Effects of elevated CO2 and temperature on cold hardiness and spring bud burst and growth in Douglas-fir (Pseudotsuga menziesii)

Tree Physiology 18, 671--679 1998 Heron Publishing----Victoria, Canada Effects of elevated CO2 and temperature on cold hardiness and spring bud burst and growth in Douglas-fir (Pseudotsuga menziesii) SUNGHEE GUAK, 1 DAVID M. OLSYZK, 2 LESLIE H. FUCHIGAMI 1,3 and DAVID T. TINGEY 2 1 Department of Horticulture, Oregon State University, Corvallis, OR 97331, USA 2 US EPA, National Health and Environmental Effects Research Laboratory, Western Ecology Division, Corvallis, OR 97333, USA 3 Author to whom correspondence should be addressed Received July 18, 1997 Summary We examined effects of elevated CO 2 and temperature on cold hardiness and bud burst of Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) seedlings. Two-yearold seedlings were grown for 2.5 years in semi-closed, sunlit chambers at either ambient or elevated (ambient + 4 C) air temperature in the presence of an ambient or elevated (ambient + 200 ppm) CO 2 concentration. The elevated temperature treatment delayed needle cold hardening in the autumn and slowed dehardening in the spring. At maximum hardiness, trees in the elevated temperature treatment were less hardy by about 7 C than trees in the ambient temperature treatment. In general, trees exposed to elevated CO 2 were slightly less hardy during hardening and dehardening than trees exposed to ambient CO 2. For trees in the elevated temperature treatments, date to 30% burst of branch terminal buds was advanced by about 6 and 15 days in the presence of elevated CO 2 and ambient CO 2, respectively. After bud burst started, however, the rate of increase in % bud burst was slower in the elevated temperature treatments than in the ambient temperature treatments. Time of bud burst was more synchronous and bud burst was completed within a shorter period in trees at ambient temperature (with and without elevated CO 2 ) than in trees at elevated temperature. Exposure to elevated temperature reduced final % bud burst of both leader and branch terminal buds and reduced growth of the leader shoot. We conclude that climatic warming will influence the physiological processes of dormancy and cold hardiness development in Douglas-fir growing in the relatively mild temperate region of western Oregon, reducing bud burst and shoot growth. Keywords: chill days, chilling requirements, climate change, climatic warming, cold hardening, dehardening, elevated carbon dioxide, thermal time, visible injury. Introduction Atmospheric concentrations of CO 2, the major greenhouse gas, are predicted to continue rising by 1.5 ppm per year (Houghton et al. 1995). The CO 2 -induced climatic warming is expected to increase global mean surface air temperatures by about 0.8 C by 2000 and by 3 to 4 C by 2100 (Carson 1996). This increase in air temperature is likely to influence the timing of bud burst in trees by altering the amounts of chilling and forcing temperatures required to induce bud burst (Cannell and Smith 1984, Kramer 1994). Some simulations indicate that climatic warming in cool regions will advance bud burst in Scots pine (Hänninen 1991), Sitka spruce (Cannell and Smith 1984, Murray et al. 1994), and apple (Cannell and Smith 1986). Early bud burst can result in frost damage, whereas late bud burst can reduce productivity (Weiser 1970, Sakai and Larcher 1987). Climatic warming may also alter the time and degree of hardening and dehardening so that trees might remain less hardy than normal during winter (Kettunen et al. 1987, Repo et al. 1996). Increased atmospheric CO 2 concentrations may also affect bud phenology (Waring 1969, Cannell 1990, Murray et al. 1994) and cold hardiness (Murray et al. 1994, Tinus et al. 1995). For example, elevated CO 2 delayed bud burst and advanced bud set of Sitka spruce (Murray et al. 1994), and increased autumn and spring cold hardiness of Douglas-fir, although winter hardiness was decreased (Tinus et al. 1995). In conjunction with climatic warming, elevated CO 2 reduced the risk of autumn and spring cold damage in Sitka spruce in Scotland (Murray et al. 1994). In contrast, elevated CO 2 negatively affected cold hardiness development in black spruce (Margolis and Vezina 1990). Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco), an important forest species in the Pacific Northwest, has three dormant states (correlative inhibition, rest, and quiescence) and has a chilling requirement of about three months, normally satisfied by mid-january (Lavender and Stafford 1984), for the rest release needed for vigorous shoot growth (Lavender 1981). The current climate of the Willamette Valley in western Oregon is relatively mild (over 200 frost-free days per year), with a fairly large number of hours with temperatures near 5 C, the temperature at which chilling requirements of perennial species are satisfied most efficiently (Lavender and Stafford 1984). It has been suggested that climatic warming could reduce the growth of Douglas-fir in mild temperate regions

672 GUAK, OLSYZK, FUCHIGAMI AND TINGEY because chilling would be insufficient (Kimmins and Lavender 1987, McCreary et al. 1990). The objective of this study was to determine how long-term exposure of Douglas-fir seedlings to climate change (i.e., elevated CO 2 and temperature) affects cold hardening and dehardening, and influences the timing and rate of bud burst in a relatively mild temperate region. Materials and methods Chambers Chambers, located at the EPA, Corvallis, OR, consisted of aluminum frames covered with 3-mm, clear Teflon film (Du Pont Electronics, Wilmington, DE) (Tingey et al. 1996). Each chamber was 2.0 m wide, 1.0 m front-to-back, and 1.5 m tall at the back sloping to 1.3 m at the front. It was possible to control and manipulate climatic and edaphic factors in the sunlit chambers while maintaining typical environmental variability (Tingey et al. 1996). Attached to the back of each chamber was a steam humidifier and an aluminum air handler containing air blowers, a cold water heat exchanger, and electric strip heaters. Each aboveground chamber was attached to a watertight soil lysimeter set below ground level. Lysimeters were filled with a coarse, loamy, mixed frigid (Typic Hapludand) soil reconstituted from A, B, and C horizons, and topped with 6 cm of forest floor litter shortly after planting. Trees Two-year-old Douglas-fir seedlings were provided by the Weyerhaeuser Company in Aurora, OR, from seed lots representing five low-elevation (< 500 m) seed zones in the southern Willamette Valley and the Cascade Mountains. The trees averaged 0.63 cm in diameter and 41 cm in height when planted in the chambers on June 7--8, 1993. Fourteen trees were planted in each chamber, in three rows of five, four, and five trees spaced 36.5 cm between rows and 39.5 cm within rows. Trees were watered with reverse-osmosis-filtered water to mimic seasonally varying soil water contents in the area (Tingey et al. 1996). No fertilizer was added to the soil to insure that system nutrients were provided only by soil biological processes. Control trees were planted outside the chambers at two natural sites: in open-air lysimeters at the EPA in Corvallis, OR; and in field plots at 537 m elevation in the Oregon Cascades at 44 23 30 N, 122 22 30 W. Trees at both control sites were planted in the same type of soil and in the same planting pattern as in the chambers. The trees in the open-air lysimeters were planted in the same type of containers as the chambered trees, whereas trees in the field plots were planted directly in the ground. Climate control and treatments The chamber experiment was based on a factorial design with two CO 2 concentrations (ambient and elevated, i.e., ambient + 200 ppm) and two temperatures (ambient and elevated, i.e., ambient + 4 C). There were three replicate chambers (n = 3) for each chamber treatment. The control sets of trees consisted of replicate open-air lysimeters (n = 2) and replicate field plots (n = 3). Ambient CO 2, air temperature, and dew point were monitored continuously near the chambers and the values used to regulate chamber CO 2, air temperature, and vapor pressure deficit (VDP) (Tingey et al. 1996). All chambers were maintained at ambient temperature and CO 2 concentration from planting until the elevated temperature and CO 2 treatments began on July 23 and August 17, 1993, respectively. The chambers effectively simulated natural seasonal and diurnal changes in atmospheric CO 2, air and soil temperatures, VPD, and soil water content (Tingey et al. 1996). Hourly air temperature data were used to calculate daily maximum, minimum, and mean temperatures for all treatments from October 1, 1995 to May 31, 1996 (Figure 1). When cold hardiness measurements were begun in October 1995, stem diameters of the chambered trees were similar in all treatments and averaged 1.88 ± 0.10 cm (± SE). Mean tree heights were 63.0 ± 6.2 cm and 78.6 ± 4.5 cm in the elevated and ambient temperature treatments, respectively. Mean stem diameter and height of control trees in the open-air lysimeters Figure 1. Mean, maximum, and minimum daily temperatures for open-air control lysimeters (OAC), ambient temperature treatment chambers (AT), and elevated temperature treatment chambers (ET) from October 1, 1995 to May 31, 1996. Arrows by daily mean values indicate sampling time of Douglas-fir seedlings at Corvallis for cold hardiness tests. TREE PHYSIOLOGY VOLUME 18, 1998

CARBON DIOXIDE, TEMPERATURE, COLD HARDINESS AND BUD BURST 673 were 2.17 ± 0.30 cm and 86.2 ± 4.6 cm, respectively; and the corresponding values for the field-grown trees were 2.33 ± 0.16 cm and 110.5 ± 8.3 cm. Sample preparation Cold hardiness was determined by scoring visible injury caused by artificially freezing current-year needles. Needles were collected from the chambers and open-air lysimeters on October 20 and December 4, 1995; and January 17, February 27, and April 2, 1996. Field samples were taken two to five days after the chamber samples. Similar procedures for sample collection and preparation, freezing test, and scoring of visible injury were used at all sampling locations. Seven intact current-year needles were collected from the middle portion of the crown of each of 14 trees (total of 98 needles) per replicate. The needles were pooled, sealed in a plastic bag, and immediately placed in an insulated box with ice for transportation to the laboratory for freezing tests. The composite sample was then rinsed twice with deionized water and divided into six subsamples of 15 needles each (one subsample for each testing temperature). The 15 needles for each treatment were randomly placed in one of five lanes (one lane per treatment, excluding the field-grown treatment) in a double-layer of moistened cheesecloth, then wrapped in aluminum foil (13 18 cm envelope). The edges of the foil were folded tightly to prevent possible dehydration of the needles during the freezing tests. All samples were kept at 4 C until the freezing test began. The average time elapsed from collecting the samples to starting the freezing test was 5 to 6 h. Buds and stems were freeze tested once in January 1996. Three current-year branches (5--10 cm long) per tree were sampled from the mid-crown and pooled to give a composite sample of 42 branches per replicate. The branches were divided into six subsamples, one for each temperature. After needles and buds (axillary) were removed, the remaining branches were cut into 1-cm-long sections after discarding about 2 cm at the terminal end of each branch. A subsample consisting of 12 buds and 12 stem pieces was placed in the lane adjacent to the corresponding needle subsample on moistened cheesecloth and wrapped in aluminum foil as described for the needle subsamples. Cold hardiness assessment Freezing was done in a freezer (Model 8770, Forma Scientific, Inc., Marietta, OH) regulated with a programer/controller (Model WEST 3750, WEST Instruments, East Greenwich, RI). Six freezing temperatures were used (five in October) to bracket the temperature ( C) at which 50% of the tissue is damaged (L t50 ). The range of temperatures varied with sampling date and always had two extremes, one for inducing no damage and the other for inducing 100% damage. This range was determined by preliminary freezing tests conducted one week before each main test. The samples wrapped in aluminum foil were placed on an aluminum plate (10-mm thick) in the freezer and incubated at 2 C overnight. The freezing rate was 3 C h 1 with samples held for 30 min at each test temperature before being removed from the freezer. The samples, which were left in the aluminum foil and moist cheesecloth envelopes, were then thawed at 4 C overnight followed by incubation for 10 days at 22--23 C. Needle samples were scored for visible injury according to the proportion of damaged area. All scores were made at 25% injury intervals. Injury of stem samples was scored according to the discoloration and disintegration of the cambium and wood tissues. Bud samples were cut longitudinally and injury to preformed shoots and bud scales assessed. To estimate cold hardiness (L t50 ), nonlinear regression (PeakFit software, v. 1.0, Jandel Scientific, San Rafael, CA), was used to fit a logistic sigmoid function (Equation 1) (Repo et al. 1996) to the data for all treatments: y = a 0 1 + e [ (x a 1)/a 2 ], (1) where y is the fitted % visible injury, x is the exposure temperature, a 0 is a constant that determines the asymptotes of the function, and a 2 is the slope of the function at the inflection point a 1. Cold hardiness (L t50 ) was interpolated from the point of 50% visible injury at the freezing-test temperature. Bud burst and growth Leader buds (14 trees per replication) were examined on March 4 and weekly from March 18 until June 3, 1996. Branch terminal buds (four trees per replication) were monitored at 3-day intervals from mid-february to mid-may. Time of bud burst was determined as the date that green needles emerged 2 mm through scales from leader buds and branch terminal buds. For leader buds, time of bud burst was determined from the mean of those buds that had burst by June 3, 1996, whereas for the branch terminal buds, mean date to 30% bud burst (DBB 30% ) was determined from regressions of % bud burst and time. New shoot growth was measured on June 3, 1996. Chill days and thermal time To determine the relationship between chill days and thermal time received to DBB 30%, the number of chill days to DBB 30% was defined as the accumulated number of days since October 1 when the mean air temperature was at or below 5 C. Thermal time required to DBB 30% was the accumulated degree days above the mean daily base temperature of 5 C from December 1 (cf. van den Driessche 1975, Lavender 1981). Statistical analysis Analysis of variance was used to determine significance of treatment effects (CO 2 and temperature) from the 2 2 factorial experiment, with a separate statistical analysis for each date. The effect of the chambers on tree response was determined with an unpaired t-test between chambered control trees and trees in the open-air lysimeters; location effect was determined with an unpaired t-test between trees in the open-air lysimeters and field-grown trees. TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com

674 GUAK, OLSYZK, FUCHIGAMI AND TINGEY Results Cold hardiness Figure 2 shows the accumulated thermal time (December to May) and chill days (October to May) for the open-air lysimeters and the treatment chambers. Elevated temperatures significantly delayed the start of cold hardening and decreased the rate of cold hardening in autumn and delayed the start of dehardening in spring, and the effects were more pronounced in ambient CO 2 than in elevated CO 2 (Table 1 and Figure 3). Hardening in the elevated temperature treatments proceeded at a mean rate of 0.15 C day 1 to maximum hardiness in mid- January compared with a mean rate of 0.24 C day 1 in the ambient temperature treatments (Figure 3). The elevated temperature treatments delayed hardening to 20 C by about 24 days. In mid-january, plants in the ambient temperature treatments were more hardy by a maximum of about 7 C than plants in the elevated temperature treatments. During the period from January 17 to Februry 27, dehardening occurred more slowly in the elevated temperature treatments than in the ambient temperature treatments (0.02 versus 0.18 C day 1 ; Figure 3). Thereafter, the dehardening rates in the elevated and ambient temperature treatments increased to 0.27 and 0.41 C day 1, respectively, until minimum cold hardiness was reached in early April. The difference in dehardening rates resulted in a significantly lower (more hardy) L t50 for trees in the elevated temperature treatment ( 13 C) compared with that for trees in the ambient temperature treatment ( 8 C, P < 0.001). Although trees in the open-air lysimeters and field-grown trees eventually reached a similar degree of maximum cold hardiness in January, the hardening rate from October 20 to December 4 was higher for field-grown trees than for trees in open-air lysimeters (Figure 3). The difference in hardening rate corresponded to the lower daily mean temperature for field-grown trees during the period (data not shown). The rate of dehardening was similar for both sets of control trees. The changing slope and temperature extremes of the visible injury curves over time (Figure 4) indicate changing sensitivity of needles to freezing stress. These curves also indicate differences in the rate of hardening and dehardening among treatments. For example, in January, when maximum hardiness was attained, the visible injury curves for needles in the elevated temperature treatments were significantly shifted to the left (as indicated by the arrows at L t50 ) relative to the curves for needles in the ambient temperature treatments, indicating that elevated temperature reduced needle cold hardiness. In contrast, in early April, visible injury curves for needles in the elevated temperature treatments were shifted to the right compared with the curves for needles in the ambient temperature treatments, indicating that warming delayed the dehardening process in the spring. In contrast, bud and stem cold hardiness, determined in mid-january, were not significantly affected by any of the Figure 2. (a) Cumulated thermal time (degree days) above 5 C and (b) number of chill days at or below 5 C (mean daily temperature) for open-air control lysimeters (OAC), ambient temperature treament chambers (AT) and elevated temperature treatment chambers (ET) at Corvallis. Table 1. Analysis of variance of effects of CO 2 and temperature on needle cold hardiness (L t50 ) of Douglas-fir. Replicates: n = 3 for all of the chamber treatments and n = 2 for the open-air control lysimeters. Abbreviations: ACAT, ambient CO 2 + ambient temperature (chamber control); OAC, open-air control lysimeter; and FIELD, field-grown controls. P values of L t50 Effect October 20, 1995 December 4, 1995 January 17, 1996 February 27, 1996 April 2, 1996 CO 2 0.589 0.100 0.250 0.006 0.018 Temperature (T) 0.113 0.047 0.018 0.071 0.001 CO 2 T 0.401 0.196 0.703 0.248 0.156 ACAT versus OAC 0.940 0.967 0.758 0.125 0.030 OAC versus FIELD 0.197 0.068 0.461 0.256 0.020 TREE PHYSIOLOGY VOLUME 18, 1998

CARBON DIOXIDE, TEMPERATURE, COLD HARDINESS AND BUD BURST 675 Figure 3. Effects of elevated CO 2 and temperature on cold hardiness of Douglas-fir needle tissue as determined by visible injury (L t50 ; mean ± standard error). Treatments: ( ) ambient CO 2 + ambient temperature (chamber control, ACAT n = 3), ( ) elevated CO 2 + ambient temperature (ECAT, n = 3), ( ) ambient CO 2 + elevated temperature (ACET, n = 3), ( ) elevated CO 2 + elevated temperature (ECET, n = 3), and ( ) open-air control lysimeters (OAC, n = 2), and ( ) field-grown controls (FIELD, n = 3). treatments, although L t50 of buds and stems tended to be lower (more hardy) in the ambient temperature treatments than in the elevated temperature treatments (Table 2). In the ambient temperature treatments, mean stem L t50 ( 29.8 C) was similar to needle L t50 ( 29.3 C) and both tissues were about 7 C more hardy than buds ( 22.8 C). In the elevated temperature treatments, the L t50 values of the tissues was in the order: stem L t50 ( 28.2 C) < needle ( 24.5 C) < bud ( 21.7 C), indicating that climatic warming might have a greater effect on needle hardiness than on stem or bud hardiness. Bud burst and growth Elevated temperature advanced DBB 30% of branch terminal buds (Table 3 and Figure 5). Elevated CO 2 in combination with elevated temperature delayed DBB 30% by about 10 days compared to that of trees in the ambient CO 2 + elevated temperature treatment. Similar enhancement of bud burst in response to elevated temperature was also observed for the leader buds; however, the elevated CO 2 treatment had no effect on the timing of burst of the leader buds. Mean dates to bud burst of the leader buds were April 18, 16, 12, 10 and 18 in the ambient temperature + ambient CO 2, ambient temperature + elevated CO 2, elevated temperature + ambient CO 2, elevated temperature + elevated CO 2 and open-air lysimeter treatments, respectively (P = 0.226). Once bud burst commenced, the rate of increase in % bud burst was greater in the ambient temperature treatments than in the elevated temperature treatments (Figure 5). Consequently, by early June, most of the leader buds in the ambient temperature treatments had burst, whereas only about 75% of leader buds had burst in the elevated temperature treatments Figure 4. Effects of elevated CO 2 and temperature on cold hardening and dehardening of Douglas-fir, as indicated by changes in the model of visible injury as a function of stress temperature at about 40-day intervals. The black dots (b) represent L t50 values for three sampling dates. The L t50 value for the ambient CO 2 + ambient temperature treatment (ACAT) is compared with the other treatments for each sampling date. The arrows indicate a statistically significant difference from the value for the ACAT control for each sampling date. Arrows to the right and left at 50% injury represent more hardy and less hardy than the ACAT control, respectively. Treatment abbreviations are as in Figure 3. (Figure 5). Many unbroken buds in the elevated temperature treatments were aborted or did not break until very late in the 1996 growing season (data not shown). New shoot growth of trees in the ambient temperature treatments was about twice that of trees in the elevated temperature treatments, regardless of CO 2 treatment (Figure 6). During the previous two seasons, mean height growth had increased by about 54 and 92% of the initial height in the elevated and ambient temperature treatments, respectively. Stem diameter increased about 198% in all of the treatments during the study period. Thermal time and chill days There was a nonlinear relationship between chill days and thermal time (accumulated degree days > 5 C) received to DBB 30% (Figure 7a). Elevated temperature treatment increased the daily mean air temperature on the mean date of 30% bud TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com

676 GUAK, OLSYZK, FUCHIGAMI AND TINGEY Table 2. Effects of elevated CO 2 and temperature on cold hardiness (L t50, mean ± SE) of needle, bud, and stem tissues of Douglas-fir determined on January 17, 1996, and analysis of variance for effects. Treatment abbreviations are as in Figure 3. Cold hardiness (L t50, C) Needle Bud Stem Treatment Mean (SE) ACAT 29.95 (1.98) 22.30 (1.41) 30.65 (0.79) ECAT 27.60 (1.65) 23.08 (0.94) 29.05 (0.74) ACET 25.15 (0.60) 22.15 (0.31) 28.45 (1.43) ECET 23.94 (1.11) 21.19 (0.66) 27.91 (1.11) OAC 30.96 (2.13) 23.00 (0.48) 29.73 (1.19) FIELD 29.47 (0.69) 23.49 (0.07) 26.98 (0.69) Effect P values CO 2 0.250 0.922 0.338 Temperature (T) 0.018 0.300 0.151 CO 2 T 0.703 0.375 0.628 ACAT versus OAC 0.758 0.806 0.541 OAC versus FIELD 0.461 0.862 0.117 burst by about 2.5 C (Figure 7b). As a result of the chamber effect, trees in the ambient temperature + ambient CO 2 treatment received about 7 fewer chill days than trees in the openair lysimeters, but there was no difference in thermal time received to DBB 30%. Discussion The elevated temperature treatments disturbed cold hardening and dehardening of Douglas-fir in western Oregon, resulting in acquisition of a lower degree of cold hardiness than at ambient temperatures. This response might be caused by an insufficient period or intensity of low temperature required by seedlings to complete the process of cold hardening in the autumn and winter. Maximum hardiness is obtained when three stages of hardening have been satisfied (Howell and Weiser 1970, Fuchigami et al. 1971). The first stage (early initiation of hardening) is associated with a shortening of the photoperiod and is favored by mild temperatures; the second stage, during which cold hardiness increases greatly, is in- Table 3. Effects of elevated CO 2 and temperature on mean dates to 20, 30, 40 and 50% bud burst of the branch terminal buds of Douglas-fir, and analysis of variance for effects. Treatment abbreviations are as in Figure 3. Treatment/Effect Mean no. Mean dates (± SE) to % bud burst Figure 5. Mean accumulated % bud burst of the leader bud and branch terminal buds of Douglas-fir in the ambient and elevated temperature treatments in the presence of ambient or elevated CO 2. Values are the mean % bud burst of three replicates for each chamber treatment, and two replicates for the trees in the open-air lysimeteres (OAC) (each replicate consisted of 14 and 4 trees for the leader buds and branch terminal buds, respectively). Treatment abbreviations are as in Figure 3. buds (± SE) 20% 30% 40% 50% Treatment ACAT 169 (13) 04/12 (3.7) 04/16 (3.8) 04/19 (4.1) 04/24 (5.1) ECAT 184 (23) 04/14 (1.1) 04/17 (1.0) 04/20 (1.1) 04/22 (1.3) ACET 150 (10) 03/27 (4.1) 04/01 (5.1) 04/04 (4.9) 04/09 (8.5) ECET 175 (16) 04/06 (2.3) 04/11 (1.2) 04/16 (0.4) 04/22 (3.8) OAC 255 (16) 04/13 (1.2) 04/16 (1.7) 04/18 (2.1) 04/21 (2.4) Effect P values CO 2 0.285 0.288 0.195 0.456 Temperature (T) 0.006 0.020 0.037 0.474 CO 2 T 0.408 0.398 0.239 0.360 ACAT versus OAC 0.870 0.901 0.903 0.792 TREE PHYSIOLOGY VOLUME 18, 1998

CARBON DIOXIDE, TEMPERATURE, COLD HARDINESS AND BUD BURST 677 Figure 6. Effects of elevated CO 2 and temperature on mean (± SE) shoot length of Douglas-fir leaders on June 3, 1996 for the leader buds that had burst. There were three replicates for each chamber treatment, except n = 2 for the ambient CO 2 + ambient temperature treatment because of aphid infestation; each replicate consisted of 14 trees. The effect of temperature across CO 2 concentrations was significant (P = 0.002). Treatment abbreviations are as in Figure 3. duced by low temperatures; and the third stage (attainment of maximum hardiness) is usually triggered by prolonged exposure to very low temperatures ( 30 to 60 C). However, some species, including Rocky Mountain Douglas-fir, are capable of being hardened to very low temperatures without being exposed to temperatures much below freezing (Burr et al. 1989). Therefore, in our study, the decreased maximum hardiness at elevated temperature suggests that warming may have disturbed the second stage of cold hardening, because the first stage is mainly mediated by photoperiod and the third stage rarely exists in western Oregon. In the elevated temperature + ambient CO 2 treatment, we observed a lag phase (December 20 to February 27) during which little dehardening took place, followed by a period of rapid dehardening. According to the degree growth stage ( GS) model of Fuchigami et al. (1982), the rate of dehardening increases gradually when plants change from the state of maximum rest toward the stage of spring bud break. Therefore, it is likely that the warming-induced lag phase of dehardening might be related to a delay in rest release as a result of elevated temperature. We found strong evidence that trees were not fully chilled in the elevated temperature treatments. For example, there was a significant reduction in leader shoot growth and a slowed increase in the accumulated % bud burst after bud burst commenced. There were also decreases in final % bud burst and number of chill days (approximately 25% of that at ambient temperature) by mid-february, the time when the chilling requirements of Douglas-fir growing in the mild temperate region are normally satisfied (Lavender and Stafford 1984). Silim and Lavender (1994) also noted that the ability of shoots of white spruce to deharden was related more closely to chilling requirements than to exposure to warm temperatures. Comparison between the rates of dehardening and bud burst (e.g., rate of increase in % bud burst) suggests that, in mild temperate regions, the rate of dehardening was more closely related to the rate of bud burst than to the timing of bud burst, because the rates of both dehardening and bud burst were dependent on the degree to which trees were chilled. In contrast, in boreal regions, the earlier loss of cold hardiness at elevated temperatures led to an earlier growth commencement (Hänninen 1991, Repo et al. 1996). This synchronization between the start of dehardening and the beginning of growth (time of bud burst) is primarily a result of increased spring temperatures when the chilling requirements of the trees are fully met. Thus, when conducting climatic warming studies, one needs to be cautious when trying to correlate the start of dehardening to growth commencement because any correlation between the two may depend on the degree to which trees are chilled. In Douglas-fir, vigorous shoot elongation occurs in the spring provided that bud dormancy induction or development is not disturbed and the trees have been fully chilled at temperatures near 5 C during winter (van den Driessche 1975, Figure 7. (a) The relationship between thermal time (accumulated degree days above 5 C) from December 1 to the mean date to 30% bud burst (DBB 30% ) of branch terminal buds and number of chilling days at or below 5 C from October 1 in Douglas-fir seedlings growing in western Oregon; and (b) DBB 30% of branch terminal buds and temperatures on the mean date to 30% bud burst (DBB 30% ) for the ambient CO 2 + ambient temperature, ambient CO 2 + elevated temperature and the open-air lysimeter treatments. Thermal time and number of chill days to DBB 30% were calculated from daily mean air temperatures. Mean temperature increase in the spring was depicted by connecting two mean temperatures (two straight lines). Treatment abbreviations are as in Figure 3. TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com

678 GUAK, OLSYZK, FUCHIGAMI AND TINGEY Lavender 1981). In the present study, the poor new shoot growth in the elevated temperature treatments (only 50% of that at ambient temperatures) suggests that warming adversely affected the endogenous dormancy sequence of Douglas-fir, possibly as a result of increased autumn air and soil temperatures. Reduced shoot growth in the elevated temperature treatments was also observed during the previous two seasons (authors unpublished observations). This result supports the findings of McCreary et al. (1990) that growth of Douglas-fir is limited when seedlings are grown at temperatures above 7 C during winter. As shown by the nonlinear relationship between thermal time to bud burst and the duration of previous chilling (Figure 7a), there was a large increase in the thermal time to bud burst with a decrease in the number of chill days in the elevated temperature treatments that probably resulted in poorly chilled trees (Murray et al. 1989). In such a case, most simulation models predict that warming will not bring about earlier bud burst (Cannell and Smith 1983, Murray et al. 1989). In our study, however, the elevated temperature treatment hastened bud burst, and this effect was more evident for branch terminal buds than for leader buds. This enhanced bud burst at elevated temperatures might have resulted from warming-forced earlier bud development on abnormally (or normally) formed winter resting buds that had received little chilling. Such forced early bud development results in poor, abnormal shoot growth in the spring because of the unsatisfied chilling requirement. Aborted buds (i.e., no bud scale, leaf, and embryonic shoot) have been found among trees growing at elevated temperatures (M. Apple, EPA, Corvallis, personal communication). In addition to enhancing bud burst, the elevated temperature treatment slowed the rate of increase in % bud burst after it had commenced, resulting in erratic bud burst among and within trees, compared to the more synchronous bud burst observed in trees at ambient temperatures. Myking and Heide (1995) also reported erratic bud burst in European white birch seedlings exposed to temperatures above 15 C. In general, elevated CO 2 concentrations negatively affected hardiness during cold hardening and dehardening. Similarly, Margolis and Vezina (1990) observed a negative effect of elevated CO 2 and temperature on the development of cold hardiness in buds of black spruce. In contrast, Tinus et al. (1995) found that high CO 2 concentrations (700 ppm) increased autumn and spring hardiness of Rocky Mountain Douglas-fir, and Repo et al. (1996) speculated that enhanced hardening in response to elevated CO 2 plus temperature could be caused by an increased cellular content of energy-rich compounds, such as carbohydrates. Elevated CO 2 in combination with elevated temperature delayed the time of bud burst of branch terminal buds, compared to that of trees in the ambient CO 2 + elevated temperature treatment. Murray et al. (1994) observed a similar effect in Sitka spruce in Scotland and concluded that increasing atmospheric CO 2 concentrations in conjunction with climatic warming was likely to improve survival from spring and autumn frosts by enhancing bud set in the autumn and delaying bud burst in the spring. Elevated CO 2 in combination with elevated temperature might have partially counteracted the earlier bud burst brought about by elevated temperatures. We predict that a major effect of climatic warming on Douglas-fir trees in mild temperate regions will reflect its detrimental effects on dormancy and cold hardiness development, which will, in turn, affect shoot and bud growth. However, risk of frost damage in the spring as a result of advanced bud burst will not increase in this particular climate. Acknowledgment The authors thank Drs. Richard Tinus and William Proebsting for valuable comments on the manuscript and Pricilla Licht for editorial assistance. The work reported this article was funded in part by the U.S. Environmental Protection Agency under purchase order 5B1136NTX to Oregon State University. It has been subjected to the Agency s peer and administrative review and approved for publication as an EPA document. References Burr, K.E., R.W. Tinus, S.J. Wallner and R.M. King. 1989. Relationships among cold hardiness, root growth potential and bud rest in three conifers. Tree Physiol. 5:291--306. Cannell, M.G.R. 1990. Modeling the phenology of trees. Silva Carel. 15:11--27. Cannell, M.G.R. and R.I. Smith. 1983. Thermal time, chill days and prediction of budburst in Picea sitchensis. J. Appl. Ecol. 20:951-- 963. Cannell, M.G.R. and R.I. Smith. 1984. Spring cold damage on young Picea sitchensis. II. Predicted dates of bud burst and probability of cold damage. Forestry 57:177--197. Cannell, M.G.R. and R.I. Smith. 1986. Climatic warming, spring bud burst and cold damage on trees. J. Appl. Ecol. 23:177--191. Carson, D.J. 1996. Modeling climate change: achievements and prospects. Commonw. For. Rev. 75:12--18. Fuchigami, L.H., C.J. Weiser and D.R. Evert. 1971. Induction of cold hardening in Cornus stolonifera Michx. Plant Physiol. 47:98--103. Fuchigami, L.H., C.J. Weiser, K. Kobayashi, R. Timmis and L.V. Gusta. 1982. A degree growth stage ( GS) model and cold hardening in temperate woody plants. In Plant Cold Hardiness and Freezing Stress. Mechanisms and Crop Implications. Eds. P.H. Li and A. Sakai. Academic Press, New York, pp 93--116. Hänninen, H. 1991. Does climatic warming increase the risk of cold damage in northern trees? Plant Cell Environ. 14:449--454. Houghton, J.T., L.G. Meira Filho, B.A. Callander, N. Harris, A. Kattenberg and K. Maskell. 1995. Climatic change 1995. In The Science of Climate Change. Cambridge University Press, Cambridge, 572 p. Howell, G.S. and C.J. Weiser. 1970. Fluctuations in cold resistance of apple twigs during spring dehardening. J. Am. Soc. Hort. Sci. 95:190--192. Kettunen, L., J. Mukula, V. Pohjonen, O. Ratanen and U. Varjo. 1987. The effects of climatic variations on agriculture in Finland. In The Impact of Climatic Variations on Agriculture, Vol. 1: Assessment in Cool Temperate and Cold Regions. Eds. M.L. Parry, T.R. Carter and N.T. Konijn. Kluwer Academic Publishers, Dordrecht, pp 513--614. TREE PHYSIOLOGY VOLUME 18, 1998

CARBON DIOXIDE, TEMPERATURE, COLD HARDINESS AND BUD BURST 679 Kimmins, J.P. and D.P. Lavender. 1987. Implications of climate change for the distribution of biogeoclimatic zones in British Columbia and for the growth of temperate forest species. In Woody Plant Growth in a Changing Chemical and Physical Environment. Ed. D.P. Lavender. University of British Columbia Press, Vancouver, B.C., pp 209--309. Kramer, K. 1994. A modeling analysis of the effects of climatic warming on the probability of spring frost damage to tree species in the Netherlands and Germany. Plant Cell Environ. 17:367--377. Lavender, D.P. 1981. Environment and shoot growth of woody plants. Research Paper No. 45, Forest Research Laboratory. Oregon State University, Corvallis, OR, 47 p. Lavender, D.P. and S.G. Stafford. 1984. Douglas-fir seedlings: some factors affecting chilling requirement, bud activity, and new foliage production. Can. J. For. Res. 15:309--312. Margolis, H.A. and L.-P. Vezina. 1990. Atmospheric CO 2 enrichment and the development of frost hardiness in containered black spruce seedlings. Can. J. For. Res. 20:392--1398. McCreary, D.D., D.P. Lavender and R.K. Hermann. 1990. Predicted global warming and Douglas-fir chilling temperatures. Ann. Sci. For. 47:325--330. Murray, M.B., M.G.R. Cannell and R.I. Smith. 1989. Date of budburst of fifteen tree species in Britain following climatic warming. J. Appl. Ecol. 26:693--700. Murray, M.B., R.I. Smith, I.D. Leith, D. Fowler, H.S.J. Lee, A.D. Friend and P.G. Jarvis. 1994. Effects of elevated CO 2, nutrition and climatic warming on bud phenology in Sitka spruce (Picea sitchensis) and their impact on the risk of cold damage. Tree Physiol. 14:691--706. Myking, T. and O.M. Heide. 1995. Dormancy release and chilling requirement of buds of latitudinal ecotypes of Betula pendula and B. pubescens. Tree Physiol. 15:697--704. Repo, T., H. Hänninen and S. Kellomaki. 1996. The effects of longterm elevation of air temperature and CO 2 on the cold hardiness of Scots pine. Plant Cell Environ. 19:209--216. Sakai, A. and W. Larcher. 1987. Frost survival of plants: responses and adaptation to freezing stress. Springer-Verlag, Berlin, 321 p. Silim, S.N. and D.P. Lavender. 1994. Seasonal patterns and environmental regulation of cold hardiness in shoots of seedlings of Thuja plicata, Chamaecyparis nootkatensis, and Picea glauca. Can. J. Bot. 72:309--316. Tingey, D.T., B.D. McVeety, R. Waschmann, M.G. Johnson, D.L. Phillips, P.T. Rygiewicz and D.M. Olsyzk. 1996. A versatile sun-lit controlled-environment facility for studying plant soil processes. J. Environ. Qual. 25:614--625. Tinus, R.W., D.H. Greer and L.A. Robinson. 1995. Effect of high CO 2 on cold hardening and dehardening of three conifers. Interior West Global Change Workshop, Fort Collins, CO, pp 48--50. van den Driessche, R. 1975. Flushing response of Douglas-fir buds to chilling and to different air temperatures after chilling. B.C. For. Serv. Res. Notes. No. 71, 22 p. Waring, P.F. 1969. The control of bud dormancy in seed plants. In Dormancy and Survival. No. XXIII. Ed. H.W. Woolhouse. Symp. Soc. Exp. Biol., Cambridge Univ. Press, pp 241--262. Weiser, C.J. 1970. Cold resistance and injury in woody plants. Science 169:1269--1278. TREE PHYSIOLOGY ON-LINE at http://www.heronpublishing.com