Epsilon Open Archive

A new method for detailed estimations of local abave-ground woody biomass is presentad. The
procedure has been developed with the aim to support large and small scale strategic planning,
from various aspects of bio energy utilisation in Sweden. lmportant features of the method are
options to deal with areas defined by the user, and to include local harvest restrictions. A
planning tool suitable for local forest owners, municipalities, and bio energy enterprises has
previously not been available since eostume-made local estimates of woody biomass have
lacked sufficient precision or been considered to costly.
In the suggested method satellite image data and a sample of field plats from the Swedish
National Forest lnventory (NFI) are combined using the "k Nearest Neighbour" method (kNN).
The results are transferred to the farestry planning system Hugin which provides estimates of
the present and future potentials of forest bio-fuels. In the Hugin system there are great
possibilities to take into account various local interests and restrictions that influence the
potential of woody biomass utilisation. As an example, the priority and the intensity of farestry
treatments such as establishment of new stands, pre-commercial thinning, thinning, and final
cut can be considered. Other factors that can be incorporated are harvest restrictions for
specific areas like nature preservation areas or riparian zones. In addition minimum
requirements for harvest amounts per hectare can be set as general restrictions. In the initial
steps geographical objects like roads, railroads, or forests at a specified distance from roads
can be excluded from further calculations.
As a demonstration example the abave-ground woody biomass was calculated for the
municipality Älvsbyn in northern Sweden. The estimates were based on traditionally performed
cutting operations and results can be presented separatsly for different species by tree
fractions (bark, needles, branches and tops). The influence of distance to road was
demonstrated in a separate estimate where the maximum terrain transport distance was set to
300 m. Calculations were made for the present state of the forest and the state during the
following five ten-year periods.
In order to get an Relative root mean squared error (RMSE) for mean wood volume of <10%
using only the NFI data, the area for which the mean value is calculated must be at least
100.000 ha. With the presented method it will be possible to estimate wood volume and
biomass volume for areas as small as 200-450 ha with acceptable RMSE.