- 期刊
蒸發散量Penman-Monteith估算方程式蒸汽壓力差計算式檢測與評估
Assessment on the Optimal Applicability of Vapor-Pressure-Deficit Calculation Methods in the Penman-Monteith Equation for Estimating Evapotranspiration
摘要
蒸汽壓力差(vapor pressure deficit, VPD)是表達大氣蒸發能力重要指標，在蒸發散量混合法或Penman-Monteith型態估算模型亦居重要成分。眾多蒸發散量估算方程式中涉及VPD之計算，由於地域性氣候環境背景特性、氣象因子選用考量、取樣資料頻率多寡、資料平均方式及資料組合方式等之不同，目前遂有超過三十種以上VPD計算式，廣爲國內外著名機構或學者推荐在世界各地應用。 本文以台灣地區各地農業氣候分區2002-2003年實際氣象觀測記錄，應用頻率分析法及統計指標法進行檢測評估。其中，頻率分析法係藉誤差、全距、優劣排序及出現頻率，並匹配權重指數予以量化方式，統計指標法係採根均方差(root mean square error, RMSE)、相關係數(coefficient of correlation, R^2)及效率係數(coefficient of efficiency, CE)三項客觀統計指標，分別進行國內外機構或學者所推荐3種型態共計26款VPD計算式，在國內本土環境應用時適用性之評估。 分析結果顯示，以頻率分析法而言，各地農業氣候分區均呈現，VPD計算式以24個時溫度日平均爲飽和蒸汽壓力，與濕球下陷量爲實際蒸汽壓力，具誤差最少、全距最小、出現頻率最多及最高權重指數，性能表現最佳，這樣的結果，對照統計指標法之檢測評估，同樣具最低RMSE、R^2值0.99以上及最大CE，性能表現優於其他所有VPD計算式，兩者，結果一致。因此，綜合上述兩項方法評估結果，本文建議24個時溫度日平均爲飽和蒸汽壓力，與濕球下陷量爲實際蒸汽壓力，堪稱最適台灣氣候環境本土化VPD計算式，可在台灣各地農業氣候分區應用，而國際性著名機構或學者所推薦VPD計算式，並不適合在台灣繼續應用。
並列摘要
Vapor pressure deficit (VPD) is a significant index for the capacity of evaporation, and also plays an important parameter for estimating evapotranspiration, particularly in the combination or Penman-Monteith type equations. Due to the difference on the environmental climate characteristic, climate variables selected, number of records sampled, averaging means and compoundable ways, there are 3 types 26 VPD calculation methods which were widespreadly used for estimating evapotranspiration around the world. Those methods for VPD calculation were analyzed and compared using 2002-2003 data from meteorological stations to determine the most appropriate method that can be applied in Taiwan. Frequency Approach (FA) and Statistical Index Approach (SIA) were used to evaluate and compare the results and applicability among these methods in Taiwan. Error, range, frequency, and weighted index (WI) were used in FA. Three objective statistical indexes, root mean square error (RMSE), coefficient of correlation (R^2), and coefficient of efficiency (CE), were used in SIA. For the FA, the results indicated that the VPD calculation using mean of 24 hourly temperature values as saturation vapor pressure, and wet-bulb depression as actual vapor pressure yielded the lowest error, the least range, and the most frequency with the highest WI. These outcomes were as well as the SIA with the lowest EMSE, R^2 higher than 0.99 and the highest CE. The performance of this VPD method was proved better than all other VPD methods. The findings were consistent with attempts at FA and SIA in this paper. Accordingly, this VPD method was recommended to apply in Taiwan. However, the other VPD calculation methods in literature were not recommended to be used in Taiwan.