Remote sensing of CDOM and DOC in alpine lakes across the Qinghai-Tibet Plateau using Sentinel-2A imagery data.
Remote sensing of CDOM and DOC in alpine lakes across the Qinghai-Tibet Plateau using Sentinel-2A imagery data.
复制标题
DOI：
10.1016/j.jenvman.2021.112231
复制
发表时间：
2021-03
影响因子：
8.7
通讯作者：
Ge Liu;Sijia Li;K. Song;Xiang Wang;Z. Wen;T. Kutser;Pierre-Andrew Jacinthe;Y. Shang;L. Lyu;C. Fang;Ying Yang;Qian Yang;Baohua Zhang;Shuai Cheng;Junbin Hou
中科院分区：
文献类型：
--
作者：
Ge Liu;Sijia Li;K. Song;Xiang Wang;Z. Wen;T. Kutser;Pierre-Andrew Jacinthe;Y. Shang;L. Lyu;C. Fang;Ying Yang;Qian Yang;Baohua Zhang;Shuai Cheng;Junbin Hou
有色溶解有机质（CDOM）和溶解有机碳（DOC）作为水环境中溶解有机质（DOM）的重要组成部分，在内陆水生系统的碳循环中发挥着重要作用。传统上，内陆沃茨中的CDOM和DOC主要通过现场观测和实验室测量来确定。过去对CDOM和DOC的大部分湖泊调查都集中在容易到达的区域，覆盖了全球一小部分湖泊。据我们所知，在青藏高原（QTP）等交通不便的地区，对湖泊知之甚少。为了应对这一挑战，光学卫星遥感可能是有用的，捕捉一个天气视图的CDOM和DOC的高频率在大尺度上，补充in-situs方法为内陆沃茨。在这项研究中，从QTP（2014-2017）的36个湖泊中收集了216个样本，以确定CDOM在350 nm处的吸收系数（a350）与Sentinel-2A多光谱仪器（MSI）图像反射率数据之间的关系。A350与B4/B2（R2= 0.78，p < 0.01）和B4/B3（R2= 0.62）呈强线性正相关。以B_4/B_2和B_4/B_3为输入变量，建立了多步回归模型（R_2 = 0.81，p < 0.01）。CDOM-DOC的相关性呈分散性（R2= 0.34，p < 0.05）。通过将内陆沃茨按盐度梯度分为4组，我们能够建立更强的CDOM-DOC相关性（R2> 0.8，p < 0.01）。利用卫星获取的CDOM和DOC，证明了淡水和咸水沃茨之间的显著差异，（0.86 ± 0.67 m-1）和低DOC在淡水中观测到的浓度为（3.76 ± 4.92 mg L−1），而青藏高原盐湖的CDOM（0.53 ± 0.72 m-1）和DOC（15.76 ± 17.07 mg L-1）则呈现相反的趋势。这项研究证实，卫星光学图像可以用于监测的CDOM和DOC的湖泊的青藏高原，这是敏感的气候变化，很少调查，由于恶劣的环境和交通不便。此外，它强调了在估算湖泊DOC的过程中结合盐度和遥感数据的重要性。
As important components of dissolved organic matter (DOM) in an aquatic environment, colored DOM (CDOM) and dissolved organic carbon (DOC) play an essential role in the carbon cycle of an inland aquatic system. Traditionally, CDOM and DOC in inland waters have been primarily determined usingin situobservations and laboratory measurements. Most of past lake investigations on CDOM and DOC focused on easily accessible regions and covered a small fraction of lakes worldwide. To our knowledge, little is known about lakes in less accessible areas like the Qinghai-Tibet Plateau (QTP). To address this challenge, optical satellite remote sensing might be useful for capturing a synoptic view of CDOM and DOC with high frequency at large scales, complementingin situsampling methods for inland waters. In this study, 216 samples collected from 36 lakes across the QTP (2014–2017) were examined to determine the relationships between CDOM absorption coefficient at 350 nm (a350) and Sentinel-2A Multi Spectral Instrument (MSI) imagery reflectance data. A strong positive linear correlation witha350was observed with B4/B2 (R2= 0.78,p< 0.01) and with B4/B3 (R2= 0.62). A multi-step regression model was established for estimatinga350with B4/B2 and B4/B3 as input variables (R2= 0.81,p< 0.01). A scattered CDOM-DOC relationship was revealed (R2= 0.34,p< 0.05) using a pooled dataset. By dividing the inland waters into four separate groups in accordance with their salinity gradients, we were able to develop much stronger relationships (R2> 0.8,p< 0.01) for CDOM-DOC. Significant differences between fresh and saline waters were demonstrated using satellite-derived CDOM and DOC, where high CDOM (0.86 ± 0.67 m-1) and low DOC (3.76 ± 4.92 mg L−1) concentrations were observed for freshwaters, while inverse trends of CDOM (0.53 ± 0.72 m-1) and DOC (15.76 ± 17.07 mg L−1) were demonstrated for saline lakes in the Tibetan Plateau. This study confirmed that satellite optical imagery can be used for the monitoring of CDOM and DOC of the lakes of the Tibetan Plateau, which are sensitive to a changing climate and are infrequently investigated due to the harsh environment and poor accessibility. Moreover, it highlighted the importance of combining salinity and remote sensing data in the process of estimating lake DOC.