Low Resolution

PAGES International Project Offi ce Sulgeneckstrasse 38 3007 Bern Switzerland Tel: +41 31 312 31 33 Fax: +41 31 312 31 68 [email protected] Text Editing: Leah Christen News Layout: Christoph Kull Antti Ojala, Guest Editor Circulation: 3400 Christoph Kull and Keith Alverson, Editors

VOL.11, N°2&3 – OCT. 2003 HOLIVAR Holocene Research

Annually banded archives: (from left to right) tree rings of Huon pine from Tasmania (Edward Cook); varves from Lake Holzmaar, Germany (Bernd Zolitschka); speleothems from Rana, Norway (J. Kihle); coral from Papua New Guinea (Sandy Tudhope); ice from GISP2, Greenland (Anthony Gow) www.pages-igbp.org Contents - Climate and Fennoscandian Tree-Rings 2 Announcements - Climate Reconstruction from Peatlands - Editorial: HOLIVAR - Climate Forcing During the Holocene - PAGES Session at the Fall AGU 2003 - Tephra, a Tool for Dating and Correlation - Inside PAGES - Simulating the Last Millenium Climate - Tales From the Field: We Need Your Input - Holocene Climate in the Tien Shan Region - Newsletters: Do You Need a Hardcopy? - Climate Variations in the Gulf of California - New on the Bookshelf - Climate and the Maya - Marie Curie International Fellowships 31 Workshop Reports 4 Program News - 2nd HOLIVAR Workshop, Belgium: - The HOLIVAR Programe Climate Modeling Holocene Climate Variability 6 National Page - Environmental Magnetism, India - Russia 35 Epilogue 7 Science Highlights - Holocene Research: Past Present Future - Developments in Age-Depth Modeling - Instrumental Climate Data 36 Last Page - Varved Lake Sediments in Scandinavia - Calendar - 14C as an Indicator of Solar Variability - A New Paleofi re Database

ISSN 1563–0803 The PAGES International Project Offi ce and its publications are supported by the Swiss and US National Science Foundations and NOAA. 2 Announcements Announcements 3

Editorial HOLIVAR (Holocene Climate Variability) is a new ESF (European Science Foundation) scientific program that seeks to bring together researchers interested in short- and long-term climate variability over the Ho- locene period. HOLIVAR is concerned with scientific issues identified as being of international importance for climate change studies by the PEPIII research community in IGBP-PAGES. It is also a contribution to the PAGES/CLIVAR intersection, which is a shared research agenda between WCRP and IGBP. The central questions addressed by the HOLIVAR program concern how climate has varied naturally on annual to centennial time-scales and determining the cause of Holocene climate variability. The main natural forcing factors under discussion are solar irradiance and volcanic activity, and a key focus of the HOLIVAR program is to determine how the various natural archives (e.g. tree rings, speleothems, lake and marine sediments, mires, and glaciers) can be used together to reconstruct past climate on these time-scales and in relation to these forcings Archives with visible annual resolution are of special importance because they provide a precise chronology and offer the best resolution as climate proxies. The relationship between climate proxies and climate variables—temperature and rainfall—is more often indirect and complex than linear over time. One of the main tasks of HOLIVAR is to combine natural proxies with instrumental records and documentary data in order to understand past natural climate variability more fully and to calibrate and test paleoclimate reconstruction methodologies. Holocene climate modeling is also of particular interest within the HOLIVAR community. Combined paleoclimate data can be used to test and improve the performance of climate models, whereas model experiments can help in the understanding of the causes of past climate variability. The papers presented in the science highlights section of this PAGES Newsletter are based on the extended ab- stracts volumes and the many fruitful discussions that took place during the first and second ESF-HOLIVAR workshops “Combining climate proxies” and “Investigating Holocene climate variability using data-model comparisons” held at Lammi Biological Station, Finland in April 2002 (http://www.gsf.fi/esf_holivar/) and Louvain-la-Neuve, Belgium in June 2002 (http://www.cru.uea.ac.uk/~timo/holivar/), respectively.

ANTTI OJALA, MATTI SAARNISTO Geological Survey of Finland, [email protected], [email protected] RICK BATTARBEE University College London, [email protected]

��

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 2 Announcements Announcements 3

38, 3007 Bern, Switzerland. Phone, 2004 Newsletters fax and email remain the same. The next Newsletter is open to Inside PAGES science highlights from all PAGES- PAGES SSC Meeting 2003 relevant research. Scientific con- The PAGES Science Steering tributions as well as workshop Committee (SSC) recently held its reports and program news are After several months of disruption annual meeting in Banff, Canada. welcome. to office routine due to changes The minutes of the meeting are If you are interested in con- in personnel, the IPO is running available on our website at http: tributing, contact Christoph Kull smoothly again. //www.pages.unibe.ch/ppeople/ ([email protected]). All submis- The Executive Director Keith sscleaders.html. sions should follow the instructions Alverson, is now supported by Julie Brigham-Grette, former for authors on our website at http: four part-time office staff: Office Vice-Chair, has taken over the role //www.pages.unibe.ch/products/ Manager (Selma Ghoneim), Web- of SSC Chair from Vera Markgraf. newsletters.html and be submitted Database Coordinator (Kaspar After serving two terms, Vera by 31 December 2003. Grathwohl), Project Officer (Leah Markgraf, Matti Saarnisto and In the interest of saving paper Christen), and Science Officer José Boninsegna will be leaving and postage, we are asking all (Christoph Kull). More information the SSC. The IPO thanks them for members if they would like to re- about the role of these people in all their hard work. Riuji Tada and ceive ‘virtual’ PAGES Newsletters. the day-to-day running of PAGES, Carole Crumley were nominated to These newsletters would be sent along with their contact details, can serve for another three-year term. by email as pdf file attachments. be found on our website at http: There are currently several com- Please contact Leah Christen (lea //www.pages.unibe.ch/ppeople/ mittee positions vacant. [email protected]) if you staff.html. The IPO would like to encour- would like to have the Newsletter age nominations from developing emailed to you rather than posted. New PAGES Office Location countries and from scientists from Back issues are always available as In September, another major office the southern hemisphere—two pdf files to download from our web- change took place when we moved areas that are currently underrep- site at http://www.pages.unibe.ch/ from our current location in the cen- resented on the SSC. products/newsletters.html. ter of the city to an office near the Please contact Keith Alverson main train station. This move was ([email protected]) for necessary because of modifications further information. to our current office building. The new address is Sulgeneckstrasse Tales from the Field: We need your Input!

Do you have an interesting and humorous story from your paleoenvironmental fieldwork? If you write it down in 500 words or less and send it to us, we will put it in PAGES news!

��

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 4 Announcements - Program News Program News 5

New on the PAGES bookshelf: ��

��

Marie Curie Incoming International Fellowships (IIF)

The European Commission’s Sixth Framework Programme includes the introduction of Marie Curie Incoming International Fellowships. These Fellowships are available to fund top-class researchers from countries outside the 15 European Union Member and Associated States to support research visits of one to two years in a host research organization within one of the EU States. Support for fellows to return to their country of origin may be included for developing countries, emerging and transition economies. EUR11 million has been allocated to a 12 February 2004 deadline.

PAGES would be happy to act as host organization for a researcher interested in applying for a Fellow- ship to conduct research in Switzerland. Such a fellowship would be mutually beneficial and would enable you to take advantage of PAGES’ international connections. We invite you to contact us with a short research proposal by 31 December 2003. Please note: You would need to apply for the Fellowship yourself but we could provide assistance, if necessary.

Further information about the Fellowship is available at: http://fp6.cordis.lu/fp6/call_details.cfm?CALL_ID=30

Please send proposals to: Christoph Kull ([email protected])

HOLIVAR (Holocene Climate Variability)

RICK BATTARBEE Environmental Change Research Centre, University College London, UK; [email protected]

HOLIVAR (http://www.esf.org/ ans and climate modelers, and the • How can an understanding of holivar) is an ESF scientific pro- program aims to stimulate research past variability improve the gram in the life and environmental on a number of key questions: performance of climate mod- sciences. It seeks to bring together els, leading to an improved European scientists interested in • How and why has climate var- prediction of future climate climate variability over the Ho- ied naturally on different time- change? locene period—broadly the last scales (annual, centennial and • How can climate models help 11,500 years of Earth history. Sci- millennial) over the Holocene to explain past climate change? entists involved in the program are period? paleoclimatologists, climate histori-

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 4 Announcements - Program News Program News 5

• How has climate variability and the nature of human society in- �� teracted during the Holocene? �� Although these are questions of �� global relevance, HOLIVAR focuses �� mainly on research in Europe and �� Africa, and is allied closely to the �� aims and objectives of the Past �� Global Changes (PAGES) PEP-III project, which is concerned with �� climate change along a pole-to- �� pole transect through Europe and �� Africa. It is also a contribution to the PAGES/CLIVAR Intersection, �� which is a shared research agenda between WCRP and IGBP. Figure 1: The HOLIVAR concept harmonization of paleoclimate 2004 Research workshop on Scientific Rationale data from the different archives, to “Holocene climate forcing”, The latest research on recent climate encourage the use of multi-proxy Zurich, Switzerland. variability is tending increasingly databases that can be used to com- 2004 Research workshop on to the view that greenhouse-gas bine data of different types at the “Climate-human society inter- forcing is becoming the dominant, regional scale, and to encourage actions during the Holocene”, though not the only, process driv- and facilitate comparisons between location to be decided. ing global warming. Consequently, data and model output (Fig. 1). 2006 Open Science Meeting on future climate change will be the “Holocene climate variability”, result of interactions between the Work Timetable London, UK effects of human-induced changes The program objectives are being and the effects of natural variability. met over a five-year period (2001- Steering Committee and There is therefore an urgent need 2005) by a series of meetings, work- Funding to understand these interactions shops, and training courses, which HOLIVAR is financed on an à la and to document and identify the will culminate in an open science carte basis by ESF member organi- characteristics of natural climate meeting in September 2005. zations in Austria, Belgium, Finland, variability. 2001 International conference France, Germany, the Netherlands, Natural variability, whether as- (co-sponsored by IGBP-PAGES) Norway, Sweden, Switzerland, and sociated with mechanisms external on “Past Climate Variability the United Kingdom. The program or internal to the earth system, is through Europe and Africa”, steering committee meets annually expressed on inter-annual, decadal, Aix-en-Provence, France. and comprises: and century time-scales, and it is 2002 Research workshop on Rick Battarbee, Chair (UK), An- variability on these time-scales ”Combining proxies”, Lammi, dré Berger (BE), Keith Briffa (UK), that climate and earth system mod- Finland. Françoise Gasse (FR), Eystein Jan- elers are seeking to simulate. Data 2002 Research workshop on sen (NO), Ingemar Renberg (SE), for only the most recent part of the “Data-model comparisons”, Matti Saarnisto (FI), Christoph Spötl historical record can be derived Louvain-la-Neuve, Belgium. (AT), Jurg Beer (CH), Bas van Geel from instrumental measurements 2003 Research workshop on (NL), Dirk Verschuren (BE), Bernd and, therefore, there is a need to “Holocene dating, chronolo- Zolitschka (DE), Svenje Mehlert extend the instrumental record gies, and age modelling”, Utre- (ESF) and Joanne Dalton Goetz backwards in time using proxy re- cht, the Netherlands. (ESF). cords of climate change contained 2003 Training course 1 on Further information is avail- in naturally occurring archives, such “Quantitative Holocene climate able on the program web site as lake and marine sediments, peat reconstruction and data-model at (http://www.esf.org/holivar) bogs, tree rings, speleothems, and comparisons”, London, UK. or by contacting Joanne Dalton ice cores. 2004 Research workshop on Goetz, ESF ([email protected]); However, none of these archives “Databases and data analysis”, Heather Binney, ECRC-UCL, UK alone is adequate to capture the Bremerhaven, Germany. ([email protected]); or temporal- and spatial-scale vari- 2004 Training course 2 on Rick Battarbee, ECRC-UCL, UK ability needed for regional com- “Quantitative Holocene climate ([email protected]). parisons with climate model output. reconstruction and data-model Consequently, the prime objectives comparisons”, location to be of HOLIVAR are to encourage the decided.

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 6 National Science Highlights: HOLIVAR 7

� ��

��

�� # �� # #

# �� # �� # # Apatity # �� Tallinn # �� # # Petrozavodsk # St.Petergourg �� # Minsk # �� Kiev # Arkhangelsk # # Petropavlovsk # Vologda �� Moscow Kamchatsky �� Yakutsk # �� Yekaterinburg # �� # �� Tb'ilisi # Tomsk # �� # Novosibirsk �� # Ulan-Ude Barnaul # �� Vladivostok �� Toshkent Bishkek # �� # # �� # ��

��

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 6 National Science Highlights: HOLIVAR 7

Developments in Age-Depth Modelling of Holocene Stratigraphical Sequences

H.J.B. BIRKS AND EINAR HEEGAARD Botanical Institute, University of Bergen, Bergen, Norway; [email protected]; [email protected]

Absolute chronology is the basis for (3) model selection and evaluation. rates of biotic change. Here one all comparisons and correlations Assuming that the 14C dates are re- needs to develop a statistical model of Holocene stratigraphical proxy liable, are not subject to problems for estimating calibrated ages for records. It is essential to develop of hard-water effects and incorpo- each sample based on a geologi- independent absolute chronologies ration of “old” carbon from other cally realistic model of sediment for such stratigraphic sequences sources and, in the case of marine accumulation. Linear interpolation before attempting comparisons, sequences, that the marine reser- between dated horizons is not ap- correlations, and syntheses, so as voir effect is known, the first step is propriate here. Age-depth models to avoid the well-known problems calibration of the radiocarbon dates. are required that (1) are statistically of visual “curve-matching” and as- Different calibration procedures can reliable and robust, (2) provide es- sociated dangers of the “reinforce- give different calibrated ages and timates of the uncertainties of the ment syndrome” and the “suck-in” very different ranges and probabil- calibrated age estimates for each phenomenon (Bennett, 2002a). A ity distributions, especially if Bayes- sample, (3) take account of the un- statistical approach for developing ian approaches are adopted. In the certainties in the datings, radiocar- age-depth models is desirable so Bayesian approach, it is possible bon calibrations, and sampling, (4) that the uncertainties in the radio- to take account of other dates from are geologically realistic, and (5) are metric age determinations can be the same sequence during the cali- easy to implement. Various meth- taken into account and confidence bration of individual radiocarbon ods such as linear interpolation, intervals for the resulting models dates. If it is known that one date cubic splines, and regression mod- estimated and displayed. is younger than another because els (e.g. polynomial least-squares As almost all Holocene strati- of their known stratigraphical re- regression) in the framework of graphic sequences (e.g. peats, lake lationship, it is possible to use this generalized linear models (GLM) sediments) are from non-laminated information to reduce the range of have been used (see Bennett, 1994, sediments, an absolute chronology uncertainties for the calibrated ages 2002b). The approach we have ad- must be based on radiometric dat- of both dates (Bennett, 2002b). The opted (Heegaard and Birks, 2003) is ing (14C, also 210Pb for the very re- irregular probability distribution weighted, non-parametric regres- cent past) and, if available, volcanic functions of calibrated dates can sion in the framework of general- tephras of known age. Radiocarbon be summarized in many ways (e.g. ized additive models (GAM) (Yee years do not equal calendar years. intercept, mode, median, weighted and Mitchell, 1991). In contrast to It is, however, possible to calibrate average) and the resulting summa- GLM, GAM models are driven radiocarbon dates (with their as- ries can have important influences primarily by the data themselves sociated uncertainties) for the on the final age-depth models. An rather than by fitting prespecified Holocene into calibrated ages, additional potentially critical ques- GLM regression models to the data. thanks to the development of the tion is whether the radiocarbon cali- Smoother functions are fitted to the internationally adopted INTCAL98 bration curve should be smoothed data with appropriate prespecified radiocarbon calibration data-set to match the potential resolution of link and error functions. Weight- and the availability of calibration the stratigraphical sequence of in- ings can be introduced for all dated software such as CALIB, OXCAL, terest. The effect of such smoothing samples, with weights inversely and BCAL. Age-depth modelling is usually to alter the uncertainties proportional to the uncertainties using calibrated ages is essential if in the resulting calibrated ages. in the calibrated ages. Well-dated correlations and comparisons are to Given a set of calibrated ages horizons (e.g. tephra layers) or the be made with proxy records based (and associated uncertainties) for a lake-sediment core-top receive high on an absolute chronology, such as stratigraphical sequence, the next weights in the regression, whereas ice-cores or tree-rings. Such model- step is to convert the sample depths calibrated ages within a radiocar- ling using calibrated ages has, how- into estimated calibrated ages prior bon plateau, and hence with a large ever, several statistical problems. to comparison with other proxy re- calibration uncertainty, receive low cords. This can readily be done by weights. Developing a Chronology simple linear interpolation with The stages in developing an ab- confidence intervals estimated for Choosing the Model solute chronology for a Holocene the resulting age estimates (Ben- After fitting a series of GAM models sequence based on a series of 14C nett, 1994, 2002b). However, it is with different error specifications dates are (1) calibration of the 14C often also necessary to estimate and smoothers using different de- dates, (2) statistical age-depth mod- sediment deposition times, fossil grees of freedom to the available elling using calibrated dates, and accumulation rates (“influx”), or calibrated ages and their associ-

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 8 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 9

that other sources of uncertainty (sampling, radiocarbon assay, etc.) remain the same. ��

� � � �

� Conclusion and Outlook

� �� In comparing and interpreting Holo- � �

� cene stratigraphical records, more �

� ��

� attention needs to be paid to the � �

� uncertainties associated with the � �

� age-depth models. The uncertainties

� �� provide a guide for the basis of any correlation (Bennett, 2002a). In the � correlation of an “event” between � �� sequences, the 95% confidence in- �� tervals of the age estimates should Figure 1: The relationship between sediment depth and calibrated years B.P. for Gåvåli lake, consistently overlap but as Ben- Dovre, Norway (data are kindly lent to us by Wenche Eide). The red line is the expected calibrated year for a particular sediment layer. The dotted lines represents the 95% confidential nett (2002a) notes, the correlation interval based on the insecurity of the observed calibrated dates and the regression line in may still be valid even if a 1 in 20 combination. Here we used a constant variance in the regression. confidence interval fails to overlap. ated uncertainties, the next step is the weighting functions used are re- However, when most or all of the to select the simplest parsimonious alistic, (3) sample age is known with confidence intervals do not overlap, model (called the “minimal ad- a larger error than sample depth, the proposed correlation should be equate model” in statistical model so we are justified in regressing rejected. Comparison and correla- building), namely the simplest sta- age on depth rather than depth tion of stratigraphical sequences tistically significant solution that on age, (4) the principle of parsi- are, in reality, hypotheses that uses the fewest terms in the model mony, (5) there are no sedimentary should be tested and falsified using and the fewest numbers of degrees disturbances in the sequence and statistical criteria (Bennett, 2002a). of freedom in the fitted smoother. there are no gaps due to coring ar- Statistically based age-depth mod- Model selection is done by exam- tifacts or errors, and (6) the mean els are valuable because they force ining regression diagnostic plots radiocarbon date is independent of the researcher to be explicit about and PRESS statistics. The point- its variance. The final choice is the the assumptions of the model and wise standard errors for the mean simplest adequate model possible the quality of the data used. How regression response are added to statistically and has the largest de- much confidence do we really have the standard error of the individual grees of freedom. It should there- in the basic radiocarbon age deter- age estimates to calculate 95% fore be relatively robust and have minations and in the subsequent confidence intervals for the entire good predictive power. However, calibrations? Age-depth modelling age-depth model. All these analy- the model is a working hypothesis is perhaps one of the weakest areas ses are implemented as a S-PLUS to be tested by independent crite- in Holocene research at present and macro (Heegaard and Birks, 2003). ria. Being a statistical model, it has it is an area that urgently requires The resulting age-depth model and uncertainties and all Holocene age further attention. Such work is cur- associated age estimates and 95% estimates generally have associat- rently in progress at the Bjerknes confidence intervals should be ed 95% confidence intervals of be- Centre for Climate Research in critically evaluated by comparison tween 50-200 calibrated years. Such Bergen. with independently derived age es- intervals can easily encompass timates from the same geographic rapid events of short duration (Ben- REFERENCES region. This GAM approach allows nett, 2002a). If age-depth modelling Bennett, K.D.,1994: Confidence intervals for age the use of calibrated 14C dates, 210Pb is applied consistently to different estimates and deposition times in late-Quaternary sediment sequences. The Holocene 4: 337-348. dates, tephra, and other indepen- proxy records at different sites, this Bennett, K.D., 2002a: Comment: the Greenland 8200 dently dated horizons in one age- should provide a robust basis for cal. yr BP event detected in loss-on-ignition pro- depth model and has been found comparing and correlating many files in Norwegian lacustrine sediment sequences. Journal of Quaternary Science 17: 97-99. to produce realistic age-depth Holocene proxy records. Future Bennett, K.D., 2002b: Documentation for psimpoll relationships for a range of lake- developments should be directed 4.10 and pscomb 1.03. (http://www.kv.geo.uu.se sediment sequences in Norway, at trying to reduce the uncertain- under link to “Software”). Uppsala University. Sweden, Finland, Austria, Switzer- ties of the resulting age estimates. Heegaard, E. and Birks, H.J.B., 2003: Age-depth modelling of late-Quaternary sediment sequences land, Greenland and the UK. Decomposition of the total model (in preparation). An age-depth model is statisti- variance suggests that to reduce Yee, T.W. and Mitchell, N.D.,1991: Generalized addi- cally based and as such includes the model uncertainties from 100- tive models in plant ecology. Journal of Vegetation Science 2: 587-602. various assumptions: (1) the dat- 200 to 25-50 years, uncertainties in ings used are reliable, (2) the radio- the radiocarbon calibration need carbon calibrations are reliable and to be reduced by 75%, assuming

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 8 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 9

Instrumental Climate Data

REINHARD BÖHM Central Institute for Meteorology and Geodynamics, Vienna, Austria; [email protected] Instrumental data form the nec- a) b)

�� essary basis for climate recon- � �

��

structions. The data are used for � �

� ��

calibration in proxy reconstruc- � �

��

� �� tions and are usually considered � �

�� to be “true” climate information by � �� proxy specialists. However, before � � � �� applying instrumental data to proxy �� climate reconstructions, a number � �

��

of questions should be addressed: � ��

�� To what extent do instrumental � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �

data reflect real climate? For which � � � � � � � � � � � regions of Europe does well-tested and homogenized instrumental Figure 2: Two examples of the spatial variability of long-term climate trends. (a) Long-term annual precipitation totals (single years and 30 years smoothed) for a grid point climate information exist? How in SE-France (4E-46N, blue) and one in N-Italy (10E-46N), both relative to the 20th century long are the existing instrumental mean. (b) Long-term annual sunshine duration in the Eastern Alps (30 years smoothed) for series and what is their spatial den- high (>2000m, blue) and low (<600m, yellow) elevation sites. sity? Last but not least, coverage by is not made random simply by Two European sub-regions different climate parameters should combining a large number of se- serve to illustrate “close to ideal” also be taken into account. That is, ries together, since in many cases instrumental climate variability “climate” reconstructions should there are systematic biases in the datasets. They show the number of be more than single-variable original series. Figure 1 shows two inhomogeneities that are detectable analyses. In most cases, climate is examples of such systematic offsets and removable, the average length represented by temperature alone. in climatic time series—one typical of homogeneous sub-intervals for Precipitation is sometimes also in- of recent decades, one of the 19th the different climate parameters, cluded but very rarely is the whole century. While homogenization is and to what extent such high- climate system analyzed. valuable, poor quality homogeniza- density regional datasets differ Experience with instrumental tion can erroneously import climate from existing large-scale datasets. climate reconstruction in Alpine information from remote regions. These sub-regions are the “Larger areas shows that homogenization It should therefore be undertaken Fennoscandian Region” (NACD is indispensable for extracting with care, using mathematical tests dataset, e.g. Moberg and Alex- the real climate signal from non- together with information about andersson, 1997; Hanssen-Bauer climatic noise. Non-climatic noise station history (meta-data) and be and Nordli, 1998) and the “Greater is caused by a number of factors, carried out in smaller sub-regions Alpine Region” (ALOCLIM and such as relocation, and changes in with a dense network of highly cor- ALPCLIM datasets, e.g. Auer et al., measuring technology, instrument related series. For an overview on 2001; Böhm et al., 2001). Both are installation and observation time, homogenization see Peterson et al. “multiple”, “long-term” and “dense”, among others. The uncertainty (1998). much time has been invested in homogenization of both instrumental a) b) �� series, and both regions are rich �� in proxy data sources for paleoclimate reconstructions. Studies �� of these regional high-density da- �� tasets (typical mean site distances � � �� of 30 to 100 km) show that existing � global-scale datasets (e.g. Jones, �� 1994; Vose et al., 1993) are deficient �� when resolving meso-scale variabil- ��

�� ity patterns. Paleo-climatologists � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �

� � � � � � � � � � � � should pay attention to this fact since most of the proxy sources Figure1: Two examples of systematic offsets in instrumental climatic time series in Central react to local to meso-scale climate Europe. (a) Sunshine duration offset (new minus old in percent of monthly totals) in the 1980s due to automation in the Austrian network (red: low elevation sites, blue: high elevation sites). and not to continental to global- (b) Temperature offset (new minus old in °C) in the late 19th century for four typical surroundings scale changes. Two examples from due to changes in observation time (red: urban, green: extra-alpine rural, light blue: inner-alpine the European Alps illustrate that <1000m asl., dark blue: inner-alpine >2000m asl.) there is not only variability in time

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 10 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 11

REFERENCES Auer, I., Böhm, R. and Schöner, W., 2001: Austrian �� long-term climate 1767-2000 – Multiple instru- �� mental climate time series from Central Europe. �� Österr. Beitr. Zu Meteorologie und Geophysik 25, �� 147 pp (plus Data- and Metadata-CD) �� Böhm, R., Auer, I., Brunetti, M., Maugeri, M., Nanni, �� T. and Schöner, W., 2001: Regional temperature �� variability in the European Alps 1760-1998 from �� homogenized instrumental time series. Interna- tional Journal of Climatology 21: 1779-1801 � Jones P.D., 1994: Hemispheric surface air tempera- � � � � � � � � � � � � � � � � � � ture variations: A reanalysis and an update to � � � � � � � � � � � � � � � � � � 1993. Journal of Climate 7: 1794-1802 Moberg, A. and Alexandersson, H., 1997: Homog- Figure 3: Development of the potential of instrumental climate data for paleo-climate recon- enization of Swedish temperature data. Part II: structions. Homogenised gridded air temperature compared but also considerable variability in the second half of the 19th century. with a subset of global air temperature since 1861. International Journal of Climatology 17: 35-54 space—horizontally (Fig. 2 left) and Since then, well-homogenized and Peterson, T.C., Easterling, D.R., Karl, T.R., Groisman, vertically (Fig. 2 right). spatially dense instrumental in- P., Auer, I., Böhm, R., Plummer, N., Nicholis, N., Figure 3 presents a timeline formation has been available to Torok, S., Vincent, L., Tuomenvirta, H., Salinger, J., Førland, E.J., Hanssen-Bauer, I., Alexandersson, that summarizes our conclusions calibrate proxy data. For the one H., Jones, P. and Parker, D., 1998: Homogeneity on the quality, availability and po- hundred years in between, both Adjustments of In Situ Atmospheric Climate Data: tential of instrumental data for the sources (instrumental and proxy) A Review. International Journal of Climatology 18: 1493-1517 reconstruction of climate. There is are reliant on each other, with only no question that before the mid- a combination of different sources For full references please consult: 18th century the very rare instru- providing reasonable information www.pages-igbp.org/products/newsletters/ref2003_2.html mental information that existed is about past climate variability. That inferior to high-quality reconstruc- period may therefore best be called tions based on proxies. The real the “consistency period”. calibration period started as late as

Climate and Environmental Reconstructions from Scandinavian Varved Lake Sediments

ANTTI E.K. OJALA1, MATTI SAARNISTO1 AND IAN F. SNOWBALL2 1Geological Survey of Finland; [email protected] ; [email protected] 2Quaternary Geology, Lund University, Sweden; [email protected]

Introduction This myriad of factors is the prin- Annually resolved proxy records, cipal cause of distinct rhythmic such as varved lake sediments, sedimentation and of the forma- are important archives of past cli- tion and preservation of varves mate variability and environmental in lakes (Renberg, 1982; Ojala et change. The primary strength of any al., 2000, Zillén et al., in press). varved archive is that it contains a Varves in small to mid-sized lakes continuous and inherent calen- in Finland and Sweden typically dar year timescale that forms a appear as clastic-biogenic couplets temporal framework for paleoen- (Fig. 1) (Petterson, 1996; Ojala et vironmental reconstruction. In ad- al., 2000), which sometimes cover dition, many physical, chemical and the entire past glacial history of the biological parameters can be used basin, up to 10,000 years or more. to accurately infer fl uxes. The struc- A typical Fennoscandian lake varve ture and composition of varves are Figure 1: Images of Lake Nautajärvi (central is composed of an allochthonous governed by a multitude of inter- Finland) clastic-organic nonglacial varves fi ne-grained minerogenic layer at AD 200 taken from the fresh sediment acting factors in different sedimen- surface (left) and thin-section, using a back- deposited as a result of the spring tary environments that, in the best scattered mode of surface scanning electron snowmelt, grading into an organic cases, can be related to external microscope (right). summer layer formed by primary forcing factors and subsequently seasonality, which causes several productivity and a thin dark-colored interpreted as records of regional months of ice cover on lakes, peak layer composed of fi ne-grained or- climate change (e.g. Hughen et al., fl oods during the spring snowmelt, ganic matter that settles during the 2000). summer (and winter) stratifi cation winter when the lake is ice-covered The contemporary annual cli- of the water column, algal blooms (Renberg, 1982). In many cases, the mate in Fennoscandia has strong and occasional autumn storms. varves are best seen after oxidation

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 10 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 11

�� the intensity of spring discharge is �� predominately determined by snow �� storage and the water equivalent of �� snow in spring (Kuusisto, 1984). Of �� course, the rate of catchment ero- ��

� sion also depends on the supply of � �� material in the catchment and tim- � � ��

� ing of the most intensive runoff in � ��

� � relation to soil thawing. In contrast � �� to many biological proxies that �� reflect summer conditions, the �� physical properties of clastic-or- �� ganic varves are related to climatic �� conditions that prevail over the �� winter months, primarily temperature and the amount of snow that Figure 2: Precisely dated high-resolution reconstructions of geomagnetic field secular varia- accumulates in the catchments of tions for the last 10,000 years, based on a network of varved lake sediments in Finland and Sweden, has resulted from cooperation between the Geological Survey of Finland and the the individual lakes. Geological Institute at the University of Lund in Sweden. The example shows a varve dated Many of the varved lake sedi- inclination master curve for Nautajärvi in central Finland (a) and a stacked inclination master ment sequences under investi- curve from Sarsjön and Frängsjön in northern Sweden (b). Both curves are based on the spheri- cal smoothing of data with a moving time-window of 150 years (Snowball and Sandgren, 2002). gation in Finland and Sweden The potential to correlate paleomagnetic features and transfer ages with an approximate error are located in areas that exhibit of less than 200 years is demonstrated by the comparison with paleomagnetic data obtained considerable year-to-year variabil- from the site of Kälksjön (c), which is located in central Sweden and contains uncounted varves. An amalgamated reconstruction of the total field vector (direction and intensity) is the ity of winter temperature and pre- aim of ongoing collaboration. This reconstruction will refine the ages of the regional features cipitation. One of the phenomena and provide a data set that can be used to distinguish solar and geomagnetic influences on known to regulate winter climate rates of atmospheric cosmogenic nuclide production. in the North Atlantic–European of the sediment, when the annual clastic-biogenic varves (thickness, region is the atmospheric circula- iron cycle is highlighted. density, composition and texture) tion pattern called North Atlantic Recent publications report long is mainly indicative of the annual Oscillation (NAO) (e.g. Hurrel, 1995; clastic-biogenic varved sequences influx of mineral matter into the Luterbacher et al., 2002). During in the lakes of Nautajärvi, Alim- basins, which depends principally years of positive NAO, cyclones mainen Savijärvi, Korttajärvi (Ojala on the rate of catchment erosion originating in the Atlantic sweep & Saarinen, 2002; Ojala et al., in and the transportation of detritus westwards across Fennoscandia press; Ojala & Tiljander, in press), during the intense spring snowmelt and the Baltic Sea region, causing Kassjön (Petterson et al., 1999), flood. In the boreal vegetation zone, relatively mild and wet winters. On Sarsjön, Frängsjön (Snowball et al., 2002; Snowball & Sandgren, 2002), ��

Furskogstjärnet, Mötterudstjärnet � �� (Zillén et al., 2002) and Kälksjön � � (Zillén et al., in press). Counting of the annual laminations indicates that the error estimate of these varve chronologies can be pushed

down to ±1%. Such accurate and � �

precise multiple chronologies form � � an excellent basis for between- � lake comparisons of paleodata and � �

provide a template for assessing � the accuracy of correlations made using other techniques, such as paleomagnetic dating and tephrochronology (Ojala and Saarinen 2002, ��

Pilcher et al., this issue, Snowball � ��

and Sandgren, 2002, Zillén et al., � 2002) (Fig. 2). �� Basis of Climate Reconstruction High- and low-frequency variabil- Figure 3: Annual mineral and organic matter accumulation in Lake Nautajärvi central south- ity of the physical properties of ern Finland since 7902 BC based on studies of relative X-ray density.

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 12 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 13

the other hand, years of negative However, due to local system dy- between the major factors that NAO are characterized by colder namics, it is evident that the rela- drive the sedimentation. and drier winters. It appears that tionships between external forcing these lakes are located at a sensi- factors and the annual accumula- REFERENCES Ojala, A.E.K. and Saarinen, T., 2002: Palaeosecular tive climatic boundary, where just tion of mineral/organic matter are variation of Earth’s magnetic field during the a few degrees warming during the neither linear nor stable over time. last 10,000 yrs based on the annually laminated winter (+NAO) results in a consid- Post-isolation catchment stabiliza- sediment of Lake Nautajärvi, central Finland. The Holocene 12: 391-400. erably shortened period of snow tion, vegetation succession, human Ojala, A.E.K. and Tiljander, M., Testing the fidelity of cover, even to the point where disturbance and extreme events sediment chronology: comparing varve and paleo- no persistent snow cover forms like forest fires exert a consider- magnetic results from Holocene lake sediments from central Finland. Quaternary Science Reviews (Vehviläinen & Lohvansuu, 1991; able influence on catchment ero- (in press). Vehviläinen & Huttunen, 1997). As sion and the nature of the varved Snowball, I., Sandgren, P. and Petterson, G., 1999: a consequence, the peak discharge records. At the end of the day, the The mineral magnetic properties of an annually in spring is severely reduced, catch- properties of clastic-organic varves laminated Holocene lake-sediment sequence in northern Sweden. The Holocene 9: 353–362. ment erosion is significantly lower in Fennoscandia reflect the annual Snowball, I.F. and Sandgren, P., 2002: Geomagnetic and less allochthonous mineral flux of mineral matter to the sedi- field variations in northern Sweden during the matter is transported and depos- mentary basins via a multitude of Holocene quantified from varved lake sediments and their implications for cosmogenic nuclide ited on the lake bottom. One period interacting factors. The main aim production rates. The Holocene 12: 517–530. of weakened spring discharge and of future work is to use a variety of Zillén, L.M, Wastegård, S. and Snowball, I.F., 2002: erosion has been identified from analytical methods, such as high- Calendar year ages of three mid-Holocene tephra layers identified in varved lake sediments in west Finnish varved sediments to have resolution digital image analysis central Sweden. Quaternary Science Reviews 21: occurred between AD 980-1250 (e.g. Tiljander et al., 2002, Ojala 1583–1591. (Saarinen et al., 2001, Tiljander & Francus, 2002, Saarinen & Pet- et al., in press). This period coin- terson, 2002), mineral magnetism For full references please consult: cides with the Medieval Climate (Snowball et al., 1999), pollen, and www.pages-igbp.org/products/newsletters/ref2003_2.html Anomaly. Its climate implications charcoal analysis to discriminate are under intensive investigation. 14C as an Indicator of Solar Variability

TOMASZ GOSLAR Adam Mickiewicz University, Faculty of Physics and Poznan Radiocarbon Laboratory, Poznan, Poland; [email protected] Introduction

The quasi-constant concentration of �

14 � C on the Earth is maintained by a ��

balance between its radioactive de- � �� cay and its production from atmo- �� spheric nitrogen. This production � ��

is possible due to the high-energy � ��

protons from cosmic radiation. At- � � �� 14 �

oms of C enter the biogeochemi- � � � � cal cycle, where carbon circulates � � �

between the atmosphere, terrestrial �� biosphere and the oceans. Since � �

the characteristic exchange rates �� between reservoirs and times of �� circulation within individual reser- �� voirs are significant with respect to the half-life of 14C, radiocarbon Figure 1: Comparison of radiocarbon production rate (Q/Qo) and records of solar activity in concentration on the Earth is not recent centuries. uniform, being the highest in the sponds to a decrease in the flux of Reconstruction of Atmospheric atmosphere and distinctly lower in galactic cosmic rays impinging on 14C Levels in the Past the deep oceans. the Earth. The intensity of cosmic Past variations in atmospheric 14C Charged particles emitted from radiation in the atmosphere is also concentration have been recon- the Sun produce a magnetic field modulated by the Earth’s magnetic structed as a “by-product” of rain interplanetary space. This field field. To first-order approximation, diocarbon calibration. During the inhibits penetration of low-energy the globally averaged 14C produc- Holocene period, this calibration is cosmic-ray protons into the center tion rate is inversely proportional to based on 14C ages of bidecadal or of the solar system. As a rule, an the square root of the geomagnetic decadal tree-ring samples, dated increase in solar activity corre- dipole moment (Lal, 1988). absolutely by means of dendro-

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 12 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 13

�� Similar maxima, between AD 1290-

� ��

� 1390, 1400-1600, and 1800-1860 are � � �

� known as Wolf, Spörer and Dalton, �

� � � � �

� � � � � � � � � � � � �� respectively. Since the nineteenth � �� century, atmospheric radiocarbon �

� ��

� concentration has been seriously � �� affected by fossil fuel combustion �

� �

� and nuclear tests. The first effect � can be taken into account in the �� calculations, based on the known history of fossil fuel consumption. Figure 2: Estimate of Sun-induced fluctuations in14 C production rate over the last 10,000 years. The second effect precludes model Right-hand side: Frequency distribution of bidecadal production rate averages. The sunspot calculations of 14C production after number scale is based on the correlation in recent centuries. 1950. chronology (Stuiver et al., 1998). (e.g. Bard et al., 1997). On the other Production of cosmogenic iso- These data document ±20‰ fluctua- hand, model considerations using topes depends on both the Earth’s tions in atmospheric Δ14C, usually archeomagnetic data lead to the and the Sun’s magnetic field. It can 100-300 years long, superimposed conclusion that the long-term de- be expressed as: on an almost monotonic downward crease in Δ14C results from secular trend from ca. 120‰ at the begin- changes in the geomagnetic field Q = C * Ageo * Asol ning of the Holocene to ca. 0‰ at (Stuiver et al., 1991). the end of the 19th century. In historical records, solar ac- where C denotes the production Over the last few centuries, tivity is quantitatively represented rate by cosmic rays not attenuated 14C concentration has also been by the number of sunspots (Wolf by solar and geomagnetic fields,

measured in single-year tree-ring number). Although correlation be- and Ageo and Asol are the respective samples (e.g. Stuiver and Brazi- tween sunspot number and past attenuation factors. The latter geo- unas, 1993). These data reveal qua- 14C changes has been discussed in magnetic factor has been modeled si-periodic Δ14C fluctuations with an numerous papers, only a few have using a reconstruction of the Earth’s amplitude as low as 2.5‰. directly compared sunspot number magnetic field, as in previous stud- with 14C production rate (e.g. Stuiv- ies (e.g. Goslar, 2001). The Relationship Between 14C er and Quay, 1980). This correlation There is a striking correlation Production and Solar Activity is presented here (Fig. 1). 14C pro- between mean sunspot number in in Recent Centuries duction rates are calculated from consecutive sunspot cycles (S) and As noted above, there is a more annual values of 14C concentrations average values of 14C production direct relationship between solar (Stuiver and Braziunas, 1993) with (Q). Although annual values of 14C activity and the rate of 14C pro- the PANDORA model of the global production also correlate signifi- duction than 14C concentration. carbon cycle (used, for example, by cantly with annual sunspot num- Fluctuations in 14C concentration Goslar et al., 2001). ber, the single-year Q-S regression can be caused by changes in 14C The maximum 14C production coefficient is distinctly lower than production; this may also reflect rate in the last millennium occurred that for cycle averages. One reason solar and/or geomagnetic variabil- between AD 1645 and 1715, during for this is that annual values of Q ity, as well as changes in the global the period when sunspots were al- are different in different minima of carbon cycle. In general, no single most lacking (Maunder Minimum). solar activity, despite very similar one of these causes can be isolated minimum sunspot numbers (i.e. from the others unless additional �� close to 0). This means that solar information is available. Such in- activity expressed by the sunspot formation is provided by records number is not unequivocally re- 10 �� 14

of Be in the ice caps of Greenland � lated to C production. Indeed, it �

and Antarctica. Like radiocarbon, � has been known for some time

10 �

Be is produced in the atmosphere � that sunspots are not directly tied �

� �

by cosmic rays, however it falls � to the magnetic properties of the out almost immediately, hence its solar wind. A more direct measure concentration in sediments reflects of solar magnetic field properties is �� mostly local cosmogenic produc- provided by Aa indices, which re- tion rates. Fine correlation between flect the magnitude of short-term � �� 100- to 300-year-long features of �� magnetic activity measured in two the 14C and 10Be records of the last antipodal laboratories. Both sun- Figure 3: Shapes of Maunder-type (red), and earlier millennia indicate that Spörer-type (blue) and Wolf-type (black) spot number and Aa index reveal these features are related to Sun- maxima in 14C production rate during the an 11-year cycle but, unlike sun- induced changes in production rate Holocene. spots, the minimum Aa differs for

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 14 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 15

different cycles and the minimum flects the changes in geomagnetic mum (AD 1290-1390), we propose values appear correlative to the field, Sun-induced changes of the referring to these shorter events as cycle-averages of sunspot numbers 14C production were calculated Wolf-type. (Fig. 1). Nevertheless, since recon- (Fig. 2). struction of Δ14C in the Holocene The distribution of Q/Qo values REFERENCES relies on bidecadal data, in using has no right-hand tail (i.e. almost no Bard, E., Raisbeck, G.M., Yiou, F. and Jouzel, J., 1997: Solar modulation of cosmogenic nuclide produc- it as a proxy of solar activity, the value exceeds 130%). This is consis- tion over the last millennium: comparison between relationship between Q and S cycle- tent with equation 1, which requires 14C and 10Be records. Earth and Planetary Science averages still seems applicable. that the attenuation not be larger Letters 150: 453-462. Goslar T., 2001: Absolute production of radiocarbon than 1 (as it was, for example, dur- and the long-term trend of atmospheric radiocar- Characteristic Features of 14C ing the Maunder Minimum). This bon. Radiocarbon 43: 743-749. Production in the Holocene consistency provides additional Lal D., 1988: Theoretically expected variations in 14 the terrestrial cosmic-ray production of isotopes. The Δ C record in the Holocene support for the interpretation of [w:] Solar-Terrestrial relationships and the Earth allows the 14C production rate to the obtained record in terms of environment in the last millennia. Societa Italiana be calculated provided the param- past solar activity. di Fisica 95: 216-233. eters of the global carbon cycle are Stuiver and Braziunas (1988) Stuiver M., Braziunas T., 1993. Sun, ocean, climate and atmospheric 14CO2: an evaluation of causal known. In previous studies (e.g. noted distinct similarities between and spectral relationships. The Holocene 3: Stuiver and Braziunas, 1988), it has fluctuations in 14C production, and 289-305. usually been assumed that those distinguished nine Maunder-type Stuiver M., Reimer P.J., Bard E., Warren Beck J., Burr G.S., Hughen K.A., Kromer B., McCormac G., parameters were constant and the and eight Spörer-type maxima. vanderPlicht J. and Spurk M., 1998. INTCAL98 same as today. The record of past Calculations using up-to-date 14C Radiocarbon age calibration, 24,000-0 cal BP. 14C production rates calculated ac- calibration data (Stuiver et al., Radiocarbon 40: 1041-1083. cording to this assumption shows 1998) suggest that some maxima, For full references please consult: 100- to 300-year-long fluctuations previously qualified as Maunder, www.pages-igbp.org/products/newsletters/ref2003_2.html superimposed on a long-term are distinctly shorter (Fig. 3). Since trend. Assuming that this trend re- this group includes the Wolf maxi-

Reconstruction of Low- and High-Frequency Summer Temperature Changes From a Tree-Ring Archive of Fennoscandian Forest-Limit Scots Pine

MATTI ERONEN1, MARKUS LINDHOLM2, SAMULI HELAMA1, JOUKO MERILÄINEN2 AND MAURI TIMONEN3 1 Department of Geology, University of Helsinki, Finland;[email protected]; [email protected] 2 Saima Centre for Environmental Sciences, Savonlinna, Finland; [email protected]; [email protected] 3 Finnish Forest Research Institute, Rovaniemi, Finland; [email protected] Material and its resolution to tree-ring data, inter- Ecogeographical Setting annual-to-decadal scale variability

��

Tree-ring samples of Scots pine �� is resolved by most methods of

��

�� (Pinus sylvestris L.) were collected �� building chronologies. Variation

��

from living trees, dead standing �� at the multi-centennial time-scale

��

�� logs, old buildings, and subfossil has recently been successfully ex- wood from peat-bogs and small �� tracted from the northern data us- ��

��

lakes (Eronen et al., 2002). �� ing specified techniques in signal

��

�

��

� �� The selected dataset containing �� extraction (Helama et al., 2002). �� 1081 tree-ring series. The latter �� Annual accuracy of dendrochrono- archive is the major source of �� logical dating is a prerequisite for

�

��

samples. �� several paleoclimate examinations,

�

The area is situated between �� including quantitative multi-disci-

68° and 70° N, 20° and 30° E, in the �� plinary comparisons, multi-proxy northern Fennoscandia (Fig. 1). Ho- reconstructions and climate forc- mogeneity of tree-ring data over the Figure 1: Sampling sites in northernmost Fen- ing studies (e.g. Ogurtsov et al., geographical distribution was dem- noscandia, with national coordinates. Present 2002). The further extension of this onstrated (Lindholm, 1996). Highly pine-limit is marked with a broken line. chronology may become possible consistent tree-ring chronologies Chronology Temporal Ranges given that there is a gap between may be built from diverse sites as At present, the northern Finnish the time of arrival of pine, as indi- well as from various age-classes of supra-long pine chronology spans cated by pollen analysis, and the pine trees (Lindholm et al., 2000; from 5520 BC to AD 2001 (Eronen beginning of the earliest tree ring Lindholm et al., submitted). et al., 2002). As dendrochrono- series (Eronen et al., 2002). On the logical cross-dating yields annual other hand, possibly conditions for

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 14 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 15

subfossil preservation were not fa- � ��

vorable in the early Holocene. � � � � � �

Reconstruction of Climate � � � �

Variables � �

� �

Radial growth (tree-ring width) � � � of northern forest-limit pines is � ��

limited by concurrent mean July � �

temperatures over the region under � � study. This particular relationship � � � � becomes weaker with increas- � � � �

ing distance from the forest limit � � �

(Lindholm et al., 2000). In previ- � � � �

ous studies, approximately 40 to � �

� ��

50% of independent mid-summer � �

temperature variations were cap- � � tured by transfer models using a �� single (ultra-long) tree-ring width �� chronology. Reconstruction of mid- Figure 2: Finnish Lapland pine tree-ring-based mid-summer reconstruction extending back summer temperatures is elongated to 5520 BC (data from Eronen et al. 2002; Helama et al. 2002) shown as annual (blue) and centennial (red) anomalies from long-term mean. over the last 7.5 millennia (Fig. 2). Helama, S., Lindholm, M., Timonen, M. and Eronen, Among climate-related variables, 13% of the total chronology length M., 2002: The supra-long Scots pine tree-ring North Atlantic Oscillation (NAO) is covered by ten or less samples. record for northern Finnish Lapland – Part 2: was reconstructed by Lindholm et interannual to centennial variability in summer temperatures for 7500 years. The Holocene 12: al. (2001) using a network of pines Future Prospects 681-687. mainly from Finland and Northwest In addition to increasing sample Lindholm, M., 1996: Reconstruction of the past Russia. replication, the intra-annual ring- climate from ring-width chronologies of Scots pine (Pinus sylvestris L.) at the northern forest limit in density record can strengthen Fennoscandia. Publications in Sciences 40. PhD Chronology Signal Strength the desired temperature signal. thesis. University of Joensuu. 169 p. The confidence of a climate signal The maximum latewood density Lindholm, M., Eggertsson, Ó., Lovelius, N., Raspopov, is measured as a function of sample variable could be used along with O., Shumilov, O. and Läänelaid, A., 2001: Growth indices of North European Scots pine record replication and inter-correlation. Ex- ring-width variables in transfer the seasonal North Atlantic Oscillation. Boreal traction of signals of low frequency functions. Interestingly, chronol- Environment Research 6: 275-284. climate variation from a given long ogy could certainly span over the Lindholm, M., Lehtonen, H., Kolström, T., Meriläinen, J., Eronen, M. and Timonen, M., 2000: Climatic chronological sequence requires a debated 8.2 k event after an elonga- signals extracted from ring-width chronologies of greater number of samples than tion of one millennium. Scots pine from the Northern, Middle and South- would be needed to obtain the ern parts of the boreal forest belt in Finland. Silva Fennica 34: 317-329. same level of reliability in signals REFERENCES Eronen, M., Zetterberg, P., Briffa, K., Lindholm, of high frequency variability. The M., Meriläinen, J. and Timonen, M., 2002: The For full references please consult: annual sample replication exceeds supra-long Scots pine tree-ring record for northern www.pages-igbp.org/products/newsletters/ref2003_2.html thirty trees over 36% of the total Finnish Lapland – Part 1: chronology construction and initial inferences. The Holocene 12: 673-680. chronology length. Approximately

Climate Reconstruction From Peatlands

DAN CHARMAN1 AND MARKKU MÄKILÄ2 1Dept Geographical Sciences, University of Plymouth, Plymouth, UK; [email protected], 2Geological Survey of Finland, Espoo, Finland; [email protected] Peatlands occur throughout the mid long, continuous sequences from with radiocarbon, especially where to high latitudes of the northern the early to mid Holocene up to the mosses dominate the peat. There- hemisphere, with huge expanses present day. Accumulation rates are fore, peatlands potentially provide in low-relief areas of the oceanic highly variable, but 0.5-1 cm per a widespread source of well-dated and boreal zones. Smaller areas decade is typical for raised bogs in paleoenvironmental data, covering occur in mountainous regions Europe. For example, the long-term most of the Holocene at decadal throughout the world, in oceanic average peat accumulation rate for resolution. Despite this, their use temperate zones in the southern 210 raised bog profiles in Finland for paleoclimatic reconstructions hemisphere, and in the tropics. is 0.6 cm per decade. The deposits has still been limited to a relatively The vast majority of the peat is are mainly autochthonous and small number of studies in a few Holocene in age, often forming relatively straightforward to date main locations. Recent innovations

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 16 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 17

that use multiple records from several sites in the same climate area to provide composite records (Fig. 2). Many of the records from northwest Europe display cyclicities with periodicities of varying lengths including possible solar-related cycles (Chambers and Blackford, 2001) and links to ocean variability on longer time-scales (Hughes et al. 2000). One key issue in current research is to resolve how long-term changes in surface wetness relate to seasonal and annual climate variability. Con- ceptual process models suggest that surface wetness changes reflect some combination of temperature, Figure 1: Vegetation and surface patterns (hummocks and hollows) of a raised bog (Kilpisuo) in precipitation, humidity and wind southern Finland, one of the sites used in studies of carbon accumulation and climate change speed. The precise nature of this re- (Mäkilä, 2001; Mäkilä, et al., 2001). lationship is unknown and is likely in techniques applicable to peat- et al., 2000), and mercury analysis to be different in different climate lands, better chronological precision (Martinez-Cortizas et al. 1999). zones. Variability in surface wet- of the record, and improvements in •Records of changing mass and ness is probably primarily related to our understanding of how peat- carbon accumulation rates. In- summer climate because peatlands lands record climate history, mean creased accumulation rates may are either saturated or frozen in the that the time is now ripe to begin reflect periods of more positive PE winter months. Comparisons with to exploit these deposits more fully. balance (Mäkilä et al., 2001; Mäkilä, documentary data suggest that both Here, we outline some of the main 2001) (Fig. 1). precipitation and temperature may approaches, current research and •The timing and rate of spread of have determined surface wetness ways in which peat records may peat initiation. The most successful in northern England over the past be developed in the future to meet of these studies have been in west- 900 years (Barber et al., 1994) (Fig. the needs of Holocene paleoclimate ern Canada using large data sets 2). Comparisons of high-resolution research. from very wide geographic areas, records with instrumental climate where peat initiation data have been data confirm that summer climate is Approaches to Reconstruct Past integrated with other paleoclimatic the main driver of wetness changes Climates Using Peat Deposits: indices and climate models (Gajews- but that the relationship with tem- •Records of changing surface ki et al., 2000), and cyclicities similar perature and precipitation probably wetness. Most techniques aim to to those found in ocean and ice core varies in different locations across reconstruct past surface wetness as records have been shown (Campbell Europe (Table 1). Under a prevail- a proxy for precipitation-evaporation et al., 2000). ing oceanic climate such as that (PE) balance. These include plant in northern England, summer pre- macrofossil analysis, peat humifica- Reconstruction of past surface wet- cipitation may be the main factor re- tion and testate amoebae analysis. ness changes is the basis of many flected in reconstructed water table The surface wetness of all peatlands existing records, including some variability. Under a more continen- is dependent to some extent on PE Table 1: Linear correlation coefficients between reconstructed water table records from two but in fact ombrotrophic peatlands European raised bogs (based on testate amoebae analysis) and instrumental climate data. (raised bogs or watershedding blan- The Butterburn Flow record covers AD 1800-1990 and is based on decadal averages (n=20 ket mires) are the only sites where for temperature, n=17 for precipitation). The Männikjärve record covers AD 1951-1995 and is there is no influence of surface run- based on 5-year averages (n=9). *p<0.05. From Charman et al. (submitted). off or groundwater whatsoever, and Climate variable Butterburn Flow Männikjärve bog precipitation and evapotranspiration England Estonia are the only components of water Temperature balance. Summer -0.046 *-0.698 •Records of temperature change. Growing season -0.127 -0.176 Recent work has shown the consid- Mean annual 0.026 -0.060 erable potential of proxies that re- Precipitation cord temperature directly. Principal Summer *0.458 0.437 amongst these are species-specific and compound-specific isotopic Growing season 0.106 0.377 analysis (e.g. Pendall et al., 2001; Xie Mean annual 0.188 *0.585

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 16 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 17

tal regime such as that in northern ��

Europe, winter precipitation stored � � �

� � �

as snow can also have an effect on � �

� � a) � � � � �

reconstructed water tables. Snow � � � � � � � melts rapidly in springtime filling � � � ��

hollows and causing heavy floods. �

� �� Summer temperature also exerts a � ��

� � ��

much stronger desiccating effect in � �� b) more continental sites, affecting the � �� moisture balance in surface peats di- �� rectly. New and improved analytical � �� c)

approaches and better understand- � � ing of the peat-climate system now � � ��

suggest a number of ways forward � ��

for maximizing the use of the peat � � �� record. � �

��

�� Key Areas for the Improvement d) �� of Climate Reconstructions: �

�� •Understanding the peatland– � ��

climate relationship. Calibration of �� records with instrumental data is �� critical in deriving empirical rela- �� tionships between changes in proxy �� data and climate parameters (see Figure 2: Mean annual water table changes for the last 1500 years, reconstructed from testate above). Development of physical amoebae analysis of three peat profiles in northern England c( ), compared with documentary process models from hydrological, index of summer wetness and winter severity (a) and ice accumulation record from GISP2 (d). Instrumental records of mean annual temperature and precipitation are also shown (b). All biological and meteorological data data are 100-year running averages. Raw documentary and water table data are also shown. would lead to greater confidence Fluctuations in the reconstructed water tables, documentary indices and ice accumulation in empirical relationships and ex- show similar trends over this time period. See Charman and Hendon (2000). trapolating calibrated records over peatland records and other proxies peatland paleoclimate records can pre-instrumental time periods. Com- such as speleothem, ice core and now be fully integrated with other parisons with other independent ocean records (e.g. Charman and Holocene paleoclimate reconstruc- climate proxies will also help test Hendon, 2000; Charman et al., 2001; tions. hypothesized surface wetness–cli- Pendall et al., 2001) are limited by mate links. chronological uncertainty. Cali- REFERENCES •Further development of new brated radiocarbon ages still often Charman, D.J., Brown, A.D., Hendon, D., Kimmel, A. and Karofeld, E., Testing the relationship between proxy climate techniques. Isotopic, mean possible 2σ errors of ±200 Holocene peatland palaeoclimate reconstruc- compound-specific and mercury years—critical in trying to establish tions and instrumental data. Quaternary Science analyses are now producing exciting whether climate variability on cen- Reviews (submitted). Charman, D.J. and Hendon, D., 2000: Long-term results. These approaches need to be tennial scales is in or out of phase. changes in soil water tables over the past 4500 explored more extensively. Wiggle-matched radiocarbon ages years: Relationships with climate and North •Establishment of additional (Mauquoy et al., 2002) and teph- Atlantic atmospheric circulation and sea surface temperature. Climatic Change 47: 45-59. records from different climatic rochronology (e.g. Van Den Bogaard Charman, D.J., Caseldine, C., Baker, A., Gearey, B.R. regimes. Much existing work has and Schmincke, 2002) both have the and Hatton, J., 2001: Palaeohydrological records been based in oceanic temperate potential to significantly improve from peat profiles and speleothems in Sutherland, northwest Scotland. Quaternary Research 55: regions. Additional work on conti- chronologies and the precision of 223-234. nental, alpine and higher latitude correlations with other records. Mäkilä, M., 2001: Climate in relation to carbon areas is needed to fully exploit the accumulation on Kilpisuo, a raised bog in southern potential peatland archive. Much of Conclusions: Finland. In: Vasiliev, S. V., Titlyanova, A. A. and Velichko, A. A. (eds.) West Siberian peatlands and this work is already underway, as is One of the main HOLIVAR themes carbon cycle: Past and present: Proceedings of an reflected in recent results from cen- is the comparison of different proxy International Symposium Noyabrsk. pp.38-40. tral Europe (Mitchell et al., 2001), climate records—with all the atten- Mäkilä, M., Saarnisto, M. and Kankainen, T., 2001: Aapa mires as a carbon sink and source during the continental North America (Booth, dant problems of different chronolo- Holocene. Journal of Ecology 89: 589-599. 2002), South America (Pendall, 2001) gies, climate parameters recorded and New Zealand (e.g. McGlone and and geographical coverage. With For full references please consult: Wilmshurst, 1999; Wilmshurst et al., the development of a wider range of www.pages-igbp.org/products/newsletters/ref2003_2.html 2002). techniques and a better understand- •Improving chronological con- ing of the proxy-climate relationship trol. Existing comparisons between linked to improved chronologies,

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 18 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 19

Climate Forcing During the Holocene

RAYMOND S. BRADLEY Climate System Research Center, Department of Geosciences, University of Massachusetts, Amherst, USA; [email protected] What caused Holocene climate variations? On the very longest, multi-millennial timescales, the main factors affecting Holocene climate change are related to orbital forcing (changes in obliquity, pre- cession and eccentricity). Decadal- to century-scale changes may be related to solar forcing (changes in total solar irradiance) and changes on shorter timescales (decadal to inter-annual) are most likely related to volcanic forcing (Bradley, 2003). Not all climatic variability seen in the Holocene can be ascribed to specific external forcings. For example, the ~8200 calendar year BP “event”, seen in many paleoclimatic archives, resulted from catastrophic pro-glacial lake drainage at the margins of the Laurentide Ice Figure 1: Schematic representation of insolation anomalies at the top of the atmosphere Sheet (Alley et al., 1997; Barber et relative to 1950 levels. Anomalies are color-coded in W m-2. The x-axis of each monthly panel al., 1999). This rapid flooding of the shows changes from 10,000 calendar years BP (left) to today (right). The y-axis shows latitude North Atlantic with freshwater had a from 90ºN (top) to 90ºS (bottom) (figure prepared by Anne Waple). significant regional climate impact, This resulted in higher summer Holocene paleoclimatic studies unrelated to any external forcing. insolation in the early Holocene at that claim solar forcing has driven Internal modes of climate system all latitudes of the northern hemi- observed changes, based largely variability (e.g. El Niño-Southern sphere (ranging from ~40W m-2 on comparisons of proxy records Oscillation, North Atlantic Oscilla- higher than today at 60ºN to 25W with the 14C anomaly series. It re- tion, Pacific Decadal Oscillation) m-2 higher at the Equator). Thus, mains to be seen how robust these that probably vary on both long July insolation (radiation at the relationships are, and what plau- and short timescales—though we top of, or outside, the atmosphere) sible mechanism might link solar know relatively little about their has slowly decreased over the last activity/irradiance variations with long-term behavior—may also pro- 12,000 years (Fig. 1). Anomalies climate in widely separated parts duce significant changes in regional during southern hemisphere sum- of the globe, from the Arabian Sea climate. And stochastic resonance mers were smaller, centered at and Africa, to the North Atlantic and in the climate system—by which a lower latitudes, and were opposite Greenland. One possibility (seen in weak quasi-periodic forcing signal in sign (that is, insolation increased some model simulations) involves may be amplified into a non-linear, over the course of the Holocene). solar variations influencing the bi-stable climate signal—may have For example, January insolation Hadley circulation (intensity and/or brought about relatively abrupt anomalies at 20ºS were ~30W m-2 extent), which then leads to tropi- changes in the past by pushing the below current values in the early cal and sub-tropical precipitation system across critical thresholds Holocene. anomalies, and to further telecon- (Ruzmaikin, 1999). Thus, the spec- Changes in solar irradiance nections to extra-tropical regions. trum of Holocene climate variabil- (energy emitted by the Sun) might It is well known, from stud- ity reflects the combined effects of be expected to affect all parts of ies of instrumental records, that all of these forcings and feedbacks the Earth equally. However, this explosive volcanic eruptions can within the climate system. is not so because the response to have short-term cooling effects on Orbital forcing involves a signifi- solar irradiance forcing is amplified overall hemispheric or global mean cant redistribution of energy, both regionally, as a result of feedbacks temperatures (Robock, 2000). This seasonally and latitudinally (Fig. 1). and interactions within the atmo- results from direct radiative effects, In the early Holocene, precessional sphere (Rind, 2002). The record of with the volcanic aerosol reducing changes led to perihelion at the solar variability in the Holocene is energy receipts at the surface, plus time of the northern hemisphere discussed further in the accompany- associated circulation changes that summer solstice (whereas today, ing article by Goslar. Here, I would may result from such effects. Most it is closer to the winter solstice). only note that there are many temperature effects are not detect-

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 18 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 19

�� recent centuries suggest that these two factors can explain most of the �� variability of temperatures over �� the last millennium (e.g. Crowley, 2000). It therefore seems likely that �� they have also played a significant role in overall Holocene forcing on �� inter-annual to century timescales. � � �� REFERENCES �� Alley, R.B., Mayewski, P.A., Sowers, T., Stuiver, M., Taylor K.C. and Clark, P.U., 1997: Holocene climatic Figure 2: Holocene record of volcanic sulfate (anomalies from background variations) recorded instability: a prominent, widespread event 8200 yr in the GISP2 ice core from Summit, Greenland (Zielinski et al., 1994). ago. Geology 25: 483-486. able after a few years, so individual eruptions in the past, such as in the Barber, D.C., Dyke, A., Hillaire-Marcel, C., Jennings, explosive eruptions only contribute period 9500-11500 calendar years A.E., Andrews, J.T., Kerwin, M.W., Bilodeau, G., McNeely, R., Southon, J., Morehead, M.D. and short-term variability to the spec- BP (Fig. 2) (Zielinski et al., 1994) Gagnon, J.M., 1999: Forcing of the cold event trum of Holocene climate. However, Furthermore, in the early Holocene of 8,200 years ago by catastrophic drainage of if eruptions were more frequent in there were many large volcanic Laurentide lakes. Nature 400: 344-348. Bradley, R.S., 2003: Climate forcing during the the past, or if they happened to oc- signals, greater in magnitude than Holocene. In: Global Change in the Holocene: cur in clusters, it is possible that those recorded after the eruption of approaches to reconstructing fine-resolution the cumulative effect of eruptions Tambora (1815), the largest erup- climate change. A.W. Mackay, R.W. Battarbee, H.J.B. Birks & F. Oldfield (eds.). Arnold, London (in could have persisted for longer, re- tion in recent centuries. Of course, press). sulting in decadal- to multi-decadal a larger sulfate signal might simply Robock, A., 2000: Volcanic eruptions and climate. scale impacts. Such effects would mean that the eruption event was Reviews of Geophysics 38: 191-219. be enhanced if the initial cooling closer to the deposition site, so we Zielinski, G., Mayewski, P.A., Meeker, L.D., Whitlow, S., Twickler, M.S., Morrison, M., Meese, D.A., Gow led to feedbacks within the climate do not have a definitive long-term A.J. and Alley, R.B., 1994: Record of explosive vol- system, such as more persistent record of the magnitude of overall canism since 7000 B.C. from the GISP2 Greenland snow and sea-ice cover, which volcanic forcing, or more specifi- ice core and implications for the volcano-climate system. Science 264: 948-952. would raise the surface albedo cally, of atmospheric optical depth and possibly alter the atmospheric and its distribution latitudinally (cf. For full references please consult: circulation. Indeed, the non-seasalt Roberston et al., 2001). Energy bal- www.pages-igbp.org/products/newsletters/ref2003_2.html sulfate record from the GISP2 ice ance and GCM studies that have pa- core does suggest that there were rameterized the effects of explosive periods of more frequent explosive volcanism and solar forcing over

Tephra, a Powerful Tool for Precision Dating and Correlation

JON R. PILCHER1, RAY BRADLEY2, LESLEIGH ANDERSON2 AND PIERRE FRANCUS2 1School of Archaeology and Palaeoecology, Queen’s University, Belfast, Northern Ireland; [email protected] 2 Department of Geography, University of Massachusetts, Amherst, MA, USA; [email protected]; [email protected]; [email protected] Summary there is probably no part of the Icelandic tephra is proving to be globe from which such deposits a valuable new chronological are absent. The use of Icelandic and stratigraphical tool for the tephras in stratigraphic studies and palaeoenvironmentalist in NW Eu- as a dating tool is well developed rope. It is particularly useful in the in NW Europe. Tephra particles are historic period where radiocarbon usually less than 100 micrometers dating seldom provides adequate (commonly 20-60µ in diameter) and precision, and also in situations are glass (Fig. 1). In all the studies where sediments are unsuitable described here, the tephra is invis- for such dating because of variable ible in the cores and sections and marine influence. must be detected microscopically. The ideal medium for detection of Geographical Distribution and minute traces far from the parent Temporal Range volcano is a highly organic matrix Tephra is the air-fall component of such as peat. The tephra can be re- ejecta from a volcano. Tephras can leased by burning sub-samples of Figure 1: Tephra (width about 40µ) from the be found embedded in deposits in the organic material and dissolv- AD1158 eruption of Hekla found in peat from the Lofoten Islands. many areas of the world, in fact ing the peat ash in dilute acid. The

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 20 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 21

clean material can then be mounted The area has been subject to ma- to lake basin. Even better, we will in resin and examined under a mi- jor relative sea level changes, with also be able to correlate with other croscope. This process is described many lakes passing through both work on a whole transect along the in Pilcher and Hall (1992). Where marine and fresh water episodes. Atlantic seaboard of Europe at key the matrix has more than about This poses special constraints on dates such as 2310 BC (Hekla 4). 1% mineral content, separation of the value of radiocarbon dating. the tephra is more difficult. Sieving Following the successes of Chemical Fingerprinting of between mesh sizes of 26 and 75µ using Icelandic tephra for chrono- Tephras will retain most of the tephra and logical control in Ireland (Pilcher Simple preparation by burning is will remove sand and clay. If this is et al., 1996; Hall and Pilcher, 2002), not suitable when chemical analy- not adequate, then a heavy liquid Scotland (Dugmore et al., 1995), sis is required, as burning changes separation has proved successful Germany (van den Bogaard and the glass chemistry. In this case, (Turney, 1999). Schminke, 2002) and other areas the organic matter is dissolved distant from Iceland, this technique in a chemical oxidizing mixture Tephra in Arctic Norway was selected for dating in Lofoten. as described by Dugmore et al. Last year, a team from the Univer- As there had been no tephra work (1995) The tephra recovered from sity of Massachusetts, Amherst and done in this area before, we select- the acid treatment is washed and mounted for electron microprobe analysis. �� Volcanic systems vary in their chemistry. The products of individ- �� ual volcanoes may also differ and �� in some cases, such as in Iceland, � �� many individual eruptions are �

� ��

� geochemically distinct. Because � �� the differences between erup-

� ��

� �� tions are quite subtle, wavelength �

� ��

� dispersive microprobe analysis is �� used rather than the simpler and more commonly available energy �� dispersive analysis. It is usual to analyze between 10 and 20 indi- � vidual tephra fragments. � �� Type material comes from pro- �� files examined in Iceland. Many Figure 2: Stratigraphic sequence of tephra analysis from a peat bog at Bøstad in the Lofoten have been analyzed (e.g. Larsen Islands. Marked layers have been identified by matching microprobe-determined major element and Thorarinsson, 1977; Haflidason chemistry of the tephra with that of historically dated layers from Iceland and elsewhere. Not et al., 2000), many more remain to all layers have yet been determined. be examined. Analyses are stored Queen’s University, Belfast carried ed several peat profiles to develop in the Tephrabase Data Bank and out exploratory fieldwork in the a local tephra stratigraphy before are available to all researchers at Lofoten Islands off the northwest applying this to more complex this URL: http://www.geo.ed.ac.uk/ coast of Norway, just inside the lake sediments. Two peat profiles tephra/tbasehom.html Arctic Circle. We hypothesize that have been analyzed and show one Not all tephras are chemically the Lofoten islands are sensitive of the best historical tephra se- distinct. For example, the AD 1947 to changes in ocean currents (fluc- quences outside of Iceland itself. and AD 1510 tephras from Hekla tuations of the North Atlantic Drift) A good range of historical tephras are very similar, as are those of and to circulation shifts associated has been identified, including Hekla 4 (2310 BC) and Hekla 5 with different modes of the North among others, tephras from (AD) (5990 BC). Normally other strati- Atlantic circulation. 1875 (Askja), 1510 (Hekla), 1362 graphic information allows us to At present, the outer islands are (Oraefajokull), 1158 (Hekla) and discriminate between these pairs bathed in waters that reach +11°C 1104 (Hekla) eruptions (Fig. 2) and of tephras. in summer (+5-6°C in winter) in also the big eruption of Hekla in spite of their high latitude. Chang- 2310 BC. The AD 1158 and 2310 BC Tephra for Correlation Between es in the strength and/or position tephras have already been used Cores, Sample Sites and of this warm water tongue will to provide chronological control Regions drastically affect the Lofoten en- in our pilot cores from lacustrine A typical eruption lasts for days vironment. The islands have many and brackish water sediments. or perhaps even months, a very deep fresh water lakes, fjords and These layers can be used to cor- short time in relation to the accu- peat bogs with Holocene records. relate core to core and lake basin mulation of typical terrestrial and

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 20 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 21

marine stratigraphic sequences. fossils may have been moved geochemically typed tephra layers Thus, an ash layer forms a tightly vertically in the sediment profile. it contains. defined time-marker. For example, if the tephra were One of the perennial problems spread over 5 cm, there would be ACKNOWLEDGEMENTS of working in bog and lake depos- little point in undertaking a pollen The Lofoten Islands project is supported by its is correlation from one core to study at a resolution of 1 cm. a National Science Foundation grant (Paleo- climatology Program) to the University of another. Analyzing tephra can be Massachusetts. an ideal technique in these circum- Tephra as a Chronology stances, since the horizontal layers Tephrochronology consists of REFERENCES are so sharply defined. knowing when eruptions took van den Bogaard, P. and Schminke, H.-U., 2002: place and then of tying a tephra Linking the North Atlantic to central Europe: a high resolution Holocene tephrachronological record Tephra as a Taphonomic Guide fall to a particular eruption by its from northern Germany. Journal of Quaternary Tephra has so far been under-uti- chemistry. From about AD 1100, a Science 17: 3-20. lized as a taphonomic guide. As the well-recorded chronology of calen- Haflidason, H., Eiriksson, J. and van Kreveld, S., tephra fall is assumed to occur at drical accuracy exists for Iceland. 2000: The tephrochronology of Iceland and the North Atlantic region during the Middle and Late least within a single year, we can Before settlement in Iceland, we Quaternary: a review. Journal of Quaternary Sci- make a good assessment of post are dependent on dating tephra ence 15: 3-22. depositional processes by looking layers by independent means. Hall, V.A. and Pilcher, J.R., 2002: Late quaternary Icelandic tephras in Ireland and Great Britain: at the vertical spread of the tephra Radiocarbon dating of a terrestrial detection, characterization and usefulness. The in a sediment core. Spreading can peat allows a precise date to be Holocene 12: 223-230. also be the result of delayed depo- applied to tephra-bearing low-or- Larsen, G. and Thorarinsson, S., 1977: H4 and other acid tephra layers. Jokull 27: 29-47. sition of particles stored for more ganic lake sediments or marine to Pilcher, J.R., Hall, V.A. and McCormac, F.G., 1996: An or less time in the watershed or brackish water sediment that is un- outline tephrochronology for the north of Ireland. the lake basin. As the particle size suitable for precision radiocarbon Journal of Quaternary Science 11: 485-494. of the typical tephra layer is of the dating. One of the ongoing proj- For full references please consult: same order as that of pollen, for ects is to use the calendrical preci- www.pages-igbp.org/products/newsletters/ref2003_2.html example, it can suggest the extent sion of the new NGRIP Greenland to which pollen and other micro- ice core to produce dates for the

Simulating the Climate of the Last Millennium

MARTIN WIDMANN1 AND SIMON F.B. TETT2 1Institute for Coastal Research, GKSS Research Centre, Geesthacht, Germany; [email protected] 2Hadley Centre for Climate Prediction and Research, Met Office, Bracknell, UK; [email protected]

Introduction tween internal variability and the under development. In between Climatic changes during the last effects of varying external forc- EBMs and GCMs are a variety of thousand years have received ings, and understanding feedback models with varying numbers of great interest. Uncertainty in cli- mechanisms. dimensions and complexity, for mate reconstruction is smaller instance, earth system models of than in earlier periods and the Types of Climate Models and intermediate complexity (EMICs), external forcing factors of the Simulations which describe the atmosphere climate system are relatively well The simplest climate models are and ocean dynamics in less detail known, making comparisons with zero- or one-dimensional energy but which place more emphasis on 20th century climate well suited to balance models (EBMs), which the role of vegetation and chemical assessing the influence of human have low computational costs processes in the climate system. activities. Numerical modelling and clear links between simulated Climate models are mainly used in complements the efforts to recon- processes and climate. At the other two different ways. In equilibrium struct past climates from proxy end of the spectrum are quasi-re- experiments, the forcing factors records. The goals of paleoclimate alistic, computationally expensive, for the climate system, such as modelling are to reduce uncer- general circulation models (GCMs), solar irradiance, the atmospheric tainties in climate reconstructions which usually feature sub-models composition, or the earth’s orbit, through consistency tests with evi- for the atmosphere, the ocean are held constant but may vary dence from proxy data, to validate and the sea ice on 3-dimensional between different runs. These climate models, to provide hypoth- grids with typical horizontal spac- simulations represent the mean eses on the climatic evolution at ings of a few hundred kilometers climate and the statistics of inter- locations or variables not covered and 15-100 vertical levels. GCM nal climate variability. Transient, by proxy data, and to improve components for the carbon cycle, forced simulations also include the process understanding. The latter chemical processes, land ice and climate response to time-varying should include distinguishing be- vegetation dynamics are currently forcings using historical estimates.

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 22 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 23

�� identical to the signal of changing greenhouse gas concentrations. �� The pronounced insolation decrease during the Dalton Mini-

� mum (DM) around 1820 caused �

� �� global cooling. Some of the recent �

� work focuses on the Late Maun- � �

� der Minimum (LMM). A transient �

� ��

� simulation with an EBM forced by �

� solar and volcanic activity, anthro- � � pogenic greenhouse gases and � �

� �� aerosols produced a global tem- �

� perature well correlated with proxy �

� reconstructions (Crowley, 2000); �

� �� in particular a cooling during the LMM was found. A LMM cooling was also found in a 1000-year-long �� run with a two-dimensional zon- �� ally averaged atmosphere-ocean �� model (Bertrand et al., 2002). It Figure 1: Time series of global-mean temperature. Shown are time series of global-mean tem- should be noted, however, that perature from ECHO-G (solid black line) and HadCM3 (dashed red line) simulations. Note that the climate sensitivity to changes ECHO-G was forced with natural and anthropogenic simulations while HadCM3 was driven in the forcing can be tuned in both only by natural forcings. Both simulations are shown as differences relative to the 1550-1800 average and represent a 25-year running average. models. A 1000-year-long run with an EMIC (Bauer et al., 2003) used Since the evolution of the climate was only moderate and deviations similar forcings but also informa- system is not completely deter- from observations became evident tion on deforestation. Northern mined by external forcings but also (Barnett, 1999). A very long simula- Hemisphere temperatures corre- contains a stochastic component, tion of 15000 years using the GFDL lated well with proxy reconstruc- even a perfect model with all forc- model with a relatively low reso- tions. During the LMM and the DM, ing mechanisms included will yield lution suggested that large-scale, pronounced cooling took place due only one out of many possible cli- multi-decadal temperature anom- to solar and volcanic forcing. Dur- mate realizations consistent with alies with very strong amplitudes ing the last half of the 19th century, the forcings. This realization will of about 6-10 standard deviations deforestation led to cooling. be different from the realization could be generated merely by in- Shindell et al. (2001) used equi- that took place in the real world. ternal processes (Hall and Stouffer, librium runs of an atmospheric Therefore any comparison be- 2001). Recently, 1000-year-long GCM with detailed ozone chemis- tween model and reconstruction equilibrium runs with higher try coupled to a slab ocean model will be probabilistic in nature. resolutions (about 300 km) were to model the difference between conducted at UKMO (Collins et al., the LMM and the period around Results 2001) and MPI. Because they are 1780. During the LMM, the surface By the mid-1990s, coupled GCM much longer than the instrumental air temperature over the continents equilibrium experiments for pre- climate record, they are well suited was colder and over some ocean industrial conditions were typically to test paleoclimatic reconstruction areas warmer than 100 years later. around 100 years long but the first methods on decadal to centennial The temperature signal could be runs with lengths of about 1000 time-scales (Zorita and González- shown to be related to a change years became available at some Rouco 2002, Zorita et al. 2003). in the AO (Arctic Oscillation)/NAO modelling centers (GFDL, MPI Many transient simulations (North Atlantic Oscillation), which and UKMO). These simulations begin in the middle of the 19th in turn is driven by variations in were used, for instance, to clarify century to investigate the climatic the meridional temperature gra- the roles of the atmosphere and effects of anthropogenic green- dient. This temperature pattern the ocean in generating internal house gases and aerosols (e.g. appears partly consistent with climate variability (Manabe and Stott et al., 2000). The climatic re- evidence from proxy data but too Stouffer, 1996), or to estimate sponse to changing solar forcing many large areas without adequate natural variability, the basis for was investigated with a coupled proxy data remain to draw definite detection and attribution of cli- GCM forced by estimates for solar conclusions on the agreement mate change (Hegerl et al., 1996). variability from 1700 to the present between simulation and observa- However, the agreement among (Cubasch et al., 1997). The spatial tions. models on the spatial structure and pattern associated with solar forc- The LMM was also included magnitude of natural variability ing was found to be similar but not in 500-year-long transient runs of

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 22 Science Highlights: HOLIVAR Science Highlights: HOLIVAR 23

�� liminary results from two GCM simulations of the last 500 years a) �� demonstrate that it is now just � �� about feasible to do this using �� state-of-the-art GCMs. This then � �� provides a good opportunity to �� test both climate models and �� proxy data by comparison with � �� one another and thus to better � �� forecast anthropogenic climate � � change during this century. It also �� allows estimates to be made of � �� the relative contribution of natural � � � and anthropogenic climate forc- � �� ings, and internal climate variabil- �� ity to total climate variability and change. A current European Union b) �� funded project called SOAP (http: �� \\www.cru.uea.ac.uk\cru\projects\ � �� soap) aims to do this and bring � �� together scientists working with �� paleodata with those working �� with models. � �� ACKNOWLEDGEMENTS �� SFBT was supported by the UK Govern- � ment Met Research contract and, partly, by � �� EU contract EVK2-CT-2002-00160 (SOAP), �� while computer time for the HadCM3 �� simulation was provided by the UK Dept. �� of Environment, Food and Rural Affairs PECD/37. The transient ECHO-G simulation has been part of the KIHZ project of the �� Helmholtz-Society. We are grateful to F.J. �� González-Rouco and E. Zorita, GKSS for providing the ECHO-G data. Figure 2: LMM cooling (1675-1710). Shown are changes in near-surface temperature from HadCM3 (a) and ECHO-G (b) for the 1675-1710 period minus the long-term average for REFERENCES 1550-1800. Contours are at ±3, 2, 1, 0.5, 0.25 and 0 K. Global-average values are −0.21 and Collins, M., Tett, S.F.B. and Cooper, C., 2001: The −0.28K. internal climate variability of HadCM3, a version of the Hadley Centre coupled model without flux fully coupled atmosphere-ocean more important.The cooling dur- adjustments. Climate Dynamics 17: 61-81. GCMs conducted by the GKSS Re- ing the LMM has a spatial structure Crowley, T.J., 2000: Causes of climate change over the past 1000 years. Science 289: 270-277. search Centre in collaboration with somewhat different from Shindell Hall, A. and Stouffer, R.J., 2001: An Extreme Climate the MPI (ECHO-G, Fischer-Bruns et et al. (2001). Most noticeable is a Event in a Coupled Ocean-Atmosphere Simulation al. 2002, Zorita et al. 2003) and strong cooling in the North-West without External Forcing. Nature 409: 171-174. by UKMO (HadCM3). Preliminary Atlantic (Fig. 2), associated with Shindell, D.T., Schmidt, G.A. Mann, M.E. Rind, D. and Waple, A., 2001: Solar forcing of regional climate results from both models show a increasing sea-ice extent in both change during the Maunder Minimum. Science clear global-mean cooling during simulations. Also apparent is a 294: 2149-2152. the LMM and the DM (Fig. 1), which smaller cooling over Europe and Stott, P.A., Tett, S.F.B., Jones, G.S., Allen, M.R., Mitchell, J.F.B. and Jenkins, G.J., 2000: External is of similar magnitude in both other regions. In ECHO-G there is control of 20th century temperature variations by models. The ECHO-G simulation widespread cooling over the entire natural and anthropogenic forcings. Science 15: was forced with the major anthro- northern hemisphere with peak 2133-2137. Zorita E., von Storch, H., González-Rouco, F.J., pogenic and natural forcings while cooling to the west of Greenland Cubasch, U., Luterbacher, J., Legutke, S., Fischer- the HadCM3 was only forced with of −2K. By contrast, HadCM3, while Bruns, I. and Schlese, U., 2003: Simulation of the natural forcings. Both groups are still having a large amount of cool- climate of the last five centuries. GKSS report 2003/12. (available from GKSS Research Centre, currently carrying out simulations ing to the west of Greenland man- Library, D-21494 Geesthacht, Germany) with natural forcings only, and ages to sustain a larger land-sea both natural and anthropogenic contrast than that of ECHO-G. For full references please consult: forcings, respectively. Note the in- www.pages-igbp.org/products/newsletters/ref2003_2.html creasing divergence between the Summary two simulations from the mid-19th In this paper we have introduced century as the simulated effect of the idea of simulating the climate anthropogenic forcings becomes of the last 500 to 1000 years. Pre-

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 24 Science Highlights Science Highlights 25

Late Holocene Paleoclimate and Paleogeography in the Tien Shan- Balkhash Region

B. J. AUBEKEROV, R. SALA, S. A. NIGMATOVA Laboratory of Geo-Archaeology, Institute of Geology, Academy of Sciences, Almaty, Kazakhstan; [email protected]

The Tien Shan-Balkhash region Late Holocene Climate and (Semirechie) (Fig. 1) is a rich natural Human Adaptation archive of paleoclimatic data but Quantitative evaluation and absolute is poorly known. Research in this dating of paleoclimatic changes for region, performed during Soviet the late Holocene have only been at- times and resumed in recent years, tempted in the last 5 years in the con- shows that changes in atmospheric text of geo-archeological research for circulation, paleoclimate, and pa- project INTAS 97-2220. Temperature leo-geographical conditions, includ- and precipitation reconstructions ing a chronology of basic events are based on palinological data and in the arid continental regions of transfer functions located in three Central Asia, can be successfully different zones: The semideserts of reconstructed. Research included Southern Pre-Balkhash, the foothills multiproxy reconstruction from geo- Figure 1: Satellite image STS059-223-042, of the Tien Shan and Jungarian logical-geomorphological, hydrologi- April 1994. Issyk-kul Lake, Northern Tien Shan ranges, and the high mountains. cal, paleontological (paleobotanical, and, 400 km north, the Balkhash Lake partly Sites with average trench depths of iced. In the upper right, the atmospheric cir- palinological), pedological (micro- culation patterns of the Siberian anticyclone 250 cm provided a 3000-3500 year morphological) and archeological are visible. Black dots indicate sites for sequence of events. Samples were studies, as well as independent palinological and geo-archeological research: collected every 7-15 cm to allow 150- 1-Tamgaly (800 m asl) and 2-Serektas (750 m) dating methods (EPR – Electronic in semidesert landscape of the Chu-Ili range; year temporal resolution. Dating was Paramagnetic Resonance; 14C). 3-Talgar (800 m) in piedmont steppes and 4- based on EPR of carbonates and on Turghen (1800 m) in mountain meadows of correlation with archeological find- Northern Tien Shan; 5-Tasbas (1600 m) in Quaternary Climatic Events middle mountain steppes on the northern ings from settlements and tombs Data collected during Soviet times slopes of the Jungarian range. Red dots indi- submitted for 14C analyses. concerning the Quaternary period cate sites of geo-archaeological research: It has been established that lo- from left to right Kuljabasi, Oi-Jailau, Assy, as a whole, are based on qualitative Saty, Charyn, Usunbulak, and Koksu. cal climatic fluctuations are well analyses and relative dating, and correlated in tendency, time and pe- provide low temporal resolution. modern mode of atmospheric circu- riodicity with global estimates. But The following results have been lation was established. relevant regional anomalies, mainly reported: Throughout the Quaternary Pe- determined by the continental loca- − The formation of the high- riod, latitudinal changes in the po- tion of the region, are also observed mountain range of the Tien Shan sitioning of the Atlantic anticyclone (phase 3), emphasizing the forc- began during the Pleistocene, 1.6-1.3 and its branches exerted a crucial ining impact of seasonal anomalies. million years BP. The rise of plateaus fluence on the climate of Kazakhstan Also, sub-regional differences exist above 3000 m and the establishment and Semirechie. It characterized the between plains and mountains be- of the first and largest glaciation oc- switch between glacial and intergla- cause of their different exposure to curred 1.2-1.1 million yrs BP. cial periods and, during interglacial drought and glacial events and the − In the second half of the Pleis- and Holocene times, between pluvial different (sometimes opposite) reac- tocene (0.5 million yrs BP), the in- and arid phases. Changes in the po- tion of their environments to climatic crease in mountain height to more sition of the Atlantic anticyclone also variations. than 5000 m blocked the path of affected the water balance of interior In Semirechie, climatic forcing on the Indian monsoons. This caused reservoirs (Caspian, Aral, Balkhash, borders of landscape zones in plains, a reduction and then total halt of Alakol, Zaisan) and the volume of ice and of vertical belts in mountains is monsoonal influence on West-Cen- deposits in the Tien Shan Mountains. strong enough to influence human tral Asia and started the process of Thus, lakes and mountains provide use of large areas. In such cases, aridization that gave the region its complementary archives for paleo- one can analyze both the correlation modern features. climatic reconstruction. In Semire- between climatic and palinological − With the fading of the mon- chie during the Holocene period, events, as well as patterns of human soonal influence, polar air masses we reconstruct regular alternation land-use and environmental adapta- from the Arctic glacial zone (Siberian between arid (hot-dry) and pluvial tions such as settlement localization, anticyclone) and eastward branches (cold-wet) phases, correlated with switches between sedentary and no- (along 45º-50º latitude) of the Atlantic high (northerly) and low (southerly) madic life or between pastoralist and anticyclone (Azores High) began to positioning of the Atlantic anticy- farming activities, or development of play a major role. In this way, the clone. irrigation schemes, etc.

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 24 Science Highlights Science Highlights 25

ing of former accumulations of ice. The borders of vegetative cover in all landscape zones and vertical belts remained close to modern. Human communities (Late Bronze culture) practiced pastoralist activities with settled villages in mountain habitats (Assy 2500 m asl, Turghen 2300 m, Tasbas 1600 m, Oi-Jailau 1400) and scattered houses in oases of pied- monts and plains (Tamgaly 800 m).

Early Subatlantic; Early Iron Age (2800-2000 BP) The average climate was pluvial— cooler and wetter than modern, with larger seasonal amplitudes of temperature and humidity. Both plains and mountains had cold-wet conditions, with some differences in periodicity: In plains the peak of cold and humidity fell around 2400 BP; in the mountains around 2200 BP, a difference probably due to prolonged ice accumulation. Glaciers on the Northern Tien Shan grew back to more than 2 km in length. Balkhash water levels reached a minimum of 338 m around 2500 BP and then peaked at a maximum of 343 m in 2200 BP. The borders of vegetal zones in the plains and mountains changed in response to climatic fluctuations. Alpine zones became glacial and were abandoned (at least during a) b) the cold season). Semidesert plains were transformed into steppes, Figure 2: Reconstruction of average temperature and precipitation in Semirechie during the last consistently increasing the potential 3200 yrs (150 yrs temporal resolution; based on palinological analyses). a) In plains (Tamgaly, areas of human activity and favoring 800 m asl, semidesert landscape); b) In mountains of Semirechie (Turghen, 2300 m asl, alpine the conversion to nomadic pastoral- meadows). Temperature (T) is indicated by red lines, precipitation (P) by blue dashed lines. The absolute chronology is provided by EPR and 14C analyses (dots) and archaeological cor- ism with seasonal occupation, new relation (segments). herd composition, horse riding, de- Modern climatic conditions in Semi- and 34. The five main phases are as mographic expansion, and relative rechie including the arid continental follows: decline in stable housing mostly plains and humid mountains are located in the piedmont zone half- moderately hot and dry, with an Late Subboreal: Late-Final way between winter and summer average continentality index of Bronze Age (3200-2800 BP) pastures. A major economical and 30. Preliminary reconstruction of Plains had arid conditions, slightly cultural change occurred (Saka and paleoclimate and human land-use warmer and drier than modern; early Wusun cultures). in plain and mountain zones of mountains had humid conditions, Semirechie during the last 3200 slightly cooler and wetter than Short Interval 2000-1800 BP years distinguishes 5 main phases, modern. Glaciers on the Northern This period must be noted for its pe- described below (Fig. 2). Tien Shan contracted to less than 2 culiarity. There are ubiquitous signs Calibration with modern stan- km in length. Balkhash water level of increased continentality, with very dards suggests that during the continued to shrink (from 344 m severe cold-dry winters and cold-wet whole interval, average tempera- asl at 3800 BP) to less than 342 m summers. The phenomenon is most ture range fluctuated between +1ºC/ asl. From this time on it fluctuated pronounced in plains where ex- -2ºC; precipitation between ±20% in around the 342 m level, due to a ceptionally cold temperatures and plains and ±30% in mountains, and complex balancing of synchronic seasonal contrasts of precipitation continentality indexes between 27 hydrological changes and the melt- caused a botanical catastrophe and

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 26 Science Highlights Science Highlights 27

cryolitic formations. Data suggest activities together with the spread climatic conditions tended progres- the establishment of an exceptional of irrigation practices, proto-urban- sively towards the warmer and drier blockage of winter atmospheric cir- ization along river courses and the conditions of modern times. culation in Central Asia for almost blossoming of the northern branch two centuries, which must corre- of the Silk Road (late Wusun and REFERENCES spond to some climatic events of Turkic periods). Aubekerov B.J., 1993: Stratigraphy and paleogeography of the plain zones of Kazakhstan during the global dimension. Steppes were late pleistocene and holocene. In: Development of reduced and human habits become Late Subatlantic; Middle and landscape and climate of Northern Asia, late pleis- more sedentary, with primitive ag- Late Middle Ages (800-200 BP) tocene and holocene. Vol 1, pp101-110. Nauka, Moscow (in Russian). ricultural practices in the piedmont Climate became cooler and wetter Aubekerov B.J., Sala R., Nigmatova S. and Shakupova fans (Wusun culture). than either the former period or the S., 2001: Paleoclimatic conditions in the arid zones modern climate. Moderate advances of Semirechie during the Bronze and Early Iron epochs. In: Natural and social problems of geog- Middle Subatlantic; Early and of ice occurred. The plains reached raphy of arid regions, pp 26-33. KAZNY, Almaty (in Mid Middle Ages (1800-800 BP) a minimum temperature around Russian). Plains and mountains were charac- 550 BP, the mountains around 250 Kremenezki K.B. and Tarasov P.E., 1994: Paleogeo- graphic study of Kazakhstan during the holocene: terized by cold-wet conditions until BP. Balkhash water levels continued palinology and stratigraphy. Nauka, pp151-159. 1000 BP when conditions close to to rise until a peak of 344 m was Moscow (in Russian). modern were established. Conti- reached in 1910 AD. A conversion to Pakomov M.M. and Chupina L.H., 1973: The general nentality indexes remained high with pastoralist practices and nomadic formation laws of the modern spore-pollen spectra in the eastern mountains of Central Asia. In: Trudi, cold-dry winters. The Northern Tien habits took place. Most likely this Mesdunarodnie Palinologiskaya Conferentsia. Shan saw a catastrophic degrada- was favored by the reestablishment Moscow (in Russian). tion of ice deposits from which they of climatic conditions similar to those Tarasov P.E. et. Al., 1992: Climatic fluctuation in the steppe zone of Kazakhstan during holocene, based never recovered. Balkhash water lev- of the Early Iron Period but also by on data from spore-pollen analyses. Vestnik MGU els decreased to a minimum of 339 political military events and the crisis series 5, Geography n 1-94, pp 99-103. Moscow m around 800 BP. Human geography in the Silk Road trade (Mongol and (in Russian). saw the expansion of agricultural early Kazakh periods). After 200 BP,

Holocene Ocean-Climate Variations in the Gulf of California, Mexico

ROBERT DOUGLAS1, OSCAR GONZALEZ-YAJIMOVICH1,2, DONN GORSLINE1, FRANCISCA STAINES-URIAS1,2AND J. FERNANDO ARREOLA-HERNANDEZ 1,3 1Department of Earth Sciences, University of Southern California, Los Angeles, California 90089-0740, USA; [email protected]; 2 Facultad de Ciencias Marinas, Universidad Autonoma de Baja California, Ensenada, Mexico; 3 Centro Interdisciplinario de Ciencias Marinas-Instituto Politecnico Nacional, La Paz, Mexico.

Despite complexities, the Gulf of �� California can be considered an �� ocean-climate system in which �� primary productivity and sedimen- �� tation are largely controlled by sea- �

sonal monsoon winds. It offers an � ��

ideal model for exploring the ocean- � � climate variations in which decadal �� to millennial scale changes can be reconstructed from sediment cores �� that sample the latest Pleistocene and Holocene. � During winter months, strong � �� NW winds flow down the length of �� the Gulf, cooling surface waters, and Figure 1: Spectra of peaks of variability found in laminated sediments and carbonate content initiating surface mixing and Ekman in core BAP96-CP, Alfonso Basin (MTM method). Major periodicities are given in Table 1. transport along the eastern margin of biogenic debris contribute to the and so-called “shade floras”, mat- and the development of a major clayey mud that predominates in forming and large centric diatoms winter-spring bloom of diatoms. slope and basin sediments. which grow in the nutricline. The Primary productivity rates in the During summer months, winds mats are dumped to the seafloor central Gulf can exceed 4 gC/m2/day are from the SE, a strong ther- in late autumn or early winter with during the winter season, making mocline develops, which reduces the onset of the NW winds. In the the Gulf one of the more productive nutrient transfer and there is a central and eastern Gulf, these mat- sites in the global ocean. The winds decline in primary productivity. The forming diatoms contribute to the also transport large quantities of biogenic flux to the seafloor is formation of sediment laminations. dust, which together with the flux dominated by pelagic carbonate The SE winds bring summer storms

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 26 Science Highlights Science Highlights 27

and, except for El Niño years, most Table 1: Spectral peaks of variability derived from separate analyses of the laminated sediments and carbonate record in core BAP96-CP in Alfonso Basin (MTM method). Ages estimated using of the precipitation in the Gulf falls average sedimentation rates derived from AMS 14C age-depth model. between July and October. The deep-water sediment record in the Gulf reflects the monsoon control of biogenic and terrigenous input, especially in the hemipelagic sediments of the slopes. In the Gulf, the Oxygen Minimum Zone (OMZ) typically ranges between about 300 and 1000 m and where it intersects the seafloor, bottom-water dissolved oxygen concentrations are colleagues of the U.S. Geological In Guaymas Basin, an increase in <0.5 ml/l to zero. Laminated varves Survey (Menlo Park, USA) have the production of biogenic silica and accumulate in these depositional focused on the diatom-rich, carbon- changes in the relative abundance environments. Preserved within the ate-poor sediments in cores DSDP of diatom floras suggest shifts in laminations is a record of the mon- 480 and JPC56 in Guaymas Basin the seasonal intensity of upwelling, soon-induced seasonal variations (central area of the Gulf), while our with spring upwelling dominating in production and sedimentation as group (listed above) has focused on in the early and late Holocene and well as longer-term ocean-climate the sediments microfabrics, stable winter upwelling dominating in the variations. isotopes and carbonate microfos- mid Holocene and today (Barron, et Holocene sediments in the Gulf sils in the more carbonate-rich al., 2000; Barron, 2003). After about of California are biogenic silica- and sediments of the western Gulf and 6 ka, spring upwelling increased on organic carbon-rich, generally car- Alfonso Basin (mouth of the Gulf). the mainland side of the Gulf. Co- bonate poor and within the OMZ, These investigations have produced incident with the shift in upwelling, intensely laminated. In contrast, complementary results and suggest the carbonate content of Guaymas late Pleistocene and Younger Dryas that the shifts in the Holocene Gulf Basin shifted, with more intense dis- sediments are carbonate-rich, are ultimately the result of an inten- solution in sediments in the eastern biogenic silica-poor and typically sification and seasonal shift in the Gulf leading to very low carbonate bioturbated. NW monsoon winds that control content, generally beneath the main Recent multi-institutional in- the patterns of upwelling, and that upwelling centers. vestigations have emphasized the changes are paced by climate Alfonso Basin is situated near the different approaches to producing rhythms that have strong modes of junction of the Gulf of California and high-resolution (10-30 years/sample) about 200 yr, 300 yr and 1500 yr. the open Pacific Ocean, which places Holocene records. John Barron and it in a sensitive location to record the variation in the monsoon-driven � �� climate in the Gulf, as well as the �� larger-scale climate variations occur- �� ring in the Pacific. The basin is silled �� (depth ca. 250 m), filled with dysox- � �� ic-anoxic waters and accumulating � � ��

� laminated sediments on the basin � � � �

� floor. Cores collected from the ba- �

��

� sin floor have recovered most of the � � �

� Holocene (7700 yBP) based on AMS � � �� radiocarbon dates and varve counts � � � produced from digitally scanned X-

�� radiograph positives. All cores were �� sampled at 1 cm intervals (approx. 15-35 years/sample). �� Spectral analysis of the lami- �� nated stratigraphy and carbonate in �� core BAP96-CP was used to identify �� millennial and centennial peaks of �� climate variability (Douglas, et al., �� 2002) (Fig. 1). For the microfab- Figure 2: Stable isotope record of Bolivina subadvena in a core collected near the center of rics analysis, gray scale images Alfonso Basin (core BAP96-CP; 1 cm sample interval). Age-depth model based on excess 210Pb were produced from X-radiograph profiles and corrected radiocarbon dates. Based on recently collected multi-cores from the prints, digitized and converted to same site, we estimate that about 200 years is missing from the top of core CP. pixel number and analyzed using

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 28 Science Highlights Science Highlights 29

the multi-tapered Method (MTM). in tree rings and 10Be records. Mil- of Central Gulf Water flooded the No filters were applied. This pro- lennial scale variations at 900 yrs basin (Douglas and Staines-Urias, duced a continuous, millimeter-scale (898±82 yrs) and 1500 yrs (1480±100 in press). If this interpretation is cor- record. The resulting power spectra yrs) are strong in both the carbonate rect, the isotopic excursions record frequency peaks were converted to and sediment lamination records. periods of intensified NW winds in calendar years based on the age- Stable isotopes measured on the the mid and late Holocene. depth model. Because of variations benthic foraminifera Bolivina subad- in the microfabric record (density of vena in Alfonso Basin (core BAP96- REFERENCES laminations, microburrows, flood CP) exhibit small variation through Barron, J., Bukry, D. and Bischoff, J., 2003: A 2000 yr long Record of Climate From the Gulf of California. layers), spectral peaks were cal- the 7700 yBP record, except in the Proceedings of the 19th Pacific Climate Workshop, culated separately for three cores 18O record. Beginning about 6ka, the PACLIM 2002, Technical Report 69 of the Inter- and for selected intervals within the record is punctuated with a series agency Ecological Program for the San Francisco Estuary, 15 pp. cores. The results were the same for of short-term excursions or events Barron, J., Bukry, D. and Bischoff, J., 2003: High-reso- all cores. of 0.5 to 0.9‰, the largest of which lution Paleoceanography of the Guaymas Basin, Strong peaks of variability were occurs between 5200 and 1400 yBP Gulf of California, during the past 15,000 years. Paleoceanography, in press. found with periodicities at about 150 (Fig. 2). The events signal a turnover Bernal Franco, G. R., 2001: Registro Paleoceano- yrs (144±18 yrs), 200 yrs (192±20 in the water-mass in the basin and grafico en los sedimentos laminados de la cuenca yrs) and 300 yrs (300±20 yrs) replacement by either much colder de La Paz, margen occidental del bajo Golfo de California. tesis Doctoral, Centro de Investigacion (Table 1). These are associated with or higher salinity water. The basin sill Cientifica y Educacion Superior de Ensenada productivity/dissolution cycles in the is located at ca. 250 m depth, near (CICESE), Mexico, 110 pp. carbonate record. All these period- the present-day lower boundary of Douglas, R., Gorsline, D., Grippo, A, Granados, I. icities are present prior to 3200 yBP Central Gulf Water—a high salinity and Gonzalez-Yajimovich, O., 2002: Holocene ocean-climate variations in Alfonso Basin, Gulf of but the 300-year cycle dominates in water-mass produced by evapora- California, Mexico. In, West, G. J. and Buffaloe, the latest Holocene. Similar production in the northern Gulf of Califor- L. (eds.), Proceedings of the18th Pacific climate tivity pulses have been identified by nia during winter months—and is Workshop, PACLIM 2001, Technical Report 68 of the Interagency Ecological Program for San Gladys Bernal-Franco (2001) in the advected at shallow depths along Francisco Estuary, pp. 7-20. biogenic record in La Paz Basin on the western margin depth of the Douglas, R. and Staines-Urias, F., Stable isotopic a 330-year period. We suspect that Gulf towards the mouth. We believe evidence for variation in the production of Central Gulf Water-Mass and turnover in Alfonso Basin the centennial variability is related the most likely explanation for the during the Holocene, Ciencias Marinas,in press. to variations in solar radiation iden- isotopic excursions is basin turn- tified in radiocarbon fluctuations over, when increased production

Climate and the Maya Gerald H. Haug1, Detlef Günther2, Larry C. Peterson3, Daniel M. Sigman4, Konrad A. Hughen5, Beat Aeschlimann2 1Geoforschungszentrum Potsdam, D-14473 Potsdam, Germany; [email protected] 2Department of Chemistry, ETH, CH-8092 Zürich, Switzerland 3Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL 33149, USA 4 Department of Geosciences, Princeton University, Princeton, NJ 08544, USA 5 Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA The relative stability of Holocene derived inputs to Cariaco Basin with strength, driven by seasonal shifts climate, at least when compared to roughly bimonthly resolution (50 µm in the position of the Intertropical the rapid, large-amplitude climate sample spacing) for the period AD Convergence Zone (ITCZ; Fig. 1) and excursions of the last ice age, has 700 to 950 (Haug et al., 2003). Ter- its associated convection (Peterson contributed to a view that the road to minal Collapse of the Classic Maya et al., 2000; Haug et al., 2001). Light- modern human civilization has been civilization in the lowlands of the colored laminae consist mostly of little affected by climate road-bumps. Yucatan Peninsula in Mesoamerica biogenic components deposited dur- However, significant variations in re- occurred during this time, one of ing the dry winter-spring upwelling gional Holocene climate have now the most dramatic events in human season when the ITCZ is located at been recognized, with some having history (Fig. 1). Our new record of its southernmost position and trade had clear societal impacts. Here we riverine input to Ocean Drilling winds along the Venezuelan coast summarize recently reported new Program (ODP) Site 1002 shows in are strong. In contrast, dark laminae data from the annually laminated unprecedented detail evidence of a are deposited during the summer-fall sediment record of the anoxic Ca- link between regional drought and rainy season when the ITCZ migrates riaco Basin off northern Venezuela, the demise of the Maya culture. to its most northerly position, nearly a now well-known paleoclimate ar- Paired light-dark laminations above Cariaco Basin. Individual chive. Using a new method for the preserved in the anoxic, non-biotur- dark laminae are rich in terrigenous measurement of bulk sediment bated sediments of Cariaco Basin grains and contain higher quantities chemistry, we have been able to are the direct result of large regional of titanium (Ti) and other lithophilic develop a record of varying river- changes in rainfall and trade wind elements. Our interpretation of bulk

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 28 Science Highlights Science Highlights 29

natural groundwater resources a) were restricted over much of the lowlands. During the Pre-Classic Period prior to about AD 150, Maya culture flourished and the first major cities were built. Between ~AD 150 and 250, the first documented historical crisis hit the lowlands and led to the Pre-Classic Abandonment of major cities (Fig. 2). However, populations recovered, cities were reoccupied and Maya culture blossomed in the b) 94 90 86 following centuries. Around AD 750, lassic Maya Collapse: C at the peak of this Classic Period, the three phases between AD 760 to 910 best estimates for the population of Chichen Itza the Maya lowlands range from 3 to Uxmal 20 13 million inhabitants. Phase III: Between about AD 750 and 950, AD 861-910 the Maya experienced a demo- Edzna graphic disaster as profound as any in human history. During this 18 Calakmul Terminal Classic Collapse, many of Palenque the densely populated urban cen- Phase I: Tikal ters were permanently abandoned AD 760-810 and Classic Maya civilization came Aguateca Caracol to an end. What caused this to hap- 16 pen? While the Cariaco Basin record cannot provide a complete expla- Phase II: nation, it supports the view that Copan AD 811-860 changes in rainfall patterns played 14 a critical role. Using a Micro-XRF technique not previously applied to sediments, we Figure 1: (a) The Pyramid of Chichen Itza in the Yucatan penisnsula lowlands. (b) Three measured the Ti content (at 50µm phases of Terminal Classic Collapse (~AD 760 to 910). Phase I: Initial abandonment of the western lowlands where rainfall was the primary source of water. Phase II: Abandonment spacing) of a slab sample from of the southeastern lowlands where freshwater lagoons provided at least some surface ODP Hole 1002D taken from the water. Phase III: Large-scale abandonment of remaining cities in the central lowlands and stratigraphic interval known to en- the north. (Modified from Gill, 2000). compass the period of the Terminal Ti content as an index of regional the climate is dry. Hence, the center Collapse. Further methodological hydrologic change, reflecting varia- of Maya civilization was located in and chronological details can be tions in the mean ITCZ position with the same climatic regime as Ca- found in Haug et al. (2003). Within time, is supported by comparison of riaco Basin, with both areas near the Hole 1002D slab sample, we ob- the Holocene Cariaco record with in- the northern limit of seasonal ITCZ served distinct Ti minima at depths dependent paleoclimatic data from movement. of ~12 mm, 38 mm, 58 mm and 78 the region. In order to inhabit the Yucatan mm (Fig. 2) which we interpret as It has been suggested that re- lowlands and to deal with normal marking multi-year drought events current patterns of drought played seasonal variations in rainfall, the that began in about AD 910, 860, 810 an important role in the complex Maya developed elaborate strategies and 760, respectively. These are su- history of the Maya (Hodell et al., to accumulate and store water. Cities perimposed on a period of generally 2001; Gill, 2000). Maya civilization were designed to catch rainfall and lower Ti content. Not counting the developed in a seasonal desert and quarries were converted into water duration of the more severe multi- was highly dependent on a con- reservoirs. The Maya also built on year droughts, the number of varves sistent cycle of rainfall to support topographic highs in order to use between drought events indicates a agricultural production. Most of the the hydraulic gradient to distribute spacing of ~40 to 47 years (± 5), a rain available to the Maya fell during water from canals into complex number that agrees remarkably well the summer, when the ITCZ sits at irrigation systems. Nevertheless, with the observation of subpeaks at its northernmost position over the despite its sophistication, the hu- about 50-year intervals in the well- Yucatan. During the winter, the ITCZ man-engineered system ultimately known Lake Chichancanab sediment is located south of the lowlands and depended on seasonal rainfall since density record (Hodell et al., 2001).

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 30 Science Highlights Workshop Reports: 2nd Holivar Workshop 31

a) role in the collapse of Classic Maya Phase III Phase II Phase I civilization, and provides a temporal

) template for comparison of archeo- % ( logical data. m u i n

a Drought conditions may also t i T have been responsible for the earlier Pre-Classic abandonment of Maya cities that occurred between ~AD 150

Slab sample (mm) and 250, intervals similarly marked b) by low sediment Ti. These periods of drought are all most probably the

) result of climatic conditions that pre- % ( vented the ITCZ and its associated m u i n

a rainfall from penetrating as far north t i T as usual. A southward displacement of the ITCZ, as indicated by low Ti in Cariaco sediments, would be Slab sample (mm) expected to lead to similar rainfall

c) ”Medieval Warm Period” reductions in the Maya lowlands. ”Little Ice Age” We suggest that the rapid expan- )

% sion of Maya civilization from AD 550 ( m u

i to 750 during climatically favorable n a t i (i.e., relatively wet) times resulted T in a population operating at the limits of the environment’s carrying capacity, leaving Maya society Age (calendar ka BP) especially vulnerable to multi-year Figure 2: (c) Bulk Ti content (3-point running mean of 2-mm-resolution measurements; Haug droughts. Given the perspective et al., 2001) of sediments from ODP Hole 1002C, Cariaco Basin during the last 2000 years. Well-known climate events and major events in Maya civilization history are shown. (b) Bulk of our long, sediment time-series, Ti content (30-point running mean of 50-µm-resolution analyses) of a 30-cm-long slab sample it would appear that the droughts from companion ODP Hole 1002D in the time interval ~1.8 to 1.0 ka. (a) Bulk Ti content (3- we have highlighted were the most point running mean of 50-µm-resolution analyses) of the slab sample during Terminal Classic Collapse. Gill’s three phases of Maya Collapse are shown. severe to affect this region in the first millennium AD. The intervals of peak Mayanists generally agree that there quality of natural water sources in drought were brief, each lasting be- is strong evidence in the archeologi- the Yucatan lowlands, a factor that tween ~3 and 9 years. However, they cal record for regional variability in would clearly come into play during occurred during an extended period the Terminal Classic Collapse. Most periods of drought. While the north- of reduced overall precipitation that would also concur that the collapse ern lowlands have the lowest annual may have already pushed the Maya occurred first in the southern and average rainfall, collapsed cenotes system to the verge of collapse. central Yucatan lowlands, and that (water-filled limestone sinkholes) in many areas of the northern lowlands this region provide the most direct underwent similar decline a century access to groundwater in the Yu- REFERENCES Gill, R.B., 2000: The Great Maya Droughts: Water, or so later. A more controversial catan. In the central lowlands, some Life and Death. University of New Mexico Press, tripartite pattern of city abandon- freshwater is available in and around Albuquerque, NM, 1-464. ment (Fig. 1) was proposed by Gill the Petén Lake district. Towards the Haug, G.H., Günther, D., Peterson, L.C., Sigman, D.M., Hughen, K.A. and Aeschlimann, B., 2003: Climate (2000) based on analysis of the last west and south, access to groundwa- and the Collapse of Maya Civilization. Science dates carved into local monuments ter is scarce and rainfall was almost 299, 1731-1735. (stelae). On this basis, Gill argues certainly the primary source of wa- Haug, G.H., Hughen, K.A., Sigman, D.M., Peterson, for three separate phases of col- ter for Maya cities. During sustained L.C. and U. Röhl, 2001: Southward migration of the Intertropical Convergence Zone through the lapse that terminated respectively drought, access to groundwater Holocene. Science, 293: 1304-1308. in ~AD 810, 860, and 910. He further was likely an important factor in Hodell, D.A., Brenner, M., Curtis, J.H. and T. Guilder- speculates that periods of drought determining which large population son, 2001: Solar forcing of drought frequency in the Maya lowlands. Science, 292: 1367-1370. triggered the Maya demise. Within centers could survive. Peterson, L.C., Haug, G.H., Hughen, K.A. and U. Röhl, the limits of chronological uncer- No one archeological model is 2000: Rapid changes in the hydrological cycle of tainty, the proposed end dates for likely to completely capture a phe- the tropical Atlantic during the last glacial. Sci- ence, 290: 1947-1951. each phase of abandonment match nomenon as complex as the Maya the three most severe drought decline. Nevertheless, the Cariaco events inferred from the Cariaco Basin sediment record provides Basin record.Considerable vari- support for the hypothesis that re- ability exists in the distribution and gional drought played an important

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 30 Science Highlights Workshop Reports: 2nd Holivar Workshop 31

Climate Modeling

ESF/HOLIVAR MEETING. LOUVAIN-LA-NEUVE, BELGIUM, 12-15 JUNE 2002 How could insolation changes, a) �� 6 ka�� - recent� �� with their modest change in annual June,�� July,�� August�� mean forcing, cause the drying of �� the northern subtropical areas—including the full desertification of ��

the Sahara—and the global cool- �� ing experienced at high northern latitudes? This question motivated �� the pioneering general circulation �� model experiments by Kutzbach and Mitchell in the eighties. They � pointed out the importance of the �� seasonal contrast of insolation in ��

governing the ocean-land tempera- �� ture gradient. Yet, twenty years later, quantitative understanding of the 6 ka - recent climate of 6,000 years ago remains b) �� June,�� July,�� August�� a priority and also a prerequisite for �� confidence in future climate predic- tions. Modelers have understood the �� need for considering feedback from �� the biosphere, the ocean mixed layer and sea-ice, and hence the �� importance of including these components in state-of-the-art models. �� While general circulation models � have been used to better under- �� stand the mean climate features of �� the Holocene, many new high tem- �� poral resolution data have become available, bringing new challenges Figure 1: The figure shows the modelled changes in June-July-August mean precipitation for the changes between the mid-Holocene (6000 years ago) and the present day for an example to climate scientists and modelers. of (a) a PMIP I simulation in which the orbital parameters were changed but the sea surface Time series with decadal and annual temperature and vegetation cover were held constant at present day, and (b) a PMIP II style resolution have shown the need for simulation in which changes in vegetation and oceanic circulation were modelled. The mo soons are considerably enhanced when these effects are included. The robustness of these long transient simulations to reach a results, and the changes in other regions and interannual variability will be examined in the better understanding of the mecha- PMIP II project. The results shown are using version 3 of the Hadley Centre model and all nisms involved in climate variability simulations should be considered provisional. and abrupt climate change during Along with these recently published the third dimension, and may be an the Holocene. Some of the first of results, 9,000-year-long time-series example of the difficulties that will these were carried out with models computed with 3-D models were appear when performing similar (of intermediate complexity) largely presented for the first time. At the experiments with state-of-the-art based on a zonally averaged struc- cost of increased computing time, models. ture with a secondary treatment of the latter models better capture both Currently, state-of-the-art mod- the longitudinal resolution. Experi- atmosphere and ocean dynamics, els can only be run for specific time ments with CLIMBER-2 (Claussen including their natural variability. slices (although this is rapidly chang- et al., 1999) suggested that the de- This sometimes yielded unexpected ing and in the next few years we will sertification of Sahara was abrupt results. The ECBILT-CLIO model (UCL, start to see the first 9000-year simu- because decreasing summer inso- Louvain-la-Neuve), for example, ex- lations with low resolution versions lation crossed a threshold in Char- hibits extreme precipitation variabil- of state-of-the art general circula- ney’s positive monsoon-vegetation ity at the century time-scale in the tion models). Coupled atmosphere- feedback. In parallel, Crucifix et al. Sahara prior to 5,000 years ago. For ocean-vegetation models can be run studied the vegetation-temperature the model, insolation is then at such for simulations up to a millennium feedback operating at high latitudes, a level that the vegetation-monsoon in length, and the annual, decadal, but concluded that this feedback was positive feedback is strongly excited and century scale variability can be not strong enough to cause abrupt by the internal variability. Even if this examined. For instance, changes in climate change, at least during the result seems unrealistic, it points out ENSO variability during the mid- Holocene. the additional complexity brought by Holocene can be examined. Such

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 32 Workshop Reports: 2nd Holivar Workshop Workshop Reports: 2nd Holivar Workshop 33

studies are important to help us cycle during the Holocene was models. For this latter problem, it is understand the controls on climate also stressed. Until recently, most also essential that we build a new variability, and to determine whether attempts to understand the cause database of the natural variability of

the current models can successfully for the CO2 increase during the Ho- climate forcing (especially solar and reproduce the observed changes. locene were based on conceptual volcanic), as well as building and Coupled atmosphere-ocean-vegeta- models. With biogeochemical cycles maintaining databases of high qual- tion modelling and natural variability being implemented in models, a ity climate proxies for the whole of will be the major foci of the second more detailed model-data compari- the Holocene period. The increasing phase of the Palaeoclimate Model son will be possible. Brovkin et al. ability of the models and improve- Intercomparison Project (PMIPII). (2002) recently illustrated the power ments in high-resolution data will As well as focusing on the mid- of such tools by analyzing the conse- ensure that the next few years will Holocene, the new project will also quences of Holocene climate change be an exciting time. perform simulations for the early and change in ocean alkalinity on the Holocene (see EOS, 1 October 2002, terrestrial and oceanic carbon stocks REFERENCES and http://www-lsee.eea.fr.pmip during the Holocene. They also simu- Brovkin, V., Hofmann, M., Bendtsen, J. and At such time-scales, it is also lated atmospheric 13C concentration Ganopolski, A., 2002: Ocean biology could control δ atmospheric d13C during glacial-interglacial cycle. crucial to account for solar forcing to compare it with data and test their Geochemistry Geophysics Geosystems 3: art.no. (i.e., variations in solar activity) and hypotheses about alkalinity and ly- -1027 volcanic eruptions. The ECBILT-MOM socline depth—even if in this par- Claussen, M., Kubatzki, C., Brovkin, V., Ganopolski, V., Hoelzmann, P. and Pachur, H.J., 1999: Simulation 13 model, developed at the KNMI, has ticular case, δ C data fail to provide of an abrupt change in Saharan vegetation in the been used to examine the phase sufficient constraint. The alternative mid-Holocene. Geophysical Research Letters 26: relationships between solar forcing approach is to use the model to 3037-2040 Collins, M., Osborn, T.J., Tett, S.F.B., Briffa, K.R. and and the extent of glaciers in the Alps. generate synthetic records. In other Schweingruber, F.H., 2002: A comparison of the Studies of the last millennium have words, attempting to reconstruct the variability of a climate model with paleotem- also highlighted the importance of a raw data seems to be a promising perature estimates from a network of tree-ring densities. Journal of Climate 15: 1497-1515 good knowledge of the natural vari- way to better understand how cli- ability in forcing mechanisms (e.g. mate worked in the past. MICHEL CRUCIFIX1 AND PAUL VALDES2 Collins et al., 2002). Model simula- The meeting suggested many 1Hadley Centre for Climate Prediction and tions must include this variability if future research directions. It also Research, UK; we are to faithfully test the models. highlighted the importance of study- michel [email protected] 2University of Reading, UK; Currently, our knowledge of this forcing both past millennium and longer [email protected] ing is generally rather poor. time-scale variability. All of the ap- The increasing importance of proaches are valuable for improv- gathering more data and modeling our understanding of climate ling expertise on the global carbon variability and for evaluating climate Investigating Holocene Climate Variability: Data-Model Comparisons

ESF/HOLIVAR MEETING. LOUVAIN-LA-NEUVE, BELGIUM, 12-15 JUNE 2002 Comparing model output with in- son methods were presented at the GCM (coupled ocean-atmosphere formation provided by proxy data workshop to address some of these general circulation model) control was discussed in detail. The com- difficulties. run, and the temperature variability parison is challenging because the The traditional inverse approach reconstructed from tree-ring density characteristics of model output and is to reconstruct some aspect of across the Northern Hemisphere. The paleodata are very different and climate from proxy data (often results were sensitive to the process- many sources of uncertainty exist. including an aggregation to larger ing of the tree-ring data but indicated Climate is represented at different spatial scales more suited for com- an underestimation of the variability spatial scales: local in proxy records, parison with model output) and by the model. This could be partly several hundred kilometers or more then compare the reconstruction to accounted for by the influence of in models. Further, the registered simulated climate data. It was dem- natural external forcing changes (so- variability in proxies is only partly onstrated during the workshop how lar irradiance or volcanic aerosols). caused by climatic variations, so that the non-climatic residuals influence Comparisons of proxy-based tem- it is necessary to isolate the climatic variance, trends and spatial patterns perature reconstructions with EBM signal using statistical methods and and that appropriate treatment of (energy balance model) simulations to represent the non-climatic residu- these residuals can avoid biased of the response to natural and an- als (perhaps by a suitable stochastic comparisons. Examples were given thropogenic forcings (updated from model). Finally, the responses of the (Collins et al., 2002) of the compari- Crowley, 2000) were presented, with proxies to the local or large-scale son, at the inter-decadal time-scale, the reconstructions used to identify climate may be non-invertible. A between levels of internally gener- the climate sensitivity that provides number of model-proxy compari- ated variability simulated during a the best fit between model and data

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 32 Workshop Reports: 2nd Holivar Workshop Workshop Reports: 2nd Holivar Workshop 33

�� •the Sub-Boreal–Sub-Atlantic tran- �� sition around 850 BC (2750 y

� BP), � ��

� •the termination of the African Hu-

� mid Period around 6–5 ky BP, � �

� and � �� the 8.2 ky BP cooling event.

� • �

� Evidence was presented from proxy �

� data that significant climate change � �� occurred during these periods. The � � � stronger signal-to-noise ratio for � �

� these events should make diagnosis �

� of climate behavior more feasible in �

� �� both data and models. It was noted �

� (Fig. 1) that model studies focusing � � on these periods should also aim to � � perform ensemble runs to capture � �� the range of inherent model variability (e.g., Renssen et al., 2002), thereby better allowing the model to encompass the single realization �� represented by reality (as registered Figure 1: Simulation of the 8.2 ky BP event (Renssen et al., 2002): response of the thermo- in proxy data). haline circulation to freshwater pulses of 20 Sv released during simulation years 550-570 in the Labrador Sea, for 5 ensemble members (each offset by 20 Sv for clarity). Other forcings Discussion made it clear that a are representative of 8.5 ky BP. hierarchy of modeling studies, in- for the past 1000 years. On longer cluding the simulation of tree rings, cluding GCMs and EMICs, was con- time-scales, both quantitative and ice cores, glacier length and local sidered essential given the uncertain qualitative comparisons of proxy sea level (e.g., Reichert et al., 2001; forcings, the need for ensemble data with the climate simulated by Weber and Oerlemans, 2002). simulations, and the complex in- GCMs and EMICs (Earth system While inverse reconstruction and teractions taking place within and model of intermediate complexity) forward modeling with proxy data al- between subsystems of the Earth were presented, using simulations low comparison with model output, system. It was proposed that the of time slices, transient simulations presentations were also made of HOLIVAR program establishes a da- of particular events (e.g., Fig. 1 and possible methods for the combina- tabase of the different model runs Renssen et al., 2002) and transient tion of proxy and model data to covering the Holocene that would simulations of virtually the whole provide improved estimates of past be accessible to the scientific com- Holocene (Crucifix et al., 2002; climate variations. One technique, munity. Brovkin et al., 2002). named DATUN (Data Assimilation Further information is available The workshop participants Through Upscaling and Nudging, at http://www.cru.uea.ac.uk/~timo/ agreed on the great potential offered Jones and Widmann, 2002), ‘nudges’ holivar/ by the forward-modeling approach, the climate towards an atmospheric in which appropriate process-based circulation state that has been re- REFERENCES Collins, M., Osborn, T.J., Tett, S.F.B., Briffa, K.R. and (physical, biological, chemical) or constructed from a set of paleodata, Schweingruber, F.H., 2002: A comparison of the empirical models are driven by while remaining nearly consistent variability of a climate model with palaeo-tem- climate model output to simulate with the model’s physics and the perature estimates from a network of tree-ring densities. Journal of Climate 15: 1497-1515. a proxy value or time series that applied external forcings. A second Crowley, T.J., 2000: Causes of climate change over is then compared with the actual method was proposed (by Mat Col- the past 1000 years, Science 289: 270-277. proxy data. This approach can deal lins), whereby many realizations of Renssen, H., Goosse, H. and Fichefet, T., 2002: Mod- explicitly with non-linear and non-in- each simulation year are generated eling the effect of freshwater pulses on the early Holocene climate: the influence of high frequency vertible proxy response to multiple (again under appropriate external climate variability. Paleoceanography 17, No. 2, climate drivers, and can also aid forcing variations) and the closest 10.1029/2001PA000649. our understanding of the processes analog to the available paleodata for responsible for the proxy behavior. that year is selected before proceed- For full references please consult: www.pages-igbp.org/products/newsletters/ref2003_2.html Insufficient development, tuning and ing to the next year of the simula- validation of process-based models tion. It was proposed to concentrate HANS RENSSEN1 AND TIM OSBORN2 for different proxy types and loca- efforts on specific periods during the 1Universiteit Amsterdam, Netherlands, [email protected] tions has restricted the application Holocene: 2 th University of East Anglia, United Kingdom, of the forward-modeling approach •the classical “Little Ice Age” (16 [email protected] to date but noteworthy progress to 19th century AD), was reported at the workshop, in-

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 34 Workshop Reports Epilogue 35

Environmental Magnetism

INDIAN INSTITUTE OF GEOMAGNETISM, NAVI MUMBAI, INDIA A one-day workshop on “Environ- mental Magnetism” was held at the Indian Institute of Geomagnetism (IIG), Navi Mumbai in honor of the donation of equipment from the Al- exander von Humboldt Foundation, Germany to IIG’s Environmental Magnetism Laboratory (EML). The equipment was formally handed Figure 1: Lonar Lake over to IIG’s EML group on 15 April field and the manifestation of con- There is a universal need to have 2003 by Ms. Gerda Winkler, Deputy sequences of these interactions in reliable constraints on the timing Consul General of Germany - Mum- the interplanetary region. of any geological formation. Radio- bai. EML was initially set up at the The scientific session was metric dating is one such authentic Colaba campus in 1996 albeit only opened with an overview of re- tool, however, the cost factor de- with basic magnetic measurement search using mineral magnetism ters scientists from widely adopt- equipment. The equipment do- being carried out in various parts ing the technique. The workshop nated by the Humboldt Foundation of India. The application of this tech- highlighted the relatively inexpen- includes a Spinner Magnetometer nique has confirmed the presence sive luminescence dating method, and a Thermal Magnetic Suscepti- of the younger dry cold event re- which can provide accurate data at bility System for temperature mea- corded in proglacial lake deposits of a lower price. Chronological results surement in the range of -200°C to Garbyang, higher central Himalaya. on Aeolian deposits of Western In- 800°C. Other equipment at EML Other sites being studied across In- dia were presented. includes an AF Demagnetizer, Dual- dia include Lonar Lake (Fig. 1), Gulf The scientific session also saw Frequency Magnetic Susceptibility of Kachh and Cambay, east coast discussion on different aspects of Meter and Pulse Magnetizer. Thus, deltas and their mangrove depos- climate change and paleoenviron- EML is now a state-of-the-art facil- its, and the Thar Desert lakes. The mental reconstruction involving ity for carrying out sophisticated usefulness of gleaning information field studies in targeted regions. environmental mineral magnetic from mineral magnetic properties Results of granulometric and ar- experiments on all types of geo- to reconstruct past climates was cheological studies in the Mahi logical materials. discussed. The importance and utili- and Sabarmati River basins, the The purpose of the workshop ty of magnetic susceptibility studies salt lakes of the Thar Desert, and was to announce the availability of in understanding the Deccan trap mangroves were presented. The EML’s high-tech equipment to the basalts was also covered in great significance of lakes as archives Indian scientific community and detail. The practical significance of of climate change was emphasized to highlight the merits of applying domain states of Deccan basalt as by many participants. The workshop mineral magnetism techniques for paleoenvironmental indicators was stressed the need to integrate mul- resolving geophysical, geological highlighted. The importance of field tidisciplinary approaches in the un- and paleoclimatological problems. studies in deciphering Quaternary derstanding of past climate change The scientific program was jointly climate and geoarcheological fea- through remote sensing and geo- convened by N. Basavaiah (IIG, tures was delineated by discussing morphologic studies. Navi Mumbai) and A.S. Khadki- specific examples covering differ- The workshop concluded with kar (Agharkar Research Institute ent parts of India, Nepal and Israel. a panel discussion that generated (ARI), Pune). Participants included It was shown how field evidence of several recommendations. Fore- scientists from Tata Institute of Fun- environmental change and paleo- most among them was a proposal damental Research, India Institute monsoon signatures are reflected to undertake a multilayered and of Technology-Powai, IIG, Bhabha through the formation of paleosols, multidisciplinary study to clarify Atomic Research Centre, ARI, Dec- fluvial landforms and desert dunes. the origins of Lonar Crater Lake in can College, Oil and Natural Gas The highlight of the meeting was Maharashtra. The calibration of in- Corporation and Maharaja Saya- an introduction to the principles struments through inter-laboratory jirao University of Baroda. and applications of Mössbauer and multi-instrument comparisons The workshop consisted of a spectroscopy in elucidating mag- was also recommended. general and a technical session. netic mineralogy. Deciphering the The general session, open to all, oxidation state of iron in magnetic N. BASAVAIAH was devoted to “Space Weather”. minerals and its general applica- Indian Institute of Geomagnetism, Navi Mumbai, India Prof. G. S. Lakhina (Director, IIG) tion to geological problems was [email protected] gave a lecture on the effects of the discussed. sun’s radiation on Earth’s magnetic

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 34 Workshop Reports Epilogue 35

Changing Perspectives for Changing Times

My love affair with the Holocene began almost 50 years ago. In those days it was called the ‘Post glacial’, or the ‘Flandrian’. At the time, my Cambridge mentors were much beguiled by the Late glacial. For me though, the main excitement came from trying to reconstruct and better understand the peopled landscapes of the past. Despite many deflections from that concern, it has returned consistently, but in so doing, it has not simply been renewed, but rather reinvented. The passage of time has transformed the priorities for Holocene research in ways that could not have been imagined half a century ago. The idea that learning about past climate variability might inform our view of impending future changes would have seemed far-fetched, and the concept of multi-proxy climate reconstructions based on biological remains, quite out of the question. In the late 1950’s, radiocarbon dating was in its infancy and, for most researchers, dates were extremely difficult to obtain. What will be the ‘state of the art’ and what will our current ‘tool kit’, look like in fifty years time? Technical innovations are among the most obvious facilitators of future changes in methodology, but for anyone lacking the relevant expertise in instrument design, they are among the least predictable. In my own experience, the interactions between those posing palaeo-environmental questions and those contributing the technology for resolving them, have been rather like a series of dead ends suddenly and unexpectedly opening out into serendipitous partnerships of great delight and excitement. Since the palaeo-scientists’ responsibility begins with the questions posed, we should turn to these. I think there are some safe assumptions – that the underlying ‘future-oriented’ research agenda for palaeo-science will not be entirely replaced by something quite new and different; the synergy between model and data communities will increase and become central to most advances in the field; computing power will serve this synergy with ever enhanced capabilities; the speed of data acquisition and assimilation will increase at an accelerating rate; the climatic calibration of environmental proxies and associated uncertainties will become increasingly quantitative; and finally, comparisons between past and present climates on a regional, zonal, hemispheric and global basis will become increasingly secure. Holocene research has strengths arising from the unique opportunities it offers for secure quantitative proxy calibration with minimum extrapolation, for exploring the antecedents of current conditions, whether of climates or ecosystems, for generating multi-archive/multi-proxy reconstructions with high temporal resolution, for elucidating the long term behaviour of current modes of variability, and for providing the longer term historical context for the detection and attribution of future environmental changes. In my view, two of the most intransigent weaknesses are in the areas of chronology and the reconstruction of the long-term dynamics of terrestrial ecosystems. Resolving many of the crucial future research questions will demand increasingly precise and accurate chronologies, and in some key environmental contexts, they remain elusive. My second point may seem harsh or ill-informed, but I make it nonetheless, for it also signals a major opportunity. If we consider the synergy between PAGES and CLIVAR, it rests in part on a shared exploitation of the continuum between the palaeo-record and the period of direct observations. Palaeo-climatology, within this context, makes a crucial contribution to dynamics rather than ‘history’. As and when the implications of future global change become more clearly and widely apparent, one corollary must surely be a growing interest in the past response of ecosystems to a variety of interacting perturbations, on a range of timescales. What if the hoped-for broader synergy between ‘palaeo’ and ‘present/future’ science fails to develop? Certainly, that threat lies implicit in the gap between ecosystem modelling communities using inputs derived from the short period of direct observations, and the terrestrial palaeo-science community that is often seen as dealing with history rather than process. When we compare the situation with that prevailing at the CLIVAR/PAGES interface, the gap is all the more unfortunate: whereas a global network of standardised and verifiable climate records goes back for a century or so, direct observations of ecosystem structure and function are much less systematic and more recent and patchy over most of the world. Closing this gap between ‘palaeo’ and ‘process’ in terrestrial ecosystem studies is an immense challenge, but one that must be met if both communities are to benefit and to make the best possible contribution to our understanding of ecosystem responses to global change. The time interval from which much of the necessary evidence will be drawn is the Holocene. During its life time, HOLIVAR will have helped to strengthen the basis for future research, both by coordinating and publicising what has been achieved so far, and by helping to create a cohort of young researchers whose skills and opportunities will, in due course, outstrip those of their mentors. That is as things should be.

FRANK OLDFIELD Visiting scientist PAGES IPO [email protected]

PAGES NEWS, VOL.11, N°2&3, OCT. 2003 PAGES NEWS, VOL.11, N°2&3, OCT. 2003 36 Last

CALENDAR 2003 / 2004

October 27 - 30, 2003, Seattle, USA April 1 - 7, 2004, Berlin, Germany First SEARCH Open Science Meeting Paleoclimate Proxy Data for Reconstructing Holocene Environments of West Asia Further information: http://www.arcus.org/search/search.html Further information: [email protected] November 2 - 5, 2003, Seattle, USA 2003 GSA Annual Meeting: Special PAGES Sessions July 4 - 9, 2004, Granada, Spain T96: Lakes and Holocene Environmental Change International Palynology Congress (IPC) T98: Low Latitude Influences on the High Latitude, or Tropical Session the Other Way Around Further information: Further information: http://www.11ipc.org/ http://geosociety.org/meetings/2003/ http://www.pages.unibe.ch/calendar/2004/topical_IPC.html

November 16 - 19, 2003, Trieste, Italy September 1 - 3, 2004, Bergen, Norway First International Young Scientists Global Change Bjerknes Centenary 2004: Climate Change in High- Conference Latitudes

Further Information: Further Information: http://www.ictp.trieste.it/%7etwas/YS_START_2003_ http://www.bjerknes.uib.no/conference2004/ announce.html

http://www.pages-igbp.org/calendar/calendar.html

��

PAGES NEWS, VOL.11, N°2&3, OCT. 2003