Riparian vegetation as an indicator of riparian condition: Detecting departures from historic condition across the North American West

Article · November 2016with437 Reads
DOI: 10.1016/j.jenvman.2016.10.054
Abstract
Floodplain riparian ecosystems support unique vegetation communities and high biodiversity relative to terrestrial landscapes. Accordingly, estimating riparian ecosystem health across landscapes is critical for sustainable river management. However, methods that identify local riparian vegetation condition, an effective proxy for riparian health, have not been applied across broad, regional extents. Here we present an index to assess reach-scale (500 m segment) riparian vegetation condition across entire drainage networks within large, physiographically-diverse regions. We estimated riparian vegetation condition for 53,250 km of perennial streams and rivers, 25,685 km in Utah, and 27,565 km in twelve watersheds of the interior Columbia River Basin (CRB), USA. We used nationally available, existing land cover classification derived from 30 m Landsat imagery (LANDFIRE EVT) and a modeled estimate of pre-European settlement land cover (LANDFIRE BpS). The index characterizes riparian vegetation condition as the ratio of existing native riparian vegetation cover to pre-European settlement riparian vegetation cover at a given reach. Roughly 62% of Utah and 48% of CRB watersheds showed significant (>33%) to large (>66%) departure from historic condition. Riparian vegetation change was predominantly caused by human land-use impacts (development and agriculture), or vegetation change (native riparian to invasive or upland vegetation types) that likely resulted from flow and disturbance regime alteration. Through comparisons to ground-based classification results, we estimate the existing vegetation component of the index to be 85% accurate. Our assessments yielded riparian condition maps that will help resource managers better prioritize sites and treatments for reach-scale conservation and restoration activities.
6 Figures
Research article
Riparian vegetation as an indicator of riparian condition: Detecting
departures from historic condition across the North American West
William W. Macfarlane
a
,
*
, Jordan T. Gilbert
a
, Martha L. Jensen
a
, Joshua D. Gilbert
a
,
Nate Hough-Snee
a
, Peter A. McHugh
a
,
b
, Joseph M. Wheaton
a
,
c
, Stephen N. Bennett
a
,
b
,
c
a
Department of Watershed Sciences, Utah State University, 5210 Old Main Hill, Logan, UT 84322-5210, USA
b
Eco Logical Research, Inc., Providence, UT 84332, USA
c
Anabranch Solutions, LLC, Nibley, UT 84327, USA
article info
Article history:
Received 18 April 2016
Received in revised form
16 October 2016
Accepted 26 October 2016
Available online xxx
Keywords:
Floodplain assessment
Condition assessment
Riparian restoration
Landscape planning
Columbia River Basin
Utah
abstract
Floodplain riparian ecosystems support unique vegetation communities and high biodiversity relative to
terrestrial landscapes. Accordingly, estimating riparian ecosystem health across landscapes is critical for
sustainable river management. However, methods that identify local riparian vegetation condition, an
effective proxy for riparian health, have not been applied across broad, regional extents. Here we present
an index to assess reach-scale (500 m segment) riparian vegetation condition across entire drainage
networks within large, physiographically-diverse regions. We estimated riparian vegetation condition for
53,250 km of perennial streams and rivers, 25,685 km in Utah, and 27,565 km in twelve watersheds of
the interior Columbia River Basin (CRB), USA. We used nationally available, existing land cover classi-
cation derived from 30 m Landsat imagery (LANDFIRE EVT) and a modeled estimate of pre-European
settlement land cover (LANDFIRE BpS). The index characterizes riparian vegetation condition as the
ratio of existing native riparian vegetation cover to pre-European settlement riparian vegetation cover at
a given reach. Roughly 62% of Utah and 48% of CRB watersheds showed signicant (>33%) to large (>66%)
departure from historic condition. Riparian vegetation change was predominantly caused by human
land-use impacts (development and agriculture), or vegetation change (native riparian to invasive or
upland vegetation types) that likely resulted from ow and disturbance regime alteration. Through
comparisons to ground-based classication results, we estimate the existing vegetation component of
the index to be 85% accurate. Our assessments yielded riparian condition maps that will help resource
managers better prioritize sites and treatments for reach-scale conservation and restoration activities.
©2016 Elsevier Ltd. All rights reserved.
1. Introduction
In semi-arid and arid environments oodplain riparian ecosys-
tems are often the dominant wetland elements in otherwise dry
landscapes (Knopf et al., 1988), providing diverse habitats and
ecosystems services. Floodplain riparian ecosystems support
disproportionately diverse plant and animal communities relative
to adjacent upland ecosystems, with manyspecies occurring only at
high abundance in riparian areas (Johnson et al., 1977; Knopf, 1985;
Soderquist and Mac Nally, 2000). Flood dynamics and the coloni-
zation and stabilization of landforms during vegetation succession
create diverse oodplain mosaics (Kleindl et al., 2015) and complex
instream habitat (Hupp and Osterkamp, 1996; Kauffman et al.,
1997) that support sh and other aquatic biota. Across the inte-
rior western U.S. however, many riparian areas have been altered or
are threatened by human impacts that directly and indirectly
impact stream hydrologic, geomorphic, and ecological processes
that shape riparian vegetation (Nilsson and Berggren, 2000;
Obedzinski et al., 2001).
Common impacts to riparian vegetation often include ow
alteration (Poff et al., 2011) from water withdrawal, diversion or
impoundment (Goodwin et al., 1997), intensive agriculture (Allan,
2004; Klemas, 2014), urbanization (Allan, 2004; Hardison et al.,
2009; Paul and Meyer, 2001), re suppression (Stone et al., 2010),
invasive plant species (Shafroth et al., 2002; Stromberg et al., 2007),
*Corresponding author.
E-mail addresses: wally.macfarlane@usu.edu (W.W. Macfarlane), jtgilbert89@
gmail.com (J.T. Gilbert), martaljensen@gmail.com (M.L. Jensen), joshuadgilby@
gmail.com (J.D. Gilbert), nate@natehough-snee.org (N. Hough-Snee), peter.a.
mchugh@gmail.com (P.A. McHugh), joe.wheaton@usu.edu (J.M. Wheaton),
bennett.ecological@gmail.com (S.N. Bennett).
Contents lists available at ScienceDirect
Journal of Environmental Management
journal homepage: www.elsevier.com/locate/jenvman
http://dx.doi.org/10.1016/j.jenvman.2016.10.054
0301-4797/©2016 Elsevier Ltd. All rights reserved.
Journal of Environmental Management xxx (2016) 1e14
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
beaver removal (Naiman et al., 1986), and upland species
encroachment (Marlow et al., 2006). One result of human distur-
bance is that as ow regimes and sediment supply are altered,
oodplains often become hydrologically disconnected from their
channels through channel narrowing or oodplain aggradation
(Pollock et al., 2014; Schumm, 1999; Simon and Rinaldi, 2006). As
oodplains and channels are decoupled, riparian plant perfor-
mance declines, reducing many riparian species' competitive abil-
ities (Scott et al., 2000).
On many oodplains, the encroachment of woody invasive
species (e.g. Tamarix spp., Elaeagnus angustifolia) or upland shrubs
(e.g. Juniperus spp., Pinus spp.) serves as a prominent indicator of
riparian habitat degradation (Harms and Hiebert, 2006; Jarnevich
et al., 2011; Wang et al., 2013). Hydrologic alteration that reduces
the magnitude, duration and frequency of oods, for example, often
precedes the expansion of Tamarix along oodplains (Dean and
Schmidt, 2011; Manners et al., 2014). Reduced ows and
increased Tamarisk abundance reduce native species' physiological
performance, shifting community composition further toward
Tamarisk (Dean and Schmidt, 2011; Manners et al., 2014). When
native riparian vegetation is replaced by invasive, woody species,
bare, alluvial oodplain landforms can become dense thickets that
rapidly accrete sediment, reducing oodplain landforms' inunda-
tion frequency and hydrologic connectivity to the channel (Dean
and Schmidt, 2011; Manners et al., 2014). When mapped, these
invasions manifest themselves as an increase in woody vegetation
cover over historic levels (Webb and Leake, 2006). Across the
interior western U.S., upland or woody invasive species' dominance
is often associated with impaired ow and sediment regimes that
limit native vegetation dispersal, establishment, growth and
competition, reducing the amount of available native, riparian
habitat (Richardson et al., 2005).
Despite widespread study of the causes and consequences of
transitions from native riparian vegetation to upland or invasive
species (Richardson et al., 2007), and the large number of vegeta-
tion change detection methodologies and techniques, utilizing
remotely sensed data (Hussain et al., 2013), regional assessments of
the magnitude and extent of riparian degradation are rare across
western North America. We attribute this largely to a lack of his-
toric data and to methodological limitations (Dunford et al., 2009;
Pert et al., 2010). While researchers have used geographic infor-
mation systems (GIS) to map riparian buffers (Aguiar and Ferreira,
2005; Apan et al., 2002; Pert et al., 2010), vegetation change (Piegay
et al., 2009), and condition (Johansen et al., 2008), most of this
research has relied on manual aerial photo interpretation or eld
visits at limited spatial extents (Goetz, 2006). For example, Dunford
et al. (2009) mapped 174 ha of the Drone River in France, while
Lawson et al. (2007) quantied vegetation change within a single
Australian catchment. To understand current ecological and phys-
ical conditions and prioritize oodplains for conservation and
restoration, Stella et al. (2013) noted that, we need to enlarge
the scope of riparian studies beyond the site and reach to a true
biogeographical perspective of the corridor, catchment, and
regional scales.
Recent advancements in image analysis software, imagery res-
olution, and the availability of accurate, free GIS data, now provide
opportunities to map changes in riparian vegetation composition,
structure, and spatial extent at unprecedented scales (Dufour et al.,
2012). These geospatial tools have evolved in parallel with similar
tools for mapping geomorphic change (Wheaton et al., 2010) and
mapping landforms (Gilvear and Bryant, 2016), allowing for
network scale evaluation and characterization of entire stream
networks, including their valley bottoms (Gilbert et al., 2016; Roux
et al., 2015). These technical advances allow for unprecedented
evaluation of hydrologic, geomorphic, and ecological change of
entire river systems. Here, we take advantage of these advances to
expand the scope of riparian condition studies to large landscapes
where human land- and water-use have altered the hydrologic,
physical, and ecological processes that historically supported native
riparian vegetation communities. We ask two questions:
(1) How does current riparian vegetation composition differ from
historic riparian vegetation composition across the western
United States?
(2) Where riparian vegetation has changed from its historic
composition, what are the causes of this transition?
We address these questions by assessing riparian vegetation
change (departure from historic condition) across the state of Utah
and within twelve watersheds of the interior Columbia River Basin
(CRB). We estimate the causes of vegetation change within discrete
reaches, mapped to entire drainage networks, and validate current
vegetation condition using eld observations. These maps of
vegetation change, and its probable causes, are presented at a
spatial resolution that can support both reach-level assessments of
current condition and watershed-scale restoration planning.
2. Methods
2.1. Riparian vegetation departure index
The riparian vegetation departure index (RVD) is a ratio that is
similar to the observedto expected(O/E) type metrics used in
environmental condition assessments (e.g., Hawkins et al., 2010).
RVD characterizes riparian vegetation condition for a given stream
reach as the ratio of existing vegetation to an estimation of pre-
European settlement vegetation coverage (Fig. 1). To numerically
calculate condition, native riparian vegetation is coded as 1and
invasive and upland classes are coded as 0in both the existing, and
pre-European settlement vegetation rasters (see supplementary
materials Table S1) and condition is calculated as the ratio of cur-
rent to historic native riparian coverage for a given reach.
To support reach-level assessments, we bound the lateral extent
of our analysis by generating analysis polygons within the valley
bottom. By denition, a valley bottom is comprised of the stream or
river channel and the associated low-lying, contemporary ood-
plain (Fryirs et al., 2015; Wheaton et al., 2015). The valley bottom is
used because it roughly represents the maximum possible extent of
riparian vegetation (Ilhardt et al., 2000). Analysis polygons are
generated in three steps. First, each valley bottom unit is split into a
series of Thiessen polygons, with centroids located at the midpoint
of each stream segment (Fig. 1). Thiessen polygons were chosen for
this process because their geometric properties guarantee that all
points within a polygon are closer to its centroid than to any other
polygons (Esri, 2016). This ensures that vegetation adjacent to the
reach is applied to the correct segment, even when working with
irregular planform geometries and valley bottoms. This is similar to
the concept of Notebaert and Piegay (2013) of breaking up the
valley bottom into discrete geographic objects (DGOs) using the
Fluvial Corridor Tool (Roux et al., 2015). Second, the valley bottom is
buffered by the pixel resolution of the vegetation data (i.e., 30-m
vegetation data is buffered by 30 m) to ensure that the relevant
vegetation data is completely contained within the valley bottom in
headwater reaches (Fig. 1). Finally, we clip the Thiessen polygon
layer to the buffered valley bottom. The resulting polygons become
the analysis features for which the RVD tool calculations are sum-
marized (Fig. 1 and see supplementary materials Fig. S1).
Within each polygon, the mean of the values (i.e. the 1s and 0s)
is calculated for both the existing and historic vegetation layers,
resulting in values that represent the proportion of each polygon
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e142
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
with native riparian cover (Eqn. (1);Fig. 1 and see supplementary
materials Fig. S1). The nal processing step is to apply the RVD
calculation from the analysis polygons to reach segments and
divide the historic proportion by the existing proportion (Eqn. (2);
Fig. 1 and see supplementary materials Fig. S1). Low values (closer
to 0) signify large departures from historic riparian coverage
whereas high values (i.e., approaching or exceeding 1.0) denote that
riparian communities are relatively intact (or even increasing). To
facilitate output display we symbolize each reach based on depar-
ture from historic cover, dened as the calculated ratio subtracted
from one, which results in a percent departure (Eqn. (3)). We
categorize negligible departureas less than 10%, minor departure
10%e33%, signicant departure33%e66% and large departure>
66%. The quality of this ratio depends both on the accuracy of the
vegetation coverage datasets, and the appropriateness of the spatial
scale (i.e., reach) at which calculations are made, relative to input
data resolution.
M
rip
¼ð00
tot
Þþð11
tot
Þ
C
tot
(1)
Prop ¼M
ex
M
hist
(2)
Dep ¼1Prop (3)
2.2. Riparian vegetation conversion type classication
While RVD provides a score of vegetation's departure from
historic condition, it provides no information regarding the
potential causes of the departure, nor does it necessarily provide a
realistic target for restoration (e.g., given contemporary constraints
on the system; Dufour and Pi
egay, 2009). The riparian vegetation
conversion type classication (RVCT) compares existing land cover
types to historic land cover types for the same location, which can
provide insights into potential causes of the departure from its
historic condition. Specically, land cover classications for the
historic and existing vegetation layers are compared on a pixel-by-
pixel basis to determine whether a conversion has occurred (e.g., a
pixel classied as riparian in the historic layer is now depicted as
agriculture in the existing layer) (see supplementary materials
Fig. S1). The output network is attributed with elds containing
proportions for each type of conversion (which consist of conifer
encroachment, conversion to agriculture, conversion to grass/
shrubland, conversion to invasive, devegetation, development and
no change) for a given reach, and can be symbolized accordingly for
displaying this information. Additionally, a raster output depicting
change on a cell by cell basis is produced.
To determine the RVCT the existing and historic riparian vege-
tation rasters were coded with unique integer scores based on
vegetation type; the values were assigned based on general land
cover types such as riparian, conifer, and upland (see
supplementary materials Fig. S2, Tables S2 and S3). For land cover
classes that exist in both the existing and historic vegetation ras-
ters, we assigned identical values (i.e. riparian vegetation types are
coded as 100 in both the existing and historic land cover). Unique
values were assigned to land cover types within the existing
vegetation raster that did not exist in the historic (e.g. agriculture,
urban, invasive vegetation). The values from the existing vegetation
raster were then subtracted from the values of the historic vege-
tation raster, resulting in new, unique values representing specic
conversion types (see supplementary materials Table S4). Within
Fig. 1. A conceptual diagram of the riparian vegetation departure index showing how mid points of the drainage network (1) are used to generate Thiessen polygons (2) and how
these polygons are buffered by the resolution of the vegetation data to ensure that vegetation data is completely contained within the valley bottom in headwater reaches (3).
Riparian vegetation departure is calculated using the ratio of existing area of native riparian vegetation (4) to historic area of native riparian vegetation (5) and the output is a
segmented drainage network containing riparian departure from historic condition scores (6).
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e14 3
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
each Thiessen polygon, the relative proportion of each conversion
type was calculated, and these values were attributed to the
drainage network output with a unique eld for each conversion
type. For symbolization, if the proportion of no changefor a reach
was 0.85 or greater, the reach was symbolized as no change.
Otherwise, it was symbolized by the next most dominant conver-
sion type. When symbolized with a conversion type, a reach was
sub-categorized into minor, moderate or signicant conversions
(e.g., minor conifer encroachment) based on the proportion asso-
ciated with the conversion. If the dominant conversion's proportion
was less than or equal to 0.25, it was categorized as minor. If the
proportion was between 0.25 and 0.5, it was categorized as mod-
erate, and if it was greater than 0.5, it was categorized as signicant
(see supplementary materials Fig. S2). We have packaged the RVD
and RVCT indexes into the RVD tool and the supplement materials
to this paper provides RVD tool documentation.
2.3. Case study application and validation
2.3.1. Study locations
Our assessment of vegetation condition focused on perennial
drainage networks across Utah (z25,600 km of streams), as well as
twelve watersheds within the CRB. Collectively, these two regions
are the focus of ongoing riparian restoration efforts that aim to
improve the status of imperiled riparian and aquatic species. Focal
watersheds within the CRB include the John Day and Upper Grand
Ronde Oregon, the Tucannon, Entiat, Wenatchee, and Asotin in
Washington, and the Upper Salmon, Yankee Fork, Lemhi, Lochsa,
Lower Clearwater, and South Fork Clearwater, Idaho (totaling
z27,565 km of streams). The CRB effort was part of the Columbia
Habitat Monitoring Program (CHaMP; http://champmonitoring.
org) which tracks the status and trend of anadromous salmonid
habitat throughout the CRB (Bouwes et al., 2011).
Utah is a physiographically diverse landscape covering
219,808 km
2
that range from alpine meadows to desert canyons
and support a wide range of riparian conditions. The state of Utah
includes three primary physiographic regions, each with unique
topographic, geologic, and geomorphic characteristics: the Colo-
rado Plateau, the Basin and Range, and the Middle Rocky Mountains
(USGS, 2016c). Elevations in Utah range from 664 m at Beaver Dam
Wash in the southwestern corner of the state to 4123 m high King's
Peak in the Uinta Mountains. Utah provides an ideal range of
landscapes across which the robustness of a riparian vegetation
departure analysis can be tested. Similarly, the CRB is comprised of
the Columbia Plateau Physiographic Province (USGS, 2016c) which
includes a diverse range of landscapes including mountains, pla-
teaus, canyons, and the rolling hills and deep soils of Washington
and Oregon's Palouse region (Fig. 2).
2.3.2. Case study data inputs
The segmented drainage network. We used the US Geological
Survey (USGS) National Hydrography Dataset (NHD), a carto-
graphically derived 1:24,000 drainage network (USGS, 2016b) that
we reduced to perennial streams and rivers (Table 1). We
segmented the drainage network longitudinally into 500 m long
segments because this was a reasonable length along which to
sample 30 m LANDFIRE vegetation data within the valley bottom to
get a representative sample of vegetation condition. The choice of
reach length here also reects a resolution useful for conservation
and restoration planning.
The valley bottom polygon. We used the Valley Bottom
Extraction Tool (V-BET) with manual editing to delineate valley
bottoms (Gilbert et al., 2016)(Table 1). V-BET is an ArcGIS Toolbox
and the source code is downloadable at https://bitbucket.org/
jtgilbert/riparian-condition-assessment-tools/wiki/Home. V-BET
requires two inputs: a DEM and a polyline drainage network. For
this regional application, only nationally available USGS National
Elevation Data (NED) 10 m DEMs (USGS, 2016a) provided the
required coverage. We used NHD cartographic 1:24,000 scale
dataset (USGS, 2016b), subset to perennial streams and rivers as the
drainage network.
Vegetation layers. For the existing vegetation layer we used
LANDFIRE EVT 2012 Version LF_1.3.0 (the latest version available), a
nationwide 30 m Landsat satellite imagery-based land cover clas-
sication (LANDFIRE, 2016a)(Table 1). For the historical vegetation
layer, we used the LANDFIRE Biophysical Settings (BpS) layer
(Table 1). The BpS layer is an estimation of the vegetation that may
have been dominant on the landscape prior to Euro-American
settlement. BpS is based on both the biophysical environment
and an approximation of the historical disturbance regime
(LANDFIRE, 2016b). LANDFIRE uses the Landscape Succession
Model (LANDSUM) a spatially explicit vegetation dynamics simu-
lation program where succession is regarded as a deterministic
process and disturbances (e.g. re, insects, and disease) are treated
as stochastic processes (Rollins, 2009).
Zhu et al. (2006) used a cross-validation technique to determine
that LANDFIRE EVT data layer accuracies were between 60 and 89%
and that LANDFIRE BpS accuracies were between 64 and 67%. A
study in Utah that reconstructed reference conditions for 11
forested sites based on trees present in 1880 using tree-ring data
found that LANDFIRE BpS data were 58% accurate compared with
the tree-ring data for each plot (Swetnam and Brown, 2010). It is
important to note that this accuracy assessment was conducted for
many more classes, (i.e., individual forest tree species) hence, a
lower classication accuracy would be expected compared to
aggregating to relevant classes (native riparian, nonnative riparian,
and upland) like we have done in this study.
A large river polygon. In some cases, there are raster cells
falling within valley bottoms that are not classied as vegetation,
either under existing (EVT) or historic (BpS) conditions, and must
be treated differently in RVD/RVCT calculations. The open water
class falls into this category and, accordingly, was coded as No Data
in large rivers whereas it was coded as 1outside of large rivers.
This coding was determined through test runs and comparisons to
eld data that revealed that if all open water was classied as a 1it
overestimated departure from historic condition, but that if all
open water was classied as No Datait underestimated departure
from historic condition. Open water cells outside of large rivers are
generally single isolated cells among various vegetation classes so
they do not have a large impact on the departure calculations.
2.4. Accuracy assessment analysis
We assessed the accuracy of existing vegetation layer input data
by estimating how well the LANDFIRE EVT classication compared
to eld observations of both vegetation extent and composition
(i.e., % of oodplain occupied and % native riparian vegetation). We
performed eld assessments in randomly selected analysis poly-
gons in the Weber watershed of northern Utah, and systematically
stratied analysis polygons in the Tucannon watershed of south-
eastern Washington. The surveys were stratied based on access,
quality of the vantage point, and USEPA Level IV Ecoregions (EPA,
2016). The eld data collection consisted of estimating native ri-
parian cover within analysis polygons from viewpoints above the
valley bottom. We surveyed 91 analysis polygons, 31 within the
Weber watershed (see supplementary materials Fig. S4) and 60
within the Tucannon watershed (see supplementary materials
Fig. S5).
Agreement between index-based and eld-based assessments
of native riparian coverage was evaluated using an error matrix
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e144
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
Fig. 2. Study locations map showing the state of Utah and the twelve watersheds within the interior Columbia River Basin that were assessed using the riparian vegetation depature
index and riparian vegetation conversion type classication. U.S. Environmental Protection Agency Level III Ecoregions are also displayed for physiographic context.
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e14 5
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
(Foody, 2002). Overall accuracy was calculated as the proportion of
points correctly classied by LANDFIRE EVT and Cohen's Kappa (K)
statistic as a measure of ground and map agreement, adjusted for
the agreement expected due to chance alone (Aronoff, 2005).
Additionally, for each vegetation class consumer accuracy (% of a
modeled class that mirrored the ground truth class) and producer
accuracy (% of a ground truth class that the index correctly iden-
tied), as well as errors of omission (% of a ground truth class the
index classied incorrectly) and commission (% of an index class
that was placed into the wrong ground truth class), were calculated.
3. Results
3.1. Region wide results
3.1.1. Statewide Utah application
Across Utah, the RVD tool revealed spatially variable patterns of
riparian vegetation departure from historic condition. Signicant to
large departures were evident for the large alluvial rivers, where
agricultural and urban land uses are common (Fig. 3). Minor to
negligible departures were common for the headwater streams
located on public lands (Fig. 3). Roughly 38% of the drainage
network throughout the state showed negligible to minor de-
partures from historic condition while roughly 62% showed sig-
nicant to large departures from historic condition (Fig. 4A and
Table 2), indicating that riparian vegetation along 15,736 km of
Utah's drainage network have been signicantly altered since Euro-
American settlement. These departure patterns are highlighted in
ecoregion-level summaries, where the less populated and less
intensively farmed regions showed lower departure scores. Our
RVD analysis thus suggests that a majority of Utah's riparian areas
are in an altered or degraded condition relative to historic
conditions.
The RVCT output for Utah shows similarly variable spatial pat-
terns of riparian vegetation conversion across the state (see
supplementary materials Fig. S6) to the RVD (Fig. 3). The tool
identied conversion to agriculture and developed land within the
most populated portions of the state (see supplementary materials
Fig. S6). Eight percent of the state's riparian areas have been con-
verted to agriculture and 13% have been converted to developed
lands. Seven percent of the network shows conversion to invasive
species, predominantly in the southern and southeastern parts of
the state where tamarisk invasion is common (Fig. 4 and see
supplementary materials Fig. S6). The not convertedbalance (39%
of all perennial riparian areas) were generally distributed
throughout mountainous, headwater portions of the state's
drainage network.
3.1.2. Columbia River Basin application. Across eleven of the twelve
watersheds assessed within the interior CRB, the RVD tool revealed
coherent patterns of riparian vegetation departure from historic
condition (Fig. 5). The notable exception was the Lemhi watershed,
where LANDFIRE EVT appears to model signicantly less riparian
vegetation than exists on the ground. As in Utah, signicant to large
departure was evident for the large alluvial rivers where bigger
valley bottoms allow for the most intensive land uses (Fig. 5). Minor
to negligible departure was common for headwater streams (Fig. 5).
Roughly half (52%) of the drainage networks in assessed water-
sheds showed negligible to minor departure from historic condi-
tion, while roughly half (48%) showed signicant to large departure
from historic condition (Fig. 6A and Table 3.), indicating that ri-
parian vegetation along 13,101 km in assessed CRB watersheds have
been signicantly altered since Euro-American settlement.
Across watersheds, RVD suggests riparian vegetation within the
Lemhi watershed is in the poorest condition, with over 85% of
assessed kilometers having a large departure from historic condi-
tion (Fig. 6A). The Tucannon showed the next largest departure
with over 58% with a large departure, which likely stems from the
sub-basin's location, southeast Washington's Palouse country, an
intensively farmed wheat-growing region, At the low end of the
impact spectrum, the John Day watershed showed the least
degradation in vegetation with only 22% of assessed km showing
large departure from historic condition. This may be due to the high
proportion of the drainage network occupying public lands, where
historic logging practices have greatly diminished and riparian
areas likely show some recovery. As in other watersheds, low
condition segments tend to occur predominantly along mainstem
reaches within broad alluvial valleys within the John Day Basin.
The RVCT outputs highlight logical spatial patterns of riparian
vegetation conversion across the CRB watersheds that largely track
contemporary land uses and the degree to which the imprint of
past land uses still persist (see supplementary materials Fig. S7).
The tool identied conversion to agriculture and developed land
along the most populated portions of the watersheds (Fig. 6B).
Twenty-seven percent of the rivers throughout the assessment
watersheds showed conifer encroachment. Across all watersheds,
over 6% of stream segments showed conversion to agriculture, 5%
showed conversion to developed and less than one percent showed
conversion to invasive vegetation (Fig. 6B and see supplementary
materials Fig. S6). Nearly half (46%) of the riparian areas showed
no detectable conversion of vegetation type.
3.2. Accuracy assessments
Error matrices of eld-observed riparian cover and LANDFIRE
EVT riparian cover for the Weber (Table 4 and see supplementary
materials Fig. S4) and Tucannon watersheds (Table 5 and see
supplementary materials Fig. S5) indicate a high overall level of
agreement. For the Weber watershed, overall exiting vegetation
classication accuracy was 84%. The Cohen's Kappa (K) statistic,
which ranges from 0 (no agreement) to 1 (perfect agreement) was
0.77. A Kbetween 0.61 and 0.80 is generally taken as evidence of
Table 1
Input data used in the riparian vegetation departure index.
Input data Criteria Source
Segmented drainage network Perennial streams and rivers USGS National Hydrography Dataset Cartographic 1:24,000 scale http://nhd.usgs.gov/
Digital Elevation Model Terrain model for delineating
valley bottom
USGS National Elevation Dataset 10 m Digital Elevation Model http://ned.usgs.gov/
Valley bottom polygon Maximum riparian extent Valley Bottom Extraction Tool (V-BET) https://bitbucket.org/jtgilbert/riparian-condition-
assessment-tools/wiki/Home
Existing vegetation LANDFIRE version
LF_1.3.0 2012 (EVT)
Existing vegetation LANDFIRE Existing Vegetation Type (EVT) data http://www.landre.gov/
NationalProductDescriptions21.php
Historic vegetation: LANDFIRE version
LF_1.3.0 2012 (BPS)
Historic vegetation LANDFIRE Biophysical Setting (BpS) depicted reference condition http://www.landre.gov/
NationalProductDescriptions20.php
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e146
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
Fig. 3. Map showing the riparian vegetation departure index across the perennial drainage network of Utah.
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e14 7
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
substantialagreement (Landis and Koch, 1977). Classication ac-
curacy in the Tucannon watershed was similarly favorable, with an
overall accuracy of 86% and Kof 0.81; Kbetween 0.81 and 1.00
indicates almost perfectagreement (Landis and Koch, 1977). Thus,
the RVD tool's input for characterizing contemporary native ripar-
ian vegetation coverage appears to accurately capture what on-the-
ground assessments revealed in these two watersheds.
4. Discussion
4.1. Extent of riparian vegetation change, causes, and future
applications
Floodplain riparian ecosystems are highly dynamic mosaics of
distinct landforms with different uvial and upland disturbance
regimes (Kleindl et al., 2015; Whited et al., 2007), environmental
stressors, and high rates of species turnover (Decocq, 2002).
Fig. 4. Pie chart showing (A) the riparian vegetation departure index and (B) riparian vegetation conversion type classication by U.S. Environmental Protection Agency Level III
Ecoregions across the perennial drainage network of Utah.
Table 2
Summary of the riparian vegetation departure index categories for the perennial drainage network of Utah.
Departure from historic condition Stream length (km) % of drainage network
Large 10,416 41
Signicant 5320 21
Minor 4697 18
Negligible 5144 20
Total 25,577
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e148
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
Riparian vegetation change reects successional processes that are
either cyclical, such as valley bottoms being reshaped by oods at a
given recurrence interval (Naiman et al., 2000), or directional such
as when water is withdrawn from a channel and effectively elimi-
nates ood-mediated disturbance, propagule transport, and soil
moisture that support riparian forest establishment and growth
(Souchon et al., 2008).
By mapping riparian vegetation departure from historic condi-
tion, we have shown that directional change away from dominant,
historic vegetation communities is a common phenomenon for ri-
parian areas in the western U.S. Although our analysis was largely
descriptive, the conversion type (RVCT) component, combined with
past work (e.g. Manners et al., 2014), provides insight on the
mechanisms that are driving and maintaining this directional shift
across otherwise dynamic oodplains. In many cases, human land
and water use have fundamentally changed disturbance regimes
and streamow dynamics (Poff et al., 2007). These changes have
pushed riparian succession toward upland species, including co-
nifers (Greene and Knox, 2014) and invasive woody species
(Stromberg et al., 2007). This vegetation transition can further
locklandforms into place and alter underlying hydrologic and
geomorphic processes that drive cyclical succession and maintain
diverse oodplain vegetation composition and structure (Dean and
Schmidt, 2011; Greene and Knox, 2014; Scott et al., 2000).
Cyclical succession was historically common along free-owing
rivers of the western U.S., as bank erosion, oods, droughts and re
(Kleindl et al., 2015) created oodplain mosaics of distinct land-
forms and riparian communities that vary with ood inundation
frequency and duration (Nakamura et al., 2007). This led to the
development of species-diverse oodplain mosaics that were
captured within the LANDFIRE potential vegetation dataset that
was used to determine historic condition. As oodplain modica-
tion occurred at many reaches, succession was no longer based on
uvial disturbance (e.g. Mouw et al., 2013), and competition be-
tween species, but instead upland disturbance, and direct conver-
sion of oodplains to other land uses became the dominant factors
shaping riparian vegetation dynamics.
Our assessment showed that conifer encroachment represents
the largest vegetation conversion type in both Utah (18%) and the
CRB (26%) suggesting that our assessment is effectively capturing
this pervasive form of land cover change. Increased wildre sup-
pression since European settlement, paired with groundwater
pumping and ow alteration, may allow conifer encroachment to
occur more rapidly than in areas where ow alteration has occurred
alone (Pettit and Naiman, 2007). Natural disturbance regimes (re,
hydrology) have been dramatically altered throughout the western
U.S. (Carlisle et al., 2011), fostering upland encroachment
throughout our study region and much of the western U.S.
(Theobald et al., 2010).
In the higher elevation forests, considerable research has shown
that grassland and shrublands are being replaced by forest (Zier
and Baker, 2006) and aspen stands are declining due to changes
Fig. 5. Map showing the riparian vegetation departure index output across the perennial drainage network of the twelve watersheds of sheries management concern of the interior
Columbia River Basin.
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e14 9
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
in disturbance regimes (Rogers, 2002). In the mid-elevations of the
Great Basin the encroachment of upland shrubs (e.g. Juniperus spp.,
Pinus spp.) is widespread and pervasive (Van Auken, 2000). Our
riparian condition index and conversion type assessment show that
this encroachment extends to wetter landformsdparticularly
where ow-mediated uvial disturbance and the water necessary
Fig. 6. Pie chart showing (A) the riparian vegetation departure index and (B) the riparian vegetation conversion type classication across the perennial drainage network of the twelve
watersheds of sheries management concern of the interior Columbia River Basin.
Table 3
Summary of the riparian vegetation departure index by category across the perennial drainage networkof the twelve watersheds of sheries
management concern of the interior Columbia River Basin.
Departure from historic condition Stream length (km) % of drainage network
Large 8984 33
Signicant 4117 15
Minor 4140 15
Negligible 10,324 37
Total 27,565
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e1410
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
to support hydrophytic vegetation have been removed. For
example, we found that within the Lemhi and Asotin watersheds
forest succession has shifted toward conifer, invasive, or other up-
land vegetation types. These converted reaches are likely to become
locked into place if hydrogeomorphic disturbance and upland
disturbance regimes remained altered, making it difcult to reini-
tiate cyclical succession that supports diverse plant communities
like those historically found along oodplains within the study
area.
A limitation of our conversion type assessment index is that it
qualies all non-conformity to historic condition as a degradation,
even though there are situations where conversion is not neces-
sarily a degradation. For example, a characteristic aspect of rivers is
the natural rejuvenation of valley bottoms by bank erosion, fol-
lowed by vegetation succession (Geerling et al., 2006). As such, the
conversion classes' devegetation and conversion to grass/shrubland
may have a natural cause: rejuvenation by a meandering river.
However, further investigation indicates that non-degradation
conversion is limited across our study areas. Using aerial photo
interpretation in Google Earth, we interrogated the devegetated and
conversion to grassland/shrubland outputs within a representative
watershed: Weber watershed, Utah. We examined the conversions
and attributed them to either (a) degradation resulting directly
(e.g., gravel mining or the construction of transportation infra-
structure) or indirectly (e.g., upland encroachment) from anthro-
pogenic disturbance, or (b) natural rejuvenation manifest as bare or
vegetating oodplain surfaces. The vast majority of devegetated
(~75%) and conversion to grassland/shrubland (~85%) conversions
were identied as degradation. The conversions identied as nat-
ural rejuvenation were limited in size and mainly restricted to
larger mainstem rivers. By excluding the active (bankfull) channel
from computations (i.e., lying within a large river polygon), our
analysis framework inherently minimizes the potential for mis-
characterizing rejuvenation. In practice, the consequences of this
limitation are minimal given that our tools are meant primarily for
human-affected landscapes, in which conversion due to
degradation rather than natural succession is the likely case; and
that restoration practitioners are likely to make on-the-ground site
visits before allocating resources to specic actions.
While our effort describes historic and current vegetation types,
future research could pair vegetation change with historic, recent,
or projected (future) stream ow records to inform hypotheses on
how ow alteration may be driving unidirectional succession away
from hydrophytic riparian vegetation. By linking our riparian
vegetation departure index with past and future models of ood-
plain hydrology and climate in specic terms, riparian vegetation
change from historic riparian habitat mosaics (Whited et al., 2007)
can be used to infer trajectories of future oodplain succession and
homogenization that shape species composition and habitat qual-
ity. Currently, our index maps the most obvious symptom of ri-
parian degradationdvegetation change at the aggregate
composition level. There are other, perhaps more subtle, changes
preceding these shifts that offer insight on mechanisms that foster
dynamic, healthy riparian ecosystems. Across the interior Pacic
Northwest and Utah, our index informs more specic, basin-level
research agendas to effectively answer these fundamental
questions.
4.2. Watershed and riparian management implications
We applied the RVD tool to two heavily altered, U.S. riverscapes
that are the focus of watershed planning and restoration cam-
paigns. To our knowledge, this is the rst region-wide effort to map
riparian vegetation on drainage networks as it departs from historic
condition while also identifying the causes of vegetation conver-
sion. Similar regional instream habitat and geomorphic network
analyses have been undertaken (Benda et al., 2007), as have ana-
lyses assessing riverscapes' capacity to support beaver dam build-
ing activity (Macfarlane et al., 2015), yet similar assessments of
adjacent riparian vegetation communities were largely lacking.
Our approach contrasts with intensive reach-scale studies that
identify vegetation change following known hydrologic or
Table 4
Field-based Weber watershed error matrix and Cohen's K score illustrating the agreement of ground based existing vegetation with LANDFIRE EVT classication. The diagonal
in bold text shows the correctly classied ground plots.
Field data Large Signicant Minor Negligible Total Producer Accuracy (%) Omission Error (%)
Large 14 14 100 0
Signicant 33 100 0
Minor 2 4667 33
Negligible 1 2 5863 38
Column total 14 6 6 5 31
Consumer aAccuracy (%) 100 50 67 100
Commission Error (%) 0 50 33 0
Overall accuracy (%) 84
Cohen's K 0.77
Table 5
Field-based Tucannon watershed error matrix and Cohen's K score illustrating the agreement of ground based existing vegetation with LANDFIRE EVT classication. The
diagonal in bold text shows the correctly classied ground plots.
Field data Large Signicant Minor Negligible Total Producer Accuracy (%) Omission Error (%)
Large 21 21 100 0
Signicant 2 13 15 87 13
Minor 1 2 921464 36
Negligible 1 910 90 10
Column Total 24 16 9 11 60
Consumer Accuracy (%) 88 87 100 82
Commission Error (%) 12 13 0 18
Overall Accuracy (%) 87
Cohen's K 0.81
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e14 11
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
geomorphic alteration (Merritt and Cooper, 2000; Scott et al., 2000)
or space-for-time studies that consider multiple eld-monitored
reaches to understand relationships between riparian vegetation
community types and their environmental correlates (Hough-Snee
et al., 2015). While reach-scale studies elucidate many processes
that shape vegetation communities, they do not provide sufcient
spatial coverage to inform landscape-scale riparian conservation. In
contrast, our analyses complement these ne-scale studies by of-
fering simple metrics (i.e., a departure index and conversion type
details) that can be rapidly quantied across entire watersheds.
This framework provides baseline data for detailed studies of cur-
rent vegetation composition, and future riparian vegetation trends,
and informs watershed planning, conservation, and restoration.
Output from RVD enables planners to identify conservation
areas that are intact and should be protected, and areas that have
been altered from their historic condition and potential candidates
for intervention. RVCT results, which characterize how current
vegetation differs from historic vegetation, provide further resto-
ration planning insight by narrowing candidate reaches to those
with reasonable recovery potential. For example, with two reaches
characterized by low but similar RVD values, yet different dominant
conversion types (e.g., to agricultural vs. developed use), restora-
tion resources may be preferentially allocated to the site with a
greater restoration potential. Despite clear utility for identifying
candidate sites for restoration, our tools do not prescribe restoration
treatments, nor do they spell out what the goals of a given resto-
ration should be. Riparian restoration potential is tied to the
remaining ecological, hydrological, and geomorphic processes
along a given stream and oodplain. Accordingly, historic reference
points such as past vegetation composition and structure, are
impractical restoration goals, and most conservation organizations
now identify realistic, process-based restoration targets rather than
compositional goals based on historic vegetation types.
Prior to this study, spatially explicit riparian vegetation data did
not exist for most of Utah. Resource managers within Utah now
have a consistent baseline assessment of riparian habitat condition
that is useful for planning restoration and conservation activities
for species listed, or considered for listing, under the U.S. Endan-
gered Species Act (ESA; e.g. greater sage grouse; NRCS, 2015). Ri-
parian areas are critical for sage grouse rearing, for example, and
the RVD tool can help managers identify high priority areas for
conservation and/or restoration that facilitates their life cycle
(Donnelly et al., 2016). Similarly, in our Columbia River Basin study
area, the RVD tool can provide managers with a consistent
assessment of riparian condition across several watersheds con-
taining ESA-listed salmon and steelhead populations. Although
salmon and steelhead recovery planning processes are in place (e.g.
Snake River Salmon Recovery Board, 2011), many restoration de-
cisions are still made at the sub-basin level and informed by the
best available data, which can often be limited to expert opinion
(Booth et al., 2016). Our work provides freely available and
consistently interpretable data that can inform future basin-wide
assessments of riparian condition and ultimately streamline
aquatic and oodplain habitat recovery planning.
5. Conclusions
The index-derived riparian vegetation departure from historic
condition data provides important baseline information on how
riparian vegetation has changed across Utah and the interior
Columbia River Basin. This approach was appropriate for coarse-
scale evaluations of riparian vegetation condition across regional
drainage networks and is exible and can be easily updated with
higher resolution or better quality inputs as they become available.
High-resolution riparian vegetation imagery may be necessary in
areas with narrow riparian corridors that cannot be effectively
captured with 30 m resolution data. The index provides informa-
tion that can guide watershed- and reach-scale riparian conserva-
tion and restoration planning.
Data availability
The outputs of this work are published in both a shapele
format useable in any GIS program and as KML les for exploring
and visualizing outputs in Google Earth. The outputs are publicly
available at: http://etal.joewheaton.org/rcat and the source code of
the Riparian Condition Assessment Tool (R-CAT) is available at:
https://bitbucket.org/jtgilbert/riparian-condition-assessment-
tools/wiki/Tool_Documentation/RVD. Data in KML format is also
available online in Appendix A.
Acknowledgements
This work was supported by the U.S. Department of the Interior
Bureau of Land Management (USU Award No. 151010), the Utah
Department of Natural Resources' Endangered Species Mitigation
Fund (USU Award No. 140600), Utah Division of Wildlife Resources'
Pittman and Robertson Fund (USU Award No. 150736), the Snake
River Salmon Recovery Board through Eco Logical Research (USU
Award No. 200239) and the Bonneville Power Administration (BPA
project numbers: CHaMP 2011-006 and ISEMP 2013-017), as part of
the Columbia Habitat Monitoring Program (http://
champmonitoring.org) through a sub-award from Eco Logical
Research (USU Award No. 150737). N. Hough-Snee was supported in
part by a STAR Fellowship awarded by the U.S. Environmental
Protection Agency (USU Award no. 91768201e0).
We are grateful to Justin Jimenez (BLM) who had the vision to
undertake a riparian assessment across the Colorado Plateau, and
built the partnerships for successful implementation. The devel-
opment of the index benetted greatly from insights and conver-
sations with Jeremy Jarnecke (BLM), Russell Norvell (UDWR), Jimi
Gragg (UDWR), Chris Keleher (UDNR), Frank Howe (USU), Justin
Shannon (UDWR), Gary O'Brien (USU), Phaedra Budy (USGS/USU),
Konrad Hafen (USU), Chris Jordan (NOAA) and the Weber River
Partnership. Chalese Hafen, Shane Hill and Chris Smith provided
GIS support. Reid Camp, Andrew Hill, Elijah Portugal, and Scott
Shahverdian eld-validated index outputs in the Tucannon and
Weber watersheds. We thank two anonymous reviewers for their
helpful feedback that greatly improved the manuscript.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jenvman.2016.10.054.
References
Aguiar, F.C., Ferreira, M.T., 2005. Human-disturbed landscapes: effects on compo-
sition and integrity of riparian woody vegetation in the Tagus River basin. Port.
Environ. Conserv. 32, 30e41.
Allan, J., 2004. Landscapes and riverscapes: the inuence of land use on stream
ecosystems. Annu. Rev. Ecol. Evol. S 35, 257e284.
Apan, A.A., Raine, S.R., Paterson, M.S., 2002. Mapping and analysis of changes in the
riparian landscape structure of the Lockyer Valley catchment, Queensland,
Australia. Landsc. Urban Plan. 59, 43e57.
Aronoff, S., 2005. Remote Sensing for GIS Managers. Esri Press Redlands, CA.
Benda, L., Miller, D., Andras, K., Bigelow, P., Reeves, G., Michael, D., 2007. NetMap: a
new tool in support of watershed science and resource management. For. Sci.
53, 206e219.
Booth, D., Scholz, J., Beechie, T., Ralph, S., 2016. Integrating limiting-factors analysis
with process-based restoration to improve recovery of endangered salmonids
in the Pacic Northwest, USA. Water 8, 174.
Bouwes, N., Moberg, J., Weber, N., Bouwes, B., Bennett, S., Beasley, C., Jordan, C.E.,
Nelle, P., Polino, M., Rentmeester, S., Semmens, B., Volk, C., Ward, M.B., White, J.,
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e1412
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
2011. Scientic Protocol for Salmonid Habitat Surveys within the Columbia
Habitat Monitoring Program. Integrated Status and Effectiveness Monitoring
Program, Wauconda, WA, p. 118.
Carlisle, D.M., Wolock, D.M., Meador, M.R., 2011. Alteration of streamow magni-
tudes and potential ecological consequences: a multiregional assessment. Front.
Ecol. Environ. 9, 264e270.
Dean, D.J., Schmidt, J.C., 2011. The role of feedback mechanisms in historic channel
changes of the lower Rio Grande in the Big Bend region. Geomorphology 126,
333e349.
Decocq, G., 2002. Patterns of plant species and community diversity at different
organization levels in a forested riparian landscape. J. Veg. Sci. 13, 91e106 .
Donnelly, J., Naugle, D., Hagen, C., Maestas, J., 2016. Public lands and privatewaters:
scarce mesic resources structure land tenure and sage-grouse distributions.
Ecosphere 7, 1e15.
Dufour, S., Muller, E., Straatsma, M., Corgne, S., 2012. Image Utilisation for the Study
and Management of Riparian Vegetation: Overview and Applications,
Advancing River Restoration and Management, Volume 1: Fluvial Remote
Sensing for Science and Management. John Wiley &Sons, Somerset, NJ.
Dufour, S., Pi
egay, H., 2009. From the myth of a lost paradise to targeted river
restoration: forget natural references and focus on human benets. River Res.
Appl. 25, 568e581.
Dunford, R., Michel, K., Gagnage, M., Piegay, H., Tremelo, M.L., 2009. Potential and
constraints of Unmanned Aerial Vehicle technology for the characterization of
Mediterranean riparian forest. Int. J. Remote Sens. 30, 4915e4935.
EPA, 2016. Ecoregions. US Environmental Protection Agency. https://www.epa.gov/
eco-research/ecoregions (accessed 10 Mar 2016).
Esri, 2016. GIS Dictionary: Thiessen Polyons. Esri, Redlands, CA.
Foody, G.M., 2002. Status of land cover classication accuracy assessment. Remote
Sens. Environ 80, 185e201.
Fryirs, K.A., Wheaton, J.M., Brierley, G.J., 2015. An approach for measuring
connement and assessing the inuence of valley setting on river forms and
processes. Earth Surf. Process. 41, 701e710.
Geerling, G.W., Ragas, A.M.J., Leuven, R.S.E.W., van den Berg, J.H., Breedveld, M.,
Liefhebber, D., Smits, A.J.M., 2006. Succession and rejuvenation in oodplains
along the river allier (France). Hydrobiologia 565, 71e86.
Gilbert, J.T., Macfarlane, W.W., Wheaton, J.M., 2016. V-BET: a GIS tool for delineating
valley bottoms across entire drainage networks. Comput. Geosci. 97, 1e14.
Gilvear, D., Bryant, R., 2016. Analysis of Remotely Sensed Data for Fluvial Geo-
morphology and River Science, Tools in Fluvial Geomorphology. John Wiley &
Sons, Ltd, pp. 103e132.
Goetz, S.J., 2006. Remote sensing of riparian buffers: past progress and future
prospects. J. Am. Water Resour. As 42, 133e143.
Goodwin, C.N., Hawkins, C.P., Kershner, J.L., 1997. Riparian restorationin the western
United States: overview and perspective. Restor. Ecol. 5, 4e14.
Greene, S.L., Knox, J.C., 2014. Coupling legacy geomorphic surface facies to riparian
vegetation: assessing red cedar invasion along the Missouri River downstream
of Gavins Point dam, South Dakota. Geomorphology 204, 277e286.
Hardison, E.C., O'Driscoll, M.A., DeLoatch, J.P., Howard, R.J., Brinson, M.M., 2009.
Urban land use, channel incision, and water table decline along coastal plain
streams, North Carolina. J. Am. Water Resour. As 45, 1032e1046.
Harms, R.S., Hiebert, R.D., 2006. Vegetation response following invasive tamarisk
(Tamarix spp.) Removal and implications for riparian restoration. Restor. Ecol.
14, 461e472.
Hawkins, C.P., Olson, J.R., Hill, R.A., 2010. The reference condition: predicting
benchmarks for ecological and water-quality assessments. J. N. Am. Benthol.
Soc. 29, 312e343.
Hough-Snee, N., Roper, B., Wheaton, J., Lokteff, R., 2015. Riparian vegetation com-
munities of the American Pacic Northwest are tied to multi-scale environ-
mental lters. River Res. Appl. 31, 1151e1165 .
Hupp, C.R., Osterkamp, W., 1996. Riparian vegetation and uvial geomorphic pro-
cesses. Geomorphology 14, 277e295.
Hussain, M., Chen, D., Cheng, A., Wei, H., Stanley, D., 2013. Change detection from
remotely sensed images: from pixel-based to object-based approaches. ISPRS J.
Photogramm. 80, 91e106 .
Ilhardt, B.L., Verry, E.S., Palik, B.J., 2000. Dening riparian areas, Forestry and the
riparian zone, Orono. Maine 7e14.
Jarnevich, C.S., Evangelista, P., Stohlgren, T.J., Morisette, J., 2011. Improving national-
scale invasion maps: tamarisk in the western United States. West. N. Am. Nat.
71, 164e175 .
Johansen, K., Phinn, S., Lowry, J., Douglas, M., 2008. Quantifying indicators of ri-
parian condition in Australian tropical savannas: integrating high spatial res-
olution imagery and eld survey data. Int. J. Remote Sens. 29, 7003e7028.
Johnson, R.R., Haight, L.T., Simpson, J.M., 1977. Endangered Species Vs. Endangered
Habitats: a Concept, Importance, Preservation and Management of Riparian
Habitat: a Symposium. USDA Forest Service General Technical Report RM-43,
pp. 68e79.
Kauffman, J.B., Beschta, R.L., Otting, N., Lytjen, D., 1997. An ecological perspective of
riparian and stream restoration in the western United States. Fisheries 22,
12e24.
Kleindl, W., Rains, M., Marshall, L., Hauer, F., 2015. Fire and ood expand the
oodplain shifting habitat mosaic concept. Freshw. Sci. 34, 1366.
Klemas, V., 2014. Remote sensing of riparian and wetland buffers: an overview.
J. Coast. Res. 30, 869e880.
Knopf, F.L., 1985. Signicance of Riparian Vegetation to Breeding Birds across an
Altitudinal Cline, Riparian Ecosytems and Their Management: Reconciling
Conicting Uses. USDA Forest Service General Technical Report, RM-120,
pp. 105e111.
Knopf, F.L., Johnson, R.R., Rich, T., Samson, F.B., Szaro, R.C., 1988. Conservation of
riparian ecosystems in the United States. Wilson Bull. 100, 272e284.
LANDFIRE, 2016a. Biophysical Setting (BPS) Layer Landscape Fire and Resource
Management Planning Tools Project. http://www.landre.gov/
NationalProductDescriptions20.php (assessed 23 March 2016).
LANDFIRE, 2016b. Existing Vegetation Type (EVT) Layer Landscape Fire and
Resource Management Planning Tools Project. http://www.landre.gov/
NationalProductDescriptions21.php (accessed 23 April 2016).
Landis, J.R., Koch, G.G., 1977. The measurement of observer agreement for cate-
gorical data. Biometrics 33, 159e174.
Lawson, T., Gillieson, D., Goosem, M., 2007. Assessment of riparian rainforest
vegetation change in tropical North Queensland for management and restora-
tion purposes. Geogr. Res. 45, 387e397.
Macfarlane, W.W., Wheaton, J.M., Bouwes, N., Jensen, M.L., Gilbert, J.T., Hough-
Snee, N., Shivik, J.A., 2015. Modeling the capacity of riverscapes to support
beaver dams. Geomorphology. http://dx.doi.org/10.1016/
j.geomorph.2015.11.019.
Manners, R.B., Schmidt, J.C., Scott, M.L., 2014. Mechanisms of vegetation-induced
channel narrowing of an unregulated canyon river: results from a natural
eld-scale experiment. Geomorphology 211, 100e115.
Marlow, C.B., Tucker, R., Sauer, B., Shea, V., 2006. Broadening the scope of prescribed
res: opportunities to rehabilitate degraded riparian zones. For. Ecol. Manag.
S169.
Merritt, D.M., Cooper, D.J., 2000. Riparian vegetation and channel change in
response to river regulation: a comparative study of regulated and unregulated
streams in the Green River Basin, USA. Regul. River 16, 543e564.
Mouw, J.E.B., Chafn, J.L., Whited, D.C., Hauer, F.R., Matson, P.L., Stanford, J.A., 2013.
Recruitment and successional dynamics diversify the shifting habitat mosaic of
an Alaskan oodplain. River Res. Appl. 29, 671e685.
Naiman, R.J., Bilby, R.E., Bisson, P.A., 2000. Riparian ecology and management in the
Pacic coastal rain forest. Bioscience 50, 996e1011.
Naiman, R.J., Melillo, J.M., Hobbie, J.E., 1986. Ecosystem alteration of boreal forest
streams by beaver (Castor Canadensis). Ecology 67, 1254e1269.
Nakamura, F., Shin, N., Inahara, S., 2007. Shifting mosaic in maintaining diversity of
oodplain tree species in the northern temperate zone of Japan. For. Ecol.
Manag. 241, 28e38.
Nilsson, C., Berggren, K., 2000. Alterations of riparian ecosystems caused by river
regulation: dam operations have caused global-scale ecological changes in ri-
parian ecosystems. How to protect river environments and human needs of
rivers remains one of the most important questions of our time. Bioscience 50,
783e792.
Notebaert, B., Piegay, H., 2013. Multi-scale factors controlling the pattern of ood-
plain width at a network scale: the case of the Rhone basin, France. Geo-
morphology 200, 155e171.
NRCS, 2015. Sage Grouse Initiative 2.0. http://www.sagegrouseinitiative.com/
(accessed 14 March 2016).
Obedzinski, R.A., Shaw, C., Neary, D.G., 2001. Declining woody vegetation in riparian
ecosystems of the western United States. West. J. Appl. For. 16, 169e181.
Paul, M.J., Meyer, J.L., 2001. Streams in the urban landscape. Annu. Rev. Ecol. S. 32,
333.
Pert, P., Butler, J., Brodie, J., Bruce, C., Honzak, M., Kroon, F., Metcalfe, D., Mitchell, D.,
Wong, G., 2010. A catchment-based approach to mapping hydrological
ecosystem services using riparian habitat: a case study from the Wet Tropics,
Australia. Ecol. Complex. 7, 378e388.
Pettit, N.E., Naiman, R.J., 2007. Fire in the riparian zone: characteristics and
ecological consequences. Ecosystems 10, 673e687.
Piegay, H., Alber, A., Slater, L., Bourdin, L., 2009. Census and typology of braided
rivers in the French Alps. Aquat. Sci. 71, 371e388.
Poff, B., Koestner, K.A., Neary, D.G., Henderson, V., 2011. Threats to riparian eco-
systems in Western North America: an analysis of existing literature. J. Am.
Water Resour. As 47, 1241e1254.
Poff, N.L., Olden, J.D., Merritt, D.M., Pepin, D.M., 2007. Homogenization of regional
river dynamics by dams and global biodiversity implications. P. Natl. Acad. Sci.
U. S. A. 104, 5732e5737.
Pollock, M.M., Beechie, T.J., Wheaton, J.M., Jordan, C.E., Bouwes, N., Weber, N.,
Volk, C., 2014. Using beaver dams to restore incised stream ecosystems.
Bioscience 64, 279e290.
Richardson, D.M., Holmes, P.M., Esler, K.J., Galatowitsch, S.M., Stromberg, J.C.,
Kirkman, S.P., Py
sek, P., Hobbs, R.J., 2007. Riparian vegetation: degradation, alien
plant invasions, and restoration prospects. Divers. Distributions 13, 126e139.
Richardson, J.S., Bilby, R.E., Bondar, C.A., 2005. Organic matter dynamics in small
streams of the Pacic Northwest. J. Am. Water Resour. As 41, 921e934.
Rogers, P., 2002. Using forest health monitoring to assess aspen forest cover change
in the southern rockies ecoregion. For. Ecol. Manag. 155, 223e236.
Rollins, M.G., 2009. LANDFIRE: a nationally consistent vegetation, wildland re, and
fuel assessment. Int. J. Wildland Fire 18, 235e249.
Roux, C., Alber, A., Bertrand, M., Vaudor, L., Pi
egay, H., 2015. FluvialCorridor: a new
ArcGIS toolbox package for multiscale riverscape exploration. Geomorphology
242, 29e37.
Schumm, S.A., 1999. Causes and controls of channel incision. In: Darby, S.E.,
Simon, A. (Eds.), Incised River Channels. Wiley, Chichester, pp. 19e33.
Scott, M.L., Lines, G.C., Auble, G.T., 2000. Channel incision and patterns of cotton-
wood stress and mortality along the Mojave River, California. J. Arid. Environ.
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e14 13
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
44, 399e414.
Shafroth, P.B., Stromberg, J.C., Patten, D.T., 2002. Riparian vegetation response to
altered disturbance and stress regimes. Ecol. Appl. 12, 107e123.
Simon, A., Rinaldi, M., 2006. Disturbance, stream incision, and channel evolution:
the roles of excess transport capacity and boundary materials in controlling
channel response. Geomorphology 79, 361e383.
Snake River Salmon Recovery Board, 2011. Snake River Salmon Recovery Plan for SE
Washington: 2011 Version. Washington Governor's Salmon Recovery Ofce,
p. 598. http://snakeriverboard.org/wpi/wp-content/uploads/2013/01/Full-
Version-SE-WA-recovery-plan-121211.pdf.
Soderquist, T.R., Mac Nally, R., 2000. The conservation value of mesic gullies in dry
forest landscapes: mammal populations in the boxeironbark ecosystem of
southern Australia. Biol. Conserv. 93, 281e291.
Souchon, Y., Sabaton, C., Deibel, R., Reiser, D., Kershner, J., Gard, M., Katopodis, C.,
Leonard, P., Poff, N.L., Miller, W.J., Lamb, B.L., 2008. Detecting biological re-
sponses to ow management: missed opportunities; future directions. River
Res. Appl. 24, 506e518.
Stella, J., Rodríguez-Gonz
alez, P., Dufour, S., Bendix, J., 2013. Riparian vegetation
research in Mediterranean-climate regions: common patterns, ecological pro-
cesses, and considerations for management. Hydrobiologia 719, 291e315.
Stone, K.R., Pilliod, D.S., Dwire, K.A., Rhoades, C.C., Wollrab, S.P., Young, M.K., 2010.
Fuel reduction management practices in riparian areas of the western USA.
Environ. Manage 46, 91e10 0.
Stromberg, J.C., Lite, S.J., Marler, R., Paradzick, C., Shafroth, P.B., Shorrock, D.,
White, J.M., White, M.S., 2007. Altered stream-ow regimes and invasive plant
species: the tamarix case. Global Ecol. Biogeogr. 16, 381e393.
Swetnam, T.L., Brown, P.M., 2010. Comparing selected re regime condition class
(FRCC) and LANDFIRE vegetation model results with tree-ring data. Int. J.
Wildland Fire 19, 1e13.
Theobald, D.M., Merritt, D.M., Norman, J.B., 2010. Assessment of Threats to Riparian
Ecosystems in the Western US. USDA Stream Systems Technology Center and
Colorado State University, Fort Collins, CO.
USGS, 2016a. National Elevation Data. US Geological Survey. http://nationalmap.
gov/elevation.html (accessed 23 July 2016).
USGS, 2016b. National Hydrography Dataset. US Geological Survey. http://nhd.usgs.
gov/ (accessed July 6 2016).
USGS, 2016c. Physiographic Divisions of the Conterminous U. S. US Geological
Survey. http://water.usgs.gov/GIS/metadata/usgswrd/XML/physio.xml
(accessed 15 May 2016).
Van Auken, O.W., 2000. Shrub invasions of North American semiarid grasslands.
Annu. Rev. Ecol. Evol. S 31, 197e215.
Wang, L., Silv
an-C
ardenas, J.L., Yang, J., Frazier, A.E., 2013. Invasive saltcedar
(Tamarisk spp.) distribution mapping using multiresolution remote sensing
imagery. Prof. Geogr. 65, 1e
15.
Webb, R.H., Leake, S.A., 2006. Ground-water surface-water interactions and long-
term change in riverine riparian vegetation in the southwestern United
States. J. Hydrol. 320, 302e323.
Wheaton, J.M., Brasington, J., Darby, S.E., Sear, D.A., 2010. Accounting for uncertainty
in DEMs from repeat topographic surveys: improved sediment budgets. Earth
Surf. Processes 35, 136e156.
Wheaton, J.M., Fryirs, K.A., Brierley, G., Bangen, S.G., Bouwes, N., O'Brien, G., 2015.
Geomorphic mapping and taxonomy of uvial landforms. Geomorphology 248,
273e295.
Whited, D.C., Lorang, M.S., Harner, M.J., Hauer, F.R., Kimball, J.S., Stanford, J.A., 2007.
Climate, hydrologic disturbance, and succession: drivers of oodplain pattern.
Ecology 88, 940e953.
Zhu, Z., Vogelmann, J., Ohlen, D., Kost, J., Chen, X., Tolk, B., 2006. Mapping existing
vegetation composition and structure for the LANDFIRE prototype project. In:
Rollins, M.G., Frame, C.K. (Eds.), The LANDFIRE Prototype Project: Nationally
Consistent and Locally Relevant Geospatial Data for Wildland Fire Management.
US Department of Agriculture. Forest Service, Rocky Mountain Research Station,
Fort Collins, CO, pp. 197e215.
Zier, J.L., Baker, W.L., 2006. A century of vegetation change in the San Juan Moun-
tains, Colorado: an analysis using repeat photography. For. Ecol. Manag. 228,
251e262.
W.W. Macfarlane et al. / Journal of Environmental Management xxx (2016) 1e1414
Please cite this article in press as: Macfarlane, W.W., et al., Riparian vegetation as an indicator of riparian condition: Detecting departures from
historic condition across the North American West, Journal of Environmental Management (2016), http://dx.doi.org/10.1016/
j.jenvman.2016.10.054
  • [Show abstract] [Hide abstract] ABSTRACT: The shape, size and extent of a valley bottom dictates the form and function of the associated river or stream. Consequently, accurate, watershed-wide delineation of valley bottoms is increasingly recognized as a necessary component of watershed management. While many valley bottom delineation approaches exist, methods that can be effectively applied across entire drainage networks to produce reasonably accurate results are lacking. Most existing tools are designed to work using high resolution topography data (i.e. > 2 m resolution Digital Elevation Model (DEM)) and can only be applied over relatively short reach lengths due to computational or data availability limitations. When these precise mapping approaches are applied throughout drainage networks (i.e. 102 to 104 km), the computational techniques often either do not scale, or the algorithms perform inconsistently. Other tools that produce outputs at broader scale extents generally utilize coarser input topographic data to produce more poorly resolved valley bottom approximations. To fill this methodology gap and produce relatively accurate valley bottoms over large areas, we developed an algorithm that accepts terrain data from one to 10 m with slope and valley width parameters that scale based on drainage area, allowing for watershed-scale valley bottom delineation. We packaged this algorithm in the Valley Bottom Extraction Tool (V-BET) as an open-source ArcGIS toolbox for ease of use. To illustrate V-BET's scalability and test the tool's robustness across different physiographic settings, we delineated valley bottoms for the entire perennial drainage network of Utah as well as twelve watersheds across the interior Columbia River Basin (totaling 55 400 km) using 10 m DEMs. We found that even when driven with relatively coarse data (10 m DEMs), V-BET produced a relatively accurate approximation of valley bottoms across the entire watersheds of these diverse physiographic regions.
    Full-text · Article · Jul 2016
Project
MorphSed, a research project funded by the Spanish Ministry, is analysing the morphosedimentary dynamics of a representative human-stressed fluvial system (suffering major local alterations due to …" [more]
Project
This project seeks to quantify the volume of water beaver dams can store at watershed scales, determine how storage capacity changes with beaver dam density, and evaluate if restoration of beaver p…" [more]
Project
To determine the effectiveness of a common stream restoration method (addition of large woody debris) to increase the geomorphic and habitat diversity of sections of Asotin Creek and to monitor the…" [more]
Article
July 2016 · Computers & Geosciences · Impact Factor: 2.05
    The shape, size and extent of a valley bottom dictates the form and function of the associated river or stream. Consequently, accurate, watershed-wide delineation of valley bottoms is increasingly recognized as a necessary component of watershed management. While many valley bottom delineation approaches exist, methods that can be effectively applied across entire drainage networks to produce... [Show full abstract]
    Article
    December 2015 · Geomorphology · Impact Factor: 2.79
      Abstract The construction of beaver dams facilitates a suite of hydrologic, hydraulic, geomorphic, and ecological feedbacks that increase stream complexity and channel–floodplain connectivity that benefit aquatic and terrestrial biota. Depending on where beaver build dams within a drainage network, they impact lateral and longitudinal connectivity by introducing roughness elements that... [Show full abstract]
      Discover more