7853

• UCREFRP Catalog Number: 7853
• Author: Auble, G. T., J. M. Friedman and M. L. Scott
• Title: Relating Riparian Vegetation To Present And Future Streamflows
• USFW Year: 1994
• USFW - Doc Type: Ecological Applications
• Copyright Material: NO

August 1994

FLOW REGIME AND RIPARIAN VEGETATION

551

detailed autecological data than are available for many
riparian species.

t

Inundation duration as the
independent variable

Inundation duration, as determined by a flow du-
ration curve, is the single environmental variable de-
termining vegetation response in our model. The sim-
plicity of this approach has both advantages and
disadvantages. Many studies in humid regions (e.g.,
Wells 1928, Teversham and Slaymaker 1976, Teskey
and Hinckley 1977, Bedinger 1979, Klimas et al. 1981,
Harris et al. 1985, Hupp and Osterkamp 1985) have
demonstrated that plant species and communities can
be distinctly arrayed along a gradient of inundation
duration. This study extends that conclusion to a semi-
arid setting (Fig. 6), and we have obtained similar re-
sults in two as yet unpublished gradient analyses of
riparian vegetation in the Colorado High Plains. Fur-
ther studies are needed, particularly of herbaceous spe-
cies, to determine whether our conclusions can be gen-
eralized to semi-arid and arid regions as a whole.
In many of the studies done in humid regions, in-
undation duration has been interpreted as a predictor
of the degree of anoxia to which roots are exposed.
However, in this western riparian application, inun-
dation of most of the riparian zone is short-lived and
anoxia is less important. Inundation duration succeeds
here as a predictor of vegetation distribution because
it is correlated with flow-related variation in a suite of
environmental variables in the riparian zone. These
include shear stress, sediment deposition and erosion,
soil moisture, and depth to groundwater, in addition
to soil oxygen concentration. For example, sites with
a high inundation duration are likely to be closer to
groundwater when not flooded, are likely to be inun-
dated to greater depths when flooded, and are likely to
be subject to greater and more frequent shear stress
than sites with a low inundation duration.
Nonetheless, the relation of other environmental
variables to inundation duration is imperfect. While
the error involved in using a single surrogate measure

al 0.8
CDU
ECD
~~
15;jj 0.6
c....
go
~ ~ 0.4
U.lll
~
0-
CD 0.2

~ I
:\ I
.\
: I"
: \
". I "
.....1 '\
\ \
\. "
\
\
\
"
"
.......-. .. .

- Reference
.. ..... Diversion
-- Diversion. increased. minimum
- - - - Moving. average

0.0
o

SO 100
Discharge (m3/s)
FIG. 7. Flow duration curves for hydrologic alternatives.

TABLE 2. Percentage of bar area occupied by each cover type
under hydrologic alternatives.

Cover types
Hydrologic Open Eleo- Equi- Hetero-
alternative Water charis setum theca
Reference 0 37 53 10
Diversion 2 15 48 35
Diversion-Increased-
Minimum 13 6 46 35
Moving-Average 2 33 37 28

may be acceptable for comparisons within a site, it
does limit transferability of information between sites.
For example, the position of a species on an inunda-
tion-duration gradient determined at a reach with
groundwater recharge may be different from the po-
sition of the same species determined at a reach with
groundwater discharge.
Although we judge inundation duration to be the
best single independent variable for our purposes, no
single variable can capture all the important aspects of
water level change in a riparian area (Mitsch and Gos-
selink 1986). Alternative parameters may be superior
in some situations. Both Hupp and Osterkamp (1985)
and Bedinger (1979) arrayed riparian vegetation along
a composite variable that combined inundation du-
ration and recurrence interval. Recurrence interval can
be defined as the average time between inundations
for a particular point. The case of a bottomland down-
stream of a dam used for power peaking is an example
of a situation suited for use of recurrence interval. In
this case a point might be inundated for half of each
day. It is essential to be able to distinguish such a point
from another that is inundated continuously for 182 d
and then dry for the next 182 d. Finally, the clear
relationship between inundation duration and present
vegetation (Table I, Fig. 6) suggests, but does not prove,
that anticipated changes in inundation duration can be
successfully used to predict vegetation change. Verifi-
cation of model predictions will be necessary to test
the adequacy of our approach.

ISO

Channel change

Like most other methods used to predict changes in
stream biota following flow alteration, the direct gra-
dient method does not explicitly represent processes
of channel change. This is not a problem if channel
morphology is constant. However, if channel geometry
changes, the calculated inundation durations of plots
are likely to be in error. If bars move, but the areas of
different surface types do not change, then the model
predictions should still be valid. Where upstream dams
are anticipated to cause major changes in channel width
and depth (Williams and Wolman 1984) it will be nec-
essary to predict their effect on the inundating dis-
charges ofa representative set of plots before using the
direct gradient method to predict vegetation change.