Subscribe to ShellEye bulletin service Feedback survey Register your interest

Observing harmful algal blooms from space

22 August 2016
Dr Andrey Kurekin and Dr Hayley Evers-King, both Earth Observation scientists at PML, give an update on their progress with satellite monitoring.
Estimation of water quality parameters from satellite images provides significant advantages over direct sampling measurements, whcih can be time consuming and expensive. The approach adopted in the ShellEye project uses satellite data to generate map of chlorophyll concentration and harmful algal blooms (HABs).

The chlorophyll maps are used to estimate water quality and areas of high chlorophyll concentrations on the maps can indicate the presence of algal bloom, which may become harmful for shellfish farms. Harmfulness of algal blooms cannot be established by analysis of the chlorophyll concentration maps alone; it requires analysis of water colour properties for harmful algal species. 

A methodology has been developed at PML for automatic analysis of water colour properties and detection of harmful algal blooms of different species. The methodology is applied to satellite images to generate HAB risk maps for UK coastal waters, showing the areas with high risk of bloom in red, harmless blooms in green and bloom areas in blue. The maps were generated for the Karenia-mikimotoi and pseudo-nitzschia algal species, by processing MODIS and VIIRS satellite images in near real-time, to deliver the most recent information about the development of algal blooms.

The quality of HAB risk maps was validated by comparing in-situ measurements taken at sampling stations located along the UK coast. We used SAMS monitoring data in 2006-2012 to validate Karenia-mikimotoi risk maps and Food Standards Agency monitoring data in 2012-2014 to validate pseudo-nitzschia risk maps.  

The HAB risk maps will be used in the project for the development of water quality bulletin service and to provide sustainability of the ShellEye service beyond the end of MODIS mission, we are working on extending the developed methodology for HAB detection to new satellite sensors, such as generating Karenia-mikimotoi HAB risk maps from processed VIIRS images, which is currently in the testing stage.

As well and using data from the MODIS and VIIRS satellites, the ShellEye project has been investigating how other sensors that measure ocean colour can be used to support water quality monitoring in the our coastal oceans. Sensors that were originally designed for land based monitoring (such as Landsat 8), can be used in the coastal ocean when there are big signals to observe, such as high biomass algal blooms or very turbid river plumes. These sensors offer some significant advantages in terms of their spatial resolution, allowing us to look much closer near shore in greater detail.  

Using Landsat 8 to estimate concentrations of suspended matter, allows for the mapping of river plumes, to understand where and when these occur. This information can then be related to other parameters, such as rain, river flow rates et.c to understand the prevalence of different biotoxins.

Using  high resolution sensors like Landsat 8 also comes with some challenges. As with all sensors that record visible colour, they can’t see through cloud. They also do not revisit the same area as frequently as medium resolution sensors like MODIS and VIIIRS. This means we do not get images from Landsat 8 as often (only every 8-16 days depending on the location). To help combat this problem, the ShellEye project is starting to work with data from sensors onboard new satellites from the EU Copernicus programme. Sentinel 2a (and the soon to be launched Sentinel 2b) will help to increase the frequency of images we are able to capture, up to every few days, once both satellites are launched and used alongside Landsat 8. Satellite data is also supporting the modelling work being conducted by University of Exeter scientists within the Shelleye project, with sea surface temperature data obtained daily from satellites as an input into models used to forecast E.coli levels.