Using satellite data to monitor vegetation in floodplains

Published: 7 December 2018

Do you wonder whether the actual development of floodplain vegetation can or will result in increased hydraulic roughness and an increased flood risk of the adjacent river? The first version of the Deltares Vegetation monitor classifies vegetation for a chosen area and satellite image and can determine the change in the vegetation with respect to a given reference classification.

This open tool for the analysis of satellite data has been designed to visualize the change in floodplain vegetation along the Dutch rivers Rhine and Meuse. Vegetation at these floodplains is valued for its ecological services and provision of habitat for other organisms, but at the same time it can result in increased hydraulic roughness of the floodplain, and an increased flood risk, while these floodplains are important for the conveyance of high river waters.

Determine priorities

Rijkswaterstaat, the responsible manager of the Dutch river system, has the challenging task to determine where and how much vegetation must be removed from the floodplains so the acceptable threshold of hydraulic roughness of the floodplains is not exceeded. For this Rijkswaterstaat uses a reference vegetation classification with acceptable levels of biomass and hydraulic roughness coefficients per dominant vegetation type. Each year, field-based assessments must be carried out to identify whether, where and how much vegetation must be removed from the floodplains. However, these assessments are a challenging task, because the floodplains comprise at least 70.000 ha and they are owned by over 12.000 individual owners.

Vegetatiemonitor webviewer openearth


Predicting and floodplains abroad

The Deltares Vegetation monitor uses Google Earth Engine based analyses of ESA Sentinel-2 images to compare the current state of the vegetation with the reference map. These analyses are visualized in a web viewer for easy access by Rijkswaterstaat and the responsible land owners who can now use this quick scan to identify areas of potential risk that might need a field-based assessment to determine the required vegetation management. The intention is to further expand the Vegetation monitor with other algorithms, for example to analyze the development in vegetation over time, and to predict the future development of the vegetation. Furthermore, the tool could also be used for the monitoring of vegetation abroad, by adaption of the vegetation classes in the Google Earth Engine algorithm.