Photosynthetic responses of Larrea tridentata to seasonal temperature extremes under elevated CO2


Author for correspondence: Elke Naumburg Tel: +1 970 377 9410 Fax: +1 970 377 9406 Email:


  • • Elevated CO2 potentially decreases the effects of temperature stress on photosynthesis. Under both freezing and high temperatures previous studies have shown that elevated CO2 can particularly enhance photosynthetic rates, although results from freezing studies are more variable.
  • • Here we show gas exchange responses of Larrea tridentata to elevated CO2 over a 6-yr period when temperature stress events may have had a significant effect on photosynthesis in the field.
  • • Nighttime freezing air temperatures decreased subsequent daytime photosynthetic rates, stomatal conductance, and the maximum yield of PSII similarly under ambient and elevated CO2. Further, we found no statistically significant relationship between leaf temperature and photosynthetic enhancement. Overall, the degree of photosynthetic enhancement under elevated CO2 was directly proportional to the response of stomatal conductance to CO2.
  • • Thus, elevated CO2 does not significantly affect apparent physiological responses of Larrea to temperature extremes. However, because of the tight relationship between stomatal conductance and photosynthetic enhancement, potential climate change effects on stomatal conductance will significantly influence Larrea performance in the future.


Anthropogenic increases in carbon dioxide have and will continue to alter the global climate (Schneider, 1989; Kacholia & Reck, 1997). Although predictions vary, global air temperatures by the middle of this century will increase by 1.3° to 5.8°C compared with pre-industrial times because of rising levels of atmospheric CO2 and other greenhouse gases (Watson et al., 1996; Crowley, 2000; Houghton et al., 2001). Global warming will likely alter synoptic weather patterns, change precipitation regimes, and increase air and soil temperatures, thus resulting in potentially dramatic effects on ecosystems if environmental factors exceed the tolerance range of species (Peters & Lovejoy, 1992). In southwest American deserts limits to species distributions relate mostly to episodic stress events (Smith & Nowak, 1990; Smith et al., 1997). Specifically, drought (Pockman & Sperry, 2000) and freezing temperatures dictate the distributional limits of many warm desert species (Beatley, 1974; Pockman & Sperry, 1997). Thus, potential shifts in plant establishment, growth, and reproduction will be intimately tied to the degree of biophysical stress in a future climate (Smith & Nowak, 1990; Larcher, 1995).

It is well established that photosynthesis decreases following episodic freezing. Numerous studies have shown that the primary site of injury is the plasmalemma, which is damaged by freeze-dehydration (Thomashow, 1999; Uemura & Steponkus, 1999; Xin & Browse, 2000; Pearce, 2001). Increasing levels of anthropogenic atmospheric CO2 may affect freezing tolerance via changes in membrane composition, dehydrins, stress tolerance proteins, nucleating agents, antifreeze compounds, or cryoprotectants (Steponkus, 1984; Thomashow, 1999; Xin & Browse, 2000). These changes could influence the impacts of freeze-induced cellular dehydration on cellular viability and photosynthetic or respiratory metabolism, as well as allocation to metabolic sinks. The responses to temperature stress are complex and vary considerably across species. For example, many species exhibit a protection of photosynthesis from freezing temperatures when grown under elevated CO2, whereas other species display decreased photosynthetic tolerance to freezing (Nobel, 1996; Wiemken et al., 1996; Guak et al., 1998; Lutze et al., 1998). Even within a species there can be both increases and decreases in physiological tolerance to freezing (Hymus et al., 2001). Many desert plants are limited in their distribution by freezing temperatures, and growth in elevated [CO2] can increase net CO2 uptake (Nobel, 1996). However, little is known about the effects of changes in atmospheric composition on thermal tolerance for plants from arid regions. For Mojave Desert species in the genus Yucca, elevated CO2 decreased both membrane and PSII damage following freezing temperatures (Loik et al., 2000). Consequently, Fv : Fm (indicating PSII efficiency) and ΦPSII (related to photosynthetic rates) were significantly higher in elevated CO2 plants.

At the other extreme, high temperatures decrease photosynthetic rates because the ratio of oxygenation to carboxylation reactions of Rubisco increase as a result of comparatively greater increases in oxygen affinity and solubility. Therefore, elevated CO2 will enhance photosynthesis relatively more at high than at moderate temperatures (Long, 1991). Elevated levels of atmospheric CO2 may also affect the relative amount of Photosystem II photochemical energy allocated to thermal dissipative processes in comparison with carbon fixation (Hymus et al., 1999). Moreover, the effects of elevated CO2 potentially interact with coincident drought as well as other atmospheric constituents such as ozone, and may lead to a reduction in photosynthetic productivity (Huxman et al., 1998; Grams & Matyssek, 1999; Hamerlynck et al., 2000a). As a consequence, rising atmospheric CO2 may effectively alter the temperature stress tolerance of plants in one or more ways.

The northern Mojave Desert represents a high-stress environment that is characterized by disjunct seasonal distributions of precipitation and temperature (Smith et al., 1997). Approximately half of the annual precipitation occurs between November and March, when low daytime temperatures and nighttime freezing are frequent. Conversely, once freezing temperatures cease, soil moisture rapidly declines and is rarely recharged until the following winter (Smith et al., 1997). As a consequence, low temperatures limit plant growth during periods of comparatively high soil moisture, while later in the season water and high temperature limitations dominate. Larrea tridentata (DC.) Cov. is an evergreen xerophytic shrub that dominates in the Mojave and other North American warm deserts and can be photosynthetically active throughout the entire year. It has shown high photosynthetic temperature acclimation, with peak rates essentially constant year-round despite large differences in air temperature (Mooney et al., 1978). As such, Larrea makes an ideal model to study the effects of rising atmospheric CO2 on the enhancement of photosynthesis as a function of air temperature. While previous studies have addressed general patterns of Larrea photosynthesis under elevated CO2 (Huxman et al., 1999; Hamerlynck et al., 2000b; Naumburg et al., 2003), none have directly addressed effects of seasonal temperature variations on Larrea photosynthesis. We tested the hypotheses that photosynthesis is enhanced for plants growing under elevated CO2 compared with plants growing under ambient, control CO2 concentrations, following exposure to subzero or hot air temperatures that occurred naturally throughout six growing seasons. Specifically, we predicted that elevated CO2 would partially alleviate the detrimental effects of subzero or hot air temperatures on photosynthesis in Larrea plants in their natural environment.


The study was conducted with mature Larrea tridentata at the Nevada Desert FACE Facility (NDFF), in southern Nevada, USA. The NDFF has been operating since 1997; it is composed of three 23-m-diameter plots operating continuously at 550 µmol mol−1 CO2, three ambient CO2 plots with the same FACE apparatus, and three nonblower controls (Jordan et al., 1999). Elevated CO2 plots were fumigated > 96% of the time when temperatures were above 4°C, with a mean elevated [CO2] of 547 µmol mol−1 during those times. The vegetation is typical of Mojave Desert shrub communities, dominated by Larrea tridentata and the deciduous shrubs Ambrosia dumosa and Lycium spp. (Beatley, 1975). The site is equipped with a weather station; soil water content is measured with TDR probes at 0–20 and 0–50 cm depths. Plants of the northern Mojave Desert, where the NDFF is located, are near the northern distributional limit of several warm desert species (Beatley, 1974) and are exposed to pronounced water and low temperature stresses. Minimum air temperature for January, averaged from 1961 to 1990, is 0°C. Episodic freezing events to −7°C are 50% probable on January 1, and events to −2°C are 50% probable until mid-March. Record extreme subzero temperatures have occurred between November and April. The Mojave Desert is also the most arid desert in North America, and receives primarily winter-spring rainfall. Summer rainfall is highly episodic, and in many years does not occur. With an arid climate, the Mojave Desert is strongly influenced by wet-dry cycles, with annual precipitation below the mean in c. 7 out of every 10 yr.

Gas exchange measurements were concentrated during the active growing season and repeated every 2–4 wk. Less frequent measurements were made during the summer dry season and winter. During four of the six growing seasons, gas exchange data were collected immediately following a subzero temperature night. For each measurement date three plants in each ambient and elevated CO2 ring were selected and their photosynthetic and stomatal conductance rates measured at ambient environmental conditions (photosynthetic photon flux density, temperature) 5–7 times during an entire day. These measurements utilized a LI-6400 open-mode portable photosynthesis system equipped with an LED light source and CO2 mixer (Li-Cor, Lincoln, NE, USA). From these daily courses, daily integrated photosynthetic rates (Aday) were calculated by linear extrapolation between data points and assuming light compensation at sunrise/sunset. For further details see Naumburg et al. (2003). Photosynthetic enhancement ratios were calculated by dividing the plot means for elevated CO2 plants by the plot mean for ambient CO2 plants. Finally, to assess the control of stomatal conductance over a range of environmental conditions, we followed the procedures described in Oren et al. (1999) where stomatal conductance (gs) is regressed against the logarithm of leaf-to-air vapor pressure deficit (VPD). The resultant equation yields a ‘reference’ conductance equivalent to gs at a VPD of 1 kPa (gs ref) and the sensitivity of gs to VPD (slope). These two parameters are related to each other in that plants with higher stomatal conductances (e.g. gs ref) are also more responsive to VPD (i.e. have a greater absolute slope value). This relationship is conserved over a wide range of species and environmental conditions (Oren et al., 1999).

Predawn water potentials were measured on the same day as gas-exchange, but were measured for one plant in each ambient and elevated CO2 ring (n = 3). Thus, gas-exchange and predawn water potentials were not measured on the same plants, which necessitated using treatment means for analyzing the relationship between these variables.

Measurements of chlorophyll fluorescence were obtained with pulse-modulated fluorometers (Models FMS1 and FMS2; Hansatech, Kings Lynne, UK). Measurements were made on the first full-size leaf down from the apical meristem. Leaves were randomly selected on the south-facing side of shrubs, and marked so that measurements could be repeated over the day and year. To assess the responses of leaves to ambient variation in air temperature, leaves were dark-adapted for 20 min to a maximum of 1 h (based on sampling time and travel between plots) using the Hansatech leaf clips. The efficiency of excitation transfer was measured between 0800 and 1000 h as Fv : Fm = (Fm– Fo) : Fm.

Statistical analyses were conducted with SAS (Version 8, SAS Institute, Cary, NC, USA). Throughout, means and standard errors are reported and P < 0.05 is considered significant.


Freezing nighttime temperatures, especially below −5°C, depressed photosynthesis in both ambient and elevated CO2 (Fig. 1). Unfortunately, data below −5°C were collected during the winter of 1998/1999, when significant precipitation did not occur until late spring. These weather patterns resulted in low soil moisture content (Nowak et al., 2004) and predawn water potentials around −4 to −5 MPa (Fig. 2). Similar conditions prevailed in subsequent years, which precluded data collection during high soil moisture and severe frost conditions. Consequently, water stress in conjunction with freezing-induced depression could have caused the low photosynthetic rates. However, for any given predawn water potential, plants that experienced freezing air temperatures had lower photosynthetic rates than plants that did not experience freezing (Fig. 2). This suggests that, even though we cannot predict at what nighttime temperature subsequent photosynthesis will be completely inhibited, freezing suppresses rates, even when plants are water stressed.

Figure 1.

Daily integrated photosynthetic rates of Larrea tridentata during spring/winter as a function of the minimum temperature recorded during the previous night. Data were collected during the course of the day and integrated to produce daily values. Data are means ± SE for n = 3 plants per ambient or elevated CO2 plots.

Figure 2.

Daily integrated photosynthetic rates of Larrea tridentata as a function of predawn water potential for days following freezing or nonfreezing nights. For clarity only data for ambient CO2 plants are shown. Freezing nights were defined as having average air temperatures below 0°C for at least 1 h.

The maximum yield of excitation energy flow through PSII, as measured by Fv : Fm, varied seasonally (Fig. 3). Specifically, PSII maximum yield was lowest in late winter and early spring during times of high water content in combination with low air temperatures and high PPFD, in contrast to early summer when soil and plant water potentials were still relatively high and maximum air temperatures were not yet considered stressful for this hot desert site. However, there was no significant difference in Fv : Fm for plants in elevated compared with ambient CO2 plots (F = 0.001, P = 0.979) across seasons. We observed a maximum Fv : Fm of 0.8 for plants under both elevated and ambient CO2 treatments during the subsequent midday measurements, suggesting that nonphotochemical quenching was able to recover subsequent to freezing nighttime air temperatures.

Figure 3.

Seasonal patterns of apparent quantum yield of photosystem II (Fv : Fm, dimensionless). Data are means ± SE for n = 3 plants per ambient or elevated CO2 plots. Circles, no freezing; squares, freezing; closed, ambient CO2; open, elevated CO2.

Daily photosynthetic enhancement under elevated CO2 was independent from nighttime minimum or daytime measurement temperatures (Fig. 4). Indeed, daily photosynthetic enhancement was closely tied to the ratio of elevated-to-ambient daily average stomatal conductance (Fig. 4). To further explore the temperature and photosynthetic enhancement relationship, we used instantaneous measurements and their associated temperatures (Fig. 5). For this data set we included only measurements where ambient CO2 photosynthetic rates exceeded 3 µmol m−2 s−1 because with lower rates, measurement errors become large, which in turn can cause large errors in ratios. Again, we found no significant relationship between measurement temperatures and photosynthetic enhancement. Even analyzing individual dates separately for a significant temperature/photosynthetic enhancement relationship did not yield consistent trends (not shown). For examples of actual gas exchange rates from a single season, see Table 1.

Figure 4.

Photosynthetic enhancement ratios for daily integrated rates as a function of average leaf measurement temperatures following freezing and nonfreezing nights (a). (b) The same data as a function of the ratio of daily average stomatal conductance under elevated compared with ambient CO2.

Figure 5.

Photosynthetic enhancement ratios for instantaneous rates (Anet) as a function of average leaf temperature. By contrast with Fig. 4, only measurements where ambient Anet exceeded 3 µmol m−2 s−1 were included to reduce the effect of measurement errors on the enhancement ratio.

Table 1.  Environmental and gas exchange data at 0900 h Pacific Standard Time for three measurement dates that included a freezing event and hot/dry conditions
 Treatment13 March 012 April 0122 May 01
  1. Tmin, minimum temperature; Tleaf, leaf temperature at time of gas exchange measurements; Ψpd, predawn water potential; Anet, photosynthetic rates at ambient environmental conditions; gs, stomatal conductance.

Night time Tmin (°C) −1.4  9.5 9.8
Ψpd (MPa)amb.−3.5 (0.3) −2.6 (0.9)−4.2 (0.1)
elev.−3.2 (0.2) −2.2 (0.1)−4.6 (0.5)
Anet (µmol m−2 s−1)amb. 3.0 (0.2)  9.1 (1.9) 3.7 (0.6)
elev. 3.3 (0.9) 16.7 (1.4) 6.6 (2.6)
gs (mmol m−2 s−1)amb.54.1 (6.9)161.7 (20.0)36.8 (7.9)
elev.40.4 (13.5)171.7 (28.8)34.6 (11.9)
Tleaf (°C)amb.14.9 (0.1) 20.2 (0.4)30.1 (0.2)
elev.15.2 (0.4) 19.6 (0.2)28.5 (0.2)

Reductions in daily photosynthetic rates subsequent to nights with freezing air temperatures were driven by stomatal responses. These responses included both a slight reduction in daily maximal stomatal conductance for a given predawn water potential (Fig. 6; F = 8.9, P = 0.005) and greater sensitivity of stomata to VPD (Fig. 7). Elevated CO2 did not affect these two stomatal responses to the environment, because even though gs appeared reduced at a given predawn water potential (Fig. 6), this was not statistically significant in a multiple regression analysis that included CO2 treatments as a dummy variable. Freezing, however, resulted in a significantly lower intercept than nonfreezing nights (88.1 vs 136.5 mmol/m2 s−1; F = 8.9, P = 0.005).

Figure 6.

Daily average stomatal conductance as a function of predawn water potential for days following freezing or nonfreezing nights.

Figure 7.

(a) Examples of the response of two different sets of Larrea tridentata leaves to daily fluctuations in leaf-to-air vapor pressure deficit (VPD) that were used to generate the relationships in (b). (b) Sensitivity of stomata to VPD (slope on y-axis) as a function of the computed ‘reference’ conductance at a VPD of 1 kPa. Circles, no freezing; squares, freezing; closed, ambient CO2; open, elevated CO2.


Contrary to expectations, we found little evidence that elevated CO2 enhanced photosynthetic tolerance of either freezing or high air temperatures in Larrea tridentata (Fig. 4). Rather, ambient and elevated CO2 plants showed similar responses to both stresses, which included the performance of photosystem II (Fig. 3), stomatal conductance (Fig. 7), and photosynthetic rates (Figs 4 and 5). Only on extremely cold and warm days was a substantial enhancement in Aday observed (Fig. 4), but these data points were too infrequent to make a definitive case for photosynthetic enhancement at elevated CO2 during extreme temperature episodes.

PSII activity did not differ for plants on elevated compared with control CO2 plots following exposure to subzero air temperature. At the same time, the maximum yield of PSII, an indicator of photosynthetic performance in response to temperature stress (Larcher, 1995; Loik et al., 2000), was lowest during spring. These results are consistent with those for Norway spruce, for which soluble carbohydrate content and freezing tolerance were not affected by elevated CO2 (Dalen et al., 2001). For several grassland species from different functional groups, Obrist et al. (2001) found an increase in soluble carbohydrates, but a decrease in freezing tolerance. Snow gum (Eucalyptus pauciflora) is one of the most freezing tolerant eucalypts, yet exposure to elevated CO2 resulted in enhanced freezing damage (Lutze et al., 1998). By contrast, our results for seedlings of the Mojave Desert species Yucca brevifolia and Y. schidigera, grown under elevated CO2 in a glasshouse, indicated an increase in freezing tolerance based on uptake of the vital stain neutral red (Loik et al., 2000).

A number of Mojave Desert species exposed to elevated CO2 in a glasshouse showed less down regulation of PSII following a high temperature episode in contrast to plants grown at ambient CO2 concentrations (Hamerlynck et al., 2000a; Taub et al., 2000). For some species, exposure to elevated CO2 can affect high- and low-temperature stress tolerance in different ways. For example, the reduction in growth caused by summertime high temperatures for yellow birch (Betula allegheniensis) is reduced, and dormant buds undergo less wintertime damage, by elevated CO2 (Wayne et al., 1998). Elevated CO2 can also lead to differential responses of photochemistry and photoinhibition. Elevated CO2 and chloroplastic ci are predicted to lead to greater levels of PSII electron flow to photochemical carbon reduction when the demand for assimilates is high, although there is an important interaction with leaf nitrogen content (Hymus et al., 2001). The patterns of PSII electron flow have a strong seasonal component – during periods of active growth, assimilation and linear electron flow are enhanced for Pinus taeda under elevated CO2 (Hymus et al., 1999). However, during February when translocation was minimal, there was less linear electron flow through PSII and greater photoinhibition for plants exposed to elevated CO2 in comparison with controls, and these patterns were the result of altered photochemical quenching, in contrast to the efficiency of energy transfer within PSII antennae (Hymus et al., 1999). The relative contributions of photochemical quenching, sink strength, and thermal dissipation may underlie the seasonal patterns of PSII activity that we observed for Larrea tridentata.

Further, we observed a reduction of maximal stomatal conductance following nighttime exposure to subzero temperatures (Fig. 6) and a greater sensitivity of stomata to VPD (Fig. 7). Both of these responses are consistent with increased hydraulic limitations that necessitate a reduction in transpiration to avoid xylem cavitation (Brodribb & Hill, 2000). Several mechanisms could help explain this observation. First, freezing can directly reduce xylem hydraulic conductivity by inducing embolisms (Sperry & Sullivan, 1992; Sperry et al., 1994). This mechanism has been linked to the northern geographic range limit of Larrea (Pockman & Sperry, 1997). However, Larrea xylem sap freezes only below −5°C and no significant embolisms occur above −7 to −14°C (Pockman & Sperry, 1997; Martinez-Vilalta & Pockman, 2002). Consequently, freezing-induced cavitation does not explain the observed responses over the entire range of air temperatures.

Second, increases in water viscosity with decreasing temperature can effectively reduce hydraulic conductivity and cause decreases in stomatal conductance (Fredeen & Sage, 1999; Brodribb & Hill, 2000). Because water is densest at 4°C, this mechanism does not explain the observed freezing effect on Larrea stomata. Further, daytime temperatures following freezing rise quickly and dramatically during the day, with an observed daytime high of 14.6°C following the coldest night of our measurements (−13.6°C). Also, soil temperatures varied little on a daily basis during spring months, when freezing temperatures are intermittent. Thus, our observations cannot be explained satisfactorily via direct effects on xylem hydraulic conductivity.

Third, low temperatures can decrease the hydraulic conductivity of living tissues as a result of lower membrane diffusivity and aquaporin functioning (Hertel & Steudle, 1997) and finally, low temperatures can act directly on stomata via an increase in Ca2+ uptake from the leaf apoplast, which triggers stomatal closure independently from hydraulic or hormonal signals, a mechanism related to species cold tolerance (Wilkinson et al., 2001). Both of these mechanisms involve temperature-induced changes in membrane structure and functioning, properties that freezing affects strongly (Larcher, 1995; Pearce, 2001; Kozlowski & Pallardy, 2002). Consequently, these mechanisms most likely explain the observed Larrea stomatal responses to freezing above −5°C.

Given the similar responses of ambient and elevated CO2 plants discussed above, it is not surprising that we found no consistent increase in photosynthetic enhancement following freezing temperatures. However, the same reasoning does not hold during high summer temperatures, where a combination of low atmospheric and soil moisture conditions induce water stress, which results in low leaf-internal CO2 concentrations in Larrea (Ogle & Reynolds, 2002). Indeed, much of the variation in photosynthesis during summer can be directly attributed to variation in stomatal conductance (Ogle & Reynolds, 2002). Under these conditions photosynthesis is Rubisco-limited, and we would thus expect elevated CO2 to strongly enhance photosynthesis for physiological reasons (Long, 1991; Ziska, 2001). However, neither instantaneous nor daily photosynthetic rates showed a significant relationship between leaf temperature and photosynthetic enhancement. Instead, photosynthetic enhancement was closely tied to the ratio of elevated-to-ambient stomatal conductance (Fig. 4). Because photosynthesis in desert species is strongly tied to water relations and mediated by stomatal conductance in both ambient and elevated CO2 (Figs 2 and 6; Hamerlynck et al., 2000b; Ogle & Reynolds, 2002), this suggests that stomatal responses override the physiological temperature effect on photosynthesis in determining the relative increase of Larrea photosynthesis to elevated CO2.

In conclusion, the effect of elevated CO2 on Larrea tridentata photosynthetic tolerance of seasonal episodes of low- and high-temperature stress was primarily additive, that is we found little functional difference between ambient and elevated CO2 plants. However, Larrea tridentata showed a wide range of photosynthetic enhancements that were largely driven by stomatal conductance in contrast to PSII function. Differences in stomatal conductance were not related to temperature but have previously been shown to interact with water status (Nowak et al., 2000; Naumburg et al., 2003). These results suggest that the response of guard cell function to environmental conditions in a future higher-CO2 atmosphere will be especially important for arid lands plant ecology as atmospheric temperature continues to increase and precipitation patterns are altered as a result of ongoing anthropogenic activities. Understanding complex anthropogenic effects on plants and ecosystems will become increasingly important in the southwestern United States as urban development continues to expand into sensitive natural habitats.


The authors thank Dene Charlett, David Housman, Travis Huxman, Deb Monical and Eric Knight for help in the field. We further thank three reviewers for their insightful comments. This research was supported by NSF grants DEB-9814358 and DEB-0212812 and DOE grants DE-FGØ2-95ER62127 and DE-FGØ3-ØØER63Ø49.