本文主要研究土壤降雨入滲動態過程中的水、熱及地球物理等不同物理機制的交互作用，並估計農田土壤降雨入滲過程中的相關參數。土壤含水量和溫度等物理量，對於降雨入滲的相關機制，具有相當重要的角色。本研究結合耦合水力與熱傳數值模式以及地電阻資料，建立自動化耦合參數估計流程，在參數最佳化過程中，利用研究場址現地的土壤含水量、溫度和視阻抗等觀測數據，校驗模型結果並由模擬與觀測量兩者建立目標函式來進行參數最佳化控制，以進行地下水文、熱傳和地球物理參數的估計。結果顯示，此方法可以很好地估計出水、熱參數，並且也能夠量化參數估計結果的不確定性，模擬的土壤水分和溫度與實際測量數據非常接近，顯示本研究所估計的水文參數具有非常高可信度，可以同時解釋多種物理獨立觀測量，但視電阻率數據的擬合效果不如前兩者。造成電阻率擬合度相對不佳的原因可能來自於測量原理及空間範圍的差異、最佳化過程中使用的一維層狀模型以及水-電關係式參數簡化對於描述真實自然界物理機制的誤差等，最終能夠透過觀測資料校驗模型，並且利用最佳參數進行降雨入滲過程中的水流、熱傳及地球物理等不同物理量的動態行為模擬。本研究除了提供了地下環境中不同物理機制耦合的交互作用理論架構，也強調了此方法需要進一步研究以提高參數估計的準確性和對複雜物理交互作用的理解，更幫助我們更進一步了解降雨入滲作用對於土壤的多重物理機制。;This study primarily investigates the interactive effects of different physical mechanisms, such as water, heat, and geophysics, during the dynamic process of soil rainfall infiltration, and estimates relevant parameters during the infiltration process in agricultural soil. This study built a parameter optimization scheme to estimate the subsurface hydrological, thermal, and geophysical parameters by using soil moisture, temperature, and apparent resistivity observation data. In this optimization procedure, the real measured datasets collected in the study site were used to calibrate the modeling results and to constrain the optimization process by the objective function. The results showed that the controlling parameters can be well estimated and the uncertainty was also realized after the inversion. The simulated soil moisture and temperature were very close to the real measured datasets. However, the apparent resistivity data fitting is not as good as the previous two. Reasons for poor fitting of electrical data may be due to measurement limitations, the use of a one-dimensional model, and an incomplete description of the petrophysical relationship. The dynamic behaviors of various physical quantities such as water flow, heat transfer, and geophysics during the rainfall infiltration process can be finally simulated using optimal parameters. This research provides valuable insights into the coupling of different physical processes in the underground environment and highlights the need for further studies to improve the accuracy of parameter estimation and the understanding of the complex interactions between different physical quantities.
