Author's personal copy Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / p a l a e o

Southern-most Nothofagus trees enduring ice ages: Genetic evidence and ecological niche retrodiction reveal high latitude (54°S) glacial refugia Andrea C. Premoli ⁎, Paula Mathiasen, Thomas Kitzberger Laboratorio Ecotono, Universidad Nacional del Comahue, INIBIOMA CONICET, Quintral 1250, 8400, Bariloche, Argentina

a r t i c l e

i n f o

Article history: Received 9 March 2010 Received in revised form 20 September 2010 Accepted 30 September 2010 Available online 7 October 2010 Keywords: Ecological niche models Isozymes Multiple glacial refugia Nothofagus pumilio South America Tierra del Fuego

a b s t r a c t Hypotheses of tree survival in refugia or long-distance postglacial migration can be disentangled combining genetic polymorphisms and ecological niche modeling. This is particularly relevant for ecologically distinct taxa which are undistinguishable by pollen morphology in fossil records. We hereby test the long-term persistence of the cold-hardy Nothofagus pumilio at high latitudes in southern South America during glacial cycles. Modern and past (LGM) ecological niche modeling (ENM) for N. pumilio was developed using current climate data and including 19 bioclimatic variables and topography. We collected fresh tissue for isozyme analysis from 14 locations within Tierra del Fuego selected on the basis of their location relative to LGM glacier extent: ice-free, ice-margin, and formerly glaciated areas. We resolved 11 isozyme loci and calculated withinpopulation genetic diversity and among-population divergence parameters. Glacial effects on population structure were analyzed by UPGMA and Bayesian models. ENM yielded distinct areas in eastern and interior Tierra del Fuego exceeding N 0.3 probability values that would have been suitable for N. pumilio during the LGM. Populations within formerly glaciated areas hold lower genetic diversity and thus were affected by genetic drift during colonization from refugia. Our data provided no evidence for wide-ranging postglacial colonization from warmer northern locales and suggested local survival of Nothofagus forests at southernmost South America. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Paleoecological records, specifically pollen records, offer key information for evaluating species' responses to past climatic changes. However, their utility can be undermined by the fact that low-density plant populations and even locally surviving scattered individuals, which are particularly relevant in genetic terms, may remain undetected by the fossil records. This problem is augmented by the fact that many ecologically distinct species cannot be differentiated by their pollen morphology. As a result, conclusions about local persistence or recent arrival for a given species cannot be teased apart. The combined use of genetic markers (e.g. Anderson et al., 2006) and ecological niche modeling predictions (e.g. Waltari et al., 2007) would complement paleoenvironmental reconstructions in the detection of past presence of local populations that may have acted as key refugia for repopulation after deglaciation. Although uncertainties and concerns exist about the implementation and interpretation of species distribution modeling this approach has been successfully applied to northern hemisphere taxa (e.g. Svenning et al., 2008). Such a combined approach of genetic analysis, ecological niche modeling, and independent paleoecological data would have profound implica⁎ Corresponding author. Tel.: + 54 2944 422111x503; fax: +54 2944 422111. E-mail addresses: [email protected], [email protected] (A.C. Premoli). 0031-0182/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.palaeo.2010.09.030

tions in the understanding of the existence and location of glacial refugia, tree migration rates, and resilience to climatic forcing. Although extensive paleoecological work has been performed in southern South America, it remains controversial if the largely unglaciated coastal areas in northeastern Tierra del Fuego and contiguous, nowadays submerged continental shelves harbored local Nothofagus forests during the last glaciations (Markgraf, 1993; Markgraf and Huber, 2010). A steppe-tundra vegetation would have expanded in proglacial environments during the Last Glacial Maximum (LGM) at ca. 18 to 20 ka BP (Rabassa et al., 2000; Mancini et al., 2008; Markgraf and Huber, 2010). The classic interpretations of Caldenius (1932) located the eastern limit of the two oldest glaciations in Tierra del Fuego beyond the coast onto the present Atlantic submerged continental shelf. More recently, consensus established that portions of land remained ice-free during major Pleistocene glaciations at least since 1.5 Ma BP (Rabassa et al., 2000; Bujalesky and Isla, 2006). These unglaciated terrains were surrounded by large outlet glaciers originating from the Cordillera Darwin ice cap in western Tierra del Fuego that flowed north and eastwards as the Magellan, Bahía Inútil, Fagnano, and Beagle lobes, from N to S, respectively, some of them onto the present Atlantic submerged shelf (Mercer, 1983; Meglioli et al., 1990; Porter, 1990; Clapperton, 1993; Isla and Schnack, 1995; Rabassa et al., 2000; Coronato et al., 2004; Rabassa et al., 2005; Sugden et al., 2005; Rabassa, 2008).

Author's personal copy 248

A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

In contrast to the wealth of support for a coastal eastern ice-free area in Tierra del Fuego, it has remained unclear whether the dominant tree species survived LGM as well as former glacial intervals at these and other suitable locations within the island. In addition, ecologically distinct species within subgenus Nothofagus inhabiting high latitudes (N. antarctica, N. betuloides, and N. pumilio) share fusca (b) pollen type (Hill, 2001) and thus are undistinguishable from records. Although Nothofagus pumilio forests are widely distributed throughout the southeastern lowlands of the island and on both sides of the Andes, it is yet unclear if full glacial climate conditions allowed the survival of small populations of Nothofagus. Markgraf (1993) suggested that tree taxa must have grown in suitable coastal locales or areas that became exposed when sea level was lower during the full glacial. Two main sources of evidence are presented by Markgraf (1993). On one hand, evidence based on pollen and fossil beetle records from different latitudes shows that tree taxa and flightless forest beetles did not show substantial migration lags, thus suggesting that they remained locally present throughout the full glacial (Ashworth and Markgraf, 1989; Ashworth et al., 1991). Local persistence of Nothofagus is suggested by one record from the steppe/forest boundary near the eastern Atlantic coast (Viamonte 54°11′S, 66°21′W, 5 m elevation; Auer 1956 recalculated by Markgraf 1993, Fig. 2b), showing continuous representation of c. 20–40% Nothofagus pollen since an earlier interstadial (N41 kyr BP). On the other hand, dominance of non-arboreal taxa, Poaceae, Asteraceae, herbaceous taxa, and traces of Nothofagus in high latitude pollen records suggests resemblance with present-day steppe and steppe-scrub with continued presence of local tree populations (Markgraf, 1993). In addition, the mire at Puerto Hambre (53°36′21″S, 70°55′53″W), 100 m inland from the Straits of Magellan in southern Patagonia contained Nothofagus pollen at an estimated 14,700 14 C yr BP (c. 17,500 cal. yr BP) at a time when eustatic sea level was at least 120 m lower than today (Kaplan et al., 2008). This result was interpreted as local presence of forests (Heusser et al., 2000) or long-distance transport (McCulloch and Davies, 2001). Genetic markers on tree species have been widely used to reconstruct historical events in Patagonia. In particular, genetic evidence is building up to suggest that cold-tolerant woody taxa survived locally in multiple glacial refugia along their current distributional range (Premoli et al., 2000). Some of the refugia of widespread taxa such as conifers including Pilgerodendron uviferum (Cupressaceae) (Premoli et al., 2002), Podocarpus nubigena (Podocarpaceae) (Quiroga and Premoli, 2010), and the broadleaf Embothrium coccineum (Proteaceae) (Souto and Premoli, 2007) were probably located towards the southern-most species' range where maritime influence is higher compared to northern Patagonia. A preliminary study on a reduced number of populations of Nothofagus pumilio using isozymes yielded levels of genetic diversity for the southern populations from Tierra del Fuego similar to those located further north, which was interpreted to indicate local survival in multiple refugia during the last glacial period (Premoli, 1998). DNA sequences using 13 universal primers of non-coding mitochondrial and chloroplast regions yielded no polymorphism within Tierra del Fuego (BIOCORES, 2003, p. 64). This was confirmed by a thorough screening of populations using sequences of three cpDNA regions throughout the entire range of N. pumilio that produced two reciprocal clades each containing populations north and south 43° S (Mathiasen and Premoli, 2010). Although the southern range of the species was genetically depauperate in chloroplast DNA this should not be taken as evidence for recent colonization of this region. Instead all the chloroplast genotypes (haplotypes) were endemic to this region, and were all more closely related to each other than to any other haplotypes. Given the very slow rates of evolution of the chloroplast genome, this provides evidence for long-term survival of the species at high latitudes in South America. Ecological niche models (ENM) are increasingly being used in conjunction with paleoecological reconstructions and phylogeo-

graphic analyses to determine past species distributions. ENMs relate known present occurrences of species to physical data in a spatial context and thus allow to infer environmental requirements or potential distribution patterns (Guisan and Thuiller, 2005). Under the assumptions of niche conservatism (Wiens and Graham, 2005; Martínez-Meyer and Peterson, 2006) and equilibrium of species with environmental conditions, these models can be used to pinpoint possible areas that may have remained suitable for the species during glacial periods, and therefore identify potential refugia. While ENMs have long been used to reconstruct plant paleodistributions of Southern Hemisphere LGM scenarios (Williams, 1991; Kirkpatrick and Fowler, 1998; Eeley et al., 1999), only recently they have started to be combined and compared with genetic/phylogeographic evidence (Carnaval and Moritz, 2008; Jakob et al., 2009). In this study we hypothesize that Atlantic coastal areas provided the appropriate conditions to sustain forest populations that may have survived throughout the Quaternary and may have served as colonization source during the interstadials. We use genetic measures based on isozyme polymorphisms and ecological niche modeling to analyze: 1. if and where ice-free areas may have remained suitable during glacial maxima for sustaining Nothofagus pumilio forests within the island of Tierra del Fuego and 2. if patterns of withinpopulation genetic diversity and among-population divergence of the dominant tree taxon N. pumilio throughout Tierra del Fuego suggest a structuring that is concordant with the paleosuitability distribution. Although microsatellites were developed for Nothofagus species within the subgenus Lophozonia, (Jones et al., 2004; Marchelli et al., 2008) their transferability and polymorphism in N. pumilio is still being optimized (M. Arbetman, Universidad Nacional del Comahue, pers.com.). For this study, population sampling design encompassed an area within the continuous LGM ice boundary as determined by geomorphological evidence (Rabassa et al., 2000), a second area outside LGM ice boundary, and a third area located within terminal moraines corresponding to different Pleistocene glacial intervals. If viable populations persisted outside the continuous ice boundary during the full glacial we expect that those populations would harbor a higher genetic diversity than populations located in areas formerly covered by the ice sheet, because the latter may have become genetically impoverished as a result of genetic drift affecting populations during recolonization from refugia. In addition, if N. pumilio populations from Tierra del Fuego represented descendents from an ancient gene pool that had established before the onset of Quaternary climatic oscillations, we also predict that refugial areas would be genetically similar to each other due to long-term local persistence in climatically stable ice-free areas. 2. Materials and methods 2.1. Study species Nothofagus pumilio is one of the most widespread South American Nothofagus species. It occurs throughout the southern Andes and in the high latitude lowlands encompassing c. 20° latitude (Fig. 1, left panel). 2.2. Ecological niche modeling To develop the ecological niche modeling (ENM) for Nothofagus pumilio, we used current (pre-industrial) climate data at 30″ (~1 km) spatial resolution from the WordlClim database (Hijmans et al., 2005). This database provides 19 bioclimatic variables that represent summaries of means and variation in temperature and precipitation, and likely summarize dimensions of climate particularly relevant in determining species distributions (Hijmans et al., 2005). In addition we used elevation, slope, and aspect derived from the Digital Elevation Model Shuttle Radar Topography Mission (SRTM-90)

Author's personal copy A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

249

Fig. 1. Distribution map of Nothofagus pumilio along the southern Andes and location of austral-most sampled populations in the island of Tierra del Fuego (right panel). Population names follow Table 1.

dataset. A subset of spatially explicit environmental and topographic datasets for the southern South American region (60–80°W, 30–56°S) was used for the development of ENMs, focusing on our study area of interest (Island of Tierra del Fuego, 64–71°W, 52.5–56°S). To create LGM climate layers for the ENMs, we used LGM bioclimatic data at 2.5′ spatial resolution that were drawn from the 21 kyr BP simulations using the Community Climate System Model (NCAR-CCSM; Collins et al., 2004). We downscaled bioclimatic surfaces to 30″ spatial resolution following Waltari et al. (2007). First, we calculated the differences between LGM and recent (preindustrial) conditions at the native coarse resolutions. These differences were then interpolated to 30″ spatial resolution using the spline function in ArcMap (ESRI, Redlands, CA) with the tension option. Finally, the interpolated difference maps were added to the WorldClim current 30″ climate data to obtain the downscaled 30″ bioclimatic dataset corresponding to LGM (21 kyr BP). NCAR-CCSM simulates a c. −5 °C reduction during LGM (compared to modern simulated data) in mean annual temperature along the Atlantic coast of southern Argentina (c. 50–55°S, S1.a). In Tierra del Fuego simulated LGM reductions in mean annual temperature increase inland from NE to SW reaching c. −8 °C (S1.a). NCAR-CCSM model has been recently validated over the Atlantic Ocean comparing simulated with multi-proxy reconstructions of mean annual and seasonal sea surface temperatures (SST; Clauzet et al., 2008), showing better agreement in the western and central part of the South Atlantic. Specifically, for the southeastern tip of Tierra del Fuego simulated and foraminifera-reconstructed mean annual SST are coincidently 0–2 °C. Over the northeastern coast of Tierra del Fuego the model slightly underestimates mean annual SST of 0–2 °C compared to proxy reconstruction (2–4 °C; Clauzet et al., 2008). Present-day occurrence data of Nothofagus pumilio forests (Fig. 1, left panel) was compiled from a variety of sources: 1:500,000 Valdivian Ecoregion Map (Lara et al., 2000), Argentinean 1st Native Forest Inventory (SAyDS, 2005), Chilean Native Forest Cadastre 1:500,000 (CONAF-CONAMA-BIRF, 1999), and Native Forest Inventory of Tierra del Fuego 1:500,000 (Collado, 2001). Raster occurrence data was vectorized to points to obtain geographic coordinates of localities, removing duplicate records coming from overlapping data

sources. To reduce spatial autocorrelation we removed from the training set those localities within 1′ latitude of each other. To construct the ENM we used the Maxent (Phillips et al., 2006) algorithm. Maxent estimates a target probability distribution by finding the probability distribution of maximum entropy (i.e., the most spread out, or closest to uniform), subject to a set of constraints that represent our incomplete information about the target distribution. When applied to presence-only species distribution modeling, the pixels of the study area make up the space on which the Maxent probability distribution is defined; the pixels with known species occurrence records constitute the sample points, and the features are climatic, topographic, or other environmental variables (Phillips et al., 2006). The Maxent probability distribution is based on a concise mathematical definition that allows interpreting how each environmental variable relates to suitability. Provided the same environmental variables are available, suitability may be projected onto past or future environmental scenarios. We first run Maxent (version 3.3.1) with 19 current bioclimatic variables and 3 topographic variables (elevation, “aspect" and slope) to assess the importance of each variable using the jacknife tests in which training gain is monitored when each variable is excluded in turn. Subsequently we used a model created with the remaining variables and finally the model is created using each variable in isolation. With the final bioclimatic variables for current and projected (LGM) data we ran Maxent using default convergence threshold (10− 5) and maximum number of iteration (500) values, using 25% of localities for model training. Maxent outputs a continuous probability value, ranging from 0 to 1, an indicator of relative suitability for the species, based on the principle of maximum entropy, as constrained by the input occurrence data. The goodness of fit between the model and training data was assessed by analyzing the area under the curve (AUC) of a receiver operating curve (ROC). A ROC plot is obtained by plotting all sensitivity values (true positive fraction) on the y axis against their equivalent (1-specificity) values (false positive fraction) for all available thresholds on the x axis. The area under the ROC function (AUC) is usually taken to be an important goodness of fit index because it provides a single measure of overall accuracy that is not dependent upon a particular classifier threshold (Fielding and Bell,

Author's personal copy 250

A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

Fig. 2. Last Glacial Maximum potential distribution of Nothofagus pumilio in southern South America using the Maxent model including extant location of sampled populations at austral-most areas of Tierra del Fuego (right panel). Symbols represent study locations as classified in relation to the ice cover during the last glacial maximum: ice-free refugial locations (red circles), leading-edge populations (blue squares), and glaciated areas (light-blue triangles).

1997). When AUC = 0.5, the performance of the model does not differ from chance classification, and as AUC approaches 1, the predictive power of the model increases. AUC indicates the proportion of times a random selection of species presences will have a higher score than a random selection of absences (Fielding and Bell, 1997).

geomorphologic evidence of Quaternary glaciations), ice-margin areas located near the LGM ice sheet margin as well as those within terminal moraines of different glacial events, and areas presumably located on formerly glaciated surfaces mostly on mountainous terrains SW of the most recent 16 kyr BP terminal moraine arcs (Figs. 1 and 2, Table 1). Despite availability of paleoreconstuction maps (Coronato et al., 1999; Rabassa et al., 2000; Glasser et al., 2008; Fig. 2) the lack of georeferenced ice sheet boundaries prevented an a priori classification of populations according to location within these areas. However, some populations clearly fell within stable nonglaciated areas (e.g. SP, ML) and others within formerly glaciated

2.3. Genetic analyses We collected samples for genetic analyses of Nothofagus pumilio at 14 locations in Argentine Tierra del Fuego and neighboring areas in Chile. Populations were sampled to include stable ice-free areas (no

Table 1 Location, isozyme diversity, and LGM information of sampled Nothofagus pumilio populations from Tierra del Fuego.

Population

Lat. (S)

Long. (W)

A

AP

AR

AT

P

HO

HE

LGM suitability

Distance to 16 kyr ice bounday (km)

Cluster*

1 2 3 4 5 6 7 8 9 10 11 12 13 14

54° 02.955' 54° 03.852' 54° 19.220' 54° 20.981' 54º 22.348' 54° 24.709' 54° 29.215' 54° 31.793' 54° 36.585' 54° 41.369' 54° 47.928' 54° 48.55' 54° 52.15' 54° 53.044'

68° 45.754' 68° 57.463' 68° 48.979' 66° 40.389' 67º 14.921' 67° 14.076' 66° 27.037' 66° 56.365' 67° 23.837' 67° 49.838' 68° 22.194' 68° 35.75' 67° 26.40' 67° 12.023'

1.5 1.4 1.4 1.3 1.4 1.5 1.3 1.5 1.5 1.2 1.5 1.3 1.5 1.3

1.66 1.44 1.44 1.33 1.55 1.55 1.33 1.66 1.55 1.22 1.55 1.33 1.55 1.33

1.41 1.22 1.22 1.19 1.30 1.38 1.19 1.43 1.44 1.22

36.4 27.3 27.3 18.2 27.3 45.4 27.3 45.4 36.4 18.2 45.4 27.3 45.4 27.3

0.067 0.073 0.057 0.031 0.06 0.021 0.043 0.094 0.088 0.017 0.045 0.041 0.015 0.037

0.071 0.071 0.057 0.035 0.075 0.032 0.044 0.080 0.091 0.031 0.054 0.055 0.032 0.045

0.16 glaciated glaciated 0.22 0.16 0.16 0.34 0.23 glaciated glaciated glaciated glaciated glaciated glaciated

1 -10 -4 36 15 10 44 12 -10 -23 -42 -50 -22 -14

O

1.43 1.33 1.49 1.33

17 15 15 14 15 16 14 17 16 13 16 14 16 14

Average

1.4

1.46

1.33

15

32.5

s.e.

0.03

0.04

0.03

0.3

PR Paso Radman LB Lago Blanco SV Sector Vicuña SP San Pablo TN Tolhuin Norte TO Tolhuin ML Maria Luisa LC La Correntina LA Los Alamos PG Paso Garibaldi VM Volcan Martial LR Lago Roca BB Bahia Brown RC Rio Cambaceres

2.64

* Last column symbols correspond to groups of populations yielded by UPGA multivariate cluster analysis.

0.049

0.055

0.007

0.005

O ∆ O O ∆ ∆ ∆ ∆ ∆

Author's personal copy A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

terrain (e.g. LR, VM, and BB). In each population, we collected a terminal twig containing fresh leaf tissue from each of 30 randomly sampled individuals. Samples were kept cold in a portable cooler until arrival to the laboratory where enzyme extraction, electrophoretic conditions, and staining protocols followed Premoli (2003). Homogenates were stored at −80 °C and were analyzed by horizontal electrophoresis on 12% P/W starch gels (Starch Art Corporation, Smithville, Texas). We used isozyme polymorphisms that were consistently resolved in all species within subgenus Nothofagus including N. pumilio (Premoli, 1996, 2003). Although a total of 14 loci were previously resolved in N. pumilio, we used only 11 given that three loci for 6-phosphogluconate dehydrogenase (6Pgd-1, 6Pgd-2, 6Pgd-3) were consistently monomorphic. Instead, we were able to resolve shikimate dehydrogenase (Skdh) which was previously used successfully in different species of the subgenus Nothofagus (Premoli, 1996; Premoli and Kitzberger, 2005; Premoli and Steinke, 2008). The scored loci were alcohol dehydrogenase (Adh-1, Adh-2), aldolase (Ald), isocitrate dehydrogenase (Idh-1, Idh-2), malate dehydrogenase (Mdh-1, Mdh-2, Mdh-3), phosphoglucoisomerase (Pgi-1, Pgi-2), and shikimate dehydrogenase (Skdh). Heterogeneity of allele frequencies across populations was tested by chi-square tests following Workman and Niswander (1970). We calculated parameters of within-population genetic diversity using standard measures. These were, A the mean number of alleles per locus, AP the mean number of alleles per polymorphic locus, AR the allelic richness as a measure of the number of alleles independent of sample size, P the percent of polymorphic loci sensu stricto, and HO and HE as the observed and expected heterozygosity under the Hardy–Weinberg equilibrium, respectively. Inbreeding coefficients were calculated by the fixation indices (F) for each locus in each population. Departure of genotypic frequencies from the Hardy–Weinberg equilibrium and the significance of locus-wise fixation indices were tested by chi-squared tests (Li and Horvitz, 1953). Average Wright's (1965) inbreeding coefficients were estimated by FIS that represents the inbreeding due to non-random mating within populations and by FST representing inbreeding due to population subdivision. Means and 95% confidence intervals were obtained by resampling schemes following Weir and Cockerham (1984). Population structure was analyzed by total (HT) and within-population (HS) genetic diversity using polymorphic loci (Nei, 1973). These were calculated by the program FSTAT version 2.9.1 (Goudet, 2002) which computes unbiased estimates. Genetic relationships among populations were analyzed by UPGMA cluster analysis using Cavalli-Sforza and Edwards (1967) chord distances. These have proved to give generally similar results to the most commonly used Nei's (1972) genetic distance (Wiens, 2000) but put greater importance on the presence of low-frequency alleles which are likely to be more informative as far as the migration history of a species is concerned (cf. Konnert and Bergmann, 1995). To portray spatial patterns of among-population genetic differentiation we plotted paired chord distance values obtained from the input cophenetic matrix as bars on a geographical map. To further detect population structure, we also ran a Bayesian algorithm that assign individuals in the sample on the basis of their genotypes at multiple loci, probabilistically, to their known cluster of origin using the software Structure v. 2.3.1 (Pritchard et al., 2000). The number of clusters (K) was selected heuristically by comparing penalized log likelihoods over 10 replicates of each independent Markov chain Monte Carlo run to make cluster assignments of individuals specifying prior information concerning sampling location into 1 to 14 clusters, i.e. studied populations, with a burn-in of 500,000 iterations and a run length of 500,000 iterations following the burn-in. To provide an estimation of the number of clusters, K, we used an ad hoc statistic ΔK which is based on the rate of change in the log probability of data between successive K values (Evanno et al., 2005). Finally, in order to test for long-distance postglacial dispersal, we combined our isozyme data with a previously published dataset on 40 populations sampled

251

along the entire range of Nothofagus pumilio (Mathiasen, 2010; Mathiasen and Premoli, 2010). If southern-most populations would have derived by long-distance migration from northern warmtemperate sources in a stepwise fashion, populations from Tierra del Fuego would be genetically similar to nearby continental populations, which were tested by UPGMA multivariate cluster analysis. 3. Results ENM for Nothofagus pumilio that maximize training gain included 19 bioclimatic variables yielding a training area under the curve (AUC) of 0.926. Nothofagus pumilio suitability during the LGM shows two major areas with suitabilities exceeding values of N0.3 (orange– red in Fig. 2) that lie outside the ice sheet: a northerly area (N of 41°S) on both slopes of the Andes and a southerly area between 54 and 55°S towards the Atlantic coast of the Island of Tierra del Fuego (Fig. 2). Between these contrasting latitudes only smaller areas of higher suitability were found towards the eastern ice sheet boundary. Suitability values within the continuous ice sheet were flawed because LGM altitude-corrected bioclimatic variables do not account for thickness of the ice sheet (higher elevation would mean lower suitability during glacial times). A close-up of the modeled N. pumilio distribution at LGM (Fig. 2 right pannel) shows high suitability for the entire eastern tip of Tierra del Fuego and Staten Island. While some of these areas (interior and east central coastal areas) are currently still dominated by N. pumilio (Fig. 1, right panel) others (eastern Tierra del Fuego and Staten Island) are currently dominated by evergreen N. betuloides forests. LGM suitability values of sampling locations that correspond to N. pumilio populations outside the ice sheet ranged from 0.16 to 0.34 with highest values for the eastern coastal populations ML, LC, and SP (Table 1). Eleven isozyme loci were reliably resolved in the southern-most populations of Nothofagus pumilio in Tierra del Fuego, nine of which were polymorphic in at least one population (Adh-1, Adh-2, Ald, Idh-1, Idh-2, Mdh-2, Mdh-3, Pgi-2, and Skdh). Out of the nine polymorphic loci, seven (Adh-1, Adh-2, Ald, Idh-1, Idh-2, Mdh-2, and Mdh-3) yielded significantly different allele frequencies across populations (chi sqare test p b 0.02). In addition, allele frequencies for Skdh were marginally significant (chi sqare test p b 0.07). Levels of genetic diversity HE ranged from 0.031 to 0.091 and polymorphism P from 18 to 45% (Table 1). Overall mean genetic diversity, HT, was 0.059, most of which was distributed within populations (HS = 0.056). Fixation indices yielded significantly positive values in 14 of 50 possible tests (Table 2). While recently founded populations within glaciated areas show signals of inbreeding, northern populations LB and SV and the eastern SP, ML, and LC yielded non significant F values which may be an indication of relatively big, stable populations (Table 2). Moreover, populations located within the ice limits, such as TN and LA, and the high elevation PG population yielded positive F values in at least 50% of possible tests and thus suggesting that inbreeding maybe a result of small population sizes and/or isolation. Hierarchical analysis of inbreeding coefficients yielded a significant positive estimate of FIS = 0.085 ± 0.049 SE (95% CI = 0.033–0.191). The inbreeding due to population subdivision resulted in a relatively low, but significant FST = 0.049 ± 0.024 SE (95% CI = 0.020–0.104). These results indicate that N. pumilio populations are genetically structured which means that significant genetic divergence exist within (FIS) and among (FST) populations. While the former most probably reflects local seed establishment that results in the formation of family groups, the latter shows restrictions for gene flow. Multivariate cluster analysis revealed two strongly supported clusters of sites. The most genetically distinct group contains four sites within the known glacial limits and one (TO) just outside, the other group representing five sites outside and three sites just within the glacial limits (Fig. 3). Multivariate cluster analyses using other distance metrics such as Nei's (1978) unbiased genetic distance and

Author's personal copy 252

A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

Table 2 Within-population fixation indices in populations of Nothofagus pumilio from Tierra del Fuego. Last column depicts percentage of significant departures to Hardy–Weinberg equilibrium tests (*). Adh-1 1 2 3 4 5 6 7 8 9 10 11 12 13 14

PR Paso Radman LB Lago Blanco SV Sector Vicuña SP San Pablo TN Tolhuin Norte TO Tolhuin ML María Luisa LC La Correntina LA Los Alamos PG Paso Garibaldi VM Volcan Martial LR Lago Roca BB Bahia Brown RC Río Cambaceres

Idh-1

Idh-2 0.14*

Adh-2

Mdh-2

Mdh-3

Ald

−0.034 0.022 0.059 0.394* − 0.017 − 0.017

1* − 0.114 0.041* − 0.085 − 0.051 −0.017

−0.071

−0.034 −0.002

−0.017

1* 1*

modified Rogers distance (Wright, 1978) yielded the same tree topology, although with lower clustering levels of b1 and 9%, respectively, and populations from areas of glacial advance were the most dissimilar. Pairwise chord distance measures showed that populations were more genetically alike to populations within each group than to other geographically close populations (Fig. 4). For example, southwestern populations within glaciated areas were more genetically similar to population TO within the Tolhuin area than to the nearby population LA (Fig. 4). This result shows that even close populations within a given area may have had distinct evolutionary histories. Bayesian analyses accurately detected the uppermost hierarchical level of structure for K = 2 (Fig. 5, inset graph). Although almost all populations contained individuals sharing most-common genotypes, assignment of individuals to groups yielded approximately the same two main clusters of populations (Fig. 5) as by UPGMA (Fig. 3) which mainly discriminated those populations within glaciated areas from the rest. Populations within the ice limits shared the most-common genotypes with population TO which in turn were different to those found in populations outside and nearby the ice (Fig. 5). Multivariate cluster analysis of the entire data set of Nothofagus pumilio showed that populations from Tierra del Fuego clustered with distant northern-most

−0.017

− 0.034 0.198

0.429* −0.085 0.124 −0.017 −0.103

−0.049

Pgi-2

Skdh

% F p b 0.05

−0.132 −0.111 −0.132 −0.075 −0.202 −0.016 −0.111 −0.091 0.649* 0.467* 0.349* 0.632* 0.649* 0.464*

0.133 − 0.106 − 0.051 0.184 0.462*

25 0 0 0 67 20 0 0 50 100 40 33 40 33

0.079 − 0.333 − 0.2

ones which in turn were more genetically distinct to continental southern populations (S1.b). Genetically distinct groups of populations from Tierra del Fuego confirmed by UPGMA cluster analysis and Bayesian analyses had a significantly different average within-population genetic diversity HE. Populations within glaciated areas had significantly lower average genetic diversity (HE = 0.0415) than that nearby and outside the ice (HE = 0.0655) (Mann–Whitney U non parametric test, z = 2.32, p = 0.02). Therefore, locations that were covered by ice have suffered from genetic drift during the postglacial colonization process resulting in reduced genetic diversity. 4. Discussion Cold-hardy Nothofagus pumilio probably survived ice ages in distinct ice-free locations along its current distributional range including southern-most areas. While ENM shows extensive areas of potential population survival towards the northern range of N. pumilio, in the south and at intermediate latitudes towards the east of the Andes, distinct areas seem to have provided habitat for tree survival during glacial cycles. This result confirms the multiple refugia hypothesis sensu Premoli (1998) and Premoli et al. (2000) which,

Fig. 3. UPGMA multivariate cluster analysis by means of isozymes of southern-most populations of Nothofagus pumilio from Tierra del Fuego. Distinct clades depict ice-free refugial areas (red circles), leading-edge populations (blue squares), and glaciated areas (light-blue triangles).

Author's personal copy A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

253

Fig. 4. Map showing pairwise isozyme distances among all studied populations of Nothofagus pumilio from Tierra del Fuego. Thickness of lines depict relative level of gene flow between population pairs. Symbols represent ice-free refugial areas (red circles), leading-edge populations (blue squares), and glaciated areas (light-blue triangles).

based on genetic data, suggested that cold-tolerant taxa were able to persist locally throughout their current ranges. Also, a wide-range genetic study of N. pumilio corroborates southern populations as or

even more genetically diverse (mean gene diversity HE = 0.08 and allelic richness AR = 1.06) than northern populations (HE = 0.044 and AR = 0.43) (Mathiasen and Premoli, 2010). Optimal current conditions

Fig. 5. Pie charts at sampled locations show the proportion of individuals that were classified into each of two homogenous groups of populations detected by the software Structure which were interpreted as inside glaciated areas (light-blue) or outside (red), respectively. Upper right inset represents the second order rate of change of the likelihood function (ΔK) with respect to K which show a clear peak at the true value of K, here two clusters.

Author's personal copy 254

A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

for growth of N. pumilio at these southern latitudes in combination with long-term regional persistence during the Pleistocene may explain the relatively high polymorphism in Tierra del Fuego. On the other hand, if populations from Tierra del Fuego would have became locally extinct and recolonized after glacial retreat, higher genetic identity would have been expected with continental southern populations. Populations inhabiting north and south extremes of the latitudinal range were more genetically alike to each other than to mid-latitude ones by means of multivariate cluster analysis. Strikingly, these mid-latitude populations coincide with the area where strongest reductions in temperature were predicted by the LGM model (S1.a). This may suggest that N. pumilio consists of an ancestral gene pool that became fragmented sometime in the past (Premoli, 1998) and that northern and southern-most populations have served as distant recolonization sources of mid-latitudes. Phylogeographic evidence by cpDNA sequences shows that allopatric divergence found between southern and northern populations had pre-dated climatic cycles of the Neogene and suggested tectonic controls affecting ancient lineages (Mathiasen and Premoli, 2010). Although more genetically diverse, similar phylogeographic structure by cpDNA was recovered for the widespread Nothofagus antarctica. Similarly to N. pumilio, populations of the cold-tolerant N. antarctica from Tierra del Fuego were as genetically diverse as elsewhere along its range, which may also suggest long-lasting persistent populations (Acosta et al., 2009). Although morphologically distinct, these two closely related species coexist in sympatry and naturally hybridize at present (Quiroga et al., 2005) as well in the past as evidenced by shared cpDNA sequences over large geographic areas (Acosta and Premoli, 2010). Whether pure species and their hybrids survived locally in sympatry during ice ages is still to be discovered as their pollen are indistinguishable. Our data suggest local persistence of Nothofagus pumilio in high southern latitudes excluding wide-ranging migrations during the Pleistocene interglacials from refugia in warmer northerly areas. These results contrast those suggested for Northern Hemisphere woody species (Hewitt, 2000). Apparently a distinct full glacial scenario existed in the Southern Hemisphere that allowed long-term survival of forests (Markgraf et al., 1995). In particular, the ENM provides evidence for the existence of ice-free areas during full glacial times suitable for local persistence of forests within Tierra del Fuego. Likewise, cpDNA data together with ecological niche modeling of the grass Hordeum provided no evidence for range shifts towards the north during the last glacial maximum and recolonization of southerly grassland habitats afterwards, but indicated in situ survival of large populations within their extant distribution ranges even in southernmost Patagonia and Tierra del Fuego (Jakob et al., 2009). However, examples exist in the literature that show contradictory results between climate models and molecular evidence (e.g. Worth et al., 2009). This suggests that a combination of different proxy data including paleoecological records, climatic simulations, and genetic screening of entire species' ranges are needed to robustly reconstruct complex biogeographic histories. Our isozyme data show that the gene pool of Nothofagus pumilio in Tierra del Fuego consists of two groups of populations. This finding together with LGM potential distribution maps suggests at least two refugial areas for N. pumilio at its southern-most range. These were located in coastal areas along the southeastern Atlantic coast and the northwestern Pacific coast at the western portion of the Straits of Magellan. Interestingly, populations from ice-free locations appear to be genetically similar suggesting that gene flow among long-lasting populations has prevented genetic differentiation during glacial episodes. Also, some populations within glacial limits, e.g. Tolhuin (TO), seem to have been the colonization source of inland populations (Fig. 4). Moreover, our data show that distinct groups of populations have been separated through time. Hence, within Tierra del Fuego two distinct areas can be identified: relatively stable areas either ice-free or

within ice limits that have persisted as such throughout glacial episodes, and glaciated areas where N. pumilio has probably become locally extinct in the past and that only recently were colonized from long-term persistent populations. Nonetheless, we cannot discard the possibility that other refugial areas may have existed during cold periods within Tierra del Fuego such as in southern coastal areas. Although genetically diverse, significant heterozygous deficits were measured at some populations within ice limits such as Tolhuin Norte (TN) and Los Alamos (LA) which may provide evidence of in situ persistence and recurrent population bottlenecks through time. Likewise, high inbreeding measured in the high elevation population Paso Garibaldi (PG) within the glaciated group may reflect the effects of drift due to altitudinal movements during climatic oscillations of the Quaternary. In particular, morphologically intermediate individuals between N. pumilio and N. antarctica were found at this location (ACPremoli, personal observation) which may suggest local survival at high elevation ice-free areas (nunataks) that might have provided refugia and foster hybridization. Significantly lower genetic diversity within glaciated areas may be indicative of founder events. Unlike some cold-intolerant tree species that migrated hundreds to thousands of kilometers during glacial times (Kitzberger et al., 2010), Nothofagus pumilio (and other cold-tolerant species elsewhere) possibly grew close or even in contact with full glacial ice margins. This ecological difference has important implications on the distribution of genetic diversity within species. In the former scenario, extant populations close to, but outside ice boundaries should be highly divergent due to founder effects from the expansion source. Additionally, genetic diversity should be a gradually declining function of distance from refugia as a consequence of successive founder events (Petit et al., 2003). In the latter scenario, however, these differences are blurred due to the existence of local refugia near the ice boundary that acted as multiple sources for recolonization into formerly glaciated areas after deglaciation. These leading-edge populations may have persisted during glacial periods as a metapopulation, and, thus should reflect low levels of interpopulation divergence due to high gene flow among them. As a result, and as found in this study as well as in European woody species (Petit et al., 2003), genetic diversity was highest within secondary contact areas. In addition, genetic diversity declined towards more persistent (in our case coastal, in other cases lower latitude) long-term refugia and towards newly colonized (in our case inland, in others higher latitude) populations located on deglaciated terrain. Forest persistence in refugia during the LGM in Tierra del Fuego has been puzzling. Although fossil records are abundant, some limitations exist in delimiting Nothofagus species based on pollen. This is mainly due to the fact that the three Nothofagus species inhabiting austral latitudes in South America belong to subgenus Nothofagus and thus share the same fusca (b) pollen type (Hill, 2001). This makes species-specific biogeographical reconstructions difficult. Therefore our findings shed light on migration rates of Nothofagus pumilio in relation to past climate changes. Most palynological records from glaciated areas show that the development of Nothofagus woodland started by 17,000 cal yr BP (Harberton: Markgraf and Huber, 2010), after 15,000 cal yr BP (Caleta Robalo and Ushuaia, Heusser, 1989, 1995) in southern Tierra del Fuego. This temporal pattern is consistent with the existence of refugia in local periglacial environments as revealed in this study and, by no means, reflects long-distance Nothofagus glacial migration to lower latitudes, north to 43°S, and subsequent southward migration following deglaciation a scenario that should imply tree migration rates of N0.5–1 km yr− 1. Such rates would by far exceed expectations. This is particularly true for N. pumilio which is known to be poorly dispersed (effective dispersal b20 m; Cuevas, 2000) only regenerating episodically in relation to a 7–8 yr masting cycle (Cuevas, 1999), which would further slow potential migration rates. Likewise, independent genetic data suggest migration rates that are three orders of magnitude lower

Author's personal copy A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

(0.5–1 m yr− 1; Mathiasen and Premoli, 2010) than those implied by long-distance postglacial migration. Palynologists have commonly interpreted low-frequency presence of pollen of forest taxa as unforested landscapes because background pollen is assumed to originate from distant sources (e.g. Sugita, 1994) rather than from sparse individuals. As a result, pollen records have largely underestimated the existence of refugia during times of low abundance such the glacial periods (local refugia). In addition, in some cases small or sparse populations may even become undetectable in pollen records (cryptic refugia; Anderson et al., 2006). Spatially explicit genetic data such presented here underscore the importance of isolated local survival areas and are useful for tracing the location of recolonization foci after climatic amelioration. Moreover, such reassessment of the location of refugia challenges the notion of extremely high Holocene tree migration rates (in the order of km yr− 1) derived from fossil evidence that invoke a lack of persistence in high latitude local refugia (Petit et al., 2008). Our results thus also demonstrate the limited capacity of trees to track possible antropogenic climate and/or land-use related changes. Supplementary materials related to this article can be found online at doi:10.1016/j.palaeo.2010.09.030. Acknowledgements We are most thankful to the personnel of the Centro Austral de Investigaciones Científicas-CONICET from Tierra del Fuego and our friend G. Lovrich for providing logistical support. We thank V. Markgraf and an anonymous reviewer for insightful comments on our manuscript, and J. Rabassa and A. Coronato for early discussions on glacial history of austral Patagonia. Funding was provided by BIOCORES European Union Contract ICA4-CT-2001-10095, Agencia de Promoción Científica PICT 25833, PIP CONICET 5066, and Universidad Nacional del Comahue B126. Molecular analyses were developed as part of a Fulbright fellowship to ACP. We are grateful to Prof. J.B. Mitton at University of Colorado, Boulder for sharing his expertise on cpDNA primer optimization and sequencing. References Acosta, M.C., Premoli, A.C., 2010. Evidence of chloroplast capture in South American Nothofagus (subgenus Nothofagus, Nothofagaceae). Molecular Phylogenetics and evolution 54, 235–242. Acosta, M.C., Mathiasen, P., Premoli, A.C., 2009. Filogeografía de Nothofagus antarctica (Nothofagaceae): XXXII Jornadas Argentinas de Botánica, Huerta Grande, Córdoba, Argentina. 5–7 October. Anderson, L.L., Hu, F.S., Nelson, D.M., Petit, R.J., Paige, K.N., 2006. Ice-age endurance: DNA evidence of a white spruce refugium in Alaska. Proceedings of the National Academy of Sciences of the USA 103, 12447–12450. Ashworth, A.C., Markgraf, V., 1989. Climate of the Chilean channels between 11, 000 to 10, 000 yr B.P. based on fossil beetle and pollen analyses. Revista Chilena de Historia Natural 62, 61–74. Ashworth, A.C., Markgraf, V., Villagran, C., 1991. Late Quaternary climatic history of the Chilean Channels based on fossil pollen and beetle analyses. Journal of Quaternary Science 16, 279–290. Auer, V., 1956. The Pleistocene of Fuegopatagonia, part I: The ice and interglacial ages: Annales Academiae Scientarum Fennicae A III 45, pp. 1–226. BIOCORES: Biodiversity conservation, restoration and sustainable use in fragmented forest landscapes, 2003. First Annual Report INCO, International Scientific Cooperation Project Contract number: ICA4-CT-2001-10095 http://www.unep-wcmc.org/forest/ BIOCORES. Bujalesky, G.G., Isla, F.I., 2006. Depósitos cuaternarios de la costa atlántica fueguina, entre Cabo Peñas y Ewan. Revista de la Asociación Geológica Argentina 61, 81–92. Caldenius, C.C., 1932. Las glaciaciones cuaternarias en la Patagonia y Tierra del Fuego. Geografiska Annaler 14, 1–164. Carnaval, A.C., Moritz, C., 2008. Historical climate modelling predicts patterns of current biodiversity in the Brazilian Atlantic forest. Journal of Biogeography 35, 1187–1201. Cavalli-Sforza, L.L., Edwards, A.W.F., 1967. Phylogenetic analysis: models and estimation procedures. The American Journal of Human Genetics 19, 233–257. Clapperton, Ch., 1993. Quaternary Geology and Geomorphology of South America. Elsevier, Amsterdam. Clauzet, G., Wainer, I., Bicego, M., 2008. Validating NCAR-CCSM last glacial maximum sea surface temperature in the tropical and South Atlantic with proxy-data. Palaeogeography, Palaeoclimatology, Palaeoecology 267, 153–160.

255

Collado, L., 2001. Los bosques de Tierra del Fuego. Análisis de su estratificación mediante imágenes satelitales para el inventario forestal de la provincia. Multequina 10, 1–15. Collins, W.D., Blackmon, M., Bitz, C., et al., 2004. The community climate system model: CCSM3. Journal of Climate 19, 2122–2143. CONAF, CONAMA, BIRF, Universidad Austral de Chile, Pontificia Universidad Católica de Chile, Universidad Católica de Temuco. 1999. Catastro y Evaluación de los Recursos Vegetacionales Nativos de Chile. Informe Nacional con Variables Ambientales. Santiago, Chile. Coronato, A., Salemme, M., Rabassa, J., 1999. Paleoenvironmental conditions during the early peopling of Southernmost South America (Late Glacial–Early Holocene, 14– 8 ka BP). Quaternary International 53 (54), 77–92. Coronato, A.M.J., Martínez, O., Rabassa, J., 2004. Glaciations in Argentine Patagonia, Southern South America, in Ehlers, J., Gibbard, P. (Eds.), Quaternary Glaciations: Extent and chronology. Part III: South America, Asia, Africa, Australia and Antarctica. Quaternary Book Series, Elsevier, Amsterdam, pp. 49–67. Cuevas, J.G., 1999. Dinámica poblacional en el límite arbóreo superior de Nothofagus pumilio (Poepp. et Endl.) Krasser en Tierra del Fuego, Chile. PhD Thesis, Facultad de Ciencias, Universidad de Chile, Santiago, Chile. Cuevas, J.G., 2000. Tree recruitment at the Nothofagus pumilio alpine timberline in Tierra del Fuego, Chile. Journal of Ecology 88, 840–855. Eeley, H.A.C., Lawes, M.J., Piper, S.E., 1999. The influence of climate change on the distribution of indigenous forest in KwaZulu–Natal, South Africa. Journal of Biogeography 26, 595–617. Evanno, G., Regnaut, S., Goudet, J., 2005. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14, 2611–2620. Fielding, A.H., Bell, J.F., 1997. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environmental Conservation 24, 38–49. Glasser, N.F., Jansson, K.N., Harrison, S., Kleman, J., 2008. The glacial geomorphology and Pleistocene history of South America between 38°S and 56°S. Quaternary Science Reviews 27, 365–390. Goudet, J., 2002. FSTAT A program to estimate and test gene diversities and fixation indices version 2.9.3.2. Dorigny, Switzerland: Université de Lausanne. Available at http://www2.unil.ch/izea/softwares/fstat.html. Guisan, A., Thuiller, W., 2005. Predicting species distribution: offering more than simple habitat models. Ecology Letters 8, 993–1009. Heusser, C.J., 1989. Late Quaternary vegetation and climate of southern Tierra del Fuego. Quaternary Research 31, 396–406. Heusser, C.J., 1995. Three Late Quaternary pollen diagrams from southern Patagonia and their paleoecological implications. Palaeogeography, Palaeoclimatology, Palaeoecology 118, 1–24. Heusser, C.J., Heusser, L.E., Lowell, T.V., Moreira, M.A., Moreira, M.S., 2000. Deglacial palaeoclimate at Puerto del Hambre, subantarctic Patagonia, Chile. Journal of Quaternary Science 15, 101–114. Hewitt, G.M., 2000. The genetic legacy of the Quaternary ice ages. Nature 405, 907–913. Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., Jarvis, A., 2005. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25, 1965–1978. Hill, R.S., 2001. Biogeography, evolution and palaeoecology of Nothofagus (Nothofagaceae): the contribution of the fossil record. Australian Journal of Botany 49, 321–332. Isla, F., Schnack, E., 1995. Submerged moraines offshore northern Tierra del Fuego, Argentina. Quaternary of South America and Antarctic Peninsula 9, 205–222. Jakob, S.S., Martinez-Meyer, E., Blattner, F.R., 2009. Phylogeographic analyses and paleodistribution modeling indicate Pleistocene in situ survival of Hordeum species (Poaceae) in southern Patagonia without genetic or spatial restriction. Molecular Biology and Evolution 26, 907–923. Jones, R.C., Vaillancourt, R.E., Jordan, G.J., 2004. Microsatellites for use in Nothofagus cunninghamii (Nothofagaceae) and related species. Molecular Ecology Notes 4, 14–16. Kaplan, M.R., Fogwilla, C.J., Sugden, D.E., Hulton, N.R.J., Kubik, P.W., Freeman, S.P.H.T., 2008. Southern Patagonian glacial chronology for the Last Glacial period and implications for Southern Ocean climate. Quaternary Science Reviews 27, 284–294. Kirkpatrick, J.B., Fowler, M., 1998. Locating likely glacial forest refugia in Tasmania using palynological and ecological information to test alternative climatic models. Biological Conservation 85, 171–182. Kitzberger, T., Souto, C., Arbetman, M., Mathiasen, P., Premoli, A.C., 2010. Niche Model and Genetics Show Opposite Responses to Glaciations in Patagonian Trees with Contrasting Cold Tolerances: VI Southern Connection Congress. Bariloche, Argentina, 15–19 February. Konnert, M., Bergmann, F., 1995. The geographical distribution of genetic variation of silver fir (Abies alba) in relation to its migration history. Plant Systematics and Evolution 196, 19–30. Lara, A., Rutherford, P., Montory, C., Bran, D., Pérez, A., Clayton, S., Ayesa, J., Barrios, D., Gross, M., Iglesias, G., 2000. Vegetación de la Eco-región de los Bosques Valdivianos Escala 1:500.000 Informe Final Proyecto Binacional Chile–Argentina UACh - INTAAPN– FVSA: Boletín Técnico Fundación Vida Silvestre Argentina, Nº 51. Li, C.C., Horvitz, D.G., 1953. Some methods of estimating the inbreeding coefficient. American Journal of Human Genetics 5, 107–117. Mancini, M.V., Prieto, A.R., Paez, M.M., Schäbitz, F., 2008. In: Rabassa, J. (Ed.), Late Quaternary vegetation and climate of Patagonia. : Late Cenozoic of Patagonia and Tierra del Fuego. Cap, vol. 17. Elsevier, pp. 351–367. Marchelli, P., Caron, H., Azpilicueta, M.M., Gallo, L.A., 2008. Primer Note: A new set of highly polymorphic nuclear microsatellite markers for Nothofagus nervosa and related South American species. Silvae Genetica 57, 83–85.

Author's personal copy 256

A.C. Premoli et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 298 (2010) 247–256

Markgraf, V., 1993. Paleoenvironments and paleoclimates in Tierra del Fuego and southernmost Patagonia, South America. Palaeogeography, Palaeoclimatology, Palaeoecology 102, 53–68. Markgraf, V., Huber, U.M., 2010. Late and postglacial vegetation and fire history in Southern Patagonia and Tierra del Fuego. Palaeogeography, Palaeoclimatology, Palaeoecology 297, 351–366. Markgraf, V., McGlone, M., Hope, G., 1995. Neogene paleoenvironmental and paleoclimatic change in southern temperate ecosystems — a southern perspective. Trends in Ecology and Evolution 10, 143–147. Martínez-Meyer, E., Peterson, A.T., 2006. Conservatism of ecological niche characteristics in North American plant species over the Pleistocene-to-Recent transition. Journal of Biogeography 33, 1779–1789. Mathiasen, P., 2010. Variación y estructura genética en Nothofagus pumilio (Poepp. et Endle.) Krasser ‘lenga’ a lo largo de diferentes gradientes ambientales. PhD Thesis, Centro Regional Universitario Bariloche, Universidad Nacional del Comahue, Bariloche, Rio Negro, Argentina. Mathiasen, P., Premoli, A.C., 2010. Out in the cold: genetic variation of Nothofagus pumilio (Nothofagaceae) provides evidence for latitudinally distinct evolutionary histories in austral South America. Molecular Ecology 19, 371–385. McCulloch, R.D., Davies, S.J., 2001. Late-glacial and Holocene palaeoenvironmental change in the central Strait of Magellan, southern Patagonia. Palaeogeography, Palaeoclimatology, Palaeoecology 173, 143–173. Meglioli, A., Evenson, E., Zeiler, D., Rabassa, J., 1990. Cronología absoluta y relativa de depósitos glaciarios de tierra del Fuego, Argentina y Chile: XI Congreso Geológico Argentino, 2, pp. 457–460. Mercer, J.H., 1983. Cenozoic glaciation in the Southern Hemisphere. Annual Review of Earth and Planetary Sciences 11, 99–132. Nei, M., 1972. Genetic distance between populations. American Naturalist 106, 283–292. Nei, M., 1973. Analysis of gene diversity in subdivided populations. Proceedings of the National Academy of Sciences of the USA 70, 3321–3323. Nei, M., 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 76, 379–390. Petit, R.J., Aguinagalde, I., de Beaulieu, J.L., Bittkau, C., Brewer, S., Cheddadi, R., Ennos, R., Fineschi, S., Grivet, D., Lascoux, M., Mohanty, A., Müller-Starck, G., DemesureMusch, B., Palmé, A., Martín, J.P., Rendell, S., Vendramin, G.G., 2003. Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300, 1563–1565. Petit, R.J., Hu, F.S., Dick, C.W., 2008. Forests of the past: a window to future changes. Science 320, 1450–1451. Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of species geographic distributions. Ecological Modelling 190, 231–259. Porter, Ch., 1990. Character and ages of Pleistocene drifts in a transect across the Strait of Magellan. Quaternary of South America and Antarctic Peninsula 7, 35–50. Premoli, A.C., 1996. Allozyme polymorphisms, outcrossing rates, and hybridization of South American Nothofagus. Genetica 97, 55–64. Premoli, A.C., 1998. The use of genetic markers to conserve endangered species and to design protected areas of more widespread species, in: International Foundation for Science (Ed.), Proceedings International Workshop “Recent Advances in Biotechnology for Tree Conservation and Management”. Universidade Federal de Santa Catarina. Florianópolis, Santa Catarina, Brasil, pp. 157–171. Premoli, A.C., 2003. Isozyme polymorphisms provide evidence of clinal variation with elevation in Nothofagus pumilio. Journal of Heredity 94, 218–226. Premoli, A.C., Kitzberger, T., 2005. Regeneration mode affects spatial genetic structure of Nothofagus dombeyi forests in northwestern Patagonia. Molecular Ecology 14, 2319–2329. Premoli, A.C., Steinke, L., 2008. Genetics of sprouting: effects of long-term persistance in fire-prone ecosystems. Molecular Ecology 17, 3827–3835. Premoli, A.C., Kitzberger, T., Veblen, T., 2000. Isozyme variation and recent biogeographical history of the long-lived conifer Fitzroya cupressoides. Journal of Biogeography 27, 251–260. Premoli, A.C., Souto, C.P., Rovere, A.E., Allnut, T.R., Newton, A.C., 2002. Patterns of isozyme variation as indicators of biogeographic history in Pilgerodendron uviferum (D. Don) Florín. Diversity and Distributions 8, 57–66.

Pritchard, J.K., Stephens, M., Donnelly, P., 2000. Inference of population structure using multilocus genotype data. Genetics 155, 945–959. Quiroga, M.P., Premoli, A.C., 2010. Genetic structure of Podocarpus nubigena (Podocarpaceae) provides evidence of Quaternary and ancient historical events. Palaeogeography, Palaeoclimatology, Palaeoecology 285, 186–193. Quiroga, P., Vidal Russel, R., Premoli, A.C., 2005. Evidencia morfológica e isoenzimática de hibridación natural entre Nothofagus antarctica y N. pumilio en el noroeste Patagónico. Bosque 26, 25–32. Rabassa, J., 2008. Late Cenozoic glaciations in Patagonia and Tierra del Fuego. In: Rabassa, J. (Ed.), Late Cenozoic of Patagonia and Tierra del Fuego. : Developments in Quaternary Science, vol.11. Elsevier. Rabassa, J., Coronato, A., Bujalesky, G., Salemme, M., Roig, C., Meglioli, A., Heusser, C., Gordillo, S., Roig, F., Borromei, A., Quattrocchio, M., 2000. Quaternary of Tierra del Fuego, Southernmost South America: an updated review. Quaternary International 68–71, 217–240. Rabassa, J., Coronato, A.M., Salemme, M., 2005. Chronology of the Late Cenozoic Patagonian glaciations and their correlation with biostratigraphic units of the Pampean region (Argentina). Journal of South American Earth Sciences 20, 81–103. SAyDS Secretaría de Ambiente y Desarrollo Sustentable, 2005. Primer Inventario Nacional de Bosques Nativos. Buenos Aires, Argentina. http://www.ambiente.gov. ar/archivos/web/UMSEF/File/PINBN/informe_nacional_pinbn.pdf. Souto, C., Premoli, A.C., 2007. Genetic variation in the widespread Proteaceae Embothrium coccineum (Proteaceae) endemic to Patagonia: effects of phylogeny and historical events. Australian Journal of Botany 55, 809–817. Sugden, D.E., Bentley, M.J., Fogwill, C.J., Hulton, N.R.J., McCulloch, R.D., Purves, R.S., 2005. Late-glacial glacier events in southernmost South America: a blend of ‘Northern’ and ‘Southern’ hemispheric climate signals? Geografiska Annaler 87A, 273–288. Sugita, S., 1994. Pollen representation of vegetation in Quaternary sediments: theory and method in patchy vegetation. Journal of Ecology 82, 881–897. Svenning, J.C., Normand, S., Kageyama, M., 2008. Glacial refugia of temperate trees in Europe: insights from species distribution modelling. Journal of Ecology 96, 1117–1127. Waltari, E., Hijmans, R.J., Peterson, A.T., Nyári, A.S., Perkins, S.L., Guralnick, R.P., 2007. Locating Pleistocene refugia: comparing phylogeographic and ecological niche model predictions. PLoS One 2 (7), e563. Weir, B.S., Cockerham, C.C., 1984. Estimating F-statistics for the análisis of population structure. Evolution 38, 1358–1370. Wiens, J.J., 2000. Reconstructing phylogenies from allozyme data: comparing method performance with congruence. Biology Journal of the Linnean Society 70, 613–632. Wiens, J.J., Graham, C.H., 2005. Niche conservatism: integrating evolution, ecology, and conservation biology. Annual Review of Ecology and Systematics 36, 519–539. Williams, J.E., 1991. Biogeographic patterns of three sub-alpine eucalypts in south-east Australia with special reference to Eucalyptus pauciflora Sieb. Ex Spreng. Journal of Biogeography 18, 223–230. Workman, P.L., Niswander, J.D., 1970. Population studies on Southwestern Indians.II. Local genetic differentiation in the Papago. American Journal of Human Genetics 22, 24–49. Worth, J.R.P., Jordan, G.J., McKinnon, G.E., Vaillancourt, R.E., 2009. The major Australian cool temperate rainforest tree Nothofagus cunninghamii withstood Pleistocene glacial aridity within multiple regions: evidence from the chloroplast. New Phytologist 182, 519–532. Wright, S., 1965. The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution 19, 395–420. Wright, S., 1978. Evolution and the Genetics of Populations, Vol. 4, Variability Within and Among Natural Populations. University of Chicago Press, Chicago.

Fig. S1. a) LGM-modern difference in mean annual temperature simulated by the NCAR-CCSM GCM. Superimposed are site locations of populations sampled for genetic analyses (this study green circles, mid-latitude magenta squares, and northerly populations yellow triangles). b) Multivariate cluster analysis based on isozyme data of Nothofagus pumilio populations along its entire range; coding corresponds to that described in pannel a).