|
200<5J Corrmwom WATER RESPONSE DuRINc DROUGHT
<br />0.01, leaf area = —32730.66 + 17007.86 *
<br />(branch diameter) + 4634.64 * (branch diame-
<br />ter — 3.03)2). This equation was applied to the
<br />diameter at the base of the live crown (DBLQ
<br />to yield an estimate of projected leaf area for
<br />each tree. To evaluate the accuracy of this
<br />approach, we compared this branch-based esti-
<br />mate with whole-canopy leaf area estimates
<br />measured on other nearby trees; these 2
<br />approaches gave similar values (data not shown).
<br />Sapwood area (SA) was estimated using tree
<br />cores for each tree, taken at the same height
<br />and aspect as the sap flux sensors (base of the
<br />live crown), and visually distinguishing between
<br />light-colored sapwood and dark-colored heart-
<br />wood.
<br />We determined whole-tree leaf loss over
<br />the course of the study for each study tree. On
<br />DOY 229, before significant drought-induced
<br />leaf loss, a litter bucket was placed under the
<br />canopy of each tree. At the end of the study,
<br />we collected litter in each bucket, dried (72
<br />hours at 70 °C) it, and then weighed it. Using
<br />the mass of each sample and the specific leaf
<br />area -value,-;, we calculated leaf area loss. Crown
<br />area of each tree was estimated using perpen-
<br />dicular rueasurenients of crown diameter and
<br />using the average of the values to calculate
<br />crown area. This value was divided by bucket
<br />area, and the result was multiplied by leaf area
<br />from each bucket to estimate total crown leaf
<br />loss during the course of the study.
<br />Air temperature and relative humidity were
<br />measured in mi open field near the study site
<br />using a Campbell Scientific CS500 air temper-
<br />ature and humidity measurement probe (Logan,
<br />UT, USA). We collected weather data every 30
<br />seconds and averaged it hourly with a Camp-
<br />bell Scientific CR LOX data logger (Logan, UT).
<br />We calculated vapor pressure deficit (VPD)
<br />from ambient temperature and relative humid-
<br />ity measurements, assuming relative humidity
<br />inside the leaves was 100% (Montictli and
<br />Unsworth 1990),
<br />All statistical analyses were done with the
<br />SAS JMP-IN statistical package (Version 4.0.4,
<br />SAS Institute, Cary, NC), with an ot of 0.05.
<br />Relationships among tree characteristics and
<br />physiological and environmental parameters
<br />were analyzed using least-squares linear re-
<br />gression. Paired t tests of overall means were
<br />used to evaluate irrigation treatment effects on
<br />physiological variables; repeated measures
<br />analyses of variance (RM ANOVAs) on daily
<br />179
<br />and weekly averages were also used to evalu-
<br />ate irrigation effects.
<br />ftsum
<br />Mean daily transpiration (Et) was similar
<br />between watered trees and unwatered trees
<br />prior to experimental water additions (P =
<br />0.95, Fig. IA), as was mean daily canopy con-
<br />ductance (C.; P = 0.93; data not shown).
<br />Gravimetric soil water content also was similar
<br />between watered and unwatered trees prior to
<br />the watering treatment (P = 0.06; Fig. 1B).
<br />Water addition significantly increased the
<br />gravimetric soil water content (P = 0.03).
<br />During the study period gravimetric soil water
<br />content under watered trees increased signifi-
<br />cantly from 5.9% (1:0.41 sk) to 22.779b (+-0.98 sj;
<br />P = 0.03), during the same period, gravimetric
<br />soil water content among unwatered trees
<br />decreased significantly from 7.0% (±0.67 sf) to
<br />6.2% (±0.41 s.f; P = 0.02, Mg. IB). Although
<br />supplemental watering was effective in increas-
<br />ing surface soil moisture, E, (P = 0.47; Fig.
<br />IA), G, (P = 0.84; Fig, 2B), and whole tree
<br />hydraulic conductance (IC),; P = 0.63) were
<br />not significantly different between watered
<br />and unwatered trees, Both Tp,,, and T.id Also
<br />were similar between watered and unwatered
<br />trees (Tpre . P = 0.83, %F,ld: P = 0.62), with
<br />Tp., averaging about —0.54 MPa and 'Fmid
<br />approximately —1.58 NIPa during the measure-
<br />ment period (Fig. IC).
<br />We found a significant inverse linear rela-
<br />tionship between TP., and G, (P = 0.02, r2 =
<br />0.44; Fig. 2A). However, there was no relation-
<br />ship between soil gravimetric water content
<br />and G, (P = 0.47; Fig. 213). Similarly, we found
<br />a significant inverse relationship between
<br />`Epee -and Ei (P = 0.04, r2 = 0.35; Fig. 20, but
<br />there was no significant relationship between
<br />gravimetric soil water content and E, (P =
<br />0.34; Mg. 2D). Relationships between VPD
<br />and El were significant (P < 0.05, Fig. 3A) for
<br />both watered trees (r2 - 0,33) and unwatered
<br />trees (r2 = 0.21), as, were relationships between
<br />VPD and G, (P < 0.01, r2 = 0.42 [watered]
<br />and 0.44 [unwatered], Fig. 3B). Slopes of re-
<br />sponse curves for G, versus VPD relationships
<br />had overlapping 95% confidence intervals be-
<br />tween watered and unwatered trees and thus
<br />were not considered different.
<br />Both E, and C. were not significantly cor-
<br />related with either DBLC (P = 0.17 and 0.20,
<br />
|