An Informed System Development Approach to Tropical Cyclone Track and Intensity Forecasting.

• Superior document: Linköping Studies in Science and Technology. Dissertations Series ; v.1734
• Author: Roy, Chandan.
• Place / Publishing House: Linköping : : Linkopings Universitet,, 2015.; {copy}2016.
• Year of Publication: 2015
• Edition: 1st ed.
• Language: English
• Series: Linköping Studies in Science and Technology. Dissertations Series
• Physical Description: 1 online resource (166 pages)

Table of Contents

Intro -- Abstract -- Populärvetenskaplig sammanfattning -- Acknowledgements -- Contents -- Chapter 1 Introduction -- 1.1 Problem description -- 1.2 Research objective -- 1.3 Studies conducted in this research -- 1.4 Contributions -- 1.5 Organization of the thesis -- Chapter 2 Background -- 2.1 TCs in the Bay of Bengal -- 2.2 Factors governing TC track and intensity development -- 2.3 Technical processes involved in TC forecasting and warning -- 2.3.1 Data collection -- 2.3.2 TC track forecasting -- 2.3.3 TC Intensity forecasting -- 2.3.4 Warning message formulation and dissemination -- 2.4 Community response to warning -- 2.4.1 Incorporation of human perception into cyclone warning -- 2.5 Cyclone early warning system in Bangladesh -- 2.6 TC track and intensity forecasting using ANN -- 2.7 Biologically based ANN techniques for image processing -- 2.7.1 Neocognitron -- 2.7.2 Convolutional Neural Networks -- 2.7.3 Saliency based visual attention models -- 2.7.4 ANNs used in this research -- 2.8 Exploratory study on TC movement direction prediction -- 2.8.1 Data and method -- 2.8.2 Training and testing of the network -- Chapter 3 Methodological considerations -- 3.1 Central and peripheral parts of this research -- 3.2 Technological paradigm of this research -- Chapter 4 Method -- 4.1 Study area -- 4.2 Eliciting providers' and receivers' views -- 4.2.1 Interview among the meteorologists -- 4.2.2 Questionnaire survey among the residents in the coastal areas -- 4.3 TC track and intensity forecasting technique development -- 4.3.1 Efficiency of the simulation tool with respect to 2D image processing -- 4.3.2 Testing assumptions -- 4.3.3 Match between network-generated and expected outputs -- 4.3.4 Datasets used for training and testing -- 4.3.5 Network structure -- 4.3.6 Information processing in the network -- 4.3.7 Training of the networks.

4.3.8 Training performance -- 4.3.9 Activation-based receptive field analysis -- 4.3.10 Testing of the networks -- Chapter 5 Results -- 5.1 Results elicited from warning providers -- 5.1.1 TC forecasting -- 5.1.2 Warning message formulation and dissemination -- 5.1.3 Limitations and future development plans -- 5.2 Results elicited from warning receivers -- 5.2.1 Warning message reception -- 5.2.2 Warning message interpretation -- 5.2.3 Response to evacuation orders -- 5.2.4 Key reasons for non-evacuation -- 5.2.5 Satisfaction with the warnings and suggestion for improvement -- 5.3 Results obtained during systematic parameter testing -- 5.4 TC intensity forecasting performance -- 5.4.1 Pattern of intensity forecasting error -- 5.5 Combined TC track and intensity forecasting performance -- Chapter 6 Discussion -- 6.1 Problems associated with cyclone early warning in Bangladesh -- 6.2 TC track and intensity forecasting using biologically based ANNs -- 6.2.1 TC Intensity forecasting -- 6.2.2 Graphically-presented TC track and intensity forecasting-ongoing work -- 6.2.3 Prediction performance improvement -- Chapter 7 Conclusions and future work -- Chapter 8 References -- Chapter 9 Articles -- Appendix A Cyclone track forecasting using biologically based ANNs -- Appendix B TC signaling system for the maritime ports -- Appendix C TC signaling system for the river ports -- Appendix D Questions used for the in-depth interview -- Appendix E Questionnaire used for the survey -- Appendix F Per-recorded warning messages -- Appendix G Generalized TC signaling system.