Estimating forest biomass components by airborne and terrestrial laser scanning

Marius Hauglin presented a lecture on a predefined topic and defended his doctoral thesis on December 7,  2012. The topic for the trial lecture was “Why and how should we use terrestrial laser scanning in operational forest inventories”  and the title of his thesis was “Estimating forest biomass components by airborne and terrestrial laser scanning”. We congratulate!

Thesis abstract

In the thesis forest inventory methods to estimate potential logging residues and branch biomass using both airborne and terrestrial laser scanning are explored. Forest inventories are essential for effective and sustainable management of forest resources. In the last ten years there has been an increased interest in the use of forest biomass for bioenergy purposes, and biomass from forests will most likely be one of several sources of energy that will have to replace fossil fuels in the future. One example of such use of forest biomass is the utilization of logging residues, biomass that would otherwise have been left in the forest during the logging. When logging residues become a commercial product from the forest, this resource should be quantified as part of the forest inventory to improve planning and management.

The four studies in this thesis were carried out in a boreal forest in the south-eastern part of Norway, with Norway spruce and Scots pine as the dominant tree species. Several sets of field reference data were used, including destructive sampling of 50 spruce trees to accurately determine branch biomass. In the first study, a method to estimate potential logging residues on an area basis using ALS data of moderate density was presented. Logging residues was in this study defined as branches and tops and are “potential” in the sense that these are the tree-components that will become logging residues in the case of a final harvest. The methodology follows the area-based approach which has previously been successfully applied to derive estimates of e. g. mean stem volume. A strong relationship was found between field measured potential logging residues and ALS variables, with a mean prediction error of 22% at the stand level.

In the three subsequent studies in this thesis methods to estimate branch biomass at the single-tree level for Norway spruce were investigated. Norway spruce is one of the two main tree commercial species in the Nordic countries. In the second study, the relationship between ALS derived features and accurately measured branch biomass was found to be strong, with a mean prediction error of 35% for the best model. A major finding in this study was that the ALS data contained more information related to branch biomass than the actual field measured tree diameter and height.

A spruce tree measured with Terrestrial Laser Scanning (TLS)

A spruce tree measured with Terrestrial Laser Scanning (TLS)

The laser ranging principles used in ALS have also been applied in scanning from fixed positions on the ground, commonly known as terrestrial laser scanning (TLS). TLS is considered a potential tool for capturing information for forest inventories, possibly replacing time-consuming manual field registrations, or even better: capture structural information unobtainable through conventional field measurements. Extraction of information related to branch biomass from remote sensing data was further explored in the third study where TLS data were used to estimate branch biomass. A strong relationship was found between TLS derived features and accurately measured branch biomass, with a mean prediction error of 32%. Although a strong relationship was found, the observed prediction accuracy was considered to be only moderate in light of the rather extensive data acquisition. In the last study, data from ALS and TLS were combined; TLS derived branch biomass estimates were used as ground reference data in ALS based estimation of branch biomass. The aim of this fourth study was to describe this methodological framework and assess the prediction accuracy using field and remote sensing data. Predictions with a mean error of 32% were obtained for single-tree branch biomass. The results show an improvement compared to an approach without the use of TLS data, and also suggest that branch biomass can be successfully included in a single-tree forest inventory.

Supervisors

Professor Erik Næsset (main supervisor) INA, UMB
Professor Terje Gobakken INA, UMB

Evaluation committee

Research Forester Dr. Hans-Erik Andersen , USDA Forest Service  – University of Washington , USA
Professor Paul M. Treitz, Department of Geography – Queen’s University  , Canada
Professor Ole Hofstad  INA, UMB

Reference

  • Hauglin, M.. ((2012). Estimating forest biomass components by airborne and terrestrial laser scanning.). PhD Thesis.
    [Bibtex]
    @PhdThesis{Hauglin2012,
    Title = {Estimating forest biomass components by airborne and terrestrial laser scanning},
    Author = {Marius Hauglin},
    School = {Norwegian University of Life Sciences},
    Year = {2012},
    Comment = {Thesis no. 54. 98 p.},
    Owner = {hanso},
    Timestamp = {2012.12.11},
    Url = {https://static02.nmbu.no/mina/forskning/drgrader/2012-Hauglin.pdf}
    }