Peer-Reviewed Journal Details
Mandatory Fields
Karki, S., Bermejo, R., Wilkes, R., Mac Monagail, M., Daly, E., Healy, M.G., Hanafin, J., McKinstry, A., Mellander, P.E., Fenton, O., Morrison, L.
2021
July
Frontiers in Marine Science
Mapping spatial distribution and biomass of intertidal Ulva blooms using machine learning and earth observation.
Published
()
Optional Fields
Sentinel-2; Landsat; Earth Observation; Macroalgal blooms; Ulva; Green tide; Mapping
8
Opportunistic macroalgal blooms have been considered to be an essential factor in determining the overall ecological and environmental status of coastal and estuarine areas. A novel approach to map green algal cover using a Normalised Difference Vegetation Index (NDVI) approach was developed using earth observation (EO) data sets. Scenes from Sentinel-2A/B, Landsat-5 and Landsat-8 missions were processed for eight different estuarine areas of moderate, poor and bad ecological status using the Water Framework Directive classification for water bodies. Cloud-free images acquired during low-tide conditions from 2010 to 2018 over 18 days of field surveys were considered. The estimates of percentage coverage obtained from EO data sets corresponded with the field survey coverage and the estimates obtained from both 10 m and 30 m visible and near-infrared (NIR) bands showed statistically significant results. The results showed that the NDVI technique could be successfully applied to map the coverage of the blooms and to monitor estuarine areas in conjunction with other monitoring activities that involve field sampling and surveys. The combination of wide-spread cloud-coverage and high-tide conditions provided additional constraints during the selection of the images. Citing these constraints, the findings showed that both Sentinel-2 and Landsat scenes could be utilised to estimate bloom coverage. Considering the importance of biomass for understanding the severity algal accumulations, an Artificial Neural Network (ANN) was trained using the in situ historical biomass samples and the combination of radar backscatter (Sentinel-1) and reflectance in the visible and near-infrared regions (Sentinel-2). The ANN model results showed that it could be used to map the biomass, and the model performance can be improved with the addition of more training samples. The developed methodology can be applied in other areas experiencing macroalgal blooms in a simple, cost-effective and efficient way. The study has demonstrated that both the NDVI-based technique to map spatial coverage of macroalgal blooms and the ANN-based model to compute biomass have the potential to become an effective complementary tool for monitoring macroalgal blooms where the existing monitoring efforts can leverage the benefits of EO application.
https://doi.org/10.3389/fmars.2021.633128
Grant Details
Environmental Protection Agency (EPA)
Project no: 2018-W-MS-32
Publication Themes
Environment, Marine and Energy