Willow cover as a stream-recovery indicator under a conservation grazing plan

Abstract

Many rangeland streams and associated fisheries have suffered from livestock grazing as a cost of upland-forage utilization. Due to damage from intensive usage, restoration of damaged streams is now a common land-management objective. The Squaw Valley Ranch of Elko County, Nevada, US, in cooperation with the US Bureau of Land Management (BLM) and Barrick Gold Corp., is attempting to improve those portions of the Rock Creek watershed negatively affected by past ranch operations. The watershed includes both historical and occupied habitat for the threatened Lahonton cutthroat trout (Oncorhynchus clarki henshawi [Richardson]). From 2003, and continuing to the present, hot-season livestock grazing on Squaw Valley Ranch private and permitted public-land riparian areas was greatly reduced. To assess the effectiveness of this conservation effort, we (1) evaluated BLM archived images of riparian photo points in the watershed, (2) tested for change over time using data from systematic, intermittent, aerial sampling that acquired 2-cm resolution images from low-altitude surveys conducted in 2003, 2004 and 2006, and (3) compared Landsat scenes of the area from before and after 2003. Willow (Salix spp.) cover was chosen as the primary ecological indictor of riparian condition and we introduce willow canopy (m2) per m of stream length in the image field-of-view, as a practical measure of willow status. Archived images from photo points show mostly low-condition riparian plant communities, often with little or no willow canopy evident before 2003, but with conspicuous improvement thereafter. This subjective perception is supported by objective analyses finding, (1) the relative increase in willow cover nearly tripled on one stream, more than doubled on three others, and increased on all but one (fire affected) and (2) a highly significant post-2003 increase in willows in the Landsat record. Thus, the post-2003 increase in willow cover documented in three complementary lines of evidence from ground, air, and space support the predicted ecological benefits of reduced hot-season riparian grazing and the utility of 2-cm imagery as a tool for assessing watershed-wide conservation benefits from a federal cost-share-eligible conservation practice. This appears to be the first use of willow measurements from an aerial survey as a particular indicator of riparian condition and trend and the first demonstration of change detection based on objective measurements from a watershed-scale riparian monitoring effort that used systematic sampling (versus subjective selection) and high sample density to address the large Type II error (false negative) risk common to conventional land-management survey efforts.

Full text

Ecological Indicators 18 (2012) 512–519
Contents lists available at SciVerse ScienceDirect
Ecological Indicators
journal homepage: www.elsevier.com/locate/ecolind
Short communication
Willow cover as a stream-recovery indicator under a conservation grazing plan
D. Terrance Bootha,∗, Samuel E. Coxa,1, Gregg Simondsb, Eric D. Santb
aUS Department of Agriculture-Agricultural Research Service, High Plains Grassland Research Station, 8408 Hildreth Rd, Cheyenne, WY 82009, USA
bOpen Range Consulting, 6315 N Snowview Drive, Park City, UT 84098, USA
article info
Article history:
Received 23 July 2010
Received in revised form
26 December 2011
Accepted 29 December 2011
Keywords:
Aerial imagery
Effectiveness monitoring
Ecological indicators
Riparian management
Objective measurements
Watershed-scale sampling
abstract
Many rangeland streams and associated fisheries have suffered from livestock grazing as a cost of upland-
forage utilization. Due to damage from intensive usage, restoration of damaged streams is now a common
land-management objective. The Squaw Valley Ranch of Elko County, Nevada, US, in cooperation with the
US Bureau of Land Management (BLM) and Barrick Gold Corp., is attempting to improve those portions
of the Rock Creek watershed negatively affected by past ranch operations. The watershed includes both
historical and occupied habitat for the threatened Lahonton cutthroat trout (Oncorhynchus clarki henshawi
[Richardson]). From 2003, and continuing to the present, hot-season livestock grazing on Squaw Valley
Ranch private and permitted public-land riparian areas was greatly reduced. To assess the effectiveness of
this conservation effort, we (1) evaluated BLM archived images of riparian photo points in the watershed,
(2) tested for change over time using data from systematic, intermittent, aerial sampling that acquired
2-cm resolution images from low-altitude surveys conducted in 2003, 2004 and 2006, and (3) compared
Landsat scenes of the area from before and after 2003. Willow (Salix spp.) cover was chosen as the primary
ecological indictor of riparian condition and we introduce willow canopy (m2) per m of stream length in
the image field-of-view, as a practical measure of willow status. Archived images from photo points show
mostly low-condition riparian plant communities, often with little or no willow canopy evident before
2003, but with conspicuous improvement thereafter. This subjective perception is supported by objective
analyses finding, (1) the relative increase in willow cover nearly tripled on one stream, more than doubled
on three others, and increased on all but one (fire affected) and (2) a highly significant post-2003 increase
in willows in the Landsat record. Thus, the post-2003 increase in willow cover documented in three
complementary lines of evidence from ground, air, and space support the predicted ecological benefits
of reduced hot-season riparian grazing and the utility of 2-cm imagery as a tool for assessing watershed-
wide conservation benefits from a federal cost-share-eligible conservation practice. This appears to be
the first use of willow measurements from an aerial survey as a particular indicator of riparian condition
and trend and the first demonstration of change detection based on objective measurements from a
watershed-scale riparian monitoring effort that used systematic sampling (versus subjective selection)
and high sample density to address the large Type II error (false negative) risk common to conventional
land-management survey efforts.
Published by Elsevier Ltd.
“Because of a lack of uniformity in grazing, certain areas may be
sacrifice areas and be overused”; “Overuse of these sacrifice areas in
valley bottoms and around water holes is justified if the manager is
Abbreviations: BLM, US Department of the Interior, Bureau of Land Management;
COST, cosine theta; FOV, field of view; GSD, ground sample distance; NAIP, National
Agriculture Imagery Program; NDVI, Normalized Difference Vegetation Index; PIF,
pseudo invariant feature; PRA, potential riparian area; SVR, Squaw Valley Ranch;
RGB, red, green, blue, the primary colors of a color digital image; TM, Thematic
Mapper, the Landsat 5 sensor.
∗Corresponding author. Tel.: +1 307 772 2433x110; fax: +1 307 637 6124.
E-mail address: Terry.Booth@ars.usda.gov (D.T. Booth).
1Current address: USDI Bureau of Land Management, Cheyenne, WY 82009, USA.
conservative as to the area involved.” (Stoddart and Smith, 1955,
pp. 144 and 279).
1. Introduction
Cattle preference for riparian areas during the hot season
leads to riparian damage (Stoddart and Smith, 1955, pp. 144 and
279; McInnis and McIver, 2009) and management that does not
specifically control this preference is linked with overuse – spe-
cially of willow (Salix spp.) (Kauffman et al., 1983; Kovalchik and
Elmore, 1992; Schulz and Leininger, 1990; Scrimgeour and Kendall,
2003). Willow loss along desert streams has critical consequences,
including increased stream temperatures (White and Rahel,
2008; Zoellick, 2004), loss of beaver and associated habitat and
1470-160X/$ – see front matter. Published by Elsevier Ltd.
doi:10.1016/j.ecolind.2011.12.017

D.T. Booth et al. / Ecological Indicators 18 (2012) 512–519 513
water-storage capacity (Hebblewhite et al., 2005; White and Rahel,
2008), loss of the well-recognized bank-armor function against
high flows (see Vincent et al. (2009), for a recent re-affirmation of
the importance of willows for reducing bank erosion along desert
streams), and adverse effects to native trout fisheries (Harig and
Fausch, 2002; White and Rahel, 2008; Zoellick, 2004). The ecolog-
ical consequences of willow deficiencies along desert streams are
well established. The current issue is how to evaluate the effective-
ness of large-area stream-recovery efforts. How can the biological
benefit be measured to assess return on investment of watershed-
scale stream conservation programs?
In 2003, the Squaw Valley Ranch (SVR) of Elko County, Nevada,
US, in cooperation with the US Bureau of Land Management (BLM)
and Barrick Gold Corporation, instituted a conservation grazing
management program to improve riparian condition in portions of
the Rock Creek watershed affected by SVR operations. The goal is
restoration of aquatic habitat for the threatened Lahonton cutthroat
trout (Oncorhynchus clarki henshawi [Richardson]) (U.S. Federal
Register, 1975; USFWS, 1995;unpublished memorandum: Biolog-
ical opinion for the 2004 through 2024 livestock grazing system for
the Squaw Valley Allotment, Elko County, NV. United States Fish
and Wildlife Service, 2004, File Number 1-5-04-F-05. Reno, NV, 68
pp; the opinion cites historic season-long grazing as a predominant
factor in trout habitat degradation). Poor riparian condition at sur-
vey stations for key streams in the watershed from 1977 through
1997 is documented in Squaw Valley Allotment Multiple Use Deci-
sion: Biological Assessment for Formal Consultation Request (6 July
1998, on file, BLM Elko District Office). Further evidence of poor
riparian condition is found in archived BLM and Landsat images of
the watershed.
To evaluate the biological outcome of their conservation-
grazing program, SVR cooperated with the US Department of
Agriculture, Agricultural Research Service in sequential aerial sur-
veys of the Rock Creek watershed. Aerial photography can be a
cost-effective means for collecting riparian data (Clemmer, 2001;
Manning et al., 2005; Marcus et al., 2003), but the value of such
assessments depends largely on the spatial resolution of the data
(Congalton et al., 2002; Davis et al., 2002; Johnson and Covich,
1997; Muller, 1997; Prichard et al., 1999). Low-altitude, 2-cm GSD
(a measure of digital-image resolution), intermittent-capture aerial
imagery allows superior riparian assessments at a cost less than half
that of ground-based methods (Booth et al., 2006a).
Given the reports of poor riparian condition in the Rock Creek
watershed cited above, we predicted (1) willow cover would be reli-
ably measured from low-altitude, 2-cm GSD, intermittent-capture
aerial imagery thereby allowing detection of willow-cover changes
over time, (2) that BLM archived photographs, and (3) the Land-
sat image record, would be consistent with cited documents for
pre-2003 conditions; therefore, that (4) the conservation effec-
tiveness of SVR’s reduced hot-season riparian grazing could be
objectively determined, and (5) that analysis of these three lines of
willow-abundance evidence would provide an objective test of the
biological benefit of the SVR conservation-grazing program. This
appears to be the first attempt to use willow measurements from
an aerial survey as a particular indicator of recovery in a degraded
stream system.
2. Methods
2.1. Study area
Aerial surveys were conducted in 2003, 2004, and 2006 over
the 330,000-ha Rock Creek watershed (41◦17N, 116◦23W) in the
Tuscarora Mountains of north-central Nevada, US (Fig. 1). The BLM
manages 66% of the watershed but 90% of riparian areas are owned
Fig. 1. Photo locations (·) by year for the major streams of the Rock Creek watershed
in north-central Nevada, shown against 100-m contour intervals. The black arrow
indicates the location of Fig. 3, Landsat TM series.
by the SVR. Watershed elevation is 1500–2400 m. Precipitation is
250–300 mm (SCAS, 2005). The riparian zones contain coyote and
yellow willow (Salix exigua (Nutt.) and Salix lutea (Nutt.), syn. rigida)
and are characterized by shallow, low-volume streams of which
Rock Creek is the primary drainage (plant nomenclature follows
NRCS, 2009). Rock Creek flow is highly variable with mean monthly
flows March through May between 2.9 and 4.2 m3s−1falling to
0.05 m3s−1in August (stream gauge 10324500 near Battle Moun-
tain, Nevada [USGS, 2006]). Sampled streams are 1500–2200 m in
elevation, and fall into eight distinct geomorphological valley bot-
tom types (unpublished 1995 report: Inventory and assessment of
riverine/riparian habitats-Rock Creek Basin, Nevada. White Horse
Associates, Logan, UT). We used White Horse Associates’ inventory
and definition of stream reaches and divided our aerial sampling
by reach.

514 D.T. Booth et al. / Ecological Indicators 18 (2012) 512–519
Table 1
Willow cover for nine streams by year. Cover data are m2m−1of stream in image ±S.D. and the relative increase factor over 2 or 3 years is in multiples of 2003 or 2004 cover.
Creek 2003 2004 2006 Relative increase
Willow cover nWillow cover nWillow cover n
Frazera2.9 ±3.1 42 3.9 ±4.5 30 7.3 ±6.0 36 2.5×
Lewis 9.5 ±4.8 30 9.9 ±6.3 21 11.0 ±6.8 46 1.2×
Middle Rocka0.9 ±1.1 28 0.2 ±0.5 36 1.0 ±2.0 51 1.1×
Nelson 7.2 ±4.3 21 8.4 ±6.7 12 16.8 ±8.8 36 2.3×
Upper Rocka18.2 ±17.0 28 10.2 ±10.1 25 10.2 ±8.0 87 −0.6×
Soldiera– – 1.2 ±1.7 21 1.7 ±2.1 26 1.4×
Toejam 3.8 ±4.0 71 6.4 ±6.4 54 8.7 ±8.8 81 2.3×
Trout – – 4.1 ±3.9 40 4.2 ±6.2 56 1.1×
Willowa1.8 ±3.2 64 3.6 ±5.8 40 5.2 ±6.9 146 2.9×
All 5.2 ±7.9 284 4.9 ±6.3 279 7.1 ±7.9 565 1.4×
aPortions of this stream were within a 2005 burn perimeter.
The stream reaches averaged 3.4 ±2.1% slope (mean ±std. dev.),
meaning channel elevation dropped 34 mkm−1, on average, across
the watershed. Isaak and Hubert (2000) reported this channel-slope
range was associated with the highest cutthroat trout (O. clarki)
density in southeastern Idaho. While 27% of Rock Creek watershed
reaches fall in the high-slope category of >4.3%, over 66% fall into the
medium slope category of 1.8–4.3%, indicating that the watershed
provides prime (potential) trout habitat.
From 2003 to the present, there has been no intentional hot-
season livestock grazing on riparian areas; however, instances
of 100–200 head of non-permitted, late-season riparian use are
known for the first 2 years. (The non-permitted use is 6–10% of
permitted grazing and began in late July 2004 and in September
2005.) Three major wildfires in 2005 and 2006 burned upland areas
surrounding 15 of the 27 stream reaches photographed, inflicting
varying degrees of damage to riparian vegetation (Table 1).
2.2. Ground images
The BLM Elko Field Office provided us with 743 landscape-
view photographs from 111 riparian photo points associated with 7
streams in the watershed. These included upstream, downstream,
and across-stream views. Image acquisition dates ranged from
1977 to 2009. Because the conservation grazing management pro-
gram was begun in 2003, we selected 92 image sets having at least
one pre-2003 image, a 2003 image, and 2 images from 2004 to 2009.
All image sets exceeded the 4-image minimum. We asked 5 people
to rate the photo-point sequences for change in willow abundance
between 2003 and earlier versus 2004 and later, using a rating
where 0 = no rating due to apparent movement of the photo-point
or other irregularity, 1 = large decrease, 2 = moderate decrease,
3 = no change, 4 = moderate increase, and 5 = large increase. We
then counted the number of ratings in each category. Among the
five observers were 3 college students between 20 and 30 years of
age, and 2 rangeland professionals over 40 years of age. Four were
females. None of the observers was a co-author on the paper or
associated with either the SVR or the BLM.
2.3. Aerial images
Color digital, 2-cm GSD images were acquired from a light sport
airplane (FAA, 2010) equipped with: (1) a navigation system; (2)
11- and 16-megapixel, single lens reflex digital cameras (RGB) fit-
ted with 100 mm f/2.0 and 840 mm f/5.6, lenses respectively; (3)
a laser altimeter; and (4) two laptop computers (Booth and Cox,
2006, 2008; Booth et al., 2006b,c). This is a sampling, not a mapping,
method. Flight plans were created using ArcView 3.3 and ArcMap
9.0 (ESRI, Redlands, CA). All images were captured with associated
time and location data (Booth et al., 2006c).
Segments of 11 streams (27 reaches), totaling 170 km, were
surveyed July 17–18, 2003, September 9–10, 2004 and September
12–13, 2006 (Fig. 1). Each stream was sampled at approximately
100-m intervals along a continuous length starting at the source
and ending at either the junction with a larger stream, or when the
stream exited the survey boundary. Because images were triggered
manually by the pilot and not pre-programmed, photographic over-
lap between years was coincidental. Target flight altitude AGL was
200 m in 2003 and 2004, and 250 m in 2006, a change made to
increase image FOV.
2.3.1. Cover measurements
We used SamplePoint (Booth et al., 2006c) to measure cover from
images of nine streams that represent the watershed’s geomorpho-
logical variation. We used 100 points per image with the following
cover-type categories: (1) non-riparian area, (2) water, (3) willow,
(4) riparian vegetation and (5) other. Vegetation color, indicating
higher moisture, defined the riparian area. Points outside the ripar-
ian area were classified as non-riparian. Only points falling inside
the riparian area were classified into the other four categories. This
method required subjective delineation of riparian boundaries, as
do all methods that measure riparian indicators. Cover percent-
ages were converted into actual area (m2), and then normalized by
dividing actual area by the stream length within the image (see
below) to allow inter-year comparison. Thus, cover is reported
in m2m−1stream. Stream length was used for normalization
because it shows higher annual consistency than riparian width
or area.
2.3.2. Repeat measurements
Random airplane movement and manual triggering make it
impossible to plan riparian aerial surveys at this resolution in a way
that will reliably capture the same piece of ground in repeat flights;
however, the acquisition of hundreds of images resulted in some
chance overlap. These partially overlapping images were used to
measure change directly for individual willow canopies using rocks
for accuracy calibration. Using paired photos in ImageMeasurement
(Booth et al., 2006b), we measured canopy diameter of individual
willow plants. Repeat samples are also valuable in tracking changes
in channel sinuosity and bank erosion.
2.3.3. Stream length
Stream length within each image used to measure willow cover
was measured on a line with a minimum of 20 segments placed
down the center of the bank full channel using ImageMeasurement.
Additionally, the distance from each sample to the upstream end
of the stream reach containing the sample was measured from
topographical maps (1:100,000-scale) in ArcMap 9.0.

D.T. Booth et al. / Ecological Indicators 18 (2012) 512–519 515
2.3.4. Statistics
A paired t-test was used to compare willow canopy on repeated
measures of 24 willow plants. Comparisons of willow cover among
years used unpaired samples because of the limited number of
repeated samples.
Spatial autocorrelation was assessed with Moran’s I z-scores
generated by ArcMap 9.2 (ESRI, Redlands, CA). Because most
data were spatially autocorrelated, we condensed the data into
means for stream reaches (personal communication: P. Chapman,
Colo. State Univ. Dept. Statistics, 2010). The condensed data had
insufficient sample size to calculate Moran’s I z-scores or fit an auto-
gregressive model to check for spatial autocorrelation, so we used
Proc Mixed in SAS (SAS v 9.1, SAS Institute, Nashville, TN, USA) to
test for a random creek effect using stream reaches to determine if
intra-stream values were more correlated than inter-stream values
where creek was tested as a random effect (P. Chapman, op cit.). We
omitted 2003 data in all cases from the random-effect test due to
limited data. Willow-cover data were square root transformed to
satisfy the equal variance assumption. The 2005 fire burned two
reaches of Upper Rock Creek, reducing willow cover for both and
resulting in loss of statistical independence between them. There-
fore, these data were averaged together and treated as one reach.
Other reaches burned, but they were not combined because we had
no evidence that these riparian areas sustained significant dam-
age. Annual change was tested using t-tests paired by years for
individual streams.
2.4. Landsat TM record
The assessment of riparian habitat over time from Landsat
imagery was accomplished by first delineating potential ripar-
ian areas (PRAs) for nine streams using Feature Analyst Software
(Visual Learning Systems, Missoula, Montana, US) with NAIP
images (1-m resolution, aerial, 2006, color infrared). The PRAs were
defined laterally by topographic limits and we used plant commu-
nity at the riparian-upland interface to deduce that topographic
limit. Thus, the PRAs were low-lying lands supporting riparian,
or remnant riparian, communities with associated upland vegeta-
tion and judged capable of supporting a larger riparian community
given a higher water table and (or) greater soil-water storage.
After the PRAs were defined on NAIP imagery, the image loca-
tions were used to ensure geographic precision of the area for
temporal analysis in sequential Landsat images. The riparian veg-
etation within the PRA for each stream reach was measured from
Landsat images by developing NDVI values and images for each
reach (Lyon et al., 1998). These were then used to determine the
percent of potential riparian area actually occupied by riparian
vegetation. This allowed the Landsat archive to be used to assess
the 1989–2003, and 2004–2009, trends in riparian condition (the
years 1995 through 1998 were not used due to cloud-cover issues
with scenes of interest). Differences in atmospheric conditions,
sun angle, and sensor calibration make it necessary to calibrate
multi-temporal imagery and we used COST and PIF normaliza-
tion to correct for these effects (Schott et al., 1988; Chavez, 1996;
Sant, 2005). The above two methods of radiometric correction were
applied to our Landsat scenes using publicly available tools (Utah
State University, 1999, 2011) to make the COST corrections, and PIF
normalization.
Between sensor differences and changes in intra-sensor cali-
bration was accounted for by applying the appropriate published
calibration values for the different Landsat TM sensors. The 1994
TM image was acquired on a clear day and is temporally in
the center of the dates represented by the TM image dataset.
Therefore, we selected it to be the master image. The rest of
the images were normalized to the master using PIF normaliza-
tion. Because there are few manmade features on our images,
we used salt flats for bright area-calibration and north facing
slopes (shadow) and water for dark area-calibration. We restricted
our scenes to early September to reduce the variation from
inter-annual precipitation. In this summer-dry climate, upland
herbaceous vegetation is senesced by September regardless of
the amount of preceding winter and spring precipitation. The
sharp contrast in herbaceous vegetation senesced on the upland
and green in the riparian zones facilitated an accurate delin-
eation of the riparian areas. Change over time of the percent
of the PRA occupied by riparian vegetation of the nine streams
was calculated from NDVI values and those values compared
across streams for the years 1989–2003 versus 2004–2009 using a
paired t-test.
2.5. Precipitation analysis
Precipitation data were downloaded for the five closest Snotel
stations (29–46 km away; NRCS, 2007), and monthly precipitation
was averaged across all stations for 2004, 2006 and for the 29-year
period from 1982 to 2010 (extent of data) by using the first 28 days
of each month to standardize the observations among months and
years.
3. Results
3.1. Ground images
The 5 observers rating the 92 image sets produced 460 total
ratings of which 4 were rated as showing moderate decrease, 103 as
showing no change, and 353 as showing either a moderate or large
increase in willow abundance for the period 2004–2009 relative to
2003 and earlier.
3.2. Aerial images
3.2.1. Data acquisition
The aerial surveys produced 723 useable images in 2003, 590 in
2004 and 959 in 2006. Mean image GSD was 1.8 cm in 2003, 2.0cm
in 2004 and 2.3 cm in 2006. Twenty image FOVs from 2006 par-
tially overlapped 2004 image FOVs, and were used for direct repeat
measurements of willow canopy. ImageMeasurement analysis accu-
racy was confirmed by comparing measurements of unchanging
objects made from the 2004 to 2006 images. Absolute mea-
surement error (average of under-measures and over-measures
without regard to sign) was 7.2 cm, or 3.8% (n= 12), which agrees
with the <10% measurement error previously reported (Booth et al.,
2006b). Stream length captured by images averaged (mean ±std.
dev.) 72.5 ±23.1 m (n= 161) and 94 ±24.3 m (n= 436) in 2004 and
2006 respectively.
3.2.2. Willows
Willow cover (m2m−1stream) increased 3.1% between 2003
and 2006 (p= 0.02, n= 12 reaches) and 2.0% between 2004 and 2006
(p= 0.004, n= 20 reaches), but there was not a significant increase
between 2003 and 2004 (p= 0.18, n= 12 reaches; Table 1). Measure-
ment of 24 individual willow plants across 20 pairs of repeat sample
images (2004–2006) revealed an average increase in canopy size of
55 ±66.2% (Fig. 5A and B).
3.3. Landsat TM record
All nine streams for which change-over-time differences in
NDVI values were calculated had riparian-vegetation increases
within PRAs for 2004–2009 as compared to 1989–2003 (p< 0.001;
Figs. 2 and 3).

516 D.T. Booth et al. / Ecological Indicators 18 (2012) 512–519
Fig. 2. Change over time in the percent of the potential riparian area occupied
by riparian vegetation of nine streams and measured from Landsat images using
Normalized Difference Vegetation Index values and images as illustrated in Fig. 4
(p< 0.001 across all streams). Error bars indicate standard error of the mean.
3.4. Precipitation analysis and stream flow
Cumulative precipitation in 2006 was above the 29-year
average preceding, and throughout, the growing season (Fig. 4),
and is reflected in measured stream flow. Annual runoff in 2006
was 118.0 ×106m3compared to 100.0 ×106, 20.9 ×106, and
5.5 ×106m3in 2005, 2004, and 2003 respectively, and to an
average 36.9 ×106m3for 1918–2006 (Bonner et al., 2005; USGS,
2006). The September 2006 flow of 1.08 m3s−1was the greatest
recorded flow for any September up to that time (USGS, 2006).
There are 6 periods prior to 2003 where cumulative water-year
(October–September) precipitation exceeded the 29-year average
of 697 mm: 1980, 1982–1984, 1986, 1989, 1993, and 1995–1998;
cumulative annual precipitation for these years ranged from 729
to 1108 mm (NRCS, 2007).
Fig. 4. Average cumulative precipitation for the five closest weather stations to the
study site, all located within the same mountain range and within 50 km of the
study site, for water years 1989, 2004, 2006 and the 29-year period 1982–2010
(full data extent for the 5 stations). We used 28 days per month to standardize the
observations among months and years.
4. Discussion
Ground and space images give supporting evidence for the poor
pre-2003 condition of key streams as documented in Squaw Val-
ley Allotment Multiple Use Decision: Biological Assessment for Formal
Consultation Request (6 July 1998, on file, BLM Elko District Office).
The pre-2003 condition contrasts sharply with post-2003 condi-
tions where the relative increase in willow cover nearly tripled
on Willow Creek during the first 3 years of the new grazing plan
and more than doubled during the same period on three other
streams (Table 1). The trend of increasing willow cover is consistent
with reduced livestock grazing (Brookshire et al., 2002; Case and
Kauffman, 1997; Holland et al., 2005; Schulz and Leininger, 1990)
and is thought to result primarily from established-plant growth
under a reduced-herbivory regime (Fig. 5).
Why should changes in willow cover be ascribed to reduced live-
stock grazing and not to above-average precipitation and greater
stream flow? We propose four reasons: (1) we found no reports of
a similar willow-growth response from a season of increased pre-
cipitation, (2) such a growth response would be phenomenal, and
inconsistent with reported annual growth rates (see Brookshire
et al., 2002), (3) there are reports (cited above) documenting
Fig. 3. A September sequence of Landsat TM scenes showing a nearly static riparian condition between 1989 and 2003, but a large improvement by 2008. The 2008 image
is consistent with the improvements documented using BLM photo-point images and 2004 and 2006 aerial surveys after the 2003 change to reduced hot-season grazing on
riparian areas. The images are true color composites using a band 7-4-3 combination. (For interpretation of the references to color in this figure legend, the reader is referred
to the web version of this article.)

D.T. Booth et al. / Ecological Indicators 18 (2012) 512–519 517
Fig. 5. Images of a particular location on Middle Willow Creek photographed in 2004 (A) and 2006 (B) with white lines indicating measurements. The willow canopy diameter
in the upper left was 1.7 m in 2004, and 2.3m in 2006, and the willow canopy in the lower right increased from 1.7 m in 2004 to 3.0 m in 2006. The width of the open water
increased from 1.3 m to 3.4 m.
a willow-growth response to multi-year protection from graz-
ing that is consistent with our findings, and (4) willows along
the streams are accessing the water table so that growth is sup-
ported in both dry and wet years unless (a) the water table drops
beyond reach of the willow roots (Bilyeu et al., 2008)or(b)
rises significantly, putting roots in an anaerobic situation (Bourret
et al., 2005). Thus the water from a year–or even a multi-year
sequence–of above-average precipitation is unlikely to stimulate
above-average growth of established willows but it could reduce
willow growth if it produced a significant sustained elevation of the
water table.
There is evidence that above-average precipitation has the
potential to bring more nitrogen into a riparian system, particularly
in overland flow from rainfall events, and that the added nitrogen
will contribute to willow growth (Schade et al., 2002; Schade and
Welter, 2005; Welter et al., 2005). The time period covered by our
ground and Landsat images includes several pre-2003 wet peri-
ods (1989 for example), and images for those periods do not show
the same increases in willow growth that occurred after 2004 (e.g.,
Fig. 4). Given our data, we doubt that nitrogen added by wet years
made a difference in pre-2003 willow cover due to the ongoing hot-
season grazing. We suspect nitrogen added by wet years is a factor
in the post-2003 willow-cover increases.
The increases in willow cover herald structural changes,
including increased stream shading, bank armor, and chan-
nel obstructions–especially the development of beaver (Castor
canadensis [Kuhl]) dams–that are consistent with long-term
progress toward slowing runoff, increasing water retention, and
promoting perennial stream flow.
This study appears to be the first use of willow measure-
ments from an aerial survey as a particular indicator of recovery
in a degraded stream system. Our results demonstrate that 2-cm
imagery and associated image-analysis tools can detect, document,
and facilitate measurement of temporal change in this, and other,
key indicators – including changes due to weather, fire, and man-
agement. We demonstrate that the 2-cm imagery fills a critical
gap between conventional field methods and the lower-resolution
imaging methods cited by Clemmer (2001),Marcus et al. (2003),
and Manning et al. (2005). We also demonstrate the benefit of using
space, aerial, and ground images to address a question.

518 D.T. Booth et al. / Ecological Indicators 18 (2012) 512–519
Riparian systems are important in all parts of the world. The
aerial methods we used are applicable to any place that can be
safely flown by a light sport airplane at 100–300 m altitude AGL,
and that have one or more indicators that, like willow cover – can
be measured from acquired imagery.
5. Conclusions
Willow cover and canopy-size increases suggest an upward eco-
logical trend for surveyed streams of the Rock Creek watershed
after 2003. Ground photographs and the Landsat record provide a
historical context for the aerial-survey data and allow us to con-
clude that willow cover, measured from 2-cm GSD aerial imagery,
increased significantly during the first 3 years of the current SVR
grazing-management program. The findings imply that reduced
hot-season grazing will increase remnant willow populations along
desert streams in the western United States where stream incision
has not substantially lowered the water table. The improvement we
measured bodes well for the restoration of flow-regulating, water-
storing components of the hydrologic system and the concurrent
recovery of habitat for Lahonton cutthroat trout. We conclude
that the low-altitude, 2-cm GSD aerial surveys, and related image
analyses allowed us to detect and quantify these riparian-system
changes better than could be done with any other currently avail-
able monitoring method or combination of methods. Aerial surveys
like those used in this study make high sample density and sys-
tematic sampling (versus subjective selection) of watersheds and
other landscape-scale units practical, thereby addressing the large
Type II error (false negative) risk common to conventional land-
management survey efforts.
Role of the funding entities
Barrick and BLM had no role in study design; or in the collection,
analysis, or interpretation of data, or in writing the report except
that S.E. Cox entered on duty in his current position in 2011, and has
participated in post-submission revision. USDA-ARS approved sub-
mission of the manuscript upon successful pre-submission external
peer review.
Disclosures
Open Range Consulting is retained by Barrick to manage the SVR.
Acknowledgments
The research was funded in part by Barrick Gold Corporation
(Barrick), Elko, Nevada; the USDA-ARS; and the Wyoming State
Office of the BLM. Carol Evans, Fisheries Biologist for BLM, Elko
Field Office, provided valuable assistance. Barrick provided aviation
facilities and crew lodging. Joe Nance, CloudStreet Flying Services,
acquired aerial images. Carmen Kennedy, USDA-ARS, provided
technical assistance. Dr. Phil Chapman, Colorado State University
Department of Statistics, consulted with us on the statistical anal-
yses, and Larry Griffith, USDA-ARS assisted in the analysis. We
thank Drs. G. Perry and D. W. Bailey, also C. Evans and M. Bishop
for reviewing an earlier version of the manuscript. Throughout
this manuscript, mention of products and proprietary names is for
information only and does not constitute an endorsement by the
authors, Barrick, USDA, or BLM.
References
Bilyeu, D.M., Copper, D.J., Hobbs, N.T., 2008. Water tables constrain height recovery
of willow on Yellowstone’s northern range. Ecological Applications 18, 80–92.
Bonner, L.J., Evetts, D.M., Swartwood, J.R., Wilson, J.W., 2005. Water resources data,
Nevada, Water Year 2004. Water-Data Report NV-04-1, U.S. Department of the
Interior, U.S. Geological Survey. <http://pubs.usgs.gov/wdr/wdr-nv-04-1/NV-
04-1.pdf> (accessed 30.07.09.).
Booth, D.T., Cox, S.E., 2006. Very-large scale aerial photography for rangeland mon-
itoring. Geocarto International 21, 27–34.
Booth, D.T., Cox, S.E., Simonds, G., 2006a. Riparian monitoring using 2-cm GSD aerial
photography. Ecological Indicators 7, 636–648.
Booth, D.T., Cox, S.E., Berryman, R.D., 2006b. Precision measurements from very-
large scale aerial digital imagery using ImageMeasurement, Laserlog, and
Merge software applications. Environmental Monitoring and Assessment 112,
293–307.
Booth, D.T., Cox, S.E., Berryman, R.D., 2006c. Point sampling digital imagery using
‘SamplePoint’. Environmental Monitoring and Assessment 123, 97–108.
Booth, D.T., Cox, S.E., 2008. Image-based monitoring to measure ecological change.
Frontiers in Ecology and the Environment 6, 185–190.
Bourret, M.M., Brummer, J.E., Leininger, J.E., Heil, W.C.D.M., 2005. Effect of water
table on willows grown in amended mine tailing. Journal of Environmental
Quality 34, 782–792.
Brookshire, E.N.J., Kauffman, J.B., Lytjen, D., Otting, N., 2002. Cumulative effects
of wild ungulate and livestock herbivory on riparian willows. Oecologia 132,
559–566.
Case, R.L., Kauffman, J.B., 1997. Wild ungulate influences on the recovery of willows,
black cottonwood and thin-leaf alder following cessation of cattle grazing in
northeastern Oregon. Northwest Science 71, 115–126.
Chavez, P., 1996. Image-based atmospheric corrections-revisited and improved.
Photogrammetric Engineering and Remote Sensing 62, 1025–1036.
Clemmer, P., 2001. Riparian Area Management – The Use of Aerial Photography to
Manage Riparian-wetland Areas. US BLM Technical Reference 1737-10. United
States Department of the Interior, Denver, CO.
Congalton, R.G., Birch, K., Jones, R., Schriever, J., 2002. Evaluating remotely sensed
techniques for mapping riparian vegetation. Computers and Electronics in Agri-
culture 37, 113–126.
Davis, P.A., Staid, M.I., Plescia, J.B., Johnson, J.R., 2002. Evaluation of Airborne Image
Data for Mapping Riparian Vegetation within the Grand Canyon. Open File
Report 02-470. United States Geological Survey, Flagstaff, AZ.
FAA (Federal Aviation Administration), 2010. Section 6. Light-sport Category Aircraft
Airworthiness Certifications, in Order 8130.2F, Airworthiness Certification of
Aircraft and Related Products. U.S. Department of Transportation, pp. 112–125,
<http://www.faa.gov/aircraft/gen av/light sport/> (accessed 21.07.10.).
Harig, A.L., Fausch, K.D., 2002. Minimum habitat requirements for establishing
translocated cutthroat trout populations. Ecological Applications 12, 535–551.
Hebblewhite, M., White, C.A., Nietvelt, C.G., Mckenzie, J.A., Hurd, T.E., Fryxell, J.M.,
Bayley, S.E., Paquet, P.C., 2005. Human activity mediates a trophic cascade caused
by wolves. Ecology 86, 2135–2144.
Holland, K.A., Leininger, W.C., Trlica, M.J., 2005. Grazing history affects willow com-
munities in a montane riparian ecosystem. Rangeland Ecology and Management
58, 148–154.
Isaak, D.J., Hubert, W.A., 2000. Are trout populations effected by reach-scale stream
slope? Canadian Journal of Fisheries and Aquatic Sciences 57, 468–477.
Johnson, S.L., Covich, A.P., 1997. Scales of observation of riparian forests and dis-
tributions of suspended detritus in a prairie river. Freshwater Biology 37,
163–175.
Kauffman, J.B., Krueger, W.C., Vavra, M., 1983. Effects of late season cattle grazing
on riparian plant communities. Journal of Range Management 36, 685–691.
Kovalchik, B.L., Elmore, W., 1992. Effects of cattle grazing systems on willow-
dominated plant associations in central Oregon. In: Clary, W.P., McArthur, E.D.,
Bedunah, D., Wambolt, C.L., Wambolt (Eds.), Symposium on Ecology and Man-
agement of Riparian Shrub Communities. General Technical Report INT-289.
United States Forest Service, Odgen, UT, pp. 111–119.
Lyon, J.G., Yean, D., Luenetta, R.S., Elvidge, C.D., 1998. A change detection experiment
using vegetation indices. Photogrammetric Engineering and Remote Sensing 64,
143–150.
Manning, M., Bates, P., Brewer, K., Finco, M., Ruefenacht, B., Guay, T., 2005. Devel-
oping an Image-based Riparian Inventory Using A Multistage Sample: Phase I.
Final Report: Image-based Wyoming Riparian Inventory Project. United States
Forest Service, Remote Sensing Applications Center, Salt Lake City, UT.
Marcus, W.A., Legleiter, C.J., Aspinall, R.J., Boardman, J.W., Crabtree, R.L., 2003. High
spatial resolution hyperspectral mapping of in-stream habitats, depths, and
woody debris in mountain streams. Geomorphology 55, 363–380.
McInnis, L.A., McIver, J.D., 2009. Timing of cattle grazing alters impacts on stream
banks in an Oregon mountain watershed. Journal of Soil and Water Conservation
64, 394–399.
Muller, E., 1997. Mapping riparian vegetation along rivers: old concepts and new
methods. Aquatic Botany 58, 411–437.
NRCS (Natural Resources Conservation Service), 2007. Nevada SNOTEL Sites.
U.S. Department of Agriculture-NRCS, National Water and Climate Cen-
ter, <http://www.wcc.nrcs.usda.gov/snotel/Nevada/nevada.html> (accessed
06.04.09.) (Contact Snow Survey and Water Supply Forecasting Program,
Portland, OR, 97232, for paper copy of archived information.).
NRCS (Natural Resources Conservation Service), 2009. PLANTS Database. U.S.
Department of Agriculture-NRCS, <http://plants.usda.gov> (accessed 06.04.09.).
Prichard, D., Clemmer, P., Gorges, M., Meyer, G., Shumac, K., Wyman, S., Miller, M.,
1999. Riparian Area Management – Using Aerial Photographs to Assess Proper
Functioning Condition of Riparian-wetland Areas (Revised). Technical Reference
1737-12. United States Department of the Interior, Denver, CO.

D.T. Booth et al. / Ecological Indicators 18 (2012) 512–519 519
Sant, E.D., 2005. Identifying Temporal Trends in Treated Sagebrush Communities
Using Remotely Sensed Imagery. MS Thesis. Utah State University, Logan, UT.
SCAS (Spatial Climate Analysis Service), 2005. PRISM Data Explorer – Gridcell Time-
Series Analysis. Prism Climate Group, <http://www.prism.oregonstate.edu/>
(accessed 08.01.10.).
Schade, J.D., Welter, J.R., 2005. Hydrologic exchange and N uptake by riparian vegeta-
tion in an arid-land stream. Journal of the North American Benthological Society
24, 19–28.
Schade, J.D., Marti, E., Welter, J.R., Fisher, S.G., Grimm, N.B., 2002. Sources of nitrogen
to the riparian zone of a desert stream: implications for riparian vegetation and
nitrogen retention. Ecosystems 5, 68–79.
Schott, J.R., Salvaggio, C., Volchok, W.J., 1988. Radiometric scene normalization using
pseudinvariant features. Remote Sensing and Environment 26, 1–16.
Schulz, T.T., Leininger, W.C., 1990. Differences in riparian vegetation structure
between grazed areas and exclosures. Journal of Range Management 43,
295–299.
Scrimgeour, G.J., Kendall, S., 2003. Effects of livestock grazing on benthic inverte-
brates from a native grassland ecosystem. Freshwater Biology 48, 347–362.
Stoddart, L.A., Smith, A.D., 1955. Range Management, 2nd edition. McGraw-Hill Book
Company, Inc., New York.
U.S. Federal Register, 1975. Threatened status for three species of trout 40 FR 29863
29864, vol. 40, no. 137, Wednesday, July 16.
USFWS (U.S. Fish and Wildlife Service), 1995. Lahontan Cutthroat Trout,
Oncorhynchus clarki henshawi, Recovery Plan. USFWS (U.S. Fish and Wildlife
Service), Portland, OR.
USGS (U.S. Geological Survey), 2006. Water-resources Data for the United
States Water Year 2006. National Water Information System: Web Interface,
<http://wdr.water.usgs.gov> (accessed 30.07.09.).
Utah State University, 1999. Atmospheric Correction Extension,
<http://www.gis.usu.edu/∼doug/SERDP/tools/ATMOS/index.html> (accessed
30.07.11.).
Utah State University, 2011. Image Standardization: At-sensor Reflectance and COST
Correction, <http://earth.gis.usu.edu/imagestd/> (accessed 30.07.11.).
Vincent, K.R., Friedman, J.M., Griffin, E.R., 2009. Erosional consequence of saltcedar
control. Environmental Management 44, 218–227.
Welter, J.R., Fisher, S.G., Grimm, N.B., 2005. Nitrogen transport and retention in an
arid land watershed: influence of storm characteristics on terrestrial-aquatic
linkages. Biogeochemistry 76, 421–440.
White, S.M., Rahel, F.J., 2008. Complementation of habitats for Bonneville cutthroat
trout in watersheds influenced by beavers, livestock, and drought. Transactions
of the American Fisheries Society 137, 881–894.
Zoellick, B.W., 2004. Density and biomass of redband trout relative to stream shading
and temperature in southwestern, Idaho. Western North American Naturalist
64, 18–26.