ARTICLE IN PRESS

Dendrochronologia 28 (2010) 37–48

1125-7865/$ - se

doi:10.1016/j.de

�CorrespondPlant Biology

Highway, Craw

Tel.: +61 8 648

E-mail addr

www.elsevier.de/dendro

ORIGINAL ARTICLE

Dendroecological potential of Callitris preissii for dating historical fires in

semi-arid shrublands of southern Western Australia

Alison J. O’Donnella,�, Louise E. Cullena, W. Lachlan McCawb,c,Matthias M. Boera,c, Pauline F. Griersona,c

aEcosystems Research Group, School of Plant Biology (M090), University of Western Australia, 35 Stirling Highway,

Crawley, Western Australia 6009, AustraliabScience Division, Department of Environment and Conservation, Brain Street, Manjimup, Western Australia 6258, AustraliacBushfire Co-operative Research Centre, Level 5, 340 Albert Street, East Melbourne, Victoria 3002, Australia

Received 13 May 2008; accepted 16 January 2009

Abstract

Historical fire regimes in the semi-arid shrublands of southern Western Australia are poorly understood, largelyowing to a lack of quantitative historical data. We sought to determine the dendroecological potential of fire-sensitiveCallitris preissii Miq. trees to date historical fires and extend the length of fire-history data available from remotelysensed imagery. We sampled C. preissii trees from known fire areas in the Lake Johnston region in southern WesternAustralia. Our objective was to assess the capacity to date historical fires using stand establishment date as a proxymeasure of time since fire. We measured stem basal diameter and height and collected stem sections of C. preissii

trees and saplings from five areas that were burnt on known dates between 1974 and 2001. We also sampled older trees(435 years), which were used to create a master chronology to assist with dating of seedlings and saplings. Tree agecould not be reliably estimated from stem basal diameter and tree height, with 95% prediction intervals of more than17 years. However, we were able to successfully determine tree age and develop a ring-width chronology using standarddendrochronological techniques. The cross-dated chronology showed a relatively high inter-series correlation in ringwidth (r=0.63) indicating consistency in growth rate among samples and sites, while mean sensitivity (0.39) signifiedhigh inter-annual variability in ring width. The age structure of C. preissii stands revealed consistent recruitment within1 year of fire occurrence and maximum intra-stand variation in tree age of 4 years. Our results confirm that C. preissii

has significant dendroecological potential to accurately date past fire events and that this approach will assist inextending fire-history records beyond recent decades for much of southern semi-arid Australia.& 2009 Elsevier GmbH. All rights reserved.

Keywords: Recruitment; Tree-ring analysis; Dendrochronology; Allometry; Tree-age estimation

e front matter & 2009 Elsevier GmbH. All rights reserved.

ndro.2009.01.002

ing author at: Ecosystems Research Group, School of

(M090), University of Western Australia, 35 Stirling

ley, Western Australia 6009, Australia.

8 7923; fax: +61 8 6488 7925.

ess: odonnell@graduate.uwa.edu.au (A.J. O’Donnell).

Introduction

Extensive wildfires are natural and common occur-rences in many landscapes worldwide and play animportant role as modifiers of ecosystem structure anddynamics (Bond and Keeley, 2005; Thonicke et al.,2001). However, historical fire regimes are poorly

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–4838

understood in many ecosystems, particularly in semi-arid and arid biomes of the southern hemisphere, largelyowing to a lack of quantitative data and limited long-term records. Where historical records are lacking,dendrochronological techniques have often been usedto reconstruct spatial and temporal attributes of past fireevents (Beniston, 2002; Pearson and Searson, 2002).Dendrochronology-based reconstructions of fire his-tories usually rely on distinguishing cambial fire scarsin the trunks of ‘fire-tolerant’ tree species and analysingtheir distribution within and among trees to estimatethe date of fire events and their extent within thelandscape (Dieterich and Swetnam, 1984; Niklasson andGranstrom, 2000; Beaty and Taylor, 2001; Heyerdahl etal., 2001; Rollins et al., 2004). However, fire-scarredtrees are not always available, particularly in environ-ments where fires tend to be stand-replacing and treesare killed outright (e.g. Bergeron, 1991; Kipfmueller andBaker, 1998; Keeley, 2006). In these situations, den-drochronological techniques may be used to examinestand age structure and establishment dates as alter-native evidence of fire occurrence and a proxy measureof the age of the last stand-replacing fire (Brown et al.,1999; Floyd et al., 2000; Grau and Veblen, 2000).

Several difficulties have so far limited the use of fire-sensitive tree species for reconstructing fire events. First,the lag in tree establishment following disturbance maybe highly variable both within and among stands; hence,establishment dates may represent only the minimumpossible age of the disturbance. Second, samplescollected at heights above the base of the stem maynot include all growth rings, since it may take severalyears for a tree to reach sampling height. Estimates oftree age from samples taken above the base of the stemmay therefore provide only a minimum potential age(Ogden, 1981; Duncan, 1989; Villalba and Veblen, 1997;Wong and Lertzman, 2001). However, lags in standestablishment following a disturbance can be estimatedby calibrating stand age against the date of knowndisturbance events, and the uncertainty of tree-ageestimates at sample height can be reduced by calibratingagainst tree-age estimates from the base of the stem oralternatively, by taking samples from the base of thestem where all years of growth are likely to berepresented. Therefore, with calibration, fire-sensitivetree species may provide a useful and accurate datasource to confirm and date historical fire occurrence inregions where fire-tolerant (fire-scarring) tree species areunavailable.

Over the last several decades, fire ecology studies haveused remote-sensing techniques to reconstruct historicalfire regimes (e.g. Minnich, 1983; Minnich and Chou,1997; Kadmon and Harari-Kremer, 1999; Minnichet al., 2000; Miller and Yool, 2002). While remotelysensed data can be used to accurately reconstruct spatialattributes of fire, the ability to determine temporal

attributes (i.e. fire dates) is limited beyond the lastseveral decades owing to large temporal gaps betweenimage dates prior to the beginning of satellite imagery.Dendrochronological techniques may be used in con-junction with remotely sensed imagery to add fire datesto spatial fire maps and extend the temporal length offire-history data beyond the most recent few decades.Furthermore, the use of remotely sensed images in fire-history research may extend the potential geographicrange of dendropyrochronology. Fire data derived usingremote-sensing techniques provides the ability to recon-struct spatio-temporal fire patterns for regions such asthe chaparral shrublands of California, USA (Zedler,1995; Keeley, 2006), subalpine forests of the RockyMountains, USA (Kipfmueller and Baker, 1998), andthe semi-arid shrublands of southern Australia (Bellet al., 1984) where stand-replacing fires are common andtherefore only time since fire data can be deriveddendrochronologically.

While dendrochronological techniques have beenwidely used to reconstruct historical fire regimes in thetemperate and boreal regions of Europe and NorthAmerica (e.g. Niklasson and Granstrom, 2000; Stephenset al., 2003; Moody et al., 2006), dendrochronology hasreceived relatively little attention in Australia (Pearsonand Searson, 2002). The adoption of dendrochronologyin Australia has been hindered by a lack of knowledgeregarding suitable tree species and several methodolo-gical problems (Ogden, 1981; Brookhouse, 2006). Todate, tree-ring research in Australia has includedinvestigations of Eucalyptus (Banks, 1987; Burrowset al., 1996; Argent et al., 2004; Brookhouse and Brack,2006), Toona (Heinrich and Banks, 2005) and severalconiferous genera: Araucaria; Agathis (Dunwiddieand LaMarche Jr., 1980b; Ash, 1983a); Phyllocladus;Athrotaxis (Dunwiddie and LaMarche Jr., 1980b);Lagarostrobus (Cook et al., 1991); and Callitris (Lange,1965; Pearman 1971; Ogden, 1978; Perlinksi, 1986;Cullen and Grierson, 2006, 2009). These studies havemostly focused on the dendrochronological character-istics of these species and the capacity to reconstructpast climates by analysing ring-width patterns.

Few Australian tree species have been examined fortheir dendroecological utility for reconstructing firehistories. Investigations of the utility of fire-tolerantspecies for dendroecological analysis have included themonocot grass tree, Xanthorrhoea preissii (e.g. Wardet al., 2001; Enright et al., 2005; Miller et al., 2007), anda small number of Eucalyptus species (e.g. McBrideand Lewis, 1984; Burrows et al., 1996; Brookhouse,2006). However, methodological difficulties associatedwith unclear ring definition, intra-annual bands andunreliable scar formation in eucalypts (McBride andLewis, 1984; Burrows et al., 1996; Brookhouse, 2006),and high frequencies of false positive and false negativefire records in Xanthorrhoea preissii (Miller et al., 2007)

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–48 39

imply that neither genus is reliable for dating past fireevents. Investigations of the utility of fire-sensitive(obligate-seeder) species to reconstruct fire histories inAustralia have been limited to several species of Banksia

(Lamont, 1985; Lamont and Barker, 1988; Jenkins et al.,2005) and Petrophile (Jenkins et al., 2005), where thenumber of growth whorls has been related to time sincefire with reasonable accuracy in young stands (o25years). However, this approach has limited capacity toextend fire records, as whorl formation may decline withincreasing time since fire and plant age (Lamont, 1985).Alternatively, the stand age structure of fire-killed trees,including Eucalyptus pauciflora in subalpine forests ofKosciuszko National Park, New South Wales (Banks,1987), and Acacia dealbata, Eucalyptus delegatensis andNothofagus cunninghamii in the forests of the CentralHighlands of Victoria (Simkin and Baker, 2008), hasshown potential for dating high intensity, stand-replacing historical fires, but the capacity of thesespecies to record a wider range of fire intensities hasnot been tested. Consequently, the application ofdendrochronological techniques to fire research inAustralia is still limited by methodological problemsand a lack of knowledge of suitable species.

Of the Australian tree species assessed to date, speciesfrom the genus Callitris have been the most intensivelystudied and have shown the most potential for use indendrochronology (Pearman, 1971; LaMarche Jr. et al.,1979; Ash, 1983b; Perlinksi, 1986; Pulsford, 1991; Cullenand Grierson, 2006, 2007, 2009). However, the den-droecological utility of Callitris species for fire-historyreconstruction has not been explored.

The genus Callitris is widely distributed throughoutAustralia but is most prominent in arid and semi-aridregions (Harris and Kirkpatrick, 1991; Bowman andHarris, 1995; Bradstock and Cohn, 2002). Callitris

species have persisted in arid and semi-arid fire-pronelandscapes intermixed with scleromorphic and pyro-genic taxa, despite being considered highly ‘susceptible’to fire (Harris and Kirkpatrick, 1991; Bradstock andCohn, 2002). Many Callitris species are regarded asextraordinarily drought-tolerant, yet they tend toexhibit disjunct distributions in semi-arid landscapes, apattern that is thought to be determined by fire(Bowman et al., 1988; Harris and Kirkpatrick, 1991;Bowman and Latz, 1993). Most Callitris species aregenerally killed by moderate to high intensity fires;hence, they are unlikely to preserve fire damage ascambial scars typically used to construct chronologies offire events (Bowman and Harris, 1995). However,Callitris species are also serotinous obligate seeders.While fire nearly always kills adult trees, it alsostimulates recruitment from prolific seed release (Harrisand Kirkpatrick, 1991; Read, 1995). Consequently,although the potential to reconstruct long chronologiesof past fires by analysing tree rings of Callitris may be

limited by their fire sensitivity, the demography anddistribution of Callitris stands may provide insight intopast fire events or fire regime variability.

An early study by Pearman (1971) suggested a potentialcorrelation between the date of fire occurrence and standestablishment (stand age) of C. preissii on Rottnest Islandnear Perth, Western Australia. More recently, McCaw(1996) used ring counts on Callitris verrucosa and Callitris

roei stems to estimate stand age and therefore fire age(time since fire) near Lake Johnston in the Goldfieldsregion of southern Western Australia, but did not attemptto verify the accuracy of the estimated stand age or fireage. While early investigations suggest a link between firehistory and the demography of Callitris stands, thereliability and accuracy of using stand age as a proxymeasure of time since fire remains unknown.

Our objective was to establish the potential of C. preissii

as a bio-indicator of fire occurrence by comparing standage structure with known fire history derived fromremotely sensed imagery. C. preissii has several character-istics pertinent to this study, namely: (1) it exhibits distinctgrowth rings, which are likely to be annual, based on thenature of the climate in the semi-arid zone (Cullen andGrierson, 2006, 2009; Sgherza, 2006); (2) it is potentiallylong-lived (up to 300–400 years old); (3) has very thin bark(o5 mm) and is therefore killed outright by most fires; (4)it possesses a poor capacity to resprout following firedamage, and (5) it retains its seed in canopy-borne conesfor several years following maturity, which are typicallyreleased following fire (Bowman and Panton, 1993;McCaw, 1996; Bradstock and Cohn, 2002; Bradstock etal., 2006). We sought to relate tree ages to known fire agesin order to assess whether stand establishment date is areliable proxy measure of time since fire. Our researchconsisted of three steps. First, we determined tree ageusing dendrochronology techniques. Second, we tested iftree age could be accurately predicted from allometricrelationships with tree height and/or basal stem diameteras a simpler alternative to dendrochronological analysis.Third, we compared stand age to known fire age todetermine the potential lag and intra-stand variability inestablishment times following a fire event.

Materials and methods

Study area

Our study focused on an extensive area of shrublandand low open woodland on the sandplains of the‘‘Lake Johnston’’ region in southern Western Australia(Australian Surveying and Land Information Group 1997)(Fig. 1). This region is remote and currently has nopermanent settlements, with vehicle access limited to twounsealed roads and a sparse network of tracks, many ofwhich are overgrown. The Lake Johnston region includes

ARTICLE IN PRESS

Fig. 1. Location of the Lake Johnston region in southern Western Australia. Dated fire areas are shown in grey tones and locations

of sampled Callitris preissii stands are indicated by white circles.

A.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–4840

ca. 1.5 million ha of eucalypt woodland, mallee andshrubland and is characterised by a semi-aridMediterranean climate with highly variable rainfall that isgenerally less than 300 mm annually (Australian Bureau ofMeteorology, 2008). Extensive lightning-caused wildfires arecommon during the summer months, and are thought toplay a major role in shaping vegetation patterns in theregion (Beard, 1976; Hopkins and Robinson, 1981). C.

preissii has a widespread distribution throughout the regionand tends to occur in stands of scattered trees in scrubheath, dominated by Allocasuarina, Banksia and Melaleuca

species (Beard, 1976).

Field data collection

We created a digital database of fires in ArcGIS 9.2(Environmental Systems Research Institute, 2006) by

mapping visible fire areas from full resolution Landsatsatellite imagery 1972–2005 (courtesy of the WesternAustralian Department of Environment and Conserva-tion) and a black and white aerial photo mosaic from 1958(Department of Lands and Surveys, Perth). Fires occur-ring after 1972 were dated by recording the firstappearance of fires in low-resolution Landsat previewimages available on the Australian Centre for RemoteSensing (2006) and United States Geological Survey EarthExplorer (2007) websites and using the image date as anindication of the fire date.

Field data collection focused on five areas burnt byfires of known dates: two fires in the summer of 1974/1975, one in January 1991, one in January 1994, andanother in February 2001. These fire events wereselected for sampling because they adequately representthe temporal and spatial range of fires that haveoccurred within the region since the early 1970s. Stands

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–48 41

of C. preissii occurring within these five fire areas wereidentified in the field and sampled for both dendrochro-nological and allometric analysis of tree age.

We tested for the effect of sample height on the accuracyof tree-age estimates to determine the most appropriateprotocol for sampling Callitris for dendropyrochronologi-cal analysis. We collected 20 cm lengths of stem from thebase (soil level) of 22 adult trees (more than 35 years old),which were subsequently sectioned at height intervals of2.5 cm into nine discs representing 0.0, 2.5, 5.0, 7.5, 10.0,12.5, 15.0, 17.5, and 20.0 cm in height from the base of thestem. The sections were cross-matched by visuallyexamining ring-width patterns. We compared the numberof rings between the base of the stem (0.0 cm) andsampling heights 2.5 to 20 cm for each tree. We found thatsamples collected from 15 to 20 cm above the base of thestem (typical coring height) could underestimate tree ageby up to 12 years and that the relationship betweensampling height and the number of absent rings was highlyvariable among trees, with a 95% prediction error of74.63 rings (Fig. 2). Consequently, we determined thatsamples collected from the base of the stem (soil level)would provide the most accurate estimates of tree age.

Within each of the five targeted fire areas, we cut basalstem sections from 10 to 15 individual C. preissii trees ina stand. Finding suitable stands (i.e. more than 50individuals) within the target fire areas was difficult asC. preissii is typically only associated with scrub-heathvegetation (Beard, 1976) and access to most of theregion is limited. We attempted to sample replicatedpopulations within each fire scar to assess the consis-tency of stand age structure across large distances (i.e.up to 30 km), but were unable to find or access viablereplicate populations within the 1991 and 2001 fire

Fig. 2. Number of rings missing from stem sections of Callitris

preissii collected from the Lake Johnston region in southern

Western Australia. The number of missing rings was calculated

as the difference between the observed number of rings at the

base of the stem and at sample heights of 2.5–20 cm. The

continuous black line represents the fitted linear model; the

dashed line represents the 95% prediction interval; n=17 trees.

areas. In total, ten stands were sampled, one from eachof the 1991 and 2001 fires, and two or three separatestands within the January 1994, and both 1974 fireareas. Overall, 96 individual C. preissii seedlings/saplings were sampled from the target fire areas andsuccessfully aged using dendrochronological techniques.A further 89 adult trees were sampled from long-unburnt areas (435 years) across the Lake Johnstonregion and used to develop a master chronology to assistwith cross-dating of seedlings and saplings. In total, 185C. preissii trees and saplings were sampled and agedusing dendrochronological techniques.

Dendrochronological analysis

Stem sections were sanded using successively finersand paper (40–600 grit) to form a smooth surface.Sections were analysed under a dissecting microscope at10� magnification and cross-dated by visually match-ing ring-width patterns both within and among sites(Stokes and Smiley, 1968). A recent study by Cullen andGrierson (2009) found that a C. columellaris chronologyfrom Lake Tay (within the Lake Johnston region) wasmost strongly correlated with autumn–winter (March–September) regional rainfall anomalies (r=0.70;po0.05; n=11). We used this relationship to aid thecross-dating of our C. preissii samples by visuallymatching ring-width patterns with autumn–winter rain-fall anomalies averaged from four weather stationssurrounding the Lake Johnston region (i.e. at Ra-vensthorpe, Salmon Gums, Norseman and Lake Grace)(Fig. 1). Ring widths were measured with an accuracyof 0.001 mm using the program MEASURE J2X(Voortech Consulting, 2004) and cross-dating waschecked using the quality control program COFECHA(Holmes, 1983). Tree age was determined by countingannual growth rings from the outermost ring (2006) tothe pith. The date assigned to the pith was assumed toequate to the germination date.

Because of the small number of rings, and hencesignature years that could be used for cross-dating, the 59samples collected from recent fires (post-1990) were visuallycross-dated, but not measured or checked with COFECHA.These samples were therefore not included in the ring-widthand correlation statistics. A ring-width chronology wasdeveloped using the remaining 39 ‘calibration’ samples(from the 1974 fires) and the 89 adult trees sampled fromlong-unburnt areas (435 years). In total, 128 samples werecross-dated, checked with COFECHA, and assessed fortheir dendrochronological characteristics.

Relationships between tree size and tree age

We measured the height and basal diameter ofall trees that were sampled for dendrochronological

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–4842

analysis of tree age. To determine the best model topredict tree age from basal diameter and tree height, wefitted linear, polynomial, exponential, logarithmic andpower functions. We determined that a linear functionwas the most suitable model, as the more complexmodels provided negligible improvement in fit (R2) androot mean square error (data not shown). We modelledthe relationship between tree age and basal diameter andtree height with simple and multiple linear functions.The models examined were:

y ¼ aþ bx1 ð1Þ

y ¼ aþ cx2 ð2Þ

y ¼ aþ bx1 þ cx2 ð3Þ

where y is the age of trees determined from tree-ringanalysis, x1 is the basal diameter (cm), x2 is the treeheight (m), a is the intercept and b and c are theregression coefficients. The linear model that best fit thedata was chosen by minimising the Akaike informationcriterion (AIC) and root mean square error (RMSE)and maximising the adjusted R2. Because we wereinterested in predicting tree age from basal diameter andtree height, we treated tree age as the dependentvariable. All analyses were undertaken using the SAS9.1 statistical software (SAS Institute Inc., 2006).

Calibration of dendrochronologically derived tree

age and fire age

We examined intra-stand variability in tree age andcalibrated for a potential lag in establishment followingfire by calculating the difference between stand estab-lishment (age of the oldest individual) determined usingdendrochronological techniques and fire age determinedfrom satellite imagery for each stand.

Fig. 3. Variation in (a) tree height, and (b) basal diameter with

age of Callitris preissii trees in the Lake Johnston region in

southern Western Australia. Solid lines indicate fitted linear

regression models, dashed lines represent 95% prediction

intervals, n=190 (including five seedlings collected from a 2003

fire that were not used for calibration or dendrochronological

analysis).

Results

Dendrochronological characteristics of C. preissii

A relatively high inter-series correlation in ring width(r=0.63) was obtained for the C. preissii chronologyfrom COFECHA, indicating consistency in growth rateamong samples and sites, some more than 100 km apart.The chronology also showed a relatively high meansensitivity value of 0.39 derived from COFECHA, whichsignifies high inter-annual variability in ring width(Fritts, 1976). Average ring width across all sampleswas 0.46 mm, but ranged from 0 mm, representing amissing ring in 2002 in some samples when annualrainfall was almost half of the 1902 to 2005 average of234 mm, to a maximum ring width of 3.41 mmcorresponding to 1992 when annual rainfall was double

the 1902 to 2005 average. Few missing or micro-ringswere recorded in C. preissii sections. Several false (intra-annual) bands were observed in most samples, but theywere generally distinguished by their incomplete bound-aries or accounted for during cross-dating.

Allometric estimation of tree age

The three linear regression models revealed that bothtree basal diameter and height were significantly relatedto tree age (po0.0001) (Fig. 3; Table 1). Model selectionbased on the Akaike information criterion (AIC) suggeststhat the multiple linear regression model based on bothtree height and basal diameter predictors (Eq. 3) wasslightly better than the simple linear models (Eqs. (1) and(2)) for predicting tree age (Table 1). However, all modelsshowed large root mean square errors (Table 1) and wide95% prediction intervals: 719.3 for the model based on

ARTICLE IN PRESS

Table 1. Regression parameters for linear and multiple linear regression models predicting the age of Callitris preissii trees in the

Lake Johnston region, southern Western Australia.

Model Eq. # a a(SE) b b(SE) c c(SE) R2Adj. RMS AIC

y=a+bx1 (1) 8.45 1.15 6.88 0.27 – – 0.78 8.95 864.18

y=a+cx2 (2) 2.52 1.48 – – 18.91 0.81 0.74 9.67 834.72

y=a+bx1+cx2 (3) 4.77 1.32 4.38 0.57 7.81 1.61 0.80 8.46 814.23

The predictive variable, x1 is basal diameter and x2 is tree height. SE is the standard error of the slope parameters (b and c) and intercept (a); RMS is

the root mean square error; AIC is the Akaike Information Criterion. n=190 (including five seedlings collected from a 2003 fire that were not used

for calibration or dendrochronological analysis).

Fig. 4. Distribution of pith dates (here, assumed to equate to establishment date) of Callitris preissii seedlings/saplings relative to the

known fire date for each of the sampled fire areas in the Lake Johnston region, Western Australia.

A.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–48 43

tree height, 717.9 years for the model based on basaldiameter, and 716.9 years for the model based on bothpredictors. This level of uncertainty limits the capacity ofthese regression models to accurately predict tree age andtherefore makes them unsuitable for dating historical fireevents across the region.

Fire-dating capacity of C. preissii

The year of stand establishment, here assumed tocorrespond to the pith date of the oldest individuals withinthe stand, was consistently equivalent to the known fireyear (Fig. 4). Variation in dendrochronologically derived

tree age within each stand was low. Between 85% and100% of trees sampled from each fire scar were dated tohave established within 2 years after the fire event, whilethe maximum post-fire recruitment lag was 4 years for allsampled stands (Fig. 4).

Discussion

Dendrochronological potential of C. preissii

Our results demonstrate that C. preissii can besuccessfully cross-dated with a high inter-series correlation

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–4844

(0.63) across a large study area, which indicates thatindividual trees (some more than 100 km apart) areresponding consistently to a regional growth limitingenvironmental variable with a strong annual cycle, mostlikely autumn–winter rainfall (Cullen and Grierson, 2009).C. preissii also showed a high mean sensitivity value (0.39)in the study region compared to standards for dendro-chronology (Fritts, 1976; Grissino-Mayer, 2001). Themean sensitivity value reported here is consistent withvalues reported by Sgherza (2006) of 0.37–0.43 forC. preissii at locations 100–200 km west of our study area.

The high inter-series correlation and mean sensitivityvalues reported here for C. preissii indicate a high levelof environmental stress and a strong climatic limit togrowth in the Lake Johnston region (Fritts, 1976).Environmental stress is also apparent in the relativelylow mean width of annual growth rings (0.46 mm). Ringwidths reported here for C. preissii from the LakeJohnston region are on average almost 10 timesnarrower than those reported by Pearman (1971) forC. preissii from Garden Island (i.e. between 2.0 and8.0 mm), which is located in a much higher rainfall zone(870 mm annually) 20 km south west of Perth off theWestern Australian coast.

Allometric estimation of tree age

Growth rates (ring widths) were highly variable amongtrees of similar age, making it impossible to accuratelypredict tree age from basal diameter or tree height. Thisvariability may reflect a high level of environmental(drought) stress, resulting in strong competition amongtrees within the same stand. Mature stands of C. preissii inthe Lake Johnston region appear to be scattered in nature(e.g. less than 10 stems per 1000 m2), while young standsthat have recently recruited after fire appear to berelatively dense (e.g. more than 100 stems per 1000 m2,in patches of up to 175 trees). It is likely that competitionamong saplings following episodic recruitment eventsresults in the death of some individuals and the subsequent‘thinning’ of the stand towards maturity, as well as largevariability in growth rate among remaining trees. In suchstands, it is unlikely that stem diameter and tree height willbe closely related to tree age (Ogden, 1981). Pearman(1971) also reported a high level of variability (200–300%)in stem diameter for C. preissii trees of similar age, whichhe attributed to differences in stand density and subse-quent competition within the stand. Several other studieshave also reported density-dependent growth variabilityand decoupling between tree size and age in Callitris stands(Horne and Robinson, 1987; Horne, 1990; Ross et al.,2008). The scattered distribution of mature Callitris standsand sparse densities of other shrubs and grasses withinstands has also been attributed to strong below-groundcompetition both among Callitris trees, and between

Callitris and other shrub species (Clayton-Greene andAshton, 1990; Harris et al., 2003).

Although several previous studies examining the agestructure of Callitris stands have utilised measurementsof stem diameter or tree height to estimate tree age (e.g.Harris and Kirkpatrick, 1991; Bowman and Panton,1993; Read, 1995), the highly variable growth ratesreported here and earlier by Pearman (1971) forC. preissii and elsewhere for other Callitris species(Ash, 1983b; Clayton-Greene and Ashton, 1990) in-dicate that allometric models based on tree size do notprovide accurate estimates of tree age and consequentlystand age in many cases and, in this instance, should notbe used to estimate time since the last fire.

Dating historical fires using C. preissii

C. preissii stands in the Lake Johnston region exhibita consistent and reliable recruitment response within 1year of fire and therefore have significant dendroecolo-gical potential for dating historical fire events. Ourresults confirm an earlier suggestion by Pearman (1971),who observed a potential correlation between the dateof past fires and the age structure of C. preissii standsnear Garden Island, Western Australia.

While C. preissii has shown significant potential as abio-indicator of fire occurrence, its use is ultimatelylimited by the longevity of individual trees. The fire-sensitive nature of the species and the fire-prone natureof the shrublands where it occurs, mean that evidence ofpast fires is readily lost in subsequent fires. Nevertheless,stands that have remained as unburnt patches aftersubsequent fires, and recently fire-killed stands may beutilised to extend the temporal length of data. Welocated several patches of C. preissii trees in areasthat have remained unburnt since at least 1940, withindividual trees aged as greater than 130 years old.C. columellaris from the same region has been recordedas reaching more than 350 years of age (Perlinksi, 1986;Cullen and Grierson, 2009).

Fire-tolerant species have been used extensively toreconstruct the frequency and season of historical fires(e.g. Brown and Sieg, 1999; Beaty and Taylor, 2001;Stephens et al., 2003; Guyette and Spetich, 2006).However, inherent problems associated with the biologyof cambial scarring mean that false negative records of fireevents (i.e. a fire-affected tree that does not form a cambialscar) can occur and result in the underestimation of actualfire frequency (Niklasson and Granstrom, 2000; Heyerdahlet al., 2001; Hessl et al., 2007; McEwan et al., 2007). Incontrast, fire-sensitive trees have not been widely utilised indendrochronology-based fire-history investigations, partlybecause they can only record the occurrence of the mostrecent fire (i.e. time since fire) and therefore provide lessinformation on historical fire regimes than fire-tolerant

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–48 45

species. However, fire-sensitive species are generally killedoutright by even low intensity fires, hence, they may haveutility for dating historical fires where they coexist withfire-tolerant species by potentially providing evidence oflower intensity fires that may not be recorded by fire-tolerant species. Fire-sensitive species that can be accu-rately aged may also be used in conjunction with remotelysensed imagery to generate fire-history information inregions where other fire records do not exist and fire-tolerant species are not available. For example, several fire-sensitive, serotinous, coniferous species that occur amongthe fire-prone chaparral shrublands of southern California(e.g. Cupressus arizonica, C. forbesii, and Pinus attenuata)and the Fynbos shrublands of southern Africa (e.g.Widdringtonia cedarbergensis) may have potential fordating historical fires (Dunwiddie and LaMarche Jr.,1980a; Midgley et al., 1995; Zedler, 1995; Keeley, 2006)and extending fire-history data in these regions.

Conclusion

This is the first study to report on a comprehensiveassessment of the dendroecological potential of Australiancypress pines (Callitris spp.) for reconstructing historic fireevents. Our results confirm that C. preissii trees can besuccessfully aged using standard dendrochronology tech-niques and that the age structure of naturally occurringstands can reveal reliable information about the timing ofhistorical fire events. Combined with the already estab-lished dendrochronological potential of other Callitris

species, our findings indicate that the dendropyrochrono-logical potential of other species within the genus shouldbe further investigated. The techniques used in this studymay be applied to C. preissii generally and potentially toother Callitris species elsewhere to validate and extend thelength of fire-history data available from remotely sensedimagery and provide a long-term perspective on thefunctioning of current and future fire regimes in a widerange of Australian landscapes.

Acknowledgements

This research is jointly funded by the Bushfire Co-operative Research Centre (An Australian GovernmentCRC Programme), and the Western Australian Depart-ment of Environment and Conservation. We are grate-ful to Jeremy Bougoure, Ciaran Sgherza, Gerald Pageand Chloe Flaherty for assistance in the field.

References

Argent, R.M., McMahon, T.A., Bowler, J.M., Finlayson,

B.L., 2004. The dendroecological potential of Eucalyptus

camaldulensis Dehnhardt (River Red Gum) from the

Barmah Forest, Victoria, Australia. Australian Geographi-

cal Studies 42, 89–102.

Ash, J., 1983a. Growth rings in Agathis robusta and Araucaria

cunninghamii from tropical Australia. Australian Journal of

Botany 31, 269–275.

Ash, J., 1983b. Tree rings in tropical Callitris macleayana F.

Muell. Australian Journal of Botany 31, 277–281.

Australian Bureau of Meteorology, 2008. Climate Data

Online. Retrieved 8 May 2008 from: /http://www.bom.

gov.au/climate/averages/S.

Australian Centre for Remote Sensing, 2006. Online Catalo-

gue. Retrieved December 2006 from: /https://acres.ga.gov.

au/intro.htmlS.

Australian Surveying and Land Information Group (AUS-

LIG), 1997. Lake Johnston, Western Australia. 1:250,000,

Australian Capital Territory.

Banks, J.C.G., 1987. Fire and stand histories in subalpine

forests on the Thredbo ski slopes, Kosciuszko National

Park, NSW, Australia. In: Jacoby, G.C., Horbeck, J.W.

(Eds.), International Symposium on Ecological Aspects of

Tree-Ring Analysis. US Department of Energy, pp.

163–174.

Beard, J.S., 1976. In: The Vegetation of the Boorabbin and

Lake Johnston Areas: Map and Explanatory Memoir.

1:250,000. Vegmap Publications, Perth, Western Australia.

Beaty, R.M., Taylor, A.H., 2001. Spatial and temporal

variation of fire regimes in a mixed conifer forest landscape,

Southern Cascades, California, USA. Journal of Biogeo-

graphy 28, 955–966.

Bell, D.T., Hopkins, A.J.M., Pate, J.S., 1984. Fire in the

Kwongan. In: Pate, J.S., Beard, J.S. (Eds.), Kwongan:

Plant Life of the Sandplain. University of Western

Australia Press, Nedlands, Western Australia, pp. 178–204.

Beniston, M., 2002. Climate modeling at various spatial and

temporal scales: where can dendrochronology help?.

Dendrochronologia 20, 117–131.

Bergeron, Y., 1991. The influence of island and mainland

lakeshore landscapes on Boreal forest fire regimes. Ecology

72, 1980–1992.

Bond, W.J., Keeley, J.E., 2005. Fire as a global ‘herbivore’: the

ecology and evolution of flammable ecosystems. Trends in

Ecology and Evolution 20, 387–394.

Bowman, D.M.J.S., Harris, S., 1995. Conifers of Australia’s

dry forests and open woodlands. In: Enright, N.J., Hill,

R.S. (Eds.), Ecology of the Southern Conifers. Melbourne

University Press, Carlton, Victoria, pp. 252–270.

Bowman, D.M.J.S., Latz, P.K., 1993. Ecology of Callitris

glaucophylla (Cupressaceae) on the MacDonnell Ranges,

central Australia. Australian Journal of Botany 41,

217–225.

Bowman, D.M.J.S., Panton, W.J., 1993. Decline of Callitris

intratropica R.T. Baker & H.G. Smith in the Northern

Territory: implications for pre- and post-European

colonisation fire regimes. Journal of Biogeography 20,

373–381.

Bowman, D.M.J.S., Wilson, B.A., Davis, G.W., 1988.

Response of Callitris intratropica R.T. Baker & H.G. Smith

to fire protection, Murgenella, northern Australia. Austra-

lian Journal of Ecology 13, 147–159.

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–4846

Bradstock, R.A., Bedward, M., Cohn, J.S., 2006. The

modelled effects of differing fire management strategies

on the conifer Callitris verrucosa within semi-arid mallee

vegetation in Australia. Journal of Applied Ecology 43,

281–292.

Bradstock, R.A., Cohn, J.S., 2002. Demographic character-

istics of mallee pine (Callitris verrucosa) in fire-prone mallee

communities of central New South Wales. Australian

Journal of Botany 50, 653–665.

Brookhouse, M., 2006. Eucalypt dendrochronology: past,

present and potential. Australian Journal of Botany 54,

435–449.

Brookhouse, M., Brack, C., 2006. Crossdating and analysis of

eucalypt tree rings exhibiting terminal and reverse late

wood. Trees 20, 767–781.

Brown, P.M., Kauffman, M.R., Shepperd, W.D., 1999. Long-

term, landscape patterns of past fire events in a montane

ponderosa pine forest in central Colorado. Landscape

Ecology 14, 513–532.

Brown, P.M., Sieg, C.H., 1999. Historical variability in fire at

the Ponderosa pine—Northern Great Plains prairie eco-

tone, southeastern Black Hills, South Dakota. Ecoscience

6, 539–547.

Burrows, N.D., Ward, B., Robinson, A.D., 1996. Jarrah forest

fire history from stem analysis and anthropological

evidence. Australian Forestry 58, 7–16.

Clayton-Greene, K.A., Ashton, D.H., 1990. The dynamics of

Callitris columellaris/Eucalyptus albens communities along

the Snowy River and its tributaries in south-eastern

Australia. Australian Journal of Botany 38, 402–432.

Cook, E., Bird, T., Peterson, M., Barbetti, M., Buckley, B.,

D’Arrigo, R., Francey, R., Tans, P., 1991. Climatic change

in Tasmania inferred from a 1089-year tree-ring chronology

of Huon Pine. Science 253, 1266–1268.

Cullen, L.E., Grierson, P.F., 2006. Is cellulose extraction

necessary for developing stable carbon and oxygen isotope

chronologies from Callitris glaucophylla? Palaeogeogra-

phy, Palaeoclimatology, Palaeoecology 236, 206–216.

Cullen, L.E., Grierson, P.F., 2007. A stable oxygen, but not

carbon, isotope chronology of Callitris columellaris reflects

recent climate change in north-western Australia. Climatic

Change 85, 213–229.

Cullen, L.E., Grierson, P.F., 2009. Multi-decadal scale

variability in autumn–winter rainfall in south-western

Australia since 1655 AD as reconstructed from tree rings

of Callitris columellaris (Cupressaceae). Climate Dynamics

33, 433–444.

Dieterich, J.H., Swetnam, T.W., 1984. Dendrochronology of a

fire-scarred ponderosa pine. Forest Science 30, 238–247.

Duncan, R.P., 1989. An evaluation of errors in tree age

estimates based on increment cores in Kahikatea (Dacry-

carpus dacrydioides). New Zealand Natural Sciences 16,

31–37.

Dunwiddie, P.W., LaMarche Jr., V.C., 1980a. A climatically

responsive tree-ring record from Widdringtonia cedarber-

gensis, Cape Province, South Africa. Nature 286,

796–797.

Dunwiddie, P.W., LaMarche Jr., V.C., 1980b. Dendrochro-

nological characteristics of some native Australian trees.

Australian Forestry 43, 124–135.

Enright, N.J., Lamont, B.B., Miller, B., 2005. Anomalies in

grasstree fire history reconstruction for south-western

Australian vegetation. Austral Ecology 30, 668–673.

Environmental Systems Research Institute Inc., 2006.

ArcGIS 9.2, Redlands, CA, USA.

Floyd, M.L., Romme, W.H., Hanna, D.D., 2000. Fire history

and vegetation pattern in Mesa Verde National Park,

Colorado, USA. Ecological Applications 10, 1666–1680.

Fritts, H.C., 1976. In: Tree Rings and Climate. Academic

Press, London 567pp.

Grau, H.R., Veblen, T.T., 2000. Rainfall variability, fire and

vegetation dynamics in neotropical montane ecosystems in

north-western Argentina. Journal of Biogeography 27,

1107–1121.

Grissino-Mayer, H.D., 2001. Evaluating cross-dating accu-

racy: a manual and tutorial for the computer program

COFECHA. Tree-Ring Research 57, 205–221.

Guyette, R.P., Spetich, M.A., 2006. Historic fire regime dynamics

and forcing factors in the Boston Mountains, Arkansas,

USA. Forest Ecology and Management 234, 293–304.

Harris, M.R., Lamb, D., Erskine, P.D., 2003. An investigation

into the possible inhibitory effects of White Cypress Pine

(Callitris glaucophylla) litter on the germination and growth

of associated ground cover species. Australian Journal of

Botany 51, 93–102.

Harris, S., Kirkpatrick, J.B., 1991. The distributions, dynamics

and ecological differentiation of Callitris species in Tasma-

nia. Australian Journal of Botany 39, 187–202.

Heinrich, I., Banks, J.C.G., 2005. Dendroclimatological

potential of the Australian red cedar. Australian Journal

of Botany 53, 21–32.

Hessl, A.E., Miller, J., Kernan, J., Keenum, D., McKenzie, D.,

2007. Mapping paleo-fire boundaries from binary point

data: comparing interpolation methods. The Professional

Geographer 59, 87–104.

Heyerdahl, E.K., Brubaker, L.B., Agee, J.K., 2001. Spatial

controls of historical fire regimes: a multiscale example

from the interior west, USA. Ecology 82, 660–678.

Holmes, R.L., 1983. Computer-assisted quality control in tree-

ring dating and measurement. Tree Ring Bulletin 43,

111–189.

Hopkins, A.J.M., Robinson, C.J., 1981. Fire induced structur-

al change in a Western Australian woodland. Australian

Journal of Ecology 6, 177–188.

Horne, R., 1990. Stand height response following variable

spacing of wheatfield white cypress pine regeneration in

New South Wales. Australian Forestry 53, 47–54.

Horne, R., Robinson, G., 1987. White cypress pine in NSW:

growth patterns and optimal thinning regimes for 60 to 80

year old stands. Australian Forestry 50, 216–223.

Jenkins, M.E., Morrison, D.A., Auld, T.D., 2005. Use of

growth characteristics for predicting plant age of three

obligate-seeder Proteaceae species. Australian Journal of

Botany 53, 101–108.

Kadmon, R., Harari-Kremer, R., 1999. Studying long-term

vegetation dynamics using digital processing of historical

aerial photographs. Remote Sensing of Environment 68,

164–176.

Keeley, J.E., 2006. South Coast Bioregion. In: Sugihara, N.G.,

Van Wagtendonk, J.W., Shaffer, K.E., Fites-Kaufman, J.,

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–48 47

Thode, A. (Eds.), Fire in California’s Ecosystems. Uni-

versity of California Press, Berkeley, CA, pp. 350–390.

Kipfmueller, K.F., Baker, W.L., 1998. A comparison of three

techniques to date stand-replacing fires in lodgepole pine

forests. Forest Ecology and Management 104, 171–177.

LaMarche Jr., V.C., Holmes, R.L., Dunwiddie, P.W., Drew, L.G.,

1979. Tree-ring chronologies of the Southern Hemisphere. vol.

4: Australia. Chronology Series V. Laboratory of Tree-Ring

Research, University of Arizona, Tucson, AZ 89pp.

Lamont, B.B., 1985. Fire responses of sclerophyll shrublands—

a population ecology approach with particular reference to

the genus Banksia. In: Ford, J.R. (Ed.), Fire Ecology and

Management of Western Australian Ecosystems: Proceedings

of a symposium held in Perth on 10–11 May 1985. Western

Australian Institute of Technology, Perth, Western Australia,

pp. 41–46.

Lamont, B.B., Barker, M.J., 1988. Seed bank dynamics of a

serotinous, fire-sensitive Banksia species. Australian Jour-

nal of Botany 36, 193–203.

Lange, R.T., 1965. Growth rings characteristics of an arid

zone conifer. Transactions of the Royal Society of South

Australia 89, 133–137.

McBride, J.R., Lewis, H.T., 1984. Occurrence of fire scars in

relation to the season and frequency of surface fires in

Eucalyptus forests of the Northern Territory, Australia.

Forest Science 30, 970–976.

McCaw, L., 1996. Fire patterns in semi-arid woodlands and

shrublands of southern Western Australia. Notes for a

seminar presented at CALM SOHQ, Como, Western

Australia.

McEwan, R.W., Hutchinson, T.F., Ford, R.D., McCarthy,

B.C., 2007. An experimental evaluation of fire history

reconstruction using dendrochronology in white oak

(Quercus alba). Canadian Journal of Forest Research 37,

806–816.

Midgley, G.F., Bond, W.J., Geldenhuys, C.J., 1995. The

ecology of Southern African conifers. In: Enright, N.J.,

Hill, R.S. (Eds.), Ecology of the Southern Conifers.

Melbourne University Press, Carlton, Victoria, Australia,

pp. 64–80.

Miller, B., Walshe, T., Enright, N.J., Lamont, B.B., 2007.

Error in the inference of fire history from grass trees.

Austral Ecology 32, 908–916.

Miller, J.D., Yool, S.R., 2002. Mapping forest post-fire canopy

consumption in several overstory types using multi-

temporal Landsat TM and ETM data. Remote Sensing of

Environment 82, 481–496.

Minnich, R.A., 1983. Fire mosaics in southern California and

northern Baja California. Science 219, 1287–1294.

Minnich, R.A., Barbour, M.G., Burk, J.H., Sosa-Ramirez, J.,

2000. Californian mixed-conifer forests under unmanaged

fire regimes in the Sierra San Pedro Martir, Baja California,

Mexico. Journal of Biogeography 27, 105–129.

Minnich, R.A., Chou, Y.H., 1997. Wildland fire patch

dynamics in the Chaparral of Southern California and

Northern Baja California. International Journal of Wild-

land Fire 7, 221–248.

Moody, T.J., Fites-Kaufman, A., Stephens, S.L., 2006. Fire

history and climate influences from forests in the northern

Sierra Nevada, USA. Fire Ecology 2, 115–141.

Niklasson, M., Granstrom, A., 2000. Number and sizes of

fires—long term spatially explicit fire history in a Swedish

boreal landscape. Ecology 81, 1484–1499.

Ogden, J., 1978. On the dendrochronological potential

of Australian trees. Australian Journal of Ecology 3,

339–356.

Ogden, J., 1981. Dedrochronological studies and the determi-

nation of tree ages in the Australian tropics. Journal of

Biogeography 8, 405–420.

Pearman, G.I., 1971. An exploratory investigation of the

growth rings of Callitris preissii trees from Garden

Island and Naval Base. Western Australian Naturalist 12,

12–17.

Pearson, S.G., Searson, M.J., 2002. High-resolution data

from Australian trees. Australian Journal of Botany 50,

431–439.

Perlinksi, J.E., 1986. The dendrochronology of Callitris

collumellaris F. Muell in arid, sub-tropical continental

Western Australia. M.A. Dissertation, University of

Western Australia.

Pulsford, I.F., 1991. History of disturbances in the white

cypress pine (Callitris glaucophylla) forests of the

lower Snowy River Valley, Kosciusko National

Park. M.Sc. Dissertation, Australian National

University.

Read, J., 1995. Recruitment characteristics of the White

Cypress Pine (Callitris glaucophylla) in arid South Aus-

tralia. Rangeland Journal 17, 228–240.

Rollins, M.G., Keane, R.E., Parsons, R.A., 2004. Mapping

fuels and fire regimes using remote sensing, ecosystem

simulation, and gradient modelling. Ecological Applica-

tions 14, 75–95.

Ross, K.A., Bedward, M., Ellis, M.V., Deane, A., Simpson,

C.C., Bradstock, R.A., 2008. Modelling the dynamics of

white cypress pine Callitris glaucophylla woodlands in

inland south-eastern Australia. Ecological Modelling 211,

11–24.

SAS Institute Inc, 2006. SAS 9.1, Cary, NC, USA.

Sgherza, C., 2006. Tree-ring width and 13C chronologies of

Callitris in the south-eastern Wheatbelt of Western

Australia. B.Sc. Hons. Dissertation, University of Western

Australia.

Simkin, R., Baker, P.J., 2008. Disturbance history and stand

dynamics in tall open forest and riparian rainforest in the

Central Highlands of Victoria. Austral Ecology 33,

747–760.

Stephens, S.L., Skinner, C.N., Gill, S.J., 2003. Dendrochro-

nology-based fire history of Jeffrey pine—mixed conifer

forests in the Sierra San Pedro Martir, Mexico. Canadian

Journal of Forest Research 33, 1090–1101.

Stokes, M.A., Smiley, T.L., 1968. In: An Introduction to Tree-

Ring Dating. The University of Chicago Press, Chicago

73pp.

Thonicke, K., Venevsky, S., Sitch, S., Cramer, W., 2001. The

role of fire disturbance for global vegetation dynamics:

coupling Dynamic Global Vegetation Model. Global

Ecology and Biogeography 10, 661–667.

United States Geological Survey, 2007. Earth Explorer.

Retrieved December 2006 from: /http://edcsns17.cr.usgs.

gov/EarthExplorer/S.

ARTICLE IN PRESSA.J. O’Donnell et al. / Dendrochronologia 28 (2010) 37–4848

Villalba, R., Veblen, T.T., 1997. Improving estimates of total

tree ages based on increment core samples. Ecoscience 4,

534–542.

Voortech Consulting, 2004. Measure J2X3.2.1. Holderness,

NH, USA.

Ward, D.J., Lamont, B.B., Burrows, C.L., 2001. Grasstrees

reveal contrasting fire regimes in eucalypt forest before and

after European settlement of southwestern Australia.

Forest Ecology and Management 150, 323–329.

Wong, C.M., Lertzman, K.P., 2001. Errors in estimating tree

age: implications for studies of stand dynamics. Canadian

Journal of Forest Research 31, 1262–1271.

Zedler, P.H., 1995. Plant life history and dynamic specializa-

tion in the Chaparral/Coastal Sage shrub flora in southern

California. In: Arroyo, M.T.K., Zedler, P.H., Fox, M.D.

(Eds.), Ecology and Biogeography of Mediterranean

Ecosystems in Chile, California, and Australia. Springer-

Verlag Inc., New York.