|
R.M. Gazal et al. /Agricultural and Forest Meteorology 137 (2006) 56-67
<br />only 0.3 m at the perennial stream site. Transpiration at
<br />the intermittent site was considerably reduced after
<br />declines of about 0.5 m (ZGw = 3.5 m, Figs. 7 and 9)
<br />Similarly, other studies have showed that cottonwood
<br />forests were vulnerable to small declines in shallow
<br />alluvial water tables (Stromberg et al., 1996; Scott et al.,
<br />1999; Horton et al., 2001b,c; Cooper et al., 2003). With
<br />less than a 0.5 m decline in ZGw, cottonwood trees
<br />exhibited reduced branch growth and little mortality
<br />while greater than 1 m decline in Zcw produced leaf
<br />desiccation, branch die back, loss of canopy vigor and
<br />substantial mortality (Stromberg et al., 1996; Scott
<br />et al., 1999).
<br />Stromberg (1993) concluded that cottonwood trees
<br />on the San Pedro river basin in Arizona do not occur at
<br />sites where Zcw is greater than —3 m. Juvenile
<br />cottonwood trees, on the other hand, usually occur in
<br />areas where Zcw ranges between 0 and 2 m (Stromberg
<br />et al., 1996). Mortality of cottonwood trees was
<br />observed when &w increased above a threshold of
<br />2.5 -3.0 m along Hassayampa River, Arizona (Horton
<br />et al., 2001b). Because the trees at the intermittent
<br />stream site were likely operating on the edge of their
<br />range of suitable habitat, the increase in ZGw during the
<br />pre- monsoon drought probably reduced the trees'
<br />access to groundwater which led to observed decreases
<br />in transpiration. During the pre- monsoon season,
<br />cottonwood trees at the intermittent stream site were
<br />also observed to experience premature shedding of
<br />some leaves due to drought stress (personal observa-
<br />tion). Flood events in the nearby stream at the onset of
<br />the monsoon season recharged the alluvial aquifer
<br />causing the groundwater table to rise and E to recover.
<br />The relationship between E and Zcw (Fig. 7) implies a
<br />greater dependency of E on fluctuations in ZGw at the
<br />intermittent stream site than at the perennial stream site
<br />where Zcw was shallow enough to sustain high E during
<br />the summer. Cottonwood, an obligate phreatophyte,
<br />obtains 80 -100% of its water from groundwater, with
<br />the proportion increasing as the soil in unsaturated zone
<br />dries out (Snyder and Williams, 2000; Horton et al.,
<br />2001 c).
<br />Cottonwood trees usually possess lateral surface
<br />roots and moderately deep roots (about 3 m) that extend
<br />into the saturated zone (Stromberg, 1993). Plants with
<br />their roots in saturated soil typically show gradual
<br />reductions in E as drought progresses while plants
<br />growing in deep water sources show large reductions in
<br />E in response to drought (Breda et al., 1993; Dawson,
<br />1996; Oren et al., 1996). Hence, transpiration of
<br />cottonwood trees at the perennial stream site likely
<br />responded less to changes in soil moisture because of
<br />65
<br />their direct access to the groundwater table (Oren and
<br />Pataki, 2001). At the intermittent stream site, however,
<br />decline in the groundwater table caused large reductions
<br />in E that may be associated with the loss of hydraulic
<br />conductivity that also facilitated a reduction in stomatal
<br />conductance. At the onset of the monsoon, E at both
<br />sites responded to groundwater recharge by precipita-
<br />tion and runoff events. Hourly maximum E at the
<br />intermittent stream site increased by 50% compared to
<br />only 15% increase in E at the perennial stream site
<br />(Fig. 2). The increase in E after the groundwater
<br />recharge may be attributed to the relaxation of hydraulic
<br />resistance in the soil and soil —root interface and reversal
<br />of stem and root embolism (Oren et al., 1996; Oren and
<br />Pataki, 2001).
<br />Changes in canopy structure during drought often
<br />alleviate the direct effects of water stress on stomatal
<br />conductance and photosynthesis (Cinnirella et al.,
<br />2002). The differences in LAI between the intermittent
<br />and perennial stream sites account for most of the
<br />differences in the magnitude of E when the water table
<br />was high (Fig. 9). Our study shows that cottonwood
<br />trees at the intermittent stream site had lower LAI than
<br />at the perennial stream site. Previous studies conducted
<br />at the same sites support this result and further reveal
<br />that LAI is more variable in intermittent stream than in
<br />the perennial stream sites (Schaeffer, 1998; Schaeffer
<br />et al., 2000). Low LAI may be attributed to the deep
<br />water table at the intermittent stream site. Hence, in
<br />order to sustain transpiration with a limited amount of
<br />water, cottonwood trees and other riparian vegetation
<br />reduce their leaf area (Cooper et al., 2003). In warm and
<br />dry sites, trees respond to increased transpiration
<br />demand by reducing the ratio of leaf area to conducting
<br />sapwood area (AL:As) (Whitehead, 1998; Cinnirella
<br />et al., 2002). Our results agreed with this finding
<br />because at the perennial stream site, having shallow
<br />Zcw and ample water supply throughout the growing
<br />season, cottonwood had significantly higher AL:As than
<br />at the intermittent stream site.
<br />Modeling E using the crop factor approach revealed
<br />that EIETO is inversely related to D and Zow at both sites
<br />(Fig. 9). EIETo is more related to variations in D and
<br />Zcw at the intermittent stream site which may imply that
<br />E at the intermittent stream site was more regulated by
<br />LAI and stomata than at the perennial stream site. The
<br />average EIETo at the perennial stream site was
<br />consistently higher than the intermittent stream site
<br />throughout the season. The differences in LAI between
<br />the two sites explained much of the differences in
<br />magnitude of EIETo. Transpiration often varies due to
<br />differences in sensitivity to drought (Oren and Pataki,
<br />
|