Powered By Docstoc
					Atlantic Basin Tropical Cyclone
     Storm Track Analysis



        Desmond Carroll
      Advisor: Frank Hardisty
                     Goals

• Analyze Tropical Cyclone susceptibility in the
  Caribbean.
• Enhance seasonal forecasts by examining
  correlation between basin wide activity and
  small island risk.
• Develop a web based visualization of the
  results.
Background
N. Atlantic Tropical Cyclone Season
  Not exclusively an American problem.
       Cuba                          Cayman




                  The Bahamas
                       The North Atlantic Basin




(Google Earth, 2010)
           The NOAA HURDAT Dataset




(NOAA, 2010)
The Extended Best Track Dataset
     Synthetic Track Generation
• Augment the existing data.
• Analyze the extreme cases.
• Compare probabilistic versus deterministic.
                  Problem
• Small densely populated islands.
• A line track may be too general.
• Neither the HURDAT nor EBT alone sufficient.
          Research Questions
• Will the inclusion of storm size in track
  modeling problems improve the quality of the
  results in the caribbean?
• Is there any link between the total number of
  named storms in the Atlantic basin and the
  susceptibility of individual islands in The
  Bahamas?
Objectives
               Buffer Analysis
• Generate a method
  for creating buffers
  from EBT data




                                 (ESRI, 1997)
Modeling Asymmetry
                 • Get closer to
                   simulating the real
                   variability in storm
                   paths.




  (NASA, 2006)
              Track Analysis
Augment the
HURDAT with the
data similar to EBT




                               (Rogers and Spirnak, 2006)
                Visualization
• Provide a public outlet.
• Place the focus on the less studied islands.
• Explore HTML5’s Canvas element.
Methods
                  Libraries
• Geospatial Libraries
  – Open Source solutions
  – Solve well known problems
  – Python bindings
                 Buffers

GDAL
  – Geospatial Data Abstraction Library
  – Data manipulation
  – Projections
Buffers




          • Shapely
            – PostGIS Inspired
            – Convex Hull
            – Cascading Union
              Algorithms
Initial Run
     Markov Chain Monte Carlo
• Unknown processes.
• Probability functions.
• Many applications.
                   MCMC Random Walk




                                      (Patil et al., 2010)
(Patil et al., 2010)
        Application to HURDAT
• Back test model for validity.
• Augment storm tracks with asymmetry.
• Extreme events can be examined in more.
  detail.
  Current Web Visualization Technologies

• Javascript client
  – HTML5 Canvas Element
• Openlayers framework
• Geoserver middleware
• PostGRESQL/PostGIS backend
                       Site Layout
http://loggedout.org
                 Future Work
•   Conduct analysis for various storm strengths.
•   Investigate temporal effects.
•   Couple with storm track model.
•   Employ more deterministic factors.
Questions?
                               References
Demuth, J., M. DeMaria, and J.A. Knaff, 2006: Improvement of advanced microwave
  sounder unit tropical cyclone intensity and size estimation algorithms. J. Appl.
  Meteor., 45, 1573-1581.
Emanuel et al. A statistical deterministic approach to hurricane risk assessment.
  Bulletin of the American Meteorological Society (2006) vol. 87 (3) pp. 299-314

Hall and Jewson. Statistical modelling of North Atlantic tropical cyclone tracks. Tellus A
    (2007) vol. 59 (4) pp. 486-498

Jiechen Wang et al. Review of Buffer Generation Algorithm Studies. Intelligent
    Information Technology Application, 2008. IITA '08. Second International
    Symposium on (2008) vol. 2 pp. 911 – 917
National Oceanic and Atmospheric Administration. (2005, December 9). National
    Hurricane Center. Retrieved May 25, 2010, from National Hurrican Center:
    http://www.nhc.noaa.gov/
                           References
Patil, A., D. Huard and C.J. Fonnesbeck. 2010. PyMC: Bayesian Stochastic
   Modelling in Python. Journal of Statistical Software, 35(4), pp. 1-81.


Rumpf et al. Stochastic modelling of tropical cyclone tracks. Mathematical
  Methods of Operations Research (2007) vol. 66 (3) pp. 475-490

Rumpf et al. Tropical cyclone hazard assessment using model-based track
  simulation. Natural hazards (2009) vol. 48 (3) pp. 383-398

Wang Jiechen et al. A Novel Method of Buffer Generation Based on Vector
  Boundary Tracing. Information Technology and Applications, 2009. IFITA
  '09. International Forum on (2009) vol. 1 pp. 579 - 582

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:1
posted:1/22/2013
language:Latin
pages:29