Monitoring Hydrological Changes Related

to Western Juniper Removal: A Paired

Watershed Approach
T.L. Deboodt, M.P. Fisher, J.C. Buckhouse, John Swanson
Abstract 225 percent (alpha > 0.05), increased days of
recorded groundwater by an average of 41 days
Since 1934, western juniper has increased its hold on (alpha > 0.05), and increased the relative availability
eastern Oregon rangelands. U.S. Forest Service of late season soil moisture at soil depths of .76 m
reports that juniper acreage has increased from 1.5 (27 in) (alpha > 0.1). Ephemeral channel flow did
million acres to over 6 million acres in 1999. not show a predictable trend following 2 yrs of post
Previous studies have shown that water use by treatment measurements. The Camp Creek project
juniper can exceed 30 gallons per day when illustrated that, for this system, managing vegetation
adequate soil moisture is present. Increased juniper for water yield may be obtainable at a much lower
dominance has been implicated in the desertification precipitation threshold than what was previously
of Oregon’s rangelands. Groundwater mitigation, understood.
reintroduction of steelhead into the upper Deschutes
River basin, and changes in laws affecting surface Keywords: paired watershed, water yield,
water right allocations have driven public policy to western juniper, range restoration
look at how water is currently being used and how
changes in water use (water law) could affect water Introduction
availability. Vegetative modeling has shown that 9
to 35 trees per acre are enough to utilize all the According to U.S. Forest Service publication PNW-
precipitation delivered to a site in a 13-in annual GTR-464, “Western Juniper in Eastern Oregon,”
precipitation zone. Earlier studies suggest that a western juniper’s dominance in eastern Oregon has
minimum of 17 in of annual precipitation is required increased 5-fold since 1934 (420,000 acres to
to measure a water yield response associated with 2,200,000 acres) (Gedney et al. 1999). The result of
vegetative manipulation. In 1993, the Camp Creek this significant shift in plant community dominance
Watershed study area was established to monitor the has been reduced forage production, increased soil
effects of juniper removal on hydrologic processes. erosion, and reduced infiltration rates. Based on
In 2005, following 12 yrs of pretreatment individual tree water use models and field
monitoring in the 2 watersheds (Mays and Jensen) observations, it has been speculated that the
all post-European aged juniper (juniper <140 years expansion of western juniper has been, at least in
of age) were cut from the treatment watershed part, responsible for the desertification of these
(Mays). Analysis indicated that juniper reduction landscapes. Based on water use models for
significantly increased late season spring flow by individual trees, the U.S. Forest Service estimates
that mature western juniper tree densities, ranging
Deboodt is an associate professor and extension agent, from 9 to 35 trees per acre, are capable of utilizing
Oregon State University Extension Service, Crook all of the available soil moisture on a given site.
County, Prineville, OR 97754. Fisher is associate Research has shown that soil loss from sites with
professor and department head, Forestry Program, Central higher than the natural variation of western juniper
Oregon Community College, Bend, OR 97701. cover is an order of magnitude greater than similar
Buckhouse is professor and extension watershed
sites that are still within their natural range of
specialist, Oregon State University, Department of
Rangeland Ecology and Management, Corvallis, OR
variation (Buckhouse and Gaither 1982).
97331. Swanson is a rangeland specialist, Bureau of
Land Management, Prineville District, Prineville, OR Established in 1993, the Camp Creek Watershed
97754. Study Area was created to monitor water quantity

The Third Interagency Conference on Research in the Watersheds, 8-11 September 2008, Estes Park, CO 227
and timing associated with juniper control. Channel relative to juniper control. Access to the site is from
morphology, hill slope erosion, and changes in the Camp Creek/Bear Creek road.
vegetation were also monitored. The project
involved the use of a paired watershed study format. Methods
The paired watershed project is located
approximately 60 mi southeast of Prineville, OR. Establishment of the study and initiation of
monitoring began in 1993. Each watershed was
Two watersheds (Mays and Jensen) were identified delineated by the location of a continuous recording
in the Camp Creek Drainage, a tributary of the flume placed in the channel at the lowest point of
Crooked River. The project consisted of the each watershed. Flow was measured and recorded
treatment (cutting juniper) of one of the paired with the aid of a data logger. Precipitation inputs
watersheds totaling approximately 250 acres with were first measured with the use of a Belfort
the other watershed serving as the untreated control. Universal Rain Gauge, and a weather station was
The U.S. Bureau of Land Management (BLM) added to each watershed in 2004 to record air
Prineville District cut approximately 200 acres of temperature, precipitation, wind speed and direction,
western juniper in Mays watershed. The cutting was solar radiation, leaf wetness, relative humidity, and
initiated in October 2005 and was completed in snow accumulation.
April 2006.
In 2004, additional monitoring was added to the
The elevation of the project area ranged from 4,500 watersheds (Figure 1). Within each watershed, a
to 5,000 ft with an average annual precipitation of 13 spring was improved and flow measured. Six
in. The historic vegetation type was mountain big shallow wells were placed across the valley bottoms
sagebrush/Idaho fescue. The site is currently of each watershed near the flume location for the
dominated by western juniper with a sparse purpose of measuring depth of groundwater. Soil
understory of shallow rooted perennial grasses and moisture and soil temperature probes were installed
forbs. Since 1993, the two watersheds have been at 2 locations within each watershed and placed at
monitored for similarities and differences. multiple soil depths.

Project objectives

• Evaluate hydrologic changes following the

cutting of post-European aged juniper (trees
established since mid-1800s).
• Evaluate changes in hill slope erosion and
channel morphology following the cutting of
post-European aged juniper.
• Evaluate changes in plant community
composition following the cutting of post-
European aged juniper.

The majority of the two watersheds consisted of

public land, administered by the BLM Prineville
District (75 percent Mays, 86 percent Jensen). The
remaining portions of each watershed were owned
by the Hatfield High Desert Ranch. The BLM—in
cooperation with the Crook County Soil and Water Figure 1. Location of monitoring stations.
Conservation District (SWCD), the permittee
(Hatfields), and the Oregon State University (OSU) All monitoring of weather, spring flow, channel
Department of Rangeland Ecology and flow, soil moisture, and depth to water was done
Management—identified the paired watersheds as an through satellite uplinks; data is available for
area of interest because of the opportunities the viewing on the website http://ifpnet.com.
study provided to monitor changes in water yields

228 The Third Interagency Conference on Research in the Watersheds, 8-11 September 2008, Estes Park, CO
Results be an indicator of watershed function (Table 2).
Increases in length would indicate an improved
Spring flow hydrologic condition. A review of the data (t-test)
indicates that changes in the average number of days
Figure 2 illustrates the differences in output between in which water was recorded in the wells increased
the two springs and the differences between years. in Mays as a result of cutting the trees (p-value =
Spring flow is dependent on timing, type, and 0.0152). Using a Wilcoxon rank test the wells in
amount of precipitation. Base flow, the flow which Mays post-treatment, recorded a greater increase in
is least likely to be influenced by a recent the number of days that water was recorded when
precipitation event or snowmelt period, is late season compared to the control watershed, Jensen
flow. Late season flow is defined as the period (p-value = 0.013).
between July and November. The first two sets of
Table 2. Comparison of average number of days of
bars represent the pretreatment years (2004–2005)
well water for the watersheds. Pre- and post-
and the last 3 sets of bars show the changes in flow
treatment years consist of 2 yrs each.
after treatment (October 2005).
Watershed Well Pre-treat Post-treat Diff.
14 Mays (treated)
Jensen (control)
Mays 1 112.5 128.5 16
12
2 119.5 135 15.5
3 195.5 285 89.5
Gallons per Minute

10

8
4 195.5 209 13.5
5 156 197 41
6

6 269.5 342.5 73
4

0
9/1/03 1/1/04 5/1/04 9/1/04 1/1/05 5/1/05 9/1/05 1/1/06 5/1/06
Date
9/1/06 1/1/07 5/1/07 9/1/07 1/1/08 5/1/08 9/1/08
Jensen 1 70 82 12
Figure 2. Differences in late season spring flow 2 78.5 89 10.5
before and after treatment. 3 283.5 296 12.5
4 314.5 361.5 47
Table 1 shows the t-test results for comparisons of 5 283.5 296 12.5
late season flow (lowest flow recorded) between the 6 167.5 141 -26.5
two watersheds and the years before (2004–2005)
and after (2006–2007) treatment. The one tailed P- Soil moisture
value is significant at alpha = 0.05**.
Observing the lowest readings of the year within
each watershed illustrated the amount of “water
Table 1. T-Test for spring flow data, lowest flow
savings” that was carried over from one year to the
recorded (GPM).
next (Figure 3). Evaluating the change in “water
Watershed savings” over years helps to see if that change was
Year Diff. Mean Variance associated only with precipitation, or if increases
Mays Jensen
2004 1.87 0.20 1.67 might have been due to the lack of deep-rooted
2005 1.90 0.13 1.77 1.720 0.00500 vegetation (the cutting of the juniper). If it was due
2006 4.80 0.23 4.57 to the removal of deep-rooted vegetation, then
2007 3.6 0.00 3.60 4.085 0.47045 excess soil moisture could move through the soil
Difference 2.365 profile and into sub-surface water storage and flow.
Standard error 0.487
t-test 4.8505805 Individual probe readings were averaged by location
One tailed P-value 0.019** within the soil profile and by site for each watershed.
ANOVA (analysis of variance) showed that the
observed increase was significant (alpha = 0.1*) for
Wells
the difference between 2006 and 2005 and for the
average increase difference of 2006–07 combined
Well data, in addition to depth measured, provides
and 2005 when comparing Mays with Jensen. Table
insight to the timing or availability of subsurface
3 shows the results of this test for the combined
water. The length of groundwater availability could
years 2006–07 compared to 2005.

The Third Interagency Conference on Research in the Watersheds, 8-11 September 2008, Estes Park, CO 229
 Mays Watershed
Soil Moisture - Lowest reading of the year
longer than Jensen. In 1998, Jensen flowed for more
50
Bottom Profile days when compared to Mays. In 2007, while length
45
2005-2006
of flow was greater in Mays, Jensen’s flow as
40
2006-2007
2007-2008 measured in accumulated cubic feet per second was
greater than Mays’ flow.
35
Percent Moisture

30

25

20
Of special note in the observation of these systems
was the winter of 2006, following the cutting of
15

10

5 juniper in Mays. The snow pack, which began its

0
0 1 2 3 4 5 6 7 accumulation in December 2005 was static at
Lower Site
Probe
Upper Site Average
approximately 16 in. December and early January
rain events saturated the snow pack. While no water
Jensen Watershed
Soil Moisture - Lowest reading of the year content measurements were taken, field notes
Bottom Profile
50 indicate that the snow was saturated and frozen on
45

40
2005-2006
2006-2007
top. Field notes also indicated that the snow pack
35
2007-2008
was solid enough for researchers to be able to walk
Percent Moisture

30 on top of the snow without breaking through. As

mentioned earlier, soil temperatures during this
25

20

15 period did not drop below 32°F for either watershed.

10

5
Channel flow in Mays began on 7 January 2006.
0 Flow was recorded through mid-June 2006. In
0 1 2 3
Probe
4 5 6 7
contrast, flow in Jensen did not begin until 1 April
Lower Site Upper Site Average
2006 and ceased to flow by early May. During this
period, all observations for both watersheds
Figure 3. Example of changes in soil moisture. 1
indicated that flow was generated exclusively from
year pre-treatment and 2 years post treatment.
bank seepage and that no evidence of overland flow
was observed for either watershed.
Table 3. Significance of end of year soil moisture
accumulation post- vs. pre-treatment.
In contrast, during the winter of 2007, very little
snow pack was accumulated. Bare ground was
Year Probe (Location) P-value
observed in both watersheds (50–70 percent of the
2006–07 vs. 05 Bottom (0.27 in) 0.1002*
landscape) with snow accumulation areas measuring
2006–07 vs. 05 Middle (0.18 in) 0.1796
less than 6 in. Soil temperatures in early February
2006–07 vs. 05 Top (0.7 in) 0.6132
were approximately 22°F. An early February storm
produced a rain on snow event. Flow was recorded
Channel flow in both watersheds and evidence was observed
which indicated the majority of channel flow
Channel flow in the two watersheds is ephemeral. originated as overland flow. Sediment movement
These channels only have flow during periods of was observed on the hill slopes and in the channels.
snowmelt and extreme summer thunderstorm Sediment deposits had to be removed from both
activity. Ephemeral channels tend to be more flumes. These two different observations help to
influent in relation to the groundwater than perennial illustrate the high variability within these systems
flows contributing to groundwater rather than and the difficulty in connecting channel flow data to
groundwater contributing to channel flow. treatment effects, especially during the first two
years following treatment.
Ephemeral channel flows or days of flow did not
show a relationship to the treatment. Recorded
channel flow occurs during the spring and early
Management Implications
summer months and is usually associated with the
A healthy, functioning watershed is one that
snow melt period. In 1996 and 2004 total annual
captures, stores and safely releases the precipitation
days of flow were greater than days of springtime
that is delivered to the site. Land management
channel flow, a result of late summer thunderstorms
decisions should include looking for ways to
and early fall rain. In all years but one, Mays flowed

230 The Third Interagency Conference on Research in the Watersheds, 8-11 September 2008, Estes Park, CO
increase opportunities for precipitation to infiltrate offset this limitation. Changes in groundwater may
into the soil profile (vegetation management), have downstream effects, delaying the time it takes
moving excess moisture into sub-surface storage and water move through the system and by adding to
groundwater, slowly releasing that water to channel or perennial stream flow downslope.
minimize the risk of soil loss and channel bank and
bed instability (Fisher et al. 2008). Hibbert (1983) By combining the upland and riparian benefits of
and others have suggested that there would be no juniper removal, the system will begin to move
water yield increase as a result of vegetation toward a watershed that is functional in its ability to
manipulation (juniper cutting) in precipitation zones capture, store, and safely release water while
where annual precipitation was less than 4,300 mm providing a site that is productive and capable of
(17 in). Any change to the water budget would only being managed for sustainable use.
yield an increase in soil moisture, improving
herbaceous vegetative production. Acknowledgments

The 30-yr average annual precipitation at Barnes This research project was initially supported through
Station (U.S. Geological Survey weather station) a grant from the U.S. Environmental Protection
located approximately 10 mi east of the study site is Agency with Dr. John Buckhouse (OSU), Mel
349 mm (13.75 in). Precipitation over the last 4 yrs George (University of California–Davis), and Sherm
on the study site has ranged from 278 mm (10.95 in, Swanson (University of Nevada–Reno) providing
80 percent of average) to 449 mm (17.68 in, 129 grant leadership. Additional funding was provided
percent of normal). Both the high and low by a Bureau of Land Management Science Grant
precipitation years occurred during the post- (thanks to Michelle McSwain, Prineville District)
treatment phase of the study. and the Oregon Watershed Enhancement Board.
Other financial support was provided by Crook
A review of the data collected over the course of the County Taylor Grazing Board, Crook County
last 13 yrs indicated that the cutting of post- Extension Service District, and the Secure Rural
European aged juniper has changed the water Schools and Community Self-Determination Act of
balance equation. Analysis of the first 2 yrs 2000: Title II. Authors appreciate the reviews of
following treatment has shown that spring flow, Tamzen Stringham and Tony Svejcar.
groundwater, and soil moisture have all increased
when compared to pre-treatment levels. References
Comparisons of ephemeral channel flows did not
show as clear a trend (data not presented here). Buckhouse, J.C., and R.E. Gaither. 1982. Potential
Ephemeral channels tend to be more influent in sediment production within vegetative communities
relation to the groundwater, contributing to in Oregon’s Blue Mountains. Journal of Soil and
groundwater rather than groundwater contributing to Water Conservation 37(2):120–122.
channel flow.
Fisher, M.P., T. Deboodt, and J. Buckhouse. 2008.
In the uplands, management implications suggest Channel geomorphologic changes and hillslope soil
that with juniper removal, herbaceous vegetation can movement following juniper treatment on Camp
create a more uniform groundcover across the Creek paired watershed study. In Building Bridges:
hillslope. Reduced bare ground results in increased Grasslands to Rangelands, Proceedings of the 2008
infiltration opportunity and decreased soil erosion. Joint Meeting of the Society for Range Management
Improved hydrologic function of the uplands can and the America Forage and Grass Council,
maintain site stability and fertility. Louisville, KY, 26–31 January 2008, Abstract no.
2402. Society for Range Management, Wheat Ridge,
Within the riparian area, management implications CO.
point to the opportunity to increase spring flow for
livestock, wildlife, and domestic use along with Gedney, D.R., D.L. Azuma, C.L. Bolsinger, and N.
some mitigation of water diversion. Late season low McKay. 1999. Western juniper in eastern Oregon.
flows limit land management alternatives. U.S. Department of Agriculture, Forest Service,
Increasing flows by cutting juniper could partially

The Third Interagency Conference on Research in the Watersheds, 8-11 September 2008, Estes Park, CO 231
Pacific Northwest Research Station, General
Technical Report PNW-GTR-464.

Hibbert, A.R. 1983. Water yield improvement

potential by vegetation management on western
rangelands. Water Resources Bulletin 19(3):375–
381.

Larsen, R.E. 1993. Interception and Water Holding

Capacity of Western Juniper. Oregon State
University. Ph.D. dissertation. Corvallis, OR. 172 p.

232 The Third Interagency Conference on Research in the Watersheds, 8-11 September 2008, Estes Park, CO