The Holocene18,5 (2008) pp. 667–677 The Mt Logan Holocene–late Wisconsinan isotope record: tropical Pacific–Yukon connections David Fisher,1* Erich Osterberg,3 Art Dyke,1 Dorthe Dahl-Jensen,5 Mike Demuth,1 Christian Zdanowicz,1 Jocelyne Bourgeois,1 Roy M. Koerner,1 Paul Mayewski,3 Cameron Wake,2 Karl Kreutz,3 Eric Steig,4 James Zheng,1 Kaplan Yalcin,6 Kumiko Goto-Azuma,7 Brian Luckman8 and Summer Rupper9 (1Geological Survey of Canada, (NRCan), 601 Booth Street, Ottawa, Ontario K1A 0E8, Canada; 2Climate Change Research Center, EOS, Morse Hall, University of New Hampshire,Durham NH 03824, USA; 3Climate Change Institute and Department of Earth Sciences, University of Maine, Orono ME 04469, USA; 4Quaternary Research Center, ESS, 19 Johnson Hall, Box 1360, University of Washington, Seattle WA 98195-1360, USA; 5Niels Bohr Institute, Juliane Maries Vej 30, University of Copenhagen, DK-2100, Copenhagen East, Denmark; 6Department of Geosciences, 102D Wilkinson Hall, Oregon State University, Corvallis OR 97331-5506, USA; 7NIPR, Kaga 1-9-10, Itabashi-ku, National Institute of Polar Research,Tokyo 173-8515, Japan; 8Department of Geography, University of Western Ontario, London, Ontario N6F 5C2, Canada; 9Department of Geological Sciences, Brigham Young University, S389, Provo UT 84602, USA) Received 13th November 2007; revised manuscript accepted 26 February 2008 Abstract: The ice core recovered from Prospector Russell Col on Mt Logan (5.4 km a.s.l.), in the Yukon spans over 20000 years. This unique record offers a Pacific view of the stable isotope and chemical record from the Lateglacial to the present. The timescale is based on seasonal counted years, the largest known volcanic acid signatures and the major shift in stable isotopes and chemistry at the end of the Younger Dryas. There are large and sustained changes in the stable isotopic record that are anti-correlated with marine and continental chemistry series. The oxygen-18 in this area is not a proxy for palaeotemperature but rather for source region. The last major isotope shift in AD1840 in δ(18O) and chemistry is compared with the Quinn’s ENSO record. During periods of more frequent La Niña (stronger tropical easterlies) there is more zonal flow of water vapour transport to the Pacific Northwest, δ(18O) values are larger and the deuterium excess dsmaller. These periods coincide with periods of lower accumulation/precipitation in southern Yukon. The Holocene δ(18O) record indicates many large shifts between the meridional (strong El Niño) and zonal (La Niña). Comparison of the Logan isotopic record and the moisture/temperature-sensitive time series of peat bog inception dates for the Northwest shows a strong correlation (0.36) that points to high accumulation rates coincident with low δ(18O) and enhanced meridional flow. Major changes in the core at 4200 BP and 7000–8000 BP point to enhanced meridional flow, which coincide with big changes in the Pacific palaeorecords of the balance between El Niño and La Niña. 4200 BP seems to have inaugurated the ‘modern’ ENSO world. Key words: Mt Logan, stable isotopes, Holocene, ENSO, peat, N Pacific, sudden change. *Author for correspondence (e-mail: [email protected]) © 2008 SAGE Publications 10.1177/0959683608092236 668 The Holocene18,5 (2008) Introduction (Clague et al., 1995; Clausen et al., 1995; Zheng et al., 1998; Yalcin et al., 2007). The unidentified 1516-event must be ‘local’, In Holocene times outside the tropics, stable isotopes of water because, although it is the largest sulfate concentration peak in the given in terms of oxygen (δ(18O)) and deuterium (δ(D)) are usu- last 550 years in the PRCol and Eclipse cores (Yalcin et al., 2007), ally correlated geographically and temporally with the precipita- it has no prominence in the eastern Arctic ice cores (Clausen et al., tion weighted temperature of the site (Dansgaard et al., 1973; 1995; Zheng et al., 1998). The White River candidate peak is Jouzel et al., 1984). The standard unitless measure of ratio within a metre of where it is expected. Thus, it is assigned the age [D]/[H], denoted δ, is defined: of the most recent large eruption (AD803) that deposited the white ash layer over most of southern Yukon (Clague et al., 1995). The δ(D) =1000* (([D]/[H]) −([D]/[H]) )/([D]/[H]) (1) timescale presented here is very close to the tentative one used in sample SMOW SMOW Fisher et al. (2004), with differences smaller than 5%. where SMOW denotes ‘standard mean ocean water’. Stable iso- Using an approximate theoretical timescale the annual layer topic ratios are also affected by the water cycle history between thicknesses were estimated and the cores cut into at least six sam- sources and the site and a generous amount of stratigraphic noise ples per year (Figure 2). The samples were measured for a host of (Fisher et al., 1985, 1996). The deuterium excess, d, is a derived chemical species that vary seasonally. Over 8000 samples were variable (d = δ (D)−8* δ(18O)) that is a conserved quantity for obtained and measured. The high resolution depth series were then each source region (Merlivat and Jouzel, 1979; Johnsen et al., adjusted for densification and the years counted. Layer counting 1989). Although it is also sensitive to water cycle history that was accomplished independently by two groups in an objective bears the moisture to the sites (Fisher, 1992), dis an indicator of way. The data series were put on an ice-equivalent depth scale and the source ocean temperature. The geographic distribution of run through a band pass filter that only passed periodicities that δ(18O) and δ(D) has been modelled with the help of GCMs could be annual in nature. For example, in the upper 40 m (ice (Jouzel et al., 1997; Sturm et al., 2005) and intermediate com- equivalent) only power in the period range 30 cm to 90 cm was plexity models (Fisher, 1990, 1992; Kavanaugh and Cuffey, passed. The near-end points were filtered with the help of adjacent 2003). The storm by storm models of Holdsworth (2008) hold the mirror reflections. The resulting smooth wave forms were easily promise of eventually unravelling the synoptic scale processes at and unambiguously countable by both groups and the resulting work in the Logan region but, for now, we will use the well-tested ages were identical and consistent with the largest volcanic hori- intermediate complexity model as our interpretive tool. zons such as Katmai (AD 1912). As a check, the earlier NWCol Previous work on two Mt Logan δ(18O) series (Holdsworth data sets of Holdsworth were passed through the same procedure etal., 1992; Fisher et al., 2004) demonstrated that it is out of phase and the resulting timescale for the NWCol series was always with other proxy temperature series, showing a ‘Little Ice Age’ within a year or two of that reported by Holdsworth et al. (1992). with ‘warmer’, more positive, δvalues. At 5400 m a.s.l., δ(18O) is Figure 2 shows the measured layer thicknesses (averaged) and a not a palaeotemperature indicator, but rather one of palaeosource theoretical layer thickness for a model that correctly places the (Fisher et al., 2004). Katamai eruption (AD1912) and the transition at 11703 BP with A number of temporal records are available from this region respect to (wrt) AD2000 (Vinther et al., 2006). Down to about 122 and their locations are given in Table 1. m depth (110 m ice equivalent) annual layers are countable. This The top of the Mt Logan massif is a 5400 m a.s.l. plateau about goes back about 300 years. Deeper than that volcanic horizons and 10 km × 25 km dotted with several ~700 m high cones, the tallest the transition are used. The insert in the upper right of Figure 2 of which is Mt Logan (Figure 1b). Between the cones are rela- shows the bottom detail and the unfortunate loss of core quality tively flat areas with low ice velocities (Fisher et al., 2004). Given marked by the grey bar near 170 m depth. The poor core quality the low annual temperature and lack of uncovered area for ice to resulted in coarser temporal resolution. The grey bar in Figure 3 expand into, this situation has probably been stable over the shows where this is manifested, ie, from 7500 to 9500 BP. Holocene. Older ice is within the bottom-most 5% of the thickness Ages beyond AD803 were initially pinned by a very clear tran- and possibly affected by flow discontinuities. sition into ice-age ice with a sudden reduction in electrical con- ductivity (ECM) coinciding with a large drop in δ(18O) (Wolff et al., 1997) and increase in various chemical species (eg, sulfate) Timescale for PRCol Logan Holocene (Figure 3e and f). This transition is assigned the Greenland GICC05-date of 11703 cal. BP (Johnsen et al., 1997; Vinther The PRCol record is placed on a timescale using layer counting, a et al., 2006 ). In between AD803 and the transition the SO−time 4 model, recent accumulation rates and the major acid peaks; eg, series was compared to the Smithsonian ‘Large Holocene Eruption’ Katmai, Laki, a large unknown peak at AD1516 and White River data base, parsed for only the eruptions in Alaska, Aleutian Islands Table 1 Specifics of the core sites Name Elevation Lat. Long. Mean temp. (°C) Main Max. age accumulation Depth reached (m) (m a.s.l.) institutionsa (yr BP) (m ice/yr) PRCol 5340 60.595 140.50 −29 GSC UMaine 30000 ~0.65 188 (bed) NWCol 5340 60.617 140.50 −29 EC 350 ~0.6 103 Eclipse 3017 60.51 139.47 ~ −5 UNH, UMaine ~1000 1.38 345 King Col 4135 60.58 140.60 −17 NIPR ~300 ~1.0 220.5 Jelly Bean Lake 800(1650) 60.351 134.805 −1 Umass, U Pitt, ~7500 n/a n/a UAF aGSC, Geological Survey of Canada; EC, Environment Canada; NIPR, National Institute of Polar Research (Japan); U Pitt, University of Pittsburg; Umaine, University of Maine; UAF, University of Alaska Fairbanks. David Fisher et al.: Mt Logan Holocene–late Wisconsinan isotope record 669 Figure 2 (a)Annual Layer thicknesses, measured is black and mod- elled is grey. The sampling interval (shown in black) down to 110 m (ice equivalent) is fine enough relative to the annual layer thickness to allow at least six samples per year and annual layer counting. Annual layer counting for PRCol and NWCol go back about 300 years, after which PRCol relies on the transition signature at 11703 BP and major Figure 1 (a) Location map of the Alaska and the Yukon sites. (b) Logan volcanic events of the N Pacific region, as described in Figure 3. The and S Yukon sites. The Logan massif is indicated by the ellipse and is about insert covers the bottom 25 m. The grey vertical bar shows where the 5400 m a.s.l. The white areas are glaciated, and the dark grey is water sampling interval became large, because of poor core quality. The res- olution in part of the Holocene is poor because of this grey zone. Its effects are seen in the grey bar shown in Figure 3 in the poor resolu- and the Kamchatka Peninsula. We compared the Smithsonian VEI tion in part of the early Holocene. (b) Age ice–depth relationship for (Volcanic Explosivity Index) for North Pacific eruptions as posted the PRCol on the internet (Smithsonian, 2007) to sulfate time series from PRCol then made small (< 10%) smooth shifts in PRCol ages to et al.’s (1997) GISP2 volcanic SO−series appears in Figure 3d. 4 accommodate the Smithsonian series. The Smithsonian time series Similar high levels are found in the tentative early Holocene of the of these large eruptions appears in Figure 3c. The ice core sources PRCol core, Figure 3b. The typical Greenland high SO−Younger 4 of SO−are from ocean salts, continental mineral dust or volca- Dryas signature combined with low δ(18O) (Stuiver et al., 1997) 4 noes. The [Ca] plotted in Figure 3b (darker shade) is taken as a appears in Figure 3e and f and is very close to that for PRCol, pre- measure of the non-volcanic input of sulfate and thus the volcanic sented in Figure 3a and b. When all the above age assignments are sulfate peaks are those for which there are no Ca peaks. These are made for the PRCol the resulting annual layer thickness profile is back high-lighted with vertical bars. The correspondence between that presented in Figure 2. major ‘swarms’ of large eruptions and volcanic SO−is good with Between ~ 4ka BP and the transition, there are few constraints 4 the exception of the time interval 7500–9500 BP (real ages), except the theoretical (smooth and monotonic) timescale that goes which corresponds to the interval of poor core quality. In particu- through the existing tie points. Nevertheless, the large changes in lar the great swarm of North Pacific eruptions from 3500 to 4000 stable isotopes and chemistry throughout the Holocene will be BP matches one of the largest SO−peaks in the PRCol middle present even if the assigned ages have errors much larger than 4 Holocene. In the early Holocene (9500 to 11500 BP) of the 10%. We estimate the errors in the upper 300 years to be ± 2 years, GISP2 Greenland core there are extremely high volcanic SO−lev- in the 300 BP to 600 BP interval about ±5%, from 600 to 4000 BP 4 els attributed by Zielinski et al. (1997) to enhanced volacanism set 15%. The tie point at the transition looks very solid but between off by mantle unloading at the end of the ice age. Zielinski 4000 and ~12000 BP we have no tie points and we assume the 670 The Holocene18,5 (2008) Figure 3 The Holocene of PRCol, Mt Logan using the timescale described in the text. (a) The oxygen 18 profile. (b) The SO−(lighter shading) and 4 Ca++concentrations in ppb (parts per billion by mass). (c) Large volcanic eruptions in Alaska, Aleutian Islands and Kamchatka Peninisula, taken from the Smithsonian internet-posted VEI listing. (d) Shows 25-yr averages of the GISP2 volcanic sulfate record (Zielinski et al., 1997). (e) GISP2 oxygen 18 record (Stuiver et al., 1997) and (f) the GISP2 total sulfate record (Zielinski et al.,1997). The grey bar in (b) indicates where core quality was poor layer thickness always decreases smoothly. The correlations length in Fisher et al. (2004) and is one of many during the between the PRCol Holocene and nearby absolutely dated palaeo- Holocene. Probably all the large Holocene changes share a similar records will be shown to be very significant. This supports the explanation. The recent 300 years of both Logan cores show this PRCol timescale. shift (Holdsworth et al., 1992; Fisher et al., 2004) and since they are both accurately co-registered they are stacked in order to reduce noise (Fisher et al., 1985, 1996). Fisher et al. (2004) noted that high- Interpretation of the PRCol Holocene resolution stable isotope records at three elevations (800 (JBL) and comparison with other records (Anderson et al., 2005), 3000(E) and 5400(PRCol) m a.s.l. (see Table 1)), show different manifestations of this shift. It is largest at The large differences between the Logan δ(18O) series Figure 3a and the top (PRCol), clear at the bottom (JBL), but not present at the that of Greenland as represented by the GISP2 series, Figure 3e, are middle elevation(E). The size of the AD1840 shift at these three sites now discussed. Figure 3a shows at the extreme left the large shift in appears as squares in Figure 4 along with the results of the model AD 1840. This AD 1840 shift in the PRCol δ(18O), is discussed at presented in Fisher et al.(2004) used to explain these differences. David Fisher et al.: Mt Logan Holocene–late Wisconsinan isotope record 671 Figure 4 The size of (a) the δ(18O) and (b) the deuterium excess dshifts in AD1840 at four sites (elevations) in the Southern Yukon. ‘PRCol’, ice core from Logan; E, ice core from Eclipse Dome; and JBL, a lake record from Jelly Bean Lake (Anderson et al., 2005); see Table 1. ML is the dif- ference from Marcella Lake 100 km SE of JBL (Anderson et al., 2007). The line indicates model predictions of the differences when water sources are switched between zonal and meridional, see Fisher et al. (2004). Grey shading is the model error range The model posits that such shifts are driven by alternations between Pacific tropical ENSO system. Zhao and Moore used ice core mainly zonal sources of moisture and meridional sources, and that series from Asia (Thompson et al.,2000) to assert that the Pacific the different elevations ‘see’ different source distributions. The tropical easterlies (La Niña) were strong prior to AD1840 and that model’s shifting between moisture sources can successfully repro- after AD1840 there was a shift to weaker easterlies and an increase duce the 1840 shift at the three elevations, but what is the physical in the ENSO oscillations. They assert that Asian monsoon became change in the Pacific climate to cause the meridional-zonal switch- weaker after 1840 as a result of the diminished easterlies. The ing? The suspected switch controlling which is the source of mois- Quinn and Neal (1992) record of El Niño appears in Figure 5d. ture is the state of the Pacific tropical Easterly winds, ie, the state of Their ENSO record is based on documentary evidence. We have balance between El Niño and La Niña, (Moore et al., 2001; Fisher given their assessment of El Niño strength numerical values (with et al., 2004; Zhao and Moore, 2006). their El Niño’s rated Very Strong = 10, Strong+ = 6, Strong = 4, What else changes in or close to AD 1840 and how does this Moderate+ = 2, Moderate = 1, Moderate − = 0.7). The 50 year relate to the Logan record? The Yukon temperature proxy records averages of the Quinn El Niño strength record is shown in Figure derived from tree rings show a very modest temperature decrease. 5e, with weak El Niño periods shaded and strong ones unshaded. For example, temperature-sensitive Yukon and Gulf of Alaska It is clear that periods of sustained weak El Niño’s (shaded) coin- tree ring chronologies suggest that the temperature prior to AD cide with the more positive δ(18O) on Mt Logan. This is made 1840 was colder (Wiles et al., 2004; Youngblut and Luckman, obvious in Figure 5f where the statistical distribution of annual 2008). What does show a very big shift in the south Yukon (near δ(18O) values is plotted for the whole series (black) and for the Kluane lake) is the precipitation rate proxy from Luckman’s shaded intervals (weak El Niño). The distribution for the whole ‘Landslide’ chronology (Luckman et al.,2002; Van Dorp, 2004) series has a double maximum and the weak El Niño δ(18O) set is (Figure 5a). The annual average Logan stack for δ(18O), Figure 5c, significantly different than the strong El Niño set. shows the shift in AD1840. The Logan δ(18O) annual series corre- The SW Yukon’s precipitation-sensitive tree ring record, Figure lates significantly (−0.21) over 300 years with the Landslide-pre- 5a, from the Kluane region is (anti-) correlated with Logan δ(18O) cipitation series. When it is dry in the south Yukon, Logan δ(18O) and matches low Yukon proxy-precipitation with reduced ENSO is more positive (and dis smaller) and according to the theory the activity as shown in Quinn’s record, Figure 5d. Figure 5b shows moisture flow more zonal for PRCol Logan (Fisher et al.,2004). the Logan sodium concentration [Na] record and demonstrates what What changes occurred in the Pacific in the mid-1800s that is even clearer throughout the Holocene record, that [Na] and δ(18O) could have caused such changes in the Yukon and Alaska? As are in anti-phase with each other (r = −0.41, see Figure 7e and 7f). suggested in Mann et al. (2000), Fisher et al. (2004) and Zhao and In the 105 year calibration interval, [Na] PRCol annual series cor- Moore (2006), there was a significant shift in the nature of the relates with its δ(18O) series (−0.17) and with the NPI (North Pacific 672 The Holocene18,5 (2008) Index measures the strength of the Aleution low) for autumn and winter (−0.38). Unfortunately, the calibration interval does not span the most recent large shift in AD1840. For example, nobody knows if the calibration-based relationships between various mete- orological indices (such as NPI) and ice core variables hold across the 1840 transition. So we have had to rely on the more fragmen- tary palaeorecords to interpret this proto-type shift. From Figure 5, the generalization is that strong La Niña (or strong easterlies) coincide with lower accumulation in the Yukon and more positive δ(18O) Mt Logan. One possible physical driver for shifts such as AD 1840 is a slight change in the ‘deep’ water temperature and/or its depth. By ‘deep’ is meant the water below the warm mixed layer, which is of the order of 100 m thick. Sun’s (2000) ENSO oscillation solution is chaotic, whereby small changes in Tc (the deep water temperature of the tropical Pacific) can push the tropical Pacific from permanent La Niña to the oscil- latory La Niña–El Niño or ENSO state. Alterations in the easter- lies of this sort have very large effects on the transport of water vapour (monsoons) (Zhao and Moore, 2006). We note in passing that the strength of the Kuroshio current, is largely powered by the coupled Pacific tropical easterlies and mid-latitude westerlies and thus sensitive to changes in these easterly winds (Munk, 1950; Stommel, 1958). Holocene comparison of PRCol δδ(18O), d and Na with peat inception rates in the northwest A rich source of information about climate is found in the peat data base assembled by A. Dyke (personal communication, 2007) and reported in (Gorman et al.,2007). These bogs presently exist over most of North America (outside the driest biomes). Peat bogs start or expand when the climate is sufficiently wet and warm. They do not start in dry and/or cold conditions and once the ice departed they started to grow. After the departure of the ice there is a necessary lag, while regional ecosystems are re-established (Gorman et al., 2007). The multithousand-year trends in the in- peat inception date series are partly caused by this lag, combined with longer climate trends (Gorman et al., 2007). There are ~2100 C14 peat inception dates in the data base. Figure 6c shows all the inception dates divided into C14 age bins, subdivided by longitude range. Looking at the whole data set of 2054 inception dates (Figure 7c, top panel), there is a clear 8.2 ka(calibrated) dry/cold event and there is a strong oscillation from very wet to dry after 4.1 ka (calibrated). Comparison of the eastern sites (Figure 6a) to the western (Figure 6b) shows the 8.2 ka event is not outstanding in the west and the 4.1 ka oscillation in the eastern record is replaced in the west by a wet/warm episode near 4.0 ka. Examination of the east-to-west data series shows the effects of the earlier departure of the Wisconsinan ice in the western regions. The maximum in the inception rate is earlier in the west and it becomes more subdued. For the Yukon and Alaska, which were largely ice free, there is very little long-term trend in the inception date series (not shown) and only the residuals are left. The residu- als we take as related to the climate of the time. Figure 6a and b Figure 5 (a) Preciptation-sensitive tree ring chronology for the shows the whole data set divided at longitude 100°W and the dates Kluane Lake region near Mt Logan (Luckman et al.,2002; Van Dorp, converted to real ages using a smooth ninth order polynomial 2004) (elevation ~ 1000 m a.s.l.). (b) Na concentration stack for the through the Stuiver et al.(1998) C14 calibration set. The PRCol PRCol and NWCol Mt Logan ice cores. (c) δ(18O) stack for the PRCol and NWcol Logan series over the common 300 years. (d) Quinn and Neal (1992) series of El Niño strengths based on documentary Figure 5 (continued) evidence and given numerical values by us. (e) 50 yr averages of distributions of the annual δ(18O) of (c) for the whole 300 values the Quinn and Neal series. The unshaded intervals are times of numer- (black) and for just the low El Niño (shaded) intervals. The signifi- ous El Niño’s and the shaded intervals those of scarce El Niño or cance of the difference between the weak El Niño (strong La Niña) more constant and stronger tropical easterlies. (f) Shows the probability δ(18O) population from the overall population is close to 100% David Fisher et al.: Mt Logan Holocene–late Wisconsinan isotope record 673 Figure 6 Peat Inception dates of peat bogs, data from Gorman et al. (2007) (a) Calibrated dates for the eastern half of North America 100> log.>50. (b) For the western half 100<long<170 W. (c) Carbon 14 dates for various longitudinal subsets of the peat inception data set Holocene should intuitively relate to the more westerly sites north this history at ML must be largely in anti-phase with that indicated of 55°N. But, in order to maintain peat sample size (N > 600) sim- by the peat records for most of the Yukon and Alaska. ML is in ilar to the ice core’s resolution, we use the whole western subset the ‘rain shadow’ and could possibly give a local history different but look only at the residuals of the inception date series. The from the peat history. It is also tempting to suggest that, since ML residuals are much the same as for the extreme western subset. receives rain and snow from a very narrow elevation band near The residual series is compared with the Logan isotope and 700 m a.s.l., its δ(18O) history should be in anti-phase to that of sodium series in Figure 7. JBL and PRCol (see Figure 4). The PRCol δ(18O) record follows that of the (meteoric water For the Holocene, Figure 7 shows that 25 year averages of fed) nearby Jellybean Lake (JBL) (Anderson et al., 2005). δ(18O) and Na are anti-correlated (−0.41). For the calibration inter- However, Anderson et al. (2007) report that δ(18O) for Marcella val, we know larger Na values go along with stronger Aleution Lake (‘ML’ at 697 m a.s.l.) 100 km SE of JBL is in anti-phase to lows, which pull in more distant moisture. In the Holocene the that of JBL and PRCol. They have argued that ML δ(18O) records lower δ(18O) goes with the higher Na, higher little-dand high peat the local (precipitation–evaporation) balance. If this is true, then bog creation rates. The model interpretation in Fisher et al.(2004) 674 The Holocene18,5 (2008) Figure 7 (a) and (b) The peat inception ages for the western sites (b, long = >100), and the extreme western sites (a, 170>long>120 W ). A low pass filter (~4500 a) is run through the western data and the residuals ploted (inverted ) in (c). The green residuals indicate wet/warm and the tan the opposite. The series in (d) is the inverted Logan-PRCol little-d series (deuterium excess). Larger d (green) indicates more distant/warmer sources for the water (Dansgaardet al.,1973; Merlivat and Jouzel, 1979; Johnsen et al.,1989; Fisher, 1990). The good coherence between Logan dand peat series (correlation 0.36, N = 400) shows that wetter/warmer conditions are coincident with more distant/warmer sources. (e) and (f) The PRCol δ(18O) and Na (salt) series suggests the high d and low δ(18O) indicates more distant water high precipitation and periods of strong El Niño ( or periods of sources via enhanced meridional water flux. We harken back to weak La Niña and weaker tropical easterlies) go together. It may the AD1840 d, and δ(18O) shifts and the relationship suggested in be the case that many of the shifts in the Holocene have these Figure 5. There we concluded that more negative δ(18O), large d, co-relations. David Fisher et al.: Mt Logan Holocene–late Wisconsinan isotope record 675 Discussion: tropical easterlies, Asian has a Younger Dryas isotopic signature that is in anti-phase with the chemical impurities. Thus, during the late Wisconsin, Logan shares monsoons and the Kuroshio current a typical multivariable signature with eastern Arctic ice cores. In the Holocene, the major isotopic and chemistry variations are in anti- It is thought that a more active period of El Niño inhibits and phase and have a dominant periodicity of ~950 years. Even given weakens the Asian monsoon (Morrill et al., 2003) that provides the possible errors in the timescale, this period must have this order water to India, China and Indo China. Stronger and persistent La of magnitude. The main physical drivers for the Logan Holocene Niñas (tropical easterlies) strengthen this monsoon. The interval record must have a timescale of about 950 years and in the chain of about 4200 BP (3700 to 5000 BP), which is defined in PRCol by physical processes there should be at least one amplifier that makes low δ(18O), high dand high [Na], has shown up in many Pacific the Logan record’s isotope changes much larger than those of and Asian proxy records as one of weak or failed Asian monsoons Greenland and eastern Arctic Canada. and drought conditions in Australia (Morrill at al., 2003, 2006; There are three main categories of suggested drivers for the Shulmeister and Lees, 1995). Chen et al. (1999) have reported millennial-scale variations in the Holocene. They have been summa- from lake studies that 4.2 ka BP, 7.6 ka BP and 11.3 ka BP were rized by Maslin et al.(2003) and are: (1) internal instability of the dates that marked beginnings of extreme desiccation events in the Greenland Ice Sheet or Arctic sea ice; (2) ocean oscillator and the arid part of China. The Dongge Cave record also records the 4.2 bipolar seesaw; (3) North Atlantic changes forced by solar variations. ka and 7 to 7.5 ka BP ‘weak monsoon’ events (as well as many Whatever combination of these physical processes produces the others that align with the Logan isotope and chemistry series) large millennial-scale variations of the PRCol ice core, we strongly (Wang et al., 2005). Wang et al. (2005) noted that a long- suspect the tropical Pacific ocean and its subsurface temperature established agriculture in central China collapsed because of the must be involved through the easterly wind system and the 4.2 ka BP desiccation, and Drysdale et al. (2006) have blamed the Kuroshio current. This will have to be explored in other papers. 4.2 ka BP event for crippling droughts in many sites in Africa and Asia. The demise of the middle-eastern Akkadian empire has also Acknowledgements been linked to a disastrous drought at 4025 BP (cal.) (de Menocal, 2001) so this ~4100 BP climate event was global in extent and probably culturally more significant than the 8200 BP event. The field and laboratory work for these projects have been gen- Booth et al.(2005) reported a North American-wide drought 4 erously supported by International Arctic Research Center ka BP and the peat bog inception study of Gorham et al. (2007) (IARC), the National Institute for Polar Research (Japan), the also captures this event (see Figure 6). A regional analysis of the Geological Survey of Canada, the Niels Bohr Institute of the peat data (Figure 6 ) shows, however, that there was in fact no 4.1 University of Copenhagen and the National Science Foundation ka drought in the extreme west coast, Yukon and Alaska but rather (NSF grants OPP-0136005, OPP-0240878, 0094587). The assis- the opposite. tance of the Kluane Research Station manager Andy Williams Other large Asian monsoon failures have been reported by (Arctic Institute of North America) was invaluable. The active Morrill et al. (2006) in 7000 to 7500 BP, which align with another help and cooperation of Gerrold Holdsworth is also gratefully coincidence of very low δ(18O), high dand high salt in the Logan acknowledged. The paper was improved through discussion with Holocene. Earlier than 7500 the Holocene has been characterized Lloyd Keigwin. as a time of strong monsoons, driven partly by the early Holocene northern insolation maximum enhancement of terrestrial tempera- References tures (Overpeck et al., 1996). The 7500 to 10000 BP period in PRCol is one of high δ(18O), small dand low salt (Figure 7), which Anderson, L., Abbott, M.B., Finney, B.P. and Burns, S.J. 2005: translates to one of zonal flow, implying strong tropical easterlies Regional atmospheric circulation change in the North Pacific during and strong Asian monsoons consistent with the interpretations of the Holocene inferred from Lascustrine carbonate oxygen isotopes, other authors. The PRCol record shows that between 10000 and Yukon Territory, Canada. Quaternary Research64, 21–35. 11300 BP there was low δ(18O), large dand higher salt content, ––— 2007: Late Holocene moisture balance variability in the south- which would lead to an inference of strong meridional flow, west Yukon Territory, Canada. Quaternary Research 26, 130–41, strong ENSO and weak Asian monsoons. Morrill et al. (2003) doi:10.1016/j.quascirev.2006.04.011. highlight a strong monsoon interval near ~11500 BP, which could Booth, R.K., Jackson, S.T., Forman, S.L., Kutzbach, J.E., Bettis, be reflected in the PRCol record. E A., III, Kreig, J.and Wright, D.K.2005: A severe centennial-scale drought in mid-continental North America 4200 years ago and appar- Shulmeister and Lees (1995) suggested that the ~4000 BP event ent global linkages. The Holocene15, 321–28. was essentially a switch-on of the modern ENSO regime and Sun Chen, F., Shi, Q.and Wang, J.1999: Environmental changes docu- (2000) suggests this could only happen when the tropical Pacific mented by sedimentation of Lake Yiema in arid China since the Late ‘deep’ water sources had cooled sufficiently from their early Glaciation. Journal of Paleolimnology22, 159–69. Holocene warmth. Looking at Figure 7d and 7e for PRCol the pre- Clague, J.J., Evans, S.G., Rampton, V.N.and Woodsworth, G.J.1995: dominantly ENSO-regime began in ~4200 BP and that, post-4200, Improved age estimates for the White River and Bridge River tephras, PRCol has had mostly meridional moisture. This beginning of a western Canada. CanadianJournal of Earth Sciences32, 1172–79. ‘modern ENSO’ regime coincides with the ‘Pulleniatina’ collapse Clausen, H.B., Hammer, C.U., Christensen, J., Schøtt-Hvidberg, of the Kuroshio current (Ujiie et al., 2002). Other authors have C.S., Dahl-Jensen, D., Legrand, M. and Steffensen, J.P., 1995: pointed to a connection between the vigour and position of the 1250 years of global volcanism as revealed by central Greenland ice cores. Nato ASI SeriesI, 30. In Delmas, R.J., editor,Ice core studies Kuroshio current and the strength of the tropical easterlies (Jacobs of global biogeochemical cycles. Springer, 175–94. et al.,1994). The basic physics of the current’s generation require Dansgaard, W., Johnsen, S.J., Clausen, H.B.and Gundestrup, N. such a connection (Munk, 1950; Stommel, 1958). 1973: Stable isotope glaciology. Meddeleser om Grønland197, 53 pp. The salient aspect of the PRCol Holocene are the size and per- De Menocal, P.2001: Cultural response to climate change during the sistence of the stable isotope and chemistry features. The Holocene Holocene. Science292, 667–73. isotopic changes are an order of magnitude larger than the ice core Drysdale, R., Zanchetta, G., Hellstrom, J., Maas, R., Fallick, A., records of the Eastern Arctic, eg, see Figure 3. Figure 3, showing Pickett, M., Cartwright, I. and Piccini, L. 2006: Late Holocene the Logan and Summit records, demonstrates that the Logan record drought responsible for the collapse of Old World civilizations is 676 The Holocene18,5 (2008) recorded in an Italian cave flowstone. Geology 34, 101–104, Maslin, M., Pike, J., Stickley, C.and Ettwein, V.2003: Evidence of doi:10.1130/G22103.1 2006024. Holocene climate variability in marine sediments. In Mackay, A., Fisher, D.A.1990: A zonally averaged stable-isotope model coupled Battarbee, R., Birks,J. and Oldfield, F., editors,Global change in the to a regional variable-elevation stable isotope model. Annals of Holocene.Arnold, 185–209. Glaciology14, 65–71. Merlivat, L.and Jouzel, J.1979: Global climatic interpretation of the ––— 1991: Remarks on the deuterium excess in precipitation in cold deuterium–oxygen 18 relationship for precipitation. Journal of regions. Tellus43B, 401–407. Geophysical Research84(C8), 5029–33. ––— 1992: Stable isotope simulations using a regional isotope model Morrill, C., Overpeck, J.T. and Cole, J.E. 2003: A synthesis of coupled to a zonally averaged global model. Cold Regions Science abrupt changes in the Asian summer monsoon since the last deglacia- and Technology21, 61–77. tion. The Holocene13, 465–76. Fisher, D.A., Reeh, N.and Clausen, H.B.1985: Stratigraphic noise Morrill, C., Overpeck, J.T., Cole, J.E., Liu, K., Shen, C.and Tang, in time series derived from ice cores. Annals of Glaciology7, 76–86. L. 2006: Holocene variations in the Asian monsoon inferred from Fisher, D.A., Koerner, R.M., Kuivinen, K., Clausen, H.B., geochemistry of lake sediments in central Tibet. Quaternary Research Johnsen, S.J., Steffensen, J.P., Gundestrup, N.and Hammer, C.U. 65, 232–43. 1996: Intercomparison of ice core δ(18O) and precipitation records Moore, G.W.K., Holdsworth, G.and Alverson, K.2001: Extra-trop- from sites in Canada and Greenland over the last 3500 years and over ical response to ENSO as expressed in an ice core from the Saint Elias the last few centuries in detail using EOF techniques. In Jones, P.D., Mountain range. Geophysical Research Letters28, 3457–60. Bradley, R.S. and Jouzel, J., editors, Climate variations and forcing Munk, W.H.1950: On the wind-driven ocean circulation. Journal of mechanisms of the past 2000 years.Springer, 297–328. Meteorology7, 79–93. Fisher, D.A., Wake, C., Kreutz, K., Yalcin, K., Steig, E., Mayewski, Overpeck, J., Anderson, D., Trumbore, S.and Prell, W.1996: The P., Anderson, L., Zheng, J., Rupper, S., Zdanowicz, C., Demuth, southwest Indian Monsoon over the last 18000 years. Climate M., Waszkiewicz, M., Dahl-Jensen, D., Goto-Azuma, K., Bourgeois, Dynamics12, 213–25. J.B., Koerner, R.M., Sekerka, J., Osterberg, E., Abbott, M.B., Quinn, W.H.and Neal, V.T.1992: The historical record of El Niño Finney, B.P.and Burns, S.J.2004: Stable isotope records from Mount events. In Bradley, R.S. and Jones, P.D., editors, Climate since AD Logan and Eclipse ice cores and nearby Jellybean Lake; water cycle of 1500.Routledge, 623–48. the North Pacific over 2 000 years and over 5 vertical kilometres; sud- Shulmeister, J.and Lees, B.G.1995: Pollen evidence from tropical den shifts and tropical connections. Géographie physique et Australia for the onset of an ENSO-dominated climate at c. 4000 BP. Quaternaire58, 337–52. The Holocene5, 10–18. Gorman, E., Lehman, C., Dyke, A., Janssens, J.and Dyke, L.2007: Smithsonian 2007: Large Holocene eruptions. http://www.volcano. Temporal and spatial aspects of peatland initiation following deglacia- si.edu/world/largeeruptions.cfm tion in North America. Quaternary Science Reviews26, 300–11. Stommel, H.1958: The Gulf Stream. University of California Press Holdsworth, G. 2008: Interpreting HO isotope variations in high and Cambridge University Press, 202 pp. 2 altitude ice cores using a cyclone model. Journal of Geophysical Stuiver, M., Braziunas, T.F., Grootes, P.M. and Zielinski, G.A. Research113, doi:10.1029/2007ID008639,2008. 1997: Is there evidence for solar forcing of climate in the GISP2 oxy- Holdsworth, G.H., Krouse, H.R.and Nosal, M.1992: Ice core climate gen isotope record? Quaternary Research48, 259–66. signals from Mount Logan Yukon AD 1700–1987. In Bradley, R.S. and Stuiver, M.P., Reimer, P.J.and Braziunas, T.F.1998: High-preci- Jones, P.D., editor,Climate since AD 1500. Routledge, 483–504. sion radiocarbon age calibration for terrestrial and marine samples. Jacobs, G.A., Hurlburt, H.E., Kindle, H.C., Metzger, J.L., Radiocarbon40, 1127–51. Mitchell, W.J., Teague, G.A. and Wallcraft, A.J. 1994: Decade- Sturm, K., Hoffman, G., Langmann, B. and Stichler, W. 2005: scale trans-Pacific propagation and warming effects of an El Niño Simulation of δ(18O) in precipitation by the regional circulation model anomaly. Nature370, 360–73. REMOiso. Hydrological Processes19, 3425–44, DOI:10.1002/hyp.5979. Johnsen, S.J., Dansgaard, W.and White, J.W.C.1989: The origin Sun, D. 2000: Global climate change and El Niño: a theoretical of Arctic precipitation under present and glacial conditions. Tellus framework. In Diaz, H.F. and Markgraf, V., editors, El Niño and 41B, 452–68. the Southern Oscillation. Cambridge University Press, 443–63. Johnsen, S.J., Clausen, H.B., Dansgaard, W., Gundestrup, N.S., Thompson, L.G., Yao, T., Mosely-Thompson, E., Davis, M.E., Hammer, C.U.H., Andersen, U., Andersen, K.K., Hvidberg, S., Henderson, K.A. and Lin, P.-N. 2000: A high-resolution millennial Dahl-Jensen, D., Steffensen, J.P., Shoji, H., Sveinbjörnsdóttir, A.E., record of the South Asian monsoon from Himalayan ice cores. Science White, J., Jouzel, J.and Fisher, D.1997: The δ(18O) record along the 289, 1916–19. Greenland Ice Core Project deep ice and the problem of possible Ujiie, Y., Ujiie, H. and Taira, A. 2002: ‘Pulleniatina minimum Eemian climatic instability. In Greenland summit ice cores.Journal of event’; the drastic change of the Kuroshio Current in the NW Pacific Geophysical ResearchSpecial Publication 102(C12), 26397–410. Ocean. Eos Transactions of the AGU Fall Meeting Supplement, 83, Jouzel, J., Merlivat, L. and Lorious, C. 1984: Deuterium and Abstract #PP71A-0381. Oxygen 18 in precipitation: modeling of the isotopic effects during Van Dorp, R. 2004: Dendrochronological studies of lake level snow formation. Journal of Geophysical Research89(D7), 11749–57. changes at Kluane Lake, Yukon Territory. BSc Thesis, University of Jouzel, J., Alley, R.B., Cuffey, K.M., Dansgaard, W., Grootes, P., Western Ontario, Geography Department. Hoffmann, G., Johnsen, S.J., Koster, R.D., Peel, D., Shuman, Vinther, B.M., Clausen, H.B., Johnsen, S.J., Rasmussen, S.O., C.A., Stievenard, M.and White, J.1997: Validity of the temperature Andersen, K.K., Buchardt, S.L., Dahl-Jensen, D., Seirserstad, reconstruction from water isotopes in ice cores. Journal of I.K., Siggaard-Andersen, M-L., Steffensen, J.P., Svensson, A., Geophysical Research 102(C12), 26471–88. Olsen, J.and Heinemeier, J.2006: A synchronized dating of three Kavanaugh, J.L.and Cuffey, K.M.2003: Space and time variations Greenland ice cores throughout the Holocene. Journal of Geophysical of δ18O and δD in Antarctic precipitation revisited. Global Research111, D13102, DOI:10.1029/2005JD006921. Biogeochemical Cycles17, 1017, doi: 10.1019/2002GB001910, 2003. Wang, Y., Cheng, H., Edwards, R.L., He, Y., Kong, X., An, Z., Luckman, B.H., Watson, E. and Youngblut, D.K. 2002: Wu, J., Kelly, M.J., Dykoski, C.A.and Li, X.2005: The Holocene Reconstruction of precipitation and temperature patterns in British Asian monsoon: links to solar changes and North Atlantic climate. Columbia and the Yukon Territory. Final Report to the Meteorological Science308, 854–57. Service of Canada. Collaborative research agreement, 2001–2002. Wiles, G.C., D’Arrigo, R.D., Villalba, R., Calkin, P.E. and April 2002, 146 pp. Barclay, D.J. 2004: Century-scale solar variability and Alaskan Mann, M.E., Bradley, R.S. and Hughes, M.K. 2000: Long-term temperature change over the past millennium. Geophysical variability in El Niño/Southern Oscillation and associated teleconnec- Research Letters31, L15203, DOI 10.1029/2004GL020050. tions. In Diaz, H.F. and Markgraf, V., editors, El Niño and the Wolff, E., Moore, J.C., Clausen, H.B. and Hammer, C.U. 1997: Southern Oscillation.Cambridge University Press, 357–412. Climate implications of background acidity and other chemistry