土壤含水量是水文循環中的重要參數，而地表濕度的空間變化對於地表交互作用之解析至關重要。然而，地面觀測的空間分布通常是受限的，遙測技術配合衛星觀測具有提供全球和區域尺度的表面土壤濕度之潛能，但目前大多數的土壤濕度產品僅具有全球尺度或低空間解析的被動或主動-被動微波感測器所反演提供。另一方面，光學感測器可針對區域尺度以更高空間解析提供與土壤濕度相關的溫度植升乾燥指數（TVDI）來估算地表乾燥程度。根據之前的研究，大多數 TVDI是針對特定或單一地物種類/土地利用的現場測量來評估的，而大範圍區域土壤濕度的空間變化則需各地物種類/土地利用之訊息，例如裸地、草地、灌木地和林地覆蓋。因此，本研究提出基於主動-被動微波反演之土壤濕度產品（SMAP），評估每各地物種類/土地利用所對應光學感測器 MODIS 的 TVDI。整體結果顯示，MODIS TVDI 與 SMAP 土壤濕度的空間分布非常吻合，且明顯受地表溫度變化之影響。在本研究提出的精細修正後，大多數土地覆蓋類型的土壤濕度與 TVDI 的相關性有顯著的提升，R = >0.50，尤其是裸土區域，R = 0.70。根據 MODIS TVDI 與 SMAP 土壤濕度之相關性，應用 TVDI 在土壤濕度的估算獲得不錯之成果，尤其是在植被稀疏的土地覆蓋類型。;Soil moisture plays an important parameter to understand the various environmental phenomenon, especially in hydrological cycles. Spatial variability of the surface soil moisture condition is essential to a more comprehensive understanding of the surface interactions. However, the spatial information from ground-based observation is generally limited. Remote sensing technology could potentially provide a global and regional scale of surface soil moisture. Most soil moisture products in metrics unit (volumetric unit) are usually provided by passive or active-passive microwave sensor which has global scale or low spatial resolution. Optical sensor remote sensing can perform surface dryness index in terms of Temperature Vegetation Dryness Index (TVDI) related to the soil moisture condition in regional scale or finer spatial resolution. According to the previous study, most of the TVDI results were evaluated by in-situ measurements for specific land cover types. The spatial variation of soil moisture within the entire region would necessitate in each land cover/land use, such as bare land, grassland, shrubland, and forest land covers. Therefore, this research proposed to evaluate the TVDI from MODIS satellite imagery for each land cover type based on active-passive microwave remote sensing of soil moisture products, i.e. Soil Moisture Active-Passive (SMAP). The overall results show the spatial distribution of TVDI quite similar with soil moisture information from SMAP and heavily influenced by the land surface temperature variation of the landcover type. After the refined correlation proposed in this study, soil moisture in most of the landcover types significantly correlated with TVDI with R = >0.50, especially in bare soil with R = 0.70. The estimated soil moisture based on TVDI represents the soil moisture quite well, especially in the early dry season. The uncertainty on the high dense vegetation canopy was included to estimate the soil moisture in the shallow soil depth and the bare soil with sparse vegetation landcover type performed a better estimation due to less uncertainty. However, this research could provide the soil moisture content in 1 km spatial resolution based on optical sensor and the possibility to achieve finer spatial resolution.
