By Topic

Study of Hurricanes and Typhoons from TRMM Precipitation Radar Observations: Self Organizing Map (SOM) Neural Network

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
K. Direk ; Colorado State Univ., Fort Collins, CO ; V. Chandrasekar

Precipitation radar (PR) on Tropical Rainfall Measuring Mission (TRMM) satellite provides high resolution vertical profile of reflectivity (VPR) of tropical storms. Three- dimensional downward-looking observations of tropical storms are very useful to study Hurricanes and Typhoons. The increased reflectivity measured in bright band (BB) region can lead to rainfall overestimate. It is also known that VPR of BB holds extensive information on the types of precipitation and their variability. Better knowledge of VPR of storms is important to understand cloud dynamics and microphysical processes, and to improve satellite retrieval algorithm. Because of a large number of VPR observation, it is of interesting to classify the VPR into characteristic profiles so that it can be useful in studying and comparing different vertical reflectivity profiles. In this study, Self Organizing Map (SOM) Neural Network is used as a method to study and classify VPR of Hurricanes and Typhoons. SOM is unsupervised neural network. It forms a non-linear mapping of the data to a two-dimensional map grid that can be used as an exploratory data analysis tool for generating hypotheses on the relationships of VPR. Similarity relationships within the VPR data and its vertical structure can be visualized and interpreted. Preparation of vertical profile of reflectivity used as input vectors of SOM algorithm is one of the most vital steps. In total eleven Hurricanes and forty Typhoons are studied. VPR of Hurricanes and Typhoons are classified into characteristic profiles. The result of classification shows a distribution that indicates location of each characteristic profile within a storm when viewed from the PR. Percentages of contribution of each characteristic profile to Hurricanes and Typhoons can also be determined. By using SOM, VPR can be classified into various numbers of classes up to one hundred. In this study, VPR is classified into four classes. Two simple operations were performed. Firstl- - y, SOM was applied to all VPR data regardless of rain type. Secondly, stratiform and convective portion of VPR was applied to SOM separately. For stratiform portion of Hurricanes and Typhoons, the bright band (BB) properties including the height of BB peak, BB thickness, reflectivity of BB peak and BB sharpness index of Hurricanes and Typhoons are investigated and compared to those of generic oceanic storm. Comparison of those BB properties of Hurricanes, Typhoons and generic oceanic storm reveals similarities and differences among them.

Published in:

2006 IEEE International Symposium on Geoscience and Remote Sensing

Date of Conference:

July 31 2006-Aug. 4 2006