-
COASTAL CHANGE
Commissioned by the European Space Agency's Coastal Erosion Project
-
COASTAL CHANGE
Commissioned by the European Space Agency's Coastal Erosion Project
Products & Services
The Coastal Change from Space is a global service developed by ARGANS to detect annual and seasonal changes along long stretches of shorelines over 25 years of observations. This can provide significant insights into determining where investments by governments should be focused and how corrective actions can be targeted, in the framework of a climate adaptation strategy.
Waterline
The waterline processor extracts a line corresponding to the instantaneous boundary between land and water. To do this, the processor extracts specific bands from a multispectral image and applies a spectral index calculation to these image bands. This creates a binary image where each image pixel corresponds to either land or water. The waterline is then extracted along this boundary between the binary pixels.
The waterline processor is incredibly versatile, allowing the use of various spectral indices and customizable processing parameters. This allows waterlines to be cleanly extracted, even in noisy areas with high wave action or sedimentation. The processor can also ingest all manner of optical images, including Landsat 5 - 8, Sentinel-2, and even VHR products such as RapidEye, Pleiades, WorldView and QuickBird. This allows for the creation of waterline datasets spanning over 26 years.
By ingesting co-registered images, the processor provides waterlines with exceptionally low positional uncertainty. This increases the sensitivity to which coastal change can be detected from any subsequent derived products. In addition to monitoring coastal erosion, the waterlines can be utilised in mapping tidal creek networks, estimating storm inundation extents, and monitoring industrial coastal development.
Shoreline
The detected waterline is further processed to take into consideration the beach profile and adjust to a tidal datum i.e., a fixed elevation corresponding to a mean or extreme value of tidal elevation.
The position of the waterline at any given moment will vary due to environmental factors such as the beach morphology, tidal forcing and even breaking waves. This variability makes it difficult to accurately measure coastal erosion using waterlines alone. The shoreline processor transforms a waterline from its instantaneous position to a fixed, corrected position. This new line represents the theoretical waterline at a selected tidal datum height, with environmental factors omitted. This fixed point of reference allows coastal erosion or accretion to be precisely measured through analysis of shorelines across a given time range. Using the 26-year long waterline record, it is possible to analyse specific erosive storm events, seasonal changes to beach volumes, and calculate historic rates of shoreline change.
The processor uses various metocean data to calculate the shoreline position including measured or modelled water level, wave data and beach surveys. It uses an innovative interpolation method to simultaneously calculate a varying water level, tidal datum height and beach slope at each point along the shoreline. This maintains consistent shoreline accuracy, even in areas with complex hydrodynamic regimes or along vast stretches of coastline where metocean information may be sparse.
Bathymorphology
To gain a better understanding of what is occurring in the intertidal region ARGANS have developed a technique to measure the amount of sunlight that is reflected from the seabed through the water column and then through the atmosphere to the over passing satellite sensors onboard the Sentinel 2 Multispectral Imager. This process works very well in clear waters, however waters near areas of potential erosion are filled with suspended sediments and organic material. ARGANS Ltd have developed sophisticated masks to enable the analysts to know when the seabed is being observed.
Land Use Classification
Coast dynamics and their response to climate change rely on the type of coastal geology and morphology. To obtain the knowledge on the type of land cover, a supervised classification is performed on coastal areas. Due to their nature, coastal features highly change through the year and depending on the season, are similar to backshore field crops or excavation. A temporal classification is thus performed using yearly acquisitions (from January to December), in order to capture and discriminate all seasonal variations of the different types of land cover.
The following example showcases a classification map of the Start Bay area, SW England, for the 2017 year.
On coastal areas, this temporal classification, as the acquisition used have been taken at different water levels; we can capture and classify the tidal area.
The waterline processor is incredibly versatile, allowing the use of various spectral indices and customizable processing parameters. This allows waterlines to be cleanly extracted, even in noisy areas with high wave action or sedimentation. The processor can also ingest all manner of optical images, including Landsat 5 - 8, Sentinel-2, and even VHR products such as RapidEye, Pleiades, WorldView and QuickBird. This allows for the creation of waterline datasets spanning over 26 years.
From the computed classification, we can extract different areas of interest in order to monitor specific environments. At ARGANS, We focus on three specific areas. Firstly, the coastal area (backshore with land features masked), with all surrounding coastal features and local environment. Then, the littoral area (composed of beaches, mudflat, tidal area, etc.), where waterline and shoreline evolve, a shoreline can be generated out of the littoral area, excepting in case of extreme weather events. Finally, the ocean area with no coastal or land features.
All coastal features
Littoral area
Water area
Full coastal area classification
Image Co-registration
Image co-registration is performed when the intention is to study two or more images in a series, typically to understand or to follow change. Over an important time interval, it is possible to use one or different satellite products, which will influence the spatial resolution. The rationale of co-registration is to ensure that the images become spatially aligned so that any feature in one image overlaps as well as possible its footprint in any other image in the series. Co-registration is normally carried out by selecting one image as the reference (base/master) to which all the other images are aligned.
The master image is supposed to be the most perfect image possible on the area, As part of the project we used very High Resolution and orthorectified data, the selection of this image can be carried out by use of the image metadata. The co-registration process requires the identification of common objects in the master and the target image (ie those to be co-registered). The locations of common objects are called tie points and are used for relocation. AROSICS the co-registation module used to generate these link points following a grid with equidistance is predefined by the user. This will then be filtered according to resemblance rules, a function is used to align the distorted image on the master.
GIS Analytical Techniques
ARGANS analysts use a variety of techniques to perform Quality Control and to enhance the raw waterline and shoreline products that are available from the Data Access Portal. These use the freely available open-source GIS platform, QGIS.
The linked document explains the steps performed to utilise this Quality Control information and to post-process the products using these techniques.
Data Sites
During Phase 1 five test areas were selected: Longue Pointe de Mingan, Qeubec, Start Bay & Perranporth, UK, the Wexford Coast south of Dublin, Ireland and the Spanish coast between El Prat South of Barcelona and the Tordera Delta. Subsequently another 15 sites have been added for Phase 2 and all sites are displayed on the map.
Data Sources / Flow
The coastal erosion project requires data from varied sources: public freely downloadable global satellite imagery along with Very High Resolution (VHR) images from commercial third-party suppliers, and regional auxiliary data including tidal predictions, or preferably in situ tide gauge measurements, wind speed, wave direction and, importantly, tidal datums from regional observatories. Some of the complexity of the data flow is captured in this infographic.
