Climatological Characteristics and Predictability of Tropical Cyclone Tracks in the Western North Pacific
- Alternative Title
- 북서태평양에서 태풍 진로의 기후학적 특성과 예측가능성에 관한 연구
- Abstract
- The characteristics of interannual and interdecadal variations on the track of tropical cyclones (TCs) that have landed in Korea for the past 54 years have been investigated. In a study on climatological characteristics of Korean Peninsula (KP) landfalling TC, the landfalling frequency of strong TC sharply increased since the late 1980s. Furthermore, the full track, recurving and landfalling locations of KP landfalling TC recently moved towards the east. This change was correlated with the recent shift of Western North Pacific High (WNPH) towards the east.
Using Change-point analysis (CPA), the phase of KP landfalling TC frequency had been divided into two phases with high landing frequency (1951-1965, 1986-2004) and one phase with low landing frequency (1966-1985). In terms of the characteristics of KP landfalling TC activity in the three phases, the landfalling position in Korea was shifted eastward from the west coast to the south coast in KP like the result of climatological characteristics. The full track also moved towards the east and thus TCs in the past had often passed through the mainland China. However, more TCs tended to hit Korea directly, their intensity became stronger.
Using fuzzy clustering method (FCM), the KP landfalling TC track could also be divided into four groups TCs of two groups that come up from the south landed in the south coast and the west coast and TCs of two groups that come from the west landed in the southern and central areas of the west coast. The former two groups included TCs since 1980s while the latter two groups mostly covered TCs before 1980s. The TCs from the former two groups were very strong when they hit Korea because most of them did not pass through the mainland China. Furthermore, the increase in the TC intensity was correlated with the increase of seas surface temperature (SST) in the western North Pacific (WNP) since 1980s.
This paper has also investigated the relationship between WNP TC track and climate variation. Even though TC was often observed in WNP during a dry season in summer or in the year with poor rainfall, it rarely landed in Korea. Because WNPH was not expanded to the mid-latitude area of East Asia during a dry season in summer or in the year with poor rainfall in Korea, after all, the steering flow towards Korea was weak. In a year with positive Arctic Oscillation (AO) and negative Tibetan Plateau snow cover (TPSC), on the contrary, WNPH expanded towards the west. Because TCs often landed in China first, they already weakened when they hit Korea.
To predict the frequency of TC which affects Korea during summer, multiple linear regression model (MLRM) has been developed. MLRM is based on a concept that if upper-troposphere Tibetan anticyclone is intensified in spring, the frequency of TCs which have an impact on Korea in summer increases. Because the intensity of upper-troposphere Tibetan anticyclone in spring is greatly influenced by the TPSC in winter and spring, the frequency of TC that affects Korea in summer was adjusted by the status of TPSC. In addition, this paper has further expanded the scope of analysis and developed MLRM to predict the frequency of TC that affect East Asia in summer. The independent variables used in this MLRM referred to 850-hPa geopotential height in spring near the Philippine Sea and Bering Sea. A concept that if north-high and south-low anomalous atmospheric pressure pattern become strong in WNP since spring, the frequency of TC that affects East Asia in summer would increase was used. Furthermore, this paper has developed MLRM that predicts summer WNP TC genesis frequency (TCGF) using teleconnection patterns in spring. The teleconnection patterns in this model are Siberian High Oscillation (SHO), North Pacific Oscillation (NPO) and Antarctic Oscillation (AAO). The prediction accuracy of these three statistical models has been verified through cross validation.
This paper has analyzed the predicted data from 15 general circulation models (GCMs) through multi-model ensemble (MME) to predict the future summer TCGF and TC track patterns. Under A1B scenario, it’s forecasted that the summer WNP TCGF would increase by 3 or 4 in a century (2071-2100) based on simple linear regression model (SLRM) and MLRM. However, it is predicted that the TCGF towards the mid-latitude area of East Asia would decrease because of weak WNPH. According to an analysis of future prediction in consideration of diverse variables (including SST) that have an impact on TC intensity, it appears that TC intensity would get stronger in the future.
- Issued Date
- 2009
- Awarded Date
- 2009. 8
- Type
- Dissertation
- Publisher
- 부경대학교 대학원
- Alternative Author(s)
- 최기선
- Affiliation
- 부경대학교 대학원
- Department
- 대학원 환경대기과학과
- Advisor
- 변희령
- Table Of Contents
- General introduction = 1
Data, methods, and definitions = 3
1. TC data = 3
2. Reanalysis data = 3
3. Observation data = 5
4. Methods and definitions = 5
Characteristics = 9
CHAPTER 1 Track of TC landfalling Korea = 9
1. Climatology = 9
1.1 Introduction = 9
1.2 Changes in the KP landfalling TCs = 11
1.2.1 Landfalling frequency = 11
1.2.2 Landfalling track = 13
1.2.3 Recurving location = 14
1.2.4 Shift of western North Pacific high = 15
1.2.5 Intensity change = 19
1.3 Change in large-scale conditions in the WNP = 19
1.3.1 Total precipitable water = 19
1.3.2 Geopotential height and zonal wind = 21
1.3.3 Sea surface temperature = 22
1.4 Summary and discussion = 24
2. Decadal variation = 27
2.1 Introduction = 27
2.2 Climatology of a KP landfalling TC = 30
2.3 Interdecadal variations related to the KP landfalling TC = 32
2.3.1 Division of regime shifts into TC frequency = 32
2.3.2 Variations in TC intensity, track, and genesis = 36
2.3.3 Variations in atmospheric circulation = 42
2.4 Concluding remarks = 45
3. Cluster analysis = 47
3.1 Introduction = 47
3.2 Fuzzy Clustering Method (FCM) = 49
3.3 Classification of landfalling tracks = 50
3.4 Characteristics of each track pattern = 51
3.4.1 Full track = 52
3.4.2 TC intensity change between before landfall and after landfall = 53
3.4.3 Long-term variability of TC landfalling frequency = 56
3.4.4 Atmospheric circulations = 58
3.5 Summary and conclusions = 58
CHAPTER 2 Relationships between climate variation and TC track = 61
4. Korean summer monsoon rainfall = 61
4.1 Introduction = 61
4.2 Interdecadal variation of Korean summer rainfall for 50 years = 64
4.3 Interdecadal variation of Korean summer rainfall for 35 years = 65
4.3.1 EOF analysis = 65
4.3.2 Large-scale environments = 71
4.3.3 TC activity = 75
4.4 Summary and conclusions = 77
5. Tibetan Plateau snow cover = 78
5.1 Introduction = 78
5.2 TC activity and Tibetan Plateau snow cover (TPSC) = 79
5.3 Large-scale circulation pattern changed by TPSC = 84
5.4 Summary = 85
6. Summer Arctic Oscillation = 86
6.1 Introduction = 86
6.2 Summer AO Index = 88
6.2.1 EOF analysis = 88
6.2.2 Definition of high-AO years and low-AO years = 89
6.3 TC activity corresponding to summer AO variation = 92
6.3.1 TC genesis frequency (TCGF) = 92
6.3.2 TC passage frequency (TCPF) = 93
6.3.3 Large-scale environment = 94
6.3.4 TC recurvature = 99
6.3.5 TC intensity = 100
6.4 Concluding remarks = 103
Predictability = 107
CHAPTER 3 Prediction of TC track using multiple linear regression model = 107
7. Track of TC affecting Korea = 107
7.1 Introduction = 107
7.2 Multiple linear regression model (MLRM) = 109
7.2.1 Definition of potential predictors by stepwise multiple regression = 109
7.2.2 Development of MLRM = 114
7.3 Analyses of validity of MLRM = 116
7.3.1 Cross validation = 116
7.3.2 Difference between the predicted positive and negative years = 117
7.4 Summary and conclusions = 122
8. Track of TC affecting East Asia = 125
8.1 Introduction = 125
8.2 Correlation between summer TC frequency and preceding spring large-scale circulation = 127
8.3 Development of MLRM = 129
8.4 Difference between the predicted positive and negative TC frequency years = 131
8.4.1 TC frequency = 131
8.4.2 Large-scale environments = 132
8.4.3 TC track pattern = 139
8.5 Summary and conclusions = 140
9. Relationship with TC genesis = 142
9.1 Introduction = 142
9.2 Development and validity of MLRM = 144
9.2.1 Development of MLRM = 144
9.2.2 Validity of MLRM using the cross validation = 150
9.3 Difference between the hindcasted positive and negative TCGF years = 150
9.3.1 TCGF = 152
9.3.2 Large-scale environments = 154
9.3.3 TC Tracks = 163
9.4 Concluding remarks = 165
CHAPTER 4 Prediction of future TC track = 168
10. Prediction of future TC track using multi-model ensemble (MME) = 168
10.1 Introduction = 168
10.2 Data and methods = 169
10.3 Future environments for TC genesis in the western Pacific = 173
10.4 Future environments for TC intensity and movement = 178
10.5 Prediction of TCGF using linear regression model = 181
10.6 Summary and conclusions = 182
General conclusions = 185
REFERENCES = 189
APPENDIX Directory of Acronyms = 209
Acknowledgments = 212
- Degree
- Doctor
- Files in This Item:
-
-
Download
Climatological Characteristics and Predictability of Tropical Cyclone Tracks in the Western North Pa.pdf
기타 데이터 / 20.07 MB / Adobe PDF
-
Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.