INTRODUCTION
The widespread occurrence of fire in the tropics has favoured the expansion of invasive species such as Pteridium caudatum (L.) Maxon. This fern is a post-fire colonizer and is widely distributed across the Yucatán Peninsula, where it forms monospecific patches in recently cleared and burnt areas (Gliessman Reference GLIESSMAN1978). In this region, P. caudatum invasion is a complex process, which involves environmental degradation, land-use strategies and fire (Schneider & Fernando Reference SCHNEIDER and FERNANDO2010). Schneider (Reference SCHNEIDER2006, Reference SCHNEIDER, Millington and Jepson2008) described the development of monodominant P. caudatum patches as starting with a few plants shaded by secondary vegetation, which expand drastically once secondary vegetation is cleared through fires or agricultural activities.
It is widely accepted that the spread of bracken stands depends on vegetative growth, but the expansion of existing populations, re-establishment after fires in already invaded fields, and colonization of new sites, could occur by spores arising from spore banks (Dyer Reference DYER, Thompson and Smith1989). Additionally, some evidence suggests that the occurrence of fire is required for P. caudatum spore germination (Gliessman Reference GLIESSMAN1978) and that conditions encountered following fire, such as elimination of competitors and creation of sterile, alkaline and nutrient-rich substrate, are very close to the optimal conditions for sporeling establishment (Page Reference PAGE1976, Reference PAGE1982, Reference PAGE, Smith and Taylor1986).
Post-fire regeneration from propagule banks in the soil largely depends on the temperatures occurring in the topsoil during fires (Auld Reference AULD1986, Bond et al. Reference BOND, LE ROUX and ERNTZEN1990, Bradstock & Auld Reference BRADSTOCK and AULD1995, Hodgkinson Reference HODGKINSON1991, Moreno & Oechel Reference MORENO and OECHEL1991a, Reference MORENO and OECHEL1991b; Schimmel & Granström Reference SCHIMMEL and GRANSTRÖM1996). Excessive heat can be lethal for seeds or spores near the soil surface (Baskin & Baskin Reference BASKIN and BASKIN1998, Keeley Reference KEELEY1977, Keeley & Zedler Reference KEELEY and ZEDLER1978, Zammit & Zedler Reference ZAMMIT and ZEDLER1988), but a brief exposure to the high temperatures caused by fire might be sufficient to stimulate the germination of buried, dormant propagules (Baeza & Vallejo Reference BAEZA and VALLEJO2006, Gashaw & Michelsen Reference GASHAW and MICHELSEN2002, Thanos & Goerghiou Reference THANOS and GOERGHIOU1988, Williams et al. Reference WILLIAMS, CONGDON, GRICE and CLARKE2004).
Because soil is an effective insulator against high fire-induced temperatures (Agee Reference AGEE1993, Whelan Reference WHELAN1995), the depth at which the seeds and spores are buried in the soil modifies the conditions that these propagules undergo during a fire (Auld Reference AULD1986, Bradstock & Auld Reference BRADSTOCK and AULD1995, Odion & Davis Reference ODION and DAVIS2000, Pickup et al. Reference PICKUP, MCDOUGALL and WHELAN2003), hence influencing their post-fire survival and germination (Bradstock & Auld Reference BRADSTOCK and AULD1995, Bradstock et al. Reference BRADSTOCK, AULD, ELLIS and COHN1992, Gill Reference GILL, Mooney, Bonnicksen, Christensen, Lotan and Reiners1981). The relationships between fire-induced temperatures, seed burial depth and seed germination have been studied in several species of angiosperms (Auld Reference AULD1986, Auld & O'Connell Reference AULD and O'CONNELL1991, Gill Reference GILL, Mooney, Bonnicksen, Christensen, Lotan and Reiners1981), but their effects on fern spores have not been evaluated. This paper evaluates the effect of temperatures generated during experimental fires on the germination of P. caudatum spores buried at three soil depths. Concretely, we hypothesized that the spores of bracken fern remain viable after fires, due to the insulating effect of soil. We also documented the temperatures reached and the duration of maximum temperatures during fires, and evaluated the impacts of these two factors upon spore germination. Finally, we quantified the proportion of spores of this species that remain viable in the soil after 2 y, to evaluate whether an effective spore bank exists.
METHODS
Study species
Pteridium caudatum, like all species of bracken, is one of the most successful invasive species in the world (Taylor Reference TAYLOR, Thompson and Smith1990). It is present in Bermuda, southern Florida, the West Indies, Central America and northern South America, mainly in deforested, agricultural and fire-prone areas (Page Reference PAGE, Smith and Taylor1986). Some of the characteristics that make this species a successful competitor, capable of creating monospecific patches, include high resistance to diseases and pests (Cooper-Driver Reference COOPER-DRIVER1990), allelopathic effects (Gliessman & Muller Reference GLIESSMAN and MULLER1976), abundant vegetative propagation (Page Reference PAGE, Smith and Taylor1986), tolerance of a broad range of climatic and edaphic conditions (Gliessman & Muller Reference GLIESSMAN and MULLER1976), elevated spore production (Conway Reference CONWAY1957), high dispersal rate (Gregory & Hirst Reference GREGORY and HIRST1957) and resistance to fire (Gliessman Reference GLIESSMAN1978). Perhaps its most prominent fire adaptation is its buried rhizome, which sprouts vigorously following fires (Fletcher & Kirkwood Reference FLETCHER, KIRKWOOD and Dyer1979). Additionally, bracken fern spores invade recently burned areas, allowing it to establish in the basic conditions created by fire. Other tropical ferns (e.g. some members of Gleicheniaceae) have a similar ecological behaviour in the Asia–Pacific region (Cohen et al. Reference COHEN, SINGHAKUMARA and ASHTON1995, Russell et al. Reference RUSSELL, RAICH and VITOUSEK1998) and other locations of the neotropics (Slocum et al. Reference SLOCUM, AIDE, ZIMMERMAN and NAVARRO2004, Walker Reference WALKER1994, Walker & Boneta Reference WALKER and BONETA1995, Zimmerman et al. Reference ZIMMERMAN, PASCARELLA and AIDE2000).
Study site
In July 2005 we collected soil, litter and P. caudatum spores from seasonally dry secondary forest at El Edén Ecological Reserve, located 38 km north-west of the city of Cancun, in the state of Quintana Roo, Mexico. The weather in this site is warm and humid, with a seasonal period of rainfall from June to December, a mean annual temperature of 24.7 °C and a mean annual precipitation of 1511 mm. Soils at this site are very young and shallow (0–4 cm), covering limestone rock. Vegetation in the area includes semi-deciduous seasonal forest, tropical savanna and secondary vegetation in different stages of succession. The study area has experienced a long history of both natural (e.g. hurricanes) and anthropogenic (e.g. fires, harvesting of timber and non-timber forest products) disturbances. There are also sites affected by recurring fires during the dry season, which are currently covered with P. caudatum.
Experimental fires
The effect of fires on germination of bracken spores was tested in 12 experimental fires conducted in the Pteridology Laboratory of the Universidad Autónoma Metropolitana. Spores used in these treatments were obtained from 15 adult sporophytes at the study site. Fertile pinnae with mature sporangia were stored inside paper bags and left to dry at environmental temperature to facilitate spore release. The contents of the bags were sifted with a metallic mesh, and the spores obtained were stored under dry conditions in glass bottles until used in experimental fires, 1 mo after collection. Each fire was achieved in a block containing 1200 cm3 of dry soil and a constant fuel load of 47.4 ± 5.3 g (± SE) of dry litter of P. caudatum. Litter and soil samples used in this experiment were dry when collected, but to ensure homogeneity of treatments, were further dried for 48 h at 70 °C before being weighed and used for the experimental blocks. The fuel load used in the above described blocks is similar to that found in P. caudatum-invaded fields at the study site, and was obtained by collecting the litter present in 30 randomly located 20 × 20-cm quadrats in the field. To test for the presence of viable spores in soils of the study area (i.e. spore bank), additional soil samples without previous treatment were used in germination trials.
Before ignition, three containers of iron mesh with 1 mg of P. caudatum spores (without previous treatment) were placed in each of the 12 blocks of soil, one on the soil surface, and the other two buried at 1 and 3 cm, respectively. To bury the containers we made a c. 2-cm diameter hole, inserted the iron mesh containers and covered the hole. In each soil block, three K-type thermocouples (Onset Computer Corporation, Pocasset, MA, USA) connected to data loggers (model U12–014, HOBO, Onset Computer Corporation, Pocasset, MA, USA) were inserted in the blocks at the same depth as the spores and adjacent to the spore containers. All sensors were activated approximately 1 min before fires were lit. Litter was subsequently ignited, and the temperatures at each depth were recorded each minute for 2 h. The maximum temperature and duration of the maximum temperature were obtained from the data loggers.
Spore germination
The spore containers were dug out 1 d after the fires, and spores were sown within the next 48 h on 5-cm diameter Petri dishes containing previously sterilized agar with Thompson's nutrients (Klekowski Reference KLEKOWSKI1969). Spores were spread on the surface of the solidified medium with a fine brush. Three replicates for each depth per block were incubated at 25 °C and illuminated with fluorescent light under a photoperiod of 12 h light/12 h darkness for 2 mo. Untreated spores were used as controls and were incubated under the same laboratory conditions.
Spore germination in Petri dishes after experimental fires was recorded weekly for 2 mo with the help of a stereoscopic microscope (American Optical, Stereo Star ZOOM, AO Company, USA). The number of germinated spores (gametophytes) was counted in each replicate in five 0.5 × 0.5-cm fields, and the results were expressed as the mean number of gametophytes developed in 0.25 cm2. We calculated the germination percentages of spores that were exposed to fire at the soil surface and of those that were buried at 1 and 3 cm in the 12 experimental fires.
To verify the presence of reservoirs of viable spores of P. caudatum in the soil at the study site, we evaluated germination of spores according to the method described by Ramírez-Trejo et al. (Reference RAMÍREZ-TREJO, PÉREZ-GARCÍA and OROZCO-SEGOVIA2004) at two different times: 13 d after the collection in the field and after 2 y. Soil samples were taken from 45 points randomly chosen within the study site. Having removed the litter layer in each point, soil samples were collected by inserting a steel cylinder (8 cm length, 5 cm diameter) vertically in the soil to a depth of 3 cm. Each soil sample was completely removed from the steel cylinder and placed in polyethylene bags for transport to the laboratory. Three days after the collection in the field a total of 45 samples of approximately 50 g of soil were placed in transparent plastic containers (5 cm diameter), covered with a transparent plastic, kept damp, and placed under solar lamps (75 W F96T12/D day light, photon flux density = 437 μmol m−2 s−1) for a photoperiod of 12 h of light, at 18 °C to induce spore germination. After 10 d, spore germination was recorded with a stereoscopic microscope (American Optical, Stereo Star ZOOM, AO Company, USA), by counting the total number of spores that germinated in each container. Results were expressed as the mean number of germinating spores cm−2 in the germination vessel. For 2 y, soil samples were kept dry at room temperature in polyethylene bags. After that time, soil samples were again cultured following the same procedure.
Statistical analyses
Maximum temperature, duration of the maximum temperature and germination percentages recorded at the different depths were compared using a blocked design analysis of variance. Post hoc comparisons were done with a Tukey's test. The effects of depth, maximum temperatures and duration of the maximum temperatures on the germination percentage of P. caudatum were determined with a stepwise regression analysis. Angular transformation of the percentage data was conducted to homogenize the variance between groups and to comply with the conditions of normality (Sokal & Rohlf Reference SOKAL and ROHLF1995). Similarly, the data for maximum temperature and duration of maximum temperature were log- and square root-transformed, respectively. All error terms are standard errors. Analyses were conducted using the SYSTAT 11 statistical package (SYSTAT Software, Inc).
RESULTS
Behaviour of the soil temperature
Mean maximum temperatures and mean duration of maximum temperatures recorded during the fires were significantly influenced by depth (F = 25.2, F = 34.1; respectively, in both cases P < 0.001) with a negative relationship of depth and maximum temperature and a positive relationship between depth and duration of maximum temperatures. In neither case was there an effect of blocks. The heat pulse generated by the passage of the flames produced a significant increase in the surface soil temperature, where a mean maximum temperature of 172 °C ± 25.8 °C was recorded. With increasing depth, the maximum temperature was significantly reduced. At 1 and 3 cm in depth, soil heating was moderate and the mean maximum temperature was 42.0 °C ± 3.5 °C and 32.7 °C ± 2.0 °C, respectively, with no significant differences between these levels. On the soil surface, there was substantially more variation in the maximum temperatures (range of 290 °C) than beneath surface. The range of maximum temperatures was 30.4 °C at 1 cm and 16.2 °C at 3 cm. Maximum temperatures were of relatively short duration at the three soil depths. On the surface, and at 1 cm, the durations were 1.2 and 1.4 min, respectively, whereas at 3 cm the average duration of the maximum temperature was significantly increased to 3.2 min.
As the flames progressed, temperatures on the surface suddenly rose, and the maximum temperatures were recorded at a mean time of 6.1 min after ignition. This brief initial heat pulse was followed by a rapid decrease in temperature (Figure 1a). At a depth of 1 cm, the maximum temperature was reached 2.7 min later than at the surface, but the cooling was comparatively slower than at the surface, and, in some cases, several minutes passed before the initial temperature was reached (Figure 1b). At 3 cm, the heat penetrated slowly and produced a minimal and gradual increase of the soil's temperature, so the maximum temperature occurred, on average, 22.3 min after the fires started (Figure 1c).
Figure 1. Temperatures recorded at the soil surface (a) and at depths of 1 cm (b) and 3 cm (c) during 12 experimental fires. Squares indicate maximum temperatures.
Effect of the temperature and burial depth on spore germination
Germination of P. caudatum spores was recorded when soil temperatures ranged between 26.5 °C to 60 °C. The highest germination percentage (89%) occurred in spores that, during fires, remained at 3 cm in depth, where a maximum soil temperature of 33 °C was reached (Figure 2). The mean germination percentage was significantly influenced by the depth at which the spores remained buried during fires (F = 44.7, P < 0.001), with no block effect. The increase in depth produced a drastic decrease in the soil's temperature; thus spores unexposed to fire had higher mean germination percentages (86%) than those exposed to fire at any of the three soil depths (Figure 3). Spores placed on the surface were severely affected by the direct exposure to flames, which resulted in a high mortality and, therefore, a low germination response (2.1% ± 1.4%), while spores buried at 1 and 3 cm showed significantly higher germination percentages (~77%), without significant difference between depths (ANOVA, Tukey's test HSD, P > 0.05; Figure 3).
Figure 2. Post-fire germination of Pteridium caudatum spores in relation to the soil maximum temperature recorded during 12 fires at different depths: soil surface (●), 1 cm (▲) and 3 cm (■). n = 36.
Figure 3. Germination percentages of Pteridium caudatum spores exposed to fire, at three depths (white bars) and unexposed controls (black bar). Different letters denote values that are significantly different at P < 0.05.
The stepwise regression analysis revealed that depth and maximum temperatures had significant effects on the germination percentage of P. caudatum spores. When these variables are accounted for, the duration of the maximum temperatures does not account for any additional variance in the model. The equation in the stepwise regression that best explained the post-fire germination of P. caudatum spores was: y = 1.55 + (0.331 × depth) – (0.745 × maximum temperature), where y is the arcsine-transformed proportion of germinated P. caudatum spores (R2 = 0.85, P < 0.001).
In terms of viable bracken fern spores, a total density of 1.8 germinating spores cm−2 germinated in soil samples evaluated 13 d after collection. After 2 y, spore germination was reduced to 0.02 germinating spores cm−2.
DISCUSSION
Our results demonstrate that superficial fires affected the germination of spores of P. caudatum located in the topsoil, as a function of the temperature reached at each depth at which spores were buried. Experimental fires produced a considerable increase in the temperature of the soil surface, lethal for most spores of P. caudatum exposed to the flames. Nevertheless, at 1 and 3 cm in depth, the soil acted as an effective insulator to moderate the temperature increase, and the buried spores retained their viability, as expressed by a germination percentage ~ 77%. The insulating properties of soil during a fire are widely documented (Agee Reference AGEE1993, Baskin & Baskin Reference BASKIN and BASKIN1998, Bradstock & Auld Reference BRADSTOCK and AULD1995, Bradstock et al. Reference BRADSTOCK, AULD, ELLIS and COHN1992, Gill Reference GILL, Mooney, Bonnicksen, Christensen, Lotan and Reiners1981, Williams et al. Reference WILLIAMS, CONGDON, GRICE and CLARKE2004). In the context of the Yucatán Peninsula, the insulating properties of the soil become more relevant because the propagule banks (spores, seeds, buds, etc.) and the rhizome system are immersed in a very thin layer of mineral soil or survive between fissures in the limestone rock (Allen et al. Reference ALLEN, VIOLI, ALLEN, GÓMEZ-POMPA, Gómez-Pompa, Allen, Fedick and Jiménez-Osornio2003), giving them unique protection against the lethal effects of high fire-induced temperatures.
To our knowledge, this is the first study that analyses the relationship between the temperature reached during a fire and the subsequent germination of fern spores, although a few studies have evaluated the effect of temperature per se on spore germination (Dyer Reference DYER and Dyer1979, Miller Reference MILLER1968, Pérez-García & Riba Reference PÉREZ-GARCÍA and RIBA1982, Raghavan Reference RAGHAVAN1980, Ranal Reference RANAL1999). Spores of Pteridium spp. are able to germinate under a wide temperature range, from 1 °C (Conway Reference CONWAY1949) to 35 °C (Dyer Reference DYER and Dyer1979). Our results show that, during the experimental fires, dry (not imbibed) spores of P. caudatum tolerated higher temperatures (e.g. at 1 cm the mean temperature was 42.0 °C), retaining their viability and producing a high germination percentage. Our results also show that fire is not a necessary requirement for spore germination (Gliessman Reference GLIESSMAN1978), as untreated spores had higher germination than fire-exposed spores at any soil depth.
Vegetative growth is unquestionably an important mechanism explaining the re-establishment of P. caudatum after fires (Dyer Reference DYER, Thompson and Smith1989). However, our results suggest that spores buried at 1–3 cm from the soil surface, could eventually germinate and become established. In this case, re-establishment would also be the result of sexual reproduction. This would imply that patches dominated by bracken fern constitute true populations, and not just one large individual ramet.
Spores of bracken can remain viable in the soil for approximately 1 y (Dyer Reference DYER, Thompson and Smith1989). Although spore viability was greatly reduced in soil samples after 2 y, our results confirm the presence of viable spores in the topsoil of the study site. Spores that maintain their viability after fires at 1–3 cm deep could be brought to the surface through any sort of soil disturbance. These spores may be sufficient in a long-lived perennial species to allow new populations to colonize and enlarge (Dyer Reference DYER, Thompson and Smith1989).
In seasonally dry tropical forests, seeds buried in dry soil might be more tolerant of heat because they are dehydrated and metabolically inactive for most of the year (Garwood Reference GARWOOD1983). This is a highly relevant point because, in the tropics, fires set for agricultural purposes take place during the dry season (Otterstrom et al. Reference OTTERSTROM, SCHWARTZ and VELAZQUEZ-ROCHA2006), the period when fires might reach their highest temperatures (Williams et al. Reference WILLIAMS, CONGDON, GRICE and CLARKE2004).
Research on a wide variety of ecosystems shows that there is great variation in soil temperatures during a fire (Cautinho Reference CAUTINHO1978, Gimeno-García et al. Reference GIMENO-GARCÍA, ANDREU and RUBIO2004, Miranda et al. Reference MIRANDA, MIRANDA, OLIVEIRA and FERREIRA1993), which largely depends on the fuel loads, fuel moisture, and fuel size distribution, which in combination with meteorological conditions and topography affect the intensity of fires (Agee Reference AGEE1993, Busse et al. Reference BUSSE, HUBBERT, FIDDLER, SHESTACK and POWERS2005, Massman et al. Reference MASSMAN, FRANK, SHEPPERD and PLATTEN2003). In our study site, fuel loads of the litter layer were 1.2 kg m−2, and the mean temperatures produced with the same loads in laboratory conditions were 172 °C, 42 °C and 32.7 °C on the surface and at 1 and 3 cm, respectively. In a seasonally dry tropical forest in Bolivia, fuel loads of 2.2 kg m−2 resulted in superficial temperatures of 225 °C, with no temperature increase at 3 cm deep, while fuel loads of 48 kg m−2 increased temperatures to 704 °C and 227 °C at the surface and 3 cm deep, respectively (Kennard & Gholz Reference KENNARD and GHOLZ2001). Fuel loads in secondary forests with 2–5 y of fallow in the Yucatán Peninsula had fuel loads of 2.1 kg m−2 (Read & Lawrence Reference READ and LAWRENCE2003). If P. caudatum spores are present in the soil of secondary forests 2–5 y in the Yucatán, it might be possible that they will remain viable after these forests are slashed and burned for agricultural use as long as they remain 1–3 cm deep in the soil. In that case, viable buried spores of P. caudatum could explain the new colonization of sites burned repetitively for agricultural purposes in the Yucatán Peninsula.
Temperatures recorded at each soil depth in this study are consistent with the results of other investigations reporting higher temperatures for shorter periods of time at the surface and lower temperatures for longer periods of time deeper in the soil (Bradstock & Auld Reference BRADSTOCK and AULD1995, Bradstock et al. Reference BRADSTOCK, AULD, ELLIS and COHN1992, Portlock et al. Reference PORTLOCK, SHEA, MAJER and BELL1990). The time interval during which the propagules are exposed to heat produced by a fire can influence germination and mortality (Lonsdale & Miller Reference LONSDALE and MILLER1993, Zabkiewicz & Gaskin Reference ZABKIEWICZ, GASKIN and Hartley1978). In this study, the duration of maximum temperatures did not have a significant effect on the germination percentage of P. caudatum spores, thus corroborating results by Keeley et al. (Reference KEELEY, MORTON, PEDROSA and TROTTER1985) and Odion & Davis (Reference ODION and DAVIS2000), who found that seeds in the soil are more sensitive to maximum temperatures than to heating duration. It is nonetheless important to consider that longer exposure times may have detrimental effects on spores and that both variables can affect different species in a variety of ways (Lonsdale & Miller Reference LONSDALE and MILLER1993).
From our experiments, we can conclude that P. caudatum spore germination does not depend on fire, but spores buried 1–3 cm in the soil maintain high germination percentage after fires. This, in addition to the fact that spores might remain viable in the soil for up to 2 y, suggests that re-establishment of P. caudatum after fires can be achieved not only by vegetative growth, but also by germination of spores and the subsequent establishment of new individuals, which implies that bracken fields consist of true populations of P. caudatum.
Some studies conducted in tropical deciduous forests of western Mexico (Miller Reference MILLER1999) and of the Yucatán Peninsula (Rico-Gray & García-Franco Reference RICO-GRAY and GARCÍA-FRANCO1992) document a drastic reduction of seed density from the soil seed banks after fire. In the Yucatán Peninsula, a significant decrease of seeds of tree species after disturbance was found. Both could create new niches that could be occupied by spores of P. caudatum, which, once established, can be capable of promoting the colonization of new sites and eventually create spore banks (Dyer Reference DYER, Thompson and Smith1989). Additionally, after fires, bracken spores maintained their viability at 1–3 cm from the soil surface, suggesting that P. caudatum spore viability in the soil might also explain partially how this species is able to colonize new sites affected by fires repeatedly in the Yucatán Peninsula.
ACKNOWLEDGEMENTS
Results shown in this paper are part of the thesis project of the first author, who is enrolled in the Biological Sciences PhD Program at the Universidad Autónoma Metropolitana. We thank Dr V. L. Barradas for the loan of the equipment for this research. We wish also to thank the Consejo Nacional de Ciencia y Tecnología (CONACyT) for the scholarship provided for this study, and for the financial support provided through the Fondo para la Formación de Doctores en Ciencias (convocatoria I0006-2006-01. AGREEMENT UAM- CONACyT: 34915). Partial support for this project came from project UNAM-PAPIIT IN227802-3 to DRPS. We also thank the technicians and administrators of the Ecological Reserve ‘El Edén’ for their logistical support and help provided during the field work. Finally, we thank three anonymous reviewers for valuable comments on the manuscript.
