Introduction
Soil seed banks play a crucial role in the population dynamics of many plant communities (Baskin and Baskin, Reference Baskin and Baskin1998). They guarantee natural regeneration and persistence of ecosystems because soil seed banks provide a stock of viable seeds either on the surface or buried in the soil (Cheib and Garcia, Reference Cheib and Garcia2012), mainly in habitats with variable disturbance regimes and unpredictable environmental conditions (Fenner and Thompson, Reference Fenner and Thompson2005), such as arid environments (Montiel and Montaña, Reference Montiel and Montaña2003; Jurado and Flores, Reference Jurado and Flores2005; Cheib and Garcia, Reference Cheib and Garcia2012; Álvarez-Espino et al., Reference Álvarez-Espino, Godínez-Álvarez and De la Torre-Almaráz2014). Global climate models forecast that arid environments might experience both increased temperatures and a decline in frequency and magnitude of rainfall events (IPCC, Reference Stocker, Qin, Plattner, Tignor, Allen, Boschung, Nauels, Xia, Bex and Midgley2013, Reference Field, Barros, Dokken, Mach, Mastrandrea, Bilir, Chatterjee, Ebi, Estrada, Genova, Girma, Kissel, Levy, MacCracken, Mastrandrea and White2014). To understand the long-term effects of climate change in plant species, it is necessary to link future environmental changes to mechanisms that control plant regeneration such as seed germination, dormancy (Ooi et al., Reference Ooi, Auld and Denham2009, Reference Ooi2012; Cochrane, Reference Cochrane2016, Reference Cochrane2017) and soil seed banks.
Succulent species are common in arid and semi-arid American environments (Reyes-Agüero et al., Reference Reyes-Agüero, Aguirre and Peña-Valdivia2000; Ortega-Baes and Godínez-Alvarez, Reference Ortega-Baes and Godínez-Álvarez2006); population studies for these plants are common (Bowers, Reference Bowers2000, Reference Bowers2005; Montiel and Montaña, Reference Montiel and Montaña2003; Cheib and Garcia, Reference Cheib and Garcia2012; Álvarez-Espino et al., Reference Álvarez-Espino, Godínez-Álvarez and De la Torre-Almaráz2014). Álvarez-Espino et al. (Reference Álvarez-Espino, Godínez-Álvarez and De la Torre-Almaráz2014) found high germination after dormant seeds of Stenocereus stellatus (Cactaceae) were buried for 6 months. However, the potential impact of small temperature increases, as projected by global warming, on seed germination and soil seed banks has been poorly assessed (Pérez-Sánchez et al., Reference Pérez-Sánchez, Jurado, Chapa-Vargas and Flores2011).
A key factor in the demographic patterns of plants is seed dormancy (Rojas-Aréchiga and Vázquez-Yanes, Reference Rojas-Aréchiga and Vázquez-Yanes2000). It is a process where germination is prevented, to help maximize the probability to seedling establishment and spread the risk of recruitment failure (bet-hedging strategy) across time (Baskin and Baskin, Reference Baskin and Baskin1998; Jurado and Flores, Reference Jurado and Flores2005) until conditions for seedling establishment are suitable (e.g. adequate moisture, light and temperature during early growth stages). Thus, dormancy is closely linked with the formation of soil seed banks (Baker, Reference Baker, Leck, Parker and Simpson1989).
Recent studies on the potential effects of global warming in succulent species have considered physiological (Aragón-Gastélum et al., Reference Aragón-Gastélum, Flores, Yáñez-Espinosa, Badano, Ramírez-Tobías, Rodas-Ortiz and González-Salvatierra2014), ecological (Aragón-Gastélum et al., Reference Aragón-Gastélum, Badano, Yáñez-Espinosa, Ramírez-Tobías, Rodas-Ortiz, González-Salvatierra and Flores2017), biochemical (Musil et al., Reference Musil, Schmeidel and Midgley2005, Reference Musil, Van Heerden, Cilliers and Schmeidel2009) and human (Martorell et al., Reference Martorell, Montañana, Ureta and Mandujano2015) aspects. Some functional traits in population and community dynamics such as the role that future temperature increases could play on seed viability, germination, dormancy and soil seed banks have been neglected.
The largest warm desert of North America is the Chihuahuan Desert (Archer and Predick, Reference Archer and Predick2008) ranging from the southwestern United States to the Central Mexican Highlands. Models of global change for this region indicate an increase in summer temperatures (June–September) of 1–2°C by 2030 (Tejeda-Martínez et al., Reference Tejeda-Martínez, Conde-Álvarez and Valencia-Treviso2008). This raises concerns because the desert harbours a high richness of succulent plants (Rzedowski, Reference Rzedowski1991), many of them protected by Mexican environmental laws (SEMARNAT, 2010).
Ooi (Reference Ooi2012) highlighted that increased soil temperatures could accelerate the decline of seed viability and compromise bet-hedging strategies of species in dryland regions. Because of the crucial role that soil seed banks and dormancy might have in population dynamics in arid and semi-arid environments, our aim in this study was to explore how warmer soils could affect the seasonal dynamics and persistence of soil seed banks in three succulent species of two common plant families from the Chihuahuan Desert: Echinocactus platyacanthus Link and Otto form visnaga (Cactaceae), Yucca filifera Chabaud (Asparagaceae), and Agave striata Zuccarini (Asparagaceae). Under current climatic conditions, emergence of succulent species in the Chihuahuan Desert coincides with summer rainfall (Mandujano et al., Reference Mandujano, Montanā, Méndez and Golubov1998).
Physiological seed dormancy is expected in these species, because it is common in their families (Baskin and Baskin, Reference Baskin and Baskin1998). Cactus seeds can enter secondary dormancy with cues from specific conditions such as darkness (Flores et al., Reference Flores, Jurado and Arredondo2006; Rojas-Aréchiga and Mandujano-Sánchez, Reference Rojas-Aréchiga and Mandujano-Sánchez2017), and at specific times, according to variation of environmental factors (Álvarez-Espino et al., Reference Álvarez-Espino, Godínez-Álvarez and De la Torre-Almaráz2014; Ordóñez-Salanueva et al., Reference Ordóñez-Salanueva, Orozco-Segovia, Canales-Martínez, Seal, Pritchard and Flores-Ortiz2017). Thus, some seeds of succulent species germinate in summer, the rainy season, but others can enter dormancy. Seed dormancy is considered a very common adaptive plant strategy in unpredictable and harsh environments, such as arid and semi-arid landscapes (Jurado and Flores, Reference Jurado and Flores2005). Improving our understanding of both the mechanistic response and the adaptive capacity of seed banks to climate change will provide a solid basis for improved predictions of future distribution of species and risk of extinction (Ooi, Reference Ooi2012).
Materials and methods
Studied species and plant material
The three target-species are endemic to México and widely distributed, and have been previously studied for germination traits (Jiménez-Aguilar and Flores, Reference Jiménez-Aguilar and Flores2010; Pérez-Sánchez et al., Reference Pérez-Sánchez, Jurado, Chapa-Vargas and Flores2011). Echinocactus platyacanthus f. visnaga is a barrel-like cactus that can reach 2 m in height and 80 cm in diameter (Jiménez-Sierra et al., Reference Jiménez-Sierra, Mandujano and Eguiarte2007). This species is specially protected by the environmental laws of México (SEMARNAT, 2010) and considered as near threatened in the Red List of the International Union for Conservation of Nature (IUCN, 2014) due to over-exploitation. Yucca filifera is one of the largest and most common Yucca species; it is 9 m tall (Irish and Irish, Reference Irish and Irish2000). Agave striata is a short rosette plant that can reach 100 cm in height and 119 cm in diameter (Irish and Irish, Reference Irish and Irish2000). This species forms dense and extensive colonies even in very dry environments (Irish and Irish, Reference Irish and Irish2000). Yucca filifera and A. striata are not included under any protection status (SEMARNAT, 2010), although they are widely distributed, and the impact of global change in their population dynamics is still unknown.
Seeds from E. platyacanthus have positive photoblasticism and small seeds (1.8 mg) (Flores et al., Reference Flores, Jurado, Chapa-Vargas, Ceroni-Stuva, Dávila-Aranda, Galíndez, Gurvich, León-Lobos, Ordóñez, Ortega-Baes, Ramírez-Bullón, Sandoval, Seal, Ullian and Pritchard2011) compared with the other two species in our study. High temperatures promote seed germination for this species, as germination after exposure to 70°C for 14 days was higher than for seeds with no exposure to heat (Pérez-Sánchez et al., Reference Pérez-Sánchez, Jurado, Chapa-Vargas and Flores2011). Seeds from Y. filifera have neutral photoblasticism and are large (70 mg) (Flores et al., Reference Flores, González-Salvatierra and Jurado2016). Agave striata seeds are also neutrally photoblastic (Jiménez-Aguilar and Flores, Reference Jiménez-Aguilar and Flores2010) and small (2.3 mg) (Ramírez-Tobías et al., Reference Ramírez-Tobías, Peña-Valdivia, Reyes-Agüero, Sánchez-Urdaneta and Valle2012).
Seed collection and storage
Ripe fruits from the three species studied were collected in the Southern Chihuahuan Desert in San Luis Potosi, Mexico in October 2012, from at least 10 mother plants for each species. This area is dominated by desert shrublands and has an annual rainfall of 300–450 mm and a mean temperature of 18–25°C (INEGI, 2002).
The harvested fruits from each species were transferred to the ecology laboratory, where seeds were removed. Seeds were left to dry and kept in paper bags for 3 months at room temperature in normal day/night conditions before sowing in the field.
Viability test
Seed viability was evaluated using a tetrazolium test (Grabe, Reference Grabe1970) with a solution of 2,3,5-triphenyltetrazolium at 1%. In order to evaluate initial viability, 30 seeds for each species were placed in three groups of 10 seeds. These seeds were soaked for 24 h in distilled water before being placed in the tetrazolium solution. In order to facilitate the entry of the solution into the seed, an incision was made with a scalpel, parallel to the hypocotyl axis.
Each experimental unit was a group of 10 seeds, and three groups were prepared. Seeds from each group were placed in a beaker with 50 ml of tetrazolium solution. Each beaker was covered and wrapped with aluminum foil to maintain the seeds in darkness. Seeds were kept at 25°C for 48 h. Seeds were transversely cut with a scalpel and embryos were observed under a stereo-microscope (Motic model MAIN). The seeds showing whole red embryos without apparent damage were considered viable (Baskin and Baskin, Reference Baskin and Baskin1998). Viability percentage was estimated for each species.
Germination test
In a second seed lot, 100 seeds of each species were selected and their initial germination assessed in a completely randomized experiment. This consisted of 10 replicates of 10 seeds placed on moistened cotton wool in Petri dishes. Petri dishes were then placed in a growth chamber set to a relative humidity of 80%, a 12 h/12 h photoperiod, and a constant temperature of 25°C. Germination was determined when radicle protrusion was at least 2 mm and recorded daily for 30 days. Germination was calculated based on initial viable seeds.
Seed burial
The potential for soil seed bank formation was evaluated by burying the seeds and subsequently examining their germination percentage at regular intervals (Cheib and Garcia, Reference Cheib and Garcia2012) following the design used by Ordóñez-Salanueva et al. (Reference Ordóñez-Salanueva, Orozco-Segovia, Canales-Martínez, Seal, Pritchard and Flores-Ortiz2017). Forty-eight batches containing 55 seeds (2640 seeds) per species were made. Each batch was placed inside translucent cloth bags of 10 cm by 10 cm sewn with thread to prevent seed loss.
Twenty-four bags per species were allocated for burial across six replicate hexagonal open top chambers (OTCs) which induced warming. The other 24 bags per species were buried in ambient conditions outside the OTCs (also across six replicate areas).
Experimental design of germination with warming treatments across seasons
This experiment was conducted in an abandoned agricultural field located in the southernmost section of the Chihuahuan Desert (22°14′11′′N, 100°51′46′′W, 1844 m), in central México. Annual rainfall in the study area is 341 mm and is concentrated in the summer months. The rainy season occurs between June and October; mean annual temperature is 17.8°C, but it can be higher than 35°C in summer (Medina-García et al., Reference Medina-García, Díaz, Loredo, Serrano and Cano2005). Vegetation is dominated by sclerophyllous shrubs, with some cacti and succulent monocots scattered among woody species. In our study area, vegetation was cleared from the surface in a 25 m × 25 m enclosure fenced with wire (2 m height) to prevent access from cattle and people. Within this enclosure, twelve experimental plots of 5 m × 5 m were outlined following a rectangular arrangement (3 plots wide × 4 plots long). Six of these plots were randomly assigned to the induced warming treatment, while the other six plots were used as a control under ambient environmental conditions (Aragón-Gastélum et al., Reference Aragón-Gastélum, Flores, Yáñez-Espinosa, Badano, Ramírez-Tobías, Rodas-Ortiz and González-Salvatierra2014, Reference Aragón-Gastélum, Badano, Yáñez-Espinosa, Ramírez-Tobías, Rodas-Ortiz, González-Salvatierra and Flores2017). Warming was implemented using hexagonal OTCs. Warming from the OTCs allowed passive heating and are a simple approach for assessing the responses of plants to warming in the field (Musil et al., Reference Musil, Schmeidel and Midgley2005, Reference Musil, Van Heerden, Cilliers and Schmeidel2009; Aragón-Gastélum et al., Reference Aragón-Gastélum, Flores, Yáñez-Espinosa, Badano, Ramírez-Tobías, Rodas-Ortiz and González-Salvatierra2014, Reference Aragón-Gastélum, Badano, Yáñez-Espinosa, Ramírez-Tobías, Rodas-Ortiz, González-Salvatierra and Flores2017). OTCs were built with UV-resistant transparent acrylic (3 mm thick; wavelength transmission 110 < 280 nm) following the design proposed by Marion (Reference Marion, Molau and Mølgaard1996). The resulting structures were 0.50 m tall, 1.5 m wide at the open top, and 2.08 m wide at the base attached to the ground. This OTC design increases air temperature during the day by between 1.9 and 5.0°C compared with external ambient conditions (Musil et al., Reference Musil, Schmeidel and Midgley2005, Reference Musil, Van Heerden, Cilliers and Schmeidel2009; Aragón-Gastélum et al., Reference Aragón-Gastélum, Flores, Yáñez-Espinosa, Badano, Ramírez-Tobías, Rodas-Ortiz and González-Salvatierra2014, Reference Aragón-Gastélum, Badano, Yáñez-Espinosa, Ramírez-Tobías, Rodas-Ortiz, González-Salvatierra and Flores2017).
To assess the magnitude with which OTCs modified microclimate, air and soil temperature as well as relative humidity were continuously monitored within and outside these structures during the entire experiment. For this, data-loggers were used (HOBO Pro v2 and HOBO Pendant, Onset Computer Corporation, MA, USA). One data logger was fixed 10 cm above the ground at the centre, while the second one was buried so the sensor would be at <0.5 cm under the soil surface in each OTC and control plot. Readings were programmed to be recorded every hour and averaged daily. These measures were conducted from 1 March 2013 to 28 February 2014, and recorded data were used to calculate the daily mean air and soil temperatures, as well as average daily mean air relative humidity in both treatments (OTCs and control). Data of daily rainfall (mm) were obtained from the National Institute of Forestry, Agriculture and Livestock (INIFAP San Luis Potosí, México), at 2 km from our study site.
Four lots of seed for each species were buried <0.5 cm under the soil surface, at the centre of each experimental plot. In plots with OTCs, the seed lots were located directly under the opening to avoid both overwarming and reduced rainfall interception by the acrylic walls. One bag per replicate for each species and treatment (six replicates per species in OTCs and six in controls) was exhumed at the end of each season (spring, summer, autumn and winter) during one year. Seeds from each bag were distributed in five Petri dishes used in the germination experiment per season for each environment; there were six replicates from OTC and six controls.
Seed viability was evaluated using a tetrazolium test, using three replicates of 10 seeds per species. Seed germination was assessed in a completely randomized experiment considering six replicates for warming and six for control. Seeds were sown in Petri dishes with moist cotton wool (five Petri dishes for each treatment) with 10 seeds in each one. Petri dishes were placed in conditions similar to those of our initial germination trial described above. Germination records were made daily for 30 days. The remaining five seeds per replicate and treatment (60 seeds for each species) were transversely cut to assess viability across seasons by exposing the embryos in the stereoscope, to see if an embryo was present and looked healthy. Mouldy and pulverized seeds were considered unviable.
Statistical analyses
Abiotic variables were summarized for each data-logger (n = 6 per treatment) and compared between OTCs and control plots with repeated measures ANOVA. Initial viability and germination were analysed by one-way ANOVA for each species. Viability and seed germination data were arcsine transformed to fulfil the normality assumption.
Germination across seasons was analysed by two-way ANOVA using each removed bag as a single replicate at the OTC level for each species, with seed exhumation season and warming (OTCs and control plots) as predictor factors, having OTC as a random factor and season as a fixed factor. Tukey's tests were used to test for significant differences between means.
Results
Abiotic variables
Total rainfall during the study period was 296.2 mm and was distributed as follows: spring, 21.7 ± 0.12 mm; summer, 103 ± 0.30 mm; autumn, 87.9 ± 0.31 mm; winter, 83.6 ± 0.43 mm.
Between 1 March 2013 and 28 February 2014, mean daily air temperature was 19.8 ± 0.16°C inside OTCs and 18.1 ± 0.02°C in control plots. This variable significantly differed between warming treatments (F (1,6) = 23.94, P < 0.0001) and season (F (3,2184) = 156.61, P < 0.0001), and the interaction was also significant (F (3,2184) = 92.17, P < 0.0001), in that mean daily air temperature was higher in the OTC plots in spring and summer than in the other treatments combined (Fig. 1a).
Daily air mean temperature and relative humidity (mean ± standard error) across of the seasons measured in control plots (grey bars) and within OTCs (black bars) used in the field experiment (± 95% CI).
Air average daily relative humidity in the control plots was 65.8 ± 0.33% and 63.3 ± 1.12% within OTCs. We found no significant effects of warming (F (3,2184) = 4.36, P > 0.05), but season (F (3,2184) = 86.92, P < 0.0001), and the interaction between both factors were significant (F (3,2184) = 99.29, P < 0.0001), in that relative humidity was higher in the control plots in autumn than in the other combined treatments (Fig. 1b).
Mean soil temperature was affected by warming (F (1,6) = 4.78, P = 0.0290), season (F (3,2184) = 16.09, P < 0.0001), and by the warming × season interaction (F (3,2184) = 115.70, P < 0.0001). The daily mean soil temperature was 23.1 ± 0.21°C in OTCs and 22.1 ± 0.15°C in control plots. In spring, we found warmer daily mean soil temperature inside OTCs (29.0 ± 0.21°C) than in control plots (27.0 ± 0.10°C; F (3,552) = 31.01, P < 0.0001; Fig. 2). There were no differences between treatments in the other seasons: summer (F (3,558) = 3.45, P = 0.0647; Fig. 2), autumn (F (3,528) = 0.099, P = 0.7529; Fig. 2) and winter (F (3,528) = 2.209, P = 0.1390; Fig. 2); and they showed a decreasing pattern over time.
Daily soil mean temperature (mean ± standard error) across of the seasons measured in control plots (grey bars) and within OTCs (black bars) used in the field experiment (± 95% CI).
Seed viability and initial seed germination
Initial viability of fresh seeds was similar for the three studied species, which was 100% for Y. filifera and E. platyacanthus and 90 ± 10% for A. striata (F = 1.00, P = 0.4218; Fig. 3a). There were no changes in seed viability across seasons except for A. striata, which had more unviable seeds (appearing pulverized) in autumn (30%) and winter (15%). Germination differed between species (F = 16.20, P = 0.0002): Y. filifera initial germination reached 83 ± 9.4%; in A. striata it reached 30 ± 6.1%; and for E. platyacanthus it reached 49 ± 4.4% (Fig. 3b).
(a) Initial viability and (b) initial seed germination (mean ± standard error) in fresh seeds of three succulent species. Different lower case letters indicate significant differences (P < 0.005).
Germination under warming treatments across seasons
We found variation in germination for the three species studied; Y. filifera had similar germination between warming treatments and seasons (P > 0.05). Germination was high in the spring (77.2 ± 5.0%) and summer (81.9 ± 2.5%). After that, 90% of seeds germinated inside the bags buried in the soil, thus there were no remaining bags for the rest of the experimental timeframe.
In contrast, germination of A. striata significantly differed between warming treatments (F (1,47) = 9.278, P = 0.005) and seasons (F (3,47) = 31.262, P < 0.0001), but the interaction between warming and season was not significant (F (3,4) = 1.549, P = 0.21). Germination of seeds from OTCs (49.89 ± 4.57%) was higher than those from control plots (37 ± 5.7%). Germination was higher in spring (76.8 ± 3.1%) than in summer (42.8 ± 4.3%), which in turn was higher than in both autumn (30 ± 5.1%) and winter (24 ± 3.3%).
Echinocactus platyacanthus germination was affected by warming (F (1,47) = 13.223, P < 0.001), and by season (F (3,47) = 52.750, P < 0.0001), but it was not affected by the warming × season interaction (F (3,47) = 2.245, P = 0.09). Germination of seeds from OTCs (50.75 ± 3.6%) was higher than those from control plots (40.28 ± 4.9%). Higher germination was found in spring (62.36 ± 2.6%) and winter (62.61 ± 4.2%) than in summer (37.5 ± 4.3%), which was higher than in autumn (19.59 ± 1.5%).
Discussion
This is the first study to explore soil temperature increase under climate change scenarios in American deserts and their impact on seeds; our approach using OTCs appears to have provided realistic temperature scenarios. We found a 1.7°C increase in mean daily air temperature in our warming treatments (OTC vs control plots) between 1 March 2013 and 28 February 2014. This was within the 1–3°C increment projected for global change by the late 21st century (IPCC, Reference Stocker, Qin, Plattner, Tignor, Allen, Boschung, Nauels, Xia, Bex and Midgley2013, Reference Field, Barros, Dokken, Mach, Mastrandrea, Bilir, Chatterjee, Ebi, Estrada, Genova, Girma, Kissel, Levy, MacCracken, Mastrandrea and White2014) in arid ecosystems worldwide. In addition, we found a small but significant decrease in mean relative humidity (2.5%) within OTCs compared with control plots. Low relative humidity decreases the atmospheric water vapour and consequently the water condensation in the soil (Matimati et al., Reference Matimati, Musil, Raitt and February2012), which has a negative impact on water availability and thus, potentially limiting adequate water uptake of some succulent species. This could have a detrimental effect in the perpetuation mechanisms and persistence for these species, although reduced relative humidity may be caused by OTCs deflecting rainfall away from the studied plot. Both air temperature increments and mean relative humidity decrements coincide with those found by Aragón-Gastélum et al. (Reference Aragón-Gastélum, Flores, Yáñez-Espinosa, Badano, Ramírez-Tobías, Rodas-Ortiz and González-Salvatierra2014, Reference Aragón-Gastélum, Badano, Yáñez-Espinosa, Ramírez-Tobías, Rodas-Ortiz, González-Salvatierra and Flores2017) in short-term experiments in the same area.
Under global warming models, a 1°C increase in air temperature could result in an associated mean soil temperature increase of 1.5°C in some environments (Ooi et al., Reference Ooi, Auld and Denham2012). This coincides with our results because we found a 1°C increase in mean soil temperature. These results agree with those of Cochrane et al. (Reference Cochrane, Hoyle, Yates, Wood and Nicotra2015) who found an analogous daily soil temperature rise in a Mediterranean ecosystem from South West Australia using similar OTCs.
In the Southern Chihuahuan Desert, soil temperature can, for short periods reach ~45°C at the hottest time of the day (Pérez-Sánchez et al., Reference Pérez-Sánchez, Flores, Jurado and González-Salvatierra2015). During the study period, we found a higher maximum temperature in OTCs in mid-spring of 47°C. We found an increase of 2°C in the warming treatment, which falls within projection from models of global change for this region (Tejeda-Martínez et al., Reference Tejeda-Martínez, Conde-Álvarez and Valencia-Treviso2008).
Our assessment of the effects of increased mean soil temperature on the formation, dynamics and persistence of soil seed banks uncovered differential responses among the studied species. For A. striata, germination was initially high and decreased throughout the year. However, higher germination was maintained in the warmed OTC plots compared with the controls, meaning seeds retained their germination capability for longer in OTCs presumably because of maintenance of higher temperatures. Under current climatic conditions, emergence of succulent species in the Chihuahuan Desert coincides with summer rainfall (Mandujano et al., Reference Mandujano, Montanā, Méndez and Golubov1998). In future warmer environmental conditions, however, seeds could germinate when rainfall or temperature would no longer be suitable for seedling establishment.
Our germination results provide evidence of seed dormancy in aged seeds, because the proportion of seeds that are viable is larger than the proportion of the ones that germinate (Jiménez-Aguilar and Flores, Reference Jiménez-Aguilar and Flores2010; Ramírez-Tobías et al., Reference Ramírez-Tobías, Peña-Valdivia, Reyes-Agüero, Sánchez-Urdaneta and Valle2012) and consequently form soil seed banks. The lower germination recorded in autumn–winter is probably due to the onset of cold temperatures. A temperature drop is often a cue for development of secondary dormancy (Baskin and Baskin, Reference Baskin and Baskin2014), and soil warming under future conditions could delay this. Delayed onset of secondary dormancy is an important issue for this species, because seeds could germinate under warming but seedlings would perhaps not tolerate the following winter.
The Agave species have developed exceptional physiological (e.g. Crassulacean acid metabolism, ‘CAM’) and ecological adaptations to high temperature and scarce water (Nobel, Reference Nobel2010; García-Moya et al., Reference García-Moya, Romero-Manzanares and Nobel2011). In A. striata, a high germination over a wide temperature range has been documented (from 15 to 30°C) (Jiménez-Aguilar and Flores, Reference Jiménez-Aguilar and Flores2010; Ramírez-Tobías et al., Reference Ramírez-Tobías, Peña-Valdivia, Reyes-Agüero, Sánchez-Urdaneta and Valle2012), which is consistent with the final germination found in fresh seeds and those exhumed in spring for this species. Furthermore, A. striata also displays a high water uptake under different water potentials (Ramírez-Tobías et al., Reference Ramírez-Tobías, Peña-Valdivia, Trejo and Vaquera2014). These authors suggest that for A. striata low water availability is required to achieve an adequate hydration to germinate. Thus, loss of viability and poor germination from summer to the end of our study in A. striata could be attributed to a low tolerance to moist conditions, which possibly promoted the attack of fungi and pathogens causing seed decay. A loss of viability in A. striata could also be the consequence of faster seed ageing. Seeds of A. striata had a lower initial viability than the other two species. Although 90% is still relatively high, even a loss of a small percentage could have a significant impact on longevity, particularly if conditions are wet and warm which might promote rapid ageing.
In a global climate change scenario, both temperature increases and atmospheric CO2 level are expected (IPCC, Reference Stocker, Qin, Plattner, Tignor, Allen, Boschung, Nauels, Xia, Bex and Midgley2013, Reference Field, Barros, Dokken, Mach, Mastrandrea, Bilir, Chatterjee, Ebi, Estrada, Genova, Girma, Kissel, Levy, MacCracken, Mastrandrea and White2014). Other studies suggest that Agave species may cope well under climate change, but they only consider the adult life-history stage (Nobel, Reference Nobel2010; García-Moya et al., Reference García-Moya, Romero-Manzanares and Nobel2011). As our results show, the effects on the potential timing of germination and maintenance of the seed bank could affect the ability of A. striata to cope due to changes to these early life-history stages.
Seed germination of Y. filifera was high in spring and summer (with no seeds remaining viable to test for germination in the other seasons) at both temperature conditions. Rapid germination could be beneficial for seedling establishment, providing an advantage in a seasonally dry environment (Flores and Briones, Reference Flores and Briones2001). The high germination we found for Y. filifera seeds was consistent with other studies (Jiménez-Aguilar and Flores, Reference Jiménez-Aguilar and Flores2010; Cambrón-Sandoval et al., Reference Cambrón-Sandoval, Malda-Barrera, Suzán-Azpiri and Díaz-Salim2013; Flores et al., Reference Flores, González-Salvatierra and Jurado2016) as well as for other Yucca species (Flores and Briones, Reference Flores and Briones2001; Pérez-Sánchez et al., Reference Pérez-Sánchez, Jurado, Chapa-Vargas and Flores2011; Flores et al., Reference Flores, Pérez-Sánchez and Jurado2017). If Y. filifera seeds germinate fast and lose viability quickly, they probably have persistence mechanisms other than germination, such as vegetative propagation (Matuda and Piña-Luján, Reference Matuda and Piña Luján1980) to rely on, more than the other species studied.
Echinocactus platyacanthus seeds had higher germination with higher mean soil temperature inside OTCs than in control plots, similar to A. striata seeds. Our findings coincide with those of Ordóñez-Salanueva et al. (Reference Ordóñez-Salanueva, Seal, Pritchard, Orozco-Segovia, Canales-Martínez and Flores-Ortiz2015), who found that projected future temperature increments would not have detrimental effects on germination in Polaskia chende and P. chichipe, two Mexican cactus species. Ours results also support the findings of Seal et al. (Reference Seal, Daws, Flores, Ortega-Baes, Galíndez, León, Sandoval, Ceroni- Stuva, Ramírez-Bullón, Dávila, Ordoñez-Salanueva, Yáñez-Espinosa, Ulian, Amosso, Zubani, Torres-Bilbao and Pritchard2017), who projected the mean temperature of the wettest quarter of the seed collection sites from 55 cactus species from the Americas, under two climate change scenarios, and predicted under the least conservative scenario (+3.7°C) that 75% of cactus species will have increased germination performance.
We also found dormancy cycling of E. platyacanthus seeds, because buried seeds of E. platyacanthus acquired secondary dormancy in the rainy seasons (summer and autumn), which was alleviated at the end of the subsequent dry season (winter), possibly because of the high variation registered in mean and minimum soil temperature at the end of winter. Physiological dormancy is the type of dormancy found in the Cactaceae (Rojas-Aréchiga et al., Reference Rojas-Aréchiga and Vázquez-Yanes2000). Seeds with physiological dormancy can cycle through a gradation of dormancy ‘states’ in response to their environment, during which the range of conditions in which the seeds are able to germinate widens and contracts (Long et al., Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015). Dormancy cycling has also been found in the cactus Polaskia chende from the Tehuacan Valley (Ordóñez-Salanueva et al., Reference Ordóñez-Salanueva, Orozco-Segovia, Canales-Martínez, Seal, Pritchard and Flores-Ortiz2017). Dormancy cycling indicates that the ungerminated seeds are viable seeds in the soil seed bank of E. platyacanthus. Reduced germination without loss of viability may contribute to a seed bank that allows population persistence in face of increased environmental fluctuations in the future (Ooi, Reference Ooi2012).
In conclusion, we found air and soil temperature increments and a lower air relative humidity within OTC plots during one year. This average soil temperature increment changed the dynamics and persistence of soil seed banks as well as dormancy in two of the three species studied, which was reflected in differential responses in germination across seasons. Under global warming projections, unchanged or greater germination of our study species might imply that early life-history stages are resilient to climate change, however greater germination of A. striata and E. platyacanthus could impact their bet-hedging strategies. Unlike Y. filifera, which showed loss of viability after summer, E. platyacanthus forms a soil seed bank and its cycling inter-seasonal dormancy/germination associated with seasonal changes could be an efficient physiological mechanism to face climate change.
Acknowledgements
Comments from the editor and reviewers greatly improved the clarity of this manuscript.
Financial support
This research was supported by Secretaría de Educación Pública – Consejo Nacional de Ciencia y Tecnología (no. CB-2010-156205).
