6

Examples of long sub-fossil ring-width chronologies presently under construction: these arefrom northern Sweden (Tornetrask) and Russia (Taimyr). The temperature-indicative data areexpressed here as standardised anomalies from the 2000-year mean, smoothed to emphasisedecadal and longer timescale variability.

DENDROCHRONOLOGYAnalysis of Dendrochronological Variability and Associated NaturalClimates in Eurasia – the last 10,000 years (ADVANCE–10K)

Over recent years, with support fromthe European Union’s Climate and En-vironment Programme and the SwissNational Science Foundation, AD-VANCE–10K has brought togethersome 30 collaborators from 17 institu-tions, located in 10 European countries.All of these workers share the commongoals of gathering, screening, and ar-chiving a mass of primarily Eurasiantree-ring measurements. The ultimateaim is to interpret these data in terms ofannually-resolved climate variability;on spatial scales that range from local tonear hemispheric and on temporalscales extending over a few hundred tomany thousands of years.

One major focus of the project is thecontinuing development of an extensivenetwork of ‘modern’ tree-core samples,all taken from relatively cool and moistsites, spread around the northern hemi-sphere, generally at latitudes above60°N, or at higher elevations in moresoutherly areas. The sites are selectedbecause of the likely sensitivity of localtree growth to climate, specifically tem-perature, variability. To date, multipletree-core collections have been madefrom some 400 sites and the samplesprocessed under the guidance of Fritz

Schweingruber at the Swiss Federal In-stitute for Forest, Snow and LandscapeResearch in Birmensdorf, Switzerland,to produce not only ring-width data,but also measurements of intra-annualring-density. These have enabled firmlydated width and density chronologiesto be built with annual resolution foreach location, typically spanning 300–400 years.

These chronologies are capable ofrepresenting detailed patterns of year-by-year temperature anomalies acrosslarge areas of the northern continents.Linear transfer functions that relate pat-terns of recent instrumentally observedsummer temperatures to those recordedin the contemporary tree-ring data,provide quantitative temperature esti-mates, the accuracy of which is demon-strated by comparison with early inde-pendent temperature records. Thetree-ring network is currently beingused to reconstruct a multi-century his-tory of summer temperature changesacross northern North America, Europeand, most recently, northern Russia.Though the present sample network ex-tends over the whole of northern Sibe-ria, the greatest concentration of sitesand the best temperature reconstruc-

tions come from the region between theFinnish border and the Ural Mountains.There presently remains vast scope forexpanding and improving the Russiandensitometric database and the pros-pects for a rapid acceleration of thiswork have been boosted recently by theestablishment of a new tree-ring densi-tometry facility, the first in Russia, at theInstitute of Forest, Krasnoyarsk, madepossible through the support of the Eu-ropean INTAS scheme.

Besides reconstructing maps of tem-perature variability, it has also been pos-sible to provide a useful indication ofvery large scale mean temperaturechanges over the last 600 years by aver-aging the tree-ring density chronologiesacross the entire network. The resultingseries provides strong evidence of theeffect of different known volcanic erup-tions, such as in 1453 (Kuwae, 1452);1601 (Huaynaputina, 1600); 1642 (Mt.Parker, 1641) and 1816 and 1817(Tambora, 1815) on Northern Hemi-sphere mean temperature. The data alsosuggest dates for probable, but as yetunattributed, large eruptions, such asthat in 1695.

At a handful of high-latitude loca-tions in northwest Sweden, northernFinland and on the Yamal and Taimyrpeninsulas, Russia, ADVANCE–10K isworking toward the construction ofcontinuous ring-width chronologiesthat will provide unique regional evi-dence of annual, decadal and centurytimescale temperature changes, all

Location of tree-ring

density chronologies

Tornetrask

0 500 1000 1500Year (A.D.)

-2

-1

0

1

2

Norm

alis

ed a

nom

aly

Taimyr

0 500 1000 1500Year (A.D.)

-2

-1

0

1

2

Norm

alis

ed a

nom

aly

The present ADVANCE-10K tree-ring densitynetwork. The red circles show the locations ofthe long sub-fossil chronologies.

7

DENDROCHRONOLOGY

1810

1813

1816

1809

1812

1815

1818

1819

1811

1814

1817

1820

Normalised density anomaly

-6 2-3 10.5-0.5-1-2-4 0

Tree-ring density anomaly maps indicating areas of relatively warm and cool summers during the early 19th century. Over the hemisphere as awhole, 1810–1820 was one of the coolest decades of the millennium, in part due to the effects of large explosive volcanic eruptions in 1809 (knownthrough ice acidity) and 1815.

8

DENDROCHRONOLOGYfirmly dated to the exact year overmany thousands of years. These arelong-term projects, which ultimatelydepend on locating sufficient suitablesub-fossil wood samples whose agesspan all periods of the full record. Con-tinuous chronologies have been con-structed for the last 2000 years at all lo-cations. In Sweden and Finland, almostall of the preceding 5000 years is alsocomplete, although a firm chronologicallink across the 3rd and 4th centuries B.C.is still to be established. In Yamal, the‘modern’ continuous chronology ex-tends over 4 millennia and much earliermaterial, as in Taimyr, is either alreadymeasured or awaits processing. Thepotential to produce 7000 to 8000-yearlong series at all of these locations willbe realized in the near future.

Further south, in central Europe, theproject has brought together manythousands of oak ring-width series fromhistorical, archaeological or sub-fossilorigins that were previously dispersedin different dendrochronology laborato-ries in Sweden, Holland, Denmark, theUnited Kingdom, Germany and Poland.Several new regional 8000-year longchronologies have been constructedusing objective clustering techniques.Temporal variability in the strength ofinterregional associations shown inthese series, as well as detailed analysisof the extent of climate control of thechronology variability are underway.

Important work comparing tree-ringdata collections located in Hohenheimand Göttingen has recently produced acorrected and extended absolutely-dated oak series reaching back morethan 10,000 years – the longest in theworld – allowing an extension and revi-sion of previous radiocarbon calibrationcurves.

Different phases of active oak germi-nation and deposition throughout theHolocene have been identified in land-grown and bog contexts. There is alsoother evidence for the existence of no-table environmental events such as apronounced decrease in oak growth inthe middle 6th century A.D. that may belinked to very cool conditions seen in thehigh-northern trees in A.D. 536. Anotherlow oak growth period lasted for about100 years between 6270 and 6040 B.C.,corresponding with the oxygen isotopeevidence for contemporaneous cold inGreenland and Antarctic ice cores.

In the immediate future, work willcontinue on important statistical issuesrelated to the processing and interpreta-tion of all of the various tree-ring collec-tions. Potential anthropogenic influ-ences on recent tree growth will becomean increasingly important focus of thework. Increased tree productivity dur-ing the 19th and early 20th centuries andpost-1950 declines in tree density trendshave recently been identified in ourdata. The extent, detail and implicationsof these phenomena have yet to be fur-ther explored. Chronology confidenceand the expression of climate forcing aremost strongly expressed on short (an-nual to century) timescales. New dataprocessing techniques are exploring thereconstruction of longer period variabil-ity. This will require the extraction ofcommon long-term growth trends frommuch larger regional data collectionsthan are currently available. The precisereconstruction of tree line movementswill become increasingly feasible andthe potential for comparison and recon-ciliation with other long-term, thoughless well resolved, climate proxies ex-tracted from ice, lake sediments,speleothems, or peat will become in-creasingly practical and important. Thismultiproxy work should contribute use-fully to the aims of PAGES PEP III re-search as well as CAPE and PEP II.

Increased collaboration with, andsupport for, our Russian colleagueswould enable a vital extension of thework in central and eastern Siberia. Fur-thermore, the development of North/South transects of tree-ring data in cen-tral Siberia and Scandinavia would en-hance the work of the IGBP terrestrial

1400 1500 1600 1700 1800 1900Year

-6

-4

-2

0

2

Nor

mal

ised

mea

n de

nsity

ano

mal

y (s

tand

ard

devi

atio

n un

its)

1400 1500 1600 1700 1800 1900

Major volcanoes:

Tree-ring density / Northern Boreal Forest

1448

1453

1495 1544 1587

1601

1641/2/3

16661675

16951698/9 1740

1783

1816/7/8

1836/7 18841912

? ? ?VEI

7

6

5

-0.8

-0.6

-0.4

-0.2

0.0

0.2

transects. Future work could also use-fully incorporate active participation ofworkers engaged in ongoing work inthe Himalayas as well as establish a firmtree-ring chronology in central Europeextending back into the Younger Dryas,some 12,000 years ago. Finally, bettercollaboration with other, including non-European, dendroclimatologists andthose synthesizing paleodata across thewhole of the high-latitude and high-elevation regions will continue to besought.

A special issue of The Holocene, de-tailing the results of ADVANCE–10K, isscheduled for publication in late 1999.All of the primary data collected underthis project, along with the regionalchronologies and associated derivedreconstructions will be made availablethrough the World Data Center-A forPaleoclimatology.

For details of the aims, participants,publications or other information relat-ing to ADVANCE–10K contact the au-thor or visit our web site. ADVANCE–10K is funded by the European Unionunder ENV4-CT95-0127: Scientific Of-ficer Ib Troen.

KEITH BRIFFASchool of Environmental Sciences, University of EastAnglia, Norwich, UKk.briffa@uea.ac.ukhttp://www.cru.ac.uk/cru/research/

Averaging temperature-sensitive chronologies from around the hemisphere provides a preciserecord of interannual temperature variability demonstrating a clear link between cool summersand large explosive volcanic eruptions.