Short communication

7000-year human legacy of elevation-dependent European fireregimes

B. Vanni!ere a, *, O. Blarquez b, D. Rius a, E. Doyen a, T. Brücher c, D. Colombaroli d,S. Connor e, A. Feurdean f, T. Hickler f, P. Kaltenrieder d, C. Lemmen g, B. Leys h, C. Massa i,J. Olofsson j

a Laboratoire Chrono-environnement e UMR CNRS 6249, Maison des Sciences de l'Homme et de l'Environnement e USR CNRS 3124, Universit"e BourgogneFranche-Comt"e, 25000 Besançon Cedex, Franceb D"epartement de G"eographie, Universit"e de Montr"eal, C.P. 6128 Succ. Centre Ville, H3C 3J7 Montr"eal (Qu"ebec), Canadac GEOMAR Helmholtz Centre for Ocean Research, Düsternbrooker Weg 20, 24105 Kiel, Germanyd Oeschger Centre for Climate Change Research & Institute of Plant Sciences, University of Bern, Altenbergrain 21, CH-3013 Bern, Switzerlande School of Earth, Atmosphere and Environment, Monash University, 3800 Clayton, VIC, Australiaf Biodiversity and Climate Research Centre (BiK-F) and Department of Physical Geography at Goethe-University Frankfurt, Senckenberganlage 25, D-60325Frankfurt Am Main, Germanyg Helmholtz Zentrum Geesthacht, Centre for Materials and Coastal Research GmbH, Institute for Coastal Research, Max-Planck-Str.1, 21502 Geesthacht,Germanyh Geography Department, Kansas State University, 118 Seaton Hall, 66506 Manhattan, KS, USAi Department of Earth and Environmental Sciences, Lehigh University, 1 West Packer Avenue, 18015-3001 Bethlehem, PA, USAj Department of Physical Geography and Ecosystem Science, LUND University, S€olvegatan 12, SE-223 62 LUND, Sweden

a r t i c l e i n f o

Article history:Received 21 August 2015Received in revised form13 November 2015Accepted 14 November 2015Available online 7 December 2015

Keywords:EuropeEcological trendsPaleofirePrehistoric land uses

a b s t r a c t

Variability in fire regime at the continental scale has primarily been attributed to climate change, oftenovershadowing the widely potential impact of human activities. However, human ignition modifies therhythm of fire episodes occurrence (fire frequency), whereas land use alters vegetation composition andfuel load, and thus the amount of biomass burned. It is unclear, however, whether and how humans haveexercised a significant influence over fire regimes at continental and millennial scales. Based on sedi-mentary charcoal records, we use new alternative estimate of fire frequency and biomass burned for thelast 16000 years (here after 16 ky) that we evaluate with outputs from climate, vegetation, land use andpopulation models. We find that pronounced regional-scale land use changes in southern Europe at thebeginning of the Neolithic (8e6 ky), during the Bronze Age (5e4 ky) and the medieval period (1 ky)caused a doubling of fire frequency compared to the Holocene average (the last 11.5 ky). Despiteanthropogenic influences, southern European biomass burned decreased from 7 ky, which is in line bothwith changes in orbital parameters leading climate cooling and also reductions in biomass availabilitybecause of land use. Our study underscores the role of elevation-dependent parameters, and particularlybiomass and land management, as major drivers of fire regime variability. Results attest a determinantanthropogenic driving-force on fire regime and a decrease in fire-carbon emissions since 7 ky inSouthern Europe.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

The lasting and pervasive influence of fire on land ecosystemshas generated great interest in understanding environmental

drivers and effects (Blarquez et al., 2010; Colombaroli et al., 2010;Leys et al., 2014a; Marlon et al., 2013; Pausas et al., 2008;Vanni!ere et al., 2008). This is particularly true in the context ofhuman-induced global change and potential impacts on ecosys-tems and carbon emissions (Kaplan et al., 2011; Schr€oter et al.,2005; van der Werf et al., 2013). For instance, paleoecologicaldata from southern-hemisphere temperate forests indicate thathuman control over fire regimes resulted in major and persistent* Corresponding author.

E-mail address: boris.vanniere@univ-fcomte.fr (B. Vanni!ere).

Contents lists available at ScienceDirect

Quaternary Science Reviews

journal homepage: www.elsevier .com/locate/quascirev

http://dx.doi.org/10.1016/j.quascirev.2015.11.0120277-3791/© 2015 Elsevier Ltd. All rights reserved.

Quaternary Science Reviews 132 (2016) 206e212

ecosystem transitions by increased landscape flammability(McWethy et al., 2013). Correlation analysis on charcoal and pollendata from central and southern Europe, for example, has providedclear evidence of human-set fires disturbance that have been amajor factor shaping interspecific relationships and vegetationchange from ca. 4.5 ka leading in particular to spread of grasslandsand scrubland, depletion of mesophytes forests, and extend of areasfor arable and pastoral farming (e.g. Tinner et al., 2005; Carri"onet al., 2010).

Continental-scale syntheses for Europe show a long-term in-crease in fire activity since the last Glacial Maximum, ca. 21 ky(Power et al., 2008). This trend has mainly been attributed toclimate change, foremost increases in temperature (Daniau et al.,2012). The use of fire by humans to open forests for agriculture orto manage pastures is clearly evident since the Neolithic(Colombaroli et al., 2008; Kaltenrieder et al., 2005, 2010; Surmelyet al., 2009; Vanni!ere and Martineau, 2005). Nonetheless, theimpact of human-induced fire on fire regime, biomass burning andfire-carbon emissions at regional to continental scale is largelyunresolved (Bowman et al., 2011; Brücher et al., 2014; Molinariet al., 2013; Marlon et al., 2013).

Most previous paleofire studies have looked at the temporalvariability of fire regime, but few have examined changes in spatialvariability including regional disparities or uniformity (Carcailletet al., 2009; Rius et al., 2011, Vanni!ere et al., 2011; Colombaroliet al., 2014; Blarquez et al., 2015). Fire regimes are mediated byseveral factors ranging from regional controls such as climate(temperature, precipitation and wind), fuel (load, connectivity andflammability) and ignition regimes (seasonality and location, nat-ural or anthropogenic) to local landscape controls (soil, slope,exposure). We hypothese that the local factors can be the pre-dominant causes of spatial heterogeneity of fire, whereas regionaldrivers may dominate occasionally, when they are for instancerelated to large-scale climate events (Seneci et al., 2013). One wayto identify past drivers of fire regimes at the regional scale is toidentify the key periods of homogenous regional patterns in firerecords, and link these to broad-scale causes or consequences(Vanni!ere et al., 2011).

Here, we selected 18 well-dated charcoal records (Fig. 1) to builda 16 ky reconstructions of biomass burned and fire frequency forsouthern Europe (see Materials and Methods; Table S1 and Fig. S1,S2 and S3). Then, we compared the charcoal time series withmodel

simulations of past human populations, vegetation and climaticchanges to explore if human activities have influenced the long-term fire regime trajectory and fire-carbon emissions over south-ern Europe. We hypothesize that humans setting fire and humanmanipulation of the landscape respectively altered fire frequencyand biomass burned creating a mosaic fire landscape. Our approachdiffers significantly from previous studies by the selection of onlyhigh-quality standardized-method macro-charcoal records, novelstatistical synthesis processes and a data-model comparisonapproach.

2. Materials and methods

Around 20 continuous macro-charcoal (>150 mm) records, ofhigh temporal resolution (<50 years per sample), and from stan-dardized method of quantification (sieving method), documentingthe last 16,000 years (hereafter 16 ky) of fire regime history, havebeen published to date for Europe. 18 of these are located in thefire-prone regions of southern European (Romania, Italy,Switzerland, France and Portugal; Fig. 1 and Table S1). The conti-nuity of this dataset is unique: all other charcoal-inferred paleofirereconstructions available to date at the Europe-scale are based onheterogeneous data-series with disparities in methods and units ofquantification and sometimes with a poor age-control (Power et al.,2008). Furthermore, numerous other paleofire records are based onmicro-charcoal quantification (<100e200 mm), not suitable forpaleofire regime parameters assessment (Conedera et al., 2009).These previous reconstructions provide only a general estimate offire activity, which limits comparisons with other environmentaldata or model outputs that document, for instance, climate vari-ability, crop land-area, burned land-fraction or carbon emissions byfire (Brücher et al., 2014). According to their quality and the initialstandardized method of quantification, all macro-charcoal recordsselected for this paper have been analysed here using themethod ofcharcoal signal decomposition developed by Blarquez et al. (2013)adapted from Higuera et al. (2009). The decomposition of char-coal records aims to infer biomass burned and fire frequency (Aliet al., 2012), i.e. ones the two main components of fire regime,rather than discuss fire activity from raw charcoal data, which is avery ambiguous term (see supplementary data for detaileddescription of methodological procedures).

Fig. 1. Location of charcoal site-records studied and classification of elevation clusters used for the interpretation (see also Fig. 3 and Table S1).

B. Vanni!ere et al. / Quaternary Science Reviews 132 (2016) 206e212 207

3. Results and discussion

The composite Charcoal Accumulation Rate (CHAR) for the 18sites shows a general increase from 15 ky to 7 ky, followed byrelative stability until 1 ky, when it starts to decline (Fig. 2a). Thispattern is coherent with the European region synthesis based onthe charcoal database (Power et al., 2008; Marlon et al., 2013),reporting both uniform patterns in fire activity during the LGM andthe Late-glacial at the continental scale, and significant regionaldifferences during the Holocene, probably caused by the complexinterplay between climate, vegetation and human impacts. Fromour macro-charcoal dataset, the density distribution of recordshighlights two time periods: significant spatial disparities in fireactivity between 16 and 10 ky, which contrasts with relative uni-formity thereafter (Fig. 2a). These temporal boundaries are wellconfirmed by the biomass burned estimation (CHAR background)with a relative decrease trend from 7 ky onwards (Fig. 2b). Firefrequency shows a complementary figure of fire history: a generalincreasing trend over the last 16 ky and only four periods (ca. 14.5,

11.5, 4.5 and 1 ky) of strong homogeneity in the signal oversouthern Europe (Fig. 2c).

From the elevation-clusters, the main trends in fire proxies canbe resumed as follows: i) a gradual decline in biomass burned sinceca. 7 ky which became significantly lower after 5 ky in the lowlands(Fig. 2e); ii) a stronger uniformity in fire frequency increase inlowland areas compare to other areas during the last 16 ky (Fig. 2f);iii) a very low biomass burned inmidland areas and highlands areasduring the Late-glacial and the early Holocene (Fig. 2h,k); iv) a risein biomass burned after 7 ky in highlands (Fig. 2k) that contrastswith midlands and lowlands (Fig. 2h,e).

Climate and biomass hypothesis. Between 16 and 10 ky, dis-parities in the charcoal accumulation rate and the burned biomasssignal (low signal density) likely reflect variability in fuel avail-ability across the study area associated to the contrasted responseof vegetation and land cover dynamics among sites to the Late-glacial climate changes (Fig. 2a,b). Complementary, the homoge-nous trends (high signal density) in fire frequency at ca. 14.5 (theGlacialeInterglacial transition) and 11.5 ky (the Younger

Fig. 2. Fire regime proxy series through time (Fig. S1, S2 and S3). Density distribution of Charcoal Accumulation Rate Z-scores (CHAR), biomass burned and fire frequency proxies forall sites (a to c), for lowland sites (d to f), midland sites (g to i) and highland sites (j to l). The color scale translates the density of the proxy site-records at each time step; cold colorsare indicative of data dispersion and thus capture the heterogeneity among sites; hot colors indicate site response homogeneity and thus spatial uniformity. The white dotted linesseparate time periods with significant differences in the proxies' mean and variance (see Materials and Methods). All ages are expressed in calibrated kilo-years Before the Present(conventionally fixed at 1950 AD).

B. Vanni!ere et al. / Quaternary Science Reviews 132 (2016) 206e212208

DryaseHolocene transition) agrees well with the major,continental-scale temperature increase (Feurdean et al., 2013, Riuset al., 2014, Fig. 2c). During the Late-glacial, fire regime appears tobe strongly driven by marked increases in summer temperatureattested at the continental scale, although the low biomass avail-ability also provides a limiting factor that varies at the spatial scale.Despite increasing temperature and precipitation, with subsequentenrichment of fuel load, fire frequency remained low during theLate-glacial (14.7e12.7 ky, Connor et al., 2012; Doyen et al., 2015;Leys et al., 2014b; Rius et al., 2014). By contrast, the establish-ment and rapid expansion of the forest cover at the onset of theHolocene (12.7e10 ky) resulted in increased fire frequency (id.).This increase in fire frequency was spatially uniform and directlyconnected with climate changes whereas the amount of biomassburned varied regionally because of the modulated impact ofclimate on fire through vegetation.

From 10 to 7 ky, increases in biomass burnedmirror (Fig. 2b) thesouthern European temperature rise (Renssen et al., 2009) andvegetation development towards maximum woody biomass(Feurdean et al., 2013; Rius et al., 2011), both characteristic of theHolocene ThermalMaximum. At ca. 8e7 ky, themaxima and spatialhomogeneity in burned biomass contrast with moderate andspatially heterogeneous patterns in fire frequency (ca. 3 fires permillennia, Fig. 2c). This probably reflects the stochastic and spatiallyvariable recurrence of severe fire episodes. This pattern contrastswith that of the Late-glacial and the onset of the Holocene, andillustrates the influence of biomass quantity and quality in medi-ating the fire regime (Blarquez et al., 2012; Higuera et al., 2009),although it remains difficult to decipher ignition sources. A humancontribution to fire activity during the early Holocene and theMesolithic period cannot be excluded, but evidence is presentlylacking. However, the amount of biomass burned maximum at ca.8e7 ky is contemporaneous with the onset of the Neolithic, theonset of agriculture and pastoralism, and human-induced land-cover changes in southern Europe, such as deforestation by slash

and burn (Colombaroli et al., 2008; Kaltenrieder et al., 2010; Manenand Sabatier, 2003; Rius et al., 2009, 2012; Vanni!ere andMartineau,2005).

To evaluate regional patterns in fire regime and its underlyingcauses, we look for spatial site clusters in the fire records. ThePrincipal component analysis, plotted in Fig. 3, indicates thatelevation is an important driver and underscores the contrastingfire histories between lowland and highland sites in particular(Fig. 2). Low amount of biomass burned in highlands before 9e7 kyreflects low biomass availability during the early Holocene, i.e.before the upshift of timberline and treeline, whereas lowlandbiomass expansion was conducive to fire activity during the sameperiod (Carcaillet et al., 2009; Kaltenrieder et al., 2010).

During the last 7 ky, a sharp and long-lasting decrease inbiomass burned is recorded at both lowland and midland sites(Fig. 2e,h). Magny et al. (2013) documented a possible Holoceneclimate reversal from ca. 7 ky resulting in a wetter and colderclimate in the northern Mediterranean Basin. This agrees well withEuropean cooling attested by glacier mass balance changes char-acteristic of the onset of the Neoglacial (Magny et al., 2013). Thistrend of decreased amount of biomass burned is consistent withNorth America fire regime records and may reflect a nonlinearclimate response to the gradual orbitally-driven decrease in sum-mer insolation (Ali et al., 2012). Simulations with the CLIMBER-2climate model suggest that Earth’s eccentricity partially controlsthe mid-to late-Holocene millennial trend towards wetter andcooler conditions (Brücher et al., 2014, Fig. 4). By couplingCLIMBER-2 with the land surface model JSBACH, which includes adynamical vegetation scheme, Brücher et al. (2014) estimated thedisturbance of vegetation by natural fire occurrence during the last8 ky. Estimates of the burned area using these model simulationsindicate a climate-driven decrease between 7 and 5 ky for southernEurope (Fig. 4 and S4). Since 7 ky, fire disturbances are accompa-nied by modifications in the forest composition: decreasing de-ciduous woodlands in temperate area, expanding coniferous forestsin the midlands, and spreading of deciduous adapted taxa in theMediterranean Basin, at the expenses of species sensitive or intol-erant to fire (e.g. Tilia, Ulmus, Abies; de Beaulieu et al., 2005;Colombaroli et al., 2007; Vanni!ere et al., 2008; Tinner et al.,2013). Furthermore, fire affected biodiversity on the long-term,being responsible of highly diverse grasslands (e.g. Blarquez et al.,2010; Colombaroli and Tinner, 2013).

Human and biomass hypothesis. Today, the fire regime in theMediterranean and southern European regions is closely linked tothe summer-dry versus winter-wet seasonality that modulates thefire season and promotes biomass growth, respectively. However,at present, about 95% of fire ignitions are anthropogenic and ac-count for over 90% of the total area burnt (Ganteaume et al., 2013).Human ignitions generally overcome natural limitations to burningand drastically increase fire activity during summer heat waves.Furthermore, human ignitions also alter fire regimes throughreducing fuel load or connectivity. By ca. 8e7 ky, farming andherding were introduced to Europe from the Near East and Ana-tolia, leading to the Neolithic transition (Manen and Sabatier,2003). Although low populations would be expected to limit theenvironmental impact by these early farmers, the high per capitaland assumed to be needed at this timemay underpin the extensiveuse of fire to open-up landscapes by shifting cultivation and slash-and-burn practices (Kaplan et al., 2011; McWethy et al., 2013;Olofsson and Hickler, 2008).

In contrast to the decreased amount of biomass burned inlowland sites after 7 ky (Fig. 2e), fire frequency has generallycontinued to increase throughout the Holocene (Fig. 2c,f,i,l). Thisimplies more fire occurrences that burn less biomass. Increasedfires occurrences were linked to humans setting fire for land

Fig. 3. Principal component analysis (PCA), performed on the CHAR, biomass burnedand fire frequency proxies datasets from each site-record (Fig. S1 and S3). Principalcomponents 1 and 2 (PC1 and PC2) respectively explain 18% and 12% of data variability.Ellipses represent 2-sigma error of the data distribution for each elevation group. ThePCA separates the sites into three distinct clusters: lowland sites (below 750 m a.s.l.),midland sites (between 1100 and 2050 m a.s.l.), and highland sites (alpine, above2050 m a.s.l.).

B. Vanni!ere et al. / Quaternary Science Reviews 132 (2016) 206e212 209

management and reduced biomass burned results from landscapeopening. Moreover, unexpected increase in the amount of biomassburned is recorded in the highlands after 7 ky (Fig. 2k). This maysuggest an intensive human use of fire during time intervals whenthe timberline was higher. Comparisons with population estimatesfor our study area from the Global Land Use and technologicalEvolution Simulator (GLUES; Lemmen and Wirtz, 2014), stronglysupports this interpretation (Fig. 4, S5 and S6). From 8 to 3 ky, thesharp increase in human population densities mirrors a relativeincrease in fire frequency and a regional decrease in biomassburned, while themodeled burned area (from CLIMBER-2e JSBACHmodel) remained stable (Brücher et al., 2014, Fig. 4). Local studiesattest in particular that the major demographic increases duringthe Bronze Age have resulted in particularly higher fire frequencyand biomass burned to open new-forested areas for land use(Surmely et al., 2009; Feurdean et al., 2013; Rius et al., 2009, 2012).

Furthermore, the charcoal-inferred fire records were comparedwith modeled vegetation dynamics by the LPJ-GUESS model forcedby climate parameters with and without of land-use effects (Kaplanet al., 2011; Klein Goldewijk et al., 2011; McWethy et al., 2013;Olofsson et al., 2013; Smith et al., 2001; Strandberg et al., 2014,Fig. 4 and S7). The comparison of simulated woody biomass and firecarbon flux with potential natural scenarios and with humanimpact, indicates an early human impact on fuel load and fire-carbon emissions from ca. 7 ky, which significantly increased byca. 5e4 ky (Bronze age), and further accentuated by ca. 3e2 ky (Ironage; Fig. 4 and S7). These key periods of vegetation and humanland-use changes are concurrent with regional decreases inbiomass burned and an increase in fire frequency, which furthersuggests humans likely influenced fire regimes during the middle-and late-Holocene. The fire frequency maximum by 1 ky in lowlandsites is accompanied by a short-term plateau in themillennial-scaletrend of decreasing biomass burnt. This may be linked to woodybiomass recovery and the return of former fire practices to manageland after the Roman period (Kaltenrieder et al., 2010; Rius et al.,2011). Midland and highland sites show spatial disparities for thelast two millennia. However, a fire activity minimum (for bothbiomass burned and fire frequency) is recorded ca. 0.5 ky in bothhighland and lowland sites and may be connected to both the LittleIce Age cooling and probably the Black Death episodes, wiping out alarge fraction of the human population in Europe and decrease offire use for land management.

4. Conclusions

A regional analysis of fire regime components (biomass burnedand fire frequency) and identification of periods of homogenousfire response highlighted fire-regime changes and forcing factors.Furthermore, our study questions the trend of increasing fire ac-tivity in southern Europe during the Holocene shown by previouscharcoal syntheses. Instead, we document a decline in biomassburned since 7 ky, implying a reduction in fire-carbon emissions insouthern Europe. With climate forcing factors only, one wouldexpect fire activity to have drastically decreased since the HoloceneThermal Maximum. Our data show that this effect was likely

Fig. 4. Comparison between fire Lowland sites biomass burned (a) and fire frequency(b) with outputs from climate, vegetation, land use and population models; (c) Eartheccentricity; (d) Annual mean temperature and (e) precipitation for southern Europemodeled by the Earth system model of intermediate complexity CLIMBER-2 (Brücher

et al., 2014; Fig. S4); (f) Natural burned area estimated by CLIMBER-2 coupled withthe land surface model JSBACH that includes a dynamical vegetation scheme (Brücheret al., 2014; Fig. S4); (g) Population density estimates for southern Europe from theGlobal Land use and technological Evolution Simulator (GLUES; Lemmen and Wirtz,2014; Fig. S5 and S6); (h) to (k) Woody carbon biomass and fire carbon flux esti-mates from LPJ-GUESS vegetation model based on two scenarios: potential natural andwith the KK10 land use scheme (Kaplan et al., 2011; Olofsson et al., 2013; Fig. S7). Allages are expressed in calibrated kilo-years Before the Present (conventionally fixed at1950 AD).

B. Vanni!ere et al. / Quaternary Science Reviews 132 (2016) 206e212210

counter-balanced by human activities, which artificially main-tained a higher-than-expected fire frequency.We conclude that thisanthropogenic fire regime created a more spatially heterogeneousfire landscape than would have occurred under natural fire condi-tions. Elevation appears to be one of the most important factorsthat explain regional patterns in fire regime history. Elevation en-compasses the nature of the climate and fuel, its structure andamount; but is also strongly correlated with land-use patterns (e.g.crops vs pastures). Thus, we propose that no clear distinction couldbe made between “natural” and “anthropogenic” fire regimes forsouthern Europe, and the Mediterranean Basin in particular, wherefire regimes have been affected by human activities since at leastthe Neolithic (5e7 ky). Our study shows the range of natural vari-ability and the extent to which vegetation legacy (since 7 ky) in-fluences the observed trends. In this context, ecosystemmanagement strategies should not only be based on disturbancethresholds and reference conditions, but also be adapted to long-term paleoecological studies and regional history including hu-man impacts (Gillson and Marchant, 2014; Morales-Molino et al.,2015). This is inherently challenging because, in southern Europe inparticular, human populations have deeply influenced ecosystembiodiversity and their trajectory for centuries to come.

5. Author contributions

B.V., O.B., D.R., E.D., D.C., S.C., A.F., P. K. and B.L. produced charcoaldata; T.B., C.L. and J.O. performmodels; B.V. O.B. and C.M. contributeto the methodological design; B.V. and O.B. analysed the data; B.V.designed the study, collected the data, and wrote the paper; Allauthors discussed the results and commented on the manuscript.

Acknowledgements

We thank W. Tinner and M. Power for critical discussions andreading the manuscript; the research leading to these results hasreceived funding from The PALEOMEX-MISTRALS project (CNRS,France), The R"egion Franche-Comt"e (OREAS project), The Universit"eof Bourgogne Franche-Comt"e, The MSHE C.N. Ledoux and The UMRChrono-environnement, and has been supported also by ThePAGES-GPWG.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.quascirev.2015.11.012.

References

Ali, A.A., Blarquez, O., Girardin, M.P., He

ly, C., Tinquaut, F., et al., 2012. Control of themultimillennial wildfire size in boreal North America by spring climatic con-ditions. PNAS 109, 20966e20970.

Blarquez, O., Bremond, L., Carcaillet, C., 2010. Holocene fires and a herb-dominatedunderstory track wetter climates in subalpine forests. J. Ecol. 98, 1358e1368.

Blarquez, O., Carcaillet, C., Elzein, T.M., Roiron, P., 2012. Needle accumulation ratemodel-based reconstruction of palaeo-tree biomass in the western subalpineAlps. Holocene 22, 579e587.

Blarquez, O., Girardin, M.P., Leys, B., Ali, A.A., Aleman, J.C., et al., 2013. Paleofirereconstruction based on an ensemble-member strategy applied to sedimentarycharcoal. Geophy. Res. Let. 40, 2667e2672.

Blarquez, O., Ali, A.A., Girardin, M.P., Grondin, P., Fr"echette, B., Bergeron, Y., H"ely, C.,2015. Regional paleofire regimes affected by non-uniform climate, vegetationand human drivers. Sci. Rep. 5, 13356.

Bowman, D.M.J.S., Balch, J., Artaxo, P., Bond, W.J., Cochrane, M.A., et al., 2011. Thehuman dimension of fire regimeson earth. J. Biogeogr. 38, 2223e2236.

Brücher, T., Brovkin, V., Kloster, S., Marlon, J.R., Power, M.J., et al., 2014. Comparingmodelled fire dynamics with charcoal records for the Holocene. Clim. Past 10,811e824.

Carcaillet, C., Ali, A.A., Blarquez, O., Genries, A., Mourier, B., et al., 2009. Spatialvariability of fire history in subalpine forests: from natural to cultural regimes.Ecoscience 16, 1e12.

Carri"on, J.S., Fern"andez, S., Gonz"alez-Samp"eriz, P., Gil-Romera, G., Badal, E., Carri"on-Marco, Y., L"opez-Merino, L., L"opez-S"aez, J.A., Fierro, E., Burjachs, F., 2010. Ex-pected trends and surprises in the Lateglacial and Holocene vegetation historyof the Iberian Peninsula and Balearic Islands. Rev. Palaeobot. Palynol 162,458e475.

Conedera, M., Tinner, W., Neff, C., Meurer, M., Dickens, A.F., Krebs, P., 2009.Reconstructing past fire regimes: methods, applications, and relevance to firemanagement and conservation. Quat. Sci. Rev. 28, 555e576.

Colombaroli, D., Marchetto, A., Tinner, W., 2007. Long-term interactions betweenMediterranean climate, vegetation and fire regime at Lago di Massaciuccoli(Tuscany, Italy). J. Ecol. 95, 755e770.

Colombaroli, D., Vanni!ere, B., Chapron, E., Magny, M., Tinner, W., 2008. Fire-evegetation interactions during the MesolithiceNeolithic transition at Lagodell’Accesa, Tuscany, Italy. Holocene 18, 679e692.

Colombaroli, D., Henne, P.D., Kaltenrieder, P., Gobet, E., Tinner, W., 2010. Speciesresponses to fire, climate and human impact at tree line in the Alps as evi-denced by palaeo-environmental records and a dynamic simulation model.J. Ecol. 98, 1346e1357.

Colombaroli, D., Tinner, W., 2013. Determining the long-term changes in biodiver-sity and provisioning services along a transect from Central Europe to theMediterranean. Holocene 23, 1625e1634.

Colombaroli, D., Ssemmanda, I., Gelorini, V., Verschuren, D., 2014. Contrasting long-term records of biomass burning in wet and dry savannas of equatorial EastAfrica. Glob. Change Biol. 20 (9), 2903e2914.

Connor, S.E., Araújoa, J., van der Knaap, W.O., van Leeuwen, J.F.N., 2012. A long-termperspective on biomass burning in the Serra da Estrela, Portugal. Quat. Sc. Rev.55, 114.

Daniau, A.-L., Harrison, S.P., Prentice, I.C., Brewer, S., et al., 2012. Predictability ofbiomass burning in response to climate changes. Global Biogeochem. Cycles 26,GB4007.

de Beaulieu, J.-L., Miras, Y., Andrieu-Ponel, V., Guiter, F., 2005. Vegetation dynamicsin north-western Mediterranean regions: Instability of the Mediterraneanbioclimate. Plant Biosyst. 139, 114e126.

Doyen, E., Vanni!ere, B., Rius, D., B"egeot, C., Millet, L., 2015. Climate and biomasscontrol on fire activity during the late-glacial/early-Holocene transition intemperate ecosystems of the upper Rhone valley (France). Quat. Res. 83,94e104.

Feurdean, A., Liakka, J., Vannie

re, B., Marinova, E., Hutchinson, S.M., et al., 2013.12,000-Years of fire regime drivers in the lowlands of Transylvania (Central-Eastern Europe): a data-model approach. Quat. Sc. Rev. 81, 48e61.

Ganteaume, A., Camia, A., Jappiot, M., San-Miguel-Ayanz, J., Long-Fournel, M., et al.,2013. A review of the main driving factors of Forest fire ignition over Europe.Environ. Manag. 51, 651e662.

Gillson, L., Marchant, R., 2014. From myopia to clarity: sharpening the focus ofecosystem management through the lens of palaeoecology. TREE 29, 317e325.

Higuera, P.E., Brubaker, L.B., Anderson, P.M., Hu, F.S., Brown, T.A., 2009. Vegetationmediated the impacts of postglacial climate change on fire regimes in thesouth-central Brooks range, Alaska. Ecol. Monogr. 79, 201e219.

Kaltenrieder, P., Tinner, W., Ammann, B., 2005. Long-term vegetation history attimberline in the Swiss Alps (Alpe d’ Essertse, VS). Bot. Helvetica 115, 137e154.

Kaltenrieder, P., Procacci, G., Vanni!ere, B., Tinner, W., 2010. Vegetation and firehistory of the Euganean Hills (Colli Euganei) asrecorded by Lateglacial and Holocene sedimentary series from Lago della Costa(northeastern Italy). Holocene 20, 679e695.

Kaplan, J.O., Krumhardt, K.M., Ellis, E.C., Ruddiman, W.F., Lemmen, C., et al., 2011.Holocene carbon emissions as a result of anthropogenic land cover change.Holocene 21, 775e791.

Klein Goldewijk, K., Beusen, A., de Vos, M., van Drecht, G., 2011. The HYDE 3.1spatially explicit database of human induced land use change over the past12,000 years. Glob. Ecol. Biogeogr. 20, 73e86.

Lemmen, K., Wirtz, W., 2014. On the sensitivity of the simulated European Neolithictransition to climate extremes. J. Archaeol. S. C. 51, 65e72.

Leys, B., Finsinger, W., Carcaillet, C., 2014a. Historical range of fire frequency is notthe Achilles’heel of the Corsican black pine ecosystem. J. Ecol. 102, 381e395.

Leys, B., Carcaillet, C., Blarquez, O., Lami, A., Musazzi, S., Trevisan, R., 2014b. Resis-tance of mixed subalpine forest to fire frequency changes: the ecologicalfunction of dwarf pine (Pinus mugo ssp. mugo). Quat. Sc. Rev. 90, 60e68.

Magny, M., Combourieu-Nebout, N., de Beaulieu, J.L., Bout-Roumazeilles, V.,Colombaroli, D., et al., 2013. Northesouth palaeohydrological contrasts in thecentral Mediterranean during the Holocene: tentative synthesis and workinghypotheses. Clim. Past. 9, 2043e2071.

Manen, C., Sabatier, P., 2003. Chronique radiocarbone de la neolithisation enMediterranee nord-occidentale. Bull. la Soci"et"e pr"ehistorique française 100,479e504.

Marlon, J.R., Bartlein, P.J., Daniau, A.-L., Harrison, S.P., Maezumi, S.Y., et al., 2013.Global biomass burning: a synthesis and review of Holocene paleofire recordsand their controls. Quat. Sc. Rev. 65, 5e25.

McWethy, B., Higuera, P.E., Whitlock, C., Veblen, Bowman, D.M.J.S., et al., 2013.A conceptual framework for predicting temperate ecosystem sensitivity tohuman impacts on fire regimes. Glob. Ecol. Biogeogr. 22, 900e912.

Molinari, C., Lehsten, V., Bradshaw, R.H.W., Power, M.J., Arneth, A., et al., 2013.Exploring potential drivers of European biomass burning over the Holocene: adata-model analysis. Glob. Ecol. Biogeogr. 22, 1248e1260.

Morales-Molino, C., Vescovi, E., Krebs, P., Carlevaro, E., Kaltenrieder, P., Conedera, M.,Tinner, W., Colombaroli, D., 2015. The role of human-induced fire and sweet

B. Vanni!ere et al. / Quaternary Science Reviews 132 (2016) 206e212 211

chestnut (Castanea sativa Mill.) cultivation on the long-term landscape dy-namics of the southern Swiss Alps. Holocene 25 (3), 482e494.

Olofsson, J., Hickler, T., 2008. Effects of human land-use on the global carbon cycleduring the last 6,000 years. Veget. Hist. Archaeobot 17, 606e615.

Olofsson, J., 2013. The Earth e Climate and Anthropogenic Interactions in a LongTime Perspective (PhD thesis). Dept of Physical Geography and EcosystemScience, Lund. ISBN 978-91-85793-35-8.

Pausas, J.G., Llovet, J., Rodrigo, A., Vallejo, R., 2008. Are wildfires a disaster in theMediterranean basin? e a review. Int. J. Wildland Fire 17, 713e723.

Power, M., Marlon, J.R., Ortiz, N., Bartlein, P.J., Harrison, S.P., Mayle, F.E., et al., 2008.Changes in fire regimes since the last glacial Maximum: an assessment basedon a global synthesis and analysis of charcoal data. Clim. Dyn. 30, 887e907.

Renssen, H., Sepp€a, H., Heiri, O., Roche, D.M., Goosse, H., et al., 2009. The spatial andtemporal complexity of the Holocene thermal maximum. Nat. Geosci 2,411e414.

Rius, D., Galop, D., Doyen, E., Millet, L., Vanni!ere, B., 2014. Biomass burning responseto high-amplitude climate and vegetation changes in southwestern France fromthe last glacial to the early Holocene. Veget. Hist. Archaeobot 23, 729e742.

Rius, D., Vanni!ere, B., Galop, D., 2009. Fire frequency and landscape management inthe northwestern Pyrenean piedmont, France, since the early Neolithic (8000cal. BP). Holocene 19, 847e859.

Rius, D., Vanni!ere, B., Galop, D., 2012. Holocene history of fire, vegetation and landuse from the central Pyrenees (France). Quat. Res. 77, 54e64.

Rius, D., Vanni!ere, B., Galop, D., Richard, H., 2011. Holocene fire regime changes frommultiple-site sedimentary charcoal analyses in the Lourdes basin (Pyrenees,France). Quat. Sc. Rev. 30, 1696e1709.

Schr€oter, D., Cramer, W., Leemans, R., Prentice, I.C., Araújo, M.B., et al., 2005.Ecosystem service supply and vulnerability to global change in Europe. Science310, 1333e1337.

Seneci, D., Lucas, A., Chen, H.Y.H., Bergeron, Y., Larouche, A., et al., 2013. Multi-millennial fire frequency and tree abundance differ between xeric and mesic

boreal forests in central Canada. J. Ecol. 11, 356e367.Smith, B., Prentice, I.C., Sykes, M.T., 2001. Representation of vegetation dynamics in

the modelling of terrestrial ecosystems: comparing two contrasting approacheswithin European climate space. Global Ecol. Biogeogr 10, 621e637.

Strandberg, G., Kjellstr€om, E., Poska, A., Wagner, S., Gaillard, M.-J., et al., 2014.Regional climate model simulations for Europe at 6 and 0.2 k BP: sensitivity tochanges in anthropogenic deforestation. Clim. Past. 10, 661e680.

Surmely, F., Miras, Y., Guenet, P., Nicolas, V., Savignat, A., et al., 2009. Occupation andland use history of a medium mountain from the Mid-Holocene: a pluri-disciplinary study performed in the south Cantal (French Central Massif). C. R.Palevol 8, 737e748.

Tinner, W., Conedera, M., Brigitta, A., Lotter, A.F., 2005. Fire ecology north and southto the Alps since the last ice age. Holocene 15, 1214e1226.

Tinner, W., Colombaroli, D., Heiri, O., Henne, P., Steinacher, M., Untenecker, J.,Vescovi, E., Allen, J., Carraro, G., Conedera, M., Joos, F., Lotter, A.F., Luterbacher, J.,Samartin, S., Valsecchi, V., 2013. The past ecology of Abies alba provides newperspectives on future responses of silver fir forests to global warming. Ecol.Monogr. 83 (4), 419e439.

van der Werf, G.R., Peters, W., van Leeuwen, T.T., Giglio, L., 2013. What could havecaused pre-industrial biomass burning emissions to exceed current rates? Clim.Past. 9, 289e306.

Vanni!ere, B., Colombaroli, D., Chapron, E., Leroux, A., Tinner, W., et al., 2008. Climateversus human-driven fire regimes in Mediterranean landscapes: the Holocenerecord of Lago dell'Accesa (Tuscany, Italy). Quat. Sc. Rev. 27, 1181e1196.

Vanni!ere, B., Power, M.J., Roberts, N., Tinner, W., Carrio

n, J., et al., 2011. Circum-Mediterranean fire activity and climate changes during the mid-Holoceneenvironmental transition (8500e2500 cal. BP). Holocene 21, 53e73.

Vanni!ere, B., Martineau, R., 2005. Histoire des feux et pratiques agraires duN"eolithique !a l'Age du Fer en R"egion Centre (France). Implications territoriales,d"emographiques et environnementales. Gall. Pre

histoire 47, 167e186.

B. Vanni!ere et al. / Quaternary Science Reviews 132 (2016) 206e212212