INTERGLACIAL-GLACIAL VEGETATION DYNAMICS – A

NATURAL LONG-TERM ECOLOGICAL LABORATORY

H. John B. Birks1,2 & Katherine J. Willis1,3

1University of Bergen, 2University College London, 3University of Oxford

International Geological Congress, Oslo, 13 August 2008

Introduction

Apologies

Glacials and interglacials

Why study interglacials?

Interglacial vegetation dynamics

General patterns

Bioclimatic affinity groups

Whole-assemblage approach

Vegetation dynamics within interglacials

Differences between interglacials

Plant extinctions in interglacial-glacial cycles

Lessons from interglacial-glacial long-term experiments

Acknowledgements

INTRODUCTION

Apologies

• Kathy Willis

• Change of title and abstract

• Contains little new information but hopefully presents some new ways of thinking about interglacial and glacial vegetation dynamics as a natural long-term ecological laboratory

• Limited geographical focus – primarily NW Europe, with brief excursions to the Balkans

Glacials and Interglacials

The Quaternary period is the past 2.4 million years of Earth’s history. A time of very marked climatic and environmental changes with multiple glacial-interglacial cycles driven by variations in orbital insolation on Milankovitch time-scales of about 400, 100, 41, and 19-21 thousand year (kyr) intervals

Glacial conditions account for up to 80% of the Quaternary

Remaining 20% consist of shorter interglacial periods during which conditions were similar to, or warmer than, present day

Oxygen isotope fluctuations. Observations made from several deep sea cores (Modified from Shackleton et al. 1993 and Shackleton & Opdyke 1976)

Climate and glacier fluctuations observed in deposits on the continent in Europe. Red: warm phases, blue: cold phases.

Andersen 2000

Major climate forcing for the last 450 kyr calculated at 60N. Global ice volume (f) plotted as sea-level, so low values reflect high ice volumes.

Jackson & Overpeck 2000

WHY STUDY INTERGLACIALS?

1. Primarily studied for stratigraphical purposes. Provide main basis for local and regional terrestrial Quaternary stratigraphies. Endless discussion and disagreements between Quaternary stratigraphies about correlations and interpretations, especially at a local geographical scale (e.g. East Anglia).

2. Shackleton & Opdyke (1973, 1976) V28-238, V28-239, etc. deep-sea cores. Multiple glacial-interglacial cycles in a complete sequence to Pliocene. ‘The Rosetta Stone of the Ice Ages’

Nick Shackleton, Phil Gibbard, and others in Cambridge, about 1974. Nick jokingly asked “Why bother with these local interglacial stratigraphies when we now have a complete global stratigraphy?”

Nick suggested why not study terrestrial interglacials as ‘natural long-term ecological experiments?’

Interglacials are, after all, a series of unique natural long-term experiments in vegetation dynamics. This aspect of interglacials has been curiously ignored in long-term ecological studies and in ecological-based Quaternary vegetational history.

Using the palaeoecological record as a natural long-term ecological laboratory has been recently emphasised by the National Research Council’s (USA) report on The Geological Record of Ecological Dynamics – Understanding the Biotic Effects of Future Environmental Change

Flessa & Jackson 2005

Three major research priorities emerge from this report:

1. Use the geological (= palaeoecological) record as a natural laboratory to explore biotic responses under a range of past conditions, thereby understanding the basic principles of biological organisation and behaviour. THE GEOLOGICAL RECORD AS AN ECOLOGICAL LABORATORY. ‘Coaxing history to conduct experiments’ Deevey (1969).

2. Use the geological record to improve our ability to predict the responses of biological systems to future environmental change. ECOLOGICAL RESPONSES TO ENVIRONMENTAL CHANGE.

3. Use the more recent geological record (e.g. mid and late Holocene and ‘Anthropocene’) to evaluate effects of anthropogenic and non-anthropogenic factors on variability and behaviour of biotic systems. ECOLOGICAL LEGACIES OF SOCIETAL ACTIVITIES.

Only 1 and 2 relevant here.

Despite Nick Shackleton and Neil Opdyke’s ‘Rosetta Stone of the Ice Ages’, problems remain in correlating terrestrial interglacial records with Marine Isotope Stages (MIS) of V28-239 and ODP 677.

Coxon & Waldren 1997

INTERGLACIAL VEGETATION DYNAMICS

General Patterns

NW Europe – many inter-glacial pollen stratigraphies, e.g. East Anglia

Ipswichian(=Eemian) MIS 5cHoxnian(=?Holsteinian) MIS 11c or 7c

West 1980

Cromerian MIS? Pastonian MIS?

Bramertonian MIS?

Ludhamian, Thurnian, Baventian MIS?West 1980

All show four main phases

1. Pre-temperate phase – boreal trees (e.g. Betula, Pinus)

2. Early-temperate phase – deciduous trees (e.g. Ulmus, Quercus, Fraxinus, Corylus)

3. Late-temperate phase – mixed deciduous and coniferous trees (e.g. Carpinus, Abies, Picea)

4. Post-temperate phase – boreal trees (e.g. Betula, Pinus, Picea) and heathland development

Bioclimatic Affinity Groups

Can group pollen types into plant functional types or bioclimatic affinity groups (BAG) on basis of plant morphology, deciduous or evergreen, and modern climate affinities.

Three BAGs represent 70% of pollen assemblages at Velay, French Massif Central with 4 interglacials in one continuous record. Longest record in NW Europe.1. Conifers2. Broad-leaved deciduous trees (mainly Quercus)3. Mesophilous temperate trees (mainly Fagus or Carpinus)

Rest are mainly herbs.

Dark green = ConifersLight green = Broad-leaved deciduous treesBlue = HerbsOrange = Mesophilous temperate trees

MIS 1 MIS 5e MIS 7c MIS 9c MIS 11c

Chedaddi et al. 2005

Chedaddi et al. 2005

Dark green = ConifersLight green = Broad-leaved deciduous treesOrange = Mesophilous temperate trees

Correlation coefficients between interglacials

Chedaddi et al. 2005

1. MIS 9c, 7c, and 5e closely correlated.

2. MIS 1 and 11c less well correlated even though they have similar precessional variations. MIS 11c is closest climatic analogue to Holocene.

3. General similar vegetation dynamics in all five interglacials even though their durations were very different and their solar insolation values were different.

Whole Assemblage Approach

Ioannina, Pindus Mountains, Greece

Tzedakis & Bennett 1995

180 m sequence, Holocene + 4 interglacials

MIS 5e

MIS 11c

MIS 9c

MIS 7c

5e 7c 9c 11c5e

7c - -9c - - - -11c - - - - - +

- climate different (solar insolation)- pollen different+ pollen similar

Tzedakis & Bennett 1995

Different interglacials have different solar insolation values and different pollen stratigraphies, except for two interglacials with similar pollen stratigraphies.

Interglacial pollen stratigraphies, unsurprisingly, generally become more similar as one goes further north

5e

?

7c or 11c

5e

7c or 11c

Donner 1995

Some general patterns exist when considered ecologically

Denm

ark La

pla

nd

Vegetation Dynamics Within Interglacials

diatoms

CaCO3LOI

Inso

lub

le

silic

ate

s +

q

uart

z

Pedi

astrum

alga

e

Erosion

High lake productivity

Other nutrientsLeaching

LIGHT DEMANDING herbs, Betula, Pinus

RICH MULL SOILS deciduous forest trees

MOR HUMUS/PEAT Picea, Sphagnum, Ericales

Lake sediment evidence

Protocratic

Mesocratic

Telocratic

(Eemian, Denmark)

Cumulative pollen curves and chemical composition for the interglacial deposit at Hollerup, DenmarkAndersen 1966

MIS 5e

Eemian IG (Inter-glacial)

5e

Holstein IG

7c or 11c

Harreskov IG

?

Average percentages of the pollen of light-plants. Black columns indicate percentages and white = percentages x10

PROTOCRATIC PHASE

LIGHT PLANTS

Andersen 1969

Eemian IG (Inter-glacial)

5e

Holstein IG

7c or 11c

Harreskov IG

?

Average percentages of the pollen of mull-plants and Alnus. Black columns indicate percentages and white = % x10

MESOCRATIC PHASE

MULL PLANTS

Andersen 1969

Eemian IG (Inter-glacial)

5e

Holstein IG

7c or 11c

Harreskov IG

?

ACID HUMUS PLANTS

Average percentages of the pollen of acid-humus plants. Black columns indicate percentages and white = percentages x10

OLIGOCRATIC AND TELOCRATIC PHASE

Andersen 1969

Zone averages for light-demanding, brown-earth mull, and acid humus pants in four interglacials. Zones joined in order from bottom to top. Consistent patterns but different relative abundances between interglacials. Andersen 1994

+ Homo sapiens phase in Holocene

Summary of Iversen’s (1958) glacial-postglacial (interglacial) cycle

What causes the change from the mesocratic with maximum forest cover and biomass and fertile soils to the oligocratic or telocratic phases with decreasing forest cover and biomass and less fertile soils?

Need long (> 6000 year) ecological successions where one can study vegetation, soil, and ecosystem properties along chronosequences of sites of different but known ages.

Wardle et al. 2004 Science 305: 509-513

Birks & Birks 2004 Science 305: 484-485

Maxim

al

ph

ase

Retr

og

ress

ive

ph

ase

Cooloola, Australia

>600,000 yr

Arjeplog, Sweden 6000 yr

Glacier Bay, Alaska

14,000 yr

Dunes Islands Moraines

Wardle et al. 2004

Maxim

al

ph

ase

Retr

og

ress

ive

ph

ase

Hawaii 4.1 x 106 yr

Franz Josef, New Zealand >22,000 yr

Waitutu, New Zealand

600,000 yr

Lava flows Moraines Terraces

Wardle et al. 2004

Tree basal area – unimodal or decreasing response with age

Wardle et al. 2004

Measured C:N, C:P, and N:P ratios for humus and litter

Significant increases in N:P and C:P ratios with age and forest retrogression Wardle et al. 2004

In the transition from the maximal forest biomass phase to the retrogressive phase, P becomes more limiting relative to N and P concentrations decline in the litter.

N is biologically renewable but P is not, as P is leached and bound in weathered soils. Over time, P becomes depleted and less available, relative to N.

Reduced rates of litter decomposition and release of P from litter and decreased activity of microbial decomposers.

Proportion of fungi relative to bacteria increases. Fungal-based food webs retain nutrients better than bacterial-based food webs.

Nutrient cycling thus becomes more closed and essential nutrients, especially P, become less available.

Long-term decline in biomass is accompanied by increasing P limitation relative to N, reduced rates of P release from decomposing litter, and reductions in litter decomposition, microbial biomass, and ratio of bacterial to fungal biomass.

Tree biomass thus declines as P becomes increasingly limiting at onset of oligocratic phase.

Birks & Birks 2004

Difference Between Interglacials

Differences in relative abundance

Picea Corylus

+Abies 3/8*Tsuga 3/8

Fagus 2/8*Pterocarya 4/8+Picea 6/8*Not native in Europe+Not native in UK

West 1980Taxon occurrences

These and other differences between interglacials may reflect biotic responses to

• differences in climate in different interglacials (insolation, precipitation, CO2, etc)

• differences in genetic variability of taxa between interglacials (e.g. Fagus, Corylus)

• differences in biotic interactions in different interglacials (plant competitions, plant-animal interactions, etc)

• differences in location of refugia between glacials

• different probabilities of extinction after different interglacials or in glacials

• different probabilities of external disturbance factors (fire, pathogens, etc)

• interactions between some of all of these factors

Embarrassingly ignorant of nearly all these potential drivers, especially when we consider interglacials

Different interglacials had different climates in terms of solar insolation values. Interglacial pollen stratigraphies are surprisingly broadly similar at the coarse plant functional or ecological group level but are, not surprisingly, different at the assemblage or individual taxon level.

In general each interglacial begins with high summer and low winter insolation, then both summer and winter insolation reach present-day values, and finally summer temperature decreases and winter temperature increases.

Main lesson from interglacials is the seemingly wide climatic tolerances of major tree taxa that dominate interglacials.

Virtually nothing is known about tree refugia prior to the Eemian (MIS 5e).

As regards the Weichselian, knowledge of tree refugia in Last Glacial Maximum (LGM) is greatly changing, thanks to an increasing emphasis on plant macrofossil remains.

Interglacial

LGM

S N

Traditional refugium model – narrow belt in southern mountains

van der Hammen et al. 1971

1. Trees in the LGM southern and Mediterranean refugia

Location of Ioannina basin in Pindus Mountains, NW Greece

Tzedakis et al. 2002

LGM

Pinus

Quercus

Fagus

Ulmus

Corylus

Alnus

Pistacia

Tilia

Betula

Abies

Bennett et al. 1991

Pollen and plant macrofossil evidence for traditional southern European LGM refugial model

2. What about trees in central, eastern, and northern Europe during the LGM?

Detection difficult

1. Low pollen values – do these result from long-distance pollen transport or from small, scattered but nearby populations?

Classic problem in pollen analysis since Hesselman’s question to Lennart von Post in 1916. No satisfactory answer.

2. Few continuous sites of LGM age

3. Pollen productivity related to temperature and some trees cease producing pollen under cold conditions

4. Pollen productivity may also be reduced by low atmospheric CO2 concentrations

5. Other sources of fossil evidence critically important – macrofossils, macroscopic charcoal, and conifer stomata

Scanning electron microscope images of wood charcoal

Willis & van Andel 2004

SloCro

Aus

Svk

Hun

CzR PolUkr

Rom

SerB&H

CzR–Czech Republic; Aus–Austria; Slo–Slovenia; Cro–Croatia; Pol–Poland; Svk–Slovakia; Hun–Hungary; Ukr– Ukraine; Rom–Romania; Ser–Serbia; B&H–Bosnia & Herzegovina

Occurrences of macroscopic charcoal in central Europe

Willis & van Andel 2004

Tree taxa that have reliable macrofossil evidence for LGM presence in central, eastern, or northern European refugia

Abies alba Pinus cembra

Alnus glutinosa Pinus mugo

Betula pendula Pinus sylvestris

Betula pubescens Populus tremula

Corylus Quercus

Carpinus betulus Rhamnus cathartica

Fagus sylvatica Salix aucuparia

Fraxinus excelsior Sorbus

Juniperus communis Taxus baccata

Picea abies Ulmus

Bhagwat & Willis 2008

Results from QUEST LGM plant macrofossil data-base for Eurasia, particularly former USSR for 25000-17000 cal yr BP

(LGM ice and shorelines follow Peltier 2004)

Some trees (Picea, Larix) very close to ice-sheet margin as in eastern North America

Alnus - 25-17kyr cal BPLarix - 25-17kyr cal BP

Pinus – 25-17kyr cal BP Picea – 25-17kyr cal BP

Binney et al. submitted

MediterraneanLGM refugia

Northerly LGM refugia

Ice sheet

Willis et al. 2008 (in press)

3. Current LGM refugium model based on all currently available fossil plant evidence

PLANT EXTINCTIONS IN INTERGLACIAL-GLACIAL CYCLES

Well known that the European flora, especially trees and shrubs, has become increasingly impoverished from Miocene to Holocene.

Temp

Arcto-Tertiary flora

8 taxa

91 taxa

van der Hammen et al. 1971

In current discussions about consequences of future global warming on population and species survival, critical to discover if past extinctions took place in glacial or in interglacial stages.

Lack basic data from glacial stage prior to the Eemian.

Losses in NW Europe

Weichselian 2 taxa ?Bruckenthalia (=Erica) spiculifolia, Picea omorika

Eemian 5 taxa Dulichium arundinaceum, Brasenia schreberi, Chamaecyparis thyoides, Cotoneaster acuticarpa, Aldrovanda vesiculosa

Holsteinian 7 taxa Azolla filiculoides, Pterocarya, Nymphoides cordata, N. peltata, Rhododendron ponticum, Osmunda cinnamomea, Osmunda claytoniana

Cromerian 6 taxa Eucommia, Celtis, Parthenocissus, Liriodendron, Tsuga, Aesculus

Many of these appear to be less frost-tolerant than currently widespread European plants. Extinction may have occurred in glacial rather than interglacial stages

Tenaghi Philippon 1.4 myr

1. More diverse flora before MIS 22-24

2. Remaining taxa extinct in MIS 16 or soon after (e.g Cedrus, Carya, Eucommia)

3. No extinctions in MIS 12 (maximum extent of ice sheets based on benthic 18O isotopes)

Tzedakis et al. 2006

Late Pliocene pollen record from Pula Maar, Hungary, 320,000 year record

Kathy Willis et al.

Pula Maar, Hungary

• sediments composed of finely laminated oil-shales, each couplet represents a single year of accumulation

• sequence contains c. 320,000 years in annual layers between c. 3.0-2.67 Myr

• analysed for pollen, isotopes, and geochemistry at interval of every 2500 years

Fossil pollen %

Palaeo-richness

Palaeo-energy (Laskar’s 2004 insolation calculations)

Palaeo-water balance (δ18O)

Willis et al. 2007

Dark grey = warm temperateWhite = herbsLight grey = cool temperateBlack = boreal

Insolation (scaled)

Taxon

om

ic

rich

ness

(sc

ale

d)

r2= 0.22

Non-linear relationship

Highest richness corresponds to intermediate insolation

Energy-based model – Laskar energy calculations vs pollen richness

Willis et al. 2007

Combined water-energy model (energy2 + water)

Combined water-energy model provides much better predictor of richness than water or energy alone

Willis et al. 2007

High amplitude fluctuations appear to lead to low taxonomic richness

Local extinctions of Sequoia, Nyssa, Parrotia, Fagus, etc. occur in periods of high amplitude fluctuations – reductions of regional species pool

Willis et al. 2007

Global extinction of Picea critchfieldii, abundant in LGM of south-eastern US, occurred during last deglaciation (16,000-10,000 yr BP), a time of rapid climate change (Jackson & Weng 1999).

Glacial stages are highly dynamic climatically with cold, dry conditions in response to Heinrich events and Dansgaard-Oeschger interstadial-stadial variability. High amplitude fluctuations during glacial stages may be responsible for extinctions.

All other late-Quaternary extinctions are thought to be the result of human activity (e.g. Easter Island palm).

Regional late-Quaternary extinctions are surprisingly rare (e.g. Koenigia islandica from central Europe in LGM, Larix sibirica from central Sweden in mid-Holocene, Picea omorika from north-west Europe).

LESSONS FROM INTERGLACIAL-GLACIAL LONG-TERM EXPERIMENTS

1. Considerable climatic fluctuations and environmental changes in the Quaternary with many glacial-interglacial cycles

2.Biotic responses to major climatic changes

• distribution changes• high rates of population turnover• changes in abundance and/or richness• extinctions• speciations• stasis

3. The interglacial–glacial ‘laboratory’ results show that biotic responses to rapid climate change were mainly redistribution of species, genera, families, and vegetation types, high turnover, abundance changes often resulting in local or regional extinctions but very rarely any global extinctions. No evidence for speciation. Vegetation types often have no convincing modern palynological analogue.

4. Biotic responses have been varied, dynamic, complex, and individualistic.

5. Interactions between climate change and biodiversity may be non-linear and possibly unpredictable. Important to identify thresholds and irreversible changes at time-scales realistic to the organism’s life-history (100-250 yr rather than 4 yrs of PhD projects, 5-10 yrs of funded projects, or 10-50 yrs of ‘long-term’ ecological plots).

6. A combined use of these ‘laboratory’ results and molecular phylogenies could help understand rates and thresholds of climate-biodiversity interactions and provide some independent tests of current models and predictions of biodiversity response to future climate change.

ACKNOWLEDGEMENTS

The late Nick Shackleton for convincing me in 1974 that interglacials are ecologically very interesting!

Kathy Willis for many stimulating discussions and ideas, access to the QUEST plant macrofossil data, and being a co-author.

Heikki Seppä and Karin Helmens for organising this symposium

Hilary Birks, Keith Bennett, Bill Watts, and Herb Wright for many helpful

discussions about interglacial vegetation dynamics

Cathy Jenks for invaluable help