USF Libraries
USF Digital Collections

Thicknesses and density-current velocities of a low-aspect ratio ignimbrite at the Pululagua Volcanic Complex, Ecuador, ...

MISSING IMAGE

Material Information

Title:
Thicknesses and density-current velocities of a low-aspect ratio ignimbrite at the Pululagua Volcanic Complex, Ecuador, derived from ground penetrating radar
Physical Description:
Book
Language:
English
Creator:
Petriello, John A
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Pyroclastic flow
Pyroclastic surge
Northern Volcanic Zone
South American Magmatic Arc
Caldera
Dissertations, Academic -- Geology -- Masters -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: The thinning trend of a low-aspect ratio ignimbrite (LARI) in a direction of increasing topographic relief at the Pululagua Volcanic Complex, Ecuador, is established by correlating continuous ground penetrating radar (GPR) profiles and radar reflector behavior with stratigraphic measurements and unit behavior. Minimum density-current and vertical (cross-sectional) velocity analyses of the LARIs parent pyroclastic density-current are performed by analyzing the exchange of kinetic energy for potential energy in an upslope direction. Continuous GPR profiles were acquired in a direction of increasing topographic relief with the intent of identifying the LARI within the GPR record and examining the relationships between the LARI and the underlying paleo-topographical surface. Stratigraphic measurements recorded throughout the field area demonstrate that the LARI thins 7.5 m in an upslope direction (over 480 m distance and 95 m elevation). Stratigraphic measurements enable correlations with GPR profiles, resulting in LARI identification. By utilizing GPR derived paleo-topographical surface elevations, minimum flow velocities of the LARI-producing parent pyroclastic density-current at the base of upslope flow are shown to be at least 25 m/s. Vertical velocity analyses based on the identification of internal GPR reflectors, interpreted as flow streamlines, yield pyroclastic surge-like cross-sectional velocity profiles of the LARIs parent density-current. Maximum density-current velocities at the base of upslope flow reach 24 m/s and diminish toward the base of the current.
Thesis:
Thesis (M.S.)--University of South Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by John A. Petriello, Jr.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 122 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001920244
oclc - 188000798
usfldc doi - E14-SFE0002115
usfldc handle - e14.2115
System ID:
SFS0026433:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001920244
003 fts
005 20080107150438.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 080107s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002115
040
FHM
c FHM
043
s-ec---
035
(OCoLC)188000798
049
FHMM
090
QE26.2 (ONLINE)
1 100
Petriello, John A.
0 245
Thicknesses and density-current velocities of a low-aspect ratio ignimbrite at the Pululagua Volcanic Complex, Ecuador, derived from ground penetrating radar
h [electronic resource] /
by John A. Petriello, Jr.
260
[Tampa, Fla.] :
b University of South Florida,
2007.
3 520
ABSTRACT: The thinning trend of a low-aspect ratio ignimbrite (LARI) in a direction of increasing topographic relief at the Pululagua Volcanic Complex, Ecuador, is established by correlating continuous ground penetrating radar (GPR) profiles and radar reflector behavior with stratigraphic measurements and unit behavior. Minimum density-current and vertical (cross-sectional) velocity analyses of the LARIs parent pyroclastic density-current are performed by analyzing the exchange of kinetic energy for potential energy in an upslope direction. Continuous GPR profiles were acquired in a direction of increasing topographic relief with the intent of identifying the LARI within the GPR record and examining the relationships between the LARI and the underlying paleo-topographical surface. Stratigraphic measurements recorded throughout the field area demonstrate that the LARI thins 7.5 m in an upslope direction (over 480 m distance and 95 m elevation). Stratigraphic measurements enable correlations with GPR profiles, resulting in LARI identification. By utilizing GPR derived paleo-topographical surface elevations, minimum flow velocities of the LARI-producing parent pyroclastic density-current at the base of upslope flow are shown to be at least 25 m/s. Vertical velocity analyses based on the identification of internal GPR reflectors, interpreted as flow streamlines, yield pyroclastic surge-like cross-sectional velocity profiles of the LARIs parent density-current. Maximum density-current velocities at the base of upslope flow reach 24 m/s and diminish toward the base of the current.
502
Thesis (M.S.)--University of South Florida, 2007.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 122 pages.
590
Advisor: Charles B. Connor, Ph.D.
653
Pyroclastic flow.
Pyroclastic surge.
Northern Volcanic Zone.
South American Magmatic Arc.
Caldera.
690
Dissertations, Academic
z USF
x Geology
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2115



PAGE 1

Thicknesses and Density-Current Velocities of a Low-Aspect Ratio Ignimbrite at the Pululagua Volcanic Complex, Ecuador, Derived from Ground Penetrating Radar by John A. Petriello, Jr. A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Geology College of Arts and Sciences University of South Florida Co-Major Professor: Charles B. Connor, Ph.D. Co-Major Professor: Sarah Kruse, Ph.D. Diana Roman, Ph.D. Date of Approval: June 8, 2007 Keywords: Pyroclastic flow, pyroclastic surge, Northern Volcanic Zone, South American Magmatic Arc, caldera Copyright 2007, John A. Petriello, Jr.

PAGE 2

Acknowledgements I would like to offer many thanks to e ach of my committee members, Dr. Chuck Connor, Dr. Sarah Kruse, and Dr. Diana Roman, for guidance and support during my time at USF. Thank you to Paul Silva and Sussa ne Ettinger for their help during the field session. Thank you to my friends at USF for their time, assistance, and support concerning all aspects (academic and personal) of the master’s process. To my family and friends who are spread throughout the country, thank you for your words of encouragement. It has been a pleasure.

PAGE 3

i Table of Contents List of Tables iii List of Figures iv Abstract vii Chapter One: Introduction 1 Chapter Two: Low-Aspect Ratio Ignimbrites 4 Chapter Three: Pululagua Volcanic Complex 9 Chapter Four: Ground Penetrating Radar 15 Application in Volcanological Studies 15 Theory 18 Introduction 18 Acquisition 18 Common Midpoint Sounding 19 Wave Propagation 21 Velocity 22 Attenuation and Absorption 22 Chapter Five: Methods 24 Data Overview 24 Stratigraphic Section Acquisition 30 GPR Acquisition 31 GPR Processing 34 GPS Acquisition 36 Chapter Six: Results 38 Study Site 38 Stratigraphy 38 Lithology 43 Stratigraphic Section 13101 44 Stratigraphic Section 13106 47 Upslope Thinning 47

PAGE 4

ii GPR Analysis 50 CMP Soundings 64 Velocity Analyses 65 Bulk Density-Current Velocity 66 Vertical (cross-sectional) Velocity 70 Chapter Seven: Discussion 74 Chapter Eight: Conclusions 77 Chapter Nine: Recommendations 79 References 80 Appendices 84 Appendix A: Stratigraphic Sections and Legend 85 Appendix B: LARI Delineations with Thic kness vs. Distance Curves 92 Appendix C: Topographically Corrected LARIs 101 Appendix D: LARI-Boundi ng Reflector Matlab Code 106 Appendix E: CMPs with LARI-Bounding Re flector Delineations 108 Appendix F: GPR Profiles 112

PAGE 5

iii List of Tables Table 1: GPR reflection profile acqui sition parameters 33 Table 2: CMP Sounding acquisition parameters 34 Table 3: GPR profile processing parameters 35 Table 4: CMP sounding processing parameters 36 Table 5: GPS measurement frequencies 37 Table 6: Stratigraphic section measurements 42 Table 7: LARI thickness correlations 43 Table 8: GPR Site thickness deductions 53 Table 9: CMP velocities 65

PAGE 6

iv List of Figures Figure 1: Highand low-aspect ratio ignimbrite topographical relations 2 Figure 2: Pyroclastic density-current cr oss-sections 5 Figure 3: Highand low-aspect ratio igni mbrite schematic 7 Figure 4: Map of Western Ecuador 11 Figure 5: Map of the Pululagua Region 12 Figure 6: Geological sketch map of the Pu lulagua Volcanic Complex 13 Figure 7: Stratigraphy and nomenc lature of the products of the Pululagua caldera-forming eruptions 14 Figure 8: Example GPR re flection profile radargram 19 Figure 9: CMP sounding acquisition schematic 20 Figure 10: CMP radargram 21 Figure 11: Pululagua syn-caldera depo sit distribution 25 Figure 12: Pululagua study area 26 Figure 13-A: GPR Sites 1 and 2 27 Figure 13-B: GPR Sites 3 and 4 28 Figure 14: View looking SE toward GPR Sites 1, 2, and 3 29 Figure 15: View looking NE toward GPR Sites 1 and 2 29 Figure 16: View looking SE towards GPR Site 3 30 Figure 17: Study area with stra tigraphic section labels, GPR site labels, and fence diagram A-A’ 39 Figure 18: Soil, Upper Surge, Upper Tephra Fall, and the LARI 40

PAGE 7

v List of Figures (Continued) Figure 19: LARI, Lower Surge Package I, Accretionary Lapilli, and Lower Surge Package II 41 Figure 20: Lower Surge Package II, Tephra Fa llout Deposits, and Paleosol 41 Figure 21: Stratigraphic Section 13101 46 Figure 22: Fence diagram A-A’ 48 Figure 23: Stratigra phic sections and associated legend relevant to Fence Diagram A-A’ 49 Figure 24: The stratified nature of the Upper Surge 50 Figure 25: GPR Profile displaying surge-like reflectors 51 Figure 26: Site 1: Line 1 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 54 Figure 27: Site 1: Line 2 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 55 Figure 28: Site 2: Line 1 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 56 Figure 29: Site 2: Line 2 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 57 Figure 30: Site 2: Line 4 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 58 Figure 31: Site 2: Line 5 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 59 Figure 32: Site 3: Line 1 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 60 Figure 33: Site 4: Line 1 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 61 Figure 34: Site 4: Line 2 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 62

PAGE 8

vi List of Figures (Continued) Figure 35: Site 4: Line 4 GPR profile with LARI delineation, topographically corrected LARI, and thickness vs. distance curve with trend line 63 Figure 36: Upslope turbidity current flow schematic 66 Figure 37: Bulk density-current velo city schematic 66 Figure 38: Bulk density-current velocities 69 Figure 39: Dividing streamline schematic 70 Figure 40: Vertical velocity profile schematic 72 Figure 41: Site 3_Line 1 streamline delineatio ns and velocity profiles 73 Figure 42: LARI-producing density current with cross-sectional profiles 78 Figure 43: Basal LARI-producing density-current with cross-sectional profiles 78

PAGE 9

vii Thicknesses and Density-Current Velocities of a Low-Aspect Ratio Ignimbrite at the Pululagua Volcanic Complex, Ecuador, Derived from Ground Penetrating Radar John A. Petriello, Jr. ABSTRACT The thinning trend of a low-aspect ra tio ignimbrite (LARI) in a direction of increasing topographic relief at the Pululagua Volcanic Complex, Ecuador, is established by correlating continuous ground penetrating radar (GPR) profiles and radar reflector behavior with stratigraphic measurements and unit behavi or. Minimum density-current and vertical (cross-sectional) velocity analyses of the LARI s parent pyroclastic densitycurrent are performed by analyzing the exchange of kinetic energy for potential energy in an upslope direction. Continuous GPR profiles were acquired in a direction of increasing topographic relief with the intent of iden tifying the LARI within the GPR record and examining the relationships between the LARI and the underlying paleo-topographical surface. Stratigraphic measurements recorded throughout the field area demonstrate that the LARI thins 7.5 m in an upslope direc tion (over 480 m distance and 95 m elevation). Stratigraphic measurements enable correlati ons with GPR profiles, resulting in LARI identification. By utilizing GPR derived paleo-topographical surface elevations, minimum flow velocities of the LARI-producin g parent pyroclastic de nsity-current at the base of upslope flow are shown to be at leas t 25 m/s. Vertical velocity analyses based on the identification of internal GPR reflectors, interpreted as flow streamlines, yield

PAGE 10

viii pyroclastic surge-like cross-sectional velo city profiles of the LARIs parent densitycurrent. Maximum density-current velocities at the base of upslope flow reach 24 m/s and diminish toward the base of the current.

PAGE 11

1 Chapter 1 Introduction An ignimbrite is a pyroclastic deposit composed predominantly of pumiceous material, emplaced as a hot, sediment-laden flow (Walker et al., 1980; Walker, 1983). Differences between conventional ignimbrites (high-aspect ratio i gnimbrites or HARIs) and less well understood lowaspect ratio ignimbrites (LARIs) imply behavioral variations within their parent pyroclastic density-currents. The conventional ignimbriteforming current is thought to be largely controlled by pre-existing topography, as indicated by relatively thic k deposits that tend to pond in topographic depressions (Walker et al., 1980; Walker, 1983; Valentin e, 1987; Druitt, 1998). The tendency of LARIs to thinly mantle pre-existing topogr aphy, combined with th e identification of LARIs in areas of high relief, has led to th e deduction that thei r parent pyroclastic density-currents are minimally influenced by pre-existing topography and are capable of surmounting substantial topogra phic barriers (Fig. 1) (Walke r et al., 1980; Walker, 1983; Valentine, 1987; Druitt, 1998).

PAGE 12

2 Figure 1: Highand low-aspect ratio ignimbrite topographical relations; VP-valley pond, typical of high aspect ratio ignimbrites (HARIs); IVD-ignimbrite veneer deposit, typical of low aspectratio ignimbrites (LARIs) (modified from Walker et al., 1980). Agreement exists about several aspects of LARI formation. LARIs tend to form during eruptions where high discharge rates ar e maintained for limited periods of time, are often initiated by the collaps e of an eruption column, and during transport their parent flows are extremely mobile, potentially reaching velocities of 300 m/s (Druitt, 1998; Valentine, 1987). The transport and em placement mechanisms of LARI-forming pyroclastic currents are poorly understood. Discrepancies largel y revolve around the nature of transport (dominantly turbulent vs dominantly laminar) a nd the ability of the flow to surmount topographic obstacles, dependent upon the thickness of the flow (expanded vs. concentrated) (Valentine, 1987; Wilson, 1985; Fisher et al., 1993; Druitt, 1998). Here I present a case study addressing th e nature of LARI-thi nning in a direction of increasing topographic reli ef with corollary velocity analyses based on thinning observations. A pyroclastic flow emplaced during a caldera-forming episode at the Pululagua Volcanic Complex (PVC), Ecuador, (Papale and Rosi, 1993; Andrade, 2002, 2006) is explored via stratigraphic obse rvations and ground-penetrating radar

PAGE 13

3 investigations. The nature of th inning of the pyroclastic flow in a direction of increasing topographic relief is explored by correlating stratigraphic measurements with continuous ground-penetrating radar (GPR) profiles of the deposit. The utilization of GPR allows for mapping in areas where the LARI cannot be directly observed (Davis and Annan, 1988; Russel and Stasiuk, 1997). GPR reflection profiles of the LARI-covere d slope were recorded approximately parallel and normal to the inferred direction of flow. The unique capability of GPR makes it possible to identify reflectors both w ithin the deposit and b ounding the deposit, and thus trace these reflectors as the deposit climbs the underlying paleo-topography. Interpretation of these reflectors as both flow contacts and bedding horizons within the flow allows estimation of flow velocity as a function of flow thic kness and depth of the bedding horizons within the unit. The majority of velocity analyses were performed using GPR data collected on profiles that climb th e topographic slope, r oughly parallel to the inferred flow direction.

PAGE 14

4 Chapter 2 Low-Aspect Ratio Ignimbrites The aspect ratio of a rock unit is defined as T/D where T represents the average thickness, and D represents the diameter of a circle covering the same planimetric area as the unit (Walker, et al., 1980; Walker, 1983; Peterson and Tilling, 2000). The classification of a rock unit using the as pect ratio is a means of quantifying and comparing unit geometry, and in this case is pertinent to ignimbrite deposits. Aspect ratios of ignimbrites have been shown to vary between roughly 10-2 and 10-4; values approaching the former are known as high-as pect ratio ignimbrites (HARI), and those approaching the latter are known as low-aspect ratio ignimbrites (L ARI) (Walker, 1983). Thus, the aspect ratio is a way of desc ribing flow mobility independent of deposit volume. Using map data from Andrade (2000), the aspect ratio of the pyroclastic flow deposit at the Pululagua Volcanic Comp lex is calculated to be between 10-3 and 10-4, approaching the LARI classification. Traditionally the term ignimbrite has been restricted to the depositional product of a pyroclastic flow, and it is in this sens e that it will be used here. The term pyroclastic flow is typically used to refer to the highl y sediment-concentrated end-member of the pyroclastic density-current spectrum, the d ilute end-member being pyroclastic surge. Fundamental differences between the pyroc lastic density-current end-members are recognized. Recognition of end-member variat ions is largely deduced from eyewitness accounts of eruptions and from deposit studi es (Anderson and Flett, 1903; Lacroix, 1904;

PAGE 15

5 Fisher, 1993; Druitt, 2002). Pyroclastic flows typically have solids concentrations on the order of tens of volume percen t, have a free surface above which the solids concentration sharply diminishes, and transport material through a variety of mechanisms, including particle-particle contact, fluidization support, matrix support, dispersive pressure, and buoyancy (Wilson and Houghton, 2000). Most mass and momentum are carried in the basal current, resulting in higher basal velo cities than the overriding cloud (Fig. 2) (Wilson and Houghton, 2000). Pyroclastic surges contain less than 0.1 to 1% volume solids, are density stratified, with higher par ticle concentrations n ear the ground surface, and transport material primarily through turbulent suspension (Wilson and Houghton, 2000). Mass and momentum in surges are more evenly distributed, therefore basal concentrations of material are derived from sedimentation from the current, and basal velocities are slower due to ground frict ion (Fig. 2) (Wilson and Houghton, 2000). Turbulence is not a principle support mechan ism in pyroclastic flows, although it may or may not be present (Wilson and Houghton, 2000). Figure 2: Pyroclastic density-current cross-sec tions. Schematic velocity and density cross-section through the dilute end-member (pyroclastic surge) and the concentrated end-member (pyroclastic flow) (modified from Wilson and Houghton, 2000).

PAGE 16

6 Classification of an ignimbrite as HAR or LAR type has implications beyond unit geometry. Unit aspect ratios are influe nced, and give clues about, pre-existing topography, initial eruptive conditions, and transport and emplacement mechanisms. Pyroclastic flows producing HARIs are of ten confined by topographically bounded valleys and plains and the resulting deposit is a relatively thick unit, varying between approximately 10 and 1000 meters (Walker et al, 1980; Dade, 2003). The tendency to pond in topographic depressions, occurrence as a relatively thick deposit, and the presence of a horizontal or gently slopi ng upper depositional surf ace are common criteria used to distinguish HARIs from other pyr oclastic deposits (W alker et al, 1980). Conversely, pyroclastic flows yi elding LARIs appear to be minimally controlled by preexisting topography, and the resultant de posits are dominated by a thin, landscape mantling veneer deposit (Walker et al., 1980) LARIs have been observed resting on slopes of up to 30 degrees, with an upper surf ace nearly parallel to the underlying surface (Walker et al., 1980). Thus, HARIs highly alte r the pre-existing landscape, essentially erasing evidence of pre-existi ng topography, and are concentr ated in topographic lows, while LARIs tend to mimic the landscape, ma intaining pre-existing topography, and have the ability to mantle steep slope s. An implication of LARI sl ope characteristics is that a LARI-producing flow can literally climb t opographic obstacles, a phenomenon that does not occur in a HARI-producing flow. It must be noted that valleyponding is associated with LARI (Walker et al., 1980; Walker, 1983); however, the majority of the LARI consists of the veneer deposit connecting isolated valley-ponded regions (Fig. 3).

PAGE 17

7 Figure 3: Highand low-aspect ratio ignimbrite sch ematic; where Fextent of pyroclastic flow deposits; Sseared zone, pyroclastic surge; VPvalley pond deposit; Vlandscape mantling veneer deposit; Gfines-depleted deposits unde rlying normal flow material (modified from Walker, 1983). Deviations in LARI characteristics from those of more common ignimbrites (HARIs) are thought to reflect initial eruptive conditions. It is believed that LARIs form during eruptions where high discharge rates are maintained for limited periods of time (Druitt, 1998). As the erupti on column ascends into the at mosphere, a dynamic interplay between column velocity, water content, ai r entrainment, and pa rticle sedimentation determine the fate of the column. If at the point when the velocity of the column approaches zero, the density of the column ex ceeds that of air, the column will collapse to form laterally flowing density currents (D ruitt, 1998). It is typi cally this phenomenon that induces ignimbrite deposition. A source of debate within the LARI lit erature concerns transport mechanisms of the pyroclastic flow after column collapse, la rgely concerning the state of the flow as it moves across the landscape, either as an expanded flow that is thicker than the topographic obstacles it traver ses, or as a dense flow that moves as a ground-hugging HARI LARI

PAGE 18

8 sheet across the landscape. In the expanded flow model, the ability of a flow to traverse topographic obstacles is attri buted to large flow thickness relative to the topographic obstacle (Fisher et al., 1993, Valentine, 1987, Druitt, 1998). In the dense flow model, this ability is due to the high momentum of th e flow (Wilson, 1985, Druitt, 1998). The ability to image a LARI with high resolution GPR da ta may help differen tiate between these transport mechanisms.

PAGE 19

9 Chapter 3 Pululagua Volcanic Complex The Pululagua Volcanic Complex (PVC) is located in the Northern Volcanic Zone of the South American magmatic arc (Fi g. 4) (Barberi et al., 1998). The caldera is about 15 km north of Quito, Ecuador (Fig. 5) The PVC is defined by an eruptive center (3 x 2 km caldera), syn-caldera deposits, and preand post-caldera domes and dome deposits (Fig. 6) (Papale and Rosi, 1993). The volcanic history of the PVC has been broken into four series by Andrade (Andr ade 2002, 2006): Series I is represented by old pre-caldera domes and deposits, Series II is represented by young pre-caldera domes and deposits, Series III is represented by syn-calde ra deposits, and Series IV is represented by post-caldera domes and deposits. The lowermost portion of Series III deposits (inception of the syn-caldera phase) have previously been studied in detail by Papale and Rosi (1993) and Volentik et al., (2005, 2006). The syn-caldera phase comprised at least ten eruptive episodes that led to caldera collapse (Papale and Rosi, 1993). Stratigraphy of the Pululagua caldera-forming eruptions has been compiled by Papale and Ro si (1993), in which the entire sequence is divided into lower, middle, and upper un its, based on deposit thicknesses and inferred eruption intensities. The LARI studied here is located in the upper portion of the middle eruptive units (near U6 & U7) as defined by Papale and Rosi (1993) (Fig. 7). The identification of additional eruptive units le d Andrade (2002) to divide the Series III sequence into four episodes, each delineated by a thin bioturbated ash layer and dated

PAGE 20

10 using carbonized wood. The basal portion of Episode 1 has been dated at 2575 45 yBP, while the lowermost Series IV unit (direc tly overlying the uppermost Episode 4 ash layer) has been dated at 2240 50 yBP (Andrade, 2002), constraining the duration of the caldera-forming phase. The LARI lies within Episode 4, in which a radiocarbon date at the base has yielded 2460 70 yBP (Andrade, 2006). Depositional sequences and unit thicknesses within the four syn-caldera episodes vary according to geographic location and proximity to the eruptive center. The lowaspect ratio ignimbrite that is the focus of this study is lo cated in Episode 4 as defined by Andrade (2002, 2006) and the upper portion of the middle eruptive units as defined by Papale and Rosi (1993); therefore it is a pr oduct of late-stage cal dera forming events. This pink pumice and ash-rich deposit is largely concentrated in the no rtheast, east, south, and southeastern portions of the caldera. De posits in the north, northwest, and southwest are generally confined to topographic lows. In the areas of greatest concentration the unit is seen mantling the surface, in cases having been deposited while the flow was traveling uphill. Thicknesses measured in the south vary between 2-9 meters. In contrast, thicknesses in the north and west vary between 7-30 meters (Andrade 2002, 2006). The field area for this study is loca ted to the south of the caldera, largely because this area is much more accessible for geophysical surveys and stratigraphic sections are exposed in quarry walls and erosional gullies.

PAGE 21

11 Figure 4: Map of Western Ecuador. Blue circlescities; Red triangles-volcanoes. Pululagua is located just north of the equator.

PAGE 22

12 Figure 5: Map of the Pululagua region. Blue ci rclescities; Red triangles-volcanoes; Green box location of the study site.

PAGE 23

13 Figure 6: Geological sketch map of the Pulula gua Volcanic Complex; where the red square indicates the location of the stratigraphic colu mn shown in Figure 7 (modified from Papale and Rosi, 1993).

PAGE 24

14 Figure 7: Stratigraphy and nomenclature of produc ts of the Pululagua caldera-forming eruptions, observed NE of the caldera (red box in Fig. 6). U1-U10-eruptive units ar e separated from each other by erosive unconformities. BF, WA, F2-F7-Pl inian and/or Subplinian pumice fallout layers. Numbers on the left-hand side of the column indicate deposit thickness in cm. The LARI is deposited near U6 & U7 in the middle eruptive un its. Nomenclature is based on Papale and Rosi (1993). (modified from Papale and Rosi, 1993).

PAGE 25

15 Chapter 4 Ground penetrating radar Application in Volcanological Studies The use of ground-penetrating radar (GPR) in volcanological studies has increased in recent years. This stems from de monstrations of GPR cap abilities outside of the volcanological realm. Some early case st udies demonstrating the effectiveness of GPR in acquiring detailed shallow subsurface information were presented by Davis and Annan (1988). They showed that GPR is e ffective to depths of 20 m both for mapping sedimentary stratigraphy and for detecting fr acture zones in igneous rock. Their GPR interpretations were confirmed by ground truthi ng. Numerous other ca se studies led to the realization that GPR is effective in acquiring detailed subsurface information, particularly in sedimentary environments, and spawned the integr ation of GPR into studies of volcanic deposits. Descriptions of GPR contri butions to volcanological i nvestigations are reviewed below. At the caldera of Volcan Sollipulli in Chile, a combined radar and gravity survey was used to determine calder a ice thicknesses along 2D profile s and provided a constraint on the buried topography within the ice-filled caldera (Gil bert et al., 1996). Antenna frequencies of 3.5 and 5.8 MHz were utilized along two survey lines, and, based on radar reflections, caldera ice thic knesses of 424 meters were inferred (with an estimated 10 m

PAGE 26

16 accuracy). Reflections from the ice-rock inte rface were detected in all but the central portion of the caldera. A GPR survey of Holocene volcanic depos its in western Canada demonstrated that GPR is effective in delineating stratig raphic contacts and has potential to aid in quantifying deposit distributions and thic knesses (Russel and Stasiuk, 1997). GPR transects were run using 100 MH z antennae in an attempt to correlate radar profiles with near-site field exposures. Four volcanic units we re studied, a 36 meter thick basalt flow, a 34 meter thick pumice-rich tephra fallout a 15 meter thick pyroc lastic flow deposit, and a 60 meter thick pumice talus deposit (Rus sel and Stasiuk, 1997). Di rect correlations between unit characteristics a nd radar response were made by performing GPR surveys overlying observable deposits. These capabilit ies will be discussed with respect to the basalt lava flow and pumice-rich tephra fallout. In the case of the basalt lava flow, a GPR profile 40 meters in length was carried out along the flow’s upper surface. A poorly indurated, irregular, scoriaceous autobreccia up to 1 meter thick makes up the lowest portion of the flow (Russel and Stasiuk, 1997) Characteristic diffraction events were evident within the radar record at depths equivalent to the stratigraphically measured lava flow base. The diffraction events were interp reted to have been either induced by basal flow irregularities or basal autobreccia. An alysis of the tephra fallout deposit was performed with similar stratigra phic control. As was the case with the lava flow, the GPR survey was performed directly above crosssectional exposures. Underlying the teprha fallout is a thick layer of colluvium. At near midpoint along the 50 meter traverse, the radar record revealed a strong reflection at 3.6 meters depth, th at was nearly equivalent to the stratigraphically-measured basal fall depth. Rapid attenuation below this level in the

PAGE 27

17 GPR profile was interpreted as an artifact of the high el ectrical conduc tivity of the underlying colluvium. Overlying the basal re flector, a general absence of coherent reflectors was interpreted to be a result of the overall massive character of the tephra fallout deposit (Russel and Stasiuk, 1997). Other GPR contributions to physical vol canology include subsurface stratigraphic analyses at the Ubehebe hydrovolcanic fiel d, Death Valley, California. GPR surveys were performed initially with antennae frequencies of 50, 100, and 200 MHz, and in a corollary study with antennae frequenc ies of 900 MHz (Cagnoli and Russel, 1999; Cagnoli and Ulrych, 2000). The lower frequency antennae permitted evaluation of base surge deposits and alluvial material thickne sses and revealed stratigraphic unconformities between base surge deposits and underlyi ng sandstones (Cagnoli and Russel, 1999). The high-frequency portion of their study provided resolutions that allowed for interpretations of climbing dune forms in the base surge deposit. Cumulatively, these studies show that the utilization of GPR in volcanic terrain can provide an understanding of the subsurface that in some cases is well beyond what is possible via measurements of exposed stratig raphic sections. Implementing this method over broad areas provides a great deal more information than do section measurements alone. Furthermore, these studies show that GPR results nicely correlate with stratigraphic data (i.e., borehol e logs, sections) taken in pr oximity of the GPR traverse. Such direct correlations can yield definitiv e radar identifications which can then be extrapolated to the entire GPR traverse.

PAGE 28

18 Theory Introduction Ground-penetrating radar (GPR) is a nea r-surface geophysical method that uses high-frequency radio waves as a means of det ecting subsurface contrast s in electrical and magnetic properties. The premise behind GPR is that as a radio wave contacts an interface separating materials of varying elec trical and magnetic properties, a portion of the wave is reflected and later received by the GPR system. A typical system consists of a signal generator, transmitting and receiving antennae, and a receiver that amplifies, digitizes, and stores the re turning signal (Davis and Annan, 1989; Reynolds, 1997). The instrument is able to determine precisely the time difference between wave transmission and arrival. With appropriate ve locity constraints, radar travel times can be converted into depth measurements, which indicate the de pth to the radar re flector (i.e., the electromagnetic contrast). Acquisition GPR acquisition is many ways analogous to seismic reflection methods. GPR systems are often run in reflection profili ng mode, with a system consisting of one transmitting antenna and one receiving antenna, with a fixed offset between the antennae. The seismic reflection analog to GPR profiling is the common offset method, the name of which derives from the equal increments in which the seismic source is progressively

PAGE 29

19 offset. The results of GPR profiles are radargra ms, showing travel time to radar reflectors (or depth) versus distance from a fixed starting point ( Fig. 8 ). A requirement for travel time to depth conversion is knowledge of rada r velocities. Velocities are determined by performing common midpoint soundings (CMPs), which need to be performed close to the time of profile collecti on, as changes in ground mois ture alter radar velocity. Inaccurate depth values will be calculated if the moisture content varies between the time of profile and CMP acquisitions. Figure 8: Example GPR reflection profile rada rgram. Data from Site 2_Line 3 (see text). Common Midpoint Sounding Common midpoint soundings (CMPs) ge nerate reflections from common midpoints in the subsurface by moving the tran smitter and receiver away from each other such that the midpoint remains fi xed (Fig. 9) (Reynolds, 1997).

PAGE 30

20 Figure 9: CMP sounding acquisition schematic; where Tx-transmitting antenna, Rx-receiving antenna, V1 and V2-unit velocity, d-depth, a nd (i, ii, iii)progressive acquisition locations (Reynolds, 1997). The seismic analog to a CMP sounding is alte rnatively referred to as the common depth point method, CMP stack, or CMP gather. An alysis of CMP soundings can also follow techniques developed for seismic surveys. On a CMP radargram of travel-time (t) versus distance (x) (Fig. 10), one can identify the ai r wave and the direct wave, both with a direct travel path from transmitter to rece iver, the former through air and the latter through near-surface materials, which plot as straight lines. Later ar rivals are reflected waves which plot as hyperbolas, a result of the increasing time requirements necessary to reach the receiver with progressively larger transmitter-receiver separations. Reflector arrival times are picked, and both arrival ti mes (t) and their relative distances (x) are squared in order to determine velocities. The result is an x2-t2 function, which translates the hyperbolic reflection into a linear segment. Root-mean-squared velocities and in turn interval velocities can be calculated according to the Di x Method, and two-way travel time-to-depth conversions can then be perfor med using a constant velocity assumption as derived from CMP analysis. GPR da ta are typically not migrated.

PAGE 31

21 Figure 10: CMP radargram; where air wave (yello w), direct wave (red), and reflection hyperbolas (blue) are highlighted. Wave Propagation Electrical and magnetic prope rties of geologic media exert fundamental controls on wave propagation, specifically velocity and attenuation. Ve locity and attenuation are dependent on subsurface properties, namely the relative permitt ivity, conductivity, and magnetic permeability. The physics of radar wa ve behavior can be very complex; however, in most geological se ttings simplified relationships (presented below) between electromagnetic properties and wave propagation are valid.

PAGE 32

22 Velocity Radar wave velocity is governed by the relative permittivity and relative magnetic permeability, r rc v / (1) where v -wave velocity (m/s), c -electromagnetic wave velocity in a vacuum (3 x 108 m/s or 0.3 m/ns), r relative magnetic permeability (dimensionless), and r -relative permittivity (dimensionless). In most geologic media, magnetic minerals (i.e., iron oxides) are not a bundant, therefore r=1, and wave velocity is essentially dependent on r. Wave velocity is inversely proportional to the relative pe rmittivity; so as r increases, wave velocity decreases. In geologic se ttings, the relative permittivity has a minimum value for air ( r =1) and a maximum value for water ( r =80), therefore the relative permittivity of the media will lie somewhere between 1 and 80. This potential r range yields a possible velocity range between 0.3 and 0.033 m/ns. Most dry geologic materials have a relative permittivity between 4 and 8, so it is largely the water content that controls wave velocity (Davis and Annan, 1985). It is a natural corollary that rock porosity influences wave behavior, as porosity controls the amount of space available for water. Determined velocities of the LARI at Pululagua vary between 0.108 m/s and 0.128 m/s. With the assumption that r=1, this indicates that r values vary between 5.5 and 7.7. Attenuation and Absorption Energy losses leading to attenuation occur as radar waves propagate into the ground for several reasons. Inherent to radar waves is attenuation due to the geometrical spreading of energy. As the wave spreads in a spherical manner throughout the

PAGE 33

23 subsurface, a reduction in energy per unit area occurs at a rate of 1/ r2 ( r is the distance traveled). Another means by which energy is lo st is scattering. Wher e the wave meets an object or contact with a dimension similar to or greater than the wave length, scattering of energy will occur, decreasing the amplitu de of the transmitted wave. Finally, the conversion of electromagnetic energy into heat also contributes to overall energy losses, a phenomenon termed absorption. Absorption is a function of material conductivity, relative magnetic permeability, and the relative permittivity (Reynolds, 1997). Wave attenuation is characterized in terms of the skin depth ( ) and its inverse, the attenuation factor ( =1/ ). Skin depth is a general EM term that can be defined as the depth in which the signal decreases in amplit ude to 37% or (1/e) of the in itial value (Rey nolds, 1997). In non-magnetic materials, it is th e conductivity that primarily controls wave attenuation. A general rule is that as conduc tivity increases, the a ttenuation factor in creases, leading to smaller skin depth values and lower penetra tion depths. Skin depths can range from the millimeter scale for clay rich substances (no penetration) to the decameter scale for limestones and granites. Penetration depths in the Pululagua study area are a maximum of 18 m.

PAGE 34

24 Chapter 5 Methods Data Overview Stratigraphic, ground penetrating radar (GPR), and complementary GPS data were acquired with the intent of analyzi ng the relationships between LARI thicknesses and paleo-topography, and to delineate the intern al structure of the LARI. All data were acquired southeast of the Pululagua Caldera, 4.8-5.4 km from the caldera’s center (Fig. 11). This area was chosen because the LARI is thin relative to areas in the North and East of the caldera, and the LARI is deposited in areas of increasing topographic relief (Fig. 12), allowing for upslope thinning and s ubsequent velocity analyses. Thirteen stratigraphic sections were measured throughout the field area. GPR data consists of CMP soundings and profiles acquired at four individual study sites, totaling seven CMP soundings and 21 profiles (Fig. 13). Distan ces between stratigraphic sections and the nearest GPR profile range from 20 meters to several hundred meters Topography is generally increasing from north to south (Figs. 14-16).

PAGE 35

25 Figure 11: Pululagua syn-caldera deposit distribution. The caldera rim and the area of study are shown.

PAGE 36

26 Figure 12: Pululagua study area. Stratigraphic labels, GPR Sites, and syn-caldera deposit distributions are displayed.

PAGE 37

27 Figure 13-A: GPR Sites 1 (above) and 2 (below); where numbers in blue correspond to the associated GPR profile and the red CMP label co rresponds to the associated CMP (red triangle).

PAGE 38

28 Figure 13-B: GPR Sites 3 (above) and 4 (below); where numbers in blue correspond to the associated GPR profile and the red CMP label co rresponds to the associated CMP (red triangle). The exact location of CMP 2 is unknown due to data loss.

PAGE 39

29 Figure 14: View looking SE toward GPR Sites 1, 2, and 3. The increasing topographic relief low on the flanks of Casitahua Volcano (Fig. 6) is visible in the background. The pink LARI can be seen in the foreground. Figure 15: View looking NE toward GPR Sites 1 and 2. The photograph was taken from a position near Stratigraphic Section 20202 (Fig. 12). The pink LARI is visible throughout the section. LARI LARI

PAGE 40

30 Figure 16: View looking SE towards GPR Site 3. The photograph was taken from a position near Stratigraphic Section 13101 (Fig. 12). Stratigraphic Section Acquisition Erosional gullies and quarry road cuts facilitated the measurement of 13 stratigraphic sections throughout the study area. Section locati ons were chosen with the intent of accurately recording variations in LARI thickness. Clear exposures throughout the study area allowed for accurate measurements. Sections were measured with a standard survey tape. At all section locations, cl ear stratigraphic contacts allowed for precise thickness measurements of nearly all visible units. Th e LARI was visible in its entirety at all stratigraphic section locations. Strati graphic nomenclature is based on a

PAGE 41

31 month_date_section number format. For exampl e, six sections were taken on January 31, 2006, and are respectively labeled 13101 to 13106. GPR Acquisition GPR site selections are dict ated by several key factors. First, vehicle accessibility is a necessity due to the high quantity of GPR equipment. Fortuna tely, many regions to the south of the caldera are quarried, provi ding access via quarry roads and allowing freedom of site selection. Second, the terr ain must allow for reasonable GPR acquisition. Surveys cannot be performed in areas of eith er abundant vegetation or extremely rugged terrain, due to the requirement that both th e GPR transmitter and receiver stay coupled with the ground. Therefore, study sites were al so chosen according to these constraints, with vegetation consisting of small grasses and sparse shrubs, a nd topographic gradients remaining somewhat mild. Third, GPR acquisi tion took place in areas where clear field exposures allowed for stratigraphic control. Se veral stratigraphic sec tions were recorded within 30 meters of a GPR profile, allowi ng for correlations between stratigraphy and radar reflections. At each of the four GPR sites, site profiles and CMP soundings were acquired within 1-2 hours of each other to ensure that CMP derived velocities would accurately represent those of th e profiles. CMP soundings were acquired within meters of GPR profiles, often crossing a profile traverse path. GPR profile distances and orientations va ry between sites, as a function of the terrain (erosional gullies). Profiles that we re acquired in an upslope or downslope

PAGE 42

32 direction typically span the largest distan ces. At Sites 1, 2, and 3, these lines were traversed from the SW to NE. At Site 4 th ese lines were traversed from N to S. GPR data were acquired using a Se nsors & Software PulseEKKO 100 GPR system. A 400-volt transmitter was used for both CMP and profile acquisitions. Reflection profiles were acquire d in bistatic antennae mode with a sampling rate of 800 ps. During profiling mode, the number of stack s for each trace was 16, and the attempted step size for each line was 25 cm or less. Su ch a step size is desired to avoid spatial aliasing. The number of stacks for each CMP tr ace was 32. Information regarding profile and CMP variables such as antenna frequenc ies, antennae separation, time windows, the number of traces per profile/CMP, and CMP step sizes are presented in Tables 1 and 2.

PAGE 43

33 Site:Line Antenna Frequency (MHz) Antennae Separation (m) Time Window (nS) Traces Profile Distance (m) Average Trace Spacing (m) Site 1: Line 1 100 1 400 919 292 0.32 Line 2 100 1 500 775 292 0.37 Line 3 100 1 400 211 81 0.38 Line 4 100 1 400 200 85 0.43 Site 2: Line 1 100 1 400 1005 272 0.27 Line 2 100 1 400 1081 267 0.25 Line 3 100 1 400 153 35 0.23 Line 4 100 1 400 671 156 0.23 Line 5 100 1 400 153 32 0.21 Line 6 100 1 400 194 57 0.29 Line 7 100 1 400 144 43 0.29 Site 3: Line 1 100 1 400 1039 197 0.19 Line 2 100 1 400 235 45 0.19 Site 4: Line 1 100 1 400 2219 443 0.19 Line 2 200 0.5 250 5138 450 0.08 Line 3 200 0.5 250 209 19 0.09 Line 4 200 0.5 250 187 17 0.09 Line 5 200 0.5 250 145 14 0.09 Line 6 200 0.5 250 252 26 0.10 Line 7 200 0.5 250 162 16 0.10 Line 8 200 0.5 250 298 25 0.08 Table 1: GPR reflection prof ile acquisition parameters.

PAGE 44

34 Site: CMP Frequency (MHz) Initial Antennae Separation (m) Step Size (m) Time Window Traces Site 1: CMP 1 100 1 0.2 400 55 CMP 2 100 1 0.2 500 55 Site 2: CMP 1 100 1 0.2 400 64 Site 3: CMP 1 100 1 0.2 400 55 Site 4: CMP 1 100 1 0.2 400 53 CMP 2 200 0.5 0.1 200 96 CMP 3 200 0.5 0.1 250 90 Table 2: CMP Sounding acquisition parameters. Th e CMP step size refers to the distance in which the transmitter and receiver are progressively offset (i, ii, iii in Fig. 9). GPR Processing All GPR data were dewowed prior to analysis to remove low frequency components. Automatic gain controls ( AGC) were applied to enhance reflector recognition. Several GPR profiles required a reversal to main tain consistency in final presentation. For example, Site 1: Line 2 was originally acquired from NE-SW, however the final format is presented as SW-NE. Ta bles 3 and 4 are a compilation of profile and CMP processing parameters.

PAGE 45

35 Site:Line Reversal AGC Window (ns) AGC Max (ns) Site 1: Line 1 40 100 Line 2 R 50 100 Line 3 50 100 Line 4 R 50 100 Site 2: Line 1 40 100 Line 2 R 40 100 Line 3 40 100 Line 4 40 100 Line 5 40 100 Line 6 R 30 100 Line 7 40 100 Site 3: Line 1 R 40 100 Line 2 30 100 Site 4: Line 1 40 100 Line 2 R 30 100 Line 3 R 40 100 Line 4 R 30 100 Line 5 R 25 100 Line 6 R 25 100 Line 7 R 30 100 Line 8 25 100 Table 3: GPR profile processing parameters; wher e R-profile reversal and (-)-Original acquisition orientation.

PAGE 46

36 Site: CMP AGC Window (ns) AGC Max. (ns) Site 1: CMP 1 30 100 CMP 2 30 100 Site 2: CMP 1 30 100 Site 3: CMP 1 30 100 Site 4: CMP 1 30 100 CMP 2 30 100 CMP 3 30 100 Table 4: CMP Sounding processing parameters. GPS Acquisition GPS data were acquired with a Leica GPS sy stem (GS20) to establish locations of stratigraphic sections, CMPs, and GPR profiles. For stratigraphic sections, GPS data were recorded on the present day surface, di rectly overlying the section. Both UTM coordinates and elevation data were recorded. For CMPs, GPS data were recorded only at the CMP center point. For GPR profiles, GPS data were acquired with the intent of spatially delineating the profile and record ing elevation changes along the profile. Both GPR profile and GPS acquisitions were perf ormed along a survey tape; therefore, GPS positions, GPR trace numbers, and relative prof ile distances as indicated on the survey tape were recorded. The frequency of G PS points along profiles varies from one recording every 8 to 40 meters, where profil es acquired in areas of larger topographic variability often have larger numbers of G PS recordings. Table 5 presents the frequency of GPS recordings with respect to distance and GPR traces.

PAGE 47

37 Site:Line Number of GPS Recordings Average Distance Increment (m) Average Trace Increment (GPR traces) Site 1: Line 1 24 12 38 Line 2 18 16 43 Line 3 5 16 13 Line 4 4 21 50 Site 2: Line 1 8 34 12 Line 2 8 33 135 Line 3 2 34 152 Line 4 6 26 112 Line 5 2 32 152 Line 6 4 14 48 Line 7 4 10 36 Site 3: Line 1 12 16 86 Line 2 5 8 47 Site 4: Line 1 11 40 201 Line 2 17 26 302 Line 3 3 6 70 Line 4 2 16 188 Line 5 2 13 146 Line 6 2 25 253 Line 7 2 15 163 Line 8 2 25 299 Table 5: GPS measurement frequencies. The num ber of GPS recordings along GPR profiles, average distance increment (m), average trace in crement (GPR traces). For those profiles where only two points were recorded (s tart and end), the values represent the length of the line and the number of traces per line.

PAGE 48

38 Chapter 6 Results Study Site A site map displaying the locations of 13 stratigraphic sections, four GPR sites, and GPR site CMP soundings and profiles is pr esented in Figure 17. Refer to Figures 13A and 13-B for individual GPR profile labels. Line A-A’ indicates the location of a fence diagram that will be discussed below. Stratigraphy Up to nine individual stratigraphic units are exposed in the field area. From the surface downward, these units consist of a so il unit, and upper surge, an upper tephra fallout, the LARI, a surge package (Surge P ackage I), an accretiona ry lapilli unit, a second surge package (Surge Package II), a combination of tephra fallout units, and a paleosol. In some parts of the field area, the upper surge and the upper tephra fallout are reworked. This sequence is visible near its en tirety at only select stratigraphic sections (Figs. 18-20), typically in areas of high reli ef relative to other pa rts of the study site (Sections 13106, 20104, and 20105).

PAGE 49

39 Figure 17: Study area with stratigraphic section la bels, GPR Site labels, and fence diagram (AA’). Symbols are consistent with Figure 12.

PAGE 50

40 Figure 18: Soil, Upper Surge, Upper Tephra Fall out, and the LARI. The photo was taken in the vicinity of Section 13106.

PAGE 51

41 Figure 19: LARI, Lower Surge Package I, Accretio nary Lapilli, and Lower Surge Package II. The photo was taken in the vicinity of Sec tion 13106. Notice the white pen for scale. Figure 20: Lower Surge Package II, Tephra Fall out Deposits, and Paleosol. The photo was taken in the vicinity of Section 13106.

PAGE 52

42 Select measurements from thirteen stratigraphic sections are presented in chronological order (Table 6). Refer to A ppendix A for complete stratigraphic sections. LARI thickness correlations between select stratigraphic sections and the closest GPR profile position are presented in Table 7. Stratigraphic Section LARI Package Thickness (m) Depth to Upper LARI (m) Upper Tephra Fallout Thickness (m) Upper Surge Thickness (m) Surface Elevation (msl) 13101 1.40 0.301.12 0.12-0.13 0-0.81 2726 13102 3.19 1.481.56 0.12-0.13 0.74-0.80 2718 13103 6.10 1.65 Reworked 0.40 Reworked 0.40 2700 13104 8.54 0.051.17 Reworked 0-0.47 Reworked 0-0.47 2680 13105 4.12 2.48 Reworked 1.38 Reworked 1.38 2685 13106 1.85 1.061.27 0.10-0.14 0.46-0.53 2733 20101 3.30 2.253.25 0.40-0.75 1.10 2712 20102 3.45 2.73.00 0.30 1.20-1.35 2688 20103 2.68 0.680.81 0.17 0.06-0.07 2704 20104 1.77 0.711.13 0.18 0.20-0.40 2794 20105 0.94 0.660.84 0.16-0.19 0.20-0.35 2747 20201 2.28 0.550.60 0.10-0.15 Eroded 2796 20202 1.13 0.380.82 0.12-0.18 0.16-0.30 2775 Table 6: Stratigraphic section measurements. The measurements are most pertinent to GPR interpretations. GPS surface elevations are also provided.

PAGE 53

43 Stratigraphic Section GPR Profile Closest GPR Profile Position (x) (m) Distance between position x and stratigraphic section (m) LARI thickness: via section (m) LARI thickness: via profile (m) LARI profile variation as % of strat. measurement* 13101 Site 2: Line 1 1 62 1.73 1.40 -19 13101 Site 3: Line 1 105 74 1.94 1.40 -28 13102 Site 2: Line 1 88 20 1.35 3.19 +136 13103 Site 2: Line 1 168 43 1.78 6.10 +242 13103 Site 2: Line 5 31 30 6.10 3.92 -35 13104 Site 2: Line 1 260 38 3.65 8.54 +134 13105 Site 2: Line 1 260 71 3.65 4.12 +13 13106 Site 1: Line 1 0 21 2.06 1.85 -11 13106 Site 1: Line 2 0 31 1.59 1.85 +16 20101 Site 4: Line 1 9 37 1.84 3.30 +79 Table 7: LARI thickness correlations. The co rrelations are between stratigraphic section measurements and GPR profile measurements at the profile position that is nearest in distance to the stratigraphic section; where %: profile th ickness < stratigraphic measurement; + %: profile thickness > stratigraphic measurement. Lithology The goal of this study was to acquire GPR and stratigraphic thickness data, with a focus on the low-aspect ratio ignimbrite. W ith this aim in mind, the inevitable time constraints that arose duri ng the field session prevented the recording of detailed lithologic information at each section. Lithologic observations were noted for only a few stratigraphic sections. Stratigraphic observat ions predominantly consisted of thickness

PAGE 54

44 measurements. Therefore, lithologic inform ation will be presented for Sections 13101 and 13106. Photographs of the respective stratig raphic sections are presented in Figures 18-20 (Section 13106) and Figur e 21 (Section 13101). Stratigraphic Section 13101 Soil : 0.18 m. Upper Surge : Eroded, but present within 5 m horizontally (0.81 m thick). Intercalated pumice layers, lithic rich, occasional pumices up to 11 cm and lithics to 5 cm. Sparse pumice bands ranging from 2-4 cm with intercalated fine-grained bands from 3-5 mm. Top of unit grades into soil. Base of upper su rge has a fine-grained white layer that is 0.7 cm thick, with grains less than 1 mm (low porosity). Upper Tephra Fallout : 0.12-0.13 m. Average pumice be tween 3-5 cm. Lithics are 20-25% in abundance. Maximum pumice of 13 cm, maximum lithic of 10 cm. Fine-grained uppermost portion (less than 1 cm thick), marking clear contact between Upper Tephra Fall and Upper Surge. LARI : 1.4 m Upper 0.70 m: Fine-grained porti on of the LARI w ith laterally pinching pumice bands. Light beige to pink in color. Pumices within bands vary from 4-5 cm. Matrix pumices and lithics from 1-5 mm. Maximum lithic up to 3 cm and maximum pumice up to 2 cm.

PAGE 55

45 Middle 0.60 m: 90% pumice with rare lithics, occasionally up to 21 cm. Thicker pumice bands than upper portion. Pumices are commonly from 6-9 cm within bands. The unit is finergrained toward the base. Basal 0.10 m: Lithic-rich, 60-65% lithics. Maximum lithic up to 9 cm. Poorly sorted. Surge Package I : 2.8 m + ? Upper 1.3 m: Uppermost portion of this 1.3 m surge is very fine-grained, varying from 9-20 cm in thickness, with grains less than 1 mm. The bulk of the lower surge contains rare pumice fragments up to 0.5 cm, and is cross-bedded with grain size in beds unifo rm. Most beds are 1-2 mm. Occasional pumices up to 3 cm can be found within these beds. The base of this unit is finely laminated with sub-mm to 3 mm pumices and sub-mm to 1 mm lithics. Lower 1.5 m + ?: Contact between uppe r and lower portion forms a bench, indicating a change in comp etency. The upper portion of this unit contains very fine laminations with parallel contacts. The laminations are seen everywhere in outcrop and are 4-22 mm in thickne ss with sub-mm grains. Lamellae thicknesses increase downward. Below this finegrained portion, the unit becomes massive, with pumice

PAGE 56

46 fragments up to 1 cm, and a 1-2 mm sandy matrix. The base of this unit is c overed by slope debris. Figure 21: Stratigraphic Section 13101. The Upper Surge is eroded here. The LARI is not pink as is typical throughout the field area.

PAGE 57

47 Stratigraphic Section 13106 Soil : 0.50-0.63 m. Base of soil consis ts of 50% pumice and lithics. Upper Surge : 0.46-0.53 m. Cross-bedded. Upper Tephra Fallout : 0.10-0.14 m. Pumices from 1220 cm. Lithics to 20 cm. LARI : 1.85 m Upper 0.27 m: Fine-grained with less pronounced pumice trains than in lower portions of the unit. Beige in color. Pumices up to 1 cm, lithics up to 5 mm Middle 1.38 m: Highly oxidized (str iking pink) with discontinuous pumice trains up to 20 cm thick and 9 cm long. Middle portion contains 5-10% lithics, with lithics up to 3 cm. Pumices are up to 8 cm. Basal 0.20 m: Lithic-rich layer. Domi nantly dacitic, angular clasts, poorly-sorted, fines-depleted. Lithics are up to 10 cm. Surge Package I : 1.50 m. Fine-grained, cross-bedded, with lenses from 1-3 mm. Upslope Thinning The fence diagram (A-A’) as indicated on Figure 17 is presented below (Fig. 22) with accompanying stratigraphic sections. Th e associated stratigraphic sections are presented with a legend and depth labels in Figu re 23. It is clear that the LARI is thinning as the deposit climbs topography.

PAGE 58

48 Figure 22: Fence Diagram A-A’. The fence (a bove) displays only the LARI. The associated stratigrapic sections and LARI thicknesses are di splayed below the fence. The LARI is the pink unit in both the fence and the stratigraphi c sections. Stratigraphic sections are scaled to each other only, not to the fence diagram.

PAGE 59

49 Figure 23: Stratigraphic sections and associated legend relevant to fence diagram A-A’. Numbers on the left side of the sections indicate the depth below ground surface.

PAGE 60

50 GPR Analysis GPR analyses consisted of LARI identif ication and delineation within profiles, depth determinations to upper and lower LARI-bounding reflectors, LARI thickness calculations, and LARI velocity determinati ons via CMP soundings in order to convert two-way travel times into depths. GPR data are presented for profiles that allowed for subsequent velocity analyses, the majority of which were acquired in an upslope direction. Remaining GPR data are presented in Appendices B and C. GPR profiles were analyzed with respect to stratigraphic section measurements and deposit behavior (particu larly stratified upper surge beha vior) (Fig. 24) to aid in LARI identification and delineations. In terpretations are largely derived from stratigraphic sections located within 30 mete rs of the closest GPR profile. Reflector geometry and behavior also contributed to interpretations, largely a result of “wavy” reflectors indicative of surgelike features (Fig. 25). Figure 24: The stratified nature of the Upper Surg e. The Upper Surge is about 0.5 m thick in the center of the photo. The LARI is the pink unit below.

PAGE 61

51 Figure 25: GPR Profile (Site 2_Line 2) displaying surge-like (“wavy”) reflectors. The red lines represent the upper and lower LARI-bounding reflectors (see text). The horizontal reflectors at the top of the profile are air and/or direct wave arrivals. 50 ns is equivalent to roughly 3.15 m. VE 4x. Upper and lower LARI-bounding reflectors we re identified within the GPR record (e.g. Fig. 25). Reflector interpretations were then extended to profiles lacking stratigraphic control. This capability was a f unction of the intersecting profile traverse paths, which exist at all sites, and allowe d for profile to profile comparisons. A usercontrolled MATLAB function was implemente d to trace the LARI-bounding reflectors (Appendix D). This was a visual process whic h required user interpretation and numerous point selections along an individual delineation. A MA TLAB assigned interpolation combined with the user selected points result ed in output of two-way travel times as a

PAGE 62

52 function of UTM coordinates, the latter of which was tran slated into distance along the profile, yielding radargrams w ith highlighted LARI-bounding re flector delineations (Figs. 26-35). Depth values for the upper and lower LARI-bounding reflector delineations were derived from a relationship between two-way tr avel time, LARI velocity (via CMPs), and transmitter and receiver antenna spacing, using the equation 2 /2 2 2a v t d (2) where, d depth (m), t 2-way travel time (ns), v LARI velocity (m/ns), and a antenna spacing (m). The identification of upper and lower LARIbounding reflectors allowed for LARI delineations in the major ity of GPR profiles. The upper-LARI contact was not associated with a di stinctive GPR reflection. The n earest bright GPR return is interpreted to be stratigra phically above the LARI, therefore the upper LARI-bounding reflector does not represent the LARIs upper su rface, and instead represents the contact between the Upper Tephra Fallout and the Upper Surge. This interpretation derives from reflector behavior that is i ndicative of pyroclastic surge bedforms (Fig. 25). The lower LARI-bounding reflector is in terpreted as the contact between the LARI and the underlying unit (Surge Package I), largely deri ved from stratigraphi c correlations. The implication is that the delineat ions are not entirely represen tative of the LARI, a function of the upper interface. To acc ount for the discrepancy arising from the difficulty of delineating the upper-LARI interface, a GPR s ite thickness is assumed for the Upper Tephra Fallout (based on stratigraphic meas urements) (Table 8). The site-assumed Upper Tephra Fallout thickness was then deducted from upper LARI-bounding reflector depths to determine an estimated depth to the upper-L ARI contact. This appr oach is believed to

PAGE 63

53 be valid because the Upper Tephra Fallout ma intains relatively consistent thicknesses over the survey area. The corrected upper-de lineation depth values and the original lower-delineation depth values were used to cr eate Elevation vs. Distance curves for each of the profiles by topographica lly correcting the depth valu es relative to GPS derived elevation values (Figures 26-35). LARI thicknesses along each GPR profile were calculated by taking the difference between corrected upper-delineati on depth values and the uncorrected lower-delineation depth valu es, yielding Thickness vs. Distance curves with a superimposed thickness trend line (F igures 26-35). GPR profiles without LARIbounding reflector delineations ar e presented in Appendix F. GPR Site Upper Tephra Fallout Deduction (m) Site 1 0.13 Site 2 0.13 Site 3 0.13 Site 4 0.30 Table 8: GPR Site thickness deductions. Upper Te phra Fall thickness deductions used to correct upper LARI-bounding reflector depth values.

PAGE 64

54 Figure 26: Site 1: Line 1 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 65

55 Figure 27: Site 1: Line 2 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 66

56 Figure 28: Site 2: Line 1 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 67

57 Figure 29: Site 2: Line 2 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 68

58 Figure 30: Site 2: Line 4 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 69

59 Figure 31: Site 2: Line 5 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 70

60 Figure 32: Site 3: Line 1 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 71

61 Figure 33: Site 4: Line 1 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 72

62 Figure 34: Site 4: Line 2 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 73

63 Figure 35: Site 4: Line 4 GPR profile with LARI delineation (top), topographically corrected LARI (middle), and thickness vs. distance curve with trend line (bottom).

PAGE 74

64 CMP Soundings CMP soundings were used to constrain GPR site LARI velocities, allow for final travel-time to depth conversions, and determ ine LARI thicknesses (as described above). Velocity analyses were run after the LAR I-bounding reflectors were identifie d within GPR profiles; this was necessary because without first delinea ting the LARI-bounding reflectors, reflectors in the CMP radargrams are stratigraphically meaningless. The CMPs were compared to the near est position on the nearest GP R profile, which allowed for identification of the LARI-bounding reflect ors within associated CMPs. The LARIbounding reflectors were used to obtain veloci ties using the Dix Method. This process resulted in two interval velocities per CMP. The first interval velocity represents a velocity for units above the upper LARI-boundi ng reflector, while the second interval velocity ideally represents the velocity of the LARI at the time of that particular CMP sounding (Table 9). In cases where more th an one CMP were performed per GPR site, the interval velocities representing the LARI were averaged, yielding a site-averaged LARI velocity. CMPs with reflector delin eations are presented in Appendix E.

PAGE 75

65 Site: CMP VInterval (m/ns) Interval Thickness (m) Site Averaged LARI Velocity (m/ns) Site1: 0.123 CMP10.141 2.26 0.123 2.70 CMP20.121 1.31 0.122 2.68 Site2: 0.128 CMP10.116 1.48 0.128 2.45 Site3: 0.108 CMP10.136 2.86 0.108 1.94 Site4: 0.110 CMP10.117 1.93 0.116 4.73 CMP20.116 2.55 0.111 1.91 CMP30.108 2.49 0.103 2.41 Table 9: CMP velocities. The bolded VInterval velocities represent individual CMP derived LARI velocities, while the bolded Site Averaged LARI Velocities represent the velocities used in final time-depth conversions for each of the four GPR sites. Velocity Analyses Bulk density-current and vertical (c ross-sectional) velocity analyses are performed to examine upslope LARI parent density-current flow using a simplified energy-based theoretical an alysis, modeled after a tu rbidity current (Muck and Underwood, 1990). Muck and Underwood (1990) examined upslope flow of unconfined turbidity currents onto a ba thymetric high by analyzing the exchange of kinetic energy for potential energy and frictional heat. The turbidity current begi ns with a mass flow down a landward trench slope a nd ends with the turbidity current passing through a break

PAGE 76

66 in slope and ascending upslope, the latter stage which is anal yzed with respect to energy losses. C1 : Center of current gravity at base of barrier (m) C2 : Center of current gravity at peak of upslope flow (m) y : C2-C1 (m); vertical change in center of gravity U : Turbidity current velocity at C1 (m/s) : Current density (kg/m3) : Density contrast between current and ambient fluid (kg/m3) g: Force of gravity (m/s2) Eloss: Fraction of kinetic energy lost as heat during upslope flow (%) Figure 36: Upslope turbidity current flow schema tic. The kinetic energy equation and variables are displayed in the figure. (modified from Muck and Underwood, 1990). Bulk Density-Current Velocity The upslope flow analysis developed by Muck and Underwood (1990) is extended to the LARI to obtain 10 bulk flow velocities of the LARI-producing pyroclastic density-current (Fi g. 37). Parameter values are lis ted in the caption, and were held constant throughout the analysis.

PAGE 77

67 ) 1 ( / 2lossE g E U U : Minimum bulk density-current velocity required at E1 to reach E2 (m/s) E : E2-E1; Change in paleo-topographical su rface elevation from base to peak of upslope flow (m) E1 : Pre-LARI surface elevation at base of barrier (m) E2 : Pre-LARI surface elevation at peak of upslope flow (m) : Density contrast between LARI-producing pyroclastic density-current and air: 999.04 kg/m3 : Density of LARI-producing current: 1000 kg/m3 g : Force of gravity: 9.8 m/s2 Eloss: Fraction of kinetic energy lost as heat during upslope flow: 10% Figure 37: Bulk density-current velocity schem atic. The associated kinetic energy equation, variable descriptions, and parameter values are also displayed. Basal and upslope peak elevation values of the pre-LARI depositional surface are used as a supplement to center of current gravities as indicated by Muck and Underwood (1990) (Fig. 36). This elevation supplement maintains dimensional continuity. It is assumed that the basal LARI contact, as deli neated in the GPR profiles, represents the surface of the land during pyroclastic densit y-current ascent and LARI deposition. This assumption is valid unless there is evidence of erosion. Basal LAR I-contact delineations yield two-way travel time a nd depth equivalents along the pr ofile (as described above).

PAGE 78

68 By linearly interpolating between GPS derive d surface elevations and later deducting the calculated basal LARI depths, elevation va lues along the basal contact are calculated. Bulk velocity analysis is not performed fo r all GPR profiles because the LARI does not in all cases flow upslope, and not all GPR profiles were acquired near parallel to the direction of inferred flow. The calculated veloc ity is interpreted as the minimum velocity required at the approach of upslope flow for the LARI-producing pyroclastic densitycurrent to reach its maximum observed elev ation (as determined via GPR and GPS).

PAGE 79

69 Figure 38: Bulk density-current veloc ities. Velocities of the LARI producing pyroclastic density-current are a function of paleo-topographical elevation change. Box colors indicate the GPR site and assoc iated numbers indicate the GPR profile.

PAGE 80

70 Vertical (cross-sectional) Velocity Site 3: Profile 1 is analyzed to determ ine vertical velocity profiles at chosen positions along the flow. This analysis is performed with the assumption that internal GPR reflectors, which start at the base of the deposit and ultimately impinge on the basal LARI-contact, are equivalent to internal st reamlines. Fluid-dynamically, a streamline is defined as a continuous line th at has the property that its ta ngent at each point coincides with the direction of the fluid velocity at th at point (Furbish, 1997), or simply the path a parcel follows in a given fluid. This analys is is based on the premise that the LARIproducing density-current is de nsity stratified. Valentine (19 87) addressed th e interaction of an internally stra tified pyroclastic surge with topogr aphy, where in a stratified flow, it is stated that there will be a level (streamlin e) above which all fluid has sufficient energy to top an obstacle and below which all fluid is either stopped (blocked) or simply moves around the obstacle with no upward motion. Valen tine (1987) referred to this level as a critical level or a dividi ng streamline (Fig. 50). Figure 39: Dividing streamline schematic. Blocki ng in a density-stratified pyroclastic current occurs below the dividing streamline as it encounters a hill. Below the streamline material cannot flow over the obstacle due to a lack of kinetic energy (modified from Valentine, 1987).

PAGE 81

71 Two internal streamlines (GPR reflectors) were delineated with a user-controlled Matlab function. The Matlab out put of two-way travel times was converted into depths, and these depths were deducted from GPS de rived surface elevation values to determine streamline elevations along the profile. The third streamline is taken to be the LARIs upper surface, as evident in the GPR profile. If only one streamline is present (Fig. 51), the change in elevation of the streamline is equivalent to the minimum velocity for the streamline parcel to travel from the base to the peak of upslope streamline elevation. At the upslope peak of the streamline (SE1b in fi g. 51), the velocity is taken to be zero, as the parcel no longer has kinetic energy and is deposited. In the Site 3: Line 1 analysis, three streamlines are analyzed, and elevati on derived velocities are calculated in segments (Fig.52).

PAGE 82

72 ) 1 ( / 2 1lossE g SE U U : Minimum velocity of density-current at the b ase of the streamline required to reach the peak of the streamline (m/s) SE1a : Streamline elevation at base of upslope flow (m) SE1b : Peak streamline elevation within flow (m) SE1 : SE1b-SE1a; Change in streamline elevation from base to peak of upslope flow (m) : Density of LARI-producing current: 1000 kg/m3 : Density contrast between LARI-produci ng pyroclastic density-current and air: 999.04 kg/m3 g : Force of gravity: 9.8 m/s2 Eloss: Fraction of kinetic energy lost as heat during upslope flow: 10% Figure 40: Vertical velocity profile schematic with associated kinetic energy equation, variable descriptions, and parameter values.

PAGE 83

73 Figure 41: Site 3: Line 1 streamline delineations and velocity profiles. This GPR profile is topographically corrected between 97 and 197 meters distance (x), where black linesLARI contacts, S1-streamline 1, S2-streamline 2, S3-s treamline 3 (upper contact) (above). The vertical blue lines indicate chosen positions for vertical ve locity analysis, and associated numbers indicate the corresponding x-value. Velocity profiles at the corresponding x-position are shown below, where black squaresstreamline position within flow.

PAGE 84

74 Chapter 7 Discussion Pyroclastic density-current end-member s are largely differentiated by sediment concentration and flow regime. Pyroclastic flows are highly sediment-concentrated, traveling in a predominantly laminar flow st ate, while pyroclastic surges are dilute, traveling in a predominantly turbul ent flow state (Wilson and Houghton, 2000). Eyewitness accounts of historical eruptions and pyroclastic deposit studies have led to the understanding and classification of density-current end-members. Eyewitness accounts at Mt. Pelee, Mart inique, Mt. Unzen, Japan, and Soufriere Hills, Montserrat, have elucidated the behavior of pyroclastic density-currents during transport. Early accounts of pyroclastic de nsity-current transport were summarized by Fisher and Heiken (1982): Flows at Mt. Pelee reportedly trav eled along topographic depressions at great velociti es in a highly-concentrated state (Lacroix, 1904), while surges flowed in a less concentrated, expa nded, and turbulent stat e, irrespective of topography (Anderson and Flett, 1903; and L acroix, 1904). Later eyewitness accounts at Mt. Unzen (Fisher, 1995) and Soufriere Hill s (Druitt et al., 2002) elaborate on similar density-current behaviors. Pyroclastic deposit studies have contri buted to the understanding of densitycurrent transport. Field observ ations and grain-size analyses of Quaternary ignimbrites from Italy and the Azores revealed that about 90% of any single ignimbrite is relatively homogeneous (massive), fine-grained, and poor ly sorted (Sparks et al., 1973; Sparks,

PAGE 85

75 1976). Unlike flow deposits, pyroclastic surge deposits tend to show cross-bedding, wavy or planar laminations, and/or dune-like structures lack fine particles, and are better sorted (Sparks et al., 1973; Sparks, 1976; Fisher 1979; Fisher and Schmincke, 1984). The combination of eyewitness accounts and pyr oclastic deposit studies led to the characterization of density-currents with re spect to particle co ncentration and flow regime, and provided the framework for schema tic cross-sectional de nsity and velocity profiles (Fig. 2). The vertical velocity gradient analysis (Figs. 52-55) quantitatively shows that a LARI parent density-current has a surge-like vertical velocity pr ofile (Fig. 2). This interpretation depends on the assumption that internal GPR reflector s are equivalent to flow streamlines. The streamline GPR reflector s (Fig. 52) are clearl y traced until the position of impingement on the paleo-topogr aphical surface (basal-LARI contact). Relationships between the streamlines and both the underlying paleo-surface and the LARI highly resemble the dividing streamlin e schematic presented by Valentine (1987) (Fig. 50). The bulk flow velocity analyses provid e the minimum velocity of the LARI producing density-current at the base of upslope flow, while the lateral velocity analysis details subsequent losses in current velocity as the flow ascends upslope. This interpretation depends on two assumptions, firs t, that the basal LARI contact represents the paleo-topographical surface, and second, that the thickness of the LARI is equivalent to the thickness of the LARI producing density-current. Acco rding to the bulk and lateral velocity models, as flow thickness expands, minimum velocities increase; therefore, bulk flow velocity and lateral velocity values represent absolute minimums, as the observed

PAGE 86

76 LARI thickness can only be equivalent to the minimum thickness of the parent densitycurrent, a function of mass balance. Stratigraphic measurements reveal that the LARI thins upslope. The majority of LARI delineations within GPR profiles show a similar thinning trend. Exceptions to the upslope thinning trend may be a result of partial LARI erosion by the Upper Surge deposit or incomplete GPR delineations, a function of the upper LARI-bounding reflector being skewed by air and/or direct wave arrivals. LARI identification within GPR profiles is a function of reflector behavior and stratigraphic correlations; with respect to the former, “wavy” reflectors are seen overlying the upper-LARI contact (Fig. 25), follow ed by a transition into less pronounced reflectors. These “wavy” reflectors are i ndicative of surge-like bedforms, and often abruptly transition downward into an area of co ntrasting reflector beha vior, interpreted as the LARI (readily apparent in figs. 28 and 29). These shallow reflector behaviors nicely correlate with the stratigraphic sequence. Ba sal-LARI contact interp retations were guided by stratigraphic measurements.

PAGE 87

77 Chapter 8 Conclusions The LARI is identified in GPR profiles via reflector behavi or and stratigraphic measurements. The maximum variation in LARI thic kness in a direction of increasing topographic relief is 7.8 m, while the maximum elevation change is 34.6 m. Minimum LARI-parent pyroclastic density-c urrent velocities at the base of upslope flow (based off kinetic energy lo ss as the flow ascends upslope) were at least 25 m/s. Vertical velocity analyses fo r Site 3: Line 1 indicate a surge-like velocity profile for the parent pyroclastic density-curre nt, with current velocities decreasing downward through the flow. At the base of upslope flow, the maximum current velocity is 24 m/s, while at the peak of upslope flow, the maximum current velocity is 2.4 m/s.

PAGE 88

78 The LARI-producing density-current lik ely has cross-sectional density and velocity characteristics similar to that of a pyroclastic flow, with a dividing streamline separating an overriding cu rrent from the basal current. Figure 42: LARI-producing density current with cross-sectional profiles Below the dividing streamline, the bulk of th e flow is density-stratified, resulting in internal flow streamlines. Below the in ternal streamline, pa rticles are deposited due to kinetic energy losses. Multiple internal streamlines represent multiple depositional regimes. Figure 43: Basal LARI-producing density-current with cross-sectional profiles.

PAGE 89

79 Chapter 9 Recommendations Concerning future aspects of the stu dy, I would like to recommend detailed analyses of internal structure within th e LARI, with the aim of relating deposit characteristics to radar reflections. This will be very helpful in the vicinity of the Site 3_Line 1 profile in which the vertical veloc ity analysis is based. Similarly, with other GPR profiles, internal radar characteristics need to be correlated with stratigraphic exposures. The Site 2_Line 1 profile will be most useful as it is directly adjacent to the erosional gully where Stratigraphic Secti ons 13101 to 13105 were acquired. Preferably, detailed measurements of deposit structur es will be made, with a concentration on variations in grain size/density (pumice tr ains, micaceous alignments, etc.) and the depth to these variations below the upper-LARI contact. It will also be of use to acquire strati graphic measurements in the SSW portion of the study area, with the aim of characteri zing the upslope extent of the LARI. An understanding of deposit variat ions between the upslope a nd downslope portions of the study area will elaborate on topographically indu ced behavioral variations of the densitycurrent.

PAGE 90

80 References Anderson, T. and J.S. Flett, 1903. Report on the eruption of the Soufri ere in St. Vincent in 1902 and on a visit to Mont agne Pelee in Martinique. Philosophical Transactions of the Royal Society of London, Ser. A, v. 200, pp. 353-553. Andrade, D. 2002, Estudio Geovolcanologico de l Complejo Volcanico del Pululahua. Tesis de ingeniero de la Escuel a Politecnica Nacional; Quito, Ecuador; 180 p. Andade, D. 2006, Pululahua Field Guide: Cities on Volcanoes 4. Barberi, F.; Coltelli, M.; Ferrara, G.; Innocen ti, F.; Navarro J.M.; and R. Santacroce. 1988. Plio-Quaternary volcanism in Ecuador. Geological Magazine. v. 125, n. 1, pp. 1-14. Cagnoli, B. and J.K. Russel. 2000, Imaging the subsurface stratigraphy in the Ubehebe hydrovolcanic field (Death Valley, Californi a) using ground penetrating radar. Journal of Volcanology and Geothermal Research. v. 96, pp. 45-56. Cagnoli, B. and T.J. Ulrych. 2001, Ground penetrating radar images of unexposed climbing dune-forms in the Ubehebe hydrovolcanic field (Death Valley, California). Journal of Volcanology and Geothermal Research. v. 109, pp. 279298. Dade, W.B. 2003, The emplacement of low-aspect ratio ignimbrites by turbulent parent flows. Journal of Geophysical Research v. 108, n. B4, pp. 1-9. Davis, J.L. and A.P. Annan. 1989, Ground-pene trating radar for high-resolution mapping of soil and rock stratigraphy. Geophysical Prospecting. v. 37, pp. 531-551. Druitt, T.H. 1998, Pyroclastic density currents. In: Gilbert, J.S. and Sparks, R.S.J (eds) The Physics of Explosive Volcanic Eruptions Geological Society, London, Special Publications, v. 145, pp. 145-182. Druitt, T.H.; Calder, E.S.; Cole, P.D.; Hoblitt, R.P.; Loughlin, S.C.; Norton, G.E.; Ritchie, L.J.; Sparks, R.S.J.; and B. Voight. 2002, Small-volume, highly mobile pyroclastic flows formed by rapid sedime ntation from pyroclastic surges at Soufriere Hills Volcano, M ontserrat: an important volcanic hazard. in The Eruption of Soufriere Hills Volc ano, Montserrat, from 1995 to 1999 Geological Society Memoirs, no. 21. The Geol ogical Society of London, pp. 263-279.

PAGE 91

81 Fisher, R.V. 1979, Models for pyroclast ic surges and pyroclastic flows. Journal of Volcanology and Geothermal Research. v. 6, pp. 305-318. Fisher, R.V. 1993, Flow transformations in sediment gravity flows. Geology. v. 11, pp. 273-274. Fisher, R.V. 1995, Decoupling of pyroclastic currents: hazard assessments. Journal of Volcanology and Geothermal Research v. 66, pp. 257-263. Fisher, R.V. and G. Heiken. 1982, Mt. Pelee, Martinique: May 8 and 20, 1902, Pyroclastic Flows and Surges. Journal of Volcanology and Geothermal Research v. 13, pp. 339-371. Fisher, R.V. and Schmincke, H.-U. 1984, Pyroclastic Rocks Springer-Verlag Berlin Heidelberg. pp. 89-230. Fisher, R.V.; Orsi, G.; Ort, M.; and G. Heiken. 1993, Mobility of a large-volume pyroclastic flowemplacement of the Campanian ignimbrite, Italy. Journal of Volcanology and Geothermal Research. v. 56, pp. 205-220. Furbish, D.J. 1997, Fluid Physics in Geol ogy: An Introduction to Fluid Motions on Earth’s Surface and Within Its Crust Oxford University Press, Inc., pp 167-176. Gilbert, J.S.; Stasiuk, M.V.; Lane, S.J.; Adam C.R.; Murphy, M.D.; Sparks, R.S.J.; and J.A. Naranjo. 1996, Non-explosive, constr uctional evolution of the ice-filled caldera at Volcan Sollipulli, Chile. Bulletin of Volcanology. v. 58, pp. 67-83. Lacroix, A., 1904. La Montagne Pelee et ses Eruptions. Masson et Cie, Paris, 662 pp. Muck, M.T. and M.B. Underwood. 1990, Upslope flow of turbidity currents: A comparison among field observations, theory, and laboratory models. Geology v. 18, pp. 54-57. Papale, P. and M. Rosi. 1993, A case of no-wi nd plinian fallout at Pululagua caldera (Ecuador): implications for m odels of clast dispersal. Bulletin of Volcanology. v. 55, pp. 523-535. Peterson, D.W. and R.I. Tilling. 2000, La va Flow Hazards. in Encyclopedia of Volcanoes Academic Press, pp. 957-971. Reynolds, J.M. 1997, An Introduction to Applied and Environmental Geophysics John Wiley and Sons, pp. 681-749.

PAGE 92

82 Russell, J.K. and M.V. Stasiuk. 1997, Character ization of volcanic de posits with groundpenetrating radar. Bulletin of Volcanology. v. 58, pp. 515-527. Rust, A.C. and J.K. Russell. 2000, Mapping poros ity variation in a welded pyroclastic deposit with signal and velocity patterns from ground-penetrating radar surveys. Bulletin of Volcanology. v. 62, pp. 457-463. Rust, A.C. and J.K. Russell. 2001, Detection of welding in pyroclastic flows with ground penetrating radar: insights from field and forward modeling data. Journal of Volcanology and Geothermal Research. v. 95, pp. 23-34. Sparks, R.S.J., 1976. Grain size variations in ignimbrites and implications for the transport of pyroclastic flows. Sedimentology v. 23, pp. 147-188. Sparks, R.S.J., Self, S. and G.P.L. Walke r. 1973. Products of ignimbrite eruptions. Geology v. 1, pp. 115-118. Valentine, G.A. 1987, Stratified flow in pyroclastic surges. Bulletin of Volcanology. v. 49, pp. 616-630. Valentine, G.A. and R.V. Fisher. 1993, Glow ing Avalanches: New Research on Volcanic Density Currents. Science, v. 259, pp. 1130-1131. Walker, G.P.L.; Heming, R.F.; and C.J.N. Wilson. 1980, Low-apsect ratio ignimbrites. Nature. v. 283, pp. 286-287. Walker, G.P.L.; Wilson, C.J.N.; and P.C. Froggatt. 1981, An ignimbrite veneer deposit: The trail-marker of a pyroclastic flow. Journal of Volcanology and Geothermal Research v. 9, pp. 409-421. Walker, G.P.L. 1983, Ignimbrite types and ignimbrite problems. Journal of Volcanology and Geothermal Research. v. 17, pp. 65-88. Wilson, C.J.N. and B.F. Houghton, Pyrocl astic Transport and Deposition. in Encyclopedia of Volcanoes Academic Press, pp. 957-971. Wilson, C.J.N. and G.P.L. Walker. 1985, The Taupo Eruptions, New Zealand I. General Aspects. Philosophical Transactions of the Ro yal Society of London. Series A, Mathematical and Physical Sciences. v. 314, n. 1529, pp. 199-228. Wilson, C.J.N. 1985, The Taupo Eruption, New Zealand II. The Taupo Ignimbrite. Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences. v. 314, n. 1529, pp. 229-310.

PAGE 93

83 Volentik, A.C.M.; Bondadonna, C.; Connor, C.B.; Rosi, M.; Ruiz, G.; and L.J. Connor. A Study of the 2450 BP Pululagua Plinian Er uption (Ecuador): Implications for Models of Tephra Dispersal. “Ame rican Geophysical Union, Fall Meeting, December 5-9, 2005. Volentik, A.C.M.; Bondadonna, C.; Connor, C.B. ; Rosi, M.; Ruiz, G.; and L.J. Connor. Total grain-size distribution and volu me of the 2450 BP Pululagua Plinian Eruption. “Cities on Volcanoes 4, January 23-27, 2006.

PAGE 94

84 Appendices

PAGE 95

85 Appendix A: Stratigraphi c Sections and Legend A-1: Stratigraphic Section 13101 and stratigraphic legend.

PAGE 96

86 Appendix A (Continued) A-2: Stratigraphic Sections 13102 (left) and 13103 (right)

PAGE 97

87 Appendix A (Continued) A-3: Stratigraphic Sections 13104 (left) and 13105 (right)

PAGE 98

88 Appendix A (Continued) A-4: Stratigraphic Sections 13106 (left) and 20101 (right)

PAGE 99

89 Appendix A (Continued) A-5: Stratigraphic Sections 20102 (left) and 20103 (right)

PAGE 100

90 Appendix A (Continued) A-6: Stratigraphic Sections 20104 (left) and 20105 (right)

PAGE 101

91 Appendix A (Continued) A-7: Stratigraphic Sections 20201 (left) and 20202 (right).

PAGE 102

92 Appendix B: LARI Delineations with Thickness vs. Distance Curves B-1: Site 1_Line 3 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 103

93 Appendix B (Continued) B-2: Site 1_Line 4 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 104

94 Appendix B (Continued) B-3: Site 2_Line 6 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 105

95 Appendix B (Continued) B-4: Site 2_Line 7 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 106

96 Appendix B (Continued) B-5: Site 4_Line 3 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 107

97 Appendix B (Continued) B-6: Site 4_Line 5 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 108

98 Appendix B (Continued) B-7: Site 4_Line 6 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 109

99 Appendix B (Continued) B-8: Site 4_Line 7 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 110

100 Appendix B (Continued) B-9: Site 4_Line 8 LARI delineation and thickn ess vs. distance curve with thickness trend line.

PAGE 111

101 Appendix C: Topographically Corrected LARIs C-1: Site 1_Line 3 depth corrected LARI. From left to right, NW to SE C-2: Site 1_Line 4 depth corrected LARI. From left to right, NW to SE

PAGE 112

102 Appendix C (Continued) C-3: Site 2_Line 6 depth corrected LARI. From left to right, NW to SE C-4: Site 2_Line 7 depth corrected LARI. From left to right, NW to SE

PAGE 113

103 Appendix C (Continued) C-5: Site 4_Line 3 depth corrected LARI. From left to right, W to E C-6: Site 4_Line 5 depth corrected LARI. From left to right, NW to SE

PAGE 114

104 Appendix C (Continued) C-7: Site 4_Line 6 depth corrected LARI. From left to right, W to E C-8: Site 4_Line 7 depth corrected LARI. From left to right, NW to SE

PAGE 115

105 Appendix C (Continued) C-9: Site 4_Line 8 depth corrected LARI. From left to right, W to E

PAGE 116

106 Appendix D: LARI-Bounding Reflector Matlab Code function LayerPicksSimpleUTM % reads in a single .mat format file % no gains applied % lets user pick arrival times of reflections % writes out reflection arrival times % this version allows no timeshifting or trend removal % this version only saves picks at positions between first and last pick % S Kruse Oct 06 % adds in UTM positions of points along line % assumes format of worksheet is (no header) % column 1 trace, column 2 cum distance, % column 3 easting, column 4 northing % S Kruse Nov 06 clear all ; close all ; %***********INPUT SECTION STARTS HERE**************************************** filein = 'PUS2L2Mg.mat' fileUTM = 'PUS2L2M_UTM.xls' sheetUTM= 'traceUTM' ; % name of worksheet in file UTM fileoutstem = 'PUS2L2M_Lower2Ign' ; irev = 1; %set to 1 if line is reversed %The suffix LP indicates layer pick to avoid confusion in the future %SET maximum time to show on screen for making time picks in ns tshow = 550; %ns, if greater than total time in record, has no effect %***********INPUT SECTION ENDS HERE**************************************** %build output file name tfile=[fileoutstem 'times.txt' ] % output times %read in data load(filein, 'A' 'x' 't' ); [nt nx] = size(A); %Pick returns and compute velocities [x,layert] = LayerTimePickSimple(tshow,filein); %layert

PAGE 117

107 Appendix D (Continued) %interpolate UTM coordinates of points along lines U=xlsread(fileUTM, sheetUTM); [nobs ncol] = size(U) for i=1:nx iobsright = 1; while (U(iobsright,2) <= x(i) && iobsright < nobs) iobsright = iobsright + 1; end iobsleft = iobsright 1; E(i) = U(iobsleft,3) + (U(iobsright,3)-U(iobsleft,3)) ... *(x(i)-U(iobsleft,2))/(U(iobsright,2)-U(iobsleft,2)); N(i) = U(iobsleft,4) + (U(iobsright,4)-U(iobsleft,4)) ... *(x(i)-U(iobsleft,2))/(U(iobsright,2)-U(iobsleft,2)); end size(E) if (irev == 1) E = fliplr(E); N = fliplr(N); end %write out layer times TimeOutUTM(x,layert,E,N,tfile);

PAGE 118

108 Appendix E: CMP LARI-Boundi ng Reflector Delineations E-1: Site 1_CMP 1 LARI-bounding reflector delineations E-2: Site 1_CMP 2 LARI-bounding reflector delineations

PAGE 119

109 Appendix E (Continued) E-3: Site 2_CMP 1 LARI-bounding reflector delineations Site 3: CMP 1 E-4: Site 3_CMP 1 LARI-bounding reflector delineations

PAGE 120

110 Appendix E (Continued) E-5: Site 4_CMP 1 LARI-bounding reflector delineations E-6: Site 4_CMP 2 LARI-bounding reflector delineations

PAGE 121

111 Appendix E (Continued) E-7: Site 4_CMP 3 LARI-bounding reflector delineations

PAGE 122

112 Appendix F: GPR Profiles F-1: Site 1_Line 1 GPR profile F-2: Site 1_Line 2 GPR profile

PAGE 123

113 Appendix F (Continued) F-3: Site 1_Line 3 GPR profile F-4: Site 1_Line 4 GPR profile

PAGE 124

114 Appendix F (Continued) F-5: Site 2_Line 1 GPR profile F-6: Site 2_Line 2 GPR profile

PAGE 125

115 Appendix F (Continued) F-7: Site 2_Line 3 GPR profile F-8: Site 2_Line 4 GPR profile

PAGE 126

116 Appendix F (Continued) F-9: Site 2_Line 5 GPR profile F-10: Site 2_Line 6 GPR profile

PAGE 127

117 Appendix F (Continued) F-11: Site 2_Line 7 GPR profile F-12: Site 3_Line 1 GPR profile

PAGE 128

118 Appendix F (Continued) F-13: Site 3_Line 2 GPR profile F-14: Site 4_Line 1 GPR profile

PAGE 129

119 Appendix F (Continued) F-15: Site 4_Line 2 GPR profile F-16: Site 4_Line 3 GPR profile

PAGE 130

120 Appendix F (Continued) F-17: Site 4_Line 4 GPR profile F-18: Site 4_Line 5 GPR profile

PAGE 131

121 Appendix F (Continued) F-19: Site 4_Line 6 GPR profile F-20: Site 4_Line 7 GPR profile

PAGE 132

122 Appendix F (Continued) F-21: Site 4_Line 8 GPR profile