Day 4

Detailed paper information

Back to list

Paper title Observing intertidal vegetation using complementary data from ESA Sentinel-2 and Unpiloted Aerial Vehicles
  1. Mark Warren Plymouth Marine Laboratory Speaker
  2. Stefan Simis Plymouth Marine Laboratory
  3. William Jay Plymouth Marine Laboratory
  4. Aser Mata Plymouth Marine Laboratory
Form of presentation Poster
  • A7. Hydrology and Water Cycle
    • A7.06 EO for monitoring water quality and ecological status in inland waters
Abstract text The European Union Water Framework Directive (and similar directives) requires countries to monitor and report on the ecological status of inland and coastal water bodies, through biological, chemical and physical indicators. Countries reporting on a large number of water bodies struggle to collect sufficient observations to represent seasonal and interannual variability, particularly in dynamic systems influenced by terrestrial runoff. Monitoring of certain indicators can be complemented using Earth observation to fill such observation gaps and inform better management. Satellite remote sensing is particularly complementary and can be achieved at relatively low cost, but water quality products from satellites are still limited to relatively wide, open water bodies.

In shallow coastal waters and intertidal zones, runoff from farming nearby land, water treatment works outflow and untreated sewage can lead to nutrient conditions in which macroalgae flourish and out-compete other beneficial plant life such as Zostera seagrasses. Such shifts can disturb local ecology and lead to loss of biodiversity, reduce important carbon sequestration and negatively affect the blue economy.

With the use of high resolution EO data such as Sentinel-2, together with image processing and machine learning techniques, we are able to observe the areal coverage of vegetation within a tidal lake in the southwest UK and estimate the seasonal variation in cover from a 5-year time series of Sentinel-2. Photography taken from unmanned aerial vehicles additionally provides a very high resolution (~4 cm) view for evaluation or creation of training data, and an estimate of macroalgae coverage itself. Quadrat surveys (which are the accepted reference method) in the region provide further information, but at limited spatial and temporal coverage. The differences in these three levels of observation (satellite, UAV, quadrat) are discussed with suggestions on how they might be reconciled in future so that the wide area and regular temporal coverage that satellites offer can be used in reporting.

Using a clustering approach with the Sentinel-2 data we were able to assign pixels within the lake to classes relating to mud, water and vegetation. Aggregating into seasonal periods suggests that the vegetation coverage within the lake ranges from approximately 5 % of the intertidal area in winter up to 60 % in summer. UAV data, which only covers a portion of the lake, suggests much lower summer coverage of 8 %, whilst the quadrat reference method reports 68 %. To be able to use EO techniques in future WFD activities these methods will need calibrating and agreement with relevant bodies so that the spatial and temporal benefits of EO data can be fully utilised.