Differences in leaf gas exchange and water relations among species and tree sizes in an Arizona pine oak forest

Transcription

1 Tree Physiology 20, Heron Publishing Victoria, Canada Differences in leaf gas exchange and water relations among species and tree sizes in an Arizona pine oak forest T. E. KOLB and J. E. STONE School of Forestry, College of Ecosystem Science and Management, Northern Arizona University, Flagstaff, Arizona , USA Received April 9, 1999 Summary We compared leaf gas exchange and water potential among the dominant tree species and major size classes of trees in an upland, pine oak forest in northern Arizona. The study included old-growth Gambel oak (Quercus gambelii Nutt.), and sapling, pole, and old-growth ponderosa pines (Pinus ponderosa var. scopulorum Dougl. ex Laws.). Old-growth oak had higher predawn leaf water potential (Ψ leaf ) than old-growth pine, indicating greater avoidance of soil water stress by oak. Old-growth oak had higher stomatal conductance (G w ), net photosynthetic rate (P n ), and leaf nitrogen concentration, and lower daytime Ψ leaf than old-growth pine. Stomatal closure started at a daytime Ψ leaf of about 1.9 MPa for pine, whereas old-growth oak showed no obvious reduction in G w at Ψ leaf values greater than 2.5 MPa. In ponderosa pine, P n and G w were highly sensitive to seasonal and diurnal variations in vapor pressure deficit (VPD), with similar sensitivity for sapling, pole, and old-growth trees. In contrast, P n and G w were less sensitive to VPD in Gambel oak than in ponderosa pine, suggesting greater tolerance of oak to atmospheric water stress. Compared with sapling pine, old-growth pine had lower morning and afternoon P n and G w, predawn Ψ leaf, daytime Ψ leaf, and soil-to-leaf hydraulic conductance (K l ), and higher foliar nitrogen concentration. Pole pine values were intermediate between sapling and old-growth pine values for morning G w and daytime Ψ leaf, similar to sapling pine for predawn Ψ leaf, and similar to old-growth pine for morning and afternoon P n, afternoon G w, K l, and foliar nitrogen concentration. For the pines, low predawn Ψ leaf, daytime Ψ leaf, and K l were associated with low P n and G w. Our data suggest that hydraulic limitations are important in reducing P n in old-growth ponderosa pine in northern Arizona, and indicate greater avoidance of soil water stress and greater tolerance of atmospheric water stress by old-growth Gambel oak than by old-growth ponderosa pine. Keywords: Gambel oak, hydraulic conductance, nitrogen, photosynthesis, Pinus ponderosa, ponderosa pine, productivity, Quercus gambelii, stomatal conductance, vapor pressure deficit, water stress. Introduction We used field measurements of leaf water potential (Ψ leaf ) and gas exchange on old-growth ponderosa pine (Pinus ponderosa var. scopulorum Dougl. ex Laws.) and old-growth Gambel oak (Quercus gambelii Nutt.), and on different size classes of ponderosa pine to examine two questions. The first concerns adaptations of Gambel oak to the stressful climatic conditions that characterize pine oak forests in northern Arizona. In this area, dry soil and low relative humidity occur in late spring at the same time that rising temperatures promote bud break and growth of deciduous trees such as Gambel oak. Precipitation in May and June is scarce, averaging 1.9 and 1.3 cm, respectively, in Flagstaff, Arizona. Afternoon relative humidity averages 21% in June, and values near 10% often occur (Western Regional Climatic Data Center 1999). This period of dry soil and high vapor pressure deficit (VPD) typically extends into July, when the onset of thunderstorms and increased atmospheric humidity reduces water stress on ponderosa pine (Feeney et al. 1998, Kolb et al. 1998). Most studies of oak adaptation to water stress have focused on species native to moister climates in eastern North America (e.g., Abrams 1990, Ni and Pallardy 1990, Ni and Pallardy 1991, Kubiske and Abrams 1993, Abrams et al. 1994, Kubiske and Abrams 1994, Loewenstein and Pallardy 1998a, 1998b). These studies have shown both avoidance and tolerance of water stress by oaks (Abrams 1990). However, few studies have examined adaptation to water stress by Gambel oak (Neilson and Wullstein 1985, Phillips and Ehleringer 1995, Williams and Ehleringer 1996). Because Gambel oak is winter deciduous with bud burst occurring in late spring in northern Arizona, a greater proportion of its assimilative activity takes place under conditions of dry soil and high VPD compared with ponderosa pine, which is evergreen. Consequently, we hypothesized that Gambel oak should exhibit stronger avoidance or tolerance of soil and atmospheric water stress than similar-aged ponderosa pine. The second question that we examined is whether leaf gas exchange rates change during tree maturation. This question is relevant to several current issues in forest ecology. For exam-

2 2 KOLB AND STONE ple, there is considerable interest in understanding the mechanisms that underlie decreased productivity of old forests (Ryan et al. 1997). A decrease in net photosynthetic rate (P n ) during tree maturation because of decreased xylem hydraulic conductance has been proposed as a mechanism (Yoder et al. 1994, Mencuccini and Grace 1996, Ryan and Yoder 1997, Hubbard et al. 1999). Moreover, changes in physiological traits during the aging of woody plants can provide insight into factors that influence the recruitment and survival of a species (Donovan and Ehleringer 1991, 1992, Franco et al. 1994, Miller et al. 1995, Donovan and Pappert 1998). Further, interest in understanding the impacts of air pollution on forests has highlighted the need for a better understanding of changes in tree water and carbon relations during tree maturation (Grulke and Miller 1994, Samuelson et al. 1996, Kolb et al. 1997, Samuelson and Kelly 1997). For example, in several species, difference in physiological response to ozone between young and old trees has been associated with differences in stomatal conductance (G w ) (Kolb et al. 1997). We hypothesized that P n decreases with increasing tree size and age of ponderosa pine in northern Arizona as has been reported for this species in eastern Oregon (Yoder et al. 1994, Hubbard et al. 1999). We also hypothesized that this response is related to differences in G w, Ψ leaf, soil-to-leaf hydraulic conductance (K l ), or leaf nitrogen concentration among tree sizes. Materials and methods Study site The study site is Camp Navajo, a U.S. Army National Guard facility located 16 km west of Flagstaff, Arizona (35 15 N, W). The site has gently rolling topography with a northwestern aspect and an elevation of 2,219 m. Soils are derived from basalt and are classified as Brolliar stony clay loam, fine, smectitic, Typic Argiborolls. Air temperature at a weather station located 5 km from the study site ranges between a mean minimum of 9.1 C in January and a mean maximum of 27.7 C in July (Western Regional Climatic Data Center 1999). The frost-free growing season at the elevation of the study site typically starts in late May and ends in early October. Mean annual precipitation is 55.5 cm with about half of this amount usually occurring as snow and the balance occurring as intense thunderstorms in July, August and September. The only tree species at the study site are ponderosa pine and Gambel oak. The stand is uneven-aged and includes sapling, pole-sized, and old-growth ponderosa pines, and old-growth Gambel oak. Oak and the smaller pines occur in openings among clumps of old-growth pine. Typical of many forests in the southwestern USA, tree density at the study site has increased over the last century because of fire suppression and other factors (Fulé et al. 1997). Leaf water relations, gas exchange, and nitrogen concentration We measured Ψ leaf, P n, G w, and leaf nitrogen concentration on a representative sample of trees of each type (Table 1) approximately every two weeks between mid-may and early November The duration of measurement was June 11 to October 10 for oak, because it is winter deciduous. In 1998, bud burst in oak started in late May and leaves were fully expanded by mid-june. Oak leaves showed no visible evidence of autumn coloration until mid-october. Measurements on pine were restricted to 1-year-old leaves (elongated in 1997). In 1998, bud burst in pines also occurred in late May, but full expansion of current-year leaves did not occur until late September after the onset of late-summer rains, so they were not measured. We first selected five representative old-growth pines in the stand, and then selected one old-growth oak, and a group of sapling and pole pines located nearest to each old-growth pine. We measured the same five individuals of old-growth pine and oak on all dates. For the smaller pines (sapling and pole), we sampled a total of 10 or 11 trees; five of these trees were Table 1. Number of trees sampled over all dates (n), and the mean and range (in parentheses) in diameter at breast height (DBH), estimated age, 1 total tree height, and sample height in the canopy for old-growth Gambel oak, and sapling, pole, and old-growth ponderosa pines at Camp Navajo, Arizona, Species/size n DBH (cm) Age (years) Total height (m) Sample height (m) Oak ( ) ( ) ( ) ( ) Sapling pine ( ) (15 29) ( ) ( ) Pole pine ( ) (58 86) ( ) ( ) Old-growth pine ( ) ( ) ( ) ( ) 1 Ages of oak, pole pine, and old-growth pine were estimated from a linear regression on DBH developed at Camp Navajo (r 2 = 0.73 oak, r 2 = 0.58 pine) (Personal communication, P.Z. Fulé, Northern Arizona University, School of Forestry). Ages of sapling pine were measured by counting tree rings at the base of the stem. TREE PHYSIOLOGY VOLUME 20, 2000

3 LEAF GAS EXCHANGE AND WATER RELATIONS IN A PINE OAK FOREST 3 randomly selected for measurement for each tree size on each date to minimize the effects of defoliation caused by sampling the small trees. All measured trees were exposed to full sun much of the day. We measured Ψ leaf with a pressure chamber (Model 1000, PMS Instruments, Corvallis, OR). On most dates, we measured Ψ leaf at predawn (0500 h), in the morning (0900 h), and in the afternoon (1300 h). However, rain prevented the full set of measurements on several dates. For oak and for pole and old-growth pines, leaves for these measurements were obtained from one twig per tree excised with a pole pruner from the mid-canopy of oak and pole pine, and from the lower canopy of old-growth pine (Table 1). Morning and afternoon samples were restricted to portions of the crown illuminated with full sun. Immediately following excision of a twig from the crown, leaves were cut from the twig with a razor blade and sealed in a plastic bag containing a damp towel and stored under dark cool conditions until measured (within 2 h). Leaves from sapling pines were excised directly from the canopy, but were otherwise treated similarly to the other sampled leaves. This procedure of sampling, storage and measurement yielded Ψ leaf values similar to those obtained when leaves were excised from the tree and measured immediately (Kaufmann and Thor 1982, authors unpublished data). For both species, we measured Ψ leaf on several leaves from each twig until three values within 0.1 MPa of each other were obtained; the mean of these values was taken as the Ψ leaf of the tree. We measured P n and G w on leaves attached to the same twigs sampled for morning and afternoon Ψ leaf measurements. Measurements were made with an LI-6200 portable photosynthesis system (Li-Cor Inc., Lincoln, NE) equipped with a 250-ml cuvette. Preliminary studies showed that P n and G w of pine and oak leaves on detached twigs did not differ from P n and G w of leaves attached to the tree when measured immediately (within 1 min) following twig excision. We measured two fascicles (six needles) per twig for pine, and one leaf per twig for oak. Fluxes of water and CO 2 through the cuvette were measured over 30 s. The cuvette was placed in full sun to ensure high irradiances. During measurement, water vapor pressure and temperature in the cuvette were generally within 10% and 3 C of ambient values, respectively. We calculated P n and G w based on total leaf surface area for both pine and oak. For pine, total leaf surface area was estimated by measuring the length and diameter of the needle portions contained in the gas exchange cuvette and assuming that each fascicle approximated a dissected cylinder (Svenson and Davies 1992). For oak, total leaf surface area was calculated as twice the projected area. For pine, we estimated soil-to-leaf hydraulic conductance (K l ) for all gas exchange measurements by dividing leaf-level transpiration rate from the cuvette gas exchange measurement (mol H 2 Om 2 s 1 ) by the difference in Ψ leaf (MPa) between predawn and the time of the gas exchange measurement (e.g., Ni and Pallardy 1990, Hubbard et al. 1999). We did not estimate K l for oak with this approach because our ventilated cuvette minimizes boundary layer resistance, which is an important component of in situ transpiration by broad leaves like oak. We measured total nitrogen concentration on all leaves used for gas exchange measurements with a carbon/nitrogen analyzer (Model NC 2100, CE Elantech, Inc., Lakewood, NJ). Soil water content We measured soil volumetric water content approximately every two weeks at 10 locations placed systematically within the study site. At each location, soil water content was measured at depths of 0 15 cm and cm in the mineral soil by time domain reflectometry (Model 6050X1, Trase System I, Soilmoisture Equipment Corp., Santa Barbara, CA) (Rundel and Jarrell 1989). Statistical analysis We compared physiological response variables between old-growth oak and old-growth pine, and among the different sizes of pine (sapling, pole, old-growth) by fixed-effects analysis of variance (ANOVA). Factors in the ANOVA were species or tree size, sample date, and their interaction. Analyses were performed separately by measurement time (predawn, morning, afternoon) because not all times were represented on all dates. The simplifying assumptions made for these ANOVAs (e.g., sample dates were independent, all tree types were sampled the same) suggest that P values for F-tests that are close to 0.05 should be interpreted with caution. When the species or tree size main effect was significant (P 0.05), but interaction between species or tree size and sample date was not significant, we used protected LSD comparisons at a threshold P value of 0.05 to test for differences among species or tree size means over all sample dates. When interaction between species or tree size with sample date was significant, we tested for differences among species or tree sizes with pair-wise t-tests (P = 0.05) at each sample date. We also used correlation and regression techniques to study relationships among physiological and environmental variables. Gas exchange and associated morning and afternoon Ψ leaf data measured at photosynthetically active radiation (PAR) < 800 µmol m 2 s 1 were deleted from these analyses to insure comparable high-light conditions. Over 90% of the data used in these analyses were measured at PAR values between 1200 and 2000 µmol m 2 s 1. Results Environment Seasonal variation in precipitation in the year of our study (1998) was typical for northern Arizona, with a pronounced dry period in May and June that occurred between winter snow and late-summer rains. In 1998, precipitation at a weather station located near the study site (5 km) was 9.5 cm in March (3.6 cm > long-term mean), 3.9 cm in April (0.6 cm > long-term mean), 3.3 cm in May (1.4 cm > long-term mean), 0.0 cm in June (1.3 cm < long-term mean), 12.0 cm in July (5.7 cm > long-term mean), 7.2 cm in August (0.1 cm < long-term mean), and 11.0 cm in September (6.1 cm > long-term mean). TREE PHYSIOLOGY ON-LINE at

4 4 KOLB AND STONE Figure 1. Mean soil water content at two soil depths (0 15 cm, cm) and leaf-to-air vapor pressure deficit (VPD) in the morning and afternoon during the study. Soil water content declined between late April (Day 120) and mid-july (Day 195) (Figure 1). Episodes of soil recharge following rains occurred between late July (Day 211) and early November (Day 311) (Figure 1). Leaf-to-air vapor pressure deficit (VPD), calculated from leaf and air temperatures measured in the cuvette and relative humidity of the bulk atmosphere, increased between early May and mid-july (Figure 1). Afternoon VPD was consistently greater than morning VPD (Figure 1). Among sample dates, variation in mean soil water content at the 0 15 cm depth was positively and significantly correlated with mean predawn Ψ leaf (Figure 2) for both species and all tree sizes (old-growth oak, r = 0.834, P = 0.020; old-growth pine, r = 0.707, P = 0.020; sapling pine, r = 0.891, P = ; pole pine, r = 0.909, P = ). In contrast, correlation between soil water content at the cm depth and predawn Ψ leaf was non-significant for both species and all tree sizes (r 0.425, P 0.340). Differences between old-growth oak and old-growth pine Species, sample date, and the species sample date interaction were significant sources of variation (P 0.002) for predawn, morning, and afternoon Ψ leaf in comparisons between old-growth oak and old-growth pine. Predawn Ψ leaf of old-growth oak was significantly higher than predawn Ψ leaf of old-growth pine on every date (Figure 2). Differences in predawn Ψ leaf between old-growth oak and pine varied over sample dates; the largest differences (0.5 to 0.7 MPa) occurred in July (Days 195 and 211) and late August (Day 241), and the smallest difference (0.2 MPa) occurred in early June (Day 162) (Figure 2). Morning Ψ leaf was significantly greater for old-growth oak than for old-growth pine in early June (Day 162) when the oak leaves were not fully expanded (Figure 2). In contrast, morning Ψ leaf of old-growth oak was significantly lower by 0.3 to 0.5 MPa compared with morning Ψ leaf of old-growth pine on several sample dates between mid-july and mid-october (Days 195, 224, 241, and 283). Afternoon Ψ leaf for old-growth oak was significantly greater than afternoon Ψ leaf of old-growth pine in early June (Day 162) before full leaf expansion by oak (Figure 2). Also, afternoon Ψ leaf for old-growth oak was significantly lower than afternoon Ψ leaf of old-growth pine in July (195 and 211). Differences in Ψ leaf between old-growth oak and sapling and pole pines were similar to differences between old-growth oak and old-growth pine (Figure 2). Species, sample date, and the species sample date interaction were significant sources of variation (P < ) for both morning and afternoon P n and G w in comparisons between old-growth oak and old-growth pine. Morning P n was significantly lower for old-growth oak compared with old-growth pine before full leaf expansion by oak in early June (Day 162) (Figure 3). However, in July, August, and September (Days 195 to 255), morning P n was two- to three-fold greater for old-growth oak than for old-growth pine. Moreover, morning P n of old-growth oak increased between early May and mid-july (Days 162 to 195), whereas morning P n of old-growth pine decreased (Figure 3). Patterns of variation and differences between old-growth oak and old-growth pine in afternoon P n were generally similar to patterns and differences for morning P n (Figure 3). Stomatal conductance (G w ) was correlated (P < ) with P n for both oak and pine for data pooled over all measurements (old-growth oak r = 0.849, old-growth pine r = 0.838, sapling pine r = 0.841, pole pine r = 0.885). This correlation resulted in similar patterns of temporal variation and similar differences between old-growth oak and old-growth pine in G w as for P n (Figures 3 and 4). Differences in P n and G w between old-growth oak and sapling and pole pines were generally similar to differences between old-growth oak and old-growth pine (Figures 3 and 4). The relationship between VPD and P n and G w differed between old-growth oak and old-growth pine (Figure 5). For oak, P n was not correlated with VPD (r = 0.066, P = 0.619), and the correlation between G w and VPD was weak (r = 0.254, P = 0.052). For old-growth pine, correlation with VPD for both P n and G w was negative and highly significant (P n : r = 0.822, P < ; G w : r = 0.735, P < ) (Figure 5). Both P n and G w were significantly related to predawn Ψ leaf for both old-growth oak and old-growth pine (Figure 6). Correlation of P n and G w with predawn Ψ leaf was stronger for old-growth pine (P n : r = 0.518, P < ; G w : r = 0.524, P < TREE PHYSIOLOGY VOLUME 20, 2000

5 LEAF GAS EXCHANGE AND WATER RELATIONS IN A PINE OAK FOREST 5 Figure 2. Mean leaf water potential (Ψ leaf ) at predawn (0500 h), morning (0900 h), and afternoon (1300 h) for old-growth Gambel oak and old-growth ponderosa pine, and for three sizes of ponderosa pine (sapling, pole, and old-growth). Significant (P 0.05) sources of variation are shown for each ANOVA (SP*DA = species date interaction, and SI*DA = tree size date interaction). The bars show one standard error of the mean for old-growth Gambel oak and ponderosa pine. Symbols: = old growth oak; X = old-growth pine; + = sapling pine; and = pole pine ) than for oak (P n : r = 0.291, P = 0.025; G w : r = 0.430, P = ). The small amount of overlap in predawn Ψ leaf for old-growth oak and old-growth pine (Figure 6) prevented a slope comparison of P n and G w responses to predawn Ψ leaf between species. Old-growth oak and old-growth pine differed in the relationships of P n and G w with Ψ leaf at the time of the gas exchange measurement (Figure 7). For old-growth oak, correlation with daytime Ψ leaf was negative and significant for P n (r = 0.303, P = 0.021), but not significant for G w (r = 0.030, P = 0.824). For old-growth pine, correlation with Ψ leaf was positive and significant for both P n (r = 0.306, P = 0.006) and G w (r = 0.314, P = 0.005). Boundary-line analysis (Chambers et al. 1985) of the upper values of G w in Figure 7 suggested that the onset of stomatal closure and reduced P n occurred at a Ψ leaf of about 1.9 MPa for old-growth pine. In contrast, for old-growth oak there was no obvious indication of stomatal closure or an effect on P n at Ψ leaf values greater than 2.5 MPa. Significant sources of variation for leaf nitrogen concentration in the comparison between old-growth oak and old-growth pine were species (P = ), sample date (P = 0.017), and the species sample date interaction (P = ). Mean leaf nitrogen concentration of old-growth oak ranged between 32.7 mg g 1 in mid-june, before full leaf expansion, to 19.3 mg g 1 in mid-october, before the onset of autumn coloration. Leaf nitrogen concentration varied less over the measurement period in old-growth pine than in old-growth oak, with no obvious systematic variation among sample dates (11.7 to 13.9 mg g 1 ). Leaf nitrogen concentration was significantly greater in old-growth oak than in old-growth pine at all sample dates (mean over all dates, 23.7 versus 12.8 mg g 1,respectively). Differences between sapling, pole, and old-growth pines Tree size, sample date, and the tree size sample date interaction were significant sources of variation (P 0.002) for predawn and morning Ψ leaf in comparisons among different-sized pines. Predawn Ψ leaf was typically lowest for old-growth pine and greatest for either sapling or pole pines depending on the sample date (Figure 2), and these differences were significant on several days throughout the measurement period (Days 131, 140, 211, 241, 255, and 297). Sapling and pole pines had similar (P 0.05) predawn Ψ leaf on most dates (Figure 2). The largest differences in predawn Ψ leaf among tree sizes (0.3 to 0.4 MPa) occurred between old-growth and sapling pines in late summer (Days 241 and 255). TREE PHYSIOLOGY ON-LINE at

6 6 KOLB AND STONE Figure 3. Mean net photosynthetic rate (P n ) in the morning (0900 h) and afternoon (1300 h) for old-growth Gambel oak and old-growth ponderosa pine, and for three sizes of ponderosa pine (sapling, pole, and old-growth). Net photosynthetic rate was calculated on a total leaf area basis for both species. Significant (P 0.05) sources of variation are shown for each ANOVA (SP*DA = species date interaction, and SI*DA = tree size date interaction). The bars show one standard error of the mean for old-growth oak and pine. Symbols: = old growth oak; X = old-growth pine; + = sapling pine; and = pole pine. Morning Ψ leaf did not differ significantly among different-sized pines between mid-may and mid-july (Days 131 to 195) (Figure 2). However, between late July and early September (Days 211 to 255), morning Ψ leaf of old-growth pine was 0.2 to 0.6 MPa lower than morning Ψ leaf of sapling pine, and these differences were significant on several Days (224, 241, and 255). Morning Ψ leaf values of pole pine tended to be intermediate between those of sapling and old-growth pines during this period (Figure 2). In November (Day 311), morning Ψ leaf was significantly lower (0.25 MPa) in old-growth pine than in sapling and pole pines (Figure 2). For afternoon Ψ leaf of different-sized pines, tree size and sample date were significant (P ) sources of variation, whereas the tree size sample date interaction was not significant (P = 0.116). Mean afternoon Ψ leaf over all sample dates was significantly higher for sapling pine ( 1.59 MPa) than for pole ( 1.73 MPa) and old-growth pines ( 1.79 MPa), which also differed significantly. Sapling pine had the highest after- Figure 4. Mean stomatal conductance (G w ) in the morning (0900 h) and afternoon (1300 h) for old-growth Gambel oak and ponderosa pine, and for three sizes of ponderosa pine (sapling, pole, and old-growth). Stomatal conductance was calculated on a total leaf area basis for both species. Significant (P < 0.05) sources of variation are shown for each ANOVA (SP*DA = species date interaction, and SI*DA = tree size date interaction). The bars show one standard error of the mean for old-growth oak and pine. Symbols: = old growth oak; X = old-growth pine; + = sapling pine; and = pole pine. TREE PHYSIOLOGY VOLUME 20, 2000

7 LEAF GAS EXCHANGE AND WATER RELATIONS IN A PINE OAK FOREST 7 Figure 5. Relationship between vapor pressure deficit (VPD) and net photosynthetic rate (P n ) and stomatal conductance (G w ) in old-growth Gambel oak and old-growth ponderosa pine, and in three sizes of ponderosa pine (sapling, pole, and old-growth). Data include both morning and afternoon measurements over all sample dates. Correlation between P n and VPD: sapling pine r = 0.776, P < ; pole pine r = 0.747, P < ; old-growth pine r = 0.822, P < ; and old-growth oak r = 0.066, P = Correlation between G w and VPD: sapling pine r = 0.762, P < ; pole pine r = 0.662, P < ; old-growth pine r = 0.735, P < ; and old-growth oak r = 0.254, P = noon Ψ leaf of all size classes on eight of the nine sample dates (Figure 2). Tree size and sample date were significant (P 0.020) sources of variation for morning and afternoon P n in comparisons among different-sized pines, whereas the tree size sample date interaction was not significant (P 0.447) (Figure 3). Mean morning P n over all sample dates was significantly higher for sapling pine (5.40 µmol m 2 s 1 ) compared with pole (4.81 µmol m 2 s 1 ) and old-growth pines (4.57 µmol m 2 s 1 ), which did not differ significantly. Differences in morning P n among tree sizes were fairly consistent over sample dates; saplings had higher morning P n than larger pines on eight of 10 sample dates (Figure 3). Patterns of temporal variation for afternoon P n were similar to patterns for morning P n (Figure 3). Mean afternoon P n over all sample dates was significantly higher for sapling pine than for pole pine (4.67 versus 4.09 µmol m 2 s 1 ). Mean afternoon P n of old-growth pine (4.39 µmol m 2 s 1 ) was intermediate between sapling and pole pines, and did not differ significantly from them. Differences in afternoon P n among tree sizes were less consistent over sample dates than differences in morning P n (Figure 3). Significant sources of variation for G w in comparisons among different-sized pines were similar to significant sources for P n, except that the tree size sample date interaction was significant in both the morning (P = ) and the afternoon (P = 0.036) (Figure 4). As for P n, mean morning G w over all sample dates was significantly higher for sapling pine (61.7 mmol m 2 s 1 ) than for pole (53.7 mmol m 2 s 1 ) and old-growth (47.2 mmol m 2 s 1 ) pines, which also differed significantly. Sapling pine had the highest morning G w of all tree sizes on eight of the 10 sample dates (Figure 4). The only notable exception to this trend was significantly higher morning G w of pole pine compared with sapling and old-growth pines on Day 255 (Figure 4). In the afternoon, mean G w over all sample dates was significantly higher for sapling pine (53.1 mmol m 2 s 1 ) than for pole (43.9 mmol m 2 s 1 ) and old-growth (44.6 mmol m 2 s 1 ) pines, which did not differ significantly. Sapling pine had the highest afternoon G w of all tree sizes on five of eight sample dates, and these differences were significant on Days 131 and 211 (Figure 4). Also, in mid-october (Day 297), afternoon G w was significantly higher for sapling and pole pines than for old-growth pine (Figure 4). Correlation with VPD for P n and G w was negative and significant (P < ) for all sizes of pine (Figure 5). Correlation with predawn Ψ leaf for P n and G w was positive and significant (P < ) for all sizes of pine (Figure 6). Slopes for linear regressions of P n and G w on VPD, and P n and G w on predawn Ψ leaf did not differ significantly (P > 0.05) among different-sized pines (data not shown). TREE PHYSIOLOGY ON-LINE at

8 8 KOLB AND STONE Figure 6. Relationship between net photosynthetic rate (P n ) and stomatal conductance (G w ) with predawn leaf water potential (Ψ leaf ) for old-growth Gambel oak and old-growth ponderosa pine, and for three sizes of ponderosa pine (sapling, pole, and old-growth). Data include both morning and afternoon measurements over all sample dates. Correlation between P n and predawn Ψ leaf : sapling pine r = 0.513, P < ; pole pine r = 0.418, P = ; old-growth pine r = 0.518, P < ; and old-growth oak r = 0.291, P = Correlation between G w and predawn Ψ leaf : sapling pine r = 0.585, P < ; pole pine r = 0.412, P < ; old-growth pine r = 0.524, P < ; and old-growth oak r = 0.430, P = Both P n and G w were positively and significantly correlated with Ψ leaf at the time of the gas exchange measurement for sapling and old-growth pines, but not for pole pine (Figure 7). As for old-growth pine, the onset of stomatal closure and a reduction in P n for sapling pine occurred at a Ψ leaf of about 1.9 MPa (Figure 7). For pole pine, this relationship was less clear (Figure 7). Soil-to-leaf hydraulic conductance (K l ) of pine varied significantly among tree sizes (P < ) and sample dates (P < 0.001) in both the morning and afternoon, whereas the tree size sample date interaction was not significant (morning: P = 0.338; afternoon P = 0.840). For both morning and afternoon, mean K l of sapling pine was significantly greater than mean K l of both pole and old-growth pines, which had similar values of K l (Table 2). For data pooled over all sizes of pine and all sample times, both P n and G w were positively and significantly correlated with K l (P n : r = 0.284, P < ; G w : r = 0.404, P < ) (Figure 8). Significant sources of variation for leaf nitrogen concentration in the comparison among different-sized pines were tree size (P = ) and sample date (P = 0.002), whereas the tree size sample date interaction was not significant (P = 0.879). Mean nitrogen concentration over all sample dates was significantly higher for pole (12.9 mg g 1 ) and old-growth (12.8 mg g 1 ) pines than for sapling pine (11.9 mg g 1 ). Discussion Between May and mid-july at our study site, soil water content decreased and VPD increased, as has been observed in other recent studies in northern Arizona ponderosa pine forests (Kolb et al. 1998, Feeney et al. 1998, Stone et al. 1999). Spring bud burst, leaf expansion, and the onset of leaf assimilation in Gambel oak typically occur during this drying trend. In contrast, greater leaf longevity (3 to 5 years) in ponderosa pine allows assimilation to occur throughout the year if environmental conditions are favorable. These factors led to our hypothesis that Gambel oak has developed more effective mechanisms for avoidance or tolerance of soil and atmospheric water stress compared with similar-aged ponderosa pine in northern Arizona. We found strong evidence that old-growth Gambel oak is more effective at avoiding soil water stress than old-growth ponderosa pine. Old-growth oak had consistently higher predawn Ψ leaf than old-growth pine between June and October, and the differences were often large (0.5 to 0.7 MPa). Although our correlation analysis suggested reduced P n and G w with decreasing predawn Ψ leaf for both old-growth oak and pine, low predawn Ψ leaf rarely occurred in oak. The old-growth oak in our study may have deeper and more extensive roots than old-growth pine. Deep roots have been associated with greater avoidance of soil water stress in comparative TREE PHYSIOLOGY VOLUME 20, 2000

9 LEAF GAS EXCHANGE AND WATER RELATIONS IN A PINE OAK FOREST 9 Figure 7. Relationship to leaf water potential at the time of the gas exchange measurement (daytime Ψ leaf ) for net photosynthetic rate (P n ) and stomatal conductance (G w ) for old-growth Gambel oak and old-growth ponderosa pine, and for three sizes of ponderosa pine (sapling, pole, old-growth). Data include both morning and afternoon measurements over all sample dates. Correlation between P n and daytime Ψ leaf : sapling pine r = 0.340, P = 0.003; pole pine r = 0.049, P = 0.67; old-growth pine r = 0.306, P = 0.006; and old-growth oak r = 0.303, P = Correlation between G w and predawn Ψ leaf : sapling pine r = 0.372, P = ; pole pine r = 0.120, P = 0.29; old-growth pine r = 0.314, P = 0.005; and old-growth oak r = 0.030, P = studies of co-occurring oak species (Abrams 1990). Also, Gambel oak is characterized by a massive underground system of roots, rhizomes, and lignotubers (Engle et al. 1983, Tiedemann et al. 1987). Tiedemann et al. (1987) estimated that the underground biomass per land area of Gambel oak is near the upper limit for all oaks. These characteristics of Gambel oak root systems suggest greater access to deep-soil water compared with other tree species. A stable isotope study of water uptake for Gambel oak and bigtooth maple (Acer grandidentatum Nutt.) in a canyon in Utah supports this suggestion (Phillips and Ehleringer 1995). However, in contrast Table 2. Mean and one standard error of the mean (in parentheses) of soil-to-leaf hydraulic conductance (K l, mol H 2 Om 2 s 1 MPa 1 ) for sapling, pole, and old-growth ponderosa pines in the morning (0900 h) and afternoon (1300 h) over all 1998 sample dates. Means in the same column followed by different letters differ significantly (P 0.05). Tree size Morning K l Afternoon K l Sapling pine 2.52 a (0.116) 2.41 a (0.097) Pole pine 1.87 b (0.113) 1.64 b (0.096) Old-growth pine 1.62 b (0.112) 1.78 b (0.112) to a lack of use of summer precipitation by Gambel oak in the canyon study (Phillips and Ehleringer 1995), our finding of a strong positive correlation between predawn Ψ leaf and soil water content in the surface soil (0 15 cm) for Gambel oak suggests some use of summer precipitation at our upland site in Arizona. We also found that old-growth Gambel oak was more able to tolerate atmospheric water stress than old-growth ponderosa pine. Both P n and G w declined sharply with increasing VPD for ponderosa pine in our study, as observed in other studies with this species (cf. Helms 1972, Feeney et al. 1998, Hubbard et al. 1999). In contrast, there was little relationship between P n or G w and VPD for Gambel oak. A possible explanation for this difference in response between species is that Gambel oak s wider, hairier leaves and greater leaf-level transpiration rate compared with ponderosa pine result in a thicker boundary layer for oak that uncouples water vapor pressure at the leaf surface from vapor pressure of the bulk air (Jarvis and McNaughton 1986). Thus compared with ponderosa pine, Gambel oak may have a higher value of the decoupling factor for vapor pressure deficit at the leaf surface to the air outside the boundary layer (Ω l ) (Jarvis and McNaughton 1986). Meinzer et al. (1997) have shown that detection of a relationship between G w and VPD for leaves with a thick boundary layer can depend on the reference point used to measure vapor TREE PHYSIOLOGY ON-LINE at

10 10 KOLB AND STONE Figure 8. Relationship between soil-to-leaf hydraulic conductance (K l, mol H 2 Om 2 s 1 MPa 1 ) and net photosynthetic rate (P n ) and stomatal conductance (G w ) in three sizes of ponderosa pine (sapling, pole, and old-growth). Data are pooled over all measurement times and sample dates. Correlation with K l : P n, r = 0.284, P < ; and G w, r = 0.404, P < pressure outside the leaf. A negative relationship between G w and VPD occurred only when this reference point was the leaf surface, not the air outside the boundary layer (Meinzer et al. 1997). Another explanation for the low sensitivity of P n and G w to VPD in Gambel oak is that the high predawn Ψ leaf of oak buffered the stomatal response to VPD. A study of Gambel oak by Williams and Ehleringer (1996) in a drier vegetation type (Pinus edulis Engelm. Juniperus osteosperma (Torr.) Little) than the ponderosa pine type in our study is consistent with this explanation. Based on leaf carbon isotope discrimination, Williams and Ehleringer (1996) concluded that G w of Gambel oak populations was highly sensitive to site variations in VPD. The oaks studied by Williams and Ehleringer (1996) were smaller (DBH 11 to 18 cm) than our oaks (DBH 23 to 38 cm), which suggests that our oaks had higher predawn Ψ leaf because of their deeper or more extensive roots at the moister site. Old-growth Gambel oak also showed greater tolerance of low daytime Ψ leaf than old-growth ponderosa pine, with no apparent negative effects on P n or G w. Our correlation and boundary-line analyses suggested little stomatal closure in oak at daytime Ψ leaf values as low as 2.5 MPa, whereas stomatal closure in old-growth pine started at a Ψ leaf of about 1.9 MPa. Also, oak maintained higher P n and G w than old-growth pine on several dates in July and August, despite having lower daytime Ψ leaf. Consistent with widespread patterns in leaf longevity and physiological traits (Reich et al. 1992, 1997), the shorter leaf-life span of oak was associated with greater foliar nitrogen concentration and greater maximum P n and G w compared with pine. Our second hypothesis was that P n decreases with increasing size and age of ponderosa pine in northern Arizona, and that this decrease is related to differences among tree sizes in G w, Ψ leaf, K l, or leaf nitrogen concentration. Consistent with this hypothesis, old-growth pine had lower morning and afternoon P n and G w, lower predawn and daytime Ψ leaf, and lower K l than sapling pine. Furthermore, predawn Ψ leaf, daytime Ψ leaf, and K l were positively correlated with P n for pine. These findings are generally similar to previous reports for ponderosa pine in eastern Oregon (Yoder et al. 1994, Hubbard et al. 1999). However, the differences in G w in our study between old (25-m-tall old-growth) and young (2-m-tall saplings) trees were smaller than the differences reported between old (30-m-tall) and young (10-m-tall) trees in Oregon. For example, G w in young trees in eastern Oregon was about 77% greater in the morning (1000 h) and about 150% greater in the afternoon (1300 h) than in old trees, compared with maximum seasonal differences of 50% in the morning (0900 h) and 45% in the afternoon (1300 h) in our study. Also, sensitivity of P n and G w to VPD was greater for old versus young trees in Oregon (Hubbard et al. 1999), whereas this response was similar for different-sized trees in our study. These differences between our study and the Oregon studies (Yoder et al. 1994, Hubbard et al. 1999) may be related to a difference in measurement location in the canopy of the old trees. Our measurements were made in the lower portion of the canopy (9 m above the ground), whereas measurements were made in the upper canopy (upper third of the canopy in a 33-m-tall tree) in the Oregon studies. The greater difference in sample height between young and old trees in the Oregon studies compared with our study would result in a greater difference in G w if G w is regulated by K l, and if K l increases with longer transport distance or more branch junctions (Hubbard et al. 1999). Differences in P n and G w between old and young ponderosa pines in the Oregon studies occurred in the absence of a difference in Ψ leaf (Yoder et al. 1994, Hubbard et al. 1999). In contrast, old-growth pine in our study often had lower predawn and daytime Ψ leaf than sapling pine. The magnitude of these differences (0.2 to 0.4 MPa predawn, 0.2 to 0.6 MPa morning, 0.4 to 0.5 MPa afternoon) exceeded the difference predicted based on a difference in sample height and gravity (0.08 MPa). This finding suggests that low K l of old-growth ponderosa TREE PHYSIOLOGY VOLUME 20, 2000

11 LEAF GAS EXCHANGE AND WATER RELATIONS IN A PINE OAK FOREST 11 pine may limit overnight equilibration of leaves with the water potential of the soil, as well as limiting water supply to transpiring leaves in some environments. Differences in light-saturated P n between old and young trees might also occur because of differences in leaf nitrogen concentration as a result of the dominant use of leaf nitrogen for photosynthetic enzymes (Field and Mooney 1986, Reich et al. 1997). In our study, old-growth and pole pines had greater leaf nitrogen concentrations than sapling pine, indicating that the higher light-saturated P n of saplings was probably not a result of a greater concentration of photosynthetic enzymes. Carey et al. (1998) reported that, in Nevada, the direction of differences in leaf nitrogen concentration among different-sized ponderosa pines depended on the growing environment: Nitrogen concentration increased with increasing tree size in a dry environment, but decreased in a moist environment. The higher leaf nitrogen concentration of pole and old-growth pines compared with sapling pine in our study is consistent with the pattern reported for dry sites by Carey et al. (1998). The hydraulic limitation hypothesis suggests that P n, G w, Ψ leaf, and K l of pole-sized pine in our study should have been intermediate between values for sapling and old-growth pines. Pole pine generally had intermediate values between sapling and old-growth pines for morning G w, and morning and afternoon Ψ leaf. For other physiological characteristics, values for pole pine were similar to sapling pine (predawn Ψ leaf ) and old-growth pine (morning and afternoon P n, afternoon G w, mean K l, mean leaf nitrogen concentration). Variability in tree size and density of neighboring trees was greater for pole than for sapling and old-growth pines in our study. These factors may have reduced our ability to clearly distinguish pole pine from sapling and old-growth pines in some physiological characteristics. In summary, old-growth Gambel oak was more effective at avoiding soil water stress and tolerating atmospheric water stress and low daytime Ψ leaf compared with co-occurring old-growth ponderosa pine at our study site in northern Arizona. This difference in response to water stress between species resulted in a difference in the seasonal pattern of P n between oak and pine. Although the seasonal patterns of P n, G w and Ψ leaf were generally similar among ponderosa pines that differed in size, old-growth pine had lower P n than sapling pine. This difference in P n between sapling and old-growth pines was associated with lower G w, K l, and predawn and daytime Ψ leaf, and higher leaf nitrogen concentration of old-growth pine. These findings for ponderosa pine growing in northern Arizona are consistent with the hydraulic limitation hypothesis for reduced P n of old trees, but do not support differences in leaf nitrogen concentration as an important causal mechanism. Acknowledgments We thank Jonathan L. Horton (Northern Arizona University, School of Forestry) for helpful comments on earlier versions of the manuscript, Zhong Chen (Northern Arizona University, School of Forestry) for assistance with the field measurements, Peter Z. Fulé (Northern Arizona University, School of Forestry) for the tree age dbh relationship, and the staff at the Camp Navajo Army Depot for facilitating the research. This research was supported in part by the U.S. Department of Defense, Camp Navajo Army Depot. References Abrams, M.D Adaptations and responses to drought in Quercus species of North America. Tree Physiol. 7: Abrams, M.D., M.E. Kubiske, and S.A. Mostoller Relating wet and dry year ecophysiology to leaf structure in contrasting temperate tree species. Ecology 75: Carey, E.V., R.M. Callaway and E.H. DeLucia Increased photosynthesis offsets costs of allocation to sapwood in an arid environment. Ecology 79: Chambers, J.L., T.M. Hinckley, G.S. Cox, C.L. Metcalf and R.G. Aslin Boundary-line analysis and models of leaf conductance for four oak hickory forest species. For. Sci. 31: Donovan, L.A. and J.R. Ehleringer Ecophysiological differences among juvenile and reproductive plants of several woody species. Oecologia 86: Donovan, L.A. and J.R. Ehleringer Contrasting water use patterns among size and life-history classes of a semi-arid shrub. Funct. Ecol. 6: Donovan, L.A. and R.A. Pappert Ecophysiological differences among growth stages of Quercus laevis in a sandhill oak community. J. Torrey Bot. Soc. 125:3 10. Engle, D.M., C.D. Bonham and L.E. Bartel Ecological characteristics and control of Gambel oak. J. Range Manage. 36: Feeney, S.R., T.E. Kolb, M.R. Wagner and W.W. Covington Influence of thinning and burning restoration treatments on presettlement ponderosa pines at the Gus Pearson Natural Area. Can. J. For. Res. 28: Field, C. and H.A. Mooney The photosynthesis nitrogen relationship in wild plants. In On the Economy of Plant Form and Function. Ed. T.J. Givnish. Cambridge University Press, London, U.K., pp Franco, A.C., A.G., de Soyza, R.A. Virginia, J.F. Reynolds and W.G. Whitford Effects of plant size and water relations on gas exchange and growth of the desert shrub Larrea tridentata. Oecologia 97: Fulé, P.Z., W.W. Covington and M.M. Moore Determining reference conditions for ecosystem management of southwestern ponderosa pine forests. Ecol. Appl. 7: Grulke, N.E. and P.R. Miller Changes in gas exchange characteristics during the life span of giant Sequoia implications for response to current and future concentrations of atmospheric ozone. Tree Physiol. 14: Helms, J.A Environmental control of net photosynthesis in naturally growing Pinus ponderosa Laws. Ecology 53: Hubbard, R.M., B.J. Bond and M.G. Ryan Evidence that hydraulic conductance limits photosynthesis in old Pinus ponderosa trees. Tree Physiol. 19: Jarvis, P.G. and K.G. McNaughton Stomatal control of transpiration: scaling up from leaf to region. Adv. Ecol. Res Kaufmann, M.R. and G.L. Thor Measurement of water stress in subalpine trees: Effects of temporary tissue storage methods and needle age. Can. J. For. Res. 12: TREE PHYSIOLOGY ON-LINE at

12 12 KOLB AND STONE Kolb, T.E., T.S. Fredericksen, K.C. Steiner and J.M. Skelly Scaling size and age responses to ozone in forest trees: A review. Environ. Pollut. 98: Kolb, T.E., K.M. Holmberg, M.R. Wagner and J.E. Stone Regulation of ponderosa pine foliar physiology and insect resistance mechanisms by basal area treatments. Tree Physiol. 18: Kubiske, M.E. and M.D. Abrams Stomatal and nonstomatal limitations of photosynthesis in 19 temperate tree species on contrasting sites during wet and dry years. Plant Cell Environ. 16: Kubiske, M.E. and M.D. Abrams Ecophysiological analysis of woody species in contrasting temperate communities during dry and wet years. Oecologia 98: Loewenstein, N.J. and S.G. Pallardy. 1998a. Drought tolerance, xylem sap abscisic acid and stomatal conductance during soil drying: a comparison of canopy trees of three temperate deciduous angiosperms. Tree Physiol. 18: Loewenstein, N.J. and S.G. Pallardy. 1998b. Drought tolerance, xylem sap abscisic acid and stomatal conductance during soil drying: a comparison of young plants of four temperate deciduous angiosperms. Tree Physiol. 18: Meinzer, F.C., J.L. Andrade, G. Goldstein, N.M. Holbrook, J. Cavelier and P. Jackson Control of transpiration from the upper canopy of a tropical forest: the role of stomatal, boundary layer and hydraulic architecture components. Plant Cell Environ. 20: Mencuccini, M. and J. Grace Hydraulic conductance, light interception and needle nutrient concentration in Scots pine stands and their relations with net primary productivity. Tree Physiol. 16: Miller, P.M., L.E. Eddleman and J.M. Miller Juniperus occidentalis juvenile foliage: advantages and disadvantages for a stress-tolerant, invasive conifer. Can J. For. Res. 25: Neilson, R.P. and L.H. Wullstein Comparative drought physiology and biogeography of Quercus gambelii and Quercus turbinella. Am. Midl. Nat. 114: Ni, B. and S.G. Pallardy Responses of liquid flow resistance to soil drying in seedlings of four deciduous angiosperms. Oecologia 84: Ni, B. and S.G. Pallardy Response of gas exchange to water stress in seedlings of woody angiosperms. Tree Physiol. 8:1 9. Phillips, S.L. and J.R. Ehleringer Limited uptake of summer precipitation by bigtooth maple (Acer grandidentatum Nutt) and Gambel s oak (Quercus gambelii Nutt). Trees 9: Reich, R.B., M.B. Walters and D.S. Ellsworth Leaf life-span in relation to leaf, plant, and stand characteristics among diverse ecosysytems. Ecol. Monogr. 62: Reich, R.B., M.B. Walters and D.S. Ellsworth From tropics to tundra: global convergence in plant functioning. Proc. Natl. Acad. Sci. USA 94: Rundel, P.W. and W.M. Jarrell Water in the environment. In Plant Physiological Ecology Today Field Methods and Instrumentation. Eds. R.W. Pearcy, J. Ehleringer, H.A. Mooney, and P.W. Rundel. Chapman and Hall, New York, pp Ryan, M.G. and B.J. Yoder Hydraulic limits to tree height and tree growth. BioScience 47: Ryan, M.G., D. Binkley and J.H. Fownes Age-related decline in forest productivity: pattern and process. Adv. Ecol. Res. 27: Samuelson, L.J. and J.M. Kelly Ozone uptake in Prunus serotina, Acer rubrum, and Quercus rubra forest trees of different sizes. New Phytol. 136: Samuelson, L.J., J.M. Kelly, P.A. Mays and G.S. Edwards Growth and nutrition of Quercus rubra L. seedlings and mature trees after three seasons of ozone exposure. Environ. Pollut. 91: Stone, J.E., T. E. Kolb and W.W. Covington Effects of restoration thinning on presettlement Pinus ponderosa in Northern Arizona. Restoration Ecol. 7: Svenson, S.E. and F.T. Davies, Jr Comparison of methods for estimated surface area of water-stressed and fully hydrated pine needle segments for gas exchange analysis. Tree Physiol. 10: Tiedemann, A.R., W.P. Clary, and R.J. Barbour Underground systems of Gambel oak (Quercus gambelii) in central Utah. Am. J. Bot. 74: Western Regional Climatic Data Center sage.dri.edu/index.html. Williams, D.G. and J.R. Ehleringer Carbon isotope discrimination in three semi-arid woodland species along a monsoon gradient. Oecologia 106: Yoder, B.J., M.G. Ryan, R.H. Waring, A.W. Schoettle and M.R. Kaufmann Evidence of reduced photosynthetic rates in old trees. For. Sci. 40: TREE PHYSIOLOGY VOLUME 20, 2000