| 摘要: | 由於溫室效應與人類過度開發影響下，現今世界上許多地區正面臨水資源枯竭的問題。土壤濕度與地表溫度的模擬為研究並瞭解這樣的問題的重要關鍵。現階段，本研究欲從氣候尺度（即區域尺度與長期分析）的觀點進行討論，以台灣石門水庫集水區為例，利用NOAH LSM進行五年模擬，討論地表溫度與土壤濕度變化，瞭解氣候變遷對水資源影響。 輸入模式的大氣變數場資料分為兩部分：一為地面測站資料，主要為經濟部水利署測站之日累積降雨量；另為美國Goddard Space Flight Center（GSFC）Global Land Data Assimilation System（GLDAS）所測估之近地面降雨、輻射、大氣變數場（氣溫、比溼、壓力、風速）。本研究亦使用GSFC GLDAS提供的地表參數，包括土地利用（植被種類）、土壤種類、坡度種類、地表植被覆蓋率及反照率。 在固定所有條件（大氣變數、地表參數）下，分別利用經濟部水利署提供2001至2005年測站觀測之日累積雨量與GSFC GLDAS提供的三小時雨量資料，進行兩組數值模擬實驗。利用經濟部水利署之霞雲流量站資料評估後，結果顯示使用測站日累積雨量資料所模擬之逕流量明顯較好。 受限於許多因素，本研究只能初步分析與討論模式模擬的地表溫度，然而我們需要更瞭解地表參數（如土壤特性）對於模擬地表溫度的重要性，因為模式計算水文與能量平衡都會受到影響。未來，研究應考慮地表參數的空間分佈不均勻性，改進目前評估方法。另外，若能取得更多測站以及衛星資料或是地面同化資料（地表溫資料），我們將可以進行更完整評估與改進，將模式應用於區域氣候的模擬研究。 Because of global warming and human over-development, now the gradually exhausted water resource in many region of the world is a serious problem we need to face. The simulated soil moisture and land surface temperature (LST) have being key points for studying and understanding this kind of problem. The first step of such a research, the global scale analysis (much larger spatial and temporal scale) is chosen as our study’s viewpoint, and we choose NOAH Land Surface Model (LSM) developed by NCEP to carry out five-year offline simulations subject to observed near-surface atmospheric forcing at Shih-Men reservoir watershed. Finally, we’ll investigate how climate change affects our water resource through discussing the variation of modeled LST and soil moisture. The atmosphere forcing data come from two sources: one is from local surface measurements like daily accumulative rainfall data measured by surface stations maintained by the Water Resources Agency; the other source is assimilated data produced by the GSFC GLDAS. Major variables in the GLDAS data set include near surface atmospheric parameters (air temperature, specific humidity, wind speed, and surface pressure), surface radiative fluxes (incident short-wave radiation, incident long-wave radiation) and surface rainfall rate. The GLDAS also provides land surface parameters. It includes land use type (vegetation type), soil type, land surface elevation, surface vegetation fraction, and albedo. We perform two numerical experiences, E1 and E2, in the selected region of Shih-Men reservoir watershed. The two experiments are subject to the same forcing data (2001-2005) and land surface parameters, except surface rainfall. E1 is forced by GSFC GLDAS rainfall data, and E2 is forced by station rainfall data. Results show that the simulated runoff by E2 is better than that of E1. At this stage, the evaluation of simulated LST in the current thesis is restricted by many factors, we can only investigate preliminary. We need to understand how important land surface parameters should be when simulate LST, however, since energy balance and water balance processes will be affected by land surface parameter in model calculation. For the next stage of the research, the spatial heterogeneous characteristic of land surface parameter should be considered to improve the evaluation approach. Besides, if more in situ and satellite observations data or high quality assimilation data are available, we can not only do more completed evaluation and refinement, but also apply it for regional climate modeling study. |
