Pit Craters Throughout the Solar System

ABSTRACT

KLING, CORBIN LYLE. Pit Craters throughout the Solar System. (Under the direction of Dr. Paul K. Byrne).

Pit craters are common features found on planetary bodies in the Solar System. Pit craters are circular to elliptical in plan view shape, and extraterrestrial examples usually have inverted conical shapes. Typically, pit craters are found in groups that are often aligned in chains, and often parallel to bounding normal faults. The frequency with which pit craters and normal faults are seen together has led to dilational faulting emerging as a leading mechanism for how pit craters form. Other possibilities for pit crater formation include lava tube collapse and phreatomagmatic eruptions above a dike tip. The occurrence of these features on many different worlds, from Earth to Mars to Venus makes understanding the factors that govern their formation important. But efforts to decipher how pits on other planetary surfaces form is often limited by image or topographic resolution of the available data. On Earth, we can investigate these landforms in the field and examine them in situ, allowing for detailed interpretations of how pits form and develop. This dissertation has three science chapters that look at pit crater assemblages in different geological environments, specifically on Earth and Mars, with additional data taken from the literature of pits on other planetary bodies. Together, these chapters offer new insight into the mechanisms of formation, morphological styles, and development of this abundant planetary landform type. Chapter 2 is a focused study on the contribution of pit craters to the complex topography at Noctis Labyrinthus, Mars, a large physiographic province east of the volcanic Tharsis rise and west of the Valles Marineris canyon system. The observations made at

Noctis Labyrinthus lead to the interpretation that volatiles played a substantial role in the structural evolution of this region of Mars. Chapter 3 analyzes pit craters at several locations across the Solar System, underpinned by fieldwork at two sites on Earth: Craters of the Moon National Monument and Preserve in Idaho and Hawaii Volcanoes National Park on the big island of Hawai'i. The fieldwork in Idaho and Hawai'i provided a useful basis for understanding how pit craters can form from multiple different processes, either tectonic or volcanic, and evolve to similar morphologies. Chapter 4 focuses on a single set of pits within The Grabens region of

Canyonlands National Park in Utah. Of the four sites studied in Chapter 4, three (Sites 1, 3, and

4) are situated along fractures (either seen in the field or interpreted from field or topographic data), implying that the formation of these pits at least has been heavily influenced by the tectonics in the region. Additionally, numerous features attributed to surficial water flow were noted near sinkholes in Sites 2, 3, Site 4, and which I interpret as secondary erosional features further contributing to the morphology of these features. This work shows that pits on Earth are much smaller than their extraterrestrial counterparts and that they can develop because of multiple processes. The resultant pit shape appears at least in part correlated with those formation mechanisms. Pits with cylindrical shapes are found in basaltic flows, which probably indicates that the basaltic flows in the walls are strong enough to maintain vertical or near- vertical walls for extended periods. Yet pits on other planetary bodies are often shaped like inverted cones, and this finding can mean one of two things: 1) that either the pits originally formed in unconsolidated material and always assumed an inverted conical shape; and/or 2) that the pits initially formed as cylinders, either in competent material or less cohesive material and subsequent, secondary processes then led the pit walls to taking on an inverted conical shape.

by Corbin Lyle Kling

A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Marine, Earth, and Atmospheric Sciences

Raleigh, North Carolina 2020

APPROVED BY:

______Paul Kevin Byrne Danielle Wyrick Committee Chair External Member

______Karl W. Wegmann Helena Mitasova

______Delwayne Bohnenstiehl ii

DEDICATION

To Alli, Mom, Dad, and Tyler. I would not be here today without all of your love and support.

“So, it’s kind of like the quest for the holy grail. Well, you know, who gives a shit what the holy grail is. It’s the quest is what’s important.” –Yvon Chouinard

iii

BIOGRAPHY

Corbin received his Bachelors of Science in Geology from the University of Georgia

(UGA) in 2014. He continued into the Masters of Science program at UGA, and finished in

2016. He joined the Marine, Earth, and Atmospheric Sciences (MEAS) Department at NC State for his Ph.D. in the fall of 2016. During his time at NC State, Corbin has participated in many extracurricular events in the department in addition to his Ph.D. coursework and research, including a year as the President of the Graduate Student Association MEAS Chapter, the NC

State Graduate School Writing Certificate, and the Preparing the Professoriate Program.

Following the completion of his Ph.D., Corbin will be joining the Center for Earth and Planetary

Studies at the National Air and Space Museum as a Smithsonian Postdoctoral Fellow at the conclusion of his Ph.D.

ACKNOWLEDGMENTS

First I would like to acknowledge the funding sources that supported my Ph.D. throughout my time at NC State. The NC State Graduate School Graduate Student Support Plan provided funding the first year of my Ph.D. The NASA Earth and Space Science Fellowship provided funding for the final three years of my Ph.D. work (grant #: 80NSSC17K0491). Two

Geological Society of America Graduate Student Grants provided funding for the Craters of the

Moon National Monument and Canyonlands National Park fieldwork. Fieldwork was completed with the assistance of multiple colleagues. Field photos of Hawaiian pit craters and locations of previously unpublished pit craters around Kilauea were generously provided by D. Wyrick prior to my first visit to those sites. I was able to visit all the pit sites in Hawaii Volcanoes National

Park for the first time due to the generosity of Dr. Bob Craddock from the Smithsonian, who provided housing accommodations near the park in April 2018. Allison Vo, Julian Chesnutt, and

Dr. Paul Byrne provided field assistance in Craters of the Moon National Monument and

Preserve in the Summer of 2018. Zach Williams provided field assistance during the

Canyonlands National Park fieldwork in October 2019. Finally, I would like to thank my committee for their unwavering support throughout my time at NC State.

TABLE OF CONTENTS

LIST OF TABLES ...... viii LIST OF FIGURES ...... ix Chapter 1: Introduction 1.1. Pit Craters...... 1 1.2. Science Chapters ...... 2 1.3. Key Findings ...... 5

Chapter 2: Tectonic Deformation and Volatile Loss in the Formation of Noctis Labyrinthus, Mars ...... 10 Plain Language Summary ...... 11 Abstract ...... 11 1. Introduction ...... 12 1.1. Noctis Labyrinthus ...... 12 1.2. Faulting within Noctis Labyrinthus ...... 15 1.3. Pit Craters...... 17 2. Methods...... 19 2.1. Datasets ...... 19 2.2. Mapping ...... 20 2.3. Topographic Analysis ...... 21 2.3.1. Fault Displacement Profile Analysis ...... 21 2.3.2. Pit Crater and Trough Analysis...... 22 3. Results ...... 23 3.1. Fault Populations ...... 23 3.1.1. Population 1 Results ...... 24 3.1.2. Population 2 Results ...... 24 3.1.3. Population 3 Results ...... 24 3.2. Pit Craters...... 25 3.3. Troughs ...... 26 4. Discussion ...... 27 4.1. Normal Faulting ...... 27 4.2. Pit Craters...... 28 4.3. Other Geomorphological Evidence for a Volatile Rich Subsurface ...... 31 4.4. Total Volume Loss in Noctis Labyrinthus ...... 32 5. Conclusions ...... 33

Chapter 3: Pit Crater Formation in the Solar System: Case Studies from Earth on the Snake River Plain, Idaho, and Kilauea Caldera, Hawaii 1. Introduction ...... 50 1.1. Pit Craters on Earth ...... 52 1.2. Craters of the Moon National Monument Study Sites ...... 54 1.3. Hawaii Volcanoes National Park Study Sites ...... 56 1.4. Scientific Rational ...... 57 2. Methods...... 58 2.1. Data ...... 58

2.2. Field Methods ...... 60 2.2.1. Uncrewed Aerial System(UASs) ...... 60 2.2.2. Processing of UAS and Handheld Photogrammetric Data ...... 61 2.3. Geospatial Methods ...... 62 2.3.1. Mapping of Tectonics and Volcanics ...... 62 2.3.2. Geospatial Analysis ...... 63 3. Results ...... 64 3.1. Field Work Results ...... 64 3.1.1. Craters of the Moon National Monument Field Areas ...... 64 3.1.1.1. King’s Bowl ...... 64 3.1.1.2. Yellowjacket Waterhole...... 65 3.1.1.3. Coyote Lake ...... 66 3.1.1.4. Coltrell’s Blowout ...... 67 3.1.2. Hawaii Volcanoes National Park Field Areas ...... 67 3.1.2.1. Kilauea Iki ...... 67 3.1.2.2. Devil’s Throat ...... 69 3.1.2.3. Twin Pits ...... 69 3.2. Remote Observations of Venus ...... 70 3.3. Other Data Collected from Literature ...... 72 3.3.1. Asteroids and Small Moons ...... 72 3.3.2. Enceladus ...... 74 4. Discussion ...... 74 4.1. Craters of the Moon ...... 74 4.1.1. King’s Bowl ...... 74 4.1.2. Yellowjacket Waterhole ...... 76 4.1.3. Coyote Lake and Coltrell’s Blowout ...... 77 4.2. Hawaii Volcanoes ...... 78 4.2.1. Kilauea Iki ...... 78 4.2.2. Devil’s Throat and Twin Pits ...... 79 4.3. Planetary Pits ...... 81 4.3.1. Noctis Labyrinthus, Mars...... 81 4.3.2. Nyx Mons, Venus ...... 81 4.3.3. Pits from the Literature ...... 82 4.4. Shapes ...... 83 4.5. Statistical Analysis of Available Pit Crater Depth/Diameter Data ...... 86 5. Conclusions ...... 88

Chapter 4: Tectonic Sinkholes and the Fate of Water in the Grabens region of Canyonlands National Park, Utah 1. Introduction ...... 115 1.1. Background ...... 115 1.2. Science Rational...... 118 2. Methods...... 118 2.1. Field Work ...... 118 2.2. GIS Analysis ...... 119

vii

2.3. Fault Analysis ...... 120 2.4. Channel Analysis Methods ...... 122 3. Results ...... 123 3.1. Field Work ...... 123 3.1.1. Site 1 ...... 123 3.1.2. Site 2 ...... 124 3.1.3. Site 3 ...... 124 3.1.4. Site 4 ...... 125 3.2. GIS Analysis ...... 125 3.3. Fault Analysis ...... 126 3.4. Channel Analysis ...... 127 4. Discussion ...... 129 4.1. Sinkholes ...... 129 4.2. Channels ...... 130 4.3. Faults ...... 134 5. Conclusions ...... 137

Chapter 5: Conclusions 5.1. Overview ...... 159 5.2. Discussion and Scientific Importance ...... 160 5.3 Future Work ...... 163

References ...... 165

Appendix A: Table of Collated Pit Crater Data ...... 175

Appendix B: Noctis Labyrinthus Pit Crater Measurements ...... 190

viii

LIST OF TABLES

Table 2.1 Pit Crater and Trough Summary ...... 47

Table 3.1 Collated Rock Strength Properties ...... 93

Table 4.1 Sinkhole Locations ...... 131

Table 4.2 Sinkhole Geometric Properties ...... 143

Table 4.3 Channel Reach Normalized Channel Steepness ...... 158

LIST OF FIGURES

Chapter 1

Figure 1.1 Examples of pit craters on different planetary bodies in the Solar System. A) Noctis Labyrinthus, Mars (centered at 14.88°S, 95.69°W), the focus of Chapter 2; B) Yellowjacket Waterhole (centered at 43.49°N, 113.51°W) in Craters of the Moon National Monument and Preserve, a field site from Chapter 3; and C) pit crater chains on the flanks of Nyx Mons, Venus (centered at 30.05°N, 49.83°E), which are further discussed in Chapter 3...... 6

Figure 1.2 Physiographic map of Noctis Labyrinthus, Mars made with the HRSC–MOLA blended digital elevation model (DEM) overlaid on the THEMIS Daytime IR global image mosaic...... 7

Figure 1.3 A) Photograph of the northeastern pit of the Twin Pits in the Ka’u Desert on the Big Island of Hawaii. B) Aerial photograph from uncrewed aerial systems operations of King’s Bowl in Craters of the Moon National Monument and Preserve in Idaho...... 8

Figure 1.4 Photograph of a pit crater just south of the The Grabens region in Canyonlands National Park, Utah...... 9

Chapter 2

Figure 2.1 A) A global map of Mars, showing the MOLA–HRSC digital elevation model overlaid on a MOLA–HRSC hillshade; the location of Noctis Labyrinthus is marked by a black box. The map is displayed in Robinson projection, centered at 0°E. B) A regional map of Noctis Labyrinthus; prominent, named features are labelled. Elevation data are from the MOLA–HRSC dataset overlaid on the THEMIS daytime IR global mosaic. The map is in equirectangular projection, centered at 9°S, 101°W...... 35

Figure 2.2 Box models illustrating different generalized methods of pit crater formation. a) Extensional tectonics leading to graben formation and void space at depth (i.e., dilational normal faulting). b) Dike intrusion with Plinian-style eruption at the tip, creating void space and subsequent collapse. c) Cryosphere interaction with pit craters formed from tectonics or diking. Figure adapted from Wyrick et al. (2004)...... 36

Figure 2.3 A portion of the southeast of Noctis Labyrinthus, to illustrate the spatial interplay between normal faults, pit craters, and troughs. The location of this figure is marked on Figure 1b as a black rectangle. This map is in equirectangular projection, centered at 14.3S, 96.9°W...... 37

Figure 2.4 Mapping results for Noctis Labyrinthus shown on top of the THEMIS day IR global mosaic. All mapped features from this study are shown. Additionally, landslide locations (classified by H/L values) from Crosta et al. (2018a) are shown. The map is in equirectangular projection, centered at 9°S, 101°W...... 38

Figure 2.5 The orientations of faults (mint green), pit craters (green blue), and troughs (orange) we mapped in this study. Orientations were measured with respect to north for all structures, resulting in half-rose diagrams...... 39

Figure 2.6 Displacement profiles for the 22 graben-bounding faults analyzed in this study. Faults were classified based on strike. a) Data for all faults studied; b) Population 1 faults (n = 8); c) Population 2 Faults (n = 4); and d) Population 3 faults (n = 10)...... 40

Figure 2.7 A depth–diameter plot of all pit craters located within the Noctis Labyrinthus region (n = 215). A linear regression line is shown in black, with 95% confidence interval of fit shown in grey. The fit of this line is r2 = 0.88, indicating a strong correlation between pit crater depth and diameter...... 41

Figure 2.8 A plot of pit crater depth–diameter values for pits in this study (from Noctis Labyrinthus), as well as for pit data from Hawaii and Nyx Mons, Venus pit data collected by the authors. Additional pits from the literature for Mars are from Wyrick et al. (2004) and for Earth are from Whitten and Martin (2019)...... 42

Figure 2.9 Histograms of depth/diameter ratios and size-frequency diagrams of depth and diameter for Noctis Labyrinthus pits and Martian pits from Wyrick et al. (2004). Probability density functions overlaid on histograms for comparison, and Noctis probability density function overlaid on Mars to show difference between Noctis pits and the rest of Mars. Negative exponential functions plotted on top of size-frequency diagrams to show tendency of pit shapes to follow that type of distribution...... 43

Figure 2.10 Examples of periglacial signatures in Noctis Labyrinthus, Mars. a) Ice wedge polygons and solifluction lobes, centered at 7.4°S, 96.3°W. b) Thermokarstic terrain identified in Rodriguez et al. (2016), centered at 6.9°S, 99.0°W. Both maps show CTX imagery in equirectangular projection at the same view scale. .... 44

Figure 2.11 A depth–diameter plot of pit craters and rampart craters in the Noctis Labyrinthus region. Crater depths were calculated using scaling ratios from Robbins et al. (2011) and Reiss et al. (2005), to bracket the different minimum depth possibilities for a notionally present cryosphere here. The minimum depth to volatiles as calculated from the rampart crater excavation depth (assumed to be 33% of the crater depth: Melosh (1996)), and maximum depth to tensile fracturing on Mars, are shown with black and brown dashed lines, respectively...... 45

Figure 2.12 An example of a landslide within Noctis Labyrinthus that has a long runout and H/L ratio that indicates potential fluidized movement. Several smaller landslides are evident immediately south of the large slide, illustrating the prevalence of mass wasting in Noctis Labyrinthus. The map is displayed in equirectangular projection, centered at 9.86°S, 95.16°W...... 46

Figure 2.13 Schematic cartoon for the formation sequence of Noctis Labyrinthus we propose in this study. a) Prior to extension, a volatile-rich stratigraphy is present. b) Extensional structures begin to form. c) Dilational faulting from material strength differences leads to pit crater formation within the graben. d) Pit craters and/or faults breach a volatile-rich layer(s) and expose it to the atmosphere, leading to sublimation or melting. e) Subsequent landsliding and other mass-wasting processes contribute to the present form of Noctis Labyrinthus...... 48

Chapter 3

Figure 3.1 A global map showing locations of several accessible pit crater assemblages on Earth. A) The Big Island of Hawaii, where pit crater formation has been documented since the mid-1900s on the flanks of the Kilauea caldera. B) The U.S. state of Idaho, where the Craters of the Moon National Monument and Preserve is located, a site of recent basaltic volcanism (~2000 years ago). C) Utah, another U.S. state, where pit craters have been documented within graben in Canyonlands National Park. D) Iceland, situated on the Mid Atlantic Ridge, where pits formed during a rifting event in the 1970s in the deltaic plain near Asbyrgi Canyon...... 92

Figure 3.2 An Agisoft Metashape™ Processing workflow for uncrewed aerial systems data to produce orthomosaic images and digital surface models...... 94

Figure 3.3 A formation diagram from Hughes et al. (2018) for King's Bowl where the structure is interpreted as a phreatomagmatic explosion pit. Ground cracking in relation to fissure width and depth is also illustrated...... 95

Figure 3.4 Lava recession textures noted in the bottom of King’s Bowl, Idaho at the northern cleft. Orange lines trace the recession streaks, handheld radio for scale...... 96

Figure 3.5 Mapping results from King's Bowl in Craters of the Moon National Monument and Preserve overlaid on the orthomosaic generated from UAS data. Fractures are shown in green and pit craters and depressions in teal. Darker areas are from cloud shadows present moving through the scene during the UAS flights. .... 97

Figure 3.6 Yellowjacket waterhole UAS DSM with pits (black) and fractures (blue) mapped...... 98

xii

Figure 3.7 Field photos from Yellowjacket Waterhole illustrating the important features identified in the field. A) A view into a pit crater, comprising unconsolidated material (probably tephra), beneath more competent lava flows. The rubble in the pit is <1m in size, and is made up of blocks of those overlying lava flows. B) A view of three pit craters in the chain looking north, the tees near pits are 2–4 m tall. C) A view of the rightmost pit crater in panel B, with a cinder cone in the background; the trees in view here are 2–4 m tall. D) A view looking south (with the pits from panel B behind the camera view) showing fractures that parallel the pit chain...... 99

Figure 3.8 Coyote lake UAS-derived DSM (left) and orthomosaic (right) used for analysis and interpretation...... 100

Figure 3.9 Field photos of the Coyote Lake feature. A) A view looking south from the northern edge of the pit crater; the rocks in the foreground are 10–30 cm across. B) A view of thin lava flows in walls of pit, flows are 15–45 meters away from camera location. C) A view looking west from the pit crater floor up at blocky rubble along the western margin. Some blocks are made up of more than one flow; the blocks in the foreground are 75–150 cm across. D) A photo showing an old lava flow on the flank of Coyote Lake...... 101

Figure 3.10 Coltrell’s Blowout UAS DSM (left) and orthomosaic (right), used for analysis and interpretation...... 102

Figure 3.11 The topography of the southeast corner of the Kilauea Caldera. Inset shows location of the field photos in Figure 11 and the maps shown in Figure 12. Earthquakes between June and July 2018 are plotted based on their magnitude (color) and depth to rupture (circle diameter)...... 103

Figure 3.12 Handheld photos of pits contained within the Kilauea Iki pit crater chain showing important findings. A) The northern end of the chain looking south; pits here are 1–2 m across. B) The view looking north of a large pit containing two smaller pits , which are 50 cm to 1 m across. C) A view looking south showing a pit with an overhang after Kilauea eruption, P. Byrne for scale (~1.8 m tall). D) A view looking north of a vegetated pit with distinct margins; the Kilauea Iki cinder cone visible behind (background center). E) A view of a brand-new pit that formed after the Kilauea 2018 eruption, looking south towards the other pits (Photo courtesy of Dr. Bob Craddock, Smithsonian Institution.) F) A close-up view of the inner WNW wall of this new pit, seen in October 2018. Note the stratigraphy in the pit walls, and the overhang to the right...... 104

Figure 3.13 Topographic maps of the Kilauea Iki pit crater chain, from June (left) and July (center) 2018. The difference between the two surveys is shown in the right panel, illustrating the formation of the new pit crater on the northern (top) end of the chain (red outline)...... 105

xiii

Figure 3.14 Handheld photos and a 3D model of Devil’s Throat pit crater on Hawaii. A) A view looking north–northeast illustrating the many stacked lava flows that make up the walls; the trees on the edge of the pit are 1.5–2.5 m tall. B) A view of the talus at the bottom of pit; rubble pieces are up to 1 m across. C) The view showing fracturing on the southern edge of pit. This fracture is 5–10 cm across. D) An additional view of the pit, looking to the ESE. E) A top-down view of the 3D model, showing camera locations determined in Agisoft Metashape. Note the shape of the pit both at the top (plan view) and within the pit crater...... 106

Figure 3.15 Handheld images and a 3D model of the western pit in the Twin Pits group located in the Ka’u Desert of Hawaii Volcanoes National Park. A) A view looking north (with Mauna Loa in the background), with fracturing along the edge of the pit; these fractures are 1–10 cm across. B) A view to the south, showing stacked lava flows within the pit wall, a younger flow cascading into pit, and talus within the pit that is 50 cm or less in size. C) View looking east showing eastern pit and cinder cone in background; Cinder cone is ~300 m away. D) Top down view of 3d model. E) Side view of the 3d model looking to the northeast, illustrating the stacked talus on the left. F) Side view of the pit, looking to the southeast, illustrating the irregular shape...... 107

Figure 3.16 Handheld images and a 3D model of the eastern pit in the Twin Pits group in the Ka’u Desert of Hawaii Volcanoes National Park. A) The view to the northwest showing the western pit, with Mauna Loa in the background. B) Looking east, with cinder cone ~150 m away. C) A view looking southwest that shows the stacked flows in the uppermost pit walls; here, the wall is ~50 m across. D) The view into the eastern pit, showing lava flows draping over the sides of the pit, as well as what appears to be an exposed conduit leading out of the pit. The conduit opening is ~10–20 m across. E) A top-down view of the 3D model developed for this pit, looking to the northeast; the conduit in the bottom of the pit is visible. F) A side view of the 3D model, showing the mostly cylindrical shape of the pit...... 108

Figure 3.17 A) A regional map of Nyx Mons, Venus, showing the pit craters I mapped (n = 312). The radar look direction is from the left. B) An inset map of one pit crater in the region, showing the radar shadow inside of one pit I used for depth estimation. C) A schematic diagram showing how the shadow measurements are used. See text for details...... 109

Figure 3.18 Histograms of available pit crater data depth/diameter ratios. Earth pits are from Whitten & Martin (2019), Mars pits are from Wyrick et al. (2004), Noctis pit and Venus pit data collected in this study. Probability density functions shown for each population (and Noctis overlaid on Mars for comparison). Mean depth/diameter ratio and uncertainty shown for each histogram...... 110

xiv

Figure 3.19 Results of an analysis of variance multicompare test. Blue points (Whitten & Martin (2019) and Wyrick et al. (2004)) indicate populations that are similar, while red points (Noctis Labyrinthus, Mars and Nyx Mons, Venus; this study) are populations with no similarity to the other three...... 111

Figure 3.20 Diameter Frequency plot for each group of pits analyzed. Negative exponential functions for each group displayed in red showing the tendency of pit crater diameter to be explained by negative exponential functions...... 112

Figure 3.21 Depth frequency plot for each group of pits analyzed. Negative exponential functions for each group displayed showing the tendency of pit crater depth to be explained by negative exponential functions like the depths...... 113

Chapter 4

Figure 4.1 Regional maps of the Grabens Region of Canyonlands National Park. A: National Elevation Database 5 m DEM; B: National Agricultural Imagery Program 2018 San Juan County Mosaic with cross sectional lines and grey boxes denoting field site figure locations; C: Structure map showing faults and sinkholes; and D: Regional watersheds and channels derived from the NED 5 meter DEM...... 140

Figure 4.2 Top: a) GPR profile by Kettermann et al. (2015) across Devil’s Lane in the northern Grabens region, b) interpreted GPR profile showing an abundance of shallow subsurface faults (black dashed lines) within alluvium, and possibly within Cedar Mesa Sandstone (which occupies the lower 3–4 m of the profiles) where reflectors are less prominent. Blue lines are the interpreted stratigraphy of the alluvial deposits. Bottom: NAIP image with GPR profile and locations of the shallowest near-surface faults marked (white dashed lines) from analysis by Kettermann et al. (2015)...... 142

Figure 4.3 Cartoon showing a) mixed-mode fracturing creating dilational fault segments at the surface and b) how a weak lubricating layer could detach additional blocks from the fault plane causing further collapse structures to form. From Kettermann et al. (2015)...... 143

Figure 4.4 Schematic illustration showing the relationship between graben-bounding faults, intra-graben faults, sinkholes, and stratigraphy in the northern Grabens Region, modified from Kettermann et al. (2015)...... 143

Figure 4.5 Site 1 regional structure map (centered at 38.03° N, 109.96° W) with sinkholes, normal faults (inferred and known) and ephemeral channels mapped. Base map is a UAS-derived DSM overlaid on the corresponding hillshade. A–A’ and A’’–A’’’ cross section lines shown are in Figure 6. Gaps are due to missed UAS photo captures during the mapping flight plan...... 145

Figure 4.6 Cross sections A–A′ and A′′–A′′′ (shown in Figure 1B) covering the sinkholes identified at Site 1. No vertical exaggeration...... 146

Figure 4.7 A) Photograph looking to the southeast, showing the central sinkhole at Site 1. B) Cartoon depiction of that same area showing the rock units and fault slip directions...... 146

Figure 4.8 Site 2 regional structure map (centered at 38.04° N, 109.96° W) with sinkholes, faults (inferred and known), and ephemeral channels mapped. Base map is a UAS-derived DSM overlaid on the corresponding hillshade...... 147

Figure 4.9 A) Photograph showing sinkholes at Site 2. B) Cartoon depiction of the same view showing rock units and fault slip directions...... 147

Figure 4.10 Photograph showing the channel intersection with an open joint surface at Site 2. The approximate location where this photograph was taken is shown on Figure 8...... 148

Figure 4.11 Sites 3 and 4 regional structure map (centered at 38.05° N, 109.95° W) with sinkholes, faults (inferred and known), and ephemeral channels mapped. Locations of C’’–C’’’, D–D’, and D’’–D’’’ cross sections shown in Figure 12 also shown. Base map is a UAS-derived DSM overlaid on the corresponding hillshade...... 149

Figure 4.12 Cross sections C–C′ and C′′–C′′′ (shown in Figure 1B) covering the sinkholes identified at Site 3. No vertical exaggeration...... 149

Figure 4.13 A) Photograph showing the western sinkhole at Site 3. B) Cartoon depiction of the same scene with rock units and nearby faults denoted...... 150

Figure 4.14 Temporal sequence of NAIP imagery from 2006 to 2018, as well as 2019 UAS orthomosaic collected in this study. Key pit outline and channel head locations from 2006 are plotted on each image for reference, and any changes are shown with red polygons or dots...... 150

Figure 4.15 A) Photograph looking to the southwest showing fluvial dendritic depression at Site 4. B) Cartoon depiction of the same view showing rock units labeled and fault slip direction...... 151

Figure 4.16 Cross sections D–D′ and D′′–D′′′ (shown in Figure 1B) covering the sinkholes identified at Site 4. No vertical exaggeration...... 151 Figure 4.17 Trunk-channel longitudinal profiles for the four catchments in The Grabens region (A–D), one unfaulted channel (Bull Wash) just outside of this region for comparison (E), and all longitudinal profiles normalized (F). Colors correspond to those used in Figure 1D...... 152

xvi

Figure 4.18 InSAR maps from Furuya et al. (2007). a) Regional mean yearly vertical displacement (shown in mm/yr), with displacement away from the spacecraft shown as positive. Faults are overlaid in white mapped originally by Mcgill & Stromquist, 1979; b) hillshade showing faults and cross-sectional lines, from Furuya et al. (2007). The location of Cross Canyon and Imperial Valley is shown with the white box...... 153

Figure 4.19 E–E’ Cross Section from Figure 17, which intersects Cross Canyon, from Furuya et al. (2007). The location of Site 3, Cross Canyon, is shown in the purple band...... 154

Figure 4.20 Elevation versus Chi plots for each of the study channels in The Grabens region (from top to bottom, Cross Canyon, Butler Wash, and Y Canyon) and Bull Wash for comparison at the bottom. Each reach selected for chi analysis is shown in red dashed lines, and the fit and R2 shown for reference. Knickpoints are marked in blue triangles, which correspond to those locations on Figure 2. Reference m/n for each catchment is 0.45...... 155

Figure 4.21 Chi (left) and normalized channel steepness (ksn) (right) of channels displayed on top of slope shade maps (slope map with transparency over a grayscale DEM) for The Grabens and surrounding region. Knickpoints marked with blue triangles. Chi map has catchment divide general outlines shown with colored polygons (colors correspond to Figure 1D). Black arrows on Chi map indicate anticipated drainage divide movement based on comparisons. Low near drainage divides indicates erosion and expansion of the drainage network at the expense of the channels on the other side...... 156

Figure 4.22 Knickpoint elevations above the Colorado River for each catchment in the study area (colored dots which correspond to Figure 1D). Elevation above Colorado river is based on the elevation of the Butler Wash outlet into the Colorado River (1185 m). Lowest knickpoint from each catchment absolute elevations shown with black circles...... 157

CHAPTER 1

Introduction

1.1. Pit Craters

Pit craters are common features found on planetary bodies in the Solar System (Figure

1.1). Pit craters are circular to elliptical in plan view shape, and extraterrestrial examples usually have inverted conical shapes, although occurrences of pit craters with straight walls and more cylindrical shapes have been noted on Earth and Mars (Okubo & Martel, 1998; Okubo, Atypical pits). The pit craters here are similar to ‘sinkholes’ formed through karstic dissolution, but this work does not focus on the karst equivalent of pit craters. Typically, pit craters are found in groups that are often aligned in chains, and often parallel to bounding normal faults. The frequency with which pit craters and normal faults are seen together has led to dilational faulting emerging as a leading mechanism for how pit craters form (Wyrick et al. 2010). Other possibilities for pit crater formation include lava tube collapse and phreatomagmatic eruptions above a dike tip (Hughes et al. 2018); all three mechanisms will be discussed in this dissertation.

The occurrence of these features on many different worlds, from Earth to Mars to Venus

(Figure 1.1), makes understanding the factors that govern their formation important. But efforts to decipher how pits on other planetary surfaces form is often limited by image or topographic resolution of the available data. For bodies such as Earth and Mars, a plethora of remotely sensed data exists that allows for detailed global mapping of these features down to decameter scale.

Additionally, on Earth, we can investigate these landforms in the field and examine them in situ, allowing for detailed interpretations of how pits form and develop. Since their formation is likely due to collapse of material into the subsurface, pits could retain information about the subsurface and help understand the geology of other planetary bodies better.

1.2. Science Chapters

This dissertation has three science chapters that look at pit crater assemblages in different geological environments, specifically on Earth and Mars, with additional data taken from the literature of pits on other planetary bodies. Together, these chapters offer new insight into the mechanisms of formation, morphological styles, and development of this ubiquitous planetary landform type.

Chapter 2 is a focused study on the contribution of pit craters to the complex topography at Noctis Labyrinthus, Mars (Figure 1.2), a large physiographic province east of the volcanic

Tharsis rise and west of the Valles Marineris canyon system. For this chapter, I mapped the pits craters, normal faults, and troughs throughout the entire Noctis region to build a better understanding of the how it came to be so structurally and morphologically complex. I found that the pit craters and fault likely breached a subsurface volatile layer, allowing those volatiles to sublimate to the atmosphere, in turn causing further collapse of the pit craters to from the large troughs that dominate the central portion of Noctis Labyrinthus. Additionally, I explored the possible role of subsurface volatiles in contributing to pit formation and evolution by mapping all the rampart craters—impact craters with clear volatile mobilization in their ejecta blankets

(Melosh 1989)—in the region down to 100 m in diameter, and also analyzed a landslide database

(from Crosta et al., 2018a) for the area. The rampart craters reveal that volatiles may have existed at depths as shallow as ~100 m when impact occurred, and landslide height-to-runout- length ratios point to possible fluidization of the landslides within the troughs at Noctis.

All of these findings lead to my overall interpretation of the formational history of Noctis

Labyrinthus, which started with abundant normal faulting and graben development from regional stresses (possibly tied to Tharsis), which in turn formed pit craters in graben floors because of

3 dilational faulting. Next, pit craters and some faults reached and exposed a stratigraphic layer(s) relatively high in volatiles (possibly pore space-hosted ice or even water), allowing them to sublimate and promoting further collapse, creating the large troughs. Finally, these troughs continued to be eroded, evidenced by large landslides throughout the mapped troughs and by the presence of periglacial features in the bases of many of the troughs. The periglacial features indicate that the bases of these troughs had substantial ice deposits and seasonal freeze–thaw cycling to produce those features. Additionally, the presence of solifluction lobes along the trough walls strongly suggest that there was enough ice within the substrate there to cause downhill slumping of that ice-laden material.

Chapter 3 analyzes pit craters at several locations across the Solar System, underpinned by fieldwork at two sites on Earth: Craters of the Moon National Monument and Preserve in

Idaho and Hawaii Volcanoes National Park on the big island of Hawai’i (Figure 1.3). Craters of the Moon is host to an assortment of pits. King’s Bowl is a pit there that formed by phreatomagmatic eruption (Hughes et al., 2018), but I interpret the nearby crater chain at

Yellowjacket Waterhole as likely related to fracturing. In Hawaii, many pits exist with cylindrical shapes and which have formed within stacked basaltic lava flow packages (Okubo &

Martel, 1998). These pits show evidence for subsurface cavities that once had lava in them, so I attribute these pits to collapse of a previously filled dike tip or lava reservoir. On the flanks of

Kilauea Iki, I identified a new pit crater within cinder deposits that formed as a result of the 2018

Southeast Rift Zone eruption after the Kilauea summit lava lake drained. This new pit is located within a pre-existing crater chain, and is ~1 m wide and deep with straight walls and a cylindrical shape. Monitoring light detection and ranging (lidar) flights during the eruption indicate that this pit formed between July and September 2018. I then obtained earthquake data for this time

4 frame, finding that this pit chain is forming above a caldera-bounding ring fault. To my knowledge, this is the first field-supported detection of a new pit crater having formed from tectonic activity. Finally, I collated all available pit crater data for asteroid, moons, and planets within the Solar System to more fully describe the populations of extraterrestrial pits. This dataset is now the largest morphometric dataset of pit craters collected to date.

In contrast to Chapter 3, my third science chapter focuses on a single set of pits within

The Grabens region of Canyonlands National Park in Utah (Figure 1.4). These pits, arguably more accurately termed “sinkholes,” are found within Quaternary alluvial and aeolian sediments in the graben floors, and most are situated within the middle of the graben. I visited and characterized four field sites just south of the Canyonlands National Park boundary to better understand how these pits form and evolve through time, and to assess the role of groundwater in their development. Evidence found in the field supports the conclusion that these sinkholes are due to intra-graben faulting that may have taken advantage of joints predating the graben. The central pit at one site has Cedar Mesa Formation lithology comprising one wall, but Quaternary alluvium the other, suggesting the presence of a fault at that location. Evidence for extensional fracturing was also found at two additional sites, where open joints were seen at the intersection of small channels and fault surface traces, indicating water and sediment drainage at that location. Analysis of aerial imagery between 2006 and 2019 shows the emergence of one new pit and channel head migration ~100 m upstream. The pits here have redirected the channels towards the fault surface, where the water drains into the subsurface. My analysis indicates that at least 58% of the sinkholes documented in The Grabens region have fluvial channel morphologies associated with them, by which I conclude that the sinkholes in this region are very effective at capturing small subsets of catchments, redirecting considerable volumes of

5 water into the subsurface. This finding has important implications for Colorado River salinity downstream from this location.

1.3. Key Findings

These three projects together lead to some key conclusions. I originally anticipated that the 3D shape of a given pit crater would uniquely correspond to a given formation mechanism, with cylindrical shapes indicating a volcanic origin, perhaps, and an inverted cone corresponding to fault control. I found instead that a pit can have a cylindrical shape even if forming due to faulting, such as the brand new pit that I identified at Kilauea Iki. Pit shape is therefore more likely a factor of time and material properties in which the pit is found, with all pits eventually becoming inverted cones as the slopes reduce to the angle of repose over time. Secondly, a pit crater can form from a multitude of processes, and sometimes those processes might be coeval, in which case it is difficult to discern the original mechanism responsible for pit crater formation.

Third, a fundamental understanding of the geological environment in which the pits are forming will lead to a better understanding of the pits themselves (e.g., the formation and evolution of pits within Noctis Labyrinthus and their interplay with the presence of inferred subsurface volatiles); my observations with remotely sensed data and from fieldwork clearly demonstrate that pits are easily dissected and influenced by secondary erosional processes once formed.

Figure 1.1. Examples of pit craters on different planetary bodies in the Solar System. A) Noctis Labyrinthus, Mars (centered at 14.88°S, 95.69°W), the focus of Chapter 2; B) Yellowjacket Waterhole (centered at 43.49°N, 113.51°W) in Craters of the Moon National Monument and Preserve, a field site from Chapter 3; and C) pit crater chains on the flanks of Nyx Mons, Venus (centered at 30.05°N, 49.83°E), which are further discussed in Chapter 3.

Figure 1.2. Physiographic map of Noctis Labyrinthus, Mars made with the HRSC–MOLA blended digital elevation model (DEM) overlaid on the THEMIS Daytime IR global image mosaic.

Figure 1.3. A) Photograph of the northeastern pit of the Twin Pits in the Ka’u Desert on the Big Island of Hawaii. B) Aerial photograph from uncrewed aerial systems operations of King’s Bowl in Craters of the Moon National Monument and Preserve in Idaho.

Figure 1.4. Photograph of a pit crater just south of the The Grabens region in Canyonlands National Park, Utah.

CHAPTER 2

Tectonic Deformation and Volatile Loss in the Formation of Noctis Labyrinthus, Mars

Chapter 2 was submitted to the Journal of Geophysical Research: Planets, with the following author list:

Corbin L. Kling1; Byrne, P.K.1; Atkins, R.M.1; and Wegmann, K.W.1

1 Marine, Earth, and Atmospheric Sciences Department, North Carolina State University,

Raleigh, NC

*Reviews received from JGR: Planets on August 17, 2020

Plain Language Summary

Noctis Labyrinthus (Lat., ‘Labyrinth of the Night’) on Mars is a topographically complex region that connects Mars’ large volcanic province, Tharsis, and large canyon system, Valles Marineris.

The Noctis Labyrinthus region has many fractures, pits, craters, and troughs, but how these features formed and interacted with one another is still not fully understood. We mapped the fractures, pits, craters, and troughs in this region to better understand the relative formation of

Noctis Labyrinthus. We conclude that extensional stresses led to pit crater formation through the drainage of material into fractures. Further dissection of the landscape was enhanced by an ice- rich subsurface, which was exposed by pit formation and then melted or sublimated to create ever larger depressions. We estimated the depth to this ice-rich subsurface from the distribution and sizes of craters with fluidized ejecta deposits. The bottoms and walls of the troughs show geomorphological evidence for sustained periglacial activity in the form of thermokarst, solifluction lobes, and ice-wedge polygons, adding to the evidence for subsurface ice having played a role in the formation for this distinctive part of Mars.

Abstract

Noctis Labyrinthus is an enigmatic and structurally complex area situated between the Tharsis

Rise and Valles Marineris on Mars. Noctis Labyrinthus is dissected by normal faults that form horst and graben, pit craters that are situated inside the graben, and large troughs that cross-cut the graben and pit craters. At the bases of the troughs there is an abundance of mass wasting and periglacial features, and some evidence for fluvial erosion, suggesting that these troughs hosted liquid and perhaps even ice at some point in the past. The mapping and analysis of these structural and morphological features in Noctis Labyrinthus helps in establishing their role in the

12 regional formation history. Fault displacement profiles, combined with morphometric data from pit craters, is used to assess how the pit craters relate to the much larger troughs in the region, and whether those troughs are likely due to extensional tectonic deformation alone. This comparative analysis suggests that some pit craters grew deeper than the amount of displacement accommodated by their bounding faults. Layers with subsurface volatiles (such as ground ice) were intersected and exposed by the larger Noctis Labyrinthus pit craters, enabling sublimation and melting that further promoted mass wasting and the growth and coalescence of pits and graben into the large troughs observed within this area of Mars. Under this scenario, subsurface volatiles played an important role in the formation of this structurally complex region and may still be present there.

1. Introduction

1.1. Noctis Labyrinthus

Noctis Labyrinthus, Mars is situated east of the Tharsis rise, at the western end of Valles

Marineris (Figure 2.1a). Despite being located between the biggest volcanic complex and largest canyon system on Mars, the exact nature and formational history of Noctis Labyrinthus is not well understood. The region sits atop a topographic high and could be considered a dome or plateau (Figure 2.1b). The region abounds with normal faults, pit craters, and large troughs that cross-cut both the normal faults and pit craters (Figure 2.1b). Additionally, hillslope modification within the troughs of Noctis Labyrinthus is evidenced by landslides and periglacial features that together are indicative of mass-wasting processes that have helped shape the troughs.

The broad structural characteristics of Noctis Labyrinthus have been recognized since the original mapping from Viking imagery by Masson (1977), who concluded that the region resembled the East African Rift. Bistacchi et al. (2004) used image data from the Mars Global

Surveyor (MGS) Mars Orbiter Camera to perform a large-scale fault kinematic analysis and recognized three different orientations of faults—two exhibiting horst-and-graben extensional structures and the third featuring strike-slip motion. Andrews-Hanna (2012a,b,c) explored the formation of Valles Marineris and concluded that extensional stresses from lithospheric-scale dike intrusion and isostatic compensation were responsible for its formation. This author further determined that Noctis Labyrinthus, lacking the isostatically compensated plateaus that characterize Valles Marineris, therefore does not share the same lithospheric-scale dike system as the nearby valley system. Leone (2014) postulated that Noctis Labyrinthus could represent a network of collapsed lava tubes. However, this 2014 study did not provide evidence for lava flows of required volume and eruptive flux to support the formation of lava tubes multiple kilometers deep and wide, nor offer examples of similarly sized lava tubes elsewhere on Mars or other planetary bodies.

Additional work has been focused on the troughs in Noctis Labyrinthus and the physiographic landforms within those troughs (Hajna et al., 2017; Rodriguez et al., 2016).

Rodriguez et al. (2016) argued for fault-controlled groundwater release and collapse in Noctis

Labyrinthus, yet there is a notable scarcity of widespread fluvial features to suggest that a great amount of water was released within Noctis. Hajna (2017) suggested that karstic features are present on the floors of the large troughs of Noctis Labyrinthus, and interpreted the presence of such features as indicative of groundwater interactions with the surficial materials present after the formation of the large troughs. This work also asserted that there could be a possible

14 connection between the size of the troughs and groundwater aquifers being breached by faulting.

However, it is not clear whether the fault-rock characteristics in the region would inhibit or permit the subsurface flow of water.

These earlier studies each proposed similar formational histories for Noctis Labyrinthus, featuring some combination of tectonic, volcanic, or volatile-driven processes. Yet the direct link between tectonic- and volatile-related processes for Noctis has not yet been fully established.

Therefore, we focused on two principal types of tectonic structure in this region—normal faults and pit craters—and characterized their morphology, distribution, and timing relations. We then compared these structures with possible depths to volatile-rich materials in the Martian subsurface. In particular, we worked to address three driving science questions:

1) Did normal faulting alone create the large troughs present in Noctis Labyrinthus?

2) What role did pit craters play in the formation of Noctis Labyrinthus?

3) What does the periglacial and mass wasting geomorphology of the troughs at Noctis

Labyrinthus indicate about the formational history of this region?

We used topographic analysis to characterize the faulting and pit craters within Noctis

Labyrinthus. Fault displacement profiles were used to investigate the geometries and growth styles of these faults; pit craters were characterized in terms of depth and major/minor axes lengths. We also used rampart craters to estimate the depth to an assumed ice-rich volatile layer using scaling ratios for crater depth and penetration depth (e.g., Reiss et al., 2005, 2006; Robbins

& Hynek, 2012). Landslides from the Crosta et al. (2018) data set were used to determine the regional landslide H/L (height/runout length) ratio, small values of which can indicate the presence of fluidized mass movement. Periglacial features are widely present in the region,

15 including solifluction lobes (downhill regolith creep promoted by melting of permafrost), ice- wedge polygons, and thermokarst (topography related to permafrost environments).

In general, we find that normal faulting in Noctis Labyrinthus predates, and likely controls, pit crater formation. Pit craters are preferentially located within graben, i.e., between two normal faults that dip towards each other. Pit crater chains are mostly situated on the periphery of the large troughs that make up the ‘labyrinth’ of Noctis Labyrinthus. In certain instances, pits grow larger (in terms of depth and width) with increasing proximity to the labyrinth. This finding supports the view that pit craters have played a vital role in the formation of the large troughs in Noctis Labyrinthus.

1.2. Faulting Within Noctis Labyrinthus

Normal faulting is indicative of extensional tectonic deformation occurring at local to regional scales. The surficial expression of normal faulting is commonly horst and graben structures, especially where extension is driven by volcanic or lithospheric sources of extensional stresses. These fault scarps tend to be well preserved on Mars because of the relatively low erosion rates (Golombek & Bridges, 2000) compared with Earth. Normal faulting on Mars is often attributed to large-scale diking, in which complexes of individual graben form over an advancing dike tip (Scott et al., 2002; Wilson & Head, 2002), such as the large-scale graben systems in the Tharsis region. Indeed, McKenzie and Nimmo (1999) discussed how the generation of large dike swarms from the Tharsis Rise could account for the spatially extensive graben systems that extend across much of the planet’s western hemisphere. These authors proposed that such large dike systems could potentially melt subsurface ice and promote the

16 formation of the large flooding systems, perhaps even playing a role in the development of

Valles Marineris (McKenzie and Nimmo, 1999).

Dilational normal faulting is a process that results when fault plane dip is changed by the mechanical stratigraphy of the rock through which the fault is penetrating (Ferrill & Morris,

2003; Ferrill et al., 2011; Sims et al., 2003; Wyrick et al., 2003; Wyrick & Smart, 2009). The dip of a fault can steepen as the structure cuts through mechanically stronger rock, becoming shallower once it encounters a weaker layer, and so producing a zone in which void space at depth can be generated. This change in fault dip angle can lead to a collapse of overlying rock mass into the void space, which is manifested on the surface as depressions such as pit craters.

Dilational faulting leading to pit cratering has been observed in Iceland (Ferrill et al., 2011;

Whitten & Martin, 2019) and in the U.S. states of Hawaii (Okubo & Martel, 1998) and Idaho

(Hughes et al., 2018), all of which feature basaltic rocks.

Normal faults around Noctis Labyrinthus have almost certainly penetrated through stratigraphic layers with different mechanical strengths. For example, the region around Tharsis likely contains interbedded lava flows, ash layers, and surface regolith (Tanaka et al., 2004), which have progressively lower rock mass ratings (Schultz, 1995). By comparison with Earth

(e.g., Ferrill et al., 2011), this type of stratigraphic layering should be readily conducive to forming dilational faults on Mars.

Tensile fractures originating at the surface should also be considered for opening void space. Numerous estimates for maximum depth of tensile fracturing in Icelandic basalts

(Gudmundsson, 1992) and mid-ocean ridge basalts (Bohnenstiehl & Carbotte, 2001) show that tensile fracturing on Earth can reach depths of hundreds of meters. On Mars, similar estimates

17 can be made for the maximum tensile fracture depth using the relation (from Gudmundsson,

1992):

푍푚푎푥 = 3푇0/𝜌푔 where Zmax = maximum depth of tensile fractures, To = tensile strength of the rocks (1–6 MPa), ρ

= density of rock mass (2300 kg/m3), and g = gravity (3.711 m/s2). This approximation yields depths ranging from ~350 m to ~2.1 km for tensile fracture depths, which could allow for substantial surface material to drain into the subsurface.

1.3. Pit Craters

Pit craters are a type of landform that are observed on a diverse range of planetary surfaces. Pits have been reported on Earth (Okubo & Martel, 1998), the Moon (Head & Wilson, 1993), Venus

(Bleamaster & Hansen, 2004), Mars (Banerdt et al., 1992; Davis et al., 1995; Wyrick et al.,

2004), icy satellites, and several asteroids (Wyrick et al., 2010). Pit craters are circular, near- circular, or elliptical depressions, distinguished from impact craters by the lack of a raised rim

(Figure 2.2). Aligned groups of pits tend to form linear crater chains, and chains can coalesce to form contiguous, trough-like structures (e.g., Crumpler & Aubele, 1978).

Numerous pit crater formation mechanisms have been proposed, including lava tube collapse (Cushing et al., 2007; Cushing et al., 2015; Leone, 2014), drainage into extensional fractures ( Banerdt et al., 1992; Tanaka & Golombek, 1989; Tanaka, 1997) and/or dilational normal faulting (Ferrill et al., 2004; Wyrick et al., 2004) (Figure 2.2a), dike intrusion (Mege &

Masson, 1996, 1997; Mege et al., 2003; Scott et al., 2002; Scott & Wilson, 2002) (Figure 2.2b), and cryosphere or hydrosphere interactions causing collapse into a void space (i.e, sinkholes)

(Figure 2.2c). Two of these mechanisms have emerged as leading contenders for the formation of pit craters: dike intrusion (Scott and Wilson, 2002) and dilational faulting.

Dike intrusion was proposed by Scott and Wilson (2002) to account for the pit craters on the slopes of the Alba Mons volcano on Mars. These authors reported two forms of pit craters on the edifice: smaller pits that are regularly shaped, less than 2 km in diameter, and typically situated within graben; and larger pit craters that are irregular in shape, often coalesced with one another, and 4–10 km in diameter. Both types of pits were inferred to have formed via collapse related to dike intrusion: the smaller type resulted from collapse into the voids left behind after volatile outgassing of dikes, whereas the larger pit craters were the aftermath of explosive volcanic eruptions from the underlying dikes, which erased all evidence for vents at those locations (Scott and Wilson, 2002).

Dilational faulting was the preferred mechanism of Wyrick et al. (2004) for pit crater formation at numerous sites across Mars, as poorly consolidated material drained into the void space created by high-angle extensional fractures and faults. Dilational faulting has also been invoked for pit craters in northern Iceland, where rifting along the North American and Eurasian plates formed dilational faults at the surface into which the overlying sediments collapsed

(Whitten & Martin, 2019, Ferrill et al., 2011).

Importantly, Wyrick et al. (2004) found that pit crater chain formation can be enhanced by, but does not require the presence of, dikes, which commonly occur in extensional settings

(e.g., Okubo & Martel, 1998). It is not clear, for example, if dike intrusion plays a role in the formation of pits on icy satellites or asteroids. Thus, for terrestrial planets, the identification of pit craters at a given site is not by itself an indicator of whether dike intrusion or dilational faulting occurred. This uncertainty is compounded by studies that considered only one or the

19 other process for a given site where either—or both—is possible, as in the case of Noctis

Labyrinthus, Mars.

2. Methods

We used image and topographic data to characterize the normal faults and pit craters that are present within and surrounding Noctis Labyrinthus. Image data were used to record locations, lengths, and areas of faults and pits. Topographic data was used to measure fault displacement profiles for isolated normal faults, from which fault displacement–length scaling ratios, pit crater major and minor axes lengths and depths, and pit crater scaling ratios were determined.

2.1. Datasets

This study utilized the Mars Odyssey THEMIS Global Day-IR mosaic (Edwards et al.,

2011), greyscale image data from the Mars Reconnaissance Orbiter (MRO) Context Camera

(CTX) (Malin et al., 2007), and Mars Express High-Resolution Stereo Camera (HRSC) greyscale imagery (Jaumann et al., 2007). These three data sets were used for structural mapping at

1:200,000 view scale. When mapping at a smaller view scale was required and, where available, we used imagery from the MRO High Resolution and Stereo Imaging Experiment (HiRISE) dataset. Additionally, topographic data from the MGS Mars Orbiter Laser Altimeter (MOLA) and HRSC merged global digital elevation model (DEM) (Fergason et al., 2018) aided efforts to map the fault scarps as precisely and accurately as possible. A portion of our mapping is shown in Figure 2.3.

2.2. Mapping

All mapping was carried out in the ArcGIS™ 10.6 environment. The nominal view scale of 1:200,000 was decided on to balance the great number of structures present with the complexity required by the map observations. The THEMIS day IR global mosaic was used as the base map for Noctis Labyrinthus, which provided sufficient spatial resolution (100 m/px) to map the major tectonic structures, pit craters, and large troughs therein. The THEMIS mosaic was supplemented with CTX (6 m/px) and/or HRSC (~20–100 m/px) images superior image resolution was required for satisfactory mapping. Mapping of normal faults followed planetary tectonics fault mapping conventions (e.g., Byrne et al., 2014; Klimczak et al., 2018; Callihan &

Klimczak, 2019), where vertices of polylines were placed along the base of the observed fault scarp in the CTX/HRSC, thermal, or topographic data (whichever resolved the fault scarp best)

(Figure 2.3).

Pit caters and troughs were mapped using polygon shapefiles, with vertices placed along the rim of the depression (Figure 2.3). Troughs and pit craters were distinguished from one another based on their appearance. If a pit crater was found to have undergone additional erosion as evidenced by landslides, scalloped rims, or an irregular planform shape, it was not included in our pit mapping and was considered a trough instead. The mapped troughs feature steeper walls than the pits, have chaotic terrain at their bases, tend to show cross-cutting relations with other structures (faults, other troughs, and/or pits), and have irregular planform shapes not consistent with the characteristic circular-to-elliptical pit crater planform. The landslide database from

Crosta et al. (2018) was clipped to the Noctis Labyrinthus area, and the height (H)/length (L) ratio from that dataset used to characterize the subset of landslides.

The MOLA–HRSC merged DEM (200 m/px horizontal resolution, 3 m vertical resolution) was used to acquire topographic measurements of the normal faults and pit craters in the region, and a hillshade from this DEM was made for visualization purposes. Displacement profiles were constructed for a selection of isolated graben (i.e., those not visibly interacting with, or affected by, other faults) distributed throughout the Noctis Labyrinthus region.

2.3. Topographic analysis

2.3.1. Fault Displacement Profile Analysis

Displacement profiles record the amount of along-fault throw as a function of fault length, and permit investigation of individual fault character and growth such as mechanical confinement (e.g., restriction of vertical fault growth because of subsurface stratigraphy)

(Ackermann et al., 2001; Manighetti et al., 2001; Polit et al., 2009; Schulz and Fossen, 2002).

Displacement profiles also provide a means to define or classify faults into discrete populations by more than simply strike (i.e., orientation) or displacement–length ratio (i.e., the ratio of maximum displacement along a fault to the corresponding length), which are common metrics for comparing fault populations (Cowie & Scholz, 1992; Clark & Cox, 1996). Specifically, when combined on one plot and normalized for length and displacement, the profiles can be used to define fault populations using the shape of the displacement profiles (Manighetti et al., 2001).

Displacement profiles were constructed for 24 graben-bounding normal faults that did not show obvious evidence of interaction with neighboring faults. The faults for displacement profile analysis were selected based on the populations described by Bistacchi et al. (2004). We measured the topographic offset along selected normal fault pairs, similar to other recent studies of faults on planetary bodies (Callihan et al., 2018; Klimczak et al., 2018), at spacings of 2 km.

That is, we assumed that the relief of a given fault scarp, in the absence of substantial erosion, represents the vertical offset of the fault itself. To measure relief at each increment, we developed a Matlab™ code to create and plot the topographic profiles, request user input to manually mark the base/top of each fault scarp, and automatically record several metrics for each profile from which measurements were made. These measurements included: profile number; distance along the fault; fault azimuth (i.e., strike); throw (vertical displacement) for both of the graben bounding faults; the slope of each graben wall; and the latitude, longitude, and elevation of both the bottom and top of each bounding normal fault scarp.

2.3.2. Pit Crater and Trough Analysis

Pit crater morphometries were calculated with the ArcGIS™ Zonal Statistics and

Minimum Bounding Geometry functions. The Zonal Statistics function provided the range of elevation for each polygon within the shapefile, which we took to indicate the maximum depth of each pit crater. The Minimum Bounding Geometry tool in ArcGIS™ 10.6 creates a bounding rectangle that self-orients to the major and minor axes of the input shape and provides measurements of each axis and strike of the major axis (measured from north, and east positive, i.e., 0–180°). This tool was used to ensure that measured pit crater dimensions were 90° from one another, and that these measurements reflected the maximum and minimum lengths of each pit without introducing mapping bias. The minimum bounding Geometry tool was used with the

‘width’ option so as to make the width as small as it could , ensuring that all pits were measured with respect to their minimum axes (which will be referred to as the diameter).

3. Results

3.1. Fault Populations

The results of our fault mapping and analysis are presented in Figure 2.4 (in which we show all mapped features) and Figure 2.5 (showing rose diagrams of fault orientations). A total of 8,249 faults were mapped with lengths ranging from 66 m to 166 km. Most of the faults are oriented N–S, but orientations of all azimuths exist, helping to describe the overall curvilinear map pattern of faults that characterizes the central part of Noctis Labyrinthus. We analyzed the faults in groups similar to the two normal fault populations that Bistacchi et al. (2004) noted, as we found those two groupings to be representative of what we mapped. In contrast to the

Bistacchi study, however, we did not bother to record evidence for strike-slip deformation because our focus was on the pit craters and their relations to the normal faulting. The faults were divided into groups based on strike. The north–south-oriented faults (associated with Noctis

Fossae) make up Population 1 (section 3.1.1) (Figure 2.4). For the displacement profile analysis, we divided the circumferential faults into two groups, Populations 2a (section 3.1.2) and 2b

(section 3.1.3) (Figure 2.4) to determine if faults in different physiographic regions of Noctis

Labyrinthus have noticeable differences in length and displacement. The fault displacement profiles we show in Figure 2.6, constructed from 22 graben-bounding normal faults, show that the displacements are no greater than ~250 m. The average width of graben for the displacement profiles we constructed is 1.5 km. We discuss our results for each of the three fault populations in detail below.

3.1.1. Population 1 Results

The dominant population of faults (n = 4,935; 59.2% of the total) in the region comprises those trending generally north–south. This group is defined by strikes ranging 331°–030° and

151°–210°. These structures are generally located in the northern part of Noctis Labyrinthus and resemble a “graben swarm” (Scott et al., 2002; Wilson & Head, 2002). Faults in this population have lengths of 130 m to 166 km (derived from the mapping data). These faults have displacements of up to 210 m at the greatest observed throw value, and generally have maximum throw values ranging from ~100 m to 150 m (derived from the displacement profile analysis)

(Figure 2.6b).

3.1.2. Population 2 Results

The second-most abundant population (n = 2,138; 25.7%) trends northeast–southwest and represents one component of the faulting that appears circumferential to the topographic high of

Noctis Labyrinthus; faults in this population have strikes of 031° to 090° and 211° to 270°, and on the basis of our mapping data are 195 m to as much as 139 km long. This population has maximum throws ranging from 75 m to 150 m (derived from the displacement profiles, Figure

2.6c).

3.1.3. Population 3 Results

The third-most abundant normal fault population (n = 1,261; 15.1%) trends northwest– southeast and is the second component of the circumferential faulting around the dome of Noctis

Labyrinthus (strikes of 091°–150° and 271°–330°). This population has maximum throws up to

150 m (Figure 2.6d) and ranges in length from 66 m to 143 km.

3.2. Pit Craters

In total, we identified and mapped 215 pit craters in the region. The average depth of pit craters is 354 m with a standard deviation of 344 m (Supplemental Table 1). If a pit crater was found to have undergone additional erosion it was considered a trough instead (see section 2.2 for mapping descriptions). In general, about half (n = 111, 52% of total) of the pit craters are situated on the periphery of the large troughs and central Noctis Labyrinthus region (Figure 2.4).

The pit craters have no pervasive orientation pattern, similar to that of the faults (Figure 2.5). In southeastern Noctis Labyrinthus, the pit craters in the chains shown in Figure 2.3 are wider and deeper with increasing proximity to the edge of the Noctis Labyrinthus troughs (i.e., those farthest from Noctis are 1–2 km wide and ~150 m deep, compared with diameters of ~ 7 km and depths of ~2 km for those pits closest to Noctis). The average diameter for all pit craters is ~2 km. The total void space volume of all pits is 938 km3. Pit depth and diameter are closely correlated, with a linear regression r2 value of 0.88 (Figure 2.7).

We plotted pit crater measurements from other Mars- and Earth-based pit crater studies

(e.g., Wyrick et al., 2004; Whitten & Martin, 2019) alongside our Noctis Labyrinthus data, as well as data we collected in a separate study for pit craters near Kilauea Iki on Hawaii, and Nyx

Mons, Venus (see supplemental material for methods). We combined these pit datasets to place those we map in Noctis Labyrinthus in a broader context of known pit craters elsewhere on Mars and Earth (Figure 2.8). The pits we map at Noctis Labyrinthus (Figure 2.8, teal circles) tend to have less depth for given pit diameters than pits in the global Mars pit crater data from Wyrick et al. (2004) (Figure 2.8, red circles). When compared with pits at Kilauea Iki (Figure 2.8, blue circles) and Iceland, from Whitten & Martin (2019), (Figure 2.8, black dots) on Earth, the

26 depressions at Noctis Labyrinthus are ~1–2 orders of magnitude larger in both depth and diameter.

The pit craters at Noctis Labyrinthus were evaluated to determine their mean depth/diameter ratio and to determine their size-frequency distributions. The Noctis pits have a mean depth/diameter ratio of 0.1333 ± 0.0042 (Figure 2.9), and little uncertainty in that mean.

The Mars pits from Wyrick et al. (2004) have a mean of 0.5867 ± 0.0042 (Figure 2.9), much higher than the mean for the Noctis Labyrinthus pits. The size frequency distributions for both depth and diameter show that the Noctis Labyrinthus pits and Mars pits from Wyrick et al.

(2004) follow a negative exponential distribution, something that has not been shown for pit craters before (Figure 2.9).

3.3. Troughs

We mapped a total of 163 troughs within Noctis Labyrinthus. These landforms were distinguished from pit craters on the basis of pronounced morphological differences (see section

2.2). The troughs have no pervasive orientation, like the pit craters (Figure 2.5). The depths of the troughs range from 100 m to 5 km; these structures have lengths of 2 km to more than 300 km for the largest examples. The average width of the troughs is 22 km. The bases of the troughs often feature mass-wasting deposits and periglacial features; of the 163 troughs we mapped, 33 of them (20%) contain at least one landslide. Evidence for periglacial processes is present in a majority of the troughs, with solifluction lobes (Price, L. W., 1974; Johnsson et al. 2012) being the dominant expression of these processes: 90% (n = 147) of all troughs mapped had solifluction features present along the walls (Figure 2.10a). What we interpret as relict ice wedge polygons (Black, R.F., 1976; Brooker et al. 2018) are present in 10% (n = 16) of the

27 troughs (Figure 2.10a), and thermokarst topographies (Rodriguez et al. 2016)(Figure 2.10b) are present in 20% (n = 32) of the troughs. The total volume of the void space within the troughs is

1.07×105 km3. The total combined volume of pit craters (section 3.2) and troughs is 1.08×105 km3.

4 Discussion

4.1. Normal Faulting

We find that normal faults within our Noctis Labyrinthus study area have average displacements of 150–250 m and lengths of up to 400 km. The average width of graben in the region are substantially less than that of the large troughs (1.5 km versus 22 km), strongly suggesting that the large troughs do not solely result from tectonic activity, but were shaped by a secondary process (or processes).

Most (n = 144, or 67% of total) pit craters in the region are situated within graben; this close spatial association suggests that normal faults play an important role in the formation of these pits. These faults may have formed due to diking (Okubo & Martel, 1998); under this scenario, the pit craters most likely developed from dike tip evacuation and collapse of the resulting void space, volcanic outgassing at the dike tip leading to material collapse, or some combination of both mechanisms. With volcanic outgassing, distinct ash fallout or plume dispersal on the ground might be observed (Scott & Wilson, 2002). With dike tip collapse, a collapse extended in the direction of the strike of the dike might be found, which would thus not lead to the formation of mostly circular pit craters. However, we do not find morphological evidence for either of these scenarios.

We therefore prefer a scenario under which pit cratering in Noctis Labyrinthus results from dilational faulting (section 1.2). Given its setting close to the largest volcanic province on

Mars, the prevalence of ash deposits, welded tuffs, and other stratigraphic layers of different strengths within the Noctis Labyrinthus region could promote the formation of dilational fault systems as faults penetrated through these different units. The fault displacement profiles, however, do not show evidence that the faults are mechanically confined. It is therefore possible that, although there may be several geological units with different strength values, these differences are sufficient for minimal deflection of fault dip angles, but not enough to restrict the vertical growth of these faults, which would manifest as clearly plateaued displacement profiles

(e.g., Cowie and Scholz, 1992; Ferrill & Morris, 2003; Ferrill et al., 2004; Ferrill et al., 2017).

4.2. Pit Craters

Comparison of the Noctis Labyrinthus pits to other pits on Mars is important to distinguish if the Noctis Pits are different in any way from previous pit assemblages on Mars.

The mean depth/diameter ratio of the pits in Noctis Labyrinthus is 0.1333± 0.0042, lower than the average depth/diameter ratio of the pits in the Wyrick et al. (2004) data set, 0.5867 ± 0.0042

(Figures 2.9, 2.8). When compared with pit crater assemblages from other planets (Figure 2.8) the Noctis Labyrinthus pits fall below the depth/diameter regions for Venus and Earth, signaling that these pits are different from pits elsewhere the Solar System, and not just Mars.

The lower ratio could be due to a few different factors. First the measurement methods differed in this study when compared to the Wyrick et al. (2004) study. This study used the

MOLA-HRSC combined DEM with 200 m/px resolution to determine the depth via the shapefiles of the mapped pits. The minimum and maximum pixel values within the shapefiles

29 were used to determine the depth of each pit. This likely underestimates the depth in some cases, as the 200 m/px resolution will not necessarily capture the absolute bottom of the pit. In contrast,

Wyrick et al. (2004) used photoclinometry to determine the pit depths with available visible imagery that had sufficient shadows. This method relies on finding such imagery with sun inclination angles that produce shadows, and accurate determination of the sun angle for each scene used to measure pit depths. Both methods have downsides; possible underestimation of pit depths via the DEM and shapefile method and finding imagery and accurate determination of sun inclination angle via the photoclinometry method. It still stands, however that the pits in Noctis

Labyrinthus have a much lower depth/diameter ratio, much less than if every pit was underestimated systematically (but the underestimation mostly affects the smaller pits that have less pixels describing their shape). This low depth/diameter ratio might be indicative of the subsurface stratigraphy which I believe hosts volatiles that further dissect pits.

The presence of pit craters within graben, and the observation that in places the diameter and depth of these pits increase with proximity to the large troughs of Noctis Labyrinthus

(Figure 3), again indicates a close link between faulting and pit formation (as discussed above).

However, the secondary process(es) that could have operated to evolve a pit crater into a large trough is not clear. Nonetheless, the prevalence of rampart craters in the region, as well as evidence given in previous work (e.g., Rodriguez et al., 2016; Hajna et al., 2017) for the presence of surface and groundwater within the troughs of Noctis Labyrinthus, support the possibility that interactions of pit craters with sub-surface volatiles (i.e., water ice) or groundwater could continue the modification and deepening of pit crater walls after initial formation by tectonic processes. Such a scenario is bolstered by global estimates for the current

Martian cryosphere depth by Clifford et al. (2010), who estimated that the cryosphere may be

30 situated 0–5 km below the surface at equatorial latitudes. It may be, then, that pit craters and normal faults were able to liberate subsurface ice within this region, resulting in the further erosion and collapse of pit craters.

To further investigate the possibility that an ice-rich layer (or layers) may have been present at relatively shallow crustal levels across Noctis Labyrinthus, we mapped regional rampart craters ≥500 m in diameter with CraterTools in ArcGIS. Rampart craters are those with a distinct, fluidized ejecta morphology attributed to the impact having featured a volatile-rich target material or bolide. Under the assumption that the rampart crater morphology here is due to target volatiles and not to impact by an ice-rich body, we calculated the approximate depth of penetration of each impact crater with two depth–diameter scaling ratios, from Reiss et al. (2005) and Robbins et al. (2012), respectively. These two calculated depths give first-order approximations for the minimum depth to a volatile-rich layer(s) excavated during these impacts.

The smallest rampart carter we mapped is ~750 m in diameter, which, per the crater scaling of

Robbins et al. (2012), combined with the assumption that excavation of crater ejecta is from the top 33% of the crater depth (Melosh, 1996), suggests a minimum penetration depth of ~208 m to reach a notional volatile-rich layer. In comparison, the Reiss et al. (2005) scaling ratio for this same ~750 m-diameter rampart crater results in a penetration depth of ~300 m. Of course, because this approach returns minimum depth estimates for single, spot locations within Noctis

Labyrinthus, larger rampart craters may have reached volatiles at this minimum depth, but their greater depths of penetration mask the interaction of the crater excavation process with volatiles in the shallow subsurface.

The main finding here is that a majority of the mapped pit craters (n = 118; 55%) extend to greater depths than the minimum depth to volatiles we calculate based on rampart crater

31 diameters (Figure 2.11). The implication of this finding is that, although not all pit craters have grown to sufficient depth to reach a volatile-rich layer derived from the minimum depth to volatiles as indicated by the population of rampart craters in Noctis Labyrinthus, some pits have.

It is possible that the larger troughs in Noctis Labyrinthus were once sets of oriented, individual pits that grew wider and deeper as they coalesced during the liberation of volatile materials from the Martian subsurface. If so, then the remaining isolated pits that we mapped have not (yet) reached a sufficient depth to interact with groundwater or ice in this manner, and/or that such subsurface volatiles were or are no longer present at shallow crustal levels at these locations.

The greatest abundance of pit craters is in the southeastern portion of Noctis Labyrinthus

(Figure 2.4); here, pit crater dimensions increase with increasing proximity to the large troughs in the center of the region (see section 3.2). This observation is consistent with the pits closest to

Noctis Labyrinthus representing a more advanced stage of collapse compared with more distant pits, supporting the view that the troughs in Noctis Labyrinthus evolved from coalesced and heavily modified pit craters that interacted with and liberated some subsurface volatile component, resulting in subsequent collapse and growth.

4.3. Other Geomorphological Evidence for a Volatile-Rich Subsurface

At the bases of the large troughs in Noctis Labyrinthus is geomorphological evidence for landslides, periglacial processes, and a minor contribution from fluvial signatures (Hajna et al.,

2017). For example, Crosta et al. (2018) identified 232 landslides in the region, all within what are mapped as troughs in this study (e.g., Figure 2.4 and Figure 2.12). From the Crosta et al.

(2018) study, the landslides present within Noctis Labyrinthus have crowns (i.e., the landslide initiation point) along the top of the troughs and have toe runouts that are long (with an average

32 height/length (H/L) ratio of 0.27), indicating a potentially fluidized (water or other volatiles) component to the landslides. The H/L value for indications of more ‘fluidized’ landslide behavior is 0.30 per the classification scheme of Crosta et al. (2018b), and the Noctis Labyrinthus landslide average is just below that (55% of landslides have H/L ≤0.30), indicating some amount of fluidization. Of note, however, these low H/L ratios alone do not definitively support the presence of volatiles during sliding, but do provide additional support that subsurface volatile- rich material may have influenced (either through initiating or facilitating) a subset of the landslides in the area.

There is a distinct lack of geomorphological evidence for widespread fluvial activity in

Noctis (e.g., deposits of material from fluvial movement, braided channels running throughout

Noctis, etc.), implying that the nature of volatile loss in this region occurred in discrete events in liquid form over short distances and/or was sublimated immediately to the atmosphere. The geomorphological evidence for periglacial processes and landforms along the trough floors includes ice-wedge polygons (Figure 2.10a), solifluction lobes (Figure 2.10a), and thermokarst signatures as reported first by Rodriguez et al. (2016) (Figure 2.10b), further supporting the presence of an ice-rich layer penetrated by the pit craters and faulting. Any fluvial activity within individual troughs, then, was probably mostly spatially isolated in nature, and not manifest as a connected system transporting water downslope. Given the presence of periglacial landforms in the bases of the trough, however, liquid or frozen water could have persisted for tens to thousands of years depending on the climatic conditions present within Noctis Labyrinthus.

4.4. Total Volume Loss in Noctis Labyrinthus

To further understand how much volume of material was lost to pit crater and trough formation (either through tectonic or secondary erosional processes), we calculated the volumes of both sets of those features in ArcGIS™. Each mapped polygon was used to create clipped triangulated irregular networks (TINs) for each pit or trough. The maximum height of each polygon shapefile was used together with the TINs to determine an upper bound on the volume of these features with the ArcGIS™ Polygon Volume tool. A summary of pit crater and trough measurements, as well as volume estimates, are given in Table 1. Our total calculated pit crater volume is 938 km3, and trough volume is 107,221 km3. The pits therefore cumulatively account for 0.9% of the volume of all the troughs throughout the region. Further, the pits constitute 3.4% of the total trough surface area. The volume of all pit craters and troughs together is ~100,000 km3; by way of illustration, this value is about 3.5% of the current estimates of the Greenland ice sheet (Nordhaus, 2019). Although some of the volumes of pits and troughs presumably reflect purely tectonic deformation, this finding implies that, if our hypothesis regarding volatile loss at

Noctis Labyrinthus is correct, a substantial amount of those subsurface volatiles in this region has been released to the Martian surface and atmosphere.

5. Conclusions

The formation of the deep, large troughs at Noctis Labyrinthus has been postulated to be from a combination of processes including lava tube collapse (Leone, 2014), magma body withdrawal (Andrews-Hanna, 2012a, b, c), and karstic dissolution of material (Hajna et al.,

2017). Based on photogeological mapping and fault and pit crater analyses, Noctis Labyrinthus likely initially developed from pit craters and normal fault systems that formed because of

34 regional extensional tectonics (Figure 2.13a-c). Some pit craters, representing drainage of material into the subsurface, then extended to and exposed a subsurface volatile deposit (e.g., an ice-rich stratum in the Martian crust) located at depths as shallow as ~200 m, with continued collapse taking place as some of this volatile material sublimated or was otherwise removed

(Figure 2.13d-e). This excavation and exposure may have taken place due to normal faults in the region as well, such that displacements on the normal faults could have exposed volatile material in the walls of the fault scarp, promoting further collapse of material, and in some cases transforming graben into troughs. This inference is consistent with the abundance of landslides

(n = 232) within the large troughs of Noctis Labyrinthus, which indicate a history of mass wasting following faulting.

Figure 2.1. A) A global map of Mars, showing the MOLA–HRSC digital elevation model overlaid on a MOLA– HRSC hillshade; the location of Noctis Labyrinthus is marked by a black box. The map is displayed in Robinson projection, centered at 0°E. B) A regional map of Noctis Labyrinthus; prominent, named features are labelled. Elevation data are from the MOLA–HRSC dataset overlaid on the THEMIS daytime IR global mosaic. The map is in equirectangular projection, centered at 9°S, 101°W.

Figure 2.2. Box models illustrating different generalized methods of pit crater formation. a) Extensional tectonics leading to graben formation and void space at depth (i.e., dilational normal faulting). b) Dike intrusion with Plinian- style eruption at the tip, creating void space and subsequent collapse. c) Cryosphere interaction with pit craters formed from tectonics or diking. Figure adapted from Wyrick et al. (2004).

Figure 2.3. A portion of the southeast of Noctis Labyrinthus, to illustrate the spatial interplay between normal faults, pit craters, and troughs. The location of this figure is marked on Figure 1b as a black rectangle. This map is in equirectangular projection, centered at 14.3S, 96.9°W.

Figure 2.4. Mapping results for Noctis Labyrinthus shown on top of the THEMIS day IR global mosaic. All mapped features from this study are shown. Additionally, landslide locations (classified by H/L values) from Crosta et al. (2018a) are shown. The map is in equirectangular projection, centered at 9°S, 101°W.

Figure 2.5. The orientations of faults (mint green), pit craters (green blue), and troughs (orange) we mapped in this study. Orientations were measured with respect to north for all structures, resulting in half-rose diagrams.

Figure 2.6. Displacement profiles for the 22 graben-bounding faults analyzed in this study. Faults were classified based on strike. a) Data for all faults studied; b) Population 1 faults (n = 8); c) Population 2 Faults (n = 4); and d) Population 3 faults (n = 10).

Figure 2.7. A depth–diameter plot of all pit craters located within the Noctis Labyrinthus region (n = 215). A linear regression line is shown in black, with 95% confidence interval of fit shown in grey. The fit of this line is r2 = 0.88, indicating a strong correlation between pit crater depth and diameter.

Figure 2.8. A plot of pit crater depth–diameter values for pits in this study (from Noctis Labyrinthus), as well as for pit data from Hawaii and Nyx Mons, Venus pit data collected by the authors. Additional pits from the literature for Mars are from Wyrick et al. (2004) and for Earth are from Whitten and Martin (2019).

Figure 2.9. Histograms of depth/diameter ratios and size-frequency diagrams of depth and diameter for Noctis Labyrinthus pits and Martian pits from Wyrick et al. (2004). Probability density functions overlaid on histograms for comparison, and Noctis probability density function overlaid on Mars to show difference between Noctis pits and the rest of Mars. Negative exponential functions plotted on top of size-frequency diagrams to show tendency of pit shapes to follow that type of distribution.

Figure 2.10. Examples of periglacial signatures in Noctis Labyrinthus, Mars. a) Ice wedge polygons and solifluction lobes, centered at 7.4°S, 96.3°W. a) Thermokarstic terrain identified in Rodriguez et al. (2016), centered at 6.9°S, 99.0°W. Both maps show CTX imagery in equirectangular projection at the same view scale.

Figure 2.11. A depth–diameter plot of pit craters and rampart craters in the Noctis Labyrinthus region. Crater depths were calculated using scaling ratios from Robbins et al. (2011) and Reiss et al. (2005), to bracket the different minimum depth possibilities for a notionally present cryosphere here. The minimum depth to volatiles as calculated from the rampart crater excavation depth (assumed to be 33% of the crater depth: Melosh (1996)), and maximum depth to tensile fracturing on Mars, are shown with black and brown dashed lines, respectively.

Figure 2.12. An example of a landslide within Noctis Labyrinthus that has a long runout and H/L ratio that indicates potential fluidized movement. Several smaller landslides are evident immediately south of the large slide, illustrating the prevalence of mass wasting in Noctis Labyrinthus. The map is displayed in equirectangular projection, centered at 9.86°S, 95.16°W.

Table 3.1: Pit Crater and Trough Summary Max Total Max Total Max Max Max surface surface n volume volume depth width length area area (km3) (km3) (km3) (km) (km) (km2) (km2) Pits 215 214 938 258 2069 1.9 12.4 38.6 Troughs 163 13,443 107,221 4509 61,598 5.6 143 387 Pit % of trough 1.6 0.9 5.7 3.4 33.9 8.7 10.0 Combined Totals 108,159 63667 Greenland Ice Sheet 2,850,000 1,700,000 (Nordhaus et al., 2019)

Figure 2.13. Schematic cartoon for the formation sequence of Noctis Labyrinthus we propose in this study. a) Prior to extension, a volatile-rich stratigraphy is present. b) Extensional structures begin to form. c) Dilational faulting from material strength differences leads to pit crater formation within the graben. d) Pit craters and/or faults breach a volatile-rich layer(s) and expose it to the atmosphere, leading to sublimation or melting. e) Subsequent landsliding and other mass-wasting processes contribute to the present form of Noctis Labyrinthus. Deep structure with white– black gradient shown in d and e is to illustrate that the material must go somewhere in the subsurface if materials drains into the subsurface, that could be a fault, a dike dip, or another void space allowing drainage of material from above.

CHAPTER 3

Pit Crater Formation in the Solar System: Insights from field on Earth and remote sensing

on other Planetary Bodies

Chapter 3 is being prepared for Submission to the journal Icarus, with the following authors:

Corbin L. Kling1; Byrne, P.K.1; Wegmann, K.W.1; Wyrick, D.Y.2; Bohnenstiehl, D.B.1,

Chesnutt, J. M.1

1 Marine, Earth, and Atmospheric Sciences Department, North Carolina State University,

Raleigh, NC

2 Southwest Research Institute, San Antonio, TX

1. Introduction

Pit craters are a type of landform that has been observed on a diverse range of planetary surfaces (Figures 1.1, 3.1). They have been reported on Earth (Okubo and Martel, 1998; Whitten et al., 2019) (Figure 2), the Moon (Head and Wilson, 1993), Venus (Bleamaster and Hansen,

2004), Mars (Banerdt et al., 1992; Davis et al., 1995; Wyrick et al., 2004), icy satellites (e.g.,

Martin et al., 2017), and several asteroids (e.g., Wyrick et al., 2010). Pit craters are circular, near- circular, or elliptical depressions, and are distinguished from impact craters by the lack of a raised rim. Aligned groups of pits tend to form linear crater chains, and chains can coalesce to form contiguous, trough-like structures (e.g., Crumpler and Aubele, 1978). Many crater chains are situated within graben, which are long extensional structures characterized by a down- dropped floor bounded by two antithetic, inward-dipping normal faults. The occurrence of pit craters in all these environments makes them a useful feature by which to understand better the near-surface properties of the planetary bodies on which they are seen.

Numerous pit crater formation mechanisms have been proposed, including lava tube collapse (Cushing et al., 2007; Leone, 2014; Cushing et al., 2015), dike intrusion (Mège and

Masson, 1996, 1997; Montési, 1999; Scott and Wilson, 2002; Scott et al., 2002), drainage into extensional fractures (Tanaka and Golombek, 1989; Banerdt et al., 1992; Tanaka, 1997), and dilational normal faulting (Ferrill et al., 2004; Wyrick et al., 2004). Two of these mechanisms have emerged as leading contenders for the formation of pit craters: dike intrusion (Mège and

Masson, 1996, 1997; Montési, 1999; Okubo and Martel, 1998) and dilational faulting (Wyrick et al., 2004; Ferrill et al., 2011). Dike intrusion was proposed by Scott and Wilson (2002) to account for the pit craters on the slopes of Alba Mons, a large volcano on Mars. These authors reported two forms of pit crater on the volcano: smaller pits that are regularly shaped, less than 2

51 km in diameter, and typically situated within graben; and larger pit craters that are irregular in shape, often coalesced with one another, and 4–10 km in diameter. Both types of pit were inferred to have formed via collapse related to dike intrusion: the smaller type resulted from collapse into the voids left behind after volatile outgassing of dikes, whereas the larger pit craters represented the aftermath of explosive volcanic eruptions from the underlying dikes that erased all evidence for vents at those locations (Scott and Wilson, 2002).

Dilational faulting was the preferred mechanism for pit crater formation by Wyrick et al.

(2004) for pits at numerous sites across Mars. This mechanism involves the creation of void space by high-angle extensional fractures and faults, into which poorly consolidated material drains (Ferrill et al., 2004, 2011; Wyrick et al., 2004). Dilational faulting has also been invoked for pit craters in northern Iceland, where rifting along the Mid-Atlantic Ridge formed voids along fractures into which the overlying sediments collapsed (Ferrill et al., 2011).

Importantly, Wyrick et al. (2004) found that pit crater chain formation can be enhanced by, but does not require, the presence of dikes, which commonly occur in extensional settings

(e.g., Okubo and Martel, 1998). (It is not clear, for example, if dike intrusion plays a role in the formation of pits on icy satellites or asteroids.) Thus, for terrestrial planets, the identification of pit craters at a given site is not by itself an indicator of whether dike intrusion or dilation faulting occurred. This uncertainty is compounded by studies that considered only one or the other process for a given site where either, or both, is possible—as in the case, for example, of Noctis

Labyrinthus, Mars (see Chapter 1).

A key understanding of pit craters might come from investigating pit shape with regard to the strength of the materials in which the pits are forming. The governing factors of slope formation (i.e., that that may have a bearing on pit crater shape) include Young’s Modulus and

52 tensile strength (del Potro and Hürlimann, 2008; Schulz, 1995), coefficient of friction (Sullivan et al., 2011), and the angle of repose (Lowe, 1976) for a given material. The tensile strength is defined as the strength at which the rock will fracture under tensile stress, whereas the uniaxial compressive strength is the stress at which a rock will fracture under a compressive load. The coefficient of friction is a controlling factor on the angle of repose for a given material.

The Young’s Modulus, tensile strength, and uniaxial compressive strength for a range in basaltic materials (which all pits in this chapter form in) can be found in Table 1. The average

Young’s Modulus of a suite of basalts collated by Schultz (1995) was 78 ± 19 GPa, whereas pyroclastic materials assessed by del Potro and Hürlimann (2005) averaged 3.4–82.3 GPa for strongly welded pyroclastics and weakly welded or interlocked pyroclastics, respectively. For comparison, del Portro and Hürlimann (2008) also included volcanic soils, which were found to have a Young’s Modulus of <1 GPa. More importantly, the tensile strengths of these materials range from 0 GPa for volcanic soils to -14.5 GPa for basalts. These strength estimates clearly demonstrate that, as expected, largely intact basaltic lava flows are the strongest materials, and weakly welded pyroclastic or volcanic soils the weakest, of any of the geological materials in which the pit craters we examined have formed.

1.1. Pit Craters on Earth

By comparing features on Earth with similar landforms we see and map on other planets, such as Mars, we can gain a better understanding remotely of extraterrestrial geology. This approach is termed “comparative planetology,” and has been applied to better understand planetary surfaces for more than four decades. This chapter is focused on what information pit

53 craters on Earth can offer to help understand the properties and formation of pit craters on other planets.

Four relatively accessible locations on Earth in which pit craters have been described

(Figure 1) include: on the island of Hawai’i, along the southeast Kilauea rift zone and to the west of Kilauea within the Ka’u Desert (Figure 1A); within Craters of the Moon National

Monument and Preserve in Idaho (Figure 1B); co-located with graben within the Needles

District of Canyonlands National Park (Figure 1C); and in northern Iceland near Ásbyrgi canyon

(Figure 1D). The Hawaiian pits are thought to be of volcanic origin, either from dike emplacement or from explosive eruptive events (Okubo and Martel, 1998). Similarly, the pits at

Craters of the Moon National Monument and Preserve are also related directly to volcanic eruptions and associated landforms, though considerable fracturing has been noted in the area surrounding the pits, which may indicate a tectonic component to the formation of pits there

(Kuntz et al., 1992; Klimczak and Byrne, 2017). Pits in the Canyonlands National Park are co- located within graben, but almost no literature is available on the pits themselves; a single study focused on the dilational nature of the normal faults there and inferred the pits (termed in the paper ‘sinkholes’) to have formed from antithetical normal faulting in the graben floors

(Kettermann et al., 2015). The pits in Iceland are thought to be a product of a rifting event along the mid-ocean ridge and occur in volcanic sediments located within a river basin (Ferrill et al.,

2011). The presence of small normal faults around the pits, observations made of exposed pit walls, and the lack of volcanic deposits directly associated with the pits, together point to a tectonic origin for these pits (Ferrill et al., 2011).

Here, I focus on two study sites, Craters of the Moon National Monument and Preserve, and Hawaii Volcanoes National Park. I conducted fieldwork at both sites to investigate the pit

54 craters therein, and use them as a basis for interpreting pit craters found on other planetary bodies the Solar System.

1.2. Craters of the Moon National Monument Study Site

Craters of the Moon National Monument and Preserve (CRMO) is located in south- central Idaho, within the Great Rift volcanic zone, which is part of the larger Snake River Plain volcanic province. CRMO is also situated at the northern edge of the Basin and Range physiographic province, defined by large-scale extensional tectonic structures that are oriented approximately north–south. The origin of the Basin and Range is due to the subduction of the

Farallon plate beneath the North American plate. The Farallon plate is subducting at a shallow angle (flat slab subduction), causing drag along the lower part of the North American plate and producing the extensional tectonic systems in response (Eaton, 1982; Humphreys, 1995). The

Snake River Plain is a 50–100-km-wide lava plain that extends from Payette, ID to Yellowstone

National Park in the east. The Snake River Plain is host to the youngest basaltic volcanism in the contiguous United States. The volcanism is youngest on the east and is progressively older to the west (Kuntz et al., 1992).

Craters of the Moon was designated a National Monument in 1924 and later expanded to include the Wapi and King’s Bowl lava flows to the south in 2002. The area comprises four component rift sets: 1) the Great Rift; 2) Open Crack Rift; 3) King’s Bowl Rift; and 4) Wapi Rift

(Prinz, 1970). The first published exploration of the area was by Limbert (1924).

The volcanism in the area ranges from ~15,000–2,000 years old (Kuntz et al., 1986).

Radiocarbon dating performed by Kuntz et al. (1986) from the numerous flows showed at least eight episodes of volcanism in the whole Craters of the Moon area, with the King’s Bowl and

Wapi lava flows relatively younger (~2300 yr. before present (B.P.)) and the Craters of the Moon

Lava field somewhat older (~2000–15,000 yr. B.P.). Kuntz et al. (1986) dated only 20 of 60 identified flows in the region, but selected samples for age dating analysis on the basis of morphology and relative age given by the degree of vegetation and color which ensured that a representative spread of flow ages would be determined.

The types of volcanic flows present at CRMO were dominantly emplaced in an effusive style and are basaltic in composition. The volcanic field is dominated by large cones and effusive pahoehoe and a’a’ lava textures. The cones are evidence of large spatter events and lava fountaining sustained enough to build constructional topography. The effusive lava represents a later stage, once gas content within the magma decreased enough to allow for low viscosity flow and effusive styles of volcanism. The occurrence of cones throughout the field, combined with the effusive lavas, was taken as evidence by Kuntz et al. (1982) that the magma source was able to re-equilibrate between eruptive episodes. Smaller spatter cones (tens of meters in diameter) can be found within portions dominated by effusive lavas and larger cones (hundreds of meters in diameter). The smaller spatter cones represent localized episodes of more gas-laden lavas being erupted (Kuntz et al., 1992; Rader et a., 2018). All four rift zones are aligned along the same strike (approximately 150°). Spatter cones and lavas are also aligned in the general strike of

150°, which corresponds to the regional slope to the south. The Sawtooth Mountains are just to the north, and there is a regional drop in elevation to the Snake river of ~450 meters, providing a sufficient gradient (1–2%) to control the flow direction of the lavas here. Fractures present throughout the flows are aligned similarly (NW–SE), and much older epicenters for the Snake

River volcanics to the west are also often similarly aligned. The progression of the magma

56 source through the region resembles an arcuate shape, with the current epicenter located at

Yellowstone National Park (Kuntz et al., 1992).

1.3. Hawaii Volcanoes National Park (HVNP) Study Site

Hawai'i Volcanoes National Park is situated on the biggest Hawaiian island, Hawai’i. The

Park encompasses a large portion of the southern and southwestern side of the island (Figure

3.1a). The landscape is largely basaltic, built from flows emanating from Kilauea, the largest caldera in the Park, and smaller associated calderas and vents in the area (Jaggar, 1947). The source for Kilauea is a hot spot underneath the Pacific tectonic plate, and the Hawaiian island chain records the movement of the plate over the hot spot. Kilauea is not the only active caldera on the island of Hawaii; to the north, the larger counterpart to Kilauea is Mauna Loa. There has been speculation in the past that Mauna Loa and Kilauea share the same magma reservoir such that when one is erupting, the other is shut off, and vice versa (Helge et al., 2012). Pit craters were first described in the area by Wilkes (1845), who gave the name to pits found along the

East Rift Zone.

Okubo and Martel (1998) were the first to collate the entirety of pit crater research for

Kilauea and use observations and interpretations to further the definition of pit craters proposed by Wilkes in 1845. Okubo and Martel (1998) measured length, width, and depth of all pit craters in the Southwest Rift Zone (SWRZ) and measured the then-current size of the pit named Devil’s

Throat in 1998. These authors incorporated measurements from previous studies (Favre, 1993;

Jaggar, 1947; Macdonald and Eaton, 1964; and Whitfield, 1980), to build a picture of how these pits were developing.

Okubo and Martel (1998) offered a generalized pit crater formation sequence for those in the Kilauea area that starts with magma moving vertically into a tall, steep, mode-I rift zone- related fracture (i.e., an opening-mode fracture). As the propagating fracture due to magma entering the cavity nears the surface, surficial fractures form in pairs, one on either side of the underlying magma body. These evenly spaced ground fractures are indicative of a dike at depth, so their occurrence is important to understanding how the pits formed in Hawaii. The mode-I fracturing allows for blocks of overlying rock to collapse into the cavity, creating an inverted cone shape as continued collapse occurs. Okubo and Martel (1998) also asserted that the ground- cracking zones promote faster collapse of material, creating elliptical shapes for pit craters. The observation that pits can become elongated perpendicular to the strike of the fracturing is uniquely opposite to pits forming in a purely tectonic sense, when the elongation is expected in the direction of fracturing.

1.4. Scientific Rationale

Observations of pits within basaltic environments are the best direct comparison to pit craters forming on other planetary bodies such as Mars and Venus, which have primarily basaltic crusts. The morphologies and formation mechanisms of the pits in similar environments on Earth can therefore help us better understand the formation of those pits on other terrestrial planetary bodies in similar basaltic environments. Therefore, the science questions addressed in this chapter are:

1) How do the length, width, and depth values for pit craters across the Solar System compare with pits on Earth for which such data are available? And what do differences in these values tell us for pits on each body, and for pit crater formation more generally?

2) What do the shapes of pit craters (i.e., inverted-conical or straight-walled) mean for the mechanism(s) and environment(s) governing their formation?

3) Are there any distinguishing morphologies or geological environments corresponding to one or more pit crater formation mechanisms that can be used to help determine pit formation on other planetary bodies?

2. Methods

The core approach for this project involved visiting these sites and taking photographs and field notes, and investigating the geological structures present to assess the formation mechanism of specific pit craters to the greatest extent possible. Additionally, unpiloted aerial systems (UAS) were utilized to obtain high-resolution color imagery and topographic data for further mapping and geographic information system (GIS) analysis. The methods for the fieldwork, UAS mapping, and GIS mapping are described below.

2.1. Data

A variety of data was used in this project, including light detection and ranging (lidar) topographic data, and UAS-derived imagery and topographic data. Lidar data was used for the topographic analyses of pit craters in Hawaii Volcanoes National Park. For temporal baseline data, the Big Island Lidar Survey from 2010 offered adequate coverage for all the pits in this project. The 2018 Kilauea Southeast Rift Zone eruption prompted lidar monitoring flights co- sponsored by the Hawaii Volcano Observatory and the National Center for Airborne and Laser

Mapping, with those flights taking place in June, July, and August 2018. Additional data collected by the USGS via a UAS Lidar system was collected in September 2018 near the

59 summit of Kilauea to track the deflation of the summit. All these lidar datasets were used to help map the pit craters, but, more importantly, to track any changes in the pit crater morphology during the increased seismic activity from the Kilauea eruption.

In addition to the lidar data for the analysis of pits in Hawaii, a handheld camera was used to create photogrammetric models to give context for the pit shape at the outcrop scale. To generate those data, handheld photos were taken around each pit from multiple look angles and heights to ensure as much overlap between images as possible, for accurate 3D reconstruction. In some instances, the entire circumference of a given pit was not accessible and so could not be fully photographed, in which case 3D reconstruction was either not possible or resulted in poor results.

The UAS data acquired for CRMO were generated from a DJI Matrice 600 UAS equipped with the DJI Zenmuse X5 camera. (See Section 2.1.1 for more details on the UAS data collection and processing methods.) The data generated by photogrammetric processing included orthomosaics and digital surface models (DSMs). The number of photos, resolution of each orthomosaic and DSM, and final file names for each UAS survey completed at CRMO are given in Appendix 1. Orthomosaic image resolutions vary between 4 cm and 10 cm/pixel, and DSM horizontal resolutions range from 8 cm and 20 cm/pixel, with a vertical error is <10 cm. Yet higher resolutions could be attained by processing the images with “Ultra-high quality” settings, but the time required to carry out such processing was >1 day, and the marginal (2–4 cm) increase in resolution was determined to not be worth the additional processing time.

Additional imagery for CRMO was provided via the National Agricultural Imagery

Program database. This imagery is collected by piloted aircraft, usually once per year, and has resolutions of up to 1 m/pixel in some places; these data were used to help plan and execute the

60 fieldwork. Additional USGS topographic quadrangle maps and imagery maps were also used during fieldwork to navigate from site to site.

2.2. Field Methods

The most important information gathered during the fieldwork included in-situ observations of these pits, their interiors, and the surrounding volcanological setting (e.g., the presence of lava flows, ashfall, cinder cones, lava squeeze ups, inflation clefts, etc.). Any pervasive fracture systems, notable lava textures, and evidence for mechanical stratigraphy visible within the walls of the pits were also noted, to help determine the controlling factors on pit growth.

2.2.1. Uncrewed Aerial Systems (UASs)

Uncrewed aerial systems (UAS) were used to supplement field observations by providing high-resolution color imagery and topographic data to map and analyze using GIS. The DJI

Matrice™ 600 hexacopter UAS equipped with the DJI Zenmuse™ X5 camera system was used in CRMO to collect these data. The craft takes six batteries, and we opted for the larger TB48S 6 cell lithium polymer batteries with 5700 mAh capacity. These larger TB48S batteries allow for up to 38 minutes of flight time with no payload, or around 30–35 minutes of flight time with the

DJI Zenmuse™ X5 camera installed. This camera has a 16-megapixel sensor and a 72° field of view (equivalent to a 30 mm lens) and only 0.40% distortion. Minimal distortion is important when conducting UAS surveys for photogrammetric processing to not have significant pixel distortion during the processing, which can cause irregularities in the data outputs.

The UAS was controlled using an Apple™ iPad and the Pix4D Capture™ application.

The Pix4D Capture™ application allows the user to define the area of interest, and determine optimal overlap, sidelap, elevation, and camera settings for the image acquisition. After defining these settings, the user launches the drone (either manually or automatically through the Pix4D

Capture™ application), uploads the flight plan to the drone via the application, and the UAS will fly the designated flight pattern, triggering the camera shutter at pre-determined locations. Once a set of batteries has been depleted, the user can install a second set of batteries and resume the mission from the position of the last captured image.

The flight missions for this work were designed with 75% overlap and 70% sidelap for the images to be acquired. The aircraft was flown at 120 m altitude above takeoff location (just under the maximum 122 m allowed per Federal Aviation Administration rules). Multiple battery swaps were required at most sites due to the size of the field areas, and sometimes missions spanned multiple days depending on the battery charge level and weather conditions (as high winds can affect how efficient the drone is during flight).

2.2.2 Processing of UAS and Handheld Photogrammetric Data

The handheld photos and UAS data were both processed using the same workflow in

Agisoft Metashape Professional™ software. To create accurate and precise data products from the UAS imagery, ground control points (GCPs) were placed in the field and marked with an

Emlid RS+ real-time kinematic (RTK) global navigation satellite system (GNSS), consisting of one base station unit and one rover unit. The Emlid system is capable of recording positions at centimeter- to sub-centimeter-scale accuracy and precision for GCPs, allowing for the same accuracy and precision in the final data products. The base station collects location readings for a

62 set amount of time, usually at least ten minutes, after which the distance between that base station and the rover unit (moved by the user) is measured in real-time as different field locations are marked with the rover. These real-time location differences are used to provide the increased resolution of measurement accuracy within the rover data. At least eight, and up to ten, GCPs were used to ensure that the final data products were well georeferenced and suitable for GIS analysis. Each GCP location was determined using the Emlid RS+ GNSS system with acceptable errors of ± 20 cm in latitude, longitude, or elevation. Sometimes multiple points were collected for a given GCP if error values were too high or not stable at the time of collection.

The Agisoft Metashape Professional™ programs produces orthorectified and georeferenced orthomosaics and DSMs from input images. This software aligns captured images using pixel matching along with the exchangeable image file format (Exif) data imparted into each JPEG (which includes GPS latitude and longitude, as well as yaw, pitch, and roll of the aircraft) to align the photos and to determine the best mosaicking for photogrammetric processing. The Agisoft software further allows users to mark ground control points (GCPs), either those targets placed in the field or natural landmarks, to increase the accuracy of the georeferencing. The general processing workflow is shown in Figure 3.2.

2.3. Geospatial Methods

2.3.1. Mapping of Tectonic and Volcanic Landforms

Mapping was conducted using the orthomosaics and DSMs generated by Agisoft

Metashape Professional™. A GIS was set up for each field site within ArcGIS ArcMap™ 10.7.1 that contained all relevant data. Each UAS-derived DSM was used to create a hillshade (a grayscale image with artificial illumination to accentuate topography) raster for each mapping

63 area through the ArcMap™ 3D Analyst Hillshade tool. Shapefiles for pit craters, fractures, and volcanic features of interest were created for digitizing on top of the orthomosaics and DSMs. Pit crater shapefiles were created as polygon files with Z-values enabled, fracture shapefiles were created as polyline files, and volcanic features were recorded as points.

Mapping of features was done on the imagery or the DSM with transparency overlaid on the hillshade to provide a pseudo-three-dimensional look, which can aid in the identification of features. Any feature that showed evidence of being or hosting a depression or cavity, that lacked a raised rim, and that was circular to elliptical in plan was mapped as a pit crater. The polygons for such features were drawn such that the vertices were placed along the edge of each depression, as indicated by the hillshade and DSM. The fractures were mapped by finding observable fissures on the imagery or DSM and hillshade combination.

2.3.2. Geospatial Analysis

Determining the morphometric properties of mapped features was done with tools in

ArcMap™ 10.7.1. Pit crater orientation and minor/major axes were measured with the Minimum

Bounding Geometry Tool in the Spatial Analyst toolbox of ArcMap™. This tool fits the smallest rectangle (based on the width or the length of the feature of interest—in this case, the minimum pit crater width) to the polygon or polyline feature inputs. This tool also measures the length and width of the fitted rectangle, and the azimuthal orientation of the rectangle (measured from north, between 0–180°). To determine the pit crater depth, the Zonal Statistics Tool was used to determine the elevation range within each polygon that defines a pit. The spheroidal length of fractures was determined using Tools for Graphics and Shapes (Jenness, 2011), ensuring that the length measurements were accurate.

Morphometric data were analyzed in R Studio™, a graphical user interface for the R statistical computing language, and in Matlab™. Pit crater diameters (i.e., major axis lengths) and depths were compared with those measurements for pits in Noctis Labyrinthus, other

Martian pit crater data from Wyrick et al. (2004), and data for pits in Iceland from Whitten &

Martin (2019). Approximations for pit geometry based on published image data for a select set of asteroids were also included (sections 3.2.1. and 3.2.2.). A linear regression was used to determine the relationship between pit depth and diameter for each planet’s population, to investigate if pit craters have a similar geometric relationship between different planets and similar scaling to faulting (Clark & Cox, 1996; Cowie & Scholz, 1992a, 1992b).

The measurements and morphometric data for each pit are given in Appendix A for all such structures studied in the field. Descriptions of each pit chain were made by compiling all field observations with GIS data to interpret a formational history for each assemblage of pits at the site visited.

3. Results

3.1. Field Work Results

3.1.1. Craters of the Moon National Monument Field Areas

3.1.1.1. King’s Bowl

King’s Bowl (centered at 42.94°N, 113.21°W) is one of the more well-studied regions within Craters of the Moon National Monument and Preserve. Hughes et al. (2018) described the formation mechanism for the main pit at King’s Bowl as a phreatomagmatic explosion resulting from groundwater interaction with freshly deposited hot rock or a dike tip in the shallow subsurface (Figure 3.3). Our fieldwork at King’s Bowl reaffirms these findings, as evidence for

65 an explosive event is present in the form of strewn rocks and boulders and as downwind ash deposits seen while we were in the field. Inside the main pit at King’s Bowl, I found evidence for the eruptive fissure feeder dike in the northern cleft, which also showed lava textures consistent with the recession of the lava (Figure 3.4). The lava flows here all exhibit horizontal layering and are <1 m thick. Minor undulations in the lava bed thickness are present.

Figure 3.5 shows the mapping results overlaid on the UAS-derived orthomosaic. We measured the pit crater and depression dimensions (major axis, diameter, and depth), which are given in Appendix A. All fractures evident on the orthomosaic or DSM and hillshade were mapped; the majority of these fractures are located approximately 1–1.5 km on either side of the vent and associated pits/ depressions.

3.1.1.2. Yellowjacket Waterhole

Yellowjacket Waterhole (Figure 3.6, centered at 43.40°N, 113.50°W) is located within the northern part of CRMO and is accessible via backcountry trails. The area is composed of numerous cinder cones and eruptive centers. Yellowjacket Waterhole is named for a waterhole within a collapsed lava tube located just south of the pits studied in this region. Due to the limitations placed by the proximity to the wilderness boundary on UAS operations, flight coverage did not include all the pits in this area.

The pits here are in mostly unconsolidated cinder and tephra deposits, likely from one of the nearby vents (Figure 3.7A). The southern end of the pit chain cross-cuts a vent (Figure 3.7B,

C), which could be the source for the cinder deposits these pits formed in. Within the pits, evidence for layering of materials with different strengths is present (Figure 3.7A). Some lava flows appear to form the uppermost layer in the pits, with more unconsolidated material below,

66 indicating a layering arrangement in which strong material near the surface overlies weak material. Deeper in the pits, other lava flows are present, but most of the material visible within the pit walls is composed of collapsed material that appears to be mostly unconsolidated (with some blocks of lava flows present).

Figure 3.6 shows the mapping results overlaid on the UAS-derived orthomosaic.

Measured pit crater and depression dimensions are listed in Appendix 2. Major axes lengths for these pits ranged from 20 to 121 m, minor axes from 14 to 73 m, and depths from 3 to 24 m.

Fractures were observed in the field on the western side of the pit chain. All fractures evident on the orthomosaic or DSM and hillshade were mapped, resulting in eight total fractures, all aligned with the pit crater chain (Figure 3.6, 3.7D).

3.1.1.3. Coyote Lake

Coyote Lake (centered at 48.19°N, 113.16°W) is a previously unnamed feature located between the main CRMO park to the north and King’s Bowl to the south. The area is heavily vegetated, suggesting that the flows here are older than those at Yellowjacket Waterhole or

King’s Bowl, which both feature much less vegetation. The appearance of Coyote Lake from aerial imagery is that of an elongated pit crater, narrow and long, with circular scalloped edges on the northern and southern ends. From field observations, the feature is raised above the surrounding plains, and the UAS-based DSM shows the convex nature of this feature (Figure

3.8, 3.9A). Thin lava flows are evident in the uppermost walls of the pit; no flows are greater than 50 cm thick here (Figure 3.9B, C). Around the edges of the pit, discrete lava flows were observed emanating from the pit walls and away from the pit in a ‘lava channel’-like form, which are not evident on the DSM due to its small relative elevation to its surrounding.(Figure

3.9D). No fracturing on either side of the pit—like that seen at King’s Bowl and Yellowjacket

Waterhole—was seen in the area during the fieldwork or identified with the UAS-derived DSM and orthomosaic during mapping.

3.1.1.4. Coltrell’s Blowout

Coltrell’s Blowout (Figure 3.10, centered at 42.99°N, 113.19°W) is located just north– northeast of King’s Bowl. Coltrell’s Blowout is in older lava flows than King’s Bowl, due to the heavy vegetation observed covering the entire area such as those seen at Coyote Lake. When viewed with aerial imagery, Coltrell’s Blowout is morphologically like that of Coyote Lake: a long linear feature with rounded and scalloped edges on the north and south end (Figure 3.10).

Thin (<50 cm) lava flows were observed within the uppermost walls of the pit. As at Coyote

Lake, draping lava flows were seen emanating from the crest of Coltrell’s Blowout down the sides of the topographic high.

3.1.2. Hawaii Volcanoes National Park Field Areas

3.1.2.1. Kilauea Iki

A small pit crater chain (with more than 10 pits) is located on the eastern flank and parallel to the Kilauea caldera and within the southern deposits of the 1959 Kilauea Iki eruption

(Figure 3.11). The deposits from the 1959 eruption are mostly unconsolidated weakly welded cinder and tephra created from the large lava fountain that was located just to the north in the

Kilauea Iki crater. I documented this site first in late April/early May of 2018. During that visit, I was able to take measurements of the pit axes (Appendix 2) and recorded the pit morphology with handheld photos (Figure 3.12A–F). Shortly after my first visit, the Kilauea lava lake

68 emptied and the Southeast Rift Zone eruption began in mid-May (Neal et al., 2018). Monitoring lidar data was flown in response to the eruption by the Hawaiian Volcano Observatory (HVO) and National Center for Airborne Laser Mapping (NCALM), and I used these data to identify changes in the pit chain (Figure 3.13).

There are at least 10 pits visible in this chain on the 2009 Big Island Lidar Survey

(Figure 3.13), although some were too vegetated to access during fieldwork. The pits in this chain are meter scale, with the largest being 6.46 m in diameter. The deepest pit is 1.73 m deep.

All the pits in this chain are generally conical in shape, although some have straight sections or overhangs near the top (Figure 3.12C). The pits are aligned with the eastern wall of the Kilauea

Caldera (Figure 3.11). Lidar-derived elevation data show a slight elevation change of a few meters in the east–west direction along the strike of the pit chain (Figure 3.13). The two monitoring lidar surveys for the 2018 eruption were collected in June and July that year. The

June survey showed no change in the pit crater chain, but the July survey showed the emergence of a brand-new pit on the north side of the chain, closest to the Kilauea Iki cinder cone (Figure

3.13B, C). The new pit is approximately 1 m by 1 m in aerial extent and ~1 m deep. The axes and depth measurements are an approximation due to lidar return issues (the pit depth is listed as

>5 m deep on the lidar data); the approximations I give here are based on viewing the pit in person during fieldwork in October 2018. The July lidar survey shows the pit larger than it is, possibly due to a multipathing problem with the laser pulses that reached the pit interior. Unlike the other conical-shaped pits in the chain, this new pit is cylindrically shaped. The locations of the nearest resolved earthquakes to the pits from the USGS Earthquake database (USGS, 2020) are shown relative to nearby faults and the pit chain itself in Figure 3.11. A total of six earthquakes was recorded in that region during the June–July 2018 timeframe, with two of those

69 earthquakes occurring within 500 m of the nearest fault to the pit chain. The magnitudes of the earthquakes range from 2.6 to 3.8, and depths range from 0.1 km to a 1.3 km deep (Figure 3.11).

3.1.2.2 Devil’s Throat

Devil’s throat is a cylindrical pit with overhangs along some of the pit rim situated 4 km to the southeast from the Kilauea summit (Figure 3.14A–E). The growth of this pit has been documented since it originally formed in the mid-1900s (Jagger, 1947). Its walls are made up of many stacked lava flows, with no clear evidence of unconsolidated material (e.g., pyroclastic layers) within the walls. The entirety of the walls are made up of horizontal to sub-horizontal basalt flows no thicker than 5 meters. The bottom of the pit has an uneven pile of blocky material

(sub-meter- to meter-sized) that appears to be talus sourced from the walls (Figure 3.14B).

Fracturing is present on both sides equidistant from the pit aligned with the west–east-trending

Eastern Rift Zone extending from Kilauea caldera to the east–southeast (Figure 3.14C). The fractures converge on the pit, appearing to bound the walls of the pit on the northeast and southwest sides for less than 10 meters along each side. Figure 3.14E shows the 3d model generated from handheld images of Devil’s Throat. Importantly, the 3d shape of the pit is cylindrical, with little deviation from a circular plan view shape at depth within the pit.

Accumulated talus is preferentially stacked on the northeast side of the pit (Figure 3.14B, E).

3.1.2.3 Twin Pits

The “Twin Pits” pit craters are located 5.5 km to the southwest of the Kilauea summit in the Ka’u Desert. Both pits are cylindrical to irregular in 3D shape and, as for Devil’s Throat, show stacked lava flows and little to no unconsolidated pyroclastic material within the walls. The

70 western pit is partially filled with unconsolidated material on the northwestern side, as well as blocks of lava flows that appear to originate from the sides of the pit (as opposed to debris from nearby explosive eruptions which would be present on the ground surface around the pit as well).

The eastern pit is unfilled with debris except for some large sub-meter- to meter-size blocks seen in the bottom (Figure 3.15A–F, 3.16A–F). Both pits are made up of ~1-2 m thick horizontal lava flows, and no noticeable strongly or weakly welded pyroclastics (as observed when I entered the southeast pit). The upper third of the eastern pit is made of stacked horizontal lava flows, whereas the bottom two thirds of the wall are covered in lava that is draping down towards the bottom of the pit (Figure 3.16D). To the immediate northeast of both pits is a cinder cone

(Figure 3.16B). In the bottom of the northeastern pit, a conduit 10–20 m across can be seen exiting the northeast side of the pit in the direction of the volcanic cone (Figure 3.16D). The 3d models generated from handheld photographs show that these two pits are mostly cylindrical in shape, with some overhangs in places.

3.2 Remote Observations of Venus

Pit craters have been noted on the surface of Venus before, with the most substantial analysis by Davey et al. (2013). They analyzed the pits found at four locations across the surface of Venus: Ganiki Planitia, Ulfrun Regio, Themis Regio and Idunn Mons and found they are generally found in weakly welded cinder deposits and were confined to graben–fissure systems, leading to the interpretation that the pits are hierarchically clustered by the locals they are found in. This work established the geologic environments that pit craters and chains on Venus are found in but did not systematically measure morphometric properties such as diameter and depth

(Davey et al. 2013).

Additionally, Sawford et al. (2015) noted the occurrence of >100 pit crater chains on the flanks of the Nyx Mons volcano on Venus, which is an 875 km-diameter volcano located within the Bell Regio region of Venus (centered at 30.2 N, 48.5 E). Previous work attributed the radial grabens, or grabens that radiate from the center of the volcanic edifice, to stresses from lithospheric flexure, an uncommon occurrence on Venus (McGovern and Solomon 1998;

Herrick, et al. 2005). This volcano was chosen for analysis due to the previous literature citing the large number of pit craters, pit crater chains, and co-existence with grabens present. Sawford et al. (2015) classified the pit crater chains on Nyx Mons by appearance and described 4 types: 1) even-sized, i.e., non-coalesced, circular to elliptical pits of similar sizes; 2) trough-type, i.e., coalesced elongated pits within the chain; 3) “tadpole”-type, i.e., one larger pit with several smaller, ancillary pits; and 4) lava-filled type. These authors also noted the occurrence of graben systems in the area, of which some are collocated with the pit crater chains, but do not bound the pit crater chains completly.

I used Magellan synthetic aperture radar (SAR) data (Ford, 1993) to map the pit craters.

The previous literature on this volcano and the opportunity to provide more insight into the pit craters present made Nyx Mons a good addition to this project. The SAR data was acquired in both left-look and right-look modes; for this mapping, I used the left-look data. The spatial resolution is ~100 m/pixel for the SAR data. There is Magellan topographic data available for this region, but its resolution is too coarse (~1 km/px) to make depth measurements on the pits.

I mapped a total of 312 pit craters in the region, of which 49 had noticeable shadows created by the Magellan radar illumination angle. Using photoclinometry due to the presence of shadows, I was able to estimate the depths of these 49 pits (Figure 3.17A–C). Photoclinometry is a method by which to approximate the topography of a feature, in this case the depth of the pit

72 craters, using the shadow lengths. The pit craters have major axes 500–5000 m in length, and, per my photoclinometry estimates, range in depth from 100–660 m. The pit chains are aligned radially downslope on the volcano and are mostly circular in planform view, although some irregular pits do exist. The 3D shapes of the pits cannot be discerned with the resolution of the

Magellan altimetric or stereophotogrammetric data, but the photoclinometry provides a base- level approximation for their depths.

3.3. Other Data Collected from Literature

To increase the breadth of this study, data was collated from the literature on suspected pit craters on small bodies and moons within the Solar System. I found data for the Martian moon Phobos, asteroid 951 Gaspra, asteroid 243 Ida, and Enceladus, the sixth largest moon of

Saturn. While the data presented here is useful in describing pit craters throughout the Solar

System, it should be noted that for some of these bodies (the asteroids in particular), the image data are scarce/low resolution, and accurate georeferencing is sometimes not available. The measurements reported here are the best estimates for the sizes of those pits or depressions.

3.3.1. Asteroids and Small Moons

Veverka et al. (1992) provided estimates for linear depressions, or grooves, on the

Martian moon Phobos, asteroid 951 Gaspra, and asteroid 243 Ida. Although the term “grooves” are not typically used to describe pit craters, these authors mapped the circular depressions making up the groves (sometimes the full groove if no circular depression were found). Phobos, a moon of Mars, has depressions with most over 8 km in length, with a maximum of as much as

13 km, and widths that range from 100 m to over 900 m (this is the extent of data reporting from

Veverka et al. 1992). The depths of these depressions on Phobos are typically 100 m, with a maximum of 200 m. On asteroid Gaspra, the groove-like depressions have maximum lengths of

2.5 km, but most are under 1.5 km. These features have a maximum width of 400 m, and depths less than 20 m (although Veverka et al. (1992) gave no definitive depth measurements).

Sullivan et al., (1996) measured grooves and depressions on Ida, a main-belt asteroid imaged during a flyby by the Galileo spacecraft. These depressions had lengths of up to 4 km, widths of up to 400 m, and depths less than a few tens of meters, but more accurate measurements were not feasible given the nature of the Galileo flyby data . Proctor et al. (2003) measured grooves and depressions on 433 Eros, a near-Earth asteroid, using data from the NEAR

Shoemaker spacecraft, which visited the asteroid in 2000. The Eros groves are often found in orthogonal sets and have scalloped edges. Due to this morphology, Proctor et al. (2003) regarded these structures as having formed by drainage of unconsolidated material into underlying fractures. With available data, those authors reported the depressions on Eros as having lengths less than 800 m, widths less than 100 m, and depths of tens of meters.

Buczkowski et al. (2015) found that lineaments on Vesta near Brumalia Tholus to closely resemble pit crater chains and attributed them to subsurface faulting. Additional pit chains on

Vesta were identified on Vestalia Terra that are unrelated to the Rheasilvia secondary impact crater lineaments that have similar strikes. Buczkowski et al. (2016) found linear the features of

Samhain Catenae on Ceres to closely resemble pit craters due to the lack of raised rim and the coalescing nature of the features. These pit craters found on Vesta and Ceres are found in chains that are often coalesced, hence no depth or diameter data is available to include in the analysis later in this chapter.

3.3.2. Enceladus

Martin et al. (2017) characterized pit crater chains on Enceladus, and found they cross- cut cratered plains interpreted to be 4.0 Ga (Kirchoff & Schenk, 2009). The pits also cross-cut terrains interpreted to be geologically young, such as the so-called “tectonized regions”, or regions of concentrated tectonic features (Crow-Willard & Pappalardo, 2015; Kirchoff &

Schenk, 2009), indicating that the pits on Enceladus could be of 0.2–2.0 Gyr old. The pits range in diameter from 300 to 1000 m, though no accurate measure of their depths can be made with currently available data. They are aligned in chains and, in some cases, in parallel sets of aligned chains. Martin et al. (2017) concluded that the most plausible explanation for the pits on

Enceladus is drainage of material into extensional fractures. These authors further attributed the occurrence of these pits in distinct parallel sets to a spatially variable stress field, which in some places was strong enough to create parallel sets.

4. Discussion

4.1. Craters of the Moon

4.1.1. King’s Bowl

King’s Bowl is an oddly shaped pit crater, neither fully circular nor elliptical but more almond shaped; this shape may be due to the proposed phreatomagmatic explosion invoked by

Hughes et al. (2018) that excavated the contents of the pit. The almond shape could be due to the relative weakness of the rock along the rift axis (for which King’s bowl is directly above), allowing the explosion to excavate more material linearly along that rift than perpendicular to the rift. The walls of King’s Bowl are relatively vertical and made up of basaltic volcanic flows, both

75 indications of higher rock strength than loose, unconsolidated material (ash or tephra), or weakly welded cinder deposits.

The condition outlined by Hughes et al. (2018) for an explosive pit origin requires that the country rock be hot enough to flash the water to steam as the water table return to the area post-eruption. This scenario thus implies a relatively short time between eruption and water table return, which may be a limiting factor of explosive pits forming in many volcanic areas. A linear volcanic system, i.e., part of a rift, such as that at King’s bowl, may allow for faster migration of water back to where rock remains hot after intrusive activity, in contrast to a centralized volcanic system (i.e., a central volcano) in which surrounding rock would be heated evenly and radially. If explosive pits (those formed through phreatomagmatic explosions, or Plinian style eruptions) all have irregular shapes, such as King’s Bowl, then an irregular shape may be a useful indicator for determining if pits on other planets where groundwater is suspected, are explosive or not.

The presence of depressions along the volcanic vent axis on which King’s Bowl is situated suggests that other episodes of magma/lava subsurface withdrawal occurred in the area but did not result in similar phreatomagmatic explosive events. The forms of these other depressions resemble those of inflation clefts or lava push-ups, which then experienced a deflation or cooling of the cleft/push-up. These depressions, even though I do not classify them as pit craters because the inflated central parts have not fully collapsed (i.e. the majority of the lava flow is intact, but fractured on the top), show that small, quasi-circular features that lack rims could also be present due to withdrawal of magma, rather than the draining of material into the subsurface along faults, and should not be overlooked when considering the environment in which pit craters have formed within a given area.

Overall, King’s Bowl is aligned with a regional rift, much like the pits found in other field sites studied here. The alignment of these pits with the regional rift supports the assertion that they are related to the formation of these volcanic vents, and are not due to a secondary factor, such as faulting unrelated to the formation of the vents (i.e., from the regional stresses in the Basin and Range). King’s Bowl is the only pit studied in this dissertation that has an explosive origin, and I attribute the almond shape to be a result of that formation mechanism.

Considering this is the only explosive pit studied, perhaps explosive pits are not as common as collapse pits, and this is important to remember when interpreting these features on other planets.

4.1.2. Yellowjacket Waterhole

Yellowjacket Waterhole has nine pits that are conical in shape and that formed in mostly unconsolidated weakly welded cinder and tephra deposits. These pits’ shapes and the material in which they formed point to weak material strength (Table 2.1), with collapse of that material infilling the pits to form an inverted cone shape and coming to rest near or below the angle of repose. The presence of thin basaltic flows at the stratigraphic tops of these pits indicates that if a strong-over-weak arrangement of layers occurs (e.g., by later lavas flowing over ash deposits), the weak layering will dictate the pit shape more than the strong layering.

The cross-cutting relations between the Yellowjacket Waterhole pits and a volcanic cone

(Figure 3.7B, C) indicate that the pits formed during or after the formation of that cone. The cone may be the source of, or at least contributed to, the cinder deposits in which the pits formed, on the basis of their proximity. The presence of the volcanic cone near the pits means that collapse of overlying material could be due to void space created from magma withdrawal, but no clear evidence was found for such withdrawal like lava coating the walls of the pits, clear lava

77 flows present throughout the pit walls, etc. A small fracture network (with <1 m horizontal displacement) is parallel to the pits on the western side of the chain (Figure 3.6, 3.7D), but the small horizontal offset of the fractures alone (~1 m) could not provide enough space to open the large pits (i.e., 10s of meters in diameter) here. If there are some underlying fractures in this area that accommodated the formation and growth of the Yellowjacket Waterhole pits, I did not observe any evidence for them.

4.1.3. Coyote Lake and Coltrell’s Blowout

Coyote Lake and Coltrell’s Blowout both have similar morphologies that I interpret to be akin to a lava lake building topography around itself over time. The presence of thin lava flows

(between ~25 cm and <50 cm in most cases) in the uppermost exposed walls of each of these long linear features (Figure 3.9B, C; 3.10C, D) suggests that most flows erupted from this depression were not large volume. Rather, it is possible that any lava lake hosted here would have overtopped its banks, creating the small (1–2 m-wide) lava channels seen in the field, and consecutively build up its levees over time. Analogous structures have been noted on the Moon in southeastern Mare Serenitatis (Brent Garry, personal communication), with linear, central, elongated pit-like depressions and raised topography relative to their surroundings.

Both structures are aligned with the regional rift axis in CRMO, further supporting the idea that they were once volcanic vents contributing to the regional lava flows. Unlike King’s

Bowl and Yellowjacket Waterhole, Coltrell’s Blowout and Coyote Lake are interpreted to be in much older lava flows due to the high level of vegetation, although there has been no independent age dating to support this assertion. The coalesced nature of these pit led me to visiting them in the field, but it is clear that they are not collapse related but rather volcanically

78 related. These features show that imagery alone can be deceiving when interpreting pits on other planets. A surface feature that appears to be a coalesced pit might have a raised rim, and without topographic data with sufficient resolution to discern that raised rim, a false interpretation of the formation mechanism would be easy to make.

4.2. Hawaii Volcanoes National Park

4.2.1 Kilauea Iki

The finding of a new addition to the pit chain at Kilauea Iki is one of the most important observations of this study. The emergence of a new pit due, presumably, to eruptive activity and associated earthquakes during the 2018 Kilauea East Rift Zone eruption is the first time that pit formation has been documented at the detail allowed by the high spatial resolution (~100 pts/m) of the lidar data flown during the eruption. (Previously, Devil’s Throat was documented with visual observations and no temporal measurements exist).

These pits are situated within cinder and tephra deposits from the 1959 Kilauea Iki lava fountain, deposits that appear mildly welded but mostly unconsolidated, allowing for the pits to be conical in shape due to the weak nature of those deposits (at least for the pre-existing pits).

The new pit is cylindrical, unlike the pre-existing pits within the chain, which is not necessarily expected given that the material in which this pit has formed is mechanically weak, welded pyroclastic material. Tracking any further growth of this pit would provide valuable information regarding pit morphology and whether, and how long, it takes to become conical.

The small earthquakes that were recorded in this area, and the proximity to multiple faults mapped by the USGS (see Figure 3.11), provide a basis for interpreting these pits as resulting from the collapse of surficial material into a ring-fault (or, more like, a set of

79 circumferential faults) bounding the Kilauea caldera. As the caldera collapse progressed during the draining and evacuation of the lava lake (May–August 2018), earthquakes became abundant around the summit in response to the deflation (Neal et al. 2019). This activity probably promoted the creation of the additional pit observed at Kilauea Iki. Earthquake data provide locations and orientations of fault planes in the area (Figure 3.11), with two falling within 500 m of the pit chain. Even though no fault is directly observed in association with these pits, it is possible they are forming above a syn- or antithetic fault splaying off one of the main faults in the area that does not have a complementary surface break to indicate their presence at depth.

4.2.2. Devil’s Throat and Twin Pits

The largest pits found near Kilauea are Devil’s Throat and the Twin Pits in the Ka’u

Desert, all of which are cylindrical shaped. Devil’s Throat is one of the most studied and well- known pit craters on Hawaii since its formation in the mid-1900s (E.G., Jaggar, 1947). The geological basis for the formation of Devil’s Throat is well laid out by Okubo and Schulz (1998), and our fieldwork yielded observations that supported their interpretation of collapse of material into a void space created by an evacuated magma chamber bound by a fracture system. The Twin

Pits appear to have a similar formation mechanism to Devil’s Throat, as the former have similar cylindrical shapes and sizes, and have formed within stacked lava flows but lack the presence of fractures around the periphery as those seen at Devil’s Throat. The cinder cone located just northeast of the Twin Pits, and the conduit seen in the bottom of the northeastern pit (Figure

3.16B), indicate that there was probably underground storage of magma just below the surface at the pits, that was ultimately drained due to the cone eruption or subsequent drainage elsewhere after eruptive activity ended.

The walls of these pits are made up of stacked basaltic lava flows (i.e., mechanically strong material, Table 2.1) with little to no observable deposits of interbedded ash, tephra, or cinder (i.e., weak material). It is therefore likely that the strong basalt walls allow for the cylindrical, nearly vertical walls of these pits. The Devil’s Throat pit is <100 years old (Okubo &

Martel, 1998) and maintains straight vertical walls, although it is clear that material has fallen off the walls and into the pit (both from the accounts of the pit widening, and visual evidence of blocky materials at the base (Figure 3.14B). The Twin Pits are similar, in that they are in strong basaltic lavas and have near-vertical walls. The western pit is partially infilled with talus and debris but overall has a mostly near-vertical appearance. The observable walls in the northeastern pit are circular in planform shape, and the 3D shape is cylindrical. Only the upper walls have circular planform shapes in the southwestern pit, and cross-sectional shape from there down is altered by the talus debris that almost reaches the top of the pit.

The timeframe since formation and shape are important factors in assessing the possibility that originally vertically-walled pits eventually become inverted cones due to mass wasting of wall materials. Strong materials like the basalt flows that make up Devil’s Throat and

Twin Pits will likely resist collapse for longer than weaker materials such as weakly welded pyroclastic deposits at Kilauea Iki and Yellowjacket Waterhole. Sustained observations of these examples, as well as for the brand-new pit in the Kilauea Iki pit chain, will provide good time- series data for how pit morphology changes over time. For example, there are stories (per communication with park officials) that Devil’s Throat was at one point small enough to jump a horse over (~1950’s), so the large straight walled pits are being modified currently, and tracking that change will be key in determining how long a cylindrical pit takes to acquire an inverted cone shape.

4.3. Planetary Pits

4.3.1 Noctis Labyrinthus, Mars

Noctis Labyrinthus, Mars is host to many pit craters (Chapter 2). The pits there formed in basaltic flood lava flows that are from the Late Hesperian (3.4 Ga–3.0 Ga) (Tanaka et al., 2014).

These pits are much older than the Earth analogues studied here, so there is a considerable gap in our temporal understanding of how pits might come to assume final shapes that are inverted cones. Noctis Labyrinthus is also dissected by hundreds of troughs I interpret as the morphological continuation of pit craters when volatiles stored at depth are released (e.g., ice sublimating due to exposure to the Martian atmosphere) (Chapter 2). However, there is no direct evidence of pit formation in Noctis Labyrinthus from volatile outgassing from the underlying sediments; instead, these pits are found in a region dissected with graben and normal faults, which likely drove the formation of the pits here.

4.3.2. Nyx Mons, Venus

The occurrence of pits on the flanks of a Venusian volcano suggests a volcano tectonic origin for these features. Many pits are aligned in chains, and some additional collapse or volcanic vent features were noted as well, indicating that these pits could be due to radial diking from Nyx Mons during the emplacement and evolution of the underlying magma system

(McGovern and Solomon 1998; Herrick et al. 2005). However, there is no evidence of lava flows emanating from the pits themselves, like those seen at King’s Bowl, ID. These Venusian pits could also have a purely tectonic origin due to extensional stresses developing during inflation of the magma chamber, but associated fracturing located near or around the pits mapped in this study was not noted on the radar data. Other studies of pit craters on volcano flanks have

82 concluded that even while in the presence of a volcanic center, many pits lack evidence for a volcanic formation (Byrne et al. 2012). Ascraeus Mons exhibits many pit craters located radially from the volcano and associated with graben, but lack any evidence for a volcanic origin, indicating that dilation faulting is the most likely explanation for the formation of the majority of the pits there (Byrne et al. 2012).

4.3.3. Pits from Literature

The pits described on small bodies and moons give useful insight into the processes that could create pit-like structures. Termed ‘grooves’ on Gaspra, Ida, and Phobos, these pits occur in chains, have circular to scalloped edges, and have been attributed to either 1) secondary cratering from a larger impactor, or 2) drainage of material into extensional fractures (Section 1.1).

Extensional fractures on Phobos have been linked to tidal stresses imparted by Mars, its parent body (Proctor et al., 2003). Gaspra and Ida both have pits that are attributed to secondary impact cratering (Veverka et al., 1992; Sullivan et al., 1996). Procktor et al. (2003) showed evidence for pit formation on Eros, a near-Earth asteroid, forming from drainage into extensional fractures.

Buczkowski et al. (2016, 2016) investigated pits linear trough features on Vesta and Ceres and determined that they also resemble drainage of material into subsurface fracturs on those bodies.

Martin et al. (2017) attributed the pits on Enceladus to also being due to drainage into extensional fractures. Together, these observations suggest that, even on small bodies, the occurrence of pits can be attributed to extensional fracturing and drainage of material into the subsurface. For Phobos, Eros, and Enceladus, fracturing was attributed to tidal and regional stresses. Depressions on Ida and Gaspra, on the other hand, might not be pit craters by the definition I am using (circular to elliptical), but the coalesced nature of the depressions and lack

83 of raised rim gives suspicion that they are coalesced pit craters and not another feature such as an impact crater. Without further knowledge of these small bodies in the form of global imagery and topographic data, the lack of the raised rims along these features lead to a tectonic origin for the depression formation. Dilation faulting on these small bodies would require there to be relatively defined stratigraphy to produce a changing dip angle—information that is not available at this time. The lack of raised rims is the most plausible evidence to use to argue that these are indeed from dilational faulting or infilling of preexisting fractures as these small bodies evolved over time.

4.4. Shapes

I have shown examples of both inverted conical and cylindrical pits, together with descriptions of where these pits are found. The shape of pits appears to be causally correlated with the geologic materials in which they exist and previous literature on modelling of pit craters comes to the same result (Wyrick & Smart, 2009; Smart et al., 2011). Smart et al. (2011) found that generally if a pit is forming in material with different mechanical strengths, it tends to be wider and deeper than if a pit is forming in a material with consistent strength, such as regolith.

Table 2.1 illustrates the different material properties for what I interpret to be the materials in which pits are forming in. The tensile strengths are most notable here, considering the sloughing of material is a tension-dominated process. Basalt is the strongest of the materials, with a tensile strength of -14.5 MPa while weakly welded pyroclastic materials only had a tensile strength of -

1.2. This illustrates how weakly welded pyroclastic are more readily eroded from pit walls, creating a more conical shape as more material falls off the pit walls and builds up talus slopes,

84 whereas basaltic lavas are 10x stronger, indicating it takes much longer to erode those materials from pit walls and onto the floor of the pit.

Of the pits visited in the field for this study, King’s Bowl, Devil’s Throat, and the Twin

Pits are all found in competent basaltic lava flows of varying thicknesses, which I take to represent the basaltic flows reported in Schultz (1995) (Table 2.1), therefore relatively strong.

On the other hand, Yellowjacket Waterhole and the pits on the flanks of Kilauea Iki I take to be better represented by weakly welded pyroclastic rocks reported by del Potro & Hürlimann

(2008), therefore relatively weak. Recognizing these differences then allowed to infer that pit shape is a least in part, due to the materials in which the pits are forming as King’s Bowl,

Coltrell’s Blowout and Coyote Lake are all more vertically walled, and Yellowjacket Waterhole are more conical. The pits at Yellowjacket Waterhole are the only ones with observable material difference within the pit stratigraphy (strong, thin lava flows over weaker welded pyroclastics), and they are the largest conical pits on Earth in this study, consistent with the findings of Smart et al. (2011) on mechanical stratigraphy within pits produces wider and deeper pits.

The observation of cylindrical pit formation in weakly welded material at Kilauea Iki does not support a direct connection between pit shape and material strength, but rather indicates that an inverted cone pit shape may be a function of time, such that newly formed pits could be cylindrical, gradually acquiring an inverted conical shape through some combination of mass wasting, weathering, and infilling. This possibility has important implications, because pits on other planets with cylindrical shapes may be geologically young, reflecting recent or even ongoing modification or formation by faulting or intrusive activity in the near subsurface.

Whereas pits with inverted cone shapes may be geologically old, or just forming in relatively incompetent material and assuming inverted cone shapes quickly.

In general, conical pits on Earth are found in mostly weakly welded pyroclastic materials with low strengths (Table 2.1), which allows for the material to come to a rest at the angle of repose faster than for pits formed within relatively strong lava flows. The finding of a new pit in mostly unconsolidated cinder and tephra deposits with a cylindrical shape drives my interpretation that pits may all form as a cylinder originally, but, with time, eventually develop walls that come to rest at the angle of repose to form inverted cones. The material that the pits form in, i.e. weak or strong, will dictate how fast the pit can become an inverted cone and assume the shape of an inverted cone. Stronger materials (i.e., lava flows) will maintain steeper walls for longer periods of time, and weaker materials will more readily fall in and create a conical shape.

To further investigate pit shape and how sediment infilling and erosion could make an originally cylindrical pit into an inverted cone pit, cylindrical and conical volumes for each pit with available depth and diameter data were calculated. These volumes are shown in Appendix

A and were used to calculate the volume of sediment required to change a cylindrical pit into a conical pit. On average, the pits from Mars and Venus, if originally cylindrical would require 5.4 km3 of accommodation space to form them, opposed to 1.5 km3 of accommodation space if originally inverted cone shaped. If a cylindrical pit were to be infilled, 3.9 km3 of sediment and debris would be required to transform that cylindrical pit into a conical pit. On Earth, cylinder pits would on average be 34 m3 in volume and inverted cone pits would be 17.8 m3 in volume, leading to an average infilling required of 16.2 m3. To compare the differences in Mars/Venus pits with Earth pits, the infilling was transformed into a unitless magnitude, derived by dividing the infilling required by the volume o f that pit if it were a cone (Appendix A). Mars and Venus

86 pits on average require 6.7x the volume of the cone to transform it from a cylinder into a cone, whereas an Earth pit only requires 2.6x the volume of the cone to transform it.

It seems unlikely from these calculations that most pits form as a cylinder originally, due to the extremely large amount of infilling required to transform that cylinder into an inverted cone shape. In terms of original accommodation space for spits to form from these calculations, cylinder require a very large amount of space to form, and seems unlikely that a dike tip or fault could create that many cubic kilometers of void space near the surface for a pit to form within.

Rather, an inverted cone takes much less accommodation space to form and is the more likely scenario. This analysis does not rule out the observations on Kilauea Iki of a cylindrical pit forming within a chain of inverted cone pits. The formation of pits must be locally dependent on the material it is forming in and if the accommodation space becomes all available at once or is made in small increments over time.

4.5. Statistical Analysis of Available Pit Crater Depth/Diameter Data

The morphometric data for pit crater shapes (depth and diameter) I collected for Noctis

Labyrinthus, Mars and Nyx Mons, Venus and data collected from the literature for Mars (Wyrick et al. 2004) and for Earth (Whitten & Martin 2019) were used to compare different assemblages to one another. These data sets were chosen because the pits were mapped out systematically in each case. The mean depth/diameter ratio for each assemblage was compared using the multcompare function in Matlab™, which takes the result from an analysis of variance test

(ANOVA) and will report which groups are statistically similar/different. The ANOVA test is only appropriate if the data is normally distributed, which I find the four groups to be (Figure

3.18). The histograms of the depth/diameter ratios are displayed along with probability density

87 functions to show the shape of the distributions. The means and uncertainty in those means are displayed on the histograms to give further confidence that these data are normally distributed.

The pits from Whitten & Martin (2019) have the highest depth/diameter ratio, 0.6375 ± 0.0353, while the pits from Noctis Labyrinthus, Mars have the lowest, 0.1333 ± 0.0042. The normally distributed data allow to make statistically informed interpretation of the different pit groups and provide discussion for why they may be similar or different. The results of the multcompare test are shown in Figure 3.19. Depth/diameter ratios for the Mars pits from Wyrick et al. (2004) and for Earth pits from Whitten & Martin (2019) are statistically similar, while the depth/diameter ratios for pits collected at Noctis Labyrinthus, Mars and Nyx Mons, Venus in this study are not similar to any of the other three groups.

Furthermore, the size frequency distributions of each the depth and diameter of the four data sets were evaluated to determine if those properties have characteristic distributions, regardless of planet or local. A size frequency distribution shows the cumulative total of pits that have a specific depth/diameter or greater, so 100% of the other pits greater than or equal to that smallest pit depth recorded. (or the same for diameter). Figures 3.20 and 3.21 show the diameter and depth frequency diagrams for the four groups of pits analyzed. The data (depth or diameter) for each group was then fitted with a negative exponential function to see if all the group have similar distributions. The negative exponential functions are shown in red on each diagram, and the equations for those are displayed as well.

The pit groups compared in this chapter display variability in pit crater depth diameter ratios from planet to planet. The only two groups that had similar depth/diameter means were

Martian pits from Wyrick et al. (2004) (0.5867 ± 0.0042) and Earth pits (in Iceland) from

Whitten & Martin (2019) (0.6375 ± 0.0353) (Figure 3.19). The two groups of pits I collected in

88 this study from Noctis Labyrinthus, Mars and Nyx Mons, Venus both had lower depth/diameter ratios than Wyrick et al. (2004) and Whitten & Martin (2019), with Noctis Labyrinthus having the lowest ratio of all of the groups (0.1333). Noctis Labyrinthus may have the lowest depth/diameter ratio in part due the conditions present within the subsurface that is discussed in

Chapter 2, some amount of volatiles that are dissecting pits of a certain depth and larger and producing troughs that are not morphologically similar to pits. The pits at Nyx Mons, Venus were measured using photoclinometry on Magellan SAR data, so further work to determine the error on these measurements will allow for better interpretation of those pits. Pits from this study at Craters of the Moon National Monument and Preserve and Hawaii Volcanoes National Park were not included because the size of those datasets was not great enough to conduct statistical analysis on, as data sets with 10’s of data points are more robust for making those comparisons.

The size-frequency distributions (Figures 3.20, 3.21) for the groups of pits studied here all can be described by negative exponential functions. This is the first time that pits have been described this way. This is important as three different planets all have negative exponential functions that explain the pit on their surfaces and it may be that pits on all planetary surfaces can all be explained with negative exponential functions, like shown in this work.

5. Conclusions

Pit craters on Earth studied here give valuable insight into the processes that can form pit craters. Pits in Hawaii Volcanoes National Park and at Craters of the Moon National Monument and Preserve in Idaho show that circular-to-elliptical-shaped pits are common in volcanic settings on Earth. Both locations have pits in strong (i.e., lava) and weak (i.e., cinder/tephra) materials. In Hawaii, the activity of Kilauea promoted active formation and modification of the

89 pits there, allowing for in situ temporal observations of pit cratering development. Additionally, that pits have been identified on asteroids and moons shows that pit cratering is a Solar System- wide process.

In Hawaii, active pit formation during the Kilauea eruption in 2018 was captured by lidar monitoring flights for the Kilauea caldera, which underwent collapse after the lava lake evacuated due to the lower east rift zone eruption. In response, a ring fault around Kilauea had multiple earthquakes in the June–July 2018 time period, causing the formation of a brand-new pit within a pit chain already present and aligned with the overall system of caldera-bounding normal faults. The new pit, in contrast to the rest in the chain, formed with a cylindrical shape; the other pits are inverted cones. This observation implies that time is an important factor in the evolution of pit crater shape. Further monitoring of this new pit may help determine how time plays a factor in pit morphology modification after formation.

Devil’s Throat and Twin Pits in Hawaii are hosted in stacked lava flows and are mostly cylindrical shaped. All three of these examples appear to have had connections with magma movement in the near subsurface. The northeast Twin Pit seems to be connected to a nearby cinder cone via a conduit at the base of the pit, and Devil’s Throat overlies a (now withdrawn) dike tip, allowing for collapse (Okubo and Martel, 1998). The cylindrical shape and circular plan view of these volcanically related pits give a basis for describing pits on the flanks of volcanoes elsewhere in the Solar System like found on Nyx Mons, Venus. These pits with nearly vertical walls and that are only tens of years old suggest that, after forming, pits in stronger materials may take some time to develop a conical shape.

King’s Bowl in Idaho is a valuable example of one of the rarer formation mechanisms, phreatomagmatic explosion, but nonetheless occurs in an environment and region where pits of

90 tectonic as well as volcanic origin also occur. King’s Bowl also gives a basis for the shapes of phreatomagmatic pits being irregular and not circular to elliptical, in contrast to that of pits formed purely from tectonic deformation or from intrusive or lava lake activity. The

Yellowjacket Waterhole pit chain is an example of larger tectonic pits that was probably influenced by the volcanic vents in the immediate area, but have classic circular shape without a raised rim, like the majority of those pits observed on other planetary bodies. Coltrell’s Blowout and Coyote Lake probably formed from perched lava lakes, or vents with low effusion rates, building up the topography around them slowly. These features could be mistaken as coalesced pits from aerial or satellite images, but do not feature an underlying dilational normal fault(s) and/or a withdrawn magma conduit. The raised rims are the only clear distinguishing characteristic from pits formed by collapse or phreatomagmatic explosion, or pits formed by evacuation of a magma body.

Pit craters found in Noctis Labyrinthus and Venus are much larger than their counterparts on Earth. Although the presence of Earth-sized pits (i.e., <100 m in planform view) on Mars or

Venus is not out of the question, the radar image data available for Venus is at a vastly different spatial resolution than what is available for Mars (75 m/px and <1 m/px, respectively). Even with

HiRISE data for Mars, however, no pits have been observed on Mars with similar scale to those found on Earth. It is possible that pits on Earth might be able to reach the sizes of those on Venus and Mars but hydrological action, erosion, and recycling of the crust does not allow for them to continue to grow to such sizes (Ferrill et al. 2004). And perhaps, on Mars and Venus, there are pits that are small like those observed on Earth, but given the spatial radar image resolution for

Venus, and spatial coverage of HiRISE on Mars (< 5%), the data has not been collected to make these observations yet. Additionally, smaller pit craters on Venus or Mars may be infilled with

91 lava, sand, etc. (with the latter, for example, during a large Mars sandstorm), since the pits are perfectly suited to act as sediment traps. Wyrick et al. (2004) discusses how older pit craters will have lower depth/diameter ratios, such that an old pit on Mars should have a lower Dpeth diameter ratio than a new pit. Though, determining the ages of single pits or a pit chains on Mars is difficult without a sufficient cratering record in the vicinity to use.

This chapter explored the many forms of pit craters throughout the Solar System, including two detailed field studies conducted at Hawaii Volcanoes National Park and Craters of the Moon national Monument and Preserve. The key new findings from this chapter are that the pit shape appears to be: 1) a factor of the material in which the pits are forming, with one potential outlier being the new pit that formed in the pit chain at Kilauea Iki, and 2) the geological environment in which the pits are forming can provide a starting point for interpreting how pits form (i.e. pits in volcanic environments tend to have volcanic related origin).

Figure 3.1. A global map showing locations of several accessible pit crater assemblages on Earth. A) The Big Island of Hawaii, where pit crater formation has been documented since the mid-1900s on the flanks of the Kilauea caldera. B) The U.S. state of Idaho, where the Craters of the Moon National Monument and Preserve is located, a site of recent basaltic volcanism (~2000 years ago). C) Utah, another U.S. state, where pit craters have been documented within graben in Canyonlands National Park. D) Iceland, situated on the Mid Atlantic Ridge, where pits formed during a rifting event in the 1970s in the deltaic plain near Asbyrgi Canyon.

Table 3.1 Collated Rock Strength Properties Uniaxial Young's Modulus(Gpa) Tensile Strength (Mpa) Compressive Strength (Mpa) Basalt (Schultz, 1995) 78 ± 19 -14.5 ± 3.3 266 ± 98 Strongly Welded Pyroclastic Rock (del 14.1 not measured 35.7–82.3 Porto & Hürlimann, 2008) Weakly Welded or interlocked Pyroclastic 3.4 -1.2 12.2 Rock (del Porto & Hülimann, 2008) Volcanic Soils (del Porto <1 0 <1 & Hürlimann, 2008)

Import photos and Build Mesh (height ground control point Build Texture field (2.5D)) GPS information

Build DEM Align Photos (low or Build Dense Cloud (Geographic, Dense medium quality) (high quality) Cloud)

Build Orthomosaic Mark ground control Align Photos (high or (Geographic, Mosaic, points manually ultrahigh quality) Refine Seamlines, Enable Hole Filling)

Optimize Cameras Optimize Cameras Export Orthomosaic (GCPs unchecked, (GCPs checked, and DEM Images Checked) Images Unchecked)

Figure 3.2. An Agisoft Metashape™ Processing workflow for uncrewed aerial systems data to produce orthomosaic images and digital surface models.

Figure 3.3. A formation diagram from Hughes et al. (2018) for King's Bowl where the structure is interpreted as a phreatomagmatic explosion pit. Ground cracking in relation to fissure width and depth is also illustrated.

Figure 3.4. Lava recession textures noted in the bottom of King’s Bowl, Idaho at the northern cleft. Orange lines trace the recession streaks, handheld radio for scale in bottom right.

Figure 3.5. Mapping results from King's Bowl in Craters of the Moon National Monument and Preserve overlaid on the orthomosaic generated from UAS data. Fractures are shown in green and pit craters and depressions in teal. Darker areas are from cloud shadows present moving through the scene during the UAS flights.

Figure 3.6. Yellowjacket waterhole UAS DSM with pits (black) and fractures (blue) mapped.

Figure 3.7. Field photos from Yellowjacket Waterhole illustrating the important features identified in the field. A) A view into a pit crater, comprising unconsolidated material (probably tephra), beneath more competent lava flows. The rubble in the pit is <1m in size, and is made up of blocks of those overlying lava flows. B) A view of three pit craters in the chain looking north, the tees near pits are 2–4 m tall. C) A view of the rightmost pit crater in panel B, with a cinder cone in the background; the trees in view here are 2–4 m tall. D) A view looking south (with the pits from panel B behind the camera view) showing fractures that parallel the pit chain.

100

Figure 3.8. Coyote lake UAS-derived DSM (left) and orthomosaic (right) used for analysis and interpretation.

101

Figure 3.9. Field photos of the Coyote Lake feature. A) A view looking south from the northern edge of the pit crater; the rocks in the foreground are 10–30 cm across. B) A view of thin lava flows in walls of pit, flows are 15–45 meters away from camera location. C) A view looking west from the pit crater floor up at blocky rubble along the western margin. Some blocks are made up of more than one flow; the blocks in the foreground are 75–150 cm across. D) A photo showing an old lava flow on the flank of Coyote Lake.

102

Figure 3.10. Coltrell’s Blowout UAS DSM (left) and orthomosaic (right), used for analysis and interpretation.

103

Figure 3.11. The topography of the southeast corner of the Kilauea Caldera. Inset shows location of the field photos in Figure 11 and the maps shown in Figure 12 . Earthquakes between June and July 2018 are plotted based on their magnitude (color) and depth to rupture (circle diameter), and faults from USGS Quaternary Fault database plotted in black.

104

Figure 3.12. Handheld photos of pits contained within the Kilauea Iki pit crater chain showing important findings. A) The northern end of the chain looking south; pits here are 1–2 m across. B) The view looking north of a large pit containing two smaller pits, which are 50 cm to 1 m across. C) A view looking south showing a pit with an overhang after Kilauea eruption, P. Byrne for scale (~1.8 m tall). D) A view looking north of a vegetated pit with distinct margins; the Kilauea Iki cinder cone visible behind (background center). E) A view of a brand-new pit that formed after the Kilauea 2018 eruption, looking south towards the other pits (Photo courtesy of Dr. Bob Craddock, Smithsonian Institution.) F) A close-up view of the inner WNW wall of this new pit, seen in October 2018. Note the stratigraphy in the pit walls, and the overhang to the right.

105

Figure 3.13. Topographic maps of the Kilauea Iki pit crater chain, from June (left) and July (center) 2018. The difference between the two surveys is shown in the right panel, illustrating the formation of the new pit crater on the northern (top) end of the chain (red outline).

106

Figure 3.14. Handheld photos and a 3D model of Devil’s Throat pit crater on Hawaii. A) A view looking north– northeast illustrating the many stacked lava flows that make up the walls; the trees on the edge of the pit are 1.5–2.5 m tall. B) A view of the talus at the bottom of pit; rubble pieces are up to 1 m across. C) The view showing fracturing on the southern edge of pit. This fracture is 5–10 cm across. D) An additional view of the pit, looking to the XXX. E) A top-down view of the 3D model, showing camera locations determined in Agisoft Metashape, and the cylindrical 3d shape of the pit.

107

Figure 3.15. Handheld images and a 3D model of the western pit in the Twin Pits group located in the Ka’u Desert of Hawaii Volcanoes National Park. A) A view looking north (with Mauna Loa in the background), with fracturing along the edge of the pit; these fractures are 1–10 cm across. B) A view to the south, showing stacked lava flows within the pit wall, a younger flow cascading into pit, and talus in the floor (~50 cm or less in size). C) View looking east showing eastern pit and cinder cone in background; Cinder cone is ~300 m away. D) Top down view of 3d model. E) Side view of the 3d model looking to the northeast, illustrating the stacked talus on the left. F) Side view of the pit, looking to the southeast, illustrating the irregular shape.

108

Figure 3.16. Handheld images and a 3D model of the eastern pit in the Twin Pits group in the Ka’u Desert of Hawaii Volcanoes National Park. A) The view to the northwest showing the western pit, with Mauna Loa in the background. B) Looking east, with cinder cone ~150 m away. C) A view looking southwest that shows the stacked flows in the uppermost pit walls; here, the wall is ~50 m across. D) The view into the eastern pit, showing lava flows draping over the sides of the pit, as well as what appears to be an exposed conduit leading out of the pit. The conduit opening is ~10–20 m across. E) A top-down view of the 3D model developed for this pit, looking to the northeast. F) A side view of the 3D model, showing the mostly cylindrical shape of the pit.

109

Figure 3.17. A) A regional map of Nyx Mons, Venus, showing the pit craters I mapped (n = 312). The radar look direction is from the left. B) An inset map of one pit crater in the region, showing the radar shadow inside of one pit I used for depth estimation. C) A schematic diagram showing how the shadow measurements are used (adapted from Wyrick et al. (2004). See text for details.

110

Figure 3.18. Histograms of available pit crater data depth/diameter ratios. Earth pits are from Whitten & Martin (2019), Mars pits are from Wyrick et al. (2004), Noctis pit and Venus pit data collected in this study. Probability density functions shown for each population (and Noctis overlaid on Mars for comparison). Mean depth/diameter ratio and uncertainty shown for each histogram.

111

Figure 3.19. Results of an analysis of variance (ANOVA) multicompare test. Blue points (Whitten & Martin (2019) and Wyrick et al. (2004)) indicate populations that are similar, while red points (Noctis Labyrinthus, Mars and Nyx Mons, Venus; this study) are populations with no similarity to the other three.

112

Figure 3.20. Diameter Frequency plot for each group of pits analyzed. Negative exponential functions for each group displayed in red showing the tendency of pit crater diameter to be explained by negative exponential functions.

113

Figure 3.21. Depth frequency plot for each group of pits analyzed. Negative exponential functions for each group displayed showing the tendency of pit crater depth to be explained by negative exponential functions like the depths.

114

CHAPTER 4

Tectonic Sinkholes and the Fate of Water in The Grabens region of Canyonlands National

Park, Utah

Chapter 4 is being prepared for submission to the journal Geomorphology, with the following authors:

Corbin L. Kling1; Wegmann, K.W.1; Williams, Z.W.1; Byrne, P.K.1

1 Marine, Earth, and Atmospheric Sciences Department, North Carolina State University,

Raleigh, NC

115

1. Introduction

1.1. Background

Canyonlands National Park in southeastern Utah (Figure 4.1) hosts a well-developed graben system, aptly named The Grabens, that contains numerous sinkholes previously identified via aerial imagery (Kettermann et al. 2015) (Table 1). The graben system here is composed of normal faults that parallel the Colorado River, in a linear to curvilinear map pattern trending

NNE. The faults are due to the dissolution of the evaporites in the Pennsylvanian Paradox

Formation, which underlies the rock strata in which the faults are manifest, viscously deforming and slumping to the NNW. This viscous movement is set up by the Monument Upwarp, which is a regional Laramide- (early Tertiary) monocline structure (Goldstrand, 1994; Condon, 1997;

Walsh & Schultz-Ela, 2003). The limbs of this fold form regional slopes that vary from 2°–4° to the NNW within The Grabens region (McGill & Stromquist, 1979; Huntoon 1982; Trudgill &

Cartwright, 1994; Walsh & Schultz-Ela, 2003; Furuya et al., 2007), promoting the slumping of the viscous Paradox Fm. as the Colorado River incises down through it. As the Paradox Fm. viscously flows NNW downdip towards the Colorado River, the overlying rock mass deforms in a brittle manner, producing the normal faults and grabens seen at the surface (McGill &

Stromquist, 1979).

The geology of this region is composed of five primary stratigraphic units, including (in increasing stratigraphic age): alluvium, Cedar Mesa Sandstone, Halgaito Shale–Elephant Canyon

Transition, Honaker Trail Fm., and the Paradox Fm. Alluvium is found on the graben floors and varies in thickness; for example, Billingsly et al. (2002) found ponded deposits of this material in the field, indicative of thicknesses up to 4 m in some areas, whereas Kettermann et al. (2015) carried out ground-penetrating radar (GPR) surveys in the area, finding average alluvium

116 thicknesses of 2 meters. The only rock unit within the study region observable at the surface is

Cedar Mesa Sandstone, which is a ~100–150 meter-thick massive cross-bedded sandstone with cherty limestone beds in some areas (Baars, 1962). Below that unit, the Halgaito Shale–Elephant

Canyon Transition is a variable-thickness shale with minor sandstone and limestone layers in the upper parts (Baars, 1962). This unit is 10–50 meters thick in The Grabens region. The Paradox

Fm. is the lowermost unit of interest for this study, and is a sequence of evaporite salts that have a major impact on the salinity of the Colorado River (Tuttle & Grauch, 2009).

The sinkholes are located in the graben floors, often near or along the graben walls, as opposed to the center of the graben (Figure 4.1). Kettermann et al. (2015) attributed the formation of these pits to Mode 1 fracturing at the surface. The Mode 1 fracturing was proposed to be due to a weak lubricating layer at depth, for which they inferred from field investigations in the northern section of The Grabens (Figure 4.2); although they did not establish the stratigraphic origin of the weak lubricating layer (Figure 4.3 and Figure 4.4). This weak lubricating layer results in dilational fault due to the changing dip angle as the fracture propagates through layers with different mechanical properties. As discussed in Chapter 2 and 3, dilational faulting results when the dip angle of a normal fault either shallows or steepens at depth (due to changes in material properties of the rock types through which the fault is penetrating) (Ferrill & Morris, 2003).

Additionally, The Grabens region is crosscut by four catchments (Butler Wash, Y

Canyon, Cross Canyon, and one unnamed canyon; Figure 4.1D) that are heavily influenced by the evolution of the faults in the area (Ollier, 1981). These catchments are considered to be ephemeral and only have flowing water during large-volume rainfall events a few times per year

(Trudgill, 2002); this region experiences 10–20 cm of rainfall annually (Nuckolls and McCulley,

117

1987). Sinkholes are often found in close association with fluvial features within the grabens

(Table 4.1). I use the term “channel” here to describe any feature that appears fluvial in nature

(i.e., streams courses, channels, and dendritically patterned ground). The sinkholes are not circular-to-elliptical in planform shape like many depressions elsewhere on Mars (Chapter 2) or

Earth (Chapter 3), but are irregular, scalloped, and often elongated in the downslope or fault strike direction. These irregular shapes and the downslope arrangement of the sinkholes implies that there is active secondary erosion within these pits, possibly due to interactions with surficial water.

The sinkholes in Canyonlands together represent one of the best known occurrence of sinkholes within well-developed grabens on Earth, much like we see on other planetary bodies

(Figure 1.1, Ch.1), and which are termed pit craters. The difference between pit craters and sinkholes lies in the fact that “sinkhole” implies a component of erosion by water, whereas “pit crater” is a catch-all term for collapse-related depressions regardless of formation mechanism.

Further, pit craters as a term is usually used in reference to tectonic or volcanic collapse pits, but is rarely used outside of the planetary science community. Here, we refer to these depressions as sinkholes, consistent with previous literature on the area and with the implied contribution by water to their present-day appearance. The similarities between the grabens and pit craters within

Noctis Labyrinthus, Mars and those within Canyonlands National Park are worth noting.

Determining the controls of non-karstic sinkhole formation and shape on Earth, especially regarding the role of fluvial and groundwater activity, helps when interpreting quasi-circular depressions on other planetary bodies.

118

1.2 Science Rational

The sinkholes in The Grabens region of Canyonlands National Park are well-preserved and accessible examples of sinkholes within a well-developed extensional system. As noted above, the sinkholes’ occurrence near or along fault traces, and their irregular shapes, help to distinguish these sinkholes from other depressions on Earth studied in Chapter 3. It is this observation of unusual depression shape that chiefly motivates the two fundamental science questions for this work:

1. Investigating the formation and evolution of these intra-graben topographic depressions is

important to understanding sinkhole formation elsewhere on Earth and in the Solar

System. What role does water play in the formation of these sinkholes?

2. The sinkholes in Canyonlands are clearly associated with and impacted by erosion and

nearby drainage networks. How do the sinkholes influence the four major catchments in

the region (and vice versa)? Do the sinkholes serve as conduits for surface water-to-

groundwater recharge in this arid setting?

2. Methods

2.1. Field Work

Field work was conducted in October 2019 to make in-situ observations, and to acquire ground- and uncrewed aerial systems (UAS)-based photographs of accessible sinkholes. Four sites were visited and documented, out of 28 identified on National Agricultural Imagery

Program (NAIP) data prior to commencing the field survey (Figure 4.1C) (Table 4.1). UAS surveys were conducted with a DJI Phantom 4 Quadcopter and supplemented with RTK GNSS photo-target ground control points (GCPs) collected with Emlid Reach RS+ differential GNSS

119 devices. At least eight to ten GCPs were placed at each field site prior to the UAS flights, and subsequently marked on UAS data to georeference the resultant orthomosaic and digital surface model (DSM). The UAS data were processed in Agisoft Metashape Professional v1.5.5 to create high-resolution orthomosaics and DSMs. Field observations of the sinkholes included observations of the stratigraphy (or lack thereof) within sinkhole walls, as well as the occurrence and properties of erosional features, tensional fractures, and vegetation in the area. Photographs were taken with handheld cameras at each sinkhole site.

2.2. GIS Analysis

NAIP aerial imagery (at a resolution of 60 cm per pixel), USGS National Elevation

Dataset (5 m/pixel), interferometric synthetic aperture radar (InSAR) data from the two

European Remote Sensing Satellites (ERS1/2) C-Band (~1 m/pixel), and UAS (10 cm/pixel) data were used in conjunction to evaluate the sinkholes and faults in the region. The NAIP imagery was used to identify sinkhole locations, and to plan field work at the four selected sites. All available historical aerial imagery were acquired and used to note any changes in sinkhole morphology through time. Historical aerial image data are available for the field study sites dating back to 1952 through the Utah Geological Survey Aerial Imagery Collection

(https://geodata.geology.utah.gov/imagery/), and include five years with little cloud cover and accurate georeferencing: 2006, 2009, 2011, 2014, and 2018. UAS DSMs and orthomosaics generated from UAS imagery collected during fieldwork were used to document changes in sinkhole shape and size between 2018 and 2019.

Topographic data from the processed UAS images (10 cm/pixel) and from the National

Elevation Database (NED) (5 m/pixel) were used to measure sinkhole depth, width, and length.

120

The NED DEM was used to extract catchment boundaries, longitudinal channel profiles, for the four main channel networks that drain the study area within the Grabens region, using the

Hydrology toolset in ArcMap 10.6. The trunk stream longitudinal profile from each catchment was used for assessment of the channel morphology relative to the spatial position of the faults and sinkholes and how both might be affecting the topographic evolution of the regional channels. Interferometric Synthetic Aperture Radar (InSAR) displacement maps (made with data from ERS1/2) from Furuya et al. (2007) were used to quantify average yearly surface displacements surrounding the sinkholes and along the nearby faults.

2.3 Fault Analysis

To test the hypothesis that a weak lubricating layer is responsible for the opening of dilatant fractures at the surface, as hypothesized by Kettermann et al. (2015), fault slip (Ts) and dilational tendency (Td) values were calculated for the stratigraphic units in the region, following the equations in Ferrill et al. (2017), which include:

푇푠 = 휏⁄𝜎푛 (1)

푇푑 = 𝜎푧푧 − 𝜎푛⁄𝜎푧푧 − 𝜎푥푥, (2) where τ is the shear stress and σn is the normal stress at the stratigraphic boundary, and σzz, and

σxx are the maximum and minimum principal compressive stresses (σ1 and σ3, respectively) for those stratigraphic boundaries in 2D space. In this region, σ1 is considered vertical (i.e., corresponding to the overburden), and σ3 is horizontal and perpendicular to fault strike. These analyses produce a value between 0 and 1, with 1 indicating that a fracture is more likely to slip, so helping to discern which rock units might be contributing to the dilation of the faults and causing subsurface drainage of material. In equation (1), slip occurs at Ts values greater than 0.6,

121 which is usually the coefficient of static friction for the rock in question—in this case the Cedar

Mesa Fm. and Halgaito Shale–Elephant Canyon Transition (Byerlee 1978; Morris et al. 1996;

Zoback et al. 2007)—because the frictional resistance to sliding must be met or exceeded to produce sliding. The frictional resistance to sliding depends on the cohesive strength of the rocks. Representative values of the required variables (τ, σzz, σxx, and σn) were calculated using equations from the literature on the geology of the area (McGill & Stromquist 1979; Moore &

Schultz 1999; Kettermann et al. 2015). Those equations are:

𝜎푧푧 = 𝜌푔푧 (3)

𝜎푥푥 = 푘𝜎푧푧 + 퐷 + 푇푏 (4) where in equation (3) ρ is the density of the rock, g is acceleration due to gravity, and z is the depth of interest for the slip and dilation tendency analyses. In equation (4), k is a constant relating horizontal stress to vertical load using the Poisson’s ratio of the rock, D is normal stress due to the downslope component of gravity, and Tb is the integrated shear stress on the base of the brittle plate (practically, the rock mass overlying the viscously deforming Paradox Fm.).

Further discussion of these equations and their use can be found in McGill & Stromquist (1979).

We utilized the fracture dip estimates (90° in the Cedar Mesa and 70° in the Halgaito Shale) that

McGill and Stromquist (1979) found as representative values of extensional structures in The

Grabens region.

Topographic profiles were drawn perpendicular to the fault traces at the locations of sinkholes to investigate fault displacement at those locations. Individual topographic profiles constructed from the 5 m/px USGS NED DEM (for longitudinal profiles) and UAS DSMs (for profiles of individual sinkholes) were used to measure the vertical offset along the fault (assumed

122 to be the displacement if fault dip = 90°) at the location of each sinkhole group for comparison with measured sinkhole depth(s) at that location.

2.4. Channel Analysis Methods

The use of Chi (χ) as a metric for comparing channels across catchment boundaries first originated with Perron and Royden (2013), who developed it to reduce noise in channel profile analysis. The analysis uses χ, which is an integral transform of the horizontal distance of a channel’s elevation profile. χ has units of length. Further description of this analysis’s mathematical constraints is available in Gallen and Wegmann (2017). Channels were analyzed with this method by utilizing ChiProfiler (Gallen and Wegmann, 2017), a set of codes for channel analysis within TopoToolbox 2 for Matlab (Schwanghart and Scherler, 2014). Chi

Profiler allows for interactive selection of channels based on an input DEM and parameters

(minimum accumulation threshold for channel flow, DEM smoothing window size, fill or carve topographic depressions, etc.). Once the channels are selected within the tool, the code prompts the user to select channel reaches on a plot of elevation versus . A channel reach for this study was considered any part of the plot with a consistent slope in -elevation space, such that faults crossing the trend of the channel profile at high angles would be captured as separate reaches for comparison. After the user selects the reaches, knickpoints, localized zones of increased channel gradient, may be selected and plotted.

After analysis of each study reach, the user can choose to export a shapefile of the channels and knickpoints to be further evaluated within a GIS. The exported shapefile for the channels includes the normalized channel steepness (ksn) and  values, which allows for a map view representation of those data. Normalized channel steepness (ksn) will be higher near

123 knickpoints or zones of active erosion. Chi values, when compared across channel catchment boundaries, may indicate which channels are expanding their catchment size through erosion

(low  values), and which channels are more likely to be overtaken by neighboring, faster- eroding channel networks (high  values).

3. Results

3.1. Field Work

We identified two types of sinkholes at the four field sties visited: (1) those with and (2) those without connections to the modern channel network. At Site 1, the sinkholes have no observable connections to an extent channel network, whereas at Sites 2–4 all sinkholes are connected to the local modern channel network.

3.1.1. Site 1

Site 1 is located in a small canyon south of the intersection of Imperial Valley and Cross

Canyon (Figure 4.5). The site has three sinkholes near the center of the graben (Figure 4.6). The easternmost sinkhole is circular in plan view, the westernmost is semi-circular, and the central sinkhole has an elongate elliptical shape with one vertical wall (Figure 4.5). The two circular-to- semi-circular sinkholes are within Quaternary alluvium, and the elliptical sinkhole shows one wall composed of Cedar Mesa Sandstone (the south side), and the northern (sloped) wall made up of Quaternary alluvium (Figure 4.7). The three sinkholes are between 0.7 and 7.5 meters deep, 5.8 to 32.6 meters wide, and 8.5 to 49.9 meters long (Table 4.2). There are no clear drainage networks or surficial water features intersecting any of the sinkholes at Site 1.

124

3.1.2. Site 2

Site 2 is located in Imperial Valley (Figure 4.1A) and has six sinkholes in a chain within, and parallel to the graben strike, and two additional sinkholes not aligned with the chain or with each other (Figure 4.8). The sinkholes are in the middle of the main graben and are just north of a smaller faulted block segment within the larger graben. The sinkholes have formed in

Quaternary alluvium and are heavily vegetated; all are partially filled with dead vegetation that limits depth estimations from the UAS DSM (Figure 4.9). The sinkholes are 0.9 to 4.1 meters deep, 3.2 to 11.6 meters wide, and 4.3 to 21.9 meters long. Nearby evidence of recent faulting within the graben floor was found along the smaller faulted bedrock block, where a joint was found with small drainages leading into it (Figure 4.10).

3.1.3. Site 3

Site 3 is the most complex and has the most incised channel system of the four investigated sites. Site 3 is in Cross Canyon and is composed of sinkholes and new channel initiation points (channel heads) (Figure 4.11). During fieldwork, three sinkholes were found at the site, and have varying shapes from elongate ellipses to a weakly dendritic outline. The sinkholes are 6.4 to 15 meters deep, 19.9 to 38.6 meters wide, and 69.8 to 136.0 meters long

(Table 4.2). The sinkholes and channel heads intersect near the middle of the canyon (Figure

4.11). On the southern side of the area, the sinkholes are elongated and parallel to the nearby fault wall (Figures 4.12 and 4.13). Farther north, the sinkholes cross the incised Cross Canyon and intersect the current Cross Canyon channel system. The Cross Canyon itself is a larger incised canyon (>100 m across), but the current channel initiation points within the site are much smaller (<5 m across).

125

Site 3 was the only site with noticeable morphological changes to sinkhole and channel shapes on the temporal NAIP and UAS imagery (Figure 4.14). Between 2014 and 2018, the western sinkhole linked up across the canyon to an area of surficial water drainage. Between

2011 and 2018, the channel heads propagated up to ~100 m upstream (Figure 4.14).

3.1.4. Site 4

Just north of Site 3 in Imperial Valley, there is a smaller group of sinkholes and drainages that are aligned parallel to the strike of their hosting graben (Figure 4.11). This site has two distinct sinkholes within a well-developed drainage network (Figure 4.15) that coalesces at the base of the relay ramp structure to the northeast. An open dilatant fracture (joint) ~1 m wide was found near the intersection of the drainages and the relay ramp. The three sinkholes are on the southern side of the drainage; two are within, and one outside, the current drainage network. All sinkholes here are within the middle of the graben and aligned parallel to the nearby graben wall

~125 meters away. (Figure 4.16).

3.2 GIS Analysis

Five regional catchments were extracted from the NED 5 m/px DEM. These catchments are known as Cross Canyon, Butler Wash, Y Canyon, and Bull Wash, catchment just to the south of The Grabens area, and a fifth, unnamed one (Figure 4.1D). Those catchment boundaries were then used together with the NED 5m DEM to extract the stream networks within each catchment, from which the trunk channels were identified. Topographic longitudinal profiles of each of the main five main channels were then plotted for qualitative comparison of their topographies.

126

The four trunk-channel longitudinal profiles from The Grabens exhibit primarily planar and convex channel segments (Figure 4.17A–D). Each of the four shows abrupt changes in channel elevation (gradient) that correlate to the locations of active faults (a relation verified by comparing the location along the longitudinal profile with the mapped faults). Figure 4.17A shows the location of sinkholes within Cross Canyon along the profile, but there is no distinct topographic feature such as a sudden elevation drop off (i.e., a knickpoint) observed in the longitudinal profile at the location of the sinkholes. Figure 4.17E, the longitudinal profile for

Bull Wash, displays a prominent knickpoint, and is more concave-up than the rest of the profiles

(Figure 4.17F).

The Cross Canyon catchment, which contains all of the sinkholes studied in this work

(Figure 4.1C, D), shows a high degree of drainage capture into subsurface fracture or sinkholes.

To quantify how much of the surface water is potentially redirected due to the presence of sinkholes at Site 3, the area of the whole catchment was compared with the area of the catchment upstream of the sinkhole locations in Cross Canyon (Figure 4.1D). These sinkholes account for the potential capture of 68% of the basin’s surface-water accumulation area.

3.3 Fault Analysis

The topographic profiles constructed for each study site were used to determine the amount of offset on the faults nearest to the sinkholes. At Site 1, the fault on the north side of the graben floor has a maximum vertical offset of 50 m (Figure 4.6). Site 2, the northern and southern faults on either side of the centrally located sinkhole chain have 110 m and 130 m of vertical offset, respectively. The fault just north of Site 3, located only 10 m from the sinkholes, has 44 m of vertical offset (Figure 4.12). The northern and southern faults on each side of Site 4

127 have 43 m and 25 m of offset, respectively (Figure 4.16). The relay ramp that the sinkholes and drainages intersect has no measureable displacement, although erosion at this location may have removed evidence for such offset.

The calculated dilational tendency for the Cedar Mesa Sandstone and the Halgaito Shale is 0.78 and 0.18, respectively, whereas the slip tendency for these two rock units is 0 and 0.89, respectively. Values closer to one indicate a higher probability for dilation or slip. Estimates for the dilational and slip tendency for the stratigraphic units in this area show that the Halgaito

Shale–Elephant Canyon Transition is more likely to slip, and the overlying Cedar Mesa

Sandstone is more likely to dilate. This is most likely due to no confining pressure at the surface allowing for vertical fractures to form.

InSAR data from Furuya et al. (2007) show evidence for subsidence along the western extent of the Grabens region (Figure 4.18a). At the location of Site 3 in Cross Canyon, a mean annual vertical displacement of ~ +1 mm/yr (positive is away from the satellite, i.e., subsidence) was observed, and the Cross Canyon area in general is trending towards more subsidence (green line 0–1mm/yr change, Figure 19) than the surrounding area along the E–E’ profile (Figure

4.18b; Figure 4.19, purple box).

3.4. Channel Analysis Results

The χ analysis allows for interpretation of the normalized channel steepness between channels since the ChiProfiler toolset makes it easy to maintain input parameters. Figure 4.20 shows the elevation versus χ plots for the trunk channels of the four catchments evaluated in the study region. The channels are demarkated by faulted segments and non-faulted segments that are visible on both aerial and topographic data. Those segments are also apparent on the

128 elevation versus χ plots in Figure 4.20, with faulted channel reaches having steeper gradients and non-faulted segments displaying shallower gradients. Table 4.3 provides the ksn values for each channel reach for easy comparison. The faulted segments have higher ksn values, with a mean of 210.4, whereas the non-faulted segments have lower ksn values with a mean of 67.7.

Figure 4.21 shows the  and ksn values plotted along the length of the channels for spatial comparison. The Chi values allow for the interpretation of how the catchments are interacting with one another, and generally, lower  values close to a catchment divide indicates that through time the low- catchment will consume the catchment with a higher  value on the otherside of the divide (black arrows on Figure 4.21 show this expected movement). The ksn increases at the locations of the knickpoints, which correspond to channel segments in proximity to normal faults (indicated by warmer colors, yellow, orange, and red). The southern catchments

(Cross Canyon and Bull Wash) have lower  values than the more northern catchments (Y

Canyon and Butler Wash).

The knickpoints can provide important information about tributary channel incision in response to base level fall of the Colorado River. Figure 4.23 shows the elevation of knickpoints above the Colorado River from the fixed elevation of 1185 m, the Butler Wash to Colorado

River confluence. The mapped knickpoints range in elevation from under 200 m (Y Canyon) to over 1 km (Bull Wash). Additionally, the faulted channels from The Grabens region all have more knickpoints than Bull Wash (except for Y Canyon).

129

4. Discussion

4.1. Sinkholes

The sinkholes investigated at the four field sites provide insight into the processes affecting both the development of the sinkholes as well as the channel systems that run through

The Grabens region. Site 1 has three sinkholes that are up to 7.5 meters deep, 32.6 meters wide,

49.9 meters long (Table 2). The central pit has Cedar Mesa Sandstone juxtaposed with

Quaternary alluvium making up the opposite walls of the pit. We interpret this sinkhole to have formed (and perhaps continuing to form) over a normal fault that is anti- or synthetic to the graben-bounding faults. That none of the three sinkholes here has any surficial water features leading into or out of them suggests that these depressions are purely tectonic in origin.

Site 2 has eight sinkholes, all within Quaternary alluvium, that are up to 4.1 meters deep,

11.6 meters wide, and 21.9 meters long (Table 4.2). We attribute the sinkholes forming here to be due to anti- or synthetic faulting (to the graben bounding faults) in the graben floor. In turn, the sinkholes have modified the channel drainage system, diverting the flow of water from a northeasterly trend through Imperial Valley to the intersection of the two interpreted faults along which the sinkholes lie (Figure 4.8), and where a joint with sediments and rocks draining into it is found.

The three sinkholes at Site 3 are the largest of all such depressions mapped at the field sites. These sinkholes are up to 15 meters deep, 38.6 meters wide, and 136.0 meters long (Table

2). The temporal analysis (Figure 4.14) reveals that the third pit (ID 4, Table 2) formed between

2014 and 2018, and that the channel heads moved up to 100 meters upstream from the pits between 2006 and 2019. The new pit clearly intersects the small channel system there (Figure

4.11). Although no sinkholes were found within the channel system where the sinkholes and

130 channels coalesce into a small alcove, I infer that a sinkhole(s) was in this location originally intercepting the above-ground flow of water.

Site 4 is much like Site 2 in that there are pits aligned with the local channel system and that have formed in Quaternary alluvium. The channel system itself appears to flow into open joints in the Cedar Mesa Sandstone (e.g., Figure 4.10). The pits here are 1.6 to 3.4 meters deep,

5.0 to 9.3 meters wide, and 5.9 to 11.4 meters long.

4.2. Channels

The longitudinal profiles for the four catchments in The Grabens region show considerable evidence for modification by faulting (Figure 4.17). Previous workers have focused on the impact of slip along the normal faults on the evolution of channel development specifically in this region (e.g., Trudgill, 2002; Commins et al., 2005). This channel analysis found that the Butler Wash and Cross Canyon catchments predate the onset of fault initiation and therefore record the entire period of faulting within the region. Evidence for numerous instances of channel rerouting because of faulting is present throughout Butler Wash, and closed basins

(i.e., a graben with no above-ground exit point for water) were also noted, showing that the faults dictate the flow paths of these channels as the slumping of the rock mass progresses down gradient towards the Colorado River (Trudgill, 2002).

The lower reaches of these four catchments are most affected by the westward movement of The Grabens region, as the Colorado River is responsible for initiating downhill flow of the

Paradox Fm. along the western limbs of the Monument Upwarp. The Bull Wash catchment, located just to the south of Cross Canyon (Figure 4.1D), is not obviously cross-cut by any faults,

131 so we chose to use it as a proxy for what the shape of a non-faulted longitudinal profile should look like in this area.

Before discussing the longitudinal profiles of the study streams from the Grabens region,

I think it is informative to evaluate the Bull Wash catchment c. 10 km to the south of Cross

Canyon (Figure 4.1D). The Bull Wash catchment serves as a null-hypothesis analog for the channels draining the Grabens region because the stratigraphy is the same, but Bull Wash is currently beyond the limit of the extensional faulting that defines the Grabens area. Specifically, then, the Bull Wash catchment allows us to test the inference that the four other catchments

(Cross Canyon, Y Canyon, Butler wash, and the unnamed canyon) are being impacted by the tectonics in the region. The Bull Wash longitudinal profile is concave up for ~20 km along its length before a knickpoint is reached, and a slot-canyon-style channel continues down to the

Colorado River (Figure 4.17). We find that the knickpoint along the Bull Wash watershed shows

800 m of topographic relief down to the Colorado River, which we take as the amount of vertical incision there. Regional incision rates for the Colorado River in this area of ~100–500 m/Myr

(Darling et al. 2012; Aslan et al. 2019) indicate that the Bull Wash knickpoint originated between 1–2 Ma. These ages are consistent with cosmogenic dating studies of fluvial river terraces along the Colorado River, which indicate that incision periods and rates along the

Colorado River itself can vary from ~1–16 Myr and from about five to more than 250 m/Myr, respectively (Aslan et al. 2019). It should be noted that the 500 m/Myr incision rate is used as an

5 6 upper bound for intermediate time-scale incision rates (~10 –10 yr), but short-time-scale (<100 kyr) incision rates can be as great as 3700 m/Myr depending on local climatic and catchment variables (Aslan et al. 2019).

132

The complexity of sinkholes and channels at Site 3 offers considerable insight into the interactions of surficial water and sinkholes in this region. Channel truncation (i.e., no further downstream flow at the surface) at the location of the sinkholes means all water upstream of that location is being drained into the subsurface there. The capture of all of the upstream surface water flow (68% of the total catchment area) at the location of the Site 3 sinkholes represents a substantial amount of surface water being redirected into the subsurface, given that the area is already dominated by many closed basins—each individual graben has the potential to be a closed basin—that will also focus water into the subsurface. There are 28 sites where sinkholes are present in the Canyonlands (Figure 4.1, Table 4.1), so it is likely that much surficial water in this region is redirected into the subsurface through these depressions. This redirection of surface water in turns probably acts to slow the process of channel development in the lower reaches of the catchments (Figure 4.17) as much less water will be available for erosion and incision there.

Transmission of the Colorado River base-level fall signal into the upper reaches of these tributary drainage networks is likely slowed or even stopped as the faults are knickpoints that are continuously growing over geological time scales (Figure 4.17).

There is no evidence from the aerial imagery or DEMs that the other three catchments (Y canyon, Unnamed Canyon, and Butler Wash) are presently experiencing topographic drainage cutoff due to sinkholes—at least, not yet. There are many other documented sinkholes in the region (Figure 4.1) that capture smaller subsets of those three catchments (see Table 4.1,

“Surface Water Features?”). The surface water that enters the subsurface through the sinkholes we examined, as well as others in the area, is either being stored in the subsurface or is flowing down to the Colorado River as part of the regional aquifer. Assuming that the Paradox Fm. is easily dissolved, it could be possible that karst is developing in the salt/gypsum of this formation,

133 allowing for faster transport of the water into and through the subsurface. This finding has important implications for dissolved salt loads in the Colorado River, where the waters entering into the subsurface in the Grabens region discharge into that river. The Colorado River dissolves a substantial amount of salt from the Paradox Fm. upstream, and desalination efforts currently are focused in the Paradox valley to reduce the water salinity before it reaches agricultural regions downstream (Tuttle and Grauch, 2009; Jochems and Pederson, 2015). The redirection of water into the subsurface and along/through the Paradox Formation could be supplying an additional salt load into the Colorado River not has not been considered before.

The analysis of -elevation and ksn indicates that the catchments within The Grabens are affected by the normal faults. They have more knickpoints, steeper channel gradients, and in map view, channel reorganization around faults or across faults is apparent. The lowest reaches (those nearest the Colorado River) of the four catchments all have similarly steep gradients (ksn > 200), and up to four knickpoints within the first 10 km in some cases (Figures 4.20, 4.21). Two specific correlations between knickpoint elevations and distance from the Colorado River are apparent, particularly at ~210 m and 3-4 km from the Colorado River outlets of Cross Canyon,

Butler Wash, and Y Canyon, indicating that those knickpoints could be related to the same base level fall signal of the Colorado River. Similarly, Cross Canyon and Bull Wash both have knickpoints at ~ 600 m above and ~ 11 km upstream from the Colorado River, which are likely tied to a common base level fall signal (Figure 4.22).

The lowest knickpoints from each catchment (Figure 4.22, black circles) do not fall at similar elevations, indicating channel reorganization control by faulting. The lack of additional correlations between knickpoint elevation groups indicates that the faults are actively modifying these channels faster than they can respond to base level fall along the Colorado River, leading to

134 catchment reorganization. This reorganization likely means that The Grabens region will be dissected more slowly and have abandoned channels when compared with a catchment such as

Bull Wash. This interpretation is supported by remotely sensed aerial imagery and topographic data. The channel reorganization within The Grabens will also lead to catchment takeover by neighboring catchments such as from Bull Wash to the south, as the channels are more developed and have less tendency to be rerouted by active faults.

4.3. Faults

The dilational tendency analysis shows that dilation most likely occurs within Cedar

Mesa Sandstone compared with the underlying Halgaito Shale–Elephant Canyon Transition Fm.

(with values of 0.89 and 0, respectively). This outcome is the result of higher fault dip angles being possible within the Cedar Mesa Sandstone due to lower overburden pressures. The

Halgaito Shale, being a less competent rock with more overburden pressures, will promote formation of shallower dips in fractures that pass through it. The shallower dip angle through the

Halgaito Shale allows Mode 1 (opening mode) fractures to form in the overlying Cedar Mesa

Sandstone.

Previous assessments of fault dips in the Grabens district of Canyonlands by McGill and

Stromquist (1979) found steeply dipping (near vertical or vertical) faults that extend from the surface to a depth of ~100 m, where their dip angle decreases to 75° until they intersect the

Paradox Fm. at a depth of about 100–200 m. The dilational tendency analysis confirms the previous evidence for dilational faults in the region (McGill & Stromquist, 1979), as the Halgaito

Shale member is more likely to slip, whereas the Cedar Mesa Sandstone is more likely to dilate due to the different fault dip angles and the negligible overburden pressure at the surface.

135

The central sinkhole at Site 1 has Cedar Mesa Sandstone juxtaposed with Quaternary alluvium on either side of its walls. This juxtaposition is interpreted to correspond to a fault in this location, which has allowed the drainage of material from the downthrown block (Figure

4.6). Sites 2 and 3 are both located in the middle of a large graben, and upon further inspection in the field were found to be situated above an anti- or synthetic fault that forms a relay ramp across the graben floor. Site 4 is the only site not found in direct association with a surface-breaking fault, although crevasses seen along the relay ramp (e.g., Figure 4.10) may correspond to a fault extending towards the sinkholes to the south.

If the formation of these sinkholes is due to slip along the large faults, we would expect pit depth and fault displacement to be correlated such that more displacement equals deeper sinkholes. To test if the sinkholes are due to the large graben-bounding faults, we compared the fault displacement along the large faults nearby the sinkholes with sinkhole depth. A confounding variable, however, is the degree of infilling of the sinkholes by sedimentation and vegetation, which is not in the current scope of this project. Additionally, the present depth of the sinkholes might be affected by surficial water runoff, for which we found evidence at three of the four field sites visited, and 16 of the 28 (57%) total sinkhole locations across The Grabens

Region. It is still unclear what contribution to sinkhole growth (i.e., widening and deepening) or infilling (i.e., shallowing) surface water plays. The above-ground flow of water could either: 1) promote more collapse and material drainage into the subsurface, increasing the width and depth of sinkholes; or, 2) transport more sediment into the sinkhole thereby clogging up the subsurface drainage pathway, and serving to decrease sinkhole depth and width. It is even possible that both outcomes are possible, operating to various extents at different times for the same sinkhole or variable across all of the different sinkholes.

136

From our measurements at the four field sites, the fault displacement and sinkhole depths do not appear correlated. The deepest sinkhole at Site 1 is 7.5 m deep (Table 2), with the nearest fault having 50 m of vertical displacement (Figure 4.6). The deepest sinkhole at Site 2 is 4.1 meters deep (Table 2), but the faults there have at least twice the vertical offset of the fault with the greatest offset from the other field sites (110 m versus 50 m, respectively). Site 3 has the deepest sinkhole (15.0 meters, Table 2), but the nearby fault (<10 m away from pit boundary,

Figure 4.11) does not have the most vertical displacement (44 m, Figure 4.12) of the faults measured. Site 4 has sinkholes 3.4 meters deep (Table 2) and the faults on the north and south sides of the graben have 43 m and 25 m of vertical displacement, respectively (Figure 4.17). The lack of any consistency in sinkhole depth versus vertical displacement along the major faults nearby suggests that those faults do not control the development of these central sinkholes.

Rather, the depressions may be controlled by secondary faults in the graben floors that have no observable surface trace. At Sites 2 and 4, there is no evidence for a surface-breaking fault that the sinkholes lie along, only indications of an open dilational fracture at both sites.

Such structures (subsurface faults with no surface expression) were observed in GPR profiles at analogous sites farther north in Canyonlands National Park by Kettermann et al.

(2015) (Figure 4.2). These structures also are suggested by the sinkhole from Site 1 that has alluvium on one wall and Cedar Mesa Sandstone on the other. These sinkholes most likely formed along a similar dilational fracture within the Cedar Mesa Sandstone just below the 1–2 m thick alluvium that covers most of the graben floors. Site 2 is located along the middle of the

Imperial Valley graben floor but is adjacent to a synthetic fault just to the south, where we identified a small chasm with drainages leading into it. This sinkhole chain probably also formed

137 along an anti- or synthetic fault (no dip information was discernable from our field observations) offset from the southern faulted block.

4.5. Conclusions

The sinkholes present in the Canyonlands region play an important role in the geomorphological evolution of the area. There are at least 28 identified sites with one or more sinkholes present within the larger Grabens region, on the basis of mapping from aerial imagery; at least 57% of these sinkholes show evidence for surface water interaction (i.e., channels running into the sinkholes) on that basis, too. And, as shown by our findings at Site 3, not only do the sinkholes drain material into the subsurface, they form local base level pour points that once formed, appear to enhance the formation of ephemeral, up-slope migrating erosional channel networks across the floors of the grabens with which they are contained (Figure 4.11).

Of the four sites visited during fieldwork, three had evidence for further modification

(that is, an obvious component of erosion not consistent with a purely tectonic origin or activity) from channels and drainages. Site 1 is the only example of sinkholes formed by purely tectonic activity based on the lack of fluvial features, supported by the juxtaposition of Quaternary alluvium and Cedar Mesa Sandstone Fm. (Figure 4.6) on either side of the central sinkhole walls

(Figure 4.7). Site 2 is host to sinkholes and surficial water features. Those surficial water features converge on a local topographic low, where a joint is situated. We interpret these sinkholes to be due to antithetic and synthetic faulting in the graben floor. Site 3, the most morphologically complex of the four visited, has the deepest sinkhole (8 m deep) as well as a well-developed channel head system propagating upstream from the sinkhole and channel intersections (Figure 4.11). This is the location of the truncation of the Cross Canyon channel by

138 the sinkholes, and all surface water arriving from upstream appears to now drain into the subsurface here. Site 4 is much like Site 2 in that it hosts sinkholes in association with surficial water features that intersect along a graben floor fault where another open tension fracture is found that acts to drain water into the subsurface. These four sites show the impact sinkholes in this region have in influencing erosion and water transport within the graben.

Detailing the interactions of the sinkholes, faults, and surficial water in this region reveals a complex interplay between these three geomorphological types of feature. The nature of the faulting in this region is heavily influenced by the stratigraphy, which facilitates Mode I fractures at the surface because of the dilatational nature of the faults at depth. On the basis of published work on pit craters elsewhere (cf. Chapters 2 and 3), we infer that these tensional fractures permit the subsurface drainage of alluvium, initiating sinkhole development from downward drainage of surficial materials into opening void space at depth. Further dissection of the sinkholes begins when locally channelized (i.e., closed basin within a graben) or regionally channelized (i.e., Cross Canyon, Butler Wash) catchments intersect the sinkholes and redirect surface water into the subsurface. This redirection may then supersede canyon erosion and development (Figure 4.17) downstream from the canyon-bottom sinks, impeding the upstream propagation of regional base level fall (river incision) set by the Colorado River as the downstream reaches of these tributary networks will lack the surface water discharge to modify their longitudinal profiles. The water being redirected into the subsurface in this region is likely interacting with the Paradox Fm. in the subsurface and ultimately contributing to the salt load of the Colorado River in this area, such as is known to occur in more northern reaches of the

Colorado River Basin (Tuttle & Grauch, 2009). This is an important finding given that desalination efforts are in effect along the Colorado River so that downstream agricultural use is

139 not impacted by high-salinity water, and to satisfy legal requirements for the dissolved salt load as the river enters Mexico (Tuttle & Grauch, 2009).

The Grabens region sinkholes are good examples of tectonically controlled pit depressions that have also undergone secondary erosion and modification from surficial water flow. These are key observations when considering the abundance of quasi-circular depressions

(“pit craters”) on other planets that appear to be related to tectonic activity. In the Grabens region, sinkholes seem to be related mostly to intra-graben faulting and not to the large graben- bounding faults that dominate the landscape. The sinkholes in the Canyonlands are small, with none deeper than 15 meters. Pit craters on Mars, in contrast, are much larger, and in some cases on the order of kilometers in depth, width, and length (Chapter 2). This stark difference in size is most likely due to erosional processes on Earth masking or removing evidence for pit craters before they can grow to such sizes.

The sinkholes in the Grabens can influence, and be heavily influenced by, the surrounding surface water courses. Future field work should focus on Butler Wash, which contains many of the catalogued sinkhole sites (Figure 4.1C, D). In addition, optically stimulated luminescence dating of the sinkhole stratigraphy, particularly at Site 3, could shed light on the absolute age of the Cross Canyon catchment, and help build a better understanding of the interaction between regional gravity-driven tectonics, dilational faulting, and the development of non-karstic sinkholes, and their importance in controlling further topographic evolution of the regional surface and groundwater networks.

140

Figure 4.1. Regional maps of The Grabens region of Canyonlands National Park. A: National Elevation Database 5 m/px DEM; B: National Agricultural Imagery Program 2018 San Juan County Mosaic, with cross-sectional lines and grey boxes denoting field site figure locations; C: Structure map showing faults and sinkholes we identified from aerial imagery (blue dots from Ketterman et al. (2015), yellow dots added in this study); and D: Regional catchmentss and channels derived from the NED 5 m/px DEM.

141

Table 4.1 Sinkhole Locations Surface Water ID UTM 12N Y UTM 12 N X Shape Features? 1 4209368 590719 y elongate/channelized 2 4209606 591066 n circular 3 4210755 590885 y circular/channelized 4 4209836 589521 n circular 5 4209795 591244 y circular 6 4210065 588864 y elongate/channelized 7* 4216770 593396 y elongate/channelized 8* 4216034 597505 y channelized 9* 4215463 596701 y elongate/channelized 10* 4215371 596640 y n.d. 11* 4214530 595810 n n.d. 12* 4214371 595736 n channelized 13* 4215468 595484 y elongate/channelized 14* 4216391 595506 n circular 15* 4216431 595634 n elongate 16* 4217023 595891 n elongate 17* 4218036 596738 y elongate/channelized 18* 4211503 591366 y elongate/channelized 19* 4211649 591550 y elongate/channelized 20* 4211852 592077 y circular/channelized 21* 4212655 592642 n elongate 22* 4212454 592596 y elongate/channelized 23* 4213942 593386 y elongate/channelized 24* 4212701 595234 y elongate/channelized 25* 4212338 594443 n n.d. 26* 4211547 593350 y elongate/channelized 27* 4211060 593742 n elongate 28* 4221884 599130 n elongate n.d. = non-descript, cannot tell shape from aerial imagery *taken from Kettermann et al. (2015) supplemental map, although those data did not include latitude or longitude values, nor comments regarding any association with surface water features nor on shape

142

Figure 4.2. Top: a) GPR profile by Kettermann et al. (2015) across Devil’s Lane in the northern Grabens region, b) interpreted GPR profile showing an abundance of shallow subsurface faults (black dashed lines) within alluvium, and possibly within Cedar Mesa Sandstone (which occupies the lower 3–4 m of the profiles) where reflectors are less prominent. Blue lines are the interpreted stratigraphy of the alluvial deposits. Bottom: NAIP image with GPR profile and locations of the shallowest near-surface faults marked (white dashed lines) from analysis by Kettermann et al. (2015).

143

Figure 4.3. Cartoon showing a) mixed-mode fracturing creating dilational fault segments at the surface and b) how a weak lubricating layer could detach additional blocks from the fault plane causing further collapse structures to form. From Kettermann et al. (2015).

Figure 4.4. Schematic illustration showing the relationship between graben-bounding faults, intra-graben faults, sinkholes, and stratigraphy in the northern Grabens Region, modified from Kettermann et al. (2015).

144

Table 4.2 Sinkhole Geometric Properties ID Site Number Width (m) Length (m) Depth (m) Strike (degrees) 13 1 16.3 49.9 7.5 63 14 1 14.2 24.7 0.7 76 15 1 5.8 8.5 2.2 65 5 2 11.6 21.9 4.1 109 6 2 7.2 7.6 3.4 30 7 2 3.3 7.1 1.2 68 8 2 6.8 19.7 3.4 77 9 2 5.0 5.2 0.9 94 10 2 9.9 11.8 2.7 25 11 2 3.2 4.3 0.9 69 12 2 6.1 8.5 0.9 29 3 3 38.6 97.4 15.0 58 4 3 19.9 136.0 6.5 39 16 3 32.6 69.8 6.4 32 0 4 9.3 11.4 2.5 0 1 4 7.5 9.6 3.4 137 2 4 5.0 5.9 1.6 16 Morphometric properties of sinkholes located at the four field sites visited

145

Figure 4.5. Site 1 regional structure map (centered at 38.03° N, 109.96° W) with sinkholes, normal faults (inferred and known) and ephemeral channels mapped. Base map is a UAS-derived DSM overlaid on the corresponding hillshade. A–A’ and A’’–A’’’ cross section lines shown are in Figure 6. Gaps are due to missed UAS photo captures during the mapping flight plan.

146

Figure 4.6. Cross sections A–A′ and A′′–A′′′ (shown in Figures 1B, 4.5) covering the sinkholes identified at Site 1. No vertical exaggeration.

Figure 4.7. A) Photograph looking to the southeast, showing the central sinkhole at Site 1. B) Cartoon depiction of that same area showing the rock units and fault slip directions.

147

Figure 4.8. Site 2 regional structure map (centered at 38.04° N, 109.96° W) with sinkholes, faults (inferred and known), and ephemeral channels mapped. Base map is a UAS-derived DSM overlaid on the corresponding hillshade.

Figure 4.9. A) Photograph showing sinkholes at Site 2. B) Cartoon depiction of the same view showing rock units and fault slip directions.

148

Figure 4.10. Photograph showing the channel intersection with an open joint surface at Site 2. The approximate location where this photograph was taken is shown on Figure 8.

149

Figure 4.11. Sites 3 and 4 regional structure map (centered at 38.05° N, 109.95° W) with sinkholes, faults (inferred and known), and ephemeral channels mapped. Locations of C’’–C’’’, D–D’, and D’’–D’’’ cross sections shown in Figure 12 also shown. Base map is a UAS-derived DSM overlaid on the corresponding hillshade.

Figure 4.12. Cross sections C–C′ and C′′–C′′′ (shown in Figure 1B) covering the sinkholes identified at Site 3. No vertical exaggeration.

150

Figure 4.13. A) Photograph showing the western sinkhole at Site 3. B) Cartoon depiction of the same scene with rock units and nearby faults denoted.

Figure 4.14. Temporal sequence of NAIP imagery from 2006 to 2018, as well as 2019 UAS orthomosaic collected in this study. Key pit outline and channel head locations from 2006 are plotted on each image for reference, and any changes are shown with red polygons or dots.

151

Figure 4.15. A) Photograph looking to the southwest showing fluvial dendritic depression at Site 4. B) Cartoon depiction of the same view showing rock units labeled and fault slip direction.

Figure 4.16. Cross sections D–D′ and D′′–D′′′ (shown in Figure 1B) covering the sinkholes identified at Site 4. No vertical exaggeration.

152

Figure 4.17. Trunk-channel longitudinal profiles for the four catchments in The Grabens region (A–D), one unfaulted channel (Bull Wash) just outside of this region for comparison (E), and all longitudinal profiles normalized (F). Colors correspond to those used in Figure 1D.

153

Figure 4.18. InSAR maps from Furuya et al. (2007). a) Regional mean yearly vertical displacement (shown in mm/yr), with displacement away from the spacecraft shown as positive. Faults are overlaid in white mapped originally by Mcgill & Stromquist, 1979); b) hillshade showing faults and cross-sectional lines, from Furuya et al. (2007). The location of Cross Canyon and Imperial Valley is shown with the white box.

154

Figure 4.19. E–E’ Cross Section from Figure 17, which intersects Cross Canyon, from Furuya et al. (2007). The location of Site 3, Cross Canyon, is shown in the purple band.

155

Figure 4.20. Elevation versus Chi plots for each of the study channels in The Grabens region (from top to bottom, Cross Canyon, Butler Wash, and Y Canyon) and Bull Wash for comparison at the bottom. Each reach selected for chi analysis is shown in red dashed lines, and the fit and R2 shown for reference. Knickpoints are marked in blue triangles, which correspond to those locations on Figure 4.21. Reference m/n for each catchment is 0.45.

156

Figure 4.23. Chi (left) and normalized channel steepness (ksn) (right) of channels displayed on top of slope shade maps (slope map with transparency over a grayscale DEM) for The Grabens and surrounding region. Knickpoints marked with blue triangles. Chi map has catchment divide general outlines shown with colored polygons (colors correspond to Figure 1D). Black arrows on Chi map indicate anticipated drainage divide movement based on  comparisons. Low  near drainage divides indicates erosion and expansion of the drainage network at the expense of the channels on the other side.

157

Figure 4.22. Knickpoint elevations above the Colorado River for each catchment in the study area (colored dots which correspond to Figure 1D). Elevation above Colorado river is based on the elevation of the Butler Wash outlet into the Colorado River (1185 m). Lowest knickpoint from each catchment absolute elevations shown with black circles

158

Table 4.3 Channel Reach Normalized Channel Steepness

Faulted Reaches ksn Unfaulted Reaches ksn

Cross Canyon Reach 1 206.8 Cross Canyon Reach 3 38.7 Cross Canyon Reach 2 306 Cross Canyon Reach 5 13.2 Cross Canyon Reach 4 94.9 Butler Wash Reach 3 49.7 Butler Wash Reach 1 124.3 Butler Wash Reach 5 27.7 Butler Wash Reach 2 388.4 Y Canyon Reach 3 19 Butler Wash Reach 4 281.8 Bull Wash Reach 1 300.2 Y Canyon Reach 1 112.5 Bull Wash Reach 2 40.2 Y Canyon Reach 2 168.3 Bull Wash Reach 3 52.4

Mean 210.4 67.7

159

CHAPTER 5

Conclusions

5.1. Overview

This dissertation describes an investigation of pit crater formation within the Solar

System, with specific emphasis on the pits found at Noctis Labyrinthus, Mars and at multiple sites on Earth. Each science chapter (Chapters 2–4) focuses on a specific locale (Chapters 2 and

4) or set of sites (Chapter 3) that represent notable pit crater assemblages that readily enable the study of how pit craters form across the Solar System.

Chapter 2 addressed how pit craters and normal faulting contributed to the formation of

Noctis Labyrinthus on Mars, a structurally complex region. I found that pit craters are more often than not associated with normal faults in this region, and both normal faults and pit craters are cross-cut by larger trough-like features. Evidence for a volatile-rich subsurface in the form of periglacial landforms and mass-wasting deposits is found in the base of the troughs, and rampart craters in the region indicate that volatiles could have been (or still are) at a stratigraphic depth of

100 m or less when those craters were formed. These observations lead to my interpretation that volatiles played a substantial role in the structural evolution of this region of Mars.

Chapter 3 detailed populations of pit crater on several different Solar System bodies, comparing them with two accessible and well-preserved pit crater locales on Earth, Craters of the

Moon National Monument and Preserve in Idaho, and Hawaii Volcanoes National Park on the

Big Island of Hawai’i. Extraterrestrial pit craters are found at a range of scales, and measurements of those pits have been made in previous literature, but never collated into a comprehensive census of pit crater morphology. The fieldwork in Idaho and Hawaii provides a

160 useful basis for understanding how pit craters can form from multiple different processes, either tectonic or volcanic, and evolve to similar morphologies.

In Chapter 4, I analyzed a different type of pit crater morphology in the Grabens region of Canyonlands National Park, Utah. The pits in Canyonlands differ from those in Chapters 2 and 3 because the Utah sinkholes are found in mostly unconsolidated alluvium and aeolian sediments and sedimentary rock packages, rather than in a volcanic or volcanotectonic setting like at Craters of the Moon National Monument and Preserve and Hawaii Volcanoes National

Park. In addition to assessing previously identified pits here, I located four additional sites in the region (from aerial imagery) where sinkholes have formed and cataloged four sites in detail with field observations. Of these four sites, three (Sites 1, 3, and 4) are situated along fractures (either seen in the field or interpreted from field and topographic data), implying that the formation of these pits at least has been heavily influenced by the gravity-driven extensional tectonics in the region. Additionally, numerous features attributed to surficial water flow were noted near sinkholes in Sites 2, 3, Site 4, and which I interpret as secondary erosional features further contributing to the morphology of these features.

These three science chapters together expand our understanding of pit crater formation by integrating remote sensing and fieldwork at known and new pit crater sites on Earth, complemented by geological data for pits on a variety of other planetary bodies.

5.2. Discussion and Scientific Importance

This work has yielded some important insights into how pit craters form and develop on different Solar System bodies. For example, I have shown that pits on Earth are much smaller than other pits studied in this project. This finding partly reflects the resolution of remotely

161 sensed data for other planetary bodies, as much of the data available for bodies other than Earth has resolutions on the decameter scale. (Although HiRISE imagery for Mars does have sub- meter resolutions, coverage for all of Mars with this dataset is <5% at time of this writing).

However, it still stands that pits on Earth from tectonic or volcanic activity are not as large as their planetary counterparts, which leads to a fundamental question about how pits form in general: what causes void space in the subsurface for surficial material to collapse into, and do the size differences between Earth and Mars, say, relate to secondary processes or the formation mechanism(s) governing pit formation on Mars?

The influence of erosion on Earth pits cannot be understated, as Earth is the only body in this study with an active hydrologic system, tectonics, and atmosphere, all of which will help shape the ultimate form of terrestrial pits. As seen in Canyonlands, pits found in unconsolidated sediments are often influenced by surficial water flow—and can even redirect that flow of water into the subsurface at some locations. This finding shows that pits can influence secondary erosion of the landscape by focusing surficial water into the subsurface, unrelated to the original collapse of material into the subsurface. The largest pits I studied on Earth, King’s Bowl in

CRMO and the Devil’s Throat and Twin Pits in HVNP, all formed in stacked basaltic lava flows with little to no unconsolidated materials making up the walls. These pits have cylindrical shapes, which probably indicates that the basaltic flows in the walls are strong enough to maintain vertical or near-vertical walls for extended periods.

Yet pits on other planetary bodies are often shaped like inverted cones, and this finding can mean one of two things: 1) that either the pits originally formed in unconsolidated material and always assumed an inverted conical shape; and/or 2) that the pits initially formed as cylinders, either in competent material (i.e., lava flows) or less cohesive material (e.g., cinder,

162 ash, or alluvium) and subsequent, secondary processes such as mass wasting then led the pit walls to evolving into an inverted conical shape.

So what controls initial pit formation, and does that mechanism(s) have an influence on the original and/or final pit shape? From the work in these three chapters, I found that pits can develop because of multiple processes, and the resultant pit shape appears at least in part correlated with those formation mechanisms. At Noctis Labyrinthus, I regard the inverted-cone- shaped pits there as tectonic in origin. In Hawaii, I attribute pits that are cylindrical in shape and form in stacked lava flows to a subsurface cavity collapsing. Those subsurface cavities are, in turn, likely to have formed because of the evacuation of lava from the subsurface, meaning that these pits are volcanically related. The one confounding observation to this interpretation for the

Hawaiian pits is the example at Kilauea Iki, which formed from the outset as a cylinder shaped void space in cinder deposits, whereas the rest of the pits in that chain are inverted cones.

Monitoring the shape of this pit over the coming years will allow for unprecedented insight into how, and how long, it takes a pit formed in unconsolidated material to acquire an inverted conical shape, as presumably it will.

The King’s Bowl pit in Craters of the Moon National Monument and Preserve is the only example of an explosive pit studied in this work, whereas the nearby Yellowjacket Waterhole pits are interpreted to be due to faulting. King’s Bowl is an abnormally shaped, lenticular pit; the

Yellowjacket Waterhole pits have inverted-cone shapes. I interpret these latter pits to be due to faulting. Finally, in The Grabens region of Canyonlands National Park, pits here have inverted conical shapes after forming in unconsolidated alluvium. I interpret these pits to have generally formed over anti- or synthetic faults in the middle of the graben floors, and so are primarily tectonic. But these pits have also been shaped by a secondary process, the erosion of surficial

163 deposits and pit flank slopes by surface water that has sometimes masked the original pit shape in favor of a dendritic channel morphology.

5.3. Future Work

This work highlights the properties of some major pit crater assemblages on Earth and elsewhere in the Solar System. Field work at other pit locations on Earth will provide additional insight into the formation and development of this ubiquitous type of landform. For example, there are pits in Siberia, Russia, that appear to be related to methane outgassing in the permafrost there that is melting. Those pits may be a useful analogue for pits at Noctis Labyrinthus, for which I am invoking subsurface volatile sublimation as a factor contributing to the secondary erosion of the pits to form troughs.

Tracking the change in shape of the pits on the Big Island of Hawaii will be both scientifically valuable and possible as this area, being volcanically active, is regularly the target for high-resolution lidar data collection and focused aerial imaging campaigns. Google Earth imagery suggests that the new pit on the flanks of Kilauea Iki has grown by a few meters since I last visited in October 2018. Detailed mapping of this enlarged pit will offer a valuable temporal sequence of how the pit shape is modifying over time.

The pits in Canyonlands provide a unique glimpse into the recent stratigraphy of the area, and therefore could be used to further investigate the fluvial geomorphology of the area through optically stimulated luminescence dating of the alluvial layers nicely exposed in pit walls. Doing so would help not only in deciphering the details of the tectonics in the region but also of the channels that are affected by the faulting therein.

164

This work provides a basis for evaluating pit craters in a variety of different geological environments, either through remote sensing, field work, or both, on rocky planetary bodies including Earth. The shapes and nature of the geological materials that the pits form in are the best indicators of the processes involved in their formation, and without direct indications of what those properties are (i.e., when only remote sensing of pits on other planets is possible), secondary observations and analog fieldwork are key to interpreting the nature of those pits correctly.

165

REFERENCES

Andrews-Hanna, J. C. (2012a). The formation of Valles Marineris: 1. Tectonic architecture and the relative roles of extension and subsidence. Journal of Geophysical Research: Planets, 117(3). https://doi.org/10.1029/2011JE003953

Andrews-Hanna, J. C. (2012b). The formation of Valles Marineris: 2. Stress focusing along the buried dichotomy boundary. Journal of Geophysical Research: Planets, 117(E4). https://doi.org/10.1029/2011JE003954

Andrews-Hanna, J. C. (2012c). The formation of Valles Marineris: 3. Trough formation through super-isostasy, stress, sedimentation, and subsidence. Journal of Geophysical Research: Planets, 117(6). https://doi.org/10.1029/2012JE004059

Aslan, A., Karlstrom, K. E., Kirby, E., Heizler, M. T., Granger, D. E., Feathers, J. K., et al. (2019). Resolving time-space histories of Late Cenozoic bedrock incision along the Upper Colorado River, USA. Geomorphology, 347, 106855. https://doi.org/10.1016/j.geomorph.2019.106855

Baars, D. L. (1962). Permian system of Colorado plateau. AAPG Bulletin, 46(2), 149–218.

Banerdt, W. B., Golombek, M. P., & Tanaka, K. L. (1992). Stress and tectonics on Mars (pp. 249–297).

Billingsley, G. H., Block, D. L., & Felger, T. J. (2002). Surficial geologic map of The Loop and Druid Arch quadrangles, Canyonlands National Park, Utah. US Geological Survey.

Bistacchi, N., Massironi, M., & Baggio, P. (2004). Large-scale fault kinematic analysis in Noctis Labyrinthus (Mars). Planetary and Space Science, 52(1–3), 215–222. https://doi.org///dx.doi.org/10.1016/j.pss.2003.08.015

Black, R. F. (1976). Periglacial features indicative of permafrost: ice and soil wedges. Quaternary Research, 6(1), 3-26.

Bleamaster, L. F., & Hansen, V. L. (2004). Effects of crustal heterogeneity on the morphology of chasmata, Venus. Journal of Geophysical Research: Planets, 109(E2), n/a. https://doi.org/10.1029/2003JE002193

Brooker, L. M., Balme, M. R., Conway, S. J., Hagermann, A., Barrett, A. M., Collins, G. S., & Soare, R. J. (2018). Clastic polygonal networks around Lyot crater, Mars: Possible formation mechanisms from morphometric analysis. Icarus, 302, 386-406.

Buczkowski, D. L., & Wyrick, D. Y. (2015). Tectonism and magmatism identified on asteroids. Geological Society Special Publication, 401, 423–441. https://doi.org/10.1144/SP401.18

166

Buczkowski, D. L., Schmidt, B. E., Williams, D. A., Mest, S. C., Scully, J. E. C., Ermakov, A. I., Preusker, F., Schenk, P., Otto, K. A., Hiesinger, H., O’Brien, D., Marchi, S., Sizemore, H., Hughson, K., Chilton, H., Bland, M., Byrne, S., Schorghofer, N., Platz, T., … Russell, C. T. (2016). The geomorphology of Ceres. Science, 353(6303). https://doi.org/10.1126/science.aaf4332

Byerlee, J. (1978). Friction of rocks. Pure and Applied Geophysics PAGEOPH, 116(4–5), 615– 626. https://doi.org/10.1007/BF00876528

Byrne, P. K., Klimczak, C., Şengör, A. M. C., Solomon, S. C., Watters, T. R., & Hauck, S. A. (2014). Mercury/’s global contraction much greater than earlier estimates. Nature Geoscience, 7(4), 301–307.

Byrne, P. K., Van Wyk De Vries, B., Murray, J. B., & Troll, V. R. (2012). A volcanotectonic survey of Ascraeus Mons, Mars. Journal of Geophysical Research E: Planets, 117(1), 1–25. https://doi.org/10.1029/2011JE003825

Callihan, M. B., & Klimczak, C. (2019). Topographic expressions of lunar graben. Lithosphere, 11(2), 294–305. https://doi.org/10.1130/l1025.1

Clark, R. M., & Cox, S. J. D. (1996). A modern regression approach to determining fault displacement-length scaling relationships. Journal of Structural Geology, 18(2–3), 147– 152.

Condon, S. M. (1997). Geology of the Pennsylvanian and Permian cutler group and Permian Kaibab limestone in the Paradox Basin, southeastern Utah and southwestern Colorado. US Government Printing Office.

Cowie, P. A., & Scholz, C. H. (1992). Displacement-length scaling relationship for faults; data synthesis and discussion. Journal of Structural Geology, 14(10), 1149–1156.

Crosta, G. B., Frattini, P., Valbuzzi, E., & De Blasio, F. V. (2018a). Introducing a New Inventory of Large Martian Landslides. Earth and Space Science, 5(4), 89–119. https://doi.org/10.1002/2017EA000324

Crosta, G. B., De Blasio, F. V., & Frattini, P. (2018b). Global Scale Analysis of Martian Landslide Mobility and Paleoenvironmental Clues. Journal of Geophysical Research: Planets, 123(4), 872–891. https://doi.org/10.1002/2017JE005398

Crow-Willard, E. N., & Pappalardo, R. T. (2015). Structural mapping of Enceladus and implications for formation of tectonized regions. Journal of Geophysical Research: Planets, 120(5), 928–950.

Crumpler, L. S., & Aubele, J. C. (1978). Structural evolution of Arsia Mons, Pavonis Mons, and Ascreus Mons: Tharsis region of Mars. Icarus. https://doi.org///dx.doi.org/10.1016/0019- 1035(78)90041-6

167

Cushing, G. E., Titus, T. N., Wynne, J. J., & Christensen, P. R. (2007). THEMIS observes possible cave skylights on Mars. Geophysical Research Letters, 34(17). https://doi.org/10.1029/2007GL030709

Cushing, G. E., Okubo, C. H., & Titus, T. N. (2015). Journal of Geophysical Research : Planets. Journal of Geophysical Research: Planets, 1252–1266. https://doi.org/10.1002/2014JE004735

Darling, A. L., Karlstrom, K. E., Granger, D. E., Aslan, A., Kirby, E., Ouimet, W. B., et al. (2012). New incision rates along the Colorado River system based on cosmogenic burial dating of terraces: Implications for regional controls on Quaternary incision. Geosphere, 8(5), 1020–1041. https://doi.org/10.1130/GES00724.1

Davey, S. C., Ernst, R. E., Samson, C., & Grosfils, E. B. (2013). Hierarchical clustering of pit crater chains on Venus. Canadian Journal of Earth Sciences, 50(1), 109–126. https://doi.org/10.1139/cjes-2012-0054

Davis, P. A., Tanaka, K. L., & Golombek, M. P. (1995). Topography of Closed Depressions, Scarps, and Grabens in the North Tharsis Region of Mars: Implications for Shallow Crustal Discontinuities and Graben Formation. Icarus. https://doi.org///dx.doi.org/10.1006/icar.1995.1071 del Potro, R., & Hürlimann, M. (2008). Geotechnical classification and characterisation of materials for stability analyses of large volcanic slopes. Engineering Geology, 98(1–2), 1– 17. https://doi.org/10.1016/j.enggeo.2007.11.007

Eaton, G. P. (1982). Origin and Tectonic Significance. Annual Review of Earth and Planetary Sciences, 10(Figure 1), 409–440.

Edwards, C. S., Nowicki, K. J., Christensen, P. R., Hill, J., Gorelick, N., & Murray, K. (2011). Mosaicking of global planetary image datasets: 1. Techniques and data processing for Thermal Emission Imaging System (THEMIS) multi-spectral data. Journal of Geophysical Research E: Planets, 116(10), 1–21. https://doi.org/10.1029/2010JE003755

Fergason, R. L., Hare, T. M., & Laura, J. (2018). HRSC and MOLA Blended Digital Elevation Model at 200m v2. Astrogeology PDS Annex, U.S. Geological Survey. http://bit.ly/HRSC_MOLA_Blend_v0

Ferrill, D. A., Wyrick, D. Y., Morris, A. P., Sims, D. W., & Franklin, N. M. (2004). Dilational fault slip and pit chain formation on Mars. GSA Today, 14(10), 4–12.

Ferrill, D. A., Morris, A. P., McGinnis, R. N., Smart, K. J., Wigginton, S. S., & Hill, N. J. (2017). Mechanical stratigraphy and normal faulting. Journal of Structural Geology, 94, 275–302. https://doi.org/10.1016/j.jsg.2016.11.010 Ferrill, D. A., & Morris, A. P. (2003). Dilational normal faults. Journal of Structural Geology,

168

25(2), 183–196. https://doi.org///dx.doi.org/10.1016/S0191-8141(02)00029-9

Ferrill, D. A., Wyrick, D. Y., & Smart, K. J. (2011). Coseismic, dilational-fault and extension- fracture related pit chain formation in Iceland: Analog for pit chains on Mars. Lithosphere, 3(2), 133–142.

Ford, J. P. (1993). Guide to Magellan image interpretation. National Aeronautics and Space Administration, Jet Propulsion Laboratory.

Furuya, M., Mueller, K., & Wahr, J. (2007). Active salt tectonics in the Needles District, Canyonlands (Utah) as detected by interferometric synthetic aperture radar and point target analysis: 1992-2002. Journal of Geophysical Research: Solid Earth, 112(6), 1992–2002. https://doi.org/10.1029/2006JB004302

Gallen, S. F., & Wegmann, K. W. (2017). River profile response to normal fault growth and linkage: An example from the Hellenic forearc of south-central Crete, Greece. Earth Surface Dynamics, 5(1), 161–186. https://doi.org/10.5194/esurf-5-161-2017

Golombek, M. P., & Bridges, N. T. (2000). Erosion rates on Mars and implications for climate change: Constraints from the Pathfinder landing site. Journal of Geophysical Research: Planets, 105(E1), 1841-1853.

Hajna, N. Z., Baioni, D., & Tramontana, M. (2017). Karst landforms Within Noctis Labyrinthus, Mars. Acta Carsologica, 46(1).

Head, J. W., & Wilson, L. (1993). Lunar graben formation due to near-surface deformation accompanying dike emplacement. Planetary and Space Science. https://doi.org///dx.doi.org/10.1016/0032-0633(93)90114-H

Hughes, S. S., Kobs Nawotniak, S. E., Sears, D. W. G., Borg, C., Garry, W. B., Christiansen, E. H., et al. (2018). Phreatic explosions during basaltic fissure eruptions: Kings Bowl lava field, Snake River Plain, USA. Journal of Volcanology and Geothermal Research, 351, 89– 104. https://doi.org/10.1016/j.jvolgeores.2018.01.001

Humphreys, E. D. (1995). Post-Laramide removal of the Farallon slab, western United States. Geology, 23(11), 987–990. https://doi.org/10.1130/0091- 7613(1995)023<0987:PLROTF>2.3.CO;2

Huntoon, P. W. (1982). The Meander anticline, Canyonlands, Utah: An unloading structure resulting from horizontal gliding on salt. Geological Society of America Bulletin, 93(10), 941–950. Jaumann, R., Neukum, G., Behnke, T., Duxbury, T. C., Eichentopf, K., Flohrer, J., et al. (2007). The high-resolution stereo camera (HRSC) experiment on Mars Express: Instrument aspects and experiment conduct from interplanetary cruise through the nominal mission. Planetary and Space Science, 55(7–8), 928–952. https://doi.org///dx.doi.org/10.1016/j.pss.2006.12.003

169

Johnsson, A., Reiss, D., Hauber, E., Zanetti, M., Hiesinger, H., Johansson, L., & Olvmo, M. (2012). Periglacial mass-wasting landforms on Mars suggestive of transient liquid water in the recent past: Insights from solifluction lobes on Svalbard. Icarus, 218(1), 489-505.

Kettermann, M., Grützner, C., Van Gent, H. W., Urai, J. L., Reicherter, K., & Mertens, J. (2015). Evolution of a highly dilatant fault zone in the grabens of Canyonlands National Park, Utah, USA - Integrating fieldwork, ground-penetrating radar and airborne imagery analysis. Solid Earth, 6(3), 839–855. https://doi.org/10.5194/se-6-839-2015

Kirchoff, M. R., & Schenk, P. (2009). Crater modification and geologic activity in Enceladus’ heavily cratered plains: Evidence from the impact crater distribution. Icarus, 202(2), 656– 668.

Klimczak, C., Kling, C. L., & Byrne, P. K. (2018). Topographic Expressions of Large Thrust Faults on Mars. Journal of Geophysical Research: Planets, 123(8), 1973–1995. https://doi.org/10.1029/2017JE005448

Kuntz, M. A., Champion, D. E., Spiker, E. C., Lefebvre, R. H., & McBroome, L. A. (1982). The Great Rift and the Evolution of the Craters of the Moon Lava Field, Idaho. Cenozoic Geology of Idaho: Idaho Bureau of Mines and Geology Bulletin, 423–437.

Kuntz, M. A., Covington, H. R., & Schorr, L. J. (1992). An overview of basaltic volcanism of the eastern Snake River Plain, Idaho. Geological Society of America Memoirs, 179, 227– 268.

Leone, G. (2014). A network of lava tubes as the origin of Labyrinthus Noctis and Valles Marineris on Mars. Journal of Volcanology and Geothermal Research, 277, 1–8. https://doi.org///dx.doi.org/10.1016/j.jvolgeores.2014.01.011

Lowe, D. R. (1976). Grain Flow and Grain Flow Deposits. SEPM Journal of Sedimentary Research, Vol. 46(1), 188–199. https://doi.org/10.1306/212f6ef1-2b24-11d7- 8648000102c1865d

Malin, M. C., Bell, J. F., Cantor, B. A., Caplinger, M. A., Calvin, W. M., Clancy, R. T., et al. (2007). Context Camera Investigation on board the Mars Reconnaissance Orbiter. Journal of Geophysical Research: Planets, 112(E5), n/a. https://doi.org/10.1029/2006JE002808

Manighetti, I., King, G. C. P., Gaudemer, Y., Scholz, C. H., & Doubre, C. (2001). Slip accumulation and lateral propagation of active normal faults in Afar. Journal of Geophysical Research, 106, 13. https://doi.org/10.1029/2000JB900471

Masson, P. (1977). Structure pattern analysis of the Noctis Labyrinthus-Valles Marineris regions of Mars. Icarus, 30(1), 49–62. https://doi.org///dx.doi.org/10.1016/0019-1035(77)90120-8

McGill, G. E., & Stromquist, A. W. (1979). The grabens of Canyonlands National Park, Utah:

170

geometry, mechanics, and kinematics. Journal of Geophysical Research, 84(B9), 4547– 4563. https://doi.org/10.1029/JB084iB09p04547

McGovern, P. J., & Solomon, S. C. (1998). Growth of large volcanoes on Venus: mechanical models and implications for structural evolution. Journal of Geophysical Research E: Planets, 103(E5), 11071–11101. https://doi.org/10.1029/98JE01046

McKenzie, D., & Nimmo, F. (1999). The generation of martian floods by the melting of ground ice above dykes. Nature, 397(6716), 231–233. https://doi.org/10.1038/16649

Mège, D., & Masson, P. (1997). Graben formation and dike emplacement on Earth and other planets. In Lunar and Planetary Science Conference (Vol. 28).

Mège, D., & Masson, P. (1996). Amounts of crustal stretching in Valles Marineris, Mars. Planetary and Space Science, 44(8), 749–781. https://doi.org///dx.doi.org/10.1016/0032- 0633(96)00013-X

Melosh, H. J. (1989). Impact cratering: A geologic process. Research supported by NASA. New York, Oxford University Press (Oxford Monographs on Geology and Geophysics, No. 11), 1989, 253 p., 11.

Moore, J. M., & Schultz, R. A. (1999). Processes of faulting in jointed rocks of Canyonlands National Park, Utah. Geological Society of America Bulletin, 111(6), 808–822.

Neal, C. A., Brantley, S. R., Antolik, L., Babb, J. L., & Etc. (2019). The 2018 rift eruption and summit collapse of Kīlauea Volcano. Science, 363(January), 367–374.

Nordhaus, W. (2019). Economics of the disintegration of the Greenland ice sheet. Proceedings of the National Academy of Sciences of the United States of America, 116(25), 12261–12269. https://doi.org/10.1073/pnas.1814990116

Nuckolls, H. M., & McCulley, B. L. (1987). Origin of saline springs in Cataract Canyon, Utah.

Okubo, C. H., & Martel, S. J. (1998). Pit crater formation on Kilauea volcano, Hawaii. Journal of Volcanology and Geothermal Research, 86(1–4), 1–18. https://doi.org/10.1016/S0377- 0273(98)00070-5

Ollier, C. (1981). Tectonics and landforms. Price, L. W. (1974). The developmental cycle of solifluction lobes. Annals of the Association of American Geographers, 64(3), 430-438.

Prinz, M. (1970). Idaho rift system, snake river plain, idaho. Bulletin of the Geological Society of America, 81(3), 941–948. https://doi.org/10.1130/0016-7606(1970)81 Prockter, L., Thomas, P., Robinson, M., Joseph, J., Milne, A., Bussey, B., et al. (2002). Surface expressions of structural features on Eros. Icarus, 155(1), 75–93. https://doi.org/10.1006/icar.2001.6770

171

Rader, E., Kobs Nawotniak, S., & Heldmann, J. (2018). Variability of Spatter Morphology in Pyroclastic Deposits in Southern Idaho, as Correlated to Thermal Conditions and Eruptive Environment. Earth and Space Science, 5(10), 592–603. https://doi.org/10.1029/2018EA000377

Reiss, D., Hauber, E., Michael, G., Jaumann, R., & Neukum, G. (2005). Small rampart craters in an equatorial region on Mars : Implications for near-surface water or ice, 32(July), 2–5. https://doi.org/10.1029/2005GL022758

Reiss, D., Gasselt, S. V. A. N., Hauber, E., Michael, G., Jaumann, R., & Neukum, G. (2006). Ages of rampart craters in equatorial regions on Mars : Implications for the past and present distribution of ground ice, 1452(10), 1437–1452.

Robbins, S. J., & Hynek, B. M. (2012). A new global database of Mars impact craters ≥1 km: 2. Global crater properties and regional variations of the simple-to-complex transition diameter. Journal of Geophysical Research E: Planets, 117(6), 1–21. https://doi.org/10.1029/2011JE003967

Rodriguez, J. A. P., Zarroca, M., Linares, R., Gulick, V., Weitz, C. M., Yan, J., et al. (2016). Groundwater flow induced collapse and flooding in Noctis Labyrinthus, Mars. Planetary and Space Science, 124, 1–14. https://doi.org/10.1016/j.pss.2015.12.009

Schultz, R. A. (1996). Relative scale and the strength and deformability of rock masses. Journal of Structural Geology, 18(9), 1139–1149. https://doi.org/10.1016/0191-8141(96)00045-4

Schwanghart, W. and Scherler, D., 2014. TopoToolbox 2–MATLAB-based software for topographic analysis and modeling in Earth surface sciences. Earth Surface Dynamics, 2(1), pp.1-7.

Scott, E. D., & Wilson, L. (2002). Plinian eruptions and passive collapse events as mechanisms of formation for Martian pit chain craters. Journal of Geophysical Research: Planets, 107(E4), 11. https://doi.org/10.1029/2000JE001432

Scott, E. D., Wilson, L., & Head, J. W. (2002). Emplacement of giant radial dikes in the northern Tharsis region of Mars. Journal of Geophysical Research: Planets, 107(E4), 10. https://doi.org/10.1029/2000JE001431

Sims, D. W., Morris, A. P., Ferrill, D. A., Wyrick, D. Y., & Colton, S. L. (2003). Physical models of pit chain formation over dilational faults on Mars. In Lunar and Planetary Science Conference Proceedings (Vol. 34).

Sullivan, R., Greeley, R., Pappalardo, R., Asphaug, E., Moore, J. M., Morrison, D., et al. (1996). Geology of 243 Ida. Icarus, 120(1), 119–139.

Sullivan, R., Anderson, R., Biesiadecki, J., Bond, T., & Stewart, H. (2011). Cohesions, friction angles, and other physical properties of Martian regolith from Mars Exploration Rover

172

wheel trenches and wheel scuffs. Journal of Geophysical Research E: Planets, 116(2). https://doi.org/10.1029/2010JE003625

Tanaka, K. L., & Golombek, M. P. (1989). Martian tension fractures and the formation of grabens and collapse features at Valles Marineris. In Lunar and Planetary Science Conference Proceedings (Vol. 19).

Tanaka, K. L. (1997). Origin of Valles Marineris and Noctis Labyrinthus, Mars, by structurally controlled collapse and erosion of crustal materials. In Lunar and Planetary Science Conference Proceedings (Vol. 28).

Tanaka, K. L., Skinner Jr, J. A., Dohm, J. M., Irwin III, R. P., Kolb, E. J., Fortezzo, C. M., et al. (2014). Geologic map of Mars.

Trudgill, B., & Cartwright, J. (1994). Relay-ramp forms and normal-fault linkages, Canyonlands National Park, Utah. Geological Society of America Bulletin, 106(9), 1143–1157. https://doi.org/10.1130/0016-7606(1994)

Trudgill, B. D. (2002). Structural controls on drainage development in the Canyonlands grabens of southeast Utah. AAPG Bulletin, 86(6), 1095–1112.

Walsh, P., & Schultz-Ela, D. D. (2003). Mechanics of graben evolution in Canyonlands National Park, Utah. Bulletin of the Geological Society of America, 115(3), 259–270. https://doi.org/10.1130/0016-7606(2003)115

Whitten, J. L., & Martin, E. S. (2019). Icelandic pit chains as planetary analogs: Using morphologic measurements of pit chains to determine regolith thickness. Journal of Geophysical Research: Planets, 124(11), 2983-2999. https://doi.org/10.1029/2019JE006099

Wilson, L., & Head, J. W. (2002). Tharsis-radial graben systems as the surface manifestation of plume-related dike intrusion complexes: Models and implications. Journal of Geophysical Research: Planets, 107(E8), 24. https://doi.org/10.1029/2001JE001593

Wyrick, D. Y., Ferrill, D. A., Morris, A. P., Colton, S. L., & Sims, D. W. (2004). Distribution, morphology, and origins of Martian pit crater chains. Journal of Geophysical Research E: Planets, 109(6). https://doi.org/10.1029/2004JE002240

Wyrick, D. Y., Buczkowski, D. L., Bleamaster, L. F., & Collins, G. C. (2010). Pit Crater Chains Across the Solar System. In Lunar and Planetary Science Conference. (Vol. 41).

Wyrick, D. Y., & Smart, K. J. (2009). Dike-induced deformation and Martian graben systems. Journal of Volcanology and Geothermal Research, 185(1–2), 1–11. https://doi.org/10.1016/j.jvolgeores.2008.11.022

173

Wyrick, D. Y., Morris, A. P., Todt, M. K., & Watson-Morris, M. J. (2015). Physical analogue modelling of Martian dyke-induced deformation. Geological Society, London, Special Publications, 401(1), 395-403.

174

APPENDICES

175

Appendix A: Table of Collated it Crater Data Mars and Venus Pits

Cylinder Cone Infilling Magnitude Diameter Major Depth Dataset Planet Volume Volume Required of Infilling (m) Axis (m) (m) (km^3) (km^3) (km^3) Required

Wyrick et al. (2004) Mars 1262 671 2.66 0.2798 2.3805 8.5 Wyrick et al. (2004) Mars 1550 1641 7.99 1.0321 6.9587 6.7 Wyrick et al. (2004) Mars 1439 1007 4.55 0.5459 4.0065 7.3 Wyrick et al. (2004) Mars 1688 1156 6.13 0.8623 5.2680 6.1 Wyrick et al. (2004) Mars 1753 674 3.71 0.5422 3.1696 5.8 Wyrick et al. (2004) Mars 1098 448 1.55 0.1414 1.4040 9.9 Wyrick et al. (2004) Mars 908 688 1.96 0.1485 1.8141 12.2 Wyrick et al. (2004) Mars 1645 661 3.42 0.4683 2.9477 6.3 Wyrick et al. (2004) Mars 1624 757 3.86 0.5227 3.3395 6.4 Wyrick et al. (2004) Mars 1164 907 3.32 0.3217 2.9950 9.3 Wyrick et al. (2004) Mars 1735 2055 11.20 1.6195 9.5816 5.9 Wyrick et al. (2004) Mars 1739 730 3.99 0.5780 3.4102 5.9 Wyrick et al. (2004) Mars 2303 2682 19.40 3.7240 15.6805 4.2 Wyrick et al. (2004) Mars 1200 312 1.18 0.1176 1.0586 9.0 Wyrick et al. (2004) Mars 1198 597 2.25 0.2243 2.0226 9.0 Wyrick et al. (2004) Mars 1424 340 1.52 0.1805 1.3405 7.4 Wyrick et al. (2004) Mars 1324 394 1.64 0.1808 1.4580 8.1 Wyrick et al. (2004) Mars 1573 438 2.16 0.2837 1.8807 6.6 Wyrick et al. (2004) Mars 1392 548 2.40 0.2780 2.1185 7.6 Wyrick et al. (2004) Mars 1616 749 3.80 0.5121 3.2905 6.4 Wyrick et al. (2004) Mars 1653 790 4.10 0.5651 3.5374 6.3 Wyrick et al. (2004) Mars 1247 332 1.30 0.1352 1.1655 8.6 Wyrick et al. (2004) Mars 1123 461 1.63 0.1522 1.4742 9.7 Wyrick et al. (2004) Mars 1334 462 1.94 0.2152 1.7209 8.0 Wyrick et al. (2004) Mars 1384 878 3.82 0.4403 3.3772 7.7 Wyrick et al. (2004) Mars 1159 455 1.66 0.1600 1.4967 9.4 Wyrick et al. (2004) Mars 1661 774 4.04 0.5590 3.4798 6.2 Wyrick et al. (2004) Mars 1519 1736 8.28 1.0487 7.2357 6.9 Wyrick et al. (2004) Mars 1465 439 2.02 0.2467 1.7738 7.2 Wyrick et al. (2004) Mars 2076 360 2.35 0.4062 1.9417 4.8 Wyrick et al. (2004) Mars 1207 316 1.20 0.1205 1.0777 8.9 Wyrick et al. (2004) Mars 987 465 1.44 0.1186 1.3233 11.2 Wyrick et al. (2004) Mars 2301 719 5.20 0.9966 4.2009 4.2 Wyrick et al. (2004) Mars 2007 1221 7.70 1.2876 6.4110 5.0 Wyrick et al. (2004) Mars 1286 317 1.28 0.1372 1.1435 8.3 Wyrick et al. (2004) Mars 1025 404 1.30 0.1111 1.1898 10.7 Wyrick et al. (2004) Mars 1063 582 1.94 0.1722 1.7714 10.3 Wyrick et al. (2004) Mars 1468 485 2.24 0.2736 1.9631 7.2

176

Wyrick et al. (2004) Mars 1310 308 1.27 0.1384 1.1292 8.2 Wyrick et al. (2004) Mars 1234 452 1.75 0.1802 1.5721 8.7 Wyrick et al. (2004) Mars 1026 329 1.06 0.0907 0.9698 10.7 Wyrick et al. (2004) Mars 1390 1181 5.16 0.5974 4.5598 7.6 Wyrick et al. (2004) Mars 1138 362 1.29 0.1227 1.1715 9.5 Wyrick et al. (2004) Mars 1131 367 1.30 0.1229 1.1811 9.6 Wyrick et al. (2004) Mars 2812 808 7.14 1.6727 5.4653 3.3 Wyrick et al. (2004) Mars 2824 1286 11.41 2.6850 8.7242 3.2 Wyrick et al. (2004) Mars 1012 639 2.03 0.1713 1.8602 10.9 Wyrick et al. (2004) Mars 1080 529 1.79 0.1615 1.6333 10.1 Wyrick et al. (2004) Mars 1045 387 1.27 0.1106 1.1599 10.5 Wyrick et al. (2004) Mars 2360 915 6.78 1.3342 5.4498 4.1 Wyrick et al. (2004) Mars 1102 395 1.37 0.1256 1.2419 9.9 Wyrick et al. (2004) Mars 2733 1059 9.09 2.0708 7.0217 3.4 Wyrick et al. (2004) Mars 1519 1168 5.57 0.7055 4.8682 6.9 Wyrick et al. (2004) Mars 1968 2124 13.13 2.1536 10.9783 5.1 Wyrick et al. (2004) Mars 2910 2228 20.37 4.9393 15.4291 3.1 Wyrick et al. (2004) Mars 1491 695 3.26 0.4045 2.8510 7.0 Wyrick et al. (2004) Mars 930 286 0.84 0.0648 0.7708 11.9 Wyrick et al. (2004) Mars 1447 284 1.29 0.1557 1.1354 7.3 Wyrick et al. (2004) Mars 1259 398 1.57 0.1652 1.4090 8.5 Wyrick et al. (2004) Mars 1242 827 3.23 0.3340 2.8929 8.7 Wyrick et al. (2004) Mars 2336 1872 13.74 2.6744 11.0638 4.1 Wyrick et al. (2004) Mars 4531 6391 90.97 34.3499 56.6231 1.6 Wyrick et al. (2004) Mars 2272 1881 13.43 2.5420 10.8840 4.3 Wyrick et al. (2004) Mars 1848 3671 21.31 3.2821 18.0305 5.5 Wyrick et al. (2004) Mars 2673 2577 21.64 4.8204 16.8199 3.5 Wyrick et al. (2004) Mars 2936 2711 25.01 6.1180 18.8875 3.1 Wyrick et al. (2004) Mars 1415 345 1.53 0.1808 1.3528 7.5 Wyrick et al. (2004) Mars 1159 864 3.15 0.3038 2.8421 9.4 Wyrick et al. (2004) Mars 2079 776 5.07 0.8781 4.1903 4.8 Wyrick et al. (2004) Mars 2062 1692 10.96 1.8834 9.0773 4.8 Wyrick et al. (2004) Mars 2097 1016 6.69 1.1697 5.5237 4.7 Wyrick et al. (2004) Mars 1843 3122 18.08 2.7762 15.3000 5.5 Wyrick et al. (2004) Mars 2123 954 6.36 1.1257 5.2371 4.7 Wyrick et al. (2004) Mars 2038 3214 20.58 3.4948 17.0830 4.9 Wyrick et al. (2004) Mars 1876 593 3.49 0.5464 2.9485 5.4 Wyrick et al. (2004) Mars 1610 582 2.94 0.3950 2.5488 6.5 Wyrick et al. (2004) Mars 1860 986 5.76 0.8930 4.8685 5.5 Wyrick et al. (2004) Mars 2052 2885 18.60 3.1803 15.4180 4.8 Wyrick et al. (2004) Mars 1259 277 1.10 0.1149 0.9807 8.5 Wyrick et al. (2004) Mars 1349 574 2.43 0.2735 2.1592 7.9 Wyrick et al. (2004) Mars 1447 1124 5.11 0.6161 4.4934 7.3

177

Wyrick et al. (2004) Mars 1305 2186 8.96 0.9746 7.9875 8.2 Wyrick et al. (2004) Mars 1415 378 1.68 0.1981 1.4822 7.5 Wyrick et al. (2004) Mars 1230 719 2.78 0.2848 2.4936 8.8 Wyrick et al. (2004) Mars 1559 2442 11.96 1.5538 10.4064 6.7 Wyrick et al. (2004) Mars 1087 466 1.59 0.1441 1.4472 10.0 Wyrick et al. (2004) Mars 1167 493 1.81 0.1758 1.6317 9.3 Wyrick et al. (2004) Mars 1072 539 1.82 0.1622 1.6531 10.2 Wyrick et al. (2004) Mars 1167 360 1.32 0.1284 1.1915 9.3 Wyrick et al. (2004) Mars 1804 2518 14.27 2.1453 12.1253 5.7 Wyrick et al. (2004) Mars 1658 639 3.33 0.4599 2.8685 6.2 Wyrick et al. (2004) Mars 1568 729 3.59 0.4692 3.1218 6.7 Wyrick et al. (2004) Mars 1356 540 2.30 0.2599 2.0405 7.8 Wyrick et al. (2004) Mars 1504 2519 11.90 1.4917 10.4104 7.0 Wyrick et al. (2004) Mars 1202 358 1.35 0.1354 1.2165 9.0 Wyrick et al. (2004) Mars 3141 272 2.68 0.7025 1.9815 2.8 Wyrick et al. (2004) Mars 2334 177 1.30 0.2524 1.0454 4.1 Wyrick et al. (2004) Mars 2145 252 1.70 0.3035 1.3946 4.6 Wyrick et al. (2004) Mars 2333 294 2.15 0.4189 1.7359 4.1 Wyrick et al. (2004) Mars 2239 300 2.11 0.3937 1.7165 4.4 Wyrick et al. (2004) Mars 1512 819 3.89 0.4902 3.4001 6.9 Wyrick et al. (2004) Mars 3268 1489 15.29 4.1632 11.1240 2.7 Wyrick et al. (2004) Mars 3716 1805 21.07 6.5253 14.5466 2.2 Wyrick et al. (2004) Mars 1346 482 2.04 0.2286 1.8096 7.9 Wyrick et al. (2004) Mars 1986 906 5.65 0.9355 4.7172 5.0 Wyrick et al. (2004) Mars 2842 1379 12.31 2.9160 9.3963 3.2 Wyrick et al. (2004) Mars 2223 1129 7.88 1.4606 6.4240 4.4 Wyrick et al. (2004) Mars 1730 808 4.39 0.6331 3.7583 5.9 Wyrick et al. (2004) Mars 2024 1732 11.01 1.8575 9.1555 4.9 Wyrick et al. (2004) Mars 3005 1404 13.25 3.3191 9.9353 3.0 Wyrick et al. (2004) Mars 2966 1611 15.01 3.7103 11.3010 3.0 Wyrick et al. (2004) Mars 1710 1064 5.72 0.8145 4.9014 6.0 Wyrick et al. (2004) Mars 1634 1162 5.96 0.8122 5.1527 6.3 Wyrick et al. (2004) Mars 1843 1126 6.52 1.0013 5.5182 5.5 Wyrick et al. (2004) Mars 1713 1047 5.63 0.8043 4.8302 6.0 Wyrick et al. (2004) Mars 1854 1293 7.53 1.1636 6.3675 5.5 Wyrick et al. (2004) Mars 1287 840 3.40 0.3643 3.0321 8.3 Wyrick et al. (2004) Mars 1544 618 3.00 0.3857 2.6120 6.8 Wyrick et al. (2004) Mars 1737 1081 5.90 0.8539 5.0451 5.9 Wyrick et al. (2004) Mars 1669 1020 5.35 0.7438 4.6043 6.2 Wyrick et al. (2004) Mars 1443 544 2.47 0.2966 2.1696 7.3 Wyrick et al. (2004) Mars 1438 853 3.85 0.4618 3.3917 7.3 Wyrick et al. (2004) Mars 1572 1061 5.24 0.6864 4.5534 6.6 Wyrick et al. (2004) Mars 1716 758 4.09 0.5843 3.5020 6.0

178

Wyrick et al. (2004) Mars 1997 1043 6.54 1.0890 5.4546 5.0 Wyrick et al. (2004) Mars 2054 1134 7.32 1.2525 6.0650 4.8 Wyrick et al. (2004) Mars 2377 1666 12.44 2.4643 9.9766 4.0 Wyrick et al. (2004) Mars 2866 2257 20.32 4.8535 15.4681 3.2 Wyrick et al. (2004) Mars 1661 988 5.16 0.7136 4.4420 6.2 Wyrick et al. (2004) Mars 1459 586 2.69 0.3266 2.3594 7.2 Wyrick et al. (2004) Mars 1749 931 5.12 0.7456 4.3699 5.9 Wyrick et al. (2004) Mars 2205 1015 7.03 1.2920 5.7392 4.4 Wyrick et al. (2004) Mars 1700 786 4.20 0.5947 3.6031 6.1 Wyrick et al. (2004) Mars 1380 769 3.33 0.3834 2.9505 7.7 Wyrick et al. (2004) Mars 1730 1357 7.38 1.0633 6.3120 5.9 Wyrick et al. (2004) Mars 1616 837 4.25 0.5722 3.6771 6.4 Wyrick et al. (2004) Mars 1587 777 3.87 0.5123 3.3616 6.6 Wyrick et al. (2004) Mars 1986 1750 10.92 1.8070 9.1116 5.0 Wyrick et al. (2004) Mars 1478 873 4.05 0.4993 3.5543 7.1 Wyrick et al. (2004) Mars 1571 1023 5.05 0.6610 4.3880 6.6 Wyrick et al. (2004) Mars 1945 1345 8.22 1.3321 6.8864 5.2 Wyrick et al. (2004) Mars 2119 2005 13.35 2.3569 10.9904 4.7 Wyrick et al. (2004) Mars 1780 915 5.12 0.7590 4.3577 5.7 Wyrick et al. (2004) Mars 1468 554 2.55 0.3126 2.2424 7.2 Wyrick et al. (2004) Mars 1304 664 2.72 0.2956 2.4246 8.2 Wyrick et al. (2004) Mars 1674 689 3.62 0.5055 3.1180 6.2 Wyrick et al. (2004) Mars 2004 922 5.80 0.9694 4.8353 5.0 Wyrick et al. (2004) Mars 2067 1196 7.77 1.3378 6.4287 4.8 Wyrick et al. (2004) Mars 1372 501 2.16 0.2469 1.9125 7.7 Wyrick et al. (2004) Mars 1592 1237 6.19 0.8208 5.3660 6.5 Wyrick et al. (2004) Mars 1586 719 3.58 0.4735 3.1090 6.6 Wyrick et al. (2004) Mars 2183 587 4.03 0.7323 3.2934 4.5 Wyrick et al. (2004) Mars 2384 1806 13.53 2.6872 10.8389 4.0 Wyrick et al. (2004) Mars 2182 887 6.08 1.1056 4.9747 4.5 Wyrick et al. (2004) Mars 2259 892 6.33 1.1917 5.1387 4.3 Wyrick et al. (2004) Mars 2433 802 6.13 1.2429 4.8872 3.9 Wyrick et al. (2004) Mars 1447 785 3.57 0.4303 3.1382 7.3 Wyrick et al. (2004) Mars 1845 696 4.03 0.6203 3.4139 5.5 Wyrick et al. (2004) Mars 2125 784 5.23 0.9268 4.3071 4.6 Wyrick et al. (2004) Mars 1410 722 3.20 0.3758 2.8224 7.5 Wyrick et al. (2004) Mars 1808 774 4.40 0.6624 3.7339 5.6 Wyrick et al. (2004) Mars 2247 972 6.86 1.2848 5.5767 4.3 Wyrick et al. (2004) Mars 2554 2022 16.22 3.4530 12.7708 3.7 Wyrick et al. (2004) Mars 2674 1155 9.70 2.1621 7.5406 3.5 Wyrick et al. (2004) Mars 2535 1303 10.38 2.1921 8.1849 3.7 Wyrick et al. (2004) Mars 1913 1029 6.18 0.9859 5.1983 5.3 Wyrick et al. (2004) Mars 2174 1632 11.15 2.0193 9.1269 4.5

179

Wyrick et al. (2004) Mars 2010 1100 6.95 1.1635 5.7826 5.0 Wyrick et al. (2004) Mars 2783 1978 17.29 4.0107 13.2830 3.3 Wyrick et al. (2004) Mars 2901 1882 17.15 4.1465 13.0056 3.1 Wyrick et al. (2004) Mars 2890 1255 11.39 2.7442 8.6502 3.2 Wyrick et al. (2004) Mars 2072 1919 12.49 2.1569 10.3346 4.8 Wyrick et al. (2004) Mars 2076 1740 11.35 1.9632 9.3850 4.8 Wyrick et al. (2004) Mars 2136 1019 6.84 1.2172 5.6208 4.6 Wyrick et al. (2004) Mars 2750 1529 13.21 3.0272 10.1824 3.4 Wyrick et al. (2004) Mars 2759 1005 8.71 2.0028 6.7082 3.3 Wyrick et al. (2004) Mars 2027 1447 9.21 1.5565 7.6580 4.9 Wyrick et al. (2004) Mars 2771 2098 18.26 4.2174 14.0464 3.3 Wyrick et al. (2004) Mars 2848 1464 13.10 3.1088 9.9900 3.2 Wyrick et al. (2004) Mars 2933 1582 14.58 3.5629 11.0141 3.1 Wyrick et al. (2004) Mars 2452 2414 18.60 3.7997 14.7958 3.9 Wyrick et al. (2004) Mars 2338 1101 8.09 1.5756 6.5113 4.1 Wyrick et al. (2004) Mars 2486 1390 10.86 2.2490 8.6069 3.8 Wyrick et al. (2004) Mars 2290 1680 12.09 2.3065 9.7799 4.2 Wyrick et al. (2004) Mars 2081 1104 7.22 1.2516 5.9659 4.8 Wyrick et al. (2004) Mars 2114 1072 7.12 1.2542 5.8653 4.7 Wyrick et al. (2004) Mars 2245 879 6.20 1.1598 5.0397 4.3 Wyrick et al. (2004) Mars 2067 870 5.65 0.9731 4.6764 4.8 Wyrick et al. (2004) Mars 2220 1732 12.08 2.2347 9.8448 4.4 Wyrick et al. (2004) Mars 2219 965 6.73 1.2440 5.4832 4.4 Wyrick et al. (2004) Mars 2245 907 6.40 1.1968 5.2002 4.3 Wyrick et al. (2004) Mars 2324 958 6.99 1.3546 5.6398 4.2 Wyrick et al. (2004) Mars 3505 1083 11.93 3.4832 8.4421 2.4 Wyrick et al. (2004) Mars 3688 1833 21.24 6.5270 14.7105 2.3 Wyrick et al. (2004) Mars 4160 1385 18.10 6.2749 11.8257 1.9 Wyrick et al. (2004) Mars 3037 1039 9.91 2.5088 7.4043 3.0 Wyrick et al. (2004) Mars 3061 1078 10.37 2.6443 7.7222 2.9 Wyrick et al. (2004) Mars 3268 2524 25.91 7.0570 18.8562 2.7 Wyrick et al. (2004) Mars 2384 1542 11.55 2.2944 9.2545 4.0 Wyrick et al. (2004) Mars 1590 619 3.09 0.4097 2.6823 6.5 Wyrick et al. (2004) Mars 2611 1371 11.25 2.4469 8.7990 3.6 Wyrick et al. (2004) Mars 3009 1069 10.11 2.5339 7.5714 3.0 Wyrick et al. (2004) Mars 2685 1315 11.09 2.4819 8.6104 3.5 Wyrick et al. (2004) Mars 2412 1122 8.50 1.7089 6.7931 4.0 Wyrick et al. (2004) Mars 2583 2112 17.14 3.6890 13.4493 3.6 Wyrick et al. (2004) Mars 2460 3005 23.22 4.7608 18.4628 3.9 Wyrick et al. (2004) Mars 1702 1277 6.83 0.9685 5.8597 6.1 Wyrick et al. (2004) Mars 1845 1329 7.70 1.1844 6.5188 5.5 Wyrick et al. (2004) Mars 1515 771 3.67 0.4633 3.2063 6.9 Wyrick et al. (2004) Mars 1148 1321 4.76 0.4558 4.3085 9.5

180

Wyrick et al. (2004) Mars 1339 652 2.74 0.3060 2.4367 8.0 Wyrick et al. (2004) Mars 1283 1775 7.15 0.7649 6.3895 8.4 Wyrick et al. (2004) Mars 2709 1488 12.66 2.8588 9.8049 3.4 Wyrick et al. (2004) Mars 2598 1981 16.17 3.5005 12.6681 3.6 Wyrick et al. (2004) Mars 2148 1287 8.68 1.5546 7.1303 4.6 Wyrick et al. (2004) Mars 2102 2287 15.10 2.6455 12.4570 4.7 Wyrick et al. (2004) Mars 2003 533 3.35 0.5598 2.7941 5.0 Wyrick et al. (2004) Mars 2070 969 6.30 1.0870 5.2145 4.8 Wyrick et al. (2004) Mars 1544 507 2.46 0.3164 2.1428 6.8 Wyrick et al. (2004) Mars 2058 1503 9.72 1.6666 8.0509 4.8 Wyrick et al. (2004) Mars 2034 720 4.60 0.7798 3.8210 4.9 Wyrick et al. (2004) Mars 1859 1511 8.82 1.3671 7.4575 5.5 Wyrick et al. (2004) Mars 1550 1015 4.94 0.6384 4.3041 6.7 Wyrick et al. (2004) Mars 1569 1043 5.14 0.6722 4.4689 6.6 Wyrick et al. (2004) Mars 1517 824 3.93 0.4964 3.4306 6.9 Wyrick et al. (2004) Mars 1452 810 3.69 0.4471 3.2478 7.3 Wyrick et al. (2004) Mars 1653 441 2.29 0.3155 1.9747 6.3 Wyrick et al. (2004) Mars 1499 1774 8.35 1.0436 7.3106 7.0 Wyrick et al. (2004) Mars 1446 517 2.35 0.2830 2.0656 7.3 Wyrick et al. (2004) Mars 1298 951 3.88 0.4195 3.4585 8.2 Wyrick et al. (2004) Mars 1057 1466 4.87 0.4288 4.4393 10.4 Wyrick et al. (2004) Mars 1086 1539 5.25 0.4752 4.7755 10.0 Wyrick et al. (2004) Mars 1218 773 2.96 0.3002 2.6576 8.9 Wyrick et al. (2004) Mars 1199 704 2.65 0.2650 2.3868 9.0 Wyrick et al. (2004) Mars 1463 1122 5.16 0.6287 4.5282 7.2 Wyrick et al. (2004) Mars 1309 1588 6.53 0.7124 5.8180 8.2 Wyrick et al. (2004) Mars 1433 728 3.28 0.3914 2.8860 7.4 Wyrick et al. (2004) Mars 1086 1167 3.98 0.3603 3.6212 10.0 Wyrick et al. (2004) Mars 1839 1030 5.95 0.9119 5.0388 5.5 Wyrick et al. (2004) Mars 1568 1075 5.30 0.6919 4.6035 6.7 Wyrick et al. (2004) Mars 1832 956 5.50 0.8400 4.6622 5.6 Wyrick et al. (2004) Mars 1834 1220 7.03 1.0743 5.9549 5.5 Wyrick et al. (2004) Mars 1557 689 3.37 0.4373 2.9329 6.7 Wyrick et al. (2004) Mars 1312 519 2.14 0.2339 1.9053 8.1 Wyrick et al. (2004) Mars 2172 1180 8.05 1.4574 6.5944 4.5 Wyrick et al. (2004) Mars 2083 1364 8.93 1.5494 7.3765 4.8 Wyrick et al. (2004) Mars 1789 826 4.64 0.6921 3.9503 5.7 Wyrick et al. (2004) Mars 1586 546 2.72 0.3596 2.3609 6.6 Wyrick et al. (2004) Mars 1296 714 2.91 0.3140 2.5931 8.3 Wyrick et al. (2004) Mars 1425 825 3.69 0.4386 3.2548 7.4 Wyrick et al. (2004) Mars 2070 892 5.80 1.0006 4.8001 4.8 Wyrick et al. (2004) Mars 2224 1009 7.05 1.3066 5.7432 4.4 Wyrick et al. (2004) Mars 1859 866 5.06 0.7835 4.2741 5.5

181

Wyrick et al. (2004) Mars 1717 968 5.22 0.7471 4.4744 6.0 Wyrick et al. (2004) Mars 1751 640 3.52 0.5137 3.0069 5.9 Wyrick et al. (2004) Mars 1725 1107 6.00 0.8624 5.1367 6.0 Wyrick et al. (2004) Mars 2074 1051 6.85 1.1836 5.6644 4.8 Wyrick et al. (2004) Mars 2065 1584 10.28 1.7683 8.5077 4.8 Wyrick et al. (2004) Mars 2117 1575 10.47 1.8480 8.6270 4.7 Wyrick et al. (2004) Mars 2099 2328 15.35 2.6852 12.6661 4.7 Kling Noctis Mars 7720 8722 1409 34.17 21.9844 12.1882 0.6 Kling Noctis Mars 2272 8002 344 2.46 0.4649 1.9905 4.3 Kling Noctis Mars 4150 11929 816 10.64 3.6792 6.9595 1.9 Kling Noctis Mars 4171 5601 810 10.61 3.6892 6.9247 1.9 Kling Noctis Mars 3117 5583 582 5.70 1.4804 4.2188 2.8 Kling Noctis Mars 2752 5072 615 5.32 1.2194 4.0977 3.4 Kling Noctis Mars 5037 9430 896 14.18 5.9514 8.2271 1.4 Kling Noctis Mars 2811 3711 447 3.95 0.9247 3.0228 3.3 Kling Noctis Mars 1683 2620 214 1.13 0.1587 0.9728 6.1 Kling Noctis Mars 3184 3732 603 6.03 1.6004 4.4313 2.8 Kling Noctis Mars 1951 2253 258 1.58 0.2571 1.3242 5.2 Kling Noctis Mars 3185 6784 535 5.35 1.4208 3.9324 2.8 Kling Noctis Mars 1642 2069 198 1.02 0.1398 0.8816 6.3 Kling Noctis Mars 1666 2907 125 0.65 0.0908 0.5634 6.2 Kling Noctis Mars 1219 1754 95 0.36 0.0370 0.3269 8.8 Kling Noctis Mars 1268 1446 94 0.37 0.0396 0.3349 8.5 Kling Noctis Mars 1499 2003 102 0.48 0.0600 0.4203 7.0 Kling Noctis Mars 2215 4059 419 2.92 0.5382 2.3775 4.4 Kling Noctis Mars 2036 2803 327 2.09 0.3549 1.7367 4.9 Kling Noctis Mars 1261 2513 167 0.66 0.0695 0.5921 8.5 Kling Noctis Mars 1173 1917 296 1.09 0.1066 0.9842 9.2 Kling Noctis Mars 1592 2523 306 1.53 0.2030 1.3274 6.5 Kling Noctis Mars 1614 2587 284 1.44 0.1937 1.2463 6.4 Kling Noctis Mars 2730 3502 459 3.94 0.8956 3.0411 3.4 Kling Noctis Mars 1764 2086 226 1.25 0.1841 1.0683 5.8 Kling Noctis Mars 1358 1653 159 0.68 0.0768 0.6016 7.8 Kling Noctis Mars 1946 3635 254 1.55 0.2518 1.3010 5.2 Kling Noctis Mars 1415 1618 171 0.76 0.0896 0.6705 7.5 Kling Noctis Mars 1066 1200 135 0.45 0.0402 0.4119 10.3 Kling Noctis Mars 805 1047 88 0.22 0.0149 0.2076 13.9 Kling Noctis Mars 1175 1373 87 0.32 0.0314 0.2897 9.2 Kling Noctis Mars 1593 2175 161 0.81 0.1070 0.6988 6.5 Kling Noctis Mars 3029 4696 522 4.97 1.2538 3.7135 3.0 Kling Noctis Mars 2734 4664 467 4.01 0.9139 3.0973 3.4 Kling Noctis Mars 2036 2157 222 1.42 0.2409 1.1791 4.9 Kling Noctis Mars 2638 3523 390 3.23 0.7105 2.5216 3.5

182

Kling Noctis Mars 1295 1689 115 0.47 0.0505 0.4174 8.3 Kling Noctis Mars 957 1147 86 0.26 0.0206 0.2379 11.5 Kling Noctis Mars 824 1099 65 0.17 0.0116 0.1567 13.6 Kling Noctis Mars 516 648 25 0.04 0.0017 0.0388 22.3 Kling Noctis Mars 600 822 26 0.05 0.0025 0.0466 19.0 Kling Noctis Mars 1717 2150 193 1.04 0.1490 0.8921 6.0 Kling Noctis Mars 3360 4057 541 5.71 1.5990 4.1117 2.6 Kling Noctis Mars 1197 1406 100 0.38 0.0375 0.3385 9.0 Kling Noctis Mars 1114 2067 88 0.31 0.0286 0.2794 9.8 Kling Noctis Mars 2687 3522 504 4.25 0.9527 3.3018 3.5 Kling Noctis Mars 1155 2213 112 0.41 0.0391 0.3673 9.4 Kling Noctis Mars 1332 1955 127 0.53 0.0590 0.4725 8.0 Kling Noctis Mars 2068 5481 332 2.16 0.3717 1.7852 4.8 Kling Noctis Mars 2458 4916 414 3.20 0.6548 2.5421 3.9 Kling Noctis Mars 741 1031 29 0.07 0.0042 0.0633 15.2 Kling Noctis Mars 954 1139 74 0.22 0.0176 0.2042 11.6 Kling Noctis Mars 1269 1467 122 0.49 0.0514 0.4349 8.5 Kling Noctis Mars 1234 1471 100 0.39 0.0399 0.3478 8.7 Kling Noctis Mars 1904 3651 156 0.93 0.1481 0.7851 5.3 Kling Noctis Mars 3258 4075 526 5.38 1.4617 3.9221 2.7 Kling Noctis Mars 1359 1641 156 0.67 0.0754 0.5906 7.8 Kling Noctis Mars 1296 1902 133 0.54 0.0585 0.4830 8.3 Kling Noctis Mars 1148 2154 138 0.50 0.0476 0.4501 9.5 Kling Noctis Mars 2307 2500 319 2.31 0.4445 1.8675 4.2 Kling Noctis Mars 2206 4082 239 1.66 0.3045 1.3519 4.4 Kling Noctis Mars 1900 2092 227 1.35 0.2145 1.1404 5.3 Kling Noctis Mars 1655 1813 138 0.72 0.0990 0.6186 6.3 Kling Noctis Mars 2012 2580 254 1.61 0.2692 1.3363 5.0 Kling Noctis Mars 1126 1196 77 0.27 0.0256 0.2468 9.7 Kling Noctis Mars 2897 4867 569 5.18 1.2502 3.9284 3.1 Kling Noctis Mars 1542 1822 142 0.69 0.0884 0.5995 6.8 Kling Noctis Mars 2701 5929 509 4.32 0.9722 3.3469 3.4 Kling Noctis Mars 5795 6651 1082 19.70 9.5127 10.1857 1.1 Kling Noctis Mars 2701 3110 575 4.88 1.0982 3.7809 3.4 Kling Noctis Mars 2306 2841 384 2.78 0.5346 2.2473 4.2 Kling Noctis Mars 1539 1823 188 0.91 0.1166 0.7924 6.8 Kling Noctis Mars 4584 6654 904 13.02 4.9731 8.0455 1.6 Kling Noctis Mars 2525 4440 454 3.60 0.7578 2.8436 3.8 Kling Noctis Mars 4087 5561 619 7.95 2.7069 5.2409 1.9 Kling Noctis Mars 1038 1863 81 0.26 0.0228 0.2413 10.6 Kling Noctis Mars 8498 13499 1864 49.76 35.2410 14.5227 0.4 Kling Noctis Mars 6340 7803 1222 24.34 12.8593 11.4801 0.9 Kling Noctis Mars 3200 3796 664 6.68 1.7801 4.8952 2.8

183

Kling Noctis Mars 12391 38642 1911 74.39 76.8143 -2.4239 0.0 Kling Noctis Mars 2450 2994 396 3.05 0.6223 2.4257 3.9 Kling Noctis Mars 5360 18817 1019 17.16 7.6643 9.4946 1.2 Kling Noctis Mars 3058 3753 497 4.77 1.2167 3.5579 2.9 Kling Noctis Mars 2226 3635 251 1.76 0.3256 1.4297 4.4 Kling Noctis Mars 2563 7081 417 3.36 0.7171 2.6405 3.7 Kling Noctis Mars 1593 3270 146 0.73 0.0970 0.6337 6.5 Kling Noctis Mars 1657 2467 234 1.22 0.1682 1.0499 6.2 Kling Noctis Mars 4663 9111 1005 14.72 5.7209 9.0016 1.6 Kling Noctis Mars 2072 3648 343 2.23 0.3855 1.8472 4.8 Kling Noctis Mars 1527 2995 181 0.87 0.1105 0.7578 6.9 Kling Noctis Mars 1958 2382 221 1.36 0.2218 1.1376 5.1 Kling Noctis Mars 1646 1932 156 0.81 0.1107 0.6960 6.3 Kling Noctis Mars 1888 3142 69 0.41 0.0644 0.3449 5.4 Kling Noctis Mars 993 1194 149 0.46 0.0385 0.4264 11.1 Kling Noctis Mars 2315 2751 367 2.67 0.5149 2.1542 4.2 Kling Noctis Mars 2664 5184 439 3.67 0.8156 2.8584 3.5 Kling Noctis Mars 2979 4925 374 3.50 0.8689 2.6313 3.0 Kling Noctis Mars 751 776 85 0.20 0.0126 0.1880 15.0 Kling Noctis Mars 1347 1557 172 0.73 0.0817 0.6462 7.9 Kling Noctis Mars 2881 8822 702 6.35 1.5254 4.8283 3.2 Kling Noctis Mars 3824 5208 741 8.90 2.8368 6.0652 2.1 Kling Noctis Mars 2669 3541 545 4.57 1.0164 3.5534 3.5 Kling Noctis Mars 2374 2631 323 2.41 0.4766 1.9324 4.1 Kling Noctis Mars 2578 3201 727 5.89 1.2649 4.6231 3.7 Kling Noctis Mars 2707 3961 494 4.20 0.9477 3.2534 3.4 Kling Noctis Mars 3311 8623 526 5.47 1.5096 3.9617 2.6 Kling Noctis Mars 3137 4251 480 4.73 1.2366 3.4939 2.8 Kling Noctis Mars 1529 1782 110 0.53 0.0673 0.4611 6.8 Kling Noctis Mars 2083 2783 234 1.53 0.2658 1.2655 4.8 Kling Noctis Mars 2270 4443 277 1.98 0.3737 1.6017 4.3 Kling Noctis Mars 1866 3619 209 1.23 0.1905 1.0347 5.4 Kling Noctis Mars 1545 2197 130 0.63 0.0812 0.5497 6.8 Kling Noctis Mars 1060 1232 27 0.09 0.0079 0.0820 10.3 Kling Noctis Mars 1282 1629 39 0.16 0.0168 0.1403 8.4 Kling Noctis Mars 2703 3759 511 4.34 0.9774 3.3619 3.4 Kling Noctis Mars 3092 10916 432 4.20 1.0813 3.1151 2.9 Kling Noctis Mars 5448 18788 1105 18.91 8.5863 10.3262 1.2 Kling Noctis Mars 1425 2358 47 0.21 0.0250 0.1854 7.4 Kling Noctis Mars 2079 3167 90 0.59 0.1018 0.4860 4.8 Kling Noctis Mars 1124 1329 47 0.17 0.0155 0.1504 9.7 Kling Noctis Mars 1579 3002 117 0.58 0.0764 0.5040 6.6 Kling Noctis Mars 1678 1743 37 0.20 0.0273 0.1678 6.2

184

Kling Noctis Mars 1380 2173 57 0.25 0.0284 0.2187 7.7 Kling Noctis Mars 1779 1887 87 0.49 0.0721 0.4141 5.7 Kling Noctis Mars 2440 2685 157 1.20 0.2447 0.9588 3.9 Kling Noctis Mars 1359 1720 22 0.09 0.0106 0.0833 7.8 Kling Noctis Mars 1631 1705 34 0.17 0.0237 0.1505 6.4 Kling Noctis Mars 1426 1770 281 1.26 0.1496 1.1093 7.4 Kling Noctis Mars 1811 2617 148 0.84 0.1271 0.7150 5.6 Kling Noctis Mars 2099 3674 526 3.47 0.6067 2.8618 4.7 Kling Noctis Mars 1500 1694 173 0.82 0.1019 0.7133 7.0 Kling Noctis Mars 1100 1416 128 0.44 0.0405 0.4018 9.9 Kling Noctis Mars 1145 1441 84 0.30 0.0288 0.2733 9.5 Kling Noctis Mars 1243 1982 235 0.92 0.0951 0.8226 8.7 Kling Noctis Mars 1124 1700 382 1.35 0.1263 1.2226 9.7 Kling Noctis Mars 1702 1989 229 1.22 0.1737 1.0508 6.1 Kling Noctis Mars 1738 2366 246 1.34 0.1945 1.1486 5.9 Kling Noctis Mars 1679 2070 140 0.74 0.1033 0.6351 6.1 Kling Noctis Mars 2470 4094 494 3.83 0.7890 3.0443 3.9 Kling Noctis Mars 1408 2972 199 0.88 0.1033 0.7770 7.5 Kling Noctis Mars 1639 5024 389 2.00 0.2736 1.7294 6.3 Kling Noctis Mars 909 1012 26 0.07 0.0056 0.0686 12.2 Kling Noctis Mars 4116 11866 826 10.68 3.6635 7.0173 1.9 Kling Noctis Mars 1085 1737 201 0.69 0.0619 0.6232 10.1 Kling Noctis Mars 2087 7569 386 2.53 0.4402 2.0907 4.7 Kling Noctis Mars 1267 2150 29 0.12 0.0122 0.1032 8.5 Kling Noctis Mars 1218 1359 55 0.21 0.0214 0.1891 8.9 Kling Noctis Mars 1473 1597 86 0.40 0.0489 0.3491 7.1 Kling Noctis Mars 944 1035 70 0.21 0.0163 0.1913 11.7 Kling Noctis Mars 561 1048 37 0.07 0.0030 0.0622 20.4 Kling Noctis Mars 1146 1295 51 0.18 0.0175 0.1661 9.5 Kling Noctis Mars 711 844 38 0.08 0.0050 0.0799 15.9 Kling Noctis Mars 2222 3277 324 2.26 0.4188 1.8429 4.4 Kling Noctis Mars 3555 26749 692 7.73 2.2896 5.4389 2.4 Kling Noctis Mars 1098 1216 70 0.24 0.0221 0.2194 9.9 Kling Noctis Mars 1969 3392 140 0.87 0.1421 0.7239 5.1 Kling Noctis Mars 1297 1662 78 0.32 0.0344 0.2835 8.3 Kling Noctis Mars 1153 1721 137 0.50 0.0477 0.4486 9.4 Kling Noctis Mars 2994 3852 322 3.03 0.7557 2.2730 3.0 Kling Noctis Mars 3427 4755 976 10.51 3.0009 7.5070 2.5 Kling Noctis Mars 4370 5472 478 6.56 2.3898 4.1726 1.7 Kling Noctis Mars 2722 3144 272 2.33 0.5276 1.7984 3.4 Kling Noctis Mars 5128 6153 628 10.12 4.3234 5.7937 1.3 Kling Noctis Mars 3395 4111 464 4.95 1.4001 3.5488 2.5 Kling Noctis Mars 4236 5434 580 7.72 2.7246 4.9939 1.8

185

Kling Noctis Mars 2451 4093 225 1.73 0.3539 1.3786 3.9 Kling Noctis Mars 1097 1172 59 0.20 0.0186 0.1847 9.9 Kling Noctis Mars 5777 6914 1167 21.18 10.1963 10.9835 1.1 Kling Noctis Mars 1362 1641 121 0.52 0.0588 0.4590 7.8 Kling Noctis Mars 1828 2113 161 0.92 0.1408 0.7837 5.6 Kling Noctis Mars 1619 1944 141 0.72 0.0968 0.6204 6.4 Kling Noctis Mars 4284 13250 770 10.36 3.6996 6.6635 1.8 Kling Noctis Mars 5894 8080 988 18.29 8.9856 9.3088 1.0 Kling Noctis Mars 1107 1329 237 0.82 0.0760 0.7482 9.8 Kling Noctis Mars 11271 12856 1809 64.05 60.1634 3.8913 0.1 Kling Noctis Mars 2386 2680 159 1.19 0.2370 0.9549 4.0 Kling Noctis Mars 3034 4160 265 2.53 0.6386 1.8872 3.0 Kling Noctis Mars 3099 3761 354 3.45 0.8901 2.5564 2.9 Kling Noctis Mars 5843 6950 855 15.69 7.6420 8.0527 1.1 Kling Noctis Mars 2455 2532 260 2.01 0.4102 1.5950 3.9 Kling Noctis Mars 2603 2957 277 2.27 0.4914 1.7738 3.6 Kling Noctis Mars 2065 12209 476 3.09 0.5314 2.5566 4.8 Kling Noctis Mars 1907 2127 89 0.53 0.0847 0.4485 5.3 Kling Noctis Mars 717 1499 52 0.12 0.0070 0.1101 15.7 Kling Noctis Mars 1299 1731 93 0.38 0.0411 0.3384 8.2 Kling Noctis Mars 1567 2390 325 1.60 0.2089 1.3910 6.7 Kling Noctis Mars 920 2480 141 0.41 0.0312 0.3763 12.0 Kling Noctis Mars 1874 2237 327 1.93 0.3006 1.6245 5.4 Kling Noctis Mars 882 1098 162 0.45 0.0330 0.4159 12.6 Kling Noctis Mars 625 766 52 0.10 0.0053 0.0968 18.2 Kling Noctis Mars 4036 4539 974 12.35 4.1536 8.1962 2.0 Kling Noctis Mars 1873 2308 205 1.21 0.1883 1.0180 5.4 Kling Noctis Mars 1827 2008 330 1.89 0.2884 1.6057 5.6 Kling Noctis Mars 2899 4895 679 6.18 1.4939 4.6900 3.1 Kling Noctis Mars 676 786 22 0.05 0.0026 0.0441 16.8 Kling Noctis Mars 407 731 79 0.10 0.0034 0.0976 28.5 Kling Noctis Mars 944 1674 50 0.15 0.0117 0.1366 11.7 Kling Noctis Mars 2508 12204 535 4.22 0.8810 3.3343 3.8 Kling Noctis Mars 1700 2069 174 0.93 0.1316 0.7976 6.1 Kling Noctis Mars 2108 2977 553 3.66 0.6433 3.0189 4.7 Kling Noctis Mars 2349 2978 410 3.03 0.5923 2.4334 4.1 Kling Noctis Mars 881 1091 82 0.23 0.0167 0.2103 12.6 Kling Noctis Mars 826 2945 191 0.50 0.0341 0.4615 13.5 Kling Noctis Mars 1053 1335 339 1.12 0.0984 1.0230 10.4 Kling Noctis Mars 1872 5840 166 0.98 0.1523 0.8240 5.4 Kling Noctis Mars 4969 5136 945 14.75 6.1086 8.6434 1.4 Kling Noctis Mars 4477 5205 887 12.48 4.6544 7.8211 1.7 Kling Noctis Mars 6347 17672 1339 26.70 14.1217 12.5776 0.9

186

Kling Noctis Mars 5426 6274 1111 18.94 8.5633 10.3751 1.2 Kling Noctis Mars 4126 4996 1122 14.54 5.0006 9.5430 1.9 Kling Noctis Mars 6245 6722 945 18.54 9.6486 8.8916 0.9 Kling Noctis Mars 1270 1588 113 0.45 0.0477 0.4031 8.4 Kling Noctis Mars 1908 2317 240 1.44 0.2287 1.2099 5.3 Kling Noctis Mars 1546 1553 174 0.85 0.1089 0.7362 6.8 Kling Noctis Mars 1035 1259 62 0.20 0.0174 0.1842 10.6 Kling Noctis Mars 1273 1544 63 0.25 0.0267 0.2252 8.4 Kling, Nyx Mons Venus 1534 2153 223 1.07 0.1374 0.9373 6.8 Kling, Nyx Mons Venus 1534 2478 142 0.68 0.0875 0.5968 6.8 Kling, Nyx Mons Venus 1867 2485 367 2.15 0.3349 1.8177 5.4 Kling, Nyx Mons Venus 1830 2047 277 1.59 0.2429 1.3496 5.6 Kling, Nyx Mons Venus 1424 2306 154 0.69 0.0818 0.6072 7.4 Kling, Nyx Mons Venus 1404 2025 157 0.69 0.0810 0.6115 7.5 Kling, Nyx Mons Venus 531 807 253 0.42 0.0187 0.4034 21.6 Kling, Nyx Mons Venus 767 1352 384 0.93 0.0591 0.8661 14.6 Kling, Nyx Mons Venus 685 788 155 0.33 0.0190 0.3145 16.5 Kling, Nyx Mons Venus 365 478 99 0.11 0.0035 0.1101 31.9 Kling, Nyx Mons Venus 819 893 244 0.63 0.0428 0.5850 13.7 Kling, Nyx Mons Venus 767 901 252 0.61 0.0388 0.5684 14.6 Kling, Nyx Mons Venus 1603 3206 499 2.51 0.3357 2.1773 6.5 Kling, Nyx Mons Venus 1231 1713 478 1.85 0.1896 1.6589 8.7 Kling, Nyx Mons Venus 1323 1450 208 0.86 0.0953 0.7692 8.1 Kling, Nyx Mons Venus 747 1160 308 0.72 0.0450 0.6778 15.1 Kling, Nyx Mons Venus 1757 2941 492 2.72 0.3976 2.3181 5.8 Kling, Nyx Mons Venus 873 1119 241 0.66 0.0481 0.6129 12.7 Kling, Nyx Mons Venus 525 919 144 0.24 0.0104 0.2271 21.9 Kling, Nyx Mons Venus 981 1364 197 0.61 0.0496 0.5575 11.2 Kling, Nyx Mons Venus 569 719 133 0.24 0.0113 0.2265 20.1 Kling, Nyx Mons Venus 1955 2365 448 2.75 0.4483 2.3033 5.1 Kling, Nyx Mons Venus 1007 1579 660 2.09 0.1752 1.9128 10.9 Kling, Nyx Mons Venus 476 781 225 0.34 0.0133 0.3231 24.2 Kling, Nyx Mons Venus 803 1254 287 0.72 0.0484 0.6756 13.9 Kling, Nyx Mons Venus 600 664 251 0.47 0.0237 0.4495 19.0 Kling, Nyx Mons Venus 872 1498 258 0.71 0.0514 0.6554 12.8 Kling, Nyx Mons Venus 668 1019 195 0.41 0.0228 0.3864 17.0 Kling, Nyx Mons Venus 646 966 243 0.49 0.0265 0.4666 17.6 Kling, Nyx Mons Venus 999 1147 547 1.72 0.1429 1.5738 11.0 Kling, Nyx Mons Venus 558 886 191 0.33 0.0156 0.3193 20.5 Kling, Nyx Mons Venus 783 1248 297 0.73 0.0477 0.6829 14.3 Kling, Nyx Mons Venus 1937 2347 345 2.10 0.3389 1.7605 5.2 Kling, Nyx Mons Venus 2622 3071 534 4.40 0.9611 3.4376 3.6 Kling, Nyx Mons Venus 3652 5235 579 6.64 2.0217 4.6213 2.3

187

Kling, Nyx Mons Venus 1154 1885 171 0.62 0.0596 0.5603 9.4 Kling, Nyx Mons Venus 1189 1583 224 0.84 0.0829 0.7538 9.1 Kling, Nyx Mons Venus 733 922 347 0.80 0.0488 0.7503 15.4 Kling, Nyx Mons Venus 1265 2179 230 0.91 0.0964 0.8177 8.5 Kling, Nyx Mons Venus 1983 3355 270 1.68 0.2780 1.4041 5.1 Kling, Nyx Mons Venus 1699 3029 291 1.55 0.2199 1.3333 6.1 Kling, Nyx Mons Venus 1579 1998 410 2.03 0.2676 1.7662 6.6 Kling, Nyx Mons Venus 675 1201 112 0.24 0.0134 0.2241 16.8 Kling, Nyx Mons Venus 1457 2066 323 1.48 0.1795 1.2990 7.2 Kling, Nyx Mons Venus 1215 1648 82 0.31 0.0317 0.2813 8.9 Kling, Nyx Mons Venus 1385 2154 277 1.21 0.1391 1.0661 7.7 Kling, Nyx Mons Venus 761 942 92 0.22 0.0139 0.2060 14.8 Kling, Nyx Mons Venus 1610 1986 373 1.89 0.2531 1.6335 6.5 Kling, Nyx Mons Venus 543 859 185 0.32 0.0143 0.3013 21.1

Earth Pits Cylinder Cone Infilling Magnitude Diameter Major Depth Dataset Planet Volume Volume Required of Infilling (m) Axis (m) (m) (m^3) (m^3) (m^3) Required Whitten & Martin Earth 4.4 2.8 38.7 14.2 24.5128 1.7 (2019), Iceland Whitten & Martin Earth 3.5 3.4 37.4 10.9 26.4810 2.4 (2019), Iceland Whitten & Martin Earth 3.6 3 33.9 10.2 23.7504 2.3 (2019), Iceland Whitten & Martin Earth 4.1 2.8 36.1 12.3 23.7431 1.9 (2019), Iceland Whitten & Martin Earth 3.2 2 20.1 5.4 14.7445 2.8 (2019), Iceland Whitten & Martin Earth 2.5 1.4 11.0 2.3 8.7048 3.8 (2019), Iceland Whitten & Martin Earth 2.7 1.8 15.3 3.4 11.8328 3.4 (2019), Iceland Whitten & Martin Earth 3.4 2.8 29.9 8.5 21.4340 2.5 (2019), Iceland Whitten & Martin Earth 3.1 3.5 34.1 8.8 25.2807 2.9 (2019), Iceland Whitten & Martin Earth 3.4 5.4 57.7 16.3 41.3371 2.5 (2019), Iceland Whitten & Martin Earth 3.1 4.2 40.9 10.6 30.3368 2.9 (2019), Iceland Whitten & Martin Earth 2.8 2.6 22.9 5.3 17.5343 3.3 (2019), Iceland Whitten & Martin Earth 3 2.6 24.5 6.1 18.3783 3.0 (2019), Iceland Whitten & Martin Earth 3.3 3.6 37.3 10.3 27.0585 2.6 (2019), Iceland Whitten & Martin Earth 2.6 2.4 19.6 4.2 15.3561 3.6 (2019), Iceland Whitten & Martin Earth 3.3 1.2 12.4 3.4 9.0195 2.6 (2019), Iceland

188

Whitten & Martin Earth 3.2 1.9 19.1 5.1 14.0073 2.8 (2019), Iceland Whitten & Martin Earth 5.5 2.4 41.5 19.0 22.4624 1.2 (2019), Iceland Whitten & Martin Earth 4.5 1.8 25.4 9.5 15.9043 1.7 (2019), Iceland Whitten & Martin Earth 3.8 1.6 19.1 6.0 13.0523 2.2 (2019), Iceland Whitten & Martin Earth 6.2 2.3 44.8 23.1 21.6529 0.9 (2019), Iceland Whitten & Martin Earth 6.7 2.5 52.6 29.4 23.2412 0.8 (2019), Iceland Whitten & Martin Earth 6.5 2.3 47.0 25.4 21.5265 0.8 (2019), Iceland Whitten & Martin Earth 5.6 1.9 33.4 15.6 17.8275 1.1 (2019), Iceland Whitten & Martin Earth 4.3 2.2 29.7 10.6 19.0700 1.8 (2019), Iceland Whitten & Martin Earth 5.3 1.7 28.3 12.5 15.8040 1.3 (2019), Iceland Whitten & Martin Earth 4.8 1.6 24.1 9.7 14.4765 1.5 (2019), Iceland Whitten & Martin Earth 3.3 1.3 13.5 3.7 9.7711 2.6 (2019), Iceland Whitten & Martin Earth 4.1 1.5 19.3 6.6 12.7195 1.9 (2019), Iceland Whitten & Martin Earth 8.2 4.3 110.8 75.7 35.0780 0.5 (2019), Iceland Whitten & Martin Earth 3.2 2.8 28.1 7.5 20.6424 2.8 (2019), Iceland Whitten & Martin Earth 2.5 1.7 13.4 2.8 10.5702 3.8 (2019), Iceland Whitten & Martin Earth 4.5 2.7 38.2 14.3 23.8565 1.7 (2019), Iceland Whitten & Martin Earth 4.4 1.9 26.3 9.6 16.6337 1.7 (2019), Iceland Whitten & Martin Earth 8.3 4.8 125.2 86.6 38.5913 0.4 (2019), Iceland Whitten & Martin Earth 3.6 2 22.6 6.8 15.8336 2.3 (2019), Iceland Whitten & Martin Earth 4.3 1.9 25.7 9.2 16.4695 1.8 (2019), Iceland Whitten & Martin Earth 7.4 2.9 67.4 41.6 25.8438 0.6 (2019), Iceland Whitten & Martin Earth 8.4 3.7 97.6 68.3 29.2922 0.4 (2019), Iceland Whitten & Martin Earth 13.3 6.9 288.3 319.5 -31.2329 -0.1 (2019), Iceland Whitten & Martin Earth 4.4 1.2 16.6 6.1 10.5055 1.7 (2019), Iceland Whitten & Martin Earth 2.2 1 6.9 1.3 5.6444 4.5 (2019), Iceland Whitten & Martin Earth 0.7 0.35 0.8 0.0 0.7248 16.1 (2019), Iceland Whitten & Martin Earth 4.4 2 27.6 10.1 17.5091 1.7 (2019), Iceland

189

Whitten & Martin Earth 5.7 2.3 41.2 19.6 21.6228 1.1 (2019), Iceland Whitten & Martin Earth 7.4 2.7 62.8 38.7 24.0615 0.6 (2019), Iceland Whitten & Martin Earth 3.2 3 30.2 8.0 22.1168 2.8 (2019), Iceland Whitten & Martin Earth 2 1.7 10.7 1.8 8.9012 5.0 (2019), Iceland Whitten & Martin Earth 2.3 2 14.5 2.8 11.6815 4.2 (2019), Iceland Whitten & Martin Earth 2.5 2 15.7 3.3 12.4355 3.8 (2019), Iceland Whitten & Martin Earth 2.8 1.7 15.0 3.5 11.4647 3.3 (2019), Iceland Whitten & Martin Earth 1.8 1.1 6.2 0.9 5.2873 5.7 (2019), Iceland Whitten & Martin Earth 3.2 2.3 23.1 6.2 16.9562 2.8 (2019), Iceland Whitten & Martin Earth 2 1.3 8.2 1.4 6.8068 5.0 (2019), Iceland Whitten & Martin Earth 2.7 2.3 19.5 4.4 15.1197 3.4 (2019), Iceland Whitten & Martin Earth 2.4 1.7 12.8 2.6 10.2542 4.0 (2019), Iceland Whitten & Martin Earth 3.2 2.8 28.1 7.5 20.6424 2.8 (2019), Iceland Whitten & Martin Earth 2 1.5 9.4 1.6 7.8540 5.0 (2019), Iceland Kling, Kilauea Iki Earth 3.8 4 0.82 9.8 3.1 6.6893 2.2 Kling, Kilauea Iki Earth 3.06 3.2 0.77 7.4 1.9 5.5147 2.9 Kling, Kilauea Iki Earth 2.9 2.93 0.67 6.1 1.5 4.6290 3.1 Kling, Kilauea Iki Earth 11.3 13.74 1.23 43.7 41.1 2.5471 0.1 Kling, Kilauea Iki Earth 4.13 4.2 1.01 13.1 4.5 8.5944 1.9 Kling, Kilauea Iki Earth 6.02 13.58 1.63 30.8 15.5 15.3622 1.0 Kling, Kilauea Iki Earth 6.46 8.34 1.73 35.1 18.9 16.2090 0.9

190

Appendix B Noctis Labyrinthus Pit Crater Measurements Minimum Maximum Mean Max Minor Major Pit Rim Rim Rim Max Volume Surface Surface Depth Bounding FID Latitude Longitude volume Axis Axis Group Elevation Elevation Elevation (m3) Area (m2) Area (m) Fault (km3) (m) (m) (m) (m) (m) (km2) 0 1 -13.596 -97.145 5096 6511 5795 34964500473 34.965 54036518 54.037 7720 8722 1409 y 1 1 -13.739 -97.268 6233 6564 6397 1602307485 1.602 11300254 11.300 2272 8002 344 y 2 1 -13.964 -97.352 5726 6540 6185 11390485679 11.390 35296213 35.296 4150 11929 816 y 3 1 -14.504 -97.699 5890 6694 6333 5990405135 5.990 17449489 17.449 4171 5601 810 y 4 1 -14.424 -97.645 6098 6680 6411 3235133423 3.235 12127498 12.127 3117 5583 582 y 5 1 -14.313 -97.560 6083 6702 6413 2111486812 2.111 7869838 7.870 2752 5072 615 y 6 1 -14.734 -97.865 5813 6703 6349 9563972393 9.564 28856587 28.857 5037 9430 896 n 7 1 -15.356 -98.284 6276 6720 6530 1291475687 1.291 7020532 7.021 2811 3711 447 n 8 1 -15.219 -98.190 6464 6669 6571 220715758 0.221 2470053 2.470 1683 2620 214 n 9 1 -15.109 -98.120 6133 6733 6497 1833981576 1.834 8259991 8.260 3184 3732 603 n 10 1 -15.067 -98.096 6461 6713 6594 253941079 0.254 2856509 2.857 1951 2253 258 n 11 1 -14.996 -98.051 6189 6719 6470 3190661217 3.191 14033104 14.033 3185 6784 535 n 12 1 -15.250 -98.213 6499 6690 6619 134766913 0.135 1901840 1.902 1642 2069 198 n 13 1 -14.926 -98.016 6590 6714 6654 134412762 0.134 2512679 2.513 1666 2907 125 n 14 1 -14.896 -97.994 6621 6709 6672 46974734 0.047 1139117 1.139 1219 1754 95 n 15 1 -14.819 -97.935 6609 6699 6656 37591990 0.038 920293 0.920 1268 1446 94 n 16 1 -14.795 -97.921 6584 6681 6637 75703845 0.076 1826108 1.826 1499 2003 102 n 17 1 -13.281 -97.224 6224 6641 6413 1185098516 1.185 5728744 5.729 2215 4059 419 y 18 1 -13.186 -97.233 6301 6622 6429 622156432 0.622 3984441 3.984 2036 2803 327 y 19 1 -13.228 -97.231 6422 6580 6483 177889036 0.178 1792217 1.792 1261 2513 167 y 20 1 -13.147 -97.230 6296 6586 6452 200961456 0.201 1588955 1.589 1173 1917 296 y 21 2 -13.899 -96.486 6039 6392 6210 420202601 0.420 2604405 2.604 1592 2523 306 y 22 2 -13.943 -96.493 6148 6414 6283 223541818 0.224 2382862 2.383 1614 2587 284 y 23 2 -13.992 -96.493 5962 6417 6186 1380666496 1.381 6205886 6.206 2730 3502 459 y 24 2 -14.038 -96.498 6137 6360 6230 228294791 0.228 2183749 2.184 1764 2086 226 y 25 2 -14.060 -96.484 6128 6282 6199 85048332 0.085 1221540 1.222 1358 1653 159 y 26 2 -14.116 -96.489 6031 6292 6168 476071403 0.476 4190070 4.190 1946 3635 254 y 27 2 -14.159 -96.502 6129 6294 6214 89550066 0.090 1171670 1.172 1415 1618 171 y 28 2 -14.189 -96.612 6247 6376 6306 38314788 0.038 527394 0.527 1066 1200 135 y 29 2 -14.230 -96.632 6249 6342 6292 16287285 0.016 322991 0.323 805 1047 88 y 30 2 -14.281 -96.650 6243 6322 6278 31574532 0.032 802093 0.802 1175 1373 87 y 31 2 -14.316 -96.665 6241 6410 6310 155399232 0.155 1731582 1.732 1593 2175 161 y 32 2 -14.431 -96.730 5912 6439 6194 2198223438 2.198 9590066 9.590 3029 4696 522 n 33 2 -14.510 -96.780 5977 6450 6243 1437399553 1.437 8055803 8.056 2734 4664 467 n 34 2 -14.575 -96.816 6174 6386 6287 177144976 0.177 2740683 2.741 2036 2157 222 n 35 2 -14.610 -96.847 5997 6384 6202 1071137352 1.071 6198136 6.198 2638 3523 390 n 36 2 -14.657 -96.881 6299 6421 6353 81349754 0.081 1286976 1.287 1295 1689 115 n 37 2 -14.685 -96.897 6335 6421 6376 21463053 0.021 517829 0.518 957 1147 86 n 38 2 -14.698 -96.911 6385 6431 6409 7512252 0.008 326700 0.327 824 1099 65 n 39 2 -14.716 -96.923 6422 6437 6430 466684 0.000 80110 0.080 516 648 25 n 40 2 -14.789 -96.965 6404 6433 6416 3611977 0.004 189064 0.189 600 822 26 n 41 2 -14.811 -96.971 6256 6449 6340 204748167 0.205 2338732 2.339 1717 2150 193 n

191

42 2 -14.859 -96.985 5924 6461 6214 2374869566 2.375 10111588 10.112 3360 4057 541 n 43 2 -14.912 -97.000 6323 6421 6374 32894582 0.033 763143 0.763 1197 1406 100 n 44 2 -15.183 -97.082 6305 6385 6341 41988260 0.042 1043293 1.043 1114 2067 88 n 45 2 -17.825 -97.237 5260 5761 5548 1312843165 1.313 6361811 6.362 2687 3522 504 n 46 2 -17.875 -97.208 5599 5715 5658 71702722 0.072 1245896 1.246 1155 2213 112 n 47 2 -17.957 -97.160 5499 5625 5561 90042586 0.090 1411126 1.411 1332 1955 127 n 48 2 -18.009 -97.129 5262 5581 5413 1205036744 1.205 7916146 7.916 2068 5481 332 n 49 2 -18.550 -96.738 4936 5356 5176 1223773117 1.224 8602752 8.603 2458 4916 414 n 50 2 -18.633 -96.699 5232 5263 5246 3720141 0.004 240219 0.240 741 1031 29 n 51 2 -18.671 -96.683 5201 5270 5233 19734082 0.020 521753 0.522 954 1139 74 n 52 2 -18.715 -96.660 5137 5259 5189 70809354 0.071 1045188 1.045 1269 1467 122 n 53 2 -18.747 -96.650 5155 5253 5196 47475135 0.047 963499 0.963 1234 1471 100 n 54 2 -18.767 -96.445 5028 5183 5105 210647996 0.211 4371613 4.372 1904 3651 156 n 55 2 -15.418 -97.194 5867 6392 6146 1880133377 1.880 8803784 8.804 3258 4075 526 n 56 2 -15.955 -97.691 6244 6396 6330 71022696 0.071 1051624 1.052 1359 1641 156 n 57 2 -15.988 -97.697 6238 6367 6307 62442658 0.062 1263187 1.263 1296 1902 133 n 58 2 -16.019 -97.708 6219 6354 6304 58918178 0.059 1197543 1.198 1148 2154 138 n 59 2 -16.064 -97.720 6040 6352 6236 443963518 0.444 3837726 3.838 2307 2500 319 n 60 2 -16.261 -97.864 6124 6358 6236 502359068 0.502 5658180 5.658 2206 4082 239 n 61 2 -16.329 -97.893 6125 6353 6244 180772267 0.181 2391628 2.392 1900 2092 227 n 62 2 -16.357 -97.900 6204 6326 6263 89891141 0.090 1713860 1.714 1655 1813 138 n 63 2 -16.439 -97.923 6066 6317 6210 325499022 0.325 3133827 3.134 2012 2580 254 n 64 2 -16.527 -97.950 6237 6311 6264 29281111 0.029 641783 0.642 1126 1196 77 n 65 2 -16.598 -97.960 5706 6273 6040 2124453971 2.124 10240821 10.241 2897 4867 569 n 66 2 -16.654 -97.967 6138 6281 6210 103170583 0.103 1529548 1.530 1542 1822 142 n 67 2 -16.743 -97.954 5769 6261 6021 2571137592 2.571 11362644 11.363 2701 5929 509 n 68 2 -16.849 -97.887 5240 6330 5778 14671483912 14.671 30028342 30.028 5795 6651 1082 n 69 2 -16.925 -97.859 5622 6190 5868 1883393228 1.883 6084486 6.084 2701 3110 575 n 70 2 -16.989 -97.814 5786 6168 5970 725394200 0.725 4119070 4.119 2306 2841 384 n 71 2 -17.011 -97.734 6033 6217 6139 101050502 0.101 1696296 1.696 1539 1823 188 n 72 2 -17.088 -97.720 5283 6191 5741 8906853909 8.907 21333141 21.333 4584 6654 904 n 73 2 -17.311 -97.624 5464 5905 5720 1098155016 1.098 7328367 7.328 2525 4440 454 n 74 2 -17.389 -97.567 5306 5928 5601 5009736820 5.010 17034456 17.034 4087 5561 619 n 75 2 -15.919 -97.669 6319 6394 6350 39168153 0.039 842847 0.843 1038 1863 81 n 76 3 -13.899 -96.486 4681 6574 5665 72040147745 72.040 86126502 86.127 8498 13499 1864 n 77 3 -13.943 -96.493 5276 6491 5866 21925487329 21.925 38237439 38.237 6340 7803 1222 n 78 3 -13.992 -96.493 5800 6458 6106 2765500684 2.766 9058243 9.058 3200 3796 664 n 79 3 -14.038 -96.498 4591 6482 5573 214188112445 214.188 257921350 257.921 12391 38642 1911 n 80 3 -14.060 -96.484 5835 6223 6046 848396797 0.848 5273371 5.273 2450 2994 396 n 81 3 -14.116 -96.489 5189 6197 5746 27219352410 27.219 66377738 66.378 5360 18817 1019 n 82 3 -14.159 -96.502 5654 6139 5901 1534393539 1.534 6732200 6.732 3058 3753 497 n 83 3 -14.189 -96.612 5874 6105 5990 413662368 0.414 4319774 4.320 2226 3635 251 n 84 3 -14.230 -96.632 5684 6099 5898 1862909432 1.863 10381755 10.382 2563 7081 417 n 85 3 -14.281 -96.650 5920 6051 5967 228617097 0.229 3135693 3.136 1593 3270 146 n 86 3 -14.316 -96.665 5813 6030 5934 183393628 0.183 2263235 2.263 1657 2467 234 n 87 3 -14.431 -96.730 5058 6042 5641 10977790015 10.978 28764833 28.765 4663 9111 1005 n 88 3 -14.510 -96.780 5668 6007 5819 890815698 0.891 5042696 5.043 2072 3648 343 n 89 3 -14.575 -96.816 5750 5924 5845 182920276 0.183 2495609 2.496 1527 2995 181 n 90 3 -14.610 -96.847 5732 5947 5817 314206318 0.314 2769257 2.769 1958 2382 221 n 91 3 -14.657 -96.881 5677 5826 5755 112448068 0.112 1897400 1.897 1646 1932 156 n

192

92 7 -14.685 -96.897 6611 6673 6658 46708440 0.047 3561308 3.561 1888 3142 69 y 93 7 -14.698 -96.911 6412 6566 6498 46916970 0.047 670174 0.670 993 1194 149 y 94 7 -14.716 -96.923 6250 6589 6417 469642613 0.470 3872066 3.872 2315 2751 367 y 95 7 -14.789 -96.965 6243 6675 6495 1518647944 1.519 9181380 9.181 2664 5184 439 y 96 7 -14.811 -96.971 6335 6699 6531 1572186342 1.572 10154987 10.155 2979 4925 374 y 97 7 -14.859 -96.985 6522 6614 6569 11400543 0.011 259202 0.259 751 776 85 y 98 7 -14.912 -97.000 6498 6664 6562 101285678 0.101 995407 0.995 1347 1557 172 y 99 7 -15.183 -97.082 6095 6737 6363 5276818501 5.277 15649605 15.650 2881 8822 702 y 100 5 -17.825 -97.237 5544 6278 5901 4925662280 4.926 14605352 14.605 3824 5208 741 y 101 5 -17.875 -97.208 5805 6322 5996 1939759537 1.940 6940269 6.940 2669 3541 545 y 102 5 -17.957 -97.160 5843 6191 5955 748255133 0.748 4411247 4.411 2374 2631 323 y 103 5 -18.009 -97.129 5765 6466 6000 2299265873 2.299 5345264 5.345 2578 3201 727 y 104 4 -18.550 -96.738 5960 6439 6178 1465475989 1.465 7143668 7.144 2707 3961 494 y 105 4 -18.633 -96.699 5914 6446 6167 4289223996 4.289 18139411 18.139 3311 8623 526 y 106 4 -18.671 -96.683 5813 6286 6038 2099432384 2.099 9314549 9.315 3137 4251 480 y 107 4 -18.715 -96.660 6357 6469 6394 97125363 0.097 1404277 1.404 1529 1782 110 y 108 4 -18.747 -96.650 6177 6414 6295 335932556 0.336 3540019 3.540 2083 2783 234 y 109 4 -18.767 -96.445 6227 6490 6379 508819313 0.509 5868373 5.868 2270 4443 277 y 110 4 -13.805 -96.301 6313 6522 6427 314925287 0.315 3856362 3.856 1866 3619 209 y 111 4 -13.955 -96.160 6390 6517 6447 126724810 0.127 1966663 1.967 1545 2197 130 y 112 4 -14.032 -96.134 6445 6473 6456 10993751 0.011 640401 0.640 1060 1232 27 y 113 4 -14.271 -95.988 6491 6530 6510 22214177 0.022 1120629 1.121 1282 1629 39 y 114 11 -14.654 -95.839 5583 6089 5850 1682170516 1.682 7461553 7.462 2703 3759 511 y 115 11 -14.810 -95.764 5658 6084 5851 4039092613 4.039 21710500 21.711 3092 10916 432 y 116 11 -14.975 -95.675 5066 6135 5609 36006047487 36.006 74711940 74.712 5448 18788 1105 y 117 6 -15.015 -95.628 6448 6492 6468 35321341 0.035 2001112 2.001 1425 2358 47 y 118 6 -15.104 -95.585 6405 6492 6437 183000279 0.183 3564150 3.564 2079 3167 90 y 119 6 -15.173 -95.526 6484 6527 6501 19620745 0.020 760839 0.761 1124 1329 47 y 120 6 -15.224 -95.487 6447 6558 6494 169513101 0.170 2804546 2.805 1579 3002 117 y 121 6 -15.355 -95.401 6534 6564 6548 24180918 0.024 1560427 1.560 1678 1743 37 y 122 6 -15.455 -95.312 6468 6525 6495 48921857 0.049 1722212 1.722 1380 2173 57 y 123 6 -15.526 -95.268 6500 6586 6531 106692773 0.107 2192504 2.193 1779 1887 87 y 124 6 -15.671 -95.173 6498 6653 6561 355363854 0.355 4340972 4.341 2440 2685 157 y 125 6 -15.697 -95.148 6591 6613 6602 14427214 0.014 1320239 1.320 1359 1720 22 y 126 6 -3.878 -99.165 6540 6577 6554 29654459 0.030 1640960 1.641 1631 1705 34 y 127 18 -3.860 -99.200 6452 6709 6600 137679356 0.138 1481319 1.481 1426 1770 281 y 128 18 -3.845 -99.219 6613 6766 6687 206799456 0.207 3037873 3.038 1811 2617 148 y 129 18 -3.806 -99.273 6574 7086 6790 1329155185 1.329 4594810 4.595 2099 3674 526 y 130 18 -3.741 -99.380 6910 7082 6978 139285287 0.139 1368632 1.369 1500 1694 173 y 131 18 -3.715 -99.425 6689 6813 6759 43233456 0.043 834458 0.834 1100 1416 128 y 132 18 -3.708 -99.441 6649 6744 6705 36575023 0.037 992162 0.992 1145 1441 84 y 133 18 -3.675 -99.528 6562 6790 6686 134340320 0.134 1366588 1.367 1243 1982 235 y 134 8 -12.850 -96.087 5849 6223 6031 186787090 0.187 986521 0.987 1124 1700 382 y 135 8 -12.909 -96.108 6237 6461 6331 214381290 0.214 1783839 1.784 1702 1989 229 y 136 8 -12.947 -96.124 6249 6473 6338 303113498 0.303 2433918 2.434 1738 2366 246 y 137 8 -12.990 -96.141 6315 6466 6377 142657608 0.143 1808483 1.808 1679 2070 140 y 138 8 -12.850 -95.799 6023 6492 6251 1416497447 1.416 6057615 6.058 2470 4094 494 y 139 8 -12.930 -95.755 6385 6567 6461 204752169 0.205 1977365 1.977 1408 2972 199 y 140 8 -13.018 -95.705 6146 6511 6372 450356640 0.450 3556924 3.557 1639 5024 389 y 141 8 -13.144 -95.669 6392 6433 6414 7560287 0.008 360445 0.360 909 1012 26 y

193

142 8 -13.178 -95.650 5889 6714 6353 9176503559 9.177 26176957 26.177 4116 11866 826 y 143 8 -13.117 -95.586 6444 6633 6519 121872303 0.122 1058425 1.058 1085 1737 201 y 144 8 -13.175 -95.550 6369 6751 6509 1793100146 1.793 7520346 7.520 2087 7569 386 y 145 8 -13.222 -95.526 6772 6799 6788 15920604 0.016 1400358 1.400 1267 2150 29 y 146 8 -13.312 -95.550 6750 6805 6770 33707946 0.034 921799 0.922 1218 1359 55 y 147 8 -13.345 -95.528 6634 6715 6679 45290170 0.045 1268682 1.269 1473 1597 86 y 148 8 -3.824 -95.935 6576 6646 6620 9600364 0.010 361273 0.361 944 1035 70 y 149 8 -3.972 -95.969 6721 6758 6740 2772238 0.003 161389 0.161 561 1048 37 y 150 8 -4.209 -96.033 6613 6664 6638 18681527 0.019 760480 0.760 1146 1295 51 y 151 8 -4.662 -99.032 6664 6700 6683 4743363 0.005 241704 0.242 711 844 38 y 152 9 -4.736 -99.034 5878 6193 6043 421432048 0.421 4012216 4.012 2222 3277 324 y 153 9 -4.867 -99.052 5721 6404 6021 22159586053 22.160 62427569 62.428 3555 26749 692 y 154 9 -4.902 -99.067 6284 6349 6327 14253874 0.014 641882 0.642 1098 1216 70 y 155 10 -4.942 -99.079 6134 6273 6199 227761978 0.228 3527584 3.528 1969 3392 140 y 156 10 -4.971 -99.098 6147 6222 6196 31641201 0.032 1163523 1.164 1297 1662 78 y 157 10 -5.016 -99.106 6167 6301 6222 74916177 0.075 1004642 1.005 1153 1721 137 y 158 10 -5.053 -99.104 5999 6324 6191 807030632 0.807 6467787 6.468 2994 3852 322 y 159 15 -4.557 -98.993 5808 6800 6257 5795515624 5.796 12025960 12.026 3427 4755 976 y 160 15 -4.484 -98.966 6370 6860 6611 4206020990 4.206 16937375 16.937 4370 5472 478 y 161 15 -5.089 -100.763 6265 6535 6388 439590900 0.440 5830058 5.830 2722 3144 272 y 162 15 -5.073 -100.736 6261 6883 6560 7439203894 7.439 23285438 23.285 5128 6153 628 y 163 15 -5.044 -100.703 6176 6634 6391 2160008714 2.160 9957088 9.957 3395 4111 464 y 164 15 -5.002 -100.653 6109 6691 6397 4189796766 4.190 15482988 15.483 4236 5434 580 y 165 15 -4.889 -100.544 6538 6760 6611 796530234 0.797 6160190 6.160 2451 4093 225 y 166 15 -4.881 -100.526 6559 6628 6587 31560996 0.032 781806 0.782 1097 1172 59 y 167 15 -4.863 -100.510 5669 6834 6184 18038531351 18.039 29449172 29.449 5777 6914 1167 y 168 16 -2.877 -100.133 6304 6429 6366 74734165 0.075 1357104 1.357 1362 1641 121 y 169 16 -2.916 -100.084 6647 6789 6687 202692286 0.203 2302425 2.302 1828 2113 161 y 170 16 -2.941 -100.061 6661 6786 6723 100011823 0.100 1710680 1.711 1619 1944 141 y 171 16 -2.980 -100.008 6201 6976 6643 13166759768 13.167 42305989 42.306 4284 13250 770 y 172 16 -3.022 -99.947 5651 6647 6010 20467060137 20.467 34440074 34.440 5894 8080 988 y 173 16 -3.060 -99.906 5849 6135 6024 95339069 0.095 825815 0.826 1107 1329 237 y 174 17 -3.109 -99.843 5047 6843 5871 104989798407 104.990 116829689 116.830 11271 12856 1809 y 175 12 -3.176 -99.681 6298 6445 6362 338912864 0.339 4148277 4.148 2386 2680 159 y 176 12 -3.254 -99.593 6303 6564 6408 1096009204 1.096 7740141 7.740 3034 4160 265 y 177 12 -3.319 -99.516 6608 6971 6748 1724972141 1.725 7938956 7.939 3099 3761 354 y 178 12 -3.364 -99.480 6232 7091 6590 14399733234 14.400 31276001 31.276 5843 6950 855 y 179 12 -3.462 -99.403 6405 6660 6552 423662748 0.424 4185227 4.185 2455 2532 260 y 180 12 -3.495 -99.433 6194 6464 6304 719187298 0.719 5164759 5.165 2603 2957 277 y 181 14 -3.539 -99.383 6297 6771 6494 3507526468 3.508 13867837 13.868 2065 12209 476 y 182 14 -3.549 -99.364 6724 6811 6760 117091111 0.117 2563867 2.564 1907 2127 89 y 183 14 -3.527 -99.401 6720 6773 6746 10953749 0.011 361192 0.361 717 1499 52 y 184 14 -3.440 -99.508 6706 6794 6746 56102129 0.056 1204715 1.205 1299 1731 93 y 185 14 -3.432 -99.522 6526 6834 6664 301662629 0.302 2333255 2.333 1567 2390 325 y 186 14 -3.431 -98.360 6650 6799 6733 72624024 0.073 1093842 1.094 920 2480 141 y 187 14 -3.473 -98.620 6408 6718 6548 410835594 0.411 2541513 2.542 1874 2237 327 y 188 14 -3.517 -98.860 6489 6646 6581 29701127 0.030 448146 0.448 882 1098 162 y 189 14 -3.311 -98.835 6485 6539 6520 3166787 0.003 161060 0.161 625 766 52 y 190 14 -3.373 -98.853 5517 6500 6014 6157925709 6.158 13878017 13.878 4036 4539 974 y 191 14 -3.402 -98.865 6230 6440 6362 139765305 0.140 2694876 2.695 1873 2308 205 y

194

192 14 -3.455 -98.877 6214 6540 6368 291752984 0.292 2400408 2.400 1827 2008 330 y 193 14 -5.166 -104.568 5966 6604 6310 2509277998 2.509 9354077 9.354 2899 4895 679 y 194 14 -5.210 -104.521 6745 6771 6759 4309207 0.004 230042 0.230 676 786 22 y 195 14 -5.235 -104.464 6630 6735 6692 1586727 0.002 40652 0.041 407 731 79 y 196 14 -5.288 -104.396 6714 6751 6734 9548113 0.010 730708 0.731 944 1674 50 y 197 14 -5.348 -104.301 6231 6765 6488 3643483479 3.643 14070993 14.071 2508 12204 535 y 198 14 -5.376 -104.234 6376 6534 6444 161681859 0.162 2086070 2.086 1700 2069 174 y 199 14 -5.427 -104.158 6151 6692 6361 1204129071 1.204 3839943 3.840 2108 2977 553 y 200 14 -5.316 -104.344 5817 6217 5984 788948449 0.789 3748246 3.748 2349 2978 410 y 201 14 -5.613 -103.861 6619 6699 6654 16960644 0.017 401969 0.402 881 1091 82 y 202 14 -5.664 -103.785 6465 6646 6537 94614530 0.095 885962 0.886 826 2945 191 y 203 14 -5.667 -103.652 6153 6461 6315 76221995 0.076 538761 0.539 1053 1335 339 y 204 14 -5.652 -103.628 6575 6735 6654 302716404 0.303 4524859 4.525 1872 5840 166 y 205 13 -5.603 -103.511 6030 6951 6416 10042327653 10.042 20441767 20.442 4969 5136 945 y 206 13 -5.716 -103.745 6021 6899 6424 7917133026 7.917 17588153 17.588 4477 5205 887 y 207 13 -5.785 -103.769 5470 6824 6066 59144947239 59.145 85178722 85.179 6347 17672 1339 y 208 13 -5.217 -103.840 5196 6363 5684 16910929039 16.911 27152782 27.153 5426 6274 1111 y 209 13 -5.562 -106.125 5866 6982 6311 9736428077 9.736 15909443 15.909 4126 4996 1122 y 210 13 -5.545 -106.205 5363 6297 5884 12235150531 12.235 30821623 30.822 6245 6722 945 y 211 1 -5.515 -106.304 6604 6714 6676 40672078 0.041 1174284 1.174 1270 1588 113 y 212 1 -5.449 -106.388 6460 6697 6595 219205280 0.219 2694395 2.694 1908 2317 240 y 213 1 -5.396 -106.579 6490 6649 6596 63575746 0.064 1412768 1.413 1546 1553 174 y 214 1 -5.415 -106.527 6571 6619 6597 12260465 0.012 520882 0.521 1035 1259 62 y 215 1 -4.733 -106.345 6645 6707 6679 29854467 0.030 1081931 1.082 1273 1544 63 n