Plant Ecology

The dynamics of three shrub species in a fire-prone temperate savanna: the interplay between the seed bank, seed rain and fire regime
--Manuscript Draft-Manuscript Number: Full Title: Article Type: Keywords: Corresponding Author: VEGE3494 The dynamics of three shrub species in a fire-prone temperate savanna: the interplay between the seed bank, seed rain and fire regime Original Research Asteraceae, obligate seeders, persistent seed bank, sprouters, transient seed bank. Michael J. Hutchings, PhD University of Sussex Brighton, UK UNITED KINGDOM

Corresponding Author Secondary Information: Corresponding Author's Institution: Corresponding Author's Secondary Institution: First Author: First Author Secondary Information: Order of Authors: Guadalupe Galindez, PhD Pablo Ortega-Baes, PhD Ana L Scopel, PhD Michael J. Hutchings, PhD Order of Authors Secondary Information: Abstract: A model was developed to assess how the seed rain and fire regime affect seed bank dynamics and seedling establishment of three native shrub species (Acanthostyles buniifolius, Baccharis pingraea and B. dracunculifolia) with different regeneration strategies, in temperate South American savanna. Seed bank and seed rain were quantified for each species under different fire regimes, and their relative roles in regeneration were evaluated. All species had short-term persistent seed banks and high annual variability in seed production. A high proportion of seeds deposited in the seed rain produced seedlings after fire; few entered the soil seed bank. Fire killed a high proportion of the seeds in the soil seed bank. Seedlings derived from the seed rain had a higher probability of surviving for two years than seedlings emerging from the soil seed bank. In the absence of fire, establishment depended on germination both from the seed rain and the soil seed bank, whereas with annual fire, establishment was primarily dependent on germination of seeds arriving in the annual seed rain, regardless of species' regeneration strategies. These results help to explain changes in the vegetation of South American temperate savannas as a result of changes in fire regime and grazing management during the last 50 years. By revealing the crucial roles of the soil seed bank and seed rain in regeneration, this study provides vital information for the development of appropriate management practices to control populations of shrub species with different regeneration strategies in South America temperate savannas. Diego Gurvich [email protected] Marcelo Cabido [email protected] Martin Oesterheld [email protected]
Powered by Editorial Manager and Preprint Manager from Aries Systems Corporation

University of Sussex

Guadalupe Galindez, PhD

Suggested Reviewers:

Mark Ooi [email protected]

Powered by Editorial Manager and Preprint Manager from Aries Systems Corporation

Manuscript Click here to download Manuscript: galindez_et_al_4-7-12.doc Click here to view linked References

1 1 2 Final Word count: 5910 Short running title: Seed bank dynamics and temperate savannas

3 4

The dynamics of three shrub species in a fire-prone temperate savanna: the interplay between the seed bank, seed rain and fire regime

5 6 7 8 9 10 11 12 13 14

Guadalupe Galndez1, Pablo Ortega-Baes2, Ana L. Scopel3 and Michael J. Hutchings4

Banco Base de Germoplasma, IRB-INTA, De los Reseros y N. Repetto, Hurlingham, 1686,

Argentina
2

Laboratorio de Investigaciones Botnicas (LABIBO), Facultad de Ciencias Naturales,

Universidad Nacional de Salta-CONICET, Av. Bolivia 5150, Salta, 4400, Argentina

Laboratorio Estacin de Biologa Sierras, FAUBA-CONICET, Hiplito Yrigoyen 354,

Crdoba, 5174, Argentina

School of Life Sciences, University of Sussex, Falmer, Brighton, Sussex, BN1 9QG, UK.

15

16 17 18 19 20 Correspondence: M. J. Hutchings Email: [email protected] Tel: (+44)-01273-872761

FAX: (+44)-01273-678433

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

Abstract A model was developed to assess how the seed rain and fire regime affect seed bank dynamics and seedling establishment of three native shrub species (Acanthostyles buniifolius, Baccharis pingraea and B. dracunculifolia) with different regeneration strategies, in temperate South American savanna. Seed bank and seed rain were quantified for each species under different fire regimes, and their relative roles in regeneration were evaluated. All species had short-term persistent seed banks and high annual variability in seed production. A high proportion of seeds deposited in the seed rain produced seedlings after fire; few entered the soil seed bank. Fire killed a high proportion of the seeds in the soil seed bank. Seedlings derived from the seed rain had a higher probability of surviving for two years than seedlings emerging from the soil seed bank. In the absence of fire, establishment depended on germination both from the seed rain and the soil seed bank, whereas with annual fire, establishment was primarily dependent on germination of seeds arriving in the annual seed rain, regardless of species regeneration strategies. These results help to explain changes in the vegetation of South American temperate savannas as a result of changes in fire regime and grazing management during the last 50 years. By revealing the crucial roles of the soil seed bank and seed rain in regeneration, this study provides vital information for the development of appropriate management practices to control populations of shrub species with different regeneration strategies in South America temperate savannas. Key-words: Asteraceae, obligate seeders, persistent seed bank, sprouters, transient seed bank.

42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

Introduction Fire opens up spaces in plant communities that can be occupied either by the re-sprouting of established plants or by the establishment of new plants from seed (Keeley 1981, Whelan 1995, Bond and van Wilgen 1996). Species that can re-colonise open ground by re-growth are commonly referred to as sprouters, and those without the capacity for sprouting are referred to as obligate seeders. Following fire, sprouters can also re-colonise open ground from seed (Keeley 1981). In general, obligate seeders have more persistent seed banks (i.e. seed banks containing a higher proportion of seeds that are viable for more than a year, Thompson 1993) because their regeneration after fire depends entirely on establishment from seed, whereas species that can re-sprout tend to have more transient seed banks (Zammit and Zedler 1988, Auld et al. 2000, Campbell and Clarke 2006). Seeds, either from a soil seed bank, or arriving as seed rain dispersed from neighbouring unburnt areas, are essential not only for the immediate regeneration and maintenance of seeder species following fire, but also for the long-term persistence of sprouters, despite their capacity for re-growth (Clarke and Knox 2002, Lamont and Wiens 2003). Although persistent seed banks play a role in the long-term maintenance of species in fire-prone habitats (Whelan 1995, Cohen et al. 2004, Auld and Denham 2006), regular disturbance by fire does not favour the accumulation of large seed banks containing many persistent seeds, because there is insufficient time to replenish such seed banks between successive fires (Lamont et al. 1991, Cohen et al. 2004, Auld and Denham 2006). The annual seed rain is therefore expected to be of greater importance for regeneration in habitats with frequent fires (Whelan 1986, Zammit and Zedler 1988, Auld et al. 2000). In recent decades, anthropogenic activities have increasingly altered the structure of South American temperate savanna ecosystems by changing their disturbance regimes, in particular by suppressing fires, eradicating native herbivores and introducing grazing

4 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 livestock and exotic trees (Silva et al. 2001; Cabral et al. 2003; Altesor et al. 2006). In the temperate savanna of El Palmar National Park (EPNP) in north-east Argentina, three native shrub species - Acanthostyles buniifolius, Baccharis pingraea and B. dracunculifolia (all Asteraceae) - have increased their distributional range and abundance following control of fire and exclusion of grazing (Goveto 2005). If fire had been completely suppressed, shrub encroachment would have been dominated by fire-sensitive species, as in other South American savannas (Moreira 2000). However, this is not the case in EPNP where, despite efforts to prevent fires, most of the park has been burnt at least once between 1968 and 2000, and some areas have been burnt on as many as 13 occasions (Goveto 2005). The three species have different regeneration strategies after fire (Galndez et al 2009a). A. buniifolius is a sprouter, B. dracunculifolia is an obligate seeder and B. pingraea is a facultative sprouter (i.e. it can regenerate both from seeds and from sprouts). We predicted that, regardless of their regeneration strategy, the establishment of all three species in environments where fire is suppressed would show strong dependence both on the annual seed rain and on germination from the soil seed bank, whereas in environments experiencing frequent fires, establishment would depend primarily on the annual seed rain. To test these predictions we quantified the dynamics of the seed bank and seed rain both with and without fire for all three species, and developed a conceptual model to assess how the annual seed rain and fire regime affected their seed bank dynamics and establishment. The results are used both to explain the encroachment of shrubs in these South America temperate savannas and to suggest appropriate management practices for their control. Materials and methods Study area The study was carried out in El Palmar National Park (31 55' S, 58 17" W), an 8,500ha area of protected habitat located in the Entre Ros province of Argentina. The climate is temperate

5 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 (mean annual temperature 19.0 3.8 C), and wet (mean annual precipitation 1346 291mm), with 58% of the rain falling during the warm season (OctoberApril). The study area was located on well-drained, fine sandy loam soils with vegetation dominated by the three shrub species under study. This area was previously subjected to accidental fires in summer 1992 and 1997. Part of the study area (referred to below as the burnt site), was also burnt during a wildfire in January 2000, whereas the remainder of the study area has not been burnt since 1997 (referred to below as the unburnt site). Study species The study species are perennial shrubs that grow up to 3m in height. All are widespread in South America (Zuloaga and Morrone 1999). Acanthostyles buniifolius is hermaphroditic whereas Baccharis pingraea and B. dracunculifolia are dioecious. Flowering typically occurs from December to March. Several achenes (hereafter referred to as seeds) are produced per capitulum, and dispersed by wind between February and May (Burkart 1974). All three species produce seeds with small mass and size (0.06- 0.1 mg and 1mm, Galndez et al. 2009b). Quantification of seed banks To determine the type of seed bank (i.e. transient or persistent) accumulated by each species, 30 plastic trays (20 20 cm in area and 5 cm in depth), each containing sterilized soil, were placed at randomly-chosen locations within the burnt site just before seed dispersal in January 2000, so that they received dispersing seeds in that year both from plants that had re-sprouted after fire and from plants growing in nearby unburnt habitat. This sample size (20 20 5 cm - i.e. 2000 cm3 of soil) was used in several different parts of the study, and is referred to below as a sampling unit. Seeds were prevented from dispersing onto the surface of these soil samples for the next two years by covering them with fine nylon mesh (0.07 0.07 mm). No

6 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 seeds remain on parent plants after dispersal ends in May, so that in each of these years, the filters were removed at the end of seed dispersal. Thus, the only seeds of the study species that accumulated in these soil samples were those dispersed in early 2000. Seedlings that emerged from these soil samples were identified and counted every month from January 2000 until December 2002, and then removed. No seedlings emerged after December 2002. Germination requirements and longevity of seeds of the three species were determined in the laboratory, using seeds collected in March 2000. Four replicates of 25 seeds per treatment per species were sown in Petri dishes on the surface of 1% agar, and incubated under either constant (25C) or alternating (30/20C) temperature regimes under either an 8 h light/16 h dark photoperiod regime, or in complete darkness. Germination was recorded daily for 30 days in the light/dark treatments. For the dark treatment the Petri dishes were wrapped in aluminium foil, and germination was recorded once, after 30 days. For all species, germination was defined as having occurred when the emergent radical was 1 mm in length. To evaluate seed longevity, seeds collected in 2000 were stored in paper bags under laboratory conditions (15-25C), and their germination tested after one and two years. Because of the results obtained from the experimental examination of germination requirements (see below), seed germination was only evaluated under alternating temperature, and under the light/dark regime, as described above. Seed bank dynamics model The dynamics of seeds in the seed bank were examined, and the effects of different fire regimes on seedling establishment were modeled for each species. The phases and transitions considered in the model are illustrated in Fig. 1, and estimation of values for each of the input variables is explained below. All values in the model were quantified in terms of seed or seedling densities m-2.

7 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 The number of germinable seeds produced (A in Fig. 1) was estimated from the mean number of capitula produced by 20 randomly selected reproductive individuals of each species at each of the study sites. The total number of flowering branches was counted for each individual, and three branches were randomly selected, harvested and taken to the laboratory to determine the number of capitula produced. Thirty capitula were selected at random from these branches, and the seeds from each one were placed in separate Petri dishes, on filter paper soaked in distilled water, and incubated at 30/20C in an 8 h light/16 h dark regime. The total number of germinable seeds produced m-2 by each species was estimated as the mean number of germinated seeds per capitulum the mean number of capitula per branch the total number of flowering branches per reproductive individual the density of reproductive individuals of the species. To estimate densities of seeds arriving as seed rain, and of seedlings emerging and seedlings establishing from the seed rain (B, C and D in Fig.1, respectively), 240 plastic trays, each containing a sampling unit of sterilized soil, were placed randomly within the burnt site before seed dispersal commenced in January 2000. Once seed dispersal had finished (May 2000), 120 of these trays were transferred to a greenhouse with a temperature of 25 5C, an 8h light/16h dark cycle and weekly watering. The number of seeds arriving as seed rain was estimated indirectly to be equal to the number of seedlings that emerged from these trays (see Cohen et al. 2004) until emergence ceased in June 2001. The other 120 trays remained in the field, and seedling emergence was recorded every month until seed dispersal began in the following year (January 2001). Seedling establishment was estimated as the proportion of these seedlings that survived in the field over two years (June 2000 to June 2002). To estimate densities of seeds in the soil seed bank, and densities of seedlings emerging and seedlings establishing m-2 from the soil seed bank (E, F and G in Fig.1, respectively), the surfaces of 240 sampling units were marked out before seed dispersal began in January 2000

8 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 in randomly chosen locations in the burnt site. Before dispersal began, 120 of these sampling units were selected at random and excavated to a depth of 5 cm. Each was placed in a plastic tray and transferred to a greenhouse where the soil seed bank was estimated indirectly by the number of emerging seedlings recorded each month until emergence ceased. Seed input into the 120 sampling units that remained in the field was prevented by covering them with fine mesh during seed dispersal (January-May 2000 and January-May 2001) and seedling emergence was recorded every month in each of these sampling units until December 2001, after which no more seedlings emerged. Seedling establishment was estimated as the proportion of emerged seedlings that survived in the field over two years (December 2000 to December 2002). The number of seeds remaining in the soil seed bank (H in Fig. 1) was estimated as the difference between (E) and (F) over two years (December 20002002). The number of seeds that died because of fire (I in Fig. 1), was estimated as the difference between E and E, where E is the number of seeds that emerged from the soil seed bank at the unburnt site, determined using the same methods as in the burnt site). The probabilities of transitions 1-8 (Fig. 1) between the phases in the model were estimated as follows: transition 1 = B/A; transition 2 = C/B; transition 3 = D/C; transition 4 = 1 - transition 2; transition 5 = F/E; transition 6 = G/F; transition 7 = 1 transition 5; and transition 8 = I/E'). Seed bank dynamics: the role of the seed rain and the fire regime To examine how seed bank dynamics and seedling establishment of each of the three species were affected by seed rain and fire regime, we used the transitional values generated above and ran a model, in which the Germinable seeds produced phase was modified according to two management scenarios, namely without fire, in which it was assumed that fire occurred only in 2000, and with annual fires, in which it was assumed that fire occurred each year from 2000 to 2002 inclusive, with neighbouring unburnt areas remaining as sources of seeds.

9 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 Germinable seeds produced in 2001 and 2002 were estimated following the methods used to determine seed production in 2000. Because seed longevity does not exceed two years for any of the species (see Results), the model only considered seeds present in the soil seed bank in 2001 and 2002, and those arriving by dispersal in the previous years (2000 and 2001) as sources for recruitment of new plants. Seeds that did not germinate in 2000, 2001 or 2002 were considered to be dead. Statistical Analysis The proportion of seeds that germinated in the laboratory under different light and temperature conditions was compared by ANOVA, after arcsineprop transformation to normalise the data, with light, temperature and their interaction as main effects. A Tukey comparison of means test was used to determine significance of differences between treatments when ANOVA results were significant. The number of germinable seeds produced in 2000, 2001 and 2002 was compared using Kruskal-Wallis tests (InfoStat 2005). Results Type of seed bank For all three species, seedlings emerged both from recently dispersed seeds and from seeds that had been in the soil seed bank for up to one year. Virtually no seedlings emerged from seeds dispersed two years previously for any of the species (Fig. 2a). For all species, freshly harvested seeds, and seeds stored for a year under laboratory conditions, showed higher germination percentages (50-90%) than seeds that had been stored for two years (< 4%, Fig 2b). Germination of seeds of all species was affected by treatment (Fig. 2c). For Acanthostyles buniifolius and Baccharis dracunculifolia germination was affected by the interaction between light and temperature ( F1, 12 = 5.4, P = 0.038 and F1, 12 = 6.2, P = 0.028,

10 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 respectively). For A. buniifolius the lowest germination occurred in darkness at constant temperature (65%), whereas for B. dracunculifolia, the lowest percentages were recorded in darkness at constant and alternating temperatures (7% and 30%, respectively). For B. pingraea, only light affected germination (F1, 12 = 126.4, P < 0.0001). Germination was higher in light (on average 76%) than in dark treatments (on average 26%). Seed bank dynamics model For all three species, the seed rain produced seedlings immediately after dispersal, in autumn (March-June), whereas seedlings arising from the soil seed bank emerged during the spring months (September-December). For Acanthostyles buniifolius (Fig. 3a), 0.08% of the germinable seeds produced in 2000 reached the soil in the form of seed rain, of which 73% produced seedlings and the rest entered the soil seed bank. Thirty six percent of the seedlings that emerged from the seed rain in 2000 were still alive in 2002 (0.83 plants m-2). Of the seeds that were already present in the soil seed bank before seed dispersal, 66% were killed by fire in 2000. Of the remaining seeds, 73% produced seedlings and 13% of these were still alive in 2002 (0.62 plants m-2). Seeds that did not germinate from the seed bank that existed prior to seed dispersal remained in the soil seed bank (1.67 seeds m-2). For Baccharis pingraea (Fig. 3b), 9% of the germinable seeds produced in 2000 were added to the soil as seed rain, of which 81% produced seedlings in the same year. Only 3% of these seedlings were still alive in 2002 (2.10 plants m-2). Of the seeds that were present in the soil seed bank before 2000, 45% were destroyed by the fire in that year. Of the remaining seeds, 83% produced seedlings and 6% of these were still alive in 2002 (1.20 plants m-2). The 17% of seeds that did not germinate remained in the seed bank (3.75 seeds m-2). For Baccharis dracunculifolia (Fig. 3c), 0.07% of the germinable seeds produced in 2000 reached the soil as seed rain, and 73% of these seeds produced seedlings in 2000. Forty

11 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 five percent of these seedlings were still alive in 2002 (5.40 seedlings m-2). Of the seeds recorded in the soil seed bank prior to 2000, 46% were killed by the fire in that year. Of the remaining seeds, 77% produced seedlings in 2000 and 28% of these seedlings were still alive in 2002 (5.60 seedlings m-2). The 23% of seeds that did not germinate remained in the seed bank (5.84 seeds m-2). Seed bank dynamics: the role of the seed rain and the effect of fire regime The number of germinable seeds produced m-2 showed significant variation between years for all three species (Acanthostyles buniifolius: Kruskal-Wallis = 54.3, P < 0.0001; Baccharis pingraea: KW = 10.7, P = 0.004, B. dracunculifolia: KW = 20.88, P < 0.0001, Fig. 4). For A. buniifolius, seed production was on average 23 times higher in 2001 (98746 germinable seeds m-2) than in the other two years (Fig. 4a). For B. pingraea it was on average 5.3 times higher in 2000 and 2002 (1085 germinable seeds m-2) than in 2001 (Fig. 4b), and for B. dracunculifolia, seed production was on average 3.6 times higher in 2000 and 2001 (24386 germinated seeds m-2) than in 2002 (Fig. 4c). In the without fire scenario, the model predicted that seedling establishment would be higher for all species in years in which seed production and seed rain were also higher (Fig. 5). For Baccharis dracunculifolia, seedling establishment also depended on seed rain in years with low seed production. In contrast, for Acanthostyles buniifolius and B. pingraea, seedling establishment in years with low seed rain depended mainly on recruitment of seedlings from the seed bank. In the annual fire scenario, the effect of fire was more evident on all species in years in which seedling establishment was lower. The magnitude of this effect was greater in years in which seedling recruitment derived mainly from the seed bank rather than from the seed rain (e.g. A. buniifolius in 2002 and B. pingraea in 2001). Discussion

12 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 Type of seed bank In fire-prone environments, persistent seed banks are important for the maintenance of populations of obligate seeders (Keeley 1987, Whelan 1995, Auld et al. 2000), and they often make a significant contribution to the long-term persistence of populations of species that resprout after fire has killed their above-ground parts (Auld and Denham 2006). Our results showed that, regardless of their regeneration strategies (sprouter, facultative sprouter, obligate seeder, respectively), seeds of Acanthostyles buniifolius, Baccharis pingraea and B. dracunculifolia can remain viable in the soil for up to two years and that a high proportion germinate in light conditions. Galindez et al. (2009b) also reported that these species have small seeds and require light for germination, and other studies have identified small seed size, and light as a requirement for germination, as characteristic of species with persistent soil seed banks (Thompson et al 1993, Auld et al. 2000; Jankowska-Blaszczuk and Daws 2007, Ortega-Baes et al. 2010, Flores et al. 2011). Seed bank dynamics model Some shrub species in fire-prone environments can produce many more seeds in a single good year than are stored in the soil, suggesting that many of the seeds produced each year are lost, either by being transported elsewhere or by being destroyed in situ (Keeley 1977). Our results support these observations. All three species produced many seeds after fire, but less than 10% of the seeds produced were incorporated into the soil, probably either because they were dispersed elsewhere because of the absence of standing biomass forming a physical barrier to seed movement after fire, or because of predation. Nevertheless, a high proportion of the seeds of all species that reached the soil germinated immediately after dispersal. In contrast, a number of other studies in fire-prone plant communities have shown that newly emerging seedlings are derived primarily from the soil seed bank rather than from seeds dispersed from neighbouring areas (Keeley and Zedler 1978, Moreno and Oechel 1992,

13 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 Alexander and Schrag 2003). However, Oliveira (1998) obtained similar results to those reported here, with seeds of many wind-dispersed woody species of savanna habitats germinating immediately after dispersal. Studies in fire-prone environments have shown that, even in low intensity fires, temperatures are high enough to kill most seeds on the soil surface, and survival and germination of seeds in the soil seed bank depends on their depth of burial (Zammit and Zedler 1988, Auld and Denham 2006). This study showed that many of the seeds in the soil seed bank were killed by fire, but (Galndez, personal observation) has also shown that a high proportion of seeds buried at depths of 1 cm are able to produce seedlings if they survive fire, presumably because germination is stimulated by factors including increased soil temperature, increased light and nutrient availability and/or removal of established plants (Keeley et al. 1985, Vil and Lloret 2000, Ooi et al. 2006). Thus, the impact of fire on the sizes of the seed banks of the study species would be especially severe when the seed rain in the previous year has been small or non-existent. This would be the case when fire has been intense and/or it has covered a large area. The probability of survival of seedlings in the field, as measured over two years, was low for all three species and it differed between seedlings that emerged from the seed rain and seedlings emerging from the soil seed bank. For A. buniifolius and B. dracunculifolia, seedlings germinating from the seed rain had higher survival than those germinating from the seed bank, whereas for B. pingraea, seedlings germinating from the seed rain had lower survival. This may be because germination of A. buniifolius and B. dracunculifolia occurred mainly in March-April, when temperature and rainfall are relatively high, whereas for B. pingraea, most seedlings emerged in May-June when conditions are colder and drier. GarcaNuez et al. (2001) reported that germination at the beginning of the summer is beneficial in savanna habitats with marked seasonality, because it maximizes the length of the first

14 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 growing period. Despite the differences in proportional survival, seedlings derived from the seed rain and seed bank yielded similar densities of surviving plants after two years in all three species. Seed bank dynamics: the role of the seed rain and the influence of fire regime Annual seed production fluctuated several-fold in all three species. Galndez et al. (2009a) reported that all three species allocate a higher proportion of their aerial biomass to seed production in years in which they first reproduce than in later years of life. Higher reproductive effort has been reported to be advantageous in environments where the timing of fire is unpredictable (Parker and Kelly 1989; Auld and Denham 2006). For A. buniifolius and B. pingraea, high seed production in 2001 and 2000, respectively, was associated with the first reproduction of re-sprouted plants in the burnt area, and for B. dracunculifolia, high seed production in 2000 was associated with the first reproduction of 3-year-old plants in the unburnt site that had previously been burnt in 1997. Early establishment is less dependent on annual seed rain in species with persistent seed banks than in those with transient seed banks (Auld et al. 2000). However, germination, and death of seeds due to loss of viability or fire (Whelan 1995, Hyatt and Casper 2000), leads to a requirement for the addition of seed rain to the soil every year in order to maintain the seed bank and contribute to the establishment of new plants, as suggested by the model presented here. When the model was run without fire, seedling establishment depended both on the number of seeds deposited on the soil surface in a given year and on the number of seeds remaining in the seed bank from the previous year. In contrast, with annual fire, and with unburnt areas remaining to provide a source of seed rain, establishment depended much more on seed rain than on the seed bank remaining after fire. This suggests that if no reproducing plants remain as seed sources in neighbouring unburnt areas following fire, establishment of new plants will be virtually dependent on germination from the seed bank.

15 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 Moreover, if there is no seed rain for two or more consecutive years, the seed banks of these species will be completely depleted, resulting in no seedling emergence. Therefore, seed rain is essential both for the maintenance of their seed banks and for the long-term persistence of their growing populations. These results explain the changes that have been observed in the vegetation of South American temperate savannas during the last 50 years. When these habitats were subject to frequent fires (i.e. every one or two years) the shrub species studied here occurred at low densities, and their persistence was strongly dependent on the germination of seeds arriving as seed rain from neighbouring unburnt areas. As fire became less frequent, these shrubs increased in density and in the area they occupied as a result of the establishment of seedlings emerging both from the seed rain and the seed bank. By revealing the crucial role of these two sources of seed in contributing to the population dynamics of shrub species with different regeneration strategies, this study provides vital information on which management practices for conservation of the temperate savannas of South America can be based. Implementation of a management regime that only restores frequent and small-scale (e.g. < 50 ha) fires will not be effective in controlling shrub encroachment, because populations can be sustained by seeds arriving as seed rain from nearby unburnt habitat. Very large-scale fire is almost impossible to apply as a management tool at El Palmar National Park because of technical and practical difficulties; consequently unburnt areas always remain that can act as seed sources for these species. Current management to prevent the encroachment of woody species in many temperate savannas includes prescribed fires in combination with other actions including mechanical and chemical control, and grazing by livestock (Redhead et al. 2012, Brudvig et al. 2011). However, grazing has been prevented at El Palmar National Park since this protected area was created in 1965, even though it would make a significant contribution to the control of shrub encroachment. In support of this contention, Galndez

16 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 (2008) has reported that B. pingraea and B. dracunculifolia are not present in areas where both burning and grazing have taken place. We therefore recommend the application of both frequent burning and grazing, to restore the savanna vegetation to its former condition. Conclusion We conclude that Acanthostyles buniifolius, Baccharis pingraea and B. dracunculifolia possess short-term persistent seed banks and high annual variability in seed production. In environments without fire, establishment depends strongly on germination both from the seed rain and the soil seed bank, whereas in environments subjected to annual fires, establishment is primarily dependent on germination of seeds arriving in the annual seed rain, regardless of species regeneration strategies. These results help to explain changes observed in the vegetation of these South American temperate savannas during the last 50 years, revealing the crucial role of the soil seed bank and seed rain in maintaining populations of shrub species with different regeneration strategies and providing valuable information for the design of management practices for their control. Acknowledgements We thank S. Shring for her assistance with statistical analysis and C. Pereyra for his assistance in field data collection. This research was supported by the Agencia Nacional para la Promocin Cientfica y Tcnica (PICT 08-08962). Additional funds were provided by the University of Buenos Aires (UBACyT G 095). References Alexander HM, Schrag AM (2003) Role of soil seed banks and newly dispersed seeds in population dynamics of the annual sunflower, Helianthus annuus. Journal of Ecology 91: 987998.

17 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 Altesor A, Pieiro G, Lezama F, Jackson R B, Sarasola M, Paruelo JM (2006) Ecosystem changes associated with grazing in subhumid South American grasslands. Journal of Vegetation Science 17: 323332. Auld TD, Denham AJ (2006) How much seed remains in the soil after a fire? Plant Ecology 187: 1524. Auld TD, Keith DA, Bradstock RA (2000) Patterns in longevity of soil seed banks in fireprone communities of south-eastern Australia. Australian Journal of Botany 48: 539548. Bond WJ, van Wilgen BW (1996) Fire and Plants. Chapman and Hall, London. Brudvig LA, Blunck HM, Asbjornsen H, Mateos-Remigio VM, Wagner SA, Randall JA (2011) Influences of woody encroachment and restoration thinning on overstory savanna oak tree growth rates. Forest Ecology and Management 262: 1409 1416. Burkart A (1974) Flora Ilustrada de Entre Ros (Argentina). Parte VI. Coleccin Cientfica del Instituto Nacional de Tecnologa Agropecuaria (INTA), Argentina Cabral AC, De Miguel JM, Rescia AJ, Schmitz MF, Pineda FD (2003) Shrub encroachment in Argentinean savannas. Journal of Vegetation Science 14: 145 152. Campbell ML, Clarke PJ (2006) Seed dynamics of resprouting shrubs in grassy woodlands: seed rain, predators and seed loss constrain recruitment potential. Austral Ecology 31: 10161026. Clarke PJ, Knox KJE (2002) Post-fire response of shrubs in the tablelands of eastern Australia: do existing models explain habitat differences? Australian Journal of Botany 50: 5362. Cohen S, Braham R, Sanchez F (2004) Seed bank viability in disturbed longleaf pine sites. Restoration Ecology 12: 503515. Flores J, Jurado E, Chapa-Vargas L, et al (2011). Positive photoblastism in cacti seeds and its relationship with some plant traits. Environmental and Experimental Botany 71: 79-88.

18 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 Galndez G, Biganzoli F, Ortega-Baes P, Scopel AL (2009a) Fire responses of three cooccurring Asteraceae shrubs in a temperate savanna of South America. Plant Ecology 202: 149-158. Galndez G.,Ortega-Baes P, Daws MI, Shring S, Scopel AL and Pritchard HW (2009b) Seed mass and germination in Asteraceae species of Argentina. Seed Science and Technology 37: 786-790. Galndez G (2008) Arbustizacin de las sabanas mesopotmicas: Estrategias de regeneracin de los arbustos nativos y sus repuestas al fuego y a la herbivora. Doctoral thesis, School of Agronomy, University of Buenos Aires, Argentina, pp 87. Garca-Nuez C, Azcar A, Silva JF (2001) Seed production and soil seed bank in three evergreen woody species from a neotropical savanna. Journal of Tropical Ecology 17: 563-576. Goveto L (2005) Ocurrencia Histrica de Fuegos en la Sabana del Parque Nacional el Palmar: Evidencias Climticas y Florsticas. MSc thesis, School of Agronomy, University of Buenos Aires, Argentina, pp 116. Hyatt LA, Casper BB (2000) Seed bank formation during early secondary succession in a temperate deciduous forest. Journal of Ecology 88: 516527. InfoStat (2005). InfoStat versin 2005. Grupo InfoStat, FCA, Universidad Nacional de Crdoba, Argentina. Jankowska-Blaszczuk M, Daws MI (2007) Impact of red: far red ratios on germination of temperate forest herbs in relation to shade tolerance, seed mass and persistence in the soil. Functional Ecology 21: 10551062. Keeley JE (1987) Role of fire in seed germination of woody taxa in California chaparral. Ecology 68: 434443.

19 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 Keeley JE, Morton BA, Pedrosa A, Trotter P (1985) Role of allelopathy, heat and charred wood in germination of chaparral herbs and subfrutescents. Journal of Ecology 73: 445 458. Keeley JE (1981) Reproductive cycles and fire regimes. In USDA Forest Service General Technical Report WO-26, Fire Regimes and Ecosystem Properties, pp. 231-277. Keeley JE, Zedler PH (1978) Reproduction of chaparral shrubs after fire: a comparison of sprouting and seeding strategies. American Midland Naturalist 99: 142 161. Keeley JE (1977) Seed production, seed populations in soil and seedling production after fire for two congeneric pairs of sprouting and nonsprouting chaparral shrubs. Ecology 58: 820829. Lamont BB, Wiens D (2003) Are seed set and speciation rates always low among species that resprout after fire, and why? Evolutionary Ecology 17: 277 292. Lamont BB, Le Maitre DC, Cowling RM, Enright NJ (1991) Canopy seed storage in woody plants. Botanical Reviews 57: 277317. Moreira AG (2000) Effects of fire protection on savanna structure in Central Brazil. Journal of Biogeography 27: 10211029. Moreno JM, Oechel WC (1992) Factors controlling postfire seedling establishment in southern California chaparral. Oecologia 90: 5060. Oliveira PE (1998) Fenologia e biologia reproductiva das especies de Cerrado. In: Sano SM, Pedrosa de Almeida S (ed) Cerrado: ambiente e flora, EMBRAPA, Planaltina, DF, Brazil, pp. 169-192. Ooi MKJ, Auld TD, Whelan RJ (2006) Dormancy and the fire-centric focus: germination of three Leucopogon species (Ericaceae) from south-eastern Australia. Annals of Botany 98: 421430.

20 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 Ortega-Baes P, Aparicio M, Galndez G, del Fueyo P, Shring S, Rojas-Archiga M (2010) Are cactus life forms related to germination responses to light? A test using Echinopsis species. Acta Oecologica 36: 339-342. Parker VT, Kelly VR (1989) Seedbanks in California chaparral and other Mediterranean climate shrublands. In: Leck MA, Parker VT, Simpson RL (ed) Ecology of Soil Seed Banks. Academic Press, San Diego, USA. Redhead J, Cuevas-Gonzales M, Smith G, Gerard F, Pywell R (2012) Assessing the effectiveness of scrub management at the landscape scale using rapid field assessment and remote sensing. Journal of Environmental Management 97: 102-108. Silva JF, Zambrano A, Farias MR (2001) Increase in woody component of seasonal savannas under different fire regimes in Calabozo, Venezuela. Journal of. Biogeography 28: 977- 983. Thompson K (1993) Persistence in soil. In: Hendry GAF, Grime JP (ed) Methods in Comparative Plant Ecology: A Laboratory Manual Chapman and Hall, London, UK., pp 252. Thompson K, Band SR, Hodgson JG (1993) Seed size and shape predict persistence in soil. Functional Ecology 7: 236241. Vil M, Lloret F (2000) Woody species tolerance to expansion of perennial tussock grass Ampelodesmos mauritanica after fire. Journal of Vegetation Science 11: 597 606. Whelan RJ (1995) The Ecology of Fire. Cambridge Studies in Ecology. Cambridge University Press, Cambridge, UK. Whelan RJ (1986) Seed dispersal in relation to fire. In: Murray DR (ed) Seed Dispersal, Academic Press, Sydney, Australia, pp 237271. Zammit C, Zedler P.H. (1988) The influence of dominant shrubs, fire and time since fire on soil seed banks in mixed chaparral. Vegetatio 75: 175187.

21 484 485 486 Zuloaga FO, Morrone O (1999) Catlogo de las Plantas Vasculares de la Repblica Argentina II. A-E. Missouri Botanical Garden, St. Louis, Missouri, USA, pp 621.

22

487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510

Figure captions Fig.1. Conceptual model of the seed bank and seedling establishment dynamics. Numbers next to the arrows are the proportions of the seeds or seedlings making the transition between the stages linked by the arrows. See text for a full description. Fig.2. (a) Mean ( SE) number of seedlings m-2 emerging at the burnt site from seeds dispersed in 2000 (), from seeds remaining in the soil seed bank for one year (), and from seeds remaining in the soil seed bank for two years ( ), for the three study species. (b) Mean ( SE) seed germination percentages of fresh seeds (), seeds stored for one year under laboratory conditions (), and seeds stored for two years under laboratory conditions ( ), for the three study species. (c) Mean ( SE) seed germination percentages under different light and temperature conditions (: 25 C with light, : 20/30 C with light, : 25 C without light, : 20/30 C without light) for the three study species. Fig.3. Seed bank and seedling establishment dynamics of (a) A. buniifolius, (b) B. pingraea, and (c) B. dracunculifolia, after the fire of 2000 at the burnt site. The values in parentheses are numbers m-2 and the values next to the arrows are the proportions of the seeds or seedlings making the transition between the stages linked by the arrows. Fig.4. Mean ( SE) number of germinable seeds produced m-2 by (a) A. buniifolius, (b) B. pingraea, and (c) B. dracunculifolia, in each of the studied years at the burnt and unburnt sites. Note differences in scale on the y-axes. Fig. 5. Seedling establishment (seedlings m-2) from the seed rain () and from the soil seed bank () in the without fire scenario, and from the seed rain () and from the soil seed bank () in the annual fire scenario , for each year of (a) A. buniifolius, (b) B. pingraea, and (c) B. dracunculifolia,

23 511 512 513 514 515 516 517 518 519 520 521 Fig. 1.
Seeds remaining in the soil seed bank (H) Seeds that died because of fire (I)
(8)

Germinable seed produced (A)

(1)

Seed rain (B)

Seedling emergence (C) (F)

(3) (6)

Seedling establishment (D) (G)

Soil seed bank (E) (E)

24

(a)
12 10

Seedlings m-2

8 6 4 2 0
A. buniifolius B. pingraea B. dracunculifolia

(b)
100

Germination (%)

80 60 40 20 0
A. buniifolius B. pingraea B. dracunculifolia

(c)
100

Germination (%)

80 60 40 20 0
A. buniifolius B. pingraea B. dracunculifolia

522 523

Fig. 2.

25 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
Germinable seeds produced
(4047) 0.266

0.0008

Seed rain
(3.12) 0.734 0.362 0.135

(a)

Seedling emergence
(2.29) (4.58)

Seedling establishment
(0.83) (0.62)

0.666

Seeds that died because of fire

Soil seed bank

0.733

0.267

Seeds remaining in the soil seed bank

Germinable seeds produced

0.0905

Seed rain
(80.42) 0.811

(b)

0.189

(65.21) (18.54) 0.065 0.448

Seedling emergence

0.032

Seedling establishment
(2.10) (1.20)

Seeds that died because of fire

Soil seed bank

0.832

Seeds remaining in the soil seed bank

Germinable seeds produced

0.0007

Seed rain
(16.25) 0.731

(c)

0.269

(11.88) (19.79) 0.283

Seedling emergence

0.454

Seedling establishment
(5.40) (5.60)

Seeds that died because of fire

0.461

Soil seed bank

0.772

Seeds remaining in the soil seed bank

Fig. 3.

576
4

575
-2

Fig. 4.

Seed number ( x 10 ) m
10 12 0 2 4 6 8
0 1 2 3 4 5

Seed number ( x 10 ) m
5

-2

Seed number ( x 104 ) m-2

0
(b)

1 (c)

(a)

2000

Year

2001
26

2002

27

20.4 20.2 20.0 4 3 2 1 0

(a)

Seedlings m -2

4 3 2 1 0

(b)

Seedlings m -2

12 11 10 9 8 7 6 5 4 3 2 1 0

(c)

Seedlings m -2

2000

2001

2002

Year
577 578 Fig. 5.