Salt marsh extract, RGB and height model

Aerial survey salt marsh quantification

August 2015

Our Small Unmanned Aircraft (SUA) was used to survey an area of salt marsh, to assess its morphology and quantify its size. In particular we needed to quantify the proportion of the marsh which had been colonised by plant species.

Approximately 90% of the marsh was inaccessible by foot. Heavily dissected by creeks and drainage ditches, the site could not easily or quickly be recorded using land survey techniques. Owing to the tides, the survey had to be conducted rapidly, so the SUA was the ideal solution.   

Salt marshes form on the banks of an estuary or river in the intertidal zone where the speed and flow of water allow sediment to accumulate, up to the point at which it can be colonised by vegetation. Changes to the coastal environment caused by natural agency or as a product of human intervention can affect the speed and flow of water around a marsh and thus its health. Monitoring of salt marsh size and changing morphology is important, as it can indicate and quantify the effect of such changes. This is increasingly important as widespread change results from rapidly rising sea levels.

The SUA completed a 20 minute photographic survey of an area of c.60,000 sq m, capturing around 300 images from a height of roughly 40m.  From this, a detailed 3D mosaic image of the marsh was created depicting the upper areas colonised by vegetation and the deeper creeks and man-made drainage ditches that lie in-between.  

To manually quantify the area colonised by the main body of salt marsh species, as opposed to pioneer species or algae within the gullies and creeks between, would have taken days owing to the fractal nature of the marsh. Instead map mathematics was applied.

Since the colonised areas were uppermost, we first derived a new surface from the original data, based solely on taking the heights of points which were colonised.  Then we subtracted that surface from the original one.  It is a bit like laying a huge piece of graph paper over the marsh, which follows the surface but does not dip into the gullies, and then asking how far away each square of that paper is from the original surface.  If a square overlaid vegetation there would be little difference, while a square suspended over a deep gully would produce a big difference. The new ‘difference’ layer produced by this comparison,  can then be interrogated to determine the vegetated area.

The approach produced a highly detailed topography of the site and the area of main salt marsh vegetation. These data sets enable future repeat survey to monitor and quantify change.

We would like to acknowledge the input of the client manager DLEnviro Limited in completing this work.