The first covering of Denmark's shallow water

Marine vegetation mapping

First large-scale overview of important coastal marine habitats

Current knowledge on the spatial distribution of underwater vegetation is to a large degree based on very time consuming techniques.

The observations come from divers as part of official environmental monitoring programs, or they are based on interpretations of aerial photos covering smaller areas.

This can provide detailed information at local scale but lacks the possibility of a large-scale synoptic overview.

Based on detailed satellite data, novel machine learning techniques and advanced data processing, DHI GRAS has created the first-ever national overview of the spatial distribution of shallow-water submerged aquatic vegetation in Denmark.

The map shows the distribution of aquatic vegetation (eelgrass and macroalgae) observed in 2018 and may form an important baseline for evaluation of any changes at large to regional scale in the coming years.

How the map is made

Technique and data

The map is made based on data from the Copernicus Sentinel-2 satellite in 10m spatial resolution from 2018.

We have combined our advanced image analysis approach with a machine learning algorithm to derive detailed information about the sea floor coverage. The classification approach is refined with the inclusion of training data from the Danish national monitoring program and other sources.

Challenges

Water clarity is one of the parameters controlling how deep we can monitor the underwater vegetation. Other factors are atmospheric clarity (the satellite observes the Earth through the atmosphere) and the amount of light (the higher the intensity the deeper the penetration in the water). In Denmark the best combination of these factors is typically found in late-spring / early-summer and input data for the mapping activity was therefore mainly from May 2018.

The output maps where validated with a subset of the training data to ensure the highest possible accuracy.

What's there to see

The map is shown with 3 different classes:

  • No Vegetation: where the sea floor has been assessed to be mainly free from vegetation
  • Sparse Vegetation: where the sea floor has been assessed to contain some vegetation
  • Dense Vegetation: where the sea floor has been assessed to contain a dense vegetation cover
Before After

No data is shown where conditions are too deep to see the sea floor, or where suspended sediment in the water column obscures the view.

Before After

The map is shown with a certain degree of smoothing. The classification approach is based on identification of objects with similar behaviour, meaning that individual image elements (in this case areas with a size of 10 x 10 m) are combined into groups of elements with similar appearance. The minimum size of these objects was set to 4 image elements (400 m2).  

Before After

What's next?!

The current map shows the conditions as they were observed in 2018 and forms an important first base-line data set. With the methodology in place we see several ways to advance the current study and get beyond the static base-line.

Year-to-year variations

Intra-annual variations

Year-to-year variations

A single map is illustrative of the conditions at the time of observation but including another time step will allow us to identify changes and detect trends.

There is a natural variability in the spatial distribution of the aquatic vegetation, but little is known about the actual behaviour. Looking at year-to-year variations will give us much better knowledge of the quantity of variability, it will allow us to see effect of environmental regulations (eelgrass distribution is an indicator of the environmental status of an area, a positive trend in vegetation coverage could suggest improved status).

Intra-annual variations

We have frequent satellite overpasses during the year, so a possibility is also to look at intra-annual variation.

This will allow us to assess the growth rate of eelgrass, help to separate eelgrass and macroalgae in a classification (eelgrass is much less dynamic than macroalgae), and give us better knowledge on how extreme events (e.g. the very warm water temperatures during the summer of 2018) affects the vegetation health and coverage.

See for yourself

We believe it is important to share this kind of information whenever it is possible. We have therefore made sure everybody has easy access to the new Danish nation-wide map of aquatic vegetation through the data browser below. It works on your mobile phone also so you can check out the map while out and about.

An initiative of DHI

At DHI we are proud to publish the first ever nation-wide overview of the distribution of aquatic vegetation in all Danish shallow waters.

The activity was made possible by a generous grant from the VELUX Foundation as part of their Environment and Sustainability activities.

DHI GRAS

gras@dhigroup.com
+45 4516 9100

Agern Alle 5,
2970 Hørsholm,
Denmark

CVR: 25621646