[CCoE Notice] MS Thesis Presentation

[Abercrombie, Irene F]

MS Thesis Defense
Predictions of Significant Wave Height in Lake Okeechobee, Florida Using Approaches Related to Simplified Stochastic Procedure and Wave Energy Spectrum
Ismat Khan
Date: Wednesday, November 7th, 2012

Location: N137, Conference Room
Time: 9:00 am

Committee Members:
Dr. Keh-Han Wang
Dr. Jerry R. Rogers
Dr. Kyle Strom
Dr. Abdusselam Altunkaynak
(Associate Professor, Civil Engineering,
Istanbul Technical University, Turkey)


Prediction of significant wave height is critically important to the physical and environmental impact study of coastal, estuarine, or large lake environments. In this study, development of predictive models for the determination of time varying significant wave heights in Lake Okeechobee, Florida using the simplified stochastic procedure and wave energy spectrum method is presented. The stochastic procedure related regression and perceptron least square method (PLSM) are utilized to formulate the Regression Model 1 (RM1), Regression Model 2 (RM2) and PLSM models. Wind speed is a major input for the RM1 and PLSM models.   A new wave spectrum based Modified Pierson-Moskowitz (MPM) Spectrum model is also developed. The model parameters are calibrated using part of data (training data) collected in 2002. All models are also tested and validated by comparisons with selected independent data. The predicted significant wave heights from each model are compared and with the Artificial Neural Network (ANN) predictions obtained by Altunkaynak and Wang (2012). Error analyses are performed to evaluate the performance of each developed model. The comparisons between predicted significant wave heights from each model and observed data indicate that the proposed RM1, RM2, PLSM and MPM are effective models that are acceptable for predicting significant wave height in Lake Okeechobee. Among the four models, RM2 (with significant wave height as an input) and PLSM (with significant wave height and wind speed as inputs) perform better than the RM1 and MPM models. All models can make good predictions at stations LZ40 and L006 near the center of the Lake. However, an underestimation of the significant wave height is observed at the station L005 close to the marshy area.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20121105/ad581874/attachment.html