Monitoring levels of cyanobacterial blooms using the visual cyanobacteria index (VCI) and floating algae index (FAI)

• Published in: ITC journal Vol. 38; pp. 335 - 348
• Main Authors: Oyama, Yoichi, Fukushima, Takehiko, Matsushita, Bunkei, Matsuzaki, Hana, Kamiya, Koichi, Kobinata, Hisao
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.06.2015
• Subjects: Chlorophyll-a; Cyanobacterial blooms; Lake; Landsat; Phycocyanin; Water amenity; Japan, Honshu, Ibaraki Prefect., Kasumigaura L; INW, Japan; Water amenity; Phycocyanin; Cyanobacterial blooms; Chlorophyll-a; Lake; Landsat
• ISSN: 1569-8432; 0303-2434; 1872-826X
• DOI: 10.1016/j.jag.2015.02.002

•We proposed a method for monitoring level of cyanobacterial bloom from Landsat data.•The visual cyanobacteria index (VCI) was used for classifying the level of the blooms.•The FAI was employed to relate the VCI and Landsat data.•The VCI classification system was appropriate when cyanobacteria form surface scums. Cyanobacterial bloom is a growing environmental problem in inland waters. In this study, we propose a method for monitoring levels of cyanobacterial blooms from Landsat/ETM+ images. The visual cyanobacteria index (VCI) is a simple index for in-situ visual interpretation of cyanobacterial blooms levels, by classifying them into six categories based on aggregation (e.g., subsurface blooms, surface scum). The floating algae index (FAI) and remote sensing reflectance in the red wavelength domain, which can be obtained from Landsat/ETM+ images, were related to the VCI for estimating cyanobacteria bloom levels from the Landsat/ETM+ images. Nine field campaigns were carried out at Lakes Nishiura and Kitaura (Lake Kasumigaura group), Japan, from June to August 2012. We also collected reflectance spectra at 20 stations for different VCI levels on August 3, 2012. The reflectance spectra were recalculated in correspondence to each ETM+ band, and used to calculate the FAI. The FAI values were then used to determine thresholds for classifying cyanobacterial blooms into different VCI levels. These FAI thresholds were validated using three Landsat/ETM+ images. Results showed that FAI values differed significantly at the respective VCI levels except between levels 1 and 2 (subsurface blooms) and levels 5 and 6 (surface scum and hyperscum). This indicated that the FAI was able to detect the high level of cyanobacteria that forms surface scum. In contrast, the Landsat/ETM+ band 3 reflectance could be used as an alternative index for distinguishing surface scum and hyperscum. Application of the thresholds for VCI classifications to three Landsat/ETM+ images showed that the volume of cyanobacterial blooms can be effectively classified into the six VCI levels.