作物蒸發散在水收支過程中扮演著相當重要的角色，亦稱為作物需水量(ETcrop)，為政府推動農地重劃或擴大灌溉服務過程，掌握灌區需求水量之重要參數，其中參考作物蒸發散量(ETo)是計算作物需水量(ETcrop)的主要步驟，快速而準確推估新灌區不同作物種植的需求水量對於灌溉計畫的制定相當重要。其中FAO56_Penman-Monteith方法為聯合國糧食及農業組織(FAO)的專家小組建議推估ETo的方法之一，亦為國際間普遍使用之方法，但該方法需要使用到許多的氣象參數，包含風速、濕度、太陽輻射等參數，實際應用上，因參數取得不易而常常受限，台灣現階段所設立的許多氣象測站也無法直接應用FAO56_Penman-Monteith方法進行推估。其次，面臨氣候變遷對於氣候、雨量及水資源的衝擊，台灣水資源使用比例以農業灌溉用水為首的情形下，作物需水量於未來不同期間的變動牽動台灣整體水資源的供需調配，因此若能事前評估未來變動趨勢，有助於國家提前因應氣象及水情變動趨勢而擬定修正產業用水及區域水資源供需及調配政策。
有鑒於此，本研究基於參考作物蒸發散量(ETo)推估之快速性、廣泛性、準確性及因應氣候變遷的未來變動趨勢預測，建立一套參考作物蒸發散量(ETo)優選模式，該模式由3種計算方式組成：1.機器學習(MLP)；2.Hargreaves and Samani(HS)經驗公式修正；3.比較HS、MLP與FAO建議的參數替代方法，並以均方根誤差(RMSE)為優選評估指標；進一步優選3種適合台灣的GCM模型(CanESM5、EC-Earth3、ACCESS-ESM1-5)，並透過機器學習的單參數方法，套用臺灣氣候變遷推估資訊與調適知識平台計畫(TCCIP)提供的未來溫度資料進行未來短中長期之4種情境條件下(SSP1-2.6、SSP2-4.5、SSP3-7.0、SSP5-8.5)之變化趨勢推估。
本研究以柯本-蓋革(Köppen-Geiger)氣候分類法為基礎將台灣劃設為4個不同氣候區；並選定嘉義、新竹、恆春及永康氣象測站作為各氣候分區的代表測站，並探討不同測站的推估結果，採用20年日資料(2004~2023年)進行評估，並以16年日資料建立模型，4年的日資料進行驗證；研究成果如下列四點：1.使用線性回歸方法(HS_adj2)修正Hargreaves and Samani(HS)會優於使用乘上單一修正係數方法(HS_adj1)，並且HS_adj2方法與HS_adj1方法可以使原始HS公式的RMSE值分別降低0~21.74 %、0~12.35 %，此2種修正方法在夏季的改善效果相當有限，並且在春季及冬季的改善效果則普遍較好，此外，可以發現不同測站透過HS_adj2方法所求得的ETo在冬季的RMSE皆較小，並且在夏季的RMSE普遍較大。2.優選模式顯示在所有情況下ETo的最佳推估方法皆為MLP模型，當氣象觀測參數僅有溫度資料時，ETo的最佳推估方法為MLP-a模型，其次則是HS_adj2模型，而HS模型與PM-1的模型性能排名則要視目標測站而定，但大多數情況下，PM-1模型會優於HS模型，也就是說原始HS公式推估結果普遍會比PM-1方法差，但透過本地校準後的HS公式(HS_adj2方法)則會優於PM-1方法。然而當測站本身因缺乏歷史觀測資料，導致MLP模型無法建立且HS公式也無法得到本地校準時，依然可透過PM-1~PM-4方法進行ETo的推估。3.交叉分析及驗證成果發現推估ETo其首要的氣象觀測參數為太陽輻射數據，而風速及濕度則較為次要，因此本研究建議未來在設置氣象觀測站時，可以優先考量設置太陽輻射測量儀器。4.氣候變遷預測成果顯示各測站應用3種GCM模型及4種模擬情境條件下均有相同趨勢，即ETo從隨著短期(2021~2040)、中期(2041~2060)至中長期(2061~2080)的這段期間呈現持續上升，並於中長期(2061~2080)或長期(2081~2100)時達到高峰，此外，在大部分情況下，所有GCM模型於SSP5-8.5的模擬情境下ETo的上升幅度最大，而在SSP1-2.6的模擬情境下ETo的上升幅度最小，凸顯氣候變遷對於台灣農業灌溉需求水量隨著時間推移產生高需求的趨勢，建議未來的水資源供需政策須提前調整因應。有鑒於此，本研究建議未來應想辦法控制溫室氣體的排放量，避免未來朝著SSP5-8.5情境發展。

Crop evapotranspiration (ETcrop) plays a crucial role in the water budget process and is also known as crop water requirement. It is an essential parameter for the government in assessing water demand within irrigation areas, especially when promoting land readjustment or expanding irrigation services. The estimation of reference evapotranspiration (ETo) is a key step in calculating ETcrop, and quickly and accurately estimating water requirements for different crops in new irrigation areas is critical for planning irrigation schemes. The FAO56 Penman-Monteith method, recommended by the Food and Agriculture Organization of the United Nations (FAO) expert panel, is one of the widely used methods for ETo estimation worldwide. However, it requires multiple meteorological parameters, such as wind speed, humidity, and solar radiation, which often pose challenges due to data accessibility. Consequently, many weather stations in Taiwan currently cannot directly apply the FAO56 Penman-Monteith method for ETo estimation. Furthermore, in the face of climate change impacts on weather patterns, rainfall, and water resources, Taiwan’s water usage is dominated by agricultural irrigation. Variations in crop water requirements across future time periods will significantly influence the balance of water supply and demand in Taiwan. Therefore, the ability to forecast future trends in water needs will assist the government in proactively adapting industrial water usage and regional water resource management policies to better respond to anticipated climate and hydrological changes. 
Therefore, this study develops an optimal model for estimating reference crop evapotranspiration (ETo) based on its rapidity, broad applicability, accuracy, and predictive value in assessing future trends under climate change. The model comprises three calculation methods: (1) machine learning (MLP), (2) an adjusted Hargreaves and Samani (HS) empirical formula, and (3) a comparison between HS, MLP, and FAO-recommended parameter substitution methods, with root mean square error (RMSE) used as the optimal evaluation indicator. Additionally, three suitable global climate models (GCMs) for Taiwan—CanESM5, EC-Earth3, and ACCESS-ESM1-5—were selected. Using a single-parameter machine learning approach, the study incorporated future temperature data provided by the Taiwan Climate Change Projection InformationPlatform (TCCIP) to estimate ETo trends across four scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5) over short-term, medium-term, medium-long-term, and long-term periods.
This study classifies Taiwan into four distinct climate zones based on the Köppen-Geiger climate classification system. The Chiayi, Hsinchu, Hengchun, and Yongkang meteorological stations were selected as representative stations for each climate zone. The estimation results for these stations were analyzed using 20 years of daily data (2004~2023). A model was developed using 16 years of daily data, with an additional 4 years of daily data used for validation. The findings of this study are summarized in the following four points：1.Using the linear regression method (HS_adj2) to adjust the Hargreaves and Samani (HS) formula outperforms the method of applying a single correction factor (HS_adj1). The HS_adj2 and HS_adj1 methods can reduce the RMSE values of the original HS formula by 0~21.74 % and 0~12.35 %, respectively. However, the improvement achieved by both methods is relatively limited during summer, while more significant improvements are generally observed in spring and winter. Furthermore, it is evident that the ETo values obtained using the HS_adj2 method show smaller RMSE values in winter across different stations, whereas larger RMSE values are generally observed in summer. 2.The optimal model selection shows that the best method for estimating ETo in all cases is the MLP model. When only temperature data is available from meteorological observations, the best method for estimating ETo is the MLP-a model, followed by the HS_adj2 model. The performance ranking of the HS model and the PM-1 model depends on the target station, but in most cases, the PM-1 model outperforms the HS model. In other words, the original HS formula generally performs worse than the PM-1 method in estimating ETo. However, after local calibration using the HS_adj2 method, the adjusted HS formula surpasses the performance of the PM-1 method. Nevertheless, in cases where historical observation data at a station are insufficient to develop MLP models or to locally calibrate the HS formula, the PM-1 ~ PM-4 methods can still be utilized for ETo estimation. 3.Cross-analysis and validation results show that the primary meteorological observation parameter for estimating ETo is solar radiation data, while wind speed and humidity are of secondary importance. Therefore, this study recommends prioritizing the installation of solar radiation measurement instruments when establishing meteorological observation stations in the future. 4.The climate change prediction results show that, under the application of three GCM models and four simulation scenarios, all stations exhibit the same trend. Specifically, ETo is projected to continuously increase from the short-term (2021~2040) through the medium-term (2041~2060) to the medium-to-long-term (2061~2080), peaking in either the medium-to-long-term (2061~2080) or long-term (2081~2100). Furthermore, in most cases, all GCM models show the largest increase in ETo under the SSP5-8.5 scenario, while the smallest increase occurs under the SSP1-2.6 scenario. This highlights the increasing trend of agricultural irrigation water demand in Taiwan due to climate change over time. It is recommended that future water resource supply and demand policies be adjusted in advance to address this issue. Therefore, this study suggests that efforts should be made to control greenhouse gas emissions to prevent the future development of the SSP5-8.5 scenario.

摘要	摘-I
ABSTRACT	摘-III
目錄	I
表目錄	III
圖目錄	V
第一章 緒論	1
1.1 研究動機與目的	1
1.2 研究流程架構	4
第二章 文獻回顧	6
2.1 資料補遺方法	6
2.2 參考作物蒸發散量(ETo)推估方法	8
2.3 機器學習方法回顧	11
2.4 氣候變遷對環境的衝擊探討	13
第三章 研究方法	15
3.1 參考作物蒸發散量(ETo)推估方法	15
3.1.1 FAO56-Penman-Monteith(FAO56-PM)方法	15
3.1.2 Hargreaves and Samani(HS)公式	20
3.1.3 機器學習-MLP模型	21
3.2 參考作物蒸發散量(ETo)優選模式之建立	23
3.2.1 FAO56-PM_氣象參數之建議替代方法	23
3.2.2 Hargreaves and Samani(HS)公式修正方法	24
3.2.3 MLP模型之建立及演算	25
3.3 氣候變遷預測	28
3.3.1 氣候變遷情境	28
3.3.2 GCM模型之選用	29
3.4 優選流程評估說明	31
3.4.1 資料補遺- 優選流程	31
3.4.2 交叉驗證方法	32
3.4.3 評估指標	34
3.5 研究區域及氣象資料概述	35
3.5.1 氣候分區理論及應用	35
3.5.2 氣象資料選用	40
第四章 結果與討論	42
4.1 資料補遺優選方法及結果	42
4.1.1 資料補遺方法優選結果	42
4.1.2 資料補遺結果	44
4.2 參考作物蒸發散量(ETo)分析成果	60
4.2.1 FAO56-Penman-Monteith推估成果	60
4.2.2 Hargreaves and Samani(HS)推估成果	63
4.2.3 FAO56-Penman-Monteith替代方法推估成果	66
4.2.4 Hargreaves and Samani(HS)公式推估及修正成果	69
4.3 機器學習(MLP)模式推估成果	80
4.3.1 MLP模型超參數選定結果(GridSearchCV)	80
4.3.2 MLP模型交叉驗證結果	85
4.4 參考作物蒸發散量(ETo)優選模式分析成果及效能評估	87
4.5 推估氣候變遷對參考作物蒸發散量(ETo)之影響	90
4.5.1 GCM模型之選用	90
4.5.2 氣候變遷對ETo的影響預測	101
第五章 結論與建議	132
5.1 結論	132
5.2 建議	134
參考文獻	136
表目錄
表3-1  FAO56-PM公式與替代方法	24
表3-2  各模型對應之輸入參數	27
表3-3  各國發展之GCM模式一覽	30
表3-4  交叉驗證分組說明	33
表3-5  柯本-蓋革(KÖPPEN-GEIGER)氣候分類定義	37
表3-6  台灣氣象測站一覽	38
表3-7  目標測站位置	40
表3-8  各測站使用之補遺測站IDW權重一覽	40
表3-9  各測站2004~2023年_資料有效筆數及缺失筆數(補遺前)	41
表4-1  各目標測站資料補遺評估結果	43
表4-2  各測站2004~2023年_敘述性統計資料(補遺前後)	44
表4-3  各目標測站經驗係數KRS之選定	67
表4-4  各目標測站歷年最高氣溫與最低氣溫之差值	67
表4-5  各目標測站FAO56-PM替代方法結果	69
表4-6  各目標測站歷年最低氣溫與露點溫度之差值	69
表4-7  各測站修正係數(按年度別修正)	70
表4-8  HS公式修正後結果(按年度別修正)	72
表4-9  嘉義測站各月份修正係數(按月分別修正)	73
表4-10 新竹測站各月份修正係數(按月分別修正)	74
表4-11 恆春測站各月份修正係數(按月分別修正)	75
表4-12 永康測站各月份修正係數(按月分別修正)	76
表4-13 各目標測站HS公式修正後結果(按月分別修正)	78
表4-14 各測站前6順位MLP-A模型網格搜索結果	81
表4-15 各測站前6順位MLP-B模型網格搜索結果	82
表4-16 各測站前6順位MLP-C模型網格搜索結果	83
表4-17 各測站前6順位MLP-D模型網格搜索結果	84
表4-18 各測站 MLP-A模型_不同交叉驗證分組方法結果	86
表4-19 各測站 MLP-B模型_不同交叉驗證分組方法結果	86
表4-20 各測站 MLP-C模型_不同交叉驗證分組方法結果	87
表4-21 各測站 MLP-D模型_不同交叉驗證分組方法結果	87
表4-22 各測站不同情境下GCM模型排名_最高氣溫	99
表4-23 各測站不同情境下GCM模型排名_最低氣溫	99
表4-24 GCM模型總排名	100
表4-25 各測站所選用模型與網格	101
 
圖目錄
圖1-1  本論文研究流程圖	5
圖3-1  MLP模型架構圖	22
圖3-2  農業站與署屬有人站分布圖	39
圖3-3  目標測站與補遺測站位置	39
圖4-1  各測站平均氣壓盒鬚圖(2004~2023年)	46
圖4-2  各測站最高氣溫盒鬚圖(2004~2023年)	47
圖4-3  各測站最低氣溫盒鬚圖(2004~2023年)	48
圖4-4  各測站平均露點溫度盒鬚圖(2004~2023年)	49
圖4-5  各測站平均風速盒鬚圖(2004~2023年)	50
圖4-6  各測站累積日射量盒鬚圖(2004~2023年)	51
圖4-7  各測站平均氣壓之逐月盒鬚圖(2004~2023年)	54
圖4-8  各測站最高氣溫之逐月盒鬚圖(2004~2023年)	55
圖4-9  各測站最低氣溫之逐月盒鬚圖(2004~2023年)	56
圖4-10 各測站平均露點溫度之逐月盒鬚圖(2004~2023年)	57
圖4-11 各測站平均風速之逐月盒鬚圖(2004~2023年)	58
圖4-12 各測站累積日射量之逐月盒鬚圖(2004~2023年)	59
圖4-13 各測站參考蒸發散量(FAO56-PM)盒鬚圖(2004~2023年)	61
圖4-14 各測站參考蒸發散量(FAO56-PM)各月盒鬚圖(2004~2023年)	62
圖4-15 各測站參考蒸發散量(HS)盒鬚圖(2004~2023年)	64
圖4-16 各測站參考蒸發散量(HS)各月盒鬚圖(2004~2023年)	65
圖4-17 各測站FAO56-PM方法與HS方法年平均結果(2004~2023年)	71
圖4-18 各測站FAO56-PM方法與HS方法月平均結果(2004~2023年)	77
圖4-19 各測站HS公式(按月份別修正)各月份結果	79
圖4-20 嘉義測站各方法之效能評估	88
圖4-21 新竹測站各方法之效能評估	89
圖4-22 恆春測站各方法之效能評估	89
圖4-23 永康測站各方法之效能評估	89
圖4-24 嘉義測站_不同情境下GCM模型與實際最高溫之RMSE(2015~2023)	91
圖4-25 嘉義測站_不同情境下GCM模型與實際最低溫之RMSE (2015~2023)	92
圖4-26 新竹測站_不同情境下GCM模型與實際最高溫之RMSE (2015~2023)	93
圖4-27 新竹測站_不同情境下GCM模型與實際最低溫之RMSE (2015~2023)	94
圖4-28 恆春測站_不同情境下GCM模型與實際最高溫之RMSE (2015~2023)	95
圖4-29 恆春測站_不同情境下GCM模型與實際最低溫之RMSE (2015~2023)	96
圖4-30 永康測站_不同情境下GCM模型與實際最高溫之RMSE (2015~2023)	97
圖4-31 永康測站_不同情境下GCM模型與實際最低溫之RMSE (2015~2023)	98
圖4-32 嘉義測站_CANESM5模型於不同情境及期間條件下之ETO	103
圖4-33 嘉義測站_ EC-EARTH3模型於不同情境及期間條件下之ETO	104
圖4-34 嘉義測站_ACCESS-ESM1-5模型於不同情境及期間條件下之ETO	105
圖4-35 新竹測站_ CANESM5模型於不同情境及期間條件下之ETO	106
圖4-36 新竹測站_ EC-EARTH3模型於不同情境及期間條件下之ETO	107
圖4-37 新竹測站_ ACCESS-ESM1-5模型於不同情境及期間條件下之ETO	108
圖4-38 恆春測站_ CANESM5模型於不同情境及期間條件下之ETO	109
圖4-39 恆春測站_EC-EARTH3模型於不同情境及期間條件下之ETO	110
圖4-40 恆春測站_ACCESS-ESM1-5模型於不同情境及期間條件下之ETO	111
圖4-41 永康測站_CANESM5模型於不同情境及期間條件下之ETO	112
圖4-42 永康測站_EC-EARTH3模型於不同情境及期間條件下之ETO	113
圖4-43 永康測站_ACCESS-ESM1-5模型於不同情境及期間條件下之ETO	114
圖4-44 嘉義測站_ 不同GCM模型於4種模擬情境及期間條件下之ETO	115
圖4-45 新竹測站_ 不同GCM模型於4種模擬情境及期間條件下之ETO	116
圖4-46 恆春測站_ 不同GCM模型於4種模擬情境及期間條件下之ETO	117
圖4-47 永康測站_ 不同GCM模型於4種模擬情境及期間條件下之ETO	118
圖4-48 嘉義測站_不同情境及期間條件下3種GCM模型模擬的之ETO	119
圖4-49 新竹測站_不同情境及期間條件下3種GCM模型模擬的之ETO	120
圖4-50 恆春測站_不同情境及期間條件下3種GCM模型模擬的之ETO	121
圖4-51 永康測站_不同情境及期間條件下3種GCM模型模擬的之ETO	122
圖4-52 嘉義測站_不同情境下3種GCM模型的ETO變化量	124
圖4-53 新竹測站_不同情境下3種GCM模型的ETO變化量	125
圖4-54 恆春測站_不同情境下3種GCM模型的ETO變化量	126
圖4-55 永康測站_不同情境下3種GCM模型的ETO變化量	127
圖4-56 嘉義測站_不同情境下3種GCM模型的ETO變化百分比	128
圖4-57 新竹測站_不同情境下3種GCM模型的ETO變化百分比	129
圖4-58 恆春測站_不同情境下3種GCM模型的ETO變化百分比	130
圖4-59 永康測站_不同情境下3種GCM模型的ETO變化百分比	131

Acuna, E., & Rodriguez, C. (2004). The treatment of missing values and its effect on classifier accuracy. In Classification, Clustering, and Data Mining Applications: Proceedings of the Meeting of the International Federation of Classification Societies (IFCS), Illinois Institute of Technology, Chicago, 15–18 July 2004, p. 639~647. Berlin, Heidelberg: Springer Berlin Heidelberg.
Ajami, H. (2021). Geohydrology: global hydrological cycle. 
Akca, S., Keskin, M. Z., Arra, A. A., & Şişman, E. (2023). Spatial and regression-based missing precipitation data imputation: Western Black Sea region. Intercontinental Geoinformation Days, 7, p.180~183. 
Allen, R. G., Pereira, L. S., Raes, D., & Smith, M. (1998). Crop evapotranspiration-Guidelines for computing crop water requirements-FAO Irrigation and drainage paper 56. Fao, Rome, 300(9), D05109. 
Allen, R. G., Pereira, L. S., Howell, T. A., & Jensen, M. E. (2011). Evapotranspiration information reporting: I. Factors governing measurement accuracy. Agricultural Water Management, 98(6), p.899~920. 
Almorox, J., Quej, V. H., & Martí, P. (2015). Global performance ranking of temperature-based approaches for evapotranspiration estimation considering Köppen climate classes. Journal of Hydrology, 528, p.514~522. 
Almorox, J., & Grieser, J. (2016). Calibration of the Hargreaves–Samani method for the calculation of reference evapotranspiration in different Köppen climate classes. Hydrology Research, 47(2), p.521~531. 
Anapalli, S. S., Ahuja, L. R., Gowda, P. H., Ma, L., Marek, G., Evett, S. R., & Howell, T. A. (2016). Simulation of crop evapotranspiration and crop coefficients with data in weighing lysimeters. Agricultural Water Management, 177, p.274~283. 
Antonopoulos, V. Z., & Antonopoulos, A. V. (2017). Daily reference evapotranspiration estimates by artificial neural networks technique and empirical equations using limited input climate variables. Computers and Electronics in Agriculture, 132, p.86~96. 
Arnell, N. W., Lowe, J. A., Challinor, A. J., & Osborn, T. J. (2019). Global and regional impacts of climate change at different levels of global temperature increase. Climatic Change, 155, p.377~391. 
Aryalekshmi, B., Biradar, R. C., Chandrasekar, K., & Ahamed, J. M. (2021). Analysis of various surface energy balance models for evapotranspiration estimation using satellite data. The Egyptian Journal of Remote Sensing and Space Science, 24(3), p.1119~1126. 
Beck, H. E., McVicar, T. R., Vergopolan, N., Berg, A., Lutsko, N. J., Dufour, A., Zeng, Z., Jiang, X., van Dijk, A. I., & Miralles, D. G. (2023). High-resolution (1 km) Köppen-Geiger maps for 1901–2099 based on constrained CMIP6 projections. Scientific data, 10(1), 724. 
Beck, H. E., Zimmermann, N. E., McVicar, T. R., Vergopolan, N., Berg, A., & Wood, E. F. (2018). Present and future Köppen-Geiger climate classification maps at 1-km resolution. Scientific data, 5(1), p.1~12. 
Bellido-Jiménez, J. A., Estévez, J., & García-Marín, A. P. (2022). A regional machine learning method to outperform temperature-based reference evapotranspiration estimations in Southern Spain. Agricultural Water Management, 274, 107955.
Boretti, A., & Rosa, L. (2019). Reassessing the projections of the world water development report. NPJ Clean Water, 2(1), 15. 
Chartzoulakis, K., & Bertaki, M. (2015). Sustainable water management in agriculture under climate change. Agriculture and Agricultural Science Procedia, 4, p.88~98. 
Chen, F.W., & Liu, C.W. (2012). Estimation of the spatial rainfall distribution using inverse distance weighting (IDW) in the middle of Taiwan. Paddy and Water Environment, 10(3), 209-222. https://doi.org/10.1007/s10333-012-0319-1 .
Chen, H., Huo, Z., Dai, X., Ma, S., Xu, X., & Huang, G. (2018). Impact of agricultural water-saving practices on regional evapotranspiration: The role of groundwater in sustainable agriculture in arid and semi-arid areas. Agricultural and Forest Meteorology, 263, p.156~168. 
Chen, Z., Zhu, Z., Jiang, H., & Sun, S. (2020). Estimating daily reference evapotranspiration based on limited meteorological data using deep learning and classical machine learning methods. Journal of Hydrology, 591, 125286. 
Chia, M. Y., Huang, Y. F., & Koo, C. H. (2022). ANN-Based Reference Evapotranspiration Estimation: Effects of Data Normalization and Parameters Selection. Proceedings of International Conference on Emerging Technologies and Intelligent Systems: ICETIS 2021 Volume 2, p.3~12. 
Córdova, M., Carrillo-Rojas, G., Crespo, P., Wilcox, B., & Célleri, R. (2015). Evaluation of the Penman-Monteith (FAO 56 PM) method for calculating reference evapotranspiration using limited data. Mountain Research and Development, 35(3), p.230~239. 
De Sousa Lima, J. R., Antonino, A. C. D., de Souza, E. S., Hammecker, C., Montenegro, S. M. G. L., & de Oliveira Lira, C. A. B. (2013). Calibration of Hargreaves-Samani equation for estimating reference evapotranspiration in sub-humid region of Brazil. Journal of Water Resource and Protection, 2013.
Dimitriadou, S., & Nikolakopoulos, K. G. (2022). Artificial neural networks for the prediction of the reference evapotranspiration of the Peloponnese Peninsula, Greece. Water, 14(13), 2027. 
Dou, X., & Yang, Y. (2018). Evapotranspiration estimation using four different machine learning approaches in different terrestrial ecosystems. Computers and Electronics in Agriculture, 148, p.95~106. 
Ferrari, G. T., & Ozaki, V. (2014). Missing data imputation of climate datasets: Implications to modeling extreme drought events. Revista Brasileira de Meteorologia, 29, p.21~28. 
Gamal, R., El-Shirbeny, M., Abou-Hadid, A., Swelam, A., El-Gindy, A.-G., Arafa, Y., & Nangia, V. (2022). Identification and quantification of actual evapotranspiration using integrated satellite data for sustainable water management in dry areas. Agronomy, 12(9), 2143. 
Gavilán, P., Lorite, I., Tornero, S., & Berengena, J. (2006). Regional calibration of Hargreaves equation for estimating reference ET in a semiarid environment. Agricultural Water Management, 81(3), p.257~281.
Goyal, R. (2004). Sensitivity of evapotranspiration to global warming: a case study of arid zone of Rajasthan (India). Agricultural Water Management, 69(1), p.1~11.
Hargreaves, G. H., & Samani, Z. A. (1985). Reference crop evapotranspiration from temperature. Applied engineering in agriculture, 1(2), p.96~99. 
Hasanpour Kashani, M., & Dinpashoh, Y. (2012). Evaluation of efficiency of different estimation methods for missing climatological data. Stochastic environmental research and risk assessment, 26, p.59~71. 
Hashemi, M., & Sepaskhah, A. R. (2020). Evaluation of artificial neural network and Penman–Monteith equation for the prediction of barley standard evapotranspiration in a semi-arid region. Theoretical and Applied Climatology, 139, p.275~285.
Hernández-Bedolla, J., Solera, A., Sánchez-Quispe, S. T., & Domínguez-Sánchez, C. (2023). Comparative analysis of 12 reference evapotranspiration methods for semi-arid regions (Spain). Journal of Water and Climate Change, 14(9), p.2954~2969.
Hossain, S., Homma, K., & Shiraiwa, T. (2014). Decadal and monthly change of an empirical coefficient in the relation between solar radiation and the daily range of temperature in Japan: implications for the estimation of solar radiation based on temperature. Plant Production Science, 17(4), p.333~341.
Jaramillo, S., Graterol, E., & Pulver, E. (2020). Sustainable transformation of rainfed to irrigated agriculture through water harvesting and smart crop management practices. Frontiers in Sustainable Food Systems, 4, 437086. 
Kanda, N., Negi, H., Rishi, M. S., & Shekhar, M. (2018). Performance of various techniques in estimating missing climatological data over snowbound mountainous areas of Karakoram Himalaya. Meteorological Applications, 25(3), p.337~349. 
Kim, S.J., Bae, S.J., & Jang, M.-W. (2022). Linear regression machine learning algorithms for estimating reference evapotranspiration using limited climate data. Sustainability, 14(18), 11674. 
Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980. 
Kingston, D. G., Todd, M. C., Taylor, R. G., Thompson, J. R., & Arnell, N. W. (2009). Uncertainty in the estimation of potential evapotranspiration under climate change. Geophysical Research Letters, 36(20). 
Kumar, M., Raghuwanshi, N., & Singh, R. (2011). Artificial neural networks approach in evapotranspiration modeling: a review. Irrigation science, 29, p.11~25. 
Lange, H., & Sippel, S. (2020). Machine learning applications in hydrology. Forest-water interactions, p.233~257. 
Liuzzo, L., Viola, F., & Noto, L. V. (2016). Wind speed and temperature trends impacts on reference evapotranspiration in Southern Italy. Theoretical and Applied Climatology, 123, p.43~62. 
López-Urrea, R., de Santa Olalla, F. M., Fabeiro, C., & Moratalla, A. (2006). Testing evapotranspiration equations using lysimeter observations in a semiarid climate. Agricultural Water Management, 85(1-2), p.15~26. 
Luo, J., Dou, X., & Ma, M. (2022). Evaluation of empirical and machine learning approaches for estimating monthly reference evapotranspiration with limited meteorological data in the Jialing River Basin, China. International Journal of Environmental Research and Public Health, 19(20), 13127. 
Malhi, G. S., Kaur, M., & Kaushik, P. (2021). Impact of climate change on agriculture and its mitigation strategies: A review. Sustainability, 13(3), 1318. 
Mendelsohn, R. (2009). The impact of climate change on agriculture in developing countries. Journal of natural resources policy research, 1(1), p.5~19. 
Modaresi, F., & Araghi, A. (2023). Projecting future reference evapotranspiration in Iran based on CMIP6 multi-model ensemble. Theoretical and Applied Climatology, 153(1), p.101~112.
Mosaffa, H., Sadeghi, M., Mallakpour, I., Jahromi, M. N., & Pourghasemi, H. R. (2022). Application of machine learning algorithms in hydrology. In Computers in earth and environmental sciences.  Elsevier, p.585~591. 
Park, S., Im, J., Jang, E., & Rhee, J. (2016). Drought assessment and monitoring through blending of multi-sensor indices using machine learning approaches for different climate regions. Agricultural and Forest Meteorology, 216, p.157~169. 
Patil, A. P., & Deka, P. C. (2016). An extreme learning machine approach for modeling evapotranspiration using extrinsic inputs. Computers and Electronics in Agriculture, 121, p.385~392. 
Peel, M. C., Finlayson, B. L., & McMahon, T. A. (2007). Updated world map of the Köppen-Geiger climate classification. Hydrology and earth system sciences, 11(5), p.1633~1644. 
Petropoulos, G. P., Srivastava, P. K., Piles, M., & Pearson, S. (2018). Earth observation-based operational estimation of soil moisture and evapotranspiration for agricultural crops in support of sustainable water management. Sustainability, 10(1), 181. 
Rodríguez, R., Pastorini, M., Etcheverry, L., Chreties, C., Fossati, M., Castro, A., & Gorgoglione, A. (2021). Water-quality data imputation with a high percentage of missing values: A machine learning approach. Sustainability, 13(11), 6318. 
Santos, P. A. B. d., Schwerz, F., Carvalho, L. G. d., Baptista, V. B. d. S., Marin, D. B., Ferraz, G. A. e. S., Rossi, G., Conti, L., & Bambi, G. (2023). Machine Learning and Conventional Methods for Reference Evapotranspiration Estimation Using Limited-Climatic-Data Scenarios. Agronomy, 13(9), 2366.
Sharma, S., Hamal, K., Khadka, N., Ali, M., Subedi, M., Hussain, G., Ehsan, M. A., Saeed, S., & Dawadi, B. (2021). Projected drought conditions over southern slope of the central Himalaya using CMIP6 models. Earth Systems and Environment, 5, p.849~859. 
Simolo, C., Brunetti, M., Maugeri, M., & Nanni, T. (2010). Improving estimation of missing values in daily precipitation series by a probability density function‐preserving approach. International Journal of Climatology, 30(10), p. 1564~1576. 
Suleiman, A. A., & Hoogenboom, G. (2007). Comparison of Priestley-Taylor and FAO-56 Penman-Monteith for daily reference evapotranspiration estimation in Georgia. Journal of Irrigation and Drainage Engineering, 133(2), p.175~182. 
Trajkovic, S. (2007). Hargreaves versus Penman-Monteith under humid conditions. Journal of Irrigation and Drainage Engineering, 133(1), p.38~42. 
Valle Júnior, L. C. G. d., Vourlitis, G. L., Curado, L. F. A., Palácios, R. d. S., Nogueira, J. d. S., Lobo, F. d. A., Islam, A. R. M. T., & Rodrigues, T. R. (2021). Evaluation of FAO-56 procedures for estimating reference evapotranspiration using missing climatic data for a Brazilian tropical savanna. Water, 13(13), 1763. 
Vishwakarma, D. K., Pandey, K., Kaur, A., Kushwaha, N., Kumar, R., Ali, R., Elbeltagi, A., & Kuriqi, A. (2022). Methods to estimate evapotranspiration in humid and subtropical climate conditions. Agricultural Water Management, 261, 107378.
Wanniarachchi, S., & Sarukkalige, R. (2022). A review on evapotranspiration estimation in agricultural water management: Past, present, and future. Hydrology, 9(7), 123. 
Westerhoff, R. (2015). Using uncertainty of Penman and Penman–Monteith methods in combined satellite and ground-based evapotranspiration estimates. Remote Sensing of Environment, 169, p.102~112. 
Wheeler, T., & Von Braun, J. (2013). Climate change impacts on global food security. Science, 341(6145), p.508~513. 
Xia, X., Zhu, X., Pan, Y., & Zhang, J. (2020). A monthly regression correction model for the Hargreaves–Samani method in Mainland China. Irrigation and Drainage, 69(4), p.880~890.
Xiang, K., Li, Y., Horton, R., & Feng, H. (2020). Similarity and difference of potential evapotranspiration and reference crop evapotranspiration–a review. Agricultural Water Management, 232, 106043.
Zaheer, R., & Shaziya, H. (2019). A study of the optimization algorithms in deep learning. In 2019 third international conference on inventive systems and control (ICISC). IEEE, p.536~539.
Zongo, B., Barbier, B., Diarra, A., Zorom, M., Atewamba, C., Combary, O. S., Ouédraogo, S., Toé, P., Hamma, Y., & Dogot, T. (2022). Economic analysis and food security contribution of supplemental irrigation and farm ponds: evidence from northern Burkina Faso. Agriculture & Food Security, 11(1), 4. 
Zounemat-Kermani, M., Batelaan, O., Fadaee, M., & Hinkelmann, R. (2021). Ensemble machine learning paradigms in hydrology: A review. Journal of Hydrology, 598, 126266. 
甘俊二、陳清田、陳焜耀(1996)，台灣地區作物需水量推估模式之合適性研究，農業工程學報，第42卷，第2期，p. 8~19。
行政院內政部，(2024)，人口(國情簡介-人民)
https://www.ey.gov.tw/state/99B2E89521FC31E1/835a4dc2-2c2d-4ee0-9a36-a0629a5de9f0。
國家災害防救科技中心(2020)，臺灣乾旱災害特性，氣候變遷災害風險調適平台
https://dra.ncdr.nat.gov.tw/Frontend/Disaster/RiskDetail/BAL0000022。
陳憲宗、李奕欣(2016)，台灣長期皿蒸發量趨勢分析及蒸發互補關係初探，農業工程學報，第 62 卷，第1期，p.12~28。
許晃雄、王嘉琪、陳正達、李明旭、詹士樑，(2024)，國家氣候變遷科學報告2024：現象、衝擊與調適，國家科學及技術委員會與環境部聯合出版。
張斐章、 張麗秋(2014)，類神經網路導論：原理與應用，滄海圖書資訊出版。
葉信富、李振誥、陳忠偉、張格綸(2008)，評估蒸發皿係數以推估台灣南部地區蒸發散量之研究，農業工程學報，第54卷，第3期，p.27~35。
葉信富、林宏奕、李振誥(2013)，台灣地區蒸發散量及皿蒸發量時空分布之評估， 農業工程學報，第59卷，第3期， p.13~21。
經濟部水利署(2022)，台灣年降雨量，水利電子報，第0515期https://epaper.wra.gov.tw/Article_Detail.aspx?s=8013&n=30173。
經濟部水利署(2022)，111年各標的用水概況，各項用水統計資料庫
https://wuss.wra.gov.tw/annuals.aspx#a。
經濟部水利署，(2024)，水利統計簡訊
https://www.wra.gov.tw/News.aspx?n=2959&sms=9058。
臺灣氣候變遷推估資訊與調適知識平台，(2024)，統計降尺度資料
https://tccip.ncdr.nat.gov.tw/tr_02.aspx?category_id=C6&faq_id=18。