Monitoring Vegetation Change in Inland Drift Sands Based on Aerial Photographs

Organised by Laboratory of Geo-information Science and Remote Sensing

Tue 9 April 2019 13:00 to 13:30

Venue Gaia, building number 101
Room 3

By Jan van Harten

Inland drift sands are a scarce and threatened landscape type in the Netherlands. The remaining drift sands are protected by Natura 2000 legislation because of the rare habitats and fauna that are still present in these areas. One of the goals described in the Natura 2000 area plan of the Veluwe, a nature area in the centre of the Netherlands including several drift sand areas, is to enlarge the quantity of the typical drift sand habitats. Therefore, it is important to monitor how the landscape is changing. Remote sensing techniques offer great possibilities to monitor vegetation changes at a large scale. The vegetation change between 2008 and 2018 at the Kootwijkerzand, one of the drift sand areas of the Veluwe, is assessed using aerial photographs and a UAV image. The aerial photographs of each year were classified in to five classes, Bare sand, Sand with grasses and juniper hair-cap moss, Sand with grasses and Campylopus introflexus, Heath and Forest, using GEOBIA. The vegetation in inland drift sand area Kootwijkerzand was classified successfully based aerial photographs, with accuracies of 70% (2008), 57% (2016) and 63% (2018). Classification of the vegetation based on the image required by the unmanned aerial vehicle (UAV) was less successful, with an accuracy of 39%. The vegetation changes rapidly, with a decrease of bare sand surface from 88 hectares in 2008 to 36 hectares in 2018, a decrease of 59%. In the same period, the tree covered area increased with 23 hectares, of which the vast majority close to the edge of the forest. The surface area of both pioneer vegetation types (with hair-cap moss and Campylopus interflexus) and heath was stable.

Keywords: Vegetation change; Vegetation classification; Vegetation succession; Aerial photographs; UAV images; Object-based image analysis; GEOBIA; Inland drift sands; Kootwijkerzand; Veluwe