It Is Hot in the Sun: Antarctic Mosses Have High Temperature Optima for Photosynthesis Despite Cold Climate

The terrestrial flora of Antarctica’s frozen continent is restricted to sparse ice-free areas and dominated by lichens and bryophytes. These plants frequently battle sub-zero temperatures, extreme winds and reduced water availability; all influencing their ability to survive and grow. Antarctic mosses, however, can have canopy temperatures well above air temperature. At midday, canopy temperatures can exceed 15°C, depending on moss turf water content. In this study, the optimum temperature of photosynthesis was determined for six Antarctic moss species: Bryum pseudotriquetrum, Ceratodon purpureus, Chorisodontium aciphyllum, Polytrichastrum alpinum, Sanionia uncinata, and Schistidium antarctici collected from King George Island (maritime Antarctica) and/or the Windmill Islands, East Antarctica. Both chlorophyll fluorescence and gas exchange showed maximum values of electron transport rate occurred at canopy temperatures higher than 20°C. The optimum temperature for both net assimilation of CO2 and photoprotective heat dissipation of three East Antarctic species was 20–30°C and at temperatures below 10°C, mesophyll conductance did not significantly differ from 0. Maximum mitochondrial respiration rates occurred at temperatures higher than 35°C and were lower by around 80% at 5°C. Despite the extreme cold conditions that Antarctic mosses face over winter, the photosynthetic apparatus appears optimised to warm temperatures. Our estimation of the total carbon balance suggests that survival in this cold environment may rely on a capacity to maximize photosynthesis for brief periods during summer and minimize respiratory carbon losses in cold conditions.


INTRODUCTION
Antarctica is considered the coldest continent on Earth, since the surface air temperature can reach annual means of -23°C (-45°C in interior regions higher than 1500 m a. s. l.) (Fortuin and Oerlemans, 1990). However, outside the Antarctic circle the meteorological conditions have been reported to be relatively milder. For instance, in the South Shetlands Islands of Maritime Antarctica the daytime mean air temperatures vary between -5°C and 13°C in the summer and only reach -30°C in winter (Convey and Smith, 2005;Pearce, 2008). In these southern latitudes, terrestrial vegetationmainly lichens and bryophytesis restricted to ice-free areas (Peat et al., 2007;Ochyra et al., 2008). Soil surface temperatures have been recorded to be much warmer than the~2 m air temperatures reported by meteorological stations, with maximum differences of 10.7°C (Schenker and Block, 1986), 25°C (Smith, 1996) or even 27°C (Matsuda, 1968) in summer. In winter, Antarctic mosses will normally be in a dormant state protected by a thick, insulating layer of snow.
Likewise, the microclimate of mosses has been described to be radically different from air temperature recorded in Antarctica (Longton, 1974;Walton, 1982;Edwards and Smith, 1988;Smith, 1988;Hovenden et al., 1994;Melick and Seppelt, 1994;Davey and Rothery, 1997;Green et al., 2005;Pannewitz et al., 2005;Block et al., 2009;Bramley-Alves et al., 2014;Zuñiga, 2016;Convey et al., 2018). Block et al. (2009) reported daily cycles of temperature ranging from 0°C to 44.4°C during the day and then to -2.2°C at night in the moss Andreaea regularis on a rock surface at Signy Island (60°S). Such a variation between the plant surface and air temperatures has been attributed to radiation, the angle of its incidence and wind speed in polar (Wilson, 1957) and alpine (Körner, 2003) environments. The fact that daily temperature variation is so high raises questions about the actual period during which Antarctic mosses have optimum conditions for physiological processes, such us carbon fixation.
Several researchers have reported a high optimum temperature in Antarctic mosses for both CO 2 uptake (Longton, 1988b;Kappen et al., 1989;Davey and Rothery, 1997;Green et al., 2000;Pannewitz et al., 2005;Block et al., 2009) and O 2 evolution (Rastorfer, 1970;Ino, 1990;Wilson, 1990;Smith, 1999;Newsham, 2010), suggesting that the studied species were not truly psychrophilic. Longton et al. (1988a;1988b) observed high optimum temperatures for Antarctic mosses but with a broad curve resulting in a positive intercept at 0°C . In contrast, some authors have considered bryophytes to generally have lower temperature optima for photosynthesis at about 5-15°C (He et al., 2016 and references therein). Although, Ino (1990) pointed out that only mosses inhabiting locations that were frequently submersed in cold water had low optimum temperatures for photosynthesis. So far, very few researchers have addressed the effect of the high specific heat capacity of water on the maximum temperatures experienced by mosses during a daily cycle (Block et al., 2009). Water availability has been suggested to be more relevant than temperature to biology (Kennedy, 1993) and to be the main factor ruling species distribution in Antarctica (Davey and Rothery, 1997;Convey et al., 2014). However, the interactive effect that water and temperature have in providing a favourable environment for positive carbon balance in Antarctica remains unknown.
The effect of temperature on photosynthesis has been thoroughly studied (Sage and Kubien, 2007;Flexas et al., 2014;von Caemmerer and Evans, 2015). Several biochemical and biophysical processes involved in photosynthesis are affected by temperature: (1) thylakoid membrane fluidity (Hirano et al., 1981), (2) kinetics of electron transport and the Calvin-Benson cycle (Holaday et al., 1992;Sage, 2002;Walker et al., 2013), (3) stomatal aperture, and (4) CO 2 diffusivity in membranes and aqueous/wall phase of mesophyll (as a sum, termed mesophyll conductance, g m ). g m is one of the more relevant limiting factors of photosynthesis, especially in bryophytes, which present the lowest values of g m of the plant kingdom (Flexas et al., 2012;Gago et al., 2019). The short-term response of g m to temperature is considered species-specific (Bunce, 2008;von Caemmerer and Evans, 2015), although most of the studied species have shown a decrease of g m at low temperatures (Bernacchi et al., 2002;Warren and Dreyer, 2006;Scafaro et al., 2011;Ubierna et al., 2017; but see also Qiu et al., 2017). The response of g m to temperature, however, is unknown for bryophytes.
As a consequence of reduced photosynthetic rates at low temperatures is that the capacity to use light energy in carbon assimilation will decrease and saturation will occur at lower irradiances (Huner et al., 1993;Ensminger et al., 2006). Many photoprotection mechanisms have been described in both vascular plants and bryophytes to avoid photodamage of photosynthetic apparatus in this situation (Robinson and Waterman, 2014). One of the more short-term dynamic mechanisms of photoprotection is the regulated heat dissipation of excess energy (estimated by a chlorophyll fluorescence parameter termed non-photochemical quenching, NPQ), which is dependent on the build-up of a gradient in pH across the thylakoid membrane and has been reported to broadly increase during acclimation to low temperatures (Hendrickson et al., 2004;Mıǵuez et al., 2015;Yang et al., 2018). To our knowledge, despite its relevance to understanding the limitations of photosynthesis at low, variable temperatures in Antarctic mosses, neither the temperature response of photoprotective heat dissipation nor g m has so far been reported for bryophytes (with the exception of unsteady-state NPQ for one Mediterranean moss species in Deltoro et al., 1999).
Thus, the objectives of this study were (1) to model the daily carbon balance of Antarctic mosses during summer based on canopy surface temperature and its effect on photosynthesis, (2) to test the interspecific differences and the possible buffering effect of water content on moss canopy temperature, and (3) to determine the temperature responses of net CO 2 assimilation, electron transport rate, g m and photoprotection mechanisms assessed by NPQ.

Study Site and Plant Material
Two Antarctic locations were included in this study: Casey station (66°16′57″S, 110°31′36″E) on Bailey Peninsula (Windmill Islands region, East Antarctica) and Fildes Peninsula (62°12′05″S, 58°57′ 44″W) on King George Island (South Shetland Islands) ( Figure 1). According to the Australian Bureau of Meteorology, mean maximum and minimum air temperature at Casey Station during the hottest month of the Antarctic summer (January) is 2.3 and -2.5°C , respectively, whereas in winter (July) mean maximum temperatures can drop to -10.8°C (data from 1989 to 2019). Fildes Peninsula in Maritime Antarctica present similar air temperatures in summer with mean max/min temperatures of 2.8/0.1°C (data from 1969-2012, for Bellingshausen Station, consistent with data reported for Frei Montalva Station by Carrasco and Gonzaĺez, 2007).
Six species of bryophytes were studied during different Antarctic campaigns (Table 1). Bryum pseudotriquetrum, Ceratodon purpureus and Schistidium antarctici were found on Bailey Peninsula near Casey Station (see Robinson et al., 2018 andKing et al., 2020 for detailed maps and site descriptions). B. pseudotriquetrum, S. antarctici, Chorisodontium aciphyllum, Polytrichastrum alpinum and Sanionia uncinata were located at Fildes Peninsula and Ardley Island near Escudero Station. Mosses were identified to species by ACK and MJW (King George Island) and SR, JBA, and JDT (Windmill Islands). Specimen vouchers of each species were deposited in either the Janet Cosh Herbarium (University of Wollongong, Australia) or the CONC Herbarium (Universidad de Concepcioń, Chile).

Microclimatic Conditions and Moisture Effect
Daily Moss Surface Temperature Near Casey Station, East Antarctica The surface temperature of East Antarctic B. pseudotriquetrum, C. purpureus, and S. antarctici was recorded during the Antarctic summer of 2003 (from 16/01/2003 to 28/01/2003) at six locations around Casey Station (ASPA 135). At each location a polyurethane sealed iBCod temperature sensor (Thermodata Pty. Ltd., Brisbane, Australia) was placed on the moss surface for 13 days of continuous recording. The frequency at which mosses experienced a determined interval of temperature (intervals of 2°C from -4 to 28°C) was calculated as: Time (%) = 100 · t/T, where t is the number of records of each interval of temperature and T is the total of records for each species.
For the same species and locations, additional spot measurements of photosynthetic photon flux density (PPFD), wind gust speed, moss surface temperature, and air temperature were collected at midday (local time, UTC +13) for a wider period (22 days distributed from 9/11/2002 to 01/02/2003). Moss surface temperature was measured with an infrared thermometer (Scotchtrack T Heat tracer IR1600L; 3M, Austin TX, USA). A high correlation (r = 0.914) was found between iBCod and infrared thermometer recordings for four random days when samples were measured with both sensors at midday (Supplemental Figure 1). PPFD was measured with a LS-C mini quantum sensor attached to the leaf clip holder of a Walz MINI-PAM Photosynthesis Yield Analyser (WALZ, Effeltrich, Germany) placed at moss surface level during measurement. Wind gust speed and air temperature were obtained from the Australian Bureau of Meteorology at Casey station.

Daily Moss Surface Temperature on Fildes Peninsula and Ardley Island, Maritime Antarctica
Daily air and moss surface temperature of maritime Antarctic B. pseudotriquetrum, C. aciphyllum, P. alpinum, and S. uncinata were recorded during the Antarctic summer of 2019 (from 08/ 01/2019 to 30/01/2019) at seven locations around Ardley Island (ASPA150). In this case, one HOBO 4-channel thermocouple datalogger (UX120-014M, Onset Computer Corporation, Bourne, MA, USA) was placed at each location, so that the moss surface temperature of three specimens and air temperature could be recorded simultaneously. The frequency with which mosses experienced a particular temperature interval

Environmental Moisture Effect on Moss Surface Temperature
In order to test the effect of water on the seasonal shift in moss surface temperature, 90 iBcod and iButton sensors (Maxim Integrated, San Jose, USA) were deployed across water gradients on top of cushions of S. antarctici at the Red Shed site at Casey Station. Three environments were described: wet, dry and intermediate. Wet environments were located adjacent to a meltwater stream ( Figure 2); dry environments were located 2 m from the water edge, with intermediate environment located in the middle. Moss turf water content (TWC) of each environment was estimated after Lucieer et al. (2014) and King (2017)

Electron Transport Rate at Casey and Escudero Station
In order to determine the optimum temperatures for electron transport rate (ETR) of the moss species, measurements of chlorophyll fluorescence were performed under lab conditions with a Walz MINI-PAM Photosynthesis Yield Analyser fitted to a Walz external halogen lamp (FL 400). Fresh samples of moss tissue were collected from sunny microhabitats on Fildes Peninsula, King George Island (January 2015) and Casey Station, Antarctica (December 2011) and measured within 2 days. Moss samples were maintained under natural sunlight and temperature levels prior to measurement. Moss cushions were divided into moss plugs of 1-2 cm 2 diameter and subjected randomly to temperatures from 5 to 40°C in groups of 5 replicates per temperature curve (from 4 to 28°C for Casey measurements) in the laboratory. Replicate plugs were maintained in aluminum cups in a water bath set to the target temperature and moss surface temperature were monitored. The aluminum cups allowed heat transfer within the water bath but prevented submergence and ensured the photosynthetic surface of the moss remained exposed to air. At each temperature, the specimens were pre-illuminated at PPFD of~100 mmol photons m -2 s -1 while moss temperature equilibrated with the water temperature. They were then maintained at temperature for 5 min before a rapid light response curve was performed. Thermocouples were used to measure temperature of the photosynthetic tissue before and after each light response curve. ETR was calculated according to Krall and Edwards (1992): ETR = f PSII · PPFD · ab, where f PSII is the yield of PSII and ab is the product of absorbance and the partitioning of absorbed quanta between PSI and PSII. f PSII was calculated according to Genty et al. (1989): where F s and F m ' are the steady-state and maximal fluorescence at light adapted conditions, respectively.
Since ab was unknown for this set of measurements, a provisional value of 0.42 was used (Maxwell and Johnson, 2000) and was assumed to remain constant with temperature. The maximum light-saturated ETR (ETR max ) was obtained by fitting each light curve to a rational model (Smith, 1936) or to the waiting-in-line model (Ritchie, 2008) by using the Microsoft Excel Solver tool (adapted to ETR light curves from Lobo et al., 2013). The lowest square sum errors of a non-photoinhibited light curve were obtained with rational model (eqn 1), while the waiting-inline model (eqn 2) was used for photoinhibited curves.
where AQE is the Apparent Quantum Efficiency, also fitted by the model.

Mesophyll Conductance and NPQ Measurement
In order to determine optimum temperature for CO 2 assimilation (A N ) at saturating light (A sat ) and g m , 6-7 samples of approximately 3 cm 2 of three study species -S. antarctici, C. purpureus and B. pseudotriquetrumwere collected near the Red Shed at Casey Station in 2014. Each sample was air-dried and stored at -20°C until analysis. In 2018, samples were thawed and rehydrated with distilled water for 10-16 h in dark conditions at 4°C prior to measurement under laboratory conditions. Prior to further measurements moss health was assessed by chlorophyll fluorescence, with high maximum quantum yield of PSII (F v /F m ) indicating full recovery.
All gas exchange measurements were performed with a LiCOR 6800 system (LiCOR Biosciences, Lincoln, NE, USA). Between 40 and 47 dark-adapted sub-samples of each species were introduced into a custom-made cuvette consisting of a gasket affixed to a piece of thin polyester stocking fabric (Supplementary Figure 3). The size of these gaskets was equal to the chamber size to ensure proper closure of the chamber and achieve a minimum CO 2 leakage (Supplementary Figure 4). CO 2 concentration was standardized at 400 mmol CO 2 ·mol -1 air, relative humidity at 60%-75% and the flow rate within the chamber was 700 μmol·s -1 . The temperature of the chamber was varied between 5-35°C in steps of 5°C (n = 6-7 for each temperature of S. antarctici, and B. pseudotriquetrum; n = 4-5 for C. purpureus). After 5 min inside the chamber in dark conditions, dark respiration (R D ) was measured and a saturating pulse was applied in order to measure basal and maximum chlorophyll fluorescence (F 0 and F m , respectively) and to calculate F v /F m = (F m -F 0 )/F m . Then, the sample was exposed to saturating red light, with a maximum emission at 625 nm (800 μmol·m -2 ·s -1 , which was determined with partial light curves performed a priori). A sat at minimum saturating light was recorded at steady-state conditions, when diffusion limitations due to interstitial water were null and biochemistry was fully light-adapted (Supplementary Figure 5). At this stage, a second saturating pulse was applied to determine f PSII (as explained above) and non-photochemical quenching (NPQ) at each temperature, the latter calculated according to Bilger and Björkman (1990): NPQ = (F m -F m ')/F m '. ETR was calculated as explained in Daily Moss Surface Temperature on Fildes Peninsula and Ardley Island, Maritime Antarctica without assuming a constant ab. For the samples where A sat was also measured, ab was determined as 4/slope of the relationship between f PSII and f CO2 ((A N + light respiration)/PPFD) obtained by varying PPFD under non-photorespiratory conditions in an atmosphere containing <1% O 2 (Valentini et al., 1995) (n = 3-5 light curves per temperature and species). Since the ratios f PSII and f CO2 increased significantly at higher PPFD, the ab of the highest PPFD (the same used for measuring A sat ) was chosen for calculating ETR. Light respiration (R L ) was calculated from the initial light-limited portion of the low-O 2 -light curves as the negative intercept of the relationship between A N and (f PSII · PPFD)/4 according to Yin et al. (2011). In order to avoid desiccation, if required, the sample was fully rehydrated before the low-O 2 -light curves were measured; by immersing in distilled water for 1-2 min with excess water removed gently with a paper tissue before placing the sample back in the chamber.
g m was estimated according to Harley et al. (1992) with the modifications of Carriquı́et al. (2019): where G* is the chloroplastic hypothetical CO 2 compensation point in the absence of respiration and C a is the atmospheric CO 2 concentration. Since stomata are absent in gametophytes of bryophytes, stomatal CO 2 concentration (C i ) is substituted by C a in Harley original formula. G* was calculated from the Rubisco specificity factor (S C/O ) as: S C/O was averaged from the bryophytes species reported by Font et al. (Font and Galmeś, 2016). The temperature coefficient Q 10 was calculated for intervals of linear g m -temperature as follow (Van't Hoff, 1884): The relative mesophyll (l m ) and biochemical (l b ) limitations to photosynthesis were calculated according to Grassi and Magnani (2005) with the modifications of Carriquı́et al. (2019): As a proxy to dA/dC c the quotient A N /C c at 400 μmol CO 2 · mol -1 air was calculated.

Estimation of Carbon Gain
Estimations of the carbon gain of East Antarctic B. pseudotriquetrum, C. purpureus, and S. antarctici were made by combining the surface temperature recorded at Casey during the summer of 2003 (see Daily Moss Surface Temperature Near Casey Station, East Antarctica) and the temperature response curves of A sat measured in lab conditions for the same species collected from the same location during the summer of 2014 (see Mesophyll Conductance and NPQ Measurement).
Surface temperature values of S. antarctici at different moist environments at Casey during the summer of 2011/12 and 2013 were also analyzed. Each interval of temperature experienced by mosses was assigned a corresponding mean value of A sat (calculated from the polynomial curve fitting of the corresponding temperature response of A sat ) and its contribution to the total net CO 2 assimilation over the study period (A sat,T ) was estimated as: A sat,T = A sat · f, where f is the frequency of time at which this interval of temperature was recorded. The balance between net carbon fixation (sum of positive A sat,T ) and carbon lost (sum of negative A sat,T ) during the studied period was calculated per species as: These calculations were done by assuming that: (1) all high temperatures are experienced under high light conditions, (2) A sat at the lowest temperatures (<4°C) is negative (as a conservative worst case scenario) and similar to a ratio of R D experienced at 5°C (three scenarios were modeled based on 50%, 33.3%, 25% and 5% of R D at 5°C), and (3) the water content of the measured specimens allowed optimum gas exchange.

Microclimate Data
Differences between species in microclimate data (see Daily Moss Surface Temperature Near Casey Station, East Antarctica and Daily Moss Surface Temperature on Fildes Peninsula and Ardley Island, Maritime Antarctica) were tested after logarithmic transformation of daily mean, maximum, and minimum temperatures by analysing a mixed ANOVA where localization of the sensors and date were considered as random variables and species as fixed factor. One-way ANOVA was performed to test the effect of the water content on the moss surface temperature (daily mean, maximum and minimum). The relationships between PPFD or wind gust speed and the difference between moss surface temperature and air temperature were tested by Pearson correlation test.

Chlorophyll Fluorescence and Gas Exchange Data
ETR max /temperature curves were fitted to a 3-degree polynomial equation. Optimum temperatures for ETR max were obtained for each sample of the studied species by determining where the 1 st derivative of fitted polynomials was zero. Then, one-way ANOVA was used to test differences between species in optimum temperature for ETR max . The effect of temperature on gas exchange derived parameters in each species was tested by two-way ANOVA. The relationships between the parameters l b , l m , ab, and F v /F m with temperature were tested by Pearson correlation test.

Microclimatic Conditions and Moisture Effect
Daily Moss Surface Temperature Near Casey Station, East Antarctica Microclimate data of B. pseudotriquetrum, C. purpureus, and S. antarctici recorded during the summer of 2002/3 near Casey Station (ASPA 135) are shown in Figure 3. No significant effect of species on daily mean, maximum, and minimum temperatures was observed (P = 0.564, 0.995, and 0.795, respectively). The highest mean temperatures of the mosses surface were obtained between 10:00 and 13:00 (local time, UTC +13) and reached values around 11°C in the three species. Although maximum mean temperatures were only on average 2.3°C higher than mean surface temperatures, absolute maximum temperatures of 19, 18, and 17°C were recorded at midday for B. pseudotriquetrum, C. purpureus, and S. antarctici, respectively. At night, mean surface temperatures remained positive in the three studied species and only absolute minimum temperatures declined to -1°C. At midday, absolute daily minimum moss surface temperature was never below +4°C. During 13 days of measurements over the peak Antarctic summer (January), the mosses experienced temperatures below +4°C for 56.6% of the time. Moss temperatures exceeded 14°C for an average of just 2.5% of the time.
The difference between moss surface temperature and air temperature at midday was predominantly driven by solar radiation (Figure 4A, P<0.05, R 2 = 0.546) and in a weaker but significant way by maximum speed of wind gust ( Figure 4B P<0.05, R 2 = 0.165). On clear days, when irradiation exceeded 1,000 μmol·m -2 ·s -1 , air temperature and moss surface temperature reached a mean maximum difference of 16.2°C for all three species, with the highest absolute maximum difference recorded for C. purpureus (22.3°C).

Daily Moss Surface Temperature on Fildes Peninsula and Ardley Island, Maritime Antarctica
Microclimate data for the maritime Antarctic species studied during summer 2019 on Fildes Peninsula and Ardley Island are shown in Figure 5. No significant differences between species surface temperatures were found (P = 0.785, 0.461, and 0.972 for mean, maximum and minimum moss surface temperature, respectively

Environmental Moisture Effect on Moss Surface Temperature
Mean maximum and minimum surface temperature of S. antarctici were significantly affected by hydration status ( Figure 6). Thus, dry canopy temperatures reached significantly higher maximum (15.0 ± 0.9 and 21.9 ± 0.9°C for data of 2011/12 and 2013, respectively) and lower minimum (-3.0 ± 0.4 and -3.1 ± 0.4°C) daily mean temperatures. While in intermediate and wet environments, extreme temperatures were more buffered. Mean temperatures were also significantly higher in dry sites during the summer of 2013 but not in the previous summer.

Temperature Responses of Chlorophyll Fluorescence and Gas Exchange
ETR light curves at each studied temperature for species from East and maritime Antarctic locations are shown in Supplementary  Figures 6 and 7. The ETR max from rational and waiting-in-line light curve models (Supplementary Figure 8) gave maximum values at temperatures between 19-26.3°C ( Table 2). The only endemic species studied, S. antarctici, had the lowest optimum temperature for ETR (19.00 ± 0.9°C), followed by C. purpureus (21.3 ± 1.9°C) a cosmopolitan species. Polytrichastrum alpinum, which is associated with polar and alpine habitats, showed the highest optimum temperature (26.3 ± 0.7°C). No significant difference was found between optimum temperatures of B. pseudotriquetrum measured at the two study sites. Figure 7 shows the change in gas exchange and associated fluorescence parameters with temperature. The highest A sat were recorded at 25-30°C by B. pseudotriquetrum and C. purpureus, meanwhile S. antarctici showed its optimum A sat at 20-25°C ( Figure 7A). Interestingly, B. pseudotriquetrum could not maintain a positive carbon balance at 5 or 10°C. R D was also strongly inhibited at low temperatures, showing reductions of around 80% at 5°C in the three studied species. Conversely maximum values were found at the highest tested temperature, 35°C ( Figure 7B).
The optimum temperatures for electron transport were also above 25°C in the three species studied ( Figure 7C). The absorbance of PSII used for calculating ETR did not vary significantly across temperature (R 2 < 0.1 in all species). Maximum NPQ occurred at 20°C or higher temperatures ( Figure  7D), whereas F v /F m varied between 0.574 and 0.794 independent of the temperatures and the species (R 2 < 0.09 in all cases).
The estimated mesophyll conductance also showed maximum values at high temperatures -B. pseudotriquetrum, C. purpureus and S. antarctici showed optimum g m values at >30, 30, and 15-35°C, respectively ( Figure 8). g m values at 5 and 10°C were not significantly different from zero. At low temperatures the diffusional limitation due to the mesophyll (l m ) increased significantly in opposition to biochemical limitation (l b , P < 0.001 in B. pseudotriquetrum, P = 0.017 and 0.008 in S. antarctici and C. purpureus, respectively). Q 10 of g m in B. pseudotriquetrum, C. purpureus, and S. antarctici were calculated as 1.38, 6.6, and 1.31, respectively (using mean values of g m between 15 and 25°C).

Estimation of Carbon Gain
Based on the temperature data from the field (Figure 3), Windmill Island moss surface temperatures only exceeded 14°C for 2.5% of the time during midsummer. However, modeling of carbon gain using temperature and gas exchange data shown in Figure 7A indicate that 36.7%, 12%, and 8.4% of total positive net CO 2 was fixed during this short period in B. pseudotriquetrum, C. purpureus, and S. antarctici, respectively (data not shown). The total balance of positive and negative net CO 2 exchanged (lost vs fixed) is shown in Table 3. A positive carbon balance -i.e. when carbon fixation exceeds 50% of the CO 2 exchanged-was only obtained when inhibition of respiration is high for temperatures below 4°C (33.3% and 25% of the R D at 5°C in C. purpureus and S. antarctici and 5% in B. pseudotriquetrum).
When carbon balance was modeled for S. antarctici across a hydrological gradient during the summer of 2011/12 and 2013 in East Antarctica (Figure 6), positive carbon balance was obtained for dry and intermediate environments under most R D scenarios (except when 50% of R D at 5°C was modeled for dry environment in 2011/12 and intermediate moist environment in 2013) ( Table 4). On the contrary, the wet environments only presented positive carbon balance when the inhibition of R D was the highest modeled (5% of R D at 5°C in 2011/12 and below 25% in 2013).

DISCUSSION
In line with previous research, our study verified that Antarctic bryophytes are not psychrophilic plants, since all the measured species presented optimum temperatures for ETR max and A sat  A B D C FIGURE 7 | Temperature response curve of three East Antarctic moss species for (A) saturating net CO 2 assimilation (A sat ), (B) dark respiration (R D ), (C) maximum electron transport rate (ETR max ), and (D) non-photochemical quenching (NPQ). Inset graph of (C) shows variation in the product of absorbance and partitioning of photons (ab) used for calculation of ETR across temperature (R 2 = 0.068, 0.063, and 0.095 for C. purpureus, S. antarctici, and B. pseudotriquetrum, respectively). Inset graph of (D) shows stability of F v /F m across temperature (R 2 = 0.086, 0.028 and 0.046 for C. purpureus, S. antarctici and B. pseudotriquetrum, respectively). P values results from two-way ANOVA are shown in the upper-left corner of each graph. Mean ± se (n = 4-6, except for the light curve derived parameter ab, where n = 3-5).  between 19 and 26.3°C, as has been reported for mesic and tropical bryophytes (Dilks and Proctor, 1975;Furness and Grime, 1982;Glime, 2011;Wagner et al., 2013;He et al., 2016). Despite the relationship between mean temperatures experienced during the growing season and temperature optima of net photosynthesis reported from polar, alpine, temperate, desert and tropical ecosystems (Wagner et al., 2013), high specific plasticity can be observed for Antarctic species ( Table 5). The lowest optimum temperature has been found in the endemic species S. antarctici between 0 and 10°C (Kappen et al., 1989;Davey and Rothery, 1997;Block et al., 2009), although in the present study ETR max of this species was obtained between temperatures of 19-30°C, also previously reported for O 2 evolution (30°C, Wilson, 1990). High optima for photosynthesis have also been found in non-endemic species with polar distributions, such as Hennediella heimii (previously Bryum antarcticum) (19°C, Rastorfer, 1970), or species such as P. alpinum which show a moderate bipolar distribution with transitional populations in alpine environments, (T opt = 26.3°C here), but note (T opt = 10°C) in a previous study by Davey and Rothery (1997). This suggests that possessing low optimum temperatures for photosynthesis is not a strict requirement for surviving in Antarctic environments (at least until 63-64°S), even in species that restrict their distributions to these habitats (e.g., S. antarctici).
The fact that all Antarctic mosses measured showed such high temperature optima for photosynthesis even when summer mean maximum temperatures are much cooler (2.3°C in January at Casey Station) suggests that moss surface temperatures must regularly exceed air temperatures. This was evidenced in our microclimate analysis where the temperature experienced by mosses was measured with iBCods, infrared thermometers and/or thermocouples. Ecological researchers have pointed out the importance of a deep description of microclimate in understanding and modeling present and future species distribution and ecosystem functioning, specially in smallstature species (Convey et al., 2018;Lembrechts et al., 2019;Lembrechts and Lenoir, 2020). In our study, around midday moss surfaces were elevated above mean air temperatures by 16.2 (22.3)°C (Figure 4) enhanced by high irradiation and low wind speed, as has been described for Arctic and alpine ecosystems (Wilson, 1957;Körner, 2003). However, most of the time Antarctic mosses experienced suboptimal conditions for photosynthesis (during the night or cloudy/windy days) such that their surface temperatures exceed 14°C only 2.5%-3% of the time. This is a lower percentage of time than reported by Smith (1988), where moss temperatures exceeded 20°C 24% of the time. Even so, the percentage of time at which the temperature of mosses allows a positive net CO 2 assimilation must be enough to compensate for loss of carbon by respiration in order to achieve the very low growth rates (average 1.33 mm per year) reported for these Antarctic mosses (Clarke et al., 2012). In future, under climate change, Antarctic mosses are also expected to experience an increase in air temperatures and this would be expected to lead to an increase in the percentage of time they spend at optimal temperatures . Our current estimation of carbon balance suggests that carbon balance can only be positive if a large reduction of carbon loss by respiration is assumed for the lowest temperatures ( Figure 3). In environments with high nocturnal temperatures, such as tropical regions, bryophytes can lose more than 50% of the CO 2 fixed during the daytime each night (Zotz et al., 1997). So, moss survival in Antarctica may be more related to an ability to  For temperatures below 4°C a conservative negative net assimilation of 50%, 33.3%, 25%, or 5% of the respiration in the dark (R D ) at 5°C was assumed. Bold shows positive carbon balance.
inhibit respiration at low temperatures, rather than having lower optimum temperatures for photosynthesis. Another factor that affects carbon gain of many bryophytes is water availability (Proctor, 1982). Both dehydration and an excessive interstitial water content can inhibit photosynthesis (Smith, 1982;Rice et al., 2011;Wagner et al., 2013;Supplementary Figure 5). Our study confirms that water content also influences the temperatures experienced by Antarctic mosses, with excess water buffering the extremes (decreasing their maximum and increasing their minimum temperatures) and, therefore, reducing the time when mosses experience temperatures higher than 14°C. This can be explained by the high specific heat capacity of water, as has been suggested previously by Pannewitz et al. (2005) and Block et al. (2009) for Antarctic mosses and by soil researchers (Campbell et al., 1995).
Thus, provided the moss cells remain hydrated and the drier the interstitial environment, the wider the window for positive net CO 2 assimilation, since maximum temperatures are closer to optimum for photosynthesis and minimum temperatures are enough to substantially inhibit the loss of CO 2 by respiration (see Figure 4 for estimation of carbon balance). However, the optimum water content for maximum carbon gain of these Antarctic mosses is still unknown, and the interaction of limitations by both temperature and water content should be analyzed in the future.
The effect of low temperatures on photosynthesis in the Antarctic species studied here was mainly driven by diffusional limitations, rather than biochemical ones, as has been reported for Antarctic vascular plants (Saéz et al., 2018) and by other important environmental stresses such as water stress (Flexas  Nakatsubo (2002) and Kallio and Heinonen (1975) are excluded because measurements of over-watered samples were clearly performed at non-steady state conditions and variation of CO 2 assimilation might be associated with a variation of interstitial water during measurements rather than with temperature. Nadal and Flexas, 2018). Variable responses of g m to temperature have been reported for vascular plants (Bernacchi et al., 2002;von Caemmerer and Evans, 2015;Xiong et al., 2015;Huang et al., 2017). In our study, B. pseudotriquetrum, S. antarctici, and C. purpureus showed increasing g m with temperature and only in the latter was a decline at supraoptimal temperatures observed. The components of mesophyll conductance that rule its response to temperature are not well understood (Shrestha et al., 2019). Since the temperature coefficient reported for CO 2 diffusion in pure water (Q 10 = 1.25, Jähne et al., 1987) is lower than that reported for g m in vascular plants (1.8-2.2) (Bernacchi et al., 2002;Yamori et al., 2006), physical diffusion alone cannot explain the variation of g m with temperature. Instead, it has been hypothesised that CO 2 diffusion through both liquid phase (cell wall, cytosol and chloroplast stroma) and membranes (plasmic and chloroplastic, facilitated by protein transporters, i.e., aquaporins) is affected by temperature (Bernacchi et al., 2002;Evans and Von Caemmerer, 2013;Walker et al., 2013;von Caemmerer and Evans, 2015). The Q 10 of C. purpureus, B. pseudotriquetrum, and S. antarctici was calculated as 6.6, 1.38, and 1.31, respectively. This suggests that the role of any facilitated process for CO 2 diffusion is highly variable and is enhanced more in C. purpureus than in either the latter two species or the reported vascular plants.
At suboptimal temperatures photosynthesis in these mosses is unlikely to be able to utilise all the absorbed light producing an energy imbalance. Despite this, photoprotective heat dissipation (here estimated by NPQ) decreased when saturating light was combined with short-term exposure to suboptimal temperatures (optimum temperatures for NPQ = 20°C for S. antarctici and 25°C for C. purpureus and B. pseudotriquetrum). Lovelock et al. (1995) reported similar results for S. antarctici, which decreased NPQ (expressed as q N ) after 2 h of 5-0°C. Only when temperatures were below the freezing point of -7°C (Melick and Seppelt, 1992), was deepoxidation-independent (dithiothreitol insensitive) NPQ significantly increased, as has been observed for Antarctic lichens (Bartaḱ et al., 2007). Previous research into short-term changes of NPQ with suboptimal temperature (both cold and heat stress) in mesic and Antarctic species has reported various results including: (1) an increase in maximum steady-state NPQ (Xu et al., 1999;Hendrickson et al., 2004;Sinsawat et al., 2004;D'Ambrosio et al., 2006;Savitch et al., 2009;Sharkey and Zhang, 2010), (2) a decrease of NPQ, as in our study, (Bilger and Björkman, 1991;Fracheboud and Leipner, 2003;Corcuera et al., 2005;Lambrev et al., 2007;Peŕez-Torres et al., 2007;Wang et al., 2009), or (3) invariable NPQ (Peŕez-Torres et al., 2007). No correlation between de-epoxidase state or zeaxanthin concentration and NPQ at low temperatures (Xu et al., 1999;D'Ambrosio et al., 2006) and a decrease in the percentage of NPQ inhibited by dithiothreitol (Xu et al., 1999) suggest that the NPQ that is enhanced in the short-term at low temperatures could consist of zeaxanthin-independent heat dissipation (Johnson et al., 2009) and/or photoinactivated PSII reaction centre heat dissipation (Krause and Weis, 1988;Lee et al., 2001;Ivanov et al., 2003). This would be consistent with the fact that violaxanthin de-epoxidase is inhibited at low temperatures (Bilger and Björkman, 1991;Szilaǵyi et al., 2007) and the associated lower electron transport rates will compromise the generation of DpH, which is required for activation of violaxanthin de-epoxidase and inhibition of zeaxanthin epoxidase (Gilmore, 1997;Goss et al., 2008). The decrease of NPQ at low temperatures in our study could be a direct consequence of the inhibitory processes described above and not/less related with zeaxanthin-independent or photoinactivated PSII heat dissipation. Furthermore, maintenance of a constitutive zeaxanthin/lutein pool, which is a common mechanism for enhancing NPQ during coldhardening (Haldimann et al., 1996;Leipner et al., 1997;Faria et al., 1998;Venema et al., 2000;Caffarri et al., 2005;Ivanov et al., 2006;Saéz et al., 2019; but see also Savitch et al., 2002), is less likely to be present in these short-term experiments. East Antarctic S. antarctici, C. purpureus, and B. pseudotriquetrum exhibit high zeaxanthin content in the field as a result of this cold hardening (Lovelock and Robinson, 2002). However, the presence of zeaxanthin alone without the generation of a pH gradient is insufficient to induce fluorescence quenching (Bilger and Björkman, 1991;Hurry et al., 1997;Hwang et al., 2003;Goss et al., 2008). Thus, questions remain about the photoprotection role of high zeaxanthin levels in Antarctic mosses at low temperatures, given they are not associated with the build-up of a transthylakoidal DpH which normally induces regulated heat dissipation. Screening and/or antioxidant roles of zeaxanthin may need to be considered (Havaux et al., 2007;Solovchenko, 2010). However, at high temperatures and high light levels, a high and sustained concentration of zeaxanthin could help to enable a rapid photoprotection response of heat dissipation at temperatures close to the photosynthetic optima.

CONCLUSION
We conclude that Antarctic mosses are not psychrophilic plants, since their photosynthetic optima occur at relatively high temperatures. A positive carbon gain can be maintained providing respiration is strongly inhibited at low temperatures. However, the interaction of limitations by both temperature and water require further study, since the moisture of the moss environment influences the temperatures at which they metabolize. At low temperatures, NPQ was not enhanced and the decline in photosynthesis was largely caused by an increase of diffusional limitations, which also suggests the existence of facilitated and variable processes for CO 2 diffusion in these mosses.

DATA AVAILABILITY STATEMENT
All data associated with this manuscript is available in Australian Antarctic Data Centre (https://data.aad.gov.au/metadata/ records/AAS_4046_TempOptima_Frontiers_Perera-Castro).