&EPA
United States
Environmental
Protection Agency
National Health and Environmental EPA/620/R-05/001
Effects Research Laboratory January 2005
Corvallis, OR 97333
HYDROGEOMORPHIC WETLAND PROFILING: AN APPROACH
TO LANDSCAPE AND CUMULATIVE IMPACTS ANALYSIS
Environmental Monitoring and
Assessment Program
-------
-------
EPA/620/R-05/001
January 2005
HYDROGEOMORPHIC WETLAND PROFILING:
AN APPROACH TO LANDSCAPE AND
CUMULATIVE IMPACTS ANALYSIS
By
J. Bradley Johnson
Department of Biology
Colorado State University
Fort Collins, CO 80523
NATIONAL HEALTH AND ENVIRONMENTAL EFFECTS RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
NOTICE
This report was funded wholly by the U.S. Environmental Protection Agency (EPA)
under Cooperative Agreement number R-82843901 to the Colorado Geological Survey. The
research reported herein was conducted in collaboration with the EPA's Environmental
Monitoring and Assessment Program (EMAP) and the Regional Applied Research Effort
(RARE) Program.
This report has been subjected to review by the EPA National Health and Environmental
Effects Research Laboratory's Western Ecology Division and approved for publication.
Approval does not signify that the contents reflect the views of the Agency, nor does mention of
trade names or commercial products constitute endorsement or recommendation for use.
The correct citation for this document is:
Johnson, J. Bradley. 2005. Hydrogeomorphic Wetland Profiling: An Approach to Landscape
and Cumulative Impacts Analysis. EPA/620/R-05/001. U.S. Environmental Protection Agency,
Washington, D.C.
-------
ACKNOWLEDGMENTS
Many people contributed their time and expertise to this project. Foremost, I would like
to thank Richard Sumner of US Environmental Protection Agency's (EPA) Office of Research
and Development (ORD), who truly made this project possible and who generously provided
untold hours of advice, support, and insightful conversation. Mary E. Kentula of EPA ORD also
was unfaltering in her assistance, offering key insights, excellent suggestions, and editorial
expertise during project development and finalization. The efforts of D. Noe and the Colorado
Geological were essential to this project and much appreciated. My thanks to D. Steingraeber
for his continued collaboration. I would like to thank R. McEldowney of Science Applications
International Corporation, who supplied invaluable wetlands data and classified the majority of
wetlands into HGM classes. Summit County government, and H. McLaughlin were generous
with both their time and resources, and their contribution is greatly appreciated. Finally, I would
like to thank the participants who attended a workshop held at EPA Region 8's Denver office and
who significantly aided in the conceptualization of this project, including (but not limited to) B.
Bedford, D. Smith, S. Leibowitz, D. Noe, M. Claffey, D. Patten, G. Reetz, W. Schweiger, and G.
Rodriguez.
11
-------
Ill
-------
TABLE OF CONTENTS
NOTICE i
ACKNOWLEDGMENTS ii
EXECUTIVE SUMMARY vi
INTRODUCTION 1
Theoretical Basis for Approach 1
Background on HGM WP Development 3
Extension of HGM WP to the Generalized Case of Cumulative Impacts Analysis 4
METHODS 6
Description of Study Area 6
Development of the Geographic Information System 8
Designation of Process Domains 10
Identification of Ecoregions 11
Classification of Impact Categories 12
Generation of HGM WPs and Detection of Cumulative Impacts 12
Methodological Assumptions 14
RESULTS 16
Identification of Ecoregions 16
Low Lands Ecoregion 20
Middle-elevation Transitional Ecoregion 20
High Mountains Ecoregion 20
Classification of Impact Categories 20
Use of HGM WPs to Detect Cumulative Impacts 24
Evaluation of Consistency of HGM WPs Within Ecoregions 24
Detection of Cumulative Impacts 28
DISCUSSION 33
Reference-based Cumulative Impacts Analysis 34
Landscape Characterization and Indexing of Cumulative Effects 35
Utility and Limitations of HGM WP 37
Critical Evaluation of the Method 40
Insensitivity to Within-Wetland Impacts 40
Lack of Appropriate Reference Standard Landscapes 40
Categorization of Impact Level 41
Problematic Wetlands 42
IV
-------
Future Applications of HGM Wetland Profiling 42
Conclusions 43
REFERENCES CITED 45
APPENDIX 1: KEY TO SUMMIT COUNTY, COLORADO, WETLAND TYPES 53
-------
EXECUTIVE SUMMARY
Cumulative impacts and their resultant cumulative effects have become an important
focus of both environmental regulation and scientific investigation because of their potentially
severe consequences. For example, the National Environmental Policy Act (38 CFR Sect.
1500.6) and Section 404(b)(l) of the Clean Water Act (40 CFR 230.11) explicitly require that
projects proposing impacts to wetlands consider the cumulative effects of the actions and not
solely the direct impacts of the project. Despite the recognized potential for significant, negative
consequences, cumulative effects are seldom sufficiently addressed in environmental
management because of the lack of effective tools to describe and assess them.
This study developed a synthetic, hierarchical and scalable approach to landscape
characterization and cumulative impacts analysis that is based on current scientific thought of
how wetlands develop and function within landscapes. Specifically, this study investigated
whether a reference-based approach to cumulative impacts analysis could be developed by using
a hydrogeomorphically-based version of landscape profiles (Brinson 1993, Bedford 1996, Gwin
et al. 1999) in conjunction with the concepts of landscape formation and processes forwarded by
Montgomery (1999), Winter (2001), and Omernik and Bailey (1997). This approach, termed
Hydrogeomorphic Wetland Profiling (HGM WP), seeks to refine coarse acreage-based
approaches by applying a functionally-based framework to provide a landscape-scale
characterization of wetland resources that is useful to regulatory, management, mitigation, and
conservation programs.
Most basically, HGM WP is a method of summarizing the abundance and diversity of
HGM wetland types within a given ecoregion or portion thereof. Three related applications of
the HGM WP approach are detailed in this report, landscape characterization, cumulative
impacts analysis, and first-order approximation of wetland-related cumulative effects.
Landscape characterization
This research asserts that HGM WP provides a valuable means of characterizing of
landscapes with regard to their wetland component and the functions occurring within. Using a
functionally-based classification, such as HGM, to summarize the abundance and diversity
wetlands provides an estimation of the potential types and magnitude of wetland functions
performed in that landscape. For instance, a landscape with a high proportion of slope wetlands
would be expected to predominately perform the functions ascribed to slope wetlands, such as
groundwater discharge, carbon retention, and maintenance of stream base flow.
Cumulative Impacts Analysis
The main focus of the HGM WP approach is to provide a practically implementable,
reference-based approach to wetland cumulative impacts analysis. Specifically, HGM WP
quantifies impacts to the abundance and diversity of wetlands resulting from outright wetland
destruction or functional conversion (i.e. from one HGM class to another). This analytical
focuses stems from the idea that these whole-wetland impacts are the primary driver or "forcing
factor" of wetland functioning at the landscape scale.
VI
-------
The key principle underlying HGM WP in a reference-based capacity is that the
abundance and diversity of wetlands is dictated by the physical setting of the landscape (Bedford
1996). Consequently, it follows that landscapes possessing similar physical attributes should
possess similar patterns of wetland abundance and diversity. That is, physically similar
landscapes would be expected to have similar HGM WPs. Conversely, physically disparate
landscapes should have distinctly different HGM WPs. If these hypotheses are true, then
landscape references appropriate for cumulative wetland impact analyses may be defined
through analysis of comparable, minimally-impacted landscapes. This study tested the validity
of these hypotheses and tested the viability of the HGM WP approach in cumulative impacts
analysis.
First-order Approximation of Cumulative Effects
While some agencies and investigators have chosen to synonymize cumulative impacts
and cumulative effects (e.g., CEQ 1997), it is important to differentiate the terms. Following
Leibowitz et. al. (1992), cumulative wetland impacts are defined as the sum of wetland impacts
that have occurred across a given landscape, while cumulative effects are the resultant
environmental ramifications.
Since wetland profiling utilizes a functionally-based wetland classification system, I
suggest that evaluation of the changes in the relative abundance of HGM wetland classes
provides an index of the loss of the specific functions associated with the wetlands destroyed.
Moreover, since loss of wetland function is the root of wetland-mediated cumulative effects,
evaluation of HGM WP alterations provides a picture of the potential types of cumulative effects
present in the landscape. The information yielded through HGM WP analyses can then be used
to guide more data-intensive, quantitative studies of actual cumulative effects.
This study began by investigating the consistency of HGM WPs within and between
ecoregions found in Summit County, Colorado. It then evaluated whether the method could
detect differences in HGM WPs between landscapes classified a priori as reference standard
(minimally-impacted) or impacted. Specifically, in this report:
1. Landscapes in Summit County, Colorado were delineated and classified according to
their physical setting and level of land cover alteration.
2. The hypothesis that landscapes have inherent, characteristic HGM WPs resulting from
their physical composition of the landscape setting was evaluated.
3. The HGM WPs found in landscapes classified a priori as reference standard or impacted
were compared to determine whether this approach is sufficiently sensitive to detect
cumulative wetland impacts.
vn
-------
4. The use of HGM WP as a means of first-order cumulative effects estimation was
discussed along with the method's limitations and its potential for further refinement.
Summit County is located in the heart of the Rocky Mountains in central Colorado. The
county is characterized by rugged terrain and strong physical and biological gradients. Much of
Summit County is undeveloped, and 76% is under the management of the U.S. Forest Service.
The county was partitioned into 95 process domain (sensu Montgomery 1999) sample units in
which physical and ecological processes were internally homogeneous. Based on cluster
analysis, the 95 process domains were aggregated into three ecoregions: 1) low lands, 2) middle-
elevation transitional, and 3) high mountains.
Since no data exist on actual wetland impacts that have occurred in Summit County, land
use/land cover (LULC) and road density were used to index the likelihood or severity of
cumulative wetland impacts. Process domains were classified into one of two, a priori impact
categories (reference standard or impacted) based on LULC and road density using cluster
analysis.
Two primary questions were evaluated during this study: 1) considering only minimally-
impacted, reference standard landscapes, are HGM WPs relatively consistent within ecoregions
and do they differ between ecoregions?; and 2) within ecoregions, do HGM WPs statistically
differ between impacted and reference standard landscapes as a result of cumulative wetland
impacts?
The results of this investigation support both hypotheses within the study area. First,
HGM WPs within the study units were found to be more similar within than between ecoregions.
Between ecoregions, profiles differed in overall shape or in the relative abundance of two or
more wetland classes. This result suggests that ecoregions have characteristic and discernable
HGM WPs. This is a key finding since there had not been an empirical evaluation of how tightly
the abundance and diversity of wetlands is tied to the physical setting. Because of the
demonstrated linkage between physical setting and HGM WP, this approach provides a plausible
means of reference-based cumulative impacts analysis.
Comparison of reference standard and impacted landscape units (process domains) within
ecoregions showed that profile alterations could be statistically detected. Profile differences
were manifested as an overall change in profile shape or in an alteration in the ratio between two
or more wetland classes. Changes to HGM WPs in impacted landscapes could be readily tied to
current and historical land use patterns.
These findings show that HGM WP is a promising method with which to characterize
cumulative wetland impacts. Further, as a result of its design, HGM WP results can be combined
with data derived through smaller-scale approaches to yield multi-scale analyses of wetland
resources. The method also seems a useful means of addressing additional facets of landscape
Vlll
-------
analysis of wetlands, including wetland-based landscape classification, threshold detection, and
synoptic analyses.
The report is presented in four major sections. The first is this executive summary which
outlines the project's approach and major findings. The second provides an introduction to the
concepts underlying hydrogeomorphic wetland profiling including background on the regulatory
and scientific context of the method. In the third section, a generalized approach to HGM WP is
described, implemented, and tested in Summit County, Colorado. Lastly, a tool for remotely
classifying wetlands is provided as an appendix.
IX
-------
INTRODUCTION
Cumulative impacts and their resultant cumulative effects have become an important
focus of both environmental regulation and scientific investigation because of their potentially
severe consequences. For example, the National Environmental Policy Act (38 CFR Sect.
1500.6) and Section 404(b)(l) of the Clean Water Act (40 CFR 230.11) explicitly require that
projects proposing impacts to wetlands consider the cumulative effects of the actions and not
solely the direct project impacts. Despite the recognized potential for significant negative
consequences, cumulative effects are seldom adequately addressed in environmental
management because of the lack of effective tools to describe and assess them. The purpose of
the research documented in this report was to explore how recent, innovative concepts in
wetland ecology could be integrated to create a practical and powerful approach to wetland
cumulative impacts analysis.
Specifically, this study investigated whether a reference-based approach to cumulative
impacts analysis could be developed by using a hydrogeomorphically-based version of landscape
profiles (Brinson 1993, Bedford 1996, Gwin et al. 1999) in conjunction with the concepts of
landscape formation and processes forwarded by Montgomery (1999), Winter (2001), and
Omernik and Bailey (1997). This approach, termed Hydrogeomorphic Wetland Profiling (HGM
WP), seeks to refine coarse acreage-based approaches by applying a functionally-based
framework that provides the landscape-scale view of wetland resources needed by regulatory,
management, mitigation, and conservation programs.
Theoretical Basis for Approach
Scientific investigations have shown that wetlands unquestionably perform important
environmental functions (National Research Council [NRC] 1995, Mitsch and Gosselink 2000)
and that different types of wetlands perform different functions or the same functions to various
degrees (e.g., Brinson 1993, NRC 1995). Thus, with loss and degradation of wetlands comes a
concomitant loss of functions generally associated with wetlands and the particular
environmental functions attributed to specific wetland types. The functional losses result in both
direct and indirect negative effects on environmental quality such as impairment of water
quality, reduction of flood flow attenuation, and loss of wildlife habitat (Hemond and Benoit
1988, Croonquist and Brooks 1991, Council on Environmental Quality [CEQ] 1997, Bedford
-------
1999, McAllister et al. 2000, NRC 2001). When taken singly, any particular wetland impact
may have little effect on overall environmental quality. But considering whole watersheds and
basins, the cumulative sum of wetland impacts may have significant additive effects, or wetland
impacts can act synergistically to produce disproportionately severe cumulative effects (Hemond
and Benoit 1988, Nestler and Long 1997).
While some agencies and investigators have chosen to synonymize cumulative impacts
with cumulative effects (e.g., CEQ 1997), it is important to differentiate the terms. Following
Leibowitz et. al. (1992), cumulative wetland impacts are defined as the sum of extant and
historical wetland impacts that have occurred across a given landscape, while cumulative effects
are the resultant environmental consequences. Cumulative wetland impacts can take the form of
outright destruction of wetland habitat, of functional conversion (e.g., converting riparian
wetlands to depressions), or of a decline in wetland functioning through mechanisms such as
sedimentation, hydrologic alteration, or logging. Alternatively, the cumulative effects of
wetland loss are manifested as a degradation of environmental quality. One way of contrasting
cumulative wetlands impacts with cumulative effects, is that cumulative wetland impacts always
occur within wetlands across a landscape, while cumulative effects are commonly manifested
outside of wetland ecosystems, in receiving waters.
Analysis of wetland acreage trends is the most basic method of characterizing cumulative
wetland impacts. Such analyses provide data on wetland losses (or gains) and thereby provide
an indication of the general types of ecosystem functions that have been lost within the study
region. While an important means of tracking broad environmental trends, such an approach
suffers from its inherent generality and insensitivity to differential wetland functioning, and it
does not provide the level of detail necessary for smaller-scale regional studies. These short-
comings are especially significant in light of the non-random distribution of wetland impacts
(Bedford 1999), which make accurate predictions about wetland-related functional losses and
cumulative effects unlikely.
The information rendered through trend analyses can be greatly increased by stratifying
surveyed wetlands into discrete categories. Wetland classifications can be structured around any
group of parameters, but recent studies suggest that hydrogeomorphic (HGM) classification
(Brinson 1993) is particularly powerful because of its physical basis and its direct ties to wetland
-------
Loss or reduction
of wetland
functional types \
(Cumulative ^
Wetland Impacts)
Loss of the
specific
___^\ functions or ^_______
\ functional suites }
!/ associated with
affected wetland
types
Wetland-
^X related
T/ cumulative
effects
Figure 1. Flow diagram showing the relationship between cumulative wetland impacts
and related cumulative effects. Hydrogeomorphic class (slope, riverine, etc.) is one way
to define functional types of wetlands.
functioning. It stands to reason that evaluation of the changes in the relative abundance of HGM
categories provides an index of the loss of the specific functions associated with the wetlands
destroyed, and that such functional losses are the causative factors behind wetland-mediated
cumulative effects (Fig. 1). Thus, evaluation of wetland trends in terms of HGM categories
seems a promising way of tracking cumulative impacts and indexing wetland-related cumulative
effects. Hydrogeomorphic wetland profiling provides a functionally-based method for
improving trend analyses and investigating cumulative impacts and their resultant effects.
Background on HGM WP Development
Bedford (1996) argued that surveys of the abundance and diversity of wetland
"developmental templates" is an important means of quantifying the wetland diversity of
landscapes. She suggested that tallies of wetland templates in a landscape could be displayed as
simple diagrams which she termed "landscape profiles"1, and that these profiles could be a
valuable means of summarizing wetland diversity and tracking cumulative impacts. Bedford did
not provide specifics about how templates should be parameterized and classified, however.
Gwin et al. (1999) empirically applied the concept of wetland profiling to the evaluation
of the effects and effectiveness of wetland mitigation resulting from Clean Water Act permit
requirements. In developing their approach, Gwin et al. (1999) took advantage of the conceptual
1 Although Bedford applied the term "landscape profiling", I will instead use the term
"hydrogeomorphic wetland profiling" throughout this report, since it makes the focus of the
technique more explicit.
-------
similarities between Bedford's templates and Brinson's HGM classification framework, using
HGM classes in place of Bedford's templates. They then compared a profile of natural wetlands
to that of mitigation wetlands. In this specific case, the natural HGM WP was taken as the
reference with which to compare the HGM WP of mitigation wetlands. The departure between
these two profiles was used as the measure of cumulative impacts due to mitigation actions.
Using this approach Gwin et al. (1999) characterized cumulative wetland impacts within a
portion of the Willamette River watershed, but more generally they showed that HGM WP was
useful and applicable in a management setting.
Extension of HGM WP to the Generalized Case of Cumulative Impacts Analysis
The key principle supporting the idea of HGM WP in a reference-based capacity is that
the abundance and diversity of wetland types is dictated by the physical setting of the landscape
(Bedford and Preston 1988, Winter and Woo 1990, Winter 1992, Bedford 1996, Halsey et al.
1997, Winter 2001). According to these studies the physical attributes most relevant to wetland
formation are local and regional climate, and basin hydrology, geomorphology, and
hydrogeology. Similar findings have been presented in regard to aquatic systems as well (Frisell
et al. 1986, Richards et al. 1996, Johnson and Gage 1997, Kratz et al. 1997, Wiley et al. 1997,
Montgomery 1999). Thus, it follows that landscapes possessing similar physical attributes
should consequently possess similar patterns of wetland abundance and diversity, when diversity
is summarized in terms of physical makeup. That is, physically similar landscapes would be
expected to have similar HGM WPs. Conversely, physically disparate landscapes should have
distinctly different HGM WPs. If these assertions are true, then landscape references appropriate
for cumulative wetland impact analyses may be defined through characterization of comparable,
minimally-impacted landscapes.
The ideas described above form the basis for the work presented in this report. This
study began by investigating the consistency of HGM WPs within and between physically-based
ecoregions. It then evaluated whether the method could detect differences in HGM WPs
between landscapes classified a priori as reference standard or impacted. Specifically, in this
report:
1. Landscapes in Summit County, Colorado are delineated according to their physical
setting and impact level;
-------
2. The hypothesis that landscapes have inherent, characteristic HGM WPs that result from
the physical composition of the landscape setting is evaluated;
3. The HGM WPs found in landscapes classified a priori as reference standard or impacted
are compared to determine whether this approach is sufficiently sensitive to detect
cumulative wetland impacts;
4. And finally, the ways in which the HGM WP can be used as a index of cumulative
effects, and the limitations and potential further extensions of the approach are discussed.
-------
METHODS
Description of Study Area
The study area for this investigation was Summit County, Colorado. Summit County
covers approximately 1,600 km2 of the Rocky Mountains in central Colorado (Fig. 2).
Following the Continental Divide in the southwest, the county boundary corresponds to that of
the Blue River watershed, although a small portion continues beyond the northern tip of the
county. The watershed is dissected by three major alluvial valleys formed by the Blue River,
Snake River and Ten-mile Creek. Rimming these valleys and nearly surrounding the county are
high mountains of the Gore, Ten-mile, and Mosquito Ranges and the William's Fork Mountains.
Peaks in these ranges reach elevations over 4,250 m.
The county is characterized by rugged terrain and strong physical and biological
gradients. The high mountains are comprised mainly of granites, granodiorite dikes and sills,
and Precambrian metamorphic rocks, that have been carved by Pleistocene glaciers. Only a few,
small active glaciers exist today. On the shoulders of the mountains, deposit! onal features
dominate the landscape mainly in the form of moraines and outwash plains. The major valleys
are dominated by alluvial landforms, and contain the topographically flattest areas.
Strong patterns in climate, vegetation, and land use follow the physiographical gradients
that are present. Long-term climate data are not available for the high mountain areas, but at
2,920 m average annual precipitation is 48.7 cm (Breckenridge Station), while at 2,359 m it is
38.8 cm (Green Mountain Dam Station). Vegetational zones present in the county range from
montane in the lower Blue River Valley, to the high alpine in the surrounding mountains (Marr
1961). As is typical in mountain environs, vegetational zonation is pronounced and tightly
controlled by topography, elevation, and the associated climatic changes. Low in the alluvial
valleys and plateau areas, montane vegetation is dominated by grasslands and sagebrush
shrublands. In the upper montane zone, open conifer woodlands and mixed aspen-conifer forests
cover most of the landscape. The subalpine zone is typical of that found across most of the
Colorado, being strongly dominated by spruce-fir (Picea engelmannii-Abies lasiocarpa) forest
(Peet 1981). Above tree-line (-3,490 m) in the alpine zone, vegetation consists of a mix of
meadows, krumholtz and low-shrub woodlands.
-------
105ฐ0'0"W
12 Kilometers
Figure 2. Map showing Colorado (inset) and the Summit County study area.
-------
Much of Summit County is undeveloped, and 76% is under the management of the U.S.
Forest Service. Only twenty-two percent of the land is privately owned, with these properties
being strongly concentrated in the open and relatively flat valley bottoms, at the Climax
molybdenum mine in the southwest, and in the immediate vicinity of the four major ski resorts
scattered across the southern half of the county.
Development of the Geographic Information System
A geographic information system (GIS) was constructed using Arc View version 8.2
software (ESRI2002) and incorporating data from numerous sources (Table 1). All GIS data
were projected into the Universal Transverse Mercator system using the North American Datum
of 1927.
Wetland polygons were derived from a compilation of three aerial photograph surveys of
Summit County (Table 1). Identified wetlands were placed the one of the HGM classes:
riverine, slope, depressional or lacustrine fringe (henceforth "fringe"). A fifth class of
"wetlands" was included in the analyses - "irrigated meadows". Broad interpretation of these
areas is problematic. Irrigated meadows are commonly associated with natural wetlands wherein
irrigation waters greatly expand the extent of naturally occurring hydric conditions. Commonly
interspersed within the this mix of natural and irrigation-supported wetland are expanses of
upland meadow not practicably discernable on aerial photographs. The ecological role of
irrigated meadows is also difficult to interpret since irrigated meadows are indicative of land use
alteration and impact, but such areas commonly perform many beneficial wetland functions.
Owing to practical limitations these sites were aggregated into an artificial category.
Wetland polygons mapped by the Whitehorse and Ward aerial photograph surveys (Table
1) were placed into HGM classes by the Science Applications International Corporation (SAIC)
using aerial photograph interpretation, GIS, and field surveys (SAIC 2000). I classified wetlands
identified by the U.S. Forest Service aerial photograph survey into HGM classes using a
dichotomous keying algorithm based on GIS-interpreted physical attributes (Appendix 1). In a
randomly chosen set of test wetlands, a 95% concurrence was found between these two methods.
Based on this comparison, the results obtained through either approach were deemed practically
equivalent and the data sets were combined.
-------
Table 1. Description and sources of geographic data included in the study's GIS.
Data layer
Description
Citation or Source
USGS 7.5' Digital raster graphics Topographical maps
(DRGs)
1:250,000 scale geologic maps Denver and Leadville Quadrangles
Soil survey geographic data
(SSURGO)
State soil geographic database
(STATSGO)
Hydrography
Roads
Land use and land cover
Digital elevation model (DEM)
Partial coverage, large-scale soils
data
1:250,000 scale soils data
US EPA-BASINS Streams and
waterbodies
All roads within Summit County as
surveyed by the county
US EPA-BASINS Anderson Level
II land cover classes
10m resolution
http://www.lighthouse.nrcs.usda.go
v/gateway/gateway home.html
http ://greenwood. cr.usgs.gov/pub/
mf-maps/mf-2347 and
http://greenwood.cr.usgs.gOv/pub/o
pen-file-reports/ofr-99-0427,
respectively.
http://www.ftw.nrcs.usda.gov/ssurg
o/metadata/co690.html
http://www.ftw.nrcs.usda. gov/stat_
data. html.
http://www.epa.gov/OST/BASINS/
Summit County Government
http://www.epa.gov/OST/BASINS/
Summit County Government
Sub-watershed boundaries
White River National Forest aerial
photograph survey
Summit County private land aerial
photograph survey
Town of Silverthorne aerial
photograph survey
Draft Hydrologic Unit Code
(HUC)12 shapefiles
Wetland polygons on USFS land
classified by vegetation
Wetland polygons on private lands
classified by vegetation coverage
Wetland polygon in and around the
Town of Silverthorne, classified by
vegetation coverage.
US Natural Resource Conservation
Service (not yet generally released)
USFS White River National Forest
Field Office, Silverthorne, CO
Whitehorse survey. Data obtained
through Summit County
Government
Ward survey. Data obtained
through Summit County
Government
-------
Topographic slope, total basin relief, and mean elevation were calculated using a 10 m
digital elevation model. Stream order was determined by Strahler's (1957) method using
1:100,000 digital line graphics (DLGs) and was completed by Summit County Government (H.
McLaughlin, personal communication).
Designation of Process Domains
Summit County is composed of a complex mosaic of landscape types. To effect a
analytical comparison of HGM WPs, the county had to be divided into ecologically relevant and
relatively homogeneous sample units. Summit County is entirely included within a single
hydrologic unit code (HUC) 8 watershed that is divided into 62 HUC 12 sub-watersheds
(henceforth HUC 12s). These HUC 12s were used to produce a preliminary division of the
county into objectively defined sampling units. The HUC 12 layer was laid over shade relief and
geologic data layers in ArcScene to provide a 3-dimensional geologic/geomorphic view of the
region. Geologic units were grouped into functional groups reflective of their effects on
hydrogeology and geomorphology (Table 2).
Table 2. Summary description of geologic functional groups.
Group Name
Lithological
types Included
Total %
coverage
Sandstones and
shales
clastic
sandstones
various
sandstones
shales
20
Sandstones,
carbonates, and
siltstones
Various
sandstones
carbonates
conglomerates
siltstones
7
Unconsolidated
Glacial
deposits
Landslides
features
Colluvium
Alluvium
23
Volcanic
Trachytic
lava
Extruded
lava
0.4
Granite and
metamorphic
Granites
Precambrian
metamorphic
Grandodiorite
dikes and sills
49
10
-------
In spite of their relatively modest areal extent, in Summit County HUC 12 sub-
watersheds commonly encompass significant physical heterogeneity which can confound
attempts to link ecological, physical, and land use patterns (Montgomery 1999). To facilitate
ecologically meaningful analyses, sub-watersheds were used in conjunction with ecologically
defined boundaries (Omernik and Bailey 1997). Often sub-watersheds were adequately
homogenous to consider as a whole, but when HUC 12s included marked physical heterogeneity,
they were subdivided into relatively homogeneous process domains (sensu Montgomery 1999).
These units could also validly be termed fundamental hydrologic landscape units following
Winter (2001), although process domain terminology is preferred here since it conveys the
dynamic nature of wetland formation and maintenance.
When the requirement of homogeneity was not met, HUC 12s were subdivided into
preliminary process domains by partitioning the sub-watersheds at marked breaks in geology and
geomorphology. Not surprisingly, geomorphic breaks followed shifts in geology in almost every
case. Principle Components Analysis was used to evaluate these preliminary process domain
boundaries based on maximum stream order, percent coverage of geologic functional units, total
basin relief, mean elevation, and mean slope in each domain. This analysis was used to identify
outliers and sub-optimal process domain boundaries. When necessary, process domain
boundaries were reevaluated and revised based on this analysis. SPSS (2003) version 12
software was used to carry out all statistical analyses.
Identification of Ecoregions
The purpose of delineating process domains was to divide the study region into units
within which the degree of topographic, geologic, hydrologic and climatic heterogeneity were
relatively even. By its nature, such a delineation presupposes that physical and ecological
differences exist among the various process domains. Cursory examination of the diversity
among process domains bore out this supposition. Thus, all of the individual process domains
formed the inclusive study population, which itself could be stratified into categories based on
predominant physical and ecological characteristics.
Process domain categories were defined using agglomerative, hierarchical clustering
utilizing Euclidian distance and group centroids (McKune and Grace 2002). The variables
employed during the analysis were the same ones used during the delineation of process
11
-------
domains. Because the process domain categories were delineated such that each possessed
similar climate, landforms, soils, potential natural vegetation, and hydrology, I call these
categories ecoregions following Bailey (1995), Omernik (1995) and others. Discriminant
analysis (DA) was used to test the robustness of this quantitative, but ultimately subjective,
ecoregion categorization. Ecoregions formed the basis of comparison and statistical analysis,
with each ecoregion being represented by several replicate process domains.
Classification of Impact Categories
No data exist on actual wetland impacts that have occurred in Summit County. In lieu of
such data, land use/land cover (LULC) and road density were used to index the likelihood or
severity of cumulative wetland impacts. Land use/land cover was the preferred index because
certain types of land cover are known to affect wetlands and water quality (NRC 2001, Johnson
et al. 1997). Land use/land cover was characterized using Anderson Level II resolution data and
classes (Anderson et al. 1976). Areal extents of LULC classes were converted to relative
proportions. Land use/land cover classes were placed into two groups, those LULCs which
commonly result in wetland loss ("impacting") and those which are essentially "natural"
management regimes ("reference standard") (Table 3). Road density was calculated from the
GIS, as meters of road per hectare.
Process domains were classified into a priori impact categories based on LULC and road
density using the cluster analysis procedures described above. Process domains were
categorized as "reference standard" if minimal alterations and development had occurred, and
"impacted" when significant land cover conversions were present. This simple, two-category
classification was found to provide the clearest, most readily interpretable results during
preliminary analyses. Discriminant analysis was used to test the robustness of this grouping.
Generation of HGM WPs and Detection of Cumulative Impacts
Recall that an HGM WP summarizes the relative abundance of HGM (functional)
wetland types in a unit of the landscape. The first step in generating the HGM WP was to tally
the areal coverage of HGM classes within each process domain using native Arc View
geoprocessing routines. Data on areal coverage by HGM class were then compiled in a
12
-------
Microsoft Excel spreadsheet and grouped by ecoregion. HGM WPs were generated for each
ecoregion and then assigned an impact status as determined by the LULC analyses.
Table 3. Grouping of Anderson Level II
land cover class into reference standard or
impacted land use categories.
Management
Class
Anderson Level II LULC classes
"Natural" land
cover types
Bare exposed rock
Bare ground
Deciduous forest land
Evergreen forest land
Forested wetland
Herb rangeland
Herb tundra
Lakes
Mixed forest land
Mixed rangeland
Mixed tundra
Mixed urban or built-up
Non-forested wetland
Shrub rangeland
Shrub tundra
Altered land
cover types
Commercial and services
Industrial
Other urban or built-up
Reservoirs
Residential
Strip mines
Transportation, communications,
utilities
Other agricultural land
Cropland and pasture
Multivariate general linear modeling (MGLM) was used to compare mean, proportional
coverage of HGM classes, i.e., to compare HGM WPs. Two types of comparisons were made:
13
-------
1) comparison of reference standard HGM WPs between ecoregions; and 2) comparison of HGM
WPs between impact classes within ecoregions. The first comparison was used to test the
hypothesis that HGM WPs are relatively consistent within ecoregions. The second comparison
tested the hypothesis that within ecoregions HGM WPs differ between reference standard
landscapes and those subjected to environmentally disruptive land uses. Statistical results were
considered significant at the p < 0.05 level. When multiple statistical comparisons were made,
significance values were Bonferroni corrected.
Methodological Assumptions
Hydrogeomorphic wetland profiling is designed to provide a functionally-based, scale-
appropriate characterization of cumulative wetland impacts. Specifically, HGM WP quantifies
impacts to the abundance and diversity of wetlands resulting from outright wetland destruction
or functional conversion (i.e. from one HGM class to another). This analytical focus stems from
the idea that these whole-wetland impacts are the primary driver or "forcing factor" of wetland
functioning at the landscape scale.
In developing the HGM WP approach it was necessary to make a number of assumptions
based on first principles and logical constructs. These assumptions listed below, have not been
explicitly tested in this study.
1). Cumulative impacts to wetlands cause degradations of environmental quality and
ecological integrity which are termed cumulative effects.
2). The particular manifestation and types of cumulative effects incurred depend on the
specific functional wetland types that are impacted.
3). Alteration of land cover and land use stemming from civil development, conversion to
agriculture and natural resource utilization results in loss and functional conversion of
wetlands.
4). Cumulative impacts have accumulated over time since settlement by Europeans,
however, the rate and spatial distribution of impacts has been uneven and episodic.
14
-------
5). The temporal distribution of wetland impacts does not change affect the manifestation or
severity of cumulative effects. For example, given an equivalent loss of wetland acreage,
it is insignificant in terms of cumulative effects whether those impacts accumulated over
100 years or 10 years.
15
-------
RESULTS
Identification of Ecoregions
Based on cluster analyses, the ninety-five process domains were grouped into three
ecoregions: 1) low lands, 2) middle-elevation transitional, and 3) high mountains (Fig. 3). In the
dendrogram presented here, the clusters of the middle-elevational transition and high mountain
ecoregions are coarsely interspersed. Non-adjacent clusters were included within the same
ecoregion based on the results of preliminary cluster analyses that used different distance
measures and which generally grouped the separated clusters together, and also on subjective
evaluation of process domain characteristics. The groupings as reported here were found to be
the most statistically parsimonious and provide the clearest, most interpretable results. For
comparison, Fig. 4 presents a shade-relief map of raw HUC 12 sub-watersheds laid over regional
geology, contrasted with a map of process domains grouped into the three ecoregions. Table 4
provides the mean values of physical variables measured in each ecoregion.
Discriminant analysis (DA) validated the final ecoregion groupings. Using proportional
coverage of geologic units, total basin relief, mean elevation, and mean slope - the same
parameters used in the cluster analysis - process domains were assigned to the a priori
ecoregions with 100% accuracy. Table 5 provides a correlation matrix of the discriminant
variables with the first two standardized discriminant functions. Mean process domain slope and
elevation are most strongly, positively correlated with function 1, whereas geologic parameters
were most highly correlated with function 2. Exposure was also highly correlated with
discriminant functions but it was not included within these analyses owing to its strong
intercorrelation with other parameters such as slope and basin relief. Note that vegetational
composition was not explicitly included within this classification framework, although there is a
strong correspondence between ecoregions and vegetational zones.
Each of the physical parameters included within ecoregion analyses have fundamental
effects on landscape hydrogeology, geomorphology, and disturbance processes, which in turn
constrain the ecological character of those areas. Each ecoregion is briefly characterized below.
16
-------
Low Lands
Low Lands
Low Lands
Low Lands
High Mountains
Middle Elevation
Transitional
High Mountains
Middle Elevational
Transitional
Figure 3. Dendrogram of a hierarchical, aglomerative cluster analysis, grouping
process domains into ecoregions based on mean elevation, mean slope, basin
relief, and geology.
-------
A
Summit_County_boundary
HUC 12 Boundaries
Water
Sandstones and shales
Sandstones, Carbonates, and Siltstones
Unconsolodated
Volcanic
Granites and Metamorphic
B
j j Summit_County_boundary
| | HUC 12 Boundaries
^^^J Middle Elevation Transitional
| | High Mountains
0 2.5 5 10 Km
I i i i I i i i I
N
Figure 4. Geology and HUC 12 boundaries (A) in comparison to process domains and ecoregions (B) in Summit
County, CO. See the Methods section for additional explanation of boundary designation.
-------
Table 4. Summary description of physical properties of the three ecoregions.
Ecoregion
n
Average process domain size (km2)
Sandstones and shales (%)
Sandstones, carbonates, and
Unconsolidated (%)
Volcanic (%)
Granite and metamorphic (%)
Area (ha)
Total basin relief (m)
Mean elevation (m)
Mean Slope (%)
Low lands
41
143
43
2
42
0
12
58,687
1,484
2,818
33
Middle-elevation High mountains
transitional
25
191
12
2
18
1
67
47,813
1,739
3,247
42
29
184
2
18
6
0
73
53,465
1,632
3,514
49
Table 5. Discriminant analysis of process domain physical
characteristics showing the pooled within-groups correlations between
discriminating variables and standardized canonical discriminant
functions.
Discriminant Function
Geomorphic
Parameters
Mean elevation
Mean slope
Basin relief
Stream order
0.879
0.332
0.145
-0.076
-0.123
0.158
-0.133
0.034
Geologic Units
Sandstones and shales
Sandstones, carbonates, and
siltstones
Unconsolidated
Volcanic
Granite and metamorphic
-0.207 0.604
0.116 0.581
0.307 -0.384
-0.161 -0.343
-0.036 -0.163
19
-------
Low Lands Ecoregion
The Low Lands Ecoregion typically includes low elevation areas with relatively shallow
topographical gradients. Although mean process domain elevation can reach to over 3000 m in
the northeastern areas, high elevation portions of these domains have more affinity to lower
elevations owing their southwest exposure (Geiger 1965). Process domains grouped in this class
are typically sited on sandstone deposits, glacial terrain, or alluvium. Although all regional
vegetational zones are represented to varying degrees in this ecoregion, it is primarily associated
with the montane zone.
Middle-elevation Transitional Ecoregion
The Middle-elevation Transitional Ecoregion is spatially and characteristically
intermediate between the low lands and high mountains, but it is more allied to the high
mountain systems. Process domains in this ecoregion are located in heavily glaciated, middle-
elevation, mountainous landscapes. Commonly, this ecoregion includes high, intermountain
valley systems. Middle-elevation transitional areas are geologically heterogeneous. Granites
and Precambrian formations dominate the geology, but glacial and sandstone deposits are also
common. Portions of process domains classified within this ecoregion may reach the alpine
zone, but the ecoregion is most strongly associated with the subalpine zone.
High Mountains Ecoregion
The High Mountains Ecoregion is found in the highest, most rugged settings in the upper
subalpine to alpine zones. Its geology is dominated by resistive Precambrian granites and
metamorphic rocks that have been subjected to extensive Pleistocene glaciation. The high
mountain areas possess the greatest basin relief and highest mean slope of the three ecoregions.
Classification of Impact Categories
The highest level division in the land use cluster analysis was used to assign process
domains into reference standard or impacted categories. Process domains classified as reference
standard have less than 10% cover of intensive land uses, such as urban and residential
development, and they generally have road densities below 20 m/ha (Fig. 5). Many of the
reference standard process domains are entirely undeveloped and essentially roadless, notably
20
-------
those in the Eagle's Nest Wilderness Area along the western edge of the county. The remainder
of process domains were classified as impacted. The validity of this classification was cross-
checked using discriminant analysis and only five of 95 samples were misclassified (94.7%
accuracy). The five "misclassified" process domains were examined and kept in their original
category as assigned by cluster analysis, although these were truly borderline cases. As
expected, impacted process domains are mainly those in or adjacent to the Blue River Valley,
which contains the most readily developable land (Fig. 6).
21
-------
Legend
Low lands
Middle Elevation Transitional
High Mountains
Impacted
N
036 12km
I i i I i i I
Figure 5. Map of Summit County, CO showing the distribution ecoregions. Hatching
indicates those ecoregions that were classified as impacted. All others are reference
standard.
-------
Figure 6. A comparison of mean percent coverage of altered land
cover types (%ACT) and road density in reference standard and
impacted process domains. Refer to Table 3 for Anderson level II
land use classes included within the ACT category. T-bars show
one standard deviation.
-------
Use of HGM WPs to Detect Cumulative Impacts
The detection of cumulative impacts was done in two steps that evaluated the hypotheses
posed in the Methods Section. First, reference standard HGM WPs were compared between
ecoregions to determine whether HGM WPs were more consistent within than between
ecoregions. Once this had been confirmed, then I determined whether cumulative impacts could
be detected by comparing HGM WPs from reference standard and impacted landscapes within
an ecoregion.
Evaluation of Consistency of HGM WPs Within Ecoregions
Multivariate general linear modeling (MGLM) shows that reference standard HGM WPs
differ significantly between ecoregions (Table 6). Examination of between-subject effects
indicates that coverage of fringe wetlands was the only parameter that did not differ significantly
between the three ecoregions (Table 7). This pattern can be seen in Figure 7 which compares the
reference standard HGM WPs developed for each of the three ecoregions.
Summit County landscapes are heavily biased towards riverine and/or slope wetlands,
with other classes being almost incidental to the profiles. Clear patterns in the occurrence of
HGM types within the ecoregions, particularly in the coverage of these two classes, are evident
in Fig. 7. To evaluate the significance of coverage patterns, Bonferroni corrected, multiple pair-
wise comparisons of wetland functional class coverages were made between ecoregions (Table
8). Coverage of riverine wetlands varied significantly between the Middle-elevation
Transitional and High Mountain Ecoregions. Although there is an apparent decline in mean
riverine coverage between low lands and middle-elevation transitional ecoregions, this
difference is not statistically significant (Table 8). The inverse trend is seen in slope wetlands,
with such wetlands becoming increasingly common in the higher elevation ecoregions. All
differences in slope wetland coverage between ecoregions were significant.
Examination of between-subject effects indicates that coverage of fringe wetlands was
the only parameter that did not differ significantly between the three ecoregions (Table 7).
24
-------
Table 6. Results of a MGLM analysis comparing the reference standard HGM
WPs between ecoregions. The Wilk's Lambda F statistic is included here, but
other multivariate statistics yielded identical significance values.
Effect
Intercept
Ecoregion
Value
0.022
0.255
F
753.404
16.391
Hypothesis df
4.000
8.000
Error df
67.000
134.000
Sig.
0.000
0.000
Table 7. Results of a MGLM analysis of effects between HGM classes. Analysis is based on
proportional coverage of HGM classes.
Source
Ecoregion
Intercept
Error
Dependent Variable
Riverine
Slope
Depressional
Fringe
Irrigated meadow
Riverine
Slope
Depressional
Fringe
Irrigated meadow
Riverine
Slope
Depressional
Fringe
Irrigated meadow
Type III Sum
of Squares
18581.941
45244.010
132.187
10.220
6965.248
101391.366
200935.979
171.211
36.113
3213.334
24296.025
21734.199
1203.916
344.640
14179.071
df
2
2
2
2
2
1
1
1
1
1
70
70
70
70
70
Mean Square
9290.970
22622.005
66.094
5.110
3482.624
101391.366
200935.979
171.211
36.113
3213.334
347.086
310.489
17.199
4.923
202.558
F
26.768
72.859
3.843
1.038
17.193
292.122
647.161
9.955
7.335
15.864
Sig.
0.000
0.000
0.026
0.360
0.000
0.000
0.000
0.002
0.008
0.000
25
-------
120
100
0)
O)
(0
0)
>
o
o
4->
ฃ
0)
o
0)
Q.
Low lands
Middle Elevation
Transitional
High Mountains
HGM Class
Figure 7. HGM wetland profiles from reference standard process domains grouped by
ecoregion. HGM classes are arrayed along the x-axis. Within individual HGM class
clusters, columns with shared letters are not significantly different. T-bars are one standard
deviation.
-------
Table 8. Results of multiple pair-wise comparisons of the proportional coverage of
each wetland class within ecoregions. Middle-elevation Transitional Ecoregion has
been abbreviated as MET. All results have been Bonferroni corrected.
Dependent
Variable
(proportional
coverage)
Riverine
Slope
Depressional
Fringe
Irrigated meadow
Ecoregion
Low lands
MET
High Mountains
Low lands
MET
High Mountains
Low lands
MET
High Mountains
Low lands
MET
High Mountains
Low lands
MET
High Mountains
Ecoregion
MET
High mountains
Low lands
High mountains
Low lands
MET
MET
High mountains
Low lands
High mountains
Low lands
MET
MET
High mountains
Low lands
High mountains
Low lands
MET
MET
High mountains
Low lands
High mountains
Low lands
MET
MET
High mountains
Low lands
High mountains
Low lands
MET
Mean
Difference
9.3
36.1
-9.3
26.8
-36.1
-26.8
-26.8
-57.9
26.8
-31.0
57.9
31.0
-2.6
0.8
2.6
3.4
-0.8
-3.4
0.1
0.8
-0.1
0.7
-0.8
-0.7
20.0
20.1
-20.0
0.1
-20.1
-0.1
Std. Error
5.5
5.0
5.5
5.6
5.0
5.6
5.2
4.7
5.2
5.3
4.7
5.3
.2
.1
.2
.2
.1
.2
0.6
0.6
0.7
0.7
0.6
0.7
4.2
3.9
4.2
4.3
3.9
4.3
Sig.
0.290
0.000
0.290
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.121
1.000
0.121
0.026
1.000
0.026
1.000
0.559
1.000
0.819
0.559
0.819
0.000
0.000
0.000
1.000
0.000
1.000
27
-------
These patterns are readily interpretable considering the hydrogeologic setting of each
ecoregion. For example, high in the mountains, hydrologic pathways are short, bedrock
fractures reside near the surface, and slope breaks which commonly result in the day-lighting of
groundwater are common. These conditions are highly favorable to the formation of slope
wetlands, and these wetlands form the headwaters of a multitude of small stream systems, which
in turn support riverine wetlands. Lower in elevation, the shift to predominantly riverine
wetlands continues as large, high-order river systems form through the coalescence of the small
streams. Sites conducive to the formation of slope wetlands become less common, as the terrain
flattens and surficial geology becomes more complex.
Relatively few depressional wetlands exist in the study region (Fig. 7). In reference
standard areas, depressional wetlands are usually associated with small pator noster lakes and
kettle ponds. The low lands and middle-elevation transitional areas do not differ in coverage of
this wetland class, apparently since the low lands include the large outwash plains where most
kettle ponds are located, and middle-elevation transitional areas have both kettle ponds and pator
nostor lakes. The coverage of depressional wetlands was significantly higher in middle-
elevation transitional areas as compared to high mountain areas, and very few depressional
systems are found high in these mountains. None of the ecoregions differed in their coverage of
lacustrine fringe wetlands. In all cases, naturally occurring fringe wetlands are rare in this
mountainous landscape.
These results corroborate the hypothesis that landscapes that are similar with respect to
hydrogeology, geomorphology, and climatic conditions possess inherent and consistent HGM
WPs and that profiles differ between disparate landscapes. These conclusions support the use of
HGM WP in reference-based evaluation of cumulative wetland impacts. They also provide
empirical support for theoretical assertions arguing that physical landscape attributes dictate the
abundance and diversity of wetlands on the landscape.
Detection of Cumulative Impacts
As demonstrated above, reference standard portions of ecoregions have characteristic
HGM WPs. It follows that cumulative impacts manifested as wetland loss or functional-type
conversion modify these characteristic HGM WPs, but it was uncertain whether such alterations
would be statistically detectable owing to naturally occurring variation.
28
-------
To determine the ability of this approach to detect cumulative wetland impacts, I
evaluated the hypothesis that HGM WPs differ between reference standard and impacted process
domains within ecoregions. For analysis, process domains were first grouped by ecoregion and
then by impact class (i.e., reference standard or impacted). Figure 8 provides a comparison of
these data. Using multivariate GLM, HGM WPs from impacted and reference standard
ecoregions were found to differ statistically in the Low Lands and High Mountains Ecoregions,
but not in the Middle-elevation Transitional areas (Table 9).
Examination of patterns of individual HGM class coverages (between-subjects effects)
show that in the Low Lands, the coverage of riverine, fringe and irrigated meadows differed
between impact classes. In the Middle-elevation Transitional ecoregion, riverine and slope
wetland coverages were statistically different, while in the High Mountains Ecoregion, riverine,
depressional and fringe wetlands differed (Fig. 8).
Examination of these data suggests two main mechanisms working to alter profiles. The
first is direct loss of wetlands through conversion to upland. Such losses are not directly
tabulated in these analyses, but are evident through changes in proportional distributions of
wetland classes. One way to evaluate changes in relative data is through ratio analysis. In the
heavily riverine and slope wetland-biased region of Summit County, those wetland classes are
the obvious choices for ratio comparisons. In all ecoregions, the ratio of riverine to slope
wetlands differed between reference standard and impacted areas (p <0.10), showing that ratio
alteration is one of the manifestations of cumulative wetland impacts. A caution is necessary in
the interpretation of ratio analyses, however - if two or more classes of wetland are impacted to
the same degree, no difference in ratio will be detected. In the case of Summit County, this
situation seems improbable, though, owing to land use patterns, and elsewhere the selective
nature of wetland impacts have been noted (Bedford 1999). Still this potentiality should be
considered in the interpretation of results.
29
-------
120
a 100
| 80
O 60
| 40
^ 20
0
Low Lands
HGM Class
Middle Elevation Transitional
HGM Class
High Mountains
i
1
HGM Class
Figure 8. Comparison of HGM wetland profiles in reference standard and impacted
process domains within each of the three ecoregioins. A * indicates a significant
difference in mean HGM class coverage between reference standard and impacted
process domains. T-bars are one standard deviation.
-------
Table 9. Results from a multivariate general linear modeling analyses
comparing HGM WPs in reference standard and impacted process domains
grouped by ecoregion. The Wilk's Lambda F statistic is included here, but
other multivariate statistics yielded identical significance values.
Hypothesis
Effect Value F df Error df Sig.
Low Lands
Intercept
Impact Level
0.139
0.620
55.651
5.528
4.000
4.000
36.000
36.000
0.000
0.001
Middle-elevation Transitional
Intercept
Impact Level
0.000
0.744
1315050.838
1.723
4.000
4.000
20.000
20.000
0.000
0.184
High Mountains
Intercept
Impact Level
0.000
0.664
67852.817
4.207
3.000
3.000
25.000
25.000
0.000
0.015
A second type of cumulative wetland impact is conversion of wetlands from one
functional class to another. The best examples of this are seen in the Low Lands Ecoregion. In
reference standard process domains, fringe wetlands are quite uncommon, but in impacted
process domains the coverage of fringe wetlands is elevated five-fold (Fig. 8). Irrigated
meadows show this same general trend. Based on the proportional reduction in riverine
wetlands, the data suggest that riverine wetlands have been converted to these two atypical or
artificial wetland types. Examination of land use patterns corroborates this assertion. Reservoir
construction and gravel mining activities have been significant in these areas and irrigated
meadows are also commonly found associated with streams, which facilitate water transport and
delivery to the cultivated lands.
Similar conclusions are made with regard to wetland functional conversion in the High
Mountains Ecoregion. In those areas, coverage of depress!onal and fringe wetlands have been
inflated, again, apparently at the cost of riverine wetlands. In the Middle-elevation Transitional
Ecoregion, impacts seem to have been focused on slope wetlands. In reference standard areas,
there is nearly an even ratio between riverine and slope wetlands, whereas in impacted process
domains, the relative coverage of riverine wetlands is inflated relative to that of slope wetlands.
Once again, this pattern can be explained in light of land use patterns. All of the impacted
process domains contain major ski resorts. As opposed to most mountain land development
which occurs in relatively flat river valleys, ski resort development expressly occurs away from
11
-------
such flat lands, on high elevation mountain sides where slope wetlands are most frequently
located. Impacts to slope wetlands through ski resort development are a commonly noted and
disturbing side-effect of the industry (US Army Corps of Engineers Omaha and Sacramento
Districts CWA, ง404 permit records).
In Fig. 8 (Middle-elevation Transitional), the profile further indicates an increase in the
percentage of riverine wetlands. This should not be interpreted as an actual increase in the
number or coverage of riverine wetlands but rather an increase in the relative percentage of
riverine wetlands. Namely, since the percentage of all wetland coverage must sum to 100%, a
decrease in the relative coverage of one class must be balanced by an increase in another.
32
-------
DISCUSSION
Despite the existence of generic frameworks and host of specific methods (e.g., Preston
and Bedford 1988, Brooks et al. 1989, Leibowitz et al. 1992, Stein and Ambrose 2001, Hauer et
al. 2001), the scientific and regulatory communities still wrestle with how to most effectively
evaluate cumulative wetland impacts and their resultant effects. Hydrogeomorphic wetland
profiling is forwarded as an means of improving cumulative wetland impacts analyses by linking
them to current concepts of wetland development and functioning. This approach provides a
means of wetland-based landscape characterization and cumulative impacts analysis that also
affords a first approximation of potential wetland-mediated cumulative effects.
By partitioning Summit County Colorado into three physio-ecologically based
ecoregions, it was determined that reference standard examples of each ecoregion had a
characteristic and statistically discernable HGM WP. This is a key finding since there had not
been an empirical evaluation of how tightly the abundance and diversity of wetlands is tied to the
physical setting (Bedford 1996, Winter 2001). Nor had there been an indication of whether
theoretical patterns in wetland occurrence could be quantitatively discerned owing to natural
variation in landscapes. This finding also underscores the necessity of using ecologically-based
units such as process domains, fundamental hydrologic landscape units or ecoregions in the
design of management strategies, assessment tools, and ecological studies alike. Just as in the
case of socio-political boundaries, ad hoc use of watershed boundaries to answer ecologically-
based questions, such as those of wetland diversity and function, aquatic life use characteristics,
habitat abundance can provide a muddled or even misleading picture of large-scale trends. This
is not to suggest that watershed-based analyses are inappropriate for addressing many other types
of management and research problems, however.
Because reference standard HGM WPs were found to be consistent within ecoregions, I
suggest that reference standard landscapes can validly be contrasted with physically comparable,
impacted landscapes to determine the extent of cumulative impacts in the central Rocky
Mountains. This finding would have little value, however, if HGM WP differences between
reference standard and impacted landscapes were not detectable owing to natural systemic
variation. In direct comparisons within ecoregions, HGM WPs from reference standard
landscapes were found to be significantly different from profiles derived from impacted
33
-------
landscapes. Thus it is concluded that the approach also is sensitive enough to detect cumulative
impacts.
Reference-based Cumulative Impacts Analysis
The concept of reference bears some additional consideration since it is so fundamental
to the HGM WP approach. To account for historical wetland losses and circumvent the lack of
historical data, HGM WP follows current thinking in landscape and wetland assessment by
adopting a reference-based approach. In particular, it compares ecologically similar landscapes
in an manner akin to the time-for-space substitutions common to investigations of vegetation
succession (White and Walker 1997, Pickett 1989). Comparative approaches such as this have
been extremely productive in such diverse pursuits as watershed analyses (McCulloch and
Robinson 1993), wetland evaluations (Brinson 1993, US EPA 1998, Gernes and Helgen 2002,
SAIC 2000), stream assessments (Karr et al. 1986, Karr and Chu 1997, Hughes et al. 1986), and
studies of forest succession (Pickett 1989).
Here, profiles from a target landscapes are contrasted with ecologically similar,
minimally-impacted (reference standard) landscapes. The main conceptual difference between
the HGM WP reference approach and that of successional studies is that the relationship between
time and accumulation of impacts cannot be described by a simple deterministic model. For
instance, severe impacts could have accumulated in an area over a long period of time, or they
could have been brought about quickly during a period of intense urban growth and
development. This technique does not differentiate between these two scenarios, and therefore
explicitly assumes that the resultant environmental effects are similar whether cumulative impact
accumulation was slow, rapid, historical or recent.
Comparative approaches such as this do possess inherent limitations, however
(MacDonald 2000, White and Walker 1997). Considering paired-watershed studies of
cumulative impacts, MacDonald (2000) challenged that "The implicit assumption is that the
[contrasted] systems are part of the same population and therefore directly comparable. The
corollary is that any differences are presumed to result from past disturbance." In particular, for
a referenced-based approach to be useful, the parameter of interest must be relatively invariant
within categories, while at the same time being detectably different among categories. This is
echoed in Rheinhardt et al.'s (1997) discussion of critical assumptions of the HGM functional
34
-------
assessment method in which they stated that "Ecological processes [must be] so similar in form
and magnitude within any narrowly defined regional subclass that they shape biotic and abiotic
components in ways characteristic for the subclass."
Thus, in order for HGM WP to be used in a referenced-based manner it must be
demonstrated that landscapes are validly comparable and that within comparable landscapes
HGM WPs are relatively consistent, being more similar to one another than they are to profiles
from physically dissimilar landscapes. Theoretical constructs support the idea of HGM WP
constancy within physically similar landscapes, since climate, hydrology, geomorphology and
hydrogeology drive the abundance and diversity of wetlands. This study provides corroborating
empirical support for this hypothesis.
In this study, landscape comparability was achieved through division of the study region
into physically similar ecoregions. Comparability was validated by calculating the HGM WP
variance within reference standard ecoregions. Unacceptably high variance within reference
standard ecoregions would have indicated that either HGM WPs are not related to
hydrogeomorphic conditions (unlikely), or that ecoregions were inappropriately designated -
that is, they consisted of a mixed landscape types or "populations". Neither was true in this case,
therefore, the uniqueness of the individual ecoregions was established.
An important offshoot of this conclusion is that, as suggested by Bailey and Omernik
(1997), predictions with a calculable error rate can be made about the relative abundance and
diversity of wetland functional types in landscapes for which no detailed wetland survey
information exists. Significantly such predictions can be derived solely through the analysis of
geographic data, most of which are readily available as digital files on the World Wide Web and
elsewhere.
Landscape Characterization and Indexing of Cumulative Effects
Ecological functions have been usefully attributed to individual wetlands, and it would
seem that analogous benefits could be derived from evaluating the wetland functioning of whole
landscapes. An obvious approach to evaluating the wetland functioning of a landscape would be
to use a functional assessment methodology such as HGM to evaluate all or a statistical
population of wetlands in a prescribed area, and subsequently model the total landscape capacity
35
-------
for each wetland function. This approach would be a simple extrapolation of the methods
described in Smith et al. (1995). But such large-scale evaluations of wetland functioning would
not generally be tractable to owing issues including labor intensiveness and cost, and they would
only be feasible when circumstances dictated a highly detailed approach.
Because wetland profiling utilizes HGM categories (here functional classes), the method
can be used to broadly characterize the wetland functions associated with various landscape
types. For instance, High Mountain areas in Summit County are dominated by slope wetlands
with relatively fewer riverine wetlands. Therefore, one can infer that such areas primarily
perform those wetland functions associated with slope wetlands such as groundwater discharge,
maintenance of stream base flow, and carbon retention (Johnson 2000). Utilization of HGM
subclasses, rather than classes, would increase the analytical resolution of such a
characterization.
Characterization of landscapes in terms of wetland functioning seems a useful endeavor
of itself, but this concept can also be extended to the realm of cumulative effects analysis. Per
the definitions used in this report, HGM WP is not a cumulative effects assessment method, but
it can provide an index of probable wetland-related cumulative effects and it can help focus
monitoring, restoration and analysis efforts on variables most likely to be impaired owing to
wetland losses. For example, if a HGM WP analysis were to show that the proportion of
depress!onal and fringe wetlands has been increased at the expense of riverine wetlands, one
could thereby infer that at the landscape-scale, the composite functional capacity of riverine
wetlands had been reduced. That is, such landscapes would be less able to perform functions
attributed to riverine wetlands including sediment export or provision of critical wildlife travel
corridors (Hauer et al. 2002, Brinson et al. 1995). At the same time, such landscapes would
have an increase in the functions provided by depressional and fringe wetlands, most notably
surface water storage (Hauer et al. 2000).
It is important to note that, like functional approaches to wetland evaluation, valuation of
landscape alterations is not considered by HGM WP. For instance, in spite of detectable
cumulative impacts and their presumed cumulative effects, such a situation may be deemed
acceptable or even desirable in a given socioeconomic setting. In the Summit County, this is
well illustrated in the case of reservoir construction, which creates fringe wetlands at the expense
of slope and riverine wetlands (Fig. 8).
36
-------
Utility and Limitations of HGM WP
Hydrogeomorphic wetland profiling provides a means of characterizing the population of
wetlands found in landscapes based on their potential functioning. Moreover, HGM WP
analyses provide a coarse-scale view of wetland functioning and cumulative impacts within
landscapes. In particular, the method can be seen as evaluating cumulative impacts to the
abundance and diversity of wetland functional types. This scale of analyses is insensitive to
fine-scale patterns, such as the condition of any single wetland. In effect the method filters such
information. This is a beneficial characteristic since fine-scale data can clutter large-scale
analyses and hinder recognition of large-scale patterns (Holling 1992).
An important ramification of this tactic is that HGM WP only provides an index of
potential landscape-scale wetland functioning; the actual functioning or condition of individual
wetlands is left implicit. While at first seeming a limitation, this strategy is intentional. It was
adopted to maintain fidelity to the method's primary focus and aim while explicitly
acknowledging the multi-scale complexities of landscape analyses.
As it has commonly been noted, the most efficient way to evaluate a complex system is
to focus analyses on the primary driver(s) acting at the desired scale (Warren 1979, Frissell et al.
1986, Brinson 1993, Bedford 1996, Montgomery 1999). In general, a particular process or
characteristic is only of preeminent importance at a single scale of observation. For instance,
basin geomorphology is a primary driver of wetland functioning (Brinson 1993). When one
considers the mechanisms creating wetland functions, however, geomorphology falls to the
background forming a contextual factor which generally only indirectly affects specific
mechanisms such as biotic interactions or nutrient cycling. The existence, diversity, and relative
abundance of wetland functional types is seen here as the primary control of wetland functioning
at the landscape scale. That is, if a wetland has been eliminated from a landscape, it can perform
no beneficial ecological functions; all other forms of impacts are subordinate to this single fact.
37
-------
This is not to say that within-wetland impacts are unimportant. Rather, such impacts are of a
secondary importance at the landscape scale.
Thus, HGM WP adheres to analysis of cumulative wetland impacts at a single
organizational scale, while at the same time recognizing that cumulative impacts or effects can
be evaluated at multiple scales and using a variety of approaches. Such a strategy has important
practical ramifications. In leaving fine-scale characteristics implicit, the methodology remains
open-ended and able to incorporate fine-scale data, facilitating multi-scaled analysis. Data
obtained through a variety of means such as regional inventories, or assessment surveys using
any one of the multitude of approaches available such as HGM, indices of biologic integrity, the
Wetland Evaluation Technique, or the Habitat Evaluation Procedure can be directly incorporated
into HGM WP diagrams. The only requirement for inclusion of such data is that surveyed
wetlands be classified into HGM categories. This is a simple and straightforward task even after
surveys have been completed (Appendix 1). A hypothetical example of multi-scale profile
construction is shown in Fig. 9.
38
-------
D Poor Condition
D Moderate Condition
DGood
^2-
HGM Class
Figure 9. A hypothetical example of how fine-scaled data could be incorporated into an
HGM WP analysis. In this example, the condition of wetlands is determined through any
number of site-based approaches such as HGM or indices of biological integrity. Wetlands
must be classified into HGM categories (here classes) before they can be integrated into
the analysis. Within each HGM class, the relative proportion of wetlands within each
conditional category are color-coded to indicate the aggregate condition of wetlands. In
this example, the majority of riverine wetlands are in poor condition. Multi-scale profiles
from target landscapes such as this one could also be compared to reference profiles.
-------
Critical Evaluation of the Method
Hydrogeomorphic wetland profiling has been shown in this study to be a powerful tool
for the wetland-based characterization of landscapes, assessment of cumulative wetland impacts,
and indexing of cumulative effects. There are certain aspects of the method which could be
improved upon with further development or which need to addressed before widespread
utilization of the method is possible.
Insensitivity to Within-WetlandImpacts
As discussed above, HGM WP only measures changes in the relative abundance and
diversity of wetland types. As a stand-alone method it is wholly insensitive to impacts occurring
within wetlands. Thus when used to evaluate cumulative wetland impacts it is susceptible to
Type II errors, when the null hypothesis is "no cumulative impacts in the target landscape". To
illustrate, given that a HGM WP does not vary statistically from its reference standard, two
situations could exist: 1) the landscape could be intact insofar as wetland functioning is
concerned; or 2) wetlands have not been outright lost or functionally converted, but there have
been within-wetland impacts that have reduced the capacity of individual wetlands to function at
their potential.
This form of Type II susceptibility is an unavoidable consequence of the method's single
scale focus, but it can be addressed through the inclusion of small-scale data as explained above.
I further speculate that such insensitivity is not a widely significant issue since it would seem
likely that landscapes possessing unaltered profiles also possess functionally intact wetlands.
The converse of this situation seems equally likely, as well.
Lack of Appropriate Reference Standard Landscapes
Summit County was in many ways an ideal area to develop and test this approach since it
encompasses a wide range of physical settings and it possesses many minimally-impacted
landscapes. Clearly these characteristics are not commonplace in many areas, particularly in
urbanized regions such as the east coast of the US. Probably the most significant impediment to
widespread application of this approach is the lack of minimally-impacted (reference standard)
landscapes. This is not a problem unique to this methodology, but to every reference-based
analytical technique. There is no easy solution to this problem, although various solutions may
be on the horizon.
40
-------
A promising means of determining reference standard conditions could be modeling
expected HGM WPs based on analysis of hydrogeological and/or hydrogeomorphic conditions.
McLaughlin (1999) for instance, developed a model to predict the occurrence of wetlands in the
Ridge and Valley province of Pennsylvania. If such techniques can successfully be developed,
models of expected profiles could be compared to the observed profiles on the landscape.
Of course, another way around this issue is to develop a contemporary HGM WP for a
given landscape and monitor it into the future, thereby avoiding the need for space-for-time
references. In this case, the reference standard condition is the extant landscape condition.
While valuable for monitoring programs this approach would not address the issue of historical
impacts.
Categorization of Impact Level
MacDonald's (2000) corollary statement that any differences are presumed to result from
past disturbance remains an explicit assumption in this approach. No long-term data exist as to
actual wetland impacts in Summit County. Even information in the form of Clean Water Act
ง404 permit applications were found to be of little worth in describing recent wetland impacts
since they were so often incomplete or contained erroneous information (H. McLaughlin,
Summit County, Colorado Government - Mapping Department, personal communication). This
was particularly true of project applications approved early in the program's existence. To
address this data gap, LULC data were used as an index of wetland impacts. Although there is
significant precedent for such an approach (e.g., Allan et al. 1997, Bolstad and Swank 1997,
Herlihy et al. 1998, Lenat and Crawford 1994), the actual relationship between land cover
alteration and wetland impacts was not explicitly formulated. Forming a tighter link between
land cover alteration and wetland impacts might significantly improve the accuracy of this
method and utilization of other indicators of wetland impacts, such as normalized difference
vegetation indices (NDVIs) may prove advantageous in this regard (Griffith et al. 2002).
Along these same lines, process domains were herein simply classified as "impacted" or
"reference standard". Obviously, impact level is not a simple binary function, but rather it
occurs as a gradient. In this study, binary classification was adequate and produced clear results;
however, with additional investigation, more refined categories could be delineated which could
improve the method's resolution and accuracy.
41
-------
Problematic Wetlands
Lastly, in Summit County the prevalence of irrigated meadows created pragmatic
difficulties. Such systems are to varying degrees artificial. But in the Rocky Mountains, these
sites are usually associated with natural wetland systems, and because of the effects of irrigation,
large portions of these meadows may be artificial wetlands, performing many of the same
functions of natural wetlands (Johnson 2000). Given the practical constraints of this project, I
did not find a wholly satisfying way to deal with these areas. Similar classification difficulties
would likely occur should the method be applied in other regions.
Future Applications of HGM Wetland Profiling
Hydrogeomorphic wetland profiling seems a promising means of tackling many facets of
landscape-scale wetlands analyses. With further development, the method seems particularly
suited to address three additional management needs landscape classification, synoptic
analyses (sensu Leibowitz et al. 1992), and threshold detection.
Hydrogeomorphic wetland profiling provides a succinct means of characterizing the
abundance and diversity of wetland types within landscapes. Because of the utilization of HGM
categories, it also provides an index of landscape-scale wetland functioning. Given these
characteristics, the approach provides a natural basis from which to construct wetland-based
landscape classification frameworks. For instance, in terms of this study, Summit County
landscapes could be qualitatively classified as slope-dominated, riverine-dominated, or slope-
riverine co-dominated. Given the quantitative nature of the approach, categories could also be
explicitly parameterized.
Next, the synoptic approach developed by EPA researchers is intended to prioritize
restoration, protection, and permitting efforts in light of almost universal resource shortages
(Leibowitz et al. 1992, Abbruzzese and Leibowitz 1997, McAllister et al. 2000). The approach
consists of a generic methodology that can incorporate data from a number of different sources
depending on availability. After being initially proposed and tested, the methodology has been
refined and implemented (McAllister et al. 2000). Being essentially an approach framework, the
method requires the inclusion of indices of wetland function, value, functional loss and
replacement potential. The exact form of these indices is not specified and can vary from
application to application. Hydrogeomorphic wetland profiling results would seem to be
42
-------
particularly appropriate for synoptic analyses, since they can index both wetland functioning and
functional loss within landscapes.
Even removed from external approaches such as the synoptic, HGM WP analyses would
seem beneficial to wetland regulatory and management programs. Consider a hypothetical
situation in which HGM WP analyses indicate that slope wetlands have been significantly
reduced in a landscape relative to appropriate references, and further that fringe wetland
coverage has been artificially increased. In such a situation, limited agency resources could be
allocated such that permit applications proposing impacts to slope wetlands would receive
priority status, whereas applications for fringe wetlands could receive lesser scrutiny. A similar
usage could be envisioned for conservation or restoration activities, as well.
Finally, identification of thresholds of wetland impacts at which cumulative effects
become significant has been a longstanding goal of scientists and managers (Preston and Bedford
1988), yet there has been little success in this endeavor. Utilizing HGM WPs in conjunction
with detailed approaches that quantitatively evaluate environmental variables such as water
quality, it seems feasible that the thresholds of profile alteration could be identified at which
measured variables exhibit a detectable, negative response.
Conclusions
Ecoregions within the mountains of central Colorado were found to possess inherent,
characteristic and discernable HGM WPs based on analysis of reference standard landscapes.
Because of this consistency, it was suggested that HGM WP may be usable in a reference-based
manner to evaluate cumulative wetland impacts occurring within developed landscapes. Within
ecoregions, profile differences between reference standard and impacted landscapes were
significant, being manifested as an alteration to the overall shape of the HGM WP, or as
differences in the relative abundance of particular functional classes.
Hydrogeomorphic wetland profiling is a promising approach to cumulative wetlands
impact analysis. The method targets a single analytical scale, while acknowledging that impacts
can be evaluated at multiple scales. As a result of this strategy and its clear format, HGM WP
results can be combined with data derived through smaller-scale approaches to yield multi-scale
analyses of wetland resources. The method also seems a useful means to address additional
43
-------
facets of landscape analysis of wetlands, including wetland-based landscape classification,
threshold detection, and synoptic analyses.
44
-------
REFERENCES CITED
Abbruzzese, B. and S. Leibowitz. 1997. A synoptic approach for assessing cumulative impacts
to wetlands. Environmental Management 21:457-475.
Ainslie, W.B., R. Smith, B. Pruitt, T. Roberts, E. Sparks, L. West, L. Godshalk, and M. Miller.
1999. A Regional Guidebook for Assessing the Functions of Low Gradient, Riverine
Wetlands in Western Kentucky. Technical Report WRP-DE-17, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Allan, J., D. Erickson and J. Fay. 1997. The influence of catchment land use on stream integrity
across multiple spatial scales. Freshwater Biology 37:149-162.
Anderson, J. R., E. Ernest, J. Roach and R. Witmer. 1976. A Land Use And Land Cover
Classification System For Use With Remote Sensor Data. Geological Survey
Professional Paper 964. U.S. Government Printing Office, Washington, D.C.
Bailey, R. 1995. Ecosystem Geography. Springer-Verlag, New York.
Bedford, B. 1999. Cumulative effects on wetland landscapes: links to wetland restoration in the
United States and Southern Canada. Wetlands 19:775-788.
Bedford, B. 1996. The need to define hydrologic equivalence at the landscape scale for
freshwater wetland mitigation. Ecological Applications 6:57-68.
Bedford, B. L. and E. Preston. 1988. Developing the scientific basis for assessing cumulative
effects of wetland loss and degradation on landscape functions: status, perspectives and
prospects. Environmental Management 12:751-771.
Bolstad, P. and W. Swank. 1997. Cumulative impacts of land use on water quality in a southern
Appalachian watershed. Journal of the American Water Resources Association 33:519-
553.
45
-------
Brinson, M. M. 1993. A hydrogeomorphic classification for wetlands. Technical report WRP-
DE-4, U.S. Army engineer waterways experiment station, Vicksburg, MS .
Brinson, M. M. and R. Rheinhardt. 1996. The role of reference wetlands in functional
assessment and mitigation. Ecological Applications 6:69-76.
Brinson, M. M., F. Hauer, L. Lee, W. Nutter, R. Rheinhardt, D. Smith, and D. Whigham. 1995.
A guidebook for application of hydrogeomorphic assessments to riverine wetlands.
Technical Report WRP-DE-11, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Brooks, R., E. Bellis, C. Keener, M. Croonquist, and D. Arnold. 1989. A methodology for
biological monitoring of cumulative impacts on wetland, stream and riparian components
of watersheds. Pp. 387-398 in Proceedings of the international symposium: wetlands and
river corridor management, Charleston, SC. Association of State Wetland Managers,
Berne, NY.
Council on Environmental Quality (CEQ). 1997. Considering Cumulative Effects Under the
National Environmental Policy Act. Council on Environmental Quality, Washington
D.C.
Croonquist, M., and R. Brooks. 1991. Use of avian and mammalian guilds as indicators of
cumulative impacts in riparian wetland areas. Environmental Management 15:701-714.
ESRI. 2002. Arc View GIS version 8.2. ESRI, Inc. Redlands, CA.
Frissell, C., W. Liss, C. Warren and M. Hurley. 1986. A hierarchical framework for
stream habitat classification: viewing streams in a watershed context. Environmental
Management 10:199-214.
Geiger, R. 1965. The climate near the ground. Harvard University Press, Cambridge, MA.
46
-------
Gernes, M and J. Helgen. 2002. Indexes of Biological Integrity (IBI) for Large Depressional
Wetlands in Minnesota. Minnesota Pollution Control Agency Biological Monitoring
Program Environmental Outcomes Division. Final Report to US EPA.
Griffith, J., E. Martinko, J. Whistler and K. Price. 2002. Interrelationships among
landscapes, NDVI, and stream water quality in the U.S. central plains. Ecological
Applications 12:1702-1718.
Gwin, S., M.E. Kentula and P. Shaffer. 1999. Evaluating the effects of wetland regulation
through hydrogeomorphic classification and landscape profiles. Wetlands 19:477-489.
Halsey, L., D. Vitt and S. Zoltai. 1997. Climatic and physiographic controls on wetland type
and distribution in Manitoba, Canada. Wetlands 17:243-262.
Hauer, F., B. Cook, M. Gilbert, E. Clarain and R. Smith. 2000. A Regional Guidebook
for Assessing the Functions of Intermontane Prairie Pothole Wetlands in the Northern
Rocky Mountains. ERDC/EL TR-02-7, U.S. Army Engineer Research and Development
Center, Vicksburg, MS. .
Hauer, F., B. Cook, M. Miller, C. Noble, and T. Gonser, T. 2001. Upper Yellowstone River
Hydrogeomorphic Functional Assessment for Temporal and Synoptic Cumulative Impact
Analyses. WRAP Technical Notes Collection (ERDCTN-WRAP-01-03), U.S. Army
Engineer Research and Development Center, Vicksburg, MS.
Hauer, F., B. Cook, M. Gilbert, E. Clarain, and R. Smith. 2002. A regional guidebook
for applying the hydrogeomorphic approach to assessing wetland functions of riverine
floodplains in the Northern Rocky Mountains. ERDC/EL TR-02-21, U.S. Army
Engineer Research and Development Center, Vicksburg, MS.
Hauer, F. R., B. Cook, M. Miller, C. Noble, and T. Gonser. 2001. Upper Yellowstone
River Hydrogeomorphic Functional Assessment for Temporal and Synoptic Cumulative
Impact Analyses. ERDC TN-WRAP-01-03, U.S. Army Research and Development
Center, Vicksburg, MS.
47
-------
Hemond, H. and J. Benoit. 1988. Cumulative impacts on water quality functions of wetlands.
Environmental Management 12:639-653.
Herlihy, A., J. Stoddard, and C. Johnson. 1998. The relationship between stream chemistry and
watershed land cover. Water, Air, and Soil Pollution 105:377-386.
Hey, D.L. 1995. Flood reduction through wetland restoration - the upper Mississippi River
basin as a case history. Restoration Ecology 3:4-17.
Holling, C. S. 1992. Cross-scale morphology, geometry, and dynamics of ecosystems.
Ecological Monographs, 62:447-502 .
Hughes, R., D. Larsen, and J. Omernik. 1986. Regional reference sites: a method for assessing
stream potentials. Environmental Management 10:629 - 635.
Johnson, J. B. 2000. Documentation of reference conditions in the slope wetlands of the
southern Rocky Mountains: reference database, site descriptions, and revised functional
models. Report Submitted to U.S. EPA , Region 8 and the Colorado Department of
Natural Resources.
Johnson, L., and S. Gage. 1997. Landscape approaches to the analysis of aquatic ecosystems.
Freshwater Biology 37:113-132.
Johnson, L., C. Richards, G. Host, and J. Arthur. 1997. Landscape influences on water
chemistry in Midwestern stream ecosystems. Freshwater Biology 37:193-208.
Karr, J., K. Fausch, P. Angermeier, P. Yant, and I. Schlosser. 1986. Assessing biological
integrity in running waters: A method and its rationale. Champaign, IL, Illinois Natural
History Survey.
Karr, J. and E. Chu. 1997. Biological monitoring and assessment: Using multimetric indexes
effectively. EPA 235-R97-001. University of Washington, Seattle, WA.
48
-------
Kratz, T., K. Webster, C. Bowser, J. Magnusun and B. Benson. 1997. The influence
of landscape position of lakes in northern Wisconsin. Freshwater Biology 37:209-217.
Leibowitz, S., B. Abbruzzese, P. Adamus, L. Hughes, and J. Irish. 1992. A synoptic
approach to cumulative impact assessment: a proposed methodology. EPA/600/R-
92/167. U.S. EPA, Environmental Research Laboratory, Corvallis, OR. 129 pp. .
Lenat, D. and J. Crawford. 1994. Effects of land use on water quality and aquatic biota of three
North Carolina piedmont streams. Hydrobiologica 294:185-199.
Levin, S. 1992. The problem of pattern and scale in ecology. Ecology 73:1943-1967.
Loftis, J., L. MacDonald, S. Streett, H. Iyer, and K. Bunte. 2001. Detecting cumulative
watershed impacts: the statistical power of pairing. Journal of Hydrology 251:49-64.
Luecke, D.F. 1993. The Mississippi flood. Environment 35:2-3.
Mac Donald, L. H. 2000. Evaluating and managing cumulative effects: process and constraint.
Environmental Management 26:299-315.
Marr, J. W. 1961. Ecosystems of the Front Range in Colorado. University of Colorado's Studies
in Biology 8.
Maxwell, J., C. Edwards, M. Jensen, S. Paustain, H. Parrot, and D. Hill. 1995. A hierarchical
framework of aquatic ecological units in North America (Nearctic Zone). General
Technical Report NC-176, USD A Forest Service, North Central Forest Experiment
Station, St. Paul, MN.
McAllister, L., B. Peniston, S. Leibowitz, B. Abbruzzese, and J. Hyman. 2000. A synoptic
assessment of prioritizing wetland restoration efforts to optimize flood attenuation.
Wetlands 20:70-83.
McCulloch, J. andM. Robinson. 1993. History of forest hydrology. J. Hydrology 150:189-216.
49
-------
McCune, B. and J. Grace. 2002. Analysis of ecological communities. MJM software Design,
Gleneden Beach, OR.
McLaughlin K. 1999. Probability of wetland occurrence characterized by geology, slope, and
stream link number: Spring Creek, White Deer Creek, and Juniata watersheds,
Pennsylvania. Undergraduate Thesis in Geosciences, The Pennsylvania State University,
University Park, PA.
Mitsch, W. and J. Gosselink. 2000. Wetlands. Van Nastrand Rienhold Co., New York .
Montgomery, D. R. 1999. Process domains and the river continuum. Journal of the American
Water Resources Association 35:397-410.
National Research Council. 1995. Wetlands: characteristics and boundaries. National Academy
Press, Washington, D.C.
National Research Council. 2001. Compensating for wetland losses under the Clean Water Act.
National Academy Press, Washington D.C.
Nestler, J., and K. Long. 1997. Development of hydrologic indices to aid cumulative impact
analysis of riverine wetlands. Regulated Rivers: Research and Management 13:317-334.
O'Neill, R., D. DeAngelis, J. Waide and T. Allen. 1986. A hierarchical concept of ecosystems.
Princeton University Press, Princeton, N. J.
Omernik, J. 1991. Usefulness of natural regions for lake management: analysis of variation
among lakes in northwestern Wisconsin, USA. Environmental Management 15:281-293.
Omernik, J. 1995. Ecoregions: a spatial framework for environmental management. In Davis W.
and Simon, T. (eds). Biological assessment and criteria: tools for water resource
planning and decision making. Lewis Publishers, Boca Raton, FL.
50
-------
Omernik, J. and R. Bailey. 1997. Distinguishing between watersheds and ecoregions. J. Amer.
Water Res. Assoc. 33:935 - 949.
Peet, R. K. 1981. Forest vegetation of the Colorado Front Range. Vegetaito 45:3-75 .
Pickett, S. 1989. Space-for-time substitutions as an alternative to long-term studies. Pp. 110 -
135 in G.E. Likens (ed.) Long-term Studies in Ecology: Approaches and Alternatives.
Springer-Verlag, New York.
Preston, E. and B. Bedford. 1988. Evaluating cumulative effects on wetlands functions: a
conceptual overview and generic framework. Environmental Management 12:565-583.
Rheinhardt, R., M. Brinson, and P. Farley. 1997. Applying reference data to functional
assessment mitigation and restoration. Wetlands 17:195-215.
Richards, C., L. Johnson, and G. Host. 1996. Landscape-scale influences on stream habitats and
biota. Can. J. Fish. Aquat. Sci. 53 (Suppl. 1):295 -311.
Science Applications International Corporation (SAIC). 2000. Summit County wetland
functional assessment. Report prepared for the Summit County Community
Development Division.
Smith, R. D., A. Ammann, C. Bartoldus and M. Brinson. 1995. An approach for assessing
wetland functions using hydrogeomorphic classification, reference wetlands and
functional indices. Technical Report WRP-DE-9, U.S. Army Corps of Engineer
Waterways Experiment Station, Vicksburg, MS.
SPSS. 2003. SPSS version 12.0. SPSS, Inc. Chicago, IL.
Stein, E. D. and R. Ambrose. 2001. Landscape-scale analysis and management of cumulative
impacts to riparian ecosystems: past, present, and future. Journal of the American Water
Resources Association 37:1597-1614.
51
-------
Strahler, A. N. 1957. Quantitative analysis of watershed geomorphology. AmericanGeophysical
Union Transactions 38:1117-1142.
US Environmental Protection Agency (EPA). 1998. Evaluating performance of wetland
restoration activities - Wetland Bioassessment Fact Sheet 8. US EPA Office of
Wetlands, Oceans and Watersheds, Report No. EPA843-F-98-001h.
Warren, C. E. 1979. Toward classification and rationale for watershed management and stream
protection. EPA-600/3-79-059 .
White, P. and J. Walker. 1997. Approximating nature's variation: selecting and using reference
information in restoration ecology. Restoration Ecology 5:338-349.
Wiley, M., S. Kohler, and P. Seelbach. 1997. Reconciling landscape and local views of aquatic
communities: lessons from Michigan trout streams. Freshwater Biology 37:133-148.
Winter, T. and M. Woo. 1990. Hydrology of lakes and wetlands. In Wolman, M. and H. Riggs,
(eds). Surface water hydrology. Geological Society of North America, Boulder, CO.
Winter, T. 1992. A physiographic and climatic framework for hydrologic studies of wetlands.
In Robarts, R. and M. Bothwell (eds.). Aquatic Ecosystems in Semi-Arid Regions:
Implications for Resource Management. N.H.R.I. Symposium Series 7, Environment
Canada, Saskatoon.
Winter, T.2001. The concept of hydrologic landscapes. Journal of the American Water
Resources Association 37:335-349.
52
-------
APPENDIX 1
KEY TO SUMMIT COUNTY,
COLORADO, WETLAND TYPES
53
-------
-------
HYDROGEOMORPHIC WETLAND CLASSES AND SUBCLASSES
IN SUMMIT COUNTY, COLORADO -
DEFINITIONS, TAXONOMIC KEYS, AND USER INFORMATION
OPERATIONAL DRAFT - VERSION 2.1
Prepared by:
J. Bradley Johnson
Department of Biology
Colorado State University
Fort Collins, CO 80523
Submitted to:
U.S. Environmental Protection Agency
NHEERL / Western Ecology Division
Corvallis, Oregon
and
Colorado Geologic Survey
November 12, 2004
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
TABLE OF CONTENTS
PREFACE 2
INTRODUCTION 4
Regional Development of HGM 5
General Applicability and Use of this Key 5
SUMMIT COUNTY WETLAND TYPES 7
Slope Wetlands 8
Riverine Wetlands 10
Depressional Wetlands 12
Lacustrine Fringe 14
Application, Development and Testing of the Keys to
Summit County Wetlands Types 14
GIS Development 15
General Rules to Decision Making 16
Testing and Evaluation of the HGM Wetland Key 18
General Approach 18
Key Testing Results 19
KEY TO WETLAND SUBCLASSES PRESENT IN SUMMIT
COUNTY, COLORADO 22
Key to Lacustrine Fringe Wetland Subclasses 22
Key to Depressional Wetland Subclasses 23
Key to Riverine Wetland Subclasses 24
Key to Slope Wetland Subclasses 24
Glossary 29
REFERENCES CITED 32
Appendix 1 - EXAMPLES OF KEYING WETLANDS USING
TOPOGRAPHIC AND GIS DATA 34
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
PREFACE
The following appendix is a provisional description and classification of the
hydrogeomorphic wetland classes and subclasses found in Summit County, Colorado. It represents
a modification and refinement of the classification devised by SAIC (2000). At the class level,
wetland descriptions and diagnostic criteria are forwarded with a high degree of confidence. The
preceding document detailing the hydrogeomorphic wetland profiling (HGM WP) approach utilized
this level of HGM classification.
Subclass criteria and definitions are included for general consideration and to serve as a
guide for HGM subclass definition in other areas. Subclasses appear generally sound having been
delineated using the best available information, applied in mapping Summit County wetlands, and
reviewed by wetlands experts. This classification has not been rigorously evaluated, however, and
there may be valid reasons for revising aspects of it. Owing to its peripheral relevance to the
development of HGM WP, it is beyond the scope of the current proj ect to provide systematic testing
and revision of the subclass classification. The classification as it stands is forwarded as a
"strawman" framework awaiting future application, testing and refinement.
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
SUMMARY
Classification of wetlands based on hydrogeomorphic (HGM) characteristics is a powerful
technique that can be used for wetland inventory, cumulative impacts analysis and functional
assessment. Few existing inventories have categorized wetlands based on HGM characteristics,
however. To utilize existing wetland inventory data for studies applying HGM, previously
identified wetlands must be reclassified into regionally defined HGM categories.
This document first provides an overview of the HGM approach and the general applicability
of this key. It then details an HGM classification for wetlands in Summit County, Colorado. The
devised classification includes four wetland classes - slope, riverine, depressional, and lacustrine
fringe. Each wetland class contains between 2 and 24 subclasses.
Along with HGM category designation, a major goal of this study was to produce an
algorithm to classify previously identified wetlands into HGM categories using geographic data.
The algorithm was produced as a dichotomous key. A guide to the application of this key and
construction of an appropriate GIS model follow the HGM category descriptions. The final section
of the report contains the dichotomous key algorithm for determining wetland class and subclass
based on hydrogeomorphic characteristics. Illustrative examples of wetland classification are
included as an appendix following the key. The characteristics used in this classification are
generally obtainable from topographic maps, geographic information system (GIS) data, or aerial
photographs.
This classification and taxonomic key are for use only in Summit County, Colorado. Summit
County was chosen as the target area for this study since considerable wetland inventory and
classification work had already been accomplished, making development and testing of this system
possible. With revision this key could be expanded or adapted to other regions of the Rocky
Mountains. The intended use of this document is for the classification of identified wetland areas
into HGM categories. It may be used in conjunction with existing wetland delineation,
classification, functional assessment, and inventory tools, but the classification framework alone
cannot be used to determine the jurisdictional status or boundary of any wetland.
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
INTRODUCTION
In 1993, Brinson introduced a hydrogeomorphic (HGM) approach to wetland classification.
HGM is a hierarchical classification method, wherein wetlands are broadly categorized into seven
classes: Slope, Riverine, Depress!onal, Mineral Soil Flats, Organic Soil Flats, Lacustrine Fringe, and
Estuarine Fringe. Classes are defined by broad similarities in hydrology, geomorphology and
hydrodynamics. Each class may be further divided into a number of subclasses. Subclasses are
defined on an ecoregional level based on intra-regional differences in HGM characteristics. Unlike
many other wetland classifications, HGM does not use vegetation as a maj or classification criterion.
Instead, HGM utilizes the assumption that wetlands possessing similar hydrologic, geomorphic and
hydrodynamic settings will perform similar wetland functions, and further, that the vegetational
composition will be a reflection of these physical attributes.
HGM provides a powerful framework for wetland classification, allowing users to group
wetlands into functionally similar guilds. Because of its enormous utility and adaptability, the HGM
classification framework has been widely applied by wetland scientists, managers and federal
regulatory agencies. Initial applications of the HGM approach have primarily focused on wetland
functional assessment. To many, in fact, the HGM method is a functional assessment technique.
This is not strictly true, however, since fundamentally the HGM approach is simply a classification
scheme.
Recently, HGM classification has been used in the inventory of wetlands across watersheds
with the goal of investigating the cumulative impacts of wetland mitigation (Gwin et al. 1999). In
that study, HGM was strictly used as a classification and inventory tool to compare the
hydrogeomorphic characteristics of natural and mitigation wetlands. Such applications of HGM
have the advantage of applying the scientifically sound HGM classification in its purest form,
without the added complications inherent in the HGM wetland functional assessment method.
A maj or impediment to applying the HGM approach on a regional basis is that most wetland
mapping programs have employed classifications other than HGM; the most notable example of this
being the National Wetlands Inventory which applies the Cowardin et al. (1979) classification. In
order to apply HGM classification across wide areas, either new surveys must be performed using
HGM, or wetlands mapped using other classifications must be re-classified into HGM categories.
Due to time and resource limitations, this second approach is the one most commonly applied.
Re-classification of mapped wetlands into HGM classes or subclasses is a somewhat
qualitative and, at times, subjective process. To bolster consistency and reduce subjectivity,
algorithms, most commonly in the form of dichotomous keys, can be used to re-classify wetlands
into HGM categories (e.g. Tiner 2000, Adamus 2001) Development of re-classification algorithms
is, of course, predicated on the existence wetland class and subclass definitions within the region
of interest.
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Regional Development of HGM
At this time, HGM is truly a work in progress rather than a finished product. Wetland
classification and functional assessment using HGM is being developed throughout the country, but
this development has taken considerably longer than originally envisioned. Since wetland
classification is inherent in any application of HGM, designation and description of the HGM
classes and subclasses is a necessary first step. As described in the Federal Register (1997, Vol. 62,
Number 119), the HGM classes and subclasses present in a region are best defined by an Assessment
Team (A-Team) comprised of experts in various aspects of wetland science, and who also specialize
in the study of the wetland systems present within an ecological region. Such a classification by
committee has not generally proven tenable, however. Rather, regional subclasses are typically
defined by small team of scientists who base their classification on the input of local wetland
experts, developed reference wetland data sets, existing literature, and professional judgement.
This document provides an HGM classification of wetlands in Summit County Colorado.
It identifies two levels of wetland classification - the class and subclass levels. Although intended
for use only in Summit County, Colorado this classification could be modified to include other
reference domains within the montane, subalpine and alpine zones of the Southern Rocky
Mountains.
This document has three primary goals: (1) to designate and describe the HGM classes and
subclasses of wetlands existing in Summit County, Colorado; (2) present a dichotomous HGM
classification key to these wetland classes and subclasses; and (3) provide a detailed users guide to
the classification of wetlands using the developed keys.
General Applicability and Use of this Key
This wetland classification and taxonomic key are for use only in Summit County, Colorado.
With revision this key could be expanded or adapted to other regions of the Rocky Mountains. The
intended use of this document is for the classification of identified wetland areas into HGM
categories. It may be used in conjunction with existing wetland delineation, classification,
functional assessment, and inventory tools, but the classification framework alone cannot be used
to determine the jurisdictional status or boundary of any wetland.
This key explicitly includes habitats that may not meet the US Army Corps of Engineers (US
ACE 1987) criteria for wetlands; this is most commonly the case in riverine areas. Wetlands
supported wholly through irrigation or other artificial hydrologic inputs are also not considered
jurisdictional wetlands by the Sacramento District of the US ACE, who oversee wetland regulation
in Summit County. Other US ACE districts or regions may approach regulation of these wetland
types differently. Further, due to recent litigation and policy changes (e.g. The US Supreme Court's
SWANCC decision), many areas that meet the strict US ACE definition of wetland are no longer
under the jurisdiction of the Corps.
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Successful application of this classification key requires the use of various types of
geographic and/or remote sensing data. At a minimum, U.S. Geologic Survey 7.5 minute
quadrangles with wetland polygons drawn on the map or an affixed to an acetate overlay are needed.
Although classification can be accomplished solely using this basic topographic data and best
professional judgement, such an approach is limited and may often make proper classification
ambiguous.
Optimally, this wetland key should be used in conjunction with Geographic Information
System (GIS) software and data. Useful data types are National Wetland Inventory (NWI) maps,
Digital Elevation Models (DEMs), soil surveys (digital or hard copy), US EPA Reach File Data,
FEMA floodplain maps, geologic maps, and aerial photos. All of these data sources are available
on the web and most are free. Digital data types will be discussed in the section, Application of this
Key.
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
SUMMIT COUNTY WETLAND TYPES
Four classes of wetland exist in Summit County: Slope, Riverine, Depress!onal, and
Lacustrine Fringe. Within each of these classes, between 2 and 24 possible wetland subclasses have
been defined (Table 1). Subclasses definitions are based on existing HGM studies (Noe et al. 1998,
Johnson 2000, SAIC 2000, Adamus 2001, Keate 2001), wetland subclass keys (Tiner 2000),
consultation with local experts and best professional judgement. Subclass categories may need
future revision as additional studies are completed. Justification and definition of HGM subclasses
are provided in the following section. Subsequent sections include dichotomous keys for identifying
wetland classes and subclasses based on hydrogeomorphic criteria.
Table 1. List of potential HGM wetland classes and subclasses found in Summit County, Co.
Wetland types are "potential" since several of the slope subclasses have not yet been identified in
the county and may not exist there.
Wetland
Class
Slope
Riverine
Wetland Subclass
Isolated, shallow gradient, mineral/organic
soil
Isolated, moderate gradient, mineral/organic
soil
Isolated, steep gradient mineral/organic soil
Through-flow, shallow gradient,
mineral/organic soil
Through-flow, moderate gradient,
mineral/organic soil
Through-flow, steep gradient
mineral/organic soil
Inflow, shallow gradient mineral/organic soil
Inflow, moderate gradient mineral/organic
soil
Inflow, steep gradient mineral/organic soil
Outflow, shallow gradient mineral/organic
soil
Outflow, moderate gradient mineral/organic
soil
Outflow, steep gradient mineral/organic soil
Low-order, steep gradient and/or confined
floodplain
Moderate-order, steep gradient and/or
confined floodplain
High-order, steep gradient and/or confined
floodplain
Wetland
Class
Riverine
Depression
Lacustrine
Fringe
Wetland Subclass
Low-order, unconfmed floodplain
Moderate-order, unconfined floodplain
High-order, unconfined floodplain
Kettle Pond Wetlands
Oxbow and Meander Scar Wetlands
Other Natural Depressional Wetlands
Other Man-made Depressional Wetlands
Natural Lacustrine Fringe
Man-made Lacustrine Fringe
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Slope Wetlands
Slope wetland subclasses are categorized based on their hydrologic connection to the
surrounding landscape, topographical gradient, and soil composition (organic vs. mineral). Slope
wetland subclass characteristics are presented in Table 2. The slope wetland subclass categories are
taken directly from Johnson (2000). Several of the designated slope wetland subclasses were not
identified during Summit County surveys. They are retained in this classification, however, since
representatives of these subclasses may be identified with future study. Also, subclasses not
identified in the current Summit County surveys may be found in other parts of the southern Rocky
Mountains. It is hoped that the retention of these categories in this key increases the adaptability
of this classification framework to areas outside of Summit County.
Subclass criteria were chosen based on the results of Johnson's (2000) study, which
considered original data, as well as the findings of previous investigations. The chosen criteria are
thought to most strongly control the hydrogeomorphic functions performed by Rocky Mountain
slope wetlands. Other related sub-classifications of slope wetlands have utilized a variety of
hydrogeomorphic, chemical or vegetational characteristics including those used in this classification.
For instance, Tiner (2000) subdivided slope wetlands solely based on the hydrologic connection to
the surrounding landscape. Keate (2001) divided the slope wetland class on the basis of soil
composition, water table depth, and soil pH and salinity. Noe et al. (1998) indirectly categorized
slope wetlands based on their hydrology, soil composition and water chemistry. Finally, the Summit
County Special Area Management Plan development team developed a slope wetland classification
based on soil composition, and topographical gradient (SAIC 2000).
Although each of the above classifications were developed in response to specific project
and management goals, a consistent philosophy exists between the various approaches. The
classification criteria chosen for use here are intended to maintain continuity with and refine existing
regional classifications for use specifically within the Summit County study area.
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Table 2. Table of Summit County slope wetland subclass characteristics, with the name (italics),
followed common examples of each (plain text). Shaded cells indicate potential subclasses which
)robably do not exist in the study area.
Hydrologic
Connection
/ Slope
Inlet Present
(no surface
outlet)
Outlet Presence
(no surface inlet)
Surface Inlet and
Outlet Present
No surface Inlet
or Outlet
'5
V)
o
'E
re
O)
a>
0 - 4 %
Inflow, shallow-
gradient, organic
soil
Fen or Carr
Outflow, shallow-
gradient, organic
soil
Fen or Carr
Through-flow,
shallow-gradient,
organic soil
Fen or Carr
Isolated, shallow-
gradient, organic
soil
Fen or Carr
>4 - -10 %
Inflow, moderate-
gradient, organic
soil
Fen or Carr
Outflow,
moderate-
Through-flow,
moderate-
organic
soil
Fen or Carr
organic
soil
Fen or Carr
Isolated,
moderate-
gradient, organic
soil
Fen or Carr
Outflow, high-
gradient, organic
soil
Forested Fen
(Swamp)
Through-flow,
high-gradient,
organic soil
Forested Fen
(Swamp)
4 - -10 %
Inflow, moderate-
gradient, mineral
soil
Wet Meadow or
Scrub-Shrub
Outflow,
moderate-
gradient, mineral
soil
Wet Meadow or
Scrub-Shrub
Through-flow,
moderate-
gradient, mineral
soil
Wet Meadow or
Scrub-Shrub
Isolated,
moderate-
gradient, mineral
soil
Wet Meadow or
Scrub-Shrub
Outflow, high-
gradient, mineral
soil
Avalanche chute
or Open-canopy
Swamp Forest
Through-flow,
high-gradient,
mineral soil
Avalanche chute
or Open-canopy
Swamp Forest
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Riverine Wetlands
The criteria used to define riverine wetland subclasses are quite consistent between various
existing HGM classifications. The most commonly used definitional criteria are: stream order,
topographical gradient, and character and extent of the floodplain (Hauer and Smith 1998, Noe et
al. 1998,Keate 1998, SAIC 2000, Tiner 2000). The classification included herein uses each of these
characteristics to differentiate riverine wetland subclasses (Table 3). Here, stream order is used to
provide an index of channel and wetland size. Although, at times potentially misleading, such as
in the case of trellis networks in which stream order does not increase despite an increase in drainage
basin area, in this region stream order seems to generally provide an accurate and practically
attainable picture of wetland character.
Definition of additional subclasses could improve the classification in general or for a
specific application. To potentially valuable revisions are provided here. First, it could be beneficial
to consider flow duration in delineation of low stream order wetland subclasses. Stream duration
classes designated by the U.S. Geological Survey, could prove useful in this regard.
Second, subclasses are currently designated solely on criteria pertaining to the physical
wetland setting. The influence of biota on wetland functioning is only implied in this classification.
Most importantly, in these mountains, beaver can have a profound effects on wetland characteristics
and functioning. The structure of this classification assumes that virtually all potential riparian
habitat has been historically or is currently utilized by beaver. It could certainly be beneficial to
divide subclasses based on the nature or extent of beaver-induced ecosystem changes. Such a goal
requires detailed attention, however, since the classification must take into account a complex
temporal and spatial mosaic of functional states. The classification presented here acknowledges
these complexities, but takes advantage of the sheer ubiquity beaver activity by implying its general
existence, and by considering the mosaic of conditions as an intrinsic constituent of mountain
riparian systems.
10
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Table 3. Tabular representation of riverine wetland subclasses.
Stream
Order
Steep Gradient and/or Confined Active
Floodplain
Moderate to Low Gradient with a
Unconfined Floodplain
Low (1st-
3rd)
Riverine system bounded by the seasonal
high water mark with little or no
floodplain. Channels may be artificially
entrenched, thereby constraining the active
channel within a broader valley.
Channel movement is generally
unconstrained but the riparian zone is
poorly developed or narrow due to low
flow volume. These wetlands often found
in association with headwater slope
wetland complexes and commonly include
beaver-pond complexes.
Moderate
(4th - 5th)
Riverine system bounded by the seasonal
high water mark with little or no
floodplain. Narrow, temporary bars and
surfaces are common. Channels may be
artificially entrenched, thereby
constraining the active channel within a
broader valley.
Well developed, active floodplains usually
present. Floodplains may have a variety of
fluvial features such as oxbow ponds and
meander scrolls. These riverine wetlands
may be created or expanded by beaver
activity and are commonly associated with
terrace-base slope wetlands. Associated
terrace-base wetlands are dominated by
groundwater discharge which drains into
the flood plain. Such wetlands should be
evaluated as slope wetlands.
High (6th -
7th)
Riverine system bounded by the seasonal
high water mark with little or no
floodplain. Temporary bars and surfaces
are common. Channels may be artificially
entrenched, thereby constraining the active
channel within a broader valley.
Well developed, active floodplains usually
present. Floodplains may have a variety of
fluvial features such as oxbow ponds and
meander scrolls. These riverine wetlands
are commonly associated with terrace-base
slope wetlands. Associated terrace-base
wetlands are dominated by groundwater
discharge which drains into the flood plain.
Such wetlands should be evaluated as slope
wetlands.
11
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Depressional Wetlands
Few types of depressional wetlands exist in Summit County. The Summit County Special
Area Management Plan (SAMP) technical development group identified three depressional wetland
types: kettle ponds including their associated wetlands; oxbow/meander scar ponds including
associated wetlands; and other types of natural and man-made wetlands (SAIC 2000). The SAMP
subclassification has been generally maintained in this framework, except that the subclass "other
types of natural and man-made wetlands" has been divided into two subclasses based on whether
the wetland was natural or man-made (Table 4).
The first two depressional wetland subclasses share similar hydrology (groundwater
discharge, surface inlet channels and/or surface runoff) and hydrodynamics (vertical water
fluctuations), differing primarily in geomorphic position. The third and forth subclasses are
artificial, including a variety of either natural or man-made wetlands. Examples of natural
depressional wetlands types included within the third subclass are snowmelt-fed depressions above
tree line or naturally dammed, landslide-scar wetlands. Man-made depressional wetlands may take
many forms and may be wholly dependent on artificial hydrologic inputs for wetland maintenance.
The two artificial subclasses were maintained in this classification since: 1) there is relatively
little information available describing these wetlands, 2) these depressional wetlands comprise only
a small percentage of the total wetland area in the region, 3) they are often removed from
developmental risk, and 4) the types of wetlands included within either of these subclasses probably
function similarly on the landscape, with many of the man-made types providing little ecological
function.
12
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Table 4. Hydrogeomorphic characteristics of depress!onal wetland subclasses.
Wetland Subclass
Hydrologic Characteristics
Hydrodynamics
Geomorphic Position
Kettle Pond Wetlands
May be isolated, or may
have inflow, outflow, or
through-flow channels.
Hydrologic inputs may
additionally come through
groundwater discharge and
surface runoff. Hydrology
in these wetlands frequently
varies seasonally.
Vertical
Found on level or
hummocky surfaces of
glacial deposits -
typically moraines or
outwash plain.
Oxbow and Meander
Scar Wetlands
Usually lacking perennial,
natural surface water inlets
and outlets, although high-
water inlet or outlet
channels may be present.
Additional hydrologic inputs
may come through
groundwater discharge and
surface runoff. Hydrology
in these wetlands may vary
seasonally.
Vertical
Located on a fluvial
surface above, or
physically isolated
from, the active
floodplain.
Other Natural
Depressional Wetlands
May be isolated, or may
have in-flow, out-flow, or
through-flow channels.
Hydrologic inputs may
additionally come through
groundwater discharge and
surface runoff. Hydrology
in these wetlands frequently
varies seasonally.
Vertical
Various
Other Man-made
Depressional Wetlands
May be isolated, or may
have inflow, outflow, or
through-flow channels.
Additional hydrologic inputs
may additionally come
through surface runoff or
groundwater discharge.
Hydrology in these wetlands
frequently varies seasonally
and may entirely artificially
sustained.
Vertical
Various
13
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Lacustrine Fringe
Lacustrine fringe wetlands are broken into two subclasses based on genesis and
hydrodynamics: natural lacustrine fringe and man-made lacustrine fringe. These subclasses are
consistent with Noe et al. (1998) and SAIC (2000), although Noe et al. (1998) included these
wetland types in the Depressional class, and in SAIC (2000) the above subclasses were divided
based on the magnitude of water table fluctuations. Table 5 describes the HGM characteristics of
the two lacustrine fringe wetland subclasses.
Table 5. Hydrogeomorphic characteristics of Summit County lacustrine fringe wetlands.
Wetland Subclass
Natural Lacustrine
Fringe
Man-made Lacustrine
Fringe
Hydrologic
Characteristics
Surface inlet and
outlet.
Surface inlet generally
with a regulated outlet.
Hydrodynamics
Vertical. Seasonal
water level fluctuations
are relatively low.
Vertical. Water level
fluctuations large and
water level may vary
sporadically due to
water management.
Geomorphic Position
Valley bottoms and
plateaus.
Various
Application, Development and Testing of the Keys to
Summit County Wetlands Types
The reader may wish to briefly scan the keys in the following two sections to become
generally familiar with the form and approach of the key before reading this section on key usage
and application. This section describes the way in which this key was actually implemented during
development and testing. This discussion should serve as a user manual for the key during future
applications. Much of the discussion below is of a technical nature. To utilize the most current
technological approaches, it is assumed that the user already possess experience in geographic
information systems (GIS). Users without GIS experience or capacity may skip technical sections
and use manual methods as described in the General Rules to Decision Making section.
14
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
GIS Development
A GIS was developed using Arc View version 8.1 software (ESRI2001). Data incorporated
into the GIS came from a variety of sources. All GIS data was projected into the Universal
Transverse Mercator (UTM) system using the North American Datum of 1927 (NAD 27). Digital
USGS 7.5 minute quadrangles in GeoTIFF format comprised the base layer. GeoTIFFs are available
from the Natural Resources Conservation Service's (NRCS) Geospatial Data Gateway1
(http://www.lighthouse.nrcs.usda.gov/gateway/gatewayhome.html): although the GeoTIFFs used
in this project were obtained from a commercial source.
Wetland polygons were obtained from three aerial photograph surveys of Summit County.
Data were provided by Summit County. Each aerial photograph data source is described in SAIC
(2002). The original wetland classifications were converted by SAIC (2000 and 2002) to
hydrogeom orphic categories based on reinterpretation of aerial photographs, topographical data, and
field surveys.
A 10 meter DEM was used to derive 6 m (20 ft.) topographical contours. Wetland
topographic gradient was determined by defining an axis within the widest or longest portion of the
wetland perpendicular to its topographical gradient. The length or width of the wetland along this
axis was then measured using the Arc View "Measure" tool. Finally, the change in elevation across
the wetland was determined by counting contour lines and multiplying by the contour interval.
When a wetland began or ended between contours, elevation was estimated based on the relative
proportion of the contour interval covered by the wetland.
Stream ordering was carried out by hand using Strahler's (1952) method. Ordering was
applied to 1:100,000 digital line graphics (DLGs) and was completed by Summit County
Government. A field containing stream order data was added to the stream digital line graphic
(DLG) attribute table.
Geologic data for the county were obtained from the Denver and Leadville 1X2ฐ
(1:250,000) Geologic Quadrangles. These data files are available through the USGS web page at
http://greenwood.cr.usgs.gov/pub/mf-maps/mf-2347 and
http://greenwood.cr.usgs.gov/pub/open-file-reports/ofr-99-0427. respectively.
Soil data were derived from two sources. Soil Survey Geographic Database (SSURGO) data
were used were available. These data are the most detailed digital soil data available in the state,
but coverage of Summit County is incomplete. SSURGO data are available at:
http ://www.ftw.nrcs.usda. gov/ssurgo/metadata/co690.html. Where SSURGO data were unavailable,
coarser State Soil Geographic Database (STATSGO) data were used. These data, compiled at a
scale of 1:250,000, are available at: http://www.ftw.nrcs.usda.gov/stat_data.html.
1 Web site addresses (URLs) are current as of time of publishing, but are subject to
change.
15
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
General Rules to Decision Making
Determination of many hydrogeomorphic traits, such as gradient or stream order, is straight
forward using GIS or even simple paper 7.5' topographical quadrangles. Other characteristics such
as local soil composition or dominant hydrologic inputs may be more difficult to determine due to
lack of site specific data or resolution limitations. What follows are some general guidelines to
wetland characterization in cases where description may not be clear cut or where data limitations
exist.
Most wetlands have somewhat mixed hydrologies. Of course all wetlands receive meteoric
inputs, but most also receive varying degrees of surface and/or groundwater inputs. Brinson (1993),
for example, provided a triangular ordination of wetland type based on the relative contribution of
precipitation, groundwater and surface water inputs to the total wetland water budget (Fig. 1).
100%
100%
H*. 33% ง7%
SURFACE FLOW
10G5E
Figure 1. Separation of wetland types based on the relative
contribution of hydrologic inputs. From Brinson (1993)
For this HGM classification, in all cases, wetlands should be grouped into classes based on
the dominant hydrologic source or process. Considering only ground and surface water inputs,
riverine wetlands should receive more than 50 % of their hydrologic input from surface water,
whereas slope wetlands should derive more than 50% of their hydrologic input from groundwater.
As detailed in the description of Summit County wetland subclasses above, water sources of
depressional and lacustrine fringe wetlands can vary.
16
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
In practical terms, separating riverine from slope wetlands is commonly quite difficult and,
often arbitrary since these two types of systems may be interdependent and spatially intermixed.
If fine scale detail is required for a particular application, large contiguous wetlands containing a
mix of wetland subclasses can be broken in multiple polygons (Partial Wetland Assessment Areas
sensu Smith et al. 1996). For larger-scale studies this level of subdivision may not be valuable.
If a data source is lacking or its resolution is insufficient, secondary indicators may provide
adequate evidence for decision making and classification. If some wetland characteristics can not
be determined, the wetland can be left partially classified. For example a slope wetland could be
hydrogeomorphically classified as "isolated, low gradient" even if soil composition is unknown.
Examples of common hydrologic, geologic and edaphic indicators are provided below to help
resolve common ambiguous situations that could arise while using these keys.
Indicators of Hydrology in Riverine, Slope and Lacustrine Fringe Wetlands
Separating riverine from slope wetlands is the greatest challenge in this method, since the
two types are often adjacent to one another, or form portions of a greater wetland complex. Slope
wetlands at the base of stream terraces - both large and small - and flow-through slope wetlands are
the hardest to distinguish and separate from riverine wetlands. In fact, these wetlands may be so
closely associated that their separation is somewhat artificial. Therefore, while a greater number or
errors may in classifying these wetlands the consequences of doing so should generally be low. As
written before, in classifying a wetland based on hydrology, the dominant hydrologic input should
be used to determine wetland class. If different water sources dominate distinctly different portions
of a wetland, that wetland should be subdivided to reflect the situation. The following
characteristics help to distinguish hydrologic types when specific hydrologic data is not available.
Flow-through slope wetlands can often be identified based on the following characteristics:
The stream channel is low order (1st to 3rd) and wetland areas extend up adjacent valley
slopes, well above the channel elevation. This characteristic shows that wetlands receive
water from sources other than the alluvial system. Presence of a flow-through slope wetland
may also be indicated by asymmetry of the wetland with respect to the channel. That is,
within a symmetric channel valley, the wetland will be relatively broad and/or wide on the
side of the channel possessing the slope wetland. The wetland on the opposite channel, will
be more narrow due to the absence of non-alluvial, groundwater input (See Appendix 1).
Presence of a slope wetland is often indicated by the existence of a wetland on a slope or at
a slope break with no other apparent source of water. Ponds may or may not be present. If
a pond is present, the wetland may require division into slope and a depressional portions.
Slope wetlands often occur at the base of riparian terraces, where an abrupt change in slope
causes groundwater to emerge. Slope wetlands that occur under these circumstances may
often be distinguished from adjacent riverine wetlands in their spatial separation from the
channel and origination at the base of terraces. They may also extend up the terrace slope
17
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
well above the alluvial aquifer, they may be physically separated from other riverine
wetlands in the floodplain, and they may develop outlet channels which flow into the main
river or stream channel.
Slope wetlands are often found on or near the shores of reservoirs or natural lakes. In the
case of reservoirs, historically expansive slope wetland complexes have frequently been
artificially flooded, and remnants of the wetlands exist on the lake fringe. If groundwater
inputs, rather than lake water levels, still dominate a wetland's hydrology, the wetland
should be classified as slope. An indication of this condition is when the wetland extends
well above the elevation of the lake's surface.
Indicators of Soil Composition in Slope Wetlands
Below 8500' a.m.s.l. it is highly unlikely that a slope wetland would possess organic soils.
It is uncommon for very steep slope wetlands to possess organic soils.
Indicators of Geologic Composition in Depressional Wetlands
Large-scale hummocky terrain and/or clustering of small, somewhat-circular depressional
wetlands is usually indicative of glacial moraines and outwash. Therefore, depressional
wetlands located in such areas should be classified as kettle ponds unless additional
contradictory evidence exists.
Testing and Evaluation of the HGM Wetland Key
General Approach
The HGM categories devised in this classification were based on pre-existing literature,
consultation with local experts and professional judgement, as explained in the Introduction section
of this document. The hierarchical dichotomous key algorithm was produced based distinguishing
between the primary attributes defining a wetland category (either class or subclass). Subsequent
to testing, this algorithm was revised to increase clarity and usability.
Quality assurance/control (QA/QC) documentation generally considers several procedural
attributes including: 1) accuracy; 2) precision; 3) completeness; and 4) representativeness (US EPA,
1995). Other common QA/QC parameters were not applicable to this proj ect since it did not contain
laboratory analyses and no comparisons to other data sets were made. Determination of
methodological accuracy, completeness and representativeness are described below. A measure of
method precision such as relative standard deviation or percent difference could not be used here
18
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
since all analyses were completed by a single individual and repeated classification of polygons
("repeated measure") would be strongly biased by the classifier's experience.
The wetland data set, assembled by SAIC for Summit County as described in the
Introduction section, was used to test the accuracy of the classification key. Wetland classes and
subclasses assigned to each wetland polygon in this data set were essentially taken as "true" values.
The exception to this was when errors in stream ordering were encountered in the Summit County
data set. Stream order calculation, which defines riverine wetland subclass, was performed "on-the-
fly" during the Summit County wetland data set construction. For keying, an updated GIS layer
containing a Digital Line Graphic showing stream order was used. When inconsistencies arose
between the data sets, stream order would be manually re-evaluated. In every case, the GIS DLG
of stream order was correct.
One hundred randomly chosen wetland polygons were selected from the Summit County data
set. Each randomly selected wetland polygon was first keyed to class and then subclass using this
classification algorithm and the GIS resources described in the Introduction section. Upon
completion of keying, the wetland category obtained using this algorithm was compared to that
contained in the SAIC data set. To assess the accuracy of this key three conditions were defined:
1) correct class and subclass designation; 2) correct class and incorrect subclass designation; 3)
incorrect class and subclass designation. The distribution of wetlands in each of these three
categories was used to determine the accuracy of the key.
Percent completeness was calculated as: %C = (v/T)*100, where v = the total number of
valid observations and T = the total number of observations (US EPA 1995). Representativeness
was insured by using a random sampling approach and using a sample size large enough such that
all wetland classes were sampled. Due to the strong disparities in the commonness of the various
wetland types, coupled with the random sampling regime, the sample size of each wetland class
differed.
Key Testing Results
The wetland class distribution of the 100 randomly selected wetland polygons is provided
in the table below. Of the 100 polygons, 13 could not be used for comparative testing since they
were left unclassified in the Summit County data set. Thus sampling completeness was 87%.
HGM Class
Depression
Lacustrine Fringe
Riverine
Slope
Unclassified
Total
Number represented in
sample
7
2
52
26
13
100
19
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
There was a remarkable correspondence between the wetland types as designated in the
Summit County data set and the wetlands classified using this keying algorithm; that is, the key was
shown to be highly accurate. Eighty-three of 87 wetlands (95%) were classified into the correct
wetland class, while 79 of 87 wetlands (91%) were correctly classified to subclass. Only about 5%
of wetlands were miss-classified at the class level. Classification errors were of three types: 1)
Slope classified as riverine; 2) Slope classified as lacustrine fringe; and 3) Riverine classified as
slope.
In the first case, flow-through, slope wetlands were classified as riverine due to their similar
appearances and landscape settings. In the second case, slope wetlands were classified as lacustrine
fringe when former slope wetland complexes were partially flooded by reservoir waters, and
marginal portions of the wetland remained beyond the reservoir shores. Lastly, one riverine wetland
was classified as slope. In this case the wetland occurred in the floodplain near a terrace base, which
is a common landscape position for slope wetlands. "Decision rules" to help determine wetland type
were generated to mitigate errors in cases such as these. These rules are provided above in the
General Rules to Decision Making section.
20
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
KEYS TO SUMMIT COUNTY WETLAND
CLASSES AND SUBCLASSES
Following the keys are illustrative examples of wetland types depicted in a GIS
21
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
KEY TO WETLAND CLASSES PRESENT IN SUMMIT
COUNTY, COLORADO
la. Wetland is found on the margin of a natural lake or reservoir larger than 0.5 ha with water
depth exceeding 2 m, or wetland is located on the margin of an island
LACUSTRINE FRINGE WETLAND
Ib. Wetland is not associated with a natural lake or reservoir 2
2a. Wetland surrounds and includes a shallow, open water area. Wetland is not located in an
active alluvial floodplain, nor is it a beaver pond (these wetlands are classified as Riverine).
Wetland is located in an area of closed contour topography and may be hydrologically
isolated, have a surface inlet, have a surface outlet, or be a through-flow system (inlet and
outlet present). Surface water inflow and outflow may be strongly seasonal
DEPRESSIONAL WETLAND
2b. Wetland possesses open-contour topography, with or without surface water inlets or outlets
3
3a. Wetland is within the 100-year floodplain of a perennial stream or river and not located at
the base of a fluvial terrace RIVERINE WETLAND
3b. Wetland is not located within the 100-year floodplain of a perennial stream, or if it is within
the 100-year floodplain, wetland is located at the base of a fluvial terrace. Groundwater
discharge dominates hydrologic inputs. Wetland may be on sloping or relatively flat terrain
(1 % gradient). Springs or seeps are usually present SLOPE WETLAND
22
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
KEY TO WETLAND SUBCLASSES PRESENT IN SUMMIT
COUNTY, COLORADO
KEY TO LACUSTRINE FRINGE WETLAND SUBCLASSES
la. Wetland is located on the shore of a natural lake. Wetland hydrology is directly controlled
by lake water levels. Water level is relatively stable with small, seasonal variations ....
Natural lacustrine fringe wetlands
Ib. Wetland is located on the shore of a man-made lake or reservoir. Wetland hydrology is
directly controlled by lake water levels. Lake water level commonly experiences large
fluctuations in water level seasonally, or sporadically through active water management .
Man-made lacustrine fringe wetlands
KEY TO DEPRESSIONAL WETLAND SUBCLASSES
1 a. Wetland includes an open-water kettle pond formed during Quaternary glacial retreat. Kettle
ponds and associated wetlands are often roughly circular in shape and are associated with
glacial till. The kettle pond wetland may include areas of open-water, floating vegetation
mats, or solid organic or mineral soils Kettle pond wetland
Ib. Wetland not as above 2
2a. Wetland located in a relic river oxbow or meander scar and is located on a geomorphic
surface above the active floodplain, or wetland has been isolated from the active floodplain
through the effects of anthropogenic structures such as dams or dikes. Isolated riverine
depressional wetlands may receive hydrologic inputs through precipitation, groundwater
inflow and surface runoff Isolated riverine depressional wetland
2b. Wetland is in a geomorphic position not described above. Wetland may be man-made such
as a cattle pond, or natural such as a snowmelt-fed wetland above tree line .... 3
3a. Wetland was formed by natural processes. These wetlands may be located in a variety of
hydrogeomorphic situations, such as in snowmelt-fed depressions above tree line or behind
dammed landslide scars Other natural depressional wetlands
3b. Wetland is man-made. These wetlands can be found in virtually any hydrogeomorphic
position and may be hydrologically maintained solely through artificial means
Man-made depressional wetlands
23
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
KEY TO RIVERINE WETLAND SUBCLASSES
1 a. Wetland and associated stream possess a high gradient (greater than approximately 8 %), or
wetland and floodplain tightly bounded by geomorphological features such as narrow valley
walls or cliff scarps. Due to the steep gradient and/or floodplain confinement, these
wetlands will typically be narrow and closely associated with the active channel 2
Ib. Wetland and associated stream possess a moderate to low gradient (between 0 and ~ 8%),
and wetland and floodplain broadly bounded by geomorphological features such as terraces
or cliff scarps. Relatively low gradients and broad floodplains allow the development of
more expansive riverine wetlands 4
2a. Wetland is located on a stream with an order between 1 and 3
Low-order, high-gradient or bounded, riverine wetland
2b. Wetland is located on a higher order stream 3
3a. Wetland is located on a stream with an order between 4 and 5
Mid-order, high-gradient or bounded, riverine wetland
3b. Wetland is located on a stream with an order between 6 and 7
High-order, high-gradient or bounded riverine wetland
4a. Wetland is located on a stream with an order between 1 and 3
Low-order, low-gradient, unbounded riverine wetland
4b. Wetland is located on a higher order stream 3
5a. Wetland is located on a stream with an order between 4 and 5
Mid-order, low-gradient, unbounded, riverine wetland
5b. Wetland is located on a stream with an order between 6 and 7
High-order, low-gradient, unbounded riverine wetland
24
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
KEY TO SLOPE WETLAND SUBCLASSES
la. No apparent surface inflow from a stream or other water body, and no apparent outflow to
a stream or water body Isolated slope wetlands sub-key
Ib. Wetland possess an inlet channel, an outlet channel, or both 2
2a. Wetland possess both and inlet and an outlet channel
Through-flow slope wetlands sub-key
2b. Wetland posses only an inlet or and outlet channel 3
3a. Wetland only has an inlet channel. Channelized flow become disperse and eventually ceases
within the wetland Inflow slope wetlands sub-key
3b. Wetland only has an outlet channel. Wetland forms a stream headwaters
Outflow slope wetlands sub-key
Slope Wetland Sub-keys
Isolated Slope Wetland Sub-key
la. Wetland topographical gradient between 0 - 4 %
Isolated, shallow gradient, slope wetland
i. Wetland only having mineral soils
Isolated, shallow-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Isolated, shallow-gradient, organic soil slope wetland
Ib. Wetland topographical gradient > 4% 2
2a. Wetland topographical gradient > 4 % to 10 %
Isolated, moderate-gradient slope wetlands
i. Wetland only having mineral soils
Isolated, moderate-gradient, mineral soil slope wetland
25
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
ii. Wetland possesses organic soils
Isolated, moderate-gradient, organic soil slope wetland
2b. Wetland topographical gradient > 10%
Isolated, high gradient slope wetlands
iii. Wetland only having mineral soils
Isolated, high-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Isolated, high-gradient, organic soil slope wetland
Through-flow Slope Wetland Sub-key
la. Wetland topographical gradient between 0-4%
Through-flow, shallow gradient, slope wetland
i. Wetland only having mineral soils
. . Through-flow, shallow-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
. . Through-flow, shallow-gradient, organic soil slope wetland
Ib. Wetland topographical gradient > 4% 2
2a. Wetland topographical gradient > 4 % to 10 %
Through-flow, moderate-gradient slope wetlands
i. Wetland only having mineral soils
Through-flow, moderate-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Through-flow, moderate-gradient, organic soil slope wetland
2b. Wetland topographical gradient > 10%
Through-flow, high-gradient slope wetlands
iii. Wetland only having mineral soils
Through-flow, high-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Through-flow, high-gradient, organic soil slope wetland
Inflow Slope Wetland Sub-key
la. Wetland topographical gradient between 0 - 4 %
Inflow, shallow gradient, slope wetland
i. Wetland only having mineral soils
Inflow, shallow-gradient, mineral soil slope wetland
26
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
ii. Wetland possesses organic soils
Inflow, shallow-gradient, organic soil slope wetland
Ib. Wetland topographical gradient > 4% 2
2a. Wetland topographical gradient > 4 % to 10 %
Inflow, moderate-gradient slope wetlands
i. Wetland only having mineral soils
Inflow, moderate-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Inflow, moderate-gradient, organic soil slope wetland
2b. Wetland topographical gradient > 10%
Inflow, high gradient slope wetlands
iii. Wetland only having mineral soils
Inflow, high-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Inflow, high-gradient, organic soil slope wetland
Outflow Slope Wetland Sub-key
la. Wetland topographical gradient between 0 - 4 %
Outflow, shallow gradient, slope wetland
i. Wetland only having mineral soils
Outflow, shallow-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Outflow, shallow-gradient, organic soil slope wetland
Ib. Wetland topographical gradient > 4% 2
2a. Wetland topographical gradient > 4 % to 10 %
Outflow, moderate-gradient slope wetlands
i. Wetland only having mineral soils
Outflow, moderate-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Outflow, moderate-gradient, organic soil slope wetland
2b. Wetland topographical gradient > 10%
Outflow, high-gradient slope wetlands
27
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
iii. Wetland only having mineral soils
Outflow, high-gradient, mineral soil slope wetland
ii. Wetland possesses organic soils
Outflow, high-gradient, organic soil slope wetland
28
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Glossary
Confined (tightly-bounded) or Unconfined (loosely-bounded) Floodplain3: The confined
floodplain is one in which the hillslopes narrowly constrain the valley floor with little or no
floodplain development. An unconfmed floodplain is one in which valley hillslopes only broadly
constrain channel movement and floodplain development. Unconfined floodplains are often, but
not always associated with broad riparian zones. A gauge of confinement is as follows:
Valley floor width < 3 channel widths wide
Valley floor width > 3 channel widths wide
Strongly to somewhat
confined
Somewhat confined to
unconfmed
Confined Reaches
Unconfined Reaches
Depressional Wetland:4 Depressional wetlands are found within topographic depressions
possessing closed elevational contours that allow the accumulation of surface water.
Hydrogeomorphic Approach: An approach to wetland classification which categorizes wetlands
based on commonalities in their geomorphology, hydrology and hydrodynamics.
Lacustrine Fringe Wetland:4 A wetland type found adjacent to lakes where the water elevation of
the lake maintains the water table in the wetland. Lacustrine fringe wetlands may also be found on
islands within lakes.
Modified from Bisson and Montgomery (1996)
4 Modified from Smith et al. (1995).
29
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Organic soils: Organic soils are those soils which meet the NRCS (1998) definition of a Histosol
which is provided below.
Histosols:
Do not have andic soil properties in 60 percent or more of the thickness between the soil surface and either a depth of
60 cm or a densic, lithic, or paralithic contact or duripan
if shallower; and
2. Have organic soil materials that meet one or more of the following:
a. Overlie cindery, fragmental, or pumiceous materials and/or fill their intersticesi and directly below these materials,
have a densic, lithic, or paralithic contact;
or
b. When added with the underlying cindery, fragmental, or pumiceous materials, total 40 cm or more between the soil
surface and a depth of 50 cm; or
c. Constitute two-thirds or more of the total thickness of the soil to a densic, lithic, or paralithic contact and have no
mineral horizons or have mineral horizons with a total thickness of 10 cm or less; or
d. Are saturated with water for 30 days or more per year in normal years (or are artificially drained), have an upper
boundary within 40 cm of the soil surface, and have a total thickness of either.
(1) 60 cm or more if three-fourths or more of their
volume consists of moss fibers or if their bulk density,
moist, is less than 0.1 g/cms; or
(2) 40 cm or more if they consist either of sapric or
hemic materials, or of fibric materials with less than
three-fourths (by volume) moss fibers and a bulk
density, moist, of 0.1 g/cmsor more.
Riverine Wetland:4 A wetland that occurs within a floodplain or riparian corridor. The dominant
water source for this type of wetland is overbank flow from the channel or subsurface hydrologic
connection to the alluvial aquifer.
Slope Wetland:4 A wetland whose hydrology is dominated by groundwater discharge. Slope
wetlands lack significant depressional water storage due to their lack of closed contours. Slope
wetlands may occur on nearly flat landscapes.
Stream order: A means of classifying streams according to their importance within the drainage
basin. The lowest order streams are the most minor tributaries. This study used the Strahler (1952)
stream ordering method.
Hydrogeomorphic Wetland Class:4 The highest level in the hydrogeomorphic (HGM) wetland
classification. This document includes four HGM classes: Slope, Riverine, Depression, and
Lacustrine Fringe. Class are based on gross difference site hydrology, geomorphology and
hydrodynamics.
Hydrogeomorphic Wetland Subclass:4 Wetlands within a region that are similar based on
hydrogeomorphic classification factors. There are multiple subclasses within each wetland class.
The actual number of wetland subclasses defined is based on the known diversity of wetlands within
a class and particular assessment objectives.
Alluvial: Structures and features produced through river-induced deposition.
30
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Fluvial: Structures, features and characteristics produced through stream action.
Scarp: A steep slope or cliff connecting the surface of a former floodplain to a lower surface.
31
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
REFERENCES CITED
Adamus, P. R. 2001. Guidebook for hydrogeomorphic (HGM)-based assessment of Oregon wetland
and riparian sites: Statewide classification and profiles. Oregon Division of State Lands,
Salem, Or.
Bisson, P. and D. Montgomery. 1996. Valley segments, stream reaches, and channel units. In:
Hauer, F. R. and G. Lamberti (eds.), Methods of Stream Ecology. Academic Press, New
York.
Brinson, M. M. 1993. A hydrogeomorphic classification for wetlands. Technical Report WRP-DE-
4, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS .
Cowardin, L. M. Carter, V. Golet, F. C. and LaRoe, E. T. 1979. Classification of wetlands and
deepwater habitats of the United States. U. S. Fish and Wildlife Service, Biological Services
Program. Report No. FWS/OBS-79/31 .
Gwin, S. Kentula, M. and Shaffer, P. 1999. Evaluating the effects of wetland regulation through
hydrogeomorphic classification and landscape profiles. Wetlands 19: 477-489.
Hauer, F. R. and Smith, R. D. 1998. The hydrogeomorphic approach to functional assessment of
riparian wetlands: evaluating impacts and mitigation on river floodplains in the U.S.A.
Freshwater Biology 40: 517-530.
Johnson, J. B. 2000. Documentation of reference conditions in the slope wetlands of the southern
Rocky Mountains: reference database, site descriptions, and revised functional models.
Report Submitted to U.S. EPA , Region 8 and the Colorado Department of Natural
Resources.
Keate, N. 1998. Utah Riverine Model: low gradient, low elevation, riverine wetlands of the
Colorado Plateau and Great Basin of Utah - Draft. Report submitted to U.S. EPA, Region
Keate, N. 2001. Utah montane slope (fens and wet meadows) functional assessment model (draft).
Report submitted to U.S. EPA, Region 8 .
Noe, D. Cooper, D. Barry, A. Pavlik, M. Kolm, K. Harper-Arabie, R. Emerick, J. Arp, C. and
Chimner, R. 1998. Characterization and functional assessment of reference wetlands in
Colorado. Report submitted to Colorado Dept. of Natural Resources and US EPA Region
VIII.
SAIC. 2000. Summit County wetland functional assessment. Science Applications International
Corporation, Lakewood, CO. Report prepared for the Summit County Community
Development Division .
32
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
Smith, R. D. Ammann, A. Bartoldus, C. and Brinson, M. M. 1995. An approach for assessing
wetland functions using hydrogeomorphic classification, reference wetlands and functional
indices. Technical Report WRP-DE-9, U.S. Army Corps of Engineer Waterways
Experiment Station, Vicksburg, MS .
Strahler, A. N. 1957. Quantitative analysis of watershed geomorphology. American
Geophysical Union Transactions 38:1117-1142.
Tiner, R. 2000. Keys to waterbody type and hydrogeomorphic-type wetland descriptors for U.S.
waters and wetlands (Operational Draft). U. S. Fish and Wildlife Service National Wetlands
Inventory Project Report, September 2000.
US EPA. 1995. Generic quality assurance project plan guidance for programs using community
level biological assessment in wadable streams and rivers. Office of Water, EPA 841-B-95-
004.
33
-------
Summit County HGM Categories and Keys
Operational Draft - Version 2.1
November 12, 2004
APPENDIX 1
EXAMPLES OF KEYING WETLANDS USING
TOPOGRAPHIC AND GIS DATA
34
-------
Figure Al. A section of map showing the setting of a large, outflow
slope wetland (lower left hand quarter of the map. The wetland,
located in a alpine cirque, forms the headwaters of a stream system
flowing to the north. A similar, but smaller, wetland can be seen in
the lower right hand corner.
Figure A2. Map showing a isolated slope wetland occurring on
the slopes of Buffalo Mountain near tree line. In this case,
groundwater that discharges in the alpine zone apparently
recharges as it flows onto the more shallowly graded subalpine
slopes.
-------
Figure A3. Four examples of through-flow slope wetlands. The asymmetrical shape and
extension of the wetland well above the channel helps to identify these sites as slope, rather
than, riverine wetlands.
-------
Mid-Order, Unconfined Riverine
Figure A4. Four low-order, narrow riverine wetlands (confined valley) oriented north-south
along the top of the figure. These streams are tributary to a mid-order (5th order) river with a
well developed (unconfined) floodplain.
-------
Figure A5 a and b. Outlined in yellow, kettle-pond depressional wetlands can
be seen on lobular, glacial terrain (A). Man-made depressional wetlands in the
form of gravel mine ponds are evident in Map B.
-------
Figure A6. Outlined in pale pink, man-made lacustrine fringe wetlands can be seen on the
margins of Dillon Reservoir.
------- |