WSP09221

• File Number: 7630.425
• Description: Wild and Scenic - Piedra River
• State: CO
• Basin: Western Slope
• Water Division: 7
• Date: 1/1/1991
• Author: Julie Stromberg Dunc
• Title: Instream Flow Requirements for Cottonwoods at Bishop Creek - Inyo County - California
• Water Supply Pro - Doc Type: Report/Study

aon079

k'

commonality in growth pattern between
reaches.
Univariate regression analysis was used
to explore relationships between annual
ring width from 1969 to 1988 and the fol.
lowing variables: (1) volume of streamflow
during the water-year and during the
growing season (May-October), (2) prior-
year flow volume during the water-year
and during the growing season, (3) pre-
cipitation during the water-year and grow-
ing season, (4) average maximum temper-
ature during the growing season and
during individual months, and (5) prior-
year ring width. Stt'pwise multiple regres-
sion analysis was then used to develop
reach-specific models that predicted an-
nual ring width of the cottonwoods from
the most appropriate variables. Prior-year
ring width was included as a variable be-
cause it serves as a biotic integrator of pri-
or-year environmental conditions (Strom-
berg and Patten 1990). Average volume of
flow was selected as the instream flow vari-
able over parameters such as the instan-
taneous minima (Kulik 1990) because most
riparian trees can withstand short periods
of drought. Flow data for reach 6 were
mea!>ured discharge values from power
plant 6, whereas flow data for the diverted
reaches were estimated based on assump-
tions on inputs and losses from overland
runolT and other sources, respective to the
nearest stream gage (Simons, Li and As-
sociates 1990).
The iostream models for reaches 2 and
5 were calibrated on chronologies devel-
oped from randomly selected groups of
trees and verified on the other half of the
sample (n = 14 and 7, respectively). Pear-
son correlations (r) between predicted and
actual ring widths were then calculated as
within-reach verification statistics (e.g..
Cook and Jacoby 1983). Models for reach
4 cottonwoods were not verified due 10
small sample size. The models were also
verified between reaches by determining
correlation coefficients between ring
widths for d given reach and the values

Elevation
dbh (em) (m)
25:t. 7 2,024
29:t. 9 2.020
29.= 7 2,015
36 ::t II 1,.508
21::t 4 lA68
19::t6 1.465
42:!: 17 1.390
45:!: 22 1.387
50::!: 7 L385
71 ::t 19 1.380
1Cnocllrpll. POFR - P.

the classes mini-

'e mounted and
rd methods (Fritts
~ between sped-
"ere rneasu red to
ng an automated
:ing-width chra-
for each tree bv
or the two repli-
individual trft's
~o a value of 1 to
rowth due to age
~d titting a curve
,ronology and di-
ing width by the
he curve. Reach
Ited by averaging
,ogies for all ma-

~ ring chronolo-
'ar-to-vear varia-
:,) was' calculated
n coefficients bE--
.vere determined
h reach as a mea-
commonality in
:orrelation coef-
mined between
measure of the

Instream Flow-Growth Models

DIi.'l'rft.J Rt'Qcllt'5. Growth of trees in the
three diverted reaches has varied over time,

January 1991

I J. C. Stromberg and D. T. Patten

(,)

,,~
a-..1l(./IO<6
1"-AFlu>c"5
1.-.R[.OC>l.
.2j"'-Or.u.c>l2
.-
I /
. -'
6~"a--"--'__,,~

/
.
-

.
,
,

'" (
\
.

"
,
'-
-
o
~
"
o

I
A~,
o. . . . . . ~/, "
OCTOIOVOl:ColAt<FUlw.Il_......y,I)N..IlA.IOlICSEP

,
r
"
r
o
>

(
,-

I') '''~. . I
" "" /\ (

: ::1., .~. IVV' \^ I
:1 ../ \/ '. '.
,:L~~~tv~hJ
~~1'nnu~H71nn~81~uB.~UB7~
",...rtR1'Uo'I

"
o
,
,
,

FIGURE 2. ,..II"~nmonthJy(A}~nd~nnu~118)
strt'~mflows in th~ w~t~r;y~~r for 1969-1988
~t four r~~ch~s on Bishop Creek.

predicted by instream models developed
for a different reach.
Univariate regression analysis was used
to develop predictive equations relating
tree canopy vigor to average ring width
over the last 10, 15, or 20 years, assuming
that present-day vigor reflects a composite
and time-lagged response to past growth
rate of the trees. This was done for each
reach (using tree values) and for the study
area as a whole (using site means). Multi-
variate regression equations were devel-
oped relating site mortality (dead trees as
percent of the total) to average site ring
width for 1969-1988 and average age of
the cottonwood trees in the stand. This was
done after statistically adjusting ring width
values to reflect that of a 35-year-old tree
b)' using curve.titting techniques (Fritts
1976). Tree age was calculated as the num-
l:>er of annual rings plus estimated time to
grow to coring height: 2 years at reaches 5
and 6, and 4 years elsewhere.

RESULTS

as evidenced by high values for mean sen-
sitivity of the chronologies (Table 2) and
as seen graphically in Figure 3. Ring-width
chronologies generally were highly cor-

5 II~,