The first question for information about trees in the Netherlands is ”what species is it?”. At a meeting of the Dutch Tree register (Boomregister) cooperative in April 2017 it was suggested that the phenological data gathered on Nature Today about leaf growth could be combined with the NDVI index shown by Groenmonitor in order to determine the species of trees. Another source for species identification could be (hyper)spectral data, available for some areas. And there is more. The available Boomregister data contains canopy outlines and height for 52.6 M canopy objects. A number of municipalities make available the species+location of trees in their care via BGT, the Dutch Topographic Base Registry (effective July 1, 2017). Many data are available for Speulder and Sprielder forest, the largest teaching forest of the Netherlands. Also the results of the National Forest Statistic (1-4) and its successor, the Dutch Forest Inventory, are available.
The subject of this Master thesis is the question if, in principle, species identification in the Netherlands is possible from the data mentioned above (phenological observations, NDVI, (hyper)spectral data, canopy+height pattern, forest data) without personally going out and identifying each tree.
If it is possible to determine species for individual trees or groups of trees or forest stands, it would be useful information for the Dutch Tree register, especially for trees that are not owned and maintained by municipalities. Species information could be useful for linking with tree growth models, which is relevant for CO2 sequestration, biomass and ecological aspects.
- The subject of this Master thesis is the question if tree species identification is a viable option.
- Research questions would be (among others):
Are there sufficient observations on Nature Today about trees of 1 or more species?
- Are there sufficient images in Groenmonitor on trees to provide an NDVI-sequence within a year that reflects the leaf growth of a (deciduous) tree?
- How could Groenmonitor NDVI sequences be linked to Nature Today observations?
- Could an automated procedure be designed for determination of 1 species?
- Could the Dutch Forest Inventories play a role in validation of the ‘calculated’ tree species?
- What is the overall assessment of viability of this approach at this moment
- What are the obstacles in this approach (technical/methodical/organisational/other)
- Nature Today: https://www.naturetoday.com/intl/nl/observations/natuurkalender/species/plants
- Groenmonitor: http://www.groenmonitor.nl/
- Dutch tree register http://boomregister.nl/
- Dutch Forest Inventory: http://www.probos.nl/rapporten-2014/1115-nbi-nbi6-mfv
- Pandey et al. 2014: Mapping tree species etc. IEEE Sensors Journal vol.14-12, p.4431-4441
- Immitzer et al. 2012 Tree species classification etc. Remote Sensing 2012-4, 2661-2693
- Schelhaas et al 2014. Zesde Nederlandse Bosinventarisatie; Methoden en basisresultaten. Wageningen, Alterra Wageningen UR (University & Research centre), Alterra-rapport 2545 (http://edepot.wur.nl/307709).
- Student should be RS knowledgeable
Theme(s): Sensing & measuring, Modelling & visualisation