結合MODIS與Formosat-2光學影像建立高屏河口及沿岸水體高時空分辨率總懸浮物質濃度估算公式

Integration of MODIS and Formosat-2 Imagery for the Development of a Reliable and High-Tempospatial-Resolution Total Suspended Matter Concentration Retrieval Model: Case Study in Goaping River Mouth

10.6574/JPRS.2013.17(1).4

楊凱傑(Kai-Jie Yang)；張智華(Chih-Hua Chang)；劉正千(Cheng-Chien Liu)；邱耀勝(Yao-Sheng Chiu)；劉祖乾(James T. Liu)

總懸浮物質 ； 福衛二號 ； MODIS ； 水質 ； 雙衛星遙測 ； Total Suspended Matter ； MODIS ； Formosat-2 ； Remote Sensing ； Multi-satellite

航測及遙測學刊

17卷1期（2013 / 03 / 01）

53 - 65

繁體中文

沿岸水體常含有大量懸浮物質(Total Suspended Matter, TSM)，主要來自陸源沖淡水團的輸送及海底沉積物的再揚起。來自陸源輸送的TSM水團因河川流量、潮汐及沿岸流場之作用使其濃度之時空分佈特性明顯，高濃度區會有水質惡化的現象，並影響海洋生態、漁業及珊瑚礁的生長。鑑於傳統監測無法充分掌握TSM濃度時空分佈特性之缺點，本研究利用Terra及Aqua衛星搭載Moderate Resolution Imaging Spectroradiometer (MODIS)感測器之多頻譜優勢，以及福衛二號(Formosat-2, FS-2)光學影像之高時空分辨率特性，以雙衛星資料建立高屏河口之TSM濃度估算公式。研究運用8幅Aqua-MODIS及Terra-MODIS影像(Band 1)及與其對應之高屏河口TSM船測數據建立250 m解析度MODIS-TSM估算公式(R^2=0.84)。其次，利用2幅MODIS-TSM所得之高低濃度TSM影像做為地真資料，將與其對應之FS-2影像以MODIS大氣參數與FLAASH模式進行大氣校正獲致海面反射率後，建立8 m解析度FS2-TSM估算公式(R2=0.86)。最後，以2011/6/7海研三號CR1542航次採得之河口到小琉球6點TSM資料進行FS2-TSM驗證。受到大氣與驗證濃度偏低的影響，驗證結果雖不理想，但福衛二號呈現之TSM空間分佈趨勢仍與實測有高度相關(R=0.8)。本研究建立之福衛二號高分辨率TSM濃度估算公式，具有多頻譜影像如MODIS推估水質濃度之可靠度，又具有高空間解析度，就遙測於民生及水資源用途而言，可提供比傳統監測更大範圍及近即時的TSM資訊，並補齊MODIS水色產品所欠缺之河口沖淡水團分佈細節。

Total Suspended Matter (TSM) concentration in nearshore waters plays a very important role to the coastal water quality management, oceanic primary production and coral reef conservation. However, traditional approach provides limited information in terms of spatial and temporal coverage for monitoring TSM concentration in nearshore waters. The MODerate resolution Imaging Spectroradiometers (MODIS) on board Terra and Aqua satellite can provide abundant spectral information for rigorous atmospheric correction and retrieving TSM concentrations from the observation of ocean color in a large scale (250 m resolution). With the advantage of daily-revisit characteristics of Formosat-2 satellite (FS-2), the Remote Sensing Instrument (RSI) onboard FS-2 provide high-tempospatial data (8 m resolution), however, insufficient spectral information to resolve the water color. With an intention to well understand the dynamic and spatial patterns of TSM concentration in the Goaping river, this study presents a multi-satellite sensor approach that integrate both the advantage of MODIS and RSI to generate high resolution map of TSM concentration on FS-2 imagery. Firstly, a 250m resolution TSM regression model named MODIS-GP-TSM was established using MODIS Band 1 reflectance and concurrent TSM concentration measured at Gaoping River mouth. Secondly, with the TSM data derived from two MODIS images and MODIS-GP-TSM, an 8m resolution TSM retrieval model named FS2-GP-TSM was developed on the corresponding FS-2 images. The FS2-GP-TSM was verified to another set of field data taken from the river mouth to Lamey Island, with a good agreement between derived and observed TSM concentrations. This study encourages the use of multi-satellite approaches to extend the application of the FS-2 imagery on monitoring costal and nearshore waters where high resolution information is required.