and organics supported closed balsam poplarwhite
The percentage of inconsistent plots (25%) was rela-
spruce forests, open white spruce forests, and open
tively high compared with the consistency of associa-
sprucepaper birch forests. The poorly drained soils on
tions (17%) obtained for Fort Wainwright (Jorgenson
the abandoned floodplain had thick organic accumula-
et al. 1999). We attribute this to the following:
tions and supported open black spruce forests and closed
The complexity of loess distribution and the glaci-
paper birch forests.
ated terrain, which made soil properties extremely
Ecosystem components
patchy.
The higher elevations, which had broader transition
We developed hierarchical relationships among eco-
system components by successively grouping data from
zones from closed canopy forests to woodland for-
survey plots by climate, physiography, soil texture, geo-
ests to alpine shrublands.
The abundance of burned areas in various succes-
morphology, slope position, drainage, permafrost, veg-
etation structure, and vegetation composition (Table 3).
sional stages.
Frequently, geomorphic units with similar texture or
genesis were grouped (e.g., loamy and organic were
During the development of generalized trends, our per-
grouped for some lowlands) to reduce the number of
spective was that it was better to preserve distinct, gen-
classes. Ecotypes then were derived from these tabular
eral trends, rather than include all the exceptions that
associations to differentiate ecotypes that have differ-
violate the trends, and thereby increase confusion
ent sets of associated characteristics (see the Ecotype
among classes. We believe that there is a limit to how
section for more detailed descriptions and analysis).
well patterns on the landscape can be described, and
This hierarchical grouping revealed that there were
that some proportion (in this case 25%) of sites cannot
close associations among soil texture, geomorphology,
readily be explained because they are transitional (eco-
slope position, drainage, and soils, and that often there
tones) or have historical factors (e.g., change in water
were several vegetation types that occurred on a geo-
levels, disturbances) that may cause the ecosystem (par-
morphic unit or soil type. These vegetation types gen-
ticularly soils) to vary unpredictably with current envi-
erally are associated because they occur along a suc-
ronmental conditions. The occurrence of these incon-
cessional sequence. For example, herbmoss, tall scrub,
sistencies provides a theoretical upper limit for the ac-
broadleaf forest, mixed forest, and needleleaf forest is
curacy of mapping of about 75%, because a certain
the typical successional sequence of vegetation devel-
portion of the landscape will not fit readily into any of
opment after fire (Foote 1983, Viereck et al. 1983).
the classes.
The successive grouping of ecosystem components
The advantage of this hierarchical approach is that,
helps differentiate many forest types. For example, as-
by combining physiography and vegetation structure,
pen generally was associated with upland areas and
the resulting classes are relatively good at differentiat-
gravelly lowlands, while balsam poplar generally was
ing soil characteristics and vegetation composition. This
restricted to riverine areas. Birch, white spruce, and
approach is particularly useful for mapping, where the
black spruce, however, were found over a wide range
interpreter can easily distinguish physiography (e.g.,
of conditions. For more detailed presentation of floris-
flat lowlands versus hilly uplands) and vegetation struc-
tic differences among ecotypes, see the discussion of
ture (e.g., needleleaf trees, broadleaf trees, shrubs, and
vegetation composition under the Ecotype section.
graminoids), whereas distinguishing tree species (e.g.,
A large question is how well these general relation-
birch versus poplar) or shrub species (e.g., dwarf birch
ships conform to the data set and whether they can be
versus willow) is difficult. Another advantage is that it
used reliably to extrapolate trends across the landscape.
links vegetation with soil characteristics. This linkage
During development of the relationships, 25% (63/252)
is particularly important for differentiating ecotypes that
of the field observations were excluded from the table
may have different sensitivities to disturbance. For ex-
because of inconsistencies among physiography, tex-
ample, lowland wet broadleaf forest (dominated by
ture, geomorphology, moisture, and vegetation. Some
paper birch) was almost always associated with ice-
of the main inconsistencies, or departure from the cen-
rich permafrost and, therefore, is susceptible to
tral concepts, included frequent grouping of rocky sites
thermokarst that can lead to irreversible development
with moist loamy ecotypes because of cutpoint prob-
of entirely different ecosystems after disturbance. In
lems associated with the 50-cm criteria used to define
contrast, upland moist broadleaf forest (also dominated
texture, occasional occurrence of moist sites in wet low-
by birch) almost always was associated with well-
land ecotypes, occasional absence of permafrost in wet
drained, thaw-stable soils and generally can recover to
lowland ecotypes, and occasional presence of perma-
similar ecological conditions a few decades after dis-
frost in moist upland ecotypes.
turbance.
18
Previous Page
