News Contact Information

Dpt of Forest Sciences
P.O. Box 27
(Latokartanonkaari 7)
FI-00014 University of Helsinki

Student Services:
Registration of credits:
Communications and Community Relations:
PhD Studies:

How can large-scale forest inventories be improved?

In a recent study by Helsinki University, published in Remote Sensing of Environment, it is demonstrated how sample-based laser scanning data can be combined with traditional satellite data and be cost-effectively utilized for estimating forest growing stock volume across large areas.

The study was conducted in collaboration with researchers from the Finnish Forest Research Institute, the Finnish Geodetic Institute, and the Swedish University of Agricultural Sciences. The demand for this type of inventories, largely based on remotely sensed data, is rapidly increasing following the demands for forest information from the global climate and biodiversity conventions. Large parts of the world's forest are located in remote areas, to which access is limited and thus field work is very expensive to conduct.

The study investigated several strategies for combining data from laser scanning, satellites, and field surveys. A novel feature was to select samples of laser scanner data by probabilities proportional to preliminary estimates based on a combination of satellite and field data. While satellite data typically are available wall-to-wall at low cost, the laser scanning technique is more expensive. Thus, for cost reasons, it cannot be applied wall-to-wall across large areas and it is of interest to evaluate how samples of laser scanning data should be selected in large-scale forest surveys.

The study showed a clear trend of increased precision from using only field plots, through wall-to-wall satellite data and laser strip samples, to use of wall-to-wall data from both laser scanning and satellites. Further, it is found that the probability-proportional-to-size selection of laser scanning samples was superior to simple random sampling of laser scanning data.
The results of this study would be important for the development of forest inventories to meet the requirements for forest information from the global conventions.

The article Model-assisted estimation of growing stock volume using different combinations of LiDAR and Landsat data as auxiliary information will be published in Remote Sensing of Environment xxx (2014) xxx–xxx (in Press).

Photo and text: Svetlana Saarela