以坡地地下水模擬為目的之水文過程演算法與參數化研究

Abstract

本研究參與坡地社區自主防災計畫，主要工作為提供坡地穩定分析所需之坡地地下水水位，方法是使用平面二維坡面寬度函數和垂直二維變飽和度地下水模式，替代三維坡地的地下水水位的耗時模擬。因為案例現場的地文、水文資料不足，因此研究重點是放在解讀思考地文、水文與土壤的有限數據，研擬模擬模式輸入和參數化所需要的坡地水文過程的分析計算與參數化方法；同樣因為資料不足，故無法檢定優化與驗證。 首先，使用平滑化地表高程的方法估計坡地地下水位和流線，劃分坡地模擬範圍；再透過側入流流量面積化與坡面寬度等價化方法，簡化模擬範圍內側入流的影響，決定平面二維坡面寬度的擬合函數。簡化坡地土壤水力特性地參數化過程中，使用單點採樣擾動土壤進行土壤試驗，採用單指數模型描述坡地土壤水力參數受風化與生物活動影響的垂直分佈。經過分析思考，認為案例中，坡地潛移和地中傾斜管地下水位歷線變化，推估監測井地下水水位已受到滑動面與坡腳溯源侵蝕之影響，並針對此現象提出土壤水力參數估計與最佳化之合理策略。 計算變飽和度地下水模式的地表邊界條件時，首先，將強降雨的小時平均雨量進行時間分配，求得時間解析度較高、降雨強度不同的組合；其次，再使用以計算植被與建築物截留、窪蓄損失為目標的「變更的SCS-CN法」，得到扣除截留、窪蓄、排水系統損失的「供應入滲的有效降雨」，供應垂直二維變飽和度地下水模式的地表邊界流量條件。計算初始條件和下邊界條件時，使用Dupuit假設估計垂直二維模式的地下水位，剖面底部的岩盤深層滲漏量(Deep Percolation)分佈，必須與使用Laplace Equation求解的深層地下水模式，迭代達到一致性。 根據Dupuit假設與穩態降雨估計穩態水分分佈，並模擬降雨入滲些許時間，以降低無法描述未飽和層的Dupuit假設，與Richards' Equation求解變飽和度地下水模式，兩者不同構成的Spin-up problem，及其對模式模擬造成的負面影響等。最後，再自然退水至近穩態狀態並作為模擬的初始條件，並以初始條件的模擬結果進行土壤水力參數與邊界條件的再調整，再模擬坡地地下水水位。 經比較模擬結果與歷史水位，說明方法論的合理性。最後，構思和說明，未來若參數資料充足的條件下，針對方法論中提及的各項假設之檢驗方法與建議。

This study participates the Slopeland Community Voluntary Disaster Prevention Project and provides the groundwater simulation for the stability analysis of slopelands as the main task. Owing to the insufficiency of the field geographical and hydrological data, the study proposes a methodology for parameterizing and analyzing the slopeland hydrologic processes with the field data of the geography, hydrology and soil. The study utilizes the couple model of the vertical two-dimensional various saturation model and the horizontal two-dimensional slopeland width function to simulate the three-dimensional groundwater. The study area is defined with the concept of stream line and smoothed elevation. Furthermore, the study simplifies the influence of the lateral inflow of the slopeland unit by transforming the lateral flow to an area and equivalent slope width and fitting the slope width function. To simplified the parameterizing process of soil parameters, the study applies the single exponential decay model to illustrate the influences of the weathering soil and biological activities on the soil characteristics and the slopeland soil hydraulic parameters distribution. By analyzing the monitoring data of the groundwater and the slope displacement, the study estimates the groundwater is impacted by the sliding surface and the head erosion from the slope foot. For applying the couple model, this study redistributes the duration of the rainfall intensity with certain ratio to obtain the rainfall data with higher temporal resolution. Besides, the SCS method is applied to estimate the rainfall loss of the surface interception unit and draining system and to calculate the effective precipitation for infiltration. The deep percolation distribution of the bedrock at the bottom boundary is estimated by the couple model of the Dupuit assumption and the Laplace Equation. The steady state distribution of groundwater is estimated by the Duapuit assumption and steady state precipitation. In order to provide reasonable initial condition efficiently with reduced Spin-up problem caused by the physical assumption difference between the Richards' Equaiton and Dupuit assumption, the study simulates the infiltration with the estimated steady state distribution of groundwater and precipitation. Afterward, the study simulates the recession until the distribution of groundwater approaches the steady state again and utilizes this distribution as the initial condition. According to results of the initial condition, the soil hydraulic parameters and boundary conditions are re-adjusted to obtain the new initial condition and utilized to simulate the distribution of groundwater in study area. In the end, the rationality of the methodology can be described by comparing the results and historical groundwater data. Moreover, the study illustrates the method for calibrating the parameter and the assumption in the methodology with the sufficient data.