A Four-Band Semi-Analytical Model for Estimating Phycocyanin in Inland Waters From Simulated MERIS and OLCI Data
A Four-Band Semi-Analytical Model for Estimating Phycocyanin in Inland Waters From Simulated MERIS and OLCI Data
复制标题
用于根据模拟 MERIS 和 OLCI 数据估算内陆水域藻蓝蛋白的四波段半解析模型
DOI：
10.1109/tgrs.2017.2761996
复制
发表时间：
2018
影响因子：
8.2
通讯作者：
Shi Kun
中科院分区：
文献类型：
--
作者：
Liu Ge;Simis Stefan G. H.;Li Lin;Wang Qiao;Li Yunmei;Song Kaishan;Lyu Heng;Zheng Zhubin;Shi Kun
现有的估算浑浊内陆水域藻蓝蛋白（PC）浓度的遥感算法存在较高的不确定性，特别是在不同浮游植物群落的低PC浓度下。本文提供了一种四波段半解析算法（FBA_PC）的理论框架，该算法可将PC吸收与黄物质和其他浮游植物色素吸收引起的二阶变异性隔离。该算法适合中分辨率成像光谱仪（MERIS）和Sentinel-3海洋与陆地颜色仪（OLCI）的波段配置。该算法的校准基于美国、荷兰和中国12个内陆水体的吸收数据，并结合实验室培养物的测量结果，这表明FBA-PC的假设是对现有三波段方法的改进。对美国、荷兰和中国7个内陆水体FBA_PC的验证表明，调整为MERIS/OLCI带配置的FBA_PC与实测PC具有较好的一致性，均方误差=27.691<inline-formula> < text -math notation="LaTeX">$\text {mg} \cdot \text {m}^{-3}$ </ text -math></inline-formula>，平均绝对百分比误差= 172.863%，决定系数(<inline-formula> < text -math notation="LaTeX">$R^{2}) = 0.730$ </ text -math></inline-formula>。FBA_PC优于先前提出的可应用于MERIS或OLCI数据的PC算法，并且在应用于更大范围的水体时有望具有更强的鲁棒性。
Existing remote-sensing algorithms to estimate the phycocyanin (PC) concentration in turbid inland waters have high associated uncertainties, especially at low PC concentrations in diverse phytoplankton communities. This paper provides the theoretical framework for a four-band semi-analytical algorithm (FBA_PC) which isolates PC absorption from second-order variability caused by yellow matter and other phytoplankton pigment absorption. The algorithm suits the band configuration of both the Medium Resolution Imaging Spectrometer (MERIS) and Sentinel-3 Ocean and Land Color Instrument (OLCI). Calibration of the algorithm was based on absorption data from 12 inland water bodies in the USA, The Netherlands, and China, combined with measurements from laboratory-grown cultures, which demonstrated that the assumptions underlying FBA-PC are an improvement over existing three-band approaches. Validation of FBA_PC in seven inland water bodies in the USA, The Netherlands, and China showed good agreement of FBA_PC adjusted to the MERIS/OLCI band configuration with measured PC, with root-mean-square error =27.691<inline-formula> <tex-math notation="LaTeX">$\text {mg} \cdot \text {m}^{-3}$ </tex-math></inline-formula>, mean absolute percentage error = 172.863%, and coefficient of determination (<inline-formula> <tex-math notation="LaTeX">$R^{2}) = 0.730$ </tex-math></inline-formula>. FBA_PC outperformed previously proposed PC algorithms that can be applied to MERIS or OLCI data, and is expected to be more robust when applied to a wider range of water bodies.