These vegetation types generally are associated
birch vs. willow) is difficult. Another advantage
because they occur along a successional sequence.
is that it links vegetation with soil characteristics.
For example, low scrub, tall scrub, broadleaf for-
This linkage is particularly important for differen-
est, mixed forest, and needleleaf forest is the typi-
tiating ecotypes that may have different sensitivi-
cal successional sequence of vegetation develop-
ties to disturbance. For example, Lowland Wet
ment after fire (Foote 1983, Viereck et al. 1983).
Broadleaf Forest (dominated by paper birch) were
The successive grouping of ecosystem compo-
almost always associated with ice-rich permafrost
nents helps differentiate many forest types. For
and thus are susceptible to thermokarst that can
example, aspen generally was associated with
lead to irreversible development of entirely differ-
upland areas and gravelly lowlands, balsam pop-
ent ecosystems after disturbance. In contrast,
lar generally was restricted to riverine areas, and
Upland Moist Broadleaf Forest (also dominated
tamarack generally occurred on lowland areas.
by birch) almost always was associated with
Birch, white spruce, and black spruce, however,
well-drained, thaw stable soils and generally
occurred over a wide range of conditions. For more
can recover to similar ecological conditions in a
detailed presentation of floristic differences among
few decades after disturbance.
ecotypes, see the Vegetation Composition section.
The main disadvantage to this approach is that
A large question is how well these general rela-
physiography or slope position is scale dependent
tionships conform to the data set and thus can
(e.g., a small raised area seen on the ground may
they be used reliably to extrapolate trends across
function as an upland even though it occurs
the landscape. During development of the rela-
within a broad lowland area) and this contributes
tionships, 17% (40/240) of the field observations
to uncertainty in classification and mapping. This
were excluded from the table because of inconsis-
problem with differentiation of physiography is
tencies among physiography, texture, geomor-
similar to that associated with the hydrogeo-
phology, moisture, and vegetation. Some of the
morphic classes (e.g., slopes, depressions, flats)
main inconsistencies, or departure from the cen-
developed by Brinson (1993). A second disadvan-
tral concepts, included frequent occurrence of
tage is that the grouping of the many ecological
moist sites (5/9) in Upland Wet Needleleaf For-
components can lead to generation of a large
est, frequent wet sites (3/9) in Riverine Moist Tall
number of classes. For practical purposes, the
Scrub, frequent moist sites (12/50) in Lowland
number of classes needs to be reduced by com-
Wet Needleleaf Forest, and problems in differen-
bining similar characteristics and ignoring unus-
tiating lacustrine (associated with pond margins)
ual plots that do not fit the simplified trends.
from lowland physiography (4/15 sites). In addi-
During the development of generalized trends
tion, we often had difficulty in differentiating low-
that ignored unusual situations, our perspective
land from upland physiography related to small
was is that it is better to preserve distinct, general
raised areas in the lowlands. Finally, we had to
trends, rather than include all the exceptions that
group thick (>40-cm) organic soils associated with
violate the trends and thus increase confusion
flat bogs and veneer bogs with thinner organic
among classes. We believe that there is a limit
deposits (grouped with loamy texture) associated
to how well patterns on the landscape can be
with abandoned floodplain cover deposits. This
accounted for. Some proportion (in our case 17%)
was because organic depths frequently were near
of sites cannot readily be explained, because they
this cut point and because we could not differen-
are transitional (ecotones) or have historical fac-
tiate the depths with aerial photography and sur-
tors (e.g., change in water levels, disturbances)
face observations.
that may cause the ecosystem (particularly soils)
The advantage of this hierarchical approach is
to be in poor adjustment with current environ-
that by combining physiography and vegetation
mental conditions. The occurrence of these incon-
structure, the resulting classes are relatively good
sistencies provides a theoretical upper limit for
at differentiating soil characteristics and vegeta-
the accuracy of mapping of about 8085%,
tion composition. This approach is particularly
because a certain portion of the landscape will not
useful for mapping, where the interpreter can eas-
fit readily into any of the classes.
ily distinguish physiography (e.g., flat lowlands
vs. hilly uplands) and vegetation structure (e.g.,
Ecosystems and landscapes
needleleaf trees, broadleaf trees, shrubs, and
Ecotypes. The final classification of ecotypes
graminoids), whereas distinguishing tree species
(local ecosystems) included 41 individual classes,
(e.g., birch vs. poplar) or shrub species (e.g., dwarf
36
Go to contents page
Previous Page
