土壤乃集水區地表的基質，土壤厚度更常為各種水文、邊坡模式所需的基本 資訊，然而臺灣山區的土壤厚度資料卻極為不足，亟需有效土壤厚度估測模式的 發展。由於地形扮演著控制各種邊坡作用的主要角色，其與土壤化育的結果和表 徵密切相關，因此本研究以地形作用模式為基礎，結合土壤生成函數 (soil production function) 與擴散模式 (diffusion model) 建構土壤厚度的估測模式，並 以陽明山國家公園七星山地區作為試驗區，透過數值高程模型 (DEM) 來進行土 壤厚度的數值模擬。 本研究在假設模擬其間地表形貌無明顯改變，試驗區邊坡土壤的搬運得以簡 單潛移 (simple creep) 加以描述，且搬運速率在稜線、山脊等坡頂處與土壤生成 速率維持局部穩定平衡狀態 (local steady state) 的條件下，率定土壤厚度估測模 式所需參數。試驗區的模擬土壤厚度約在0 至2 公尺之間，主要受到地形曲率控 制，呈現高度空間變異性。野外實測的土壤厚度在凸坡、平緩處與模式估計值較 為吻合，在坡度較陡、集水面積較大的區域則呈現較大的估計誤差，此證實其他 邊坡作用在此等地形區的影響。整體而言，利用局部穩定平衡假設所率定的參 數，可以有效估測以潛移為主要地形作用地區的邊坡土壤厚度。對於顯著受到其 他邊坡作用影響地區，其土壤厚度的估計，則須進一步釐清各種作用的物理機 制，並將其納入考量，方能有效進行模式推估。

Soil thickness is one of the fundamental components in many hydrological and slope stability models. Since landform critically influences many slope processes, soil thickness which is related to various slope processes is also highly affected by topography. In this study, a soil production function is coupled with a simple diffusion model to form a soil-thickness prediction model. The Chiching Shan area in the Yang-Ming-Shan National Park is chosen as a study area to test the validity of the model. It is assumed that the topography has not had any drastic change during the simulation time period, the soil is moving downhill according to simple creep law, which sets the transport flux equal to a linear function of local slope gradient, and its transport rate is in equilibrium with the soil production rate at ridge lines. Model parameters are calibrated under this local steady-state assumption, and used to estimate the soil thickness within the study area. The simulated soil depths in the study area range from 0 to 2 meters, and their spatial variations are controlled by topographic curvature. Field verification shows that the model performs well on the upper parts of hillslope where low gradient convex slope forms dominate. It is discovered that in area where slope processes other than diffusion prevail, significant disparity appears between model simulated and the measured soil depth. Generally speaking, the adoption of the local steady state assumption facilitates parameter calibration and model application. However, in order to further improve the model’s applicability, mechanisms involved in other hillslope processes must be incorporated in future model development.