Journal of Ecology 2007 95, 771–779

A field experiment on climatic and herbivore impacts on post-fire tree regeneration in north-western Patagonia

Blackwell Publishing Ltd

NORLAN TERCERO-BUCARDO, THOMAS KITZBERGER, THOMAS T. VEBLEN* and ESTELA RAFFAELE Laboratorio Ecotono, Universidad Nacional del Comahue, Quintral 1250, 8400 Bariloche, Argentina, and *Department of Geography, University of Colorado, Boulder, CO 80309-0260, USA

Summary 1 Wildfires are predicted to increase in many ecosystems in relation to globally increasing temperatures but future patterns of post-fire vegetation change are largely unknown, particularly when there are synergistic effects from introduced biota. In the late 1990s northern Patagonia, Argentina, experienced extreme droughts which led to severe wildfires affecting a range of Andean ecosystems. 2 We experimentally examined how variations in moisture, temperature and herbivory by livestock affect post-1999 fire patterns of the three main tree species. Over two years we monitored, in three forest types, the survival and growth of tree seedlings in a factorial warming (+2 °C)/livestock exclosure/watering experiment. 3 Seedling survival in the warmed treatments and in the controls was nil for the evergreen Nothofagus dombeyi and the conifer Austrocedrus chilensis at the two lowelevation experimental sites. Survival of the subalpine Nothofagus pumilio in the warmed treatments at high elevation tended to be lower than in the control; for all treatments of warming alone there were no significant differences compared with the controls. 4 In all three forest types, increased water availability was essential for higher rates of tree seedling survival. Doubling water availability during the growing season resulted in up to fourfold increases in seedling survival and up to threefold increases in seedling biomass. 5 In the subalpine forest, livestock reduced seedling survival by c. 30% in non-watered treatments compared with watered treatments, probably due mainly to soil desiccation and to consumption of or damage to facilitating plants. In contrast, at lower elevation, where livestock pressure was lower, seedling survival of N. dombeyi and A. chilensis tended to be higher in unfenced sites, possibly due to reduced competition from highly palatable shrub species. 6 General circulation models predict a warming–drying trend in northern Patagonia during the twenty-first century. The resulting increase in wildfire is likely to be followed by inadequate tree regeneration and conversion from forest to shrubland cover types. This and similar studies suggest that under relatively slight changes in regional climate, increased fire occurrence interacting synergistically with moisture limitations will result in long-lasting displacements of forest by more xeric vegetation shrublands. Key-words: Austrocedrus, climate change, disturbances, fire, livestock, Nothofagus, Patagonia, seedling survival, tree regeneration Journal of Ecology (2007) 95, 771–779 doi: 10.1111/j.1365-2745.2007.01249.x

Introduction

© 2007 The Authors Journal compilation © 2007 British Ecological Society

In the context of global warming, climatically sensitive disturbances such as fire are likely to play major roles in determining future landscape patterns and trajectories Correspondence: Norlan Tercero-Bucardo (tel./fax +54 2944 422111; e-mail [email protected]).

(Franklin et al. 1992; Gardner et al. 1996). In forested landscapes, fire is likely to accelerate shifts to new species assemblages under different climates because of the sensitivity of tree regeneration to climatic factors (Brubaker 1986; Dunwiddie 1986; Asselin & Payette 2005). Thus, although the structure and composition of a mature forest may respond slowly to a new climatic regime, the pattern of tree regeneration following

772 N. Tercero-Bucardo et al.

© 2007 The Authors Journal compilation © 2007 British Ecological Society, Journal of Ecology, 95, 771–779

coarse-scale disturbance is more likely to be sensitive to climatic variation. The importance of climatic conditions to post-fire patterns of vegetation change has long been recognized (Davis 1986; Overpeck et al. 1990) but there have been few systematic studies of how climatic conditions affect post-fire tree regeneration (but see Arseneault & Payette 1992; Landhäusser & Wein 1993). Based on analysis of vegetation changes following a severe fire that occurred in 1999 in north-western Patagonia, Argentina, we provide one of the first experimental studies of how altered climate may influence post-fire tree regeneration. In many ecosystems it is difficult to disentangle the effects of climatic variation from the influence of altered land uses, such as livestock raising (Peters et al. 2006), which are also major drivers of global environmental change (Sala et al. 2000). Herbivory, especially by livestock, is well recognized as a major determinant of post-fire vegetation trajectories (Payette et al. 2000; Fuhlendorf & Engle 2004; Vandvik et al. 2005). Similarly, post-disturbance patterns of herbivory are believed to be key to future vegetation changes under altered climates (Lenihan et al. 1998; Keane et al. 1999; Schmoldt et al. 1999). However, we generally lack empirical studies that simultaneously consider the effects of herbivory and climatic variation on post-fire patterns of vegetation change. In north-western Patagonia, previous research has shown that synergistic relationships between fire and grazing by cattle are important drivers of vegetation dynamics at both broad (Kitzberger et al. 2005b) and local spatial scales (Gowda & Raffaele 2004; Blackhall 2006). Previous research in this landscape has also documented profound ecological effects of climatic variability, including a marked trend toward warmer and drier conditions during the late twentieth century, and specifically includes effects on forest ecosystems mediated through changes in fire severity (Kitzberger et al. 2005a), fire extent (Veblen et al. 1999) and landscape controls on fire behaviour (Mermoz et al. 2005). Previous studies based on tree-ring reconstructions (Villalba & Veblen 1998; Daniels & Veblen 2004) and seedling watering experiments (Kitzberger et al. 2000) have demonstrated the importance of annual variation in temperature and moisture to tree seedling establishment and survival near the lower treeline at the steppe–forest ecotone and near the upper treeline at the forest– tundra ecotone. However, the potentially synergistic effects of climatic variation and herbivory by livestock on tree regeneration in recently burned forest environments have not been examined previously. In the current study, we exploit the occurrence of severe forest fires in 1999 to examine experimentally how variation in moisture, temperature and herbivory by livestock affect post-fire patterns of the three principal tree species of this region. Specifically, we examine effects of watering, warming and livestock on the survival and growth of tree seedlings in a series of post-fire forest types from low-elevation to subalpine sites.

Materials and methods study sites and species The study was conducted in southern Nahuel Huapi National Park, Argentina (NHNP; c. 41°8′S, 71°19′W) during the two growing seasons (2002–03, 2003–04), in two sites severely burned in January 1999: Lago Los Moscos (LM, c. 41°21′65″S, 71°38′54″W; 850 m) and Cerro Donat (CD, c. 41°26′19″S, 71°36′18″W; 1150 m; Fig. 1a). Precipitation in this region occurs mainly from April to September, as snow and rain, whereas December to February is usually dry. At this latitude, mean precipitation decreases abruptly from c. 4000 mm year–1 on the western side of the Andes to less than 500 mm year–1, only 80 km to the east (De Fina 1972). Mean precipitation in the study area is approximately 1700 mm year–1 (Barros et al. 1983). Species composition of the vegetation changes sharply along this west-to-east decline in precipitation and the decline in elevation from the Andean cordillera to the Patagonian plains (Veblen et al. 1992). Subalpine forest of deciduous Nothofagus pumilio (Poepp. & Endl.) Krasser occurs above 1000–1100 m over the entire area. In the wetter area the lowland rain forests are mainly dominated by the evergreen Nothofagus dombeyi (Mirb.) Oerst. In the intermediate part of the precipitation gradient, at low elevations, N. dombeyi forms monospecific mesic forests or mixed stands with the conifer Austrocedrus chilensis (D. Don) Pic. Serm. & Bizzarri at drier sites; in the eastern area A. chilensis forms relatively open woodlands. In the western and central areas forest understoreys are typically dominated by dense and tall (> 2 m) populations of the bamboo Chusquea culeou Desv. Tall and dense shrublands occur throughout the west-to-east precipitation gradient at sites that are either not edaphically suitable for development of tall forest or as successional communities that develop after burning of tall forest (Mermoz et al. 2005). Following fire, tall forests dominated by the obligate seed reproducers (N. dombeyi, N. pumilio and/ or Austrocedrus) sometimes are replaced by shrub communities dominated by woody species, including C. culeou, that vigorously resprout after fire. In the three stands studied, the vegetation prior to burning consisted of (i) evergreen forest dominated by N. dombeyi with a typical dense understory of C. culeou; (ii) a mixed forest of the evergreen conifer A. chilensis and the evergreen N. dombeyi; and (iii) a subalpine forest formed by the deciduous N. pumilio (see more detail in Gallopín 1978). The first two stands are located at Lago Los Moscos and the third stand is on Cerro Donat (Fig. 1a). As expected for ecosystems recovering from a recent fire, in the experimental sites understorey plant cover and richness (number of species) increased approximately twofold between 2002 and 2004. Total plant cover changed from 40% to 80% and mean richness from 4 to 10 species in 2-m2 plots.

773 Climatic and herbivore impacts on tree regeneration

Fig. 1 (a) Severely burned Nothofagus pumilio subalpine forest at Cerro Donat, one of the three forest types used in this study; photo by Guillermo Amico. (b) The hexagonal Open Top Chambers (OTCs) used to raise the temperature of plots experimentally.

experimental design

© 2007 The Authors Journal compilation © 2007 British Ecological Society, Journal of Ecology, 95, 771–779

In the three forest types noted above, we selected areas that were severely and homogeneously burned in 1999 to install livestock exclosures and controls. In each forest type four 25 × 25 m plots (macro-plots) were fenced against livestock while four areas were left unfenced. The macro-plot exclosures and controls constitute a paired plot experimental design (Sokal & Rohlf 1981). Within each experimental macro-plot eight 2-m2 smallplots (experimental units) were placed, giving a total of 96 experimental plots in the three forest types. Water and temperature were experimentally manipulated in the small-plots (see detail below) at each of the three forest types. We applied a factorial split-plot experimental design (Potvin 1993; Littell et al. 2002). This consisted of four macro-plots or replications, the macro-plot factor with two experimental treatments (FE = fenced and UF = unfenced), and two small-plot experimental factors (WT = watered and WR = warmed). The following combinations of watering and warming treatments were applied: (i) only WT, (ii) only WR, (iii) WT + WR and (iv) control (non-watered or warmed). In each small plot in October 2002 (2002 cohort) we planted 15–20 seedlings of the tree species that had dominated the site before burning. Seedlings of N. pumilio and N. dombeyi were transplanted from natural stands, and seedlings of A. chilensis were

transplanted from a glasshouse. At the initiation of the experiment, seedlings of N. pumilio were 2 years old and seedlings of N. dombeyi and A. chilensis were 2 months old. Prior to planting in the field, all seedlings were transplanted into plastic pots and acclimatized over 1 month in a glasshouse. Due to low survival of seedlings of N. dombeyi and A. chilensis in the first year (0–10%), a second cohort of seedlings of N. dombeyi and A. chilensis was planted in October 2003 (2003 cohort) to ensure a second year of survivors. The seedlings of the 2003 cohort were planted in the same small-plots as the 2002 cohort but were analysed separately. Planted seedling survival was monitored monthly and percentage survival was calculated annually (survival %). In addition, at the end of the study the planted seedlings were harvested and dried separately to obtain total seedling biomass for each small-plot as an estimation of seedling growth under different treatments. Individual seedling cohorts (2002 and 2003) were harvested and dried separately. Watering treatment of small-plots (WT) was achieved by installing an automatic watering system; microsprinklers were located in each watered small-plot. These delivered c. 450 mm of water during the growing season (i.e. approximately twice the mean growing season precipitation). Temperature in warmed plots (WR) was raised by installing 48 hexagonal Open Top Chamber (OTC) portable field glasshouses (Fig. 1b);

774 N. Tercero-Bucardo et al.

each has an effective ground surface of 2 m2, with an open top (69 cm diameter) made of 3-mm-thick plexiglass (LEXAN® Plazit, 2UV-protected polycarbonate), a prototype known to produce a homogeneous c. 2 °C increase in air temperature within the small-plot. To avoid damage during the winter dormant season, the watering system was disabled and glasshouses were removed. In the three recently burned forest types, the experiment compared tree regeneration responses in unfenced and fenced (i.e. livestock excluded) areas to a doubling of water availability and to a c. 2 °C warming during the growing season, both separately and in combination. Temperature (T°) and relative humidity (RH) were monitored with 12 thermal and relative humidity sensors installed from November 2003 to April 2004 at 5 cm above the soil surface in small-plots inside OTCs, outside OTCs and in the controls. Temperature sensors (HOBO® TMC6-HA) were connected to a HOBO® H8 data logger and data were stored in a non-volatile EEPROM memory device. Air temperature and relative humidity measurements were taken at 2-h intervals. Solar radiation was estimated by taking a hemispherical photograph in each small-plot following the procedures of Rich (1990). Photographs were taken at a height of 0 cm (level with the floor) using a levelled digital camera (CoolPix 995 digital camera, Nikon, Japan) aimed at the zenith, using a fish-eye lens with a 180° field of view (FCE8, Nikon). All photographs were taken in overcast weather to ensure homogeneous illumination of the overstorey canopy and adequate contrast between canopy and sky.

data analyses

© 2007 The Authors Journal compilation © 2007 British Ecological Society, Journal of Ecology, 95, 771–779

Daily temperature and relative humidity were averaged by pooling data from sensors in the OTC and the control plots, respectively. Photographs were analysed using WinSCANOPY™ for hemispherical image analysis software, estimating the following groups of parameters: (i) direct site factors fraction (DSF) and indirect site factors fraction (ISF), which are defined as the proportion of direct and diffuse radiation received below the canopy as a fraction of that received above the canopy (Rich 1990); (ii) effective leaf area index (LAI) estimated as half the total leaf area per unit ground surface area, based on the ellipsoidal leaf angle distribution; and (iii) the average growing season direct and diffuse photosynthetically active flux density. All the lightrelated parameters estimated were highly correlated with DSF; therefore DSF was used for statistical analysis (ancovas), as a covariable of treatment effects when model results were statistically significant or nearly so. When covariables were not significant in ancovas, anovas were performed. The General Linear Model (GLM) was used to perform the data analysis following a split-plot experimental design (anovas and ancovas) (Sokal &

Rohlf 1981; Littell et al. 2002). A split-plot experimental design is often applied in a factorial experiment when one factor is more readily applied to a large experimental unit, or main plot (e.g. presence or absence of livestock), and another factor is more easily applied to smaller units within the large unit (e.g. watering and warming). anovas and ancovas were carried out to test the effect of treatment on tree seedling survival (survival %) and biomass (dry weight).

Results During the day (08:00–20:00 h), the OTCs increased the mean air temperatures by c. 2.2 °C compared with the control plots and reduced relative humidity by c. 10% (Fig. 2). At night (21:00–07:00 h), temperature and relative humidity differences between OTCs and controls were neglible.

interactive effect s of watering, warming and herbivory on tree seedlings For seedlings of the 2002 cohort, survival increased in the watered plots (Fig. 3), but there was a significant interaction with fencing (i.e. livestock removal) (WT × FE, P < 0.05; Table 1). For all three species the highest increase in seedling survival was in the watered and unfenced plots (Fig. 3). The largest differences were found in N. pumilio plots where watering increased seedling survival significantly (P < 0.01; Table 1), from 58% to 88% in fenced plots, and from 26% to 83% in unfenced plots (Fig. 3). In A. chilensis plots, watering increased seedling survival from 0 to 15% in fenced plots and from 0 to 20% in unfenced plots, whereas in N. dombeyi plots it increased survival from 0 to 4% in fenced and from 0 to 20% in unfenced plots (Fig. 3). In general, livestock increased differences in survivorship for all tree species between watered and non-watered treatments (Fig. 3). However, the livestock effect was different for the different tree species. For N. pumilio,

Fig. 2 Mean air temperature (T°) and relative humidity (RH) during the day (08:00–20:00 h) and night (21:00–07:00 h) measured at 5 cm above ground in the Open Top Chambers (OTCs) and in control plots. Two-hour means, based on daily measures from November to April, are shown (n = 12).

775 Climatic and herbivore impacts on tree regeneration

Fig. 3 Mean monthly survival of seedlings (%) from the 2002 cohorts of Nothofagus pumilio, Nothofagus dombeyi and Austrocedrus chilensis during two growing seasons (2003 and 2004). The left panel corresponds to fenced (i.e. livestock excluded) and right to unfenced plots. Table 1 Results of analysis of variance (anova) and analysis of covariance (ancova) for a split-plot design, showing the effects of watering (WT), warming (WR) and livestock (FE) on the survival of transplanted seedlings after two growing seasons (2002 cohort) and after one growing season (2003 cohort). No cohort of Nothofagus pumilio was planted in 2003. Data on percentage survival were transformed using an arcsine transformation; variances were homogeneous at P > 0.05 (Cochran’s Q-test). Significant covariable direct site factor fraction (DSF) was used to perform ancovas for the 2002 N. pumilio cohort and 2003 N. dombeyi. DSF results are only shown for ancovas. The table includes only the main factors, covariables and significant interactions. Bold values indicate significant results. Error A = whole-plot error and Error B = split-plot error (Littell et al., 2002) N. pumilio

© 2007 The Authors Journal compilation © 2007 British Ecological Society, Journal of Ecology, 95, 771–779

N. dombeyi

Source

d.f.

MS

F

P

2002 cohort WT WR FE FE × WT DSF Error A Error B

1 1 1 1 1 3 17

2.692 0.121 0.230 0.316 0.276 0.032 0.071

38.66 1.745 7.183 4.546 3.968

< 0.01 0.20 0.07 < 0.05 0.06

2003 cohort WT WR FE DSF Error A Error B

d.f.

MS

1 1 1 1

0.196 0.001 0.049 0.250

3 18

0.007 0.049

1 1 1 1 3 17

8.402 0.110 0.095 0.273 0.021 0.088

A. chilensis F

4.824 0.049 7.605 5.066

95.99 1.256 3.829 3.126

P

< 0.05 0.89 0.07 < 0.05

< 0.01 0.28 0.11 0.09

d.f.

MS

F

P

1 1 1 1

2.433 0.017 0.017 0.526

30.01 0.217 2.373 4.206

< 0.01 0.64 0.22 < 0.05

3 18

0.072 0.081

1 1 1

5.316 0.033 0.001

98.43 0.610 0.004

< 0.01 < 0.01 0.44 0.95

3 18

0.116 0.054

776 N. Tercero-Bucardo et al.

Table 2 Results of analysis of variance (anova) and analysis of covariance (ancova) for a split-plot design, showing effects of watering (WT), warming (WR) and livestock (FE) on the total biomass (dry weight) of planted seedlings. For Nothofagus pumilio no cohort was planted in 2003. Data were analysed only for biomass datasets that were complete, and consequently some cohorts are not included. Data on dry weight of seedling biomass were transformed using a log transformation; variances were homogeneous at P > 0.05 (Cochran’s Q-test). Significant covariable direct site factor (DSF) was used to perform ancovas in N. dombeyi 2003. Bold values indicate significant results N. pumilio 2002 Source

d.f.

WT WR FE DSF Error A Error B

N. dombeyi 2003

MS

F

P

d.f.

MS

F

P

1 1 1

4.582 0.618 0.172

24.25 3.273 0.957

< 0.01 0.08 0.39

< 0.01 0.63 0.37 < 0.05

0.179 0.189

1.333 0.008 0.046 0.172 0.079 0.036

37.66 0.239 0.882 4.852

3 17

1 1 1 1 3 17

livestock decreased survivorship in non-watered plots. By contrast, for N. dombeyi and A. chilensis, livestock increased survivorship of seedlings in watered plots (Fig. 3). For the 2002 cohort, all three tree species show a significant main effect of watering (P < 0.05) that results in the highest increase in seedling survival (Table 1, Fig. 3). For seedlings of the 2003 cohort, watering produced significant increases (P < 0.01) in seedling survival in both species analysed (N. dombeyi and A. chilensis; Table 1). For these species, similar results were found when analysing survival during the first year of the 2002 cohort. Watering increased survival of N. dombeyi from 10% in non-watered plots to 84% in watered plots, while for A. chilensis seedling survival increased from 0.5% in non-watered to 56% in watered plots. Experi-

© 2007 The Authors Journal compilation © 2007 British Ecological Society, Journal of Ecology, 95, 771–779

mental warming alone did not show significant effects on seedling survivorship for any cohort (Table 1). Seedling biomass was only evaluated for the N. pumilio 2002 cohort and the N. dombeyi 2003 cohort (Table 2). High seedling mortality rates in the 2002 cohorts of A. chilensis and N. dombeyi and in the 2003 cohort of A. chilensis precluded the analyses of seedling biomasses. The watering treatment elevated seedling biomasses for the 2002 cohort of N. pumilio (P < 0.01) and the 2003 cohort of N. dombeyi (Table 2, Fig. 4). Watering significantly increased the seedling biomass of N. pumilio from 4.4 g in non-watered plots to 15.1 g in watered plots. The seedling biomass of N. dombeyi increased from 0.14 g in non-watered plots to 0.59 g in watered plots (Fig. 4). Warming and fencing did not significantly affect seedling biomasses (Table 2).

Fig. 4 Total biomass (mean dry weight and SE) of planted seedlings from the 2002 cohort of Nothofagus pumilio (top) and seedlings from the 2003 cohort of Nothofagus dombeyi (bottom). Means associated with different letters are significantly different (P < 0.01) based on analysis of variance (anova) for a split-plot design, showing effects of watering (WT), warming (WR) and livestock (FE).

777 Climatic and herbivore impacts on tree regeneration

© 2007 The Authors Journal compilation © 2007 British Ecological Society, Journal of Ecology, 95, 771–779

Discussion Among all variables manipulated in this study, addition of water was clearly the factor that produced the greatest changes in post-fire responses of northern Patagonian plant communities. Doubling water availability during the growing season resulted in up to a fourfold increase in seedling survival and up to a threefold increase in seedling biomass of the three tree species in the three different forest types burned in 1999. Post-fire seedling survival was not altered by warming either at the high-elevation N. pumilio site or at the low-elevation N. dombeyi and A. chilensis sites. The age of N. pumilio seedlings planted in the experiments was older than that of N. dombeyi and A. chilensis, and probably accounts for the higher biomass and survival rate of the former. Despite the difference in the age of seedlings, the trend of the experimental response was the same for all three species. The positive responses of tree seedlings to experimentally increased moisture availability shown are consistent with the results of retrospective studies in similar nearby sites that have used tree rings to relate seedling establishment and survival to interannual climatic variation. Nothofagus pumilio and A. chilensis seedlings tend to be associated with micro-sites that have the highest soil water availability within their respective forest types. Nothofagus pumilio seedlings survive better in shaded than in sunny portions of canopy gaps (Heinemann et al. 2000); A. chilensis survives better beneath trees that survived fires (Kitzberger et al. 2005b) and beneath nurse shrubs (Kitzberger et al. 2000). Establishment pulses of these two species have been related to multi-year periods of increased moisture availability (Villalba & Veblen 1997; Daniels & Veblen 2004; Kitzberger et al. 2005b). Even for N. pumilio at the upper treeline, where higher temperatures might be expected to favour seedling survival, tree ages do not reflect an increase in seedling establishment or survival during the strong warming trend after the mid-1970s (Daniels & Veblen 2004). Thus, both the current results and retrospective studies consistently show an overriding importance of moisture availability for successful N. pumilio establishment and survival. The radial growth of N. pumilio, even at high elevation, is typically limited by moisture availability except at the wettest sites where higher temperatures enhance growth (Villalba et al. 1997). However, the Cerro Donat study site is a relatively dry site located well to the east of the wet forest zone where higher temperature has been shown to have a favourable influence on radial growth of N. pumilio. Given that temperature interacts with precipitation through evaporation and evapotranspiration to determine moisture availability, previous non-experimental approaches have not been able to distinguish clearly the relative importance of precipitation vs. temperature on either radial growth or seedling survival of the dominant tree species in northern Patagonia. Here, however, we show that, with the partial

exception of N. pumilio, where seedling growth tended to be reduced by warming (possibly by increasing water losses), tree seedling survival and growth were sensitive to changes in water input but not to changes in growing season temperature. Our results show that under both present and doubled watering regimes, a 2 °C temperature increase did not substantially alter either survival or growth rates of tree seedlings regenerating after severe fire. This suggests that increased temperatures do not substantially increase existing water deficits of seedlings so that desiccation risk is not increased. With the exception of N. pumilio, for which warming decreased growth rates, temperature increases did not modify net carbon uptake of seedlings, possibly suggesting that increases in carbon gain due to higher biotemperatures are counterbalanced by higher evapotranspiration and stomatal closure. In no case did warming increase survival or growth, which implies that none of the ecosystems studied is limited by temperature. Herbivory by cattle in post-fire areas substantially amplified the differences in seedling survival between watered and non-watered treatments in the three forest types. In the subalpine N. pumilio forest, grazing reduced seedling survival by c. 30% in non-watered treatments compared with watered ones. Because grazing did not influence survival in the watered treatments we assume seedling mortality was not related to physical damage or trampling. Instead we suggest that grazing may be inducing indirect soil desiccation by consumption of, or damage to, facilitating (i.e. nurse) plants (Raffaele & Veblen 2001), which in turn induces higher seedling mortality in non-watered plots located in the more open cattle-affected understoreys. Facilitation effects of understorey plants on tree seedling survival in treefall gaps in nearby xeric N. pumilio forests has previously been experimentally demonstrated (Heinemann & Kitzberger 2006). Thus, cattle grazing in postfire subalpine forests may prevent tree regeneration both directly by browsing and trampling and also indirectly through removal of potential nurse plants and their generally desiccating influences on a site. In contrast to the inhibitory influences of livestock on the regeneration of N. pumilio, seedling survival of N. dombeyi and A. chilensis tended to be higher in unfenced sites. Presence of livestock in the N. dombeyi forest type increased seedling survival in watered treatments. Either the potential reduction in shading from understorey plants or exposure of bare mineral soil by cattle may benefit initial tree seedling survival. Similarly, Veblen et al. (1989) showed more abundant N. dombeyi seedling establishment in deer-impacted areas due to a reduced competition from shrubs such as Aristotelia chilensis or Maytenus boaria. However, despite the initial increase in abundance of small seedlings, continued browsing impeded the development of N. dombeyi into sapling and tree size classes. In our system, given high browsing differences between fenced and unfenced plots (Blackhall 2006; E. Raffaele et al. unpublished

778 N. Tercero-Bucardo et al.

data), it is likely that abundant seedlings facilitated by cattle will in the long term be highly browsed by cattle, thus impeding appropriate regeneration. However, we also note that the results of this study are dependent on the level of livestock impacts at the studied sites. In the N. pumilio forest type there was a substantially greater presence of livestock than in the N. dombeyi and A. chilensis forest type, making it difficult to compare herbivore impacts between the high- and low-elevation sites. A warming of 2–3 °C in northern Patagonia has been predicted by five coupled GCMs (Global Climate Models) for the mid twenty-first century (IPCC 1996). There is also a general agreement among IPCC-AR4 model predictions that annual precipitation will decrease from the late twentieth to the second half of the twenty-first century in north-western Patagonia (Vera et al. 2006). Given that our experiment did not reduce water inputs, it is not clear to what extent warming could drive, through increased evapotranspiration, changes in plant communities under these drier scenarios. What is clear is that reducing precipitation inputs below present levels will prevent adequate post-fire tree regeneration. Furthermore, increases in temperature have been shown to increase fire probability (Veblen et al. 1999; Kitzberger & Veblen 2003; Villalba et al. 2005). Eventually, super-imposed warming will further accentuate this effect so that regeneration will only be possible if climatic variability produces sporadic precipitation events that overcome the increased evapotranspiration induced by warming. Disturbances such as fire are likely to play major roles in determining future landscape patterns and trajectories (Franklin et al. 1992; Gardner et al. 1996; Payette et al. 2000). Climatic variation and changes in land uses such as livestock are the major drivers of global environmental change (Sala et al. 2000), and livestock are well recognized as a major determinant of post-fire vegetation trajectories (Fuhlendorf & Engle 2004; Vandvik et al. 2005). Here we provide experimental evidence of how the effects of altered climate and livestock may influence post-fire tree regeneration and determine post-fire vegetation trajectories. Although the results of the present study are limited to the effects of climatic variation on vegetation trajectories following severe fires, the likelihood of increased wildfires in conjunction with warming trends suggests a synergistic influence of climatic variation through increased severe fire events as well as decreased post-fire tree regeneration.

Acknowledgements © 2007 The Authors Journal compilation © 2007 British Ecological Society, Journal of Ecology, 95, 771–779

Research was funded by the National Science Foundation of the USA (Award No. 0117366) and the Agencia Nacional de Promoción Científica y Tecnológica of Argentina (Award PICT 97-01-2268). N.T.-B. is a CONICET doctoral fellow and T.K. and E.R. are researchers for CONICET.

References Arseneault, D. & Payette, S. (1992) A postfire shift from lichen-spruce to lichen-tundra vegetation at tree line. Ecology, 73, 1067–1081. Asselin, H. & Payette, S. (2005) Late Holocene opening of the forest tundra landscape in northern Quebec, Canada. Global Ecology and Biogeography, 14, 307–313. Barros, V., Cordón, V., Moyano, C., Méndez, R., Forquera, J. & Pizzio, O. (1983) Cartas de precipitación de la zona oeste de las Provincias de Río Negro y Neuquén. Report UNCCONICET, Facultad de Ciencias Agrarias, Universidad Nacional del Comahue, Cinco Saltos, Río Negro, Argentina. Blackhall, M. (2006) Efecto del ganado sobre la regeneración temprana post-fuego de un bosque mixto de Nothofagus dombeyi y Austrocedrus chilensis. Degree thesis, CRUB, Universidad Nacional del Comahue, Bariloche, Río Negro, Argentina. Brubaker, L.B. (1986) Responses of tree populations to climatic change. Vegetatio, 67, 119–130. Daniels, L.D. & Veblen, T.T. (2004) Spatiotemporal influences of climate on altitudinal treeline in northern Patagonia. Ecology, 85, 1284–1296. Davis, M.B. (1986) Climatic instability, time lags, and community disequilibrium. Community Ecology (eds J. Diamond & T.J. Case), pp. 269–284. Harper & Row, New York. De Fina, A.L. (1972) El clima de la región de los bosques Andino-Patagónicos. La Región de Los Bosques AndinoPatagónicos, Sinopsis General (ed. M. Dimitri), pp. 35–58. Instituto Nacional de Tecnología Agropecuaria, Buenos Aires, Argentina. Dunwiddie, P.W. (1986) A 6000-years record of forest history on Mount Rainier, Washington. Ecology, 67, 58–68. Franklin, J.F., Swanson, F.J., Harmon, M., Perry, D.A., Spies, T.A., Dale, V.H., McKee, A., Ferrell, W.K., Gregory, S.V., Lattin, J.D., Schowalter, T.D. & Larsen, D. (1992) Effects of global climatic change on forests in northwestern North America. Global Warming and Biology Diversity (eds R.L. Peters & T.E. Lovejoy), pp. 244 –257. Yale University Press, New Haven. Fuhlendorf, S.D. & Engle, D.M. (2004) Application of the fire–grazing interaction to restore a shifting mosaic on tallgrass prairie. Journal of Applied Ecology, 41, 604–614. Gallopín, G.G. (1978) Estudio ecológico integrado de la cuenca del río manso superior (Río Negro, Argentina). I. Descripción general de la cuenca. Anales Parques Nacionales (Argentina), XIV, 161–230. Gardner, R.H., Hargrove, W.W., Turner, M.G. & Romme, W.H. (1996) Climate change, disturbances and landscape dynamics. Global Change and Terrestrial Ecosystems (eds B. Walker & W. Steffen), pp. 149–172. Cambridge University Press, Cambridge, UK. Gowda, J. & Raffaele, E. (2004) Spine production is induced by fire: a natural experiment with three Berberis species. Acta Oecologica, 26, 239–245. Heinemann, K. & Kitzberger, T. (2006) Effects of position, understorey vegetation and coarse woody debris on tree regeneration in two environmentally contrasting forests of north-western Patagonia: a manipulative approach. Journal of Biogeography, 33, 1357–1367. Heinemann, K., Kitzberger, T. & Veblen, T.T. (2000) Influences of gap microheterogeneity on the regeneration of Nothofagus pumilio in a xeric old-growth forest of northwestern Patagonia, Argentina. Canadian Journal of Forest Research, 30, 25 –31. IPCC (1996) Climate Change 1995. The Science of Climate Change Contribution of Working Group I to the Second Assessment Report of the Intergovernmental Panel on Climate Change (eds J.J. Houghton, L.G. Meiro Filho, B.A. Callander, N. Harris, A. Kattenberg & A.K. Maskell). Cambridge University Press, Cambridge.

779 Climatic and herbivore impacts on tree regeneration

© 2007 The Authors Journal compilation © 2007 British Ecological Society, Journal of Ecology, 95, 771–779

Keane, R.E., Morgan, P. & White, J.D. (1999) Temporal patterns of ecosystem processes on simulated landscapes in Glacier National Park, Montana, USA. Landscape Ecology, 14, 311–329. Kitzberger, T., Raffaele, E., Heinemann, K. & Mazzarino, M.J. (2005a) Effects of fire severity in a North Patagonian subalpine forest. Journal of Vegetation Science, 16, 5–12. Kitzberger, T., Raffaele, E. & Veblen, T.T. (2005b) Variable community responses to hervivory in fire-altered landscapes of northern Patagonia, Argentina. African Journal of Range and Forage Science, 22, 85 –91. Kitzberger, T., Steinaker, D.F. & Veblen, T.T. (2000) Establishment of Austrocedrus chilensis in Patagonian forest-steppe ecotones: facilitation and climatic variability. Ecology, 81, 914 –1924. Kitzberger, T. & Veblen, T.T. (2003) Influences of climate on fire in Northern Patagonia, Argentina. Fire and Climatic Change in Temperate Ecosystems of the Western Americas (eds T.T. Veblen, W.L. Baker, G. Montenegro & T.W. Swetnam), pp. 296–321. Springer, New York. Landhäusser, S.M. & Wein, R.W. (1993) Post-fire vegetation recovery and tree establishment at the Arctic treeline: climatechange vegetation-response hypotheses. Journal of Ecology, 81, 665–672. Lenihan, J., Daly, C., Bachelet, D. & Neilson, R.P. (1998) Simulating broad-scale fire severity in a dynamic global vegetation model. Northwest Science, 72, 91–103. Littell, R.C., Stroup, W.W. & Freund, R.J. (2002) SAS for Linear Models. 4th edn. SAS Institute Inc, Cary, NC. Mermoz, M., Kitzberger, T. & Veblen, T.T. (2005) Landscape influences on occurrence and spread of wildfires in Patagonian forest and shrublands. Ecology, 86, 2705–2715. Overpeck, J.T., Rind, D. & Goldberg, R. (1990) Climate induced changes in forest disturbance and vegetation. Nature, 343, 51–53. Payette, S., Bhiry, N., Delwaide, A. & Simard, M. (2000) Origin of the lichen woodland at its southern range limit in eastern Canada: the catastrophic impact of insect defoliators and fire on the spruce-moss forest. Canadian Journal of Forest Research, 30, 288–305. Peters, D.C., Bestelmeyer, B.T., Herrick, J.E., Fredrickson, E.L., Monger, H.C. & Havstad, K.M. (2006) Disentangling complex landscapes: new insights into arid and semiarid system dynamics. Bioscience, 56, 491–501. Potvin, C. (1993) ANOVA: experiment in controlled environments. Design and Analysis of Ecological Experiments (eds S.M. Scheiner & J. Gurevitch), pp. 46 – 68. Chapman & Hall, New York. Raffaele, E. & Veblen, T.T. (2001) Effects of cattle grazing on early postfire regeneration of matorral in northwest Patagonia, Argentina. Natural Areas Journal, 21, 243 –249. Rich, P.M. (1990) Characterizing plant canopies with hemispherical photographs. Remote Sensing Reviews, 5, 13 –29.

Sala, O.E., Chapin, F.S.R., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., Huber-Sanwald, E., Huenneke, L.F., Jackson, R.B., Kinzig, A., Leemans, R., Lodge, D.M., Mooney, H.A., Oesterheld, M., Poff, N.L., Sykes, M.T., Walker, B.H., Walker, M. & Wall, D.H. (2000) Global biodiversity scenarios for the year 2100. Science, 287, 1770–1774. Schmoldt, D.L., Peterson, D.L., Keane, R.E., Lenihan, J.M., McKenzie, D., Weise, D.R. & Sandberg, D.V. (1999) Assessing the Effects of Fire Disturbance on Ecosystems: a Scientific Agenda for Research and Management. Report no. PNW-GTR-455. U.S. Department of Agriculture Forest Service Pacific Northwest Research Station, Portland, OR. Sokal, R.R. & Rohlf, F.J. (1981) Biometry The Principles and Practice of Statistics in Biological Research, 2nd edn. Freeman, New York. Vandvik, V., Heegaard, E., Maren, I.E. & Aarrestad, P.A. (2005) Managing heterogeneity: the importance of grazing and environmental variation on post-fire succession in heathlands. Journal of Applied Ecology, 42, 139–149. Veblen, T.T., Kitzberger, T. & Lara, A. (1992) Disturbance and forest dynamics along a transect from Andean rain forest to Patagonian shrubland. Journal of Vegetation Science, 3, 507–520. Veblen, T.T., Kitzberger, T., Villalba, R. & Donnegan, J. (1999) Fire history in northern Patagonia: the roles of humans and climatic variation. Ecological Monographs, 69, 47–67. Veblen, T.T., Mermoz, M., Martin, C. & Ramilo, E. (1989) Effects of exotic deer on forest regeneration and composition in northern Patagonia. Journal of Applied Ecology, 26, 711– 724. Vera, C., Silvestri, G., Liebmann, B. & Gonzalez, P. (2006) Climate change scenarios for seasonal precipitation in South America. Geophysical Research Letters, 33. doi: 10.1029/2006GL025759. Villalba, R., Boninsengna, J.A., Veblen, T.T., Schmelter, A. & Rubulis, S. (1997) Recent trends in tree-ring records from high elevation sites in the Andes of northern Patagonia. Climatic Change, 36, 425 – 454. Villalba, R., Masiokas, M., Kitzberger, T. & Boninsegna, J.A. (2005) Biogeographical consequences of recent climate changes in the southern Andes of Argentina. Global Change and Mountain Regions: an Overview of Current Knowledge (eds U.M. Huber, K.M.H. Bugmann & M.A. Reasoner), pp. 157–168. Series: Advances in Global Change Research, Vol. 23. Springer, New York. Villalba, R. & Veblen, T.T. (1997) Regional patterns of tree population age structure in northern Patagonia: climatic and disturbance influences. Journal of Ecology, 85, 113–124. Villalba, R. & Veblen, T.T. (1998) Influences of large-scale climatic variability on episodic tree mortality in Northern Patagonia. Ecology, 79, 2624 –2640. Received 5 December 2006; revision accepted 27 March 2007 Editor: John lee