(Johnson and Vogel 1966; Nieland 1975; Van Cleve et
range of possible vegetation types over a short distance
al. 1983, 1986; Van Cleve and Viereck 1983; Viereck
(about 1 km). Ground-reference plots for ecosystem
descriptions (8-12 per transect) were located in dis-
et al. 1983, 1993). Taiga ecosystems are dominated by
open, slow growing spruce interspersed with occasion-
tinct vegetation types or spectral signatures identifiable
ally dense, well-developed forest stands and treeless
on aerial photographs. At each plot, we gave a basic
bogs. On the warmest, well-drained sites, the forests
descriptions of geology, hydrology, near-surface soil
consist of closed spruce-hardwood stands: white spruce
stratigraphy, permafrost occurrence, and vegetation.
(Picea glauca), paper birch (Betula papyrifera), and
Plots were located on aerial photography and coordi-
quaking aspen (Populus tremuloides). Productive for-
nates were obtained with a GPS. Field data sheets and
ests of balsam poplar (Populus balsamifera) and white
photos are archived at ABR, Inc.
spruce form along floodplains. On poorly drained sites,
Topographic profiles for each transect were obtained
including those underlain by permafrost and on north-
by measuring relative elevations at topographic breaks
facing slopes, the dominant forest species is black
along the length of the transects. Measurements were
spruce (Picea mariana). Bogs vary from rich sedge
made with an auto-level and rod or with a total station.
types to oligotrophic sphagnum bogs. Sedge-tussock
Because the transects were in remote locations, approxi-
meadows, with co-dominant low and dwarf shrubs, are
mate datums were obtained from the USGS maps. At
prevalent.
each sampling station, notations were made describing
surface-form and microrelief.
Hydrological observations included classification of
METHODS
the origin of water, water depth, depth to saturated soil
when water was not present in soil sampling pit, pH,
Field survey
electrical conductivity (EC), and temperature. Water
Field sampling in 1996 and 1998 was done accord-
quality measurements were made with Oakton or Cole-
Palmer pocket meters calibrated to standards within the
ing to two different sampling designs. Initially, in Sep-
range of use at regular intervals in the field. When water
tember 1996, we sampled 74 ground-reference plots
(approximately 100 m2) on seven transects (topo-
was not present, pH and EC were determined in a satu-
rated paste in a soil sample taken from 1020 cm depth.
sequences) using a gradient-directed sampling scheme
Soil stratigraphy was described from soil plugs dug
(Austin and Heyligers 1989). This design optimized the
with a shovel to approximately 50 cm using standard
likelihood of sampling the complete range of eco-
methods (SSDS 1993). Where possible, a soil core or
logical conditions and provided the spatial relationships
tile probe was used to extend the description and to
necessary for interpreting ecosystem development.
determine the depth to underlying gravel, if present.
Transect locations were stratified using the ecodistrict
Descriptions for each profile included the texture and
map to allocate the sampling to a range of physiographic
color of each horizon, the depth of organic matter, the
conditions. An additional 89 ground-reference plots
depth of thaw, the type and percentage of coarse frag-
were sampled subjectively in sites not represented along
ments, and the presence and character of mottling. All
the transects. In August 1998, we used a preliminary
profiles were photographed. To aid analyses, textural
unsupervised spectral classification of the Landsat
differences within a soil profile were grouped into a
image (see the Mapping section) to stratify sampling
single simplified texture (i.e., rocky, sandy, loamy,
of 126 less intensive verification plots. In addition, 89
clayey, or organic) for a site based on the dominant tex-
more-intensive ground-reference plots were established
ture in the top 50 cm.
to sample ecotypes that were under-represented in 1996.
Vegetation structure and composition were assessed
This sampling system was designed to over-sample rare
semi-quantitatively. Percentage cover of individual spe-
types and under-sample common types. Data from the
cies in a vegetation type was estimated visually to the
ground-reference plots were used for classifying eco-
nearest 5% if over 10% and to the nearest 1% if below
systems, identifying ecological relationships, and map-
10%. Dominant species were noted and a species list
ping. Data from the map verification plots were used
was assembled. Total cover of growth-form types (e.g.,
only for mapping.
tall shrubs, low shrubs, graminoids, etc.) was evalu-
The seven toposequences in the various ecodistricts
ated independently of individual species and cross-
were selected to cross the dominant geomorphic units
checked for accuracy. All sites were photographed. Most
in the study area: fluvial deposits (glaciofluvial outwash
species were identified in the field, and taxonomic no-
and other floodplains), glacial deposits (young and old
menclature followed Viereck and Little (1972) for
moraines), lowland eolian and retransported materials
shrubs, Hultn (1968) for other vascular plants, and Vitt
(lower slopes), upland slopes, and alpine tundra.
et al. (1988) for mosses and lichens. Unknown species
Transects were located in areas that maximized the
6
Previous Page
