The Holocene http://hol.sagepub.com/ Extension of the New Zealand kauri (Agathis australis) chronology to 1724 BC Gretel Boswijk, Anthony Fowler, Andrew Lorrey, Jonathan Palmer and John Ogden The Holocene 2006 16: 188 DOI: 10.1191/0959683606hl919rp The online version of this article can be found at: http://hol.sagepub.com/content/16/2/188 Published by: http://www.sagepublications.com Additional services and information for The Holocene can be found at: Email Alerts: http://hol.sagepub.com/cgi/alerts Subscriptions: http://hol.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://hol.sagepub.com/content/16/2/188.refs.html >> Version of Record - Feb 1, 2006 What is This? Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 The Holocene 16,2 (2006) pp. 188- 199 Extension of the New Zealand kaur (Agathis australis) chronology to 1724 BC Boswijk,1* Anthony Jonathan Palmer2 and John Fowler,1 Ogden' ('School of Geography and Environmental Science, The University of Auckland, Private Bag 92019, Auckland, New Zealand; 21 0. Box 64 Tai Tapu, Canterbury, New Zealand) Received 29 March 2005; revised manuscript accepted 4 October 2005 A HOLOCENE RESEARCH PAPER Abstract: Long tree-ring chronologies have been constructed in the Northern Hemisphere for dendroclimatology and palaeoenvironmental studies, radiocarbon calibration and archaeological dating. Numerous tree-ring chronologies have also been built in the Southern Hemisphere, primarily for dendroclimatology, but multimillennial chronologies are rare. Development of long chronologies from the Southern Hemisphere is therefore important to provide a long-term perspective on environmental change at local, regional and global scales. This paper describes the extension of the New Zealand Agathis australis (kauri) chronology from AD 911 to 1724 BC. Subfossil (swamp) kauri was collected from 17 swamp sites in the upper North Island. Kauri timbers were also obtained from an early twentieth century house on the University of Auckland campus. Twelve site chronologies and 11 independent tree-sequences were constructed and crossmatched to produce a 3631-yr record, which was calendar dated to 1724 BC -AD 1907 against the modern kauri master chronology. A new long chronology, AGAUc04a, was built by combining the modern kauri data with house timbers and subfossil kauri. This new chronology spans 1724 BC-AD 1998. It is of similar length to chronologies from Tasmania and South America and is the longest tree-ring chronology yet built in New Zealand. The greatest significance of the long kauri chronology lies in its potential as a high-quality palaeoclimate proxy, especially with regard to investigation of the El Ninio-Southern Oscillation phenomenon. The chronology also has application to investigation of extreme environmental events, dendroecology, archaeology and radiocarbon calibration. Key words: Dendrochronology, tree-ring, long chronology, kauri, Agathis australis, New Zealand, late Holocene. Introduction Long tree-ring chronologies, some spanning almost all of the Holocene, have been constructed in the Northern Hemisphere for dendroclimatology and palaeoenvironmental studies (eg, Briffa and Matthews, 2002), calibration of the radiocarbon curve, and archaeological dating (eg, Baillie, 1995). Numerous tree-ring chronologies have also been developed in the Southern Hemisphere, primarily for dendroclimatology, but multimillennial-length chronologies are rare (Luckman, 1996; Villalba, 2000). Notable records include a 3622-yr Fitzroya cupressoides (Alerce) chronology from northern Patagonia (Lara and Villalba, 1993), and a subalpine Lagarostrobos franklinii (Huon Pine) chronology from Tasmania, which extends back to 2146 BC (Cook et al., 2000). The extension of existing chronologies and/or development of additional long chronologies from the Southern Hemisphere is therefore *Author for correspondence (e-mail g.boswijkgauckland.ac.nz) important to provide a long-term perspective on local- and regional-scale environmental change, as well as contributing to the global network of tree-ring chronologies used for dendroclimatological and environmental research. In New Zealand, two chronologies span over 1000 years. These are a Lagarostrobos colenosii (Silver pine) chronology from the Oroko Swamp, on the West Coast, South Island developed by Cook et al. (2002), which spans AD 816-1998, and a 1088-yr Agathis australis D. Don Lindley (kauri) chronology (AD 911-1998) from the upper North Island (Fowler et al., 2004). Kauri has long been recognized as a species from which a multimillennial chronology could be constructed (eg, Dunwiddie, 1979: 260; Ogden, 1982: 102). The trees are long-lived, reaching ages in excess of 1000 years, and there are numerous swamps in the kauri growth region that contain well-preserved stumps and trunks. In addition, kauri was one of the major timber types used in New Zealand during the late nineteenth and early twentieth centuries for buildings and a wide variety of other structures, leaving logging remnants () 2006 Edward Amold (Publishers) Ltd Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 10.1 191/0959683606hl919rp Gretel Boswijk et al.: Extension of the New Zealand kauri chronology 189 in forests and a potential tree-ring resource in colonial-era (and later) buildings. This paper describes the development of new subfossil (swamp) kauri and house timber chronologies, and the linking of these chronologies to the 1088-yr modem kauri chronology, to create a continuous kauri record extending back to 1724 BC. It assesses the statistical quality of the long chronology as a prelude to palaeoclimate investigations, and considers other palaeoenvironmental and archaeological applications the long kauri chronology may have. Pukekapia (PUKE), were obtained from sites in the Waikato Lowlands (Figure 1) during the early 1980s by Martin Bridge and John Ogden of the University of Auckland. A single sitechronology was developed for PUKE and radiocarbon dated to c. 3500-3000 BP (Bridge and Ogden, 1986) but a site chronology was not developed at that time from the FNSR assemblage. Sampling of subfossil kauri resumed in the late 1990s. Assessment of the temporal spread of Holocene-age radiocarbon dates from swamp kauri listed by Ogden et al. (1992) indicated that whilst dates as old as c. 7400 BP had been obtained, most radiocarbon dates from kauri occurred in the past 4000 years. This information, combined with the radiocarbon date span of Bridge and Ogden's (1986) Pukekapia chronology, suggested that construction of long floating subfossil chronologies was a realistic goal. There was, however, some doubt as to whether linking the modem and subfossil record would be possible. Background Modem kauri The recent history of kauri dendrochronology has been described in detail by Fowler et al. (2004). The first modern kauri site chronology was constructed in the late 1970s (La Marche et al., 1979). By the mid 1980s there was a network of 15 modem site chronologies distributed throughout the natural growth range of kauri and chronological coverage extended back to the AD 1SOOs (Ahmed and Ogden, 1985). In the late 1990s four modem site chronologies were updated and extended so that the kauri record spanned AD 1269-1998. The analysis and crossdating of a single logging relic extended the record to AD 911. Statistical analysis of modern chronologies identified a strong common signal between all sites, justifying averaging data into a single master kauri chronology (AD 911-1998) (Fowler et al., 2004). Sample collection Swamp kauri To date, 133 samples of Holocene-age kauri have been collected from 17 swamp sites (Table 1). The sites are clustered mainly in the lower Waikato Lowlands and in Northland (predominantly in westerly locations near Dargaville) (Figure 1). Four samples have also been collected from former swamps in South Auckland. The collection of subfossil kauri from the swamp sites has been opportunistic as the extraction costs preclude targeted sampling. In recent years swamp kauri has become a valuable commodity used for fumiture and wood tuming. Extraction is often undertaken as a commercial venture by contractors, sawmillers and landowners. Only as a result of their support, has it been possible to obtain significant quantities of swamp kauri for analysis. Subfossil kauri Holocene-age subfossil kauri was first collected for dendrochronology from sites near Hamilton, Waikato (Figure 1), during the 1970s by researchers from the Laboratory of Tree Ring Research, University of Arizona (Dunwiddie, 1979). To our knowledge, no chronologies were derived from these samples. Two wood assemblages, Fumiss Rd (FNSR) and NEW ZEALAND PACIFIC OCEAN Noh Aucklan N I,land ..H .il.. 40'S *W...ellington South Island ARANAq.-.:'PAR0 *s0wna Incgarei .chnstchurch Din 500km 173°E ARMatlak HARD Kauri Musewm TIMI POur liand Tasman Sea a )Auckland WPAP FNSRAI Tauranga 100Okm Figure 1 Location of kauri sites in the upper North Island, New Zealand. Full site names are listed in Table 1. Large solid circles indicate subfossil kauri sites where multiple samples were collected. Small solid circles indicate sites with single samples only. Shaded squares (not labelled) mark the position of modern kauri sites. The house at 26/28 Wynyard Street was in the centre of Auckland City Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 190 The Holocene 16 (2006) S CG S 0 Cd a) 0 ~0 0 ca) r5 cO 14 c. O -o 4- _- -D -e -o 2._-a) a , -e Sa) a) 1-0 0 .0 a0 a C~~~~~~C -,, Pa) ._ u r COS= *0 Cc -o Cf) Cl 0 -~CO0 3" CO= _A 0) c S Cd 0 +-o 0 U u> 0 0 c ucc, s.0 E0C.- c' cS C Ld 0 ,-c a) a ) o Cc ~ ~ UoC'.4 a) a)0 fo r),-, f. ~0 z 00 -o 00 Cu a o~~ o CA V- C,_ . a) oa) -a) 0 C ._ a-r) cOd CCd Q', o >X 3 :~~~~~_;< C S 0 0o~~~,~ 0 zC) 0 a)m 0 cO"' a) 0 C - x-_Oa cn 0 -C 00 CO C- 0 00 CO ,20 C1 U,CO IC Z j2 p.v CO . QC = CO C._ 0 -A z Z cd o 0~O a) CO 4n -4 (: -0% Cf r -" -- '" - £ NT n - otu n" _" Wn "T cn cd COC o ' A -o a. Cl ClC14~~~~~~C Cl C- cil Cl CD Cl Cl CA cq N 0 N 100 00 t- " a. C o7o Cl Cl4 Clo a0 6 t- O CO CO CO C> 4o 00 Co w I tn 00 OC 0 O 0 0 0 _ ZCO 0 '-a) Cl rA tnk 0 CO 0 N N 00 - CO 'CO 0 CD 34 6 0 CCO 'IC Cl N CY C> N 0 COS Cl) en CO) C Cl r1 " Cl 9 0 N IC 'IO Cl4 C - 00 C) CD Eto C5Esi EIr000 00 'CO 00000 0000 o CY 0 IO o0 z-O -0 Cl4 0 Cl 0 Cl 10 0 00 r0 en 0 'C) c OF W w - CO t 00 - "'0 I C) c- 04 _ C) Cl -oC 0 O N a) a) 00 - 0% 00 Ch00 N o IC .% e, 00 0%00 T C - 00 00 - OC Cl -6a) -' NCn: 00 a) Z z 0 C- N4C'-O)Nf N- - tn Z: IC N ci-) r- ~N 00 N N -) 0 r- r- r- N zz a) C- I Cl 0 o -oC cOcO 00 a) CO Oq p 00 00 a) to aI) r- r-) m 0 0 U)U CQ 0o I~ e CO W) en U) En rA V) t - Cl I- 00 t c- 0 o~ ~~ CO CO~em 0: Z: rC'-) C' r- ci) CO ) 0 0 Cl Cf Clf CO4 IC' cn r- c-C- a) Q0 -oC c-- ~o0 0> a) CA ,0 0a) l £ 0 ,'- 00 -o I roq Cl 10 C.- a) 0 .0 tf) r- r C') I" F. -o 4 0 < 44 0 :t._ u <D < 0~~~~~~~~ HCA - to 0 C 00 :3 UCd 7- CO .2 cn es 5 r:0 Cd~~~~~~~~~~~~~~~~~~~~~~~~~~~C toC "'~;.2 Cd Cd ~ 00 Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 ~ ~ Gretel Boswijk et al.: Extension of the New Zealand kauri chronology 191 The kauri trunks had already been removed from peat and, in most cases, were at a timber mill when sampled, so there is no stratigraphic or other environmental information directly associated with the samples. It is known that all the kauri were recovered from drained and cleared swamps in or adjacent to river valleys mostly less than 20 m above sea level. However, it is not clear whether the kauri were growing directly on peat, or on adjacent Pleistocene sand hills. Museum samples Eight longitudinal swamp kauri slabs displayed at The Matakohe Kauri Museum, Northland (Figure 1) were sampled by trimming the top 5 cm from each slab. The slabs were from logs recovered from different sites in Northland. Two slabs were from the same locations as two other swamp kauri samples (POUT and TIKI) collected from sawmillers. Radiocarbon dates for seven slabs, obtained when the trees were extracted, indicated that the timber ranged in age from c. 7600 yr BP to 740 yr BP. The eighth slab was described as 'modem' and was cut from a 'sinker' log. Up until the early twentieth century, waterways were used to transport kauri logs to the timber mills (Reed, 1964). Logs were often lost during floods, and some of these logs sank. These 'sinkers' are occasionally recovered today. House timbers Kauri timbers were collected from an early twentieth century house (28 Wynyard Street; WYND), and a shed on an adjacent property (26 Wynyard Street) on the University of Auckland campus, which were being demolished. All the timber from the house and shed was in salvage piles on the demolition site. Samples were cut using a chainsaw from a range of timber types, including weatherboards, sarking (wall lining) and joists. In total, 93 kauri samples were obtained from the house and shed. Sampling of house timbers was undertaken to determine if such material could be a useful tree-ring resource. Because quite large trees (> 1 m diameter) were being converted into small timbers (eg, weatherboard which measures 150 mm x 22 mm in cross-section), the ring sequences are likely to be short compared with those derived from living tree cores or swamp kauri samples. However, Erne Adams (1986: 10) reports that trees reputed to be up to 4000 years old were being felled. Therefore, although the houses were constructed in the late nineteenth and early twentieth centuries, the age of kauri being logged and milled means that temporally 'old' material could have been used in a structure. Tree-ring analysis Slices (or 'biscuits') were cut from recovered swamp kauri logs (Figure 2) which were reduced to radii. Between one and six radii were cut per sample, depending on the shape of the sample. The subfossil, museum and house timbers samples were trimmed to c. 2-6 cm height to fit on a measuring stage and sanded to a fine finish to reveal the rings clearly. Short series length was not an issue with subfossil kauri, where some series had over 1000 rings, but 20 house-timber samples had less than 50 rings, so were discarded. Ring widths were measured to an accuracy of 0.01 mm using a measuring stage linked to a computer. Same-tree radii from the subfossil samples were crossmatched and then averaged together into a tree-sequence, which was subsequently used for intertree crossmatching and chronology development. Because of the small size of the house timbers, one radius was measured from each sample. Single series were crossmatched to develop a site chronology. Site chronologies were produced by averaging raw ring width measurements from each crossmatched treesequence or radii. These were used for comparison between different sites. The standardization of data after chronology building was complete is described below. The CROS programs (Baillie and Pilcher, 1973; Munro, 1984) included within the Dendro for Windows suite (Tyers, 2004) were used to aid crossmatching. The CROS programs measure the correlation coefficient between samples at every position of overlap. Significant matches are reported as a Students t value, to take account of the length of overlap. Pilcher et al. (1995) report that for Scots Pine (Pinus sylvestris) t values over 4.00 may be considered significant, and t values over 6.00 are likely to be very significant. This criterion was applied to kauri. When searching for overlaps, a minimum acceptable length of 50 years was applied. All statistically suggested matches were checked visually using line plots. It was found that significant t values could be reported, particularly with multicentury-length series, even if there was a ring error towards the end of the series. Kauri can produce false rings, where the annual ring is divided by an apparent boundary, or locally absent rings, where the annual ring is not complete around the entire circumference of the tree. Rigorous crossmatching within and between trees can usually resolve individual problems. However, clusters of locally absent rings occasionally occurred on some of the swamp kauri, affecting the suitability of the sample for crossmatching. In these cases, series were truncated to span only the reliable section(s), or excluded from further analysis. Figure 2 Nelson Parker cutting subfossil kauri samples from Hardings (HARD) at his timber mill. The logs had only recently been extracted and transported to the timber yard. Photo: A. Lorrey, 2002 Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 192 The Holocene 16 (2006) computerized crossmatching techniques. The chronologies Whangape and Furniss] were built from samples collected in the 1980s and late 1990s from the same swamp site. In the course of crossmatching, three almost identical tree-sequences were identified from the FNSR and WHAN assemblages, indicating that the same kauri logs had been sampled twice. It should be noted that when Whangape was built it was known Results Crossmatching between subfossil kauri From 133 swamp kauri samples, 98 were crossmatched against other kauri samples (discussed below). Duplicate samples were identified in the Furniss (FNSR) and Whangape (WHAN) groups, Maitahi (MAIT) and Yakas (YAKK) reducing the total number of crossmatched tree-sequences to 86 (Table 2). Tree-sequences were built from 17 samples, but none of these could be crossmatched to any other kauri samples. In this case, the tree-sequences may have had undetected ring problems preventing crossmatching or were not contemporary with any other kauri. Radii from five samples were measured, but these series had such a disturbed growth pattern that intratree crossmatching was not possible. Eleven samples were unsuitable for analysis as the rings were quite wide and complacent, or had extreme wedging and suppression. to have a weak period of 5 years when one sample had an ambiguous ring, and all other samples had at least one locally absent ring. Every care was taken with crossmatching, but independent replication of the time period was thought necessary to resolve this issue. Seven site chronologies were constructed from assemblages collected near Dargaville (Table 2). The Chitty (CHIT), Harding (HARD) and YAKK assemblages produced two site chronologies each and one chronology was created from the MAIT assemblage. Compared with the Waikato groups, sample depth in each Dargaville site-chronology is low (three to five trees only). Chronology length across all subfossil sites ranges from 582 years (Chitty2) to 1319 years (Chittyl). Six chronologies span more than 1000 years. The development of millennial-length site-chronologies was aided in part by inclusion of long individual series. The average length of crossdated subfossil sequences was 403 years. Seven samples had more than 800 rings and two exceeded 1000 rings. Interestingly, almost all of the long samples were from the Dargaville assemblages. The majority of Waikato ring sequences were between 200 and Subfossil chronologies Between 2001 and 2004, ten subfossil kauri site chronologies constructed (Table 2). Three well-replicated site chronologies were established from the Waikato assemblages. These include a new version of the Pukekapia (PUKE) chronology (Fowler et al., 2001). Compared with Bridge and Ogden's original 1986 chronology, sample depth was increased from five trees (9 radii) to ten trees (41 radii), and the chronology was extended from 491 to 803 years. The improvements in sample depth and length are likely due to advances in were Table 2 Details of the subfossil kauri and house-timber site chronologies and tree-sequences Referencesb Code Chronology/ tree-sequence Length (years) Trees/ radii Average Sensitivitya growth rate (mm) Date span Furniss Rd Pukekapia Rd FNSR PUKE Furnissi 1084C Pukekapia WHAN 10/42 10/41 1/4 27/65 1.18 1.11 0.63 1.17 0.42 0.32 0.43 0.31 315 BC-AD 769 1724 BC-922 BC 39 BC -AD 192 1180 BC-131 BC 1 2 Whangape 803 231 1050 Dargaville Chitty CHIT HARD Hoanga Maitahi HOAN MAIT Pouto Tikinui Yakkas POUT TIKI YAKK YAK008 Yakkas2 5/13 3/7 1/3 5/17 3/13 1/3 1/3 3/16 1/5 1/7 5/13 1/2 4/7 1.18 1.32 1.33 1.40 1.16 1.06 0.90 0.71 1.48 0.88 0.63 1.26 1.61 0.37 0.35 0.43 0.30 0.33 0.38 0.34 0.29 0.52 0.31 0.33 0.35 0.31 477 BC-AD 842 1257 BC -676 BC 1718 BC -1511 BC AD 124-AD 1152 1466 BC -437 BC AD 1093-AD 1660 1362 BC-944 BC 576 BC-AD 427 1315 BC -741 BC AD 67-AD 730 AD 304-AD 1273 128 BC -AD 117 1547 BC -961 BC 4 Harding 1319 582 208 1029 1030 568 419 1003 575 664 970 245 587 MAU401 POU401 TIK401 TOM401 230 97 527 374 1/2 1/2 1/1 1/1 1.80 2.02 1.31 1.32 0.40 0.35 0.28 0.36 AD 888-AD 1117 AD 702-AD 792 38 BC-AD 489 AD 800 -AD 1173 TRL WYND28A WYND28B 444 442 - /13 - /18 0.84 1.00 0.27 0.23 AD 940-AD 1383 1466 -AD 1907 10 10 Site Waikato PUKOO Whangape Chittyl Chitty2 CHI008 Harding] Harding2 HOAOOI MA 1005 MAITAHI Museum Matakohe Museum MAUN POUT TIKI TOMO House 26/28 Wynyard St Subfossil/houses WYND POUOOJ TIKOO Yakkasl 86/298 AD 3 5 6 7 6 8 9 unpublished data 1724 BC-AD 1907 aThe sensitivity statistic is based on that of Fritts (1976) bl, Boswijk et al. (2001); 2, Fowler et al. (2001); 3, Boswijk and Palmer (2003); 4-8, Boswijk (2004a,b), Boswijk (2005a,b,c); 9, Boswijk and Palmer (2004); 10, Lorrey et al. (2004). Furnissi originally spanned 724 years, but was extended to 1084 years by the addition of a new sample C* Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 in 2004. Gretel Boswijk et a!.: Extension of the New Zealand kauri chronology 193 400 years long and only one tree-sequence had more than 800 rings. The majority of Dargaville ring sequences were between 400 and 700 years long. Further comparison of age and size classes may indicate whether this is an artefact of preservation or if there are real differences in age and size between the Waikato and Dargaville trees. On average, the Waikato and Dargaville chronologies have similar growth rates: 1.15 mm/yr for Waikato, 1.16 nmm/yr for Dargaville. The Waikato chronologies have similar mean ring width values (1.11-1.18 mm) whilst the Dargaville site chronologies have a greater range of variability (0.63-1.61 mm), reflecting differences in the age-structure of the chronology groups. Comparison with modem site chronologies suggests that on average the subfossil trees tend to be slower growing than crossdated kauri from the modem sites (1.58 mm). The subfossil kauri also appear to be more sensitive than kauri from modem sites. Although these results are similar to findings by Bridge and Ogden (1986) and Ogden et at. (1993) for Holocene and ancient (c. 40 000 yr BP) kauri, such comparison should be approached with caution as the influence of sampling bias has yet to be determined. all the wood samples were rechecked. This determined that a false ring had been included in the Waikato record. The Whangape series were amended and the site chronology rebuilt. House timber chronologies Two chronologies were built from the Wynyard Street house and shed, which were crossmatched against the modem kauri master (Table 3; Lorrey et al., 2004). WYND28A and WYND28B were calendar dated to AD 940-1383, and AD 1466-1907, respectively (Table 3; Figure 4). Although only 38% of samples collected were dendrochronologically dated, the analysis indicated that house timbers have good potential for chronology building. The length of crossdated series ranged from 53 years to 222 years. Finding that series of a few hundred years could be obtained from house timbers, such as weatherboard, and that long site chronologies could be constructed was a pleasant surprise. Calendar dating the subfossil chronologies As the subfossil data base developed, the approximate temporal position of the chronologies was established by radiocarbon dates (discussed below). It became increasingly clear from these that the Waikato and Dargaville records extended across the first millennium AD and into the second millennium AD. This was confirmed by crossmatching of the subfossil sample HOAOOJ, which was younger than anticipated (AD 1093-1660), to the modem kauri master and WYND28A (Table 3; Figure 4). HOAOOJ and WYND28a also crossmatched to Yakasi. The period of overlap between the subfossil, house and modem chronologies was further replicated by two museum samples, MAU401 and TOM401. The weaker crossmatch between Yakasl, Harding] and the modem master probably relates to low sample depth - between AD 911 and 1269 the modem master is based on one treesequence only - and perhaps comparison of records derived from old trees (> 1000 years in the case of Yakasl) with the inner (young) rings of a single tree. The link between the subfossil and modem records is considered to be robust, permitting calendar dating of the subfossil kauri. The subfossil record spans 3384 years, between 1724 BC and AD 1660. The entire kauri record, including swamp kauri, house timbers and modem sites, now extends from 1724 BC to AD 1998. Intersite crossmatching of subfossil chronologies All the chronologies and unmatched tree-sequences were compared against each other to identify contemporary series. Within the Waikato group, overlaps were identified between Pukekapia and Whangape (258 years) and Whangape and Furnissi (185 years) (Figure 3). The crossmatching was statistically significant (Table 3) and visually acceptable. The Dargaville chronologies cluster in two groups, linked by Maitahi (Figure 3). In addition, six single tree-sequences from the Dargaville assemblages and four museum samples were independently crossmatched against other site chronologies and tree-sequences (Figure 3). Despite low sample depth, the crossmatching statistics are consistently good within each group with very high values obtained between long site chronologies, especially Hardingi and Yakasi (Table 3). The strength of the cross-correlations, and good visual agreement between sites, indicates that there is a strong common signal between the sites. The Dargaville record replicates 2487 years of Waikato, confirming the periods of overlap between the three Waikato chronologies. Significantly, three Dargaville chronologies spanned the weak period where the Whangape chronology had a problematic ring, at which time the chronologies went out of sequence by one year. Crossmatching between all treesequences (and radii) from Waikato and Dargaville crossing the year in question was reviewed within and between sites, and Group 1724 BC Waikato Pukekapia Quality assessment of AGAUcO4a After crossmatching was complete, the raw kauri data were standardized and a long chronology, AGAUcO4a, was con.......................I........I............AD 166 Whangape Dargaville 1:~ IFurniss Yakas2 Harding2 Maitahi Chitty I Hard, I IYakasl1i Dargaville treesequences 0 O POUOOM TIK001 HOA001 Museum TlK401 [30U401 I TOM4A 1l Figure 3 Crossmatched position of the Waikato and Dargaville site chronologies and independent tree sequences. The scale is calendar years BC/AD. The bars represent the entire length of the chronology/tree-sequence. Bars with full names, eg, Whangape, are site chronologies. Bars with short names, eg, CHI008, are independent tree-sequences Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 194 The Holocene 16 (2006) E oo a--6 _ --------- - -- t' 0 6~~~~~~~~~~~~~~~~~~~~~~0 i e0 -: - C) .; - - - - - - Coo exs 0 a _I _ <46 <46 O _- -- --- - oM \ o5o of 00 ~~~ - - cs, n _ - Ko >< C> ._ - -C - - - - - o:° o - n C,4 CZ C> ____ allO> o Ci;oo _ - -- - - - - - N-- ----- (UN - <> ¢d 006 I CDC 4 ooo K md o mr,u u 0iQN*F mbo___ ___ U~ or- ~ OO~a, _' ~ ~~~~~a 'd 00 3 U;=^2UEU m " m m - - - - - 6°F2 Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 Gretel Boswijk et al.: Extension of the New Zealand kauri chronology 195 AD 1 AD 1 000 AD 500 AD 1500 t I W AGAUmO4a 8167 709 4.00 !0t w WND2*i .43 05.7 Other chronologies l 10,68 i4430 WYND28B AOO1* 76 IHardingl AD 2000 7150 23109 1017 E TOM4O1 4 57 1 X3 PJU401 Figure 4 Overlap between the modern kauri master, house timber chronologies, museum tree-sequences and the subfossil chronologies. tvalues are listed for the key links. Scale is calendar years AD structed. Standardization followed the techniques applied to modem sites by Fowler et al. (2004). The program ARSTAN (Holmes et al., 1986) was used to fit a spline with 50% variance cut-off at 20 years (spline20) to all series. The chronology AGAUcO4a (Figure 5) was constructed by averaging standardized tree-sequences from all subfossil sites, the house timbers and the same modern series used by Fowler et al. (2004) to build the modern kauri master. Such 2.0- a) 0 "0 1.5 - 1.0 - 0.5 -2000 -1 900 -1800 -1700 -1600 -1500 -1400 -1300 -1200 -1100 1.0 c a -o 100 - - 10_ _=__'__ 0.8 0.6 0. Eco C) :0.2 ---- . ........................ .............. CD -1900 -2000 2.0 ~0a)cn 1.5 1.0 -1700 -1800 -1500 -1600 -1400 -1300 -1200 -1100 -400 -300 -200 -100 L _u_-SI _ 0.5 0.( -1000 -800 -900 -700 -600 -500 1.0 0. 10) 10 .. .. ........ 10 ........... ..... . , --.:-.... ... -400 -300 .. 0.8 0.6 U0 0. 0.4 WLI 0.2 .. 0.0 -1000 -900 -800 -700 -600 -500 -200 -100 2.1'O l 1. 5 .C 1.1 ,0 _._ 2.! 5 1, 0.0 _ 100 m 100 , I 200 300 400 500 600 700 800 900 _ _- 0.6_ 10 -. 0.4 m E 1000 1.0 0.8 cnsD a. u] -_ 2.0 -T 1.5 -+ :a z 1.0 -II 0.5 0.0 - 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 1.0 a 100 0 _.... - ..........................0.8 0.6 0.4 0.2 _ V 10 10 0) 1100 1200 1300 1400 1500 1600 1700 1800 1900 U) W - 0.0 2000 Calendar years Figure 5 The upper diagram in each pair shows the standardized (Spline 20) chronology, AGAUcO4a. The lower diagram illustrates the running EPS statistic (upper solid line) and sample depth (radii, dotted line; and tree-sequences, dashed line) for the same time period Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 196 The Holocene 16 (2006) chronologies are considered suitable for crossmatching purposes and high-frequency climate applications (Fowler et al., 2004). The statistical quality of AGAUcO4a was assessed using the Expressed Population Signal statistic (Briffa and Jones, 1990). EPS is derived from correlations within and between trees, and sample depth, and has a possible range from 0 to 1. Fowler et al. (2004) suggest that for kauri, approximately ten trees are required to produce a high quality site-chronology, although this varies depending on the standardization applied to the data. Based on analysis of the evolving quality of the modern kauri master, they also suggested that a relatively small number of trees from multiple sites may be more valuable for constructing a master chronology than a few high quality sites that include numerous trees. In this respect, low sample depth in the Dargaville site chronologies (which do not meet the c. ten-tree per site criteria) is compensated for by pooling subfossil data from all sites. This also supports inclusion of independent tree-sequences in a master chronology. Overall, AGAUcO4a is a high quality record with an EPS of 0.991. However, calculation of evolving EPS (following Fowler et al., 2004) indicates that signal quality varies over time (Figure 5). The period of highest and consistently high, statistical quality occurs between AD 1583 and AD 1996 (modem and house timbers), when EPS is over 0.900. Between AD 1511 and 1321 EPS drops below 0.800. At this time the chronology is based on few modem sites and archaeological material. Improvement in EPS prior to AD 1300 coincides with the transition to a record dominated by subfossil chronologies and there are three high quality periods during the subfossil record when EPS attains levels approaching those of the modem trees. Separate calculation of the EPS statistic for the Waikato record and the Dargaville record (not shown) suggests that the subfossil signal is dominated by the Waikato chronologies, as expected from the greater number of samples from the Waikato. Assessment of changing signal strength over time indicates that further work is required to boost sample depth to yield a consistently high quality long chronology. In particular, additional samples are required between AD 1300 and AD 1500, 137 BC and 288 BC and prior to 1341 BC. Radiocarbon dates Radiometric radiocarbon dates were obtained from the Waikato Radiocarbon Dating Laboratory for six crossmatched Waikato kauri and three crossmatched Dargaville kauri, while the subfossil chronologies were being developed (Table 4). The radiocarbon dates provided a guide to the temporal position of the floating subfossil chronologies relative to the modem kauri master, prior to the subfossil kauri being calendar dated. The linking of the modem and subfossil records by dendrochronological dating established precisely the calendar date of each '4C sample from a crossdated kauri. Because the exact length of gap between the nine samples is known, we 'wiggle-matched' the 14C dates to the new Southem Hemisphere calibration curve, SHCaIO4 (McCormac et al., 2004), to (a) refine the calibrated age ranges for the nine dates listed in Table 4 and (b) compare how well the 14C dates agree with the actual calendar date for each sample. SHCalO4 is comprised of calibration data from the Southem Hemisphere to AD 950 (McCormac et al., 2002) and, before AD 950, the internationally agreed calibration curve (IntCalO4) adjusted with a modelled offset, which varies from 55 to 58 years (McCormac et al., 2004). Therefore only one 14C date from a kauri sample falls within the period spanned by calibration data from the Southem Hemisphere. All nine 14C dates have very good fit to the calibration curve. The calibrated date spans for each wiggle-matched 14C date at 95% confidence levels also agree very closely with the actual calendar date of each kauri sample (Table 4). For example, the mid-points for calibrated and actual calendar date ranges for the eight decadal samples are within 1 year of each other. From this, we can infer that very accurate calibrated age ranges can be obtained by wiggle-matching radiocarbon dates (spaced at known intervals) to SHCalO4, over the last c. 3700 years at least. Radiocarbon dates were also obtained for five samples that were not dendrochronologically dated in order to determine if the samples were contemporary with or older than the long chronology (Table 4). All were older than 1724 Bc. The dates ranged between 2470 - 2205 cal. Bc (CHI006; WK 15530; 3933 +39 BP) and 5468 - 5206 cal. Bc (CHI015; WK 15532; 6397 +47 BP). The temporal position of the 14C sample from the latter series, and that from HAROO9, suggests that these Table 4 Radiocarbon dates from subfossil kauri in date order (youngest to oldest) for dendrochronologically dated and floating series Site code Crossdated series HARD TIKI FNSR FNSR WHAN WHAN PUKE CHIT PUKE Sample HAROlO TIKOOI WaikO06 WaikO05 WaikO04 WaikO03 WaikO02 CHI008 WaikOOl 14C date (2o) Year submitted Waikato code (BP) 2004 2003 2003 2003 2003 2003 2003 2004 2003 WK15535 WK13489 WK13901 WK13900 WK13899 WK13898 WK13897 WK15531 WK13896 1034+35 1365+44 1707+38 2073+38 2443 +38 2738 +40 3119+40 3390+38 3441 +41 988 -1042 cal. AD 669-723 cal. AD 337-391 cal. AD 73-20 cal. BC 483-430 cal. BC 893-840 cal. BC 1303-1250 cal. BC 1558-1505 cal. BC 1713-1660 cal. BC 2004 2004 2004 2004 2004 WK15530 WK15533 WK15536 WK15534 WK15532 3933 +39 4582+43 5006+37 6286+50 6369+47 2470-2205 3368-3033 3906-3644 5316-5042 5468-5206 Calibrated date Dendro-calendar date AD 101 -1020 AD 692-721 AD 360-369 51-42 BC 461-452 BC 870-861 BC 1280- 1271 BC 1535-1526 BC 1690-1681 BC Floating series CHIT HARD HARD HARD CHIT CHI006 HAR008 HAROl 1 HAROO9 CHI015 cal. cal. cal. cal. cal. BC BC BC BC BC Ten-year blocks of rings were submitted to the Waikato Radiocarbon Dating Laboratory for all samples except TIKOO 1, which was a 30-yr block. All calibrated date ranges were produced using Southern Hemisphere atmospheric data from McCormac et al. (2004) and OxCal v3. 10 (Bronk Ramsey, 1995, 2001). Radiocarbon dates for crossdated series with known time intervals were wiggle-matched to constrain the age ranges. Standard calibrated date-ranges are presented for the floating series. Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 Gretel Boswijk et al.: Extension of the New Zealand kauri chronology 197 two sequences could overlap. However, HAR009 has a quite disturbed growth pattern and intratree crossmatching was problematic. Therefore no crossmatching has been identified between these two kauri samples. Discussion Different applications of a long kauri chronology have been identified since the early days of kauri tree-ring research in the late 1970s (Dunwiddie, 1979; Ogden, 1982; Bridge and Ogden, 1986). Potential applications include dendroclimatology, dendroecological and environmental studies, archaeology and calibration of the Southern Hemisphere radiocarbon curve. Dendroclimatology The principal reason for development of subfossil kauri chronologies was to produce long, annually resolved records for dendroclimatology Recent dendroclimatological investigation of the modern kauri site-chronologies using response function analysis and frequency analysis has identified a strong climate signal common to all sites (Buckley et al., 2000; Fowler et al., 2004). Analyses by Buckley et at. (2000) established that kauri growth over the last 150 years tends to be enhanced by cool-dry conditions during the growing season, particularly in spring/early summer, and suppressed during warm-wet conditions. Fowler et al. (2000) observed that (based on Mullan, 1995) in northern New Zealand, cool-dry conditions are characteristic of El Nifio, and warm-wet characteristic of La Nifia, and that ENSO teleconnections had been identified as strongest in September-November by Mullan (1995). The identification of a consistent relationship between kauri growth and ENSO led Fowler et al. (2000) to conclude that kauri is a potentially useful proxy for investigation of past ENSO events, particularly in the context of multiproxy reconstructions. Current research is focused on furthering understanding of the connection between kauri growth and ENSO and developing reconstructions. The long kauri chronology has significantly extended temporal coverage by about an order of magnitude, which has implications with respect to investigation of millennial-scale climate change. As described above, AGAUcO4a has high overall statistical quality, but there are periods where sample depth should be increased. The level of replication also varies through time as sites and independent tree-sequences start and end (Figure 5). The long kauri record also changes from being based on widely distributed modern sites, located between 80 m and 480 m above sea level and with a predominantly northerly aspect, to archaeological material of unknown origin, and finally, to swamp kauri where the growth environment is unknown and where the low-altitude sites have a predominantly western distribution (Figure 1). The latter group account for approximately three-quarters of the length of the long chronology. The uniformitarianism issues related to the changing composition of the long chronology are currently being investigated. Dendroecology Bridge and Ogden (1986) considered that subfossil kauri chronologies would be of value for palaeoecological studies. Although kauri is widely preserved in (and recovered from) swamps, little is known about the age and population structure of the buried kauri stands (Ogden et al., 1992). The development of site chronologies from the Waikato Lowlands and Dargaville region enables investigation of age, germination and mortality trends within, and between, sites, set within a precise chronological framework. This will assist in refining our understanding of the late Holocene history and ecology of kauri. The information can also be applied in conjunction with other palaeoenvironmental evidence, such as palynology, to further understanding of environmental change in the late Holocene, especially development of the swamps. Initial observations are that at least two generations of kauri have been preserved at most sites (Figure 3). At CHIT and HARD, kauri has been present since 5000-6000 BC and continued to be preserved in the swamps until as recently as the twelfth century AD. Kauri was also being preserved at the Waikato site, WHAN/FNSR, until at least late in the first millennium AD. There, the kauri are contemporary with a rise in kauri pollen observed at other sites in the wider Waikato region, which has been suggested as indicating expansion of kauri on to marginal ground near lakes, and on, or adjacent to, swamps and oliogotrophic raised bogs (McGlone et al., 1984; Newnham et al., 1989). Extreme events Palaeoenvironmental reconstruction also extends into investigation of regional- and global-scale extreme environmental events that have left a signal in the tree-ring record. For example, contemporaneous periods of suppressed growth and narrowest ring events, anomalous rings and frost rings, have been identified in long tree-ring chronologies predominantly from the Northern Hemisphere (eg, La Marche and Hirschboeck, 1984; Baillie, 1994; D'Arrigo et al., 2001). These events have been associated with climatic cooling after major volcanic eruptions (Stothers, 1999). An alternative hypothesis is that the trees contain a signal of environmental impacts associated with close encounters with comets (Baillie, 1999). Construction of the long kauri chronology provides an opportunity to investigate whether trees in the southwestern Pacific also contain significant signals at, or around the same time, which would lend weight to the contention that some of these events had a global impact. Archaeology Since the early days of dendrochronology, tree-ring analysis has been a valuable archaeological dating technique. In New Zealand, archaeologists also became interested in the potential of tree-rings for dating archaeological sites (Lockerbie, 1950; Golson, 1955). Early efforts applied dendrochronological techniques in an attempt to crossmatch wood from Maori sites, such as palisade posts (Scott, 1964) or used tree-ring counts to date the deliberate removal of bark from totara (Podocarpus totara) trees, used for containers (Batley, 1956). However, these and other attempts at crossmatching were unsuccessful (eg, Cameron, 1960) and since then tree-ring research in New Zealand has been focused on dendroclimatology and ecology. The new subfossil kauri chronologies have increased the number of reference curves available, improving prospects for dating kauri of unknown age, either from natural deposits such as other swamp sites, or archaeological timbers. Late nineteenth- and early twentieth-century buildings (built predominantly of kauri) are unlikely candidates for dendroarchaeology as the buildings may be less than 100 years old and usually have known construction dates. However, as demonstrated by the Wynyard Street assemblage, long site chronologies can be constructed from house timbers. These chronologies have value for improving sample depth of the long chronology, particularly in the early second millennium AD. WYND28b also included samples that retained the final growth ring below bark edge, unexpectedly providing felling dates for the trees (Lorrey et al., 2004). As a consequence of this work, analysis of timbers from a larger set of kauri Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 198 The Holocene 16 (2006) buildings is now being undertaken to establish whether dendrochronology can be applied to aid interpretation of such buildings; to identify what information can be derived from timbers, apart from dates, such as tree age and timber conversion; and to investigate whether provenancing of the timber is possible. Radiocarbon calibration Dunwiddie (1979), Ogden (1982) and Bridge and Ogden (1986) highlighted the potential of a very long kauri chronology in the calibration of a Southern Hemisphere radiocarbon curve. At the time, studies on the temporal variation in 14C using dendrochronologically dated material had not been carried out extensively in the Southern Hemisphere (Dunwiddie, 1979). Because of the longevity of kauri, and extensive swamp kauri deposits, it was seen to have the potential to generate a calibration curve for the Southern Hemisphere covering a time span similar to the Northern Hemisphere (Ogden, 1982). Southern Hemisphere calibration data for the past 1000 years have now been produced using dendrochronologically dated wood from other New Zealand species (Libocedrus bidwilliei (cedar) and Silver pine (Hogg et al., 2002), as well as data from Chile and South Africa (McCormac et al. 2002, 2004). Calibration of radiocarbon dates older than 1000 BP is dependent on' comparison with Northern Hemisphere data, adjusted with an offset (McCormac et al., 2004). Based on the results presented above, accurate calibrated age ranges are produced when wiggle-matching 14C dates to SHCal04. However, McCormac et al. (2004) point out that 'calibration of Southern Hemisphere measurements is best achieved using securely dendrochronologically dated wood'. In this respect, the construction of the long kauri chronology could have application to extension of the Southern Hemisphere calibration back further in time. Conclusion The development of the long kauri chronology exceeded initial expectations in that not only were numerous long subfossil chronologies constructed, but they were also linked to the modern, calendar-dated record. This produced a continuous kauri chronology (AGAUcO4a) over 3700 years long. Previously, only one kauri radiocarbon date of c. 5790 BP from the Dargaville region had been reported by Ogden et al. (1992). The 14C dates for currently unmatched kauri samples from the Dargaville region suggest that it may be possible to develop new floating subfossil chronologies, or even extend the long chronology, provided more wood of a similar age can be found. The kauri record is the longest tree-ring chronology yet produced in New Zealand. It is similar in length to the long chronologies from Tasmania (Cook et al., 2000) and South America (Lara and Villalba, 1993), and is a new contribution to the global network of chronologies greater than 1500 years long (Luckman, 1996). The greatest significance of the long chronology probably lies in its potential as a high quality palaeoclimate proxy, especially in the context of reconstruction of past ENSO events. However, several other applications are also evident including: (a) extending understanding of kauri ecology; (b) assisting reconstruction of palaeoenvironmental change in New Zealand during the late Holocene; (c) seeking specific short-term environmental events that are recorded predominantly in northern Hemisphere chronologies; (d) archaeology, with particular reference to late nineteenth- and early twentieth century buildings; (e) radiocarbon calibration. It is clear from the above that the development of the long kauri chronology represents an important advance for New Zealand tree-ring research and will be a valuable contribution to the Southern Hemisphere and global tree-ring network. Acknowledgements Financial support for this work was provided by the Royal Society of New Zealand Marsden Fund (grant UOA108) and The Foundation for Research, Science and Technology (grant UOAXO13). We are grateful to G. and C. Chitty, K. Morris, T. Newlove, and in particular, N. Parker for supplying kauri samples. The kauri logs were usually recovered by swamp kauri contractor M. Randell. Some samples were also cut by G. Frost. The following landowners, B. Furniss, J. Hammond, P. Langsford (Waikato Lowlands), R. and D. Harding, N. Hilliam, A. Yakas, and R. and C. Yates (Dargaville), provided access to their properties. P. Crossley was indispensable on fieldtrips and in the workshop where, with 'Sven', he prepared the samples for analysis. He also had the wit to collect the first pieces of house timber, just to have a look. M. Bridge checked the long chronology. We thank two anonymous reviewers for their comments on this paper. References Ahmed, M. and Ogden, J. 1985: Modem New Zealand tree-ring chronologies III: Agathis australis Salisb. - kauri. Tree-ring Bulletin 45, 11-24. Baillie, M.G.L. 1994: Dendrochronology raises questions about the nature of the AD 536 dust-veil event. The Holocene 4, 212-17. 1995: A slice through time: dendrochronology and precision dating Batsford, 176 pp. 1999: Exodus to Arthur: catastrophic encounters with comets. Batsford, 272 pp. Baillie, M.G.L. and Pilcher, J.R. 1973: A simple crossdating program for tree-ring research. Tree Ring Bulletin 33, 7-14. Batley, R.A. 1956: Some practical aspects of dendrochronology in New Zealand. Journal of Polynesian Society 65, 232-44. Boswijk, G. 2004a: Tree-ring analysis of a buried kauri (Agathis australis) tree from Tikinui, north Kaipara Peninsula. School of Geography and Environmental Science Working Paper 21, University of Auckland. 2004b: Tree-ring analysis of buried kauri (Agathis australis) from Hoanga Road, near Dargaville and Pouto, north Kaipara Peninsula. School of Geography and Environmental Science Working Paper 24, University of Auckland. 2005a: Tree-ring analysis of sub-fossil kauri (Agathis australis) from the Harding's Farm, Hilliam Road, near Dargaville, Northland. School of Geography and Environmental Science Working Paper 26, University of Auckland. 2005b: Tree-ring analysis of sub-fossil kauri (Agathis australis) from near Maitahi Road, Dargaville, Northland School of Geography and Environmental Science Working Paper 28, University of Auckland. 2005c: Tree-ring analysis ofsub-fossil kauri (Agathis australis) from the Chitty's Farm, Colville Road, Dargaville, Northland School of Geography and Environmental Science Working Paper 27, University of Auckland. Boswijk, G. and Palmer, J.G. 2003: Tree-ring analysis of sub-fossil kauri (Agathis australis) from near Lake Whangape, Huntly West, Waikato. School of Geography and Environmental Science Working Paper 19, University of Auckland. 2004: Tree-ring analysis of sub-fossil kauri (Agathis australis) from Yakkas Farm, Babylon Coast Road, near Dargaville, Northtand. School of Geography and Environmental Science Working Paper 22, University of Auckland. Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013 Gretel Boswijk et al.: Extension of the New Zealand kauri chronology 199 Boswijk, G., Fowler, A. and Bridge, M. 2001: Tree-ring analysis of kauri (Agathis australis)ftom Furniss Road, Waikato. Department of Geography Working Paper 15, University of Auckland. Bridge, M.C. and Ogden, J. 1986: A sub-fossil kauri (Agathis australis) tree-ring chronology. Journal of the Royal Society of New Zealand 16, 17-23. Briffa, K. and Jones, P. 1990: Basic chronology statistics and assessment. In Cook, E.R. and Kairiukstis, L.A., editors, Methods of dendrochronology: applications in the environmental sciences. Kluwer Academic, 137-52. Briffa, K.R. and Matthews, J.A. 2002: ADVANCE-IOK: a European contribution towards a hemispheric dendroclimatology for the Holocene. The Holocene 12, 639-42. Bronk Ramsey, C. 1995: Radiocarbon calibration and analysis of stratigraphy: the OxCal program Radiocarbon 37, 425-30. 2001: Development of the Radiocarbon Program OxCal. Radiocarbon 43, 355-63. Buckley, B., Ogden, J., Palmer, J., Fowler, A. and Salinger, J. 2000: Dendroclimatic interpretation of tree-rings in Agathis australis (kauri). 1. Climate correlation functions and master chronology. Journal of Royal Society of New Zealand 30, 282-75. Cameron, R.J. 1960: Dendrochronology in New Zealand. Journal of the Polynesian Society 69, 37-38. Cook, E.R., Buckley, B.M., D'Arrigo, R.D. and Peterson, M.J. 2000: Warm-season temperatures since 1600 BC reconstructed from Tasmanian tree rings and their relationship to large-scale sea surface temperature anomalies. Climate Dynamics 16, 79-91. Cook, E.R., Palmer, J.G., Cook, B.I., Hogg, A. and D'Arrigo, R.D. 2002: A multi-millennial palaeoclimatic resource from Lagarostrobos colensoi tree-rings at Oroko Swamp, New Zealand. Global and Planetary Change 33, 209-20. D'Arrigo, R., Frank, D., Jacoby, G. and Pederson, N. 2001: Spatial response to major volcanic events in or about AD 536, 934 and 1258: frost rings and other dendrochronological evidence from Mongolia and Northern Siberia: comment on R.B. Stothers, 'Volcanic dry fogs, climate cooling, and plague pandemics in Europe and the Middle East' (Climatic Change 42, 1999). Climatic Change 49, 239-46. Dunwiddie, P.W. 1979: Dendrochronological studies of indigenous New Zealand trees. New Zealand Journal of Botany 17, 251-66. Erne Adams, J.G. 1986: Kauri: a king among kings. Wilson and Horton Group, 120 pp. Fowler, A., Palmer, J.G., Salinger, J. and Ogden, J. 2000: Dendroclimatic interpretation of tree-rings in Agathis australis (kauri): 2. Evidence of a significant relationship with ENSO. Journal of the Royal Society of New Zealand 30, 277-92. Fowler, A., Boswijk, G. and Bridge, M. 2001: Tree-ring analysis of sub-fossil kauri (Agathis australis) from Pukekapia Road, Waikato. Department of Geography Working Paper 16, University of Auckland. Fowler, A., Boswijk, G. and Ogden, J. 2004: Tree-ring studies on Agathis australis (kauri): a synthesis of development work on late Holocene chronologies. Tree-Ring Research 60, 15-29. Fritts, H.C. 1976: Tree rings and climate. Academic Press. Golson, J. 1955: Dating New Zealand's prehistory. Journal of the Polynesian Society 64, 113 -36. Hogg, A.G., McCormac, F.G., Higham, T.F.G., Reimer, P.J., Baillie, M.G.L. and Palmer, J.G. 2002: High-precision radiocarbon measurements of contemporaneous tree-ring dated wood from the British Isles and New Zealand: AD 1850-950. Radiocarbon 44, 633-40. Holmes, R.L., Adams, R.K. and Fritts, H.C. 1986: Tree-ring chronologies of Western North America: California, eastern Oregon and northern Great Basin. Chronology Series VI. Laboratory of Tree-Ring Research, The University of Arizona, 50-65. La Marche, V.C.J. and Hirschboeck, K.K. 1984: Frost rings in trees as records of major volcanic eruptions. Nature 307, 121-26. La Marche, V.C.J., Homer, R.L., Dunwiddie, P.W. and Drew, L.G. 1979: Tree-ring chronologies of the Southern Hemisphere 3: New Zealand. Laboratory of Tree-Ring Research, University of Arizona, 76 pp. Lara, A. and Villalba, R. 1993: A 3620-year temperature record from Fitzroya cupressoides tree rings in southern South America. Science 260, 1104- 106. Lockerbie, L. 1950: Dating the Moa Hunter. Journal of the Polynesian Society 59, 78-79. Lorrey, A., Lux, J., Boswijk, G. and Crossley, P. 2004: Dendrochronological analysis of salvaged kauri timber from 26 and 28 Wynyard Street, The University of Auckland. School of Geography and Environmental Science Working Paper 20, The University of Auckland. Luckman, B. 1996: Dendrochronology and global change. In Dean, J.S., Meko, D.M. and Swetnam, T., editors, Tree rings, environment and humanity: proceedings of the international conference, Tucson, Arizona, 17-21 May 1994. Radiocarbon, Department of Geosciences, The University of Arizona, 3-24. McCormac, F.G., Reimer, P.J., Hogg, A.G., Higham, T.F.G., Baillie, M.G.L., Palmer, J. and Stuiver, M. 2002. Calibration of the radiocarbon time scale for the Southern Hemisphere: AD 1850 - 950. Radiocarbon 44, 641-51. McCormac, F.G., Hogg, A.G., Blackwell, P.G., Buck, C.E., Higham, T.F.G. and Reimer, P.J. 2004: SHCalO4 Southern Hemisphere calibration, 0- 11.0 cal KYR BP. Radiocarbon 46, 1087-92. McGlone, M.S., Nelson, C.S. and Todd, A.J. 1984: Vegetation history and environmental significance of pre-peat and surfical peat deposits at Ohinewai, Lower Waikato lowland. Journal of the Royal Society of New Zealand 14, 233-44. Mullan, A.B. 1995: On the linearity and stability of Southern Oscillation-climate relationships for New Zealand. International Journal of Climatology 15, 1365-86. Munro, M.A.R. 1984: An improved algorithm for crossdating treering series. Tree-ring Bulletin 44, 17-27. Newnham, R.M., Lowe, D.J. and Green, J.D. 1989: Palynology, vegetation and climate of the Waikato lowlands, North Island, New Zealand, since c. 18,000 years ago. Journal of the Royal Society of New Zealand 19, 127-50. Ogden, J. 1982: Australasia. In Hughes, M.K., Kelly, P.M., Pilcher, J.R. and La Marche, V.C., Jr, editors, Climate from tree rings. Cambridge University Press, 90-103. Ogden, J., Wilson, A., Hendy, C. and Newhnam, R.M. 1992: The late Quaternary history of kauri Agathis australis in New Zealand and its climatic signficance. Journal of Biogeography 19, 611-22. Ogden, J., Newham, R.M., Palmer, J.G., Serra, R.G. and Mitchell, N.D. 1993: Climatic implications of macro- and microfossil assemblages from late Pleistocene deposits in northern New Zealand. Quaternary Research 39, 107-19. Pilcher, J.R., Baillie, M.G.L., Brown, D.M., McCormac, F.G., MacSweeny, P.D. and McLawrence, A.S. 1995: Dendrochronology of subfossil pine in the north of Ireland. Journal of Ecology 83, 665-71. Reed, A.H. 1964: The new story of the Kauri. Reed, 363 pp. Scott, S.D. 1964: Notes on archaeological tree-ring dating in New Zealand. New Zealand Archaeological Association Newsletter 7, 34-35. Stothers, R.B. 1999: Volcanic dry fogs, climate cooling, and plague pandemics in Europe and the Middle East. Climatic Change 42, 713-23. Tyers, I. 2004: Dendro for Windows program guide 3rd edition. ARCUS report 500B. Villalba, R. 2000: Dendroclimatology: a southern hemisphere perspective. In Smolka, P.P and Volkheimer, W, editors, Southern Hemisphere palaeo- and neo-climates: key sites, methods, data and models. Springer-Verlag, 27-57. Downloaded from hol.sagepub.com at The University of Auckland Library on July 22, 2013