利用Cryosat-2監測青藏地區湖水位變化:以重複軌道測高衛星及衛星影像驗證

Abstract

青藏高原平均海拔高於4000公尺，當地氣候和環境險峻，人煙稀少，湖泊眾多，湖泊水位受人類活動影響較少，故能真實反映近代氣候變遷，然而當地湖水位難以實地觀測，因此利用衛星測高觀測湖水位為一常用手段，如TOPEX/Poseidon(T/P)或ERS系列測高衛星廣泛應用於青藏地區湖水位監測，因此兩系列測高衛星之軌道設計關係，可觀測的青藏地區湖泊有限，但Cryosat-2具有高空間覆蓋率，故可觀測到西藏地區大部分的湖泊，其搭載新型感測器SIRAL(干涉/合成孔徑雷達測高儀)，該觀測模式不同於傳統測高衛星(T/P和ERS系列)。本研究利用Cryosat-2與SARAL/AltiKa計算青藏地區(75˚E-100˚E、28˚N-37.5˚N)之湖泊水位，其中Cryosat-2可獲得61個湖泊之水位變化(2010-2016年)；SARAL/AltiKa可獲得32個湖泊之水位變化(2013-2016年)。此外，因Cryosat-2不同cycle經過的湖泊位置不同，造成不同cycle觀測量個數多寡不一與觀測量間的大地起伏有公尺級差異，故本研究嘗試四種不同計算方式來估算Cryosat-2觀測之湖水位變化趨勢，(1)不考慮觀測量多寡與大地起伏差異、(2)考慮觀測量多寡並給予觀測量不同的權、(3)考慮大地起伏差異並進行湖面梯度修正與(4)考慮觀測量多寡與大地起伏差異，最後與SARAL/AltiKa觀測之湖水位變化趨勢進行比較。由於Cryosat-2目前僅提供6年的觀測數據，有些湖泊水位變化不明顯，單一衛星的觀測時間可能難以判斷水位變化趨勢，因此本研究利用Jason-2將Cryosat-2與ICESat之觀測資料進行串連，進而得到長達十年以上的湖水位變化。此外，本研究根據Cryosat-2的計算結果發現庫賽湖、卓乃湖和鹽湖於2011年有劇烈的水位變化，利用Landsat系列衛星影像發現湖泊間的連通並計算庫賽湖、卓乃湖和鹽湖的水體面積，結合兩者進一步估算水體體積變化後將更有利了解湖泊間的交互作用。

The mean elevation of the Qinghai-Tibet Plateau (QTP) exceeds 4000 m. Lake levels in the QTP are less affected by human activities than elsewhere, and may better reflect the state of contemporary climate change. Here ground-based lake level measurements are rare. Repeat altimeter missions, particularly TOPEX and ERS series of altimetry, have provided long-term observations of lake levels in the QTP, but their large cross-track distances allow only few lakes to be monitored. In contrast, the Cryosat-2 altimeter, equipped with the new sensor SIRAL (interferometric / synthetic aperture radar altimeter), provides a much finer spatial coverage than repeated missions, and can detect water level changes over a large number of lakes in the QTP. In this study, Cryosat-2 data are used to determine lake level changes in the QTP (75˚E-100˚E, 28˚N-37.5˚N), where Cryosat-2 provides data over 60 lakes and SARAL / AltiKa over 32 lakes from 2013 to 2016. Over a lake, Cryosat-2 in different cycles can pass through different spots, making the numbers of observations non-uniform and requiring corrections for lateral variations in geoidal undulation. Four cases are used to cope with these: (1) neglecting inconsistency in data volume and geoid-based lake slope (2) considering data volume, (3) considering lake slope only, (4) considering both data volume and lake slope. Because Cryosat-2 is available only from 2010 to 2016, Jason-2 data are used to fill gaps between the time series of Cryosat-2 and ICESat (2003-2009) to obtain >10 years of lake level changes.The Cryosat-2 result is then compared with the result from the SARAL to determine the best case. The Cryosat-2 result shows dramatic lake level changes in Lakes Kusai, Zhuoaihu and Salt in 2011.Landsat satellite imagery is used to assist the determination of volume changes over these lakes..