![]() |
|
以作者查詢圖書館館藏 、以作者查詢臺灣博碩士 、以作者查詢全國書目 、勘誤回報 、線上人數:85 、訪客IP:3.137.187.250
姓名 陳鈞澤(Jiun-Tze Chen) 查詢紙本館藏 畢業系所 資訊工程學系 論文名稱 整合多來源資料與機器學習方法於弱綜觀午後對流即時預報
(Integrating multiple data sources for weak synoptic afternoon thunderstorm nowcasting with machine learning algorithms)檔案 [Endnote RIS 格式]
[Bibtex 格式]
[相關文章]
[文章引用]
[完整記錄]
[館藏目錄]
至系統瀏覽論文 (2028-7-28以後開放)
摘要(中) 在台灣夏季,弱綜觀條件下的主要降雨來源為午後對流事件,午後對流事件有著短延時強降雨的特性,本研究期望整合雷達資料與多來源氣象資料,利用機器學習演算法針對對流肇始及後續發展進行一小時即時預報,我們蒐集了 2017、2018、2019 年弱綜觀午後對流事件(共 33 天)的向日葵八號衛星、GPS 測站可降水量、氣象局地面測站以及QPESUMS 雷達回波,並利用機器學習演算法 Dense Neural Network (DNN)以及Convolutional LSTM Network (ConvLSTM)進行預報。結果顯示在 ConvLSTM 上採用
Focal loss α = 0.8, γ = 2有較好的結果,在2019測試資料集上點對點ETS分數為0.1181、鄉鎮市 ETS0.3569,和 Python 雷達迴波外延方法 Pysteps-OF 使用的光流法相比,ConvLSTM 在對流初始階段以及發展階段都有比較好的結果。本研究亦針對 20190524 以及 20190617 的案例,分析了 ConvLSTM 預報的特性,結果顯示在對流初始及消散期間,ConvLSTM 會有延遲預報的現象,而在對流發展期間則有最好的預報表現,本研究亦探討在雷達資料上加入不同變數資料的影響,結果顯示加入測站資料以及衛星資料,在對流初始的預報上都有更好的結果,顯示多元的氣象資料對於午後雷暴事件的預報是十分重要的。摘要(英) In the summer of Taiwan,, the primary source of rainfall under weak synoptic conditions stems from Afternoon Thunderstorms (ATs) convective events characterized by short-duration heavy rainfall. This study aims to integrate radar data with various meteorological data to employ machine learning algorithms for real-time one-hour nowcasting of ATs convection. We collected data from the Himawari-8 satellite, precipitable water vapor from GPS stations, CWB surface stations, and QPESUMS radar echoes for weak synoptic ATs events (a total of 33 days) in 2017, 2018, and 2019. We utilized machine learning algorithms Dense Neural Network (DNN) and Convolutional LSTM Network (ConvLSTM) for forecasting. The results demonstrate superior performance when applying Focal loss with α=0.8, γ=2 in ConvLSTM.With a point-to-point ETS score of 0.1181 and a district ETS of 0.3569 on the 2019 test dataset,ConvLSTM outperforms the optical flow method used in Pysteps-OF, especially in the initial and development stages of convection. This study also examines the characteristics of
ConvLSTM forecasts using case studies from May 24 and June 17, 2019. It was found that ConvLSTM tends to exhibit delayed forecasting during convective initiation and dissipation, yet it performs best during the development phase. Furthermore, the study investigates the effects of combining different variables with radar data. The results show that incorporating data from weather stations and satellite data significantly improves the initial forecast of convection. This highlights the importance of diverse meteorological data for forecasting ATs events.關鍵字(中) ★ 午後對流
★ 機器學習
★ 卷積長短時記憶網路關鍵字(英) ★ Afternoon Thunderstorms
★ Machine Learning
★ ConvLSTM論文目次 中文摘要.................................................................................................................................................. i
Abstract ................................................................................................................................................... ii
誌謝........................................................................................................................................................ iii
Table of Contents................................................................................................................................... iv
List of Figures........................................................................................................................................ vi
List of Tables.......................................................................................................................................... ix
Chapter 1 Introduction ............................................................................................................................ 1
1-1 Background................................................................................................................................ 1
1-2 Related Works............................................................................................................................ 2
1-3 Motivation and Goals................................................................................................................. 3
Chapter 2 Materials and Methods ........................................................................................................... 5
2-1 Data Sources and Preprocessing ................................................................................................ 5
2-1-1 Column Maximum Radar Reflectivity ............................................................................... 6
2-1-2 Satellite ............................................................................................................................... 8
2-1-3 CWB Station Observation ................................................................................................ 10
2-1-4 GPS Precipitable Water Vapor......................................................................................... 15
2-2 Methods ................................................................................................................................... 17
2-2-1 Dense Neural Network(DNN).......................................................................................... 17
2-2-2 Convolutional LSTM Network(ConvLSTM)................................................................... 18
2-3. Evaluation Metrics..................................................................................................................... 19
Chapter 3 Model Training and Optimization ........................................................................................ 21
3-1 Model Training Strategy .......................................................................................................... 21
3-1-1 DNN Training Strategy..................................................................................................... 21
3-2-1 ConvLSTM Training Strategy.......................................................................................... 21
3-2. Data Imbalance Problem............................................................................................................ 22
3-2-1 DNN Data-level Method .................................................................................................. 23
v
3-2-2 DNN Algorithm-level Method ......................................................................................... 25
3-2-3 ConvLSTM Data-level Method........................................................................................ 27
3-2-4 ConvLSTM Algorithm-level Method............................................................................... 29
3-3 Feature Sets & Model Structure............................................................................................... 31
3-4 Thresholds................................................................................................................................ 34
Chapter 4 Results .................................................................................................................................. 35
4-1 Evaluations on Validation Dataset (Grid Result)..................................................................... 35
4-2 Evaluations on Validation Dataset (District Results) .............................................................. 36
4-3 20190524 Case Analysis.......................................................................................................... 37
4-4 20190617 Case Analysis.......................................................................................................... 39
4-5 CI Times vs Non-CI Times...................................................................................................... 41
Chapter 5 Discussions and Conclusions................................................................................................ 43
5-1 Discussions ............................................................................................................................ 43
5-1-1 Predictability and Features Analysis ................................................................................ 43
5-1-2 Feature Combinations Analysis........................................................................................ 45
5-2 Conclusions.............................................................................................................................. 46
5-3 Future Works ........................................................................................................................... 47
References............................................................................................................................................. 48
Appendices............................................................................................................................................ 50
A-1 20190524 Convection Result Figures................................................................................... 50
A-2 20190617 Convection Result Figures................................................................................... 57參考文獻 1. Jou, B., et al., Observational characteristics and forecast challenge of Taipei flash flood
afternoon thunderstorm: Case study of 14 June 2015. Atmos. Sci, 2016. 44(1): p. 57-
82.
2. Chen, J.-T., 利用 IBM_VDRAS 分析 TASSE 實驗期間劇烈午後雷暴降水事件:
2019 年 7 月 22 日個案分析. 2021, National Central University.
3. Chen, T.-C., S.-Y. Wang, and M.-C. Yen, Enhancement of afternoon thunderstorm
activity by urbanization in a valley: Taipei. Journal of Applied Meteorology and
Climatology, 2007. 46(9): p. 1324-1340.
4. Lin, P.-F., et al., Warm season afternoon thunderstorm characteristics under weak
synoptic-scale forcing over Taiwan Island. Weather and forecasting, 2011. 26(1): p. 44-
60.
5. Chen, T.-C., et al., Impact of afternoon thunderstorms on the land–sea breeze in the
Taipei basin during summer: An experiment. Journal of Applied Meteorology and
Climatology, 2014. 53(7): p. 1714-1738.
6. Ashesh, A., et al., Accurate and Clear Quantitative Precipitation Nowcasting Based on
a Deep Learning Model with Consecutive Attention and Rain-Map Discrimination.
Artificial Intelligence for the Earth Systems, 2022. 1(3): p. e210005.
7. Chen, I.-H., et al., Improving afternoon thunderstorm prediction over Taiwan through
3DVar-based radar and surface data assimilation. Weather and Forecasting, 2020.
35(6): p. 2603-2620.
8. Chang, H.-L., et al., Nowcast guidance of afternoon convection initiation for Taiwan.
Weather and Forecasting, 2017. 32(5): p. 1801-1817.
9. Pulkkinen, S., et al., Pysteps: an open-source Python library for probabilistic
precipitation nowcasting (v1. 0). Geoscientific Model Development, 2019. 12(10): p.
4185-4219.
10. Shi, X., et al., Convolutional LSTM network: A machine learning approach for
precipitation nowcasting. Advances in neural information processing systems, 2015. 28.
11. Shi, X., et al., Deep learning for precipitation nowcasting: A benchmark and a new
model. Advances in neural information processing systems, 2017. 30.
12. Sønderby, C.K., et al., Metnet: A neural weather model for precipitation forecasting.
arXiv preprint arXiv:2003.12140, 2020.
13. Germann, U. and I. Zawadzki, Scale-dependence of the predictability of precipitation
from continental radar images. Part I: Description of the methodology. Monthly
Weather Review, 2002. 130(12): p. 2859-2873.
14. Chang, P.-L., et al., An operational multi-radar multi-sensor QPE system in Taiwan.
Bulletin of the American Meteorological Society, 2021. 102(3): p. E555-E577.
15. Bessho, K., et al., An introduction to Himawari-8/9—Japan’s new-generation
geostationary meteorological satellites. Journal of the Meteorological Society of Japan.
Ser. II, 2016. 94(2): p. 151-183.
16. Rumelhart, D.E., G.E. Hinton, and R.J. Williams, Learning internal representations by
error propagation. 1985, California Univ San Diego La Jolla Inst for Cognitive Science.
17. Ioffe, S. and C. Szegedy. Batch normalization: Accelerating deep network training by
reducing internal covariate shift. in International conference on machine learning. 2015.
pmlr.
18. Nair, V. and G.E. Hinton. Rectified linear units improve restricted boltzmann machines.
in Proceedings of the 27th international conference on machine learning (ICML-10).
49
2010.
19. Shi, C., et al., Learning multiscale temporal–spatial–spectral features via a multipath
convolutional LSTM neural network for change detection with hyperspectral images.
IEEE Transactions on Geoscience and Remote Sensing, 2022. 60: p. 1-16.
20. Schaefer, J.T., The critical success index as an indicator of warning skill. Weather and
forecasting, 1990. 5(4): p. 570-575.
21. Kingma, D.P. and J. Ba, Adam: A method for stochastic optimization. arXiv preprint
arXiv:1412.6980, 2014.
22. Kotsiantis, S., D. Kanellopoulos, and P. Pintelas, Handling imbalanced datasets: A
review. GESTS international transactions on computer science and engineering, 2006.
30(1): p. 25-36.
23. Lemaître, G., F. Nogueira, and C.K. Aridas, Imbalanced-learn: A python toolbox to
tackle the curse of imbalanced datasets in machine learning. The Journal of Machine
Learning Research, 2017. 18(1): p. 559-563.
24. Pedregosa, F., et al., Scikit-learn: Machine learning in Python. the Journal of machine
Learning research, 2011. 12: p. 2825-2830.
25. Lin, T.-Y., et al. Focal loss for dense object detection. in Proceedings of the IEEE
international conference on computer vision. 2017.指導教授 洪炯宗 廖宇慶(Jorng-Tzong Horng Yu-Chieng Liou) 審核日期 2023-7-27 推文 plurk
funp
live
udn
HD
myshare
netvibes
friend
youpush
delicious
baidu
網路書籤 Google bookmarks
del.icio.us
hemidemi
myshare