Public defence in geoinformatics, M.Sc. (Tech.) Petri Palviainen

Title of the thesis: Measuring spectral signatures of forest floor and trees

Thesis defender: Petri Palviainen
Opponent: PhD Marco Celesti, European Space Agency - ESTEC, The Netherlands
Custos: Professor Miina Rautiainen,Aalto University School of Engineering, Department of Built Environment

This work develops field- and laboratory methods for measuring reflectance spectra of forest floor, trees, and plant parts.

This dissertation presents, evaluates, and refines ground based methods for collecting high quality spectral data of different types of forest vegetation. The dissertation constitutes one of the most comprehensive investigations to date into how forest vegetation—specifically trees and forest floors—reflects light across the optical spectrum, from the visible to the near infrared and shortwave infrared wavelengths. So far, the shortwave scattering properties of individual forest components are largely unexplored with a particular lack of knowledge on scattering anisotropy. By linking the structural and compositional properties of vegetation to features in their reflectance spectra, the collected spectral libraries serve both as reference data for drone, airborne, and satellite observations of forests and help to develop more accurate forest reflectance models.

The measurements were made under both controlled laboratory conditions and field conditions in boreal, hemiboreal, and temperate forests across Europe. By combining the hyperspectral resolution of a portable spectrometer with goniometers and handheld probes, the research reveals details on the scattering characteristics of plants and plant components that satellites cannot currently resolve due to the inherent structural complexity of forests. Through five studies, the dissertation demonstrates that, when appropriate methods are applied and measurements are made with care, it is possible to collect spectral data at different spatial scales from individual leaves to whole trees. Based on the results, forest spectra exhibit high diversity across vegetation types, plant species, and forest floor composition, which are further influenced by phenological stage and tree canopy structure. Moreover, the spectral signatures of vegetation show a strong dependence on view angle.

These insights help to clarify why forest pixels in satellite images are so difficult to interpret, and how integrating precise ground measurements with broad satellite observations can help to address this difficulty. The results highlight the important role of ground-based spectral measurements in assessing and improving the measurement accuracy of remote sensing methods. More accurate forest mapping and monitoring methods support multidisciplinary research, practical work of forest owners, work of authorities and legislators, and businesses.

Keywords: remote sensing, forest reflectance spectrum, goniometer

Thesis available for public display 7 days prior to the defence at Aalto University's public display page. 

Contact information: petri.palviainen@aalto.fi