不同計算模式應用在蓮華池土壤水力傳導度之比較

Comparison of different calculation model's application to hydraulic conductivity in Lienhuachih area

陳虹巧(Hung-Chiao Chen)

指導教授 ： 陳明杰

國立臺灣大學/生物資源暨農學院/森林環境暨資源學研究所/碩士(2015年)

https://doi.org/10.6342/NTU.2015.00329

摘要

本研究引用前人建立之GE模式（Gardner, 1958）、BC模式（Brooks and Corey , 1966）、GP模式（Gardner, 1965）與VGM模式（van Genuchten, 1980）四種飽和與不飽和水力傳導度關係式，應用於蓮華池地區的森林土壤對於水力傳導度隨壓力水頭變化的模擬效果。為了方便結合蓮華池地區的相關研究結果進行討論，本研究選定蓮華池四號及五號集水區作為研究對象，沿著兩集水區的山脊各選定四個試驗點，每個試驗點分為表層與20 cm兩種土壤深度，使用張力滲透計進行現地滲透試驗。比較解聯立法、迴歸分析法與數值模擬法三種不同數學方法於求解模式參數的適用性，並計算誤差評估指標RMSE供決定最佳參數解。將獲得最佳參數解的各模式模擬不同壓力水頭條件下的水力傳導度值與實測值作比較，可探討各模式是否適用於描述水力傳導度隨壓力水頭的變化趨勢。現地土壤在張力滲透計實驗後待乾燥一段時間，採取不擾動土壤試體供分析其物理性質，對照現地滲透試驗結果，以說明土壤物理性質對於水分滲透流動的影響。其次，於模式決定過程中，導入蔡彥邦（2013）使用雙環入滲計所測得之飽和水力傳導度，藉此修正模式於近飽和狀態的水力傳導度推估值。
研究結果顯示，相較於五號土壤，四號不同樣點之間的土壤物理性質呈現較高的均質性，且五號不同深度之土壤有較大差異，四號不同深度的土壤差異較小。此結果影響了現地水力傳導度之量測，張力滲透計的實驗結果顯示，四號不同深度之土壤在各壓力水頭下的水力傳導度值的分散程度皆小於五號。在飽和與不飽和水力傳導度關係模式的決定上，四號各樣點有較相近的土壤質地結構參數與飽和水力傳導度估計值，五號的變異程度則較大。此外，在重力作用大於毛細作用的較大壓力水頭條件下，由於水分快速通過部分中、大孔隙而無法有效填滿土壤孔隙，可能因而低估水力傳導度，但若試驗前日降雨量越多，表層土壤孔隙越容易被水分填滿而連續導水，水力傳導度測值較高。三種數學方法中以數值模擬法能使模式的推估值最接近實測值。而四種模式中則以GP模式與VGM模式的模擬效果較佳，誤差評估指標RMSE都可低於10-6。利用GP模式與VGM模式所得到之土壤質地結構參數介於12 m-1~36 m-1，蓮華池地區森林的土壤有良好的透水性。

關鍵字

土壤孔隙結構 ； 張力滲透計 ； 雙環入滲計 ； 蓮華池研究中心 ； 飽和與不飽和水力傳導度關係式

並列摘要

The study applied four saturated-unsaturated hydraulic conductivity models, which were Gardner exponential model (GE), Boorks and Corey model (BC), Gardner rational power model (GP) and van Genuchten-Mualem model (VGM), to the simulation of hydraulic conductivities under varied water pressure head condition. Three mathematical methods including solving the simultaneous equations, regression analysis, and numerical simulation were adopted to decide the parameters in the four models. Study area was at Lienhuachih watershed No. 4 and No. 5. Four locations were selected along ridge, respectively from the two watersheds, and each location comprised soil surface and the depth of 20 cm for field infiltration test by tension infiltrometer. As the soil became desiccated after the infiltration test, undisturbed soil samples of the infiltration test location were excavated for analyzing their physical properties, that to understand how the soil physical properties affect the measured hydraulic conductivity. Besides, the saturated hydraulic conductivities at the locations which had been measured by the double-ring infiltrometer by Yeng-Bang Tsai (2013) were used in the study to modify the hydraulic conductivities that simulated by the models under near-saturated condition.
According to the analyzed data, soil physical properties of watershed No. 4 were more homogeneous than that of watershed No. 5, and the discrepancy of different depths were also smaller than that of watershed No. 5. The data of field infiltration test showed that the dispersion of hydraulic conductivities of the soil surface and the depth of 20 cm of watershed No. 4 were both smaller than that of watershed No. 5. Therefore, at the results of establishing saturated-unsaturated hydraulic conductivity models, the parameters including soil text/structure parameter and the saturated hydraulic conductivity were more similar between the locations of watershed No. 4 than that of watershed No. 5. In field infiltration test, the condition of higher water pressure head that usually has more gravitational force but less of capillary force would make water only flow through parts of large soil pores instead of filling up whole pores that probably lead to underestimate the hydraulic conductivity. In the condition of more rainfall to the soil surface before the field infiltration test, it revealed that the soil pore could be effectively filled and therefore enhance the hydraulic conductivity.
The mathematical method of numerical simulation could get the lowest index of error, RMSE, which meant the method was the best way to estimate the parameters in four models. Besides, the RMSE results of GP model and VGM model were both lower 10-6, better than the results of GE model and BC model. According to GP model and VGM model, calculated values of the soil text/structure parameter were between 12 m-1 to 36 m-1, which meant that the soil of research site at Lienhuachih was well-structured, coming up to the analyzed results of the soil physical properties.

並列關鍵字

Double-ring infiltrometer ； Lienhuachih Research Center ； Saturated-unsaturated hydraulic conductivity models ； Soil porosity structure ； Tension infiltrometer

參考文獻

2. 洪志遠 (2007) 蓮華池集水區不飽和土壤的水力特性。國立台灣大學森林環境暨資源學系碩士論文。

3. 洪志遠、陳明杰 (2009) 應用張力滲透計於林地土壤孔隙水通量之探討。中華林學季刊42(4)：541-555。

8. 陳明杰、黃襦慧 (2006) 張力滲透計應用於蓮華池五號集水區土壤水力傳導度測定分析。中華林學季刊39(2)：207-220。

5. 陸象豫、唐凱軍、古秀宇、黃惠雪 (2000) 林業試驗所各林區氣候狀況。台灣林業科學15(3)：429-440。

7. 陳明杰、洪志遠 (2007) 蓮華池四號集水區土壤物理性質與水力傳導度的特性。台大實驗林研究報告21(4)：293-306。