An Absorption-specific Approach to Examining Dynamics of Particulate Organic Carbon from VIIRS Observations in Inland and Coastal Waters
An Absorption-specific Approach to Examining Dynamics of Particulate Organic Carbon from VIIRS Observations in Inland and Coastal Waters
复制标题
一种吸收特定方法，用于检查内陆和沿海水域 VIIRS 观测中颗粒有机碳的动态
DOI：
10.1016/j.rse.2019.01.032
复制
发表时间：
2019
影响因子：
13.5
通讯作者：
Hongtao Duan
中科院分区：
文献类型：
--
作者：
Guangjia Jiang;Steven A. Loiselle;Dingtian Yang;Changjun Gao;Ronghua Ma;Wen Su;Hongtao Duan
采用基于吸收的方法测定内陆和沿海沃茨的表面颗粒有机碳（POC）浓度。这种方法的改进性能是基于本地POC吸收特性的规范的基础上占主导地位的POC来源，浮游植物或碎屑为基础。该规范使用新的POC指数（PI）制定，该指数在时间和空间异质性内陆和沿海水体的POC（300- 10，000 mg/m3）条件下进行了开发和测试。POC模型基于Rrs（751）、Rrs（488）和Rrs（R/G）四个波段的遥感反射率（Rrs，sr-1），其中R为碎屑为主的沃茨的红色波段[Rrs（672）]，G为浮游植物为主的沃茨的绿色波段[Rrs（555）]。模型的R2值为0.74，rRMSE值为42.0%，N = 136，p< 0.01。巢湖和长江口的独立数据验证产生了较大的正偏差（R2= 0.59，rRMSE = 83%，δ= 634 mg/m3，S= 0.63，I = 1439 mg/m3），但与当地数据调整后，偏差减小（R2= 0.80，rRMSE = 45%，δ= 72 mg/m3，S= 0.81，I = 327 mg/m3）。此外，HydroLight模拟提出了PI和CDOM条件之间的独立相关性，以及本研究中开发的新方法的合理POC估计。使用可见光红外成像辐射计套件（VIIRS）在一系列光学复杂条件下的数据对该方法进行了测试，以量化碳动力学。我们指出了在存在多种有机碳源的生态系统中使用这种方法的优势和挑战。
An absorption-based approach was used to determine surface particulate organic carbon (POC) concentrations in both inland and coastal waters. The improved performance of this approach was based on the specification of local POC absorption characteristics based on dominant POC sources; phytoplankton or detritus based. This specification was made using a new POC-Index (PI), developed and tested across a range of POC (300–10,000 mg/m3) conditions in temporally and spatially heterogeneous inland and coastal waterbodies. The POC model was based on remote sensing reflectance (Rrs, sr−1) in four wavebands:Rrs(751),Rrs(488) andRrs(R/G), whereRis the red band [Rrs(672)] for detritus dominated waters andGis the green band [Rrs(555)] in the phytoplankton dominated waters. The model provided a high R2(0.74) and relatively lowrRMSE(42.0%,N= 136,p< 0.01). Validation with independent datasets from Chaohu Lake and the Yangtze River Estuary produced a larger positive bias (R2= 0.59,rRMSE= 83%,δ= 634 mg/m3,S= 0.63,I= 1439 mg/m3); nevertheless, the bias was reduced when tuned with local data (R2= 0.80,rRMSE= 45%,δ= 72 mg/m3,S= 0.81,I= 327 mg/m3). Additionally, HydroLight simulations presented an independent correlation between PI and CDOM conditions and reasonable POC estimates from the new approach developed in this study. The approach was tested using data from Visible Infrared Imaging Radiometer Suite (VIIRS) in a range of optically complex conditions to quantify carbon dynamics. We indicate the advantages and challenges of using this approach in ecosystems where multiple organic carbon sources are present.