Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Multi-proxy studies of climate variability in central China: Subdecadal to centennial records in stalagmite from Budda Cave
(USC Thesis Other)
Multi-proxy studies of climate variability in central China: Subdecadal to centennial records in stalagmite from Budda Cave
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
INFORMATION TO U SERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. ProQuest Information and Learning 300 North Zeeb Road, Ann Arbor, M l 48106-1346 USA 800-521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. M ulti-proxy studies of climate variability in Central China: Subdecadal to Centennial records in stalagm ite from Budda Cave. by Dorte Eide Paulsen A Thesis Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree MASTER OF SCIENCE (GEOLOGICAL SCIENCES) August 2000 Copyright 2000 Dorte Eide Paulsen Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 1405251 ___ ® UMI UMI Microform 1405251 Copyright 2001 by Bell & Howell Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. Bell & Howell Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA The Graduate School U niversity Park LOS ANGET.ES, CALIFORNIA 90089-1695 Thi s t hesi s, wri t t en by ZDQPLTC E lD g 1>aULSENf___________ Under the di r ect i on o f h s xL . . T h e s i s Commi t t ee, a n d approved by all its m e m b e r s , has been present ed to and accepted by T h e Graduat e S c h o o l , i n partial ful fi l l ment o f requi rement s for t he degree o f MASTER. OF 5Q E W C E _______________ Dean o f Graduate Studies D a te A u g u s t 4 , 2 0 0 0 _______ T H E S I S C O MMI T T E E Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Acknowledgements This thesis work contributes to the study of paleoclimate variability in China. The work was done at the Department of Earth Sciences, University of Southern California. The research was funded in parts by the Department of Earth Science at USC and under a US National Science Foundation grant entitled “High resolution Climatic Variability in Eastern China during the Last Two Millennia” (ATM-9818407). The constructive comments and thorough review by my advisors Dr. Teh-Lung Ku and Dr. Hong-Chun Li are much appreciated. Dr. Lowell D. Stott and Dr. Douglas E. Hammond served on the Thesis Committee and provided many constructive comments. Dr. Hong-Chun Li collected the stalagmite SF-1 that was used in this work. The stable isotope measurements were made in the laboratory of Dr. Lowell D. Stott to whom I am most grateful. The trace element analysis was performed at the XRAL Laboratories, Ontario, Canada. Finally, I would like to thank Dr. Carl Christian Liebe and Riley Duren for reviewing this thesis for spelling and grammatical eiTors. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table of Contents Acknowledgements ......................................................................................................... ii List of Tables....................................................................................................................vi List of Figures..................................................................................................................vii Abstract.............................................................................................................................x 1 . Introduction....................................................................................................................I 1.1. Discussion of Climatic setting in Central China..................................... 5 1.2. Structure of this thesis................................................................................ 8 2. Chronology of stalagmite SF-1................................................................................. 11 2.1. Annual Nature of Lamination Deposition................................................ 13 2.2. Lamination counting based on transmission of visible light....................14 2.3. Lamination Counting Utilizing Luminescence.........................................28 2.4. Growth rate determination based on lamination counting......................31 2.5. Summary...................................................................................................... 34 3. Paleoclimate Variability from Lamination Counting.............................................. 36 3.1. The basic mechanism of carbonate deposition..........................................37 3.2. Climatic factors controlling calcite deposition.........................................41 3.3. Instrumental records of precipitation and temperature in the vicinity of Budda cave in central China..........................................................................41 3.4. Correlation between climatic factors and the calcite precipitation of SF-1...................................................................................................................... 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3.5 Correlation between laminae thickness and the combined temperature and precipitation data..................................................................... 55 3.6. Discussion..................................................................................................... 56 3.7. Summary....................................................................................................... 59 4. Paleoclimate Variability from Stable Isotope Records............................................ 61 4.1. 8 IjC as a Paleoclimatic Indicator................................................................65 4.2. Sl80 as a Paleoclimatic Indicator................................................................69 4.3. Isotopic Equilibrium Deposition of Calcite with Drip W ater..................78 4.4. Climatic Interpretation of the Isotopic Records of SF-1 (0-16.1 ka): Centennial resolution...........................................................................................85 4.5. Climatic Interpretation of the Isotopic Records of SF-1 (0-1300a): Subdecadal resolution..........................................................................................96 4.6. Summary........................................................................................................103 5. Paleoclimate Variability from Trace Element Records.............................................106 5.1. Parameters controlling trace element incorporation into stalagmite calcite.....................................................................................................................107 5.2. Factors controlling the Mg/Ca and Sr/Ca ratios of drip water................. I l l 5.3. The climatic interpretation scheme for the Mg/Ca, Sr/Ca and Mg/Sr ratios in stalagmite SF-1.......................................................................................116 5.4. Correlations between the Mg/Ca, Sr/Ca and Mg/Sr ratios....................... 118 5.5. The records of Mg/Ca, Sr/Ca and Mg/Sr from stalagmite SF-1 and their climatic interpretation................................................................................. 1 2 0 5.6. Summary........................................................................................................127 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 . Summary.................. 129 References........................................................................................................................... 135 Appendix Appendix 1 A. Photos from the Budda Cave area............................................. 142 Appendix 2A. 2 1 0 Pb dating...................................................................................143 Appendix 2B. Determination of the 2 1 0PbtO tai activity from the a-counting o f 2 0 9 Po and 2l0P o ..................................................................................................148 Appendix 2C. TIMS 2 3 0Th/2 3 4 U dating of SF-1................................................ 151 Appendix 2D. Manual for the microscope, the digital camera and the Adobe software.............................................................................................. 152 Appendix 2E. Matlab Script used for identifying dark bands..........................153 Appendix 2F. Manual for the Luminoscope..................................................... 154 Appendix 3 A. Bilinear regression.......................................................................156 Appendix 4A. Using the micro miller to acquire samples of a stalagmite.... 161 Appendix 4B. Matlab script to calculate power spectrum............................... 166 Appendix 4C. Power spectra for the 8 l3C and Sl80 records............................167 Appendix 5A. Temperature and rainfall indices...............................................169 Appendix 5B. XRD spectra for SF-1................................................................. 170 V Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List of Tables Table 3.1. Station data......................................................................................................43 Table 3.2. R 2 values for the temperature correlations.................................................. 52 Table 3.3. R 2 values for the precipitation correlations................................................. 55 Table 3.4. R2 values for the temperature/precipitation correlation with lamina thickness.............................................................................................................................56 Table 4.1. Interpretation of the 5 1 3Cc signal.................................................................. 69 Table 4.2. S1 8 Oc interpretations.......................................................................................78 Table 4.3. Calculated 5l3Cc values based on different soil/limestone and C 3/C4 plant distributions..................................................................................................84 Table 4.4. Periods identified in the power spectra.......................................................101 Table 5.1. Typical molar ratios of Mg/Ca, Sr/Ca and Mg/Sr in natural reservoirs... 113 Table 5.2. Climate interpretation of changing Mg/Sr, Sr/Ca and Mg/Ca ratios.......118 Table 5.3. Temperature and humidity index for trace element data.............................122 Table 5.4. Temperature and humidity index based on stable isotope data..................123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List o f Figures Figure 1.1. Inside the Budda Cave................................................................................. 4 Figure 1.2. Stalagmite SF-1 in growth position........................................................... 4 Figure 1.3. Surface pressure, pressure systems and winds over Asia..........................7 Figure 2.1. The thin section of stalagmite SF-1............................................................15 Figure 2.2. The Zeiss Axiotech microscope and the SPOT color digital camera setup...................................................................................................................................16 Figure 2.3. The spatial positions of the tracks............................................................... IT Figure 2.4. Microscope image of the youngest part of track 1 (lowest resolution).. IS Figure 2.5. Microscope image of the youngest part of track 2 (lowest resolution).. 19 Figure 2.6. The monochrome version of Figure 2 .4 .................................................... 20 Figure 2.7. The vertical projection of Figure 2.6................................................ 21 Figure 2.8. The gradient of Figure 2 .7 ...........................................................................22 Figure 2.9. The identified local minima (stars) in Figure 2 .6......................................23 Figure 2.10. Microscope image showing part of track 1 (lowest resolution).............24 Figure 2.11. Microscope image showing part of track 2 (lowest resolution).............25 Figure 2.12. Microscope image showing part of track 2 (highest resolution)............26 Figure 2.13. Identified banded patterns from track 2 (highest resolution)................. 2 1 Figure 2.14. The Olympus luminoscope laboratory setup...........................................29 Figure 2.15. Correlation between band thickness along trackl and track2................ 32 Figure 2.16. Growth rate for track 1...............................................................................33 Figure 2.17. Depth/age model for the first 50 years of track 1 ....................................34 vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.1. The mechanisms of stalagmite and stalactite deposition..........................40 Figure 3.2. Map of the Budda cave location and the weather stations........................42 Figure 3.3. Contour plot of A) the temperature (°C) and B) the precipitation (mm/y) in the proximity of the Budda C ave..................................................................45 Figure 3.4. Average monthly precipitation at Ankang, Xian and Lanzhou................46 Figure 3.5. Comparison between instrumental data from Ankang and Lanzhou...... 46 Figure 3.6. Interpolated instrumental record at the Budda Cave................................. 48 Figure 3.7. Record of lamina thickness for Stalagmite SF-1.......................................49 Figure 3.8. Interpolated annual average temperature at the Budda cave and lamination thickness.......................................................................................................... 50 Figure 3.9. The correlation between average annual temperature at the interpolated Budda cave Station and lamination thickness.......................................... 51 Figure 3.10. The interpolated total annual precipitation and lamination thickness .. 53 Figure 3.11. Correlation between the total annual precipitation and laminae Thickness............................................................................................................................ 54 Figure 4.1. a) Single-layer analysis of 5 1 8Oc in stalagmite SF-1 and b) S1 3Cc versus 5 1 8Oc .......................................................................................................................80 Figure 4.2. 8 1 3Cc of carbon species involved in calcite precipitation.........................82 Figure 4.3. Isotopic composition of stalagmite SF-1.................................................... 8 6 Figure 4.4. Power spectrum for the 8 lsO record............................................................8 8 Figure 4.5. Power spectrum for the S1 3 C record.............................................................8 8 Figure 4.6. Red noise spectra for different a ................................................................. 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.7. Subdecadal isotope record for stalagmite S F -1 ......................................... 97 Figure 4.8. Reconstructed climate records......................................................................100 Figure 5.1. Rain water history.......................................................................................... 112 Figure 5.2. Correlation between Mg/Ca and Sr/Ca........................................................119 Figure 5.3. Correlation between Mg/Ca and Mg/Sr.......................................................120 Figure 5.4. Molar ratios of Mg/Ca, Sr/Ca and M g/Sr....................................................122 Figure 5.5. The temperature and rainfall indices based on trace elements and isotope data for 0-5000a....................................................................................................125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Abstract Paleoclimate during the past 16,100 years in central China was determined from a high- resolution study of growth rate, stable isotopic composition, and trace elements in a stalagmite from Budda Cave. Lamination counting was performed for the last -150 years. No correlation was found between growth rate and instrumental records of precipitation and temperature. The interpretation of the 8 I3C and 8 lsO records defined several climatic periods: a dry period centered around 7ka, the Medieval Warm period, the Little Ice Age, and the coldest historically recorded decade within the last 500 years. Climatic cycles with periods of 1567, 33, 24, 12, 9.6, and 7.2 years were also identified. A simplistic model to interpret varying Mg/Sr, Sr/Ca, and Mg/Ca ratios in terms of climate variability was consistent with the isotopic interpretations. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1. Introduction Any investigation on what causes climate changes or any assessment of the impact of antropogenic activities on the climate requires records of past climate. Reconstructing the paleoclimates enhances our appreciation of natural climate variability and may enable the construction of climate models to simulate potential variability of future climates. Very few instrumental records extend more than a few hundred years back in time. High-resolution paleoclimate proxies provide details about climate beyond the instrumental records. Paleoclimate proxies are natural phenomena that form over time and incorporate information about changing climate conditions in their structure. An important characteristic of paleoclimate records is that they can be dated fairly accurately, they span over much longer times than instrumental records, and they can be resolved with high resolution. Previous paleoclimate reconstructions have utilized a wide variety of paleoclimate records involving ice cores, ocean sediment cores, lacustrine deposits, loess, tree rings, pollen, insects, corals, and speleothems. Archived in ice cores are information about paleoprecipitation rates, surface air temperatures, atmospheric composition and past variation in solar activity. This information can be extracted from the stable isotopic composition (SI8 0 ) of the ice, the gaseous content of air bubbles in the ice, the amount of dissolved and particulate matter in the fim and ice, and the physical characteristics of the fim and ice (Bradley, 1999). Ocean sediment cores can be used to infer changes in the sea surface temperature (SST), the salinity, the surface water productivity and the oceanic circulation system. This information is obtained l Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. from the study of the Sl 8 0 and trace element content in tests of fossil marine fauna. Lacustrine deposits provide information about paleoclimatic conditions through the studies of lake-level variations. Variation in lake-level may be inferred from the study of the stratigraphy of the lacustrine deposits and from grain size analyses, measurements of organic and inorganic geochemical contents, and the magnetic susceptibility of the sediment. Loess records may provide knowledge about variations in past rainfall based on measurements of magnetic susceptibility. Furthermore, studies of the grain size distribution of the loess will reflect atmospheric circulation (and wind speed). Studies of the width and density of tree rings (and the cells within the rings) are used to obtain records of paleotemperature and paleohydrology (wet and dry periods). Pollen records are constructed based on the assemblages of pollen in stratigraphic sequences. Based on these records interpretations of past changes in temperature, precipitation and vegetation can be made. Changing climatic conditions can also be inferred from the distribution of insect remains in stratigraphic sequences. Corals can be used to infer changes in sea level, sea surface temperature, rainfall, river runoff and ocean and atmospheric circulation based on trace elements and stable isotopes incorporated in their skeletons. Historical records are also frequently used to acquire paleoclimatic information (Bradley, 1999). Stalagmites (and other speleothems) are another natural climate recorder used in studies of the terrestrial paleoclimate. Limestone caves containing speleothems are located at many places around the world. Stalagmites have been found with continuous records spanning 500ka. Utilizing current technology (microsampling and accurate 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. dating methods), it is possible to date individual stalagmite layers to a stunning one-year resolution. Several different proxy signals of past climate are contained within a stalagmite. These proxy signals include isotopes, trace elements, organic materials and optical properties. Often, several climatic and environmental factors influence an individual proxy signal, and it may be difficult to isolate the individual factors based on a single proxy. However, using a combination of more than one proxy can result in a more reliable climatic interpretation. In this thesis work, a multi-proxy study of paleoclimate variability has been performed on a stalagmite (SF-1) from Central China. The stalagmite, SF-1 was collected from the Budda Cave (33°40? N, 109°05’E) located at the southern flank of the Qin Ling Mountains approximately 80 km south of Xian at an altitude of -500 meters. The Budda cave formed in carbonate rock of late-Pleistocene age about 50 m below the present land surface. The carbonate bedrock is covered by a 6 - 8 meter thick Quaternary loess deposits of which the upper 3-5 meters are well-developed into soil. The present land surface is covered with vegetation (refer to Appendix A1 for pictures of the Budda Cave area). Figure 1.1 shows a picture of the inside of Budda Cave. For a map of the location of the cave, refer to Chapter 3. The stalagmite SF-1 grew -200 meters from the cave entrance and was in a growth position when it was collected in 1996. The stalagmite has been dated to 16,100 years B.P. Figure 1.2 depicts SF-1 in growth position at the time of collection. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 1.1. Inside the Budda Cave. Figure 1.2. Stalagmite SF-1 in growth position Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The objective of this work was to investigate the centennial to subdecadal paleoclimatic signals archived in the Budda cave stalagmite using stable isotopes, trace elements chemistry and variations in growth rates. More specifically, an attempt was made to reconstruct high-resolution records of paleotemperature and paleoprecipitation. This work will provide important information about the paleoclimate and cyclicity in paleoclimate in Central China. The resolution of this record is unprecedented for this area. It is anticipated that the knowledge gained from this work can be combined with records from other sites to increase our understanding of the monsoonal circulation systems in East Asia, their teleconnections to other climate systems and potential forcing factors. This kind of data is also essential for constructing climate models for the region. 1.1 Discussion of Climatic setting in Central China. Close to 20% of the world’s human population lives in China. It is in this area that the largest death tolls due to climatic disasters have occurred. An estimated 10-13 million people died in the 1877-1878 famine in China as a result of drought. Flooding of the Huang He River has repeatedly decimated the Chinese population (Bryant, 1997). Historic documents show that 7 million people drowned when the river flooded in 1332. 10 million people died subsequently from famine and diseases. The latest major floodings were in 1887 and 1938, each resulting in 1 million lost lives. Additionally, an estimated 11 million people died in 1938 from flood related famine (Bryant, 1997). Clearly the ability to predict floods, droughts, frost, typhoons, etc. in this densely 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. populated area is crucial. Climate predictability is also of great importance to the agriculture in this area. Three monsoon systems have a strong influence on the climatic conditions within China. There are three seasonal wind circulation systems, the equatorial Indian (summer) monsoon, the East Asian (summer) monsoon, and the Siberian-Mongolian (winter) monsoon. Variations in the strength of the summer and winter monsoons control the amount of precipitation and cause changes in the temperature. The rainfall associated with the summer monsoons is the main source of moisture to China and is crucial for sustaining life in this densely populated (agricultural) region. Failure of the summer monsoons can result in drought and disastrous famines (hunger), whereas strong summer monsoons can cause catastrophic flooding. The timing and the strength of the winter monsoon, however, control the onset and the duration of the dry season. The monsoons (seasonal wind) are caused by the difference in heating and cooling of the ocean and land in the summer and the winter months. In its simplest form, the monsoon systems behave as regional sea breezes in the summer and regional land breezes in the winter. During the summer months, the Asian continent warms up and low-pressure systems are formed which cause uplift and instability in the lower atmosphere. The ocean/sea adjacent to the continent (the South China Sea, the Bay of Bengal, the Yellow Sea and the West Pacific) are relatively cooler (because of the higher heat capacity of water compared to land). Hence, high-pressure cells are formed above the water and centers of convection (monsoon low) are formed over land. As the pressure gradient intensifies, warm moist air from the oceans/seas is drawn onto land from a 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. southwesterly direction in the Indian Ocean (the Indian Monsoon system) and from a southeasterly direction in the Pacific Ocean (the East Asian Monsoon system). Weather data have shown that the migration of the North Pacific Subtropical High plays the most important role in the advance (northward movement) and retreat (southward movement) of the summer monsoon (Li and Ku, unpublished data). During winter, the continent cools, while the oceans/seas tend to remain warmer. A significant high-pressure cell (the Siberian High) is formed by the subsidence of air above the land and this produces outflow of cold, dry air off the Asian continent to the low-pressure systems that form above the ocean. The wind direction is reversed, and the air-masses flow from the continent toward the ocean (in the Siberian-Mongolian Monsoon). Figure 1.3 shows the average surface pressure, pressure systems, and the resulting wind over Asia for a) January and b) July. a) January b) July Figure 1.3. Surface pressure, pressure systems and winds over Asia. (Anthes, 1997). The monsoon systems have a strong influence on the temperature and the precipitation. A stronger Siberian High, for example, causes a stronger winter monsoon that brings about a windier and colder winter. A stronger North Pacific Subtropical 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. High causes a stronger East Asian M onsoon and a wetter-than-normal year (Li and Ku, unpublished data). Temperature and precirpitation records are therefore, a direct result of variability in monsoons strength. 1.2 Structure of this thesis In this thesis, the paleoclimatic reconstruction based on stalagmite SF-I will be described. The report consists of four m ajor chapters. Chapter 2 describes how the chromology of stalagmite SF-1 was established. A combination of data obtained from three dating methods: TIMS U-series dating, 2I0Pb dating, and lamination counting were msed for the chronology. The TIMS U-series dating and 2l0Pb dating had been performed previously by Li and Ku (unpublished data), and their results are described briefly. The chronology based on the average growth rate assumption from TIMS 2 3 0T h /2 3 4 U and 2 1 0 Pb dating are not accurate enough to reveal very short time cycles (<10 years). To investigate the high frequency fluctuations of the isotope signals (Chapter 4) for the last 150 years (where a resolution of one year was possible) lamination counting was performed. The lamination counting was based on transmission of visual light through a thin section. Counting based on luminescence was also attempted. The details of visual lamination counting will be discussed. The objective of Chapter 3 is to investigate if there is a correlation between climatic factors (precipitation and temperature) and the growth rate of stalagmite SF-1. A thorough description of the processes that govern calcite deposition is given. Instrumental data for the region have been extracted from two databases. Based on these 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. data sets, different records related to temperature and precipitation have been calculated for the Budda Cave site. These records are compared to the growth rate of the stalagmite. Chapter 4 discusses the factors that control the 51 3 C and 8 lsO isotopic composition of a stalagmite and their relationship to paleoclimate. Before records of 5i3C and 5lsO are used for climate interpretations, it must be determined whether the stalagmite calcite deposited in stable isotopic equilibrium. If the stalagmite deposited in calcite isotopic equilibrium a relatively simple climatic interpretation may be possible. The precipitation of calcite in isotopic disequilibrium does not negate the use of the isotope records for paleoclimatic reconstructions. It does however, complicate the interpretation that rely on isotopic variability, hi this study it will be investigated whether stalagmite SF-1 was deposited in isotopic equilibrium. Two records were sampled from stalagmite SF-1 and analyzed for their isotopic compositions. One series (-16.1 ka record) provides centennial resolution, the other (-1.3 ka record) subdecadal resolution. The results of the isotopic records will be presented and a climatic interpretation will be discussed. Spectral analysis will be utilized to identify cycles in the records. The identified cycles will be compared to other paleoclimatic studies from this region. Chapter 5 investigates the potential use of trace elements contents in stalagmites as recorders of paleoclimate. A simplified model that explains variations in ratios of Mg/Sr, Sr/Ca and Mg/Ca in terms of climate variability will be constructed. This model will be applied to a record of similar length and resolution as the centennial isotope 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. record. Interpretations based on the trace elements contents will be compared to the interpretations based on stable isotopes. Chapter 6 will summarize this thesis work. 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2. Chronology of stalagmite SF-1 Accurate time scales are essential in all paleoclimatic studies - especially when the goal is to identify periodic patterns in the records or correlate certain events between records. Stalagmites (and other drip stones) can be dated in several ways, including 2 1 0Pb-dating, TIMS 2 3 0Th/2 3 4 U dating, and lamination counting. 2 1 0Pb-dating is limited to relatively young samples (<100 yr) and determines the average growth rate of a stalagmite (Baskaran and Iliffe, 1993; Li et al. 1996 and Tanahara et al., 1998). This method was used on 12 samples from the most recently deposited part of stalagmite SF-1 by Li and Ku (unpublished data). Appendices 2A and 2B explain the details. The results of Li and Ku (unpublished data) showed that SF-1 was in an active growth position when it was collected and grew with an average growth rate of ~ 0.066 ± 0.005 mm/yr at the time of collection. TIMS 2 3 0Th/2 3 4 U dating can be applied to samples with ages of 0-500ka (Li et al., 1989). A total of 5 samples from SF-1 were dated by Li and Ku (unpublished data) at Prof. R. L. Edwards’ laboratory, University of Minnesota (refer to Appendix 2C). The TIMS dating showed that the total age of SF-1 was 16.1 ka and a growth hiatus existed from -9.4 -16.0 ka. The TIMS dating also showed that the growth rates were: -0.089 mm/year in the upper part (0-1.6 cm), -0.0163 mm/yr in the middle part (1.6- 16.5 cm) and -0.72 mm/yr in the lower part (16.5-22.2 cm) of the stalagmite. The average growth rates determined by the TIMS 2j0Th/2 3 4 U dating and the 2 1 0 Pb dating will provide the chronology for the isotope and trace element records with 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. centennial resolution described in chapter 4 and 5 as well as the 1266 years isotope record with sub-decadal resolution (-3 years). One of the objectives of this thesis work was to obtain a very high-resolution stable isotope record, corresponding to 1 sample/year for the last 150 years. For this record, the chronology based on the average growth assumption from TIMS 2 3 0Th/2 3 4 U and 2 1 0 Pb dating would not be accurate enough to reveal short time cycles. This is because the growth rate of a stalagmite varies between years as a function of the climatic and environmental factors (Railsback et al., 1994; Shopov et al., 1994). For the part of the stalagmite SF-1 with very high-resolution (~1 year), lamination counting was used as a complementary method. When examining a stalagmite with a microscope (and possibly under UV light), a pattern of alternating dark and bright bands may appear. Lamination counting in stalagmites is based on the assumption that bands are deposited annually as the stalagmite grows. Banding is the result of varying amounts of clay and organic matter in the drip water source or varying amounts of dust in the cave atmosphere during the year. The annual bands in a stalagmite may be detected using either luminescence or transmission of visible light through a thin section. The spacing between the individual bands can be measured and utilized to construct a high precision depth/age relationship. This relationship can be used as a high-resolution chronostratigraphic tie between the radiometric dating and the high-resolution records (e.g. records of S 1 3C, 5l80 and trace elements) obtained from speleothems. This chapter includes a discussion of the annual nature of laminae deposition within SF-1 and a detailed description of the lamination counting performed on SF-1 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. using visual transmission. Lamination counting based on luminescence has also been attempted. Finally, the determination of the growth rate variability for the last ~ 50 years of growth and a construction of a depth/age relationship based on the width of the individual annual laminae has been made. 2.1 Annual Nature of Lamination Deposition The underlying assumption of lamination counting is that the lamination deposition process in speleothems is of annual nature. Many authors have previously reported that the deposition of layers within stalagmites are annual (Baker et al., 1993 and 1998; Baker and Smart, 1996; Broecker et al., 1960; Li et al., 1996; Railsback et al., 1994; Shopov et al., 1994) and others. Stalagmite SF-1 was sampled in the Budda Cave (109°05’E, 33°40’N) south of Xian in Central China. This area experiences highly seasonal rainfall. According to data from the IAEA/WHO global network database, approximately 75% of the precipitation that falls in this area occurs between April and September (NOAA, 2000). Highly seasonal climate is, according to Broecker et al. (1960), most likely a requirement for defined annual layering. This strengthens the assumption of the annual lamination deposition pattern in SF-1. Baker et al. (1993) postulate that annual banding may be limited and that it depends strongly on the type of water flow feeding the speleothem. Soil water and unsaturated zone storage need to be adequate to maintain flow during the summer, while winter residence times remain sufficiently short that summer and winter waters are not completely mixed. 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. To determine whether the observed bands in the stalagmite SF-1 were deposited annually, a comparison between the average growth rate determined from lamination counting (assuming that the laminae are annually deposited) and the average growth rate previously determined by Li and Ku (unpublished data) were made. Li and Ku (unpublished data) used radiometric 2I0Pb dating to calculate the average growth rate for the upper 100 years. The results of the comparison will be discussed in the conclusion of this chapter. It will be shown that the agreement between average growth rates from the two methods is very good. It is therefore assumed in the following that the lamination bands in SF-1 are annual in nature. 2.2 Lamination counting based on transmission of visible light Many speleothems exhibit varvelike sub-millimeter-scale bands of varying light transmission. These variations can be observed under visible light. The bands are the result of the presence of variable amounts of clay and organic matter, which coprecipitated or adsorbed onto the surface of the stalagmite from drip waters that passed through the soil on their way to the cave or in some cases dust in the cave atmosphere that adsorbed onto the stalagmite surface (Shopov et al., 1994). The thin section of stalagmite SF-1 that was used for lamination counting is shown in Figure 2.1. This thin section has an approximate thickness of 0.5 mm and was cut along the main growth axis. Burnham Petrographies, Monrovia, California, prepared the thin section that had a full length of 233 mm. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25 Scale in cm Figure 2.1. The thin section of stalagmite SF-1. The thin section was illuminate*! from the back and the laminae were observed under a standard Zeiss Axiotech microscope. A SPOT color digital camera was connected to the microscope and digital photos were acquired using the Adobe PhotoShop computer software. Refer to Appendix 2D for instructions on how to use the microscope, the digital camera and the software. Figure 2.2 shows the laboratory setup. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.2. The Zeiss Axiotech microscope and the SPOT color digital camera setup. Visual observation showed that no single track1 had banding continuously throughout the full length of the stalagmite. The tracks were often interrupted by the presence of small breaks and distortions in the calcite deposit that may be due to the cutting of the stalagmite. Consequently, several tracks were utilized to obtain a continuous uninterrupted record. Information is extracted from different tracks and the information is fused to construct an uninterrupted record. Digital images for the upper =4.2-mm of the stalagmite were acquired along the four tracks shown in Figure 2.3. Images with two different resolutions were acquired. An image covers either an area of 1.6mm x 2.4mm or an area of 0.3mm x 0.45mm. The reason for using two different 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. resolutions will be discussed later. The emphasis was focused on track 1, along which the samples for the isotope data were taken. Track 2 is placed at a distance of about =10mm from track 1. This track was selected because it had the longest uninterrupted sequence of bands. Figure 2.3. The spatial positions of the tracks. Samples for 51 3 C and SI8 0 records (Chapter 4) were acquired along track 1. Figure 2.4 shows a microscope image of the youngest part of track 1. Visual counting of the black bands is a relatively trivial process for human eye. The black band identification was performed independently by two test persons to check for reproducibility. The identified bands are shown as lines. The stalagmite SF-1 was in growth position when collected in 1996, so that the uppermost layer presumably precipitated during the year 1995-1996. Each successive band has been numbered one year less than the previous year. In Figure 2.5 is shown a image mosaic of the youngest part of track 2 , using the same procedure applied to track 1 . 1 A track is defined as a line in the growth direction, perpendicular to the growth layers. 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.4. Microscope image of the youngest part of track 1 (lowest resolution). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. v ; Figure 2.5. Microscope image Of the youngest part of track 2 (lowest resolution). 19 Reproduced with permission A computer program was written and used as a reference to identify growth bands. This was achieved with the following procedures. A picture consists of a grid of small picture elements called pixels. The pictures from the SPOT digital camera are in color. This means each picture consists of 3 different images (representing red, green and blue) each with 383 x 255 pixels. Three different images are acquired while the camera mechanically rotates a red, green and blue filter in front of the camera lens. The 3 images are used to control the 3 individual intensities (red, green and blue) of each pixel. Since no information was available about the spectral characteristics of the camera filters it was decided to transform the color image into a monochrome image. This was simply done by summing the 3 different images, and displaying the result as a monochrome image. The monochrome version of Figure 2.4 is shown in Figure 2.6. 20 40 60 C/2 O X 80 'c L 100 120 140 50 TOO 150 2 00 250 300 3 5 0 pixels Figure 2.6. The monochrome version of Figure 2.4 (series 2). The image has been cropped. 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Since the bands are nearly vertical and aligned with the pixels, it was possible to project the image onto the x-axis. Basically this meant adding all pixel values in the same columns. The result is shown in Figure 2.7, which is the vertical projection of the image. x 10* 3.5 2.5 0.5 200 400 0 50 100 150 250 300 350 pixels Figure 2.7. The vertical projection of Figure 2.6. A high value represents a bright vertical line and a low value represents a dark vertical line. The centers of the black bands were then located. The center of a black band is characterized by the location where the brightness function is first decreasing and then increasing. Therefore the local minima of the brightness function were sought. The first step in locating these minima was to calculate the gradient of the brightness function (dBrigthness/dx). The gradient of the vertical projection (Figure 2.7) is shown in Figure 2.8 . 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10000 8000 6000 4000 C A C/3 S J C 2000 CO •2000 400 300 350 100 150 200 250 0 50 d(pixels) Figure 2.8. The gradient of Figure 2.7. In the gradient domain, a local minimum is characterized by a gradient w ith a negative value followed by a gradient with a positive value. More specifically, the centers of the black bands are located where the gradient crosses the x-axis w ith a positive slope. The zero crossings in Figure 2.8 are identified in Figure 2.9 w ith stars imposed on the picture of Figure 2.6. It should be noted that the procedure described above requires that the lamination bands are vertical. Otherwise, the image needs, to be rotated. A Matlab script that does the band identification is shown in Appendix 2E. Unfortunately, not all microscope images show good contrasts like the ima_ges in Figure 2.4 and Figure 2.5. Figures 2.10 and Figure 2.11 show the continuation of track 1 and track 2. It is observed that there is a large area with insufficient contrast and magnification. It was not possible to do lamination counting in this region based o n the low-resolution images. To do the lamination counting in the region with little contrast, 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. additional high microscope images were acquired of track 2. This is shown in Figure 2.12. High-resolution lamination counting was performed on the high-resolution image mosaic shown in Figure 2.12. This was done independently by two test persons. The result is shown in Figure 2.13. SO 100 150 200 250 300 350 pixels Figure 2.9. The identified local minima (stars) in Figure 2.6. They are to be compared to the manually identified black bands in Figure 2.4. 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. S c a le i n mm Figure 2.10. Microscope image showing part of track 1 (lowest resolution) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.11. Microscope image showing part of track 2 (lowest resolution) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 9 7 8 7 7 76-? Figure 2.12. Microscope image showing part of track 2 (highest resolution). 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Track 2 - Series 3 Figure 2.13. Identified banded patterns from track 2 (highest resolution) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.3 Lamination Counting Utilizing Luminescence Many speleothems show strong luminescence when exposed to ultraviolet (UV) light. Luminescence of calcite seems to be caused by excitation of organic matter (salts of humic and fulvic acid) that are occluded in the growing speleothem (Baker and Smart, 1996). The amount of dissolved fulvic acid in the speleothem feed water varies with the seasons during the year. Fulvic acid is produced in photosynthesis and released through plant roots. It dissolves readily in the ground water and the level is anticipated to be high during the growing season and lower the rest of the year (Shopov et al., 1994). It is therefore likely that the fulvic acid causes an annual pattern that is recorded in the speleothem. Humic acid, a product of organic decomposition, however, dissolves slowly and is therefore believed to provide longer cyclic components to the banded nature of some speleothems. Baker et al. (1993) and Shopov et al. (1994) have previously reported luminescence bands that were parallel to the visible color banding and annual in nature. According to Shopov et al. (1994) it is possible to resolve luminescence banding to a much finer degree than the visible banding. Therefore it may be particularly useful to look for luminescence instead of visible banding when growth rates are slow and when there are high concentrations of mineral matter that obscure the annual signal from organic matter variations (Baker et al., 1998). However, Baker et al. (1993) observed that luminescence banding only occurred in a few samples (5 out of 43 sampled). They proposed that this might be due to the use of a mercury lamp source, which had a low-energy and broad spectrum output, and therefore might not have been powerful enough to excite weak luminescence banding. 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. To study potential luminescence in stalagmite SF-1 an Olympus BX60 luminoscope was used. An impinging electron beam (up to 30 KeV) excited the stalagmite SF-1 sample and the luminescence light emitted by the relaxation to lower energy states was observed in the attached microscope. Although parts of the sample showed more florescence than other parts of the sample no banded pattern was observed. Single grains showed up. A photograph of the Olympus luminoscope is shown in Figure 2.14. For instructions on how to use the luminoscope refer to Appendix 2F. Figure 2.14. The Olympus luminoscope laboratory setup. The Olympus luminoscope utilized electrons to excite the sample. This was different from the methods used by Baker et al. (1993, 1996 and 1998) and Shopov et 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. al. (1994) where energetic photons (UV) were used to excite the atoms. The problem with using electrons in lamination counting is the insufficient penetration depth of the electrons in the material. When electrons traverse in the sample, they will interact through Coulomb forces with millions of other electrons and nuclei in the sample. This means that the electron continuously looses energy and makes small deflections in its trajectory in the material. The energy that the electron loses per traversed distance is called the Linear Energy Transfer (LET). The LET for a 30 KeV electron in silicon is 17.5 MeV/cm and increases at lower energies (Silicon is very well characterized due to its use in semiconductors). A conservative estimate of the LET of a 30 KeV electron in calcite is 10 MeV/cm. This means that the radiation length is 30 KeV/lOMeV/cm = 30 microns. The real number is most likely less than 10 microns due to the fact that the LET increases as the electron loses energy. The luminescence observed in the luminoscope is therefore only generated in the upper 1 0 microns. Baker et al. (1993, 1996 and 1998) and Shopov et al. (1994) used back (or front) illumination with UV light. The wavelengths of the photons were typically in the 320- 420 nm range. Photons with this energy penetrate 1 mm to 5 mm thick samples (Shopov et al., 1994). Baker et al. C1993 and 1998) used a Zeiss Axiotech microscope with a mercury and a neon light source, respectively, to excite luminescence whereas Baker et al. (1996) scanned their sample in front of an UV laser beam (HeCd, 325 nm). In all of these methods luminescence was generated from the bulk samples. The bulk sample is considered to be composed of many thin layers, all of which carry information about the lamination. The signal-to-noise ratio between dark band and light band is proportional to the square root of the number of layers (the square root of 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the thickness of the sample). This means that the signal contrast with UV illumination might be 1 0 to 2 0 times better than using charged particles. If luminescence counting is to be explored further at USC, it is suggested to procure some accessories to the Zeiss Axiotech microscope. The current configuration of the Zeiss Axiotech is that the sample is illuminated from the back with visible light (black body radiator). This light has an insignificant component in the desired UV region. However, it is possible to attach a bright UV light source to the microscope to excite luminescence. A high pass filter should be added to the UV light source so that only light with wavelength less than 420nm would get into the sample. The microscope optics should also include a band pass filter so that only light from the luminescence would be detected in the microscope. Zeiss has a comprehensive homepage discussing these techniques (Zeiss, 2000). 2.4 Growth rate determination based on lamination counting It was assumed that the distance between two successive dark bands represents the annual growth of the sample (annual growth rate). Distances between the lines in the previous figures (Figure 2.4, Figure 2.5 and Figure 2.13) were measured. Unfortunately, there was not enough information in track 1 to do the lamination counting for the last 50 years. Therefore, lamination counting was performed on track 2, and the information was projected onto track 1 . To justify the projection of lamination counting from one track to another track, the correlation between track 1 and track 2 was investigated for the last 17 years. Figure 2.15 shows a plot of band thickness for track 1 and track 2 as functions of depth. 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0.25 C orrelation coefficient R = 0.89 T rack 2 | 0 .1 5 ■ « M e c J C u c < 8 m 0.1 • 0 .05 - T ra c k ! 1 992 1994 1988 1990 1996 1 9 8 2 1986 1978 1980 1984 Calendar year Figure 2.15. Correlation between band thickness along track 1 and track 2. In Figure 2.15 it is observed that the thicknesses of bands in track 2 are approximately 42% larger than the bandwidths in track 1, but that the variation in bandwidths in track 2 correlates well with the variation in bandwidths in track 1 (the correlation coefficient R is 0.89). It is therefore reasonable to project growth rates from different tracks, as long as the absolute overall distances are preserved. The final result of the lamination counting for the latest 50 years for track 1 is shown in Figure 2.16. As seen in Figure 2.16, there are significant variations in the growth rate. The growth rate varies a factor ~5 from 0.029 mm/year to 0.138 mm/year. 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0.25 >. 0.15 ■ E o a 0 .0 5 - a v e ra g e: 0.0828 m m /yr 0 1940 1950 196G) 1970 calendar year 198 0 1990 2000 Figure 2.16. Growth rate for track 1. The growth rate corresponding to the year 1996 is believed to be an artifact of being the last layer in the counting. The average growth rate determined from the lamination counting of track 1 is 0.0828 mm/year. This rate is consistent with the average growth rate of 0.066 ± agreement between the growth rates confirms the annual nature of the layers. Figure 2.17 shows the high precision depth/age relationship that has been constructed for track 1. It is observed that utilizing a constant growth rate assumption may introduce errors. In this example offsets as large as 5 years will be introduced on a record length on 20 years. For fuiture work it is therefore critical to use lamination counting as the high precision chronological control when climate variability on annual scales is to be obtained from stalagmites. 0.005mm/year determined by Li armd Ku (unpublished data) based on 2l0Pb dating. The 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. calendar yaar 0 1 2 4 average growth rate: 0.096 rmVyr 5 6 Figure 2.17. Depth/age model for the first 50 years of track 1. 2.5 Summary Accurate chronological control is essential in high-resolution paleoclimate studies and it is believed that lamination counting may provide chronology with annual accuracy. M ost references claim that deposition of annual banding frequently occurs in areas that experience highly seasonal rainfall. The Budda Cave in Central China is situated in such an area. It was therefore likely that the observed banding in the stalagmite was annually deposited. Banding can be measured in two different ways, either by visual inspection of darker bands under visible light or by detection of luminescence. Luminescence is typically excited by illumination with UV light. Unfortunately, no UV light source was 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. available for this work. It was therefore attempted to excite the calcite sample with electrons, but the penetration depth in the sample by electrons was insufficient to produce measurable luminescence. The lamination counting was performed visually with a Zeiss Axiotech microscope along several tracks, with two countings done in each track. A computer program was developed and used as a reference to detect dark bands. A depth/age relationship was constructed for track 1 (a track normal to the growth layers at a position where high resolution sampling of 5I3C and 5I80 has been performed). The lamination counting discussed in this chapter dates 50 years. The average growth rate for track 1 was determined to be 0.084 mm/yr. This is consistent with radiometric growth rate determination: 0.066 ± 0.005 mm/year previously determined by Li and Ku (unpublished data). The agreement between the growth rates confirms the annual nature of the layers. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3. Paleoclimate Variability from Lamination Counting In the previous chapter, lamination counting for stalagmite SF-1 was used to establish a chronology with an annual resolution to A.D. 1947. This chapter will explore lamination counting further. Using visual lamination counting, the chronology has been extended back to A.D. 1853. The deposition of calcite layers in cave deposits is a complex process that depends on different climatic and environmental factors. This makes paleoclimatic interpretations that are based on records of laminae thickness difficult. Typically, it is very complicated to determine which factors are most significant in controlling the calcite deposition at a specific site. Variation in layer thickness has been interpreted in different ways in the literature (Baker et al., 1993 and 1998; Ming et al., 1997; Railsback et al., 1994; Shopov et al., 1994; Tungsheng et al., 1997). Ming et al. (1997) and Tungsheng et al. (1997) compared a record of laminae thickness in a stalagmite from NE China with historical drought and flood records from the last 500 years and found a high correlation in the data. Low stalagmite growth rates matched years of drought and high stalagmite growth rates matched years of flood. Railsback et al. (1994) measured the thickness of composite annual layers of aragonite and calcite from a stalagmite collected from Drotsky’s cave in the northern Botswana and compared it with instrumental records of temperature and rainfall. They found that the individual aragonite and calcite layers correlated with mean summer temperature and annual rainfall respectively. Baker et al. (1998) also investigated the relationship between growth rate and drip water rate for four stalagmites from the Brown’s Folly Mine, 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wiltshire, England but did not find good correlation in the data. Baker et al., 1998 observed that stalagmite growth rates are most sensitive to variations in the Ca2 + concentration in the drip water, which in turn depends on several factors including soil temperature, soil moisture condition, soil depth, vegetation type above the cave and bedrock purity. Finally, Shopov et al. (1994) have argued that calcite growth rate is controlled by changes in the rate of drip water as well as changes in soil productivity (soil Pco2 ) as the latter will affect the degree of supersaturation of the drip water and thus the calcite deposition rate. In this chapter the theoretical model for carbonate deposition developed by Dreybrodt (1980) and Buhmann and Dreybrodt (1985) will be described briefly. This model predicts that calcite deposition is indirectly controlled by climatic factors such as temperature and precipitation rate. Hence, lamination thickness is correlated with temperature and precipitation rate, and the potential exists for using stalagmites to reconstruct past climate. Similar to the work of Railsback et al. (1994), the potential for using lamination counting as a recorder of paleotemperature and paleoprecipitation in Central China will be assessed by comparing the variations in band thickness of the uppermost layers of stalagmite SF-1 with instrumental records of temperature and precipitation. 3.1 The basic mechanism of carbonate deposition. Figure 3.1 depicts the basic physical and chemical processes that lead to the deposition of calcium carbonate on the surface of a stalagmite. As rainwater (that has equilibrated with atmospheric CO2 ) percolates through the soil horizon at the Earth’s 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. surface, it absorbs CO2 that has been generated in the soil by decomposition of organic matter. The partial pressure of CO 2 in the soil can reach high values, Pco2.soii ~ 0.1 atm (Dreybrodt, 1980), compared to the partial pressure of CO2 in the atmosphere, Pco2 . a tm ~ 0.0003 atm. By absorbing CO2 the water becomes acidic and will effectively dissolve limestone on its way from the soil horizon through cracks and fissures in the bedrock to the cave below. As the water (still having a relatively high partial pressure of CO 2) enters the cave and drips on the stalagmite surface, most of the water splashes quickly off the stalagmite. The remaining water forms a thin stagnant film on top of the stalagmite. Because of the difference in partial pressure of CO 2 of the cave atmosphere and the CO 2 dissolved in the water in the film, CO2 will diffuse to the surface of the film where it degasses. The film solution becomes supersaturated with respect to CaCC>3 and CaCC>3 starts to precipitate and the stalagmite grows on the floor of the cave. Stalactites (drip stones hanging from the cave ceiling) are formed by deposition of CaCC>3 from similar film solutions residing at the ceiling. The chemical reactions involved are also shown in Figure 3.1. Dreybrodt (1980) investigated the mechanisms of carbonate deposition and developed a more complicated theoretical model than the one described above to predict the growth rate of a stalagmite. This model was later revised and expanded by Buhmann and Dreybrodt (1985). Their model includes three independent rate-determining processes: ( 1 ) the kinetics of surface reactions (the rate of dissolution and precipitation) at the stalagmite calcite surface as a function of the activities of the species Ca2+ , Hfi, HCO3' and H2CO3 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (2) the kinetics of the reaction CO2 + H 2 O <-» H 2CO3 H* + HCO 3' (3) the molecular diffusion of the dissolved species Ca2+ , HCO 3 ', CO3 2', CO2 and H2 CO3 across the phase boundaries (between the CaCC>3 and the water film, and between the water film and the cave atmosphere, respectively, as shown in Figure 3.1). Buhmann and Dreybrodt (1985) found that precipitation rates (R) can be approximated by the linear function: R = a (C — Ceq) (3.1.1) where a is a kinetic constant [cm-s'1 ] that depends on the cave temperature, drip water Pco2 > the thickness of the water film covering the CaCC>3 surface, and hydrodynamic conditions of the flow (laminar versus turbulent flow), a increases with increasing temperature, with increasing drip water Pco2 and with increasing film thickness. Rates of deposition for turbulent flow might be an order of magnitude higher than for laminar ■ 7 flow. C is the calcium concentration in the water film [mmol-cm’ ], and Ceq is the equilibrium concentration of calcium [mmol-cm' ] with respect to calcite. Ceq depends on the temperature of the cave air: the higher the temperature, the lower the value of C eq. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chemical reactions: Figure 3.1. The mechanisms of stalagmite and stalactite deposition. 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3.2 Climatic factors controlling calcite deposition The described theoretical model predicts that calcite deposition is indirectly controlled by climatic factors such as temperature and rainfall rate. Higher cave temperatures will cause the chemical reaction rates to be faster and CaCC>3 to be deposited more rapidly on the stalagmite. Higher cave temperatures will also result in increased evaporation from the water film solution. The film solution becomes supersaturated in CaCC> 3 and more CaCC>3 will precipitate than at lower temperatures. The amount of rainfall controls the hydrodynamic flow conditions and affects the amount of dissolved species Ca2+ , HCO 3 ' CO32 ', CO2 and H 2CO 3 . Increasing precipitation rates will cause faster drip rates and thicker film solutions on top of the stalagmite. A thicker film will cause calcite to precipitate more rapidly. Lamination thickness is potentially correlated with changes in temperature and precipitation rates, and it may be possible to use stalagmites to reconstruct past climate, based on the growth history of the stalagmites. Previous studies have shown a correlation with temperature and/or precipitation and used it to reconstruct climate variability in the past (Ming et al, 1997; Railsback et al., 1994; Tungsheng et al., 1997). The correlation between these two climate factors and the annually deposited layers in stalagmite SF-1 from the Budda Cave in central China will be investigated below. 3.3 Instrumental records of precipitation and temperature in the vicinity of Budda cave in central China. The Budda cave (33°40’N, 109°05’E) is located on the south flank of the Qin Ling Mountains, at an altitude of approximately 500 m and about 80 km south of Xi’an 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. in central China, as depicted in Figure 3.2. Unfortunately, no instrumental records (of significant duration) of precipitation and temperature exist in the very close proximity of the cave. However, records from 21 stations within a radial distance of ~ 500 km from the Budda cave are available from the software program “World WeatherDisc” (WeatherDisc Associates, 1994) and the internet database “Climate Data Bases of the People’s Republic of China for 1841-1988 CDIAC TR055”, (CEOS IDN datasets, 1995). Beijing . Tianjin Baotou Yinchuan Taiyuan iini Xining Jinan 36 Zhengzhoj izhou Xian 34 - - Nanjing Hefei # Anqing Suzhou 32 ■ ■ Wuhan Chengdu Hanzhou 30 ■■ Chongqing Changde Nanchang Zunyi 28 ■■ Changsha 1 1 0 118 120 116 1 1 2 102 104 106 108 114 [°E] Figure 3.2. Map of the Budda cave location and the weather stations. The gray stars indicate station locations — the white star indicates the cave site. Stations extracted from these sources are shown in Table 3.1 and Figure 3.2. These two databases contain the most comprehensive instrumental climate data sets for China. 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The entire data set for each station was used to calculate a mean station temperature and a mean station precipitation, also shown in Table 3.1. These values are plotted as a function of their geographical position in Figure 3.3 A) and B). Figures 3.3 A) and B) are made as contour plots, so it is easy to perceive spatial variations in temperature and precipitation. Table 3.1. Station data. Station Altitude [m] Av. Temp. [°C] Ann. Prec. [mm] Dist.to cave [km] 1. Yanchi (37°47’N, 107°24’E) 1347.8 7.8 342.3 483 2. Yanan (36°36’N, 109°30’E) 957.6 9.4 565.3 329 3. Pingliang (35°33’N, 106°40’E) 1346.6 8 . 8 517.6 305 4. Qingyang X (35°44’N, 107°38’E) 1421.9 8 . 6 575.3 266 5. Yuncheng (35°02’N, l l l o01’E) 376.0 13.9 530.2 234 6 . Lushi (34°03’N, 111°02’E) 568.8 1 2 . 6 648.1 185 7. Luoyang (34°40’N, 112°25’E) 154.5 14.6 597.6 327 8 . Nan yang (33°02’N, 112°35’E) 129.2 14.9 785.1 333 9. Wan yuan (32°04’N, 108°02’E) 674.0 14.7 1 2 0 0 . 6 2 0 2 10. Ankang (32°43’N, 109°02’E) 290.8 15.6 796.7 105 11. Yunxian (32°51’N, 110°49’E) 201.9 16.2 825.1 185 12. Laohekou (32°23’N, 111°40’E) 90.0 15.3 848.2 280 13. Bazhong (31°51’N, 106°46’E) 360.0 17.0 1144.8 295 14. Zhongxiang (31°10’N, 112°34’E) 65.8 15.9 934.8 429 15. Enshi (30°17’N, 109°28’E) 457.1 16.3 1439.4 377 16. His-Ning*} (36°36’N, 109°48’E) 2295 6.9 376.0 333 17. Lanzhou * (36°06’N, 103°54’E) 1518 9.4 322.0 545 18. Tien-Shui** (34°36’N, 105°36’E) 1 2 0 2 1 1 . 6 550.3 337 19. Xian** (34°18’N, 108°54’E) 398 13.8 583.0 73 20. Hanzhong** (33°00’N, 107°00’E) 508 14.3 903.9 206 21. Yichang** (30°42’N, 111°06’E) 131 17.0 1128.6 380 indicates that the data are from the WWDisc, otherwise data is from the CEOS-IDN. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It is observed in Figure 3.3 A) that the annual average temperatures at the stations in the SE comer of Figure 3.2 are higher than those in the NW comer. This is primarily because the former stations are located at lower latitudes. Inspecting the altitudes of the stations (also listed in Table 3.1) reveals that the stations in the SE comer lie at lower altitudes which also contributes to the higher temperatures. The southeasterly winds in the East Asian (Summer) Monsoon are responsible for the majority of the rainfall in this area (Li and Ku, unpublished data). The stations in the SE comer of Figure 3.3 B) generally receive more precipitation than the stations in the NW comer. This is because the former stations are located earlier on the monsoon system’s pathway (the farther from the water source of the Pacific and South China Sea, the more precipitation has rained out) and is also due to the orogenic effect. The influence of the monsoon and its associated heavy rainfall in the months April through October, with peak values from July to September, is present throughout the area. Refer to Figure 3.4 for plots of average monthly precipitation at Ankang, Xian and Lanzhou. The goal of the instrumental data is to have a long and accurate record of temperature and precipitation at the cave site. The existing data (although the most comprehensive) are scattered in location and covering different and relatively short time intervals. It would be tempting to use the longest record, from the Lanzhou station, as instrumental record for the Budda Cave. However, when comparing the records from Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Lanzhou with Ankang (the latter is the closest station to the Budda Cave but temporally short), little correlation is found between the two, as shown in Figure 3.5. 39 Taiyuan Yindiuan 38 37 36 Zhengzhou Lanzhou 35 34 33 32 Chengdu 31 Chongqing 30 Changde 29 108 109 114 103 104 105 107 H O 111 112 113 lQ E| 39 Taiyuan Yinchuan 38 37 36 Zhengzhou Lanzhou 35 Xian 600 34 Budda cave 33 32 Chengdu 31 30 Changde 29 103 105 107 108 109 110 11 2 113 114 106 II I [°E1 Figure 3.3. Contour plot of A) the temperature (°C) and B) the precipitation (mm/y) in the proximity of the Budda Cave. 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 6 0 ■A nkang El Xian S L anzhou Month no. Figure 3.4. Average monthly precipitation at Ankang, Xian and Lanzhou. 1 8 1 6 1 4 ■ 12 - 10 8 • 6 1 9 2 0 L a n z h o u 1 9 3 0 A n k an g \ a v . : 1 5 . 6 C L a n zh o u a v .: 9 .4 C A n k a n g av .: 7 9 6 .7 m m /y r 1 9 4 0 a v .: 3 2 2 .0 m m /yr 1 9 5 0 1960 C a le n d a r y e a r 1 9 7 0 1 9 8 0 1990 Figure 3.5. Comparison between instrumental data from Ankang and Lanzhou. 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Total annual precipitation (mm At times the records are anticorrelated. Therefore, to generate the longest and most reliable record, a bi-cubic interpolation was constructed. The advantage of using a bi-cubic interpolation is twofold. Primarily, it is possible to generate a data value, whenever 3 weather stations are forming a triangle around the Budda Cave. Secondly, it is possible to interpolate the data value to the exact spot of the Budda Cave. The bi cubic interpolation will include all available stations and the importance of each station will be weighted according to its distance from the Budda Cave. The mathematical program “Matlab” made the bi-cubic interpolation using the calculated average annual temperature and total annual rainfall records from the 21 stations. Figure 3.6 shows the resulting interpolated record. The annual precipitation and mean temperature above the Budda cave are approximately 519 mm and 15°C according to Li and Ku (unpublished data). The values based on the interpolated “Budda cave” station are 703 mm and 14.3°C. These values are reasonable, when compared to the average values for the other stations shown in Table 3.1 and to the contour plots shown in Figure 3.3 A) and Figure 3.3 B). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0 £ 1 1 6 - ® Q_ E ® 1 4 - B u d d a <s 3 c e 0 1 c 0 1 0 1 s a v .: 1 4 .3 C 10 - - 1200 - - 1000 a v .: 7 0 3 m m /yr - 8 0 0 -- 6 0 0 -• 4 0 0 -200 1 9 7 0 1980 1 990 1 9 6 0 1 9 3 0 195 0 1 9 2 0 1 9 4 0 E E C a le n d a r y e a r Figure 3.6. Interpolated instrumental record at the Budda Cave. 3.4 Correlation between climatic factors and the calcite precipitation of SF-1. The lamination counting of individual annual layers of deposition in stalagmite SF-1 was continued back to A.D. 1854. The history of net calcite accumulation was estimated based on the thickness of these layers, as shown in Figure 3.7. The calcite layers range in thickness from 0.029 to 0.278 mm1 . The thickness of the calcite layers corresponds to the theoretical values of about 0.05-1.0 mm/year which are found using the model of Dreybrodt (1980) for cave water with Ca+ 2 concentrations of 0.5 to 2.6 Two relative thick layers in the beginning and end o f the record (1853, 0.328 mm) and (1996, 0.217 mm) have been omitted the Figure 3.7, since it is unknown whether these layers actually deposited within a single year, or the thick layers are artifacts of the layers being in the end and beginning o f the record. 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. mmol/1 (Railsback, 1994). Only 3 of the 144 layers fall outside of this range, being thinner than 0.05 mm. Figure 3.7 show a peculiar shift in the laminae thickness around 1930 which may be related to human activity (reduced vegetation cover above the cave), rather than climatic factors. This will be discussed later in this chapter. 0.35 0.3 • 0.25 0.2 av.: 0 .1 1 1 - 0.15 0.1 0.05 - 1850 1870 1890 1910 1930 1950 1970 1990 calendar year Figure 3.7. Record of lamina thickness for Stalagmite SF-1. Figure 3.8 shows a comparison of the temperature records from the interpolated Budda cave Station and lamina thickness. No obvious correlation is found upon visual inspection. Plotting the temperature versus the lamina thickness also reveal no statistically significant correlation between the two (R2= 0.007). This is graphed in Figure 3.9. 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. T em p. @ B udda 14 a v .: 0.08 mm Lam ination - 0.1 • - 0.05 1990 1 955 1975 1980 1985 1940 1 9 5 0 1970 1945 1960 1965 Calendar year Figure 3.8. Interpolated annual average temperature at the Budda cave and lamination thickness. Attempts were also made to find a correlation between lamina thickness and average temperature for the summer monsoon season (April through September), because it is assumed that the majority of the stalagmite deposition occurs during this season. Correlations were sought not only for the interpolated Budda Station but also for the Lanzhou and the Ankang stations for comparison. None of the temperature records for any of the stations were found to correlate well with the laminae thickness, 7 7 as shown by the R -values for the temperature correlations listed in Table 3.2. Uncertainties can be 2 R2 is the square o f the PEARSON product momentum correlation coefficient though data points of known x’s and known y ’s. For more information see the help function in Excel. 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 1 6 • £- 14- y = 2 .70 05 X + 14.123 R2 = 0.0072 ai 1 3 - 12 - 0.04 0.12 0.14 0.16 0.06 0.08 0.1 0 0.02 Laminae thickness [mm] Figure 3.9. The correlation between average annual temperature at the interpolated Budda cave Station and lamination thickness. introduced with lamination counting. These uncertainties include: 1) layers could potentially have been neglected or a smeared layer might have been counted as two layers, 2 ) lag time between precipitation events and calcite deposition may be involved, because of the residence time of water in soil and limestone, and 3) the precise collection time of the stalagmite is unknown and the first layer may represent calcite deposition that occurred in 1996 or in 1997 (Li, pers. com.). Therefore, an attempt was made to insert and delete existing years at selected places in the record. Attempts were also made to offset the lamination counting record a few years in either direction. These efforts did not improve the Revalues. 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A correlation between the temperature records with 5 years running average and the lamination thickness was also investigated. No significant correlations were found. The R2 values are also shown in Table 3.2. Table 3.2. R 2 values for the temperature correlations. Tem perature Station Factor R 2 R 2 (5 yr r.av). Budda Annual mean temperature 0.007 0.0126 Mean temperature April through September 0.029 0.0217 Lanzhou Annual mean temperature 0.0009 0.0165 Mean temperature April through September 0.0005 0.0328 Ankang Annual mean temperature 0.003 0 . 0 0 0 1 Mean temperature April through September 0 . 0 0 2 0.0092 It should be emphasized that the temperature record used for this correlation is a record of the station temperature (at Earth’s surface) and not a record of the actual temperature inside the cave. Station temperatures are used because a cave temperature record does not exist. If the station temperature does not accurately represent the temperature variations in the cave under which calcite precipitated, then the station temperature will introduce errors. According to Li and Ku (unpublished data) the annual mean temperature above the cave is generally the same as the annual mean cave temperature. It is therefore reasonable to use annual mean station temperature as an approximation for annual mean cave temperature. Figure 3.10 shows a comparison of the annual total precipitation from the interpolated Budda cave Station and laminae thickness. Again, no obvious correlation is seen. Plotting the precipitation for the interpolated Budda cave station versus the lamination thickness (Figure 3.11) does not show any statistically significant correlation 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (R2 = 0.063). Inspired by Baker et al. (1998) and Railsback et al. (1994), a correlation was sought with other precipitation factors, including the total precipitation from April to September, the monthly maximum precipitation during the year, and the number of months with rainfall above the average value. The reasoning for this was that more than 75 % of the rain in this area falls in April through September. Since carbonate deposition is assumed to be higher during the wet months, lamina thickness might correlate better with precipitation amount during these months. The maximum monthly precipitation and the number of months with rainfall above the average could potentially provide information about the strength and the duration of the monsoon system. 1200 Precipitation @ B udda 1000 • av.: 7 0 3 mm E, 800 ■ * 0 .1 5 Z Lam ination av.: 0.08 mm - 0.1 t- ■ - 0.05 197 0 1980 199 0 1950 1960 1940 1930 Calendar year Figure 3.10. The interpolated total annual precipitation and lamination thickness. 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Correlations were sought for the interpolated Budda cave station and for the Lanzhou and the Ankang stations. None of the precipitation records for any of the stations were found to correlate well with the lamination thickness. R values for the precipitation correlations are shown in Table 3.3. Again, inserting or deleting an extra year in the lamination record or offsetting the lamination counting record a few years in either direction did not improve the Revalues. An attempt was made to find a correlation between the precipitation records with 5 year moving average and the lamination thickness. No significant correlations were found here either. The R 2 values are shown in Table 3.3. 1200 1000 E 800 y = -1145.4X + 796.89 R2 = 0 .0 6 2 200 0.16 0.18 0.04 0.06 0.08 0.1 0.12 0.1 4 0.02 0 L a m in a e th ic k n e s s [m m ] Figure 3.11. Correlation between the total annual precipitation and laminae thickness. 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 3.3. R 2 values for the precipitation correlations. Precipitaltion Station Factor R 2 R 2 (5 yr r.av.) Budda Total annual precipitation 0.063 0.0303 Total precipitation April through September 0.031 0 . 0 0 0 2 Monthly max precipitation 0.051 0.0063 Number of months above average 0.039 0.0548 Lanzhou Total annual precipitation 0 . 0 2 1 0.007 Total precipitation April through September 0.025 0.0004 Monthly max precipitation 0 . 0 1 0.0154 Number of months above average 0.0004 0.0042 Ankang Total annual precipitation 0.038 0.0044 Total precipitation April through September 0.036 0.085 Monthly max precipitation 0 . 0 1 1 0.1891 Number of months above average 0.030 0.2762 3.5 Correlation between laminae thickness and the combined temperature and precipitation data An alternative attempt was made to find a correlation between the laminae thickness and the combination of temperature and precipitation data. This correlation was investigated by performing a bi-linear regression to the annual mean temperature and precpitation data from the Budda, Lanzhou and Ankang stations respectively. The bilinear regression equation was: ccP + /3-T + y = L (3.5.1) where P is a vector containing the precipitation data, T is a vector containing the temperature data and L is a vector containing the laminae thickness. The three vectors have N elements. 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The constants a, P and y were found by minimizing k in the following equation, utilizing the Excel solver function (for a more complete description on how to utilize the solver function in Excel, refer to Appendix 3 A): 2 ( a • Pi + $ • tt + y - )1 = k (3.5.2) i = i The same procedure was applied to data for the average temperature and total precipitation for the summer monsoon season and for records of 5 years running averages. Table 3.4 lists the result. The R 2 value for the correlation between (a-pi+P-tj+y) and (I,) was also calculated as shown in Appendix 3A. Table 3.4. R 2 values for the temperature/precipitation correlation with lamina thickness. Combined temperature and precipitation Station Factor R 2 R 2 (5 yr r. av.) Budda Annual mean 0 . 1 2 1 0 . 0 1 2 April through September 0.051 0 . 0 2 1 Lanzhou Annual mean 0 . 0 2 2 0.019 April through September 0.034 0.068 Ankang Annual mean 0.068 0.003 April through September 0.030 0.004 It is observed in Table 3.4, that utilizing a bi-linear regression does not improve the R2 - values significantly. 3.6 Discussion According to the theoretical model of Buhmann and Dreybrodt (1985), the growth rate of a stalagmite should increase with increasing temperature and precipitation. Nonetheless, no significant correlation has been found between the 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. growth rate of stalagmite SF-1 and temperature or precipitation. The most likely explanation is the great complexity of the calcite deposition process. Besides temperature and precipitation, other factors such as soil and drip water Pco2 , drip water Ca2+ concentration, variations in the amount of detrital material incorporated in the growing calcite, and potential stalagmite dissolution (and erosion) impose important controls on the growth of the stalagmite. Factors may add or subtract. Only when few factors predominantly control the calcite deposition will it be feasible to separate the different factors and interpret variations in layer thickness in a climatic sense. Furthermore, the factors that control calcite precipitation are site specific. As an example, Railsback et al. (1994) found good correlation between thickness of calcite layers of a stalagmite from a cave in Botswana and rainfall, suggesting that calcite precipitation was largely dependent on the quantity of water supplied to the stalagmite. Baker et al. (1998) on the other hand were not able to find correlation between rainfall and layer thickness for four stalagmites from a cave in England. The growth rate of stalagmite SF-1 may have been largely dependent on variations in Pco2 in the soil above the cave and in the drip water. Soil Pco2 is created as plant material decomposes. The amount of CO2 released in the soil depends on the amount of decaying material which in turn is related to the type and density of the vegetation above the cave. Vegetation type and density might fluctuate naturally as a response to changing climate and environment. If a significant change occurs and the vegetation density is reduced, less CO2 will be released into the soil. Consequently less CO2 will be dissolved in the groundwater. Lower acidity of the groundwater will cause less limestone to be dissolved during the water’s flow towards the cave. Hence, the 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. concentration of Ca2+ in drip water will be less, and this will lower the subsequent calcite deposition in the cave. This phenomenon might explain the shift in the laminae thickness observed around 1930 as mentioned earlier. This shift may be related to lumbering of the vegetation at the cave site (Li, pers. com.). This activity would change Pco2 as described. Addition of detrital material has probably influenced the growth rate of the stalagmite. When the lamination counting was performed on stalagmite SF-1 several darker bands that were visible to the human eye were observed. Dark bands were also found when microscope images were examined. One explanation for the different optical properties of the CaC0 3 is the inclusion of detrital material in the calcite precipitation. The factors controlling the amount of detrital material incorporated into the calcite are very complex. The amount of incorporated detrital material depends on the flow path and also on the magnitude of single precipitation events. Extreme amounts of rainfall may cause the flow to change its path resulting in significantly more (or less) detrital material than usual. It can be argued that the theoretical model by Buhmann and Dreybrodt (1985) fails to consider the possibility that high water flow may actually cause dissolution of the stalagmite surface rather than increased calcite precipitation. If the flow through the overlying limestone is too fast for the percolating water to establish equilibrium with the carbonate rock the water may be under-saturated in CaCC>3 as it enters the cave and dissolution may occur. Also, for very high drip rates the residence time of water on the stalagmite surface may be less than the time required for degassing of CO2 . This will lower the super-saturation of the film solution (in CaCC>3) and the stalagmite will grow 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. at a slower rate. Finally, enhanced physical erosion of the stalagmite due to a high flow rate is also a possibility. In conclusion, it is very difficult to interpret climate change (changes in temperature and precipitation) based exclusively on growth rate variations. The growth rates of stalagmite SF-1 cannot (by itself) be used to infer paleo-climate variability. However, by obtaining multiproxy records (e.g. growth rate, isotopic composition S1 3 C and 5lsO, and trace elements) from the same stalagmite, a more reliable reconstruction of changes in the paleo-climate (and environment) may be accomplished. More research into lamination growth is needed before it can be concluded if lamination thickness, temperature and precipitation amount correlate in a useful manner. In the specific case of stalagmite SF-1, it appears that lamina thickness does not correlate with temperature or precipitation. 3.7 Summary In this chapter the basic mechanisms for precipitation of stalagmite calcite have been discussed. Many different factors influence the growth rate of a stalagmite and this complicates the potential use of variation in laminae thickness as a recorder of paleoclimate. Instrumental records from the area were studied in order to examine the correlation between lamination thickness of stalagmite SF-1 (from the Budda Cave in Central China) and climate variability in the proximity of the cave. Data from 21 weather stations were compiled from two databases and records for the “Budda cave station” were constructed by bi-cubic interpolation of the station data. The records 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. include mean annual and summer temperatures, annual total precipitation, total summer precipitation, monthly maximum precipitation, and the number of months with precipitation above the average. No significant correlation of lamination thickness with any of these climatic factors was found. Changing amounts of CO 2 in the soil and drip water, varying amounts of detrital material incorporated in the growing stalagmite, and increased stalagmite dissolution are among the factors complicating the deposition history. These additional factors could be the reason that no correlation was found. The growth rate of stalagmite SF-1 does not provide enough information to be used as a measure of paleo-climate variability. It might, however, be used together with other proxy records from the same stalagmite in paleo-climate reconstruction. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4. Paleoclimate Variability from Stable Isotope Records This chapter describes the reconstruction of paleoclimate (paleotemperature and paleoprecipitation) variability in central China based on high resolution records1 of § 1 8Oc and S1 3Cc obtained from the stalagmite SF-1 from the Budda Cave. The factors that influence the stable isotopic composition of cave calcite and their individual effects and importance will be discussed in this chapter. Based on these discussions, climatic interpretation schemes for 5l3Cc and 5l8Oc will be presented. The interpretation of 8 IS Oc will be supported by the establishment of a “temperature- precipitation-5l8Op-relationship” based on modem measurements of surface temperatures, precipitation amounts, and 5l8Op from the weather stations at Xian and Zunyi in the proximity of the Budda Cave. The data from these weather stations have been acquired from the IAEA database. A number of climatic factors will influence the 8 I3C and 8 lsO isotopic composition of a stalagmite (8 l3Cc and S1 8Oc). It can be difficult to determine which of these factors play important roles and consequently should be included in the interpretation. A climatic factor may effect the isotopic composition in multiple ways, which will complicate the interpretation. As an example, the mean annual surface temperature which is approximately the same as the mean annual cave temperature controls both the 8 lsO isotopic fractionation between drip water and the stalagmite calcite, and it also affects the 8 lsO isotopic composition of the precipitation (Sl8Op). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Many areas that experience moderate amounts of precipitation (mid and high latitudes) exhibits a positive correlation between 5 I8Op and temperature (Gascoyne, 1992; McDermott et al., 1999). On the other hand, water-calcite 1 80-ffactionation is negatively correlated with temperature (Gascoyne, 1992; McDermott et al., 1999). Hence, temperature will affect the 5I8Oc in the opposite direction. If the temperature dependent water-calcite fractionation dominates over the temperature dependent 5isO of the precipitation, then a negative correlation will exist between temperature and 5I8Oc. However, if the temperature has a larger influence on the S1 8 0 of precipitation compared to the water-calcite fractionation, then a positive correlation will exist. There are indications that in areas of extremely heavy or light precipitation (in the tropics and subtropics), the 5lsO of precipitation does not appear to be correlated with the temperature (Li and Ku, unpublished data). Therefore, in these areas the 5l8Oc may be negatively correlated with temperature due to the water-calcite fractionation effect. The site-specific nature of the temperature dependence of 5 18Oc has resulted in opposing interpretations of 5 18Oc in different parts of the world. As an example, Gascoyne (1992) interpreted higher 5 1 8Oc values in cave deposits from Vancouver Island (Canada) as indicating warmer conditions, whereas he interpreted higher 5I8 Oc values in cave deposits from northern England as indicating colder conditions. Li and Ku (unpublished data) have compared the relationship between Sl8Oc and temperature in recent cave studies throughout the world and found that a positive relationship in 1 The 8-value is defined as: 5 in % o = [ ( R S ; m p i = - R s n i i d n r d ) / R s t M d a r d ] ‘ 1000, where RS a m p ie represent the measured isotope ratio and R sum dan i is the isotope ratio for a universal standard. For carbon: R= 3 C/I2 C and 62 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. general exists at mid and high latitudes regions of the Northern Hemisphere. A negative correlation was typically found in the tropics, desert areas, and the Southern Hemisphere. Besides temperature, the amount of rainfall has been found to be important in controlling the 8 lsO of cave calcite. Some controversy exists in the literature on how to interpret the 8 1 3 C isotopic composition of stalagmites. The principle behind the use of 8 I3C records to interpret climate change is that the S1 3 C composition of the soil pCC>2 (some of which eventually becomes incorporated into the cave calcite) is influenced by the type of vegetation cover above the cave. Different types of plants have different S1 3 C compositions. Plants adapted to colder and wetter climates typically have a lower 8 1 3 C isotopic composition than plants adapted to a warmer and drier climate (Cerling, 1984; Cerling et al., 1989; Hoefs, 1997; Ku et al., 1998). Utilizing this basic concept, 8 1 3Cc has been used in many paleoclimate studies to infer cold/wet and warm/dry climate conditions (e.g. Bar- Matthews et al., 1997; Dorale et al., 1992 and 1998; Ku et al., 1998). However, the extent of host rock dissolution above the cave will also affect the SI3Cc. Baker et al. (1997) and Bar-Matthews et al. (1999) argue that changing residence times of water in the soil above the cave may influence the 8 l3Cc significantly, since it will change the degree of limestone dissolution above the cave. Furthermore, Baker et al. (1997) argue that changes in the route of the drip water during its passage through the aquifer may influence the S1 3Cc of the cave deposit. Therefore caution is needed before 8 1 3Cc is interpreted in terms of changing paleo-vegetation. for oxygen R=l80 / 1 6 0 . 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Isotope records of 5l3Cc and Sl8Oc have a simple climatic interpretation for stalagmite calcite that formed in calcite isotopic equilibrium with the cave water. Isotopic equilibrium deposition means that fractionation of calcite during precipitation is exclusively controlled by the temperature. Stalagmite calcite is rarely deposited in complete isotopic equilibrium. Varying drip water rates, pH-values of the drip water, exchange of CO2 between the drip water and the cave atmosphere, and drip water evaporation may kinetically influence the isotope values. The criteria that will be used to indicate whether the 5 1 8Oc of the stalagmite calcite formed in stable isotopic equilibrium are the criteria introduced by Hendy (1971). A model is developed here to estimate the range of 8 I3Cc values that would be the result of calcite deposited in isotopic equilibrium. These estimates are compared to the measured 5 1 3Cc values. A disagreement between the predicted range of values and the observed range of values is indicative of kinetically dominated non-equilibrium calcite deposition. The precipitation of calcite in isotopic disequilibrium does not negate its use in paleoclimate reconstructions. It does however, complicate interpretation that rely on isotopic variability. The interpretation schemes for SI8Oc and 8 1 3Cc will be applied to two high resolution records: one record with -100 years resolution that lasts from 0 to 9.4 ka and another record with a resolution of -1-3 years covering 0 to1266 a. Both records are from the stalagmite SF-1 acquired from the Budda Cave in Central China. Samples from these records were collected using a computer controlled high precision micro miller (for a manual on how to use the micro miller, refer to Appendix 4A) and they 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. were analyzed in a VG Prism II mass spectrometer with a common acid bath automated preparation system. The 5I3C and the 8 lsO will be reported relative to the PDB standard2. Between every ~7 samples, a working standard was analyzed. Based on the reproducibility of the standards, a one-sigma measurement error of ± 0.15%o for 5 1 8Oc and ± 0.l%c for 8 1 3Cc was measured. The reproducibility ^vas also determined based on 6 duplicates from stalagmite SF-1 to be ± 0.1%o for Sl8O c and ± 0.1%o for S1 3Cc. The chronologies of the records are based on TIMS 2 3 0Th/234U dating, 2 1 0 Pb dating, and lamination counting (the latter two for the part of the staLagmite that was younger than 150 years). For each isotopic record, specific climatic events are identified and compared with findings of other groups doing paleoclimate studios in the same time period. Finally spectral analysis was performed on the isotope record to recognize if any significant cycles are imbedded in the data. 4.1 81 3 C as a Paleoclimatic Indicator Stalagmite calcite is rarely deposited in complete isotopic equilibrium with respect to 5 1 3C. Kinetic effects often have an important influence on the isotopic composition of the 8 I3C of calcite. Potential kinetic effects that may have influenced the deposition of stalagmite SF-1 will be discussed later in this chapter. The following description of 8 1 3 C as a Paleoclimatic Indicator is based o n the assumption that kinetic 2 The PDB standard (The Belemnitella americana from the Creataceous Peedee formation, South Carolina) is a commonly used international reference standard (Hoefs, 1997). 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. effects are not dominating the 5 1 3Cc signal and that the 5 l3 Cc reflects the isotopic composition of the carbon species present in the drip water (for the basic mechanism of carbonate deposition, please refer to Chapter 3). The carbon isotopic composition of drip water is primarily controlled by 1) the 8 l3C of the initial CO2 dissolved in the soil water above the cave and 2 ) the degree of dissolution of the host rock Oimestone) above the cave and its influence on the distribution of CO2 species dissolved in the cave water. These two factors are discussed in detail below. The S1 3 C of soil CCb. The Sl3C of the soil CO2 above the cave will influence the S!3C of the drip water and the 5I 3Cc of the calcite that precipitates from this water. The isotopic composition of the soil CO2 depends on the type of vegetation that is prevalent above the cave. Therefore, a change in 5 l3C c is indicative of a change in the vegetation cover above the cave. Different plants use different photosynthetic pathways (= enzymatic processes) to fix carbon. The three m ajor photosynthetic pathways for terrestrial plants are (Cerling, 1984; Cerling et al., 1989; Hoefs, 1997): 1) The C3 (Calvin-Benson) pathway: used by approximately 90% of the Earth’s terrestrial plants. The fractionation between the plant carbon and atmospheric CO2 is ~ 20-22%o which results in plants having a SI3C of between -34%c to -23%c PDB. Plants that use the C3 pathway include trees, most shrubs, herbs, and cool seasonal grasses. These plants typically grow in temperate climates (relative cold and wet). 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2) The C4 (Hatch-Slack) pathway: results in 5I3C values of — 23 %o to -6 % o PDB. This photosynthetic pathway is typically utilized by maize, sorghum, subtropic prairie grasses and savanna grasses. These plants are more adapted to warmer and drier conditions than C 3 plants. 3) The CAM (Crassulacean Acid Metabolism) pathway: uses a combination of C3 and C4 metabolism and results in 5IjC values of -33%o to — 11 %o PDB. This photosynthetic pathway is utilized in plants, like cacti and euphorbs, which are adapted to very arid environments. When vegetation above the cave decomposes, C O 2 that has a Sl3C value characteristic for the plant material is released into the soil and dissolved in the percolating ground water. The carbon isotopic composition of the ground water pCC> 2, therefore depends on the vegetation-type (C3, C4 or CAM). It is assumed that any refractory carbon has the same isotopic value as the CO 2 dissolved in the ground water. Variations in the isotopic composition of the ground water pC0 2 will be reflected in the S1 3 C of the stalagmite although fractionation in the precipitation of the stalagmite calcite will offset the value. Therefore, a record of 5I3Cc can be used to deduce changes in the relative abundance of plants utilizing the different photosynthetic pathways. As an example, relatively higher 5 1 3Cc values in general will indicate a plant community dominated by plants utilizing the C4 pathway relative to those utilizing the C3 pathway and, therefore a relative warm and dry climate prevails. According to Schleser et al. ( 1999), a drier and warmer climate would not only decrease the C3/C4 plant ratio but also cause the Sl3C of the C3 plants to increase. This is because these climate conditions 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. would restrict the uptake rate of CO2 and cause relatively more C 13C >2 to be absorbed by the plant. A decrease in vegetation density caused by changing climate conditions or by human deforestation will also be reflected in the 8 l3Cc of a stalagmite. A lower vegetation density will decrease the input of isotopically light carbon (from decomposing plant material) to the soil and ground water. Consequently, the input of isotopically light carbon to the stalagmite would diminish. Dissolution of limestone above the cave The degree of limestone dissolution above the cave will also influence the 8 1 3Cc of the stalagmite calcite. The carbon isotopic composition of limestone is generally relatively uniform with a value near ~ 0% o (Hoefs, 1997). Dissolution of limestone therefore tends to increase the carbon isotopic composition of the ground water. The more limestone dissolution, the more enriched the stalagmite calcite will be. The degree of limestone dissolution depends on the residence time of the ground water and other factors such as Pco2 , pH and temperature. On decadal or shorter time scales, increased precipitation will shortening the residence time of the seepage water in the overlying limestone, leading to a reduced amount of dissolved bedrock CaC0 3 (Ku et al, 1998). The CaCOs dissolution will also be less because of the reduced Pco2 in the seepage water. Therefore, wetter climates may result in lower 5l3Cc values. Baker et al. (1998) suggest that the 51 3 C of a stalagmite may reflect changes in flow routing above the cave. However, Li and Ku (unpublished data) argue that 68 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. significant changes in the drip-water flow route cannot occur unless tectonic activities alter the plumbing system of the cave. It is unlikely that the Budda cave has experienced any significant earthquake activity within the last 20 ka. Therefore, it is assumed that the hydrological conditions above the cave have remained the same. Changes in the hydrological conditions are therefore disregarded in the interpretation of the S1 3Cc record from the stalagmite SF-1. In summary, the S1 3Cc is assumed to reflect a combination of vegetation changes and changes in the dissolution of limestone as summarized in Table 4.1. It is observed in the table that drier and/or warmer climate leads to an increase of 51 3 C in stalagmites, whereas wet and/or cold climates result in lighter 5l3C. Table 4.1. Interpretation of the 8 1 3Cc signal. Climate change Effect of the climate change Isotopic signal Sl3Cc Drier and/or Warmer Climate C3/C4 plant ratio ? 8 I 3 Cc T 8 1 3C o fC 3 plants? Vegetation density ? Water residence tim e T and limestone dissolution T Wetter and/or Colder Climate C3/C4 plant ratio T Sl3Cc i 8 1 3 C of C3 plants ? Vegetation density T Water residence tim e ? and limestone dissolution ? 4.2 5lsO as a Paleoclimatic Indicator The amount of lsO eventually incorporated into a stalagmite precipitating from a drip water source is determined by: 1 ) the S1 8 0 isotopic composition of the drip water Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and 2) the 1 80 / 1 6 0 fractionation between the drip water and stalagmite calcite. These two factors are discussed in detail below. The 5l8Q isotopic composition of the drip water The 8 lsO isotopic composition of drip water in a cave has been shown to generally reflect the weighted annual mean of the SI80 of the precipitation, 5 1 8Op (Bar- Matthews et al. 1996; Yonge et al., 1985). This is assuming that the 8 lsO isotopic composition of precipitation is not changed significantly by evaporation processes at the surface level and that the water reaches the cave in a relative short time (<3 years), which is generally valid for areas that are neither arid or semiarid. The propagation delay between the time of rainfall and the arrival of drip water in the Budda cave is estimated to be less than 3 years. This estimate is based on measurements of 3 H in rainwater and dripwater. It only takes a couple of days for the cave drip water to respond to surface precipitation (Li, pers. com.). Li and Ku (unpublished data) measured the 8 lsO isotopic composition of drip water at the Budda Cave and compared it with the weighted mean 8IS 0 of annual precipitation at the Xian weather station in the proximity. Their results indicate that the cave water 5lsO represents the 5I8 0 value of the precipitation. The isotopic composition of precipitation 5 1 8Op is controlled by several factors: the isotopic composition of the vapor source and the resulting cloud that produces the precipitation the air mass history (change in moisture source) 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the amount of rain-out above the cave the condensation temperature The isotopic composition o f the vapor source and the resulting vapor The major water vapor source of precipitation over the Budda Cave is sea surface water in the N. Pacific, the South China Sea, and the Indian Ocean. The SlsO of these surface waters has remained the same (<0.2%o change) during the last lOka with a 8 lsO value ~ 0%o (Bradley, 1999; Fairbanks, 1989). However, prior to 12 ka, prior to the meltback of the northern hemisphere ice sheets, the S180 of these source waters would have been approximately 1 %o heavier (Bradley, 1999). l 8 0 fractionation occurs when the water evaporates, because of the higher vapor pressure of H 2 1 6 0 compared to HDieO and H 2 lsO. The degree of fractionation between water vapor and water is a function of the evaporation temperature. The equation is: Ctiiquid-vapor = R liq u ic /R v a p o r^ 1.0112-8.2678*10 5 *T°C, (4.2.1) where R iiquid = ( l8 0 / 160 ) i iquid and R vap o r = ( 180 / 160 ) vapor (Ku and Li, 1998). Hence, for an evaporation temperature of 20°C, the initial vapor source is ~9.5% o lighter than the 8 lsO of the mean surface ocean source. The annual average evaporation temperature does not vary substantially over time scales of a few thousand years. However, the evaporation temperature might have been ~2.5°C lower at the end of the last glacial age (Wang et al., 1999) causing the isotopic fractionation to be -0.2%o greater than today. The combined effects of heavier surface water and lower 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. evaporation temperature can therefore account fo r ~ (L%o — 0 . 2 % o) = 0 . 8 %o heavier 51 8 0 at around - 1 2 ka. In conclusion, changes in the vapor source region are believed to be minor and can therefore be disregarded. Furthermore, chamges in the 5lsO of the vapor source and the evaporation temperature can be neglected fo r records younger than 1 0 ka, because these factors have remained nearly constant over this time interval. Air mass history (change in moisture source) The amount of rain-out that occurs as the moist air migrates away from the source towards the cave, will influence the value; of § I 8 0 of the precipitation at the cave. As the air mass travels away from its source, the= heavy water molecules tend to rain-out first. The remaining water vapor therefore becomes lighter as the distance to the source increases according to the Rayleigh condensa_tion model (Hoefs, 1997). The trend towards lighter isotopic values of the precipitation may be reinforced by recycled continental moisture that is added to the atmosphere (Koster et al., 1993; Njitchoua et al., 1999). Storm tracks may advect air masses along warmer and cooler paths, causing different degrees of rain-out and changes in the S lsO of the precipitation (McDermott et al., 1999). In this work, it is assumed that the moisture source of the monsoon system in Central China has been relative stable over the last lOka (Li and Ku, unpublished data). Consequently, the air mass history is not believed to have caused the 8 1 8Op to change significantly during this time period. 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The amount o f rain-out above the cave The amount of rain-out above the cave will also affect the isotopic composition of the rain in a similar way, as rain-out along the rain’s path from the source (as described in the previous section). In general, the lowest 8 lsO values are found with the most intense rain fall events and the highest 5lsO values are found when the rain is less intense (Njitchoua et al., 1999 and Rindsberger et al., 1983). More than 75% of the annual precipitation in China is associated with the monsoons during the summer months. Furthermore, years of heavy rainfall in the summer correlates generally with years of high total annual precipitation. Hence, due to the rain-out effect, years with high amounts of precipitation are expected to have relatively low 5 1 8Op values whereas years with little rainfall will have higher 5l8 Op values. The rain-out effect above the cave is believed to be one of the most important factors controlling the 51 8 0 of a stalagmite. The condensation temperature The phase transformation that occurs during condensation causes isotopic fractionation. On a global scale, rain formation typically occurs in the clouds at approximately 800mb’s pressure equivalent to ~ 2 km above mean sea level (Rindsberger et al., 1983). A relationship between the temperature at 800 mb and the surface temperature exists. The surface is ~20°C warmer according to the temperature/elevation relationship: ~1°C cooling per lOOm’s increase in altitude. 73 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Therefore, a correlation between surface temperature and 5 1 8Op is expected. A linear relation of 0.6%c/°C has been found empirically between the surface temperature and the 8 lsO of the precipitation in areas with moderate rainfall (mid and high latitudes) (Dorale et al., 1998; Gascoyne 1992, among others). Yonge et al. (1985) also found a positive correlation of 0.5%c/°C between the 5i80 of dripwater (reflecting the 5isO of precipitation) and cave temperature (reflecting the mean annual air temperature above the cave) for -5°C to 25°C in caves from Mexico to Canada. McDermott et al. (1999) use a coefficient of 0.59%c/°C for European sites. However, in regions of extreme high or low rainfall (tropical and desert areas) and in some areas in the Southern Hemisphere, no correlation between temperature and 51 8Op exists. This is due to the dominance of the precipitation amount effect (Li and Ku, unpublished data). To investigate the relationship between the rain amount, the temperature, and the weighted annual mean 5 1 8Op in the Budda cave area, a bi-linear regression was performed with data from the Xian and Zunyi weather stations. Data from both of these stations were used to reduce the uncertainty related to the limited length of the data recording 1985-1992. The annual mean temperatures, the 5 1 8Op, and the yearly precipitation data for these stations were calculated based on data from the IAEA/WMO database3 (NOAA, 2000). The following paragraph describes how the bilinear regression was found based on these data sets. Assume that the P vector contains that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. precipitation data, the T vector contains the temperature data and the D vector contains the weighted average Sl80 of precipitation and N is the number of data points: ' P iN r o ' V Pz h \ N J d2 p = , T = and D = K " J The bilinear expression for D based on P and T can be written as: a - P + p f + y = D (4.2.3) The constants a, (3 and y were found by minimizing k in the following equation, utilizing the Excel solver function (For a more complete description on how to utilize the solver function in Excel, refer to Appendix 3A): 2 ( a - P i + f i - t ' + y - d t)2 = k (4.2.4) 1=1 The following equation was the result of the bi-linear regression: S lsOp (%oSMOW) = 0.510%c°C_ 1 • TC C) -1 .5 4 • 10" 3% 0 yr/m m - P[mm/ yr\ - 14.30%o (4.2.5) 3 The IAEA/WMO database is a global network for isotopes in precipitation. This database contains monthly temperature, monthly precipitation amounts and the isotopic composition o f the precipitation from 29 stations in China, covering the period from 1985 to 1992. 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. where T is the mean annual temperature in Celsius degrees and P is the total yearly amount of precipitation in mm/yr. S1 8Op is given in % o relative to SMOW4. The equation shows the inverse relationship between the precipitation amount and the 5 1 8Op as described in the previous paragraph. The equation also finds the temperature coefficient to be 0.510%o°C*1 , which is consistent with the findings of the other groups. I8 0 / l 6 0 fractionation between the drip water and stalagmite calcite. Calcite precipitation from drip water also causes isotopic fractionation. Assuming that the system is in isotopic equilibrium (isotopic equilibrium will be discussed later in this work), the fractionation between drip water and calcite depends only on the temperature. A higher temperature causes a lower 5lsO isotopic composition of the calcite. For the temperature range 0-30°C, the 8 lsO fractionation between the stalagmite calcite and the cave drip water can be described by the linear relationship (R2 =0.99) (Li and Ku, unpublished data): S l*Oc(%oPDB) = 0.97 S lsOw(%oSMOW) — 0.2272%o°C~lT (°C) + 4.2712%o (4.2.6) where the 5l8Oc is the isotopic composition of the calcite, 5 1 8Ow is the isotopic composition of the drip water and T is the temperature in Celsius degrees. 4 The SMOW-standard (Standard Mean Ocean Water) is (like PDB) a commonly used international reference standard. 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The temperature coefficient in the equation is -0.2272 %o°C~x. The negative temperature coefficient is compensated by the positive effect of temperature on S1 8Op (and on SI8Ow ): The total temperature coefficient for both fractionations will then be: (0.97-0.510%o°C"1 -0.2272%o°C'1 ) = 0.268%o °C '1 . Therefore, a net shift towards heavier 5 18Oc is expected with increasing temperatures. This is consistent with the positive 5 18Op — T (and 5 1 8Oc-T) relationship found by Li and Ku (unpublished data) based on measurements o f 5I8Oc versus cave temperature for recent forming stalagmites in nine caves throughout China. In summary, the variations in the 5 1 8Oc of stalagmite SF-1 will be indicative of both changes in the surface temperature and changes in the amount of precipitation above the Budda cave. The warmer and drier the climate, the more enriched the 5 1 8Oc becomes. Cold and wet climate will result in light 5l8Oc values. It is difficult to separate the effects of rainfall variations from temperature variations. Ku and Li (1998) suggest that the mean annual temperature varies relative little over short time scales and therefore will only influence the 5I8Oc on longer time scales (>50 years). On the other hand, the mean annual precipitation amount fluctuates on a yearly basis and will therefore be the primary factor controlling the SI8Oc in high-resolution records ( < 1 0 years). For the S1 8Oc interpretation it is assumed that the present-day “temperature- precipitation-51 8Op“ relationship also was valid in the past. Table 4.2 summarizes the 5I 8Oc interpretation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4.2. 5I8Oc interpretations. Isotopic signal Resolution Climatic pattern 5l8OcT Low (> 50 years) Warm High (< 10 years) Dry 5 1 8Oc1 Low (> 50 years) Cold High (< 10 years) Wet The interpretations of the isotopic records will focus on the determination of climatic patterns: warm/dry, warm/wet, cold/dry and cold/wet. This will be done by combining the information obtained from the trends of the isotopic signals 5l3Cc and 5l8Oc according to the interpretation methods shown in Table 4.1 and Table 4.2. 4.3 Isotopic equilibrium and kinetic effects during calcite precipitation Before the isotope records are used for climate interpretation, it should be determined whether the stalagmite calcite formed in stable isotopic equilibrium with the cave drip water. Stable isotopic equilibrium means that the isotopic fractionation between the heavier and lighter isotopes during the deposition of calcite depends on the thermodynamic conditions. Kinetic effects often have an important influence on the isotopic composition of the calcite. For 5 1 8Oc > it is important to assess the significance of periods of changing rates of evaporation. This is because evaporation will tend to increase the isotopic value of 5l8Oc > Interpretations of paleotemperatures from the 8 I8Oc signal under the assumption that the calcite deposited in isotopic equilibrium will therefore give erroneous results. To establish whether the 5 1 8Oc of stalagmite SF-1 was influenced by evaporation and potentially other kinetic effects, the criteria used by Hendy (1971) was applied. The 78 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. criteria specifies that: 1 ) 5 1 8Oc must remain constant along a single growth layer and 2 ) S1 3Cc and 5 1 8Oc values must be uncorrelated. Figure 4.1 a) shows the different SI 8Oc values from the same growth layer. The samples were drilled on both sides of the center axis at 3.5mm, 13.5mm, 23.5mm and 35.5 mm’s depth in stalagmite SF-1. Accurate sampling along the same growth band was difficult for the three uppermost layers where the bands were very narrow. The Sl8Oc values along a single growth layers were found to be relatively constant on both sides of the center axis for the 13.5mm, 23.5mm and 35.5 mm’s depth. A directional increase was found in the samples from the layer at 3.5mm’s depth which indicated that kinetic processes (evaporation) probably took place at the time of deposition (Gascoyne, 1992). Also a relatively high correlation coefficient for S1 3Cc and 5 1 8Oc was found for the youngest layer (R2 = 0.87), Figure 4.1 b). This indicates that kinetic effects (such as rapid CO2 degassing) probably influenced the deposition the 5 1 8Oc composition of this layer since such effects would shift the 5 1 3Cc and 5 18Oc in the same manner with a mass dependent offset. Relatively low correlation coefficients (average R 2 = 0.24) for 8 1 3Cc and 5l8 Oc was found for the 3 oldest layers (refer to Figure 4.1 for the individual R 2 values). This indicated that kinetic effects probably not influenced the 8 1 8Oc composition of the calcite in these layers significantly. Historical records (Li, pers. com.) reveal that the cave was opened by humans around 1850. In a closed cave the relative humidity is typically around 100% (Hendy, 1971). When the cave is opened the air humidity will decrease significantly due to convection. The lower humidity in the cave atmosphere will increase the rate of 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. evaporation. This may be a reason for the disequilibrium of 5 1 8Oc deposition for the last 150 years. - 8.0 - 8.0 . -8.5 -8.5 § -9.0 CL £ , -9.5 O 5, -1 0 -0 . -9.0 . -9.5 . -L0.0 -10.5 -10.51 - 1 0 •5 1 0 -10 0 5 [mm] from central axis 81 3 C [% o PDB] Figure 4.1. a) Single-layer analysis of 5 1 8Oc in stalagmite SF-1 and b) 5 1 3Cc versus 5 1 8Oc a) is shown on the left side of the figure and b) is shown at the right side of the figure. ■ = 3.5 mm (R2= 0.87), a = 13.5 mm (R2 = 0.02), . = 23.5 mm (R = 0.60), ando = 33.5 mm (R2 = 0.11). The SIjCc in stalagmite calcite is much more complicated due to the gaseous nature of CO2 involved in the deposition of calcite. Kinetic effects typically cause fractionation of 1 3 C during the deposition of calcite. To assess the range of values to be expected for 8 1 3Cc under the assumption that the S1 3Cc deposition occurred in isotopic equilibrium a simplified model has been constructed. In the model the input of CO 2 to the ground water from decomposing organic matter, the degree of limestone dissolution and the isotopic equilibrium fractionation between C C > 2(g ) and CaCOs exclusively control 8 1 3Cc. It is recognized that exchange of CO 2 between the drip water and the cave atmosphere and other kinetically controlled processes may influence 8 1 3Cc. However, this is not included in the model. 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The 5 1 3Cc values that can be expected for stalagmite calcite deposited in isotopic equilibrium for different distributions of C3 to C4 plants above the Budda Cave and for different ratios of soil CO2 to limestone CO2 is estimated in the following. The estimate of 5 1 3Cc is based on the assumption that there are two main sources of 8 I3C to the stalagmite calcite: 1 ) input from organic matter decomposition and 2 ) input from limestone dissolution. The equation that governs the S1 3Cc is: ^ f soil! limestone ^ soilC O z (.g) ^calcite— so ilC O ^lg )') (4.3.1) f soil! lim estone^ ^ limestone ^TCO -,— limestone ^calcite— TCOz ) Where: fsoii/iimestone is the fraction of carbon in the stalagmite that stems from organic soil carbon 8 1 3CS O ii co2( g) is the isotopic composition of the CO2 gas present in the soil. e C aictte-soiico2(g) is the isotopic enrichment factor between the stalagmite calcite and the gaseous CO 2 SI 3Ciim estoneis the isotopic composition of the bedrock (limestone). STC02-iimestone is the isotopic enrichment factor between the total carbon species and limestone. £caic ite -T C 0 2 is the isotopic enrichment factor between calcite and the total carbon species. This equation is based on the assumption, that there is no isotopic exchange between the CO2 in the cave atmosphere and carbon solutions flowing over the stalagmite. The situation is sketched in Figure 4.2. 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C, plants p nlants Figure 4.2. 51 3 C of carbon species involved in calcite precipitation. Unfortunately, no soil samples that would indicate the 5ljC of soil CO2 for the Budda Cave area were available for this work. However, based on the present day vegetation in the area, Li (pers. com.) has estimated the C3/C4 plant distribution to be approximately 70/30%. An average 51 3 C ~ -27%o for C3 plants and SI3C ~ -13%o for C4 plants (Hoefs, 1997) is adopted. If it is assumed that the CO2 gas released to the soil from the decomposing plant material is isotopically identical to that of the plant material, then the resulting 8 1 3 C of the CO 2 gas released to the soil will be: S n C C O M = f C3tCi . S " C C 3 -K l - f c , , c t ) - S n C Ct = (4.3.2) 0.7 • (— 27%o) + 0.3 • (-13%o) = - 22.8%o 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The CO 2 gas from the decomposition of organic matter will become dissolved in the percolating ground water and cause the water to become acidic. Due to the acidity of the water the carbonate bedrock (limestone) will dissolve. The 8 L 3 C of the limestone is ~0 % 0 . The values for e caicite-soiico2(g) for calcite deposition in isotopic equilibrium depends slightly on the temperature. Dulinski and Rozanski (1990) lists the numbers ScaIctte-soiIC02(g) = 10.19%oat5 C and Scalcite-soiIC02(g)— 9.52%c at 15 C. For isotopic equilibrium deposition the enrichment coefficient & rco2-iimestone is very close (within 1% difference for all carbon species involved) to - e C aisite-TC02 (Dulinski and Rozanski 1990). Therefore these two terms cancel out in equation 4.3.1. The fraction of carbon in the stalagmite that stems from organic soil carbon and the fraction that stems from limestone dissolution are unknown. For a soil/limestone distribution of 75/25%, a temperature of 15°C and the C3/C 4 plant distribution of 70/30%, the expected S1 3Cc is: S l3Cc = 0.75 • (— 22.8%o + 9.52%o) + 0.25 • 0 % o = -9.9 6 % o (4.3.3) For this combination of temperature, soil/limestone and C3/C 4 plant distribution, the calculated SI3 Cc value is in the range of the actual measured values. Changing the temperature by 10°C, will change the S1 3Cc value -0.5% o and the effect of temperature is therefore insignificant. In Table 4.3 are shown different values of 5 1 3Cc as function of different soil/limestone and C3/C 4 plant distributions. 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 4 .3 . Calculated 5l3Cc values based on different soil/limestone and C3/C4 plant distributions. C3/C4 plant Soil/limestone S1 3Cc [% 0] 25/75% 25/75% -1.75 25/75% 50/50% -3.49 25/75% 75/25% -5.24 50/50% 25/75% -2.62 50/50% 50/50% -5.23 50/50% 75/25% -7.85 75/25% 25/75% -3.49 75/25% 50/50% -6.98 75/25% 75/25% -10.47 It is observed in Table 4.3 that the Sl3Cc values are very sensitive to changes in C3/C4 plant distributions. The 8 1 3Cc record will be presented later in this chapter. Comparing the observed 8 1 3Cc values and the estimated 8 1 3Cc values from the simplified model shows that the values are in the same range. Had the values been out of range, it could have been concluded that kinetic effects dominated the deposition. Since the observed values are in the predicted range, kinetic effect did probably not dominate the calcite deposition, although some influence from kinetic effects may have taken place. This work has not attempted to model kinetic isotope effects that may have influenced the cave deposit. However, it is likely, that climatic trends will still be detectable in the signal. Kinetic effects are typically controlled by varying drip water rate, pH-values of the drip water, exchange of CO 2 between drip water and cave atmosphere, and drip water evaporation. These factors are in turn related to certain climatic variability and tend to reinforce the climate interpretations. As part of ongoing research kinetic effects are being evaluated. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.4 Climatic Interpretation of the Isotopic Records of SF-1 (0-16.1 ka): Centennial resolution Figure 4.3 shows the records of 8 1 3Cc and 5 1 8Oc for the last 16.1ka (a gap exists from 9.4 to 16ka). The chronology for these records are based on a TIMS ^ ^ h / 2 3 4 !! dating performed by Li and Ku (unpublished data) and a constant growth rate assumption (refer to Appendix 2C for Li and Ku’s results). The sampling resolution was ~I0 years from present day (1997) to 144a, -100 years from 144a-9.4ka and -3 years from 16.0- 16.1 ka5, involving a total of 142 samples. In addition to the “raw” sample values, a 5 sample moving averages is also shown in Figure 4.3 to aid the interpretation. Observations A simple visual inspection of Figure 4.3 reveals that isotope values of 5I3 Cc and S1 8Oc in the upper section (0-9.4ka) of the record are different from the values at the bottom (16.0-16.1 ka) of the record. Both the S1 3Cc and the Sl8Oc values are significantly higher in the bottom than in the upper section. For the time interval 0-9.4ka the S13Cc values vary between — 11.67%o and -6.15%o (average: -9.6%o) and the 5 1 8Oc values vary between -11.75%o and — 7.89%o (average: -9.9%o). The lowest values occur at -1950a and -7070a for 5I3Cc and 5I8Oc, respectively. The highest values occur at 0a and 15a (i.e. recent time). 5 A very high growth rate of 0.72 mm/yr for the interval 16.0-16.1 ka is the reason for the relative high resolution (~3 yrs) corresponding to a sampling interval of 2 mm. 85 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. [aad‘ oo/o] o-et-miap m o b-gi smBiw iunuji;do o i; b l u !|o auaoojoH pouad lujb/w |B ^!p a vM a6v Q O | aiun o o ( Q « o > » CD CO 0 5 I I I I [sad ‘ oo/o] o-8L-ei|3p Figure 4.3 Isotopic composition of stalagmite SF-1. 86 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. For the time interval 16.0-16.1 ka the 5I3Cc values vary between — 7.7%o and — 4.1 % o (average: -6.3%o) and the Sl8Oc values vary between — 7.71%o and — 5.19%o (average: - 6.63%a). Cycles in the centennial record Further inspection o f the 8 l 3Cc and the SI8Oc shows that the climate during the Holocene was subject to oscillations. A quasi-periodic cycle of -1500 years for the 8 l8Oc record is indicated by the thick arrows in Figure 4.3. For 5 I3Cc, a prominent cycle with a quasi-period of -700-800 years is also observed. The existence of these cycles were confirmed by spectral analysis (Fourier Transform (FT)) performed on the data series (refer to Appendix 4B for the Matlab script to perform the spectral analysis). In addition to the 1567-year cycle for the 8 lsOc record, the spectral analysis also disclosed cycles with periods of 1175 years and 469 years. Figure 4.4 graphs the power spectrum of the SI8Oc record. For the 8 l3Cc record, the FT analysis showed cycles with periods of 3135, 725, and 469 years as seen in Figure 4.5. The reason why these cycles are not obvious by inspection of Figure 4.3 is because of additive and subtractive interaction of the different cycles. It should be emphasized that a period of a cycle can only be determined by the FT analysis if the length of the cycle is short compared to the length of the record. Also, dating of SF-1 involves some uncertainty. It is noted that the 725- year and the 3135-year cycles in the 8 I 3Cc record are close to sub-harmonics and multiple (50% and 200%) of the 1567-year cycle observed in the 8 I8Oc record. For a 87 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. discussion on the sub-harmonics and multiples (harmonics) in spectral analysis, please refer to Bryant (1997). Power Spectrum for delta-O-18 record (09.4 ka) 1.2 Number* indicate panada in years 1175 * 0-6 0.2 5.006*03 6.006-03 3006-03 i ooe-03 Frequency (cyciea/year) Figure 4.4. Power spectrum for the S1 8 0 record. Power Spectrum for delta-C-13 record (0-9.4Sca) 1.2 Number* indicata periods in years 725 3135 * 0.6 1.006*03 2.006-03 3.006-03 4.006*03 6006-03 7006*03 Frequency (cycles/year) Figure 4.5. Power spectrum for the 51 3 C record. 88 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The 1567-year cycle of the 8 lsOc record (as well as its harmonics in the 5 1 3Cc record) matches a -1500-year cycle that has been observed in the Indian Monsoon system (Siroco et al., 1996). This cycle was also identified in the Greenland Ice Sheet Project (GISP) 5lsO record by Grootes and Stuiver (1997). The correlation between records from distant locations (teleconnection) suggests that an internal forcing factor (possible a long-term oscillations in the global thermohaline circulation system) may be influencing both the monsoonal climate in Asia and the arctic climate in the Northern Hemisphere. It has not been possible to match the 469-year cycle6 and the 1175-year cycle to any climatic cycles previously described in the literature. Whether these cycles also originate from a larger global climate system, represent local climate cyclicity, or simply result from random variations in the record is unknown. More research in the area of climatic cycles from paleoclimate studies throughout China (and the rest of the world) is warranted. The criteria that was used to determine if a peak was significant enough to be identified as a cycle was based on the assumption that the stalagmite deposition had a red noise spectrum. Figure 4.6 shows red noise spectra for different values of the spectrum parameter a 7. A red noise spectrum is found in many geological time series and means that the majority of the signal is at low frequencies (Bryant, 1997). For equations for the 6 The closest match to this cycle is a 300-400 year cycle found in the S1 3 C and Sl80 records from a stalagmite from Shihua Cave in Northeastern China (Ku and Li, 1998). 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. distribution refer to Torrence and Compo (1999). The exact characteristic (the parameter a) of the red noise spectrum in stalagmites is presently unknown. Therefore, no exact mathematical equation describing the significance of the peaks has been derived. Peaks were identified perceptually in the power spectra using the rule that high frequency peaks had greater significance. Only when low frequency cycles were very strong were they identified. High frequency peaks were identified even though they were relatively weak. This is the reason why some of the peaks in the power spectra were not identified. 0.8 S 0.6 0.4 0.2 ■ 0 m ax frequency F r e q u e n c y Figure 4.6. Red noise spectra for different a. Two major periods with different climate regimes are observed in Figure 4.3. The transition between the two regimes is at ~4.6ka. The youngest regime is characterized 7 a is a measure o f the degree o f “redness” o f a spectrum. a=0 means white noise, i.e. the same amount o f noise at all frequencies. a = l means that all energy is at freq=0. For different a values between 0 and 1, see Figure 4.6. 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. by larger fluctuations of the records and relative stronger correlation between the S1 3 C and the 5lsO. The second regime is characterized be smaller fluctuations and no clear correlation between SI3 C and the 8I8 0 . A regime transition at 4.6ka is also observed in the trace element record. This will be discussed further in Chapter 5. Climatic interpretation of the upper section of the centennial record Below is a climatic interpretation of the first 0 to 9.4ka of the S1 3Cc and Sl8Oc records from stalagmite SF-1. Current understanding of how changes in temperature and precipitation influence the isotopic signals 5 1 3Cc and 5 1 8Oc are summarized in Tables 4.1 and 4.2. At the bottom o f the record (16.0-16.1 ka) isotope values of SljCc and 5 1 8Oc are quite different from the values in the upper section and indicate that the deposition regime (the hydrological conditions of the cave) was very different from today. The bottom part of the record cannot be interpreted using the same interpretation scheme as for the rest of the record; it will be separately discussed later. Decreasing S1 8Ocand increasing S1 3Cc values in the SF- 1 record indicate that the climate became colder and drier climate from 9.4— 8 .6 ka. The cooling is part of the cool/warm oscillations that characterizes the climate in Central China as depicted by the 1567-year cycle in the S18Oc record. Sequential warmer and wetter conditions followed at about 8 .6 ka (as indicated by the increasing § 1 8Oc and decreasing 5 1 3Cc values). Then at -7.7 ka a trend towards a cooler and drier climate began and this trend lasted until -7 ka. From -7.2 to 5 ka the “coldest” values (that is the lowest Sl8Oc values) of the entire record is observed. During the first part of this cold period (from -7.2 to 6.2 ka) the 91 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. climate became drier and in the latter part (from 6.2 to 5 ka) the climate became wetter. The dry period at ~7ka is consistent with periods of increased aridity that have been found in other mid-latitude regions in the Northern Hemisphere based on lake sediment records (Bradley, 1999 and Bryant, 1997). The period of cold and wet climate from ~ 6.2ka to 5ka does not match the climate that has been identified in other parts of the world at this time. Pollen records and tree-line limits on mountains have indicated that the mean global temperature from 6-4.5 ka was 1-2°C warmer than present (Bryant, 1997) and this warm period has been named the Holocene Climatic Optimum. The Holocene Climatic Optimum was a time when the summer heating of the Tibetan Plateau was at its maximum during the past ~ 23 ka (orbital forcing). At that time, the winter insolation over the plateau was lower than today. High summer insolation and low winter insolation caused the seasonality in this part of the world to be greater than today. The increased seasonality probably influenced the strength of the monsoon systems, causing the summer monsoons to increase in magnitude and consequently making the climate in Central China much wetter, an interpretation that is supported by the SF-1 isotope record. The SF-1 isotope record does not support that the temperature was ~1-2°C warmer from 6-4.5ka, instead it indicates that the temperature was the coldest within the last 9.4 ka. A series of alternating cold-wet and warm-dry periods each lasting an average of -600 years is observed in the 5l8Oc and 5I3Cc records from ~5ka to 1.7ka. Then a trend toward warmer and drier climate began at 1700a which lasted for about 1100 years and culminated in the Medieval Warm Period (also referred to as the Little Climatic Optimum) from about 1000a-600a. The Medieval Warm Period was a warming period 92 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. in northern Europe, e.g. it was during the Medieval Warm Period that the Vikings colonized Greenland and Iceland and established regular trade across the North Atlantic (Bryant, 1997). The Medieval Warm Period was succeeded by the Little Ice Age, identifiable in the S,8Oc record from the exceptionally depleted values from ~550a-150a. The low 5i8Oc values, combined with a decreasing 8 IjCc value in the first part of the Little Ice Age, indicates that the climate was cold and wet. Increasing 8 ljCc in the later part of the Little Ice Age indicates a shift towards a drier climate. Finally, for the last 150 years a trend towards warmer and drier climate is indicated by the increasing SI8Oc and 8 l j> Cc values. Some researchers attribute this trend to anthropogenic activities (e.g. burning of fossil fuel) that increase the amount of green house gases in the atmosphere that trap heat and cause the Earth to warm up (Kerr, 2000a and 2000b). Climatic interpretation of the bottom of the centennial record. The isotopic composition of stalagmite SF-1 changes drastically below the hiatus (-16 ka). At ~16ka both the 8 I3Cc and the 8 1 8Oc values are much higher compared to the top section. The ~16ka part of the record belongs to a different deposition regime. Interpreting the high S 18Oc values as ~5°C higher temperatures than present would be inconsistent with other temperature proxies from other parts of the world at this time. All other proxies indicate that the climate was colder than today (Bradley, 1999; Bryant, 1997, and others). The shift towards higher S1 8Oc values should therefore be explained by a significantly different hydrological system above the cave. The higher 8 I 3 Cc and 8 l8 Oc values could also be the result of diagenesis (secondary 93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. calcite deposition) involving kinetic fractionation and possibly enrichment of the initially deposited calcite. The injection of melt water into the ocean at the end of the last glacial combined with an evaporation temperature of about 2.5°C lower than today can account for an increase of about ~ 0.8% o in the 8 1 8Oc values. The observed shift in Sl8Oc is approximately 3%o. Such a dramatic shift in 5 1 8Oc is best explained by a change in the source of the cave drip water and/or by a significant change in the seasonality of the precipitation. Hypothesis fo r the radical shift in the $ 3Cc and the $ 8Oc records at 16ka The following scenario provides a possible explanation for the radically different Sl3 Cc and S1 8Oc values at ~16 ka based on the assumption that the hydrological system above the cave were significantly different at -16 ka. Assume that the surface above the cave where SF-1 started to grow was covered by ice prior to 16.1ka (Li, pers. com.). Then as the global average temperature rose at the end of the last glacial maximun ~18ka, melt water began flowing into the cave. Vegetation in the area above the cave increased and soil pCC>2 rose as the result. 5lsO of the m elt water source above the cave could be significantly different from the present day S 1 8 0 of precipitation in the area. The nature of monsoon circulation during ice ages could be weaker due to increased albedo (reflectance of sun light) caused by the partially ice-covered continent in Asia and the smaller differential heating of land and ocean during summer and winter seasons (Bryant, 1997; Clemens and Prell, 1991). As a result o f the weaker monsoon, 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the precipitation above the Budda Cave and the ice cover in Central China may have been enriched in 5lsO compared to today (the amount of precipitation may have been less and resulted in precipitation that was isotopically heavier). As the ice started to melt at -18 ka the calcites that eventually began to form in the cave — 16 ka may have had a relatively heavy S1 8 0 value. The heavy 5 l3Cc signal observed at ~16ka may also be a reflection of the sparse vegetation above the cave, because only very small amounts of depleted carbon will be supplied to the drip water system from the organic matter in the soil. As the ice cover retreated and a greater percentage of the drip water originated from precipitation associated with the stronger summer monsoon systems (that began to be established), the 51 8 Oc signal became more similar to that of today. Also, as the vegetation density increased above the cave, the 81 3 Cc signal converged towards present values. Unfortunately, this transition zone from 15.9 ka to 9.4 ka is not recorded in the stalagmite because the stalagmite did not grow in this time interval. This discontinuity in growth was most likely not related to climate changes but caused by a temporary shift in the “mouth” of the drip water. An alternative explanation of the gap in the record could be dissolution or erosion of the stalagmite. Dissolution can happen if the drip water is under-saturated in CaCOs as it enters the cave and drips on the stalagmite. Erosion can occur if drip rates are excessive. It should be emphasized that the explanation for the enriched 51 3 Cc and 5l8Oc levels at ~16ka is just one plausible explanation. Other (and better) explanations may exist. 95 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.5 Climatic Interpretation of the Isotopic Records of SF-1 (0-1300a): Subdecadal resolution. Figure 4.7 shows the profiles of 478 pairs of 8I3Cc and SI8Oc analyzed for SF-1 with subdecadal resolutions (~l-yr resolution for the last 143 years and -3-4 year resolution during 143a-1267a). The chronology for these profiles was constructed using a combination of the TIMS 2 3 0 Th/234U dating, 2 1 0 Pb dating, and lamination counting. For the ~l-year resolution part of the record, the depth-age model obtained from the lamination counting and described previously was used. The Medieval Warm Period and the Little Ice Age were identified in the centennial resolution records. They also show up in great detail here. The Medieval Warm Period is identified in the record as a time interval from 1030a to 520a where the 51 8 Oc is above the overall 1266 years average (indicated by the dashed line in Figure 4.7) except for a very short time span of 15 years between 760a and 745a with lower values. The Little Ice Age is identified in the record as the time interval from 520a where the 5I8Oc values start to drop and become lower than the overall 1266 years average 51 8 Oc until 150a where the SI8Oc values start to increase again. The l3Cc record indicates that the Medieval Warm Period consisted of 3 parts. The first part lasted -120 years (from 1030a to 910a) and was relatively dry. The ensuing period lasted approximately -165 years and was wet. The last -225 years of the period were dry. 96 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. -7 -8 — 0 3 a a. •9 8 ■ a A -11 | -12 0 2 0 0 400 6 0 0 800 1 0 0 0 1200 Years ago [a] Figure 4.7. Subdecadal isotope record for stalagmite SF-1. About 520a the climate suddenly became much cooler, marking the onset of the Little Ice Age. According to the SI3 Cc record, the first part of the Little Ice Age (~520a- 350a) was relatively wet, whereas the remaining part (from 350a-150a) was relatively dry. This wet to dry trend is also found in a stalagmite from the Shihua Cave near Beijing (Ku and Li, 1998), as shown in Figure 4.8a. The 5l8Oc record of SF-1 suggests that the coldest peaks during the Little Ice Age were 355-447a. (-1550-1642 A.D.), 251-32la (-1676-1746 A.D) and 168-205a. (-1792-1829 A.D.). These peaks correspond well with time intervals when the temperatures in Mongolia were ~1-2°C 97 W arm Warm Little Ice Age M edieval W arm P eriod "coldest d ecad e" Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. below the present day temperature based on tree-ring measurements (Jacoby et al., 1996), i.e., 1565-1615, 1670-1750, and 1785-1900 A.D (Figure 4.8b). Ice core oxygen isotope data from the Dunde Ice Cap in NW China exhibit a similar cold trend for the period 1560-1660 (Mosley-Thompson et al., 1993). However, later cold periods as well as the Medieval warm period are not recorded in the Dunde Ice Core, as shown in Figure 4.8c. This indicates that despite the relatively close proximity between the Budda Cave and the Dunde Ice Cap (~1100 km), these two areas are influenced by different climate systems. The 51 8 Oc record indicated that the coldest period of the Little Ice Age occurred -275-297 years ago (-1700-1722 A.D.). A compilation of historical records of river and lake freezing in central and southern China showed that 1711-1720 was the coldest decade over the last 480 years (Bradley, 1999). When comparing the Mongolian Tree Ring record (Figure 4.8b) with the 5l8Oc record of SF-1 (Figure 4.7), it is observed that Central China recovered from the Little Ice Age approximately 50 years earlier than areas further to the north. A similar observation is made by comparing the stalagmite records of the Budda Cave and the Shihua Cave near Beijing in the NE China (Figure 4.8a). In NE China, the Little Ice Age lasted until -1900 A.D., whereas temperatures in Central China returned to present day values at -1850 A.D. This comparison suggests that the onset of the Little ice age could be about 80 years earlier in Central China than in NE China. Since, the chronologies for the younger parts of SF-1 and for the Mongolian three ring records are accurate to a few years, these observations are most likely not an artifact of uncertainty in the dating method. 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Keeping in mind, that the primary factor controlling the 51 8 Oc high resolution records (<10 years) is precipitation, both the 81 8 Oc and the 8I3Cc records also indicate that the wet versus dry years alternated more frequently during the last ~ 200 years than in previous intervals. Potentially, this could be an artifact of the increased, resolution of the last 144 years of the record. Instrumental records of precipitation are not long enough to verify this possibility. The range of variability of the isotope records for the last 144 years are within the range of variability present in the earlier part of the cave record. Since no continuous trend towards either isotopically lighter or heavier values are observed in the 8I3 Cc signal, the cyclic nature of the 8I3Cc record probably reflects natural variability (i.e. the natural plant distributions response to climate related cycles) rather than the influence by agricultural activities in the area. The approximate +1 % o shift in 81 8 Oc values that is observed in the record at ~150a and which lasts until the present may be due to the opening of the cave by humans around 1850 A.D. Opening of the cave may have increased the rate of drip water evaporation and caused the 8I8Oc of the stalagmite to become heavier. Figure 4.8d) shows a comparison between the 8I8Oc record for the last 144 years and the (interpolated) instrumental precipitation record at Budda Cave filtered with a 5- year moving average. The two records correlate well with each other, reinforcing the assumption that wet and dry climates can be inferred from short time (<10 years) variations in SI8Oc. 99 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Isotople composition of ths stalagmite S312 (Ku and L .I, 1998) • T Little Ice Ago •e ■o i o o o 2000 1900 n I " 0 o T a n / a a s t a y R i n g D e t e ( U e c o b y a t a l , 1 0 9 6 > ! ’ Cbkl Cold Cokl o I** 0>1 8 record from Dunda lee C ap, China (Moal»y-7 at el., t 093) 3 oo o c I * O — r Warm Cold Warm >0 I Cold |\Vann I Cokl Cokl .s o 1 2 200 1 1 O O 1 1 O O 1 740 i s a o i © o o 1 O O O Calandar yeer [A.O.] Comparison botwaan dalta-1 S-O raoord and praolp. at Budda Cava >o •e v 20 0 - is o 20 OO o o o Figure 4.8. Reconstructed climate records from data in: a) Ku and Li, 1998, b) Jacobi et al., 1996, c) Hughes et al., 1994 and from: d) instrumental precipitation data from the Budda Cave Area. 100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. On top of the climate trends of the Medieval Warm Period and the Little Ice Age are several shorter cycles. Spectral analyses identified the periods listed in Table 4.4 (refer to Appendix 4C for the actual power spectra). Spectral analyses performed on the 0-144 years record and the 0-1266 years record separately enable us to identify short cycles present only in the 0-144 year record. It should be emphasized that spectral analysis is best at identifying constant cycle signals as opposed to quasi-periodic signals. Also, as mentioned earlier the FT analysis is best at identifying signals with periods that are short compared to the time span of the record. Table 4.4. Periods identified in the power spectra. 51 3 C 0-144 years record: (48), 24, 12'; and 9.6 } years cn •a o 0-1266 years record: (422), 140, 51, 44 and 33*) years -C o a* 5lsO 0-144 years record: (36), 14.4, 12"\ 9.6*} and 7.2 years 0-1266 years record: (633-422), (253), 115 and 33*J years * Period shown in both S1 3 C and 5I8 0 . Parenthesis means that period is too long compared to the record length to be accurate Most of the identified cycles in Table 4.4 have also been identified in other paleoclimate records (Bradley, 1999, Bryant, 1997, Hughes et al., 1994; Ku and Li, 1998). These cycles are described below. Previous work has not reported the cycles of 140, 115, 51, and 44 years. Either these cycles have not been discovered or they might be an artifact of the random nature of the signal. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The 24-year period (and its simple harmonic of 12 years8 ) is probably the 22- year cycle that has been identified in the historical drought and flood data in the central part of the Yangtze River (Ku and Li, 1998). The 22-, 11- and 9.6-year cycles have been found in climate records throughout the world, and they reflect features of Sun spots or the lunar orbit (Bryant 1997). The 36-year cycle (as well as the 33-year cycle) matches the 36.7-year cycle that has previously been found in historical climate data in the eastern part of the Yangtze River Basin as well as in SW China (Bradley, 1999). This cycle has also recently been found in a stalagmite record from the NE China (Ku and Li, 1998). According to Bradley (1999) this cycle is probably related to large, synoptic- scale pressure anomalies over eastern Asia and adjacent equatorial regions. Interestingly, a 35-year cycle was also recognized in Europe by Bruckner in 1890 (Ku and Li, 1998). Whether these cycles in fact represent the same cycle and what causes the potential teleconnection remain unclear. The 7.2-year cycle (and its multiple 14.4-year cycle) found in the 5lsO 0-144a record of SF-1 is similar to a quasiperiodic 7.2-year cycle that was found in a precipitation record based on tree ring studies in the Qinling Mountains in Central China by Hughes et al. (1994). Although the origin of this cycle is uncertain, finding this cycle in two independent studies (and using two independent paleo-precipitation proxies) increases the probability that the cycle actually exists. s Harmonics occur when a strong cycle displays a weaker cycle at one-half o f its wavelength. 102 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.6 Summary In this chapter, two high resolution 81 3 Cc and 81 8 Oc records of stalagmite SF-1 from the Budda Cave covering two time intervals have been used to infer changes in paleoclimate in Central China during the Holocene. The first record that covers the last 0-9.4ka has a resolution of about 100 years. The second record lasts from 0-1266a and has a resolution of ~l-3 years. By applying Hendy’s criteria (Hendy, 1971), it was found that the deposition of 8l8Oc most likely occurred in isotopic equilibrium until -150 years ago. A model to precict the 8l3Cc for calcite precipitated in isotopic equilibrium was constructed. The predicted and the observed 81 3 Cc values were within the same range. This indicated that the kinetic effects probably did not dominate the deposition with respect to 8I3Cc. It was therefore assumed that the record could be interpreted in terms of climate variability (~ changes in vegetation and degrees of limestone dissolution), although some influence from kinetic effects may have taken place. The 81 3 Cc and SI8Oc data were interpreted in terms of different climatic factors that influence the isotopic composition of a stalagmite. It was found that increasing Sl3Cc values in general represent a shift towards drier (and warmer) climate whereas decreasing SI3Cc values indicate that the climate became wetter (and colder). Short term (<10 years) trends in 8 I8Oc can be used to infer changes in precipitation. Long term trends (>50 years) in 8l8Oc can be used to infer changes in temperature. Hence, depending on the time scale, increasing Sl8Oc values will indicate that the climate is 103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. getting warmer or drier. Colder or wetter climate can be inferred from decreasing 51 8 Oc values. The interpretation of 5l8Oc was supported by the establishment of a “temperature-precipitation-51 8 Op -relationship” based on modem measurements of surface temperatures, precipitation amounts, and 5I8O p from weather stations in the proximity of the Budda Cave. From this relationship it was found that the positive effect of temperature on 6l8Op (0.516%o°C~l) more than offsets the negative temperature coefficient for water-calcite fractionation of -0.2272%o°C'\ so that a positive relationship between 5 l8Oc and temperature for stalagmite formation in caves in Central China exists. Specific climatic events that have been reported elsewhere were identified from the 9.4ka record including a very dry period centered at ~7ka, the Medieval warm Period (1030-520a), and the Little Ice Age (520-150a). The Holocene Climatic Optimum (a period where the climate was unusually warm and wet) at ~6-4.5ka did not appear in the SF-1 record. The Medieval warm period and the Little Ice Age were also identified in the 0- 1266a record. The 1-3 year resolution gave very detailed knowledge about climate fluctuations within the last millenium in Central China. As an example, the coldest decade was identified and matched historical records. Spectral analyses were performed on the SF-1 records. A 1567-year cycle was identified. This cycle has previously been identified in the monsoonal climate records from Asia (the India monsoon system) and in the Arctic climate in the Northern 104 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Hemisphere. T his globally identified cycle could be caused by an internal forcing factor (oscillations in the global thermohaline circulation system). Cycles with periods of 22, 12 and 9.6 years 'were identified in the isotope records. These cycles are found in other records and reflect the influence of Sun spots and the lunar orbit on Earth’s climate. A cycle with a period of 33 years was also identified in the data. It is likely that this is the Bruckner cycle, earlier identified in European climate. It is believed that this cycle is caused by large synoptic-scale pressure anomalies. Furthermore, a period on 7.2 years was identified. Tree ring data from the same area in central China also exhibits this period. The origin of this local cycle is unknown. Finally, cycles of 469, 140, 115, 51 and 41 years w ere identified in the record. It was not possible to match these with cycles in the literature. These cycles have not been identified previously and their origin is unknown. A majority of the climatic events and cycles in the isotope record have been identified by other research groups. This is a strong indication that the climatic interpretation of the stalagmite is correct. This study has provided an unprecedented high-resolution record for the climate variability in central China during the last ~10ka. This has proven the value of stalagmites as continental paleoclimate recorders. They provide longer records than tree rings and have higher resolution than ice cores and marine records. Furthermore, stalagmites are widely distributed on the continents. The isotope record from SF-1 can be used to improve the understanding of the climate dynamics in central China. The identified cycles will open new opportunities for predicting and. modeling the future climate. 105 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5. Paleoclimate Variability from Trace Element Records The variations in trace element content in a stalagmite can potentially be used as a recorder of paleoclimate. More specifically, variations in the Mg/Ca, Sr/Ca and Mg/Sr molar ratios may be utilized to infer paleotemperature and paleoprecipitation variability. Reconstructing paleoclimate variability from trace elements has been explored by Bar-Matthews et al. (1999), Gascoyne (1983), Ku and Li (1998), Roberts et al. (1998) among others. The mechanisms that leads to the incorporation of varying amounts of trace elements in the cave deposits are still subject to controversy. Different groups are interpreting the trace element records in different ways. Some of these interpretations will be discussed in this chapter. In this chapter an interpretation model will be developed that is based on considerations of the processes that govern the incorporation of trace elements into stalagmites. Many simplifications will be made. The model will assume that the only factor that influences the Mg/Sr ratio is the cave temperature. Changes in the Sr/Ca and Mg/Ca ratios will be assumed to reflect changes in the soil water residence time above the cave, which in turn is determined by changes in the amount of precipitation. The model will be applied to trace element records from SF-1. Interpretations based on the trace element records will be compared to interpretations based on the isotope records (Chapter 4). Samples for the trace element analysis were acquired along the center axis of stalagmite SF-1 for every 1.5 mm in the growth direction for the upper 9.8 cm. The remaining part of the stalagmite was sampled with intervals of 2 mm. The samples were 106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. submitted to the XRAL laboratories, Ontario, Canada and analyzed for trace element content utilizing the “ICP70 Aqua Regia” method1 . The ICP70 Aqua Regia method is based on weak acid digestion of the samples. The samples are subsequently analyzed in an Inductively Coupled Plasma Mass Spectrometer (Anbar, 1998; Bradford, 1997; White, pers. com.). 5.1 Parameters controlling trace element incorporation into stalagmite calcite The processes that govern the incorporation of trace elements into stalagmite calcite are complex. Several parameters control the trace element behavior and cause the trace element ratios to vary over time. The identification and relative importance of these parameters are still subject to controversy (Burton and Walter, 1990; Mucci, 1987). Important parameters include the cave temperature, the drip water rate, and site- specific hydrological conditions such as the mineralogy of the limestone through which the drip water percolates (the extent of dolomitization) and changes in the water flow path (Gascoyne, 1992; Goede and Vogel, 1991; Katz, 1973; Roberts et al, 1998). This chapter will show that information about temperature and precipitation above the cave where the stalagmite is growing can be derived from a simplified model. The composition of the bedrock above the cave is assumed to have remained the same over the relatively short geological time that is considered. Furthermore, since the stalagmite grew continuously at the same position on the cave floor (except for the time interval 9.4ka to ~16ka, where the stalagmite had a growth hiatus) the drip-water 1 In the ICP70 Aqua Regia method a sample is analyzed for the following elements: Al, Sb, As, Ba, Be, Bi, Cd, Ca, Cr, Co, Cu, Fe, La, Pb, Mg, Mn„ Mo, Ni, P, K, Sc, Ag, Na, Sr, Sn, Ti, W, V, Y, Zn, Zr. 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. “mouth” may have remained static over time. Therefore, it is reasonable to assume that the water may have flowed through cracks and fissures in the overlying limestone following the same path. Consequently, the effect of alternating flow routing is neglected. Several studies have shown that the incorporation of Mg into calcite is temperature dependent. The incorporation is caused by three different processes (Gascoyne, 1983): • By occlusion of detrital material in the interstices between growing crystals. • By adsorption onto the surface of growing crystals. • By direct substitution for Ca2 + ions. The latter process is probably the most important for Mg incorporation into stalagmite calcite. The homogeneous distribution coefficient D m is used to quantify the degree of trace element (M) incorporation into calcite. In the case of magnesium, DMg, is defined: L JCaCOj Dm = - ^ -------- (5.1.1) rM g_, L J so lu tio n L a DM g describes the partitioning of Mg between calcite and solution. The concentration of Mg is measured relative to the element it is replacing, namely Ca. The solution from which the calcite precipitates is the cave drip water. Thermodynamic considerations predicts that DMg for inorganic precipitated calcite is temperature dependent (Garrels and Christ, 1965). The Mg/Ca ratio of ideal calcite will increase exponentially with ~3%/°C in the temperature interval between 0°C and 30°C (Lea et al., 1999). The temperature dependence of Mg in calcite has been 108 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. confirmed in both inorganic and organic calcite experiments, which will be discussed below. Roberts et al. (1998) list several equations for the homogenous partition coefficient Dm8 based on measurements of inorganic precipitated calcite in caves and in seawater. The equation DMg = 0.0017T + 0.0052 was found empirically by analyzing the Mg/Ca content of cave drip waters and the Mg/Ca content of deposited stalagmites utilizing Atomic Absorption Spectrophotometry (AAS). The equations DMg = 0.0007-T + 0.0033, DM g = 0.00042 T + 0.009, and DM g = 0.00043-T + 0.0011, were based on laboratory experiments of magnesian calcite overgrowth of seeds in standard seawater. The measurements were done utilizing AAS (Roberts et al., 1998). The homogenous distribution coefficient DMg has also been determined by measuring the incorporation of magnesium into the tests of different foraminifera species under controlled laboratory conditions (Delaney et al., 1985; Elderfield and Ganssen, 2000, Lea et al., 1999, and Numberg et al., 1996). Delaney et al (1985) pioneered the use of laboratory culture experiments to determine the relationship between the amount of Mg incorporated in the tests and temperature. Their study also included analysis of foraminiferal tests acquired from sediment traps and sediment core samples. The Mg/Ca contents of the tests were measured by AAS and found to be positively correlated with calcification temperature in all sediment traps and sediment core data sets, although the temperature-Mg/Ca relationship was not the exact same for different data sets. Their culture experiments gave similar results for the mean Mg/Ca ratio incorporated at 20°C and 30°C. Elderfield and Ganssen (2000) used Inductively Coupled Plasma (ICP) optical emission spectrometer measurements of foraminifera 109 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. picked from core-top samples. They correlated Mg/Ca with estimates of the calcification temperature (obtained from simultaneously recorded 51 8 Oc - 5 1 8 Ow ) and found the empirical equation: Mg/Ca = 0.52-e(0 '1T). Lea (1999) and Niimberg (1996) also used laboratory culture experiments of surface seawater to determine the processes that controls the uptake of Mg in the calcific tests of forminifera. Their studies showed that temperature is the leading mechanism controlling the Mg/Ca ratio. Lea et al. (1999) used an ICP mass spectrometer for the analysis and found the empirical equations: Mg/Ca = 1.36-e(0 '0 8 5 T ) and Mg/Ca = 0.53-e( O '1 C K r ) for two different types of foraminifera. Niimberg et al. (1996) used an alternative high-resolution electron microprobe analytical technique to measure the Mg/Ca ratios, and obtained the temperature relation Mg/Ca = 0.000284-10(00354T). It is unknown to what extent the homogeneous distribution coefficients, D M gS, from inorganic and organic precipitated calcite are comparable. However, all equations show that DM g is positively related with temperature either through a linear or an exponential relationship. Measuring variations in the Mg/Ca ratio of a stalagmite is therefore a potential recorder of temperature (assuming that the Mg/Ca ratio of the drip water has remained constant over time). Sr is incorporated into calcite via the same mechanisms as Mg. However, in contrast to Mg, the incorporation of Sr into calcite is independent of temperature (Katz et al, 1972)2. Goede and Vogel (1991) recognized that if other factors (e.g. salinity changes and temporal changes in the trace element concentration of the drip water) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. influence the incorporation of Mg and Sr into the stalagmite calcite in the same way, then the Sr content in calcite can be used to correct for the these factors. Goede and Vogel (1991) therefore suggested that the Mg/Sr ratio be used instead of the Mg/Ca content as a recorder of changes in temperature. 5.2 Factors controlling the Mg/Ca and Sr/Ca ratios of drip water The amount of Mg and Sr that will be incorporated into the stalagmite calcite depends on the Mg/Ca and Sr/Ca ratios of the cave drip water. The Mg/Ca and Sr/Ca ratios of the drip water in turn depends on the history of the rain water that falls above the cave and the interactions of the water with the soil and the bed rock above the cave as it percolates downwards to the cave. This is sketched in Figure 5.1. In general, both Mg2 + and Ca2 + are major ions dissolved in rainwater whereas Sr2 + accompanies these ions as a minor constituent. Mg2+ , Ca2 + , and Sr2 + in rainwater are derived primarily from particles in the air. The major sources of these atmospheric particles are: sea-salt aerosols (from marine sources) and soil dust (from terrestrial sources). The composition of sea-salt aerosols is similar to that of ocean water since no significant chemical fractionation occurs from the time of sea salt formation. Table 5.1 shows typical molar ratios Mg/Ca, Sr/Ca and Mg/Sr in different natural reservoirs. Sea- salt aerosols are enriched in Mg and Sr relative to Ca when compared to soil dust. Soil dust, on the other hand, contributes relatively more Ca to the rainwater than sea salt. 2 It should be noted that the partitioning of Sr into calcite may indirectly be dependent on temperature, since the temperature will affect the calcite precipitation rate, which has been shown to affect DS r (Lorens, 1981). However, in the simplified model this effect is neglected. I l l Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.1. Rain water history. 112 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The composition of the precipitation that reaches the ground is determined by the aerosol concentration during the rainout process (reactions between water and aerosols within the cloud during condensation) and the washout process (reactions between water and aerosols within the cloud during the fall to the ground). The average residence time in the atmosphere for sea salt is three days (Berner and Berner, 1987). Sea-salt aerosols can therefore be transported considerable distances to and over the continents. A decrease in the ratio (sea salt/soil dust) will be observed further inland. Mg2+ dominates over Ca2+ in rain near the ocean whereas Ca2+ dominates over Mg2 + further inland (compare the Mg/Ca ratio found in Table 5.1 for marine and continental rain). Likewise, the Sr/Ca ratio is expected to be higher in marine rain than in continental rain (unfortunately, no data value was found in the literature). Table 5.1. Typical molar ratios of Mg/Ca, Sr/Ca and Mg/Sr in natural reservoirs. Reservoir M olar ratio Mg/Ca Sr/Ca Mg/Sr Ocean water 3.11 0.019 164 Sea salt aerosols (3-11) (0.019) (164) Marine rain 2.67 — — Continental rain 1.13 0.0012 930 Carbonate 0.50 0.00067 750 Silicate 0.71 0.0029 250 Soil and soil dust 0.55 — — Surficial rock 0.60 — — Drip water 0.05 0.0005 100 Stalagmite Calcite 0.007 0.00013 54 Data compiled from: Baker et al., 1998; Bemer and Berner, 1987; Gascoyne, 1983; Negrel, 1993; Nozaki, 1997. When rainwater reaches the ground, it may react with substances and its composition may be altered. As an example, dust that has settled out on trees and other Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. types of vegetation is washed off and dissolved or carried as suspended material by the water. This is called throughfall. The content of Ca2+ may increase 2-3 times compared to the original rain. This estimate is based on throughfall in mountainous areas in New Mexico (Bemer and Bemer (1987)). Relatively less Mg and Sr are supplied by throughfall compared to Ca. Throughfall therefore tend to decrease the Mg/Ca and Sr/Ca ratios in the water. Some of the surface water becomes surface runoff, some of the water is recycled in evapo- transpiration, and some of the water infiltrates the soil as soil water and eventually reach the bedrock (we are only interested in this latter part). The process is also sketched in Figure 5.1. On the soil water’s journey through the soil it dissolves additional trace elements and carries with it suspended material that contains trace elements. The infiltrating water will alter its composition depending on the actual content of Mg, Sr and Ca in the minerals contained in the soil. In general, the Mg/Ca and Sr/Ca ratios will decrease as relatively more Ca is added to the soil water than Mg and Sr. After percolating through the soil the water comes into contact with the bedrock (in this case limestone above the cave). Once in the bedrock, chemical weathering will take place and limestone will be dissolved. The dissolution of limestone is facilitated by the acid that was created as a product of the organic matter decomposition in the soil. The chemical weathering results in significant changes in the composition of water. It is observed in Table 5.1 that the Mg/Ca and Sr/Ca ratios are decreased significantly in the resulting drip water (compared to the original rainwater). This is because huge amounts of Ca are added to the water from the limestone dissolution. The actual Mg/Ca and Sr/Ca ratios for the bedrock in the Budda Cave Area is not known but it is assumed that 114 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the composition of the bedrock is close to that of typical carbonate listed in Table 5.1. As the water enters the cave and becomes drip water, additional trace elements may finally be dissolved in the watcer from dust in the cave atmosphere. It is observed in T able 5.1 that increased inputs of salts to the atmosphere from sea-spray will tend to decrease the Mg/Sr ratio in rain water (typical continental rain water has a Mg/Sr ratio of 930 whereas the Mg/Sr ratio in sea salt aerosols is -164). Increased weathering of silicates and subsequent input of greater amounts of detrital material (dust and clay) to tine water will decrease the Mg/Sr ratio (silicates have a Mg/Sr ratio of -250). Varying amounts of carbonate (limestone) dissolution may also affect the final Mg/Sr ratio o f the drip water since the Mg/Sr ratio in carbonate is -750 compared to -930 in typical continental precipitation. There is disagreement in the literature about the result of varying ground water residence time on drip w ater Sr/Ca. Increased ground-water residence times can generally be expected when the climate is drier, whereas short residence times are related to more humid/rainy climate. Robert et al. (1998) suggest that increased residence time of ground watter above the cave will increase the Mg/Ca ratio due to incongruent dissolution of dolomite in the presence of calcite and because of differences in the dissolution rates of doLomite and calcite. At the same time, they suggested that the Sr/Ca ratio of the wateir would decrease as dolomite generally has lower Sr concentrations than calcite. According to Bar-Matthews et al. (1999) increasing groundwater residence time w ill cause a higher Sr/Ca ratio of the water entering the cave since larger amounts of .Sr will have time to accumulate in the soil water and be brought to the cave by the slow er flowing drip water. 115 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It should be noted that soils may have a large buffering capacity to attenuate variations in rainfall/weathering. 5.3 The climatic interpretation scheme for the Mg/Ca, Sr/Ca and Mg/Sr ratios in stalagmite SF-1 The simplified interpretation model that will be used in this chapter is based on the assumption that only two factors control the Mg/Sr, Sr/Ca and Mg/Ca ratios of SF-1. These two factors are the cave temperature and the residence time of the water in the soil above the cave. It is assumed that the temperature is positively correlated to the Mg concentration. On the other hand, it is assumed that temperature does not affect the Sr or the Ca concentration significantly. Furthermore, it is postulated that the Mg and Sr concentrations are affected by changing soil water residence times in a similar way. Therefore, changes in the Mg/Sr ratio will only reflect temperature changes. Increased Mg/Sr ratios will indicate increased temperatures. The simplified model assumes that increasing water residence times will increase the Sr concentration of the stalagmite calcite. Since the stalagmite consists of very pure carbonate, that is the Ca concentration remains almost constant in the entire stalagmite: [Ca]=39-40% (Li, pers. com.), the Sr/Ca ratio of the stalagmite presumably is positively correlated with water residence time. This is similar to the interpretation used by Bar-Matthews et al. (1999). Longer residence time is assumed to be the result of a drier climate and shorter residence time is assumed to be the result of a wetter climate. Increasing Sr/Ca ratios 116 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. can therefore be used to infer that the climate became drier and decreasing Sr/Ca ratios can be used to infer that the climate became wetter. Finally, the Mg/Ca ratio is positively correlated with both increasing temperatures and decreasing precipitation amounts. The correlation between the Mg/Sr and Mg/Ca ratios is approximately as strong as the correlation between the Mg/Ca and Sr/Ca ratios. This indicates that precipitation and temperature probably have approximately the same influence on the Mg/Ca ratio. This will be discussed further in the next section. Table 5.2 summarizes the climatic interpretation scheme for changing Mg/Sr, Sr/Ca and Mg/Ca values. Intuitively, one would expect that a combination of three factors would give 2-2-2=8 different possible combinations. However, this is not the case. Mathematically, if Mg/Ca decreases and Sr/Ca increases, then it is impossible for Mg/Sr to increase. Similarly, if Mg/Ca increases and Sr/Ca decreases, then it is impossible for Mg/Sr to decrease. As an example of the interpretation scheme presented in Table 5.2, assume that the Mg/Sr ratio increases, the Sr/Ca ratio decreases, and the Mg/Ca ratio increases. The increasing Mg/Sr ratio indicates that the climate got warmer. The decreasing Sr/Ca ratio indicates that the climate got wetter. The fact that the Mg/Ca ratio increases although Sr/Ca decreases indicates that the warming trend probably was more pronounced than the trend towards wetter climate. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.2. Climate interpretation of changing Mg/Sr, Sr/Ca and Mg/Ca ratios. Bold __________ _____ typeface means that factor is more dominant.________________ Interpretation: Mg/Sr T Sr/Ca T M g/C at Warmer and dryer Sr/Ca 1 M g/C at Warmer and wetter Mg/Ca i Warmer and wetter Mg/Sr -I Sr/Ca T M g/C at Colder and dryer Mg/Ca -I Colder and dryer Sr/Ca 1 Mg/Ca 1 Colder and wetter 5.4 Correlations between the Mg/Ca, Sr/Ca and Mg/Sr ratios Figure 5.2 shows the correlation between the Mg/Ca and Sr/Ca ratios of SF-1. It is observed here that different processes govern the relationship between Mg/Ca and Sr/Ca over time. Therefore, the time series is divided into four different climatic regimes, represented by the four different series shown in Figure 5.2. The regime representing the time period 40-4800a has a high correlation between Mg/Ca and Sr/Ca. It is therefore likely that these ratios are influenced by the same climatic factor in this regime. The same applies for the regime representing the time interval 8840-9400a. The model assumes that this factor is changing soil-water residence times, which in turn reflect changes in precipitation. The time intervals from 4800-8840a and from 15,990- 16,065a do not show a strong correlation, indicating that these ratios may have been influenced by different climatic factors. 118 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.3 depicts the correlation between the Mg/Ca and Mg/Sr ratios of SF-1. Again, the time series is divided into the same four climate regimes. The regimes representing the time periods 40-4800a, 8840-9400a, 15990-16065a all show a good correlation between Mg/Ca and Mg/Sr. It is therefore likely that these ratios are influenced by the same climatic factor in these regimes. In the model this factor is assumed to be temperature changes. The 4800-8840a interval does not show a strong correlation. This indicates that factors other than temperature may have dominated the Mg/Ca ratio. M g /C a versus S r/C a 0.00045 0.0004 R* « 0.1837 0.00035 ♦ 40*4800 □ 4 8 0 0 -8 8 4 0 O 8 8 4 0 -9 4 0 0 A 1 5 9 9 0 -1 6 0 6 5 0.0003 S 0.00025 3 0.0002 0.00015 ♦ ♦ R* « 0.6438 0.0001 -0.2551 0.00005 0.008 0.009 0.01 0.004 0.007 0.003 0.005 0.006 0 0.001 0.002 Mg/Ca [molar ratio] Figure 5.2. Correlation between M g/Ca and Sr/Ca. 119 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Since the correlation coefficient found in Figure 5.1 and Figure 5.2 are comparable for the time period 40-4800a, it can be assumed that temperature and changes in soil water residence times (precipitation) are of approximately equal importance in controlling the Mg/Ca ratio during this time. M g /C a versus M g /S r 70 • 60 * 7T 5 0 - ♦ 4 0 -4 8 0 0 □ 4 8 0 0 -8 8 4 0 O 8 8 4 0 -9 4 0 0 A 1 5 9 9 0 -1 6 0 6 5 CR2 = 0.0539 30 - R2 = 0.979 20 - 10 - = 0.7608 0.0 0 8 0.01 0.002 0.004 0 .0 0 6 0 M g/Ca [m o la r ratio ] Figure 5.3. Correlation between Mg/Ca and Mg/Sr. 5.5 The records of Mg/Ca, Sr/Ca and Mg/Sr from stalagmite SF-1 and their climatic interpretation. Figure 5.4 shows a record of the molar ratios Mg/Ca, Sr/Ca (actually 40*(Sr/Ca) and Mg/Sr. Figure 5.4 also depicts three vertical lines indicating separations between the four different climatic regimes. Strong oscillations and a good correlation between 120 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the Mg/Ca and Mg/Sr ratio and the Mg/Ca and Sr/Ca ratio (as shown in Figure 5.2 and Figure 5.3) characterize the time period from 40-4800a. It is likely that the Mg/Ca, Mg/Sr, and Sr/Ca ratios are depending strongly on the climate. The time period from 4800-8840a is characterized by small fluctuations around the same Mg/Ca, Mg/Sr, and Sr/Ca ratios as in the previous regime. It is therefore likely that the hydrological system has remained the same. The correlations between the different ratios are smaller (as shown in Figure 5.2 and Figure 5.3). This indicates that the influence from the climate was diminished and other factors had larger influence on the Mg/Ca, Mg/Sr, and Sr/Ca ratios than changes in temperature and soil water residence time. The third climate regime includes a transition where the ratios change radically. This is probably due to instability in the hydrological system. During this time period the record possibly shows changes in the hydrological system and does not reflect climatic changes. Finally, in the last climate regime, it appears the hydrological system was in another state. A possible reason for the changed hydrological conditions will be discussed later in this chapter. During this time period, the Mg/Sr and the Mg/Ca ratios could potentially be used to infer temperature changes as the correlation between these two ratios are good. However, the correlation between the Mg/Ca and the Sr/Ca in this interval is weak. Therefore, these ratios might not reflect changes in the same climatic factor (and hence not varying precipitation amounts). The climatic interpretation of the trace element records in Figure 5.4 was done using true/false logic. Initially, 5 sample-moving averages for Mg/Ca, Sr/Ca, and Mg/Sr were calculated to minimize the noise in the data. Differences between consecutive data points were calculated (e.g. AMg/Ca=((Mg/Ca)k+i-(Mg/Ca)k) and assigned the 121 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. value “+” if the difference was positive and assigned the value “ if the difference was negative. Temperature and rainfall index (+ and -) were then assigned to the molar ratios based on Table 5.3. Climate Regime 1 Climate Regime 2 C. R. 3 Climate Regime 4 < > < ► < > < > 60 0.02 0.018 ■ •70 0.016 ■ 60 < • O M g/Sr 0.014 ■ « o E • 50 0.012 - 40 O 0.01 - CO ■g 0.008 - 30 « < Q s 0.006 • 2 0 0.004 - 0.002 - 1000 2 000 3 000 4 0 0 0 5 0 0 0 6 0 0 0 7000 8 000 9 000 159 5 0 16000 16050 16100 0 ya Figure 5.4. Molar ratios of Mg/Ca, Sr/Ca and Mg/Sr. Table 5.3. Temperature and rainfall index for trace element data. AMg/Sr ASr/Ca AMg/Ca Temperature index Rainfall Index Climate trend + + + + - Warmer and drier - + + - - Colder and drier + - + + + Warmer and wetter - + - - - Colder and drier + - - + + Warmer and wetter - - - - + Colder and wetter 122 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.5 shows a plot of the temperature and rainfall indices for the time period 0 to 5000a based on the described trace element interpretation. Figure 5.5 also shows the temperature and rainfall index based on interpretation of the stable isotope data. Temperature and rainfall index for the stable isotope records were based on table 5.4 that were constructed based on Table 4.1 and Table 4.2. The temperature and rainfall indices were found by applying true/false logic to the differences in A51 3 C and AS1 sO on the 5 sample moving averages of the stable isotope data. A plot of the raw (coarse resolution) stable isotope data for stalagmite SF-1 was shown in Figure 4.2 in Chapter 4. Ta jle 5.4. Temperature and rainfall indices based on stable isotope data. ASl3C A8 lsO Rainfall index Temperature Index Climate trend + + - + Drier and warmer + - - - Drier and colder - + + + Wetter and warmer - - + - Wetter and colder Figure 5.5 shows that the interpretations of the climate trends based on trace elements content and stable isotopes correlate well. The temperature index based on isotope and on trace element are consistent 63% of the time. The precipitation index based on isotope and on trace element are consistent 78% of the time. The fact that the interpretations of the two proxies are correlated is an indication that both interpretation schemes are based on reasonable assumptions. A rough estimate of the temperature range for the last 5000 years can be calculated based on the M g/Cavite data shown in Figure 5.4, using the assumption that 123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the drip water Mg/Ca has a constant value of 0.05 and the equation DMg = 0.0017-T + 0.0052 (described earlier). The observed range of the Mg/Cacaicite ratio is 0.0040-0.0093. This corresponds to 106-44°C = 62°C. This is far beyond the range of expected temperatures. Obviously other factors than temperature (e.g. precipitation) influences the Mg/Ca significantly and complicates a quantitative interpretation. A similar attempt to calculate the temperature range based on the measured 5I80 data for the last 5000 years using the temperature relation 0.268%o°C'1 results in a temperature range of AT=11°C. This is beyond the expected range (it was not attempted to infer absolute temperature values from the 5I80 -T relationship). Evidently, an other factor(s) also influence the SI80 signal significantly. Hence, temperatures will only be inferred qualitatively (relative: cooling and warming) in this study. In a few cases, where neither the temperature nor the rainfall index agreed, a climatic interpretation was omitted. The temperature and rainfall indices are plotted for the time interval 5000 to 9400a in Appendix 5A. The correlation between the interpretations based on isotope data and trace element data is not very strong during this time interval. In the time interval 5000 — 8400a, the temperature index based on isotope and on trace element are consistent 50% of the time. The precipitation index based on isotope and on trace element are consistent 65% of the time. In the time interval 8400-9400 these percentages are 0% for temperature and 75% for precipitation. As discussed previously, the influence from the climate on the trace element ratios was diminished in climate regime 2. Hence, it is expected that the correlation should be weak. Climate regime 3 includes a transition due to changed hydrological conditions. 124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The influence from the climate is therefore probably insignificant. This is consistent with the lack of correlation found between interpretations from trace elements and stable isotopes. It should be noted that climate regime 3 consists of only 8 data points. T em perature and ra in fa ll In d ices fro m trace elem ents and Isotope d ata Index: W=Warming C=Cooling 1= Increasing rainfall D= Decreasing rainfall ■ temperature (isotope) • rainfall (isotope) g a temperature (trace) ♦ rainfall (trace) w & o 7 A T 5 00 ▲ ▲▲ W & D ■■■■■■ A l A i A A O • • ♦ ♦ 1000 • • • • • • ♦♦♦♦♦♦ 1500 ■ ■ A. A WjC & !& d Id i.A * • • • • ♦ ♦♦♦♦ 2000 A . A A A A A A A A ♦ 2$00 C & I ■■■■■■ A A A A A A A A 3 0 0 0 A A A A A 441 ♦ ♦4 W & D 5 00 400 0 • • • • 444444 \ A k A A A ▲ ▲▲ 4 5 'X) W & D • 4 4 A m 5000 AR [a ] Figure 5.5. The temperature and rainfall indices based on trace elements and isotope data for 0-5000a. The stalagmite SF-1 was not analyzed for trace elements in the interval corresponding to 70-470 because this part of the stalagmite had already been used for TIMS-U-series dating and no material was left. As described earlier, there is no significant correlation between Mg/Ca and Sr/Ca for climate regime 4 (= calcite deposition prior to the hiatus). The large shift towards increased values of Sr/Ca and the apparent lack of correlation between Mg/Ca and Sr/Ca indicates that the hydrological system probably was in another state at this 125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. time. This is supported by the large shift observed in the SIjC and S1 8 0 at the same time. The increased Sr/Ca values for the part of the stalagmite that was older than 16,000 years was confirmed by XRD3 performed on samples collected from each side of the hiatus (as shown in Appendix 5B). The XRD revealed that aragonite crystals formed in the stalagmite below 16.4 cm’s depth from the surface (-16 ka) (although the majority of crystals were calcite - refer to Table l.B in Appendix 5B). The formation of aragonite is more likely when the Sr concentration in the drip water is higher (Li, pers. com.). Both weathering of carbonate and silicate as well as an increased salt input can increase the absolute content of Sr in drip water. A large increase in the relative content of Sr to Mg, however, is most effectively obtained by weathering of silicates or increasing the sea water portion, since these have higher Sr/Mg ratios than carbonates. This is consistent with the observed increase in the amount of Na-ions (which is the result of both increased silicate and salt input) from about 0 .0 1 % after the hiatus (< ~16ka) to 0.02% before the hiatus (> 9.4 ka). The fact that the stalagmite contains a iarger amount of aragonite before the hiatus and is composed of less pure calcite (Appendix 5B, Table 5B.1) points to a greater input of suspended detrital material being the reason for the increased Sr/Ca ratio. An increased amount of detrital material should also cause a higher Al content, which was not observed (Al was always below 0.01%). This can be an artifact of the ICP70 Aqua Regia method. The ICP70 Aqua method is based on weak acid digestion, which does not necessarily dissolve all detrital material in a carbonate sample, causing less Al to be analyzed than what is actually present in a 3 XRD (X-ray diffraction) is a method that can be used to determine the arrangement of atoms within a crystal, by measuring the patterns of scattered X-rays after they pass through a crystalline substance minerals can be identified according to their “specific fingerprints”. 126 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sample (White, pers. com.). More research is needed before the cause for the huge change in Sr/Ca can be fully understood. The large change in Sr/Ca has also been observed in speleothem records from other parts of the world at about 16 ka (Bar- Matthews et al., 1999). It is suggested that the strontium isotope ratio: 8 7 Sr/8 6 Sr could be investigated in the future. The 8 7Sr/8 6 Sr ratio in speleothems are powerful tracers of the origins of paleo-waters, reflecting changes in environmental factors such as storm tracks, hydrological routing, and soil-water-rock interactions (Banner et al, 1996; Herut et al, 1993). 5.6 Summary This chapter has investigated the use of trace element ratios (Mg/Ca, Sr/Ca, and Mg/Sr) incorporated into stalagmites as an alternative proxy to stable isotopes to infer changes in paleoclimate from cave deposits. A relatively simple interpretation model was developed in which temperature changes were assumed to be the only factor controlling the Mg/Sr ratio, whereas changes in soil water residence were assumed to be the only factor controlling Sr/Ca. In the model, the Mg/Ca ratio is controlled both by changes in temperature and soil water residence time. Changes in soil water residence time in turn would reflect changes in the amount of precipitation falling above the cave. Residence times are assumed to be longer in a drier climate and shorter in a wetter climate. The model was used on a record of trace element ratios obtained from stalagmite SF-1. It was realized that trace element deposition in SF-1 has occurred in four different regimes. Different climate dependence and different hydrological conditions characterized the regimes. The strongest correlation between the climate and the trace 127 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. elements were found from 0-4800a. A correlation between the interpretation based on the isotopic records and the trace element records exist for the temperature 63% of the time and for the precipitation 78% of the time in this climate regime. The reliability of paleoclimate reconstruction is expected to be enhanced by the use of a multiproxy approach. It seems that agreement between the trace element and isotopic records should raise the confidence of paleoclimate interpretations. It was also observed that variations in trace element content could be caused by factors other than temperature and rainfall. A significant change in the hydrologic conditions in the cave area earlier than 16ka was indicated by a significant shift in the observed relationship between Mg/Ca and Sr/Ca and by a dramatic increase in the Sr/Ca ratio. Increased amounts of dissolved salts and/or suspended detrital material carried by the drip water are suggested as possible causes for these changes. More research is required to identify all hydrologic parameters and their relative importance on influencing the incorporation of Mg and Sr into the stalagmite. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6. Summary A multi-proxy study of climate variability in central China has been made. This study was based on high-resolution growth rate, isotopic composition, and trace elements in the stalagmite, SF-1, from Budda Cave. Accurate chronologies are prerequisites for all high-resolution climatic studies. TIMS 2 3 0 Th/2j4U dating and 2 1 0 Pb dating have been performed previously on stalagmite SF-1. Based on the radiometric data and a constant growth rate assumptions, the following chronology was established: the total age of SF-1 was 16.1 ka and a growth hiatus existed from -9.4 -16.0 ka. The TIMS 2 3 0 Th/2 3 4 U dating showed that the growth rates were: -0.089 mm/year in the upper part (0-1.6 cm), -0.0163 mm/yr in the middle part (1.6-16.5 cm), and -0.72 mm/yr in the lower part (16.5-22.2 cm) of the stalagmite. The 2I0Pb dating showed that the upper ~16mm grew with an average rate of -0.066 ± 0.005mm/yr. Previous work has shown that stalagmites that form in areas similar to the Budda Cave area will typically exhibit an annually banded structure. Counting annual lamination was therefore used as a complimentary method for dating since it had the potential of providing a chronology with a one-year resolution. Annual laminae can either be counted using a normal microscope with visible light transmission through a thin section or with UV light to excite luminescence in a sample. Lamination counting of SF-1 was performed under visible light for the upper 144 layers. The counting showed that the growth rate varied between 0.029 mm/yr and 0.278 mm/yr (average: O.lllmm/yr). The assumption that layers formed annually was supported also from a 129 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. close match between the average growth rate estimates from the lamination counting and the 2l0Pb dating method. The growth rate of a stalagmite varies from year to year as a function of climatic and environmental factors. Although many different factors may control the deposition of stalagmite calcite, it was important to compare the lamination thickness with instrumental data of temperature and precipitation to investigate if a simple correlation exist between either of these factors. Unfortunately, no instrumental data exists for the cave site. Furthermore, an investigation showed that the weather could not be directly compared between stations at various distances and directions from the cave. A bilinear interpolation was made to obtain the longest and most accurate record for the actual cave site. Data from the two databases WWdisc and IAEA WMO were utilized. The actual factors investigated were: average annual temperature, average summer temperature, total annual precipitation, total summer precipitation, maximum monthly precipitation, and numbers of month with precipitation above average. No correlation was found with any of these factors. Therefore, it was concluded that an interpretation of paleo-climate variability could not be made based on the growth history of the stalagmite SF-1 alone. However this information could potentially be fused with other proxies. Two high-resolution isotope records (51 3 Cc and 5l8Oc) from stalagmite SF-1 have been used to infer changes in paleoclimate in Central China. The first record that covers the last 0-9.4ka has a resolution of -100 years. The second record has a resolution of -1-3 years and spans the interval from 0-1266a. 130 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The temperature interpretation of SI8Oc was supported by the establishment of a “temperature-precipitation-Sl8Op -reIationship” based on modem measurements of surface temperatures, precipitation amounts, and 8I8Op from weather stations in the proximity of the Budda Cave. From this relationship it was found that the effect of temperature on 51 8 Op was greater (0.516%o°C*1 ) than the generally accepted theoretical effect of temperature on water-calcite fractionation (-0.2272%o°C'1 ). Therefore, the positive relationship between 51 8 Oc and temperature for stalagmites forming in caves in Central China was confirmed. It is postulated that drier and warmer climate will increase the 8I8 Oc value, whereas wetter and colder climate will tend to decrease the 8I8Oc value. Short term changes (<10 years) in SI8 Oc values reflect changes in precipitation and longer term changes (>50 years) reflect changes in temperature. Furthermore, it was assumed that increasing 81 3 Cc values represent a shift towards drier (and warmer) climate whereas decreasing S1 3 Cc values indicate that the climate is getting wetter (and colder). Specific climatic intervals were identified from the isotope records including a very dry period centered at ~7ka, the Medieval warm Period (1030-520a), and the Little Ice Age (520-150a). The Medieval warm period and the Little Ice Age were also identified in the 0-1266a record. The 1-3 years resolution provided very detailed knowledge of the climate fluctuations during the Medieval warm period and the Little Ice Age. As an example, the coldest decade within the Little Ice Age was identified. This decade matched historical records. 131 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The SF-1 isotopic records had several cycles imbedded within them. The longest cycle was a 1567 year cycle that has previously been identified in the monsoonal climate in Asia (in the India monsoon system) and in the Arctic climate in the Northern Hemisphere. It is believed that this cycle is of global extent and is caused by an internal forcing factor (probably oscillations in the global thermohaline circulation system). Cycles with periods of 22, 12, and 9.6 years were present in the isotope records. These cycles are found in many other paleoclimate records and reflect the influence of Sun spots and the lunar orbit on Earth’s climate. A cycle with a period of 33 years was also identified in the records. It is likely that this is the Bruckner cycle identified in the European climate records. It is believed that this cycle is caused by large synoptic-scale pressure anomalies. A period on 7.2 years was also included in the record. Tree ring data from the same area in central China also contains this period. The origin of this local cycle is unknown. Finally, cycles of 469, 140, 115, 51, and 41 years were identified in the record. It was not possible to match these with cycles in the literature. These cycles have not been identified previously and their origin is unknown. The incorporated amount of Mg, Ca, and Sr in stalagmite SF-1 was used as an alternative proxy to stable isotopes to infer changes in paleoclimate. The ratios Mg/Ca, Sr/Ca, and Mg/Sr were interpreted based on a very simple model. This model assumes that the only factor controlling the Mg/Sr ratio is temperature. An increasing Mg/Sr ratio indicates a warmer climate and a decreasing Mg/Sr ratio indicated a cooler climate. It is also assumed that the only factor controlling the Sr/Ca ratio is changing soil water residence time (equivalent to changing precipitation amounts). Increasing 132 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sr/Ca ratio indicates a drier climate and decreasing Sr/Ca ratio is a reflection of a wetter climate. The Mg/Ca ratio is influenced by both temperature and precipitation variations. The trace element interpretation of SF-1 showed that the trace element deposition had occurred in four different climatic regimes. The strongest correlation between the climate and the trace elements were found from 0-4800a. A correlation between the interpretation based on the isotopic records and the trace element records existed for the temperature in 63% of the time and for the rainfall in 78% of the time in this climate regime. It was also observed that variations in trace element content could be caused by factors other than temperature and rainfall. A significant change in the hydrologic conditions in the cave area earlier than 16ka was indicated by a significant shift in the observed relation between Mg/Ca and the Sr/Ca and by a dramatic increase in the Sr/Ca ratio. Increased amounts of dissolved salts and/or suspended detrital material carried by the drip water are suggested as possible causes for these changes. The combination of the climate interpretations based on the isotope and the trace element records from SF-1 can be used to improve our understanding of the paleoclimate variability in central China. The resolution and the time-span of the records are unprecedented in this region of China. Stalagmites have proven to be excellent continental paleoclimate recorders and are widely distributed on the continents. Furthermore, they are superior in some respects to other achives (e.g., longer records than tree rings and higher resolution than ice cores and marine records). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Combining the results of this high-resolution study with paleoclimate records from other parts of China will open new opportunities for predicting and modeling climate in this region. 134 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. References Anbar, A., 1998. ICP-MS Primer. 05-Nov-1999 <http://www.earth.rochester.edu/icpms/Primer.html>. Anthes, R.A., 1997. Meteorology. 7th edition, Prentice Hall, New Jersey, 214pp. Baker, A., Smart, P.L., Edwards, R.L., Richards, D.A., 1993. Annual growth banding in a cave stalagmite. Nature 364, 518-520. Baker, A., Barnes, W.L., Smart, P.L., 1996. Speleothem luminescence intensity and spectral characteristics: Signal calibration and a record of paleovegetation change. Chemical Geology 130, 65-76. Baker, A., Ito, E., Smart, P.L., McEwan, R.F., 1997. Elevated and variable values of t3C in speleothems in a British cave system. Chemical Geology 136, 263-270. Baker, A., Genty, D., Dreybrodt, W., Barnes, W.L., Mockler, N.J., Grapes, J., 1998. Testing theoretically predicted stalagmite growth rate with recent annually laminated samples: Implications for past stalagmite deposition. Geochimica et Cosmochimica Acta 62, 393-404. Baker, R.G., Gonzalez, L.A., Raymo, M., Bettis IH, E.A., Reagan, M.K., Dorale, J.A., 1998. Comparison of multiple proxy records of Holocene environments in the midwestem United States. Geology 26, 1131-1134. Banner, J.L., Musgrove M., Asmerom, Y., Edwards, R.L., Hoff, J.A., 1996. High- resolution temporal record of Holocene ground-water chemistry: Tracing links between climate and hydrology. Geology 24, 1049-1053. Bar-Matthews, M., Ayalon, A., Matthews, A., Sass, E., Halicz, L., 1996. Carbon and oxygen isotope study of the active water-carbonate system in a karstic Mediterranean cave: Implications for paleoclimate research in semiarid regions. Geochimica et Cosmochimica Acta 60, 337-347. Bar-Matthews, M., Ayalon, A., Kaufman, A., 1997. Late Quaternary paleoclimate in the Eastern Mediterranean region from stable isotope analysis of speleothems at Soreq Cave, Israel. Quaternary Research 47, 155-168. Bar-Matthews, M., Ayalon, A., Kaufman, A., Wasserburg, G.J., 1999. The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq cave, Israel. Earth Planetary Science Letters 166, 85-95. 135 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Baskaran, M., Hiffe, T., 1993. Age determination of recent cave deposits using excess 210Pb — A new technique. Geophysical Research Letter 20, 603-606. Bemer, E.K., Bemer, R.A., 1987. The global water cycle, Geochemistry and environment. Prentice-Hall, Inc., New Jersey, 397pp. Bradford, T., Cook, M.N., 1997. Inductively Coupled Plasma (ICP). Daniel Gallagher 22-Dec-1997 <http://www.ce.vt.edu/enviro2/smprimer/icp/icp.html>. Bradley, R.S., 1999. Paleoclimatology, reconstructing climates of the Quaternary. Academic Press, 2n d edition, San Diego, California, 613 pp. Broecker, W.S., Olson, E.A., Orr, P.C., 1960. Radiocarbon measurements and annual rings in cave formations. Nature 185, 93-94. Bryant, E., 1997. Climate process and change. Cambridge University Press, New York, 209pp. Buhmann, D., Dreybrodt, W., 1985. The kinetics of calcite dissolution and precipitation in geologically relevant situations of karst areas, 1. Open system. Chemical Geology 48, 189-211. Burton, E.A., Walter, L.M., 1991. The effects of PCO 2 and temperature on magnesium incorporation in calcite on seawater and MgCli-CaCh solutions. Geochimica et Cosmochimica Acta 55, 777-785. CEOS IDN datasets, 1995. Climate data bases of the People’s Republic of China for 1841-1988 CDIAC TR055. CEOS-IDN 24-Nov-1999 <http://www.neonet.nl/ceos-idn/datasets/CDIAC_TR055.html>. Cerling, T.E., 1984. The stable isotopic composition of soil carbonate and its relationship to climate. Earth and Planetary Science Letters 71, 229-240. Cerling, T.E., Quade, J., Wand, Y., Bowman, J.R., 1989. Carbon isotopes in soils and paleosols as ecology and paleoecology indicators. Nature 341, 138-139. Clemens, S., Prell, W., 1991. Late Quaternary forcing of Indian summer-monsoon winds: a comparison of Fourier model and general circulation model. Journal of Geophysical Research 96, 22,683-22,700. Delaney M., Be, A., Boyle, E., 1985. Li, Sr, Mg, and Na in foraminiferal calcite shells from laboratory culture, sediment traps, and sediment cores. Geochimica et Cosmochimica Acta 49, 1327-1341. 136 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Dorale, J.A., Gonzalez, L.A., Reagan, M.K., Pickett, D.A., Murrell, M.T., Baker R.G., 1992. A high-resolution record of Holocene climate change in speleothem calcite from Cold Water Cave, Northeast Iowa. Science 258, 1626-1630. Dorale, J.A., Edwards, R.L., Ito, E., Gonzalez, L., 1998. Climate and vegetation history of the midcontinent from 75 to 25 ka: A speleothem record from Crevice Cave, Missouri, USA. Science 282, 1871-1873. Dreybrodt, W., 1980. Deposition of calcite from thin films of natural calcareous solutions and the growth of speleothems. Chemical Geology 29, 89-105. Dulinski, M., Rozanski, K., 1990. Formation of 13C/12C isotope ratios in speleothems: A semi-dynamic model. Radiocarbon 32, 7-16. Elderfield, H., Ganssen, G., 2000. Past temperature and 8lsO of surface ocean waters inferred from foraminiferai Mg/Ca ratios. Nature 405, 442-445. Fairbanks, R.G., 1989. A 17,000 year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature 342, 637-642. Garrels, R.M., Christ, C.L., 1965. Solutions, minerals and equilibria. Freeman, Cooper & Co., San Francisco, 450pp. Gascoyne, M., 1983. Trace-element partition coefficients in the calcite-water system and their paleoclimatic significance in cave studies. Journal of hydrology 61, 213-222. Gascoyne, M., 1992. Paleoclimate determination from cave calcite deposits. Quaternary Science Reviews 11, 609-632. Goede, A., Vogel, J.C., 1991. Trace element variations and dating of a late Pleistocene Tasmanian speleothem. Palaeogeography, Palaeoclimatology, Palaeoecology 88, 121-131. Grootes, P.M., Stuiver, M., 1997. Oxygen 18/16 variability in Greenland snow and ice with 10'3 to 105-year time resolution. Journal of Geophysical Research 102, 26455-26470. Hendy, C.H., 1971. The isotopic geochemistry of speleothems — I. The calculation of the effects of different modes of formation on the isotopic composition of speleothems and their applicability as palaeoclimate indicators. Geochimica et Cosmochimica Acta 35, 801-824. 137 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Herut, B., Starinsky, A., Katz, A., 1983. Strontium in rainwater from Israel: Sources, isotopes and chemistry. Earth and Planetary Science Letters 120, 77-84. Hoefs, J., 1997. Stable Isotope Geochemistry, 4th edition, Springer-Verlag Berlin, 201pp. Hughes, M.K., Xiangding, W., Xuemei, S., Garfin, G.M., 1994. A preliminary reconstruction of rainfall in North-central China since A.D. 1600 from tree-ring density and width. Quaternary Research 42, 88-99. Jacoby, G.C., D’Arrigo, R.D., Davaajamts, T., 1996. Mongolian tree rings and 20th - century warming. Science 273, 771-773. Katz, A, Sass, E., Starinsky, A., Holland, H.D., 1972. Strontium behavior in the aragonite-calcite transformation: an experimental study at 40-98°C. Geochimica et Cosmochimica Acta 36, 481-496. Katz, A., 1973. The interaction of magnesium with calcite during crystal growth at 25- 90°C and one atmosphere. Geochimica et Cosmochimica Acta 37, 1563-1586. Kerr, R.A., 2000a. Global warming: Draft report affirms human influence. Science 288, 589a-590a (in news of the week). Kerr, R.A., 2000b. Greenhouse warming: Dueling models: Future U.S. climate uncertain. Science 288, p. 2113 (in news of the week). I R Koster, R. D., de Valpine, D.P., Jouzel, J., 1993. Continental water recycling and Ha O concentrations. Geophysical Research Letters 20, 2212-2218. Ku, T-L., Li, H.-C., Stott, L.D., 1998. Decadal climatic reconstruction from stable isotope records of speleothem in Shihua Cave, China: A summer monsoon domain. In: Wilson, R.C., Tharp, V.L. (Eds.), Proceedings of 14th Annual Pacific Climate (PACOM) Workshop, April 6-9, 1997. Interagency Ecological Program, Technical Report 57. California Department of Water Resources, pp. 1-14. Ku, T.-L., Li, H.-C., 1998. Speleothems as high-resolution paleoenvironment archives: Records from northeastern China. Proceedings of the Indian Academy of Science (Earth Planetary Science) 107, 321-330. Lea, D., Mashiotta, T., Spero, H., 1999. Controls on magnesium and strontium uptake in planktonic foraminifera determined by live culturing. Geochimica et Cosmochimica Acta 63, 2369-2379. 138 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Li, H.-C., Ku, T.-L., Chen, W.-J., Jiao, W.-Q., Zhao, S.-S., Chen, T.-M., Li, T.-Y., 1996. Isotope studies of Shihua cave, Beijing -II: Radiocarbon dating and age correction of stalagmite. Seismology and Geology 18, 329-338. Lorens, R.B., 1981. Sr, Cd, Mn and Co distribution coefficients in calcite as a function of calcite precipitation rate. Geochimica et Cosmochimica Acta 45, 553-561. McDermott, F., Frisia, S., Huang, Y., Longinelli, A., Spiro, B., Heaton, T., Hawkesworth, C.J., Borsato, A., Keppens, E., Fairchild, I. J., Klass van der Borg, Verheyden, S., Selmo, E., 1999. Holocene climate varibility in Europe: Evidence from 6lsO textural and extension-rate variations in three speleothems. Quaternary Science Reviews 18, 1021-1038. Ming, T., Dongsheng, L., Xiaoguang, Q., Hua, Z., Tieying, L., Shusen, Z., Li, H.C., Jinbo, L., Xiangyang, L., 1997. Preliminary study on the data from microbanding and stable isotopes of stalagmites of Beijing Shihua Cave. Carsologica Sinica 16, 1-10 (in Chinese with English abstract). Mosley-Thompson, E., Thompson, L.G., Dai, J., Davis, M., Lin, P.N., 1993. Climate of the last 500 years: high resolution ice core records. Quaternary Science Reviews 12, 419-430. Mucci, A., 1987. Influence of temperature on the composition of magnesian calcite overgrowths precipitated from seawater. Geochimica et Cosmochimica Acta 51, 1977-1984. Negrel, P., Allegre, C.J., Dupre, B, Lewin, E., 1993. Erosion sources determined by inversion of major and trace element ratios and strontium isotopic ratios in river water: The Congo Basin case. Earth and Planetary Science Letters 120, 59-76. Njitchoua, R., Sigha-Nkamdjou, L., Dever, L., Marlin, C., Sighomnou, D., Nia, P., 1999. Variations of the stable isotopic compositions of rainfall events from the Cameroon rain forest, Central Africa. Journal of Hydrology 223, 17-26. NOAA, 2000. General paleoclimatology data sets. NOAA Paleoclimatology Program 27-July-2000 <http://www.ngdc.noaa.gov/paleo/miscpcl.html>. Nozaki, Y., 1997. A Fresh Look at Element Distribution in the North Pacific Ocean. American Geophysical Union, EOS Transactions 78, no. 21, p. 221. Niimberg, D., Bijma, J., Hemleben, C„ 1996. Assessing the reliability of magnesium in foraminiferal calcite as a proxy for water mass temperatures. Geochimica et Cosmochimica Acta 60, 803-814. 139 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Railsback, L.B., Brook, G.A., Chen, J., Kalin, R., Fleisher, C.J., 1994. Environmental controls on the petrology of a late Holocene speleothem from Botswana with annual layers of aragonite and calcite. Journal of Sedimentary Research A64, 147-155. Rindsberger, M., Margaritz, M, Carmi, I., Gilad, D., 1983. The relation between air mass trajectories and the water isotope composition of rain in the Mediterranean Sea area. Geophysical Research Letters 10,43-46. Roberts, M.S., Smart, PX ., Baker, A., 1998. Annual trace element variations in a Holocene speleothem. Earth and Planetary Science Letters 154, 237-246. Schleser, G.H., Helle, G., Liicke, A., Vos, H., 1999. Isotope signals as climate proxies: the role of transfer functions in the study of terrestrial archives. Quaternary Science Reviews 18, 927-943. Shopov, Y.Y., Ford, D.C., Schwarcz, H.P., 1994. Luminescent microbanding in speleothems: High resolution chronology and paleoclimate. Geology 22, 407- 410. Sirocko, F., Garbe-Schonberg, McIntyre, A., Molfino, B., 1996. Teleconnections between the subtropical monsoons and high-Iatitude climates during the last deglaciation. Science 272, 526-529. Tanahara A., Taira, H., Yamakawa, K., Tsuha, A., 1998. Application of excess 2 1 0 Pb dating method to stalactites. Geochemical Journal 32, 183-187. Torrence, C., Compo, G., 1999. A practical guide to wavelet analysis. Bulletin of the American Meteorological Society 79, 61-78. Tungsheng, L., Ming, T., Xiaoguang, Q., Shusen, Z., Tieying, L., Jinbo, L, De er, Z., 1997. Discovery of microbedding in speleothems in China and its significance in the study of global change. Quaternary Sciences 1, 41-53 (in Chinese with English abstract). Wang, L., Samthein, M., Erlenkeuser, H., Grootes, P.M., Grimalt, J.O., Pelejero, C., Linck, G., 1999. Holocene variations in Asian monsoon moisture: a bidecadal sediment record from the South China Sea. Geophysical Research Letters 26, 2889-2892. WeatherDisc Associates, 1994. World WeatherDisc-software. Diskette. Vers. 3.1. Inc. 4584 NE 89th - Seattle, WA 98115 USA. Yonge, C.J., Ford, D.C., Gray, J., Schwarcz, H.P., 1985. Stable isotope studies of cave seepage water. Chemical Geology 58, 97-105. 140 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Zeiss, 2000. Microscopy Group. Carl Zeiss 1996-99 <http ://www.zeiss .com/micro/products/>. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 1A Photos from the Budda Cave area Figure 1A.1. The Budda Cave Area. Figure 1A.2. Agricultural activities in the Budda Cave Area. 142 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 2A 2 1 0 Pb dating The 2I0Pb dating technique can be used to determine accurate growth rates of young stalagmites (<100 years) (Baskaran and Iliffe, 1993; Li and Ku, 1996 and Tanahara et al., 1998). 2l0Pb is a radioactive daughter product in the decay chain of 2 3 8 U: “ O’ ......."> “ J & W , " > ........M ° ^ ( W 1 3 8 , rf) ....... The measurable total 2 1 0 Pb concentration at depth (x) in a recently deposited CaCC>3 stalagmite is the sum of parent-supported 2 1 0 Pb and excess 2I0Pb: 2l0Pb'o'Ax)= 2 1 0 s u p p o r te d + ^ K scess^ ) (2A1) Parent-supported 2 1 0 Pb is produced from the decay of 2 2 6 Ra which is incorporated into the crystal lattices during the growth of the stalagmite. Excess 2 1 0 Pb is incorporated into the crystal lattices of CaCC>3 during the growth of the stalagmite, after production by the (in-situ) decay of the noble gas nuclide Rn in cave air and in drip water. Hence, 2 1 0 Pb£ X t- eji- is not in equilibrium with 2 2 6 Ra in the CaCOs precipitate. There T O T are two sources for Rn in cave air: 1) Rn that has diffused into the cave from the bedrock where it was produced from 2 2 < 5 Ra decay, and 2) 2 2 2 Rn that has outgassed from waters flowing or dripping into the cave. The latter is according to Baskaran and Iliffe (1993) and Li and Ku (1996) presumably the most important source of 2 2 2 Rn in cave air. Ground and drip waters have very high concentrations of 2 2 2 Rn derived from the recoil of Rn atoms into the aqueous phase during its production from the U-series decay 143 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. chain. Baskaran and Iliffe (1993) measured 2 1 0 Pb concentrations in cave soil and found concentrations that were negligible to those of young speleothems. They concluded that the contribution of 2 1 0 Pb from 22 2 Rn present in cave air is insignificant compared to the contribution from 2 2 2 Rn in the drip water. The 2l0Pbexces s incorporated into the CaC0 3 lattice of the stalagmite is assumed to be constant over time. By measuring the total activity of 2 1 0 PbW tai in subsequent layers of the stalagmite and subtracting a constant contribution of 2l0Pbsu p p o rted from 2 1 0 Pbtotal, information about the activity of 2 1 0 Pb„c e ^ is obtained as a function of depth. This knowledge can be used to derive growth rates of stalagmites that grew within the last 100 years. The activity of 2l0Pbt. . rC e^- should theoretically decrease with depth in an exponential manner according to the function: 0 0 = 2 l0 Pb0 _ „ • e-“ (2A2) with the activity of 2 1 0 Pbe . V C e^ being close to zero at a depth in the stalagmite that formed 5 half lives (-110 -115 years) earlier. In equation A2.2 ‘a’is : a = — ----- (2 A3) T • ? M 2,Pb— 2\Q ° 2l0Pb^c„i(x) is the activity of 2 1 0 Pbe x c e ^ in the layer at depth (x), 2 1 °Pbo, ex cess is the activity of 2 1 0 Pbe . rc e^- in the uppermost layer, Ti/2, p b - 2 1 0 is the half life of 21 0 Pb (=22.3 yr.), and S is the growth rate (mm/yr.). 1 2I0Pb supported by the decay o f 2 2 6 Ra is assumed to be independent with depth in the stalagmite. 144 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The growth rate for the stalagmite can be found by fitting an exponential function to the data for the 2 l0P b exress versus x. The application of 21 0 Pb dating technique on SF-1. The excess 2l0Pb dating method was used on stalagmite SF-1 from Budda cave in Central China by Li and Ku (unpublished data). Twelve CaCOs samples with weights ranging from 0.52 to 1.43 g were sampled from the upper 29 mm of the stalagmite. Each CaC0 3 sample was leached with 7 M H N 03 2 and a known amount of 2 0 9 Po spike was added as a yield monitor. The solutions were then evaporated to dryness on a hot plate. The residues were dissolved with 10 ml 1M HC1 and ascorbic acid was added to the solutions. Silver planchets were placed in the solution and 2 0 9 Po and 2 1 0 Pb were plated at a temperature of about 80°C for approximately 3 hours. Later, the silver planchets were rinsed with DDIW, dried, and counted in EG&G Ortec a-spectrometers coupled to a multi channel analyzer. The 2l0Pb activities at the different depths in the upper part of the stalagmite were subsequently found according to the calculations shown in Appendix 2B and the results are given in Table 2B.1 (in Appendix 2B). Figure 2A.1 shows the activity of 2 1 0 Pbto ta i as a function of depth in stalagmite SF-1. The 2l0Pbt0 tai profile exhibits an exponential decrease to nearly constant value found below a depth o f about 9mm as the 2 l 0P b exCess has completely decayed away and only parent-supported 2 X Q l?b su p p o rted is left. Figure 2A.2 shows the fit of the equation: y (mO) = ml ■ e"m 2 m 0 + m3 (2 A4) 145 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to the 2 1 0 Pbto ta l data to obtain 2loPbo,«c ^ (= n il), a (=m2) and 2I0Pbsu p p o rted (=m3) in equations 2A. 1 and 2A.2. This fit was performed using the Kaleidagraph software. The result was: 2I0Pb o , ex cess = 1-38 dpm/g ± 0.073, a = 0.47 ± 0.033 and 2 1 0 Pbsu p p o rted = 0.052 dpm/g ± 0.0072. Hence, the average growth rate (S) of the upper part of stalagmite SF-1 determined from equation A2.3 was: S = --------- [mm/ yr] — 0.066 ±0.005mm/yr (2A.5) 22.3-0.47 — 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 - supported Pb-210 0 5 1 0 20 25 30 15 Depth from the top of Stalagmite SF [mm] Figure 2A.1. 2I0Pbfo fa / as a function of depth in SF-1. 2 It was observed that no residue remained which indicated that the powder consisted o f pure carbonate 146 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Total Pb-210 (dpm /g) 1 0.8 0.6 0.4 0.2 0 0 5 10 15 20 25 30 Depth from the top of Stalagmite SF-1 (mm) Figure 2A.2. The fit of equation: y = (m l-e '^ ^ + m S ) to the 2 1 0 Pbfo fa ; data. y = ml 'exp(-m2'm0)+m3 Value 0.073063 0.46996 0.051642 0.0071955 1-47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 2B Determination of the 2 1 0 P b t o t a i activity from the a-counting of 2 0 9 Po and 21 0 Po. The 21 0 Pbt0tai activity of the CaC0 3 sample is equal to that of the 2I0Po activity, since 2I0Po and 2I0Pb can be assumed to be in secular equilibrium. The 2 1 0 Po activity, and hence the 2I0Pb activity, can be found from the equation: C - A . _ . _ ^ P o -2 l0 .n et ^ P o - 2 1 0 Pb— 210 Jotal ~ A P o-2l0jotal C . E . p o -2 0 9 ' F po-2 1 0 ' 1 7 1 s a m p le _ C p o -2 l0 .n e t ^ V o -2 0 9 m spike ^ P o - 210 a H O A ~ r W rn 2 ' A P o -im ( 2 B . 1 ) Po-2Q9.net r Po- 210 m sample / V o - 2 0 9 Where: C.E.po-2 0 9 is the counting efficiency for 2 0 9 Po (the counting efficiency for 21 0 Po is assumed to be the same as that for 2 0 9 Po), Cp0 - 2 0 9, n et is the net count in the ROI (region of interest) for 2 0 9 Po: C po-209.n et = Cpo -209.total-Cpo-2 0 9 .b k g d , (2B.2) Cpo-2 1 0, net is the net count in the ROI for 2l0Po: Cpo-21 0 .rtet= Cpo-2\0,total~C P o-2l0,bkgd j (2B.3) FP o -209 and Fpa.2 1 0 are the fractions of the 2 0 9 Po and the 21 0 Po nuclides, respectively, that have decayed within the counting time, t3 , see below. m spike is the mass of the spike, fn sa m p le is the mass of the CaC 0 3 sample Ap 0 -2 0 9 is the decay constant for 2 0 9 Po [y‘l], X po - 2 0 9 = 0.0068 y'1 , J.P o -210 is the decay constant for 2l0Po [y'1 ], AP o -2 io = 1-8280 d''and 148 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A p o -209 is the spike activity at the time of plating, see below. FPo -209 and Fp0 . 2 1 0 can be found from: c* 2 /7 * . / 2 .r^ -2 0 9 ^ ^ ~ “ /^ * l/2 .P o -2 0 9 )* (*3/l*W 0 + N ) / 3 6 5 (OH / f \ ” P o-209 ~~ e ~ e \ L D . L¥) and n ”( 1 ° 2 /7 * ,,2 ^ - 2 1 0 M * 2) — 2/ri/2.po-2io )'^ f3 + f 2 ) /'n n < * \ r P o— 2 1 0 — ^ e {ZB.D) and Apo - 2 0 9 can be found as: a po-209 = A> • e - (Ia2/ri':-'--*wK'l/365) ( I B .6) where 7’ //2. Po -209 is the half life [y] for 2 0 9 Po, T//2. po -209 = 102 y, Tia, p 0 - 2 1 0 is the half life [d] for 2 0 9 Po, Tm. p0- 2 1 0 = 138.4 d, ti is the time [d] between the calibration date and the plating tj is the time [d] between the plating and the counting ti is the counting time [min], and Ao is the activity of the 20 9 Po spike [dpm/g] on the calibration date 3/13/1997 (Ao= 24.1 dpm/g) The standard deviation for Apb- 2 1 0 is: std .d e v .A Pb_2l0 — d: A P o_2 0 g • + (2B .1) V ^ Po-2D9.net ^ P o -2 l0 .n et Where Cp0-209, net is the net count in the ROI for 209Po and Cp0-2\o. n et is the net count in the ROI for 2 1 0 Po. 149 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2B.1. D ata fo r 2l0P b dating o f SF-1. Std. Dev. dpm/g 0 . 0 2 2 CO O O o o CO o o o 0 . 0 1 0 CO o o d Is- o o d 0 . 0 0 5 0 . 0 0 3 0 . 0 0 7 Is- o o o o o o e a. o < M < dpm/g 05 CO O t o t* - CM O CO o CO CO o o s o o CO t o o d o d 05 CM o d CO C M o o T f CO o d CD o o TT o d £ < dpm/g 2 3 . 9 8 CO 05 CO CM CO 05 CO CM 2 3 , 9 8 2 3 . 9 8 ] 2 3 . 9 8 2 3 . 9 8 2 3 , 9 8 2 3 , 9 8 2 3 , 9 8 2 3 . 9 8 2 3 . 9 8 u i o ' 0 . 1 4 5 0 . 2 2 8 0 . 1 7 5 CO d 0 . 0 9 5 05 CM o * to CO T” d CO CM d CO TT C M o 0 . 1 3 6 CO CO d 0 . 1 7 5 £ O LL CO o o M 1 CO o d T f CO o o CO o © 0 . 0 3 4 0 . 0 3 4 CO o o CO o o CO o o 0 . 0 5 3 0 , 0 5 3 0 . 0 5 3 £ S LL 0 , 0 0 0 1 3 3 CO CO o o o d 0 . 0 0 0 1 3 3 CO CO o o o d CO CO © o o o 0 . 0 0 0 1 3 3 0 . 0 0 0 1 3 3 CO CO o o o o 0 . 0 0 0 1 3 3 0 . 0 0 0 2 0 8 CO o C M o o o o CO o CM o o © o O CL 3 7 4 C D z 1 6 7 4 . 8 CO o CM CO CO CO Is- CO CO c o c o CO CO o CO 05 CO CO co’ CO C O a> to CO C M * o> C M K to C M o C M O Q . 3 7 4 ■o 05 JQ ( 0 CD CO CO CO CO CO CO C O CO CO CO o CL 3 N 5 O CO CD r - CM CO CO CO o CM c o r - N * T f CO CO C O C O CM o T— r - C O o CO o CL S 7 4 O c 7 7 7 5 . 6 1 1 9 0 7 . 6 8 7 5 3 . 6 4 4 6 3 . 6 2 7 1 4 . 6 3 4 4 3 . 5 3 4 2 3 . 5 3 7 6 3 . 5 6 8 5 7 , 5 5 4 0 1 . 5 to CO CO o CO 6 8 7 3 . 5 o CL 3 X * o 05 CM CM CM CM CM CM CM CM C M C M T— C M C M O CL § X iS O CO CO Is- r - 1 1 9 2 0 CO CO h - GO CO M ’ & CM CO t o CO CO CO TT CO CO Is- Is- CO o r - C O C O T** CM TT to CO in o CO CO 05 CO CD § .n c E o 05 05 o 05 05 o 55 05 o 05 05 o 5 05 o 05 05 o r — 05 05 o T*- 05 05 o 05 05 o 05 05 o 05 05 o 05 05 -S’ c 2 1 0 2 5 5 o CO CM O 1 0 2 6 5 1 0 2 7 3 1 0 2 7 7 1 0 2 8 3 1 0 2 8 7 1 0 2 9 1 1 0 2 9 7 1 6 0 8 6 1 6 1 0 5 1 6 1 1 1 Q lO t o t o t o t o t o t o t o to to t o t o * o O Is- CM o Is- CM o h - CM o r - . CM o CM o r - CM o r - C M o r - . CM o Is- C M o CM o C M o CM e 1 E O 0 . 2 1 8 7 0 . 2 1 2 5 0 . 2 0 3 6 00 o © 0 . 1 1 5 8 0 . 1 0 7 9 C M CO o o 0 . 1 1 9 2 0 . 1 1 4 2 0 . 1 0 2 8 CO CO o 0 . 1 0 1 6 « t 0 E 0 1 . 4 3 9 9 0 . 5 2 0 9 00 CO o Is- o 1 .0 6 3 5 CO CO d 0 .6 0 2 1 0 .6 8 3 1 CO o C5 1 .0 6 1 1 0 . 6 9 4 5 0 . 6 7 2 3 1 .0 1 8 7 Depth E 2 CO CM t o CO t o CM C O * t o r - CO t o CM t o r - CO t o C M co' t o r - CO to C M w t o Is- CO C M to r - CD C M to h - 00 CM Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 2C TIMS 2 3 0 Th/23 4 U dating of SF-1 The TIMS 2j0Th/23 4 U dating of stalagmite SF-1 was performed by Li and Ku (unpublished data) according to the data given in Table 2C.1 and shown in Figure 2C.1. From the TIMS 2 3 0 Th/2 3 4 U dating it was found that the stalagmite deposition rate was ~ 0.089 mm/year in the upper part (0-1.6 cm), ~ 0.0163 mm/yr in the middle part (1.6-16.5 cm), and ~ 0.72 mm/yr in the lower part (16.5-22.2 cm) of the stalagmite. Table 2C.1.T1MS 2 3 0 Th/2 3 4 U dating of SF-1. Sample ID Depth [mm] TIMS Age [a] SF-1 1.4-1.8 180 ±2900 SF-2 5.5-6 3110 ± 1165 SF-3 12-12.5 7090 ±310 SF-4 16.2-16.8 15990 ± 110 SF-5 22-22.5 16070 ± 225 I s Figure 2C.1. Positions where samples for TIMS 2 3 0 Th/2 3 4 U dating were taken. Original drawing by H.C.Li. CM10*I465«) SM ( t& 9 » 0 * ttO a } Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 2D Manual for the microscope, the digital camera and the Adobe software Original M essage----- From: Adam D. Woods <awoods@earth.usc.edu> To: <dpaulsen@earth.usc.edu> Sent: Friday, March 24, 2000 2:50 PM Subject: Re: Microscope, digital camera, etc. > > To use the digital camera: > > 1. Turn on the spot diagnotics box which runs the camera (it is to the > right o f the scope). > > 2. Turn on the computer. > > 3. Push the bar to the left and above the oculars all the way in (the bar > has a black plastic ball on the end). This will stop light from entering > into the scope through the oculars (all o f the light will go to the camera). > > 4. Open Adobe Photoshop. > > 5. Under the File Menu choose [import" and then "Twain Aquire..." > > 6. A box will come up which says "Spot Twain Driver" at the top. Choose > Preview. The camera will make lots o f clicking noises, and a blurry image > will appear in the box. > > 7. Focus the image by choosing "Focus" (upper left-hand comer). A box > will appear. Move the box over a portion o f the blurry image which has a > great deal o f contrast between light and dark color. Double-click on the > box, and another box will appear with a black and white image. Click > "begin", and move the focus on the microscope until the black and white > image comes into focus (this may take awhile until you get the hang o f it). > Click on "stop" and then "close". You are now ready to aquire your image. > > 8. Click on "Acquire". The camera will take the photo, and transfer it to > Photoshop where you can save the image, or manipulate it. To take another > snapshot, repeat steps 5-8. > > Note: > > To increase the speed o f photo acquisition, turn up the brightness o f the > light entering the scope. Please do not turn the knob past 10. Also, > please be sure to turn it all the way down when you are finished. Please > do not change o f the settings in the Twain Acquire program without > consulting with me first. It takes a long time to get the settings right. > > Adam Woods > Department o f Earth Sciences > University o f Southern California > Los Angeles, CA 90089-0740 > (213)740-7653 >(213)740-8801 FAX 152 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 2E Matlab Script used for identifying dark bands % Dorte Eide Paulsen, 4/28/2000 % Matlab script to identify local minima in microscope image % clear workspace clear % read tif image into memory a=imread(bil I flip, tif): % generate the red, green and blue image r=a(:,:,l); g=a(:,:,2); b=a(:,:,3); % add the 3 colors into a monochrome image bw=double(r)+double(g)+double(b); % crop the image bw I =bw(75:225,:); % blurr the image a little = decrease the resolution of the image m = [ 0.25 0.25 : 0.25 0.25]: bw2=conv2(bw 1 ,m); % display the image in monochrome grayscales imagesc(bw2); colormap(gray); % make the vertical projection pro=sum(bw2); % calculate the gradient dpro=diff(pro); % identify’ zerocrossings that are not noise for i=l00:382, if dpro(i)<0&dpro(i+l)>0&abs(dpro(i))/pro(i)>0.003, % if it is a local minimum, put a red star on the image hold on p lot(i+ l,80,’ r*’ ): end end % plot the vertical projection figure plot(pro) % plot the gradient of the vertical projection figure plot(dpro) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 2F Manual for the Luminoscope Original M essage----- From: Adam D. Woods <awoods@earth.usc.edu> To: <dpaulsen@earth.usc.edu> Sent: Wednesday, March 29, 2000 11:50 PM Subject: Luminoscope > CL Scope Operation > > 1. Check if the microscope is set up for CL work, and make sure the CL > chamber is locked firmly into place. > > 2. Check to make sure the Whitey valve (i.e., the valve from the pump > to the chamber) is closed (closed = perpendicular to the wall behind the > microscope). > > 3. Open sample chamber, and insert sample, rock side up. Close > chamber, and check to make sure that the sample is in focus. If not, > insert a glass slide between the sample and one of the metal spacers. > Do not place the large glass slides back to back, as air may get trapped > this way. > > 4. Close chamber, and turn on pump. > > 5. Switch microscope from off to pumpdown; pressure in chamber should > be one ATM. > > 6. Check to make sure that the purge valve is closed, then open the > Whitey valve by turning 90° while simultaneously holding the chamber in > place. > > 7. Pressure in chamber should slowly drop. Wait until it reaches 20-30 > millitorr. > > 8. Switch to "Regulated" or "Manual" mode, and turn on the beam. To > optimize the position o f the beam, adjust the position of the magnets on > the steel frame. > > To change objectives: > > Raise current objective to maximum height. Remove steel frame which > holds the magnets, and switch objectives. > > To change samples: > > Turn off the beam, and switch instrument to pumpdown. Close Whitey > valve. Vent chamber with purge valve, and open chamber. Change sample, > and begin from #2, above. > 154 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. > Shut Down Procedure: > > 1. Turn beam off, and set kV and mA to zero. > > 2. Switch instrument to pumpdown mode. > > 3. Close Whitey valve. > > 4. Remove samples. > > For complete shutdown: > > Complete all o f the above steps, turn instrument off, and turn off > vacuum pump. > 155 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 3A Bilinear regression This appendix explains how a bilinear regression is performed. Assume that the P vector contains the precipitation data, the T vector contains the temperature data and the L vector contains the laminae thickness and N is the number of data points: M Cl \ p 2 t2 . h p = , T = t v K N J and L = J N; (3A1) The actual data sets (from the Budda, Lanzhou, and Ankang stations) used for the bilinear regressing contained between 30 and 53 data point, but for simplicity only 5 data points are considered in this example: P = r511y '15.9^ '0.111' 569 15.7 0.074 749 , f = 11.9 and L = 0.092 888 11.8 0.073 585 v 7 14.5 V . 0.083 'v J (3A.2) The bilinear expression for L based on P and T can be written as: a P + j3-f + y = [ (3A3) where a ,3 and y should be found by minimizing K the following equation utilizing the excel solver function: 156 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. a-Pi+P-tt+y-l,)2 =fc (3A4) /= ! In the example from before, the P vector values are entered in cell A9 - A13. The T vector values are entered in cell B9 — B13 and the D vector is entered in cell C9 - C13. The values of a, (3 and y are unknown at this point, but they are located in cell B3 - B5. The r; values are defined as: ri =a-p, + p ti + y (3A.5) and are calculated in cell D9 — D13. The values: (Z/-(a-p,.+£-r,.+r))2 (3A6) are calculated in cells G9 - G13. Finally, the value: (3A7) i= l is calculated in cell G15. An image of the window, using the Excel spread sheet is shown in Figure 3A.1. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. S l g i i l l l i s l M * ! GIS =SUM(G9:G13) S M S *!r^ A 5 S « S m Dorte Eide Paulsen 7/19/2000. USC, Bilinear regression example ----------------------- ------ ------ --.... - - ... - 151 .s: r : alpha beta gamma 1 ......... ..." 1" ..... . ...... ..... 1 " 7 1 .......... .7 ....... M 177 — --------- -------------------------- — -------- ............- -...... - — - S 3 . __________________ I m 8', p vector t vector 1 vector ri=alpha*pi+heta'ti+gam m a (ri-li)*2 ........... ’" . . . ?! W . S I O ? 517' 569 15.9 0.111 533.9' 15.7 0.074 585.7 284930.7 342957.8 *1 -11 12 749 888 11.9 0.092 761.9 11.8 0.073 900.8 580351.4 811309.1 >13 514s 585 14.4 0.083 600.4 360380.5 1 5 .18 47 ... --------- s u m or G9 to G13 1 23799301 - ... v .- : 18 19 20; ........... ..... -..- .......-.... — - ...- - - ..... - .......... -.............- - ---- ---------- $ 2 2 23: m . --- — ----- -----...-....- ----------- ----- --- - -...- ---------------- ------------------- ... '* 1 3 ) ' p 75 261 ..........-........... - - ----- ------ - - ---- - ......- -...................... - - - £ ... . , , A . m 28 ........- .......- .... - -....................... - - - - - ..... I mh I ► ( Wft S heetl /-S h e e ts / SteetS Figure 3A.1. Screen dump of the Excel spread sheet used to calculate the bilinear regression. The remaining task is to find 3 different numbers for cc, (3 and y (located in cell B3 to B5) that will minimize cell G15. At this point the number “1” has been entered in cell B3 to B5, because the values have not been calculated yet. The Excel solver function (located in the tools menu) is used for this. Figure 3A.2 shows the solver window. 158 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Solvei Paiameteis E2E3 Figure 3 A.2. The solver menu. In the “set target cell” is entered $G$15. This is the cell that needs to be minimized. The box stating “min” is selected, indicating that the target cell should be minimized. In “by changing cells” is entered $B$3:$B$5, which are the o c, |3 and Y that are sought. Then the “solve” button is pressed. The solver function will enter the best values of a, (3 and y in B3 — B5. This is shown in Figure 3A.3. As observed in Figure 3A.3, the calculated values for a , (3, and y are: -1.5T0-4, - 8.18-10'3, and 0.299. In conclusion, the result is: — 1.5-10"4 • p + 8.18 -10-3r + 0.299 = I (3A8) Figure 3A.3 also shows how to calculate the R2 value by using the Excel build- in function “RSQ”. For the correlation between the observed laminae thickness and the laminae thickness based on equation (3A.8) the R2 is determined to be: Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Rz = RSQ(c3: c8,d3 : dS) = 0.35 (3A9) X M i c r o s o f t E x c e l - i e g r e s s t o n e x 3 m G i7 r | - ^ * | =RSQ(C9:C13.D9:D13) j - * ■ A ' 1 B :J - C I ~ D ! J; ' 1 1 - -'v-iL - T j T. Dorte Eide Paulsen 7/19/2000. USC, Bilinear regression example ............. -........ 3 *4* alpha -0.00315 beta 43.00818 ------------------------- ---- - - -...... .................... ..... - .....- - - - - ... - - - - - .................. ......... - - - ---- - -1 . 5 : m .7 " '8 . 9-: 110; gamma 0.299326 p vector t vector 517 15.9 569 15.7 I vector ri=alpha*pi+beta'ti+gam m a 0.111' 0.092143.................... ... 0.074 0.088023 (ri-li)*2 ~ ' ' 0.000356 0.000145 - . - - - ' . - - I 11 i1£| 749 11.9 888 11.8 0.092 0.090263 0.073 0.070349 3.02E-06 7.03E-06 - 13: 21$ 585 14.4 0.083 0.094272 0,000127 s a s sum or G9 to G13 0.000637 1 7 ; as. a s 20 -21: ^22: 23 24' 25.-I ;2E' 27- m ......................— - ........ r-squared I 0.3509281 ■ ..... . ' " - i K - ... . . _ „ a i<f< 1 ► l> ll\S h eetl /S h eet2 /Sheet3 / .- V Figure 3A.3. The result of the solver function. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 4A. Using the micro miller to acquire samples of a stalagmite A 3-axis computer controlled micro miller was used to acquire samples for the isotopic and trace element analysis. The ISEL Automation micro miller was connected to a PC though the serial port. The operating system on the PC was DOS. The program used to upload instructions to the micro miller was Techno-isel-PAL. Programming details on the PAL program is given in the programmer’s manual1 . The micro miller setup is shown in Figure 4A.1. x-axis z-axis stalagmite drill Figure 4A.1. The micro miller setup. 1 Techno/Isel, 2101 Jericho Turnpike, Box 5416, New Hyde Park, NY 11042-5416, USA: PAL Programmer’s Manual, October 1994. 161 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Two programs were constructed for sampling the stalagmite. The program "init" moves the drill to the position where the sample needs to be drilled. The program "samp 100" drills the actual sample. The two programs will be described in details below. Init program In Figure 4A.2 is shown a screen dump after the PAL program has been started and the init program has been loaded. The PAL program automatically asks during startup for a file to load. To create a new unnamed file, just press <enter>. | 7 x 12 3 ^ 1 @ B Frle: A:\INIT f 10 for pulldown menu S <) x l s x y ' ; r e f e r e n c e x y / ; S d e f m e in 1 t pus 1 9't _ 1 ( 1 [M l 0 ) , 1 3 U . 0 ( 11 ) II I!) ,!)./( 1 IIII 0 ) , II ( 1 0 II 0 ) ; -d e f i n e s p eed ( ‘ jlIIJU); -units m ; n o n e l n 11 p i; s ; stop. Figure 4A.2. Screen dump of the init program. 162 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. For the specific details of all the program lines see the programmer’s reference manual. The “init” program moves the drill from its origin position to the position (x, y, z)=(194.1mm, 130.0mm, 8.2mm). These numbers should be chosen depending on the actual position of the stalagmite. In program line 4: “#define initpos 194.1(1000),130.0(1000),8.2(1000),0(1000);” the first two parameters define the x and y coordinates. The third and fourth parameters are the z coordinate in the downward and upward directions, respectively. The movement in the x and y directions are determined by the sign of the coordinate (-194.1 would move the drill 194.1 mm in the opposite direction). The numbers in the parentheses specify the velocities (in steps/second). When the program has been entered save the program in the “file“ pull-down menu by executing the “save As” item. The program is then compiled by executing the “compilE to disk” item in the “windows” pull-down menu. Finally, upload the instructions from the disk by choosing “transfeR from disk” in the “windows” pull down menu. At this point the instructions is transferred from the PC memory to the memory of the microprocessor controlling the micro miller. The program can be executed by pressing the “start” button located on the micro miller. Sample program. In Figure 4A.3 is shown a screen dump after the PAL program has been started and the “samp 100” program has been loaded. The PAL program automatically asks during startup for a file to load. To create a new unnamed file, just press <enter>. 163 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. F i l e : A:\SfiMF100 F10 fur puilduun neiiu n o u p U( 1UIJU) , 0( 1 0 0 0 ) , M J . fi< MJIIIJ) , IK 1 0 0 0 ) ; n o u p IK 101)11) ,IK 11)011) , ‘ j . IJ( 1 0 0 0 ) , iJ C 1 illltJ) ; n o up IK 1IJ0D) ,!!( 1 00 0 ) , 0 . IK 'Kill) , IK 10110) ; n o u e 0 ( 1 0 0 0 ) , 0 ( 1 0 IJ I)) , 1 . 0 M 0 0 ) , I) M U 0 0 ) ; ! 1) p <) l ii 0 r i 1 1 i n ■ ) I n o u p IK 10110) , U( 1 1I O il ) , 1 . S( Ml > , IK 1 0 I!) ; n o u e IJC 1 UIJIJ) , U . '*( M J) , IK 1UU) ,1H 1 IJU ) ; n o up IK 1 0 0 0 ) , 0 . < K MJ) , U( 1IIIJ) , IK 1 0 0 ) ; n o up IK 1 0 0 0 ) , IK 1IJ110) , I K 1 00 ) , I . K I D I D ; n o u r IK 1 0 II) , IK 1 0 0 ) , IK 1 0 0 ) , 1 . I K 1 I I I I ) ; n o up IK 10 0 ) ,U ( 1IJ0) , IK 1 I!II) , - h . IK IJ 0 ) ; n o u p IK 10110) , IK 1 0 0 ) , IK 1 0 0 ) , U r> - IK I- Oil II1 ; n o u p I). 1 ( 1 0 0 0 ) , 1J( 1110) , [J( 1 1J IJ) ,LK 1 0 0 ) ; s t II |) . Figure 4A.3. Screen dump of the “samp 100” program. For the specific details of all the program lines see the reference manual. The program consists of four parts: 1) Lines 1-4: The drill is moved 50mm down. First 40mm at a velocity of 2000 steps/s, then 5mm at 1000 steps/s, then 4mm at 500 steps/s and finally 1 mm at 100 steps/s. 2) Lines 5-9: The drill is moved down another 1.5 mm at 50 steps/s. Then the drill is moved 0.4 mm in the y-direction at 50 steps/s and back again. These two steps drill the actual sample. 3) Line 10-12: The drill is moved 50 mm up again, in reverse order of step 1. 164 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4) Line 13: The drill is moved 0.1 mm (100 microns) forward in the x-direction and the drill is ready for next sample. When the program has been entered save the program in the “file“ pull-down menu by executing the “save As” item. The program is then compiled by executing the “compilE to disk” item in the “windows” pull-down menu. Finally, upload the instructions from the disk by chosing “transfeR from disk” in the “windows” pull-down menu. At this point the instructions is transferred from the PC memory to the memory of the microprocessor controlling the micro miller. The program can be executed by pressing the “start” button located on the micro miller. Since the program automatically advanced the drill 0.1-mm, the next sample can be drilled by just pressing the “start” bottom located on the micro miller again. Between samples the stalagmite was blown clean by utilizing an air gun with compressed N2 . A photograph of the actual setup with the air gun and N2 tank is shown in Figure 4A.4. Figure 4A.4. Photograph of the setup. 165 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 4B Matlab script to calculate power spectrum % ---------------------------------------------------------------------------------------------------------- % 6/15/2000 Dorte Eide Paulsen % Department of Earth Sciences, USC % Matlab script to calculate power spectrum for delta 13C signal % For details see page 6-33-6-35 in Mathworks, Using Matlab, Version 5 % ---------------------------------------------------------------------------------------------------------- % clear workspace clear % load isotope data load long.txt % extract, time, carbon and oxygen x=long(:,1); c=long(:,2); o=long(:,3); % make new time axis that is equally spaced at 1 year xi=l:9398 ' ; % make interpolation of the raw carbon data to the new time axis ci=interpl(x,c,xi); % show a plot of the C13 measurements plot(xi,ci); % transform the C13 data to the fourier domaine fc=fft(ci); % remove the average value in the fourier domaine fc(l) = [] , - % calculate the power of the fourier transformed signal N=length(fc); power=abs(fc(1:N/2)).A2; % calculate the corresponding frequencies nyquist=l/2; freq=(1:N/2)/ (N/2)*nyquist; % make a plot of the power spectrum figure plot(freq,power); axis([ 0 0.01 0 6e6]); 166 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX 4C Power spectra for the 513C and 8lsO records 0-144 a P o w e r s p e c t r u m C -1 3 (0-1 4 4 a) 1.2 Numbers incficate periods in years 48 0.8 sC 0 .6 • 0.4 24 9.6 0.2 • 0.3 0 0.05 0.1 0.15 0.2 0.25 frequency [cycJes/year] 0-1266 a P o w e r s p e c t r u m f o r d e lta -1 3 -C (0 -1 266a) 1.2 Numbers indicate periods in years 422 0.8 • i C L 0.4 140 0.2 • 33 5.00E-03 1.00E-02 2.50E-02 2.00E-02 Frequency [cycies/year] Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0-144 a P o w e r S p e c tru m f o r delta-1 8 -O re c o r d (0-14 4 a) Numbers indicate periods in years 36 1 0.8 g 0.6 0.4 0.2 9.6 72 0 0.25 0.3 0 0.05 0.15 0.1 Frequency (cycies/year) 0-1266 a P o w e r s p e c t r u m f o r 0 - 1 8 re c o rd (0-12 6 6 a ) 1.2 Numbers indicate periods in years 633-422 0.8 g 0.6 0.4 • 0.2 • 253 115 33 3.00E-02 2.50E-02 Frequency [cycies/year] 168 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. R eproduced with permission of the copyright ow ner. Further reproduction prohibited without p erm issio n . AT Temperature and rainfall indices from trace elements and isotope data Index: W=Warming C=Cooling 1= Increasing rainfall D= Decreasing rainfall ■ tem perature (Isotope) • rainfall (isotope) A tem perature (trace) ♦ rainfall (trace) • i ♦ ♦ 5C D O 5500 • m ♦ ■ ■ A A A A w A A • • ♦ ♦ ■ ■ W 6)00 ■ ■ 6500 • • c & D 7000 • • • ♦ ♦ ♦ 7$00 • • • w 8000 • 0 ♦ ♦ 85 • • ♦ ♦ ♦ 00 900C • < * • ♦ • • • ♦ ♦ ♦ 9500 10000 H to a a a s B (2 S a n B . S' a 8 AR R Lh > [ a ] O S VO APPENDIX 5B XRD spectra for SF-1 2500 — 2000 H ft § _ 1500 - 1000 - 500 O 10 1 "f" ■ 15 1560X SF Stalagmite n f ,n 20 25 30 35 40 Angle (degrees) 45 50 55 1580X SF Stalagmite 1200 lO O O 1 —1 s 800 s C A 600 400 200 O Angie (degrees) 1600X SF Stalagmite 3500 - 3000 - 2500 - 3 re 2000 - C A « < 1500 - 1000 - 500 - O - 25 30 35 40 Angle (degrees) 55 170 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1640 GAP SF Stalagmite 2500 - 2000 - E L 3 1500 - —- *7 1000 - 500 - Lv, j A k A i 30 35 40 Angle (degrees) 55 1680X SF Stalagmite 1 0 0 0 9 0 0 8 0 0 Z O O 6 0 0 5 0 0 3 0 0 200 100 5 5 1 5 3 5 4 0 5 0 10 20 2 5 3 0 4 5 Angle (degrees) 171 O X SF Stalagmite 1200 8 0 0 6 0 0 200 10 1 5 5 5 20 3 0 3 5 4 0 50 2 5 4 5 Angle (degrees) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1980X SF Stalagmite 1600 1200 - 1000 - 8 0 0 6 0 0 - 200 - 5 5 2 5 3 0 3 5 40 4 5 5 0 10 1 5 20 Angle (degrees) Table 5B.1. Regions of interest in XRD spectra Sam ple ID ROIs (integrated counts) Aragonite/ Max. count C aldte purity- C aldte lAragonite Calcite Caldte peak 1560X 643 18 0.028 2316 1580X ’ 379 68 Relative puxre 1600X 714 10 0.014 2903 1640 GAP 595 26 0.043 2137 1680 385 54 0.140 894 1710 439 58 0.132 1134 L ess p u re 1980 536 27 0.050 1487 *) The noisy spectrum for this sample is believed to be an artifact of the sample being too small for the XRD. Sample 1580X was therefore discarded. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Silicic and germanic acids: Laboratory determination of their molecular diffusivities and field study of their benthic fluxes along the California margin
PDF
A proxy for reconstructing histories of carbon oxidation in the Northeast Pacific using the carbon isotopic composition of benthic foraminifera
PDF
Oxygen isotopic evidence for fluid infiltration in the Mount Stuart batholith, Washington
PDF
Petrologic and geochronologic study of Grenville-Age granulites and post-Granulite plutons from the La Mixtequita area, state of Oaxaca in southern Mexico, and their tectonic significance
PDF
Numerical simulation of whole core squeezer radon pore water profiles: Methodological considerations and evaluation of benthic fluxes and rates of bio-irrigation and advection
PDF
Magma mingling/mixing in a heterogeneous, multi-pulse magmatic system: An example from the Jackass Lakes pluton, central Sierra Nevada batholith
PDF
Modeling of continuous tiltmeter data from the 1984 rifting event at Krafla Volcano, Iceland
PDF
The origin of enigmatic sedimentary structures in the Neoproterozoic Noonday dolomite, Death Valley, California: A paleoenvironmental, petrographic, and geochemical investigation
PDF
Paleoecology and paleoenvironments of early Triassic mass extinction biotic recovery faunas, Sinbad Limestone Member, Moenkopi Formation, south-central Utah
PDF
Reinterpreting the tectono-metamorphic evolution of the Tonga Formation, North Cascades: A new perspective from multiple episodes of folding and metamorphism
PDF
Management of large earthquake data using the Antelope Relational Database and seismicity analysis of the 1999 Turkey earthquake sequences
PDF
Investigating the influence of atmospheric changes on the variability of the North Pacific using paleoproxy data and a fully coupled GCM
PDF
Radium isotopes in San Pedro Bay, California: Constraint on inputs and use of nearshore distribution to compute horizontal eddy diffusion rates
PDF
Fourier grain-shape analysis of quartz sand from the Santa Monica Bay Littoral Cell, Southern California
PDF
Biotic recovery from the end-Permian mass extinction: Analysis of biofabric trends in the Lower Triassic Virgin Limestone, southern Nevada
PDF
Oxygen-related biofacies in slope sediment from the Western Gulf of California, Mexico
PDF
Paleoenvironments and paleoecology of the disaster taxon Lingula in the aftermath of the end-Permian mass extinction: Evidence from the Dinwoody Formation (Griesbachian) of southwestern Montana...
PDF
Structural evolution of the southwestern Daqing Shan, Yinshan Belt, Inner Mongolia, China
PDF
Three-dimensional reconstructions of fold growth using growth strata: An example from the Catalan Coastal Ranges, northeast Spain
PDF
Helicoplacoid echinoderms: Paleoecology of Cambrian soft substrate immobile suspension feeders
Asset Metadata
Creator
Paulsen, Dorte Eide
(author)
Core Title
Multi-proxy studies of climate variability in central China: Subdecadal to centennial records in stalagmite from Budda Cave
Degree
Master of Science
Degree Program
Geological Sciences
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
geochemistry,OAI-PMH Harvest
Language
English
Contributor
Digitized by ProQuest
(provenance)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-35899
Unique identifier
UC11341055
Identifier
1405251.pdf (filename),usctheses-c16-35899 (legacy record id)
Legacy Identifier
1405251.pdf
Dmrecord
35899
Document Type
Thesis
Rights
Paulsen, Dorte Eide
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
geochemistry