Geophysical applications in compressional orogens
Item Type text; Dissertation-Reproduction (electronic)
Authors Libarkin, Julie Carol
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Download date 04/10/2021 15:26:40
Link to Item http://hdl.handle.net/10150/289021 INFORMATION TO USERS
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 overiaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.
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.
Bell & Howell Information and Learning 300 North Zeeb Road, Ann Artx>r, Ml 48106-1346 USA 800-521-0600
GEOPHYSICAL APPLICATIONS IN COMPRESSIONAL OROGENS
by
Julie Carol Libarkin
A Dissertation Submitted to the Faculty of the
DEPARTMENT OF GEOSCDENCES
In Partial Fulfillment of the Requirements For the Degree of
DOCTOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
1 999 UMI Number: 9946823
UMI Microform 9946823 Copyright 1999, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized copying under Title 17, United States Code.
UMI 300 North Zeeb Road Ann Arbor, MI 48103 2
THE UNIVERSITY OF ARIZONA ® GRADUATE COLLEGE
As members of the Final Examination Committee, we certify that ve have read the dissertation prepared by jniio r. riharkln entitled Geophysical Applications in Compressional Orogens
and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy
n H.,, Robert Butl Date I
Date
Date
susan Beck . Date 1, J999 Michelle Hall-Wallace Date ^
Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.
P __ _ _iH w • 191 Dissertation Director pobert Butler I 3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.
Brief quotations firom this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.
SIGNED bu 4
ACKNOWLEDGMENTS
This dissertation has taken many unforeseen turns and changed many times over the past five years. First and foremost, I would like to thank my advisor, Robert Butler for giving me the freedom to indulge my scientific interests in a variety of different arenas. His willingness to introduce me to the world of paleomagnetism and teach me the tricks of the trade was invaluable for my successful paleomagnetic study of Bolivia. As well, Clement Chase and Jay Quade deserve thanks for including me in their innovative cosmogenic isotope studies and encouraging my own theoretical meanderings. Comments made by these and the other members of my dissertation committee, Peter Coney, Susan Beck, and Michelle Wallace, vastly improved this manuscript. The paleomagnetic study of Bolivia would not have been possible without the field assistance of David Richards, Thierry Sempere, and Robert Butler. Thierry Sempere also provided valuable insight into the geology of Bolivia. Laboratory work completed by Brigette Martini, Scott Balay, and Rob Feyerharm helped speed this project along. Finally, conversations with Rob Van der Voo were very useful to our final data interpretation. The 2lNe cosmogenic work would not have been possible without the assistance of Jane Poths and Bill Phillips of Los Alamos National Laboratory and Nat Lifton of the University of Arizona. Both Bill and Nat were invaluable in the early sample preparation stages. Bill and Jane spent many hours training me on the noble gas mass spectrometer and in processing and interpreting my data. All three were willing participants in many enthusiastic discussions about cosmogenic isotopes in general and my project in particular. Dr. Laurie Brown and Dr. Peter Lipman participated in several insightful discussions of the San Juan Volcanic Field and Fish Canyon Tuff of Colorado. Finally, William Hart provided field assistance without which I would not have been able to collect all of my samples. I would like to thank my family and friends for their wholehearted encouragement and support. I would not have been able to complete this dissertation without their motivation. Finally, I thank Michael G. Moore for his unwavering encouragement and for acting as an enthusiastic sounding board for all of my ideas. 5
TABLE OF CONTENTS Page
ABSTRACT 6
1. INTRODUCTION 7
1.1. Cosmogenic Isotopes and Their Uses 8
1.2. Cosmic Radiation 9
1.2.a. Galactic cosmic rays 11 1.2.b. Solar cosmic rays 14
1.3. History of Cosmogenic Isotope Studies 15
2. PRESENT STUDY 19
2.1. General Conclusions: Appendix A 19
2.2. General Conclusions: Appendix B 20
2.3. General Conclusions: Appendix C 21
APPENDIX A. COSMOGENIC 2lNE IN QUARTZ AS A POTENTIAL PALEOALTIMETER: CASE STUDY FROM THE 28 MA FISH CANYON TUFF, COLORADO 23 APPENDIX B. TWO-ISOTOPE METHOD FOR ESTIMATING PALEOELEVATION 52
APPENDIX C. TERTIARY REMAGNETIZATION OF PALEOZOIC ROCKS FROM THE EASTERN CORDILLERA AND SUB-ANDEAN BELT OF BOLIVIA 76 REFERENCES 115 6
ABSTRACT
Geologic endeavor is a continual search for windows into the past. Our ability to
characterize ancient geological processes is only as good as the techniques we have at our
disposal; often the desire to uncover new information drives the development of new
geologic methods or the modification of old techniques. This dissertation is composed of a
series of projects focused on gaining new insights into the history of orogenic systems
through the application of existing techniques or the development of new methods. Three
primary projects were undertaken, the first a paleoelevation study aimed at determining the elevation of the Colorado Rockies 28 million years ago (Ma). This study was an attempt to both constrain the paleoelevational geometry of North America and test a new
paleoaltimetry technique. This "one-isotope" technique relies on the relationship between cosmogenic isotope production rates and elevation; while a paleoelevation for the Colorado
Rockies was not derived, the technique should prove useful in other geologic situations.
From this initial technique, a two-isotope technique was derived which bypasses some of the difficulties inherent to the one-isotope method. A complete theoretical development of this two-isotope technique is included. Finally, a project documenting the wide-spread remagnetization of a suite of pre-Cenozoic Bolivian red sedimentary rocks reflects the impact orogenesis can have on an entire region. Taken as a whole, these projects focus on both the effects of mountain-building events and the techniques we can use to better understand them. 7
CHAPTER 1. INTRODUCTION
Hutton's Principle of Uniformitarianism states that the geological processes we see operating today occurred in the same way throughout the Earth's history. Geologic research is based upon this idea; we infer the evolution of the Earth by making comparisons between the rock record and present-day geological processes. Making these connections requires many assumptions, primary of which are assumptions about the morphology and timing of deformation of the earth's landmasses. To truly be able to interpret the rock record, geologists need to know, at the very least, the paleotopography of the major continents.
In this work, cosmogenic isotopes and magnetization preserved in rocks are used as a means for determining paieoelevation and understanding the processes associated with mountain-building, respectively. Two techniques are developed which rely upon the relationship between cosmic ray flux and elevation. The one-isotope technique yields a cosmogenic isotope production rate calculated from measured cosmogenic isotope concentrations and a known paleosurface exposure time. Comparing this production rate to a known sea level rate and using the production rate-elevation relationship will give a paieoelevation for the studied surface. This technique was tested on the 28 Ma Fish
Canyon Tuff in Colorado (see Appendix A for full details). Alternatively, two cosmogenic isotopes with different production mechanisms (such as muon capture and neutron spallation) can be used to estimate a paieoelevation. Paieoelevation is estimated by comparing a ratio of the concentrations of the two measured isotopes to predicted values
(see Appendix B for a development of this theory). The third and final study contained in this dissertation documents a wide-spread remagnetization event in Bolivia which is associated with the formation of the Andes Mountains (Appendix C).
This chapter serves as a preface to the three papers appended to this dissertation.
Taken as a whole, these manuscripts represent a series of projects aimed at refining our 8 understanding of and ability to determine the Earth's history. Data presented in this dissertation were acquired primarily by the author. Paleomagnetic samples were collected by me, D.R. Richards, and R.F. Butler. Paleomagnetic analyses were conducted by me and laboratory technicians at the University of Arizona Paleomagnetics Laboratory.
Samples of the Fish Canyon Tuff were collected by me, C.G. Chase, J. Quade, and B.
Hart. All ^iNe measurements were made with the assistance of Drs. J. Poths and W.M.
Phillips of Los Alamos National Laboratory. The ^'Ne data presented in Appendix A was gathered by Dr. J. Quade (Run #1) and me (Runs #2 and #3). Finally, all analyses were conducted by Dr. W.C. Mcintosh of the New Mexico Geochronology
Research Laboratory of the New Mexico Bureau of Mines at New Mexico Tech.
Manuscripts were prepared by me; co-authors supplied text specific to their area of expertise and assisted in manuscript revision. A brief synthesis of the history, uses, and theory behind cosmogenic isotopes is included in this introduction, as a preface for the studies which make up the bulk of this work (Appendix A and B). Appendix C is a self- contained published manuscript related to paleomagnetism and requires little inU^oduction.
1.1. Cosmogenic Isotopes and Their Uses
Isotopes produced through the interactions of cosmic ray particles and matter in the atmosphere or at the Earth's surface are known as cosmogenic isotopes. Cosmogenic isotopes produced from interactions with primary cosmic ray protons were originally studied in meteorites. Later, once the nature of the secondary cosmic ray flux began to be deciphered, physicists and geologists realized that isotopes produced from cosmic ray neutrons in the atmosphere may contain useful geologic information. Originally, these cosmogenic nuclides were proposed as a means of determining sediment chronologies, hydrologic mixing and circulation rates, and glacial ages (Lai and Peters, 1967). An entire field of study has since evolved out of this beginning and thirty years later, cosmogenic 9 isotopes have an ever-widening arena of uses, including the dating of geomorphic surfaces
(Cerling and Craig, 1994), calculation of erosion rates (Phillips, 1998), and determination of paleoelevations (Appendices A and B, this dissertation). Additionally, original conceptions of "useful" cosmogenic isotopes usually included only radioactive species, such as or '^Be, produced in the atmosphere as the result of neutron interactions.
Today, a wide variety of cosmogenic nuclides are studied, including: 1) radioactive and stable species, such as ^He and ^iNe; 2) isotopes produced both in the atmosphere and at the Earth's surface, in situ ; and 3) isotopes produced from interactions with other cosmic rays, specifically muons (Stone et al., 1998).
1.2. Cosmic Radiation
Cosmic radiation is a term used to describe a stream of ionizing radiation which originates in space and is continuously interacting with the Earth and its atmosphere.
Primary cosmic radiation comes in two forms: 1) galactic cosmic rays, produced from supemovae; and 2) solar cosmic rays, associated with solar flares and solar wind. The energy of this cosmic radiation is highly variable, ranging from 10 MeV to 10^0 eV for galactic cosmic rays and 1-100 MeV for solar rays (Friedlander, 1989). A brief discussion of each is provided below.
Primary cosmic radiation incident at the top of the Earth's atmosphere is composed of roughly 87% protons (hydrogen nuclei), 13% alpha particles (helium nuclei), and a small amount of heavier nuclei, muons, and electrons (Friedlander, 1989). When this radiation encounters the Earth's atmosphere, interactions between matter and the energetic cosmic rays results in a cascade of secondary cosmic rays. These secondary particles are predominantly neutrons; by the time cosmic radiation reaches the Earth's surface, the proton component is nearly extinct. Reactions between these secondary neutrons and isotopes in the atmosphere and at the Earth's surface are the dominant, although not only. 10
production mechanism for the cosmogenic isotopes studied by geologists. Spallation, the
splitting of a target nucleus into three or more particles, is the primary mechanism for
cosmogenic nuclide production by neutrons (Fabryka-Martin, 1988).
Cosmogenic isotopes can also be produced through the interaction of primary
cosmic ray muons with matter. Muons are heavy electrons, with equivalent magnitude
charges, masses about 207 times heavier than a normal electron, and a mean life of 2.2 |is.
Muons can be either positively or negatively charged; however, because positive muons are
repelled by the nucleus of an atom, they will have litde involvement in the production of
cosmogenic isotopes. Rather, a negative muon slowing down in matter will either decay
into an electron and energetic neutrinos, D^, or be captured by a proton within the nucleus
(Z,A). When the negative muon is captured, the proton will change into a neutron such
that:
|l+(Z,A) —> (Z-1,A)*+'U^
where the neutrino carries off most of the freed energy. The resulting nucleus, (Z-1,A)* is
de-excited by the emission of neutrons, protons, or alpha-particles, yielding an isotope
(Z-l-Xp,A-(Xn+Xp)) where Xn and Xp represent the number of emitted neutrons and
protons, respectively. The most likely scenario for de-excitation is the emission of one
neutron, such that Xn=l and Xp=0 (Charalambus, 1971; Friedlander, et al., 1964; see
Fabryka-Martin, 1988 for an excellent review). Muons make up about one-half of the cosmic ray flux at sea level, although the majority of cosmogenic isotope producdon is due to energetic neutrons (Stone et al., 1998). For some isotopes, such as '^Be or 36ci, production by muons may account for up to 10% of the total production at sea level
(Brown et al., 1995). Additionally, cosmic ray muons attenuate much more slowly in matter than cosmic ray neutrons. As a result, muon production quickly dominates over 11
neutron production with depth; almost all production below three meters depth at sea level
results from muon interactions.
1.2.a. Galactic cosmic rays
The energy releasing transition of an ordinary star into a much smaller mass is
called a supernova. This energy release is responsible for the production of cosmic
radiation (Friedlander, 1989). Although there are two classes of supemovae, only Type II
supemovae, those occurring in stars with a small, dense core composed of heavy elements,
are energetic enough for cosmic ray production. In a Type II supernova, an extensive,
hydrogen-rich sphere with a maximum radius of IQl'^ cm surrounds the inner core. A
supernova occurs when the nuclear burning of heavy elements within the core ceases,
resulting in a decrease in core pressure. When the core collapses, one percent of the
released gravitational energy is transferred to the hydrogen envelope as a shock wave. This
energy heats and ejects hydrogen and heavier elements from the outer core away from the
central core collapse (Friedlander, 1989). This ejected hydrogen, helium, and other
elements makes up the majority of galactic cosmic radiation.
Type n supemovae emit 10^^ ergs of energy. 99% of this energy is emitted as fast-
moving neutrinos (Dorman, 1974). Neutrinos are stable particles with a discernible energy
and momentum, but little or no mass; as a result, neutrinos have little effect on matter and
are geologically unimportant. The remaining 1% of the energy emitted by a supernova is consumed in the emission of gamma rays, x-rays, and cosmic rays (Fesen, 1992; Draganic et al, 1990). When a supernova occurs, this radiation is emitted radially in all directions
(Meyer, 1992). The propagation path for this radiation is of consequence to some cosmogenic isotope studies and can take two forms: 1) difftision along galactic magnetic field lines; or 2) by the Shockwave effect. 12
The galactic magnetic field is a weak, non-uniform field existing throughout the
Milky Way. The field is strong enough to affect the path of most cosmic ray particles.
Although particles with energies above lO'^^-lO'^ eV will remain largely unaffected by the galactic magnetic field, the majority of cosmic ray particles will diffuse along galactic magnetic field lines (Friedlander, 1989; Laster et al, 1968). Changes in the orientation or strength of the galactic magnetic field will result in alterations in the direction and intensity of galactic cosmic ray flux. Because the galactic magnetic field is not uniform across the galaxy, the Earth will encounter magnetic fields of different strengths and orientations as it passes through space. Theoretically, this could affect the flux of primary cosmic radiation entering the Earth's atmosphere.
Supernova shock waves can also produce changes in the primary cosmic ray flux influencing the Earth. Some researchers have hypothesized that supernova radiation may propagate as a shock wave rather than as diffuse particles (Sonett et al, 1987). Shock waves are large amplitude waves traveling at high velocities; supernova radiation would then be contained in the wave as a wave packet. Supernova derived shock waves will envelop the solar system and impinge on the Earth only if the supernova occurs within about 535 light years (ly) of the earth (Terry and Tucker, 1968). Within this distance, the shock wave front will strike the Earth, essentially causing the supernova radiation to reach the Eanh in a single burst (McHargue, 1994). Cosmogenic isotope production will increase dramatically over the short interval of shock wave-Earth interaction. Shockwaves would appear in the cosmogenic nuclide record as large amplitude, short-lived peaks in isotope concentration. Such peaks have been observed in '°Be inventories in ice cores and have been interpreted by some researchers as the result of supernova (McHargue, 1994).
The flux of primary cosmic rays will also be affected upon entering the geomagnetic sphere of influence. Cosmic ray particles can be described by their rigidity, r = pc/z, where p = particle momentum; c = speed of light; and z = particle charge. In order to enter the 13
Earth's atmosphere, a particle rigidity must be great enough to overcome the deflecting force of the geomagnetic field. This minimum required rigidity is known as the cut-off rigidity and is a direct function of the strength of the Earth's magnetic field. When discussing cosmic ray particles, it is easiest to think in terms of vertical cut-off. In attempting to quantify the effect of the geomagnetic field on cosmic ray flux, it is useful to generalize and assume that all particles enter the atmosphere orthogonal to the Earth's surface (or, in other words, vertically). The force exerted by the geomagnetic field on a cosmic ray particle will be perpendicular to both the regional field and the direction of cosmic ray flux. At the equator, the regional geomagnetic field is parallel to the Earth's surface. For the generalized case of vertical incidence at the equator, all incident particles are perpendicular to the Earth's surface and, therefore, to the regional geomagnetic field.
The vertical cut-off rigidity will be highest at the equator. Similarly, the force exerted at either pole, where the regional geomagnetic field is perpendicular to the earth's surface, will be zero for vertically incident particles.
The effect of this geomagnetic force on total cosmic ray flux is a function of the energy spectrum of the incident primary cosmic rays. The particles in cosmic radiation can have energies ranging from 10 MeV to 10-° eV (Friedlander, 1989). This variation in energy is a direct function of variations in particle velocity. Small decreases in velocity will result in large decreases in particle energy. When a cosmic ray particle encounters the geomagnetic field, the field deflects the particle perpendicular to the direction of particle motion and away from the Earth. If the particle is U-aveling fast enough (and therefore has a high enough energy), then the original direction of motion will be a large component of the particle's resultant motion. This geomagnetic deflection results in a cosmic ray "latitude effect".
At sea level, this latitude effect translates to 1.7 times more cosmic radiation at the poles than at the equator (Rose et al, 1956). This effect increases with elevation (Simpson 14 and Fagot, 1953; Simpson, 1951). A smaller but similar "longitude effect" has also been observed. Eastern Hemisphere cosmic ray flux at the equator is roughly 4-5% smaller than fluxes in the Western Hemisphere (Bowen et al, 1936). This disparity is likely a function of localized fluctuations in the geomagnetic field known as non-dipole field features.
The longitude effect is unimportant for the majority of cosmogenic isotope studies.
Atmospheric circulation mixes cosmogenic isotopes produced in the atmosphere, negating both the latitude and longitude effects. In situ cosmogenic isotopes, those produced at the
Earth's surface, will, however, experience short-term variations in production rate with sampling locality as a function of both the longitude and latitude effects. Fortunately, for most studies, the longitude effect is negligible. Most in situ cosmogenic isotope studies deal with isotopes that have been exposed for tens of thousands of years (Niedermann et al., 1994; Nishiizumi et al., 1989), with only a few studies dealing with samples with exposure ages <1000 years (Kurz et al., 1990). Measured cosmogenic isotope concentrations represent the production rate integrated over the length of exposure. As a result, short time scale variations in the cosmic ray flux, such as the 11-year cyclicity due to solar flares (Raisbeck et al, 1990), will average out of the data. Similarly, non-dipole field features, with lives on the order of 1000s years (Butler, 1992), will average out for long exposure times.
1.2.b. Solar cosmic rays
Solar radiation will not result in a net increase in cosmic ray flux incident on the atmosphere and Earth's surface. Rather, periods of increased solar activity are marked by a decrease in cosmogenic nuclide production by secondary neutrons. Solar radiation consists predominantly of low energy particles, primarily protons, which will have little impact on secondary neutron production. Additionally, during times of increased solar activity, these protons will interact with higher energy galactic cosmic ray protons, removing them from 15 the primary flux to the Earth. Therefore, solar flares, although increasing the total number of protons entering the atmosphere, change the energy spectrum of the primary cosmic ray flux. The depletion of cosmic radiation of high energy protons results in a decrease in secondary neutron, and therefore cosmogenic isotope, production (Oescheger and Beer,
1990; O'Brien, 1979). Increases in solar activity occur on a roughly 1 l-year cycle which has been correlated to measured decreases in galactic cosmic ray flux (Raisbeck et al,
1990).
1.3. History of Cosmogenic Isotope Studies
Cosmic radiation was first discovered in the early 1900s (see Hillas, 1972 for an in- depth review of the history of this branch of research). Cosmogenic isotopes, on the other hand, did not become widely studied until the early 1950s and 1960s. During this time, researchers attempted to quantify the cosmic ray flux as a function of many different parameters: geomagnetic latitude, longitude, elevation, and energy spectrum, to name a few. Data from ship-bound surveys at sea level (Rose, 1956) or balloon flights at mountain and upper atmosphere elevations (Lord, 1951; Simpson, 1951; Simpson and
Fagot, 1953) were used as a means of characterizing the cosmic ray flux. Models for the latitudinal and altitudinal variation in cosmic ray flux were developed; these were eventually used to determine associated variations in cosmogenic isotope production.
In order to compare in situ cosmogenic isotope studies conducted at different localities around the Earth, geologists had to develop a means by which to scale production rates as a function of the latitude and altitude of the sampling locality. As explained in section 1.2.a., cosmic ray flux attenuates with latitude due to the influence of the geomagnetic field. Attenuation with altitude, on the other hand, is due to interactions between cosmic rays and matter in the atmosphere. Finally, cosmic radiation will also 16
attenuate with depth in the Earth; this becomes important for in situ cosmogenic isotope
studies.
The production rate of an isotope, P(xi,yi), due to cosmic rays at any elevation,
yi, and depth within the Earth, xi, as a fianction of production at the surface of the earth at
sea level, P(Xoyo), can be written as:
where L and A are the attenuation pathlength, in g cm"^, in the atmosphere and in rock of
density, p, in g/cm^, respectively (Brown et al., 1991; Lai, 1991). This equation allows
cosmogenic isotope production rates to be scaled to any altitude and depth within the Earth.
The attenuation pathlength variables, L|j, Ln, A^, and An, have different values for
neutrons, n, and muons, |i. Ln ranges in value from 135 and 170 g cm-2 for sea level to
mountainous altitudes (Brown et al., 1991). Most researchers choose an intermediate value
for Ln of about 150. L^ is not as well measured, but has been estimated at about
250 g cm-2 (Lai, 1988). An and A^ are dependent upon the lithology of the target rock.
An will be very similar to Ln (Lai, 1991). Alternatively, muon attenuation is much slower
in rock than in air, corresponding to a A^ of about 1300 g cm-^ (Bilokon, et al., 1989).
Attenuation of muons and neutrons at ground level show a similar attenuation as a function
of latitude (Bilokon, et al., 1989).
Today, cosmogenic isotope production rate data is scaled to standard conditions,
usually sea level and high latitude (>60°). Researchers primarily use three methods to scale cosmogenic production rates with latitude and altitude. Since altitudinal attenuation of cosmic radiation is much larger than latitudinal, some researchers choose to ignore
latitudinal variation (e.g. Brown et al, 1991). Attenuation is then characterized solely by 17
equation (1). Because this treatment of attenuation does not take the latitude effect (the ratio
between cosmic ray flux at >60° and flux at 0°) into account, researchers using only this
attenuation equation primarily deal with samples exposed at high latitudes, where the
latitude effect is diminished.
Yokoyama et al. (1977) and O'Brien (1979) present alternative cosmic ray
attenuation calculations based on nuclear cascade models. These models, while based upon
similar assumptions, are in significant disagreement with each other. Additionally, small
modifications to any of the estimated nuclear cascade parameters will result in large errors
in calculated production rates (Lai, 1991). Accordingly, few researchers rely on these
models for scaling cosmogenic production rates.
The final and most popular method used by researchers to scale production rates
stems from a set of best-fit polynomials calculated from nuclear disintegration rates as a
function of altitude and geomagnetic latitude (Lai and Peters, 1967; Lai, 1991). These
polynomials are used almost uniformly by researchers to scale calculated production rates
to the sea level, high latitude standard. These polynomials were developed from neutron
flux and nuclear disintegration data gathered in the 1940's and 50's (Rose et al, 1956;
Simpson and Fagot, 1953; Simpson, 1951; Lord, 1951) and Lai (1991) predicts a 10% error on the latitude-altitude variation of this data. The perceived low errors resulting from
these polynomials resulted in them becoming the preferred method for scaling cosmogenic
production rates over the other published methods.
To scale cosmogenic production rates using the nuclear disintegration polynomials, rese2u:chers must first assume that cosmogenic production rates are proportional to nuclear disintegration rates. This assumption is not unwarranted in the lower atmosphere if we ignore neutron flux effects which occur close to the earth's surface (Bethe, 1940). To scale in situ cosmogenic production rates to sea level and high latitude, a nuclear disintegration rate at the sampling altitude and latitude. Si, is calculated from the polynomial best-fits of 18
Lai (1991). Dividing the measured production rate, P(yi,X,i), by this value and
multiplying by the nuclear disintegration rate at sea level and high latitude. So, will give the
scaled production rate, P(yo,^)- That is:
P(yo,^o) = P(yi,>.i)(So/Si) (2)
Although currently the best available technique, this scaling method relies on only a few
data points, some of which show a clear longitude effect (Simpson and Fagot, 1953;
Simpson, 1951). The influence of the longitude effect may distort the latitudinal scaling.
At low latitudes, especially for high elevation sampling localities, this may result in significant errors for production rates scaled to high latitude. 19
CHAPTER 2. PRESENT STUDY
The majority of this dissertation is contained within a series of appendices which
take the form of published, submitted, or progressing manuscripts. Each manuscript
synthesizes the information gathered during the course of the study and presents the most
interesting and relevant aspects as a coherent whole. The most important findings
documented in each of these manuscripts are presented below.
2.1. General Conclusions: Appendix A
This manuscript, entitled "Cosmogenic-'Ne in quartz as a potential paleoaltimeter:
Case study from the 27 Ma Fish Canyon Tuff, Colorado" was co-authored by myself, Drs.
C.G. Chase and J. Quade of the University of Arizona, Dr. J. Poths of Los Alamos
National Laboratory, and Dr. W.C. Mcintosh of the New Mexico Geochronology
Research Laboratory of the New Mexico Bureau of Mines at New Mexico Tech. This
manuscript will be submitted to Earth and Planetary Science Letters .
This paper documents the development of a new technique for determining
paleoelevation. Inherent assumptions and geologic constraints of available paleoelevation
techniques prohibits their applicability to all but a narrow range of geological situations.
The developed technique relies on the exponential relationship between altitude and cosmic
ray flux and is applicable in a wider range of geologic settings. The method was tested
through application to the 27 Ma Fish Canyon Tuff in the Colorado Rocky Mountains, allowing comparison with previously calculated Rockies paleoelevation. Although erosion of the Fish Canyon Tuff paleosurface prohibited a paleoelevation calculation for the
Colorado Rocky Mountains, we develop the feasibility of such an approach and provide guidelines for the application and improvement of the technique for use in other settings.
In particular, we were able to show that it is possible to distinguish cosmogenic -'Ne preserved in quartz from other components of-'Ne, even in a 28 Ma tuff. Estimating the 20
paleoelevation of a paleosurface from this cosmogenic component, while simple in theory,
requires the combination of several factors. A paleoelevation can only be calculated using
cosmogenic ^iNe inventories if background and modem cosmogenic levels of ^'Ne are
low; if the length of exposure of the ancient surface is measurable; and if the erosion
history of the study locality is well understood. In our case, erosion of the Fish Canyon
Tuff paleosurface restricted our ability to calculate a paleosurface. Units with relatively short exposure times and resistance to erosion will likely have the highest probability of success in determining paleoelevation. Finally, samples should be processed in a specific order to maximize project success. Exposure times should be calculated first to ensure a
time long enough for cosmogenic ^'Ne accumulation but short enough to limit erosion effects. Analysis of background and modem -'Ne levels will indicate the ability to isolate ancient cosmogenic -'Ne for in specific rocks.
2.2. General Conclusions: Appendix B
This manuscript, entitled "Two-isotope method for calculating paleoelevation" was prepared by myself. I anticipate that it will be submitted to Earth and Planetary Science
Letters for publication.
This manuscript presents the theoretical development of a modification of the paleoelevation technique presented in Appendix B. The two-isotope technique for estimating paleoelevation is based upon the measurement of two cosmogenic isotopes produced through different mechanisms, in this case neutron spallation (^'Ne) and negative muon capture (^SAr). Muons and neutrons attenuate at different rates both in the atmosphere and below the Earth's surface. As a result, the ratio between cosmogenic production resulting from muons and neutrons will change with changes in altitude and rock depth. Measuring two cosmogenic isotopes produced through different mechanisms preserved in a paleosurface and then comparing a ratio of these concentrations to expected 21
values will give a corresponding paleoelevation. Furthermore, the ^iNe technique
described in Appendix A requires several project constraints, primary of which is a
determination of the length of a surface's initial exposure to cosmic radiation. This
constraint severely limits the geological applicability of this technique as it requires a
surface surrounded by two datable units. The two-isotope technique presented here does
not require knowledge of exposure time, vastly increasing the range of application of the
technique. The estimation of paleoelevation from the measurement of multiple isotopes
removes several key assumptions and constraints from the one-isotope paleoelevation
technique described in Appendix B, including the determination of a surface exposure time.
For the two-isotope technique to truly become a viable technique, further analysis of
isotope production rates from negative muons at sea level must be conducted to clarify
current understanding of muon-induced cosmogenic isotope production.
2.3. General Conclusions: Appendix C
This manuscript, entitled "Tertiary remagnetization of Paleozoic rocks from the
Eastern Cordillera and sub-Andean Belt of Bolivia" was co-authored by myself, Drs. R.F.
Butler and D.R. Richards of the University of Arizona, and Dr. T. Sempere of Mision
Orstom en el Peru. This manuscript was published by the American Geophysical Union
and appears in Journal of Geophysical Research, v. 103, no. 312, p. 30,417-30,429,
1998.
This manuscript represents a comprehensive analysis of the magnetization carried
by eight Devonian to Permian sedimentary units in the Eastern Cordillera and the sub-
Andean Belt of Bolivia. The initial impetus for this study was the determination of
Paleozoic paleomagnetic poles for Gondwana from South American strata. Our analyses, however, revealed a wide-spread remagnetization which prohibits paleopole determination.
This documentation of such a pervasive remagnetization in an orogenic zone highlights the 22 importance of using multiple field tests for paleomagnetic stability, especially in deformed regions. In particular we were found that Devonian to Permian sedimentary units throughout the Eastern Cordillera and sub-Andean Belt of Bolivia have experienced extensive remagnetization. We believe that this remagnetization is associated with the
Andean orogeny, as evidenced by its syn-folding nature as revealed by the fold test.
Several studies of the same rock units assigned much older ages to this preserved magnetization; in light of this we conclude that any paleomagnetic study must utilize multiple tests of magnetic stability to ensure that a remanence is primary, especially in deformed regions such as the sub-Andean Belt. Finally, because of this wide-spread remagnetization, we suggest that future attempts to determine paleopoles from Paleozoic rocks in South America should focus on undeformed regions, such as the forebuige region of the active foreland basin. 23
APPENDIX A
COSMOGENIC 21NE IN QUARTZ AS A POTENTIAL
PALEOALTIMETER: CASE STUDY FROM THE 28 MA FISH CANYON
TUFF, COLORADO
J.C. Libarkin, J. Quade, C.G. Chase, J. Poths, W. Mcintosh 24
Abstract
Here we develop and test a new quantitative technique for determining paleoelevation. This technique relies on the exponential relationship between cosmogenic isotope production rates and elevation. We propose that paleoelevation can be calculated for a paleosurface of known exposure time through the measurement of cosmogenic 2lNe.
The ratio between ^iNe concentrations and paleoexposure time gives a production rate; this in turn can be used to calculate paleoelevation. To test this method, we measured the 2lNe inventory contained in the 27 Ma old Fish Canyon Tuff from Colorado. In detail, we had to consider several factors to ensure isolation of the component of 2lNe resulting from ancient exposure to cosmic radiation. First, because the Fish Canyon Tuff has been exposed in Colorado for the past few million years, we needed to ensure that significant modern cosmogenic -'Ne production had not occurred in our samples. To do this, we chose sampling localities for high rates of modem erosion and took samples of the Fish
Canyon Tuff exposed at the modern surface. We then measured the modern -'Ne concentrations in these samples and subtracted this component from our other sample inventories. Second, because the Fish Canyon Tuff is an old unit, we needed to account for any nucleogenic -'Ne contained in our samples. We processed the samples to remove as much of this component as possible. Additionally, we took several samples which were shielded from cosmic radiation during both the original and modem Fish Canyon Tuff exposures. ^'Ne inventories measured in these samples represent a background which was also subtracted from other sample concentrations. Finally, erosion of the Fish Canyon
Tuff surface during exposure 27 Ma is of major concern. Based on corrected sample inventories and a "^^Ar/^^Ar paleoexposure time of 506±13 kyr, we calculated a production rate (scaled to latitude > 60°) of 6.85±0.74 ^'Ne atoms (g Si02)*' a*'. This rate is much lower than the modem sea level production rate of 21±3.6 ^'Ne atoms (g Si02)"' a*'. This can best be explained by extensive erosion of the Fish Canyon Tuff during the interval 25 before Carpenter Ridge deposition. Although we were not able to accurately determine a paleoelevation for the Fish Canyon Tuff, we can suggest that future use of this technique will be most successful for paleosurfaces on erosion-resistant lithologies, with shorter paleoexposure times, or with known erosion histories.
1. Introduction
The uplift history of the Colorado Rockies is debated by many researchers. Until recent quantitative calculations of Colorado Rockies paleoaltitude, uplift to present elevations was thought to be a solely Miocene or younger phenomenon [1]. The presence of a Late Eocene erosion surface in the Rockies and the idea that regional erosion surfaces only form at or near sea level led most workers to argue that erosion must have kept pace with Laramide uplift. Consequently, the present 3-4 km's of elevation of the Eocene erosion surface was thought to result from a post-Laramide, possibly Pliocene, orogeny.
Paleobotanical evidence from the 35 Ma old Florissant flora of Colorado, however, yielded a paleoelevation estimate of 2.5 km, indistinguishable from present elevation [2, 3]. We set out to both test the contrasting theories for the timing of Rocky Mountains uplift and the potential for studying cosmogenic noble gases in pre-Quatemary rocks.
In situ produced cosmogenic noble gases are being used increasingly by geologists to date catastrophic events, determine exposure ages, and to evaluate erosion rates [e.g. 4,
5, 6, 7, 8, 9, 10]. Cosmogenic noble gas research has concentrated almost entirely on rocks younger than a few million years. In contrast, our approach involves measuring the inventory of cosmogenic ^iNe in buried weathering horizons on the upper surface of the 27
Ma old Fish Canyon Tuff. ^^Ar/^^Ar dating of the Fish Canyon and overlying Carpenter
Ridge Tuffs yielded a maximum exposure time for these sampled horizons. A paleoelevation can be calculated by dividing the measured cosmogenic ^iNe concentrations 26 by the exposure age. We attempted here to measure the 2lNe inventory in quartz separates from the Fish Canyon TufT exposed in south-central Colorado.
The flux of cosmic radiation increases exponentially with increasing altitude (Fig.
la). In general, the amount of ^'Ne produced from 28si(n,2a)2lNe reactions in quartz will be proportional to cosmic ray flux. If the length of unit exposure to cosmic radiation can be measured, the determined 2lNe concentration can be translated into an average annual 2lNe production rate. Compared to a sea level production rate of 21±3.6 2lNe atoms (g Si02)"' a*' [11], this value should constrain the altitude at which the sampled unit was exposed (Fig. la). This sea level production rate is due in the most part to cosmic ray neutrons, although, roughly 1 2lNe atoms (g Si02)"' a"' will result from muon interactions in quartz [12]. This muon production rate is too small relative to neutron-production to be of importance here.
Neon can also be produced from non-cosmogenic sources. Production of nucleogenic -^Ne from a particles [•S0(a,n)2iNe] produced by the decay of U and Th may be high in pre-Quaternary rocks, potentially complicating the -'Ne signal.
Additionally, 22Ne can be produced through reactions with F, '^F(a,n)22Ne [13].
However, this production of a-derived isotopes will only occur in the outer 40 ^m of the target grains [14]. Chemical etching of this outer rind should remove nearly all of the non- cosmogenic component of the isotope being studied. Remaining non-cosmogenic gas can be accounted for through the measurement of samples which have been shielded from cosmic rays. "Background" ^^Ne can be subtracted from measured ^'Ne concentrations, yielding the value of solely cosmogenically produced ^iNe.
The retention of noble gases in the target mineral is also of concern to researchers.
Mafic minerals, such as olivine and pyroxene, have been shown to completely retain ^He, while felsic minerals such as quartz and feldspar do not [8, 10]. 21 Ne, however, has a 27
6
5
4 Px^o 0 0.2 0.4 0.6 0.8 1.0 3
2
1
0 0 30 40
P/Po
Fig. 1: Attenuation of cosmogenic isotope production witii a altitude and b depth in the earth, where Pq is the production rate at the earth's surface at sea level, Py is the production rate at an elevation, y, in g/cm-, and Px is a production rate at a depth, x, in cm [21]. a Py = Poexp[(yo-y)/Ly] where yo is the atmospheric depth at sea level, taken as 1033 g/cm-, and Ly is the atmospheric attenuation pathlength, taken as 170 g/cm- b Px = Poexp[px/Lx] where p is bulk density in g/cm^ and Lx is the attenuation pathlength in the rock, taken as 165 g/cm-. 28
much larger atomic radius than ^He and is therefore less likely to "leak" out of quartz and
feldspar.
2. Geologic and Tectonic Setting
The San Juan Volcanic Field of central and southern Colorado contains an
impressive record of post-Laramide ignimbrite volcanism. The Fish Canyon Tuff is one of
the largest recognized ash flows in the world, with a volume of >3000 km^ [15]. The Fish
Canyon Tuff is mineralogically homogeneous from top to bottom. Extremely quartz rich
for a tuff, quartz and feldspar each make up one-third of the minerals in a typical Fish
Canyon hand-sample. At our study localities, the Fish Canyon Tuff is overlain by the
Carpenter Ridge Tuff, or by an intermediate sedimentary unit, the Los Pinos Formation,
which lies below the Carpenter Ridge Tuff (Fig. 2). The Los Pines Formation sampled in
this study contained abundant alluvial quartz. Because the Fish Canyon Tuff is the only
unit in the area which contains abundant quartz and because the Los Pinos Formation lies directly above the regionally extensive Fish Canyon Tuff, the Los Piiios Formation
probably contains only redeposited Fish Canyon quartz.
We chose to sample the Fish Canyon Tuff for several reasons. First, as described, this tuff contains unusually abundant quartz, the target mineral for cosmogenic -'Ne production. Second, the Fish Canyon Tuff has been extensively dated previously by
'^OAr/^^Ar [16]. Further ^OAt/^^At analysis of the Fish Canyon and Carpenter Ridge Tuffs by us yields a measurable time interval between Fish Canyon and Carpenter Ridge eruptions. Third, the Fish Canyon's homogeneity becomes important as we attempt to remove the nucleogenic component of ^^Ne. Finally, it is 7 my younger than the Florissant fossil beds, making it useful for comparison with a paleoelevation determined by independent means. 29
1 CR cliff / CR cliff
20 / 21 vitrophyre
/ nB 4. rjel . cm'^ jSl A /86 cm
Pink, soft, large pumices, Pink, soft, large pumices, some biotite. Fine, ashy matri.\. CR |CR some biotite. Fine, ashy matrix. slope of contact 5 cm of calcium carbonate, 14-25 cm of sand and silt, clay-rich changes from i I roots (modem) i 1 zones throughout, biotite, quartz, LP 15-30* ~ -iy_ flat fragments of FC White, abundant biotite. FC White, abundant biotite, quartz eyes. Ashymatri.x. quartz eyes. Ashymatriv FC
Fig. 2: Schematic of sampling localities. FC = Fish Canyon Tuff; CR = Carpenter Ridge Tirff; LP = Los Finos Formation. The dashed line refers to the contact between the FC and the overlying unit (either CR or LP), a Sites 01, 20-22 were located on a hill slope overlooking Lime IGln Creek. Site positions are shown relative to each other. Site 01 was located 6 vertical meters below site 20. b Orientation of sampling localities with the ground surface for sites 20 and 22. About one meter of overlying material was removed from above the Fish Canyon Tuff 27 Ma surface, c Generalized strati graphic sections for sites 20 and 22. 30
The present-day elevation at our study localities ranges from 2.4 to 2.8 km (Table
1). The contact between the Fish Canyon and Carpenter Ridge Tuffs is abrupt and obvious in the field (Fig. 2). Likewise, contacts with the Los Pinos Formation are easily differentiated. There is a notable lack of soil formation on the surface of the Fish Canyon
Tuff, suggesting that either the tuff was exposed under poor soil forming conditions or experienced erosion which out-paced soil forming processes. X-ray diffraction of samples from the surface of the Fish Canyon Tuff indicates the presence of clay minerals, suggesting that bentonization has occurred on this surface.
3. Sampling
Fish Canyon Tuff and Los Pirios Formation samples were taken from eleven localities in the south-central area of Colorado, at about 37°N, 106°W (Fig. 3, Table 1).
Fish Canyon samples were taken at 10±2 cm intervals below the Fish Canyon-Carpenter
Ridge or Fish Canyon-Los Piiios contact. The predicted exponential decrease in cosmogenic production with depth in a rock should be evident in these profiles (Fig. lb).
Additional samples of the Los Pinos Formation were also taken at 10±2 cm intervals.
Several steps were also taken to ensure that cosmogenic 2lNe produced during modem exposure did not contaminate the study. First, sampling localities were chosen from areas with high rates of modem erosion. Talus accumulation, little vegetation, and high slope gradient all suggested high erosion rates and, therefore, minimal exposure of the sampled section to modem cosmic radiation. Once the Fish Canyon-Carpenter Ridge contact had been identified, pits were dug about one meter below the modem surface to the contact to again decrease the likelihood of modem exposure to cosmic radiation (Table 2; Fig. 2). 31
Locality Locality Name Lat (°N) Long (°W) Elev. (m) Type
01 Lime Kiln Creek 37.608 106.324 2536±51 MS
7 Across from Road 3 ICC 38.175 106.479 2695±55 FC
8 38.194 106.489 2554±37 SS
9 38.14 106.501 2723±44 SS
13 38.169 106.406 2805±45 FC
20 Lime Kiln Creek 37.608 106.324 2536±51 FC
21 Lime Kiln Creek 37.608 106.32 2516±85 FC
22 Lime Kiln Creek 37.606 106.329 2705±51 FC
24 Road 3 ICC 38.178 106.476 2472±144 FC, LP
27 Taylor Canyon 38.169 106.488 2773±69 FC
Locality is the sampling locality; Locality Name refers to a local name for the site or nearby feature; Lat., Long, and Elev. are the latitude, longitude, and modern elevation of the sampling locality, respectively; Type refers to the unit sampled where MS = Fish Canyon Tuff samples taken at the modem surface; LP = Los Piiios Formation samples; FC = Fish Canyon Tuff samples taken near the ancient Fish Canyon surface; SS = Shielded Fish Canyon Tuff samples. 32
QO 40 N Denver
• Florissant
38 N
Gunnison
ii-C ,7 24 106 107 14S 09
Del Noi
01 20 21 22
SCALE Kilometers 0 20 40 60 80 100 1 I I I I I
Fig. 3: Map of the study area and sampling localities. Open circles are shielded Fish Canyon Tuff sampling localities; filled circles are other Rsh Canyon Tuff or Los Pinos Formation sampling localities. 33
3.1. Shielded samples
We collected three shielded samples (8, 9, and 14) that should not contain any
cosmogenic 20,2i,22is}e produced during Fish Canyon time or during recent exhumation.
Any neon contained in these samples must have been produced nucleogenically [13]. We
refer to the ^iNe contained in these samples as ^^Ne*background- Sample 8 was collected
from the back of a small cave, 2.5 horizontal meters from the modem slope. Sample 9 was
also collected from the back of a small cave and was completely shielded from the horizon.
Sample 14 was taken from a recent roadcut and was completely buried during Fish Canyon and recent time.
3.2. Modem Surface Samples
Samples of Fish Canyon Tuff exposed at or very close to the modem surface were processed for ^^Ne. These samples were taken six vertical meters below the Fish Canyon-
Carpenter Ridge contact and were therefore shielded from cosmic radiation during Fish
Canyon exposure 27 Ma. Any 2lNe above background levels contained within these samples must be the result of recent exposure. These samples allow us to quantitatively evaluate and remove modem cosmogenic ^^Ne (^'Ne*modeni) from our sample inventories.
Two samples of the modem Fish Canyon Tuff surface were taken. Sample 01A consisted of the first 10 cm of the exposed Fish Canyon; sample OIB represents the 10-20 cm depth interval. 40 cm of overlying colluvium were removed before this site was sampled; however, the colluvium was not well-developed and is not believed to have acted as shielding during the majority of the exposure presently affecting this site.
4. Analytical Methods
4.1 Quartz Separates
Quartz separates (grain diameter = 0.5 to 1.0 mm) were prepared using standard sieving, heavy liquid and magnetic techniques. Separated grains were ultrasonically 34 cleaned at room temperature for at least two 90-180 minute periods in a solution of 2% nitric acid (HNO3) and 1% HF [17]. This ultrasonic bath results in the removal of the outer 40 |im of the quartz grains and the dissolution of any feldspar grains remaining in the separate. Grains with visible inclusions are likely to contain nucleogenic 2lNe [14], so care was taken to remove grains with visible inclusions before ultrasonic cleaning. Finally, samples were ultrasonically rinsed in a 1% hydrochloric acid (HCL) and 2% peroxide
(H2O2) solution for one hour at room temperature.
4.2. Bulk Density
The bulk density of the Fish Canyon Tuff was measured using a standard liquid paraffin method [18]. Dry specimens from Fish Canyon Tuff samples from three sampling localities were weighed in air and then dipped in melted 60'C paraffin with density of 0.9 g/cm^. Specimens were again weighed in air (allowing determination of the weight, and volume, of the paraffin) and then placed in water and weighed a third time. The weight of the displaced water can be calculated by determining the difference between the weight of the specimen in air and in water. The volume of the specimen (plus paraffin) is equal to the weight of the displaced water. Subtracting the volume of the paraffin from this value and dividing the result into the weight of the specimen in air gives bulk density. The average of several specimen bulk densities was used as the Fish Canyon Tuff bulk density.
4.3. Mass Spectrometry
A double-focusing, 9.53 cm radius noble gas mass spectrometer at Los Alamos
National Laboratories was used to measure the neon amounts and isotopic composition in our samples. Neon was extracted by heating samples under high vacuum in the presence of a hot SAES getter. An activated charcoal getter cooled with liquid nitrogen was used to separate out noble gases heavier than neon. An air standard was analyzed after each sample 35 run to determine mass spectrometer sensitivity. Neon analyses are corrected for blanks measured before each sample run, electron multiplier mass discrimination, and double charge effects due to and CO2. '^^Ar contributes 5-19% to ^ONe values while CO2 contributes about 3% to 22Ne [19].
For samples analyzed during RUN #1, neon was extracted by heating the samples to 900°C for 90 min. Additionally, all of the samples in RUN #I were crushed in a liquid nitrogen environment before healing. High ^ONe concentrations in some of the samples from RUN #1 prompted us to modify our procedures for subsequent runs. All remaining samples, with the exception of 21A-900''C, were heated to 800°C and were not crushed, with the exception of 21A-cr. The agreement of the 2lNe* values for the crushed and uncrushed samples of 21A and 9 implies that crushing is not required to completely degas the sample. Additionally, the absence of excess 2lNe* in the reheat of sample 21A to
900°C indicates that extraction is completed with heating to only 800°C.
Cosmogenic ^'Ne^corr was calculated by subtracting 2lNe* measured in modern surface and shielded samples from total 2tNe*. ^'Ne*corr is excess cosmogenic ^'Ne calculated as:
2'Ne*corr — ^'Ne*t(jtai - (•'Ne*background + ~'Ne*tTiodem) (1) where 2lNe* is
2lNe* = 20Ne-[(2lNe/20Ne)n,eas - (2lNe/20Ne)air)] (2)
4.4. '*°Ar/^Ar Dermination of Maximum Exposure Time
The time interval between eruption of the Fish Canyon Tuff and the overlying
Carpenter Ridge Tuff was measured by single-crystal laser-fusion ''°Ar/^'Ar analysis.
Sanidine phenocrysts were prepared using standard crushing, magnetic, and high density liquid techniques, followed by etching in 15% HF to remove adhering matrix. Individual crystals were fused using a 10 watt CO2 laser and analyzed by the MAP 215-50 mass 36
spectrometer using methods similar to those described by Mcintosh and others [20].
Because uncertainty in the neutron flux received during irradiation is the single most
important factor limiting precision in studies of mid-Cenozoic sanidines, considerable
attention was paid to neutron monitoring. The Fish Canyon Tuff was used as the monitor
mineral, thereby avoiding some additional uncertainty that would inevitably be introduced if
a third nnineral (other than the Fish Canyon or Carpenter Ridge Tuffs) were used to monitor
neutron flux. In addition to being one of the "unknowns" in this study. Fish Canyon Tuff sanidine is internationally recognized and widely used as a flux monitor for '•^Ar/^^Ar dating. We accept 27.84 Ma as an eruption age for the Fish Canyon Tuff [16], although there is some variation in the age assigned to the Fish Canyon Tuff by different laboratories
[21].
Samples were irradiated at the Texas A & M reactor facility in two separate batches.
In the first irradiation (NM-66), sanidine crystals from the Fish Canyon Tuff at sites 14 and
22 and from the Carpenter Ridge Tuff at sites 24 and 27 were placed in alternating holes in standard aluminum irradiation trays. In the second irradiation (NM-76), sanidine crystals from a Fish Canyon Tuff standard and the Carpenter Ridge Tuff at site 27 were mixed together in the same holes, in an attempt to improve the precision of the neutron flux monitoring. This approach was made possible by the results of the first irradiation, which demonstrated that individual crystal age determinations from the two ignimbrites did not overlap (Fig. 4). In the end, this second approach to irradiation was not found to measurably improve precision relative to the first approach, so results from the two irradiations were pooled.
5. Noble Gas Results
Samples of the modem Fish Canyon surface contain the most ^'Ne*, followed by
Fish Canyon samples taken just below the contact with the Carpenter Ridge. Samples 37
intermediate depths yield intermediate values, with shielded samples containing the least
2lNe* (Fig. 5). These relationships are ideal for determination of a paleoelevation and are
concordant with the expected decrease in cosmogenic isotope production with depth in a
rock (Fig. lb). Low background 2lNe* values indicate that nucleogenic production does
not exceed cosmogenic production. Additionally, scaling of the modern
concentrations to the sampling depth below the modem surface yields modem 2lNe*
values which are below the values measured in the uppermost Fish Canyon. This indicates
that ancient cosmogenic 2lNe* is preserved in our samples.
5.1. Shielded samples (8, 9, 14)
Fish Canyon Tuff which was shielded from modem and ancient cosmic radiation
sampled at localities 8, 9, and 14 yielded the lowest 2lNe* inventories of all the Fish
Canyon Tuff samples (Fig. 5; Table 2). The total concentrations of 2lNe* in samples 8
and 14 were 10.95±0.56 and 5.29±0.27 x 10^ atoms (g Si02)"', respectively. Two
separate measurements of sample 9 yielded concentrations of I0.00±0.51 and 9.29±0.47 x
105 2lNe atoms (g Si02)''. Sample 14 received twice as much time in HF than the other
samples in the study; this is likely the cause of the decreased 2lNe concentrations relative to
the other shielded samples. The average 2lNe*background value calculated from samples 8
and 9 is taken as 10.01±0.30 x 10^ atoms (g Si02)"'. This value is subtracted from each sample 2lNe* to correct for 2lNe*background-
5.2. Modem surface samples (01 A, 01B)
Two samples of the Fish Canyon Tuff exposed at the modem surface were obtained at Site 01. Samples 01A and OIB yielded ^iNe* values of 67.78±3.88 and 46.88±2.81 x
105 2lNe atoms (g Si02)'' after background correction, respectively and at their respective sampling depths. Scaling these values to a surface concentration and 38
2-)——.—^^
Carpenter Ridge Tuff Fish Canyon Tuff 27334 ±0.013 27.835 ±0.011 Ma
! 1 ' 1 1 27.0 27.2 27.4 27.6 27.8 28.0
Age (Ma)
Fig. 4: Single-crystal laser-fusion ''^Ar/^^Ar results for sanidine phenocrysts from Fish Canyon Tuff and Carpenter Ridge Tuff. These results indicate a yield a 506±13 kyr interval between Fish Canyon Tuff and Carpenter Ridge Tuff eruptions. Only data points from irradiation NM-66 are shown, to minimize overlapping of data points. 39
2iNe* (x 10® atoms g*') 2 5 -t- OlA OIB 21A 21A-cr ^ ^ J ^ / / / / / ^ f/7~r7-7i
20A 22D SSSSSSSXXS3 27C SSSSSSSSS3 22C SSSXSSSSS3 13A SSSSSS3SS3 FC samples 7A sxsssxssa • Modem Surface 22B SSSSSSSSSl 13C sssssssss Ai conact with CR 20F SSSSSSS3 Below contact with CR 7 (90 cm) SSSSS3 24C sSSSSI Shielded 24D sSSSN=: 24A • LP samples 8 w v^ 9 WVVj
9 vv^ 24X 14
Rg. 5: Excess ^^Ne (^^Ne*) measured in all 22 samples. Samples have not been corrected for background or modem 21Ne. 40 averaging them yields an average ^'Ne*niodern of 70.71±3.47 x 10^ 2lNe atoms
(g Si02)''. This value is scaled to an average sampling depth below the modem surface and subtracted from each sample 2iNe* to correct for modem ^^Ne contamination.
The equation used to scale cosmogenic isotope concentrations with depth is
C(x) = C(o) exp[-px/Lx] (1)
where C(x) is concentration measured at depth, x, in cm; C(o) is concentration at the surface; p is the bulk density of the target rock, in g/cm^; and L* is the attenuation pathlength in the rock, taken here to be 165 g/cm-. Samples 9, 20A, and 27C yielded densities of 2.32, 2.74, and 1.91 g/cm^, respectively, giving an average density of 2.32 g/cm^. This average density was used as the density of the Fish Canyon Tuff throughout this study.
5.3. 27 Ma Fish Canyon Tuff surface samples
Maximum concentrations of ^'Ne* for the 27 Ma Fish Canyon Tuff surface occur in samples 20A and 21 A. One measurement of 20A and two of 21A yielded ^iNe* values of 26.6712.71, 28.57±2.81, and 30.08±2.88 x 10^ 2lNe atoms (g Si02)-', respectively
(Table 2, Fig. 5). As with sample 9, the agreement of the 21A measurements within error demonstrates the reproducibility of our results. Other samples of the 27 Ma surface yielded
2lNe* values well below these maximum concentrations.
5.4 .Los Pinos Formation
The Los Pinos Formation and Fish Canyon Tuff sampled at site 24 yielded low
2i{vie* values of 8.87-13.36 x 10^ ^ijsie atoms (g Si02)"' (Fig. 5, Table 2). After correction for ^^Ne*(jackground ^nd ^^Ne^moderir these samples were found to contain no 41 excess ^'Ne*corr- Either these samples contain no ancient cosmogenic ^iNe or the
2lNe*n,odem value used is an overestimate for this sample locality.
6. '*®Ar/3'Ar Results
'^OAr/^^Ar results from single-crystal laser fusion analysis of sanidine from the Fish
Canyon Tuff (n = 303 crystals) and the overlying Carpenter Elidge Tuff (n = 201 crystals) are shown in Figure 4. Results from each unit form tightly grouped Gaussian distributions that do not overlap. The mean age for the Fish Canyon Tuff samples analyzed as unknowns, relative to the assumed Fish Canyon Tuff monitor age of 27.84 Ma, is 27.835
± 0.011 Ma (all errors quoted at ± 2 a), which serves to demonstrate the reproducibility of the method. If an age of 27.840 Ma is accepted for Fish Canyon Tuff, then the mean age for Carpenter Ridge Tuff is 27.334 ± 0.013 Ma, indicating that 506±13 kyr elapsed between eruption of these two regional ignimbrites. Because "^^Ar/^^Ar dating is a relative method, with precision much higher than the established accuracy of monitor ages [16], the
506 ± 13 ka determined for the Fish Canyon Tuff/Carpenter Ridge age difference is more accurate than the absolute ages determined for either ignimbrite.
6. Discussion
A 22Ne/20Ne to 2lNe/20Ne three-isotope plot can be used to determine the true cosmogenic component of the 2lNe* measured in our samples (Fig. 6; [22]). Cosmic ray spallation reactions will produce each of the three neon isotopes in roughly equal amounts, resulting in a shift from the base air value along a line of slope = 1 (Fig. 6, solid line). Production of neon from (a, n) reactions in oxygen will result in ^iNe production, little 20isfe production, and no 22Ne. This contribution to the neon system in our samples will result in a shift towards the horizontal on the three-isotope diagram, (a, n) reactions in fluorine and magnesium will produce 22Ne exclusively and a shift towards the vertical. A three- 42
isotopw plot of all 23 samples after conection for background and modem 2lNe* shows
sample 14, 21A-900''C, and site 24 samples as outliers. A linear regression (dashed line)
through all other data yields a best-fit slope of 1.65 (Fig. 6). This value is above maximum
values accepted by other researchers (i.e. 1.24 in [22]). However all samples except for
the three samples processed in Run #3 (20F, OlA, and OIB) and sample 14 fall within
analytical error along the air-pure cosmogenic mixing line, indicating that we have isolated
cosmogenic 2lNe* in these samples after we make corrections for background and modem
contamination [22]. Sample 14 received the most HP treatment, suggesting a connection
between outlying samples and HP interference. HF can interfere with ^ONe; it is likely that
these samples have moved horizontally away from the ideal mixing line due to excess
20Ne. Excess ^ONe will result in a low calculated ^^Ne concentration; in fact, sample 14 contained anomalously low background levels of 2lNe relative to other shielded samples.
Unfortunately, the modem surface samples are also outliers; the calculate modem -'Ne component is likely a minimum value. In general, the analytical methods used, primarily sample etching and measurement of -'Ne in shielded and modem-exposed samples, were sufficient to allow us to isolate the cosmogenic ^'Ne component in most samples, although
HF is likely interfering with 20Ne in some cases.
We can use the measured modem 2lNe* concentration of 70.71±3.47 x 10^ ^ijsje atoms (g Si02)*' to determine the length of exposure of the Fish Canyon Tuff to modem cosmic radiation. Assuming an average elevation of 2.5 km, and therefore a production rate of -130 2lNe atoms (g Si02)''a-', an exposure time of -55 kyr is calculated.
Although we chose our sampling localities for geomorphic features which implied rapid erosion (such as steep slopes or little vegetation), there is still an important component of modern ^iNe"* present in our samples. Researchers must be careful not to mistake 43
.65x +^0.098 0.115
^ 0.105 • Air value
A 8,9,9 •OIA.OIB
0.095 A 14 • 24A. 24C, 24D. 24X O 21A-900*C O All other samples I I I 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
2»Ne/20Ne
Fig. 6: Three-isotope diagram for neon. All 22 samples plus the 21A-900°C reheat are plotted with error bars. The gray square is the air ratio, where (-lNe/20Ne)air = .00296 and (--Ne/^ONe)air = 0.1020. The gray triangles are shielded samples, 8 and two measurements of 9, used to calculate -^Ne*backEround- The open triangle is the depleted shielded sample, 14. The open circle is 21A-900 C. The black diamonds are the samples used to calculate -^Ne*niodem. OlA and OIB. The open squares are samples from site 24. The solid line is the ideal line along which cosmogenic neon should fall (slope = 1). The dashed line is the best fit linear regression through all samples except samples from site 24 and 14, with slope = 1.65. 44 this component for an ancient signal.
In order to calculate a production rate, the length of exposure of the target surface must be measurable. This requires a unique situation in which both the over- and under lying units can be dated. We chose to sample the Fish Canyon Tuff surface in part because both the Fish Canyon and Carpenter Ridge Tuffs are datable. We were entirely successful at generating an "exposure age" using '^OAr/^^Ar. In fact, the less than 5% error on this measurement suggests that this technique will be feasible in other study localities as well.
Calculated production rates, Px, from the 8 samples which contain 2'Ne*corr much lower than expected for production at sea level (Fig. 7b). It is obvious, therefore, that these samples are depleted in ^'Ne^corr- The low 2'Ne*corr values for all of the samples indicate that erosion deeply affected the Fish Canyon Tuff surface during its half a million years of exposure. The magnitude of the length of exposure is itself a likely indicator of erosion and was much longer than we had initially suspected. Attempting this technique on a surface with a much shorter exposure time (< 100 kyr) should dramatically reduce the effects of erosion on the studied unit. Additionally, 2>Ne inventories in units similar to the
Los Pinos Formation can be used to generate erosion models for the target surface.
As described in Section 2., we believe the Los Pinos Formation contains quartz grains eroded out of the Fish Canyon Tuff. Although our samples of the Los Piiios
Formation did not contain any 2'Ne*corr. similar sedimentary samples can be used to reconstruct the relationship between 2lNe concentrations and erosion rates [19,23]. The shape of the ^iNe profile expected in a preserved 3 m thick sedimentary unit is a key to erosion rates (Fig. 8). The profile shown here results from the erosion of a target unit, such as the Fish Canyon Tuff, and eventual redeposition of the eroded material. Here, we treat the erosion and deposition as continuous phenomena. Four erosion scenarios are presented, with total erosion of the parent unit ranging from 0.1 to 40 m for the duration of exposure to cosmic radiation, in our case 500 kyr. The position of the maximum ^^Ne 45
Px/Po
0.1 0.2 0.3 0.4 —H- —I—
10 --
20 + o. u Q 30 00 c I 40 + CO C/3 50 f
60 .. ^
Fig. 7: Calculated production rates versus depth below the Fish Canyon-Carpenter Ridge contact for eight samples which yield excess -^Ne*coiT- Po is taken as 21 ±3.6 -^Ne atoms (g Si02)"^ a"l [11]. a Production rates versus sampling depth, b Calculated production rates plotted with curve showing attenuation of cosmogenic productions rates witfi depth in the earth. 46 concentrations changes as total erosion of the parent unit changes; low erosion rates result in maximum ^iNe build-up. Increased rates of erosion result in lowered 2iNe build-up and a shift of the point of maximum ^iNe towards the bottom of the section. Using this technique to calculate average erosion rates can reveal erosional information which can have substantial effects on this paleoelevation technique.
An ideal study using the ^iNe paleoelevation technique should ensure that:
1) Cosmogenic 2lNe inventories are abundant enough for mass spectrometric analyses;
2) Background levels of a-produced 2lNe are low relative to cosmogenic 2lNe; 3) Modem cosmogenic 2lNe build-up is low relative to ancient 21 Ne; 4) The exposure age of the target surface is measurable using ^^Ar/^^Ar or a similar technique; and 5) The erosion history of the study locality is well understood or measurable.
7. Conclusions
Successful application of this paleoelevation technique requires that ancient cosmogenic ^iNe be distinguished from other types of ^iNe. Every attempt should be made to remove modem cosmogenic and nucleogenic components of -'Ne from the sample. This includes removing at least one m of overlying material from the surface before sampling and chemical etching of quartz grains before mass spectrometric analysis.
Contaminating ^^Ne components can not be completely removed, however, requiring analysis of modem surface as well as shielded samples. Additionally, erosion during the initial exposure will greatly affect the ancient ^'Ne signal. It is important to ascertain the extent of erosion and erosion rates before a paleoelevational estimate can be made.
Measured ^^Ne concentrations from 22 Fish Canyon Tuff and Los Pinos Formation samples yielded a maximum inventory of 30.08±2.88 x 10^ atoms (g Si02)"'. The resulting production rate (scaled to latitude > 60°) of 6.8510.74 ^iNe atoms (g Si02)"' a"' is much lower than the modern sea level production rate of 21±3.6 ^iNe 47
Sedimentary unit
Bedrock Total erosion of parent unit 0.1 m
— — — 3 m — • — • — 40 m
2 4 6 8 10 2'Ne concentration (atom §*0
Fig. 8: Effect of continuous deposition over 500 kyr on the preserved exponential profile in a sedimentary unit. Four scenarios are shown; erosion of the parent unit (in our case the Fish Canyon Tuff) ranges from 0.1 to 40 m. 48
atoms (g Si02)"' a*' and thus no paleoelevational information was obtained. This can best
be explained by extensive erosion of the Fish Canyon Tuff during the interval before
Carpenter Ridge deposition.
The extensive erosion which cx:curred on the Fish Canyon Tuff surface results from
both the erodable nature of the tuff itself and the extended period of exposure. We suggest
that a similar study conducted on less erodable, extrusive felsic rocks with a minimal
exposure time should provide the needed insight into the history of Colorado Rockies
uplift. Finally, future studies should collect and process samples in a careful order to
ensure project success. Initial determination of exposure time followed by mass
spectrometric analysis of shielded and modem surface samples will ensure early on that
conditions for a successful paleoelevation determination have been met.
Acknowledgments
We would like to thank Dr. Laurie Brown for providing preliminary rock magnetic
analyses and Dr. Peter Lipman for assistance in the field and his invaluable expertise in the
San Juan Volcanics. Dr. Bill Phillips provided assistance with mass spectrometric analyses and along with Dr. Nat Lifton provided encouragement and insight during the course of
this study. This research was part of a Ph.D. thesis completed by Julie Libarkin at the
University of Arizona. 49
REFERENCES
[1] W.M. Davis, Tlie Colorado Front Range: a study on physiographic presentation, Association of American Geographers Annals 1 (1991) 21-83.
[2] K.M. Gregory and C.G. Chase, Tectonic significance of paleobotanically estimated climate and altitude of the late Eocene erosion surface, Colorado, Geology 20 (1992) 581-585.
[3] H.W. Meyer, Lapse rates and other variables applied to estimating paleoaltitudes from fossil floras, Palaeogeography, Palaeoclimatology, Palaeoecology 99 (1992) 71-99.
[4] J. O. Stone, G.L. Allan, L.K. Field, R.G. Cresswell, Cosmogenic chlorine-36 from calcium spallation, Geochim. Cosmochim. Acta 60 (1996) 679-692.
[5] T.E. Cerling and H. Craig, Geomorphology and in-situ cosmogenic isotopes, Ann. Rev. Earth Planet. Sci. 22 (1994) 273-317.
[6] R.J. Foreda and T.E. Cerling, Cosmogenic neon in recent lavas from the western United States, Geophys. Res. Lett. 19 (1992) 1863-1866.
[7] T. Graf, C.P. Kohl, K. Marti and K. Nishiizumi, Cosmic-ray produced neon in Antarctic rocks, Geophys. Res. Lett. IS (1991) 203-206.
[8] M.D. Kurz, D. Colodner, T.W. Trull, R.B. Moore and K. O'Brien, Cosmic ray exposure dating with in situ produced cosmogenic ^He: results from young Hawaiian lava flows, Earth Planet. Sci. Lett. 97 (1990) 177-189.
[9] D. Lai, In situ-produced cosmogenic isotopes in terrestrial rocks, Ann. Rev. Earth Planet. Sci. 16 (1988) 355-388.
[10] T.E. Cerling, R.J. Poreda and S.L. Rathbura, Cosmogenic ^He and ^iNe age of the Big Lost River flood. Snake River Plain, Idaho, Geology 22 (1994) 227-230.
[11] S. Niedermann, T. Graf, J.S. Kim, C.P. Kohl, K. Marti and K. Nishiizumi, Cosrnic-ray-produced 2lNe in terrestrial quartz: the neon inventory of Sierra Nevada quartz separates, Earth Planet. Sci. Lett. 125 (1994) 341-355. 50
[12] S. Charalambus, Nuclear transmutation by negative stopped muons and the activity induced by the cosmic-ray muons. Nuclear Physics A166 (1971) 145-161.
[13] I. Yatsevich, and M. Honda, Production of nucleogenic neon in the Earth from natural radioactive decay, J. Geophys. Res. 102 (1997) 10,291-10,298.
[14] J.L. Turner, Location of uranium and thorium in granites-radiogenic influence on cosmogenic isotope abundance, Geol. Soc. Amer. Abstracts with Programs 25 (1993) A-89.
[15] J.A. Whitney and J.C. Stormer, Mineralogy, petrology, and magmatic conditions from the Fish Canyon Tuff, Central San Juan Volcanic Field, Colorado, J. of Petrology 26 (1985) 726-762.
[16] P.R. Renne, C.C. Swisher, A.L. Deino, D.B. Karner, T.L. Owens, D.J. DePaolo, Intercalibration of standards, absolute ages, and uncertainties in 40Ar/39Ar dating, Chem. Geol. 145 (1998) 117-152.
[17] C.P. Kohl, K. Nishizumi, Chemical isolation of quartz for measurement on in-situ- produced cosmogenic nuclides, Geochim. Cosmochim. Acta 56 (1992) 3583-3587.
[18] J.R. Johnston, An accurate method for determining volume of soil clods, Soil Science 58 (1945) 449-452.
[19] W.M. Phillips, E.V. McDonald, S.L. Reneau, J. Poths, Dating soils and alluvium with cosmogenic ^^Ne depth profiles: case studies from the Pajarito Plateau, New Mexico, USA, Earth Planet. Sci. Lett. 160 (1998) 209-223.
[20] Gather, S.M., Chamberlin, R.M., Chapin, C.E., and Mcintosh, W.C., Stratigraphic consequences of episodic extension in the Lemitar Mountains, central Rio Grande Rift in Basins of the Rio Grande Rift; structure, stratigraphy, and tectonic setting, Randy K.G. and Gather, S.M. (eds.), GSA Special Publication 291 (1994), p. 157-170.
[21] A.L. Deino, R. Potts, Single-crystal ^^Ar/^^Ar dating of Olorgesaille Formation, southern Kenya Rift, J. Geophys. Res. 95 (1990) 8453-8470.
[22] S. Niedermann, T. Graf, K. Marti, Mass spectrometric identification of cosmic-ray produced neon in terrestrial rocks with multiple neon components. Earth Planet. Sci. Lett. 118 (1993) 65-73. [23] W.M. Phillips, Applications of noble gas cosmogenic nuclides to geomorphology, Ph.D., University of Arizona, 1998.
[24] D. Lai, Cosmic ray labeling of erosion surfaces: in situ nuclide production rates and erosion models. Earth Planet. Sci. Lett. 104 (1991) 424-439. APPENDIX B
TWO-ISOTOPE METHOD FOR ESTIMATING PALEOELEVATION
J.C. Libarkin 53
Abstract
Currently available paleoelevation techniques are inadequate for the task of determining paleoelevations for a wide range of geological situations and time periods.
Assumptions inherent to available methods and the conditions necessary for success limit their applicability. Here, a new paleoelevation technique is developed which utilizes a minimal set of assumptions and which can be applied under a variety of conditions. This two-isotope technique relies upon the relationship between cosmogenic stable isotope production and elevation. The ratio between 2lNe, produced through neutron interactions with quartz, and ^SAr, produced from muon capture in sanidine, is used as a proxy for paleoelevation. This technique is an improvement upon one-isotope cosmogenic paleoaltimetry which requires an exposure time calculation and assumptions regarding primary cosmic ray flux and erosion. The ^iNei^^Ar technique does not require knowledge of exposure time and changes in the primary flux. Additionally, erosion of the exposed surface can be accounted for. Although 3®Ar production rates can be theoretically calculated, measurement of this rate in geologic strata will decrease the uncertainty associated with the technique.
1. Introduction
Recent work by Libarkin and others (1999) suggests that cosmogenic isotopes are a potential source of paleoelevation information. They developed a method for calculating paleoelevation using cosmogenic ^'Ne concentrations in quartz. This new paleoelevation technique relies on the exponential relationship between elevation and cosmogenic isotope production rate. Like other methods for calculating paleoelevation (Gregory and Chase,
1992; Sahagian and Maus, 1994; Meyer, 1992; Forest et al., 1995), this one-isotope technique is itself limited to a small range of geological situations. The potential for large errors in calculated paleoelevations using any of these paleoelevation techniques requires 54 application under specific geologic conditions. Clearly, a need exists for a paleoelevation technique which is not as geologically restricted.
2. Paleoelevation Techniques
2.1. Previous Methods
Few techniques have been developed for determining paleoelevation quantitatively.
Leaf morphology, basalt vesicularity, and cosmogenic isotope production rates are all correlatable to atmospheric phenomena which are in turn related to elevation. The accuracy and feasibility of each paleoelevation technique depends upon the altimetric property being calculated. Leaf morphologies of ancient plants can be used to determine both paieotemperature (Gregory and Chase, 1992; Meyer, 1992) and paleoenthalpy (Forest et al., 1995), which each vary with elevation. Basalt vesicularity studies can be used as a means of calculating paleopressure (Sahagian and Maus, 1994). Cosmogenic isotopes are used to calculate cosmogenic production rates (Libarkin et al., 1999). Each of these techniques have large errors associated with them and are geologically restricted.
Paleoelevation can be calculated from temperature-elevation relationships and paleobotanical data (Gregory and Chase, 1992; Meyer, 1992). Paleoelevations are calculated from the observation that leaf morphology characteristics (shape, size, leaf margin type, etc.) change with temperature. Paleobotanically calculated paleotemperatures are then correlated with elevation. Reported errors in paleoelevation derived from botanical paieotemperature studies can range up to 1 km (Gregory and Chase, 1992). This error does not take climatic effects into account; the temperature-elevation relationship may change as climate changes.
The difference between enthalpy, H, at sea level and study site enthalpy, as derived from leaf morphology, offers an improvement of the basic paleobotanical technique (Forest et al., 1995). Formal errors in paleoelevation derived from enthalpy calculations are less 55
than those calculated from estimated paleotemperatures. Although potentially superior to
the paleobotanical methods of Gregory and Chase (1992) and Meyer (1992), here too
assumptions affect this technique's applicability. Primarily, the assumption that the total
specific energy content of air, h (= H + gZ; where g = gravitational acceleration and Z =
elevation), is independent of longitude must be made. This invariance of h implies that
large-scale atmospheric flow is zonally symmetric (trends in one direction within a
latitudinal band). While this is true on a large scale, topographical features can perturb this
flow symmetry. Therefore, if the area under study had a high paleoelevation, the disruption in atmospheric flow patterns would result in non-correlation between
paleoenthalpy and elevation. Finally, all paleobotanical methods, whether temperature or energy driven, are limited to elevations below the tree line (Sahagian and Maus, 1994).
Paleoelevation can also be inferred from paleopressure calculations. Basalt
vesicularity studies are based upon the assumption that the size of basalt vesicles is a
function of atmospheric pressure (Sahagian and Maus, 1994). Specifically, the ratio
between vesicle volumes at the top and the bottom of a basalt flow is a function of atmospheric pressure at emplacement, lava flow thickness, and lava density. Pressure calculations can only be resolved to 0.1 bar, translating to errors in paleoelevation of 1 km or more (Sahagian and Maus, 1994). Additionally, to avoid the possibility of an external source of, or disturbance of, vesicle formation, only lava flows with simple emplacement histories can be analyzed using this method. Finally, method sensitivity is dependent upon flow thickness. At sea level, this means that optimal flow thickness is 3 m (Sahagian and
Maus, 1994). The desired thickness of the flow changes with changes in elevation; however, if a paleoelevation is being calculated, the desired flow thickness is unknown. 56
2.2. Cosmogenic one-isotope technique
Libarkin and others (1999) develop a technique for calculating paleoelevation which
relies upon the relationship between cosmic ray flux and elevation. A cosmogenic isotope
production rate can be calculated from measured cosmogenic isotope concentrations and a
known paleosurface exposure time. Comparing this production rate to a known sea level
rate and using the production rate-elevation relationship will give a paleoelevation for the
studied surface. The production rate of an isotope (usually by neutrons), P(xi,yi), due to
cosmic rays at any elevation, yi, and depth within the earth, xj, as a function of production
at the surface of the earth at sea level, P(Xoyo), can be written as:
/•(xi.).,) = /'(.Vg, ). (exp[(y„ - j., )/tj) . (exp[-x,p/A]) (I)
where L and A are the attenuation pathlength, in g/cm^, in the atmosphere and in rock of
density, p, in g/cm^, respectively (Brown et al., 1991; Lai, 1991). Applying this equation
to paleoelevation studies requires the assumption that the primary cosmic ray flux rate is
constant through time.
In order to use this relationship to calculate the paleoelevation of a surface, several
geological conditions must be met (Libarkin et al., 1999). Primarily, the studied surface
must contain measurable amounts of the cosmogenic isotope, requiring that the surface be exposed for at least a minimum amount of time; contributions to the isotope inventory from
modem cosmogenic and non-cosmogenic sources must be small relative to ancient cosmogenic contributions; the exposure time of the surface must be measurable with a high degree of precision; and erosion of the studied surface during its initial exposure must be low or measurable. 57
These conditions and the assumptions inherent to the elevation-cosmogenic
production rate relationship can result in several potentially large sources of error, namely:
1) Fluctuations in cosmic ray flux will cause changes in the production rate versus
paleoelevation curve, these fluctuations resulting from changes in the galactic cosmic ray
flux or changes in the geomagnetic field, for instance; 2) Discrepancies in the actual
paleoexposure time of the sampled paleosurface and the measured paleoexposure time can
result from limitations of available dating techniques; and 3) A relatively small amount of
erosion of the sampled paleosurface will yield isotope concentrations which are far below
predicted values.
I have developed a theoretical technique which avoids some of the issues central to the cosmogenic one-isotope paleoelevation method. Rather than using one isotope to determine paleoelevation, two isotopes with different modes of production are used.
Paleoelevation is estimated by comparing a ratio of the concentrations of two measured isotopes to predicted values. In using this ratio, the question of exposure time becomes unimportant; exposure time disappears from the two-isotope paleoelevation equation.
Additionally, changes in the magnitude of the cosmic ray flux will have only a small effect on the predicted isotope ratio. Finally, this technique also allows a paleoelevation estimation under steady-state erosion conditions. Since steady-state erosion is more likely than a case of minimal to no erosion, this is advantageous. If exposure time exceeds a required length, then the rate of steady-state erosion becomes unimportant. Since calculations of erosion rate carry their own sources of uncertainty, removing this rate from our consideration results in a decreased uncertainty for the paleoelevation estimation. 58
3. Theory
3.1. Two-isotope method for calculating paleoelevation
From (1) it is seen tiiat a relationship exists between cosmogenic production rate and elevation. Assuming that this isotope is produced by two different cosmic rays, negative muons, |1, and neutrons, n, the production rate P(xiyi) would then be:
(2)
For the remainder of this paper, the exponential changes in production rate due to depth in the earth, x, or elevation, y, will be expressed as /(x,r) and /(y,r), respectively, r denotes a cosmic ray, in this case either |I or n. The attenuation pathlength variables, L^, Ln, A^, and An, have different values for neutrons, n, and muons, p.. Ln ranges in value from 135 and 170 g cm*-for sea level to mountainous altitudes (Brown et al., 1991). Most researchers choose an intermediate value for Ln of about 150. is not as well measured, but has been estimated at about 250 g cm*- (Lai, 1988). An and A^ are dependent upon the lithology of the target rock. An will be very similar to Ln (Lai, 1991). Alternatively, muon attenuation is much slower in rock than in air, corresponding to a A^ of about 1300 g cm^-
(Bilokon, et al., 1989).
Given that production rate, P, can be expressed as isotope concentration, C, divided by the time, t, over which production has occurred, we can write (2) as:
Clt= f{x,n)* f{x,n) (3) 59
If we measure two isotopes, A and B, each of wiiich are produced by negative
muons, p. and neutrons, n, then the ratio of their concentrations can be written as:
A simplified notation will be used throughout the rest of this paper, such that:
(5) Cg(x,,yJ Cg D3 •/(>-,/:)+ D^./(>.,n)
Therefore, the paleoelevation, yi, in g cm-^, of a surface containing two cosmogenically produced isotopes, A and B, can be expressed as a function of their concentrations (Fig. 1):
" • In Hi -In D,-D^. kJ (6)
Notice that this method for determining paleoelevation does not require the determination of an exposure age. Changes in the magnitude of the primary cosmic ray flux will have little effect on the CA:CB-elevation relationship. This method cannot, however, account for changes in the energy spectrum of the cosmic ray flux. The independence of this technique from exposure time and changes in the magnitude of the primary cosmic ray flux 60
7.1 0 km
V 6.1 X — — 1 km > 2 km 5.1 N S ——3km 4.1 r-. ^ % I 'X. - — 4 km 2 L ^C«3 3.1 5 km k1 ^— ^ — - 6 km
200 Depth (cm)
Figure 1. Ratio of cosmogenic ^iNe (in quartz) to cosmogenic 38Ar (in sanidine) for elevations of 1-6 icm and depths within the earth of 0-200 cm for a paleosurface which does not experience erosion during its paleoexposure. Variables in eq. (6) are set such that: Lji, Ln, A^, and An are 250, 150, 1300, and 165 g cm*-, respectively (see text for explanation); p = 2.32 g cm-3; PA(Xoyo)|i. P.Ji(Xoyo)n- PB(Xoyo)(i. and PB(Xoyo)n are 0.95, 21, 10.3, and 1 atoms g-'(mineral) yr', respectively. A^, An, and p are lithology dependent variables. 61
implies that the two-isotope method can potentially give more constrained paleoelevation
estimates than the method of Libarkin, et al. (1999).
3.2. Two-isotope method under steady-state erosion
The two-isotope paleoelevation technique can be modified to incorporate steady-
state erosion conditions during the paleoexposure. If a surface undergoes erosion, then the concentration, C(x), of a stable isotope per gram of material measured at a depth, x, below
that surface will be (modified from Lai, 1991 and Brown et al, 1995):
(7)
where e is the erosion rate in cm yr'. If erosion occurs for a long time, such that t» A • (pe)"', then the second term in equation (7) will go to zero:
(8) pe
After this exposure time, the isotope will reach secular equilibrium, such that the profile of isotope concentrations with depth is unchanging with time. The minimum required exposure time, then, is inversely proportional to the rate of erosion which affected the studied surface during its initial exposure. Rapid erosion requires a shorter exposure time than slow erosion (Fig. 2). Using the relationship (8), equation (5) can be modified such that: fl, •/(>.") (9) D^.f(y,n) 62
Muons leutrons;
H
I I X W I 103 •• E c
ia-» 10-1
Erosion rate (cm/yr) Figure 2. Minimum exposure time required for 2iNe:38Ar to reach secular equilibrium for various steady-state erosion rates. 2'Ne;38Ar will reach equilibrium after time, t, such that t» A • (pe)-'. Exposure times shown here are for t = A • (pE)-i. Variables are set such that: A^, and An are 1300, and 165 g cm*-, respectively, and p = 2.32 g cm-3. A^, An, and p are lithology dependent variables. 63
- 4 km
0.1 i 50 100 150 200
Depth (cm)
Figure 3. Ratio of cosmogenic 2iNe (in quartz) to cosmogenic 38Ar (in sanidine) for elevations of 1-6 km and depths within the earth of 0-200 cm for a paleosurface which experiences steady-state erosion during its paleoexposure. ^iNei^^Ar has reached secular equilibrium. Time required to reach secular equilibrium is a function of erosion rate (Fig. 2). Variables in eq. (6) are set such that: L^., Ln, A^, and An are 250, 150,1300, and 165 g cm-2, respectively (see text for explanation); p = 232 g cm-^; PA(Xoyo)ji. PA(Xoyo)a. PB(^yo)ny and PB(Xoyo)n are 0.95,21, 10.3, and 1 atoms g-^mineral) yr', respectively. A^, An, and p are lithology dependent variables. 64
The addition of steady-state erosion results in a decrease in the Ca^Cb predicted for
a specific elevation (Fig. 1; Fig. 3). The relative difference between Ca'Cb at high
elevations and at sea level is much smaller when erosion effects are added to the
paleoelevation equation. This suggests that although erosion effects can be accounted for,
paleoelevation calculations will be most accurate for surfaces which have experienced no
erosion.
4. Production Rates
The two-isotope method requires the measurement of two isotopes with very
different modes of production. To increase the sensitivity of the paleoelevation calculation,
isotope A should be predominantly neutron produced; isotope B, predominantly muon
produced. The relationship between elevation and CaICb will be most pronounced if
isotopes A and B have these dominant modes of production. If PA(Xoyo)n » PA(Xoyo)n
and PB(Xoyo)n » PB(Xoyo)n- then (5) can be approximated as:
£i,M4 <.o,
In order for this paleoelevation technique to be applicable in a wide range of geological settings and times, it is necessary to choose A and B such that they 1) have target elements common to many rocks; 2) are long-lived; and 3) are themselves rare.
Stable, cosmogenically-produced noble gases have been extensively studied as a
potential source of geological information (Poreda and Ceding, 1992; Graf et al., 1991;
Kurz et al., 1990; Cerling et al., 1994). One such isotope, ^^Ne, has already been used as a paleoelevational indicator (Libarkin et al., 1999). ^iNe is produced in quartz, almost exclusively by 28si(n,2a)2'Ne. Neon is rare, with only 18.18 ppm neon in air. ^iNe 65 makes up 0.27% of all neon, yielding 49 ppb ^iNe in air (De Bievre and Barnes, 1985).
2'Ne meets the required conditions for a useful paleoelevation isotope and is used in the rest of this paper as isotope A; however, other stable or long-lived isotopes, such as ^He, which are also predominantly neutron produced exist and could be substituted. ^He is a useful cosmogenic isotope to study in quartz-poor environments, such as in basalt flows.
Only a few studies have been done regarding muon-production of cosmogenic isotopes (Takagi et al., 1974; Kirsten and Hampel, 1977; Florkowski et al., 1988; Brown et al., 1995; Stone et al., 1998). These center around radioisotopes, with little mention of muon production of stable isotopes. While some of these radioisotopes are long-lived (e.g.
'2^^I), their target isotopes are rare in most geological situations. As with neutron-produced isotopes, noble gases are the most likely isotopes which can be used as isotope B. Given the constraint of a geologically common parent isotope, 38Ar from 39K(^,n)38Ar is the most likely candidate for B, measured in a K-rich mineral such as sanidine. Additionally,
^^Ar is a rare isotope, making up 0.063% of all argon and 5.67 ppm in air (De Bievre and
Barnes, 1985). Finally, note that -'Ne and 3®Ar meet the conditions PA(Xoyo)n »
PA(Xoyo)n and PB(Xoyo)n » PB(Xoyo)n such that the ratio -'Ner^SAr will have a high degree of sensitivity to changes in elevation.
4.1. Production due to neutrons
The production rate of ^'Ne from 28si(n,2a)2'Ne in quartz has been calculated by
Niedermann and others (1994) as 21±3.6 atoms (g SiOi)'' a*' at sea level and high latitude. The production rate of ^SAr from K-isotopes is poorly constrained. K as a target has been ignored in the few cases where production from neutrons is discussed (Lai,
1991; Freundel et al, 1986). Fabryka-Martin (1988) calculates zero-probability for production for the reaction '*OK(n,t)38Ar. Finally, cross-section data for K to ^^Ar 66 reactions at cosmic ray neutron energies is unavailable. Here an estimate of 1 atoms g"' yr' is used as the 39.40,41^(11,np)38Ar production rate.
4.2. Production due to muons
Very little experimental data for isotope production firom muons exists. Data which do exist refer only to a small energy band and is not representative of the cosmic ray muon flux (e.g. Eckhause et al., 1965; Zinov et al., 1966). The theory of muon capture, however, is well developed (Fermi and Teller, 1947; Charalambus, 1971; von Egidy and
Hartmann, 1982). Until production rates of isotopes from muon capture are measured experimentally, production rates must be calculated using this theory (Charalambus, 1971;
Stone et al., 1998).
A negative muon slowing down in matter will either decay into an electron and energetic neutrinos, Dii, or be captured by a proton within the nucleus (Z,A). When the negative muon is captured, the proton will change into a neutron such that:
ji +{Z,A)^{Z-lA)* + v^ (11)
where the neutrino carries off most of the freed energy. The resulting nucleus, (Z-1,A)* is de-excited by the emission of neutrons, protons, or alpha-particles.
The production rate, P, in atoms g-' sec*', of an isotope (Z-l-Xp,A-(Xn+Xp)), where Xn.p = 0, 1, 2..., from muon capture in a compound can be written as:
P = I (/i) • f • f J • f • f • (12) ' c d n I 67
where I^(h) is the stopping rate of negative muons at depth, h, in muons g-^ sec'; fc is the
chemical compound effect which indicates the fraction of muons stopped by the target
element in the compound; fj is the fraction of muons captured from die Is muonic level by
the target nucleus; fi is the isotopic abundance of the target element; and fn is the probability
of production of the nuclide (Z-l-Xp,A-(Xn+Xp)) from de-excitation of the target nucleus
(Z-I,A)* by the emission of Xn neutrons and/or Xp protons. The most likely scenario for
de-excitation is the emission of one neutron, such that Xn=l and Xp=0 (Charalambus,
1971; Friedlander, etal., 1964; Fabryka-Martin, 1988).
Negative muon stopping rates, I^(h), have been described by physicists for a wide
range of energies and depths within the earth (see Stone et al., 1998 for a review). Stone
(1998) develops a method for calculating I^(h) for 20-5500 g cm-2 depth by fitting negative muon stopping rate data to a fifth order polynomial. These calculated stopping rates drop from 175 stopped muons g"' yr' near the surface to almost nothing at about 50 m depth
(the stopping rate is a function of density, so depths are usually reported in g cm-2, which is simply depth, in cm, multiplied by density, in g cm-^). We are interested in the ratio of a predominantly neutron-produced isotope to a muon-produced isotope. Since neutron flux is negligible (6% of surface values for a rock density of 2.32 g cm-^) below 2 meters depth, samples taken more than 2 m below the surface will yield mostly the muon- produced isotope. Since negative muon stopping rates are nesirly constant for this interval, an average stopping rate of 175 stopped muons g"' yr' is adopted (Stone et al., 1998).
This value is very similar to the value of 177 calculated by Charalambus (1971), even though this calculation was based on an inaccurate depth-intensity curve (Bilokon et al,
1989). This indicates that the stopping rate of negative muons is invariant for very shallow depths (i.e. <2 m). However, if a paleosurface undergoes erosion during its exposure. 68
muon production below 2 m depth will become important. An exact I^(h) under erosive
conditions can be calculated using Stone et al. (1998) only if the erosion rate is known.
The chemical compound effect, fc, is not well-constrained experimentally and must
be calculated. As shown by Stone et al (1998), fc can have a wide range of values depending upon the theory used. Predictions for percent capture by Ca of muons stopped
in calcite range from 34-40%, while experimental data suggests values below 30%.
Research in binary compounds suggests that fc is periodic with Z (Zinov et al, 1966; von
Egidy and Hartmann, 1982) rather than following the simple formulation of Fermi and
Teller (1947). The validity of extending fc calculated from the theory for binary compounds to more complex compounds was addressed by von Egidy and Hartmann
(1982). Comparisons between theoretical values and measured fc in glasses and complex alloys shows agreement for most elements within 10-20%. Likewise, theoretical calculations from the Fermi-Teller Law also agree with von Egidy and Hartmann's formulation within this error range. Until a more precise calculation of fc is made for minerals such as sanidine, the Fermi-Teller Law will give fc values within 20%;
(13)
where n is the target element; a is the atomic concentration in the molecule; and Z is the number of protons in the element. For K in sanidine (KAlSisOg) and Si in quartz (SiOa), fc = 0.138 and 0.467, respectively.
fn, the probability of producing a specific nuclide (Z-l-Xp,A-(Xn+Xp)), is also poorly consu-ained. In essence, fn is the probability that the excited nucleus, (Z-1,A)*, resulting from muon capture will emit Xn neutrons, n, and Xp protons, p, (sometimes in the form of an alpha particle, a). The emission of p or a is extremely unlikely since the 69
electrostatic potential (or Coulomb) barrier prohibits both the entry and exit of positively-
charged particles (see Fabryka-Martin, 1988 for a review). Rarely, protons or a can
overcome the repulsive Coulomb barrier due to a small permeability in the barrier
(Friedlander, et al., 1964). fn will be 0.01-0.05 for most product nuclei with Xp> 0
(Fabryka-Martin, 1988). Muon-production of ^'Ne in quartz can only result from the
unlikely 28si(|j.,3p4n)2iNe reaction, fn (^'Ne from 28si) is taken as 0.02 from
Charalambus (1971); this is probably an upper limit. The production of in sanidine
will result from 39K(fj,,n)38Ar. The emission of one neutron after muon capture is the
most likely occurrence; fn (^®Ar from ^^K) has been calculated by Charalambus (1971) as
0.558.
fd and fj are well-constrained, fd, the fraction of stopped muons captured by the
nucleus, falls in the range 0-1; fd is close to zero for low Z and approaches 1 for high Z
(Charalambus, 1971; Weissenberg, 1967). Weissenberg (1967) calculates the total
probability, X, that a muon will disappear as the sum of the nuclear capture, X^, and decay,
Xd. probabilities, fd, the fraction of muons which are captured by nucleus of Z protons can
be calculated as Xc/X. Weissenberg (1967, Table 3.10) lists the values for and X; fd can
be calculated for any element from these values. For Si (Z = 14) and K (Z = 20), fd is
0.63 and 0.82, respectively. The final factor, fj, is simply the isotopic abundance of the target element. For and 28s i, fj is 0.93 and 0.92, respectively.
Putting each of these values into (12) yields muon induced production rates for
-'Ne and 3®Ar of about 0.95 and 10.3 atoms g-' yr', respectively. The muon-production rate for -'Ne is likely no more than this since proton emission is uncommon. For ^^Ar production by muons, fc is the limiting factor; modifying fc by ±20% yields 38ar production rates from 8.2 to 12.3 atoms g"' yr'. Although this is a large error, the average calculated values for -'Ne and 38Ar must suffice until muonic production rates can be more accurately determined. 70
4.3. Sources of excess ^'Ne and
2'Ne and 38Ar resulting from other sources must be subtracted from sample inventories. First, because this technique will be used on rocks which were exposed to cosmic radiation, buried, and then exhumed, it is important to account for contamination from modem cosmogenic production. The details for this procedure and recommended sampling methods are described in Libarkin et al (1999). The only other significant source of excess ^iNe is the production of nucleogenic ^iNe from a particles ['80(a,n)2iNe] produced by U and Th decay. This mechanism will be important in older rocks, so it is important to calibrate measured sample inventories by measuring ^iNe in shielded samples.
This component is then subtracted from sample inventories, yielding purely cosmogenic
2lNe (Libarkin et al, 1999).
^^Ar has many more potential sources of contamination. Production paths which may affect ^^Ar concentrations in sanidine or other K-rich minerals are: 1) Negative muon capture: •'0Ca(^,2nP+)38Ar; 2) Thermal or fast neutrons: '•-Ca(n,na)38Ar; '•'K(n,a,p-
)38Ar; ''0Ar(n,np,p-)38Ar; 37ci(n,p-)38Ar; and 3) Radioactive decay: 38K—> 38Ar + p-.
Negative muon capture by "^^Ca atoms in sanidine crystals will be negligible. The proportion of''^Ca in sanidine will be minimal and fc calculated from the Fermi-Teller Law will approach zero, giving a production rate well below 1 atoms g*' yr'. The other muon- production parameters for '^oCa are large (fn = 0.199; fd = 0.86; fj = 0.97), suggesting that comparing 38Ar concentrations in calcium-rich minerals to -'Ne in quartz will also yield paleoelevation information. However, Lai (1991) uses cross-section data to approximate a neutron production rate for ^^Ar from '^^Ca of 24 atoms g-^ yr' at sea level. This rate violates the condition PB(Xoyo)n » PB(Xoyo)n; the ^iNer^SAr relationship will not be as sensitive to changes in elevation if Ar is produced from ^^Ca rather than 71
Production of ^SAr and 38k from neutrons is not well documented. and have very low isotopic abundances of 0.65% and 6.7%, respectively. Production from these isotopes should be low in sanidine. However, "^^Ar and ^^ci are isotopically significant, with abundances of 99.6% and 24.2%, respectively. ^''Cl will not likely be an important species in sanidine, but it may prove significant in other K or Ca rich minerals.
'^OAr will likely be a significant isotope, especially in sanidine. As such, the production of
38Ar from ^^^Ar from thermal neutrons should be investigated. Finally, 38k is a short-lived radioactive isotope of K which will be produced from spallation reactions with heavier nuclei. 38k decays into 38Ar; however, because 38k is so short-lived, relatively little data exists regarding production cross-sections or rates. Each of these sources of 38Ar may potentially be significant, although the lack of existing data prohibits incorporation of these sources here.
5. Conclusions
The two-isotope paleoelevation technique represents an improvement of the paleoaltimetric technique developed by Libarkin et al (1999). The applicability of the two- isotope technique to surfaces which have undergone extensive erosion is illustrated by comparison to Libarkin et al (1999). They measured -'Ne concentrations in the 27 Ma
Fish Canyon Tuff of Colorado in an attempt to calculate a paleoelevation for the Colorado
Rockies. This attempt was unsuccessful, primarily due to erosion which effected the surface of the Fish Canyon Tuff during its initial exposure. Because erosion can be accounted for by the two-isotope method, measurement of 38Ar concentrations in sanidines from the Fish Canyon Tuff should yield 2iNe:38Ar values which can be correlated to paleoelevation. For instance, the maximum ^'Ne concentrations measured by Libarkin et al
(1999) were 30.08±2.88 x 10^ atoms (g Si02)"'. Comparing this to predicted sea 72 level 2iNe:38Ar in the case of steady-state erosion yields an expected concentration of
9.4-11.4 38Ar atoms (g KAlSisOg)*'. At 2.5 km, the paleoelevation calculated for the 35
Ma Florissant Lake Beds (Gregory and Chase, 1992), the predicted ^Sat concentration is
16.6-20.1 38At atoms (g KAlSiaOg)"'. With the caveat that these predicted ranges are dependent upon an uncertain 38Ar muon-induced production rate of 10.3 (±20%) atoms g'' yr', the predicted Fish Canyon 38Ar concentrations for sea level and 2.5 km are unique.
This technique, then, has great potential for paleoelevation determinations even in areas which have experienced large scale erosion. Finally, because production rate attenuates with latitude as well as altitude, this technique may potentially yield paleolatitudinal information.
Acknowledgments
I would like to thank Clem Chase, Jay Quade, and Nat Lifton for their insightful comments over the course of this project as well as suggestions during the writing phase.
Nat Lifton and Bill Phillips were willing participants in numerous discussions about cosmogenic isotopes. Jane Poths and Bob Reedy at Los Alamos National Labs also offered helpful suggestions. This research was part of a Ph.D. thesis completed by Julie
Libarkin at the University of Arizona. 73
REFERENCES
Bilokon, H., Castagnoli, G.C., Castellina, S., Piazzoli, B.D., Mannocchi, G., Meroni, E., Picchi, P., Vemetto, S., 1989, Flux of vertical negative muons stopping at depths 0.35-1000 hg/cm^: Journal of Geophysical Research, v. 94, p. 12,145- 12,152.
Brown, E.T., Edmond, J.M., Raisbeck, G.M., Yiou, F., Kurz, M.D., and Brook, E.J., 1991, Examination of surface exposure ages of Antarctic moraines using in situ produced '^Be and 26a1: Geochimicaet Cosmochimica Acta, v. 55, p. 2269-2283.
Brown, E.T., Bourles, D.L., Colin, F., Raisbeck, G.M., Yiou, F., and Desgarceaux, S., 1995, Evidence for muon-induced production of '"Be in near-surface rocks from the Congo: Geophysical Research Letters, v. 22, p. 703-706.
Ceding, T.E., Poreda, R.J., and Rathbum, S.L., 1994, Cosmogenic ^He and ^iNe age of the Big Lost River flood. Snake River Plain, Idaho: Geology, v. 22, p. 227-230.
Charalambus, S., 1971, Nuclear transmutation by negative stopped muons and the activity induced by the cosmic-ray muons: Nuclear Physics, v. A166, p. 145-161.
De Bievre, P. and Barnes, L.L., 1985, International Journal of Mass Spectrometry and Ion Processes, v. 65, p. 211-230.
Eckhause, M., Siegel, R.T., and Welsh, R.E., 1966, Muon capture rates in complex nuclei: Nuclear Physics, v. 81, p. 575-584.
Fabry ka-Martin, J.T., 1988, Production of radionuclides in the earth and their hydrogeologic significance, with emphasis on chlorine-36 and iodine-129 [Ph.D. thesis]: Tucson, University of Arizona, 400 p.
Fermi, E. and Teller, E., 1947, The capture of negative mesotrons in matter: Physical Review, v. 72, p. 399-408.
Florkowski, T., Morawska, L., and Rozanski, K., 1988, Natural production of radionuclides in geological formations: Nuclear Geophysics, v. 2, p. 1-14.
Forest, C.E., Molnar, P., and Emanuel, K.A., 1995, Palaeoaltimetry from energy conservation principles: Nature, v. 374, p. 347-349. 74
Freundel, M., Schultz, L., and Reedy, R.C., 1986, Terrestrial siJCr-Kr ages of Antarctic meteorites: Geochimica et Cosmochimica Acta, v. 50, p. 2663-2673.
Friedlander, G., Kennedy, J.W., and Miller, J.M., 1964, Nuclear and Radiochemistry (2nd ed.): New York, John Wiley and Sons, 585 p.
Graf, T., Kohl, C.P., Marti, K., and Nishiizumi, K., 1991, Cosmic-ray produced neon in Antarctic rocks: Geophysical Research Letters, v. 18, p. 203-206.
Gregory, K.M. and Chase, C.G., 1992, Tectonic significance of paleobotanically estimated climate and altitude of the late Eocene erosion surface, Colorado: Geology, v. 20, p. 581-585.
Kirsten, T. and Hampel, W., 1977, Weak radioactivities induced by cosmic ray muons in terrestrial minerals: in Proceedings of the International Conference on Low- Radioactivity Measurements and Applications: The High Tatras, Czechoslovakia, p. 427-435.
Kurz, M.D., Colodner, D., Trull, T.W., Moore, R.B., and O'Brien, K., 1990, Cosmic ray exposure dating with in situ produced cosmogenic 3He: results from young Hawaiian lava flows: Earth and Planetary Science Letters, v. 97, p. 177-189.
Lai, D., 1991, Cosmic ray labeling of erosion surfaces: in situ nuclide production rates and erosion models: Earth Planet. Sci. Lett., v. 104, p. 424-439.
Libarkin, J.C., Chase, C.G., Quade, J., Poths, J., and Mcintosh, W., 1999, Cosmogenic 2lNe in quartz as a potential paleoaltimeter: Case study from the 27 Ma Fish Canyon Tuff, Colorado: in this dissertation. Appendix A.
Meyer, H.W., 1992, Lapse rates and other variables applied to estimating paleoaltitudes from fossil floras: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 99, p. 71-99.
Niedermann, S., Graf, T., and Marti, K., 1993, Mass spectrometric identification of cosmic-ray produced neon in tenestrial rocks with multiple neon components: Earth and Planetary Science Letters, v. 118, p. 65-73.
Poreda, R.J. and Cerling, T.E., 1992, Cosmogenic neon in recent lavas from the western United States: Geophysical Research Letters, v. 19, p. 1863-1866. 75
Sahagian, D.L. and Maus, J.E., 1994, Basalt vesicularity as a measure of atmospheric pressure and paleoelevation: Nature, v.372, p.449-451.
Stone, Evans, J.M., Fifield, L.K., Allan, G.L., and Cresswell, R.G., Cosmogenic chiorine-36 production in calcite by muons: Geochimica et Cosmochimica Acta, V. 63, p. 433-454.
Takagi, J., Hampel, W., and Kirsten, T., 1974, Cosmic-ray muon-induced in tellurium ores: Earth and Planetary Science Letters, v. 24, p. 141-150. von Egidy, T. and Hartmann, F.J., 1982, Average mucnic Coulomb capture probabilities for 65 elements: Physical Review A, v. 26, p. 2355-2360.
Weissenberg, A.O., 1967, Muons: New York, John Wiley and Sons, p. 158-169.
Zinov, V.G., Konin, A.D., and Mukhin, A.I., 1966, Atomic capture of negative muons in chemical compounds: Soviet Journal of Nuclear Physics, v. 2, p. 613-618. 76
APPENDIX C
TERTIARY REMAGNETIZATION OF PALEOZOIC ROCKS FROM THE
EASTERN CORDILLERA AND SUB-ANDEAN BELT OF BOLIVIA
J.C. Libarkin, R.F. Butler, D.R. Richards, T. Sempere, 1998
Published in: Libarkin, J.C., Butler, R.F., Richards, D.R., and Sempere, T., Journal of
Geophysical Research, v. 103, p. 30,417-30,429,1998. Copyright by the American
Geophysical Union. 77
JOURNAL OF GEOPHYSICAL RESEARCH. VOL. 103. NO. Bli. PACES 30.-l|7-30..i:9. DECE.MBER 10. IW8
Tertiary remagnetization of Paleozoic rocks from the Eastern Cordillera and sub-Andean Belt of Bolivia
J. C. Libarkin. R. F. Butler, and D. R. Richards Dcpanmcnt of CeoKicnccs. University of Ariioiu, Tucson
T. Sempere Misidn Orttom en el Peni. Linu
Abstract. Paleomagneiic samples were collected from 9S sedimentary horizons in eight different Devonian to Permian sedimeniar>' units at eight localities in the Eastern Cordillera and (he sub- Andean Belt of Bolivia. For 77 sites, thermal demagnetization allowed determination of a charac* teristic magnetization (ChRM) with site*mean 95rr confidence limit. 095, £ 15^. The ChRM is carried predominanil)' or entirely by hematite. Fold and reversal tests from two of (he sampled localities indicate that the characteristic magnetization is synfolding. likely acquired during the ear* liest stages of deformation. Additionally, a modified conglomerate test at one locality and the nearly uniform direction of ChRM across the Devonian to Permian age units clearly reveals the secondary nature of the characteristic magnetization. Finally, the ChRM directions are discordant from any expected Paleozoic directions. Paleomagnetic poles calculated from the ChRM directions fall near the Cenozoic portion of the apparent polar wander path for South America. We interpret these obser>-ations to indicate widespread chemical remagnetization of these Paleozoic strata during. but prior to completion of. Cenozoic Andean folding.
1. Introduction of these rocks can be documented. Less thorough paleomag netic study could lead to the erroneous conclusion that a pri> Paleozoic paleogcotfraphers jenerally accept that i.'ie south* ern continents (South Atnerica. Africa. India. Amarciica. and mary Paleozoic NR.M has been retained by these strata in the Australia) were united as the supercontinent Condwana between Eastern Cordillera and sub*Andean Belt of Bolivia. Late Cambrian and Late Triassic times. Howe\er. two contnst* in$ inierpreiationi of the Cond^ana apparent polar wander 2. Tectonic and Geologic Setting (APW) path have yet to te rwjolved [Van der Voo. 1938: Sedimentation in Bolivia was almost exclusively r.arine Briiltn, 19921. Because palejma :netic data from A;:stralia from the Late Cambrian through the Permian, and predomi* [HurUy c.nd V*«/i der Voo. 19371 ar.J Africa et c/.. nantly continental in the Mesozoic anJ Cenozoic 19S7] dominate the discuislon. we 3ttemp:s»l to determine 19951. Ettenj/onal deformation of western South[Stmptrt. A.-nerica Palcozoiw* pa[eom.i):netic poles for GonJwana through pateo* began in the early Mesozoic. Relatively little compressional magnetic study of South Ar.erica. Sampler were taksn from tectonic activity occurred before Andean*related shortening Paleozoic red sedimentary strata across the Eastern Cordillera began along the Pacific margin -S9 .Ma. Oeginning in itt Late a.rd sub-Aniiean Belt of Bolivia. .Acquisition of a larii paleo* Cretaceous, a varying compressional regime along this margin majnvtic data set from this region, however, forced us to the resulted in the development of a number of foreland basins, as ccr.wlusion t.'iat the sampled strata have teen affected by is evidenced by typical foreland successions in the A.ndes. wiJcsprtfad chemical remajr.etizaticr. probably durin: or just Overall, the development of an external foreland basin con* preceding Cenozoic foldir;. i.-olled by nearby easterly propagating deformation, beginning Thtf results of this stuJ> are of sisnitlcar:,-.* to pileoma^* as early as the Paleogene. is a major feature of the Cs:^ozoic ni'.jc analyses unJ^rtaken elsewheri- NVe report heri well* evolution of the Andes 1997J. This deforma detirminrJ site^mcan cha.-3c:en$!ic d;:'.'Ction» cf natural rema* tion continues today, with[Sonptrt an activ« et at.. foreland basin existing r.jnt rracnetism (.VRM; frcn some Paleozoic red sedrmentary east of t.^e Andes. rcc<» -Ahich haNC teen tn:erprcte.J pr
American Geophysical Union 2000 Florida Avenue. NW Washington, DC 2CC09 Tel: 11-202-462-6000 Fai: •1.?0^-.128-05i56
August 3, 1999
Julie C. Libarkin Departmenl of Geoscienccs Uiiiver.";iiy of Arizona Tucson, AZ 85721
Dear Ms. Libarkin;
Wc arc pleased to grniu pennission for the use of the muierial requested for inclusion in your thesis, including microfilm edition.s thereof. Permission is rcstrictcd to the use stipulated. The original publication must be appropriately citcd. The credit line should read; "author.':, journal or book title, VOUTOC number, p.ige numbers, year," aiid the phrase "Copyright by the American Geophysical Union." Substitute the lost phrase with "Puhlixhcd hy the American Geophysical Union" if the paper is not subject to U.S. copyright -- see tlie copyright line on the first page of the published paper for such cla.-isification.
Please feel free to contact ms again if you need further assistance. Thank you.
Sincerely,
Piimelii Ciilliham Publications Adniinistruiion Coordinator 79
Abstract. Paleomagnetic samples were collected from 98 sedimentary horizons in eight
different Devonian to Permian sedimentary units at eight localities in the Eastern Cordillera
and the sub-Andean Belt of Bolivia. For 77 sites, thermal demagnetization allowed
determination of a characteristic magnetization (ChRM) with site-mean 95% confidence
limit, 095, < 15°. The ChRM is carried predominantly or entirely by hematite. Fold and
reversal tests from two of the sampled localities indicate that the characteristic magnetization
is synfolding, likely acquired during the earliest stages of deformation. Additionally, a
modified conglomerate test at one locality and the nearly uniform direction of ChRM across
the Devonian to Permian age units clearly reveals the secondary nature of the characteristic
magnetization. Finally, the ChRM directions are discordant from any expected Paleozoic directions. Paleomagnetic poles calculated from the ChRM directions fall near the
Cenozoic portion of the apparent polar wander path for South America. We interpret these observations to indicate widespread chemical remagnetization of these Paleozoic strata during, but prior to completion of, Cenozoic Andean folding.
1. Introduction
Paleozoic paleogeographers generally accept that the southern continents (South
America, Africa, India, Antarctica, and Australia) were united as the supercontinent
Gondwana between Late Cambrian and Late Triassic times. However, two contrasting interpretations of the Gondwana apparent polar wander (APW) path have yet to be resolved
[Van der Voo, 1988; Briden, 1992]. Because paleomagnetic data from Australia [Hurley and Van der Voo, 1987] and Africa [Hargraves et ai, 1987] dominate the discussion, we attempted to determine Paleozoic paleomagnetic poles for Gondwana through paleo magnetic study of South America. Samples were taken from Paleozoic red sedimentary strata across the Eastern Cordillera and sub-Andean Belt of Bolivia. Acquisition of a large so
paleomagnetic data set from this region, however, forced us to the conclusion that the
sampled strata have been affected by widespread chemical remagnetization probably durin§
or just preceding Cenozoic folding.
The results of this study are of significance to paleomagnetic analyses undertaken elsewhere. We report here well-determined site-mean characteristic directions of natural
remanent magnetism (NRM) from some Paleozoic red sedimentary rocks which have beeri
interpreted previously to possess a primary magnetization [Ernesto etai, 1988]. Through comparison of paleomagnetic data from widely distributed collecting locations, application of a modified conglomerate test, and observation of consistent characteristic directions throughout a Devonian through Permian age section, the remagnetization of these rocks can be documented. Less thorough paleomagnetic study could lead to the erroneous conclusion that a primary Paleozoic NRM has been retained by these strata in the Eastern Cordillera and sub-Andean Belt of Bolivia.
2. Tectonic and Geologic Setting
Sedimentation in Bolivia was almost exclusively marine from the Late Cambrian through the Permian, and predominantly continental in the Mesozoic and Cenozoic
[Sempere, 1995]. Extensional deformation of western South America began in the early
Mesozoic. Relatively little compressional tectonic activity occurred before Andean-related shortening began along the Pacific margin -89 Ma. Beginning in the Late Cretaceous, a varying compressional regime along this margin resulted in the development of a number of foreland basins, as is evidenced by typical foreland successions in the Andes. Overall, the development of an external foreland basin controlled by nearby easterly propagating deformation, beginning as early as the Paleogene, is a major feature of the Cenozoic evolution of the Andes [Sempere et ai, 1997]. This deformation continues today, with an active foreland basin existing east of the Andes. 81
The formation of the Bolivian orocline has been an active focus of research for the past decade and modifies the record of simple compressional deformation in South America
[e.g. Roperch and Carlier, 1992; Butler et ai, 1995; Randall et ai, 1996]. Crustal scale block rotations, counterclockwise to the north and clockwise to the south, have been documented in Bolivia, Chile, and Peru [Randall et ai, 1996]. Rotations documented through paleomagnetic study of Upper Cretaceous to Tertiary strata have been observed in the region (Figure 1). Documented rotations within the study area range in magnitude from
5° to 65° [Butler et al., 1995].
Eight stratigraphic units ranging in age from Early Devonian to Early Triassic were sampled for this project. Precise site localities and unit ages are listed in Table 1. Although the sampled strata were primarily red beds, the depositional histories of these units are distinct, with depositional environments ranging from fluvial to glacio-marine. Site locations were selected on the basis of grain size (fine-grained units are preferable for paleomagnetic analysis) and red to purple pigmentation. A general description of the depositional and lithologic nature of sampled units is provided here. Descriptions of
"local" lithologies refer to the nature of the formation at the sampling locality. Figure 2 illustrates formation names, ages, and lithologic correlations for the Eastern Cordillera and sub-Andean Belt.
2.1. Vilavila Formation
The Lower Devonian (Lochkovian) Vilavila and Santa Rosa formations were deposited in fluvial to shallow marine environments. The Vilavila Formation has been mapped over a much smaller area than the Santa Rosa, namely, west and northeast of
Oruro, and consists of mainly red, cross-bedded to massive, coarse to fine sandstones intercalated with red siltstones and mudstones. The Santa Rosa Formation consists of 82
70°W 200 km
Bolim
SWilwE
Paraguay
Argentina
62''W
Figure 1. Map showing site localities and locations for which vertical axis rotations have been documented in and near the study area. Stippled area. Eastern Cordillera; lined area, sub-Andean Belt. Black squares are site localities for this study. PO, Pongo; IH, Inca Huasi; SW, Samaipata West; SE, Samaipata East; AL, Alarache; NZ, Narvaez; AT, Abra Tapecua; VM, Villamontes. Refer to Table 1 for a list of the formations sampled at each of these localities. Gray circles are localities from which paleomagnetic data are available to constrain vertical axis rotations [Randall et al., 1996; Hartley et al., 1992; Butler et al., 1995]. Abbreviations are qh, Quebrada Honda; si, Salla; pag, Paciencia Group; psg, Purilactis Group; ta, Tiupampa; lu, Laguna Umayo; Ip, La Palca; qu, Quehua; ce, Cerdas. 83
TABLE 1; Sample Localities and Formation Names Location Lat. Long. Unit Name Unit Age Site Names (•s) (*W)
Villamontes 21.27 63.54 Itacua Fm D rroo 1-003
Villamontes 21.27 63.54 Tupambi Fm M TPOOl-017
Villamontes 21.27 63.54 San Teimo Fm upper M STO10-022
Villamontes 21.27 63.54 Taiguati Fm M TG023-025
Villamontes 21.27 63.54 Vitiacua Fm P-Tr VTO10-022
Abra Tapecua 21.43 63.93 Ipaguazu Fm Tr IZOG1-002
Narvaez 21.44 64.26 San Telmo Fm upper M STOO1-006
Alarache 22.24 64.60 Taieuati Fm M TGOO1-003
Alarache 22.24 64.60 Vitiacua Fm F-Tr VTOO1-004
Samaipata-East 18.78 63.82 Escarpment Fm M ECOO1-003
Samaipata-West 18.15 63.93 Taieuati Fm M TG004-009
Inca Huasi 19.83 63.73 Taiguati Fm M TG020-022
PongG 17.71 66.58 Vilavila Fm D WOOl-015
Lat. and Long, are the latitude and longitude of the sampling locality, respectively. Unit Age lists regional ages of sampled units where D = Devonian, P = Permian, Tr = Triassic, and M = Mississippian. 84
CHACO BASIN EASTERN (Sub-Andean Belt CORDILLERA chrondogy and lowlands)
Castelun Ravelo Tapecua
Ipaguazul Mdde
Chutani yitlacual
Cangapi
Kasinwvian
Moscovian
Bashkirian
Serpukhovian San Telniol EscarpnnentI
vsean fTaiguati Chorro ttacuami Tournaisian [Tupambi] Cumana tacua Famennian Coilpacucho Frasnian
bivetian Sicasica Eifelian LosMonos Emsian Huamamparrpa Bel^n Pragian Ida
l^ilavHa Lochkovian pnrasn Catavi Tarabuco Ludlow Kirusillas Weniock
Llandovery Llallagua Ashoill Cancanin Cancaniri
Figure 2. Stratigraphic names and correlation for the sub-Andean and Andean zones of Bolivia. 85 similar sedimentary facies but does not show red pigmentation. Slumped beds occurring in both the Vilavila and Santa Rosa, in the study area and elsewhere, are indicative of syndepositional tectonic activity [Sempere, 1995].
2.2. Mississippian Units
We sampled the Mississippian Itacua, Tupambi, Taiguati, Escarpment and San
Telmo formations within the central and southern sub-Andean Belt of Bolivia. These stratigraphic units were deposited in a glacio-marine environment while the region was in a period of tectonic instability [Sempere, 1995].
The uppermost Famennian to lower Mississippian Itacua Formation consists of predominantly coarse, moderately-well lithified red sandstones with a few interbeds of poorly lithified fine sandstone to mudstone. Regardless of grain size, sandstones from the
Itacua commonly have a red muddy matrix. The Itacua is approximately 50 m thick in the section sampled at Villamontes.
The lower Mississippian Tupambi Formation consists of coarse to fine sandstones, commonly conglomeratic, deposited in cross-bedded to massive channels and alternating with subordinate and thinner red fine sandstone to siltstone beds. The sandstones are porous and permeable and are known to form hydrocarbon reservoirs in the sub-Andean
Belt [Sempere, 1995]. At Villamontes, the Tupambi Formation is about 350 m thick. The surfaces of many sandstone beds in the Tupambi Formation contain abundant visible hematitic zones. These zones are probably due to alteration of this unit caused by fluids circulating through the sandstones.
The middle Mississippian Taiguati Formation consists of generally reddish sandy to muddy sediment derived predominately from underlying formations. At Villamontes, this 86 unit is approximately 100 m thick and appears mostly as a series of well-cemented bluish- gray mudstones.
The Escarpment and San Telmo formations are now considered to be Late
Mississippian [Sempere, 1995]. The Escarpment Formation consists of thickly bedded, commonly conglomeratic, massive to cross-bedded sandstones that were deposited as debris flows, grain flows, turbidites, and slumps. The San Telmo Formation consists of generally thinner and finer sandstones alternating with red siltstones to mudstones. Water escape structures are commonly observed in the sandstones of the lower and middle parts of the unit, which also frequently display turbidite facies. Sandstones in the upper part of the unit were deposited in shallower environments with fluvial facies present at
Villamontes. Because of the good porosity and permeability of these sandstones, both the
Escarpment and San Telmo formations serve as important hydrocarbon reservoirs in the sub-Andean Belt. At Villamontes, some beds of the San Telmo Formation contain hematitic spots similar to those observed in the Tupambi Formation.
2.3. Vitiacua Formation
The Upper Permian-Lower Triassic Vitiacua Formation was deposited in a mainly resuicted-marine basin. The limy mudstones of the lower Vitiacua Formation mark a major transgression which affected much of Gondwana [Sempere et ai, 1992]. The upper
Vitiacua Formation consists predominantly of cherty dolomitic carbonate. At Villamontes this unit is a very thinly bedded to laminated succession of dominantly pink siliceous carbonate, including gray, blue, and white cherts and interbedded with red fine-grained siliceous layers. Sampling was concentrated within the pink to red layers or encompassed entire red laminations within the lighter cherty beds. Red siliceous concretions present within the upper part of the section were also sampled. 87
2.4. Ipaguazii Formation
The Triassic Ipaguazu Formation is recognized mainly in the southern sub-Andean
Belt, where it can be observed up to 400 m thick in places. It consists of red-brown
mudstones, siltstones, and fine to medium-grained sandstones, with thick evaporites
locally. These facies were deposited in alluvial to playalake environments. The sampled strata occur in a relatively thick succession of alternating thickly bedded muddy sandstones and finely bedded to laminated mudstones.
3. Paleomagnetism
3.1. Procedures and Rock Magnetic Analyses
Paleomagnetic samples were obtained from Devonian and Late Paleozoic strata at eight localities within the Eastern Cordillera and sub-Andean regions of Bolivia (Figure 1 and Table 1). Using standard paleomagnetic coring methods, oriented samples were collected from a total of 98 sedimentary beds (= paleomagnetic site with generally eight samples per site) within eight geologic units. Samples were oriented using both magnetic compass and solar compass. No significant deviations were noted between these orientation methods. An additional 14 samples were taken from slumped blocks at one particular level of the Vilavila Formation. The preservation of recognizable bedding in these slump blocks allowed us to use these samples in a modified conglomerate test.
Following sample preparation, all specimens were stored and analyzed in a magnetically shielded room with background field intensity <200 nT. NRM was measured with a three-axis cryogenic magnetometer (2G model 755R). Thermal demagne tizations at 14 to 18 steps up to 680°C were accomplished using furnaces with magnetic field intensity less than 10 nT in the sample region. Results of thermal demagnetization 88
were analyzed by principal component analysis [Kirschvink, 1980], and site-mean directions were determined using the statistical methods of Fisher [1953].
Analyses of coercivity spectra of 11 representative samples were performed using the methods of Dunlop [1972] and Lowrie [1990]. Representative results of isothermal remanent magnetism (IRM) analyses are illustrated in Figure 3. The majority of the samples showed IRM acquisition and thermal demagnetization behaviors similar to those illustrated in Figure 3. For all coercivity fractions, IRM unblocking temperatures are concentrated between 600°C and 680°C, indicating that hematite is the dominant or exclusive ferromagnetic mineral. Ten thin section analyses confirm the presence of hematite in all of the sampled units.
3.2. Paleomagnetic Results
Results of progressive thermal demagnetization of NRM were observed to be highly variable between formations and between localities. The Tupambi and San Telmo formations at Villamontes typically exhibited univectorial behavior with unusually linear decay of vector end points to the origin of vector component diagrams (Figure 4a).
Maximum angular deviation (MAD) of line fits to these demagnetization data showed very small values often less than 1° [Kirschvink, 1980]. More typically, MAD values for line fits to the thermal demagnetization data were < 5° (Figure 4b). Rarely, thermal demagnetization yielded erratic behavior from which a characteristic magnetization (ChRM) could only be extracted by fitting lines to results from the highest three or four temperature steps (Figure 4c). These line fits resulted in larger MAD values, ranging from 10° to 15°.
Specimens exhibiting MAD greater than 15° were removed from fiirther analysis. Finally, specimens from the Vitiacua Formation at Villamontes invariably produced very erratic behavior upon thermal demagnetization. No ChRM directions could be confidently MACNimZING FIELD (mT) 0 100 200 900 400 SOO 600 700 TEMPERATURE (C)
Figure 3. Coercivity spectrum analyses for representative samples from the Alarache locality, (a, c) Acquisition of isothermal remanent magnetism (IRM) as a function of magnetizing field, (b, d) Thermal demagnetization of three coercivity fractions as a QQ function of temperature. Figures 3a and 3b are from the Taiguati Formation. Figures 3c and 3d are from the Vitiacua Formation. ^ Up.N Up.N 682 L678 Vertical Plane >NRM 1673 Horizontal Plane 395 300
1-450 673 ^98 530 402 654
617 597 1580 563 3QP 669 lAOO IX 10A/tn ^RM 391 1630 >535 402 r505 664 650 645 654 300 646' NKMi 395 1639 I ; 1 "'^97 4501 530 JOX 645 498 630 [53 -- U lo ' Aim 563'! 550 1492 600 I X IC' A/in 391 65? 664
Figure 4. Typical progressive thermal demagnetization results in stratigraphic coordinates, (a) Good result. ST020B from the San Telmo Formation, Villamontes locality, (b) Average result. VV052A from the Vilavila Formation, Pongo locality, (c) o Marginal result. WOO IF from the Vilavila Formation, Pongo locality. Open squares, vertical-plane projection; closed squares, horizontal-plane projection. 91 identified from this formation and, accordingly, sites from the Vitiacua Formation collected at Villamontes were dropped from further consideration.
Two criteria were used to reject sites with poorly determined site-mean ChRM directions. Sites with fewer than four single polarity specimens yielding ChRM directions with MAD < 15° were rejected; sites with site-mean 95% confidence limit, a95, greater than 15° were rejected. Seventy-four site-mean ChRM directions pass these criteria and are listed in Table 2. For all but seven of the accepted sites, six or more sample ChRM directions were available for calculating the site-mean direction. Examples of within-site clustering of ChRM directions and ChRM site-mean directions are shown in Figure 5.
Some sites yielded unusually well grouped ChRM directions with correspondingly small confidence limits (Figure 5a). Typical sites had site-mean ChRM directions with 095 between 5° and 10° (Figure 5b). Amongst the sites passing the selection criteria, the largest site-mean 095 was 12° (Figure 5c).
Of the 74 accepted site-mean ChRM directions, only 11 contain normal polarity directions; seven of these sites are from Pongo, two are from Samaipata West, and two are from Villamontes. Site TG008 from Samaipata West contains dual-polarity magnetization with normal and reversed polarity ChRM directions preserved in each of four samples
(Table 2). Mean directions of ChRM from the polarity subsets of samples from these two sites are sufficiently well determined to warrant treating each polarity subset as a site-mean
ChRM direction in subsequent analyses.
3.3. Field Tests
Systematic examinations of field tests for paleomagnetic stability were fundamental to documenting the secondary nature of the characteristic magnetizations. Particularly critical was application of the conglomerate and fold tests [Graham, 1949] using the 92
N N
Average result; ST018 from the San Telmo Formation, theVitiacua Formation,,Rfo Rfo Villamontes. 1=42.0% Bermejo. I = 4«.0°,46.0°, D = 182.0°,182.0° D = 152.0°, 095 = 5.6°, k = 144 E 095 = 2.0°, k = 744
Marginal result: VVOOl from the Vilavila Formation, Pongo. I = -33.0', D = 5.4°, 095 = 12.0°, k = 26
Figure 5. Equal-area projections of sample ChRM directions and site mean ChRM direction, and 095 in stratigraphic coordinates for three sites, (a) Good result. VTOOl from the Vitiacua Formation, Alarache locality, (b) Average result. ST018 from the San Telmo Formation, Villamontes locality, (c) Marginal result. VVOOl from the Vilavila Formation, Pongo locality. Circles, sample Chi^ directions; squares, site-mean ChRM directions. Solid symbols are in the lower hemisphere, while open symbols are in the upper hemisphere. TABLE 2. Site-mean Directions for Characteristic Magnetizations Site Strike Dip N J 0195 k c eograp hie Coorc inates 8090 unfoldin g 10-3 (•) I D VGP VGP I D VGP VGP (•) (•) Lat. CN) Long. CE) (•) (•) Lat. CN) Long. CE)
IZOOl 013 24 8 4.20 9.8 32.8 -6.5 163.4 -58.0 83.6 -8.3 243.2 -22.3 190.2
IZ002 013 24 7 8.70 14.6 18.0 12.7 149.9 -56.0 53.1 25.3 155.0 -63.9 46.8
ECOOl 311 10 8 3.30 96.3 1.3 68.8 185.9 -55.7 289.7 61.9 194.8 -62.7 272.2
EC002 306 12 7 8.10 40.0 3.2 46.6 157.2 -66.9 357.1 41.1 164.8 -75.1 4.9
EC003 227 12 6 3.80 99.0 1.4 43.1 187.5 -80.2 252.4 36.6 181.6 -87.8 252.9 wool 159 41 7 0.64 12.0 26.1 -9.3 19.2 84.9 24.9 -28.7 9.6 80.5 10.2
VV002 159 41 8 2.30 3.4 272.6 9.3 211.3 -72.8 221.7 34.0 202.3 -68.8 209.4
VV003 159 41 6 4.60 7.7 76.6 9.1 207.4 -76.8 218.9 32.2 198.2 -72.7 205.4
VV004 159 41 8 1.30 3.1 312.7 15.8 210.9 -75.7 242.1 39.7 198.4 -72.1 222.1
VV005 159 41 7 1.30 4.0 231.9 15.8 211.0 -75.5 241.9 39.8 198.6 -71.9 222.2
VV006 159 41 7 8.90 3.9 236.3 15.3 213.8 -72.3 238.7 40.4 202.1 -68.6 221.6
VV007 159 41 8 1.10 5.5 100.8 -11.9 18.1 87.4 45.7 -30.5 7.2 83.0 13.8
VV008 159 41 8 4.00 2.3 562 -16.8 19.6 85.8 111.7 -37.9 4.0 84.8 67.4
VV009 337 12 7 6.40 3.0 402.2 -43.3 5.0 76.8 37.9 -38.2 12.1 78.0 43.7
WO 10 332 18 8 4.50 3.1 317.9 -45.3 13.6 67.6 26.3 -34.9 22.1 69.0 31.0 Site Strike Dip N J 0195 k G eograp hie Coorc inates d09 9 unfoidin g 10-3 (•) I D VGP VGP I D VGP VGP (•) (•) Lat. CN) Long. CE) C) C) Lat CN) Long. CE) won 348 12 8 5.90 3.3 287.4 -41.9 8.2 73.5 36.2 -38.0 15.6 74.9 39.8 wo 12 337 12 8 4.60 2.2 639.4 -42.8 8.5 73.2 30.5 -36.4 14.9 75.7 35.9
W013 086 03 6 2.70 5.6 143.4 39.5 198.6 -70.5 226.9 41.7 199.4 -70.8 225.9
WO 14 086 03 7 2.30 9.2 44.1 47.7 188.4 -74.1 263.7 49.9 189.0 -74.6 262.9
TG004 010 70 5 2.70 7.3 110.3 -29.9 326.2 57.7 204.6 50.9 197.1 -69.5 250.4
TG005 010 70 4 3.00 2.8 1056.6 8.0 137.7 -46.3 39.7 46.7 161.9 -70.7 352.3
TG006 010 70 7 2.70 4.6 173.4 8.3 134.2 -43.1 37.7 49.4 158.7 -67.7 261.0
TG007 010 70 4 2.90 8.2 126.8 2.1 126.1 -34.4 37.7 49.9 143.0 -54.2 358.4
TG008A 358 75 4 12.0 3.3 779 10.3 136.6 -45.7 37.6 40.8 165.0 -75.0 3.2
TG008B 358 75 4 3.40 6.7 187.7 6.5 293.9 21.5 217.2 -45.8 306.5 40.2 185.4
TG009 358 75 5 2.20 6.2 155.3 11.2 298.9 25.2 221.7 -39.2 308.4 41.6 192.2
TGOOl 209 82 8 18.0 7.5 55.7 13.9 270.6 -2.1 212.1 61.3 225.2 -47.5 244.6
TG002 209 82 8 16.0 5.0 124.2 18.4 258.8 -13.8 210.1 52.2 210.1 -61.4 233.6
TG003 209 82 7 6.00 3.8 248.5 26.8 248.0 -25.4 211.0 44.3 194.6 -76.2 224.1 Site Strike Dip N J 095 k G(eograp hie Coorc inates S09 ') unfoldin g 10-3 (A/m) (•) 1 D VGP VGP I D VGP VGP (•) (•) Lat. CN) Long. CE) C) C) Lat CN) Long. CE)
VTOOI 198 83 8 7.10 2.0 745.3 15.9 243.6 -27.5 203.1 47.7 200.0 -70.8 229.5
VT002 195 83 7 1.60 4.1 218 21.3 239.6 -32.1 204.7 48.2 188.9 -79.4 248.3
VT003 198 83 8 6.20 2.3 592.3 13.2 239.6 -30.7 199.9 43.1 201.9 -69.8 217.8
VT004 196 81 8 1.40 4.1 187.6 13.1 243.5 -27.0 201.5 48.3 204.7 -66.7 228.2
STOOl 030 18 8 7.50 4.0 188.2 34.8 172.2 -82.3 41.5 42.7 182.0 -86.2 267.2
ST002 007 17 8 19.0 6.3 79.3 38.1 156.6 -68.2 21.5 43.9 167.4 -111 3.1
ST003 027 28 7 4.00 4.3 196.6 44.1 165.0 -75.6 5.1 55.7 188.5 -73.4 271.0
ST004 027 28 8 5.70 8.2 46.3 28.4 171.2 -79.5 61.3 39.5 184.6 -85.6 219.3
ST005 027 28 8 7.80 5.1 117.5 27.1 155.4 -65.6 38.5 42.3 165.4 -76.2 10.3
ST006 029 20 6 5.80 8.6 61.3 46.7 130.8 -45.1 6.4 62.2 136.5 -47.9 343.8
TG020 001 88 8 3.90 23.8 6.4 -7.7 114.7 -21.6 39.0 -54.1 313.8 46.8 176.4
TG02i 001 88 7 7.30 2.3 661.9 -8.8 125 -30.8 44.4 -46.1 323.8 56.1 186.3
TG022 001 88 7 3.90 3.6 285.3 -7.3 118.8 -25.5 40.6 -51.6 319.2 51.6 179.0
ITOOl 202 31 6 1.40 12.5 29.9 56.5 203.8 -64.0 249.1 48.0 147.0 -59.2 5.1
STOlO 220 46 6 32.0 11.6 34.5 60.6 240.2 -35.9 243.2 53.0 180.0 -11.1 296.5 Site Strike Dip N J <*95 k eograp Ilic Coorc inates 809 9 unfoldin g 10-3 (Aym) (•) I D VGP VGP I D VGP VGP (•) (•) Lat. CN) Long. CE) C) C) Lat. CN) Long. CE)
STOll 214 46 6 42.0 5.0 177.8 42.6 214.3 -58.4 219.4 33.0 185.3 -84.0 174.1
ST012 214 46 6 33.0 6.0 124.4 48.6 206.9 -64.4 230.9 33.5 175.9 -85.1 63.2
ST013 214 46 5 160 6.8 127.1 49.6 197.2 -72.1 240.7 29.8 169.7 -78.9 53.2
ST014 201 42 6 170 4.3 240.9 60.9 227.2 -45.5 245.4 57.8 166.1 -69.1 328.2
ST015 212 44 6 130 7.0 91.9 57.9 216.3 -54.4 243.8 45.7 171.3 -80.1 348.3
ST016 205 42 6 76.0 5.9 129.5 48.6 211.2 -60.8 229.4 41.6 176.6 -85.9 345.5
ST017 205 42 6 51.0 4.9 188.9 44.8 213.4 -59.1 222.7 40.1 181.6 -87.8 253.3
ST0I8 205 42 6 80.0 5.6 144.4 63.3 206.4 -58.0 259.9 48.6 157.8 -68.3 359.4
ST019 205 42 6 49.0 6.6 105.1 48.0 226.2 -47.7 226.8 48.8 186.4 -79.8 263.5
ST020 211 44 6 87.0 7.6 78.4 60.0 204.4 -61.5 255.6 42.4 163.3 -74.3 11.3
ST021 208 44 5 62.0 9.3 69.3 44.1 220.7 -52.6 221.7 41.3 186.8 -83.3 229.0
ST022 209 46 5 92.0 8.0 92.5 43.5 226.7 -47.1 221.3 43.1 190.2 -79.9 230.5
TPOOl 208 31 6 9.20 7.4 83.0 49.6 209.1 -62.4 231.8 44.1 182.6 -84.8 269.5
TP002 208 31 6 7.60 5.5 151.4 38.8 196.0 -75.1 211.7 30.1 179.7 -84.9 113.2 Site Strike Dip N J 0195 ll G(eograp hie Coorc inates 809h unfoldin g 10-3 (A/m) (•) I D VGP VGP I D VGP VGP (•) (•) Lat. CN) Long. CE) C) C) Lat. CN) Long. CE)
TP003 208 31 6 29.0 4.7 203.6 35.0 198.3 -72.8 203.2 27.6 183.6 -82.5 144.4
TP004 208 31 6 25.0 4.6 210.8 43.2 205.4 -66.4 220.8 37.4 184.5 -85.8 202.5
TP005 208 31 6 15.0 1.9 1283.5 46.1 202.5 -68.6 227.7 38.8 180.3 -89.3 272.7
TP006 208 31 6 7.50 5.1 174.9 46.6 204.3 -67.0 228.0 39.9 181.3 -88.4 256.3
ITOO? 208 31 6 4.30 1.5 1935.4 47.4 213.4 -58.9 226.8 44.0 187.6 -81.7 240.8
TP008 208 31 6 7.60 4.6 208.9 47.8 205.7 -65.6 229.8 41.4 181.5 -87.1 267.9
TP(X)9 208 31 6 2.50 7.5 80.4 46.0 206.8 -64.9 225.9 40.3 183.6 -86.2 234.3
TPOlO 208 31 6 26.0 2.1 1043.7 49.6 204.5 -66.2 234.1 42.4 179.2 -86.6 309.0
'rpon 202 31 6 13.0 8.1 70.0 57.6 207.8 -60.7 248.0 52.1 172.2 -76.6 326.0
TP012 202 31 6 4.20 7.3 84.4 39.6 202.0 -69.5 213.9 35.4 182.9 -86.8 174.8
TP013 202 31 6 6.80 10.9 39 -52.3 41.8 51.3 53.0 -53.6 8.0 75.3 89.4
TP014 202 31 6 1.60 11.0 38.1 -39.3 2.4 87.5 51.1 -27.8 347.5 76.5 233.1
TP015 202 31 6 15.0 10.1 45.3 41.9 187.1 -82.8 231.6 31.7 169.7 -79.4 47.6
1T0I6 202 31 6 10.0 3.3 402.1 53.6 205 -64.5 242.1 48.0 174.4 -80.7 328.4 Site Strike Dip N J 1*95 k G eograp hie Coorc inates 8090 unfolditi>g 10-3 (A/m) (•) I D VGP VGP I D VGP VGP (•) (•) Lat. CN) Long. CE) (•) (•) Lat. CN) Long. CE)
TG023 210 36 6 71.0 5.5 147.7 69.8 181.6 -57.6 294.7 48.0 147.0 -59.2 5.1
TP017 202 31 6 12.0 10.8 39.8 50.2 218.1 -54.6 230.4 50.5 187.3 -78.1 264.7
Site indicates site number; Strike and Dip were measured using the right-hand rule; N is the number of samples from the site used for determinations of site-mean directions; J is the geometric mean of the intensity of the characteristic component of the magnetization; 095 is the 95% confidence limit for the mean direction computed from Fisher (1953) statistics; k is the best estimate of Fisher's precision parameter; I and D are inclination and declination of the site-mean direction, respectively; Lat. and Long, are the latitude and longitude of the virtual geomagnetic pole calculated from the site-mean direction, respectively. Results are listed in both geographic and 80% unfolding coordinates.
00 99
statistical methods of Watson [1956], Watson and Irving [1957], and McFadden [1990].
Plunge corrections were not required for localities where field tests were applied because
folds in the study area extend tens to hundreds of kilometers along strike. Results of the
modified conglomerate test applied to the sedimentary layer containing laminated slump
blocks within the Early Devonian Vilavila Formation are shown in Figure 6. Fourteen
samples from slumped blocks containing recognizable intemal bedding yielded ChRM
directions. These ChRM directions are tightly grouped in geographic coordinates but
disperse upon restoring bedding within each slump block to horizontal. This dispersion of
ChRM directions is statistically significant at the 99.9% confidence level and constitutes a
negative conglomerate test [Watson, 1956]. Because we are unable to determine the extent
of vertical axis rotations of these slump blocks, an inclination-only test of dispersion is
quantitatively more meaningful than this modified conglomerate test. Accordingly, an
inclination-only test, as developed by McFadden and Reid [1982], was applied.
Application of bedding tilt corrections to the 14 slump block ChRM directions results in a
decline in inclination grouping, with a decrease in the estimate of Fisher's precision
parameter k from 36.0 in situ to 4.4 after tilt correction. Clearly, the ChRM was acquired
after slumping of a laminated and fairly lithified sediment and is therefore a secondary
magnetization. Additionally, the in sina directions observed in the slumped blocks are very similar to the normal polarity Vilavila Formation average (Figure 7). Finally, the ChRM directions in geographic coordinates are in directions expected for a Cenozoic magnetization and are unlike those expected for a Devonian magnetization.
Local folding of the Vilavila Formation at the Pongo locality allowed application of the fold test to results from this location. Results of progressive unfolding applied to site- mean ChRM directions are illustrated in Figure 7. A minimum fold test statistic (SCOS) and a maximum k were observed for 80% unfolding, with unfolding values from -65% to just less than 90% yielding statistically indistinguishable SCOS values [McFadden, 1990]. 100
N
w- Ggngraphic roordiniitfs I = •43.0," D = 11.1% a95 = 5.4°,k = 56
w Stratigraphir rnordinates I =-33.5°, D= 17.7°, a95 = 20.4Mt = 5
S
Figure 6. Conglomerate test results from the Vilavila Formation, Pongo locality. Equal- area projections show sample ChRM directions from slump blocks in (a) geographic coordinates and (b) stratigraphic coordinates. Solid squares are lower-hemisphere directions, while open squares are upper-hemisphere directions. 101
N
Geographic Coordinates I = 26^', D = 200.7', w a95 = 9.r,k = 20.1
s N
80% unfolding I = 37.4*, D = 195J*, 095 = 3.5*, k= 130
to T 13)
.. 100
k «0
40 oSCOS 1 —^95% ilmit
0 0 to » 40 SO 00 70 ao 90 too % UNFOLDING
Figure 7. Fold test results from the Vilavila Formation, Pongo locality. Equal-area projections of site-mean ChRM directions are shown in (a) geographic coordinates and (b) at 80% unfolding, (c) SCOS and k are the fold test statistic and the estimate of Fisher's precision parameter, respectively [McFadden, 1990]. Solid circles are normal polarity site means reflected through the origin. Solid squares are reverse polarity site means. Squares are location mean directions. Ovals of 95% confidence are plotted about the location means. 102
The SCOS value for 100% unfolding falls well above the 95% confidence limit, indicating that the ChRM directions were not acquired when the strata were horizontal or even uniformly tilted. Instead, the ChRM appears to be a synfolding magnetization as observed for many Paleozoic formations of eastem North America [Scotese et ai, 1983; McCabe and
Elmore, 1989]. Apparently the Vilavila Formation acquired its characteristic magnetization at -80% unfolding when -20% of the currently observed folding had occurred. These fold test results confirm the secondary nature of the ChRM and further suggest that the Vilavila
Formation acquired its characteristic magnetization early in the development of folding.
Because this folding regionally affects Paleogene strata, this magnetization is certainly of
Cenozoic age.
Analysis of progressive unfolding applied to site-mean ChRM directions from the
Villamontes and Narvaez localities are illustrated in Figure 8. Because of the proximity of these localities, a remagnetization of these strata would likely have been roughly coeval.
Minimum SCOS was observed at 80% unfolding, with unfolding values from -65% and
-95%, yielding statistically indistinguishable SCOS values [McFadden, 1990]. However, the SCOS value for 100% unfolding falls above the 95% confidence limit. These results clearly indicate a secondary origin for the ChRM and suggest that the Mississippian strata in the Villamontes and Narvaez area also acquired their characteristic magnetization during early stages of folding.
The ChRM directions at 80% unfolding from the Early Devonian Vilavila
Formation in the Eastem Cordillera at Pongo and the Mississippian strata at Villamontes and Narvaez are in close agreement (Figures 7 and 8). This observation suggests that the synfolding remagnetization is roughly similar in age across the Eastem Cordillera and southern sub-Andean Belt of Bolivia (Figure 1). As can be seen in Figure 1, the Villa montes and Pongo localities are the southeastern- and northwestern-most sampling localities, respectively. Although field tests cannot be applied to ChRM directions for the 103
N
G«oyraphlc Coordiimes w
s N
80% unfolding Mean value for all sites combined. I = 433', D = 176.8", B9S = 3J', k = 50.7
C
OSCOS k — -95% limit
% UNFOLDING
Figure 8. Fold test results for the San Telmo Formation, Viliamontes and Narvaez localities in (a) geographic coordinates and (b) at 80% unfolding, (c) SCOS and k are the fold test statistic and the estimate of Fisher's precision parameter, respectively [McFadden, 1990]. Site mean directions are shown by circles for Viliamontes and squares for Narvaez. Gray-filled symbols are normal polarity site means reflected through the origin. Solid symbols are reverse polarity site means. Gray symbols are formation means for both localities. Ovals of 95% confidence are plotted about the location means. 104 intervening localities, the relationship between deformation and remanence acquisition evidenced at the southeastern- and northwestern-most localities suggests that all or most of the magnetizations were acquired during the initial stages of deformation. Accordingly, all directions have been unfolded to 80% (Table 2). Because the magnetization is believed to be acquired contemporaneously across formations in any given locality, location means were calculated by averaging the directions from all of the sites within that locality. The locality mean directions were then used to calculate locality mean paleopoles (Table 3).
3.4. Previous Results
Two paleomagnetic studies have been conducted on some of the same rock units and in similar localities as studied in the current project. Creer [1970] measured an NRM from the Taiguati Formation near the town of Samaipata, which is located between our two
Samaipata sampling localities. This magnetization was interpreted as primary, although it is nearly concordant with Tertiary field directions. Ernesto et al. [1988] studied a series of
Carboniferous sections across the sub-Andean Belt, including one section near
Villamontes. They concluded that these rocks carry a primary Carboniferous magnetization on the basis of both agreement with the Creer [1970] pole and increased clustering upon tectonic correction. We applied the McFadden [1990] fold test to the data acquired by
Ernesto et al. [1988] and found no statistically significant changes in SCOS values at any degree of unfolding. Additionally, their "Carboniferous" pole falls close to the Tertiary
APW path for South America, supporting our belief that these data are representative of
Tertiary, not Carboniferous directions. TABLE 3: Locality Means in Geographic and 80% Unfolding Coordinates and Calculated Paleopoles }0% UIifolding Coordin ales Geog raphic Coordinal es Locality I D "95 k Lat. Long. A95 I D 095 k Lat. Long. A9S Sites C) (°) C) CN) CE) C) C) C) C) CN) (°E) n
Villamonles 41,0 177.5 5.0 26.98 -82.8 19.0 3.3 50.1 208.0 3.5 52.35 -62.7 234.5 4.1 32
Samaipala-Wcst 48.3 153.3 12.9 22.82 -70.6 324.0 17.0 5.9 130.2 13.1 22.03 -39.2 37.2 10.0 7
PonRo 37.4 195.3 3.5 129.96 -76.9 227.3 3.7 26.2 200.7 9.1 20.10 -70.6 197.6 6.8 14
Narvncz 48.8 172.6 12.2 31.26 -78.2 310.2 15.7 37.3 159.4 11.9 32.9 -70.4 22.3 12.7 6
Alarache 49.7 202.5 6.9 76.46 -82.3 265.4 6.9 17.8 249.1 9.1 44.56 -22.8 206.2 8.8 7
Abra Tapecua 11.7 201.6 -65.4 62.0 30.4 3.1 156.7 53.3 24.1 -57.9 67.9 37.0 2
Inca Huasf 50.7 139.2 7.9 244.9 -66.3 333.1 5.4 -8.0 119.5 8.0 241.45 -26.0 41.2 8.0 3
Samaipa(a-East 47.1 178.4 25.8 23.90 -80.1 296.1 27.1 53.7 175.5 27.2 21.62 -72.0 306.6 32.3 3
Sites indicates the total number of sites used in the calculations; a95 and A95 are the 95% confidence limit for the mean direction and the mean pole, respectively, computed from Fisher [1953] statistics; k is the best estimate of Fisher's precision parameter; 1 and D are inclination and declination of the site-mean direction, respectively; Lat. and Long, are the latitude and longitude of the virtual geomagnetic pole calculated from the site-mean direction, respectively. Results are listed in both geographic and 80% unfolding coordinates. 106
4. Discussion
The paleomagnetic analyses detailed above indicate that all Paleozoic strata sampled in this study have been remagnetized, probably during early stages of Andean deformation.
Our sampling localities cover large areas of the exposed mid-Paleozoic strata of Bolivia
(Figure 1), and we find no evidence that primary magnetizations have been retained by these rocks. Instead, detailed application of field tests for paleomagnetic stability document the secondary origin of the characteristic magnetization. In other regions, such as the
Appalachians of North America, where remagnetizations have been documented [McCabe and Elmore, 1989; Stamatakos et ai, 1996], analyses of paleomagnetic pole positions determined from remagnetized rocks have been used to determine an age of remagnetization. For several reasons, age assignment for the remagnetization of Paleozoic strata in Bolivia is possible with only limited precision.
Factors limiting the precision with which the characteristic directions can be used to infer age of remagnetization and tectonic disturbance subsequent to remagnetization include the following: (1) The major variation in bedding attitudes within the Pongo sampling area allows the percent of folding at remagnetization to be fairly tightly constrained to 20% fold ing (Figure 7). However, the major sampled stratigraphic sequence at Villamontes is nearly homoclinal and that section has only a modest contrast of bedding attitude with other sampling localities in that region. Accordingly, assignment of percent folding at the time of remagnetization of Paleozoic strata at Villamontes lacks precision (Figure 8). (2) The mag nitude of APW as viewed from South America is modest during Mesozoic and Cenozoic time [Irving and Irving, 1982; Besse and Courtillot, 1991]. Therefore only very low resolution is possible when inferring ages of magnetization by matching observed paleomagnetic poles with the South American APW path. (3) Cenozoic Andean deformation has involved tectonic vertical axis rotations which affected large regions of the 107
Andes. Paleomagnetic studies have begun to reveal the timing, magnitudes, and spatial
extent of these rotations [Roperch and Carlier, 1992; Butler et ai, 1995; Randall et ai,
1996]. However those patterns are not simple, and vertical axis rotations in some areas,
most notably the sub-Andean Belt, remain to be determined.
While accepting the restrictions and uncertainties outlined above, we believe basic
inferences about age and mode of remagnetization of Paleozoic strata in the Eastern
Cordillera and sub-Andean Bell of Bolivia can be drawn from our paleomagnetic results
when coupled with geologic observations. From two lines of evidence there seems little
doubt that the observed remagnetization is Cenozoic. First, the synfolding nature of the
remagnetization connects the timing of remagnetization with that of fold and thrust belt
development in the Andes. Major phases of fold and thrust bell development began during
late Oligocene in the Eastern Cordillera and continued into the Pliocene or Quaternary in the
sub-Andean Belt [Sempere et ai, 1990; Baby et ai, 1994]. Second, the paleomagnetic
pole positions determined from the characteristic paleomagnetic directions at 20% folding
generally agree with the Cenozoic APW path for South America. Figure 9 illustrates the
Cenozoic APW path for South America, as well as expected Late Devonian and Late
Carboniferous poles. Only two paleomagnetic poles determined from our data set,
Villamonles (VM) and Pongo (PG), have small enough confidence limits to rigorously
compare with the South American APW path (Figure 9). Both of these poles, as well as
the virtual geomagnetic poles (VGP) from other areas, fall well away from expected
Paleozoic poles (Table 3 and Figure 9). The paleomagnetic pole from Villamonles falls on
the 60-70 Ma portion of the APW path while the pole from Pongo is nearest the 10 Ma
portion of the path. The discordance of the Pongo pole suggests a small amount of rotation at this locality, but again we are reluctant to be more quantitative based upon this data set alone. 108
Figure 9. Hatched circles are 95% confidence limits (A95) for locality mean paleopoles from Villamontes (VM) and Pongo (PG) localities. Open circles are the A95 for the Late Carboniferous (LC [Irving and Irving, 1982]; -9.2°N, 21.0°E, A95 = 5.0°) and Late Devonian (LD [Hurley and Van der Voo, 1987]; -13.9°N, 339.8E, A95 = 8.0°) poles. Stippled circles are the A95 for the poles which make up the Late Cretaceous-Tertiary Apparent Polar Wander (APW) path [Butler et ai, 1995; Roperch and Carlier, 1992; Raposo and Ernesto, 1995]: 10 Ma, -86.9°N, 273.5°E, A95 = 3.0°; 30 Ma, -80.5°N, 287.3°E, A95 = 2.7°; 50 Ma, -78.9°N, 307.0°E, A95 = 4.3°; 60 Ma, -80.5°N, 340.7°E, A95 = 4.2°; 80 Ma, -78.9°N, 3.0°E, A95 = 3.1°; 130 Ma, -83.8°N, 60.3°E, A95 = 4.3°. Mid-Paleozoic paleopoles have been rotated from African into South American coordinates by counterclockwise rotation of 57° about an Euler pole at 44°N, 329.4°E [Irving and Irving, 1982]. 109
Along with the above arguments for a Cenozoic age of remagnetization of Paleozoic strata in Bolivia, it is also important to point out what we believe should not be inferred from our observations. A general age assignment of remagnetization to Cenozoic time is reasonably clear. However, attempts to resolve age differences in remagnetization between the sampled areas and perhaps infer details of fluid migration are beyond the resolution of the available data. Uncertainties in the percent unfolding at the time of remagnetization and possible vertical axis rotations simply prevent any meaningful inferences beyond our first- order conclusion that the remagnetization occurred during the Cenozoic. Given the sparse current knowledge of vertical axis rotations in the Andes coupled with structural and age uncertainties for the remagnetization directions reported here, it would be stepping well beyond the available data to use the observed paleomagnetic directions to infer magnitude or even sense of vertical axis rotations subsequent to remagnetization.
Drawing upon paleomagnetic, rock magnetic, structural geologic, and geochemical studies of the remagnetization which affected the Appalachian region of North America during the late Paleozoic time, we conclude with speculations on the process of remagnetization of the Paleozoic strata of Bolivia. In North America, fluids propagating from the fold and thrust belt of the Appalachian Orogeny remagnetized siliciclastic and carbonate rocks across the orogen and onto the craton [McCabe and Elmore, 1989].
Orogenic fluid migration, including water and hydrocarbons, in the Appalachians may have been driven by either a "squeegee" effect of the developing thrust belt [Oliver, 1986] or in response to a gravitationally driven hydraulic gradient [Garven and Freeze, 1984; Lawrence and Comford, 1995]. The synfolding remagnetization of Paleozoic strata over large areas of Bolivia during Andean deformation could have followed a similar sequence of pro cesses. All Paleozoic strata sampled in this study were at some time buried below the active Andean foreland basin as evidenced by geologic relationships and hydrocarbon accumulations present in some units. Progressive west to east propagation of the fold and 110
thrust front from a late Oligocene position in the Eastern Cordillera to its present location
within and east of the sub-Andean Belt drove fluid migration with attendant chemical
remagnetization. The suggestion from the paleomagnetic data that remagnetization occurred
when the beds were only 20% deformed is explained by a ChRM which remained stable
during subsequent fold and thrust deformation. This stability is likely if orogenic fluid composition did not change dramatically during subsequent deformation [Lawrence and
Comford, 1995].
5. Conclusions
The paleomagnetic data presented here indicate that Paleozoic units across the
Eastern Cordilleran and sub-Andean Belt of Bolivia have been remagnetized during
Cenozoic deformation. The lack of preservation of original magnetizations and the pervasive remagnetization documented here suggest that primary Paleozoic magnetizations are probably not preserved in sedimentary strata in this region. Future attempts to obtain primary paleomagnetic directions from Paleozoic rocks of the Andes should focus on less deformed regions, perhaps in the forebulge region of the active foreland basin. Finally, the documentation of this remagnetization reaffirms the critical importance of multiple field tests of paleomagnetic stability when applying paleomagnetic methods to rocks in orogenic zones.
Acknowledgments. We thank Brigette Martini, Scott Balay, and Rob Feyerharm for major technical assistance with laboratory measurements and Tom Moore for assistance with computer programs. Larry Marshall provided valuable field assistance during early stages of this project. Critical logistical assistance in Bolivia was provided by Ramiro
Suarez. Genaro Montemurro, Jaime Oiler, and Miguel Cirbian also assisted during our Ill time in Bolivia. The practiced eye of Rob Van der Voo first spotted the possibility of synfolding magnetizations in our data. We thank Rob for urging us to carefully evaluate the case for synfolding magnetization which turned out to be such fertile territory. Norm
Meader provided major assistance in preparation of the camera ready copy. This research was supported by National Science Foundation grant EAR 9316414 and National
Geographic Society grant 5743-96. 112
References Baby, P., B. Guillier, J. Oiler, E. Mendez, G. Montemurro, and D. Zubieta, Sintesis estructural del Subandino boliviano, in XI Congreso Geologico Boliviano, pp. 161-169, Santa Cruz de la Sierra, 1994.
Besse, J., and V. Courtillot, Revised and synthetic apparent polar wander paths of the African, Eurasian, North American and Indian plates, and true polar wander since 200 Ma, J. Geophys. Res., 96, 4029-4050, 1991.
Briden, J. C., Paleozoic a.p.w. paths and continental reassemblies: A case of faltering progress (abstract), Eos Trans. AGU, 75(14), Spring Meet. Suppl., 51, 1992.
Butler, R. F., D. R. Richards, T. Sempere, and L. G. Marshall, Paleomagnetic determinations of vertical-axis tectonic rotations from Late Cretaceous and Paleocene strata of Bolivia, Geology, 23, 799-802, 1995.
Creer, K. M., A palaeomagnetic survey of South American rock formations, Philos. Trans. R. Soc. London, Ser. A., 267, 457-558, 1970.
Dunlop, D. J., Magnetic mineralogy of unheated and heated red sediments by coercivity spectrum analysis, Geophys. J. R. Astron. Soc., 27, 37-55, 1972.
Ernesto, M., L. A. Diogo, and A. C. Rocha-Campos, Paleomagnetism of the Subandean Carboniferous ("Gondwana") sequence in Bolivia: Tectonic implications, in Gondwana Seven, edited by H. Ulbrich and A. C. Rocha-Campos, pp. 615-636, Sao Paulo, Brazil, 1988.
Fisher, R. A., Dispersion on a sphere, Proc. R.. Soc. London, Ser. A, 217, 295-305, 1953.
Garven, G., and R. A. Freeze, Theoretical analysis of the role of groundwater flow in the genesis of stratabound ore deposits. Am. J. Sci., 284, 1125-1174, 1984.
Graham, J. W., The stability and significance of magnetism in sedimentary rocks, J. Geophys. Res., 54, 131-167, 1949.
Hargraves, R. B., E. M. Dawson, and F. B. Van Houten, Palaeo-magnetism and age of mid-Palaeozoic ring complexes in Niger, West Africa, and tectonic implications, Geophys. J. R. Astron. Soc., 90, 705-729, 1987. 113
Hartley, A. J., E.J. Jolley, and P. Turner, Paleomagnetic evidence for rotation in the Precordillera of northern Chile: Structural constraints and implications for the evolution of the Andean forearc, Tectono-physics, 205, 49-64, 1992.
Hurley, N. F., and R. Van der Voo, Paleomagnetism of Upper Devonian reefal limestones, Canning basin, western Australia, GeoL Soc. Am. Bull., 98, 138-146, 1987.
Irving, E., and G. A. Irving, Apparent polar wander paths Carboniferous through Cenozoic and the assembly of Gondwana, Geophys. Surv., 5, 141-188, 1982.
Kirschvink, J. L., The least-squares line and plane and the analysis of palaeomagnetic data, Geophys. J. R. Astron. Soc., 62, 699-718, 1980.
Lawrence, S. R., and C. Comford, Basin geofluids. Basin Res., 7, 1-7, 1995.
Lowrie, W., Identification of ferromagnetic minerals in a rock by coercivity and unblocking temperature properties, Geophys. Res. Lett., 17, 159-162, 1990.
McCabe, C., and R. D. Elmore, The occurrence and origin of Late Paleozoic remagnetization in the sedimentary rocks of North America, Rev. Geophys., 27, 471^94, 1989.
McFadden, P. L., A new fold test for palaeomagnetic studies, Geophys. J. Int., 103, 163- 169, 1990.
McFadden, P. L., and A. B. Reid, Analysis of palaeomagnetic inclination data, Geophys. J. R. Astron. Soc., 69, 307-319, 1982.
Oliver, J., Fluids expelled tectonically from orogenic belts: Their role in hydrocarbon migration and other geologic phenomena. Geology, 14, 99-102, 1986.
Randall, D. E., G. K. Taylor, and J. Grocott, Major crustal rotations in the Andean margin: Paleomagnetic results from the Coastal Cordillera of northern Chile, J. Geophys. Res., 101(B1), 15,783-15,798, 1996.
Raposo, M., and M. Ernesto, An Early Cretaceous paleomagnetic pole from Ponta Grossa dikes (Brazil): Implications for the South American Mesozoic apparent polar wander path, J. Geophys. Res., i00(B10), 20,095-20,109, 1995. 114
Roperch, P., and G. Cariier, Paleomagnetism of Mesozoic rocks from the Central Andes of Southern Peru: Importance of rotations in the development of the Bolivian orocline, J. Geophys. Res., 97(B12), 17,233-17,249, 1992.
Scotese, C. R., R. Van der Voo, and C. McCabe, Paleomagnetism of the Upper Silurian and Lower Devonian carbonates of New York State: Evidence for secondary magnetization residing in magnetite, Phys. Earth Planet. Inter., 30, 385-395, 1983.
Sempere, T., Phanerozoic evolution of Bolivia and adjacent regions, in Petroleum Basins of South America, edited by A. J. Tankard, R. Suarez, and H. J. Welsink, AAPG Mem., 62, 207-230, 1995.
Sempere, T., G. Herail, J. Oiler, and M. G. Bonhomme, Late Oligocene-early Miocene major tectonic crisis and related basins in Bolivia, Geology, 18, 946-949, 1990.
Sempere, T., E. Aguilera, J. Doubinger, P. Janvier, J. Lobo, J. Oiler, and S. Wenz, La Formation de Vitiacua (Permien moyen a superieur-Trias ?inferieur, Bolivia de Sud): stratigraphie, palynoiogie et paleontologie, Neues Jahrb. fuer Geol. Palaeontol. Abh., 185{2), 239-253, 1992.
Sempere, T., R. F. Butler, D. R. Richards, L. G. Marshall, W. Sharp, and C. C. Swisher m. Stratigraphy and geochronology of Upper Cretaceous-lower Paleogene strata in Bolivia and northwest Argentina, Geol. Soc. Am. Bull, 109, 709-727, 1997.
Stamatakos, J., A. M. Hirt, and W. Lowrie, The age and timing of folding in the central Appalachians from paleomagnetic results, Geol. Soc. Am. Bull., 108, 815-829, 1996.
Van der Voo, R., Paleozoic paleogeography of North America, Gondwana, and intervening displaced terranes: Comparisons of paleomagnetism with paleoclimatology and biogeographical patterns, Geol. Soc. Am. Bull., 100, 311- 324, 1988.
Watson, G. S., A test for randomness of directions, Mon. Not. Geophys. J. R. Astron. Soc., 7, 160-161, 1956.
Watson, G. S., and E. Irving, Statistical methods in rock magnetism, Mon. Not. Geophys. J. R. Astron. Soc., 7, 289-300, 1957. 115
REFERENCES
Bilokon, H., Castagnoli, G.C., Castellina, S., Piazzoli, B.D., Mannocchi, G., Meroni, E., Picchi, P., Vemetto, S., 1989, Flux of vertical negative muons stopping at depths 0.35-1000 hg/cm^: Journal of Geophysical Research, v. 94, p. 12,145- 12,152.
Bowen, I.S., Millikan, R.A., Korff, S.A., and Neher, H.V., 1936, The latitude effect in cosmic rays at altitudes up to 29,000 feet: Physical Review, v. 50, no. 7, p. 579- 581.
Brown, E.T., Edmond, J.M., Raisbeck, G.M., Yiou, F., Kurz, M.D., and Brook, E.J., 1991, Exaniination of surface exposure ages of Antarctic moraines using in situ produced ^^Be and 26a1: Geochimicaet Cosmochimica Acta, v. 55, p. 2269-2283.
Butler, R. F., 1992, Paleomagnetism: Magnetic Domains to Geologic Terranes: Boston, Blackwell Scientific Publications, 319 p.
Ceding, T.E. and Craig, H., 1994, Geomorphology and in-situ cosmogenic isotopes. Annual Reviews of Earth and Planetary Science, v. 22, p.273-317.
Draganic, Draganic, and Adloff, 1990, Radiation and Radioactivity on Earth and Beyond: Boca Raton, CRC Press, p. 142-147.
Fabryka-Martin, J.T., 1988, Production of radionuclides in the earth and their hydrogeologic significance, with emphasis on chlorine-36 and iodine-129 [Ph.D. thesis]; Tucson, University of Arizona, 400 p.
Friedlander, M.W., 1989, Cosmic Rays: Cambridge, Harvard University Press, p. 83- 107.
Lai, D., 1988, In situ-produced cosmogenic isotopes in terrestrial rocks. Annual Reviews of Earth and Planetary Science Letters, v. 16, p. 355-388.
Lai, D., 1991, Cosmic ray labeling of erosion surfaces: in situ nuclide production rates and erosion models: Earth Planet. Sci. Lett., v. 104, p. 424-439.
Laster, H., Tucker, W.H., and Terry, K.D., 1968, Cosmic rays from nearby supernova: biological effects: Science, v. 160, p. 1138-1139. 116
McHargue, L.R., 1994, A 50,000 year '^Be record from Gulf of California sediments [Ph.D. thesis]: Tucson, University of Arizona, 229 p.
O'Brien, K., 1979, Secular Variations in the production of cosmogenic isotopes in the Earth's atmosphere: Journal of Geophysical Research, v. 84, p. 423-431.
Oeschger, H., and Beer, J., 1990, The past 5000 years history of solar modulation of cosmic radiation from '^Be and studies: Philosophical Transactions of the Royal Society of London A, v. 330, p. 471-480.
Phillips, W.M., McDonald, E.V., Reneau, S.L., and Poths, J., 1998, Dating soils and alluvium with cosmogenic 2lNe depth profiles: Case studies from the Pajarito Plateau, New Mexico, USA, Earth and Planetary Science Letters, v. 160, p. 209- 223.
Raisbeck, G.M., Yiou, F., Jouzel, J., and Petit, J.R., 1990, '^Be and d-H in polar ice cores as a probe of the solar variability's influence on climate: Philosophical Transactions of the Royal Society of London A, v. 330, p. 471-480.
Rose, D.C., Fenton, K.B., Katzman., J., and Simpson, J.A., 1956, Latitude effects of the cosmic ray nucleon and meson components at sea level from the Arctic to the Antarctic: Canadian Journal of Physics, v. 34, p. 968-985.
Simpson, J.A., 1951, Neutrons produced in the atmosphere by the cosmic radiations: Physical Review, v. 83, p. 1175-1188.
Simpson, J.A. and Fagot, W.C., 1953, Properties of the low energy nucleonic component at large atmospheric depths: Physical Review, v. 90, p. 1068-1072.
Sonett, C.P., Morfill, G.E., and Jokipii, J.R., 1987, Interstellar shock waves and from ice cores: Nature, v. 330, p. 458-460.
Stone, J.O.H., Evans, J.M., Fifield, L.K., Allan, G.L., and Cresswell, R.G., Cosmogenic chlorine-36 production in calcite by muons: Geochimica et Cosmochimica Acta, V. 63, p. 433-454.
Terry, K.D. and Tucker, W.H., 1968, Biologic effects of supemovae: Science, v. 159, p. 421-423.