臺灣山區表土層厚度推估方法

本研究以地形因子為基礎，探討各因子與表土層(風化殘留土層及崩積層)厚度之關係。本研究蒐集全臺灣之鑽探及土鑽資料總計1199處，其中鑽探330處及土鑽869處。經研究比對後，篩選出鑽探287處及土鑽853處表土及崩積層資料與地形因子進行分析。前者之鑽探資料應用數值地型(DEM)及崩塌目錄進行微地形判釋，分為擾動地區198處及未擾動地區89處。擾動地區指該案例取樣位置位於崩塌地或已具有崩崖、滑動體或明顯滑動面等特徵之邊坡；未擾動地區則代表該案例在地形上並未發現崩塌潛勢特徵，且岩盤較為完整，無滑動面產生，地層較穩定者。最後，本研究建立臺灣本土之山區表土層厚度推估模式。首先，本研究利用坡度、高程、曲率及地形濕度指數等因子，與前述三類鑽探之表土層厚度進行分析。研究成果顯示在未擾動區鑽探之表土層平均厚度與坡度因子之相關性佳，R2達0.51；土鑽者之平均表土層厚度與坡度因子之相關性最佳， R2達0.88；而在擾動區鑽探之崩積層厚度則與高程差因子得出之相關性最佳，此處之高程差因子是採用距離冠部之高程計算，取平均值後R2達0.73。綜整研究成果，本研究進一步建立臺灣山區表土層厚度推估方法。首先，未擾動區之表土層厚度推估公式如下。其中x為坡度，y為崩積土層厚度。 1.崩積土層: y = 9.3824e-0.038x，R2=0.57。 (1)砂土: y = 8.7078e-0.043x，R2=0.73。 (2)粉土: y = 10.188e-0.034x，R2=0.7。 2. 風化殘留土 (1)黏土: y = 10.158e-0.102x，R2=0.41。其次，山區擾動區之崩積層厚度推估公式如下。其中x為距冠部高程，y為崩積層厚度。 1.頭部區: y = 0.1685x，R2= 0.54。 2.主崩體: y = 0.0565x，R2= 0.4。 3.坡腳區: y = 0.0683x，R2 = 0.49。最後，本研究選以高雄六龜山區三處邊坡進行現地表土厚度量測，驗證實測及推估厚度結果顯示，表土層厚度迴歸式分析結果為最佳。其中風化殘留土層迴歸式為y = 0.8843e-0.015x；崩積層迴歸式為y = 9.4514e-0.05x。比對結果差值有約80％為0.5m以內，最小0m，最大為1.1m，平均誤差為0.3m。

關鍵字

地形因子；表土層；崩積層

並列摘要

This study explores the relationship between geomorphological characteristics and topsoil thickness, the topsoil layer composed of residual soil and colluvium. This study collected 1199 geologic drilling data in the mountain region of Taiwan, which contained 330 geologic boring logs and 869 earth auger data. A total of 287 geologic drilling data and 853 auger data were verified and adopted in the correlation analysis with geomorphological factors. The geologic boring log data was further classified into 198 cases in the disturbed area and 89 cases in the undisturbed area based on the Digital Elevation Model (DEM) and landslide inventory. The disturbed area indicated that the site was located in the landslide or the hillslopes with features such as cliffs, sliding mass, or slip surface. The undisturbed area means no landslide susceptibility and evidence in the site, and the bedrock is relatively complete. The study finally established an estimation model for the thickness of topsoil and colluvium in the mountain region of Taiwan. The study conducts the geomorphology analysis with four factors, including Slope Gradient, Elevation, Curvature, and Topographic Wetness Index, to correlate to the topsoil thickness. Results show that the average topsoil thickness and the slope factor achieved the best correlation in the undisturbed area of geologic drilling data, R2 up to 0.51. Average topsoil thickness from the earth auger data also gained an excellent correlation to the slope factor, R2 up to 0.88. Meanwhile, the average colluvium thickness in the disturbed area showed a good correlation to the elevation factor. The elevation factor calculates the elevation difference between the drill location from the crown, R2 up to 0.73. At last, this study establishes an estimation model on the topsoil thickness for Taiwan's mountainous region. The formula for estimating the thickness of the topsoil in the undisturbed area is as follows equations. Where x is the slope and y is the thickness of the topsoil in the equation. 1.Colluvial soil：y = 9.3824e-0.038x，R2 = 0.57。 (1)Sand：y = 8.7078e-0.043x，R2 = 0.73。 (2)Silt：y = 10.188e-0.034x，R2 = 0.7。 2.Residual soil (1)Clay：y = 10.158e-0.102x，R2 = 0.41。 For the case of the colluvial thickness in the disturbed area, where x is the elevation from the crown, and y is the thickness of the colluvium. 1. Head of landslide：y = 0.1685x，R2 = 0.54。 2. Main slid body：y = 0.0565x，R2 = 0.4。 3. Foot of landslide：y = 0.0683x，R2 = 0.49。 Finally, three field tests in the Liugui District of Kaohsiung County were applied to test the soil thickness estimation model. The results showed that the residual soil thickness plus colluvium thickness combination model achieved the best performance. The difference between the estimated thickness and the measured value was 70% within 0.5m, with an average of 0.49m, the minimum difference is only 0.1m, the maximum is 1.4m, and the R² up to 0.5, which shows that the effect of the estimation model is good.