Speleothem Science: From Process to Past Environments

Table of Contents

Cover

Title page

Copyright page

Preface

Acknowledgements

I: Scientific and geological context

CHAPTER 1: Introduction to speleothems and systems

1.1 What is all the fuss about?

1.2 How is this book organized?

1.3 Concepts and approaches of system science

1.4 The speleothem factory within the karst system

CHAPTER 2: Carbonate and karst cave geology

2.1 Carbonates in the Earth system over geological time

2.2 Lithologies of carbonate host rocks

2.3 Carbonate diagenesis and eogenetic karst

2.4 Speleogenesis in mesogenetic and telogenetic karst (with contributions from John Gunn and David J Lowe)

2.5 Cave infilling

2.6 Conclusion

CHAPTER 3: Surface environments: climate, soil and vegetation

3.1 The modern climate system

3.2 Water isotopes in the atmosphere

3.3 Soils of karst regions

3.4 Vegetation of karst regions

3.5 Synthesis: inputs to the incubator

II: Transfer processes in karst

CHAPTER 4: The speleothem incubator

4.1 Introduction to speleophysiology

4.2 Physical parameters and fluid behaviour

4.3 Water movement

4.4 Air circulation

4.5 Heat flux (authored by David Domínguez-Villar)

4.6 Synthesis: cave climatologies

CHAPTER 5: Inorganic water chemistry

5.1 Sampling protocols for water chemistry

5.2 The carbonate system

5.3 Weathering, trace elements and isotopes

5.4 Carbon isotopes

5.5 Evolution of cave water chemistry: modelling sources and environmental signals

CHAPTER 6: Biogeochemistry of karstic environments

6.1 Introduction

6.2 Organic macromolecules

6.3 Pollen and spores

6.4 Cave faunal remains

6.5 Synthesis and research gaps

III: Speleothem properties

CHAPTER 7: The architecture of speleothems

7.1 Introduction

7.2 Theoretical models of stalagmite growth and of stalagmite and stalactite shapes

7.3 Geometrical classification of speleothems

7.4 Mineralogy and petrology

7.5 Synthesis

CHAPTER 8: Geochemistry of speleothems

8.1 Analysis and the sources of uncertainty

8.2 The growth interface

8.3 Trace element partitioning

8.4 Oxygen and carbon isotope fractionation

8.5 Evolution of dripwater and speleothem chemistry along water flowlines

8.6 Process models of variability over time

CHAPTER 9: Dating of speleothems

9.1 Introduction

9.2 Dating techniques

9.3 Age–distance models

9.4 Conclusions

IV: Palaeoenvironments

CHAPTER 10: The instrumental era: calibration and validation of proxy-environment relationships

10.1 Available instrumental and derived series

10.2 Methodologies

10.3 Case studies of calibrated speleothem proxies

10.4 Questions raised and future directions

CHAPTER 11: The Holocene epoch: testing the climate and environmental proxies

11.1 A brief overview of the Holocene

11.2 The past millennium

11.3 Holocene environmental changes: speleothem responses

11.4 Questions raised and future directions

CHAPTER 12: The Pleistocene and beyond

12.1 Pleistocene proxy records (ice-age climate fluctuations defined and drawn)

12.2 Insights into pre-Quaternary palaeoenvironments

12.3 Questions raised and looking to the future

APPENDIX 1: Archiving speleothems and speleothem data

References

Index

Color Plates

This edition first published 2012 © 2012 by Ian J. Fairchild, Andy Baker.

Wiley-Blackwell is an imprint of John Wiley & Sons, formed by the merger of Wiley’s global Scientific, Technical and Medical business with Blackwell Publishing.

Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell’s publishing program has been merged with Wiley’s global Scientific, Technical and Medical business to form Wiley-Blackwell.

Registered office: John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial offices: 9600 Garsington Road, Oxford, OX4 2DQ, UK

The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

111 River Street, Hoboken, NJ 07030-5774, USA

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell.

The right of the author to be identified as the author of this work has been asserted in accordance with the UK Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data

Fairchild, Ian J. (Ian John)

 Speleothem science : from process to past environments / Ian J. Fairchild and Andy Baker.

p. cm.

 Includes bibliographical references and index.

 ISBN 978-1-4051-9620-8 (cloth)

 ISBN 978-1-4443-6106-3 (epdf)

 ISBN 978-1-4443-6107-0 (epub)

 ISBN 978-1-4443-6108-7 (mobi)

1. Speleothems. 2. Paleoclimatology. I. Baker, Andy, 1968– II. Title.

 GB601.F35 2012

 551.44'7–dc23

2011046018

A catalogue record for this book is available from the British Library.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Preface

This book is a response to the explosion of interest in speleothem archives of environmental change and our experience of being scientifically stretched to understand the processes that form them. Hence, in this volume we have constructed a broad syllabus, including much new material from our own research and scholarship, as an attempt to match the requirements of a core text for anyone starting research in speleothem science. We also trust that the book will be useful in relation to subjects such as Quaternary science, geochemistry and carbonate geology: for teachers, researchers and students. Quaternary science is multidisciplinary and likewise the science of speleothems draws on many subjects. Chapter 3 summarizes many facets of the global climate system and Chapter 9 focuses on issues in dating, several of which are common to other archives. Chapter 10 provides material on the calibration of proxies which should also be of general interest, while Chapters 11 and 12 draw out the special and often unique contributions made to Quaternary science by speleothem studies. Speleothem formation also illustrates and integrates many fundamental principles of geochemistry. Chapters 2, 6, 7 and 8 provide a complement to the classic marine emphasis of carbonate geology texts and provide exemplars for many core concepts in mineralogy and geochemistry. Students of water chemistry and hydrogeology will find an updated summary of many issues related to carbonate aquifers in Chapters 2, 4 and 5. There is also plenty of material here on general scientific principles, e.g. the emphasis on systems and material transfer, for example in Chapters 1 and 4, that could be effectively used as case examples in undergraduate degrees in physical geography and environmental science. Readers can find the illustrations from this book on-line at the publisher’s website http://www.ncdc.noaa.gov/paleo/paleo.html including additional use of colour, and we have also made available on-line on our own website www.speleothemscience.info the spreadsheets that were used to produce many of the new graphics. We would be grateful for readers pointing out errors that can be corrected on-line or in a future reprinting. Finally, we hope that this book will inspire some undergraduates towards the research frontier wherever they may find it in the environmental (geo)sciences.

This volume is a tangible outcome of the lively 6 years that we were co-located at the University of Birmingham. Here, we developed a common understanding of speleothem science while also pursuing our other research agendas. We planned the book in 2008 and both benefitted from periods of study leave in 2009: AB at University College, Durham, while holding an Institute of Advanced Studies Fellowship, and IJF at the University of Newcastle, Australia, financially supported by the Leverhulme Trust. Andy moved on to the University of New South Wales (UNSW), Australia, in January 2010, and the continuation of the Leverhulme Study Abroad Fellowship allowed IJF to catch up with AB in Sydney’s Northern Beaches in September 2010. Over the past 15 years at Birmingham, Exeter, Newcastle, Keele and UNSW, we have had the pleasure of supervising the work of many fine research students and fellows on speleothem-related work, several of whom are now forging their own research careers. The Natural Environment Research Council, the European Community, the UK’s Royal Society, the Australian Research Council and the Leverhulme Trust have supported our work through several projects. We are grateful to several colleagues with connections to Birmingham who have provided input or specific sections to this book; their contributions are listed in the contents. We salute the generations of speleologists who made so many of the discoveries that provided a foundation for our science, several of whom went on to be professional scientists. AB especially thanks Pete Smart, Larry Edwards, Dominique Genty, John Gunn, Tim Atkinson and Paul Williams for their advice, encouragement and debate. IJF made his underground scientific debut in September 1994 in the company of Silvia Frisia and Andrea Borsato, who have remained good friends and collaborators ever since, as subsequently have been Frank McDermott, Christoph Spötl, Dave Mattey and Pauline Treble, as well as those previously mentioned, and others we should have done. We have found the global speleothem community to be highly supportive and forward-looking, and are grateful for all the insight and scientific inspiration we have found there, including through the Climate Change—the Karst Record meetings, the sixth of which was held in Birmingham in June 2011. We are indebted to Ian Francis for his encouragement and to his colleagues at Wiley-Blackwell for their helpfulness and efficiency in the production process. We also thank the publisher’s three reviewers, Denis Scholz, Maurice Tucker and Ming Tan, and Gregoire Mariethoz and Bryce Kelly for their comments on different parts of our text. We are indebted to Anne Ankcorn and Kevin Burkhill of the School of Geography, Earth and Environmental Sciences at the University of Birmingham, for their excellent work on drafting many of the figures in this book. Both of us depend on the enthusiasm and support of our wives Sue and Jo. While Sue and Ian simply have a non-aggression pact regarding horses and caves, we have Jo to thank particularly for proof-editing the first drafts of our text. We trust the final version will be well received, although we are only too aware of the shortcomings and biases that books written during short and busy lives bring.

Ian J. Fairchild and Andy Baker

Acknowledgements

The authors are grateful to the copyright holders for permission to reproduce the following material:

Acta Carsologica (Figures 1.2 and 2.32)

American Association for the Advancement of Science (Plate 3.7 (part), 12.1, Figures 11.7, 12.4c)

American Association of Petroleum Geologists (Figure 2.35)

American Chemical Society (Plate 5.1, Figures 5.8, 5.9, 5.16)

American Geophysical Union (Plates 3.6 and 12.2; Figures 3.1, 3.4, 3.7, 3.9, 3.1, 3.12, 3.13, 4.23, 4.35, 11.16, 11.20)

Andrea Dutton (Figure 7.29)

Annual Reviews (Figure 4.31)

Australian Government (Plate 3.2)

British Cave Research Association (Figure 5.3a)

Chaoyong Hu (Plate 3.5, right)

CRC Press (Figures 3.8 and 5.18)

Eligio Vacca (Plate 7.9)

Elsevier-Pergamon (Plates 2.2, 2.3, 2.4, 7.6c, 11.2 and 11; Figures 1.1, 1.4 (left), 1.B3, 2.4, 2.5a, 2.6, 2.7, 2.11, 2.14, 2.17, 2.24, 2.28, 2.29a, 4.1b, 4.7, 4.15, 4.16a, 4.16c, 4.16d, 4.19, 4.22, 4.26, 4.29, 5.5, 5.6, 5.10, 5.14, 5.22, 5.23, 6.3, 6.4, 6.B1, 6.B4, 6.B5, 7.1, 7.5b-f, 7.20, 7.21, 7.26, 7.27, 7.28, 7.30, 8.2a, 8.3, 8.6, 8.7a, 8.8, 8.9, 8.12, 8.13, 8.17, 8.18, 8.19, 8.20, 8.23, 8.24, 8.25, 8.26, 9.2, 9.3, 9.5, 9.6, 9.7, 10.10, 11.2, 11.8, 11.9a, 11.10, 11.19, 12.2, 12.4a, b, 12.6, 12.7

European Mineralogical Union and the Mineralogical Society (Figure 5.17)

Fabrizio Antonioli (Plate 7.7)

Geological Association of Canada (Figures 2.19 and 2.20)

Geological Society of America (Figure 2.20, 2.22, 2.25, 3.16a, b, 6.5, 11.9b, c, 12.5)

Geological Society of London (Plate 7.3, Figures 1.5b, 2.1, 5.24)

International Glaciological Society (Figure 5.12)

International Journal of Speleology (Plate 7.5, 8.22, 11.13)

IPCC Report Climate Change 2007 The Physical Science Basis their Figure 7.5 (Figure 1.B1)

Jacqueline Shinker (Plate 3.1)

Journal of Geology (Figure 3.16d)

Jud Partin and PAGES (Figure 11.11)

Karst Waters Institute (Figure 1.10)

Mineralogical Society of America (Figures 5.7 and 8.5)

Ming Tan (Plate 7.6a; Figure 1.2a)

National Speleological Society (Figures 2.26 and 7.2)

Nature publications (Figure 11.15)

Otto de Voogd (Plate 7.1 (right))

Oxford University Press (Figure 2.3)

Paul Williams (Figure 2.2)

Phil Hopley (Figure 12.1)

Royal Society for Chemistry (Figure 6.B2)

Sage Publishers (Plate 7.4)

SEPM (Society for Sedimentary Geology) (Figures 2.16, 2.21, 7.12, 7.13, 7.17, 7.19, 7.22, 7.23)

Springer (Plate 3.7 (part), 2.15, 2.23, 3.B1, 4.32, 7.4b)

Stan Robinson (Plate 7.1 (left))

University of Chicago Press (Figure 3.3)

Verband Österreichischer Höhlenforscher (Figure 7.3)

Wedgwood Museum, Barlaston, Staffordshire (Plate 1.1 insets)

Wiley-Blackwell (Figure 2.5b, c, 2.12, 2.18, 2.27a, 2.33, 2.34, 3.14, 3.15, 3.16c, 3.17, 3.21, 4.13, 4.16b, 4.25, 4.33, 5.11, 7.11b, c, 7.18, 10.6, 11.4, 11.6)

I: Scientific and geological context

CHAPTER 1

Introduction to speleothems and systems

For paleoclimate, the past two decades have been the age of the ice core. The next two may be the age of the speleothem.

Gideon Henderson (Science, 2006)

1.1 What Is All the Fuss About?

Moore (1952) recognized a need to specify an unambiguous term for mineral deposits that grew within caves and proposed ‘speleothem’ (Greek: spelaion, cave; thema, deposit). In recent years speleothems have been established as one of the most valuable resources for understanding Earth surface conditions in the past, from times when glaciers waxed and waned and our human ancestors emerged, to the present day. By ‘conditions’ we mean not only the local context (soil, vegetation, landscape instability and climate), but also the regional to global patterns of change that characterize former environments and climates (Fig. 1.1).

Fig. 1.1 Speleothems as underground recorders of signals related to parameters of the external Earth system (Fairchild et al., 2006a).

Because no two speleothems are identical (Plate 1.1), they were formerly thought to be too complex to generate reliable archives of the past. It has taken a considerable effort over the past 40 years both to show that reliable records can be obtained which, in some cases, display global phenomena, and to demonstrate that speleothems form by a set of processes that can be rationalized and understood. Even in the late 1990s, textbooks on palaeoclimates and palaeoenvironments treated speleothems only briefly. Now it can be argued that certain long and well-dated records provide the most definitive archives of the global environmental system. Speleothems also provide an enormous resource for future research on past changes, with an ultimate dynamic range of eight orders of magnitude from days to a million or more years.

The rest of section 1.1 provides a summary of the essentials of speleothem science for the benefit of all those, from undergraduates to specialists in related fields, who want to get to first base. In section 1.2, we explain how the rest of the book is organized.

1.1.1 What Types of Speleothem Are Useful for Generating Climate Archives?

Many different forms of speleothem can be distinguished (Hill & Forti, 1997). Here we consider primarily two types of deposit (dripstones) that grow from dripping water: stalactites (Greek: stalaktós, dripping) growing down from the cave roof and stalagmites (Greek: stalagmós, dropping) building up from the cave floor. Also pertinent are more continuous deposits (flowstones) that accrete beneath thin sheets of water on cave walls and floors (Fig. 1.2a). Finally, we consider some long records obtained from crystalline deposits that form underwater. Flowstones tend to have fairly continuous layers and so it is possible to duplicate records by sampling (e.g. by coring) in different places. However, the layers can show a small-scale topography reflecting the ponding of surface water (Fig. 1.2a). Stalagmites are more commonly used to generate archives than stalactites because their internal structure is simpler. Fig. 1.2b illustrates a stalagmite cross-section illustrating that there is internal layering which tends to be flat on the top of the sample, allowing a set of observations representing different time periods in the past (time series) to be generated along a sub-vertical line. However, the sample illustrated in Fig. 1.2b displays lateral shifts in its growth axis, related to changes in the landing position of water drops. It is also commonly found that growth may pause for an extended period and then resume, and that these characteristics reflect either changes in climate or local processes within the aquifer in different cases.

Fig. 1.2 (a) Speleothems in Grand Roc Cave, Dordogne, France. The cave floor is covered with a flowstone on which have developed walls (terracettes) and intervening gour pools. A columnar stalagmite a few centimetres wide occurs to lower right and stalactites in the upper left. (b) Cross-section through a stalagmite illustrating visible growth layers (laminae) and the shifting position of the growth axis which represents the ideal position to generate the most continuous time series.

1.1.2 Where Do Speleothems Occur?

The speleothems used to find out about the past are almost all calcareous (made largely of calcium carbonate, CaCO3) and composed of the minerals calcite and/or aragonite. Such speleothems occur within carbonate rocks, typically limestones (CaCO3) and/or dolomite (CaMg(CO3)2. These rocks store, transmit and yield water readily and hence are aquifers. The permeability of aquifers refers to the ease with which they transmit water. Carbonate aquifers occur in karstic regions (Gunn, 2004; Ford & Williams, 2007) where there is little surface water because it drains readily into the bedrock (Fig. 1.3). The cavities in the bedrock typically range from tiny pores, through enlarged fissures to conduits which used to host, or still contain, underground streams. Over time, the drainage system develops and tends to lower the depth of the water table, the regional surface below which cavities are water-filled. The water table divides the aquifer into the underlying phreatic and overlying vadose zones, although in karst terrains the division is not always simple. Pragmatically, in this book we refer to the vadose zone, from which dripwater originates, as part of the aquifer, although conventionally the term is often restricted to the phreatic portion of the rock. Caves in vadose zones tend to cut vertically downwards, whereas in the phreatic zone they develop as passages elongated in the direction of water movement. Speleothems are normally formed in abandoned passages (Fig. 1.3) in the vadose zone and are fed by water passing through the soil into the uppermost karst, which is typically a zone of significant water storage (the epikarst). Speleothems form part of a long and complex history of cave spaces. Hence they can be interlayered with particulate sediments and are ultimately may be later broken by earthquakes or human disturbance, buried, flooded, eroded or dissolved.

Fig. 1.3 The complex structure of an active karst aquifer in which speleothems typically occur in caves that are no longer being developed and are well above the water table

(modified from Smart and Whitaker, 1991).

Modern calcareous speleothems occur in nearly all continental regions and are principally limited by the availability of karstic host rocks and liquid water. However, they tend to be only weakly developed in frigid regions where there is less chemical drive for them to form (Fig. 1.4b; section 1.1.3). They are found in many modern desert regions, having formed during more humid conditions in past millennia. Speleothems that originally formed not far above the water table can be found on mountain tops in tectonically uplifting regions. Conversely, precipitates that formed during previous ice ages when sea level was much lower are observed beneath the sea and can be recovered for study by divers.

Fig. 1.4 (a) The normal chemical pathway leading to speleothem formation. High carbon dioxide partial pressures ( c01ue005 ) arise in the soil (e.g. point A) owing to respiration and decomposition of organic matter. When percolating water with a high c01ue006 reaches carbonate minerals, they will be dissolved, increasing the calcium concentration in solution. If there is no renewal of CO2, the water follows a ‘closed system’ path to saturation point B, whereas if CO2 is replenished to maintain a constant c01ue007 then saturation will be reached at point C (both B and C are specific to 10 °C as shown). As the water descends the karst system, at some point it may encounter an air space with a lower c01ue008 than the original soil. The water degasses CO2 and enters the oversaturated field and tends to precipitate CaCO3 (e.g. dashed line C–D).

Modified from Kaufmann (2003).

(b) Holocene speleothems in a cave above the Arctic circle, Norway: slow growth limited by low c01ue009 in the soil zone.

1.1.3 How Do They Form?

The essential chemical processes that lead to speleothem formation are illustrated in Fig. 1.4a and described in its legend. Essentially there is a carbonate dissolution region in the soil and upper epikarst and an underlying vadose region of CaCO3 precipitation. These correspond to CaCO3-undersaturated and -oversaturated solutions respectively. The amount of calcite and dolomite that can be dissolved by water depends on its acidity, normally expressed as the partial pressure of carbon dioxide ( c01ue001 ) with which the water is stable. Although strong acid (e.g. from oxidation of sulphide minerals such as pyrite, FeS2) can be important locally, normally dissolution is enhanced mainly by high c01ue002 values of up to several per cent (0.01–0.1 atm) generated by respiration and organic decomposition in soils. As the water descends through the karst, it ultimately encounters a gas phase with a lower c01ue003 compared with that which it has previously encountered. This causes degassing of CO2 from the solution and precipitation of CaCO3. Figure 1.4 shows that the difference in c01ue004 between soil and cave is a key control on the quantity of calcium removal from the water, and hence the rate of calcium carbonate precipitation. Growth rate, i.e. upward extension rate in the case of stalagmites, can be of the order of a millimetre per year in humid, warm regions, but is more typically less than 100 µm per year in cool temperate regions. Precipitation can happen above the site where the observer is based and such prior calcite precipitation leads to reduced oversaturation and can be identified by a characteristic evolution of water composition. Long-term continuous growth as slow as 1 mm per thousand years has been documented from weakly oversaturated waters.

Speleothem growth requires quite specific conditions of availability of water to supply the ingredients of growth and circulation of air to take away the waste product carbon dioxide. We use the metaphor of the speleothem incubator in this book to describe this life-support system that maintains speleothem growth. Although conditions in cave interiors are much more constant than above ground, they are not completely static. Temperature varies near entrances, the humidity and carbon dioxide content of the air change laterally and through the year, and the quantity of infiltrating water is a function of the passing seasons. These changing conditions normally impart an annual visible or chemical lamination within a speleothem if it grows quickly enough to be resolved. This lamination contains information about the seasonality of the climate during deposition. The commonest type is an annual couplet (Fig. 1.5a), reflecting warmer–cooler or more usually wetter–drier alternations, but a discrete thin impulse lamina (Fig. 1.5b), characteristic of seasonal influx of soil-derived material, is common in cool temperate climates.

Fig. 1.5 (a) Sectioned speleothem that has grown around a bottle (black) in Proumeyssac Cave, Dordogne, France. Annual growth couplets in calcite represent alternations of clear calcite (darker) and calcite containing fluid inclusions (lighter). (b) Plate 7.2, thin section of the top of stalagmite Obi84, Obir Cave, Austria, showing growth from late 1998 to the end of 2002 by bright annual impulse laminae (plane polarized light). Black area lower left is air-filled inclusion and growth zones show that this occupies a depression on the growth surface. Hiatus surface (hi) represents a brief in-year pause in growth. Inset shows a scanning electron microscope view of the crystal growth surfaces (after Fairchild et al., 2010).

1.1.4 How Do We Date Them?

A time series of observations is of limited use unless we can assign real ages to it. In the case of modern, actively accumulating speleothems, their rate of growth can be determined by direct observation, e.g. on human artefacts (Fig. 1.5a). Another successful technique on these materials is to demonstrate a distinct signal within the stalagmite of enhanced levels of radiocarbon (¹⁴C) resulting from atmospheric nuclear tests of the 1960s: this has been done for the sample illustrated in Fig. 1.5b (Smith et al., 2009). In both cases, these observations show that the growth lamination displayed is annual and so the duration of growth of older speleothems formed in the same setting can be derived by counting laminae. However, it would be unwise to rely solely on such a method because hiatuses in growth may not have been detected, and in any case growth may not have continued up to the present day. Hence normally an absolute dating method is used which allows assignment of a speleothem layer to a particular calendar age. The most common methods are radiometric, that is they rely on decay of a radioactive species from a defined starting point. By far the most commonly used method for samples between a few hundred and a few hundred thousand years in age is the uranium-thorium disequilibrium method, but a variety of radiometric and other methods can be used as a check or to extend the dating range to millions of years.

The development of techniques for precise and accurate U–Th dating on small amounts of sample is the single most important factor that has allowed speleothem science to become so prominent in recent years (Edwards et al., 1987; Hoffmann et al., 2007). Work that used to be laboriously undertaken on tens of grams of sample in the 1980s can now be performed on milligrams following the successive development of thermal ionization mass spectrometry (TIMS) and multi-collector inductively coupled plasma mass-spectrometry (MC-ICPMS). The core principle is that as the speleothem grows, it incorporates some uranium from aqueous solution, but fails to incorporate the insoluble element thorium. The nuclide ²³⁰Th accumulates over time, by alpha-decay from ²³⁴U, after that particular speleothem growth layer has been deposited. The half-life (time taken for half the radioactive nuclide to decay) is around 245,000 years and the method can be used to date samples up to around 500,000 years in age. Cheng et al. (2009b) achieved astonishingly good analytical precision (2σ) of 100 years or less on samples over 120,000 years old, which represents the state-of-the-art.

Because each radiometric data refers to one particular growth layer, what is then needed is an age-model: that is a continuous function of age versus distance on the sample. Fig. 1.6 illustrates data from several Chinese stalagmites. Stalagmite DA (Wang et al., 2005) displays nearly linear growth for the past 9000 years, whereas sb10 shows a slightly less regular growth with a hiatus of several thousand years. In both cases, the dates are close together and have sufficiently small errors that there is no ambiguity about the age-model, but in other cases there would a choice of how exactly to draw the lines between the dates (Scholz & Hoffmann, 2011). Sample D4 displays several step-changes in the rate of growth, but growth tends to be linear between these steps. All of these records are perfect in the sense that there are no age-reversals (stratigraphically younger samples with older ages). In practice, many samples are less ideal for dating because they can contain a lot of original detrital Th, as shown by the presence of ²³²Th. Although this can be corrected, it is not always known which ²³⁰Th/²³²Th ratio to use for correction and so errors can be much larger and the data can display age-reversals. Other samples may be difficult to date simply because of low U content. Where annual laminae are present, the optimal strategy is to combine lamina-counting with the U-series chronology.

Fig. 1.6 Age models for some Chinese stalagmites included in the compilations of Plate 1.2 (Dykoski et al., 2005; Wang et al., 2005; Dong et al., 2010). DA shows continuous quasi-linear growth, sb10 illustrates a hiatus, and D4 illustrates data points defining a series of growth periods with differing linear growth rates.

1.1.5 What Are the Proxies for Past Environments and Climates?

Proxies are parameters that can be measured in an archive (e.g. speleothems, ice cores, etc.) and which stand in, or substitute for, an environmental variable (e.g. mean annual temperature, seasonal monsoon intensity, or vegetation type). Examples of speleothem proxy parameters are growth rate, Mg concentration, δ¹⁸O or δ¹³C signature (see Box 5.3 for isotope definitions). In palaeoenvironmental analysis, the behaviour of proxy variables over time is used to interpret the changing environments or climates. In some cases this can be done quantitatively by means of an equation or process known as a transfer function (Fig. 1.7). Fairchild et al. (2006a) attempted to show systematically how the environmental signal becomes encoded in order to improve our understanding of how the proxy signals work. They drew attention to five realms in which signals were generated or modified:

1 Atmosphere (input of energy and matter; e.g. amount or δ¹⁸O composition of rain, temperature variability).

2 Soil and upper epikarst (processes of organic decomposition, mineral dissolution, water flow and mixing; e.g characteristic δ¹³C of vegetation, typical trace element to Ca ratio from bedrock dissolution).

3 Lower epikarst and cave (degassing and prior calcite precipitation, evaporation; e.g. changes in trace element to Ca ratios or δ¹³C signature; influence of solution saturation state on growth rate).

4 CaCO3 precipitation (partitioning or fractionation of elements and isotopes between water and CaCO3 possibly dependent on growth rate or other effects; changes require use of transfer functions to derive water compositions).

5 Secondary change (e.g. change of aragonite to calcite; exchange of ¹⁸O between calcite and water in inclusions).

Fig. 1.7 Preservation of environmental (e.g. climatic) signals in speleothems and the use of transfer functions to inverse model the original environmental condition. An annual cycle of temperature and rainfall variability is used as an example. This is encoded in a speleothem by means of its proxy parameters. At one extreme, there may be a reasonably faithful representation, or a rectified version, of the original signal, whereas at the other, a characteristic constant value of a parameter is produced in the speleothem. Information on the original conditions can be recovered by reversing the coding process; where this is done quantitatively, an equation called a transfer function is used.

In practice, certain parameters in particular contexts have proved especially valuable. The most commonly used proxy is δ¹⁸O (McDermott, 2004; Lachniet, 2009), which has been found to be most powerful in cases where the changing atmospheric δ¹⁸O composition of the original dominates the variation in δ¹⁸O in CaCO3; this is thought to be the case, for example, in the Chinese monsoon records in Plate 1.2. In other cases, the effects of temperature per se on rainfall composition and on the transfer function between water and speleothem composition can be opposed, and complicated by seasonal changes in rainfall composition: here difficulties in interpretation arise because of the local context. Likewise for δ¹³C, there are certain cases where the strong difference in isotopic composition between the aridity-tolerant C4 grasses and C3 vegetation is directly reflected in changing δ¹³C signatures over time in speleothems. On the other hand, there are several other controls on δ¹³C, which has meant that such records are often left uninterpreted. Aridity can also be reflected in covarying δ¹³C, Mg and Sr signatures of speleothems, because of varying importance of prior calcite precipitation (Chapters 5 and 8). However, more generally many different patterns of variation of trace elements and isotopes with each other occur, only some of which are currently interpretable. Some specific signals such as high contents of colloid-transported elements make sense in terms of reflecting high rainfall and infiltration, but they are not expected to yield quantitative rainfall records. It is also worthwhile to mention that growth rate variations, determined from the thickness of annual layers, have been shown to reflect climatic variables in several cases and have been applied particularly prominently to modern climate calibrations and understanding the climatology of the last millennium (see Chapters 10 and 11).

1.1.6 How Do Speleothems Compare with Other Archives?

The properties of Quaternary archives in general are well summarized in several texts (e.g. Bradley, 1999; Battarbee & Binney, 2008). The strengths of the speleothem archives lie in the following characteristics:

1 The common occurrence of continuous episodes of growth, thousands of years in duration, and the preservation of information representing timescales from days up to a million years. This is matched by ice cores, but other archives tend to have strengths either in their high resolution (e.g. tree rings) or long duration (e.g. deep marine sediments).

2 The excellent chronologies that can be obtained by U-series dating. This is far superior to the other proxies in which long records develop, although some other materials (e.g. coral terraces) can also be dated using the same techniques.

3 Speleothems contain several proxy parameters that can be used singly or in combination. Multiple proxies are being developed in numerous other archives too, including peat cores, lacustrine and marine sediments, ice cores, and indeed in individual components, such as foraminifera. It is the ability to record these parameters at very high-resolution which is a particular strength of speleothems and compares with marine coral records and tree rings for example.

4 They are widespread in inhabited continental areas and so their records have a direct relevance to regional climates and environments, and may also contribute to archaeological investigations. A limiting factor here may be the uniqueness of individual stalagmite samples, which makes conservation a concern (see Appendix 1), whereas other continental archives such as ice cores and lake records are typically more aerially extensive, as are flowstones.

5 Speleothems are physically and chemically robust and are relatively protected from erosion. This contrasts with other continental records, which are accordingly often not available before the Holocene. Certain lake records provide the strongest competition in this category and often have complementary information.

The weaknesses of speleothem science, in its current state of development, can be summarized as follows:

1 Insufficient inter-comparisons with other archives in well-constrained (e.g. co-located) contexts in order to improve the robustness of interpretation of both speleothem and other archives.

2 Insufficient understanding of the meaning of proxy variables. Many studies rely on one proxy and do not present data for others. Sometimes interpretations are unduly speculative or too simplistic in expecting a parameter to directly reflect a climate parameter.

1.1.7 What Next for Speleothem Science?

In recent years there has been a rapid advance in analytical techniques to improve sensistivity and accuracy which has had the effect of reducing the destructiveness of sampling. Accordingly, even without further technical developments, there is an enormous amount to do in terms of applying state-of-the-art techniques to existing samples. In addition, there are many research frontiers, including the following:

1. The development of U–Pb methods to extend records throughout the Quaternary and into the Pliocene.

2. A more sophisticated understanding of how speleothem records contribute to understanding climatic drivers and teleconnections of climate.

3. Continued development of new types of proxy and the more effective use of multiproxy approaches.

We come back to these issues at the end of the book.

1.2 How Is this Book Organized?

The first three chapters provide an introduction and context. Chapter 1 sets out our overall approach. The dynamic nature of speleothem-forming environments both on the human timescale and in deep (geological) time makes it attractive to apply the interdisciplinary concepts of system science as an aid to model development, as we discuss in the next section. We set out an explanatory framework for speleothems, firstly by introducing two new terms as metaphors: the speleothem incubator, the place in which speleothem growth is nurtured, and which lies within the speleothem factory, which also contains the delivery system for the raw materials for growth. We lay out the types of change occurring over different timescales and the parameters by which these changes are delivered; our task in this book is to provide a coherent connection between them. Chapter 2 focuses on the development and infilling of caves within human to geological timescales, emphasizing the various issues in carbonate geology. We outline the typical properties of aquifer carbonate rocks and their pores, the diversity of the cave systems, and the place of speleothems in the evolution and ultimate infilling or collapse of the cave. In Chapter 3, we show how the cascading and interlinked systems of climate, atmospheric moisture, soils and vegetation control many of the variables that determine the properties of speleothems. The different features of the climate system that might be recognizable in the past from speleothem studies are illustrated, as are the extent of variability of these climatic modes. We review how the biotic response to climate, modulated by the geological substrates, is reflected in the distinctive physicochemical properties of soils and their associated surface and subsurface ecosystems.

The second part of the book (Chapters 4–6) deal with the transfer processes in karst that characterize its dynamics and composition. Chapter 4 develops in detail the case for thinking of cave and karst processes as a speleothem life-support system: the speleothem incubator. We deal separately with the issues of water, air and heat movement in subsurface karst, illustrating the extent to which these can be understood quantitatively. We summarize by reviewing the integrative property of cave climatology and its influence on speleothem properties. Chapter 5 reviews the inorganic chemical components of the karst environment and its aqueous solutions in relation to speleothem composition. We deal with the fundamentally important carbonate system before considering each chemical variable, most of which have already been applied as palaeoenvironmental proxies when preserved in speleothems. Chapter 6 introduces the relatively neglected issues of the biogeochemistry of karst environments, illustrating the range of organic components that can be mobilized and incorporated into speleothems, and describing what we currently understand to be the important molecular transformations imparted on organic matter and the effects of organic assimilation, breakdown and transport on inorganic species.

The third part of the book focuses on the speleothems themselves, but we restrict attention to calcareous examples because these are overwhelmingly the types that are used in palaeostudies (Calaforra et al. (2008) discussed gypsum speleothems), using other analogues where appropriate. Chapter 7, on the architecture of speleothems, first discusses the fundamental controls on their shape and internal structure leading to a geometrical classification. A synthesis of their internal mineralogical, crystallographic and geometric structures sets an agenda for what needs to be explained in a given sample. Chapter 8 gives an overview of the geochemistry of speleothems, starting with the macroscopic setting, moving to the details of the water-speleothem interface and systematically describing the fractionations that occur between water and calcium carbonate during speleothem formation. This leads to two types of conceptual model: those that reflect patterns of seasonal change and those that reveal change in space within the cave environment. Chapter 9 on dating is particularly critical because a significant competitive advantage of speleothems in relation to other palaeoclimate proxies arises from one’s ability to determine the absolute age of formation of a given speleothem layer. We deal systematically with the radiometric methods (those that depend on decay of specific radioactive isotopes), palaeomagnetism and annual-layer counting, before concluding with an analysis of age-depth models and uncertainties.

The most pressing argument for developing speleothem studies is their capability for contributing to palaeoscience (Roberts, 1998; Bradley, 1999; Batterbee & Binney, 2008), i.e. for recording information on past environments and climates, with an unparalleled combination of age-resolution and range of environmental proxies. Further, they are located in continental regions whose future climatic evolution is the subject of intense scrutiny. Study of the past boosts our confidence in forecasting (Skinner, 2008), particularly providing data for testing the general circulation models (GCMs) and other models used for prediction of future climate. We explore these aspects in the final section of this book, comprising Chapters 10–12. Chapter 10 tackles issues of calibration and validation of speleothem climatic proxies using young speleothems that have formed within the era of instrumental meteorological measurements. This is a rapidly developing field in which a range of approaches from related proxy materials are being used and there is a currently a major international effort to identify the key locations which can answer important climatological questions. In Chapter 11, we focus on the post-glacial Holocene Epoch, summarizing the range of relatively subtle climate drivers, key target time periods, quasi-periodic signals and model comparisons that have been investigated, with a particularly detailed focus on the last millennium. The Holocene is a period of immense archaeological interest as well as potentially one where disturbance by human activity is a key influence. Speleothems have already provided some intriguing pieces in the jigsaw, as well as more complete and meaningful records in some regions. Finally, in Chapter 12, we look at the longer Pleistocene Period (and beyond) in which the major Cenozoic glacial–interglacial cycles have occurred as has been documented primarily through study of ocean cores, ice cores, and some long lake and loess sections. Here speleothem records have been demonstrated in particular regions to be excellent integrators of aspects of hemispheric or global climate and have been used to refine timescales used by workers on other Quaternary proxies.

1.3 Concepts and Approaches of System Science

Systems science covers a diverse range of approaches to understanding the workings of natural and anthropogenic environments (Kump et al., 2009). Understanding the formation of speleothems requires a range of approaches, and systems thinking has two roles to play. Firstly, a qualitative systems analysis is useful to clarify our thinking about relationships of factors that may influence speleothem formation. Secondly, a quantitative approach allows us to test our understanding by building models that simulate aspects of the processes that lead to speleothem growth and initial steps in this direction are discussed in Chapter 10.

The science of systems has diverse origins, but threads derived from the mathematical modelling of regulatory processes (cybernetics, Wiener, 1948) and from organismal biology (von Bertalanffy, 1950, 1968) were highly influential in the post-war period. The detailed concepts of system theory were subsequently applied to geomorphic environments by the influential geographers Chorley and Kennedy (1971), contributing to geography’s ‘quantitative revolution’, and were further developed by Bennett and Chorley (1978), although with no reference to karst.

Ford and Williams (1989, 2007) identified the development of structured networks of caves and drainage networks in unconfined karst as a case of a cascading system, which Huggett (1985) defined as ‘interconnected pathways of transport of energy or matter, or both, together with such storages of energy and matter as may be required’. Cascading systems are amenable to box-modelling (Box 1.1). In addition, we observe the development of morphological forms of defined geometry, which also show self-organization (the spontaneous development of ordered structures, such as calcareous deposits developing cascades of pools with rims). We might further aim to build a process–response model (Chorley & Kennedy, 1971) that brings together the quantitative mass or energy flow information with morphological parameters to give a more sophisticated understanding of the system.

Box 1.1 Box models and feedback

This box contains some basic material on flows of matter within simplified systems. Kump et al. (2009) give an accessible introduction, whereas Rodhe (1992) and Chameides and Perdue (1997) provide a more rigorous mathematical treatment.

A reservoir (also known as a box) contains matter (mass M) and can be considered homogeneous.

A flux (F) is the amount of material transferred from one reservoir to another per unit time.

Many systems are approximately linear, that is Fi (the flux or fluxes from reservoir i) is directly proportional to Mi, the mass in reservoir i (volume Vi can also be used instead of Mi for fluids at constant temperature and pressure).

If all relevant fluxes and masses remain constant over time, the system must be in kinetic or dynamic balance, and is said to be in steady-state. Under this condition, the underlying kinetic constants can be readily calculated.

A rate constant can be defined for each flux as follows:

(I) c01e001

where Fij refers to the flux from reservoir i to reservoir j and kij is the corresponding rate constant.

(II) c01e002

where τij is the residence time, corresponding to the mean time that a given molecule resides in reservoir i before it is removed to reservoir j.

A relevant example is provided in Fig. 1.B1 by the comparison of carbon reservoirs and fluxes from the pre-industrial period (black), when the system was presumed to be in steady-state (black) with the changing situation in the 1990s (grey). Here the residence time (lifetime) of fossil fuels equals the reservoir size (3700 Gt C) divided by the current usage rate (6.4 Gt yr−1), i.e. 580 years.

Fig. 1.B1 The carbon cycle (IPCC, 2007) showing reservoir (boxes) and fluxes (arrows). Figures in black are from the pre-industrial era whereas those in grey (red) are the changes resulting from human activity in the Anthropocene.

However, interpreting residence times when there are multiple fluxes out of a reservoir (i.e. multiple sinks) needs more careful discussion. For example, radioactive isotopes such as ¹⁴C produced during atmospheric nuclear testing in the 1950s and early 1960s have provided important insights into element cycling. Half of the excess radiocarbon (¹⁴C) disappeared in 10 years (Fig. 3.24). This observation has been used by climate-change deniers as evidence for the short lifetime of carbon in the atmosphere. Actually the turnover time of a reservoir (Rodhe, 1992), the time it would take to empty the atmosphere of ¹⁴C if the sinks remained constant and the sources were zero, is even shorter. In Fig. 1.B1, using the figures (in black) from the steady-state pre-industrial period:

Atmospheric carbon turnover time = mass/total outwards fluxes = 597/(0.2 + 120 + 70) = 3.1 yr.

The key point here is that ¹⁴C is temporarily stored in vegetation and the surface ocean and is rapidly returned to the atmosphere, so that a more meaningful figure for the residence time of carbon is given by the ratio of the atmospheric mass (597 Gt C) to the ‘permanent’ sink in sediments (0.2 Gt C yr−1), i.e. around 2500 years. Such a long period is the reason that our current carbon emissions are a cause for concern.

If the system is then perturbed by a change in the sources or sinks of matter, the rate constants can be used to estimate how the system will change. Using the pre-industrial figures and applying eqn. (I), the rate constant for transfer of carbon from the atmosphere to the surface ocean is 120/597 = 0.201 yr−1. Given the enhanced mass of atmospheric C typical of the 1990s (597 + 165 = 762 Gt), application of this rate constant suggests that the flux of C to the surface ocean should then be 762 × 0.201 = 153 Gt C yr−1. In fact the figure is just 92.2 (70 + 22.2), implying that more complex types of behaviour need to be considered, e.g. by more detailed consideration of the processes involved (Rodhe, 1992; see also the carbonate system in Chapter 5). Nevertheless, the approach clearly indicates the sense of change that is expected.

The above case example is from (large-scale) Earth system science, but there are analogous examples in terms of temporary sediment storage in geomorphic systems and in terms of carbon storage in soils for speleothem studies (see Chapter 8). Changing carbon dioxide concentrations in air can also be used to provide a good example of how to apply linear systems theory in a more local context. An experiment was conducted in which an investigator entered a previously well-ventilated empty office, closed the door and window, and monitored the changing carbon dioxide levels. The pattern of change is shown in Fig. 1.B2, increasing from an initial value close to that of the external atmosphere at around 400 ppm, but clearly flattening off after 3 hours towards a new steady-state value. This can be rationalized in terms of a box model (inset in Fig. 1.B2) with the person respiring CO2 as reservoir 1, and the office (reservoir 2) exchanging air with the infinite external reservoir 3.

Fig. 1.B2 Results of an experiment to illustrate the concept of a system evolving to a new steady-state after a change in boundary conditions (see text for explanation).

The mass of CO2 in the office rises at the following rate:

(III) c01e003

It can be seen that the negative feedback that results in a stabilization of the carbon dioxide level is an inherent part of the structure of the system: as the concentration of CO2 in the office rises, it is more efficiently removed in each parcel of air that exchanges with the exterior. This is a perfect linear system where the mass of carbon dioxide removed to the exterior is directly proportional to the mass in the room, because the rate of air exchange is constant. Ultimately a new steady state is reached where dM2/dt is zero:

(IV) c01e004

F12 can be estimated from general adult physiology based on the typical CO2 content (3.6%) and volume (8 litres) of exhaled breath per minute. With V2 = 32,000 litres, this would lead to an initial increase in c01ue014 of 9 ppm min−1 whereas the actual observed figure was 7 ppm min−1 (the subject was a slow breather!). The rate of air exchange through the window frame was not known at the start of the experiment, but can be calculated as 420 litres min−1 by simulating the experiment stepwise on a spreadsheet (see www.speleothemscience.info) and adjusting parameters to fit the observed curve. Hence the approach allows a complete quantification.

Mass balance approaches have many applications in geomorphic and geochemical studies. One key example in cave environments is the CO2 level in cave air because this directly control rates of speleothem growth, but normally varies seasonally. Figure 1.B3 illustrates the main fluxes of CO2 in Ernesto, a small Italian cave (Frisia et al., 2011). The rate of decay of transient high CO2 levels from visitors was used to estimate cave air residence times and exchange rates with the external atmosphere during steady-state summer conditions. The exchange times were used to demonstrate a large flux of carbon dioxide from air-filled fissures in the surrounding epikarst, much larger than that which could be released from dripwater. When considering the transition from summer to winter conditions, additional information is needed because all the fluxes are liable to change with changing temperatures and aquifer conditions. Although the system does not remain linear, the system structure remains unchanged, and enables quantitative solutions to be derived (Faimon et al., 2006; Kowalczk & Froelich, 2010; Frisia et al., 2011). A key feedback turns out to be a reduction in epikarst CO2 flux because of an increase in water saturation of the aquifer in winter. These issues are explored further in Chapters 4, 5 and 8.

Fig. 1.B3 Structure of box model for carbon dioxide fluxes

derived from a study of Ernesto Cave, Italy (Frisia et al., 2011).

F, flux; V, volume; C, concentration.

A central issue in system science is the interactive nature of system components and in particular the occurrence of feedback behaviour, whether positive (reinforcing the applied change) or negative (resisting the change). Box 1.1 takes the well-known example of the carbon cycle to illustrate feedback concepts, showing how the issues can be downscaled to caves.

A variety of system concepts that refer to the nature of change over time (Chorley and Kennedy, 1971; White et al., 1992; Phillips, 2009) are illustrated in Box 1.1 and Fig. 1.8, and will be applied to karst in the next section. A system parameter may show an evolution over time, in which case it is described as non-stationary (Fig. 1.8a). Alternatively it may be stationary, that is, having a consistent mean value averaged over the period of interest. Within this period, variation can be periodic (cyclic) as shown in Fig. 1.8b, or may be subject to non-periodic fluctuations. Cyclicity can be studied by a variety of statistical techniques that are introduced in Chapter 10. In Fig. 1.8b, the sequence is deterministic as the cycles are all exactly the same, whereas in nature there is a random component and hence the systems are non-deterministic or stochastic.

Fig. 1.8 Types of system behaviour. Diagrams (a)–(d) and (f) illustrate changes in a system property over time. In (a) there are a series of transient forcings; in (b) there is an implied cyclic forcing (although internal system dynamics can also give rise to this type of behaviour); in (c) the behaviour illustrates the addition of random forcings in a system with some memory effects; in (d) and (f) there is a single transient forcing with very different results. Diagram (e) illustrates the concepts of threshold and sensitivity in relation to a persistent forcing. See text for further explanation.

Where there is memory of previous system states, stochastic processes can give rise to apparent structured trends (Fig. 1.8c) without a corresponding systematic driving force. This is a really important concept for palaeoscientists because there is a strong temptation to over-interpret plots of variables against time (e.g. Fig. 1.8d), e.g. to consider all ‘wiggles’ on plots of past variables as meaningful (Fairchild et al., 2006a). The property is referred to as autocorrelation and statistical models can be created to describe it. Such models have an order which represents the number of previous states influencing the current state (Huggett, 1985). For example, in Fig. 1.8c, the upper line indicates a sequence of random numbers Y1, Y2, Y3 … (within the range 0–1) whereas the lower (spiky) line is a random walk representing an example of an autocorrelative function (Z) of order 1 that is related to the random series Y at time t as follows:

(1.1)

Figure 1.8c shows that apparent systematic variations are observed with the same degree of noise as the random number series. In real systems, it may be a major research task to distinguish purely random from systematic behaviour.

Stationary systems show a degree of resilience to a time-limited disturbance, as illustrated in Fig. 1.8d where the original system property is eventually recovered. An external forcing leads to system change after a reaction time and the system behaviour responds over a characteristic period known as the relaxation time. This contrasts with that shown in Fig. 1.8a where the system’s response to a series of transient forcing events is to move closer to a threshold beyond which it rapidly moves to a different state.

In the geomorphic literature, the term sensitivity has been related to the probability that a persistent change will result from a disturbance (Brunsden and Thornes, 1979; Phillips, 2009); that is, a sensitive system (Fig. 1.8e) is opposite to the resilient system shown in Fig. 1.8d. In physical science literature more generally, sensitivity usually refers to a gradient between a system property and a long-lasting forcing function (Fig. 1.8f).

The behaviour of a system over time could show simple linearly scaled responses to disturbances (e.g. double rainfall produces double discharge), but real systems are more likely to be complex where exact physical solutions are unavailable: all we can then do is to ascribe probabilities for behaviour. A special case of complex systems are those described as chaotic, where there is hypersensitivity to the initial conditions such that the final result varies wildly and cannot be predicted (Fig. 1.8e). Phenomena driven by fluid turbulence, such as weather systems are of this type. Chaos theory and its relative catastrophe theory deal with nonlinear dynamics, that is, mathematically they refer to systems that cannot be described by systems of linear equations. Whereas a non-chaotic system can have a specific defined state (a defined solution to a mathematical equation) that can be described as an attractor, in chaos theory the (strange) attractors are fuzzy. For example, climate is the (strange) attractor for weather.

Catastrophe theory deals with system discontinuities, or bifurcations, where a small additional forcing can lead to a major change in system state that can be difficult to reverse such as the modelled behaviour of oceanic circulation in the Atlantic (Rahmstorf, 1995). The growth history of many speleothems reveals irreversible changes.

Currently, the dominant reference to systems in environmental science is Earth system science (ESS), which seeks to understand and predict the future changes in the interactions between the different major components of the Earth (geosphere, biosphere, atmosphere, hydrosphere and cryosphere), together with human interactions (Ehlers & Krafft, 2001). ESS makes most sense in the context of large-scale phenomena that ideally can be sensed remotely and which are sufficiently well understood as to be amenable to quantitative simulation: global biogeochemical cycling is an example (Chameides & Perdue, 1997). In this field, it is common to use industrial metaphors to describe the workings of the Earth, for example the text of Raiswell et al. (1980): Environmental Chemistry: The Earth–Air–Water Factory. The development of improved GCMs for climate underpins much large-scale research in ESS and is also highly relevant as a context for the interpretation of speleothem palaeoenvironmental research, as discussed in Chapters 10–12. To the extent that speleothems may contain encoded information on temperature, rainfall and vegetation, and are available over much of the land area of the Earth, they can also contribute to the testing of Earth system models, in which biogeochemical modules supplement the earlier purely physicochemical content of earlier models.

However, system concepts also apply at the local scale: what Richards and Clifford (2008) refer to playfully as LESS (local environmental systems science). Identifying the behaviour of individual geomorphic systems, their process-interactions, feedbacks and emergent properties provides a holistic basis for site-specific and generic studies of particular environments. As we will see, this turns out to be an enormously important insight for speleothem science.

1.4 The Speleothem Factory Within the Karst System

In this book, we propose a nested three-fold conceptual scheme (Fig. 1.9) in terms of the following:

1 The karst system as a whole where the emphasis is on its role as an aquifer and hence is on a much larger spatial scale than that relevant for speleothems.

2 A broader speleothem factory which operates as a cascading system, supplying the raw ingredients from the epikarst, soil-ecosystem, and adjoining atmosphere and transferring them to the cave environment. We develop this analogy further in this section.

3 A speleothem life-support system which we term the speleothem incubator. This consists of those parts of the karst and cave environment that maintain the growth of speleothems during a ‘lifetime’ when they are active. This maintenance includes the regulation of air and water flows and is described in detail in Chapter 4.

Fig. 1.9 The nested domains of (1) the karst system whose primary drive is to maximize the efficient transport of water; (2) the speleothem factory which supplies raw materials and generates speleothems; and (3) the speleothem incubator which is the cave and adjacent karst environment which maintain the conditions for speleothem growth.

Klimchouk & Ford (2000a) define the karst