Canadian Forest Service Publications

A strategy to map biomass of Canadian northern boreal forests using multi-sensor/multi-resolution imagery: approach, methods and preliminary results. 2007. Beaudoin, A.; Guindon, L.; Hall, R.J.; Luther, J.E.; Fournier, R. 5 pages in Proceedings of ForestSat 2007: Forests and Remote Sensing: Methods and Operational Tools, Nov. 5-7, 2007. Montpellier, France.

Year: 2007

Available from: Northern Forestry Centre

Catalog ID: 37062

Language: English

CFS Availability: PDF (download)

Abstract

The mapping of aboveground forest biomass is necessary for reporting on the state of Canada's forests. Undertaking this task over the non-inventoried regions of Canada's northern boreal forests is challenging due to its remoteness, vast area, and relative lack of available inventory information. To address this problem, we developed a biomass mapping strategy that was predicated on the integration of field and multi-sensor/multi-resolution satellite data. The strategy entails generating information about stand attributes and biomass from small numbers of ground sample plots that are subsequently related to the shadow fraction generated from high spatial resolution images. This process results in an image map from which satellite sample plots can be derived and scaled to coarser resolution imagery such as Landsat Thematic Mapper using a kNN process. This strategy was applied to pilot regions in 3 ecozones located in the Northwest Territories, Quebec, and Newfoundland and Labrador. Shadow fraction models of forest biomass were relatively similar across the 3 pilot regions and independent validations reported RMSE that ranged from 18 to 21 t/ha with bias less than 10 t/ha. These values are comparable or even less than those obtained from the combined use of coarser resolution Landsat and inventory data. Ongoing work is testing method applicability with other sensors and extending application of the strategy over additional species and larger geographic areas.

Date modified: