- Email: [email protected]
Climatic pacing of Mediterranean fire histories from lake sedimentary microcharcoal
Global and Planetary Change 63 (2008) 317–324
Contents lists available at ScienceDirect
Global and Planetary Change 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 / g l o p l a c h a
Climatic pacing of Mediterranean fire histories from lake sedimentary microcharcoal R. Turner a,1, N. Roberts a,⁎, M.D. Jones b a b
School of Geography, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK School of Geography, University of Nottingham, University Park, Nottingham, NG7 2RD, UK
a r t i c l e
i n f o
Article history: Received 26 March 2008 Accepted 25 July 2008 Available online 29 July 2008 Keywords: microcharcoal fire oxygen isotopes Mediterranean pollen
a b s t r a c t The microcharcoal content (particles b 180 µm) of overlapping sedimentary sequences from two crater lake basins in central Turkey are used to reconstruct the regional fire history of the East Mediterranean oak–grass parkland zone from the Last Glacial Maximum to the present-day. These results are correlated with stable isotope and pollen data from the same cores in order to assess the changing role of climate, vegetation and human activity in landscape burning. This indicates that climatically-induced variation in biomass availability was the main factor controlling the timing of regional fire activity during the Last Glacial– Interglacial climatic transition, and again during Mid-Holocene times, with fire frequency and magnitude increasing during wetter climatic phases. Spectral analysis of the Holocene part of the record from Eski Acıgöl indicates significant cyclicity with a periodicity of ~ 1500 years that may be linked with large-scale climate forcing. Although proto-agricultural societies were established in this region as early as 10,000 years ago, it is only during the last two to three millennia that the pacing of wildfire cycles appears to have become decoupled from climate and linked instead to human-induced changes in land cover and fuel load availability. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Under natural conditions fire, climate and biomass are inextricably linked (Pausas, 2004; Whelan, 1995). The nature of the climate (e.g. humid vs. arid) influences the structure and species composition of a vegetation community and therefore fuel availability, which is a powerful influence on fire regimes. In turn, fire is itself a major determinant of seasonally dry ecosystems including the global distribution of tree cover (Bond et al., 2005). People modify the natural patterns of biomass burning through their manipulation and management of vegetation communities, including via practices of fire suppression or promotion (Clark and Royall, 1995), and through accidental or deliberate ignition of “natural” vegetation. Eastern Mediterranean parklands are a classic example of an ecosystem where fire, grazing and other forms of disturbance are fundamental to the maintenance and regeneration of the ecosystem (Grove and Rackham, 2001; Naveh, 1974). In addition, but in contrast to other summer-dry Mediterranean-type regions of the world (Köppen type Cs; e.g. California, S. African Cape), the Mediterranean has a long history of human occupation and impact. There is archaeological evidence of semi-sedentary advanced hunter–gath-
⁎ Corresponding author. Tel.: +44 1752 585965; fax: +44 1752 233054. E-mail address: [email protected] (N. Roberts). 1 Current address: Higher Education Learning Partnerships CETL, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK. 0921-8181/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.gloplacha.2008.07.002
erers during the Last Glacial–Interglacial transition (LGIT), and the Eastern Mediterranean subsequently became one of the earliest centres where Neolithic agriculture developed. Therefore this might be anticipated to have been a region where anthropogenic agencies joined, or even overtook, natural ones in exerting control over regional fire activity at an early date (Roberts, 2002). Charcoal particles have been widely used to reconstruct fire events over multiple timescales in different parts of the world. For example, microscopic charcoal analysis has been central in understanding vegetation change in the North American forests (e.g. Clark and Royall, 1995; 1996) and forest succession in the Western Mediterranean (e.g. Carrión and van Geel, 1999; Múgica et al., 1998; 2001; Sadori and Giardini, 2007; Colombaroli et al., 2007; Vannière et al., 2008), the history of human fire use by pre-European populations in Australia and the Americas (Clark, 1983; Clark and Royall, 1995; Kershaw et al., 1997; Trabaud et al., 1993) and understanding changes in the global carbon cycle with increased biomass burning (e.g. Carcaillet et al., 2002; Crutzen and Andreae, 1990). However, there has been limited research into long-term fire dynamics in the Eastern Mediterranean, and this has generally had low temporal resolution (e.g. Yasuda et al., 2000). In this paper we present a high resolution sedimentary record of regional fire history since the Last Glacial Maximum (LGM) from two nearby crater lake basins located in the oak parkland zone of Central Turkey. In order to assess the changing roles of climate, biomass and people on burning regimes, we compare microcharcoal results with other multi-proxy (pollen and oxygen isotope) data from the same stratigraphic sequences, along with archaeological evidence from the study region.
318
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
2. Study area The principal data we report here derive from Eski Acıgöl (38°33'01”N, 34°32'41”E; 1270 metres above sea level (masl)), a small, former brackish lake, drained in 1972, with no inflow streams and a surface catchment not much larger than the lake itself (Roberts et al., 2001). Its sedimentary record extends back to around the time of the LGM, although it is poorly resolved for the last two millennia. Consequently we overlap this during the late Holocene with a shorter, but highly resolved, sequence from Nar crater lake, located 25 km away. Nar lake (38°22′N, 34°27′E; elevation 1363 masl) is still extant with a maximum water depth of 26 m (Jones et al., 2005, 2006; England et al., in press). Nar is also relatively small (~ 0.7 km2) with a watershed catchment of ~4 km2, and the lake is stratified and forms annual varve layers. Both sites are hydrologically closed in terms of surface outflows, and are sensitive to climatic changes. These lakes are located in the oak parkland zone that covers the Anatolian plateau and western Iran (Zohary, 1973) mostly between 900 and 1500 masl (Fig. 1). The surrounding area would comprise Quercus pubescens, Q. cerris, Pistacia, Crataegus, Prunus, Pyrus, and Juniperus with Festuca, Poa and other grasses, and Artemisia and chenopods common at drier, low-elevation sites (Woldring and Bottema, 2003). Most woodland is degraded today, partly due to grazing by sheep and goat herds, and almost all fertile soils have been converted to agriculture. The regional climate is modified Oro-Mediterranean, being summer-dry, with precipitation ranging between ~ 300 and ~ 600 mm y− 1. Average summer temperatures range between 20 °C and 27 °C, while those in winter fall to 3 °C to −3 °C, reflecting the continentality and elevation of the area. The study sites lie in Cappadocia, which has a well-documented archaeological record of continuous human habitation since Neolithic times. The excavated proto-agricultural Neolithic village of Aşıklı Höyük lies to the south west of Eski Acıgöl and dates to between 10,000 and 9500 cal. years BP (Esin, 1991) (all dates are given in calendar years). The region was important throughout later prehistory for raw materials such as obsidian, it lay close to the centre of the Hittite Empire during 2nd millennium BC, and became a Roman province in 17 AD. During the last two millennia, Cappadocia was especially important during early to mid-Byzantine times, and the again during the early Turkish Selçuk period (England et al., in press). 3. Methods At Eski Acıgöl, a 1566 cm composite sequence was obtained from parallel, overlapping cores taken with a modified Livingstone piston
corer and Eijkelcamp corer with a percussion Cobra motor (ESK 96/99). Palynological analysis was conducted on a parallel sequence taken in 1992 (ESK92) using a Dachnowsky corer (which reached 1420 cm) (Woldring and Bottema, 2003). The ESK96/99 and ESK92 sequences were cross correlated through a series of well-constrained tie points, e.g. tephra horizons. Because the ESK92 sequence is shorter than ESK96/99, published pollen data are not available for the period prior to the Late Glacial Interstadial. At Nar, coring was undertaken in the deepest part of the lake using Glew, 3 m Mackereth and Livingstone lake sediment corers. Multi-proxy studies, including δ18O (measured on precipitated authigenic carbonate) and pollen analyses, have been conducted on cores taken from both crater lake systems, with dating based on Useries (Eski Acıgöl) and varve counting (Nar); (see England et al., in press; Jones et al., 2005; Roberts et al., 2001 for methodological details). 3.1. Microcharcoal analytical methods Fire events produce a pulse in charcoal which is transported away from the fire site and can be incorporated into lake sediments. Identifying individual fire events is dependent on sampling the peaks in charcoal concentration within the sedimentary record; a contiguous sampling strategy was therefore adopted in this study. This differs from microcharcoal counts made on pollen slides which are normally derived from samples taken at discrete depth intervals within a core sequence (e.g. at Eski Acıgöl, pollen samples were taken on average every 18.5 cm; Woldring and Bottema, 2003). For microcharcoal analysis from Eski Acıgöl, contiguous 1 cm3 sediment samples measuring 4.0 × 1.0 × 0.25 cm were extracted from the full length of the ESK 96/99 composite core sequence, apart from 0–76 and 1267–1294 cm for which there was no core recovery. A higher sampling resolution of 2.0 × 1.0 × 0.5 cm was applied to selected periods of major climatic change, e.g. Late Glacial–Holocene transition. Contiguous samples were also analysed at Nar, but in this case by laminae age rather than depth interval. For the last 100 varve years, thin blocks of 10 laminations were used (i.e. 1–10, 11–20 VY, etc), while the rest of the sequence sampled contiguously every 20 laminations (i.e. 101–120, 121–140 VY, etc). After adding Lycopodium tablets to enable estimation of charcoal concentrations, samples were prepared using density separation (Turner, 2007). This preparation method was found experimentally by Turner et al. (in press) to have a higher recovery than other published methods for the fine charcoal fraction, and more than ten times the recovery of standard pollen preparation method. Samples
Fig. 1. Location of study sites. Distribution of parkland ecosystem from Zohary (1973).
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
were initially treated with 10% HCl to disaggregate the sediments and then passed through a 180 µm mesh to remove (and retain) the larger, potentially more fragile charcoal particles that could fragment during the preparation and result in a potential overestimation of the charcoal concentration of a sample. Lithium heteropolytungstate with a specific gravity of 2.5 was used to separate the microscopic charcoal particles from the sediment matrix. These were then mounted on a standard microscope slide, and all charcoal particles within each field of view were counted using a BX100 high power light microscope at 200× magnification, until a count of 100 Lycopodium spores was reached. Microcharcoal particles were identified based on a set of diagnostic criteria, primarily that they had to be jet black in colour, with straight edges, and with the presence of a blue hue on edges (Turner et al., in press). 4. Results At Eski Acıgöl the ESK 96/99 core sequence recovered nonlaminated carbonate-rich lake sediments in the upper 607 cm, underlain by annually laminated limnic deposits to at least 1566 cm. 14 C dates from this sequence have been aged by ~ 3000 years due to volcanic outgassing, so that the chronology is primarily based on a series of U-series dates obtained from the composite core sequence (ESK 96/99) (see Roberts et al., 2001 for details and discussion). These are supported by preliminary higher precision U-Th Thermally Ionised Mass Spectrometer age determinations (J. Dorale, pers comm.), by varve counting and by correlations with other 14C dated pollen and stable isotope records from central Anatolia. They show that the core sequence covers the Holocene and the LGIT. Changes in the proxy records correspond to events similar in time and magnitude to the European Bölling–Alleröd and Younger Dryas events. However, it is
319
currently not possible to make firm conclusions as to synchroneity with these events, which are therefore named here as the Late Glacial Interstadial and Late Glacial Stadial (Jones et al., 2007). According to the age model used here, the onset of the Holocene at Eski Acıgöl is dated to 12,080 BP, which compares with an age of 11,700 BP from Greenland ice cores (Rasmussen et al., 2006) and other sources. It is therefore possible that the Eski Acıgöl core chronology overestimates true age by about 3–4%. In this study a linear age-depth sediment accumulation rate of ~12 yr / cm− 1 is used between 1.6 (175 cm) and 16.2 ka BP (1409 cm); before this, sedimentation rate appears to be significantly slower. We use a basal age of 23 ka BP for the ESK 96/99 sequence, but the precise age of the lowest part of the sequence is not well defined and we do not consider the period prior to 16 ka BP in detail here. The potential effects of using different age models are discussed in relation to time series analysis below. The 376 cm long sediment sequence from Nar lake is laminated throughout and consists of 1725 couplets of summer precipitated calcium carbonate and diatom-rich organic layers deposited between autumn and spring. 210Pb and 137Cs dating of the top 50 cm of this sequence, together with the analysis of modern sediments in seston traps show these couplets to be annual (Jones et al., 2005). Dating of the rest of the core sequence is therefore based on laminae counts given in varve years before AD 2001 converted to a calendar time scale (see Jones et al., 2006 for further details). δ18O values recorded from the two crater lake sequences have been linked to changes in regional water balance with negative δ18O values reflecting periods of overall wetter climate. Jones et al. (2005, 2007) used isotope mass balance modelling to calibrate these isotope data climatically. For Eski Acıgöl, they calculated that precipitation was 20– 40% higher than at present during the Late Glacial Interstadial and again in the early Holocene, and 25–40% lower than present during
Fig. 2. Microcharcoal influx data for Eski Acıgöl, Turkey, compared to summary pollen and oxygen isotope data (from Woldring and Bottema, 2003 and Roberts et al., 2001, respectively). Stars indicate U-series dating control points. Two short-lived early Holocene isotopic excursions are not shown (see Roberts et al., 2001 for discussion).
320
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
glacial and Late Glacial Stadial times given temperatures 5 to 6 °C below present. During the Mid-Holocene there is a positive shift in the δ18O values reflecting climatic aridification towards modern conditions. This trend in the δ18O values coincides with a fall in lake levels whereby laminated sediments cease to form. A diatom-inferred increase in the salinity of the lake waters occurred between ca.7000 and 4000 BP (Roberts et al., 2001). At Nar lake, δ18O data show more positive values, inferred to indicate drier climatic conditions, from 1700–1500 BP and again 600 to 40 BP, with more negative isotopic values, and a wetter climate, between 1440–1250 BP, 1000–600 BP, and since AD1960 (Jones et al., 2006). Pollen analyses from Eski Acıgöl and from other sites (e.g. Akgöl, Konya; Bottema and Woldring, 1984) show that the Anatolian plateau was dominated by largely treeless Artemisia-chenopod steppe during Late Glacial times. The onset of the Holocene was associated with the rapid expansion of grasses, an equally abrupt reduction in the proportion of Artemisia and chenopods, and a more gradual increase in Quercus robur-type pollen. The Early Holocene was characterised by open grass parklands including Pistacia as an important element, with continued expansion of oak and some mesic tree species. The maximum extent of tree cover was not reached until ca.5500 BP, after which it declined to be replaced by steppic herbs and Poaceae (Fig. 2). At Nar, pollen analyses carried out at 20 varve year intervals (England et al., in press) indicate four principal land-use phases during the last 1725 years: (i) an agrarian landscape characterised by cereals and tree crops prior to 1330 BP, marking the later part of the so-called Beyşehir Occupation phase, (ii) a period of secondary woodland establishment from 1330–1050 BP coinciding with a “late Antiquity Dark Age”, (iii) renewed deforestation with the re-establishment of cereal agriculture and pastoralism after 1050 BP, and (iv) agricultural intensification during modern times (AD 1830 to present)(Fig. 3).
4.1. Microcharcoals Relative to other SW Asian sequences analysed using the same methods (Turner 2007), overall charcoal concentration and influx values are low at both Nar and Eski Acıgöl. This is inferred to represent mainly long distance transport of charcoal particles to the lakes from the wider environment. The absence of surface inflows means that charcoal will have entered the lakes primarily through aeolian transport processes which usually sample a regional fire signal (Clark, 1998). The inferred regional fire history signal is further supported by a paucity of macroscopic charcoal particles in the N180 µm fraction, which would be associated with fires occurring within the immediate vicinity of the lakes. The lowest charcoal values at Eski Acıgöl were recorded during the Glacial and the Late Glacial Stadial whereas charcoal concentrations increased during the Late Glacial Interstadial and the Holocene (Fig. 2). Charcoal concentration and influx values have fluctuated in quasicyclical fashion throughout the Holocene, with the maximum concentration values occurring 12,000 to 9000 BP and 6000 to 5000 BP. Microcharcoal influx values at Nar prior to the 20th century are comparable to those at Eski Acıgöl during the late Holocene, with lowest values from 1725 to 1370 BP and 550 to 250 BP (Fig. 3). Microcharcoal counts on pollen slides from the Nar sequence show a broadly similar pattern of change to those using heavy liquid separation prior to ~ 200 BP, although absolute values are lower (England et al., in press). In order to explore quantitatively the multi-millennial record of landscape burning, the microcharcoal data for Eski Acıgöl have been subject to time series analysis. Spectral analysis was undertaken to examine periodicities in this microcharcoal time series using PAST (PAlaeontological STatistics; Hammer et al., 2001). Although a number
Fig. 3. Late Holocene microcharcoal influx for Nar, compared to summary pollen and oxygen isotope data from the same sequence (from England et al., in press; Jones et al., 2006, respectively). Pollen zones are from England et al. (in press).
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
of time windows have been analysed, we focus on the Holocene part of the record which spans the period back to 12,080 BP according to the age model used here. This comprises an unbroken series of 213 contiguous samples each representing an aggregated signal of regional biomass burning over 46 ± 15 years. Spectral analysis of this time series shows significant periodicities (those with periodicities below one-third of the total time series and at the 95% confidence limit or better) at (1) 3065, (2) 2018, (3) 1561, and (4) 1034 years (Fig. 4). Of these, cycle (3) is most prominent. We have tested the validity of the periodicities found in this microcharcoal record using a series of age models to track the duration and strength of different cycles, and have found cycle (3) to be robust using all realistic age models. Peak to trough microcharcoal concentration values in the Eski Acıgöl record typically fluctuate by an order of magnitude, which is too large to be explained by any quasi-periodic changes in sediment accumulation rate. There is also no evidence of long-cycle variations in sedimentation rate from varve thickness measurements for the early-Mid-Holocene (unpublished data). Because the onset of the Holocene is very clearly marked in the Eski Acıgöl record on the basis of lithology, pollen and stable isotopes, an alternative age model might use an age of 11,700 BP rather than 12,080 BP for Late Glacial: Holocene transition. In this case, the dominant length for cycle (3) would be reduced to 1431 years, assuming that the age estimate at 175 cm (1600 BP) is also correct. The strength of this ~1500 year cycle implies that the frequency and intensity of wildfire burning in the central Anatolian oak parkland ecosystem has been non-random through time during the Holocene. 5. Discussion Throughout these two overlapping microcharcoal sequences there is evidence of a close relationship between regional-scale fire activity and pollen-inferred shifts in vegetation composition and biomass. During the LGIT the Eski Acıgöl record shows an inverse relationship between the abundance of pollen indicative of cold steppe (Artemisia spp and Chenopodiaceae) and wildfire frequency (Fig. 2). During the Glacial period, and again during the Late Glacial Stadial, charcoal concentrations in the lake sediments were low, indicating that few wildfires were occurring on the Anatolian plateau. Although precise modern analogues are lacking for the vegetation at this time (Peyron et al., 1998), Freitag (1977) proposed a broad similarity with that found today in the semidesert regions of High Asia. The Glacial-age cold steppe appears to have supported only low levels of biomass, and therefore fuel availability, in turn linked to cold and dry climatic conditions. Pollen data show that the vegetation elsewhere in the Eastern Mediterranean at this time was also dominated by steppic herbs and that tree cover was restricted to isolated pockets of favourable conditions (van Zeist and Bottema, 1991). Infrequent wildfires coupled with low biomass and therefore fuel load availability may have been characteristic of the region as a whole at and immediately after the LGM. By contrast, there is a positive correlation between grass pollen and microcharcoal frequency during the LGIT (r = +0.429, p = 0.01, for
Fig. 4. Spectral analysis of the Holocene microcharcoal record for Eski Acıgöl. Horizontal dotted lines represent 0.05 and 0.01 confidence limits. Numbered peaks represent cycles at (1) 3065, (2) 2018, (3) 1561, and (4) 1034 years.
321
the period 15 to 8 ka BP), implying that summer grass fires accounted for a significant part of the regional atmospheric charcoal flux during the Late Glacial Interstadial and early Holocene. This would be consistent with Wright (1993) who argued that the current summerdry precipitation regime did not occur in glacial times, but became reestablished during the Late Glacial Interstadial and again, more permanently, at the start of the Holocene. Microcharcoal concentrations in Eski Acıgöl increased abruptly at the start of the Holocene, with an equally rapid switch from herb-steppe to grassland, and — almost certainly — a significant increase in biomass. The initial early Holocene peak in microcharcoals preceded the establishment of Neolithic settlement in Cappadocia (~10,300 Cal BP), and occurred at a time when there is very little archaeological evidence for human occupation of the central Anatolian plateau (Gerard and Thissen, 2002). It is therefore unlikely that this earliest Holocene wildfire maximum can be attributed to deliberate anthropogenic landscape burning (cf. Roberts, 2002). On the other hand, the positive covariance between microcharcoal concentration values and Poaceae pollen during the first four millennia of the Holocene, along with the slow increase in oak pollen, would be consistent with wildfires maintaining grassland at the expense of woodland. Isotope-inferred precipitation at this time was ~20% higher than at present (Jones et al., 2007), and landscape burning may consequently have acted to slow down what could otherwise have been a rapid post-glacial encroachment of trees across the Anatolian plateau. An Early Holocene peak in fire activity is also indicated in microcharcoal counts from the Van and Ghab pollen records in eastern Anatolia and northwest Syria, respectively (Wick et al., 2003; Yasuda et al., 2000). Following this phase, charcoal concentration values at Eski Acıgöl declined between 9500 and 6000 years BP, equivalent archaeologically to the ceramic Neolithic and Chalcolithic periods. Around this time there were changes in the vegetation community which affected fuel availability and therefore fire occurrence. This involved a continued expansion of oak woodland that put increased pressure on water availability within the vegetation community resulting in the partial replacement of the Poaceae with steppic herbs (Woldring and Bottema, 2003), possibly also linked to grazing pressure by domestic livestock. During the Mid-Holocene oak rather than grass pollen has the clearest positive relationship to microcharcoal flux, especially during maximum Q. robur-type values between 5 and 6 ka BP, and implying a switch in wildfire fuel source from grass to woody biomass. After ~ 5000 BP, there was a permanent decline in arboreal pollen values which coincided both with an overall reduction in regional fire occurrence and an increase in both steppic herbs and ruderal plant species. This has been inferred to reflect increased cultivation and grazing of the landscape during Bronze Age times (Woldring and Bottema, 2003) but it also coincided with a stepwise aridification of the climate as indicated in stable isotope, lake level and salinity indicators from Eski Acıgöl (Roberts et al., 2001). Both factors would have led to changes in the composition of the vegetation community and a reduction in fuel availability. If pollen and archaeological indicators of human-induced land-use change fail show to show a clear and direct relationship to landscape burning activity in Cappadocia during the Early and Mid-Holocene, what of climatic variations, as measured by δ18O data? During the LGIT, microcharcoal influx peaks coincide with negative δ18O values indicative of higher precipitation, notably during the Late Glacial interstadial and at the start of the Holocene (Fig. 2). Greater moisture availability in turn led to an increase in plant biomass and to more frequent and intense wildfire events. However, the Eski Acıgöl oxygen isotope record also shows a longer-term trend that may be linked to the ~22 ka precessional cycle. In order to remove this long-term trajectory, both microcharcoal and δ18O data have been detrended (although in practice, detrending has little effect on the former). The resulting pattern of residuals (Fig. 5) shows that peaks coincide for the two indicators, not only during the LGIT, but also for significant portions of
322
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
the Holocene. A 4 ka moving-window correlation between detrended microcharcoal influx and δ18O exhibits statistically significant r values N −0.5 for all time periods centred between 18 and 10 ka BP, and again shows r values around −0.5 for the Mid-Holocene period centred on 4.5 ka BP. The weaker statistical relationship for the intervening period may be partly linked to gaps in the δ18O record during the early Holocene. The fact that periods of wetter climate and increased landscape burning have coincided at centennial–millennial timescales does not imply that the nature of any causal relationship was constant through time. In contrast to the LGIT, for example, the Mid-Holocene period was most likely dominated by woody biomass as the principal fuel source for wildfires. It is also conceivable that periods of wetter climate encouraged an expansion of human settlement and rural landuse, which may have impacted vegetation cover and landscape burning frequency. None the less, data from Eski Acıgöl are consistent with the hypothesis that wet-dry oscillations in climate acted as the pacemaker for biomass burning intensity in the oak–grass parkland biome of central Anatolia up until at least ~3000 BP. This record of biomass burning is of further significance because of the clear ~1500 year cyclicity that is displayed through the Holocene, which in turn implies a similar dominant periodicity for the underlying climatic forcing. A ~ 1470 year climatic periodicity was recognised by Bond et al. (1997, 2001) from the flux of ice-rafted debris (IRD) in North Atlantic deep-sea cores, and has been widely discussed elsewhere in the literature (e.g. Oppo, 1997; Crowley, 2002), in some cases linked to changes in solar variability. However, it is previously little known from Holocene terrestrial records in the Mediterranean (e.g. Allen et al., 2002). In reality, North Atlantic IRD variations during the Holocene were quasi-periodic rather than truly cyclical in nature and have not been easily replicated (Moros et al., 2006). Additionally, a re-analysis by Debret et al. (2007) showed the
IRD data series to be non-stationary through time, comprising composite cycles of 1450 and 2400 years in length. None the less, a well-established causal relationship exists between the mode of the North Atlantic Oscillation and Mediterranean precipitation (e.g. Enzel et al., 2003; Türkeş and Erlat, 2005). Precipitation in central Anatolia today falls predominantly in winter and spring seasons linked to mid-latitude westerly depressions, ultimately originating in the North Atlantic. Cullen and deMenocal (2000) suggested that Holocene cooling events, such as those associated with Bond “cycles”, could explain millennial-scale climate variability in the East Mediterranean region, notably periods of drought. If North Atlantic cooling events, marked by increased IRD flux, did lead to reduced moisture availability in central Anatolia, then — conversely — IRD minima should coincide with periods of wetter climate, increased biomass availability and more frequent wildfires. Comparison of IRD minima with microcharcoal peaks from Eski Acıgöl, shown in Fig. 5, is hindered by chronological uncertainties in both records. Plausible matches exist for IRD minima at 2.0–2.6, 3.6, 8.0, 9.8 and 10.9 ka BP in the Turkish wildfire record, especially if the Eski Acıgöl chronology is matched to that for the Greenland ice cores (i.e. 11.7 ka for the start of the Holocene). On the other hand, no such matches are evident for the Holocene IRD minima at 7.1 and 8.9 ka BP, while the major peak in wildfire activity at around 5.5 ka BP coincides with an IRD maximum, rather than minimum. A causal linkage between the two therefore seems inconclusive at present. Willis et al. (in press) compared eight microcharcoal records of Holocene wildfire frequencies from different localities worldwide and failed to find consistent dominant spectral periodicites or any clear relationship with inferred solar maxima or minima. Notwithstanding this, the Eski Acıgöl fire history record does give support to the increasing array of evidence for a ~ 1500 year periodicity in proxy-climate data from
Fig. 5. Detrended microcharcoal and oxygen isotope residuals for Eski Acıgöl using 2nd order polynomials, along with moving-window correlation coefficients over 4 ka intervals. Stars indicate the timing of N Atlantic IRD minima (from Bond et al., 2001).
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
northern mid-latitudes during the Holocene (Wanner et al., in press, Figs. 5 and 19). At what point in time did wildfires become decoupled from climate and linked instead to human-induced changes in land cover and fuel load availability? Here, the Late Holocene record from Nar lake is informative. As shown in Fig. 3, the phase of highest microcharcoal influx prior to the 20th century occurred during local pollen zone III (1320–1150 BP; AD 680–950), which saw rural land abandonment and re-afforestation by oak woodland (England et al., in press). This led to an increase in woody biomass and fuel load availability for combustion in wildfires. However, the δ18O record from the same core sequence shows that this phase was not directly coincident with a period of greater moisture availability, as were microcharcoal influx peaks earlier in the Holocene at nearby Eski Acıgöl. In other words, by 1700 years ago, landscape burning in this region was no longer coupled directly to climate variations. Although the timing of the switchover from a nature-dominated to a humandominated wildfire system is not known precisely (c.f. Messerli et al., 2000), it may well have coincided with the onset of the Beyşehir Occupation phase in this region (Eastwood et al., 1998; Kaniewski et al., 2007). This period of human landscape transformation is clearly recognised in pollen records throughout western Anatolia and commenced around 3000 BP. The Nar record also highlights the fact that microcharcoal influx, and by inference landscape burning, was greater during the last century than at any time during the preceding 20,000 years. 6. Conclusions Microcharcoal analysis from Eski Acıgöl and Nar crater lake sediment cores provide a ~20 ka long record of regional-scale fire events with multi-decadal precision. Pollen and microcharcoal influx into the lakes was dominated by regional-scale atmospheric inputs, so that the sites consequently represent a spatially aggregated record of land cover and biomass burning across the oak parkland zone in which they are located. Throughout both sequences there is evidence of a close relationship between fire activity and the composition of the vegetation community, which in turn reflects fuel availability. In this biome, wildfires appear to have been fuel-limited throughout the period since the LGM. This was also true in Iberia during the Last Glacial period (Daniau et al., 2007) and in the North American prairie where clear links have been identified between phases of greater humidity promoting vegetation growth, increasing fuel load and fire occurrence (Brown et al., 2005; Camill et al., 2003). This relationship stands in contrast to that in better watered and more wooded environments in Mediterranean regions, such as Italy, where microcharcoal records indicate that landscape burning was most frequent during climatically arid phases of the Holocene, showing that here wildfires were not fuel-limited (Sadori and Giardini, 2007; Vannière et al., 2008). By employing a multi-proxy approach, we have been able to compare directly proxies for fire (microcharcoals), land-use (pollen) and climate (δ18O), without the uncertainty that can result from correlating separate records. In that regard, our data show a close statistical correlation between centennial to millennial-scale climatic fluctuations (wet to dry) and biomass burning, not only during the LGIT, but also — and more surprisingly — during the Mid-Holocene. Microcharcoal flux during the Holocene has been strongly cyclical in character, with spectral analysis showing a dominant 1500 year periodicity that is likely climatically controlled. As the Holocene progressed, the impact of people upon regional fire activity became increasingly pronounced. However, and notwithstanding abundant archaeological evidence for agricultural and pastoral land-use changes from 10 ka BP onwards, the switch from a climate- to a humanregulated fire regime appears to have taken place only late in the Holocene.
323
Acknowledgements We are grateful for the generous assistance from the late Sytze Bottema, to whom this paper is dedicated, and from Dan Charman, Warren Eastwood, Ann England, and Henk Woldring, and also from Tim Absalom for assistance with the final drafting of the diagrams. This work was supported by Leverhulme grant F/00568/1 and by the British Institute in Ankara.
References Allen, J.R.M., Watts, W.A., McGee, E., Huntley, B., 2002. Holocene environmental variability—the record from Lago Grande di Monticchio, Italy. Quaternary International 88, 69–80. Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., deMonocal, P., Priore, P., Cullen, H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in North Atlantic Holocene and Glacial climates. Science 278, 1257–1266. Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N., Showers, W., Hoffman, S., LottiBond, R., Hajdas, I., Bonani, G., 2001. Persistent solar influence on north Atlantic climate during the Holocene. Science 294, 2130–2136. Bond, W.J., Woodward, F.I., Midgley, G.F., 2005. The global distribution of ecosystems in a world without fire. New Phytologist 165, 525–538. Bottema, S., Woldring, H., 1984. Late Quaternary vegetation and climate of southwestern Turkey, Part II. Palaeohistoria 26, 123–149. Brown, K.J., Clark, J.S., Grimm, E.C., Donovan, J.J., Mueller, P.G., Hansen, B.C.S., Stefanova, I., 2005. Fire cycles in North American interior grasslands and their relation to prairie drought. PNAS 102, 8865–8870. Camill, P., Umbamhowar, C.E., Teed, R., Geiss, C.E., Aldinger, J., Dvorak, L., Kenning, J., Limmer, J., Walkup, K., 2003. Late-glacial and Holocene climatic effects on fire and vegetation dynamics at the prairie-forest ecotone in south-central Minnesota. Journal of Ecology 91, 822–836. Carcaillet, C., Almquist, H., Asnong, H., Bradshaw, R.H., Carrion, J.S., Gaillard, M.J., Gajewski, K., Haas, J.N., Haberle, S.G., Hadorn, P., Muller, S.D., Richard, P.J., Richoz, I., Rosch, M., Sanchez Goni, M.F., von Stedingk, H., Stevenson, A.C., Talon, B., Tardy, C., Tinner, W., Tryterud, E., Wick, L., Willis, K.J., 2002. Holocene biomass burning and global dynamics of the carbon cycle. Chemosphere 49, 845–863. Carrión, J.S., van Geel, B., 1999. Fine-resolution Upper Weichselian and Holocene palynological record from Navarrés (Valencia, Spain) and a discussion about factors of Mediterranean forest succession. Review of Palaeobotany and Palynology 106, 209–236. Clark, J.S., 1998. Particle motion and the theory of charcoal analysis: source area, transport and sampling. Quaternary Research 30, 67–80. Clark, J.S., Royall, P.D., 1995. Transformation of a northern hardwood forest by aboriginal (Iroquois) fire: charcoal evidence from Crawford Lake, Ontario, Canada. Holocene 5, 1–9. Clark, J.S., Royall, P.D., 1996. Local and regional sediment charcoal for fire in presettlement north-eastern North America. Journal of Ecology 84, 365–382. Clark, R.L., 1983. Pollen and charcoal evidence for the effects of Aboriginal burning on the vegetation of Australia. Archaeology in Oceania 18, 32–37. Colombaroli, D., Marchetto, A., Tinner, W., 2007. Long-term interactions between Mediterranean climate, vegetation and fire regime at Lago di Massaciuccoli (Tuscany, Italy). Journal of Ecology 95, 755–770. Crowley, T.J., 2002. Cycles, cycles, everywhere. Science 295, 1473–1474. Cullen, H.M., deMenocal, P.B., 2000. North Atlantic influence on Tigris–Euphrates streamflow. International Journal of Climatology 20, 853–863. Crutzen, P.J., Andreae, M.O., 1990. Biomass Burning in the Tropics: impact on atmospheric chemistry and biogeochemical cycles. Science 250, 1669–1678. Debret, M., Bout-Roumazeilles, V., Grousset, F., Desmet, M., McManus, J.F., Massei, N., Sebag, D., Petit, J.-R., Copard, Y., Trentesaux, A., 2007. The origin of the 1500-year climate cycles in Holocene North-Atlantic records. Climate of the Past 3, 679–692. Daniau, A.-L., Sánchez-Goñi, M.F., Beaufort, L., Laggoun-Défarge, F., Loutre, M.-F., Duprat, J., 2007. Dansgaard–Oeschger climatic variability revealed by fire emissions in southwestern Iberia. Quaternary Science Reviews 26, 1369–1383. Eastwood, W.J., Roberts, N., Lamb, H.F., 1998. Palaeoecological and archaeological evidence for human occupance in southwest Turkey: The Beyşehir Occupation Phase. Anatolian Studies 48, 69–86. England, A., Eastwood, W.J., Roberts, C.N., Turner, R., Haldon, J.F., in press. Historical landscape change in Cappadocia (central Turkey): a palaeoecological investigation of annually-laminated sediments from Nar lake. The Holocene. Enzel, Y., Bookman, R., Sharon, D., Gvirtzman, H., Dayan, U., Ziv, B., Stein, M., 2003. Late Holocene climates of the Near East deduced from Dead Sea level variations and modern regional winter rainfall. Quaternary Research 60, 263–273. Esin, U., 1991. Salvage excavations at the pre-pottery site of Aşıklı höyük in Central Anatolia. Anatolica 17, 123–164. Freitag, H., 1977. The Pleniglacial, Late Glacial and early Postglacial vegetations of Zeribar and their present-day counterparts. Palaeohistoria 19, 87–95. Gerard, F., Thissen, L., 2002. The Neolithic of Central Anatolia: internal developments and external relations during the 9th–6th millennia CAL BC. Proceedings of the International CANeW Table Ronde, Istanbul, 23–24 November 2001, Istanbul, Ege Yayınlarları. Grove, A.T., Rackham, O., 2001. The nature of Mediterranean Europe: an ecological history. Yale University Press, London.
324
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: PAleontological STatistics software package for education and data analysis. Version 1.36. http://folk.uio.no/ohammer/ past/. Jones, M.D., Leng, M.J., Roberts, N., Türkeş, M., Moyeed, R., 2005. A coupled calibration and modelling approach to the understanding of dry-land lake oxygen isotope records. Journal of Paleolimnology 34, 391–411. Jones, M.D., Roberts, N., Leng, M.J., Türkeş, M., 2006. A high-resolution late Holocene lake isotope record from Turkey and links to North Atlantic and monsoon climate. Geology 34, 361–364. Jones, M.D., Roberts, C.N., Leng, M.J., 2007. Quantifying climatic change through the LGIT based on lake isotope palaeohydrology from central Turkey. Quaternary Research 67, 463–473. Kaniewski, D., Paulissen, E., De Laet, V., Dossche, K., Waelkens, M., 2007. A highresolution Late Holocene landscape ecological history inferred from an intramontane basin in the Western Taurus Mountains, Turkey. Quaternary Science Reviews 26, 2201–2218. Kershaw, A.P., Bush, M.B., Hope, G.S., Weiss, K.-F., Goldammer, J.G., Sanford, R., 1997. The contribution of humans to past biomass burning in the tropics. In: Clark, J.S., Cachier, H., Goldammer, J.G., Stocks, B. (Eds.), Sediment records of biomass burning and global change. NATO ASI Series, vol. 1 51. Springer-Verlag, Berlin, pp. 413–442. Messerli, B., Grsojean, N., Hofer, T., Pfister, C., 2000. From nature-dominated to humandominated environmental changes. Quaternary Science Reviews 19, 459–479. Moros, M., Andrews, J.T., Eberl, D.D., Jansen, E., 2006. Holocene history of drift ice in the northern North Atlantic: evidence for different spatial and temporal modes. Paleoceanography 21, PA2017. doi:10.1029/2005PA001214. Múg ica, F.F., Antón, M.G., Ollero, H.S., 1998. Vegetation dynamics and human impact in the Sierra de Guadarrama, central system, Spain. Holocene 8, 69–82. Múgica, F.F., Antón, M.G., Ruiz, J.M., Juaristi, C.M., Ollero, H.S., 2001. The Holocene history of Pinus forests in the Spanish North Meseta. Holocene 11, 343–358. Naveh, Z., 1974. Effects of fire in the Mediterranean Region. In: Kozlowski, T.T. (Ed.), Fire in Ecosystems. Academic Press, New York, pp. 401–434. Oppo, D., 1997. Millennial climate oscillations. Science 278, 1244–1246. Pausas, J.G., 2004. Changes in fire and climate in the eastern Iberian Peninsula (Mediterranean Basin). Climatic Change 63, 337–350. Peyron, O., Guiot, J., Cheddadi, R., Tarasov, P., Reille, M., de Beaulieu, J.-L., Bottema, S., Andrieu, V., 1998. Climatic reconstruction in Europe for 18,000 yr BP from pollen data. Quaternary Research 49, 183–196. Rasmussen, S.O., Andersen, K.K., Svensson, A.M., Steffensen, J.P., Vinther, B.M., Clausen, H.B., Andersen, M.-L.S., Johnsen, S.J., Larsen, L.B., Bigler, M., Röthlisberger, R., Fischer, H., Goto-Azuma, K., Hansson, M.E., Ruth, U., 2006. A new Greenland ice core chronology for the last glacial termination. Journal of Geophysical Research 111. doi:10.1029/2005JD006079. Roberts, N., 2002. Did prehistoric landscape management retard the post-glacial spread of woodland in Southwest Asia? Antiquity 76, 1002–1010. Roberts, N., Reed, J., Leng, M.J., Kuzucuoğlu, C., Fontugne, M., Bertaux, J., Woldring, H., Bottema, S., Black, S., Hunt, E., Karabıyıkoğlu, M., 2001. The tempo of Holocene climatic change in the eastern Mediterranean region: new high-resolution craterlake sediment data from central Turkey. Holocene 11, 721–736.
Sadori, L., Giardini, M., 2007. Charcoal analysis, a method to study vegetation and climate of the Holocene: The case of Lago di Pergusa (Sicily, Italy). Geobios 40, 173–180. Trabaud, L., Christensen, N.L., Gill, A.M., 1993. Historical biogeography of fire in Temperate and Mediterranean ecosystems. In: Crutzen, P.J., Goldammer, J.G. (Eds.), Fire in the environment: the ecological, atmospheric and climatic importance of vegetation fires. John Wiley & Sons, London, pp. 277–296. Türkeş, M., Erlat, E., 2005. Climatological responses of winter precipitation in Turkey to variability of the North Atlantic Oscillation during the period 1930–2001. Theoretical and Applied Climatology 81, 45–69. Turner, R., 2007. Late Quaternary fire histories in the Eastern Mediterranean region from lake sedimentary micro-charcoals. Ph.D. Thesis, University of Plymouth, Plymouth, UK. Turner, R., Kelly, A., Roberts, N., in press. A critical assessment of the methods used to extract and experimental comparison of microscopic charcoal from lake sediments. In: Fiorentino, G., Magri, D. (Eds), Charcoals from the past: cultural and palaeoenvironmental implications. Proceedings of the Third International Meeting of Anthracology, Cavallino (Lecce), June 2004. BAR International Series, Archaeopress, Oxford, UK. Vannière, B., Colombaroli, D., Marchetto, A., Tinner, W., Chapron, E., 2008. Climate versus human-driven fire regimes in Mediterranean landscapes: the Holocene record of Lago dell'Accesa (Tuscany, Italy). Quaternary Science Reviews 27, 1181–1196. Wanner, H. et al., in press. Mid- to late Holocene climate change — an overview. Quaternary Science Reviews. Whelan, R.J., 1995. The Ecology of Fire. Cambridge University Press, Cambridge. Wick, L., Lemcke, G., Strum, M., 2003. Evidence of Lateglacial and Holocene climatic change and human impact in eastern Anatolia: high-resolution pollen, charcoal, isotopic and geochemical records from laminated sediments of Lake Van, Turkey. Holocene 13, 665–675. Willis, K.J., Bennett, K.D., Haberle, S., in press. Solar influence on Holocene fire history. In: Fiorentino, G., Magri, D. (Eds), Charcoals from the past: cultural and palaeoenvironmental implications. Proceedings of the Third International Meeting of Anthracology, Cavallino (Lecce), June 2004. BAR International Series, Archaeopress, Oxford, UK. Woldring, H., Bottema, S., 2003. The vegetation history of East-Central Anatolia in relation to Archaeology: the Eski Acıgöl pollen evidence compared with the Near Eastern environment. Palaeohistoria 43/44, 1–34. Wright Jr., H.E., 1993. Environmental determinism in near eastern prehistory. Current Anthropology 34, 458–469. Yasuda, Y., Kitagawa, H., Nakagawa, T., 2000. The earliest record of major anthropogenic deforestation in the Ghab Valley, northwest Syria: a palynological study. Quaternary International 73/74, 127–136. van Zeist, W., Bottema, S., 1991. Late Quaternary vegetation of the Near East. Dr. Ludwig Reichert Verlag, Wiesbaden. Zohary, M., 1973. Geobotanical foundations of the Middle East. Fischer, Amsterdam.
Contents lists available at ScienceDirect
Global and Planetary Change 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 / g l o p l a c h a
Climatic pacing of Mediterranean fire histories from lake sedimentary microcharcoal R. Turner a,1, N. Roberts a,⁎, M.D. Jones b a b
School of Geography, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK School of Geography, University of Nottingham, University Park, Nottingham, NG7 2RD, UK
a r t i c l e
i n f o
Article history: Received 26 March 2008 Accepted 25 July 2008 Available online 29 July 2008 Keywords: microcharcoal fire oxygen isotopes Mediterranean pollen
a b s t r a c t The microcharcoal content (particles b 180 µm) of overlapping sedimentary sequences from two crater lake basins in central Turkey are used to reconstruct the regional fire history of the East Mediterranean oak–grass parkland zone from the Last Glacial Maximum to the present-day. These results are correlated with stable isotope and pollen data from the same cores in order to assess the changing role of climate, vegetation and human activity in landscape burning. This indicates that climatically-induced variation in biomass availability was the main factor controlling the timing of regional fire activity during the Last Glacial– Interglacial climatic transition, and again during Mid-Holocene times, with fire frequency and magnitude increasing during wetter climatic phases. Spectral analysis of the Holocene part of the record from Eski Acıgöl indicates significant cyclicity with a periodicity of ~ 1500 years that may be linked with large-scale climate forcing. Although proto-agricultural societies were established in this region as early as 10,000 years ago, it is only during the last two to three millennia that the pacing of wildfire cycles appears to have become decoupled from climate and linked instead to human-induced changes in land cover and fuel load availability. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Under natural conditions fire, climate and biomass are inextricably linked (Pausas, 2004; Whelan, 1995). The nature of the climate (e.g. humid vs. arid) influences the structure and species composition of a vegetation community and therefore fuel availability, which is a powerful influence on fire regimes. In turn, fire is itself a major determinant of seasonally dry ecosystems including the global distribution of tree cover (Bond et al., 2005). People modify the natural patterns of biomass burning through their manipulation and management of vegetation communities, including via practices of fire suppression or promotion (Clark and Royall, 1995), and through accidental or deliberate ignition of “natural” vegetation. Eastern Mediterranean parklands are a classic example of an ecosystem where fire, grazing and other forms of disturbance are fundamental to the maintenance and regeneration of the ecosystem (Grove and Rackham, 2001; Naveh, 1974). In addition, but in contrast to other summer-dry Mediterranean-type regions of the world (Köppen type Cs; e.g. California, S. African Cape), the Mediterranean has a long history of human occupation and impact. There is archaeological evidence of semi-sedentary advanced hunter–gath-
⁎ Corresponding author. Tel.: +44 1752 585965; fax: +44 1752 233054. E-mail address: [email protected] (N. Roberts). 1 Current address: Higher Education Learning Partnerships CETL, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK. 0921-8181/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.gloplacha.2008.07.002
erers during the Last Glacial–Interglacial transition (LGIT), and the Eastern Mediterranean subsequently became one of the earliest centres where Neolithic agriculture developed. Therefore this might be anticipated to have been a region where anthropogenic agencies joined, or even overtook, natural ones in exerting control over regional fire activity at an early date (Roberts, 2002). Charcoal particles have been widely used to reconstruct fire events over multiple timescales in different parts of the world. For example, microscopic charcoal analysis has been central in understanding vegetation change in the North American forests (e.g. Clark and Royall, 1995; 1996) and forest succession in the Western Mediterranean (e.g. Carrión and van Geel, 1999; Múgica et al., 1998; 2001; Sadori and Giardini, 2007; Colombaroli et al., 2007; Vannière et al., 2008), the history of human fire use by pre-European populations in Australia and the Americas (Clark, 1983; Clark and Royall, 1995; Kershaw et al., 1997; Trabaud et al., 1993) and understanding changes in the global carbon cycle with increased biomass burning (e.g. Carcaillet et al., 2002; Crutzen and Andreae, 1990). However, there has been limited research into long-term fire dynamics in the Eastern Mediterranean, and this has generally had low temporal resolution (e.g. Yasuda et al., 2000). In this paper we present a high resolution sedimentary record of regional fire history since the Last Glacial Maximum (LGM) from two nearby crater lake basins located in the oak parkland zone of Central Turkey. In order to assess the changing roles of climate, biomass and people on burning regimes, we compare microcharcoal results with other multi-proxy (pollen and oxygen isotope) data from the same stratigraphic sequences, along with archaeological evidence from the study region.
318
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
2. Study area The principal data we report here derive from Eski Acıgöl (38°33'01”N, 34°32'41”E; 1270 metres above sea level (masl)), a small, former brackish lake, drained in 1972, with no inflow streams and a surface catchment not much larger than the lake itself (Roberts et al., 2001). Its sedimentary record extends back to around the time of the LGM, although it is poorly resolved for the last two millennia. Consequently we overlap this during the late Holocene with a shorter, but highly resolved, sequence from Nar crater lake, located 25 km away. Nar lake (38°22′N, 34°27′E; elevation 1363 masl) is still extant with a maximum water depth of 26 m (Jones et al., 2005, 2006; England et al., in press). Nar is also relatively small (~ 0.7 km2) with a watershed catchment of ~4 km2, and the lake is stratified and forms annual varve layers. Both sites are hydrologically closed in terms of surface outflows, and are sensitive to climatic changes. These lakes are located in the oak parkland zone that covers the Anatolian plateau and western Iran (Zohary, 1973) mostly between 900 and 1500 masl (Fig. 1). The surrounding area would comprise Quercus pubescens, Q. cerris, Pistacia, Crataegus, Prunus, Pyrus, and Juniperus with Festuca, Poa and other grasses, and Artemisia and chenopods common at drier, low-elevation sites (Woldring and Bottema, 2003). Most woodland is degraded today, partly due to grazing by sheep and goat herds, and almost all fertile soils have been converted to agriculture. The regional climate is modified Oro-Mediterranean, being summer-dry, with precipitation ranging between ~ 300 and ~ 600 mm y− 1. Average summer temperatures range between 20 °C and 27 °C, while those in winter fall to 3 °C to −3 °C, reflecting the continentality and elevation of the area. The study sites lie in Cappadocia, which has a well-documented archaeological record of continuous human habitation since Neolithic times. The excavated proto-agricultural Neolithic village of Aşıklı Höyük lies to the south west of Eski Acıgöl and dates to between 10,000 and 9500 cal. years BP (Esin, 1991) (all dates are given in calendar years). The region was important throughout later prehistory for raw materials such as obsidian, it lay close to the centre of the Hittite Empire during 2nd millennium BC, and became a Roman province in 17 AD. During the last two millennia, Cappadocia was especially important during early to mid-Byzantine times, and the again during the early Turkish Selçuk period (England et al., in press). 3. Methods At Eski Acıgöl, a 1566 cm composite sequence was obtained from parallel, overlapping cores taken with a modified Livingstone piston
corer and Eijkelcamp corer with a percussion Cobra motor (ESK 96/99). Palynological analysis was conducted on a parallel sequence taken in 1992 (ESK92) using a Dachnowsky corer (which reached 1420 cm) (Woldring and Bottema, 2003). The ESK96/99 and ESK92 sequences were cross correlated through a series of well-constrained tie points, e.g. tephra horizons. Because the ESK92 sequence is shorter than ESK96/99, published pollen data are not available for the period prior to the Late Glacial Interstadial. At Nar, coring was undertaken in the deepest part of the lake using Glew, 3 m Mackereth and Livingstone lake sediment corers. Multi-proxy studies, including δ18O (measured on precipitated authigenic carbonate) and pollen analyses, have been conducted on cores taken from both crater lake systems, with dating based on Useries (Eski Acıgöl) and varve counting (Nar); (see England et al., in press; Jones et al., 2005; Roberts et al., 2001 for methodological details). 3.1. Microcharcoal analytical methods Fire events produce a pulse in charcoal which is transported away from the fire site and can be incorporated into lake sediments. Identifying individual fire events is dependent on sampling the peaks in charcoal concentration within the sedimentary record; a contiguous sampling strategy was therefore adopted in this study. This differs from microcharcoal counts made on pollen slides which are normally derived from samples taken at discrete depth intervals within a core sequence (e.g. at Eski Acıgöl, pollen samples were taken on average every 18.5 cm; Woldring and Bottema, 2003). For microcharcoal analysis from Eski Acıgöl, contiguous 1 cm3 sediment samples measuring 4.0 × 1.0 × 0.25 cm were extracted from the full length of the ESK 96/99 composite core sequence, apart from 0–76 and 1267–1294 cm for which there was no core recovery. A higher sampling resolution of 2.0 × 1.0 × 0.5 cm was applied to selected periods of major climatic change, e.g. Late Glacial–Holocene transition. Contiguous samples were also analysed at Nar, but in this case by laminae age rather than depth interval. For the last 100 varve years, thin blocks of 10 laminations were used (i.e. 1–10, 11–20 VY, etc), while the rest of the sequence sampled contiguously every 20 laminations (i.e. 101–120, 121–140 VY, etc). After adding Lycopodium tablets to enable estimation of charcoal concentrations, samples were prepared using density separation (Turner, 2007). This preparation method was found experimentally by Turner et al. (in press) to have a higher recovery than other published methods for the fine charcoal fraction, and more than ten times the recovery of standard pollen preparation method. Samples
Fig. 1. Location of study sites. Distribution of parkland ecosystem from Zohary (1973).
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
were initially treated with 10% HCl to disaggregate the sediments and then passed through a 180 µm mesh to remove (and retain) the larger, potentially more fragile charcoal particles that could fragment during the preparation and result in a potential overestimation of the charcoal concentration of a sample. Lithium heteropolytungstate with a specific gravity of 2.5 was used to separate the microscopic charcoal particles from the sediment matrix. These were then mounted on a standard microscope slide, and all charcoal particles within each field of view were counted using a BX100 high power light microscope at 200× magnification, until a count of 100 Lycopodium spores was reached. Microcharcoal particles were identified based on a set of diagnostic criteria, primarily that they had to be jet black in colour, with straight edges, and with the presence of a blue hue on edges (Turner et al., in press). 4. Results At Eski Acıgöl the ESK 96/99 core sequence recovered nonlaminated carbonate-rich lake sediments in the upper 607 cm, underlain by annually laminated limnic deposits to at least 1566 cm. 14 C dates from this sequence have been aged by ~ 3000 years due to volcanic outgassing, so that the chronology is primarily based on a series of U-series dates obtained from the composite core sequence (ESK 96/99) (see Roberts et al., 2001 for details and discussion). These are supported by preliminary higher precision U-Th Thermally Ionised Mass Spectrometer age determinations (J. Dorale, pers comm.), by varve counting and by correlations with other 14C dated pollen and stable isotope records from central Anatolia. They show that the core sequence covers the Holocene and the LGIT. Changes in the proxy records correspond to events similar in time and magnitude to the European Bölling–Alleröd and Younger Dryas events. However, it is
319
currently not possible to make firm conclusions as to synchroneity with these events, which are therefore named here as the Late Glacial Interstadial and Late Glacial Stadial (Jones et al., 2007). According to the age model used here, the onset of the Holocene at Eski Acıgöl is dated to 12,080 BP, which compares with an age of 11,700 BP from Greenland ice cores (Rasmussen et al., 2006) and other sources. It is therefore possible that the Eski Acıgöl core chronology overestimates true age by about 3–4%. In this study a linear age-depth sediment accumulation rate of ~12 yr / cm− 1 is used between 1.6 (175 cm) and 16.2 ka BP (1409 cm); before this, sedimentation rate appears to be significantly slower. We use a basal age of 23 ka BP for the ESK 96/99 sequence, but the precise age of the lowest part of the sequence is not well defined and we do not consider the period prior to 16 ka BP in detail here. The potential effects of using different age models are discussed in relation to time series analysis below. The 376 cm long sediment sequence from Nar lake is laminated throughout and consists of 1725 couplets of summer precipitated calcium carbonate and diatom-rich organic layers deposited between autumn and spring. 210Pb and 137Cs dating of the top 50 cm of this sequence, together with the analysis of modern sediments in seston traps show these couplets to be annual (Jones et al., 2005). Dating of the rest of the core sequence is therefore based on laminae counts given in varve years before AD 2001 converted to a calendar time scale (see Jones et al., 2006 for further details). δ18O values recorded from the two crater lake sequences have been linked to changes in regional water balance with negative δ18O values reflecting periods of overall wetter climate. Jones et al. (2005, 2007) used isotope mass balance modelling to calibrate these isotope data climatically. For Eski Acıgöl, they calculated that precipitation was 20– 40% higher than at present during the Late Glacial Interstadial and again in the early Holocene, and 25–40% lower than present during
Fig. 2. Microcharcoal influx data for Eski Acıgöl, Turkey, compared to summary pollen and oxygen isotope data (from Woldring and Bottema, 2003 and Roberts et al., 2001, respectively). Stars indicate U-series dating control points. Two short-lived early Holocene isotopic excursions are not shown (see Roberts et al., 2001 for discussion).
320
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
glacial and Late Glacial Stadial times given temperatures 5 to 6 °C below present. During the Mid-Holocene there is a positive shift in the δ18O values reflecting climatic aridification towards modern conditions. This trend in the δ18O values coincides with a fall in lake levels whereby laminated sediments cease to form. A diatom-inferred increase in the salinity of the lake waters occurred between ca.7000 and 4000 BP (Roberts et al., 2001). At Nar lake, δ18O data show more positive values, inferred to indicate drier climatic conditions, from 1700–1500 BP and again 600 to 40 BP, with more negative isotopic values, and a wetter climate, between 1440–1250 BP, 1000–600 BP, and since AD1960 (Jones et al., 2006). Pollen analyses from Eski Acıgöl and from other sites (e.g. Akgöl, Konya; Bottema and Woldring, 1984) show that the Anatolian plateau was dominated by largely treeless Artemisia-chenopod steppe during Late Glacial times. The onset of the Holocene was associated with the rapid expansion of grasses, an equally abrupt reduction in the proportion of Artemisia and chenopods, and a more gradual increase in Quercus robur-type pollen. The Early Holocene was characterised by open grass parklands including Pistacia as an important element, with continued expansion of oak and some mesic tree species. The maximum extent of tree cover was not reached until ca.5500 BP, after which it declined to be replaced by steppic herbs and Poaceae (Fig. 2). At Nar, pollen analyses carried out at 20 varve year intervals (England et al., in press) indicate four principal land-use phases during the last 1725 years: (i) an agrarian landscape characterised by cereals and tree crops prior to 1330 BP, marking the later part of the so-called Beyşehir Occupation phase, (ii) a period of secondary woodland establishment from 1330–1050 BP coinciding with a “late Antiquity Dark Age”, (iii) renewed deforestation with the re-establishment of cereal agriculture and pastoralism after 1050 BP, and (iv) agricultural intensification during modern times (AD 1830 to present)(Fig. 3).
4.1. Microcharcoals Relative to other SW Asian sequences analysed using the same methods (Turner 2007), overall charcoal concentration and influx values are low at both Nar and Eski Acıgöl. This is inferred to represent mainly long distance transport of charcoal particles to the lakes from the wider environment. The absence of surface inflows means that charcoal will have entered the lakes primarily through aeolian transport processes which usually sample a regional fire signal (Clark, 1998). The inferred regional fire history signal is further supported by a paucity of macroscopic charcoal particles in the N180 µm fraction, which would be associated with fires occurring within the immediate vicinity of the lakes. The lowest charcoal values at Eski Acıgöl were recorded during the Glacial and the Late Glacial Stadial whereas charcoal concentrations increased during the Late Glacial Interstadial and the Holocene (Fig. 2). Charcoal concentration and influx values have fluctuated in quasicyclical fashion throughout the Holocene, with the maximum concentration values occurring 12,000 to 9000 BP and 6000 to 5000 BP. Microcharcoal influx values at Nar prior to the 20th century are comparable to those at Eski Acıgöl during the late Holocene, with lowest values from 1725 to 1370 BP and 550 to 250 BP (Fig. 3). Microcharcoal counts on pollen slides from the Nar sequence show a broadly similar pattern of change to those using heavy liquid separation prior to ~ 200 BP, although absolute values are lower (England et al., in press). In order to explore quantitatively the multi-millennial record of landscape burning, the microcharcoal data for Eski Acıgöl have been subject to time series analysis. Spectral analysis was undertaken to examine periodicities in this microcharcoal time series using PAST (PAlaeontological STatistics; Hammer et al., 2001). Although a number
Fig. 3. Late Holocene microcharcoal influx for Nar, compared to summary pollen and oxygen isotope data from the same sequence (from England et al., in press; Jones et al., 2006, respectively). Pollen zones are from England et al. (in press).
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
of time windows have been analysed, we focus on the Holocene part of the record which spans the period back to 12,080 BP according to the age model used here. This comprises an unbroken series of 213 contiguous samples each representing an aggregated signal of regional biomass burning over 46 ± 15 years. Spectral analysis of this time series shows significant periodicities (those with periodicities below one-third of the total time series and at the 95% confidence limit or better) at (1) 3065, (2) 2018, (3) 1561, and (4) 1034 years (Fig. 4). Of these, cycle (3) is most prominent. We have tested the validity of the periodicities found in this microcharcoal record using a series of age models to track the duration and strength of different cycles, and have found cycle (3) to be robust using all realistic age models. Peak to trough microcharcoal concentration values in the Eski Acıgöl record typically fluctuate by an order of magnitude, which is too large to be explained by any quasi-periodic changes in sediment accumulation rate. There is also no evidence of long-cycle variations in sedimentation rate from varve thickness measurements for the early-Mid-Holocene (unpublished data). Because the onset of the Holocene is very clearly marked in the Eski Acıgöl record on the basis of lithology, pollen and stable isotopes, an alternative age model might use an age of 11,700 BP rather than 12,080 BP for Late Glacial: Holocene transition. In this case, the dominant length for cycle (3) would be reduced to 1431 years, assuming that the age estimate at 175 cm (1600 BP) is also correct. The strength of this ~1500 year cycle implies that the frequency and intensity of wildfire burning in the central Anatolian oak parkland ecosystem has been non-random through time during the Holocene. 5. Discussion Throughout these two overlapping microcharcoal sequences there is evidence of a close relationship between regional-scale fire activity and pollen-inferred shifts in vegetation composition and biomass. During the LGIT the Eski Acıgöl record shows an inverse relationship between the abundance of pollen indicative of cold steppe (Artemisia spp and Chenopodiaceae) and wildfire frequency (Fig. 2). During the Glacial period, and again during the Late Glacial Stadial, charcoal concentrations in the lake sediments were low, indicating that few wildfires were occurring on the Anatolian plateau. Although precise modern analogues are lacking for the vegetation at this time (Peyron et al., 1998), Freitag (1977) proposed a broad similarity with that found today in the semidesert regions of High Asia. The Glacial-age cold steppe appears to have supported only low levels of biomass, and therefore fuel availability, in turn linked to cold and dry climatic conditions. Pollen data show that the vegetation elsewhere in the Eastern Mediterranean at this time was also dominated by steppic herbs and that tree cover was restricted to isolated pockets of favourable conditions (van Zeist and Bottema, 1991). Infrequent wildfires coupled with low biomass and therefore fuel load availability may have been characteristic of the region as a whole at and immediately after the LGM. By contrast, there is a positive correlation between grass pollen and microcharcoal frequency during the LGIT (r = +0.429, p = 0.01, for
Fig. 4. Spectral analysis of the Holocene microcharcoal record for Eski Acıgöl. Horizontal dotted lines represent 0.05 and 0.01 confidence limits. Numbered peaks represent cycles at (1) 3065, (2) 2018, (3) 1561, and (4) 1034 years.
321
the period 15 to 8 ka BP), implying that summer grass fires accounted for a significant part of the regional atmospheric charcoal flux during the Late Glacial Interstadial and early Holocene. This would be consistent with Wright (1993) who argued that the current summerdry precipitation regime did not occur in glacial times, but became reestablished during the Late Glacial Interstadial and again, more permanently, at the start of the Holocene. Microcharcoal concentrations in Eski Acıgöl increased abruptly at the start of the Holocene, with an equally rapid switch from herb-steppe to grassland, and — almost certainly — a significant increase in biomass. The initial early Holocene peak in microcharcoals preceded the establishment of Neolithic settlement in Cappadocia (~10,300 Cal BP), and occurred at a time when there is very little archaeological evidence for human occupation of the central Anatolian plateau (Gerard and Thissen, 2002). It is therefore unlikely that this earliest Holocene wildfire maximum can be attributed to deliberate anthropogenic landscape burning (cf. Roberts, 2002). On the other hand, the positive covariance between microcharcoal concentration values and Poaceae pollen during the first four millennia of the Holocene, along with the slow increase in oak pollen, would be consistent with wildfires maintaining grassland at the expense of woodland. Isotope-inferred precipitation at this time was ~20% higher than at present (Jones et al., 2007), and landscape burning may consequently have acted to slow down what could otherwise have been a rapid post-glacial encroachment of trees across the Anatolian plateau. An Early Holocene peak in fire activity is also indicated in microcharcoal counts from the Van and Ghab pollen records in eastern Anatolia and northwest Syria, respectively (Wick et al., 2003; Yasuda et al., 2000). Following this phase, charcoal concentration values at Eski Acıgöl declined between 9500 and 6000 years BP, equivalent archaeologically to the ceramic Neolithic and Chalcolithic periods. Around this time there were changes in the vegetation community which affected fuel availability and therefore fire occurrence. This involved a continued expansion of oak woodland that put increased pressure on water availability within the vegetation community resulting in the partial replacement of the Poaceae with steppic herbs (Woldring and Bottema, 2003), possibly also linked to grazing pressure by domestic livestock. During the Mid-Holocene oak rather than grass pollen has the clearest positive relationship to microcharcoal flux, especially during maximum Q. robur-type values between 5 and 6 ka BP, and implying a switch in wildfire fuel source from grass to woody biomass. After ~ 5000 BP, there was a permanent decline in arboreal pollen values which coincided both with an overall reduction in regional fire occurrence and an increase in both steppic herbs and ruderal plant species. This has been inferred to reflect increased cultivation and grazing of the landscape during Bronze Age times (Woldring and Bottema, 2003) but it also coincided with a stepwise aridification of the climate as indicated in stable isotope, lake level and salinity indicators from Eski Acıgöl (Roberts et al., 2001). Both factors would have led to changes in the composition of the vegetation community and a reduction in fuel availability. If pollen and archaeological indicators of human-induced land-use change fail show to show a clear and direct relationship to landscape burning activity in Cappadocia during the Early and Mid-Holocene, what of climatic variations, as measured by δ18O data? During the LGIT, microcharcoal influx peaks coincide with negative δ18O values indicative of higher precipitation, notably during the Late Glacial interstadial and at the start of the Holocene (Fig. 2). Greater moisture availability in turn led to an increase in plant biomass and to more frequent and intense wildfire events. However, the Eski Acıgöl oxygen isotope record also shows a longer-term trend that may be linked to the ~22 ka precessional cycle. In order to remove this long-term trajectory, both microcharcoal and δ18O data have been detrended (although in practice, detrending has little effect on the former). The resulting pattern of residuals (Fig. 5) shows that peaks coincide for the two indicators, not only during the LGIT, but also for significant portions of
322
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
the Holocene. A 4 ka moving-window correlation between detrended microcharcoal influx and δ18O exhibits statistically significant r values N −0.5 for all time periods centred between 18 and 10 ka BP, and again shows r values around −0.5 for the Mid-Holocene period centred on 4.5 ka BP. The weaker statistical relationship for the intervening period may be partly linked to gaps in the δ18O record during the early Holocene. The fact that periods of wetter climate and increased landscape burning have coincided at centennial–millennial timescales does not imply that the nature of any causal relationship was constant through time. In contrast to the LGIT, for example, the Mid-Holocene period was most likely dominated by woody biomass as the principal fuel source for wildfires. It is also conceivable that periods of wetter climate encouraged an expansion of human settlement and rural landuse, which may have impacted vegetation cover and landscape burning frequency. None the less, data from Eski Acıgöl are consistent with the hypothesis that wet-dry oscillations in climate acted as the pacemaker for biomass burning intensity in the oak–grass parkland biome of central Anatolia up until at least ~3000 BP. This record of biomass burning is of further significance because of the clear ~1500 year cyclicity that is displayed through the Holocene, which in turn implies a similar dominant periodicity for the underlying climatic forcing. A ~ 1470 year climatic periodicity was recognised by Bond et al. (1997, 2001) from the flux of ice-rafted debris (IRD) in North Atlantic deep-sea cores, and has been widely discussed elsewhere in the literature (e.g. Oppo, 1997; Crowley, 2002), in some cases linked to changes in solar variability. However, it is previously little known from Holocene terrestrial records in the Mediterranean (e.g. Allen et al., 2002). In reality, North Atlantic IRD variations during the Holocene were quasi-periodic rather than truly cyclical in nature and have not been easily replicated (Moros et al., 2006). Additionally, a re-analysis by Debret et al. (2007) showed the
IRD data series to be non-stationary through time, comprising composite cycles of 1450 and 2400 years in length. None the less, a well-established causal relationship exists between the mode of the North Atlantic Oscillation and Mediterranean precipitation (e.g. Enzel et al., 2003; Türkeş and Erlat, 2005). Precipitation in central Anatolia today falls predominantly in winter and spring seasons linked to mid-latitude westerly depressions, ultimately originating in the North Atlantic. Cullen and deMenocal (2000) suggested that Holocene cooling events, such as those associated with Bond “cycles”, could explain millennial-scale climate variability in the East Mediterranean region, notably periods of drought. If North Atlantic cooling events, marked by increased IRD flux, did lead to reduced moisture availability in central Anatolia, then — conversely — IRD minima should coincide with periods of wetter climate, increased biomass availability and more frequent wildfires. Comparison of IRD minima with microcharcoal peaks from Eski Acıgöl, shown in Fig. 5, is hindered by chronological uncertainties in both records. Plausible matches exist for IRD minima at 2.0–2.6, 3.6, 8.0, 9.8 and 10.9 ka BP in the Turkish wildfire record, especially if the Eski Acıgöl chronology is matched to that for the Greenland ice cores (i.e. 11.7 ka for the start of the Holocene). On the other hand, no such matches are evident for the Holocene IRD minima at 7.1 and 8.9 ka BP, while the major peak in wildfire activity at around 5.5 ka BP coincides with an IRD maximum, rather than minimum. A causal linkage between the two therefore seems inconclusive at present. Willis et al. (in press) compared eight microcharcoal records of Holocene wildfire frequencies from different localities worldwide and failed to find consistent dominant spectral periodicites or any clear relationship with inferred solar maxima or minima. Notwithstanding this, the Eski Acıgöl fire history record does give support to the increasing array of evidence for a ~ 1500 year periodicity in proxy-climate data from
Fig. 5. Detrended microcharcoal and oxygen isotope residuals for Eski Acıgöl using 2nd order polynomials, along with moving-window correlation coefficients over 4 ka intervals. Stars indicate the timing of N Atlantic IRD minima (from Bond et al., 2001).
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
northern mid-latitudes during the Holocene (Wanner et al., in press, Figs. 5 and 19). At what point in time did wildfires become decoupled from climate and linked instead to human-induced changes in land cover and fuel load availability? Here, the Late Holocene record from Nar lake is informative. As shown in Fig. 3, the phase of highest microcharcoal influx prior to the 20th century occurred during local pollen zone III (1320–1150 BP; AD 680–950), which saw rural land abandonment and re-afforestation by oak woodland (England et al., in press). This led to an increase in woody biomass and fuel load availability for combustion in wildfires. However, the δ18O record from the same core sequence shows that this phase was not directly coincident with a period of greater moisture availability, as were microcharcoal influx peaks earlier in the Holocene at nearby Eski Acıgöl. In other words, by 1700 years ago, landscape burning in this region was no longer coupled directly to climate variations. Although the timing of the switchover from a nature-dominated to a humandominated wildfire system is not known precisely (c.f. Messerli et al., 2000), it may well have coincided with the onset of the Beyşehir Occupation phase in this region (Eastwood et al., 1998; Kaniewski et al., 2007). This period of human landscape transformation is clearly recognised in pollen records throughout western Anatolia and commenced around 3000 BP. The Nar record also highlights the fact that microcharcoal influx, and by inference landscape burning, was greater during the last century than at any time during the preceding 20,000 years. 6. Conclusions Microcharcoal analysis from Eski Acıgöl and Nar crater lake sediment cores provide a ~20 ka long record of regional-scale fire events with multi-decadal precision. Pollen and microcharcoal influx into the lakes was dominated by regional-scale atmospheric inputs, so that the sites consequently represent a spatially aggregated record of land cover and biomass burning across the oak parkland zone in which they are located. Throughout both sequences there is evidence of a close relationship between fire activity and the composition of the vegetation community, which in turn reflects fuel availability. In this biome, wildfires appear to have been fuel-limited throughout the period since the LGM. This was also true in Iberia during the Last Glacial period (Daniau et al., 2007) and in the North American prairie where clear links have been identified between phases of greater humidity promoting vegetation growth, increasing fuel load and fire occurrence (Brown et al., 2005; Camill et al., 2003). This relationship stands in contrast to that in better watered and more wooded environments in Mediterranean regions, such as Italy, where microcharcoal records indicate that landscape burning was most frequent during climatically arid phases of the Holocene, showing that here wildfires were not fuel-limited (Sadori and Giardini, 2007; Vannière et al., 2008). By employing a multi-proxy approach, we have been able to compare directly proxies for fire (microcharcoals), land-use (pollen) and climate (δ18O), without the uncertainty that can result from correlating separate records. In that regard, our data show a close statistical correlation between centennial to millennial-scale climatic fluctuations (wet to dry) and biomass burning, not only during the LGIT, but also — and more surprisingly — during the Mid-Holocene. Microcharcoal flux during the Holocene has been strongly cyclical in character, with spectral analysis showing a dominant 1500 year periodicity that is likely climatically controlled. As the Holocene progressed, the impact of people upon regional fire activity became increasingly pronounced. However, and notwithstanding abundant archaeological evidence for agricultural and pastoral land-use changes from 10 ka BP onwards, the switch from a climate- to a humanregulated fire regime appears to have taken place only late in the Holocene.
323
Acknowledgements We are grateful for the generous assistance from the late Sytze Bottema, to whom this paper is dedicated, and from Dan Charman, Warren Eastwood, Ann England, and Henk Woldring, and also from Tim Absalom for assistance with the final drafting of the diagrams. This work was supported by Leverhulme grant F/00568/1 and by the British Institute in Ankara.
References Allen, J.R.M., Watts, W.A., McGee, E., Huntley, B., 2002. Holocene environmental variability—the record from Lago Grande di Monticchio, Italy. Quaternary International 88, 69–80. Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., deMonocal, P., Priore, P., Cullen, H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in North Atlantic Holocene and Glacial climates. Science 278, 1257–1266. Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N., Showers, W., Hoffman, S., LottiBond, R., Hajdas, I., Bonani, G., 2001. Persistent solar influence on north Atlantic climate during the Holocene. Science 294, 2130–2136. Bond, W.J., Woodward, F.I., Midgley, G.F., 2005. The global distribution of ecosystems in a world without fire. New Phytologist 165, 525–538. Bottema, S., Woldring, H., 1984. Late Quaternary vegetation and climate of southwestern Turkey, Part II. Palaeohistoria 26, 123–149. Brown, K.J., Clark, J.S., Grimm, E.C., Donovan, J.J., Mueller, P.G., Hansen, B.C.S., Stefanova, I., 2005. Fire cycles in North American interior grasslands and their relation to prairie drought. PNAS 102, 8865–8870. Camill, P., Umbamhowar, C.E., Teed, R., Geiss, C.E., Aldinger, J., Dvorak, L., Kenning, J., Limmer, J., Walkup, K., 2003. Late-glacial and Holocene climatic effects on fire and vegetation dynamics at the prairie-forest ecotone in south-central Minnesota. Journal of Ecology 91, 822–836. Carcaillet, C., Almquist, H., Asnong, H., Bradshaw, R.H., Carrion, J.S., Gaillard, M.J., Gajewski, K., Haas, J.N., Haberle, S.G., Hadorn, P., Muller, S.D., Richard, P.J., Richoz, I., Rosch, M., Sanchez Goni, M.F., von Stedingk, H., Stevenson, A.C., Talon, B., Tardy, C., Tinner, W., Tryterud, E., Wick, L., Willis, K.J., 2002. Holocene biomass burning and global dynamics of the carbon cycle. Chemosphere 49, 845–863. Carrión, J.S., van Geel, B., 1999. Fine-resolution Upper Weichselian and Holocene palynological record from Navarrés (Valencia, Spain) and a discussion about factors of Mediterranean forest succession. Review of Palaeobotany and Palynology 106, 209–236. Clark, J.S., 1998. Particle motion and the theory of charcoal analysis: source area, transport and sampling. Quaternary Research 30, 67–80. Clark, J.S., Royall, P.D., 1995. Transformation of a northern hardwood forest by aboriginal (Iroquois) fire: charcoal evidence from Crawford Lake, Ontario, Canada. Holocene 5, 1–9. Clark, J.S., Royall, P.D., 1996. Local and regional sediment charcoal for fire in presettlement north-eastern North America. Journal of Ecology 84, 365–382. Clark, R.L., 1983. Pollen and charcoal evidence for the effects of Aboriginal burning on the vegetation of Australia. Archaeology in Oceania 18, 32–37. Colombaroli, D., Marchetto, A., Tinner, W., 2007. Long-term interactions between Mediterranean climate, vegetation and fire regime at Lago di Massaciuccoli (Tuscany, Italy). Journal of Ecology 95, 755–770. Crowley, T.J., 2002. Cycles, cycles, everywhere. Science 295, 1473–1474. Cullen, H.M., deMenocal, P.B., 2000. North Atlantic influence on Tigris–Euphrates streamflow. International Journal of Climatology 20, 853–863. Crutzen, P.J., Andreae, M.O., 1990. Biomass Burning in the Tropics: impact on atmospheric chemistry and biogeochemical cycles. Science 250, 1669–1678. Debret, M., Bout-Roumazeilles, V., Grousset, F., Desmet, M., McManus, J.F., Massei, N., Sebag, D., Petit, J.-R., Copard, Y., Trentesaux, A., 2007. The origin of the 1500-year climate cycles in Holocene North-Atlantic records. Climate of the Past 3, 679–692. Daniau, A.-L., Sánchez-Goñi, M.F., Beaufort, L., Laggoun-Défarge, F., Loutre, M.-F., Duprat, J., 2007. Dansgaard–Oeschger climatic variability revealed by fire emissions in southwestern Iberia. Quaternary Science Reviews 26, 1369–1383. Eastwood, W.J., Roberts, N., Lamb, H.F., 1998. Palaeoecological and archaeological evidence for human occupance in southwest Turkey: The Beyşehir Occupation Phase. Anatolian Studies 48, 69–86. England, A., Eastwood, W.J., Roberts, C.N., Turner, R., Haldon, J.F., in press. Historical landscape change in Cappadocia (central Turkey): a palaeoecological investigation of annually-laminated sediments from Nar lake. The Holocene. Enzel, Y., Bookman, R., Sharon, D., Gvirtzman, H., Dayan, U., Ziv, B., Stein, M., 2003. Late Holocene climates of the Near East deduced from Dead Sea level variations and modern regional winter rainfall. Quaternary Research 60, 263–273. Esin, U., 1991. Salvage excavations at the pre-pottery site of Aşıklı höyük in Central Anatolia. Anatolica 17, 123–164. Freitag, H., 1977. The Pleniglacial, Late Glacial and early Postglacial vegetations of Zeribar and their present-day counterparts. Palaeohistoria 19, 87–95. Gerard, F., Thissen, L., 2002. The Neolithic of Central Anatolia: internal developments and external relations during the 9th–6th millennia CAL BC. Proceedings of the International CANeW Table Ronde, Istanbul, 23–24 November 2001, Istanbul, Ege Yayınlarları. Grove, A.T., Rackham, O., 2001. The nature of Mediterranean Europe: an ecological history. Yale University Press, London.
324
R. Turner et al. / Global and Planetary Change 63 (2008) 317–324
Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: PAleontological STatistics software package for education and data analysis. Version 1.36. http://folk.uio.no/ohammer/ past/. Jones, M.D., Leng, M.J., Roberts, N., Türkeş, M., Moyeed, R., 2005. A coupled calibration and modelling approach to the understanding of dry-land lake oxygen isotope records. Journal of Paleolimnology 34, 391–411. Jones, M.D., Roberts, N., Leng, M.J., Türkeş, M., 2006. A high-resolution late Holocene lake isotope record from Turkey and links to North Atlantic and monsoon climate. Geology 34, 361–364. Jones, M.D., Roberts, C.N., Leng, M.J., 2007. Quantifying climatic change through the LGIT based on lake isotope palaeohydrology from central Turkey. Quaternary Research 67, 463–473. Kaniewski, D., Paulissen, E., De Laet, V., Dossche, K., Waelkens, M., 2007. A highresolution Late Holocene landscape ecological history inferred from an intramontane basin in the Western Taurus Mountains, Turkey. Quaternary Science Reviews 26, 2201–2218. Kershaw, A.P., Bush, M.B., Hope, G.S., Weiss, K.-F., Goldammer, J.G., Sanford, R., 1997. The contribution of humans to past biomass burning in the tropics. In: Clark, J.S., Cachier, H., Goldammer, J.G., Stocks, B. (Eds.), Sediment records of biomass burning and global change. NATO ASI Series, vol. 1 51. Springer-Verlag, Berlin, pp. 413–442. Messerli, B., Grsojean, N., Hofer, T., Pfister, C., 2000. From nature-dominated to humandominated environmental changes. Quaternary Science Reviews 19, 459–479. Moros, M., Andrews, J.T., Eberl, D.D., Jansen, E., 2006. Holocene history of drift ice in the northern North Atlantic: evidence for different spatial and temporal modes. Paleoceanography 21, PA2017. doi:10.1029/2005PA001214. Múg ica, F.F., Antón, M.G., Ollero, H.S., 1998. Vegetation dynamics and human impact in the Sierra de Guadarrama, central system, Spain. Holocene 8, 69–82. Múgica, F.F., Antón, M.G., Ruiz, J.M., Juaristi, C.M., Ollero, H.S., 2001. The Holocene history of Pinus forests in the Spanish North Meseta. Holocene 11, 343–358. Naveh, Z., 1974. Effects of fire in the Mediterranean Region. In: Kozlowski, T.T. (Ed.), Fire in Ecosystems. Academic Press, New York, pp. 401–434. Oppo, D., 1997. Millennial climate oscillations. Science 278, 1244–1246. Pausas, J.G., 2004. Changes in fire and climate in the eastern Iberian Peninsula (Mediterranean Basin). Climatic Change 63, 337–350. Peyron, O., Guiot, J., Cheddadi, R., Tarasov, P., Reille, M., de Beaulieu, J.-L., Bottema, S., Andrieu, V., 1998. Climatic reconstruction in Europe for 18,000 yr BP from pollen data. Quaternary Research 49, 183–196. Rasmussen, S.O., Andersen, K.K., Svensson, A.M., Steffensen, J.P., Vinther, B.M., Clausen, H.B., Andersen, M.-L.S., Johnsen, S.J., Larsen, L.B., Bigler, M., Röthlisberger, R., Fischer, H., Goto-Azuma, K., Hansson, M.E., Ruth, U., 2006. A new Greenland ice core chronology for the last glacial termination. Journal of Geophysical Research 111. doi:10.1029/2005JD006079. Roberts, N., 2002. Did prehistoric landscape management retard the post-glacial spread of woodland in Southwest Asia? Antiquity 76, 1002–1010. Roberts, N., Reed, J., Leng, M.J., Kuzucuoğlu, C., Fontugne, M., Bertaux, J., Woldring, H., Bottema, S., Black, S., Hunt, E., Karabıyıkoğlu, M., 2001. The tempo of Holocene climatic change in the eastern Mediterranean region: new high-resolution craterlake sediment data from central Turkey. Holocene 11, 721–736.
Sadori, L., Giardini, M., 2007. Charcoal analysis, a method to study vegetation and climate of the Holocene: The case of Lago di Pergusa (Sicily, Italy). Geobios 40, 173–180. Trabaud, L., Christensen, N.L., Gill, A.M., 1993. Historical biogeography of fire in Temperate and Mediterranean ecosystems. In: Crutzen, P.J., Goldammer, J.G. (Eds.), Fire in the environment: the ecological, atmospheric and climatic importance of vegetation fires. John Wiley & Sons, London, pp. 277–296. Türkeş, M., Erlat, E., 2005. Climatological responses of winter precipitation in Turkey to variability of the North Atlantic Oscillation during the period 1930–2001. Theoretical and Applied Climatology 81, 45–69. Turner, R., 2007. Late Quaternary fire histories in the Eastern Mediterranean region from lake sedimentary micro-charcoals. Ph.D. Thesis, University of Plymouth, Plymouth, UK. Turner, R., Kelly, A., Roberts, N., in press. A critical assessment of the methods used to extract and experimental comparison of microscopic charcoal from lake sediments. In: Fiorentino, G., Magri, D. (Eds), Charcoals from the past: cultural and palaeoenvironmental implications. Proceedings of the Third International Meeting of Anthracology, Cavallino (Lecce), June 2004. BAR International Series, Archaeopress, Oxford, UK. Vannière, B., Colombaroli, D., Marchetto, A., Tinner, W., Chapron, E., 2008. Climate versus human-driven fire regimes in Mediterranean landscapes: the Holocene record of Lago dell'Accesa (Tuscany, Italy). Quaternary Science Reviews 27, 1181–1196. Wanner, H. et al., in press. Mid- to late Holocene climate change — an overview. Quaternary Science Reviews. Whelan, R.J., 1995. The Ecology of Fire. Cambridge University Press, Cambridge. Wick, L., Lemcke, G., Strum, M., 2003. Evidence of Lateglacial and Holocene climatic change and human impact in eastern Anatolia: high-resolution pollen, charcoal, isotopic and geochemical records from laminated sediments of Lake Van, Turkey. Holocene 13, 665–675. Willis, K.J., Bennett, K.D., Haberle, S., in press. Solar influence on Holocene fire history. In: Fiorentino, G., Magri, D. (Eds), Charcoals from the past: cultural and palaeoenvironmental implications. Proceedings of the Third International Meeting of Anthracology, Cavallino (Lecce), June 2004. BAR International Series, Archaeopress, Oxford, UK. Woldring, H., Bottema, S., 2003. The vegetation history of East-Central Anatolia in relation to Archaeology: the Eski Acıgöl pollen evidence compared with the Near Eastern environment. Palaeohistoria 43/44, 1–34. Wright Jr., H.E., 1993. Environmental determinism in near eastern prehistory. Current Anthropology 34, 458–469. Yasuda, Y., Kitagawa, H., Nakagawa, T., 2000. The earliest record of major anthropogenic deforestation in the Ghab Valley, northwest Syria: a palynological study. Quaternary International 73/74, 127–136. van Zeist, W., Bottema, S., 1991. Late Quaternary vegetation of the Near East. Dr. Ludwig Reichert Verlag, Wiesbaden. Zohary, M., 1973. Geobotanical foundations of the Middle East. Fischer, Amsterdam.