News from the Botanical Institute, University of Bergen November 1999

NORPAST news from the Botanical Institute, University of Bergen November 1999

by John Birks with contributions from Hilary Birks, Aage Paus, Sylvia Peglar, Steve Juggins, and Mike Hughes

Quantitative Reconstuctions of Holocene climate: An Oslo to Trondheim transect

John Birks and 4 colleagues from Bergen, London, and the British Antarctic Survey did field work from August 10 - 24 1999. They cored 10 lakes on a 430+ km transect from near Halden to central Dovre and collected complete cores of Holocene lake-sediments from these lakes. These cores provide the basic material for the NORPAST project on Quantitative Reconstructions of Holocene climate: An Oslo to Trondheim Transect (5.1.4). The lakes were chosen to have comparable water depths, morphometry, and size but to lie in the different major vegetational regions (e.g. nemoral-Boreal, southern Boreal, middle Boreal, northern Boreal, low-alpine, middle-alpine) along the strong climatic gradient in summer and winter temperatures from south-east Norway to Dovre. The altitudes of the lakes cored vary from 130 m to 1261 m. Great care was taken to find lakes above the highest marine limit in the Oslo area and thus contain a full undisturbed record of Holocene lake sedimentation. The lakes cored in Dovre link directly with the study sites in NORPAST project 5.1.1. In addition surface-sediment and water samples were collected from a further 12 lakes to add to the surface-sediment collection in Bergen that is used to develop organism - climate and organism - lake-water chemistry transfer functions (5.5.1).

Haugstjern37.jpg (27766 bytes)
Bruce Terry (British Antarctic Survey) and Wenche Eide (Bergen) extruding a basal sediment core from Haugtjern, a lake at 338 m elevation in the southern Boreal vegetation zone west of Hamar. This lake is part of NORPAST subproject 5.1.4.

Hornsjoen.jpg (26725 bytes)

Hornsjøen, Dovre. This lake (altitude 1261 m) was cored in August 1999 as part of NORPAST subprojects 5.1.1 and 5.1.4. It will also be studied by Wenche Eide as part of her VISTA stipendiat and in the University of Bergen NFR Strategic University Project NORPEC.

High spatial and temporal resolution, multi-proxy reconstructions of late Weichselian and early Holocene climate

Hilary Birks, Aage Paus, John Birks, and colleagues from Bergen and London did field work from June 10 - 20 (south-west Norway) and from July 2 - 15 (northernmost Norway). They cored over 27 sites and examined a further 35 sites in the search for well-developed late-glacial sequences needed for the NORPAST project Lake sediments as archives of past climate along a north-south gradient in coastal Norway: High spatial and temporal resolution, multi-proxy recontructions of late Weichselian and early Holocene climate (5.1.2). They found one excellent site near Bjerkreim in south-west Norway and one very promising site near Hammerfest in the far north. This is probably the northernmost known late-glacial lake sequence in the world. In addition, important sites were found further east on Varanger and further west on Ringvassøya. These sites will provide the basic material for studying the last glacial - interglacial transition, and for linking the terrestrial response to rapid climatic change with changes detected in the marine and fjord records (5.4.1 and 5.4.2). A major research question relates to the timing of the onset of Gulf Stream influence. This can be addressed by the west - east transect of sites from Ringvassøya to Varanger. Full cores have been obtained from the site in south-west Norway and analyses of these will begin in January 2000 when project 5.1.2 starts. Full, large diameter cores from sites in the far north will be obtained in April 2000.

Rompetrollvatnet.jpg (22954 bytes)

Rumpetrollvatnet near Hammerfest. This lake was cored by Hilary Birks et al. in July 1999 as part of NORPAST subproject 5.1.2. The sediments have a characteristic tripartite sedimentary lithology and fossil composition suggestive of a late-glacial sequence. This and other lakes in the Hammerfest area will be fully cored next April from the ice

Transfer functions

During 1999, considerable progress has been made in expanding the existing data sets required for transfer functions that are essential for quantitative palaeoclimate reconstructions of Holocene and late-glacial biological assemblages from lakes, fjords, and oceans (5.5.1). Plans have been made to integrate the development of user-friendly computer software for the NORPAST community by combining forces with the EU project EDDI (European Diatom DataBase Initiative). Steve Juggins (Newcastle) and John Birks have designed the software and have each contributed their own programs (CALI, WAPLS, WACALIB, RMSEP, MAT, ANALOG). These will now be integrated into a single program (CALI2) with a user-friendly interface, a range of input facilities, data error checks, all the appropriate transfer function methods, cross-validation procedures, model and reconstruction diagnostics, sample-specific errors of prediction, graphics, etc. This work will start in April 2000 and continue to October 2001. In addition a program for the display and analysis of multi-proxy core data (STRAT) will be developed as an extra for NORPAST.

Steve Brooks (Natural History Museum, London) on field work on the Varanger Peninsula in July 1999. He is collecting modern chironomid midge larvae to compare with the chironomid head capsules preserved in the surface sediments of the lake behind him. This work is part of NORPAST subproject 5.5.1.

Synthesis of existing quantitative palaeoclimatic data
Synthesis of existing quantitative palaeoclimatic data for the last glacial - interglacial transition in Norway has started (5.6.1). Much of these data come from the NFR-funded inter-disciplinary project at Kråkenes led by Hilary Birks. These data are being stored in a relational data-base that can be directly linked to ARC-VIEW, the GIS software that will be used in the major data-synthesis part of NORPAST (5.6.1). Plans are well developed for developing a GIS-based data-base (5.6.1) in conjunction with Mike Hughes (University College London) that can be used directly by climate modellers (e.g. REGCLIM) for comparison with climate model hindcasts for past climates. Links with climate modelling groups in the UK and Denmark with keen interest and expertise in hindcasting past Holocene climates and model validation with proxy palaeoclimatic data are being actively pursued by John Birks and Atle Nesje. In Norway and NORPAST there are now the following quantitative transfer functions available for quantitative palaeoclimate reconstructions.For each data set the range of the environmental variable is given, along with the root-mean-square-error of prediction (RMSEP) based on leave-one-out cross-validation using either linear partial least squares or non-linear weighted averaging partial least squares.

Transfer functions – Current status in Norway

Method		Summer SST, ^oC	Summer salinity, ⁰/₀₀	Winter SST, ^oC	Winter salinity, ⁰/₀₀	Ice cover, months	CO₂, ppmv	July, ^oC	January, ^oC	Annual precipitation, mm
Marine diatoms Koç and Birks (146 samples)
	Range	0-20		-1-15.5
	RMSEP	0.91		1.03
Marine dinoflagellates Grøsfjeld, Birks et al. (371 samples)
	Range	-1.2-22.7	20-36.4	-1.9-16.2	27-36.3	0-11.9
	RMSEP	2.6	1.89	1.84	1.24	1.86
Atmospheric CO2 concentrations and Salix herbacea leaves - Hilary Birks et al. (54 samples)
	Range						203-365
	RMSEP						21.9
Terrestrial pollen - Birks, Peglar, Odland, and Seppa 292 samples from Norway (191 samples) and Finland (101 samples).
	Range							7.7-16.4	-17.8-1.1	300-3234
	RMSEP							1.03	2.56	417

Chironomids - Brooks and Birks (111 samples)
	Range							3.5-15.6
	RMSEP							0.89