Impacts of Vegetative Manipulations on Sediment Concentrations from Pinyon- Juniper Woodlands

advertisement
This file was created by scanning the printed publication.
Errors identified by the software have been corrected;
however, some errors may remain.
Impacts of Vegetative Manipulations on
Sediment Concentrations from PinyonJuniper Woodlands
Vicente L. Lopes
Peter F. Ffolliott
Malchus B. Baker, Jr.
Abstract-This paper reports on relationships between suspended
sediment concentrations and streamflow discharges from pinyonjuniper watersheds subjected to cabling treatments and applications of herbicides. These relationships are compared to relationships from a control watershed that represents untreated conditions
to provide a basis for describing the effects of pinyon-juniper
conversions on sedimentation processes. The effects are further
analyzed by separating the data sets on the basis of streamflow
generation mechanisms, that is, snowmelt-runoff events and highintensity, short-duration convectional rainfall events. Findings
from the study indicate that suspended sediment concentrations,
above a threshold discharge, increased as a result of the cabling
treatment, while no change in sediment concentrations was observed as a result of the herbicide treatment. These results improved on earlier evaluations of the impacts of conversion treatments in the pinyon-juniper woodlands on suspended sediment
discharges.
Pinyon-juniper woodlands occupy large areas in the Interior West ofthe United States. Management of these ecosystems has been controversial, however, because their hydrology has been a concern. Rates of erosion have accelerated in
recent years and, as a consequence, extensive areas have
unsatisfactory soil and water conditions (Gottfried and others 1995). This assessment has been largely based on evidence of surface and gully erosion, soil compaction, and
vegetative indices.
The decline in watershed condition has been attributed to
increases in overs tory tree densities, and corresponding
decreases in understory grasses, forbs, and half-shrubs,
which provide a protective cover on soil surfaces (Gottfried
and others 1995). Overstory trees have been removed from
many sites in the hope of remedying this situation by
encouraging the production of understory plants. Another
possibility is that the observed erosion is a residual of past
land-use practices, and causes of erosion are more complex.
The controversy over the hydrology and erosional dynamics in pinyon-juniper ecosystems has generated a number of
In: Monsen, Stephen B.; Stevens, Richard, comps. 1999. Proceedings:
ecology and management of pinyon-juniper communities within the Interior
West; 1997 September 15-18; Provo, UT. Proc. RMRS-P-9. Ogden, UT: U.S.
Department of Agriculture, Forest Service, Rocky Mountain Research
Station.
Vicente L. Lopes is Assistant Professor, and Peter F. Ffolliott is Professor,
School of Renewable Natural Resources, University of Arizona, Tucson, AZ
85721. Malchus B. Baker Jr. is Research Hydrologist, Rocky Mountain
Research Station, USDA Forest Service, Flagstaff, AZ 86001.
302
research studies (Gifford 1975, Wright and others 1976,
Baker 1982, Wilcox 1994, Lopes and others 1996). As part of
a general research effort to study the hydrologic and sediment-transport regimes of watersheds, the relative magnitudes of sediment exports are being analyzed in relation to
watershed condition, streamflow-generation mechanisms,
and land management activities. Sediment rating curves,
developed to describe relationships between the amount of
sediment in suspension and streamflow discharge, can be
used to estimate the effects of management activities on
suspended sediment (Brooks and others 1997, Lopes and
Ffolliott 1993)
This paper reports upon the development of sediment
rating curves for watersheds in pinyon-juniper woodlands of
the Interior West. It presents an analysis of suspended
sediment concentration -streamflow discharge rela tionshi ps
for two pinyon-juniper watersheds subjected to vegetative
conversion treatments, and a control watershed that represents untreated conditions. These comparisons provide a
basis for determining the effects of the conversion treatments on sedimentation processes. Differences between
snowmelt-runoff and convectional rainfall streamflowgeneration events were also considered.
Study Area _ _ _ _ _ _ _ __
The watersheds studied are located about 80 km south of
Flagstaff, Arizona, a tributary of the Verde River, in the
Colorado Plateau physiographic province. These watersheds, established by the USDA Forest Service to evaluate
the effects of vegetative conversion treatments on multiple
use values, supported stands of Utah juniper (Juniperus
osteosperma) and pinyon (Pinus edulis var. fallax). Descriptions of overstory tree species compositions and density
conditions have been presented by Clary and others (1974),
Ffolliott and Thorud (1974), and Baker (1982) and, therefore, will not be presented here.
The watersheds average 1,600 m in elevation. Springerville
very stony clay soils derived from basalt and cinder parent
materials predominate (Williams and Anderson 1967). The
clays are primarily montmorillonite which swell and shrink
during each wet and dry cycle (Baker 1984). Infiltration
rates for this soil series range between 2.0 and 6.4 cmlhr
(Baker 1982). Stream channels on the watersheds have a
southwesterly orientation. Annual precipitation average
450 mm, occurring largely in two seasons. The most important precipitation from a streamflow-generating standpoint
is that originating from the frontal storms of October
USDA Forest Service Proceedings RMRS-P-9. 1999
through April, when about 60 percent ofthe annual precipitation (both rain and snow) falls. A second precipitation
season is July through early September, when high-intensity, short-duration, localized convectional storms are common. Average winter water yields (22.9 mm) account for 85
percent of the total annual water yield (Baker 1982). Suspended sediment accounts for 75 to 80 percent of the total
annual sediment discharge from the watersheds studied.
Vegetative Conversion
Treatments ___________
A cabling treatment was applied to a 131-ha watershed
(WS 1) in 1963. Larger trees were uprooted by a heavy cable
pulled between two bulldozers. Smaller trees missed by
cabling were hand-chopped, slash- was burned, and the
watershed seeded with a mixture of forage species. The
treatment did not result in significant increases in annual
water yields. In a second conversion treatment, a mixture of
picloram (2.8 kg/ha) and 2,4-D (5.6 kg/ha) was applied by
helicopter to 114 ha of a 147 -ha watershed (WS 3) in 1968.
The remaining 35 ha were either not treated or the trees
were sprayed with a backpack mist-blower. The intent of
this treatment was to reduce transpiration losses by killing
trees, reduce evaporation losses from the soil by leaving the
dead trees standing, and reduce the amount of overland
water flow trapped in the pits created when trees were
uprooted by cabling (Baker 1984). The herbicide treatment
resulted in an increase in annual water yields of about 15.5
mm. The third watershed was a 51-ha control (WS 2) against
which the cabling and herbicide treatments were evaluated.
Conditions on this control watershed represented those
obtained through custodial management, that is, through
the use of minimal managerial inputs.
Procedures _____________
A total of 191 paired suspeilded sediment concentrationstreamflow discharge measurements (in excess of 0.05 m/s)
obtained from 1975 through 1982 (12 years after the cabling
treatment in one watershed and 7 years following the herbicide treatment in the other) were used in deriving sediment
rating curves. Either grab samples or integrated samples
obtained with a DH-48 were analyzed by filtration to determine suspended sediment concentrations. Streamflow was
measured in concrete trapezoidal flumes (Baker 1986). When
a sample of suspended sediment was collected, the time was
indicated on a digital tape on continuous water-level recorders at the gauging stations. The sediment data used in this
analysis were collected at time intervals greater than 1 hr to
avoid the possibility of correlation among the paired data
sets.
Two types of events served as the basis for studying the
effects of streamflow-generation mechanisms on suspended
sediment concentrations:
• Type 1. Snowmelt-runoff events not proceeded by precipitation; relatively slow response time to peak
streamflow discharge; streamflow duration of several
days or weeks; occurs in late winter-early spring.
USDA Forest Service Proceedings RMRS-P-9. 1999
• Type 2. High-intensity, short-duration, mostly convectional rainfall events; rapid response time to peak
streamflow discharge; streamflow duration of hours or
a few days; occurs in late summer-early fall.
Rain-on-snow events, which represented less than 10 percent of the streamflow-generation mechanisms on the watersheds studied, were excluded from the analysis.
Sediment rating curves most frequently take a power
function form, such as:
(1)
where C = suspended sediment concentration (mglL), Q =
streamflow discharge (m 3/s), and a, b = constants for a
particular stream. Sediment rating curves to be derived as
a power function are commonly approximated by leastsquare linear regressions of logarithmic-transformed data
(Walling 1977). Therefore, this transformation was used to
develop the sediment rating curves in this study. The coefficient of determination, R2, was used to compare the goodness-of-fit of the sediment rating curves. To correct for the
dependence of goodness-of-fit on degrees of freedom, an
2
adjusted coefficient of determination, Ra , was used:
Ra 2 = R2 - [P(l - R2)/(N - P - 1)]
(2)
where N = number of observations, and P
independent variables = 1.
= number
of
Results and Discussion
A summary of statistics for suspended sediment concentration and streamflow discharge for the three watersheds
are shown in table 1. Analysis of suspended sediment concentrations began with the complete data sets from each of
the watersheds studied. These data sets were subsequently
partitioned into the type of streamflow-generation mechanisms (snowmelt-runoff or convectional rainfall). The "Chow
test," details of which are presented by Kmenta (1986), was
performed in this second step to test the null hypothesis that
the parameters of the curves (a and b) had not changed
significantly at the 95 percent level of significance.
Measurements made during periods of high-streamflow
discharges (Type 2 events) were assigned the same weight as
meas uremen ts made during low -streamflow discharges (Type
1 events) in deriving sediment rating curves for each watershed. The parameters "a" and "b" of the relationships are
presented in table 2 with the 95 percent confidence limits,
fitted standard errors, coefficients of determination, and F
statistic.
There was a difference in the "a" and "b" val ues between
the sediment rating curve for the cabled watershed (WS 1)
and that for the untreated control watershed (WS 2) at
streamflow discharges greater than 0.15 m 3/s. This difference indicates that there are higher suspended sediment
concentrations from the cabled watershed than from the
3
control for streamflow discharges greater than 0.15 m /s.
These higher concentrations are likely a reflection of the soil
disturbances caused by uprooting trees in the cabling treatment (Clary and others 1974). However, there was not a
statistical difference for sediment rating curves derived
for the watershed treated with herbicides (WS 3), which
303
Table 1-Summary of statistics for suspended sediment concentration and streamflow discharge.
WS 1
Sediment Concentration (mg/L)
StDv
Min
Max
3
Streamflow (m /s)
StDv
Min
Event type 2
n4
Mean
ALL
117
104
12
8.88
8.97
5.50
13.33
13.64
5.38
0.39
0.38
1.17
73.37
73.07
15.99
0.102
0.101
0.104
0.056
0.056
0.030
0.058
0.058
0.058
0.261
0.261
0.127
32
21
10
7.87
7.74
7.25
6.94
8.21
2.59
2.10
2.10
4.29
40.17
40.17
12.01
0.093
0.096
0.073
0.035
0.030
0.011
0.058
0.060
0.058
0.216
0.216
0.084
42
41
10.55
10.18
13.69
13.64
0.74
0.74
80.17
80.17
0.088
0.087
0.028
0.027
0.059
0.058
0.151
0.151
2
2
2
2
ALL
1
2
3
3
ALL
1
3
=
=
=
=
Mean
Max
1WS 1 cabled watershed, WS 2 control watershed, and WS 3 herbicide watershed.
2ALL complete data set for each watershed; 1 = snowmelt-runoff events, 2 convenctional rainfall events.
3Event type 2 is not represented in the data set for WS 3.
4Sample (n) for the streamflow-generation mechanisms do not add up to the n for the complete data set for each watershed studied because of the inclusion of 1 frontal
rainfall event occurring in late-fall in the complete "data sets.
=
Table 2-Sediment rating curve parameters with the 95 percent confidence limits, standard errors, coefficients of determination, and F
statistics.
Event
type 2
n
4
a
95 Percent
Confidence
limits 5
b
95 Percent
Confidence
limits 5
Fitted
Error
R2
F
ALL
1
2
117
104
12
639.04
697.71
356.71
421.70 - 968.28
448.95 - 1084.17
55.41 - 2296.24
2.10
2.13
1.98
1.77- 2.44
1.77 - 2.48
0.54 - 3.43
0.23
0.23
0.29
0.58
0.57
0.43
158.32
140.17
9.42
2
2
2
ALL
1
2
32
21
10
107.26
196.01
616.84
53.63 - 214.53
69.34 - 554.09
220.64 - 1724.20
1.17
1.48
1.71
0.58 - 1.76
0.63 - 2.34
0.63 - 2.79
0.18
0.22
0.05
0.33
0.38
0.58
16.36
13.15
13.41
3
3
ALL 3
1
42
41
435.75
344.52
135.94 - 1397.76
103.53 - 1146.69
1.74
1.65
0.70 - 2.27
0.57 - 2.72
0.38
0.39
0.21
0.18
11.58
9.59
=
=
=
1WS 1 cabled watershed, WS 2 = control watershed, and WS 3 herbicide watershed.
2ALL complete data set for each watershed; 1 snowmelt-runoff events, 2 convenctional rainfall events.
3Event type 2 is not represented in the data set for WS 3.
4Sample (n) for the streamflow-generation mechanisms do not add up to the n for the complete data set for each watershed studied because of the inclusion of 1 frontal
rainfall event occurring in late-fall in the complete data sets.
5Confidence limits have been retranstormed to original units for interpretation.
=
experienced little soil disturbances as a result oftreatment,
and the untreated watersheds.
Sediment rating curves were developed to represent suspended sediment responses to the type of streamflow-generation mechanisms. There were no consistent differences in
sediment rating curves when the data sets used to derived
the equations were partitioned according to streamflowgeneration mechanisms. This finding differs from that reported for a higher-elevation ponderosa pine (Pinus ponderosa) watershed on Beaver Creek, where a significant increase
in suspended sediment concentration occurred due to
streamflow-generation mechanisms (Lopes and Ffolliott
1993, Dong 1996).
Management Implications
There was a significant difference in the suspended sediment concentrations when the cabled and control watersheds are compared. Soil disturbance caused by the uprooting of trees on WS 1 is the likely reason for this change.
304
=
Earlier studies on Beaver Creek had indicated that uprooting oftrees by cabling can increase the potential for soil loss
by overland water flow (Skau 1960, 1961). Sedimentation
parameters can also change, in general, because ofincreased
streamflows, although this did not occur after this cabling
treatment (Clary and others 1974). The difference in results
can be attributed to the fact that finer sediments (silt and
clay) require less energy to remain in suspension and, as a
consequence, can be transported by smaller water flows
than would be needed to more larger particles.
This result, and those of Clary and others (1974) and
Baker (1982), are different than might be expected from the
hypothesis that removal of overstory trees to increase production of the herbaceous plants forming a protective cover
will cause less soil erosion. No changes in suspended sediment concentration occurred on WS 3 when herbicides were
applied in a conversion treatment. The soil surface on this
watershed was not disturbed by the treatment, since most of
the herbicide was applied by helicopter.
Large-scale vegetative conversions of pinyon -juniper woodlands have been less frequent in the Interior West than in
USDA Forest Service Proceedings RMRS-P-9. 1999
recent years, due largely to the increased environmental
concerns and the increasing emphasis being placed on more
holistic management ofthe pinyon -j uniper ecosystems (Shaw
and others 1995). Little changes in suspended sediment
concentration because of conversion treatments, therefore,
are likely to occur in the near future.
Conclusions --------------------------------Sediment rating curves have been developed for three
pinyon-juniper watersheds in northern Arizona that had
been subjected to treatments ranging from custodial management to vegetative conversions by cabling and application of herbicides. The cabling treatment resulted in increased suspended sediments at specified streamflow
discharges because of the soil disturbances caused by the
uprooting of trees, while the herbicide treatment did not
result in a change. This finding differs from previous studies
(Clary and others 1974, Baker 1982), and could be helpful to
managers in responding to questions about the impacts of
vegetative manipulation on sediment concentrations from
pinyon-juniper woodlands with similar soils and precipitation regimes throughout the Interior West.
References ----------------------------------Baker, M. B., Jr. 1982. Hydrologic regimes offorested areas in the
Beaver Creek watershed. Gen. Tech. Rep. RM-90. Fort Collins,
CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station.
Baker, M. B., Jr. 1984. Changes in streamflow in a herbicide-treated
pinyon-juniper watershed in Arizona. Water Resources Research.
20:1639-1642.
Baker, M. B., Jr. 1986. A supercritical flume for measuring sediment-laden streamflow. Water Resources Bulletin. 20:847-851.
Brooks, K. N.; Ffolliott, P. F.; Gregersen, H. M.; DeBano, L. F. 1997.
Hydrology and the management of watersheds. Ames, IA: Iowa
State University Press.
Clary, W. P.; Baker, M. B., Jr.; O'Connell, P. F.; Johnsen, T. N., Jr.;
Campbell, R. F. 1974. Effects of pinyon-juniper removal on
natural resource products and uses. Res. Pap. RM-128. Fort
Collins, CO: U.S. Department of Agriculture, Forest Service,
Rocky Mountain Forest and Range Experiment Station.
USDA Forest Service Proceedings RMRS-P-9. 1999
Dong, C. 1996. Effects of vegetative manipulations on sediment
concentrations in north-central Arizona. MS Thesis. Tucson, AZ:
University of Arizona.
Ffolliott, P. F.; Thorud, D. B. 1977. Water yield improvement by
vegetation management. Water Resources Bulletin. 13:563-571.
Gifford, G. F. 1975. Approximate annual water budgets of two
chained pinyon-juniper sites. Journal of Range Management.
28:73-74.
Gottfried, G. J.; Swetnam, T. W.; Allen, C. D.; Betancourt, J. L.;
Chung-MacCoubrey, A. L. 1995. Pinyon-juniper woodlands. In:
Finch, D. M.; Trainer, J. A, tech. coords. Ecology, diversity, and
sustainability of the middle Rio Grande Basin. Gen. Tech. Rep.
RM-268. Fort Collins, AZ: U.S. Department of Agriculture, Forest
Service, Rocky Mountain Forest and Range Experiment Station.
95-132.
Kmenta, J. 1986. Elements of econometrics. New York, NY:
Macmillian Publishing Company.
Lopes, V.L.; Ffolliott, P. F. 1993. Sediment rating curves for a
clearcut ponderosa pine watershed in northern Arizona. Water
Resources Bulletin. 29:369-382.
Lopes, V. L.; Ffolliott, P. F.; Gottfried, G. J.; Baker, M. B., Jr. 1996.
Sediment rating curves for pinyon juniper watersheds in northern Arizona. Hydrology and Water Resources in Arizona and the
Southwest. 26:29-33.
Shaw, D. W.; Aldon, E. F.; LoSapio, C., tech. coords. 1995. Desired
future conditions for pinon-juniper ecosystems: Proceedings of
the symposium. Gen. Tech. Rep. RM-258. Fort Collins, CO: U.S.
Department of Agriculture, Forest Service, Rocky Mountain
Forest and Range Experiment Station.
Skau, C. M. 1960. Some hydrologic characteristics in the Utah
juniper type of northern Arizona. PhD Dissertation. East Lansing, MI: Michigan State University.
Skau, C. M. 1961. Some hydrologic influences of cabling juniper.
Res. Note 62. Fort Collins, CO: U.S. Department of Agriculture,
Forest Service, Rocky Mountain Forest and Range Experiment
Station.
Walling, D. E. 1977. Assessing the accuracy of suspended sediment
rating curves for a small basin. Water Resources Research.
13:531-538.
Wilcox, B. P. 1994. Runoff and erosion in intercanopy zones of
pinyon-juniper woodlands. Journal of Range Management. 47:
285-295.
Williams, J. A; Anderson, T. C. 1967. Soil survey on Beaver Creek
area, Arizona. Washington, DC: U.S. Department of Agriculture,
Forest Service and Soil Conservation Service, in cooperation with
the Arizona Agricultural Experiment Station.
Wright, H. A; Churchill, F. M.; Stevens, W. C. 1976. Effects of
prescribed burning on sediment, water yield, and water quality
from dozed juniper lands in Texas. Journal of Range Management. 29:294-298.
305
Download