R&D highlights edition 2019

PROJECTS Ecosystems and environmental quality 26 F loodplain vegetation in the Dutch Rhine and Meuse river branches must be strictly monitored to ensure that overall roughness in the floodplains does not reduce water discharge capacity. The flood risk increases if there is too much vegetation in the floodplains and so grazing or cutting is required. Given the presence of approximately 12,000 landowners on the floodplains of the Dutch Rhine and Meuse, the challenge for Rijkswaterstaat is to direct management in a good dialogue with the landowners, who are responsible for timely vegetation management. The aim of this project is to develop an easy screening tool to help direct this effort by classifying readily available satellite images in dominant vegetation types. Another aim is to predict the development of vegetation in the next 5-10 years using habitat modelling in the Google Earth Engine. This will help Rijkswaterstaat with their future planning for flood risk management Machine learning algorithms were used to classify Sentinel2 images in combination with information about vegetation height from Lidar data in the Google Earth Engine cloud- based platform. The resulting classification was compared with the statutory vegetation map used in all formal flood risk analyses. Where vegetation development is rougher than allowed according to this map, the area is flagged in red. This information is presented in an easy access web tool for assessments by the end users. The classified images can also be downloaded for further detailing in a GIS environment. The vegetation monitoring tool is the first tool to provide up- to-date information about the current status of vegetation in the Rhine and Meuse floodplain areas. The vegetation status provided by the tool provides added value because of the frequency of the Sentinel2 data, which is much higher than the annual aerial photograph survey (which is still used for verification). Rijkswaterstaat has found that the monitoring tool helps significantly in the dialogue with nature organisations responsible for the maintenance of the vegetation in large parts of the flood plains. The tool was used on tablets during the annual field check-ups in areas identified as being rougher than required by the statutory map. Rough spots such as quick developments of willow storage along the shore or at dried up gullies were easily identified. Although the tool classification was not always correct given the fact that this is a first version, the feedback from RWS was very positive and the tool has also been welcomed also by other end users such as Staatsbosbeheer, Natuurmonumenten, and other cultural and nature foundations. Contact: Ellis Penning, Ellis.Penning@deltares.nl , t +31 (0)88 335 868 Gertjan Geerling, Gertjan.Geerling@deltares.nl Further reading : https://www.openearth.nl/vegetatiemonitor/ The vegetation monitor allows for the assessment of the current status of vegetation in floodplains using remote sensing analysis in the Google Earth Engine. This assessment is required annually before the winter to ensure that vegetation does not block river discharge. THE VEGETATIONMONITOR: REMOTE SENSING TO FACILITATE FLOODPLAIN MANAGEMENT Screenshot of the vegetationmonitor comparing the statutory vegetationmap with the current situation Testing the vegetationmonitor with end users

RkJQdWJsaXNoZXIy Mjc4NjU=