3002

Recruitment patterns following a severe drought: long-term compositional shifts in Patagonian forests Marı´a L. Suarez and Thomas Kitzberger

Abstract: Severe droughts have the potential of inducing transient shifts in forest canopy composition by altering speciesspecific adult tree mortality patterns. However, permanent vegetation change will occur only if tree recruitment patterns are also affected. Here, we analyze how a massive mortality event triggered by the 1998–1999 drought affected adult and sapling mortality and recruitment in a mixed Nothofagus dombeyi (Mirb.) Blume – Austrocedrus chilensis (D. Don) Flor. et Boult. forests of northern Patagonia. Comparing drought-induced and tree-fall gaps, we assessed changes in forest composition, microenvironments, and seedling density and survival of both species. Drought-kill disturbance shifted species composition of both canopy and sapling cohorts in favour of A. chilensis. Drought gaps were characterized by a shadier and more xeric environment, affecting the recruitment pattern of N. dombeyi seedlings. The seedling cohort was composed mostly of A. chilensis, and its survival was always higher than that of N. dombeyi. Additionally, A. chilensis seedlings showed higher plasticity than N. dombeyi seedlings, increasing its root to shoot ratios in drought gaps. The results suggest that extreme drought itself is a strong driving force in forest dynamics, with important imprints on forest landscapes. Future climate-change scenarios, projecting an increased in frequency and severity of droughts, alert us about expected longterm compositional shifts in many forest ecosystems. Re´sume´ : Les se´cheresses se´ve`res sont capables de provoquer des changements passagers dans la composition du couvert forestier en modifiant le patron de mortalite´ des arbres adultes propre a` chaque espe`ce. Cependant, un changement permanent de la ve´ge´tation se produira seulement si les patrons de recrutement des arbres sont aussi affecte´s. Dans cet article, nous analysons comment un e´pisode de forte mortalite´ de´clenche´ par la se´cheresse de 1998–1999 a affecte´ la mortalite´ des arbres matures et des gaules ainsi que le recrutement dans des foreˆts me´lange´es de Nothofagus dombeyi (Mirb.) Blume et d’Austrocedrus chilensis (D. Don) Flor. et Boult. dans le nord de la Patagonie. En comparant les troue´es cause´es par la se´cheresse d’une part et la chute des arbres d’autre part, nous avons e´value´ les changements dans la composition de la foreˆt, les microenvironnements et la densite´ des semis ainsi que la survie des deux espe`ces. La perturbation cause´e par la se´cheresse et la mort des arbres a modifie´ la composition en espe`ces de la canope´e et des cohortes de gaules en faveur de d’A. chilensis. Les troue´es cause´es par la se´cheresse e´taient caracte´rise´es par un milieu plus ombrage´ et plus xe´rique, ce qui a affecte´ le recrutement des semis de N. dombeyi. La cohorte de semis e´tait surtout compose´e de semis d’A. chilensis et la survie de cette essence e´tait toujours plus e´leve´e que celle de N. dombeyi. De plus, les semis d’A. chilensis ont de´montre´ une plus grande plasticite´ que ceux de N. dombeyi en augmentant leur rapport racines:tige dans les troue´es cause´es par la se´cheresse. Les re´sultats indiquent qu’une se´cheresse se´ve`re constitue par elle-meˆme une puissante force de changement dans la dynamique forestie`re et qu’elle laisse des marques importantes sur les paysages forestiers. Les sce´narios de changements climatiques futurs, qui pre´voient une augmentation de la fre´quence et de la se´ve´rite´ des se´cheresses, nous sensibilisent aux changements a` long terme auxquels on devrait s’attendre dans la composition de plusieurs e´cosyste`mes forestiers. [Traduit par la Re´daction]

Introduction Increases in drought frequency and severity are becoming of particular interest in the context of global change because drought mortality has the capacity to cause rapid and pronounced vegetation shifts in dry woodlands (Allen and BreReceived 7 March 2008. Accepted 14 August 2008. Published on the NRC Research Press Web site at cjfr.nrc.ca on 15 November 2008. M.L. Suarez1 and T. Kitzberger. Laboratorio Ecotono, Centro Regional Universitario Bariloche, Universidad Nacional del Comahue INIBIOMA-CONICET, Quintral 1250, 8400, Bariloche, Argentina. 1Corresponding

author (e-mail: [email protected]).

Can. J. For. Res. 38: 3002–3010 (2008)

shears 1998; Swetnam and Betancourt 1998; Hanson and Weltzin 2000; Mueller et al. 2005) and mesic forests (Hursh and Haasis 1931; Elliott and Swank 1994; Williamson et al. 2000; Condit et al. 2004). Numerous studies have documented how moderate seasonal droughts negatively influence tree seedling performance (Fotelli et al. 2001; Gilbert et al. 2001); however, whenever drought is severe enough to kill adult trees, large compositional changes can occur by at least two mechanisms: (i) direct shifts in canopy dominance owing to differential species-specific susceptibility to drought and (ii) postdrought biotic and abiotic changes affecting tree seedling recruitment performance (Clinton et al. 1993) as well as other ecosystem properties affecting forest dynamics (Breshears et al. 2005). Regeneration processes involved after dieoff and cessation

doi:10.1139/X08-149

#

2008 NRC Canada

Suarez and Kitzberger

of drought are key in determining whether increased climate-induced mortality will lead to a relatively permanent vegetation change spanning the following decades or if it simply mimics other natural mortality causes and just increases gap turnover rates. In other words, mortality per se may not be a sufficient condition for ensuring long-term vegetation shifts. In addition, the effect of drought on finescale environmental heterogeneity could be another driving force for subsequent vegetation development, but only if mortality has reduced propagule input or has changed microenvironments in such way that regeneration of the dieoff species is not successful or if new conditions are more favourable for competitors of the receding species. Despite the disproportionate emphasis that has been given to susceptibility and immediate responses of forests to drought, few studies are available on the consequences of postdrought scenarios for further community dynamics. Postdrought forest conditions may differ substantially from those of typical natural tree falls, and forest dynamics related to each type of disturbance might be quite different too. Particularly, drought-induced mortality gaps differ from tree-fall gaps in several aspects that can affect subsequent tree regeneration patterns (Clinton et al. 1993). First, the mode of tree mortality (standing dead trees) implies little or no damage to the understory; thus, microhabitats based on soil alteration are not produced. The forest floor remains intact, acting as a physical barrier to seedling establishment, particularly for species that require mineral soils for seed germination. Second, lack of coarse woody debris may disfavour species whose reproductive strategies include ‘‘nurse logs’’ (Clinton et al. 1993). Moreover, in drought gaps, eventual branch fall is the main source of mechanical damage and coarse woody debris, and in contrast with tree falls, understory saplings may be favoured over new seedling establishment by branch fall. Third, gap–understory interactions could also influence the nature of the resource limitation on seedling growth and survival (Heinemann et al. 2000). Diffuse shade from dead-standing crowns may relatively benefit species using temporary sunflecks at the expense of species using direct radiation such as preexisting understory. In the absence of soil disturbance, preexisting herbs and shrubs may increase in response to that condition and inhibit seedling establishment through competition (Davis et al. 1998, 1999; Valladares and Pearcy 2002). In this paper, we will focus on how subsequent drought gap environments differ from tree-fall gaps because gap environments are directly related to seedling establishments, and their future development could influence the way of successional pathways inside gaps. It has been well documented that fine-scale disturbance, like tree-fall gaps, maintain many broadleaf temperate forests in compositional equilibrium (Veblen 1989; Rebertus and Veblen 1993; Lertzman 1995). In contrast, the role of drought gaps on long-term forest composition is less clear in most forest ecosystems. A dynamical integration of mortality processes triggered during drought and establishment opportunities after drought cessation is necessary to fully understand the role of these increasingly important forest disturbances. Here, we analyze how a massive mortality event triggered by an extreme drought occurred in 1998– 1999 in relation to a strong La Nin˜a event affected adult

3003

and sapling mortality and seedling recruitment of mixed Nothofagus dombeyi (Mirb.) Blume – Austrocedrus chilensis (D. Don) Flor. et Boult. forests of northern Patagonia. During the period March 1998 – February 1999, rainfall in Bariloche was 2.6 standard deviations below the historical mean, producing the most severe drought of the 20th century (Servicio Meteorolo´gico Nacional). In addition, unusual midsummer heat conditions prevailed with mean daily temperatures during January 1999 7 8C above the long-term mean and with maximum daily temperatures >28 8C over a continuous 19 day period. Under these conditions, 10% of canopy trees in Nahuel Huapi National Park were killed over an area of 11 000 ha of forest (*10% of the N. dombeyi forests) and 25% of trees were killed over 680 ha (Bran et al. 2001). Although drought affected a large area of temperate forests, tree mortality was extremely patchy, creating a high number of canopy gaps. This disturbance event was a unique opportunity to study adult and advance regeneration mortality and seedling regeneration dynamics associated with drought-induced gaps. We propose that drought events may cause forest to deviate from gap compositional equilibrium by differential adult and sapling mortality among woody species and by differential species recruitment owing to physical and biotic changes associated with drought gap compared with tree-fall gaps. Related to these, we predict that more drought- and shade-tolerant species (A. chilensis) may be relatively benefited by immediate canopy and juvenile mortality effects of drought on N. dombeyi. In addition, saplings of the more shade-tolerant species may be relatively benefited owing to the understory filter effect.

Materials and methods Study area and sampling The study was conducted in Nahuel Huapi National Park, northern Patagonia, Argentina. Precipitation in Nahuel Huapi National Park is highly seasonal with approximately 60% falling during May–August. At this latitude, mean precipitation decreases abruptly from *3000 mm
year–1 at the main Andean cordillera to <500 mm
year–1 over an approximately 80 km west to east distance (De Fina 1972). Layers of volcanic ash covered the glacial topography, and soil throughout the region is derived from these parent materials (Andosol type soils). Along this strong precipitation gradient, vegetation changes rather abruptly from temperate rain forests, dominated by tall the evergreen N. dombeyi, to semiarid steppe dominated by cushion shrubs and bunchgrasses. At precipitation levels of *1500 mm
year–1, following stand-devastating fires, N. dombeyi and A. chilensis both form even-aged populations; N. dombeyi typically occurs exclusively in the main canopy, whereas A. chilensis is partially suppressed in the subcanopy (Veblen and Lorenz 1987). We conducted this research throughout the eastern distributional limit of N. dombeyi where tree mortality as a consequence of the 1998–1999 drought was severe enough to produce massive mortality. In early spring (September) 2003, along the 1400 mm
year–1 isohyet, we randomly selected 16 gaps (eight tree-fall gaps and eight gaps produced by the 1998–1999 drought) within N. dombeyi – A. chilensis #

2008 NRC Canada

3004

mixed forests. All gaps were separated by at least 1 km and encompassed northwest, southwest, northeast, and north orientations and slopes between 28 and 208. Tree-fall gaps were sampled as controls of drought gaps because the shadeintolerant nature of tree species prevents seedlings from establishing under closed forest. To avoid the effect of treefall age and area in the analysis of survival, we selected recent tree-fall gaps (presence of bark on the gap maker) with an area of a. 250 m2. Drought-induced gaps included up to five dead trees in a 15 m  15 m plot resulting in openings of 225 m2. Additionally, we extensively sampled throughout the study area to evaluate tree regeneration status 5 years after the drought event and also the drought effect on adult composition. Drought effects on saplings and adults To assess density and mortality of predrought (saplings) and postdrought (seedlings) natural regeneration, forty-five 15 m2 plots were randomly placed throughout the study area and encompassing drought-affected forests. In each plot, natural regeneration of seedlings was determined by counting individuals (height <0.10 m) within three 1 m2 quadrats. Live and desiccated saplings (stems <0.05 m diameter at breast height and <2 m height) were counted along a 2 m wide by 5 m long transect in the center of the plot. The same density data were taken under undisturbed forest canopy (22 plots) as in the drought-disturbed canopy gaps (23 plots). Regeneration abundance beneath undisturbed forest canopy represents a good control because of the different light behaviours of both species. Additionally, to assess drought impact on adult composition, thirty-four 400 m2 plots were placed throughout the area most affected by the 1998–1999 drought. Fourteen plots were placed under canopy with <25% dead trees (low-mortality plots) and 20 plots were placed under canopy with >50% dead trees (high-mortality plots). In each plot, all N. dombeyi and A. chilensis adult trees were registered as dead or alive and measured at breast height. All dead trees corresponded to trees that were killed by the 1998–1999 drought. Drought versus tree-fall gap microclimate Topographic characteristics of each of the 16 canopy gaps (aspect, elevation, and slope) were measured. In tree falls, gap area was estimated as an ellipse based on the two longest perpendicular axes in the gap (Runkle 1981). Understory light environment was quantified using hemispherical photographs taken at 0.3 m heights and throughout north– south directions inside gaps. Photographs were taken using a leveled north-oriented fisheye lens and were analyzed using WinScanopy (version 2003b,c) (Regent Instruments Inc.). We calculated direct, indirect, and total site factors to calculate the amount and quality of light reaching the forest floor during the growing season (September–March in the Southern Hemisphere). Inside each gap, we measured percentage of water content using a hand-held three-rod ML2x Theta Probe (Delta-T Devices). Measurements were made twice at nine positions inside gaps: north (three), center (three), and south (three). All measurements were made during February when moisture deficits are more intense. Drought versus tree-fall gap natural regeneration Survival of naturally established seedlings was monitored

Can. J. For. Res. Vol. 38, 2008

during two growing seasons (2003–2004 and 2004–2005) only in eight of the 16 gaps. In midspring (October) 2003, gaps selected were divided into three zones of similar area following a north–south direction and gap positions were located. In each zone, three 1 m2 circular plots were located and all N. dombeyi and A. chilensis seedlings (<0.10 m tall) tagged. For tree-fall gaps, monitoring was carried out inside the gap effective area (sensu Runkle 1981) and no plots fell near the edge of the gap. Marked seedlings were revisited in November 2003, March 2004, November 2004, and March 2005. During each census, each seedling was recorded as alive or dead. Drought versus tree-fall gap seedling survival experiments Survival monitoring was supported by a manipulative experiment carried out in all selected gaps. Fifteen A. chilensis seedlings per pot were obtained in a greenhouse by germinating 40 seeds per pot after cold–wet stratified for 45 days. Nothofagus dombeyi seedlings are difficult to obtain in a greenhouse owing to low viability. Thus, naturally germinated seedlings were used to conduct the experiment. Seedlings belonging to the cohort of spring 2003 with presence of both cotyledons but without true leaves were collected from the forest floor from sites near the experimental area. Fifteen seedlings were transferred into each pot and remained well watered in the greenhouse from late September to the beginning of October. In the field, in each gap, three within-gap positions (north, center, and south) following the transect at the center of the gap were identified. In early November 2003, two pots (30 seedlings) per species were transplanted in each position. A small amount of mortality occurred in the first week and was replaced with extra seedlings to reach 30 initial seedlings per species. Seedlings were monitored in each plot at the beginning and end of two growing seasons (November 2003 – March 2004 and November 2004 – March 2005). Seedlings that were alive at the end of the second growing season were considered as established seedlings. At the end of the second growing season, all seedlings growing in each plot were harvested and basal diameter, height, and numbers of leaves were measured. Biomass samples were divided into leaves, stems, and roots and oven-dried at 50 8C to constant mass and weighed. Total biomass, root to shoot ratios, and leaf allocation (fraction of aboveground production allocated to leaves) were then calculated. Data analyses Regeneration density of seedlings, saplings, and adults of both species sampled was analyzed using relative abundance and compared using Mann–Whitney’s U test. Differences in light levels (direct, indirect, and total site factors) and topsoil water content were analyzed using split-plot ANOVA with gap type as the main-plot factor and position as the split-plot factor. To avoid pseudoreplication, samples from each subplot within each gap were averaged prior to conducting the ANOVA. To analyze the effect of postdrought scenarios on seedling survival, gap type (GAP), within-gap position (POS), and species (SP) were used as fixed factors in a split-split-plot ANOVA. The main plot was gap type and the two splits were gap position and species. As well, #

2008 NRC Canada

Suarez and Kitzberger

to analyze seedling survival after the effects of light are removed, an additionally split-split-plot ANCOVA was used. The lack of seedlings of both species in several subplots prevented evaluation of the effect of within-gap position. The response variable was the proportion of seedlings alive at the end of the second growing season (arcsine square-root transformed). Because final survival could have been affected by the initial abundance of seedling in each subplot, we used the number of initial seedlings as covariable. The final analysis was a split-plot ANCOVA (F[1,3] = 21.68, P = 0.02). Mean final seedling biomass, mean seedling leaf allocation, and mean seedling root to shoot ratio were the response variables in the analysis of drought gap on seedling performance. We used a split-split-plot ANOVA with the same fixed factors used in the survival analysis. To reach normality, response variables were log-transformed. All results were considered significant if the probability of a Type I error was <0.05. Multiple comparisons of treatment means were conducted using a post hoc test with Bonferroni corrections. We used R software (version 2.4.1, 2006) (R Development Core Team) in all analyses.

Results Drought-induced mortality of adult and saplings Drought caused marked differences in mortality of N. dombeyi and A. chilensis adult trees. In high-mortality plots, drought caused an abrupt reduction in N. dombeyi density (*57%), while in low-mortality plots, drought killed only 11% of the trees (Fig. 1a). In contrast, A. chilensis evidenced a low impact of drought by reducing its density in high-mortality plots by only 5%. Additionally, no A. chilensis trees were killed by drought in low-mortality plots (Fig. 1a). Thus, density of live N. dombeyi trees was more than halved by the drought in highly affected plots and composition of adults shifted from a N. dombeyi dominated forest (*65%) to a nearly equally represented N. dombeyi – A. chilensis forest. Similarly, sapling mortality was markedly higher in N. dombeyi than in A. chilensis (Fig. 1b). In high-mortality plots, almost 30% of N. dombeyi saplings were killed by drought, and little mortality was also evidenced in low-mortality plots (7%). However, in all plots, A. chilensis saplings remained mostly alive (Fig. 1b). This differential mortality raised A. chilensis dominance in the sapling cohort by at least 10% (from 50% to 60% being A. chilensis saplings). Drought versus tree-fall gaps Microenvironments Drought and tree-fall gaps were equally distributed among mountain aspects (c2 = 2.39, df 7, P < 0.93) and slopes (c 2 = 1.58, df 3, P < 0.66). Thus, environmental differences between gap types were largely attributed to gap origin. Direct radiation reaching the ground was overall lower in drought gaps than in tree-fall gaps (F[1,14] = 57.23, P << 0.0001) (Table 1). In contrast, spatial variation in radiation levels across north–south transects located within gaps was homogeneous (POS: F[2,23] = 0.72, P = 0.49) and similar between gap types (POS  GAP: F[2,23] = 1.15, P = 0.33) (Table 1).

3005

Although drought gaps were characterized as shady environments, there was no evidence of more water content in the soil. On the contrary, these environments were more xeric than tree-fall ones. Late-summer (February) volumetric topsoil water content in drought gaps was half that in treefall gaps (3.9 ± 0.7%, drought gap = 1.4 ± 0.5%, F[1,7] = 5.25, P = 0.05) (Table 1). Again, in contrast with the pattern found in canopy gaps, we could not find pronounced north– south changes in topsoil water content either within gaps (POS: F[2,14] = 0.74, P = 0.49) or between gap types (POS  GAP: F[2,14] = 1.29, P = 0.30) (Table 1). Tree seedling recruitment, survival, and performance By summer 2004, 5 years following the drought, N. dombeyi density of naturally established seedlings was fivefold higher in the undisturbed forest (64 700 individuals
ha–1) compared with the drought gaps (12 100 individuals
ha–1) (Mann–Whitney U test, P < 0.03). In contrast, A. chilensis seedlings were as abundant beneath undisturbed canopy (45 000 individuals
ha–1) as in drought gaps (63 800 individuals
ha–1). Thus, within drought gaps, the seedling bank was composed mainly of A. chilensis (84%) compared with a more evenly composed seedling bank in the undisturbed forest understory (41% A. chilensis). Survival of naturally established A. chilensis seedlings was overall higher than that of N. dombeyi seedlings independently in both drought and tree-fall gaps (SP: F[1,3] = 171.34, P = 0.0009). After 2 years, only 29% of initial N. dombeyi seedlings survived compared with a 62% A. chilensis survival (Fig. 2). However, interspecific differences in survival depended on gap type (GAP  SP: F[1,3] = 6.85, P = 0.08) with a higher difference in favour of A. chilensis seedlings in drought gaps (P = 0.01). This pattern was more apparent during the second growing season (Fig. 2). Survival of transplanted seedlings at the end of the two growing seasons was highly species dependent and was affected by gap type and within-gap heterogeneity (Table 2; Fig. 3). Similarly to naturally regenerated seedlings, transplanted A. chilensis showed overall higher survivorship than N. dombeyi seedlings (SP: P < 0.001). Seedlings tended to survive better in gap centers compared with gap edges irrespective of species (POS: P < 0.002). While both species tended to use tree-fall gap centers, only N. dombeyi behaved selectively within drought gaps, preferentially using gap centers. In contrast, A. chilensis showed high survival across the entire drought gap (Fig. 3). The effect of gap type on seedling survival was not evident until the influence of differential light levels on survival was removed (GAP: P < 0.04) (Table 2). When doing so, only A. chilensis increased its survival in drought gaps (GAP: F[1,14] = 4.54, P = 0.05), whereas N. dombeyi survival was independent of gap type. In general, and related to species inherent characteristics, N. dombeyi showed higher final individual seedling dry biomass than A. chilensis (SP: F[1,34] = 4.20, P = 0.05) but seedling biomass was unaffected by gap type. In addition, we found no negative relationship of seedling total dry biomass as a function of number of surviving seedlings in a subplot. Despite any effect on total biomass, gap type differentially modified root to shoot allocation pattern among both tree seedling species. Austrocedrus chilensis seedlings showed a marked plasticity, increasing root allocation from #

2008 NRC Canada

3006

Can. J. For. Res. Vol. 38, 2008

Fig. 1. Drought effects on Nothofagus dombeyi and Austrocedrus chilensis (a) adult and (b) sapling density. Predrought density (solid bars) is based on surviving plus drought-dead individuals. Postdrought density (open bars) is based on surviving individuals. Low-mortality plots and undisturbed canopy correspond to sites with low or even no drought mortality. High-mortality plots and drought-dead canopy correspond to sites where drought caused severe mortality.

50% in tree-fall gaps to *60% in drought gaps (GAP: F[1,14] = 3.43, P = 0.08) (Fig. 4a). In contrast, N. dombeyi seedlings showed more constant root to shoot ratios in both gap types. While N. dombeyi tended to increase proportional allocation to leaves in drought gaps, A. chilensis showed reduced allocation to leaves (GAP  SP: F[1,32] = 4.08, P < 0.05) (Fig. 4b).

Discussion The results of our study show that the severe drought that affected northern Patagonia in 1999 had immediate and subsequent effects on N. dombeyi – A. chilensis forest composition. First, high adult and sapling mortality of N. dombeyi caused an immediate reduction in its dominance in favour of the drought-tolerant A. chilenisis. Second, lower recruitment and survival rates of N. dombeyi, possible because of physical microenvironment changes and biotic filters impose by understory release response in drought-induced gaps, have subsequent effects favouring present and future recruitment of the more drought- and shade-tolerant species. Mortality ratios were highly different between dominant

tree species. Adult trees of N. dombeyi were far more affected by drought than those of A. chilensis, possibly owing to their shallow root system and lower water use efficiency in relation to A. chlensis (Kitzberger 1994). This occurred in sharp contrast with other dominant disturbances affecting these forests such as fire and tree-fall gaps (Veblen et al. 1996). Stand-devastating fires affect both species owing to high fire sensitivity so that propagule availability and similar regeneration requirements produce mixed postfire cohorts (Veblen and Lorenz 1987). In the case of tree-fall gaps, Veblen (1989) has shown that the relative proportions of tree deaths and potential successors (saplings) of these two species in the gaps imply that gap-phase regeneration is maintaining this stand in compositional equilibrium. In addition to shifts in canopy tree composition, the drought also reduced the density of potential successors of N. dombeyi under death canopy, favouring A. chilensis successors. However, it could be possible that even lower sapling density may assure gap filling owing to higher growth rates of N. dombeyi as found after fire and in tree-fall gaps (Veblen and Lorenz 1987; Veblen 1989). However, this would not be the case under drought-killed canopy because both species #

2008 NRC Canada

Suarez and Kitzberger

3007 Table 1. Tree-fall and drought gap microenvironment characteristics (mean ± SE) measured as water content (volumetric topsoil content in the first 15 cm) and light levels reaching the forest floor and throughout three gap positions: north, center, and south. Gap microenvironment Water content Tree-fall gap Drought gap Light level (direct site factor) Tree-fall gap Drought gap Light level (indirect site factor) Tree-fall gap Drought gap

North

Center

South

0.0323±0.006 0.006±0.002

0.0299±0.011 0.015±0.005

0.0211±0.008 0.007±0.004

0.21±0.017 0.11±0.014

0.27±0.030 0.13±0.019

0.23±0.022 0.12±0.016

0.20±0.013 0.12±0.011

0.21±0.056 0.11±0.014

0.19±0.010 0.12±0.013

Fig. 2. Mean seedling survival rates of monitored Nothofagus dombeyi and Austrocedrus chilensis natural regeneration in tree-fall gaps and in drought-induced gaps. November and March correspond to the beginning and end of the growing season, respectively. Solid triangles, N. dombeyi growing in tree-fall gaps; open triangles, A. chilensis growing in tree-fall gaps; solid circles, N. dombeyi growing in drought gaps; open circles, A. chilensis growing in drought gaps.

increased their growth in response to the opening in a similar fashion (Suarez et al. 2004). After drought, substantial changes were observed in forest understory microenvironments compared with tree-fall gaps, affecting tree seedling recruitment patterns. Drought gap microenvironments were substantially shadier and extremely xeric compared with tree-fall gaps, generating a new kind of environment in the forest. Deeper shades found in the drought gap floor could be attributed to a faster and more pronounced post-1999 growth response of preexisting shrubs or perennial herbs. For example, higher percentages of cover of Galium spp. and Maytenus chubutensis (Speg.) Lourteig & O’Donell & Sleumer were found in drought-induced gaps (M.L. Suarez, unpublished data). In addition, xeric conditions found in drought gaps could be related to intrinsic site conditions. A previous study preformed in the same area showed that steeper slopes associated with high rockiness and shallow soils predisposed the forest to suffer high mortality during the 1998–1999 event (Suarez et al. 2004). In

this summer-dry climate, soil moisture content of highly insolated steep slopes is closely related to topographic conditions, which could in turn determine xeric conditions prevailing inside drought-induced gaps. Our transplant experiments as well as monitoring of naturally established seedlings show that both tree species have different survival responses in relation to the two gap types studied, with drought gaps favouring the survival of A. chilensis and disfavouring that of N. dombeyi seedlings. Microenvironmental conditions (deep shades and drier soils) in this gap type and seedling performance are in accordance with current ecophysiological knowledge of the two species. Nothofagus dombeyi is considered a highly shade-intolerant species but a relatively moisture-demanding species that only is able to successfully establish in large insolated gaps but higher rainfall levels (Veblen et al. 1996). Austocedrus chilensis, in contrast, is considered to be relatively more shade tolerant, often using nurse plants for establishment (Kitzberger et al. 2000), and far more drought tolerant, extending its range into low-rainfall areas (Veblen et al. 1995). Furthermore, some of our results suggest that this different response may more likely be related to other gap conditions not directly related to differences in radiation between each gap type. Thus, xeric conditions inside drought gaps strongly disfavour N. dombeyi seedlings. These biased establishment rates between N. dombeyi and A. chilensis are in contrast with previous studies of small tree-fall gaps (Veblen 1989) that show that both species are able to coexist and mixed forest composition is maintained by a dynamic equilibrium based on higher seedling densities with lower survival rates of N. dombeyi counterbalanced with lower seedling densities with higher survival rates shown for A. chilensis. Seedling allocation patterns found could also help to explain differential survival rates found in drought gaps. While the response of A. chilensis to drought gaps was highly plastic, increasing root to shoot ratios, N. dombeyi seedlings did not show a plastic response in root to shoot allocation ratios. Furthermore, and in contrast with A. chilensis, these latter seedling responded plastically to shadier conditions of drought gaps, allocating more biomass to leave tissue, making the seedling overall more prone to desiccation. Apparently owing to intrinsic physiological constraints, N. dombeyi is largely unable to tolerate either low light levels or low water levels, whereas A. chilensis seedlings are #

2008 NRC Canada

3008

Can. J. For. Res. Vol. 38, 2008 Table 2. Split-split-plot ANOVA and ANCOVA on Nothofagus dombeyi and Austrocedrus chilensis seedling experimental survival. ANOVA Source Indirect site factor GAP Error: Rep (GAP) POS GAP  POS Error: POS  Rep (GAP) SP GAP  SP POS  SP GAP  POS  SP Error: SP  POS  Rep (GAP)

df

F .

1 14 2 2 27 1 1 2 2 41

. 0.67 . 7.96 0.67 . 12.14 0.42 0.009 0.98 .

ANCOVA P . 0.427 . 0.002 0.521 . 0.001 0.521 0.990 0.384 .

df 1 1 14 2 2 23 1 1 2 2 38

F 4.09 5.07 . 4.77 0.89 . 10.51 0.62 0.85 2.29 .

P 0.05 0.04 . 0.02 0.12 . 0.02 0.25 0.16 0.11 .

Note: Factors: GAP, gap type; POS, within-gap position; SP, species. Indirect site factor is the covariable. The error term used to test each factor is indicated below each main factor and their interactions (Rep is gap type replicate). Significant effects are shown in bold.

Fig. 3. Effect of gap type and position on seedling survival after two growing seasons (November 2003 – March 2005). Survivor data correspond to survival experiment. Solid bars, Nothofagus dombeyi; open bars, Austrocedrus chilensis. Data are observed means ± SE; *P < 0.05, **P < 0.01.

able to withstand prolonged extreme dry conditions under the shady canopy of nurse plants (Kitzberger et al. 2000). Our results are in concordance with those of others (Clinton et al. 1993, 1994; Delissio and Primack 2003; Olano and Palmer 2003; Archaux and Wolters 2006) and suggest that in forest ecosystems, drought disturbance creates a new type of canopy gap and should not be interpreted the same as common background canopy opening processes. Several features make drought gaps fundamentally different. (i) Drought gaps are highly species selective; therefore, strong influences on canopy composition are expected to occur (Fensham and Holman 1999). In the case of tree-fall gaps, some selectivity is possible based on architectural or structural characteristics of species. (ii) As shown in this study and others (Clinton et al. 1993, 1994; Beckage et al. 2000), microenvironment conditions within drought gaps are not equivalent to those in tree-fall gaps. Presence of

Fig. 4. Effects of gap type and species on (a) root to shoot ratio and (b) leaf allocation after two growing season (November 2003 – March 2005). Solid bars, Nothofagus dombeyi; open bars, Austrocedrus chilensis. Data are observed means ± SE; **P < 0.01.

dead standing trees and lack of soil disturbance tend to favour preexisting understory responses that induce shadier microenvironments. Also, higher understory density in drought-induced gaps could lead to drier condition because #

2008 NRC Canada

Suarez and Kitzberger

of an increase in water use in the upper horizons. (iii) Microenvironments are spatially less variable. The diffuse nature of gap boundaries and lack of coarse woody debris make drought gaps internally more homogeneous and with less microsite variation than that of tree-fall gaps (Poulson and Platt 1989; Gray and Spies 1997; Clinton and Baker 2000). Additionally, the lack of coarse woody debris would not provide safe sites for recruitment in drought-induced gaps, as could be in tree-fall gaps by the mechanism of increasing water-holding capacity. (iv) Differential canopy tree mortality of one of the species may create dispersal restrictions in drought-affected forests by reducing the amount of possible seeders in the surrounding area, while in tree-fall gaps, generally all species are able to reach the disturbed area. Surprisingly few studies together with ours have used true drought gaps (Clinton et al. 1993, 1994). Most of the studies analyze drought effects on future forest composition based on irrigation experiments (Fotelli et al. 2001; Marod et al. 2004; Sack 2004; Bunker and Carson 2005; Lloret et al. 2005). In general, the effects of water deficit on seedling survival are interpreted in the context of loss of diversity by species-specific mortality or low growth. Other studies have simulated drought effects using herbicides on canopy vegetation (Beckage et al. 2000; Clinton 2003). But to date, few studies dealing with drought consequences on forest composition have taken into account the possibility that the new environment left by drought might or might not be influential on the composition of future regenerating seedlings. Given the small extent and artificial selection of sites at which these studies are conducted, results should be interpreted cautiously. Extreme events such as droughts are strong drivers of forest dynamics in all forest types, with important influences and imprints on the landscape and feedbacks on future disturbance regimes. In many forests, the combined effects of biotic (pathogens and insects) or abiotic (fire) agents have the potential to extend the mortality effects of drought with important feedback on regeneration patterns (Savage 1997; Allen and Breshears 1998; Hanson and Weltzin 2000). Here, we show that drought itself without any other contributing factor can also be a broadscale disturbance with the potential to cause important vegetation changes. Long-term effects of droughts can drive not only species-specific mortality but also major changes in habitat for further regeneration and succession. Because increases in the magnitude and frequency of extreme climatic events are projected in several climate-change scenarios (e.g., El Nin˜o – Southern Oscillation frequency and amplitude), important compositional and structural changes such as those observed in Patagonian Nothofagus forests can be expected to occur in many forest ecosystems.

Acknowledgments We are grateful to Idea Wild (Fort Collins, Colorado) for providing fundamental equipment to this study. We acknowledge logistic support from the personnel from the Nahuel Huapi National Park and M. Bastidas and F. Cuassolo for their volunteer field assistance. Also, we are grateful to Andrew N. Gray and two anonymous reviewers for helping to improve the final version of the manuscript.

3009

References Allen, C.D., and Breshears, D.D. 1998. Drought-induced shift of a forest-woodland ecotone: rapid landscape response to climate variation. Proc. Natl. Acad. Sci. U.S.A. 95: 14839–14842. doi:10.1073/pnas.95.25.14839. PMID:9843976. Archaux, F., and Wolters, V. 2006. Impact of summer drought on forest biodiversity: what do we know? Ann. For. Sci. 63: 645– 652. doi:10.1051/forest:2006041. Beckage, B., Clark, J.S., Clinton, B.D., and Haines, B.L. 2000. A long-term study of tree seedling recruitment in southern Appalachian forests: the effects of canopy gaps and shrub understories. Can. J. For. Res. 30: 1617–1631. doi:10.1139/cjfr-30-10-1617. Bran, D., Pe´rez, A., Ghermandi, L., and Barrios Lamunie`re, S. D. 2001. Evaluacio´n de poblaciones de coihue (Nothofagus dombeyi) del Parque Nacional Nahuel Huapi, afectadas por la sequı´a 98/99, a escala de paisaje (1:250.000). Informe sobre mortalidad: distribucio´n y grado de afectacio´n, Bariloche, Argentina. Breshears, D.D., Cobb, N.S., Rich, P.M., Price, K.P., Allen, C.D., Balice, R.G., Romme, W.H., Kastens, J.H., Floyd, M.L., Belnap, J., Anderson, J.J., Myers, O.B., and Meyer, C.W. 2005. Regional vegetation die-off in response to global-change-type drought. Proc. Natl. Acad. Sci. U.S.A. 102: 15144–15148. doi:10.1073/ pnas.0505734102. PMID:16217022. Bunker, D.E., and Carson, W.P. 2005. Drought stress and tropical forest woody seedlings: effect on community structure and composition. J. Ecol. 93: 794–806. doi:10.1111/j.1365-2745.2005. 01019.x. Clinton, B.D. 2003. Light, temperature, and soil moisture responses to elevation, evergreen understory, and small canopy gaps in the southern Appalachians. For. Ecol. Manage. 186: 243–255. doi:10.1016/S0378-1127(03)00277-9. Clinton, B.D., and Baker, C.R. 2000. Catastrophic windthrow in the southern Appalachians: characteristics of pits and mounds and initial vegetation responses. For. Ecol. Manage. 126: 51–60. doi:10.1016/S0378-1127(99)00082-1. Clinton, B.D., Boring, L.R., and Swank, W.T. 1993. Canopy gap characteristics and drought influences in oak forests of the Coweeta Basin. Ecology, 74: 1551–1558. doi:10.2307/1940082. Clinton, B.D., Boring, L.R., and Swank, W.T. 1994. Regeneration patterns in canopy gaps of mixed-oak forests of the Southern Appalachians: influences of topographic position and evergreen understory. Am. Midl. Nat. 132: 308–319. doi:10.2307/2426587. Condit, R., Aguilar, S., Herna´ndez, A., Pe´rez, R., Lao, S., Angehr, G., Hubbell, S.P., and Forster, R.B. 2004. Tropical forest dynamics across a rainfall gradient and the impact of an El Nin˜o dry season. J. Trop. Ecol. 20: 51–72. doi:10.1017/ S0266467403001081. Davis, M.A., Wrage, K.J., and Reich, P.B. 1998. Competition between tree seedlings and herbaceous vegetation: support for a theory of resource supply and demand. J. Ecol. 86: 652–661. doi:10.1046/j.1365-2745.1998.00087.x. Davis, M.A., Wrage, K.J., Reich, P.B., Tjoelker, M.G., Schaeffer, T., and Muermann, C. 1999. Survival, growth, and photosynthesis of tree seedlings competing with herbaceous vegetation along a water–light–nitrogen gradient. Plant Ecol. 145: 341–350. doi:10.1023/A:1009802211896. De Fina, A.L. 1972. El clima de la regio´n de los bosques AndinoPatago´nicos. In La regio´n de los bosque andino-patago´nico, Sinopsis General. Edited by M. Dimitri. Instituto Nacional de Tecnologı´a Agropecuaria, Buenos Aires. pp. 35–58. Delissio, L.J., and Primack, R.B. 2003. The impact of drought on the population dynamics of canopy-tree seedlins in an aseasonal Malaysian rain forest. J. Trop. Ecol. 19: 489–500. doi:10.1017/ S0266467403003547. #

2008 NRC Canada

3010 Elliott, K.J., and Swank, W.T. 1994. Impacts of drought on tree mortality and growth in a mixed hardwood forest. J. Veg. Sci. 5: 229–236. doi:10.2307/3236155. Fensham, R.J., and Holman, J.E. 1999. Temporal and spatial patterns in drought-related tree dieback in Australian savanna. J. Appl. Ecol. 36: 1035–1050. doi:10.1046/j.1365-2664.1999. 00460.x. Fotelli, M.N., Gebler, A., Peuke, A.D., and Rennenberg, H. 2001. Drought affects the competitive interactions between Fagus sylvatica seedlings and an early seccessional species, Rubus fruticosus: responses of growth, water status and d13C composition. New Phytol. 151: 427–435. doi:10.1046/j.1469-8137.2001. 00186.x. Gilbert, G.S., Harms, K.E., Hamill, D.N., and Hubbell, S.P. 2001. Effects of seedling size, El Nin˜o drought, seedling density, and distance to nearest conspecific adult on 6-year survival of Ocotea whitei seedlings in Panama´. Oecologia (Berlin), 127: 509– 516. doi:10.1007/s004420000616. Gray, A.N., and Spies, T.A. 1997. Microsite controls on tree seedling establishment in conifer forest canopy gaps. Ecology, 78: 2458–2473. Hanson, P.J., and Weltzin, J.F. 2000. Drought disturbance from climate change: response of United States forests. Sci. Total Environ. 262: 205–220. doi:10.1016/S0048-9697(00)00523-4. PMID:11087027. Heinemann, K., Kitzberger, T., and Veblen, T.T. 2000. Influences of gap microheterogeneity on the regeneration of Nothofagus pumilio in a xeric old-growth forest of northwestern Patagonia, Argentina. Can. J. For. Res. 30: 25–31. doi:10.1139/cjfr-30-125. Hursh, C.R., and Haasis, F.W. 1931. Effects of 1925 summer drought on southern Appalachian hardwoods. Ecology, 12: 380– 386. doi:10.2307/1931640. Kitzberger, T. 1994. Fire regime variation along a northern Patagonian forest–steppe gradient: stand and landscape response. Ph.D. dissertation, Department of Geography, University of Colorado, Boulder, Co. Kitzberger, T., Steinaker, D.F., and Veblen, T.T. 2000. Effects of climatic variability on facilitation of tree establishment in northern Patagonia. Ecology, 81: 1914–1924. Lertzman, K.P. 1995. Forest dynamics, differential mortality and variable recruitment probabilities. J. Veg. Sci. 6: 191–204. doi:10.2307/3236214. Lloret, F., Pen˜uelas, J., and Estiarte, M. 2005. Effects of vegetation canopy and climate on seedling establishment in Mediterranean shrubland. J. Veg. Sci. 16: 67–76. doi:10.1658/1100-9233(2005) 016[0067:EOVCAC]2.0.CO;2. Marod, D., Kutintara, U., Tanaka, H., and Nakashizuka, T. 2004. Effects of drought and fire on seedling survival and growth under contrasting light conditions in a seasonal tropical forest. J. Veg. Sci. 15: 691–700. doi:10.1658/1100-9233(2004) 015[0691:EODAFO]2.0.CO;2. Mueller, R.C., Scudder, C.M., Porter, M.E., Talbot Trotter, R., III,

Can. J. For. Res. Vol. 38, 2008 Gehring, C.A., and Whitham, T. 2005. Differential tree mortality in response to severe drought: evidence for long-term vegetation shifts. J. Ecol. 93: 1085–1093. doi:10.1111/j.1365-2745.2005. 01042.x. Olano, J.M., and Palmer, M.W. 2003. Stand dynamics of an Appalachian old-growth forest during a severe drought episode. For. Ecol. Manage. 174: 139–148. doi:10.1016/S0378-1127(02) 00033-6. Poulson, T.L., and Platt, W.J. 1989. Gap light regimes influence canopy tree diversity. Ecology, 70: 553–555. doi:10.2307/ 1940202. Rebertus, A.J., and Veblen, T.T. 1993. Structure and tree-fall gap dynamics of old-growth Nothofagus forests in Tierra del Fuego, Argentina. J. Veg. Sci. 4: 641–654. doi:10.2307/3236129. Runkle, J.R. 1981. Gap regeneration in some old-growth forests of the eastern United States. Ecology, 62: 1041–1051. doi:10.2307/ 1937003. Sack, L. 2004. Responses of temperate woody seedlings to shade and drought: do trade-offs limit potential niche differentiation? Oikos, 107: 110–127. doi:10.1111/j.0030-1299.2004.13184.x. Savage, M. 1997. The role of anthropogenic influences in a mixedconifer forest mortality episode. J. Veg. Sci. 8: 95–104. doi:10. 2307/3237247. Suarez, M.L., Ghermandi, L., and Kitzberger, T. 2004. Factors predisposing episodic drought-induced tree mortality in Nothofagus — site, climatic sensitivity and growth trends. J. Ecol. 92: 954–966. doi:10.1111/j.1365-2745.2004.00941.x. Swetnam, T.W., and Betancourt, J.L. 1998. Mesoscale disturbance and ecological response to decadal climatic variability in the American southwest. J. Clim. 11: 3128–3147. doi:10.1175/ 1520-0442(1998)011<3128:MDAERT>2.0.CO;2. Valladares, F., and Pearcy, R.W. 2002. Drought can be more critical in the shade than in the sun: a field study of carbon gain and photo-inhibition in a Californian shrub during a dry El Nin˜o year. Plant Cell Environ. 25: 749–759. doi:10.1046/j.1365-3040. 2002.00856.x. Veblen, T.T. 1989. Nothofagus regeneration in treefall gaps in northern Patagonia. Can. J. For. Res. 19: 365–371. doi:10.1139/ x89-055. Veblen, T.T., and Lorenz, D.C. 1987. Post-fire stand development of Austrocedrus-Nothofagus forest in northern Patagonia. Vegetatio, 71: 113–126. Veblen, T.T., Burns, B.R., Kitzberger, T., Lara, A., and Villalba, R. 1995. The ecology of the conifers of southern South America. In Ecology of the southern conifers. Edited by N. Enright and R. Hill. Melbourne University Press, Parkville, Australia. Veblen, T.T., Hill, R.S., and Read, J. 1996. The ecology and biogeography of Nothofagus forests. Yale University Press, London, UK. Williamson, G.B., Laurance, W.F., Oliveira, A.A., Delamoˆnica, P., Gascon, C., Lovejoy, T.E., and Pohl, L. 2000. Amazonian tree mortality during the 1997 El Nin˜o drought. Conserv. Biol. 14: 1538–1542. doi:10.1046/j.1523-1739.2000.99298.x.

#

2008 NRC Canada