Use of remote sensing for mapping of non-native conifer species

In a project founded by the Norwegian Environment Agency we have exemplified how remote sensing can be used in the management of non-native species.

Non-native species are by many considered a threat to local biodiversity. In Norway, conifer species have been introduced in order to find species with better timber production than the native species. Several of these introduced species have been considered to be invasive, and put on an official «blacklist». Thus, from a management perspective, more information about the extent, occurrences and potential dispersal are important information. To gather such information solely based on field surveys are time-consuming and costly, and it has therefore been suggested to develop methods based on remote sensing. In this report we review different types of remote sensing data and how these can be used to map and monitor non-native species.

Natural species distributions of Norway spruce and Scots pine were created based on available literature and existing remote sensing-based forest maps. The same maps were used to create a non-native species map, i.e. a map of areas where spruce occur outside its natural distribution. We evaluated the accuracy of the map by photo-interpretation, and assessed the consistency with other occurrence data. We further estimated the area of non-native species on a county and national level in Norway. The area covered by non-native species outside the natural distribution of spruce was estimated to be 1200 km2, with a standard error of 275 km2.

A specific challenge when using remote sensing for mapping of non-native species in Norway is to separate species of the same genera. We therefore conducted a study in Fusa and Tysnes municipalities where we evaluated the ability to discriminate between Norway spruce and Sitka spruce using different types of remote sensing data. Data from Landsat 8 satellite images, aerial imagery and airborne laser scanning were tested. Slight to moderate ability to separate between the two species were found, with a best overall accuracy of 78%. The results suggest that Landsat 8 imagery can be used to discriminate between stands dominated by Norway spruce and Sitka spruce. Additional data from airborne sensors contributed not substantially in this case.

Based on our own analyses and a review of relevant literature we discuss a possible establishment of a national mapping and monitoring programme for non-native tree species.

Complete report:

  • Ørka, H. O., & Hauglin, M.. (2016). Use of remote sensing for mapping of non-native conifer species (INA fagrapportNo. 33). Norwegian University of Life Sciences, Faculty of Environmental Science and Technology, Department of Ecology and Natural Resource Management.
    [Bibtex]
    @TechReport{Oerka2016,
    author = {Ørka, Hans Ole and Hauglin, Marius},
    institution = {Norwegian University of Life Sciences, Faculty of Environmental Science and Technology, Department of Ecology and Natural Resource Management},
    title = {Use of remote sensing for mapping of non-native conifer species},
    year = {2016},
    number = {33},
    type = {INA fagrapport},
    keywords = {Remote sensing, non-native species, species distribution maps, satellite imagery, aerial imagery, airborne laser scanning, tree species classification.},
    owner = {hanso},
    timestamp = {2016.10.19},
    url = {http://www.umb.no/statisk/ina/publikasjoner/fagrapport/if33.pdf},
    }