Title Page
Abstract
Preface
Contents
Chapter 1. Introduction 20
1.1. Background 20
1.2. Objectives 25
1.3. Outline of the dissertation 27
Chapter 2. Analysis methods based on the reported meteotsunami events 28
2.1. Intensity of air pressure disturbance - Occurrence of hazardous meteotsunamis coupled with pressure disturbance traveling in the Yellow Sea, Korea 28
Abstract 28
2.1.1. Introduction 29
2.1.2. Methods 32
2.1.3. Results and discussion 33
2.1.4. Conclusions 41
2.2. Meteotsunami arrival time - Determination of accurate arrival time of meteotsunami event in the Yellow Sea 42
Abstract 42
2.2.1. Introduction 43
2.2.2. Methods 45
2.2.3. Results 47
2.2.4. Discussion 53
2.2.5. Conclusions 55
2.3. Meteotsunami propagation - Method for detection of meteotsunami propagation in the Yellow Sea: Reported cases 56
Abstract 56
2.3.1. Introduction 57
2.3.2. Methods 59
2.3.3. Results 62
2.3.4. Discussion 66
2.3.5. Conclusions 67
Chapter 3. Detection of meteotsunami generation 69
3.1. Pilot test during the period of March to April 2018 - Real-time pressure disturbance monitoring system in the Yellow Sea: Pilot test during the period of March to April 2018 69
Abstract 69
Abbreviations 70
3.1.1. Introduction 70
3.1.2. Real-time pressure disturbance monitoring system 75
3.1.3. Results 81
3.1.4. Discussion 105
3.1.5. Summary and conclusions 113
3.2. Progress report since 2019 - Progress report on addressing meteotsunami risk in the eastern Yellow Sea 116
Abstract 116
3.2.1. Introduction 117
3.2.2. Tsunami-like waves of atmospheric origins 121
3.2.3. Progress report on a meteotsunami monitoring system 124
3.2.4. Discussion and conclusions 133
3.3. Towards observation- and atmospheric model-based early warning systems - Towards observation- and atmospheric model-based early warning systems for meteotsunami mitigation: A case study of Korea 137
Abstract 137
3.3.1. Introduction 138
3.3.2. Meteotsunami warning system 143
3.3.3. Results and discussion 151
3.3.4. Conclusions 171
Chapter 4. Meteotsunami propagation and amplification mechanisms 174
4.1. Double resonance effect at a beacon station - Double resonance effect in Daeheuksando port caused by air pressure disturbances in Yellow Sea on March 31, 2007 174
Abstract 174
4.1.1. Introduction 175
4.1.2. Methods 177
4.1.3. Results 179
4.1.4. Discussion 184
4.1.5. Conclusions 186
4.2. Meteotsunami-induced harbor seiches at multiple stations - Propagation and amplification of meteotsunamis in multiple harbors along the eastern Yellow Sea coast 187
Abstract 187
4.2.1. Introduction 188
4.2.2. Data and methods 191
4.2.3. Results 196
4.2.4. Discussion 218
4.2.5. Summary and conclusions 230
4.3. Historical meteotsunami events - Occurrence of pressure-forced meteotsunami events in the eastern Yellow Sea during 2010-2019 232
Abstract 232
4.3.1. Introduction 233
4.3.2. Meteotsunami monitoring system 239
4.3.3. Pressure-forced meteotsunami events 240
4.3.4. Occurrence characteristics of meteotsunamis 253
4.3.5. Discussion and conclusions 265
Chapter 5. Concluding remarks 272
5.1. Major findings and conclusions 272
5.2. Limitations and implications 276
5.3. Future work 278
Appendices 281
Appendix A. Resonant periods and amplification of pre-amplified waves 281
Appendix B. The number of meteotsunami events 284
Appendix C. Other atmospheric disturbances during meteotsunami 285
Appendix D. Sensitivity of meteotsunami to wavelet basis functions 287
Appendix E. Pressure-forced meteotsunami of 4 May 2008 288
Bibliography 289
Publications 324
Chapter 2 19
Table 2.1.1. Maximum magnitude and arrival time of the meteotsunami and pressure disturbance on each date of the accident. 37
Table 2.2.1. Meteotsunami arrival time of each method on March 29, 2007 and March 31, 2007. 51
Table 2.2.2. Meteotsunami propagation derived from each method on March 29 and 31, 2007. 53
Table 2.3.1. Preliminary studies on the criteria for meteotsunami arrival time. 59
Table 2.3.2. Order of arrival time, direction, and speed of the meteotsunami events. 65
Chapter 3 19
Table 3.1.1. Generation coefficient according to the pressure jump events and latitude bands. 103
Table 3.1.2. Spatial characteristics and propagation pattern derived from the pilot test of the pressure jump monitoring system from March to April 2018. 111
Table 3.3.1. Correlation coefficients for low- and high-frequency parts (5 h cut-off period) of observed and predicted air pressure at 17 AWSs in the precaution zone. 156
Table 3.3.2. Intensity of 99% percentile of predicted air pressure jump from 16:00 to 18:00 KST on March 4, 2018 (see Fig. 3.3.8). 163
Chapter 4 19
Table 4.1.1. Details of the grids and the physics options used in WRF model. 178
Table 4.2.1. Details of tide gauges and AWSs presented in Fig. 4.2.1. 195
Table 4.2.2. Geometric features of basins with the tide gauges installed in the harbor and breakwater. 228
Table 4.3.1. Maximum wave height of the 42 pressure-forced meteotsunami events. The known events since 2010 are marked by a superscript. The event dates... 248
Table 4.3.2. Average intensity and occurrence rates for air pressure jumps and meteotsunamis during 11 extreme (widespread) meteotsunami events. The event... 255
Chapter 1 15
Figure 1.1.1. Photographs and a CCTV image captured during and after the meteotsunami events on March 30-31, 2007 and May 4, 2008. 22
Figure 1.1.2. A sketch illustrating the physical mechanism responsible for formation of the catastrophic meteotsunami at Nagasaki Bay (Japan) on 31 March 1979. 23
Figure 1.1.3. Left panel: A sketch illustrating the generation of long wave sea-level oscillations at remote tide gauges after the 1883 Krakatau explosion (Monserrat et al.,... 24
Figure 1.1.4. Examples of collaborative efforts of the KMA to address meteotsunami risk. The images were downloaded from the KMA presentation materials (Eom et al.,... 26
Chapter 2 15
Figure 2.1.1. Maps showing bathymetry and locations of CCTV, AWS (Automatic Weather System), and TS (Tidal Station) used to analyze the meteotsunami case in... 30
Figure 2.1.2. CCTV images which captured (a) the propagation of sudden waves (red rectangle) on 4 May 2008 (12:37:51 KST, M - 20 s), and (b) significant human... 32
Figure 2.1.3. Raw tide level and 2 h high-pass filtered tide level from the tidal stations located near the damaged area during the reported meteotsunamis of (a) 30-31 March... 35
Figure 2.1.4. The rate of sea-level pressure change at KR (1), BR (2), and DH (3) AWS during (a) 30-31 March 2007, (b) 3-5 May 2008, and (c) 25-27 April 2011.... 36
Figure 2.1.5. Rain rate of the composite KMA radar images which represents the maximum reflectivity in each column on (a) 31 March 2007, (b) 4 May 2008, and (c)... 39
Figure 2.2.1. Map of the west coast of Korea showing the positions of the AnHeung (AH), BoRyeong (BR), JangHang (JH), GunSan (GS), WiDo (WD), YeongGwang (YG) and... 43
Figure 2.2.2. Tide level anomaly from north to south. Red ellipses encircle the meteotsunami events on March 29, 2007 and March 31, 2007. 47
Figure 2.2.3. Methods for meteotsunami arrival time with artificial time series (M1-3); (a) Nonstationary time series similar to the tide level on March 31, 2007. (b)... 48
Figure 2.2.4. High-pass filtered (2 h) tide level and Scale Averaged Wavelet Power (SAWP) of the filtered tide level. AH and BR tide level data which had too many... 50
Figure 2.2.5. Meteotsunami propagation of each arrival time method (a1-3, b1-3) derived from propagation algorithm of meteotsunami and radar... 52
Figure 2.3.1. Location of the study area, showing the locations of the tide gauges (red squares) and propagation direction (black arrows) of the reported meteotsunami cases. 58
Figure 2.3.2. Method for detection of meteotsunami propagation. 62
Figure 2.3.3. High-pass filtered sea-level records (2 h) and arrival time (red circles) of the reported meteotsunami cases. (a) 31 March 2007 case, (b) 4 May 2008 case,... 63
Figure 2.3.4. Propagation map and arrival time of the reported meteotsunami cases. (a) 31 March 2007 case, (b) 4 May 2008 case, and (c) 26 April 2011 case. Black... 64
Chapter 3 15
Figure 3.1.1. Location and bathymetry of the Yellow Sea, observation stations for 89 automatic weather stations (AWSs), and 15 tide gauges (TG). Red crosses and green... 74
Figure 3.1.2. Illustration of pressure disturbance (Šepić and Vilibić, 2011) and time constraints of the preliminary caution short message service (SMS) in the real-time... 79
Figure 3.1.3. Real-time pressure jump monitoring process for the preliminary caution SMS (SMS #1) and the propagation warning SMS (SMS #2): (a-d) mean sea level... 82
Figure 3.1.4. Monitoring results of pressure jump propagation for 4-5 March 2018 (KST): (a-b) isochrone map of the pressure jump during each period; red and green... 83
Figure 3.1.5. As in Fig. 3.1.4, but for April 5-8, 2018 (KST). 86
Figure 3.1.6. As in Fig. 3.1.4, but for April 10-11, 2018 (KST). 89
Figure 3.1.7. Time series of the high-frequency sea-level oscillations (2 h high-pass filtered) during the pressure jump events. (a) 4-5 March 2018 (KST); (b) 5-6 April... 92
Figure 3.1.8. Arrival time and propagation map of the meteotsunamis for : (a) March 4-5, 2018 (KST) and (b) April 10-11, 2018 (KST). Black squares mark TG where... 96
Figure 3.1.9. Magnitude of the pressure disturbances and the sea-level oscillations during the pressure jump events according to five latitude bands; A (37-38°N), B... 102
Figure 3.1.10. Scatter diagram between the maximum intensity of the pressure disturbance and the maximum amplitude of the high-frequency sea-level oscillations... 107
Figure 3.2.1. (a) Bathymetric map of the Yellow Sea and area where additional buoy installations are planned (red dashed box). (b) Observation systems comprising 89... 120
Figure 3.2.2. Characteristics of pressure-forced meteotsunamis on 30-31 March 2007. Note that the time series of air pressure and sea level were 2 h high-pass filtered using... 123
Figure 3.2.3. Methods for analyzing air pressure disturbance using (a) a range of pressure change over 60 min (Method #1), and (b) a rate of pressure change (Šepić... 128
Figure 3.2.4. Example of false alarms reported during test operation of the monitoring system based on Method #1 (see Fig. 3.2.3a). During a false detection on 16 April... 129
Figure 3.2.5. Progress report of the meteotsunami monitoring system: propagation of air pressure jumps using Method #2 (see Fig. 3.2.3b) and resultant meteotsunamis on... 131
Figure 3.2.6. Current observation system and plan for observation system expansion. (a) 89 AWSs, ocean buoys with pressure sensors, and marine forecast zone. (b)... 134
Figure 3.3.1. High-resolution atmospheric model and observation system over the Korean Peninsula. (a) Domain and horizontal resolution of the local data assimilation... 143
Figure 3.3.2. (a) Air pressure, air pressure disturbance, and air pressure jump at the DH (15) AWS in the precaution zone during the pressure-forced meteotsunami event... 145
Figure 3.3.3. (a) Early warning system for meteotsunamis and (b) LDAPS predictions (E - 21 h: blue, E - 15 h: green, E - 09 h: red) before the onset of the meteotsunami... 149
Figure 3.3.4. Snapshots of atmospheric disturbances at 12:00, 18:00, and 20:00 KST on March 4, 2018. (a) Predicted air pressure (E - 09 h). The dashed circles enclose... 153
Figure 3.3.5. (a) Time series comparison of observed and predicted air pressure at four AWSs in the precaution zone, where the strongest air pressure jumps were... 154
Figure 3.3.6. Comparison of (a) observed and (b-d) predicted air pressure disturbance during the meteotsunami event at 17 AWSs in the precaution zone (see Fig. 3.3.1b... 157
Figure 3.3.7. Propagation pattern of air pressure jump using (a) LDAPS prediction (E - 09 h) and (b) that observed at 89 AWSs, based on the real-time monitoring protocol.... 158
Figure 3.3.8. Application of LDAPS prediction (E - 09 h) in early warning system from 15:00 to 20:00 KST on March 4, 2018. Graded colors on the map denote the... 162
Figure 3.3.9. (a) Observed air pressure at HT (17) AWS in the precaution zone and (b-d) sensitivity of air pressure disturbance for each output time interval. The gray... 168
Chapter 4 17
Figure 4.1.1. Study area (Domain 3) and satellite image of DP captured from Google Earth. Red cross markers indicate locations of KR and DP. Gray dashed lines are... 177
Figure 4.1.2. Tide level, wavelet scalogram of the tide level, and residual tide level (high-pass filtered, 4 h) and Scale Averaged Wavelet Power (SAWP) of the residual... 180
Figure 4.1.3. Horizontal distribution of the air pressure disturbances at the bottom layer derived from WRF 1 min interval results by using rate of air pressure change... 182
Figure 4.1.4. The evidence of harbor resonance at DP. (a) Demeaned, linearly detrended sea-level pressure (top) and rate of sea-level pressure (bottom) at KR (red)... 183
Figure 4.2.1. Map showing observation locations and bathymetry of the Yellow Sea. The square and circle symbols show 10 tide gauges and 126 automatic weather... 193
Figure 4.2.2. Sea level (black lines) and 2 h high-pass filtered residual sea level (red lines) from the AH, WD, YG, DH, CJ, and MS tide gauges during 25-26 April 2008.... 198
Figure 4.2.3. Wavelet power spectrum (WPS) for the sea level. The y-axis represents the Fourier period (min). The thick gray lines enclose regions of 5% significance level... 199
Figure 4.2.4 Comparison of power spectral density in the residual sea level. The red lines and the black dashed lines indicate background noise of less than 1 h during the... 200
Figure 4.2.5. Air pressure of the AD, AM, MD, YG, AJ, DH, SY, and KS AWSs during days 25-26 within the meteotsunami event: (a) demeaned air pressure, (b) air pressure... 204
Figure 4.2.6. Wind variations at the outer-located AD, MD, DH, and KS AWSs on day 25 within the meteotsunami event. The black and red lines denote along- (north... 206
Figure 4.2.7. Evolution of the air pressure oscillations derived from all AWSs. (a) Power Hovmöller diagram of less than 5 h scale-averaged wavelet power (SAWP) in... 208
Figure 4.2.8. Correlation between air pressure oscillations and resultant sea-level oscillations in the time-frequency domain at the DH AWS and tide gauge. (a) 5 h... 212
Figure 4.2.9. Propagation of air pressure disturbances (rate of pressure changes over 10 min) and meteotsunamis from 18:18 to 00:49 KST on April 25-26, 2008. (a)... 215
Figure 4.2.10. Meteorological images of the atmospheric disturbances from 18:00 to 22:30 KST on April 25, 2008: (a) rain rate on radar images and (b) temporal... 217
Figure 4.2.11. Surface pressure charts illustrated every 3 h from 15:00 to 00:00 KST on April 25-26, 2008. The images were provided by the Korea Meteorological... 221
Figure 4.2.12. Inverted barometer response, maximum meteotsunami amplitude, and generation coefficient (Šepić and Rabinovich, 2014) at each AWS and tide gauge (red... 223
Figure 4.2.13. Google Earth satellite images of the semi-closed basins in which the tide gauges (red squares) are located. The red vector shows the propagation direction... 225
Figure 4.2.14. Spectral ratios (red line), meteotsunami (days 25-27)/background (days 23-25), of the residual sea level (see Fig. 4.2.4). The dominant period bands of... 227
Figure 4.3.1. Observation system including 89 automatic weather stations (AWSs) and 16 tide gauges along the coast of the eastern Yellow Sea, with depth contours (m).... 238
Figure 4.3.2. Characteristics of the pressure-forced meteotsunami from the DaeHeuksando (DH) harbor during the 26 April 2011 meteotsunami event: (a) 36 h... 243
Figure 4.3.3. Process flow diagram showing classification of pressure-forced meteotsunami events. 245
Figure 4.3.4. Percentage of meteotsunami event types. 249
Figure 4.3.5. Temporal pattern of meteotsunami occurrences: (a) number of events and (b) distribution of wave height per month. 250
Figure 4.3.6. Spatial pattern of meteotsunami occurrences: (a) number of events per year and (b) total number of events. Gray lines indicate average occurrences at the... 252
Figure 4.3.7. Latitude-band-averaged intensity of extreme meteotsunami events: (a) air pressure jump and (b) meteotsunami. 256
Figure 4.3.8. Propagation of the air pressure jump on 4 April 2015. (a) Isochrone map. Circles mark multiple AWSs where the air pressure jump arrived from 12:00 to 18:00... 257
Figure 4.3.9. Propagation pattern of air pressure jumps during the classified meteotsunami events (speed, direction, and occurrence rate). Colors of circles and... 262
Figure 4.3.10. (a) Scatterplot of wave period and wave height of the meteotsunami events. The binned distributions of wave height and period are shown. The... 263
Figure 4.3.11. (a) Schematic diagram of harbor meteotsunamis, (b) the amplification process of ocean waves locked to air pressure jumps in the time domain, and (c) the... 270
Figure A.1. Red and blue: modeled sea-level oscillations (<2 h) with the LDAPS air pressure and wind forcing. Black: observations. The dashed lines indicate the arrival... 281
Figure A.2. Modeled resonant periods and amplification of the pre-amplified waves from the open Yellow Sea to the DH harbor using the LDAPS air pressure prediction... 282
Figure A.3. Wavelet power spectrum (WPS) for the observed and modeled sea-level oscillations (<5 h) at the DH tide gauge. The y-axis represents the Fourier period... 283
Figure B.1. The pressure-forced meteotsunami case and exceptional cases in the eastern Yellow Sea during 2010-2019 ( see Chapter 4.3.3 ). 284
Figure C.1. Atmospheric disturbances during the meteotsunami period on April 25-26, 2008 (see Chapter 4.2.3). Evidence of a squall line passage is observed at the... 285
Figure C.2. During the squall line passage, sea-level oscillations accompanied by air pressure and wind oscillations propagated with a similar range of direction and speed... 286
Figure D.1. Sensitivity test of meteotsunami power to wavelet basis functions based on time series of the reported meteotsunami case of 31 March 2007. (a) Demeaned... 287
Figure E.1. Air pressure and air pressure disturbance on May 4, 2008. (a) Air pressure at the HamPyeong (HP) AWS. (b) Calculated air pressure disturbance (see Chapter... 288