隨著科技及電腦運算資源快速的進展，大氣及水文模式已能夠模擬更高解析度的現象，更突顯出地表特性的重要性如：地形、土地利用形態、植被覆蓋率及土壤質地等，地表可藉由能量及水文的循環和儲存影響短期天氣的事件及長期氣候的變異。土壤中的水分改變地表能量的收支，並影響著數值天氣預報的表現，土壤濕度與地表水文過程息息相關，並受到地形、土地利用及土壤質地空間異質性的分佈而有所改變。因此在時間及空間上對於土壤濕度的觀測及模擬，具有很高的挑戰性。此篇研究主要目的，在於瞭解地表水文過程對土壤溼度及大氣模擬的影響，並改善模式土壤初始化、土壤形態及次網格解析度的物理水文過程。 本研究使用WRF氣象模式，探討次網格水文過程及土壤濕度與大氣間的交互作用，並主要對於2013年8月29日至9月11日的個案進行討論。首先為了解並改善土壤初始場, 使用全球地表同化系統(Global Land Data Assimilation System, GLDAS) 0.25度的土壤溫度及土壤濕度來取代原先提供自再分析場的資料做為初始值。GLDAS相較於再分析場的資料提供了較乾的土濕，也與觀測值較為接近，透過近地表潛熱與可感熱的重新分配，進而影響溫度變化、邊界層、低層雲的發展，並調節了區域風場的結構。模擬的土壤濕度空間分佈也與給定的土壤質地形態的相關土壤參數有關，因此進一步針對台灣土壤質地分類進行更新，藉由野外實地的調查資料，呈現出更為接近真實的土壤形態空間分佈，結果顯示土壤濕度在空間上因土壤粒徑大小及水分傳導速率的不同而有所變化。 另外，目前WRF-Noah模式的架構下，有以下幾項缺點：(1)土壤水分只考慮了垂直的輸送作用;(2)地表逕流只考量了超入滲過程，缺乏過飽和逕流的演算機制、地表積水的再入滲過程，並且未考量飽和地下水的土壤結構。對於台灣陡峭的地形而言，大部份降雨沿山區流入平地，在高解析度的模擬下，由地形產生水平方向水分的交換，在局部水文收支平衡中，是不可忽視的。因此，我們建置WRF大氣-地表-水文雙向耦合模式系統，以改進模式次網格解析度的逕流過程，以及近地表水文過程。由於側向水的流動，土壤水分在山區減少，平地則有所增加，並進而改變地表與低層大氣間的反饋作用。本研究闡明地表水文過程和土壤濕度的重要性及其對大氣的交互作用。 ;Heat and water storages of soil conditions memorize the signals of atmospheric information including short-term weather events and long-term climate anomalies. The amount of soil water is critical in determining the performance of numerical weather predictions, and affects the surface energy budgets. The spatial patterns of soil moisture are not only the results of land surface hydrological processes including precipitation, evapotranspiration, groundwater, and surface and subsurface runoff processes, but also are affected by the heterogeneity of topography, soil properties, and land cover characteristics as well. In order to present the effects of land surface hydrological processes on soil moisture and its feedback on the meteorological characteristics, the two-way coupled WRF-Hydro modeling system is applied from 29 August to 11 September 2013. Due to the topography-induced lateral transport of overland flow and saturated subsurface flow, the results from WRF-Hydro model show the increase of soil moisture content in the low-lying areas, while the soil moisture content is reduced over the mountainous regions, and consequently inducing hydrometeorological responses. In addition, the better initial soil states from Global Land Data Assimilation System (GLDAS) and updated soil textures based on field investigations over the Taiwan areas are used to improve the simulations and explore the effects of the GLDAS analysis on the atmospheric model prediction. The soil initialization from GLDAS products can better reproduce the amplitude of temporal variations with observations. Simulated results reveal the spatial distribution of soil moisture is extremely relevant to the soil properties of soil types prescribed in WRF model and surface water budget is analyzed to understand the physical processes. These results show the improvement and heterogeneity of simulated soil moisture though the better representation of initial soil states, soil physical and hydraulic properties, and sub-grid scale hydrological responses. This study sheds light on the importance of fine-scale hydrological processes on soil moisture and its subsequent impact on the coupling of soil moisture-atmosphere interactions in Taiwan.
