Key Laboratory of Watershed Geographic Sciences, Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 210008, China; School of Biological and Environmental Science, University of Stirling, Stirling FK9 4LA, United Kingdom.
State Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou 310012, China.
Water Res. 2024 Feb 15;250:121034. doi: 10.1016/j.watres.2023.121034. Epub 2023 Dec 19.
Remote sensing monitoring of particulate organic carbon (POC) concentration is essential for understanding phytoplankton productivity, carbon storage, and water quality in global lakes. Some algorithms have been proposed, but only for regional eutrophic lakes. Based on in-situ data (N = 1269) in 49 lakes across China, we developed a blended POC algorithm by distinguishing Type-I and Type-II waters. Compared to Type-I, Type-II waters had higher reflectance peak around 560 nm (>0.0125 sr) and mean POC (4.65 ± 4.11 vs. 2.66 ± 3.37 mg/L). Furthermore, because POC was highly related to algal production (r = 0.85), a three-band index (R = 0.65) and the phytoplankton fluorescence peak height (R = 0.63) were adopted to estimate POC in Type-I and Type-II waters, respectively. The novel algorithm got a mean absolute percent difference (MAPD) of 35.93 % and outperformed three state-of-the-art formulas with MAPD values of 40.56-76.42 %. Then, the novel algorithm was applied to OLCI/Sentinel-3 imagery, and we first obtained a national map of POC in 450 Chinese lakes (> 20 km), which presented an apparent spatial pattern of "low in the west and high in the east". In brief, water classification should be considered when remotely monitoring lake POC concentration over a large area. Moreover, a process-oriented method is required when calculating water column POC storage from satellite-derived POC concentrations in type-II waters. Our results contribute substantially to advancing the dynamic observation of the lake carbon cycle using satellite data.
中国 49 个湖泊的 1269 个原位数据表明，需区分 I 型和 II 型水体来构建适用于中国大面积湖泊的混合 POC 算法。与 I 型水体相比，II 型水体在 560nm 附近的反射率峰值更高（>0.0125 sr），POC 均值更高（4.65±4.11 vs. 2.66±3.37mg/L）。此外，由于 POC 与藻类生产力高度相关（r=0.85），因此分别采用三波段指数（R=0.65）和浮游植物荧光峰高（R=0.63）来估算 I 型和 II 型水体中的 POC。新算法的平均绝对百分比误差（MAPD）为 35.93%，优于 MAPD 值为 40.56-76.42%的三种现有公式。然后，将新算法应用于 OLCI/Sentinel-3 图像，首次获得了中国 450 个湖泊（>20km）的全国 POC 分布图，呈现出“西部低、东部高”的明显空间格局。总之，大面积遥感监测湖泊 POC 浓度时应考虑水体分类，并且在计算 II 型水体中卫星反演 POC 浓度的水柱 POC 储量时，需要采用面向过程的方法。本研究结果为利用卫星数据进行湖泊碳循环动态观测提供了重要依据。
