跳到主要內容

臺灣博碩士論文加值系統

(44.200.194.255) 您好!臺灣時間:2024/07/18 14:26
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

: 
twitterline
研究生:游禹唐
研究生(外文):YOU, YU-TANG
論文名稱:類神經網路應用於格陵蘭冰蓋質量平衡之多步預測
論文名稱(外文):Application of Neural Networks for Multi-step Forecasting of Greenland Ice Sheet Mass Balance
指導教授:陳詩雯陳詩雯引用關係
指導教授(外文):CHEN, SHIH-WEN
口試委員:白敦文林寬仁陳詩雯
口試委員(外文):PAI, TUN-WENLIN, KUAN-JENCHEN, SHIH-WEN
口試日期:2023-07-19
學位類別:碩士
校院名稱:國立臺北科技大學
系所名稱:自動化科技研究所
學門:工程學門
學類:機械工程學類
論文種類:學術論文
論文出版年:2023
畢業學年度:111
語文別:英文
論文頁數:36
中文關鍵詞:類神經網路格陵蘭冰蓋質量平衡多步數預測
外文關鍵詞:Greenland Ice SheetMulti-step ForecastNeural NetworkMass Balance
相關次數:
  • 被引用被引用:0
  • 點閱點閱:95
  • 評分評分:
  • 下載下載:16
  • 收藏至我的研究室書目清單書目收藏:0
近年來,格陵蘭冰蓋(Greenland Ice Sheet)正在加速融化,對氣候、海洋以及生物多樣性造成影響。因此,準確預測格陵蘭冰蓋質量平衡的變化,有利於風險管理,以應對全球氣候異常的挑戰。然而,考慮到氣候與冰蓋質量平衡之間存在著非線性關係以及時間序列的特性,本研究提出應用神經網路技術及使用多變量多步數時間序列方法於格陵蘭質量平衡預測。同時比較四種神經網路模型性能,分別是長短期記憶神經網路(LSTM)、循環神經網路(RNN) 、門控循環單元(GRU)以及類神經網路 (ANN)。預測模型使用溫室氣體濃度以及相關氣象因子數據,數據取自2009年4月到2021年12月。本研究使用皮爾森相關係數進行特徵選擇,得到7個高相關的特徵,包括全球二氧化碳濃度、海平面變化、甲烷、六氟化硫、格陵蘭地區二氧化碳濃度、一氧化二氮及海洋表面溫度異常值。實驗結果顯示,在步數一至步數六的冰蓋總質量平衡預測下,LSTM模型提供最可靠和準確的結果。LSTM模型預測的誤差指標Mean Absolute Percentage Error (MAPE)介於2.05%到2.65%之間。
The accelerated melting of the Greenland Ice Sheet (GrIS) has heightened impacts on global systems. Therefore, accurately predicting the total mass balance of the GrIS can aid in risk management. However, due to the nonlinear relationship between mass balance and climate, as well as the characteristics of time series. This study proposes a neural network methodology for forecasting the total mass balance (TMB) of the GrIS using a multivariable multi-step ahead approach. We compare the performance of four different neural network models: Artificial Neural Network (ANN), Long Short-Term Memory (LSTM) Neural Network, Gated Recurrent Unit (GRU), and Recurrent Neural Network (Simple RNN). These models are driven by greenhouse gas and other relevant meteorological factors collected from April 2009 to December 2021. Feature selection is performed using the Pearson correlation coefficient, which identified 7 highly correlated features: XCO2, Sea level, CH4, SF6, globalXCO2, N2O, and Global SST anomalies. The results demonstrate that the LSTM model outperforms the other model in terms of reliability and accuracy between 1-step to 6-step prediction. The Mean Absolute Percentage Error (MAPE) of the LSTM model’s predictions ranges from 2.05 % to 2.65%.
摘要 i
ABSTRACT ii
致謝 iv
Table of Contents v
List of Tables vii
List of Figures viii
Chapter 1 Introduction 1
Chapter 2 Data Sources and Imputation of missing values 3
2.1 GRACE/GRACE-FO 4
2.2 OCO-2 and GOSAT TANSO-FTS 5
2.3 NOAA GML 6
2.4 CRUTEM5 and HadSST4 7
2.5 NASA’s Goddard Space Flight Center 7
Chapter 3 Methodology 10
3.1 Pearson Correlation Coefficient 10
3.2 Normalization 10
3.3 Sliding Window 11
3.4 Neural Network 12
3.4.1 Artificial Neural Network 12
3.4.2 Recurrent Neural Network 13
3.4.3 Long Short-term Memory Neural Network 13
3.4.4 Gated Recurrent Unit 15
3.5 Performance Metrics 16
Chapter 4 Model Results and Comparison 17
4.1 Feature Selection 17
4.2 Feature Trend Analysis 18
4.3 Model Building 24
4.4 Model Results and Comparison 26
Chapter 5 Conclusion and Future Work 33
References 34
[1]I. Velicogna, "Increasing rates of ice mass loss from the Greenland and Antarctic ice sheets revealed by GRACE," (in Undetermined), Geophysical Research Letters, vol. 36, no. 19, pp. 0094-8276, 2009. [Online]. Available: https://escholarship.org/uc/item/6k84k0xp.
[2]A. V. Julián et al., "Synergistic impacts of global warming and thermohaline circulation collapse on amphibians," Communications Biology, vol. 4, no. 1, pp. 1-7, 01/01/ 2021, doi: 10.1038/s42003-021-01665-6.
[3]E. F. Driesschaert, T.; Goosse, H.; Huybrechts, Philippe; Janssens, I.; Mouchet, A.; Munhoven, G.; Brovkin, V.; Weber, S. L. and E. Q3W; Driesschaert, Fichefet, T. , Goosse, H. , Huybrechts, P. , Janssens, I. , Mouchet, A. , Munhoven, G. , Brovkin, V. and Weber, S. L. , "Modeling the influence of Greenland ice sheet melting on the Atlantic meridional overturning circulation during the next millennia," (in Undetermined), Geophysical Research Letters, vol. 34, p. L10707, 2007. [Online]. Available: https://doi.org/10.1029/2007GL029516.
[4]U. Carolina et al., "Evaluation of animal and plant diversity suggests Greenland’s thaw hastens the biodiversity crisis," Communications Biology, vol. 5, no. 1, pp. 1-12, 09/01/ 2022, doi: 10.1038/s42003-022-03943-3.
[5]R. B. Alley, P. U. Clark, P. Huybrechts, and I. Joughin, "Ice-sheet and sea-level changes," Science, vol. 310, no. 5747, pp. 456-60, Oct 21 2005, doi: 10.1126/science.1114613.
[6]M. W. Morlighem, C. N.; Rignot, Eric; An, L.; Arndt, Jan Erik; Bamber, Jonathan L.; Catania, G.; Chauché, N.; Dowdeswell, Julian A.; Dorschel, Boris; Fenty, Ian;, "BedMachine v3: Complete bed topography and ocean bathymetry mapping of Greenland from multi-beam echo sounding combined with mass conservation," (in Undetermined), Geophysical Research Letters, vol. 44, no. 21, pp. 11051-11061, 2017. [Online]. Available: https://doi.org/10.1002/2017GL074954.
[7]M. R. van den Broeke et al., "On the recent contribution of the Greenland ice sheet to sea level change," Cryosphere, vol. 10, no. 5, pp. 1933-1946, 2016, doi: 10.5194/tc-10-1933-2016.
[8]S. Brown et al., "Global costs of protecting against sea-level rise at 1.5 to 4.0 °C," Climatic Change: An Interdisciplinary, International Journal Devoted to the Description, Causes and Implications of Climatic Change, vol. 167, no. 1-2, 07/01/ 2021, doi: 10.1007/s10584-021-03130-z.
[9]T. Wahl, S. Jain, J. Bender, S. D. Meyers, and M. E. Luther, "Increasing risk of compound flooding from storm surge and rainfall for major US cities," Nature Climate Change, vol. 5, no. 12, pp. 1093-1097, 2015, doi: 10.1038/nclimate2736.
[10]J. Fyke, S. Price, O. Sergienko, M. Löfverström, and J. T. M. Lenaerts, "An Overview of Interactions and Feedbacks Between Ice Sheets and the Earth System," (in English), Reviews of Geophysics, vol. 56, no. 2, pp. 361-408-408, 06/01/ 2018, doi: 10.1029/2018RG000600.
[11]P. J. Irvine, D. J. Lunt, E. J. Stone, and A. Ridgwell, "The fate of the Greenland Ice Sheet in a geoengineered, high CO2 world," (in Undetermined), Environmental Research Letters, vol. 4, p. 45109, 2009. [Online]. Available: https://doi.org/10.1088/1748-9326/4/4/045109
[12]Andrea J. Pain, Jonathan B. Martin, Ellen E. Martin, Åsa K. Rennermalm, and S. Rahman, "Heterogeneous CO2 and CH4 content of glacial meltwater from the Greenland Ice Sheet and implications for subglacial carbon processes," Cryosphere, vol. 15, no. 3, pp. 1627-1644, 2021, doi: 10.5194/tc-15-1627-2021.
[13]Z. Fang, N. Wang, Y. Wu, and Y. Zhang, "Greenland-Ice-Sheet Surface Temperature and Melt Extent from 2000 to 2020 and Implications for Mass Balance," Remote Sensing, vol. 15, no. 4, p. 1149, 2023, doi: 10.3390/rs15041149.
[14]E. Hanna et al., "Greenland surface air temperature changes from 1981 to 2019 and implications for ice-sheet melt and mass-balance change," (in English), International Journal of Climatology, vol. 41, no. S1, pp. E1336-E1352-E1352, 01/01/ 2021, doi: 10.1002/joc.6771.
[15]J. Bolibar, A. Rabatel, I. Gouttevin, C. Galiez, T. Condom, and E. Sauquet, "Deep learning applied to glacier evolution modelling," Cryosphere, vol. 14, no. 2, pp. 565-584, 2020, doi: 10.5194/tc-14-565-2020.
[16]D. Steiner, A. Walter, and H. J. Zumbühl, "The application of a non-linear back-propagation neural network to study the mass balance of Grosse Aletschgletscher, Switzerland," Journal of Glaciology, vol. 51, no. 173, pp. 313-323, 2017, doi: 10.3189/172756505781829421.
[17]J. Bolibar, A. Rabatel, I. Gouttevin, H. Zekollari, and C. Galiez, "Nonlinear sensitivity of glacier mass balance to future climate change unveiled by deep learning," (in English), Nature Communications, vol. 13, no. 1, 12/01/ 2022, doi: 10.1038/s41467-022-28033-0.
[18]R. Anilkumar, R. Bharti, D. Chutia, and S. P. Aggarwal, "Modelling the Point Mass Balance for the Glaciers of Central European Alps using Machine Learning Techniques," EGUsphere, vol. 2022, pp. 1-27, 2022, doi: 10.5194/egusphere-2022-1076.
[19]R. Sellevold and M. Vizcaino, "First Application of Artificial Neural Networks to Estimate 21st Century Greenland Ice Sheet Surface Melt," Geophys Res Lett, vol. 48, no. 16, p. e2021GL092449, Aug 2021, doi: 10.1029/2021GL092449.
[20]J. Che et al., "Reconstruction of Near-Surface Air Temperature over the Greenland Ice Sheet Based on MODIS Data and Machine Learning Approaches," Remote Sensing, vol. 14, no. 22, 2022, doi: 10.3390/rs14225775.
[21]D. N. Wiese, D.-N. Yuan, C. Boening, F. W. Landerer, and M. M. Watkins. "JPL GRACE and GRACE-FO Mascon Ocean, Ice, and Hydrology Equivalent HDR Water Height RL06.1M CRI Filtered Version 3.0, Ver. 3.0, PO.DAAC, CA, USA." http://dx.doi.org/10.5067/TEMSC-3MJ62 (accessed 1 April, 2023).
[22]B. D. Tapley et al., "Contributions of GRACE to understanding climate change," Nat Clim Chang, vol. 5, no. 5, pp. 358-369, Apr 15 2019, doi: 10.1038/s41558-019-0456-2.
[23]X. D. Lan, E. J.; Mund, J. W.; Crotwell, A. M.; Crotwell, M. J.; Moglia, E.; Madronich, M.; Neff, D.; and Thoning, K. W. "Atmospheric Methane Dry Air Mole Fractions from the NOAA GML Carbon Cycle Cooperative Global Air Sampling Network." https://doi.org/10.15138/VNCZ-M766 (accessed 1 April, 2023).
[24]T. J. Osborn et al., "Land Surface Air Temperature Variations Across the Globe Updated to 2019: The CRUTEM5 Data Set," Journal of Geophysical Research: Atmospheres, vol. 126, no. 2, 2021, doi: 10.1029/2019jd032352.
[25]J. J. Kennedy, N. A. Rayner, C. P. Atkinson, and R. E. Killick, "An Ensemble Data Set of Sea Surface Temperature Change From 1850: The Met Office Hadley Centre HadSST.4.0.0.0 Data Set," Journal of Geophysical Research: Atmospheres, vol. 124, no. 14, pp. 7719-7763, 2019, doi: 10.1029/2018jd029867.
[26]B. Z. Beckley, N.P.; Holmes, S.A.;Lemoine, F.G.; Ray, R.D.; Mitchum, G.T.; Desai, S.; Brown, S.T.. "Global Mean Sea Level Trend from Integrated Multi-Mission Ocean Altimeters TOPEX/Poseidon, Jason-1, OSTM/Jason-2, and Jason-3 Version 5.1. Ver. 5.1 PO.DAAC, CA, USA.," 2021. [Online]. Available: https://doi.org/10.5067/GMSLM-TJ151.
[27]J. Benesty, J. Chen, Y. Huang, and I. Cohen, "Pearson Correlation Coefficient," in Noise Reduction in Speech Processing, (Springer Topics in Signal Processing, 2009, ch. Chapter 5, pp. 1-4.
[28]S. Hochreiter and J. Schmidhuber, "Long short-term memory," Neural Computation, vol. 9, no. 8, p. 1735, 1997, doi: 10.1162/neco.1997.9.8.1735.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top