Canadian Forest Service Publications

Variation in understory and canopy reflectance during stand development in Finnish coniferous forests. 2015. Kuusinen, N.; Stenberg, P.; Tomppo, E.; Bernier, P.; Berninger, F. Can. J. For. Res. 45:1077-1085.

Year: 2015

Issued by: Laurentian Forestry Centre

Catalog ID: 36080

Language: English

Availability: PDF (request by e-mail)

Available from the Journal's Web site.
DOI: 10.1139/cjfr-2014-0538

† This site may require a fee

Mark record


Inherent variability in the spectral properties of boreal forests complicates the retrieval of canopy properties such as canopy leaf area index (LAI) from satellite images. Understanding the drivers of this variability could help provide better estimates of desired canopy cover properties. Field plot data from the Finnish National Forest Inventory (NFI) and Landsat TM images were used to investigate the variation in canopy and understory reflectance during stand development in coniferous boreal forests. Spectral data for each plot were obtained from the Landsat pixel within which the plot center coordinates fell. Nonlinear unmixing was used to estimate the bidirectional reflectance factors (BRFs) of the “sunlit understory” and “canopy and shaded ground” components by site fertility and stand development classes. A forest albedo model was used to estimate the contribution of diffuse radiation reflected downwards from the canopy to the sunlit understory component. The sunlit understory BRF in the near-infrared (NIR) spectral band decreased as the site fertility decreased and the forest matured, whereas the BRFs in the red and shortwave-infrared (SWIR) spectral bands concurrently increased. The BRFs of the canopy and shaded ground component decreased slightly during stand development, mostly in NIR. Adding the diffuse contribution to the sunlit understory component changed the estimated component BRFs only little (0.1-1.7%) compared to those obtained using a linear mixing assumption. This effect was largest in the NIR and smallest in the red spectral band. For spruce plots, the measured and estimated forest variables were well correlated with the BRFs in all the studied spectral bands, but for the pine plots the correlations were notably weaker. Results show a greater importance of the fraction of visible sunlit understory on forest reflectance in pine than in spruce.

Plain Language Summary

This study conducted in the boreal forest of Finland demonstrated that the reflectivity of light varies according to the developmental stage of stands and classes of soil fertility. This reflectivity, known as albedo, was measured for tree crowns and the forest floor visible between the trees. This effect also varies between species (pine and spruce). Forests influence climate by absorbing light and converting most of it into heat. By making albedo measurements, it is therefore possible to assess the importance of this climate impact.

Satellite imagery is the most effective way to measure albedo, but because a pixel (i.e., image “grain”) often covers several hundreds or thousands of square metres, it is difficult to attribute the measurement taken on a pixel to any of the types of surfaces found on the pixel. The study therefore helped to simplify the distribution of albedo measured on these various surface types.

These findings pave the way to albedo mapping of entire forests and, possibly, to the inclusion of albedo as a measurement of the interaction between forest management and climate.