Sky exposure, crown architecture, and low-temperature photoinhibition in conifer seedlings at alpine treeline


  • M. J. GERMINO,

    1. Botany Department, Box 3165, University Station, University of Wyoming, Laramie, Wyoming 82071–3165 and,
    Search for more papers by this author
  • W. K. SMITH

    1. Botany Department, Box 3165, University Station, University of Wyoming, Laramie, Wyoming 82071–3165 and,
    2. Department of Biology, Box 7325, Wake Forest University, Winston-Salem, North Carolina, 27109–7325, USA
    Search for more papers by this author

Dr William K. Smith Department of Biology, Box 7325, Wake Forest University, Winston-Salem, North Carolina, 27109–7325, USA. E-mail:


In the alpine-treeline ecotone of the Snowy Range in Wyoming, USA, microsite sky exposure of Englemann spruce (Picea englemannii) and subalpine fir (Abies lasiocarpa) seedlings (< 5 years) was associated with the avoidance of low-nocturnal temperatures and high insolation, factors which appeared to result in low-temperature photoinhibition. In a field experiment, light-saturated photosynthesis (Asat) in current-year seedlings (newly germinated) of fir increased significantly (approximately seven-fold) in response to increased long-wave irradiance at night (warming), solar shading (approximately five-fold), and the combination of the two treatments (approximately eight-fold). Asat in current-year spruce remained unchanged in response to all treatments, but was over four-times higher than fir in control plots. These results indicated substantial low-temperature photoinhibition, and were supported by similar Asat trends in natural seedlings. Increased needle inclination and clustering in more exposed microsites for both species implicates the possible role of structural adaptations for decreased sky exposure and warmer leaf temperatures at night.


While the potential importance of seedling biology to the boundary dynamics of landscapes is recognized (e.g. Clark et al. 1998), few studies have investigated seedling dynamics at alpine treeline, one of the most distinct boundaries separating different plant communities. The transition from forest to treeless alpine meadow in the Snowy Range of the Medicine Bow National Forest in south-east Wyoming occurs over a relatively level area covering many square kilometres (Oosting & Reed 1952). Within this broad treeline ecotone, the dominance of abiotic forces on plant patterns is readily apparent (Smith & Knapp 1990). The codominant tree species, Englemann spruce (Picea englemannii (Parry) and subalpine fir (Abies lasiocarpa (Hook.) Nutt.), occur in a patchy mosaic of clonal tree islands (Oosting & Reed 1952; Shea & Grant 1985) that may be a major landscape feature influencing the microclimate of seedling establishment (Billings 1969). Tree islands can alter potential microsites through their influence on wind and snowdrift formations (Billings 1969), as well as solar and long-wave radiation exposure (Jordan & Smith 1995a). With increasing distance from the closed-canopy, subalpine forest, these tree islands become further apart and reduced in stature due to leaf abrasion by wind-driven snow crystals, needle desiccation and death (Hadley & Smith 1987). Thus, differences in snow accumulation, moisture, sun, and numerous other micrometeorological parameters occur across a distinct gradient in tree structure from the adjacent subalpine forest to the treeless alpine tundra.

The potential importance of sunlight exposure, daytime high temperatures, and night-time low temperatures to growth in spruce seedlings (Ronco 1970; Hellmers, Genthe & Ronco 1970) and alpine vegetation (Körner 1994) has been recognized, previously. However, the combined influence of these factors has been interpreted only recently within the context of low-temperature photoinhibition (Ball 1994; Jordan & Smith 1994; Ball et al. 1997). Night-time air temperatures frequently approach freezing at the upper timberline, and radiational cooling (due to low values of long-wave radiation from cold, clear skies) often results in leaf temperatures that are well below air temperature (Smith 1985; Hadley & Smith 1987; Jordan & Smith 1994, 1995a,b). Microsites with high sky exposure receive more incident photosynthetic photon flux density (PPFD) (0·4–0·7 μm) and have higher leaf temperatures during the day, but lower leaf temperatures during the night compared with microsites that have less sky exposure. In addition, decreased wind at night and cold-air drainage cause cold air to accumulate next to the ground, increasing frost occurrence at ground level where the young seedlings reside (Jordan & Smith 1994). There can be as few as five frost-free days throughout the growing season in high elevation habitats (Jordan & Smith 1995a). Dewfall is even more common (Smith & McClean 1989). Typically, clear skies lead to the strongest radiational cooling at night, in addition to high solar irradiance the next morning, a natural, and probably frequent, regime in many habitats.

The purpose of the current study was to evaluate the potential relationship between low-temperature photoinhibition and the distribution pattern of young conifer seedlings. We hypothesized that susceptibility to low-temperature photoinhibition is an important limitation to conifer seedling establishment at the upper treeline. To test this hypothesis, we examined the distribution and range of microsite sky exposure for spruce and fir seedlings growing naturally in the field, photosynthetic responses to PPFD and temperature for natural and potted seedlings in the field, and crown architectural differences which could ameliorate microsite sky exposure for natural seedlings.


Seedling distribution throughout the treeline ecotone was measured by sampling along transects established in four major zones within the ecotone. The radiation microenvironment of young seedlings was characterized by quantifying the percentage sky exposure in the upper hemisphere above individual seedlings. To test the hypothesis that spruce and fir seedlings are susceptible to low-temperature photoinhibition, the effects of sunlight and night-time minimum temperatures on photosynthesis were evaluated independently. Photosynthetic gas exchange was measured for current-year seedlings (newly germinated) in experimental treatments that either reduced the incident PPFD only, increased long-wave radiation at night only, or did both. A control treatment with natural sunlight and temperature was also included. The crown structure for seedlings growing in sites that differed in sky exposure was also evaluated because of known effects of leaf orientation and arrangement on both long-wave (night-time) (Jordan & Smith 1994; Campbell & Norman 1997) and solar radiation interception (e.g. Carter & Smith 1985). Experiments began on 10 July 1995, just after most of the winter snowpack had melted and only snowdrifts remained, and ended on 12 October 1995 with the return of the winter snowpack.

Site description

Research sites utilizing natural seedlings were located at 3230 m altitude in the alpine-treeline ecotone of the Snowy Range in the Medicine Bow National Forest of south-eastern Wyoming, USA (N 41°22'25", E 106°14'00"). This area consisted of ribbon-like tree islands (Billings 1969) that each covered a maximum ground area of nearly 1000 m2 and had mature trees approximately 15 m in height. Adjacent meadows were typically greater than 2000 m2, with herbaceous ground-cover less than about 0·25 m in height. Potted seedling experiments were conducted 5 km south-east of this site at 3260 m altitude in a slightly more exposed area of the ecotone, where even larger meadow areas separated tree islands and mats within the ribbon forest.

Distribution of natural seedlings

Four zones within the treeline ecotone were identified for sampling, based on the relative height and spacing of the conifer tree islands and mats (Fig. 1). The lowest elevation and least wind-exposed portion of the ecotone had the largest trees and the smallest average distance between tree islands, whereas the most exposed sites were characterized by low-lying krummholz mats. Zones 2 and 4 differed from 1 and 3 in that the influence of multiple krummholz patches or tree islands on ground-level seedling microsites was more likely. In each of the four ecotone zones, a baseline was established in a perpendicular orientation (north–south) to the prevailing winds (west), and transects were established in an easterly direction at regular intervals along this baseline. One square-metre quadrants were placed at either 5, 10, or 20 m intervals along the transects, based on a randomized sampling scheme (Table 1). The occurrence of seedlings less than 5 cm in height was recorded in each sample quadrant along each transect.

Figure 1.

. Division of the timberline ecotone into four zones based on height and spacing of tree islands, factors that exert a strong influence on ground-level microclimate. The typical tree structure is represented in each frame. In transect 1, the krummholz mats were less than 1 m tall, and spaced approximately 10 m or more apart. In transect 2, the islands were 1–3 m tall, and 2–10 m apart. In transect 3, the trees were 10–15 m in height, and 20–50 m apart, and in transect 4 the spacing decreased to 5–25 m. The prevailing wind direction in each box is from right to left. Increased spacing between tree islands in transect 3 is due to the ribbon effect (Billings 1969).

Table 1.  . Distribution of young seedlings (2–5-year-old) across the timberline ecotone. See Fig. 1 for structural characteristics of each transect site. The total area is the area bound by each baseline and series of transects. The sample area is the total area examined in the quadrants Thumbnail image of

To estimate the relative amounts of night-time long-wave radiation incident on a seedling microsite, the percentage sky exposure (%SKY) was determined by hemispherical photography (Jordan & Smith 1994). The quantity %SKY was measured with a fisheye lens mounted on a 35 mm camera that was positioned directly above each seedling. Black and white film negatives were developed into prints and then scanned using Applescan (Apple Computer, Inc., Cupertino, CA, USA). The program SOLARCALC (Chazdon & Field 1987) was used to calculate the percentage of the digitized image occupied by sky as described in Jordan & Smith (1994).

Experimental treatments on potted seedlings in the field

To separate the effects of low night-time temperatures versus high incident sunlight, clear plastic sheeting that transmitted nearly full-sun irradiance, including UV, was used for the night-time warming treatment (NW). Solar shade bands were used for the daytime shade treatment (DS) that had minimal influence on %SKY and, thus, long-wave irradiance from the sky (LSKY) and nocturnal temperatures. The clear plastic was 0·102 mm thick (Great Western Bag Co., St. Louis, MO, USA) and was elevated 0·5 m off the ground on frames made of thin (3·75 cm × 1·75 cm) wood strips to minimize shading. The distance between the frame and the ground allowed natural wind-flow and convection. Individual plastic sheets were large enough (1 m2) to eliminate penetration of LSKY from low angles.

For the DS treatments, hemispherical shade bands (0·4 m × 4 m) were used that consisted of strips of two layers of neutral-density screen oriented to block the sun's direct beam throughout the day. Polyvinyl chloride (PVC ) tubing was used to create a sturdy structure that could be curved to coincide with the trajectory of the sun. This open-top, half-moon design was used in order to minimize any perturbation of natural convection, and to provide as little thermal mass in the upper hemisphere of the pots as possible. The shade bands were only as wide as required to shade experimental plants, and were continually adjusted for the duration of the experiment as the solar elevation changed. The four treatment levels of radiation were the plastic sheet for warming at night (NW), the solar band to create shade during the day (DS), a combination of the two treatments (DS + NW), and a control plot with natural radiation regimes (CONT).

To evaluate the effectiveness of all treatments for controlling the designated parameters, microclimate measurements of temperature, solar, and long-wave radiation were made periodically throughout the study period. All radiation and temperature measurements were recorded using a datalogger (Model 21X; Campbell Scientific Instruments, Logan, UT, USA) with 15 min averages recorded for samples taken every 2 min. The microclimate parameters were evaluated at one replicate of the four treatments and included LSKY, two replicates of air temperatures adjacent to needles, and both integrated and spectral PPFD (0·4–0·7 μm). LSKY was measured using all-wave total hemispherical radiometers (Model Q7; Radiation and Energy Balance Systems, Inc, Seattle, WA, USA) and PPFD was measured using quantum sensors (Model Li-190; LiCor, Lincoln, NE, USA). To ensure that the spectral distribution of sunlight was not altered by the treatments, spectral measurements of sunlight passing through the treatment materials were compared with unfiltered sunlight using a spectro-radiometer (Model Li-1800; LiCor). Both the solar and long-wave radiation effects were evaluated on a total of seven days and nights, and the spectral effects during one afternoon.

Needle and air temperatures were measured using fine-wire thermocouples (0·023 mm diameter, type T, copper-constantan, Omega Engineering, Stamford, CT, USA.). The thermocouples were positioned directly adjacent to the conifer needles under each treatment due to the difficulty of maintaining direct contact. Needle dimensions were less than 0·4 mm in diameter (slightly larger than the soldered thermocouple sensing junction) and less than 1 cm in length. Comparisons of needle temperatures with thermocouples attached directly to needles, versus thermocouples placed within a few millimetres of needles showed an error of less than ± 0·4 °C (Hadley & Smith 1987). All radiation and temperature measurements were recorded using a datalogger (Model 21X; Campbell Scientific Instruments, Logan, UT, USA) with 15 min averages recorded for samples taken every 2 min.

Photosynthesis of potted germinants

Photosynthetic performance was measured for potted germinants of spruce and fir located under the treatments described above from 29 August 1995 to 16 September 1995. Seeds were collected from high elevation forests in the Arapaho National Forest south of the research site (Swift Seed Co., Jaroso, CO, USA), allowed to germinate in a glasshouse, and were immediately moved to the treatments prior to the emergence of cotyledons. Newly germinated seedlings from the adjacent forest understory were also transplanted in an identical manner. All seedlings were contained in pots (23 cm wide × 21 cm deep) covered by thin wire mesh with 1 cm2 holes to exclude meadow voles and were watered regularly. Unfertilized potting soil was used that consisted of equal volumes of humus, peat, and perlite. The positions of the pots within each treatment were rotated to reduce any heterogeneity in microclimate for individual treatments.

Mean photosynthesis under full sun (PPFD >1800 μmol m–2 s–1) was computed for each pot by averaging multiple measurements of randomly selected entire seedlings (n = 3 to 12), and all photosynthetic measurements were made in full sun. The response of photosynthesis to light was determined using neutral density screens and allowing 10 min for acclimation to each PPFD level. All photosynthetic measurements were made using a portable photosynthetic gas exchange system (Model 6200; LiCor) modified to improve measurement accuracy for low rates of photosynthesis and leaf area within the sample cuvette. Specifically, a cuvette was designed with reduced volume (200 cm3), and a hinged door on the bottom of the chamber facilitated enclosing entire seedlings within a few centimetres of the ground.

The sunlit leaf area for naturally oriented seedlings was used for computing photosynthetic flux density as recommended by Smith et al. (1991) under full sunlight conditions. Sunlit leaf area was determined for each seedling at the end of the experiment using a video area meter (Decagon Instruments, Pullman, WA, USA) as described in Carter & Smith (1985). For a given measurement, photosynthesis was measured with the entire seedling enclosed in the cuvette (i.e. both cotyledons and primary needles simultaneously). Care was taken to ensure a natural orientation to the sun at the time of measurement.

Crown architecture

The ratio of sunlit to total leaf area, referred to as ‘STAR’ (Carter & Smith 1985; Oker-Blom & Smolander 1988) was used to compare needle arrangement and orientation for spruce and fir seedlings growing in microsites that differed in %SKY. Seedlings were oriented at 70° (the maximum solar elevation for this site) to a video area meter to determine the sunlit leaf area as described above. Total needle surface area was estimated by multiplying the sum of the sunlit leaf area for all individual needles by a constant derived from cross-sectional geometry (Jordan & Smith 1993). Needle inclination from horizontal, a determinant of STAR, was measured with a clinometer. Seedlings were sampled from ‘exposed’ sites with high values of %SKY, and ‘sheltered’ sites with surrounding herbaceous vegetation, logs, low adult conifer branches, and other cover which generated reductions in %SKY. The sampled seedlings were approximately 4 to 6 years old and all were less than 5 cm in height.


A two-way analysis of variance (ANOVA), with species and microsite as the main factors, was used to compare STAR for natural seedlings of spruce and fir found in sheltered or open microsites. The interactions between the species and microsites for STAR were evaluated using the SAS LSMEANS mean separation (all possible t-tests) (Release 6·09; SAS Institute, Raleigh, NC, USA). For the experimental treatments, the microclimate data were analysed using one-way ANOVA with replicates in time and the least significant difference (LSD) mean separation (α = 0·05). Mean photosynthetic responses under each treatment (Asat, light-saturated photosynthetic CO2 flux density) were transformed using Taylor's Power Law (Southwood 1978) and then analysed using a randomized complete block design ANOVA (Hoshmand 1993) with the SAS LSMEANS (α = 0·1) mean separation for evaluating the species and treatment interaction. All ANOVA was computed with the General Linear Model in SAS. Two replicates (blocked) of the experiment, each containing one set of the four treatment levels of the independent variable (CONT, DS, NW, DS + NW), were positioned at adjacent sites within the ecotone. For each treatment, there was one pot per species, initially containing 12–20 current-year seedlings, but ending with as few as five. Seasonal averages were computed for the potted seedling microclimate and seedling photosynthesis data.


Young seedling distribution and microsite sky exposure

Of the four zones identified here for the full-breadth of the treeline ecotone (Fig. 1), seedlings were found mainly in the two transects (zones 3 and 4) established in the lowest portion of the ecotone (Table 1). Only one fir seedling was found in the second transect within a total, randomized sample area of 110 m2 within a total plot size of 2500 m2. The largest number of seedlings for both species occurred in the zone immediately adjacent to the closed subalpine forest (transect 4) where 74 spruce and 21 fir were recorded in a sample area of 80 m2 and a total plot size of 10 000 m2. Thus, seedling occurrence ranged from near zero in the upper ecotone transects (zones 1 and 2) to about one seedling/m2 in the zone closest to the adjacent subalpine forest. In addition, spruce was 3.5-fold more abundant than fir. The maximum percentage of seedlings was found at microsites with 40–80% sky exposure, while only 22% of seedlings were found at microsites with > 80% or < 40% sky exposure (Fig. 2). Only fir seedlings were found in the lowest (20–40%) category of %SKY represented by seedlings, while mostly spruce (except for one fir) were found in the highest (80–100%) category of sky exposure.

Figure 2.

. Range of sky exposure (%SKY) determined from fisheye photos for all seedlings found along transects extending from the conterminous forest into subalpine meadows and patch-forest at the Brooklyn Lake study site. The solid bars represent spruce; open bars represent fir. The seedlings were approximately 5- to 10-year-old and less than 5 cm in height.

Validation of experimental treatments

The microclimates of each of the four experimental treatments were monitored to evaluate and quantify their effectiveness in regulating target parameters (Tables 2 & 3). For the night-time warming treatment (NW), the polyethylene sheets increased the LSKY term by about 30 W m–2 above the control mean of 245 W m–2 (P = 0·0008), resulting in a mean 0·99 °C increase in estimated minimum needle temperature at night (P = 0·0111) over the total growing season (10 June 1995 to 12 October 1995). As a result, the number of nights with at least one 15 min period of below 0 °C estimated needle temperature decreased from 34% of nights to 14%. The daytime solar shading treatment (DS) decreased PPFD to about 40% of full sun values (to the approximate PPFD required for light saturation of photosynthesis of natural seedlings) and the plastic panels of the NW enclosures reduced PPFD to about 80% of full sunlight values (P = 0·0001). There were no measured effects on spectral changes in the visible or UV wavelengths in either treatment. In addition, daytime estimate needle temperature means were 2·1 °C, and 1·8 °C lower under the DS and combined DS + NW treatments than the control (P = 0·0111).

Table 2.  . Validation of treatment effects on microclimate. The letters denote significantly different groups (LSD, α = 0·05). DS represents the daytime shading treatment, NW is nighttime warming, DS + NW is the combined treatment, and CONT represents the control treatment. LSKY is long-wave radiation from the upper hemisphere; PPFD is photosynthetic photon flux density. Temperatures are mean minimum and maximum values for 15-minute measurement intervals. See Table 3 for ANOVA results Thumbnail image of
Table 3.  .ANOVA results for the validation of the experimental microclimate treatments, photosynthesis of potted germinants in the experimental treatments, and crown architecture for natural seedlings growing in microsites that differ in sky exposure. LSKY is long-wave radiation from the upper hemisphere; PPFD is photosynthetic photon flux density Thumbnail image of


The Asat responses to the experimental treatments in cotyledon-bearing, current-year seedlings were different between the species and treatments (P = 0·0551). The value of Asat for spruce in the DS, NW, or DS + NW treatments was not significantly different from Asat in the CONT treatments. In fir seedlings, however, Asat in the NW and DS treatments resulted in 6·8-fold (P = 0·0053) and 4·5-fold (P = 0·0218) increases over Asat in the CONT treatment, respectively, compared with an 8·2-fold increase in Asat in the DS + NW treatment (P = 0·0028). Asat in the CONT seedlings was four-fold higher in spruce than fir (P = 0·0262), and Asat was 55% higher in fir for the DS + NW seedlings (P = 0·0615) (Fig. 3). There was no difference in Asat between spruce and fir in the NW and DS treatments (P = 0·7254 and P = 0·9699, respectively). The highest and lowest Asat for all species and treatments occurred in the fir seedlings under the combined DS + NW treatment (5·5 μmol m–2 s–1) and in the CONT treatment (0·6 μmol m–2 s–1), respectively. Higher photosynthesis for fir germinants in the DS + NW treatment over the control is supported by higher photosynthesis at light saturation in 4Fig. 4a. The value of Asat was higher for natural 4–6-year-old seedlings of spruce than fir in conditions similar to the CONT treatment, supporting higher photosynthesis in current-year seedlings of spruce than fir in the experiment (Fig. 4b).

Figure 3.

. Mean photosynthetic CO2 flux density for spruce and fir seedlings under saturating light during late August and September 1995. The open bars represent spruce seedlings; hatched bars represent fir. See Table 3 for ANOVA and text for LSMEANS results. (n = 7 to 10, and error bars are SE).

Figure 4.

. Photosynthetic light response of (a) potted first-year seedlings of fir under the control (•) and combined DS + NW (shaded + warmed) treatments (O) (n = 3) and (b) naturally occurring 4- to 6 years-old spruce (•) and fir (O) seedlings under 3 cm height (n = 6) in microclimate conditions similar to the CONT treatment. Curves are hand-drawn; error bars indicate SE.

Crown architecture

In sky-exposed versus sheltered microsites, the ratio of sunlit to total leaf area (STAR), decreased by 12% to 0·22 ± 0·048 for fir (P = 0·925) and by 32% to 0·15 ± 0·021 in spruce (P = 0·0001) in open (high %SKY) versus sheltered (low %SKY) sites (Fig. 5). The lowest STAR values (0·15) were for spruce seedlings in microsites with high sky exposure (P = 0·0001). Correspondingly, the mean needle angles increased 57% (44·3°± 4·76 to 69·3°± 2·20) from sheltered to open sites in spruce and 100% (34·4°± 8·31 to 69·1°± 2·65) in fir (Fig. 5).

Figure 5.

. Sunlight interception efficiency (STAR, ratio of sunlit to total leaf area) and needle inclination for fourth to sixth year spruce and fir seedlings growing in microsites with high sky exposure (exposed) or low sky exposure (sheltered). P = 0·0002 for differences in STAR in open versus sheltered sites, and P = 0·0001 for differences in STAR between the species. See Table 3 for ANOVA results. (n = 7 to 9. and error bars are SE).


In the current study, a relatively large number of previously unreported young conifer seedlings (< 5-year-old) were found in this alpine-treeline ecotone (see Billings & Mark 1957). Also, the high seedling mortality (> 90%) reported for current- and second-year seedlings in the adjacent forest understory (Cui & Smith 1991) may indicate that the strongest selective pressure of any life stage in these long-lived conifer tree species occurs in current-year, emergent cohorts. Moreover, changes in emergent seedling distribution and mortality patterns could provide a year-to-year measure of treeline dynamics (Hansen-Bristow & Ives 1984; Smith & Knapp 1990; Slatyer & Noble 1992; Rochefort et al. 1994) which is not available from tree-ring studies of annual growth or age. The evidence presented here points to the potential importance of low-temperature photoinhibition as a mechanism influencing the establishment of conifer seedlings in this alpine-treeline ecotone, and possibly elsewhere. Thus, structural and microsite factors associated with sky exposure (daytime sun plus cold night skies) could be important for avoiding both the low needle temperatures and high incident sunlight that lead to low-temperature photoinhibition of photosynthesis.

%SKY and seedling distribution

The value of %SKY directly above an individual seedling can have important influences on radiation exchange, both at night and during the day (Ball 1994; Jordan & Smith 1994, 1995b). Leaf temperatures at night are strongly influenced by %SKY because of its strong influence on LSKY, especially on calm nights and minimal convective heat exchange (Jordan & Smith 1994, 1995a,b). Leaves exposed to the night sky will experience low values of LSKY and corresponding leaf temperatures well below air temperature (up to 7 °C below), often resulting in dew formation and/or radiation frost. Jordan & Smith (1994), working at the same research site, showed that the association between LSKY and %SKY was linear. For example, during the growing season, seedlings in microsites with 100% sky exposure received about 250–400 W m–2 of LSKY, while seedlings experiencing 50% sky exposure received about 50 W m–2 more incident LSKY on a clear night (Jordan & Smith 1994). During the 1995 growing season, this would have resulted in an approximately 4 °C warmer mean needle temperature at night [based on Table 2 and Jordan & Smith (1994)]. In addition to low-nocturnal temperatures, the highest %SKY values and clear skies also generated the highest incident sunlight levels the following morning. Thus, the combined stresses of low-nocturnal temperature and subsequent high-sunlight exposure may represent a relatively common situation that occurs naturally in the field and which could generate low-temperature photoinhibition (Jordan & Smith 1994).

In the current study, both spruce and fir seedlings (2- to 5-year-old) were most abundant at locations that had %SKY values (including sky-occluding features at all scales > 20 cm high) between 40 and 80% (Fig. 2). Only fir seedlings occurred in microsites with the lowest %SKY, and only spruce seedlings occurred at microsites with the greatest %SKY. The lack of seedlings at microsites with the least %SKY could reflect a lack of adequate sunlight for growth. The low abundance of seedlings at microsites with the greatest sky exposure may indicate a greater susceptibility to radiation frost and, in combination with high sunlight levels, a high degree of low-temperature photoinhibition. The occurrence of young seedlings only in the lowest portion of the ecotone, a greater abundance of spruce seedlings (Table 1), and greater %SKY for spruce seedlings (Fig. 2) is supported by Resor, Germino, and Smith (unpublished) for current-year germinants, and Daly & Shankman (1985) and Weisberg & Baker (1995) for older saplings in a similar treeline in Colorado.

Low-temperature photoinhibition

Long, Humphries, & Falkowski (1994) defined photoinhibition as the light-dependent depression in photosynthesis that results in reduced apparent quantum yield, decreased convexity in the light response of photosynthesis, and decreases in maximum photosynthesis. Although the deleterious effect of simultaneous exposure to low temperature and high sunlight on photosynthesis is recognized (reviewed by Krause 1994), many plants are also subjected to the low temperatures at night followed by high sunlight levels the following morning. Despite an abundance of greenhouse and laboratory investigations on low-temperature photoinhibition in plants, little is known about the importance of this temporal separation of minimum temperatures and high sunlight, probably a frequent occurrence in natural habitats during the growing season.

The responses of Asat to the DS, NW, and DS + NW treatments suggest that current-year seedlings of fir were most sensitive to low-nocturnal temperatures and high sunlight. Low-nocturnal temperatures and high sunlight each appeared to be deleterious to Asat in fir, but not in spruce. In fir, Asat increased substantially with protection from high sunlight or higher nocturnal leaf temperatures, and the combination of warming and shading resulted in the largest increase in Asat over control plants. The combined effects of low-nocturnal temperature and high sun also appeared to be more deleterious to photosynthesis than either stress alone. Finally, photosynthesis in current-year fir seedlings was higher than in spruce in the DS + NW treatment, but much less than for older spruce under natural conditions similar to the CONT treatment. Thus, fir seedlings appeared much more susceptible to low-temperature photoinhibition than spruce seedlings.

There is general agreement among the few studies that have dealt with the importance of low temperatures and high insolation to conifer seedlings and the results of our study. Also, the importance of low temperatures to photosynthesis in conifers is well established (e.g. Leverenz & Öquist 1987). Hellmers et al. (1970) found that minimum night-time temperatures were more limiting than daytime temperatures to the growth and survival of greenhouse Englemann spruce seedlings, although freezing temperatures were not investigated. They observed this trend even when day temperatures were lower than night-time temperatures, and the highest mortality of these greenhouse seedlings occurred under the combination of the lowest night-time temperatures and highest daytime temperatures measured. DeLucia & Smith (1987) also found that photosynthesis in adult Englemann spruce was correlated with the minimum air temperature of the previous night. Light frosts above –2·5 °C caused slight and reversible depressions in stomatal conductance and photosynthesis, while hard frosts below –4 °C induced more severe depressions lasting over 24 h.

The term ‘solarization’ has been used to refer to the photosynthetic depressions and visible chlorophyll degradation in Englemann spruce seedlings that appeared to result from the combination of water stress and high sunlight (Ronco 1970). More recently, low-temperature photoinhibition was detected in other planted conifer species in response to low-nocturnal temperatures (Lundmark & Hällgren 1987). In contrast, alpine plants can be biochemically adapted to low-temperature and high-light stress and, thus, tolerant of low-temperature photoinhibition (Streb, Feierabend & Bligny 1997). In Australian timberlines, Ball (1994, 1997) found that exposure to low night-time temperatures and high insolation were associated with differences in distribution patterns of Eucalyptus pauciflora due to low-temperature photoinhibition. Eucalyptus seedlings were found only in a limited space under mature eucalyptus trees where the solar and night-time LSKY shadow overlapped.

Crown architecture

The arrangement and orientation of cotyledons, or needles, can have important influences on sunlight absorption during the daytime, LSKY at night and, thus, leaf temperature at day or night (Smith & Brewer 1994). In the present study, seedlings which occurred at microsites with high %SKY also had needles that were more inclined from the horizontal and more clustered (Fig. 5). Needle clustering can lead to substantial daytime warming (due to reduced convective heat dissipation), even with less incident solar radiation (Smith & Carter 1988). This adjustment in crown architecture could be a distinct advantage for reducing low-temperature photoinhibition during cold days. Needle inclination away from the horizontal should also decrease sky exposure, resulting in warmer needle temperatures at night. Thus, the higher abundance of spruce versus fir seedlings in exposed sites with lower values of LSKY and higher values of incident sunlight may be related to the 2·6-fold larger reduction in STAR values measured for spruce seedlings at exposed microsites than for fir.


The current study provides data showing the presence of significant numbers of young conifer seedlings (2- to 5-year-old) in an alpine-treeline ecotone. The proximity of seedlings to mesotopographic structures such as tree islands may implicate the importance of snow accumulation, wind protection, or snow-melt moisture to seedling survival (Billings 1988, Slatyer & Noble 1992), in addition to the low-nocturnal temperatures and high insolation implicated here and in Ball (1994). Greater occurrence of fir seedlings in microsites with reduced sky exposure, plus the greater susceptibility of fir to low-temperature photoinhibition, may indicate species-specific differences leading to seedling establishment in this marginal habitat. Evidence that sunlight can be a physiological stress challenges the general view that sunlight alone is the primary limiting resource for plants in natural habitats (Tillman 1988). Night-time sky exposure and minimum temperatures, in combination with daytime sun exposure, may represent multiple stresses that limit the establishment and growth of subalpine trees in these high elevation habitats, and probably elsewhere.


Support was provided to M. J. Germino by a NASA fellowship administered by the Wyoming Space Grant Consortium and through a NSF EEP grant awarded to W. K. Smith. Statistical advice was provided by Dave Legg.