WILLIAMS COLLEGE LiBRARIES COPYRICJI IT ASSIGNMENT AND INSTRUCTIONS FOR A STUDI~NT TIIESIS Your unpublished thesis, submitted for a degree at Williams College and administered by the Williams College Libraries, will be made available for research use. You may, through this form, provide instructions regarding copyright, access, dissemination and reproduction of your thesis. The College has the right in all cases to maintain and preserve theses both in hardcopy and electronic format, and to make such copies as the Libraries require for their research and archival functions. ~Thefaculty advisor/s to the student writing the thesis claims joint authorship in this work. __ 1/we have included in this thesis copyrighted material for which 1/we have not received permission from the copyright holder/s. If' you do not secure copyright permissions by the time your thesis is submitted, you will still be allowcclto submit. However, if' the necessary copyright permissions arc not received, c-posting of' your thesis may be a fleeted. Copyrighted material may include images (tables, drawings, photographs, figures, maps, graphs, etc.), sound files, video material, data sets, and large portions of text. l. COPYRIC!IIT An author by law owns the copyright to his/her work, whether or not a copyright symbol and elate arc placed on the piece. Please choose one of' the options below with respect to the copyright in your thesis. ~llwe choose not to retain the copyright to the thesis, and hereby assign the copyright to Williams College. Selecting this option will assign copyright to the College. II' the author/swishes later to publish the work, he/she/they will need to obtain permission to do so fi·om the Libraries, which will be granted except in unusual circumstances. The Libraries will be f'rcc in this case to also grant permission to another researcher to publish some or all of' the thesis. If you have chosen this option, you do not need to complete the next section and can proceed to the signature line. __ 1/we choose to retain the copyright to the thesis for a period of ___ years, or until my/our death/s, whichever is the earlier, at which time the copyright shall be assigned to Williams College without need of further action by me/us or by my/our heirs, successors, or representatives of my/our estate/s. Selecting this option allows the author/s the flexibility ol'retaining his/her/their copyright l'or a period of'ycars or f(l1' life. Signatures Redacted USING GEOPHYSICAL TECHNIQUES IN THE CRITICAL ZONE TO DETERMINE THE PRESENCE OF PERMAFROST By Gabriel M. Lewis Professor David P. Dethier, Advisor A thesis Submitted in partial fulfillment of the requirements for the Degree of Bachelor of Arts With Honors in Geosciences Williams College Williamstown, Massachusetts May 18, 2013 ABSTRACT th th Global warming during the 20and 21centuries has increased air temperatures in alpine areas of the southern Rocky Mountains sufficiently to melt large areas of previously frozen ground, referred to as permafrost. Previous studies used geomorphological, hydrological, and GIS techniques to infer the distribution of frozen ground and ice lenses on Niwot Ridge and in adjacent Green Lakes Valley, Colorado Front Range. Predictions of permafrost occurrence have not previously been verified in the field by subsurface geophysical measurements, although permafrost, ice lenses, and temperature profiles beneath active gelifluction lobes were documented in several studies during the 1970s along Niwot Ridge. Electric Resistivity Tomography (ERT) and Ground Penetrating Radar (GPR) are geophysical techniques that have been utilized worldwide to study the evolution of alpine permafrost and ice lenses. Combining these geophysical methods maximizes the accuracy of each method while reducing their inherent ambiguities and limitations. Green Lakes and 4th of July Valleys offer ideal locations to verify the existence of ice masses within rock glaciers, where models predict they exist. This study reports: (1) interpreted results of 16 ERT and 2 GPR lines totaling 815 m that were collected to test permafrost predictions in alpine zones; (2) soil temperature profiles and morphology in several pits excavated to saprolite along ERT lines; (3) energy modeling of water temperatures for Como Creek, a small alpine creek on Niwot Ridge, compared with measured temperatures for nearby Martinelli Stream; and (4) computer modeling of subsurface temperatures from surface temperatures on Niwot Ridge, Colorado, according to a model calibrated using data from Hopkins Memorial Forest, Williamstown, MA. Analysis of seven ERT lines from elevations of 3500 to 3900 m on Niwot Ridge demonstrates that my study area lacks permanent ice lenses (resistivity of approximately 200-1000 kOm) beneath a surface layer of coarse, blocky debris (resistivity of ~ 20 kOm). Gelifluction lobes, as well as nearby snow field areas, may contain seasonal ice lenses that are misinterpreted as permafrost features, but ice often melts completely by late summer. Soil and water temperatures reveal that the subsurface is too warm to permit the development of permafrost, and heat-flow models confirm this hypothesis. Inactive periglacial deposits within the Boulder Creek Watershed support evidence of a climate through the late Pleistocene that produced and supported permafrost and permanent ice lenses. During the last glacial maximum, temperature and precipitation values supported gelifluction and permafrost at elevations as low as Gordon Gulch. Better understanding of the present distribution of permafrost and active periglacial features helps predict changes to alpine landscapes as permafrost disappears and has implications for quantity of runoff in the near future. ACKNOWLEDGEMENTS I appreciate the guidance, patience, and helpful counsel my advisor Dr. David P. Dethier provided over the past eight months, offing countless hours of support and advice which surpassed all expectations I had for a thesis advisor. Thank you for continuing to challenge, motivate, and stimulate my constant wonder and amazement throughout the research process. My gratitude extends to Dr. Matthias Leopold for his assistance in the field and technical advice during data processing throughout this school year. His assistance processing my geophysical data via long distance phone conversations from Australia was invaluable. I would like to thank Bud Wobus for all his assistance proofreading my paper and providing valuable comments and suggestions for improving my writing as a scientist. Dr. Will Ouimet’s guidance and patience both in the field and while writing my thesis provided an example of the ideal role model I strive to become. I am also grateful for Ian Nesbitt’s instrumental work assisting me in the field last summer. I would also like to thank the Williams College Geosciences Department for all the support and assistance during my four years in college. Next, I would like to thank the KECK Geology Consortium for providing an invaluable resource for research and education to liberal arts institutions nationwide. The research was partly funded through the Boulder Creek Critical Zone Observatory sponsored by the US National Science Foundation (NSF-0724960), by the NSF-supported Niwot Ridge LTER project (DEB­0423662), and the University of Colorado Mountain Research Station. Lastly, I would like to thank my friends and family for supporting me throughout my time at Williams. The essential role you have all played in shaping my life as a student, scientist, and citizen are irreplaceable. I am eternally grateful for your stability, compassion, leadership, and love. Thank you. TABLE OF CONTENTS ABSTRACT........................................................................................................................ i ACKNOWLEDGEMENTS ............................................................................................ iii LIST OF FIGURES ....................................................................................................... viii LIST OF TABLES ......................................................................................................... xvi LIST OF EQUATIONS................................................................................................ xvii LIST OF APPENDICES ............................................................................................. xviii INTRODUCTION..............................................................................................................1 Introduction .........................................................................................................1 The Critical Zone.................................................................................................2 Electrical Resistivity Tomography.....................................................................6 History of ERT............................................................................................... 6 ERT Background ........................................................................................... 7 ERT Physics................................................................................................... 8 Depth of Investigation ................................................................................. 13 Heat Flow ...........................................................................................................13 Climate................................................................................................................28 Previous Geophysical Studies on Niwot Ridge ...............................................29 Periglacial Processes and Deposits...................................................................30 Solifluction and Gelifluction ....................................................................... 31 Paleo-Periglacial Conditions...................................................................... 41 Sorted Circles.............................................................................................. 42 Rock Glaciers .............................................................................................. 45 SETTING..........................................................................................................................49 Location and Topography ................................................................................49 Niwot Ridge ................................................................................................. 52 Green Lakes Valley ..................................................................................... 55 4th of July Valley.......................................................................................... 56 Gordon Gulch.............................................................................................. 58 Geologic Background ........................................................................................60 Bedrock........................................................................................................ 60 Surficial Deposits ........................................................................................ 63 Climate................................................................................................................64 Historic Climate .......................................................................................... 64 Paleoclimate................................................................................................ 70 METHODS .......................................................................................................................78 Field ....................................................................................................................78 ERT.............................................................................................................. 78 GPR ............................................................................................................. 81 Temperature Measuring ...................................................................................82 Temperature Modeling .....................................................................................83 Laboratory .........................................................................................................86 ERT.............................................................................................................. 86 GPR ............................................................................................................. 88 RESULTS .........................................................................................................................90 Surface Deposits.................................................................................................90 Temperature modeling......................................................................................91 Modeling Hopkins Memorial Forest Subsurface Temperatures ................. 93 Soil Pit Temperature Measurements on Niwot Ridge, CO.......................... 96 Water Temperatures Map.......................................................................... 101 HFLUX Modeling............................................................................................102 ERT...................................................................................................................107 GPR...................................................................................................................115 DISCUSSION .................................................................................................................118 Surface Deposits...............................................................................................118 Temperature Modeling ...................................................................................119 HFLUX Modeling............................................................................................123 ERT...................................................................................................................125 GPR...................................................................................................................128 Paleoclimatic Conditions.................................................................................128 Future ...............................................................................................................130 CONCLUSIONS ............................................................................................................131 REFERENCES...............................................................................................................132 APPENDICES................................................................................................................141 Appendix A.......................................................................................................141 Basic Physics Review ................................................................................ 141 Point Current Source ................................................................................ 142 Two Current Electrodes ............................................................................ 143 Two Potential Electrodes .......................................................................... 145 A Single Horizontal Interface.................................................................... 146 Appendix B.......................................................................................................148 Appendix C.......................................................................................................153 Appendix D.......................................................................................................169 Appendix E.......................................................................................................170 LIST OF FIGURES Figure 1. Cross-section of the CZ showing bedrock and regolith layers along with associated geomorphic processes for each. The rock parcel trajectory (black dashed line), and water flow paths (blue dashed line) are presented in a system with minimal solid material added at the surface (image from Anderson et al., 2007, Figure 1, p.2)............... 3 Figure 2. Wenner configuration showing two current electrodes placed outside of the two potential electrodes (image modified from Burger et al., 2006, Figure 5.9, p. 278) .......... 9 Figure 3. The dipole-dipole configuration showing two potential electrodes (left) and two current electrodes (right) spaced a distance “a” from each other. Lines of constant voltage are shown (dotted) in relation to electric field lines (solid).............................................. 10 Figure 4. Electrode geometries used in this resistivity surveying study. a. Wenner. b. Wenner/Lee. c. Schlumberger. d. Dipole-dipole (image modified from Burger et al., 2006, Figure 5.9, p. 278)................................................................................................... 11 Figure 5. Resistivity values used for interpretation of resistivity-depth profiles generated in RES2DINV (image from Leopold et at., 2013, p. 13).................................................. 12 Figure 6. Temperature-depth profiles from 14 drill wells throughout Cape Thompson, Barrow, and Prudhoe Bay, Alaska. All of these profiles show a significant departure from the expected linear profile within the upper 100-200 m (image from Lauchenbruch, 1986, Figure 6, p. 234)................................................................................................................ 15 Figure 7. Temperature depth profile of a borehole at East Teshekpuk Lake, Alaska over the last three decades as part of the Global Terrestrial Network-Permafrost deep borehole array (image modified from Anderson and Anderson, 2010, Figure 9.9, p. 280) ............ 16 Figure 8. Permafrost cross-section showing continuous, discontinuous, and sporadic zones of permafrost overlain by the active layer (image from Bierman and Montgomery, 2013, Figure 9.12, p. 31)................................................................................................... 18 Figure 9. Temperature-depth profile showing variations of the near-surface region for a MAT of -10 °C and temperature fluctuation of 15 °C (image from Anderson and Anderson, 2010, Figure 9.7, p. 278) ................................................................................. 20 Figure 10. Soil temperatures between 9/5/1996 and 7/2/1997 at the Tundra Research Station, Niwot Ridge, taken at a depth of 30 cm showing thick snow cover (top) and thin snow cover (bottom) (images from Hamann, 1998, Figures 5.2 and 5.4, pp. 42 and 44) 24 Figure 11. A model for permafrost in the Colorado Front Range displaying a 50-75% probability of permafrost occurring (light purple) and 76-100% chance of permafrost existing (dark purple) (image from Janke, 2005, Figure 5, p. 383) .................................. 25 Figure 12. The center and terminus of the Arapaho rock glacier showing transverse ridges and furrows (image from White, 1971, Figure 4, p. 48) ........................................ 27 Figure 13. Schematic diagram of a N-S cross-section at Niwot Ridge portraying the composition of the CZ and lateral thickness variation of the periglacial slope deposits. Vertical exaggeration of the relief is approximately 4 times (image from Leopold et al., 2008, Figure 6, p. 89)........................................................................................................ 31 Figure 14. ERT Line 7 on Lobe 45 showing saturation from water supported by layer of silt within the lobe. During the winter, ice lenses accumulate below the surface and cause the lobe to move downhill (away from the camera). ........................................................ 33 Figure 15. Oblique view of an active gelifluction lobe on Niwot Ridge showing how the bulging, overhanging lobe front contrasts with the subdued terrace on which it has developed. Flowing water was present at the surface of the lobe throughout the summer when the picture was taken in 1962 (image from Benedict, 1970, Figure 14, p. 177) .... 35 Figure 16. The gelifluction process depicted on a slope showing the path of a particle experiencing frost heave (image modified from Washburn, 1970, p. 126) ...................... 36 Figure 17. a. Vertical profile of a lobe on Niwot Ridge showing how velocity in the subsurface varies with depth. b. Directions of soil movement as inferred from the tilting of stakes and orientation of plant roots. Arrow lengths are proportional to velocities (images from Benedict, 1970, Figures 20 and 19, pp. 182 and 181)................................ 38 Figure 18. Soil profile of turf-banked terrace lobe 499 on Niwot Ridge. 1 – Stones, drawn to scale. 2 – Dark grayish brown gravelly sandy loam. 3 – Dark grayish brown to yellowish brown sandy loam, becoming finer toward the front of the lobe. 4 – Rose diagrams summarizing the long-axis orientations of 200 sand grains, measured in thin section representing frequencies of 10% (image from Benedict, 1970, Figure 21, p. 183) ........................................................................................................................................... 39 Figure 19. Aerial view of Niwot Ridge showing location of turf-banked gelifluction lobes that propagated downslope (image from Benedict, 1970, Figure 2, p. 168) ........... 41 Figure 20. a. Photograph of the Murtél rock glacier indicating position of borehole (dot) and ERT monitoring line (bold). b. Inverted image of ERT measurement from August 2006 showing active layer (red and orange) overlying an ice block (blue) (image from Hilbich et al., 2009, Figure 1, p. 272)............................................................................... 42 Figure 21. Patterns of inferred motion within the fine centers and coarse perimeters most compatible with sorted circle formation (image modified from Anderson and Anderson, 2010, Figure 9.16, p. 286)................................................................................................. 44 Figure 22. a. N. Caine standing in the center of a sorted circle in Green Lakes Valley (image from D. Dethier, personal communication, January 9, 2013). b. Sorted circles 1-2 m in diameter which developed in raised beaches of Svalbard. (Photo by Anderson and Anderson, 2010, Figure 9.15, p. 285) ............................................................................... 45 Figure 23. Conceptual model of the Green Lakes 5 rock glacier based on water sources and flow paths (image from Leopold et al., 2011, Figure 1, p. 2) .................................... 48 Figure 24. a. LiDAR image of the Boulder Creek CZO showing the location of 4th of July Valley, Green Lakes Valley, Niwot Ridge, and Gordon Gulch. The outline of Gordon Gulch is shown in blue, the sixteen Electrical Resistivity Tomography lines are shown in red, and the background is a LiDAR hillshade from Anderson et al., 2012. Red outlines are the insets for Niwot Ridge ERT lines (Fig. 24b), Gordon Gulch ERT lines (Fig. 24c), and 4th of July ERT lines (Fig. 24d) ............................................................... 50 Figure 25. Oblique aerial view showing relative locations of the Saddle on Niwot Ridge, Green Lakes Valley and 4th of July Valley (image from Google Earth©) ........................ 53 Figure 26. Photographic overview looking south across the Fahey site in May 2010. Indicated are the outer boundary of the lobe (white dotted line), location of ERT lines I­III from Leopold et al. (2013), and the location of two boreholes equipped with temperature loggers (image from Leopold et al., 2013, Figure 3, p. 6)............................ 54 Figure 27. Late winter view of the Fourth of July Valley looking eastward towards four lobate rock glaciers (image from Benedict, 1981, Figure 77, p. 95) ................................ 58 Figure 28. a. A north facing slope in Gordon Gulch covered in lodgepole pine showing ERT line 15. b. Cross section of north facing slope showing an interpretation of mobile regolith, saprolite, weathered bedrock and fresh bedrock from shallow seismic reflection data (image from Befus, 2011, Figure 1, p. 916).............................................................. 59 Figure 29. Digital combination of the Denver West 30’x60’ Geologic Quadrangle (Kellogg et al, 2008) and the Estes Park 30’x60’ Geologic Quadrangle, North Central Colorado (Cole and Braddock, 2009) showing the location of 4th of July Valley, Niwot Ridge (N.R), Gordon Gulch (G.G.) and Boulder, CO for reference. See text for geologic descriptions ....................................................................................................................... 61 Figure 30. Ward 7.5 minute Quadrangle showing the proportion of quartz-bearing syenite (blues) and Pinedale glacial till (speckled grays) compared with other rocks (image from Gable and Madole, 1976)............................................................................. 62 Figure 31. Vegetation zones and climatic data for a transect of the Front Range from the Continental Divide to the Colorado Piedmont near Boulder, CO, showing the areas investigated in this study. Elevation ranges are typical for the vegetation zones shown, where gaps between elevation ranges represent ecotones between the zones (image modified from Birkeland et al., 2003, Figure 5, p. 85)..................................................... 65 Figure 32. Map of Niwot Ridge and Green Lakes Valley showing the locations of the D­1 Meteorological Station, Saddle Research Station, and the Albion Meteorological Station ............................................................................................................................... 66 Figure 33. Timeline of the annual sum of positive and negative degree days, calculated by summing the mean of the daily minimum and maximum ± 3.7 °C for each day of the year. A positive trend is seen for positive degree days with R2 value of 0.35, but there is no obvious trend for the negative degree days (from Leopold et al., 2013, Figure 2, p. 5) ........................................................................................................................................... 67 Figure 34. Timeline of the recorded (solid line) and reconstructed (dashed) mean annual precipitation at D1 station from 1965 to 2010. Recorded data obtained from culter.colorado.edu. Estimated precipitation from Greenland, 1989 ................................ 68 Figure 35. Sample sites and 10Be exposure ages shown on 1 meter DEM for upper GLV. The glacial limit from the LGM is shown in light blue in the North Boulder Creek drainage. Insert shows high resolution topography emphasizing the distinct boundaries of lateral moraines (image from Dühnforth and Anderson, 2011, Figure 2, p. 528) ............ 72 Figure 36. Plot comparing June insolation at the top of the atmosphere at 30 °N and 60 °N to glacial events in the Colorado Front Range (data from Benedict, 1970 and 1981). Also shown are the glacial periods throughout the past 30ka (image from Muhs and Benedict, 2006, Figure 10, p. 127).................................................................................... 76 Figure 37. a. I. Nesbitt and M. Leopold collecting data at Electrical Resistivity Line 5 in front of the electrodes, stakes, and connecting ribbon cable. b. Data collection process from field laptop (black) and 4punktlight hp (orange box) .............................................. 79 Figure 38. M. Leopold squatting in the excavated pit near ERT Line 11 as he measures the resistivity of openwork gravel 150 cm below the surface .......................................... 81 Figure 39. Map of GPR Lines 1 and 2 on Niwot Ridge showing their location in relation to the inactive gelifluction lobe and Tundra Research Station ......................................... 82 Figure 40. a. G. Lewis measuring the GPS location of a water temperature data point. b. Thermometer (red arrow) measuring temperature of cold-water spring at GPS point 150 ........................................................................................................................................... 83 Figure 41. a. Incoming summer solar radiation computed in ArcGIS© between March 20 and September 22 for an average year. b. Incoming winter solar radiation computed in ArcGIS© between September 22 and March 20 for an average year................................ 85 Figure 42. Surface deposits map of Green Lakes Valley and Niwot Ridge showing periglacial terrace and lobe crests, stone banked terrace deposits, talus and glacial deposits, periglacial deposits, diamict, and gelifluction lobes all above granitic bedrock91 Figure 43. Modeled temperature variations in the near-surface region for a MAT of -3.7 °C and ideal temperature oscillations with an amplitude of 15 °C. This plot shows temperatures over a two year time period at one meter intervals from 0-7 m.................. 92 Figure 44. a. Google Earth© map showing the location of Hopkins Memorial Forest (black) in relation to Williamstown, MA. b. Oblique view showing the location of the Taconic Crest in relation to Hopkins Memorial Forest (black) from River Tools© ......... 94 Figure 45. Daily temperature measurements in Hopkins Memorial Forest, MA, from Jan 1, 2012 through Dec 31, 2012 for the surface air temperature, soil temperature at depths of 10 cm and 80 cm, and water temperature at 5.15 m depth ........................................... 95 Figure 46. Best fit sinusoidal functions fit to daily temperature measurements in Hopkins Memorial Forest, MA for the surface air temperature, soil temperature at depths of 10 cm and 80 cm, and water temperature at 5.15 m depth .......................................................... 95 Figure 47. Average of observed daily maximum and minimum temperature at D1 station between 1952 and 2010 in blue. Modeled temperature based on MAT of -3.7 °C and amplitude of oscillation of 15 °C in red (data from culter.colorado.edu)......................... 97 Figure 48. a. Temperature measurements from the 150 cm deep pit excavated near ERT Line 11. b. Temperature measurements fit to model to compute the parameters Q/k and .. ........................................................................................................................................... 98 Figure 49. Annual temperature-depth profile near ERT Line 11 interpolated from temperature measurements (Eqn. 4). Black dots are temperature measurements from the soil pit on July 24 and the red line is their calculated temperature profile. Thin black lines are calculated from running this model at 10 day intervals to create an outline for the daily temperature profile to oscillate through............................................................. 99 Figure 50. Calculated annual temperature-depth profile in Hopkins Memorial Forest interpolated from temperature measurements (Eqn. 4). Red lines indicate temperature measurements at depths of 10, 80, and 515 cm. Thin black lines calculated from running this model at 10 day intervals to create an outline for the daily temperature profile to oscillate through.............................................................................................................. 100 Figure 51. Measured annual temperature-depth profile in Hopkins Memorial Forest interpolated from temperature measurements at depths of 10, 80, and 515 cm from Eqn. 9, Eqn. 10, Eqn. 11, and Eqn. 12 at 10 day intervals...................................................... 100 Figure 52. Map of surface water temperatures on Niwot Ridge plotted on a LiDAR background from Anderson et al., 2012.......................................................................... 101 Figure 53. Map of surface water temperature measurements in relation to the periglacial terrace and lobe crests on Niwot Ridge. Also shown are the winter snow drift heights computed by D. Dethier (unpublished data, 2013)......................................................... 102 Figure 54. a. Oblique view of a distance vs. temperature vs. time surface representation of Como Creek as modeled by HFLUX. b. Two dimensional temperature vs. time profile of Como Creek over the course of 24 hours ................................................................... 104 Figure 55. Temperature measurements of Martinelli Stream (black) from midnight August 4 to noon August 5 showing time of temperature modeling (red) ..................... 105 Figure 56. Different energy fluxes taken into consideration in the HFLUX modeling program showing solar (shortwave) radiation, longwave radiation, latent heat flux, streambed conduction, sensible heat flux, and the total heat flux in the system over a 24 hour time period.............................................................................................................. 106 Figure 57: Temperature-depth profile (red) near the head of Como Creek at 6:00 am on August 5, 2013 showing that the temperature below 145 cm remains below 4 °C ........ 107 Figure 58. Map showing the location of ERT Lines 3, 4, and 10 as well as GPR Line 1. ERT Line 15 is located in Gordon Gulch (Fig. 24c). ..................................................... 108 Figure 59. a. Profile of ERT Line 3 on the north slope of Green Lakes Valley showing a low resistivity upper 2-3 m with high resistivity (150 kO) layer below. b. Interpretation of ERT Line 3 showing scree and surface boulders, a seasonally frozen layer, and underlying frozen ice within bedrock ............................................................................. 109 Figure 60. Oblique view of M. Leopold recording data from ERT Line 3 (black) on a steep scree face in upper Green Lakes Valley ................................................................ 110 Figure 61. a. Profile of ERT Line 4 adjacent to the GL5 rock glacier showing a low resistivity upper 2-3 m with high resistivity (150 kO) layer below. b. Interpretation of ERT Line 3 showing scree and surface boulders, a seasonally frozen layer, and underlying frozen bedrock containing low to high ice concentrations ........................... 111 Figure 62. Oblique view of ERT Line 4 (black) showing flowing surface water, still surface water, and areas without any surface water. RG5 is just to the right of this image ......................................................................................................................................... 111 Figure 63. a. Profile of ERT Line 10 on Niwot Ridge showing surface vegetation, unsorted Quaternary diamict, and a layer of open work gravel below 140 cm. b. Interpretation of ERT Line 10 showing organic A Horizon, B Horizon, C Horizons I and II; inset shows expanded view of east end of line .......................................................... 112 Figure 64. Oblique view of ERT Line 10 (black) showing surface boulders and tundra vegetation above the head of Martinelli and Saddle Stream........................................... 113 Figure 65. a. Profile of ERT Line 15 in Gordon Gulch showing areas of high resistivity (>2500 Om) and low resistivity (<50 Om). Black lines represent soil pits along the line. b. Interpretation of ERT Line 15 showing surface boulders, toeslope deposits, saprolite, weathered bedrock, and bedrock..................................................................................... 114 Figure 66. Oblique view of ERT Line 15 (black) in Gordon Gulch showing a dense lodgepole pine forest with a thick mat of organic material on the forest floor............... 114 Figure 67. M. Leopold collecting GPR Line 1 across the front of the stone-banked gelifluction lobe. Image taken looking north.................................................................. 115 Figure 68. a. Cross section of gelifluction lobe along GPR Line 1 (after removing ground and air reflections) showing parallel and subparallel subsurface layers. b. Interpretation of GPR Line 1 showing the modern lobe surface, vegetated layer, inactive gelifluction lobe surface, reflections from underlying boulders, and buried older gelifluction lobes ....... 116 Figure 69. Cross section along the top of the gelifluction following GPR Line 2 (after removing ground and air reflections) showing parallel and subparallel subsurface layers ......................................................................................................................................... 117 Figure 70. 3D view of GPR Lines 1 (right) and 2 (left) intersecting in a Google SketchUp© environment showing the distance along the top from the beginning of the line and time in nanoseconds for the signal to reflect back to the probes along the vertical axis .................................................................................................................................. 117 Figure 71. Depth of Investigation indices showing the accuracy of ERT inversion results for a. ERT Line 3. b. ERT Line 4. c. ERT Line 10. d. ERT Line 15.............................. 126 Figure 72. An illustration of the configuration used to determine the potential at P1 for a single point source of current C1. Equipotential surfaces are shown; the outer two are separated by a distance dr ............................................................................................... 143 Figure 73. Diagram illustrating the configuration used to determine potential at P1 given current source C1 and current sink C2 (image modified from Burger et al., 2006, Figure 5.5, p. 272) ...................................................................................................................... 144 © th th Figure 74. RES2DINVinversion image showing a. 4of July 1-2 Wenner array. b. 4of July 1-2 dipole-dipole array........................................................................................ 153 Figure 75. RES2DINV© inversion image showing a. 4th of July 3 Wenner-half array. b. 4th of July 3 Wenner array............................................................................................... 154 Figure 76. RES2DINV© inversion image showing a. 4th of July 4 Wenner-half array. b. th thth 4of July 4 dipole-dipole-half. C. 4of July 4 Wenner array. d. 4of July 4 dipole-dipole array ..................................................................................................................... 155 Figure 77. RES2DINV© inversion image showing Gordon Gulch 1 Wenner array...... 156 Figure 78. RES2DINV© inversion image showing a. Gordon Gulch 2 Wenner array. b. Gordon Gulch 2 dipole-dipole array............................................................................... 157 Figure 79. RES2DINV© inversion image showing a. Green Lakes Front RG5 Wenner array. b. Green Lakes Front RG5 dipole-dipole array .................................................... 158 Figure 80. RES2DINV© inversion image showing a. Green Lakes Moraine Wenner-half array. b. Green Lake Moraine Schlumberger array ........................................................ 159 Figure 81. RES2DINV© inversion image showing a. Green Lakes North Slope Schlumberger array. b. Green Lakes North Slope dipole-dipole array........................... 160 Figure 82. RES2DINV© inversion image showing a Green Lakes South Slope Schlumberger array. b. Green Lakes South Slope Wenner array ................................... 161 Figure 83. RES2DINV© inversion image showing a. Niwot 1 Wenner array. b. Niwot 1 dipole-dipole array.......................................................................................................... 162 Figure 84. RES2DINV© inversion image showing a. Niwot 2 Wenner array. b. Niwot 2 dipole-dipole array.......................................................................................................... 163 Figure 85. RES2DINV© inversion image showing a. Niwot 3 Wenner array. b. Niwot 3 dipole-dipole array.......................................................................................................... 164 Figure 86. RES2DINV© inversion image showing a. Niwot 4 Schlumberger array. b. Niwot 4 Wenner array. c. Niwot 4 dipole-dipole array Niwot 4 .................................... 165 Figure 87. RES2DINV© inversion image showing a. Niwot 5 Wenner array. b. Niwot 5 dipole-dipole array.......................................................................................................... 166 Figure 88. RES2DINV© inversion image showing a. Niwot 6 Wenner array. b. Niwot 6 dipole-dipole array.......................................................................................................... 167 Figure 89. RES2DINV© inversion image showing a. Niwot 789 Wenner array. b. Niwot 789 dipole-dipole array................................................................................................... 168 LIST OF TABLES Table 1. Thermal properties of soil constituents at 20 °C and 1 atm for quartz, soil minerals, organic matter, water, and air (from Wilson, 1999).......................................... 21 Table 2. . Thermal and physical properties of liquid water (from Wilson, 1999)............ 21 Table 3. Thermal and physical properties of ice (from Wilson, 1999) ............................ 22 Table 4. Classification of landforms produced by downslope soil movements along Niwot Ridge (from Benedict, 1970) ................................................................................. 34 Table 5. Estimated altitude of the equilibrium line in Green Lakes Valley during the past 20ka................................................................................................................................... 71 Table 6. List of the ......and temporal lag constants summarizing calculated parameters from Hopkins Memorial Forest, MA using Eqn. 4 ........................................ 96 LIST OF EQUATIONS Eqn. 1 .................................................................................................................................. 9 Eqn. 2 .................................................................................................................................. 9 Eqn. 3 ................................................................................................................................ 22 Eqn. 4 ................................................................................................................................ 86 Eqn. 5 ................................................................................................................................ 87 Eqn. 6 ................................................................................................................................ 87 Eqn. 7 ................................................................................................................................ 92 Eqn. 8 ................................................................................................................................ 92 Eqn. 9 ................................................................................................................................ 94 Eqn. 10 .............................................................................................................................. 94 Eqn. 11 .............................................................................................................................. 94 Eqn. 12 .............................................................................................................................. 94 Eqn. 13 .............................................................................................................................. 96 Eqn. 14 .............................................................................................................................. 96 Eqn. 15 .............................................................................................................................. 97 Eqn. 16 ............................................................................................................................ 141 Eqn. 17 ............................................................................................................................ 142 Eqn. 18 ............................................................................................................................ 142 Eqn. 19 ............................................................................................................................ 143 Eqn. 20 ............................................................................................................................ 144 Eqn. 21 ............................................................................................................................ 144 Eqn. 22 ............................................................................................................................ 145 Eqn. 23 and Eqn. 24........................................................................................................ 146 Eqn. 25 ............................................................................................................................ 146 Eqn. 26 ............................................................................................................................ 146 Eqn. 27 ............................................................................................................................ 146 LIST OF APPENDICES Appendix A.....................................................................................................................141 Appendix B .....................................................................................................................148 Appendix C.....................................................................................................................153 Appendix D.....................................................................................................................169 INTRODUCTION Introduction Mountain permafrost is an excellent indicator of climate change and has been rapidly disappearing through the past several decades in both Europe and North America as a direct consequence of rising air temperatures (Leopold et al., 2013). Permafrost has been extensively studied in the Arctic through the 20th century, and is defined as rock or soil that may contain water which has a temperature continuously at or below 0 °C for at least two consecutive years. The shallow subsurface has recently become an especially popular topic of geologic research in both the United States and several Critical Zone Observatories throughout the world (Leopold et al., 2008). Large-scale melting of frozen ground has brought the decline of permafrost into the global spotlight; the melting of permafrost alone could raise global temperatures by 0.2 -0.8 °C by the year 2100 (Leopold, 2010; Romm, 2012). Niwot Ridge is located at the threshold of a permafrost supporting environment at an elevation of approximately 3,600 m in the Front Range of Colorado. Conclusive evidence of frozen ground was recorded at several locations along Niwot Ridge in the 1970s (Benedict, 1970), but results of my thesis indicate that permanent ice lenses no longer exist at this location. I used non-invasive geophysical techniques such as ERT and GPR for detailed observation of the Critical Zone, allowing for thorough mapping and visualization of the subsurface in environmentally sensitive areas (Leopold et. al, 2008). To better understand the processes and conditions that favor permafrost, I outline the basics of the geophysical techniques used in this study before establishing a thorough review of the location’s past and present climate, and investigate its influence on the thermodynamics of the near-subsurface. The Critical Zone Plants, animals, and human beings live in a dynamic environment where the solid earth interacts with atmospheric gases, meteoric water, and the biosphere in a zone extending from the outer vegetation canopy to the base of active groundwater (Anderson et al., 2007; Anderson et al., 2008). This life-supporting region, termed the Critical Zone (CZ), provides the primary habitat for terrestrial life and has been increasingly affected by human actions on both local and global scales (Fig. 1). Interactions among plants and soils, microbes and mineral surfaces, as well as megafauna and the earth’s surface are important everyday occurrences within the CZ (Anderson et al., 2008). Humans get all nutrients from the dynamic systems of the CZ, so preserving its balance is vital to our survival. The CZ can be thought of as a well-mixed feed-through reactor that advances the weathering front downward and brings unweathered rock into the machine. Precipitation provides necessary fluids, biologic and physical processes stir the upper-most layers, and both the Earth’s internal energy and solar radiation provide energy to sustain life and alter the bedrock (Anderson et al., 2007; Anderson et al., 2008). Soil hydrology, nutrient and gas flux, temperature regimes, and pedogenesis are among the processes that influence the biogeosphere and reflect and control various geomorphic processes in the CZ. Weathering and soil-forming processes, periglacial activity, soil creep, forest fires, flash floods, and other mass movement events slowly transform and mix the subsurface over the longer time periods (Leopold et al., 2008). The CZ is composed of slope deposits of various origins (i.e. soil profiles, vegetation, animal burrows etc.), each with their own thickness, age, and composition. Figure 1. Cross-section of the CZ showing bedrock and regolith layers along with associated geomorphic processes for each. The rock parcel trajectory (black dashed line), and water flow paths (blue dashed line) are presented in a system with minimal solid material added at the surface (image from Anderson et al., 2007, Figure 1, p.2) Both chemical and physical weathering of the CZ facilitate the transportation of mass from the land surface and the interior of the CZ. Chemical weathering is driven by thermodynamic disequilibrium compared to the conditions under which various minerals form, whereas mechanical weathering and physical erosion reflect stress gradients that break and move material depending on the magnitude of each stress (Anderson et al., 2007). Each of these processes physically alters the rock, transforming the CZ to its present state (Anderson et al., 2007). The CZ can be modeled as a steady state in which “rock materials are removed at the same rate that they are replenished,” a condition in which regolith transport equals regolith production rate (Brantley, 2008, p. 1454). Human activates have made maintaining this equilibrium an increasing difficult challenge as industrialization and agricultural practices continues to alter the land’s surface (Wilkinson, 2007). These practices provide interesting statistics regarding the amount of anthropogenically eroded surfaces and their contribution to global soil denudation (Wilkinson, 2007). Understanding the various processes operating and shaping the CZ has been recognized as a fundamental interdisciplinary problem by geochemists, soil scientists, hydrologists, biologists, entomologists, and geomorphologists (Anderson et al., 2008). In 2007, the National Science Foundation funded the Southern Sierra Critical Zone Observatory (University of California, Merced), Boulder Creek Critical Zone Observatory (University of Colorado, Boulder), and Susquehanna-Shale Hills Critical Zone Observatory (Penn State University) to advance understanding of the complex interactions that occur in Earth’s shallow surface environments. During 2009, funding for an additional three Critical Zone Observatories (CZO’s) was granted by the NSF for research locations in Jemez River Basin (Arizona/New Mexico), Christina River Basin (Delaware/Pennsylvania) and Luquillo (Puerto Rico) (“CZO about Us,” 2012). These American CZO’s are in collaboration with six other CZO’s around the world including the United Kingdom, Czech Republic, and Switzerland (Lin, 2011). The Boulder Creek CZO is primarily focused on understanding how erosion and weathering produce different CZ architectures, and how their interactions control the biological and hydrological functions of the CZ (Anderson et al., 2008). Research focuses on the origin and future development of the structure, and the physical and chemical parameters of the subsurface alpine area. The Boulder Creek CZO is an ideal location to study permafrost due to its accessibility and relatively mild climate compared to other alpine regions of the world. These factors have enabled researchers to conduct studies on permafrost along Niwot Ridge since the 1960s and 1970s. This location provides the ideal length and quality of climatic and geologic records for the area, allowing scientists to compare and contrast data collected over the past 50 years using modern technology and techniques of permafrost detection to further understand global warming and climate change in the near future (Leopold et al., 2010). Worldwide focus on climate has dramatically increased recently due to scientific understanding of global warming and its effects on civilizations around the world. Due to the precision and duration of climate measurements on Niwot Ridge since 1968, the past century’s climate is better understood in the Boulder Creek Watershed than almost anywhere else in the world. The LTER site on Niwot Ridge, the Boulder Creek CZO (czo.colorado.edu), and the Green Lakes and 4th of July Valleys offer ideal conditions to study changes in mountain permafrost. The mission statement of the LTER and CZO programs states that they intend to study the genesis and future development of the region as well as the physical and chemical parameters of the subsurface, while offering a chance to combine older data with modern technology of permafrost detection and monitoring in order to discuss the concept of climate change (Leopold et al, 2010). Studies in the upland Front Range have focused in considerable detail over the past 50 years on various geomorphic, hydrologic, climatic, and biogeochemical aspects of the snowmelt and rainwater discharge in local streams (colorado.edu/mrs/mrspubs.html). Due to the limited accessibility and thickness of regolith, it is difficult to physically observe the material that lies beneath the surface of gelifluction lobes and periglacial regions. Benedict (1970) dug profiles though several gelifluction lobes and computed orientations for hundreds of clasts within these lobes. Today it remains possible to study the subsurface in two or three dimensions using various physical methods (manually digging pits or trenching slopes) or much less intrusive geophysical methods such as ERT and GPR. Electrical Resistivity Tomography History of ERT The concept of using electrical prospecting dates back to the 1830s when Robert Fox began to experiment with natural currents associated with sulfide ore deposits in Cornwall, England. Many other physicists of the time also investigated the effect of direct current of many different ores and minerals. In the early 1910’s Conrad Schlumberger in France and Frank Wenner in the United States (for whom two ERT arrays used in this paper are named after respectively) applied current to the ground and measured the resulting voltage potential, establishing the direct current electrical resistivity method. In 1914, Schlumberger discovered a rich ore deposit in Serbia using his self-potential method, and, in 1917, H.R. Conklin introduced the electromagnetic method. Schlumberger investigated telluric currents in the 1920’s and 1930’s to explore and map ore deposits. In 1928, two American geologists mapped high-resistivity bedrock during an investigation for a proposed dam location in what was one of the earliest noncommercial applications of ERT (Burger et al., 2006). Russian mathematician Andrey Tikhonov discovered a procedure for solving the mathematical inversion of an electrical resistivity tomogram using a simple 2-layered medium. Without the benefit of computers, Tikhonov and a team of Russian geophysicists discovered a large deposit of copper using this technique, earning them the State Prize of the Soviet Union (Deidda et al., 2003). The invention and development of computers in the second half of the 19th century allowed ERT mathematical inversion processes to both easily and quickly be computed. Interpretation software has become much more powerful with the development of increased processor speeds and computer vector graphics editing programs such as Adobe Illustrator©, allowing for accurate subsurface analysis. Since the 1940’s technological progress, development of a solid theoretical basis, and improvement of the interpretation process have enhanced the applications of geophysical practicality. Field geophysical hardware is continually being improved by more industrial, robust, light-weight electrode chains that currently last through several field seasons (Burger et al., 2006). ERT Background When properly used, ERT allows detailed measurement and visualization of individual subsurface layers’ various DC electrical resistivities. Resistivity values can then be correlated with various material types such as soil, saprolite, and ice to produce a detailed map of the subsurface (Leopold et al., 2008). This technique generates an electric field using two electrodes to introduce a direct current (or low-frequency alternating current) into the ground surface. Electrical current flows through the subsurface, and potential voltage is measured at two additional electrodes on the surface (Mühll, 2002; Burger et al., 2006). By systematically changing which electrodes are the current and potential electrodes, increased ground penetration depth can be achieved (Mühll, 2002). Variations in resistance to current flow at depth cause distinctive variations in the potential measurements, providing information about subsurface structure and composition (Burger et al., 2006). Electrical resistivity is the most common electric geophysical method applied to shallow subsurface investigations (Burger et al., 2006). Electrical Resistivity Tomography can measure the resistivity and material properties of deposits in sensitive alpine environments without extensive excavation, which would otherwise be required to locate permafrost. These disturbances alter the biogeosphere and can alter the results of future studies (Leopold, 2008). ERT Physics In two-dimensional ERT surveys, such as those discussed in this paper, electrodes are configured along a straight line at equal intervals, all attached by a multi-conductor cable (Mühll, 2002). All possible combinations of two current electrodes and two potential electrodes are measured and recorded using the Wenner configuration, which uses two potential (measuring) electrodes placed inside the two current electrodes (Fig. 2). The Wenner array is well suited to changes in the vertical resistivities, thus it is widely used to map horizontal structures such as sills or sedimentary layers (Loke, 2012). This array allows for higher current and potentials, as well as a low signal-noise ratio compared to other arrays. Figure 2. Wenner configuration showing two current electrodes placed outside of the two potential electrodes (image modified from Burger et al., 2006, Figure 5.9, p. 278) For the Wenner array, the equation used to measure the apparent resistivity, .. (measured in .m), using the measured current, I (measured in amps), electrode separation, a (measured in meters), voltage drop, ..(measured in volts), is ..... ...(Eqn. 1) . This equation can be expanded to equal . .... . . .... . .......... ... . (Eqn. 2) ... .... . . .. .. ... .. ...... ...... .. A dipole-dipole array (Fig. 3) is used to test for vertically layered structures such as dikes and cavities because of the low electromagnetic coupling between the current and potential electrodes (Loke, 2012). The dipole-dipole array places two current electrodes on one side of the line, and two potential electrodes at the other end of the line, allowing the current to travel further through the subsurface. This configuration is most sensitive to resistivity changes below the electrodes in each dipole pair (Loke, 2012). In addition to Wenner and dipole-dipole arrays, both Schlumberger and Wenner-Half (Wenner-Lee) arrays were used along ERT Lines 1, 2, 3, and 8 (Fig. 4). Figure 3. The dipole-dipole configuration showing two potential electrodes (left) and two current electrodes (right) spaced a distance “a” from each other. Lines of constant voltage are shown (dotted) in relation to electric field lines (solid) Figure 4. Electrode geometries used in this resistivity surveying study. a. Wenner. b. Wenner/Lee. c. Schlumberger. d. Dipole-dipole (image modified from Burger et al., 2006, Figure 5.9, p. 278). Temperature changes at the triple point of water impact ERT’s sensitive measurement ability (Leopold et al., 2013). Resistivity values vary by orders of magnitude among soil (0-800 .m), saprolite (600-1000 .m), and ice (10 k.m-2000 k.m), allowing effective determination of frozen and unfrozen ground near the surface (Mühll, 2002; Leopold et al., 2013). Interstitial air and buried boulders in the subsurface provide difficult paths for DC electricity to follow, creating high resistivity values comparable with those of ice. A complete table of samples and their associated resistivity values (Fig. 5) allows interpreted of ERT images in this study (Lewis et al., 2013). Throughout the past few decades, ERT resolution has dramatically increased with improvements in technology, from simple four-electrode arrays with one-dimensional inversion modeling to intricate three-dimensional arrays using thousands of electrodes and data points in areas of complex geology (Mühll, 2002). Application of ERT in permafrost environments is a relatively new concept but has been extremely successful so far in ice-rich areas. Figure 5. Resistivity values used for interpretation of resistivity-depth profiles generated in RES2DINV (image from Leopold et at., 2013, p. 13). The ERT procedure sends electric current from the surface at one location to the surface at another location, but in order to attain a layered structure representing the true resistivity at certain depths, field resistivity values must be mathematically manipulated. This process requires complex equations (performed via computer software) in order to invert the data before analysis. Electrical resistivity-depth profiles are difficult to interpret because of the necessary specific knowledge of the region at hand. In addition, the variance of resistivity reflects both the temperature signal, which suggests the presence or absence of ice, and the superposed structure and geology of the subsurface (Loke, 2012). Additional geophysical techniques such as GPR or Shallow Seismic Reflection (SSR) can be performed along the same lines to help reduce the error and ambiguities involved with interpreting ERT (Loke, 2012; Leopold, 2008). Depth of Investigation Oldenburg and Li (1999) introduced the Depth of Investigation (DOI) index and first used it for permafrost sites studied by Marescot et al. (2003). The term “depth of investigation” refers to the depth below the surface at which the measured data becomes insensitive to the value of the physical properties of the earth. Estimates of the DOI for ERT surveys are essential because any results below this depth are not sufficiently accurate to be interpreted with confidence (Oldenburg and Li, 1999). Typically, a DOI index value of 0.2 or more, as suggested by Oldenburg and Li (1999), indicates less reliable regions of ERT models. Low DOI values indicate that the model is triggered by the field data and not by errors in the inversion software. Using RES2DINV©, it is possible to compare ERT results by calculating Eqn. 5 and Eqn. 6 for both high and low background models to create a DOI images for each line. These images will determine areas of the subsurface that can be interpreted accurately, and those that can not. Heat Flow Air temperatures depend on a local thermal environment, which changes hourly and seasonally, whereas subsurface temperatures in cold permafrost environments represent a systematic running mean of recent temperature history at the surface. Since temperature changes at the surface require a certain amount of time for the effect to propagate downward into the earth, deeper temperature measurements are uniformly and continuously integrated over a longer time interval from earlier in the surface temperature history. It is possible to obtain climatic history information for the past century by drilling and examining a 200 m permafrost core at any suitable location where the information is desired. Short-term noise is progressively filtered out at depth, allowing only the subsurface to retain information from major events and to reveal long-term climate trends. Heat transfer in thick permafrost is mostly by conduction because no free water is able to move through the system, simplifying the thermal physics of the subsurface dramatically. Conduction of thermal energy strongly depends on the medium through which energy travels, since some subsurface lithologies are more effective at conducting thermal energy than others. In areas of thin layers of permafrost, however, melt water from seasonal ice lenses provides thermal energy as it cools to the surrounding subsurface temperature, gradually drawing energy away from the system. In addition, solar radiation drives both diurnal and seasonal surface temperature oscillations, providing exposed south-facing areas with more thermal energy than shaded regions on north-facing slopes. Models of climate warming predict climate change will be greatest in the Arctic region, where most of Earth’s permafrost resides, creating a prudent demand for understanding heat flow in permafrost areas. Geothermal data from oil exploration wells in Alaska from the 1980s show a substantial change in the heat balance of permafrost on a multi-decadal scale. This long-term change is indicating that early signs of anthropogenic effects are already affecting the ground sub-surface temperatures (Anderson and Anderson, 2010; Lauchenbush, 1986). While the steady-state geotherm is an excellent approximation for long-term depth to permafrost, borings in northern Alaska and Canada reveal a persistent and ubiquitous departure from the expected linear temperature profiles. In all cases, the upper hundred meters of these profiles show an anomalous curvature, reflecting temperatures warmer than one would expect from projecting the geothermal gradient up from the base of the permafrost. The warm near-subsurface of these profiles has been attributed to the long-term warming of the regional climate of this arctic environment. The 100 m-deep perturbation that is commonly observed in thick permafrost temperature profiles (Fig. 6) corresponds to a century of warmer surface temperatures, approximately the same amount of time in which humans have significantly altered the greenhouse gas content of the atmosphere. This perturbation can only be explained by layers with different thermal conductivities, creating different geothermal heat gradients in the subsurface. Although the estimates for temperature changes throughout the last century contain large uncertainties, it is impossible to treat this data as random temperature variations of northern Alaska and Canada. Figure 6. Temperature-depth profiles from 14 drill wells throughout Cape Thompson, Barrow, and Prudhoe Bay, Alaska. All of these profiles show a significant departure from the expected linear profile within the upper 100-200 m (image from Lauchenbruch, 1986, Figure 6, p. 234) Repeated measurements from a borehole at East Teshekpuk Lake, Alaska, over the last three decades show substantial warming of the upper 40 m of the subsurface, while temperatures below 40 m stay constant (Fig. 7) (Clow and Urban, 2002). The profiles show little change through the late 1970s and 1980s, but temperatures have increased dramatically since the mid 1990’s (Clow and Urban, 2002). Over thirty years ago, Fahey (1975) predicted that if a temporary shift to cooler climates became more severe, longer periods of seasonal frost would reduce the frequency of diurnal freeze-thaw processes and hinder the movement of sorted circle development. Throughout the study of periglacial landscapes and permafrost, we must consider the potential impacts of this present day warming and their impacts on currently active geomorphic processes. Figure 7. Temperature depth profile of a borehole at East Teshekpuk Lake, Alaska over the last three decades as part of the Global Terrestrial Network-Permafrost deep borehole array (image modified from Anderson and Anderson, 2010, Figure 9.9, p. 280) Permafrost Permafrost was first defined in the Arctic by F. Bruemmer (1987), forms when the ground surface cools during the winter enough to create a subsurface frozen zone that is able to persist throughout the following two summers (Williams and Smith, 1989). When temperatures allow the ground to remain frozen for a longer time period, freezing temperatures penetrate deeper into the subsurface and create a thick layer of laterally extensive permafrost. The near surface of many regions throughout the world undergoes annual freezing and thawing cycles in which the water content, soil strength, and bulk density of the mobile regolith change dramatically with diurnal and seasonal cycles. A permanently frozen ground referred to as permafrost may exist below this seasonally thawed “active layer.” Permafrost is highly sensitive to changing air temperatures because warm temperatures alter the thawing depth of the annual active layer as well as the duration of the refreezing process, making permafrost an ideal indicator of global warming (Williams and Smith, 1989). Ice lenses and periglacial activity may also depend on the available soil moisture content as much as temperatures, since the freezing and thawing process of regolith materials must have access to water in order for much material to be transported. As this regolith undergoes temperature fluctuations, it brings profound geomorphic consequences to the structure of subsurface materials. Heat loss from Earth’s interior and solar energy input at the surface reach a steady state equilibrium that controls temperatures and locations where permafrost can form. The balance between these forces is extremely delicate and temperature dependent, which allows permafrost extent to range from small discontinuous patches (1-10 m2) to large continuous zones millions of square kilometers in area (Mühll et al., 2002). Permafrost can be classified as: continuous, where it underlies >90% of the surface; discontinuous, where it underlies 50-90% of the surface; sporadic, where it underlies 10-50% of the surface; or isolated, where it underlies less than 10% of the surface (Fig. 8) (Anderson et al., 2010; Leopold, 2010). Figure 8. Permafrost cross-section showing continuous, discontinuous, and sporadic zones of permafrost overlain by the active layer (image from Bierman and Montgomery, 2013, Figure 9.12, p. 31) Smith and Riseborough (2002) present a map of the worldwide distribution of permafrost that predicts as much as 20-25% of the Earth’s land surface overlies permafrost (French, 2007). Permafrost mostly exists in Greenland, Russia, Alaska, Canada, and other high latitude regions with a plentiful supply of water (Janke, 2005). A small percentage of permafrost is classified as mountain permafrost, which exists in high altitude, mid-and high-latitude regions such as the Himalayas in Asia (Jin et al, 2000), the Alps in Europe (Harris et al., 2009), and the Rocky Mountains in North America (Janke, 2005), in addition to other high altitude mountain ranges (Harris et al., 2009). Most mountain permafrost in the United States is located in Alaska because of the high latitude, low temperatures, and low solar radiation. In addition, there are areas of frozen ground in the Front Range of Colorado as well as sporadically throughout Wyoming and the Sierra Nevada’s in California (Leopold et al, 2010). In order for permafrost to exist in such low-latitude locations, a list of precise conditions must be met. Studies agree that cold air temperatures must be maintained for sufficient time periods to keep ground temperatures remaining below 0 °C. Ground surface temperature (which itself depends on air temperature), vegetation type, vegetation density, drainage, soil type, aspect, slope, elevation, thickness of snow cover, solar radiation, and site-specific internal heat flow are poorly understood factors that influence the spatial distribution, thickness, and temperature of permafrost (Leopold et al., 2010). Conditions therefore can change within short distances depending especially on topography, creating a complex mosaic of permafrost (Muhll et al., 2002). Many of the physical parameters that characterize permafrost are not well understood because the processes affecting the distribution of permafrost are complex and have only been thoroughly studied for 40 years (Janke, 2005; Janke, 2012). The temperature structure required to form permafrost is complex (Fig. 9), so the ground subsurface is divided into several horizontal layers to better understand how heat flows from each layer to the next. The active layer is the thermal boundary layer in which temperatures shift over diurnal and seasonal cycles reflecting atmospheric and solar energy. The depth of the active layer is important to engineers designing structures located in periglacial regions because free-thaw processes will make the near subsurface unstable for large buildings. The magnitude of diurnal and seasonal fluctuations decrease gradually with increased depth until the mean annual temperature is effectively maintained. The active layer controls the hydrology of the landscape and extends to a depth at which the annual maximum temperature reaches 0 °C. Any water that penetrates through the base of the active layer through a frost crack will begin to freeze. Since the frozen ground is so solid, it acts as a barrier, preventing penetration of roots and defining the base of the biologically active (during the summer at least) sod or peat layer. Since the surface water has nowhere to drain, the ground often contains a soggy layer, varying in thickness from 0.2 m in wet areas to 2 m in dry areas, above the frozen ground (Lauchenbruch, 1986; Anderson and Anderson, 2010). Figure 9. Temperature-depth profile showing variations of the near-surface region for a MAT of -10 °C and temperature fluctuation of 15 °C (image from Anderson and Anderson, 2010, Figure 9.7, p. 278) Below the active layer, average annual temperatures increase monotonically following the geothermal gradient (dT/dz) at that particular latitude and location. At some depth below the active layer, the temperature rises above 0 °C again because of heat escaping from within the earth, defining the base of the permafrost. In order to determine what features of the rock and climate control the depth of the base of permafrost, the local geothermal heat flow is defined, Q (measured in W m-2), and the local conductivity of the rock, k (measured in W/m K). Both the local geothermal heat flow and conductivity are functions of the material in which thermal energy is conducted through. Materials such as quartz, organic matter, air, liquid water, and ice have thermal conductivities that range several orders of magnitude ( Table 1, Table 2, and Table 3). Material Density, . (kg/m3) Specific heat, c (kJ/kg °C) Thermal conductivity, k (W/m K) Latent heat of evaporation (kJ/kg) Quartz 2650 .732 8.4 43 Many soil minerals * 2650 .732 2.9 43 Soil organic matter * 1300 1.925 0.25 15 Water 1000 4.184 0.6 1.42 Air 1.2 1.004 .026 0.21 Material Density, . (kg/m3) Specific heat, c (kJ/kg °C) Thermal conductivity, k (W/m K) Latent heat of evaporation (kJ/kg) Quartz 2650 .732 8.4 43 Many soil minerals * 2650 .732 2.9 43 Soil organic matter * 1300 1.925 0.25 15 Water 1000 4.184 0.6 1.42 Air 1.2 1.004 .026 0.21 Table 1. Thermal properties of soil constituents at 20 °C and 1 atm for quartz, soil minerals, organic matter, water, and air (from Wilson, 1999) * Approximate average values 21 Table 2. . Thermal and physical properties of liquid water (from Wilson, 1999) Temperature, T (°C) Density, . (kg/m3) Specific heat, c (kJ/kg °C) Thermal conductivity, k (W/m K) Latent heat of evaporation (kJ/kg) -10 997.94 4.268 --­ 2523 -5 999.1 8 --­ --­ --­ 0 999.87 4.2150 0.561 2499 5 1000.00 --­ --­ --­ 10 999.99 4.1995 0.574 2487 15 999.73 4.1894 0.596 2476 Table 3. Thermal and physical properties of ice (from Wilson, 1999) Temperature, T (°C) Density, . (kg/m3) Specific heat, c (kJ/kg °C) Thermal conductivity, k (W/m K) Latent heat of sublimation (kJ/kg) -20 920 1.958 2.433 2836 -10 919 2.029 2.319 2835 0 917 2.105 2.24 2833 In the steady-state profile, the MAT at the surface, ..., has remained constant for an extended period of time. The equation for the mean thermal structure is therefore: . ........... (Eqn. 3) . Many different field techniques including excavations, Bottom Temperature of Winter Snow (BTW), and various geophysical methods have been used to map and investigate permafrost in recent decades (Janke, 2005). These field methods often are time consuming, costly, and limited in spatial extent; requiring modeling techniques to estimate the distribution of permafrost using Geographic Information Systems and variables from Digital Elevation Models of the area such as solar radiation, aspect, slope, elevation etc. (Janke, 2005). In order to better study subsurface temperature changes in high mountain regions, extensive monitoring programs have been established in regions such as Switzerland and Colorado (Noetzli and Mühll, 2010). Permafrost extent and thickness in Rocky Mountain National Park may have decreased due to rising temperatures over the last few decades, bringing awareness and concern to regional water managers. These directors control the amount of water that large populations such as Boulder and Denver, CO can receive from annual snowmelt and melt out of seasonal ice lenses (Janke et al., 2012). Periglacial Landforms Many locations with permafrost supporting thermal conditions also have environments capable of supporting periglacial landforms. Large-scale periglacial landforms and deposits such as gelifluction lobes, sorted circles, and rock glaciers, are widely distributed although active development of these features is restricted to sites containing substantial amounts of water (Benedict, 1970; Benedict, 1981). Inactive permafrost is defined as permafrost which formed in the past and persists in areas where it would not form today due to differences in moisture and climate. Active periglacial action is promoted most where the ground is unprotected by deep snow, but spring snow melt is available for lubricating the subsurface and freezing at night (Marker, 1990). Even though freeze-thaw processes were active on Niwot Ridge in the 1970s and 1980s, few locations have current periglacial processes sufficient to maintain movement of large features such as transverse lobes and large polygons (Marker, 1990). Periglacial landforms are developed best in the open tundra because of the lack of trees and roots, which inhibit ground movement and regulate diurnal temperature variations which limit the extent of freeze-thaw processes. High velocity winds blow snow away from the ground of many active periglacial regions, leaving the surface free of snow cover. By removing the ground’s normal winter layer of insulation and allowing low temperatures to penetrate deep into the ground, these areas become susceptible to severe freeze-thaw and frost action. Cold subsurface temperatures allow ice lenses to accumulate during the winter and spring before slowly melting out throughout the rest of the year. Snow drifting is restricted in the tundra by intermittent krummholz (small groups of trees in clumps), which create large snow drifts and prevent snow from blowing downwind (Marker, 1990). Permafrost is believed to exist (Marker, 1990) throughout regions of Colorado and New Mexico above approximately 3,400 m, but data from our study shows that this might not be entirely correct. A thick layer of snow on the surface provides sufficient insulation to keep the subsurface above the permafrost threshold for much of the year. When snow is blown away from the surface, the ground is able to become much colder and maintain permafrost-suitable conditions for many consecutive months. Studies on Niwot Ridge have shown that blowing snow occurs over 50% of the year, with 95% of the days in January having blowing snow of some intensity. The average amount of snow -1 -1 transported is 25 g ms under typical blowing snow conditions, but occasionally ranges -1 -1 up to 480 g ms , reducing visibility to less than 15 m (Berg, 1986). Figure 10. Soil temperatures between 9/5/1996 and 7/2/1997 at the Tundra Research Station, Niwot Ridge, taken at a depth of 30 cm showing thick snow cover (top) and thin snow cover (bottom) (images from Hamann, 1998, Figures 5.2 and 5.4, pp. 42 and 44) Greenstein (1983) concluded that elevations above 3500 m contain permafrost and lower elevations (3470 to 2930) do not contain permafrost, which corresponds well with the model Janke (2005) presents. Based on topoclimatic information from within this area, Janke (2005) created a permafrost distribution model of the Front Range, and claims that “the [permafrost extent] model accurately represents regional distribution of permafrost,” (Fig. 11), although Leopold et al. (2010) suggest otherwise. Janke (2005) indicates that continuous permafrost extends to 3550 m on north facing slopes and 3600 m on south facing slopes, whereas discontinuous permafrost extends to 3200 m on north facing slopes and 3300 m on south facing slopes. Ives et al. (1971) places the lower level of discontinuous permafrost at 3500 and 3750 m, which corresponds with the -1 °C MAT isotherm on Niwot Ridge. Figure 11. A model for permafrost in the Colorado Front Range displaying a 50-75% probability of permafrost occurring (light purple) and 76-100% chance of permafrost existing (dark purple) (image from Janke, 2005, Figure 5, p. 383) Combined evidence from ground and air temperatures, periglacial deposits, topography, and ground cover indicates that permafrost might underlie over 80% of the Green Lakes Valley, although this is most likely largely overestimated (Janke, 2005). One study along Trail Ridge Road between 3500-2700 m in Rocky Mountain National Park compared three different permafrost distribution models, but the spatial distribution only has an 8% overlap (Janke et al., 2012). Temperature loggers down to 7 m did not provide any indication for freezing conditions throughout the winter, indicating that permafrost models might be incorrect (Janke et al., 2012). Periglacial freeze-thaw action has significantly modified many high and mid-altitude slopes, isolated mountains, and ridges above 3,300 m in elevation throughout the Front Range (Marker, 1990; Benedict 1966, 1970; Ives and Fahey 1971). Permafrost studies continue today with improved technology such as LiDAR, satellite imagery, and digital thermometers, as well as increased funding from the National Science Foundation (colorado.edu/mrs). Periglacial and permafrost features on Niwot Ridge have been used as model examples in textbooks to ensure that other studies are able to accurately recognize similar landforms elsewhere (Marker, 1990). Permafrost measurements in the Colorado Front Range date back to the mid­1800s when miners staking claims for gold and other precious minerals claimed that they had to dig through permafrost to excavate their quarries. Construction workers during the excavation of Silver Lake dam (elevation of 3124 m) and Left Hand Reservoir dam (elevation of 3250 m) state that they found continuous permafrost sometimes exceeding 60 m in thickness during the late 1800’s (Ives and Fahey, 1971). The first scientific studies of permafrost on Niwot Ridge were completed in the 1970s by Ives and Fahey (1971) and Ives (1973). Although these two studies described many sites as permafrost locations using field methods, Janke (2005) predicted many of the same locations using computer modeling (Ives and Fahey, 1971; Ives, 1973; Janke, 2005). Greenstein (1983) hypothesized an even more extensive distribution of permafrost on Niwot Ridge in locations similar to Ives and Fahey (1971) by comparing a predictive model based on MAT, Degree Days, and aspect with BTW measurements and freeze-thaw indices (Janke, 2008). Many of these models rely on elevation, MAT, solar radiation, aspect, hydrological and other physical parameters as well as GIS techniques. These models rely on ground-based data such as excavated pits, core drillings, boreholes, and evidence of periglacial landforms from previous studies as opposed to collecting new data at the time each model was produced. The study of rock glaciers in Colorado began in the 1940s, when Jack Ives described the characteristics and locations of several rock glaciers throughout the Front Range. White (1971) was one of the first to study the Arapaho rock glacier, a small drift formation separated from the stagnant Arapaho Glacier by an ice-cored terminal moraine and several lateral troughs (Fig. 12). White and Benedict discovered interstitial ice within the rock glacier, which they claimed was responsible for the features movement downslope (White, 1971). Climate Climate change provides first-order control on the existence and extent of permafrost in many locations. Exact climatic conditions, however, may be difficult to calculate both presently and in the past. Large-scale melting of permafrost has brought the decline of frozen ground due to climate change into the global spotlight over the past several decades. In a recent article published by ThinkProgress.org, the author postulates that “the single most important carbon-cycle feedback is the melting of the permafrost” (Leopold, 2010; Romm, 2012). While my study does not focus on climate change and global warming, it is necessary to introduce some of these concepts in order to fully understand the extent of permafrost during the past and present. Climate is broadly defined as long term averages (usually over a period of 30 years) of temperature, humidity, atmospheric pressure, wind, precipitation, and other meteorological measurements. Although often confused in everyday nomenclature, weather refers to the present condition of the aforementioned variables on a week-to­week basis. Deciding which factors to include in climate modeling (MAT, subsurface temperatures at certain depths, snowpack depth, bottom of the snow temperature, etc.) has proven difficult for the scientific community, since each combination of variables yields somewhat different results. Comparing proxy records of climate from ice cores, 10Be exposure ages, pollen data, and glacial moraines has allowed scientists to track long term atmospheric conditions, which influence factors such as ice cap extent, sea level rise, and mountain permafrost distribution. Climate change and its associated effects on ground surface morphology, landscape dynamics, and natural hazards are raising the importance of mapping permafrost distribution on both the global and local scale. Rising air temperatures and warming permafrost have been studied in the European mountains (Harris et al., 2009) and high altitude regions of China (Li et al, 2008). These studies give clear evidence of rising air temperatures and warming permafrost. Climate warming of the past several decades is expected to increase permafrost thawing and the occurrence of themokarsts in periglacial regions (Gooseff, 2009). The failure by mass wasting of overlying soil and vegetation results from the thawing of permafrost on hillslopes, which has been studied in both continuous and discontinuous permafrost regions of northern Canada (Gooseff, 2009). Previous Geophysical Studies on Niwot Ridge Initial geophysical work was performed during the summer of 2005 using GPR to study the CZ at various sites along Niwot Ridge (Leopold et al., 2008). The study conducted by Leopold et al. (2008) gave the first indications that previously indicated permafrost locations lacked permanently frozen ground and that seasonal ice lenses completely melted in late August (Leopold et. al, 2010; Lewis et al., 2013). Additional studies (Leopold, 2013) introduced new geophysical data from both December, 2009, and May, 2011, to further document the lack of permafrost and melting of seasonal ice lenses. The data did not indicate the existence of large ice lenses, which would be expected if permafrost existed at the study locations (Leopold et al., 2013). Whether this change in describing subsurface conditions is due to misinterpretation of subsurface conditions leading to an overestimate of permafrost, or represents a transformation in high altitude permafrost conditions over the last thirty years (as suggested by Caine, 2010) continues to be a major theme in alpine studies. My report is unique in that it combines field based observations and measured resistivities with geophysical techniques designed to detect permafrost. As high resolution satellite orthoimagery and LiDAR become more readily available in the coming decades, other permafrost features such as active patterned ground, summer ponds, or springs on ridge flanks, could provide more criteria about land cover and more detail about the distribution of permafrost without having to visit the locations in person (Janke, 2008). Periglacial Processes and Deposits Processes driven by frost and freezing temperatures produce unique features shaped by frost-heave cycles and soils which alternate between solid and semi-liquid states (Anderson et al., 2010). Since much of the land’s surface is either currently subject to seasonal freezing temperatures (up to 35% of the terrestrial surface) or has been shaped by freezing conditions in the past, the study of periglacial formations is relevant to nearly half of the earth’s land surface (Anderson et al., 2010). At the freeze-thaw margin, the phase change between water and ice can drive processes not found anywhere else on Earth (Anderson et al., 2010). During the 1970s, glaciologists estimated that nearly 100,000 km2 of alpine permafrost was present in the western United States between Washington and Arizona, much of which has melted due to global warming throughout the 20th century (Péwé, 1983). Inactive periglacial features extend 500 to 750 m vertically below the present periglacial limit on Niwot Ridge, well into the forest belt (Marker, 1990). Periglacial activities are driven mainly by frost heave and creep, the ratchet-like downslope movement of particles heaving perpendicular to the slope and subsequent vertical settling upon thawing due to gravity (Washburn, 1967). Periglacial landforms within my study area include gelifluction lobes, sorted circles, and rock glaciers. Subsurface deposits are variable in their depth to bedrock, composition, and water content (Fig. 13). composition of the CZ and lateral thickness variation of the periglacial slope deposits. Vertical exaggeration of the relief is approximately 4 times (image from Leopold et al., 2008, Figure 6, p. 89) Solifluction and Gelifluction Solifluction is the slow motion of saturated regolith from higher to lower elevation above an impermeable layer due to gravity, but this is not necessarily restricted to a periglacial environments nor frost-creep motion (Andersson, 1906). Gelifluction, on the other hand, can be described as “solifluction associated with frozen ground” (Washburn, 1967). The prominence of gelifluction in cold climates is due to diurnal and seasonal thawing of soil and ice, which provides moisture to the subsurface. This water becomes trapped above the frost or permafrost table, unable to penetrate downward, and saturates the upper layer of soil (Fig. 14) (Washburn, 1970; Benedict, 1970; Smith, 1988; Anhert, 2009). Gelifluction then transports the saturated material downhill above a frozen substrate (Andersson, 1906). Sufficient water availability, gradients steeper than 2°, and a fine interstitial matrix allow frost heave to slowly move lobes and terraces downhill (Washburn, 1973). This process depends on low air temperatures and precipitation, allowing it to be highly susceptible to warmer winters driven by climate change (see “Climate” section). The interaction between gelifluction, solifluction, and frost heave varies from site to site and year to year, and has fluctuated on a much larger scale during Holocene climatic oscillations. Gelifluction is an effective means of transporting debris downslope that interacts with underlying material, often transforming relatively weak bedrock into loose regolith over one or two cycles of lobe migration. By comparing the special extent of surface water, it is possible to dismiss the interaction between topography and the water table as the source of these pools. Gelifluction can occur under cold saturated conditions on gradients as little as 2° and can form many types of features (Washburn, 1970). These include 1) gelifluction sheets, which form as smooth surfaces with lobate lower margins; 2) gelifluction benches, characterized by their pronounced terrace shape that often forms with the longest axis parallel to the contour; 3) gelifluction lobes, characterized by their tongue-like appearance with the longest axis perpendicular to the contours; and 4) gelifluction streams, characterized by pronounced linear features also perpendicular to the contours (Washburn, 1970). Transverse gelifluction lobes occur at elevations higher than longitudinal lobes, although both are composed of similar material (Marker, 1990). There exists a strong altitudinal separation between specific types of landforms such as gelifluction lobes, cirque glaciers, tree line, scree fields, and permafrost as fully described in Marker (1990). Several different lobe and terrace morphologies exist along Niwot Ridge, including turf-and stone-banked terraces, and turf-and stone-banked lobes (Table 4). Turf-banked terraces (also referred to as “soil terraces,” “solifluction terraces,” and “nonsorted terraces” in the literature) are bench-like accumulations of granular material that lack visible sorting (Embleton and King, 1968). Turf-banked lobes (also referred to as “soil lobes,” “soil tongues,” “solifluction lobes,” “nonsorted lobes” and “nonsorted gelifluction lobes”) are similarly defined as lobate accumulations of moving soil that lack visible sorting (Embleton and King, 1968). Stone-banked terraces (also referred to as “block-banked terraces,” “boulder steps,” and “sorted terraces”) are defined as terrace-like accumulations of stones and boulders overlying a relatively stone-free moving subsoil (Embleton and King, 1968). Stone-banked lobes (also referred to as “stone­banked flow earth cones” and “stone-banked solifluction lobes”) are defined as lobate masses of rocky debris overlying relatively stone-free, fine-textured, moving soil (Embleton and King, 1968). For more detailed descriptions, see Benedict (1970). Table 4. Classification of landforms produced by downslope soil movements along Niwot Ridge (from Benedict, 1970) No surface expression Lobate Terrace-like Nonsorted Nonsorted sheet Turf-banked lobe Turf-banked terrace Sorted Blockfield Stone-banked lobe Stone-banked terrace Gelifluction lobes are easy to identify while fresh and active (Fig. 15), but become difficult to distinguish from other deposits after movement has terminated and erosion has modified the surface (Benedict, 1975). Long-term average rates of active frontal advance measured throughout the world range from 0.6 to 3.5 mm/year but rates have been measured from 0.6 to 12 cm/year on the same slope depending on relative saturation (Washburn, 1970; Benedict, 1975). The total horizontal component of downslope movement depends on the potential frost heave, downslope movement due to gelifluction, and retrograde movement due to settling (Fig. 16). In addition to moisture content, gradient is an important factor determining the magnitude of mean annual movement, since the component of gravity acting parallel to the slope increases as the sine of the gradient. As gelifluction lobes move downhill, water entrained within the front of the lobe seeps into the underlying saprolite and promotes short-term extensive freeze-thaw fracturing of the regolith and bedrock. Material within the regolith, such as gravel and coarse sand, has high porosity and permeability so it is able to promote excellent drainage and favors saturated flow. Silt is highly susceptible to flow since it lacks the cohesion of clays (Washburn, 1970). On the other hand, fines tend to remain wet longer than coarse grains so they do not promote flow as easily (Washburn, 1970). As liquid water slowly drains out from below a gelifluction lobe, it is able to carry fine material with it because the ground is poorly compacted from the passing of the lobe. This process leaves a local abundance of coarse open-work gravel with little fine grained material beneath the surface. Open-work gravel is locally abundant beneath the surface in the alpine region of Niwot Ridge; loose pebbles and cobbles with few fine particles holding them together make up most of the subsurface material. Caine (1968) found a three-layered structure underneath gelifluction features in Tasmania: a surface layer tens of centimeters thick (consisting of blocks held up by freeze-thaw action), an intermediate layer between 10-30 cm thick (characterized by interstitial hummock mud between the blocks), and a basal layer lying on bedrock (consisting of open-work gravel blocks lacking most interstitial filling). Even in high gradient saturated regions, frost heave is not uniform throughout the slope. Motion may be restricted in layers more than 50 cm below the surface because of the weight of the overlying material, limit of free-thaw efficiency, and absence of shear planes to accommodate the motion (Benedict, 1976; Harris et al., 2008). The thin active surface layer corresponds to the region in which the axes of elongate stones become predominately oriented downslope as the lobe migrates downhill. The plane of elongate stones aids the movement of the lobe by providing planes of weakness for material to slide upon (Rudberg, 1958). The direction of movement within gelifluction lobes has been inferred from the orientations of plant roots that penetrate deeply into the moving soil (Fig. 17). To determine rates of soil movement in the subsurface on Niwot Ridge, Benedict (1970) installed eight cement rods, each of which was 2.2 cm in diameter and 7.6 cm in length) into a vertical column of a gelifluction lobe in October, 1963 and recovered them in August, 1967 (Benedict, 1970). Some gelifluction deposits retain their original shapes as they proceed downhill; these lobate features may even coalesce to form pseudo-terraces. Therefore, gelifluction lobes may become confusing to interpret in the field without the assistance of Geographical Information Systems. Detailed LiDAR analysis allows for careful aerial inspection of the surface to ascertain the feature’s large-scale characteristics. Aerial photographs of Niwot Ridge show many gelifluction lobes in a downhill procession (Fig. 18). The magnitude of velocities of the layer of moving soil on Niwot Ridge was comparable to similar processes in other alpine areas. In the 1960s and 1970s, velocities ranged from 0.4 cm/year to 4.3 cm/year within the upper 50 cm of soil (Benedict 1970; Benedict, 1975). Figure 18. Soil profile of turf-banked terrace lobe 499 on Niwot Ridge. 1 – Stones, drawn to scale. 2 – Dark grayish brown gravelly sandy loam. 3 – Dark grayish brown to yellowish brown sandy loam, becoming finer toward the front of the lobe. 4 – Rose diagrams summarizing the long-axis orientations of 200 sand grains, measured in thin section representing frequencies of 10% (image from Benedict, 1970, Figure 21, p. 183) Clast orientation measurements reveal details about depositional environments and fluvial, glacial, and mass-movement processes, providing useful information for identifying mechanisms of movement and emplacement. On Niwot Ridge, clast measurements were obtained from several paired horizontal and vertical exposures in turf-banked gelifluction lobes (Millar and Nelson, 2001). Most samples were found to form moderately strong, upslope-plunging clusters aligned with the local slope orientation (Millar and Nelson, 2001). Data from Niwot Ridge represent a series of composite samples of clasts affected by orientation mechanisms which vary in nature and intensity over the vertical extent of the excavation pits (Millar and Nelson, 2001). The range of particle velocities in vertical profiles through turf-and stone-banked lobes and terraces on Niwot Ridge was found to be larger than those measured laterally across the surface of each feature (Millar and Nelson, 2001). These relationships agree with Benedict (1970)’s velocities measurements of a cross section (Fig. 17a) on a solifluction lobe of Niwot Ridge moving faster than those measured in plan-view (Fig. 17b). Regions of Niwot Ridge where snow remains on the ground through the winter, and is not blown away by fierce winds, become saturated as spring and late-summer snow thaws. Significant gelifluction occurs only in areas where the water table remains sufficiently high during the fall freeze-thaw cycle to permit thick ice-lens development (Benedict, 1970). Both seasonal and weather-driven temperature variations cause temperatures to drop below freezing, causing freeze-thaw cycles during both the spring and summer. Leopold et al. (2008) suggest that on Niwot Ridge seasonal ice lenses may form beneath discrete gelifluction lobes instead of permafrost. These lenses may be overlain by water ponded on the surface of low-permeability layers in the gelifluction lobes (Leopold et al., 2008). Benedict (1970) proposed two generations of turf-banked lobe and terrace development (Fig. 19) across the Colorado Front Range based on topographic differences and the presence or absence of patterned ground (see “Climate” section). Paleo-Periglacial Conditions Regions that experienced paleo-periglacial conditions may now be forested due to warmer temperatures and lower precipitation. These warmed regions become covered in vegetation and often restrict the movement of regolith downhill since trees trap blowing snows and protect the ground from extreme temperature variations (Washburn, 1970; Marker, 1990). On the other hand, at lower elevations such as Gordon Gulch, where erosion has overwhelmed small amounts of freeze-thaw action for the past several thousand years, the hill slopes are likely to be smoother and contain material that is slowly being transported downhill. The results of periglacial processes and deposits in the alpine and in the subalpine regions of Niwot Ridge are clearly visible by geophysical methods as anomalies within otherwise consistent terrain. Since frozen water has such a high resistivity, ERT is an excellent tool to map permafrost in both arctic and alpine regions (Fig. 20) (Cross Borehole Electrical Resistivity Tomography, 2003). Combinations of several types of periglacial processes are able to rework the subsurface so that many meters of regolith might have been created above solid bedrock. and ERT monitoring line (bold). b. Inverted image of ERT measurement from August 2006 showing active layer (red and orange) overlying an ice block (blue) (image from Hilbich et al., 2009, Figure 1, p. 272) Sorted Circles A variety of patterned ground morphologies are found in periglacial environments around the world, where geometric or repeated patterns develop naturally on the ground surface (Anderson and Anderson, 2010). The best examples of patterned ground occur in remote areas above treeline in alpine mountain regions and in the high Arctic (Ray et al., 1983). Widespread sorted circles occur in the Seward Peninsula, Alaska, the Laramie Range in Wyoming, Signy Island off the coast of Antarctica, and various mountain systems in Norway (Fahey, 1975). Causes of patterned ground formation include desiccation cracking, seasonal frost cracking, permafrost cracking, and freeze-thaw action (Ray et al., 1983). In particular, sorted circles may form from the development of convection cells, which provide a mechanism for the occurrence of regularity in the active layer (Ray et al., 1983). On Niwot Ridge and in Green Lakes Valley, sorted circles are prevalent in exposed snow-free locations, particularly frontal areas of turf-banked lobes and areas which are under temporary lakes during the winter. Wind speeds up to 11 m/s during December and January scour and remove the protective mat of alpine turf, allowing the surface to fall below freezing for many months of the year (Fahey, 1975). During these cold months, the fine silt centers form numerous small ice lenses that heave significantly upon freezing. The magnitude of the frost-heave is greatest in the center of the circle, where the fine-grained layer is thickest, which causes the outward slope to steepen over time. The coarse grained perimeter does not heave as significantly, allowing the coarse particles to remain mainly in place through the winter. As the thaw progresses downward from the surface during the spring and summer, the soil becomes weak when the ice lenses melt, causing the soil to creep radially down the slightly steepened slope. The coarse border rolls inward as the base of this heaved material is sheared away, ejecting coarse stones from the fine matrix as ice melts. After hundreds of freeze-thaw cycles, patterned ground remains with coarse stones on the border with no means of returning to the middle of the circle (Fig. 21) (Anderson and Anderson, 2010). Diameters of sorted circles in Green Lakes Valley normally range between 0.5 and 1.0 m, although occasionally two or three individuals have coalesced to form composite units up to two meters in diameter. compatible with sorted circle formation (image modified from Anderson and Anderson, 2010, Figure 9.16, p. 286) Active sorted circles were documented over a wide area during the 1970s in the alpine tundra of the Colorado Front Range, especially on Niwot Ridge and in Green Lakes Valley (Fahey, 1975). The sorted circles in Green Lakes Valley remains under a shallow extension of Green Lake 4 through the winter, allowing frost heave to easily push cobbles and boulders to the circle’s exterior while keeping sand and gravel in the center. A limited amount of silt in the system preserves the sorted circles from cracking when water is drained during the spring (Fig. 22). The distinct sorting and sharp microtopography of active sorted circles must be maintained or else needle ice, wind, gravity, and erosion smooth the features away after thousands of years. Both freeze-thaw cycles and diurnal frost heave decrease with increasing elevation at snow-free sites on Niwot Ridge, limiting the elevation range at which sorted circles can form. Diurnal frost heave cycles seldom penetrated deeper than 10 cm, even in sorted circles above treeline, allowing diurnal sorted circles to form with diameters ranging from 10-50 cm (Fahey, 1975). (image from D. Dethier, personal communication, January 9, 2013). b. Sorted circles 1-2 m in diameter which developed in raised beaches of Svalbard. (Photo by Anderson and Anderson, 2010, Figure 9.15, p. 285) Rock Glaciers Rock glaciers are completely covered with rubble or regolith below their accumulation zone and move at rates of centimeters to decimeters per year (Bierman and Montgomery, 2013). This cover provides significant insulation to keep temperature oscillations at a minimum (Fig. 10). Rock glaciers may be composed of massive layers of ice beneath their surface, or interstitial filling of ice between rubble fragments. These periglacial features extend outward from cirques or cliffs and are elongate downslope. While active, they often have lobate margins and steep termini as they move down-slope, deforming under the force of gravity (Bierman and Montgomery, 2013) Knowledge of the internal structure of rock glaciers is limited due to the difficult nature of drilling vertical boreholes through unconsolidated rubble and ice. In addition, access to active rock glaciers is limited, and undisturbed cross sections are rare. In recent years geophysical methods such as ERT, GPR, and SSR have allowed subsurface information to be acquired on rock glaciers without extensive disturbance of the rock glacier itself (Hilbich et al., 2009). Information about the internal structure of rock glaciers is useful for better understanding of how rock glaciers are reacting to ongoing climate change, since surface snow, ice fields, and mountain permafrost are beginning to thaw earlier in the season (Harris et al., 2003; Leopold et al., 2010). Many studies conducted on rock glaciers have focused on their movement, origin, and climatic significance, while few studies have focused on their enclosed ice. Reletively little known about how meltwater flows through active rock glaciers, and about the geochemical content of the water that drains from the rock glaciers (Williams et al., 2006). Giardino et al. (1992) remarked that rock glaciers physically and chemically alter the water which passes through them, acting as a concentrating mechanism rather than a filtering device for suspended ions. Hydrological analysis of the rock glacier near Green Lakes 5 suggests that snow is the dominant water source in June, melted water trapped in soil pores is the prominent source in July, and internal ice melt is the dominant source in September (Leopold, 2011). Williams et al. (2006) characterized the geochemical and isotopic content in the outflow of RG5, focusing on the 2003 season from the initiation of snow melt to the start of the next snow accumulation season. Because the bulk snowpack temperature remains below 0 °C until late spring, there is a significant temporal lag in the hydrological cycle due to concentration from the release of snow melt into a short intense period of runoff (Caine, 1996). Between June and August the geochemical signature of outflow from RG5 did not substantially differ from that of the surrounding surface waters in the Green Lakes Valley. However, while most of the snow was melting adjacent to the rock glacier 2+ 2+ in August and September, concentrations of Mgand Cain the outflow of RG5 increased significantly from other measurement sites (Williams et al., 2006). These results led Williams et al. (2006) to postulate that GL5 contains an internal ice core surrounded by interstitial ice within a coarse debris mantle. Geophysical interpretation of this rock glacier resulted in a four layer model (Fig. 23); fine sediments and soils at the surface (down to 2-3 m depth) overlie a zone of coarse debris containing large air-filled voids. Between 1-3 and 4-5 m depth the material changes to finer and wetter sediments, which are unfrozen in late summer, representing the lowest part of the active layer during recent years. Below 4-5 m, GPR and ERT data suggest a zone within the rock glacier that contains high ice content, possibly a solid ice core during the winter and early spring. Bedrock was detected at a depth of 16-18 m, consistent with its exposure on the adjacent valley floor (Leopold, et al., 2011). Janke (2008) hypothesized that rock glacier temperature information can be used to adequately represent the distribution of permafrost in the Front Range, and that rock glaciers could provide a unifying criterion for permafrost models. Because of both the amount of insulation above internal ice cores, and the slow rate at which rock glaciers progress downhill, it is possible to infer the climatic influence on permafrost by studying the thickness of ice cores within rock glaciers. The rock glaciers within Green Lakes Valley will most likely remain active far longer than the gelifluction lobes atop Niwot Ridge, both because of their minute insolation exposure and thickness of trapped ice (Williams et al. 2006; Janke, 2008). In my study, I examine two rock glaciers situated in similar locations in the Front Range of Colorado: the RG5 rock glacier near Green Lakes 5 in Green Lakes Valley and a rock glacier near the head of 4th of July Valley. Using ERT lines both on and in front of these rock glaciers allows us to visualize the subsurface of each feature, while water temperature measurements from outlet streams provide temperature information that helps us reconstruct subsurface thermal conditions. Using photographic evidence and descriptions from Benedict (1981), I could monitor developments in the extent and topography over the past 30 years. SETTING Location and Topography All four of my field sites are in the upper part of the Boulder Creek catchment within Boulder County, Colorado, in the Boulder Creek Critical Zone Observatory (CZO) (Anderson et al., 2008). The basin covers 1160 km2, stretching eastward 50 km from the Continental Divide to the Piedmont, and ranging in elevation between 4120 and 1480 m (Anderson et al., 2006). I specifically worked in four smaller drainages within this watershed: Green Lakes Valley, Niwot Ridge, Gordon Gulch, and the 4th of July Valley (Fig. 24). Figure 24. a. LiDAR image of the Boulder Creek CZO showing the location of 4th of July Valley, Green Lakes Valley, Niwot Ridge, and Gordon Gulch. The outline of Gordon Gulch is shown in blue, the sixteen Electrical Resistivity Tomography lines are shown in red, and the background is a LiDAR hillshade from Anderson et al., 2012. Red outlines are the insets for Niwot Ridge ERT lines (Fig. 24b), Gordon Gulch ERT lines (Fig. 24c), and 4th of July ERT lines (Fig. 24d) The Boulder Creek CZO extends from the peaks of the Continental Divide down to the Colorado Piedmont, draining water into the South Platte River and downstream into the Mississippi River system. Tectonic quiescence since the end of the Laramide orogeny (~40 Ma) has allowed erosion and weathering to produce the different Critical Zone architectures exposed at Gordon Gulch, Green Lakes Valley, Niwot Ridge, and 4th of July Valley (Anderson et al., 2006). Many periods of glacial advance and retreat during the past million years have carved and shaped valleys in the Front Range and deposited material from the eastern slopes onto the Colorado Piedmont (Anderson et al., 2006). Approximately 14,000 years ago, late-Pleistocene glaciers retreated from the Front Range after following U-shaped valleys close to the Continental Divide. As glaciers sculpted the landscape, periglacial features such as gelifluction lobes and sorted circles modified areas adjacent to glacier ice such as Niwot Ridge and those lower in elevation such as the currently forested areas within Gordon Gulch. As these periglacial features were forming, rivers in the Front Range episodically switched from depositing alluvium to incising the surface they pass through, eventually lowering the plains by hundreds of meters (Anderson et al., 2006). As a result, rivers have carved deep canyons with steep knickpoints along the eastern edge of the Front Range. A wide range of glaciated landscapes exists within the Front Range as a result of repeated glaciations through the Pleistocene, most evident in the deeply carved alpine valleys near the Continental Divide. Throughout western North America, many alpine glaciers and ice caps have undergone multiple phases of advance and retreat due to oscillating climate conditions. The Front Range is no exception; glaciers dominated during pre-Bull Lake time (500-300ka), Bull Lake (200ka-130ka), Pinedale (23-12ka), and the Neoglacial periods (4.5-2.7ka and 1.9-1ka) (White, 1971; Mears, 1981). Numerous studies have focused on dating glacial moraines and modeling climate through the Pleistocene (Marker, 1990; Mears, 1981). Although there is little agreement among paleo-climate models, most studies focus on temperature and precipitation during either the late Pleistocene or within the Holocene. Niwot Ridge, Green Lakes Valley, Gordon Gulch, and 4th of July Valley are a series of slopes and surfaces just east of the Continental Divide. The Boulder Creek Critical Zone Observatory is responsible for Niwot Ridge, the Boulder Creek Watershed owns Greens Lakes Valley, and the National Forest Service manages Gordon Gulch and the 4th of July Valley in the Indian Peaks Wilderness (Birkeland et al., 2003; Leopold et al., 2010). Each of these areas has interesting and highly variable topography, climate, and periglacial features. Niwot Ridge Niwot Ridge is an east trending spur on the east side of the Continental Divide, 35 km northwest of Boulder, Colorado (Fig. 25). The alpine zone along the ridge is approximately 5 km long and 1 to 1.5 km wide, covering approximately 7.85 km2. Niwot Ridge extends east from the Continental Divide from 3800 to 3450 m and includes several knolls that stand above the ridge along the central axis (Berg, 1986). Tree line is located between 3350 and 3500 m along the ridge, with the exception of many patches of krummholz located as high as 3550 m (Greenland, 1989). The subalpine climatic zone is dominated by subalpine firs and Engelman spruce, which decline in density with increasing elevation (Birkeland et al., 2003). Above tree line (approximately 3400 m), steep rock faces and talus slopes cover two thirds of the area while the rest is composed of alpine tundra vegetation growing on regolith (Caine, 1995). Nearly all of the ridge’s bedrock is hidden by a thick mantle of periglacial deposits typical of many other alpine ridge systems in the Front Range, with exposed outcrops only on the southern edge of the ridge, (Benedict, 1970). Since only four months out of the year have average temperatures above freezing, much of the area is rocky and most of the vegetation above tree line consists of grasses, sedges, and low perennial herbs (Greenland, 1989). Marr (1961) refers to the local tundra as “a mosaic of stand-type ecosystems” where individual units are often less than 1 m2 in area. Stunted Engelmann spruce (Picea engelmanni), subalpine fir (Abies lasiocarpa), and limber pine (Pinus flexillis) are able to persist to elevations above tree line in small clumps (Benedict, 1970). The Fahey Site on Niwot Ridge was the location of B.D Fahey’s doctoral thesis (see Ives and Fahey, 1971); it consists of a 50 meter wide by 70 meter long turf-banked gelifluction lobe with partially patterned ground and ponds which persist throughout the summer (Fig. 26). Patterned ground may be inactive, but fresh mud boils appear in the center of stone rings during the early melting season and do not display any characteristic large stones. The Fahey Site includes two cased boreholes with temperature loggers at depths down to five meters (both of which experienced technical problems during the summer of 2012). Using thermistor measurements, Ives and Fahey (1971) and Ives (1973) examined the extent of permafrost on Niwot Ridge and reported that they found scattered permafrost beneath a two meter active layer at elevations above 3500 m. Areas containing permafrost had been blown free of snow during the winter, but nearby snowbanks provided moisture during the summer and maintained ground temperatures near the -1 °C MAT isotherm. These two studies also indicate that they found extensive wet permafrost at elevations from 3750-3800 m and continuous permafrost sometimes exceeding 60 m thick above 4400 m (Ives and Fahey, 1971). From current knowledge about permafrost extent, it is likely that many south-facing and dry regions on Niwot Ridge below 3800 m did not actually contain permafrost in the 1970s, but rather Ives and Fahey were measuring seasonal ice lenses during the summer which had not yet melted. The temporal lag between surface temperatures and subsurface temperatures, discussed in the “Heat Flow” section above, allow ground at depths of 5-10 m to remain frozen well after surface air temperatures have risen above 0 °C. Green Lakes Valley A high-relief zone consisting of major glacially cut steps and flats extends eastward from the Continental Divide. This zone was eroded during the Pleistocene, with surficial deposits accumulating since deglaciation during the mid-Holocene (Williams, 2006). Cirques ornament the head of each alpine valley and hide in the steep shadows just below the Continental Divide. This high-relief region contains both Green Lakes 22 th Valley (total area of 7.1 km, of which 2.2 kmare above Green Lakes 4) and 4of July Valley (total area of 27.4 km2, of which 2.98 km2 are above ERT lines 12, 13, and 14) (Benedict, 1981; Benedict, 2005). Green Lakes Valley (GLV) is located just south of Niwot Ridge and ranges in elevation 4808 to 3515 m. This glacially carved valley faces eastward and is named for a series of shallow paternoster lakes that are the headwaters of North Boulder Creek (Birkeland et al., 2003; Williams, 2006). The floor of GLV is a series of lakes separated by steep steps and is bordered by narrow arêtes between glaciated cirques (Williams, 2006). Below the steep slopes, the alpine slopes fit the cliff-talus-sub-talus model, which is a factor that explains the coarse rock-dominated debris system proposed by Caine (1974). Green Lakes Valley contains the Arikaree glacier (surface area of 9-ha), which drains into a series of five lakes and eventually into North Boulder Creek. The valley displays evidence of past glaciations, but, given present climate, no large glaciers should occupy a significant portion of the valley (Dühnforth and Anderson, 2010). Although mass balance measurements from 1968 to 1998 show that the balanced net budget has been consistently negative, the Arikaree Glacier is most likely preserved due to its shaded location in an east-facing cirque near the Continental Divide. The glacier’s surface has been lowered by up to 6.8 m during that time period (Caine, 2006). 4th of July Valley The 4th of July Valley is located at the head of the North Fork of Middle Boulder Creek and is rimmed by the Continental Divide to the north and west and a steep arête to the south. This valley drains from 3745 to 3200 m in elevation along its 50 km eastward trend down to the Colorado Piedmont (Benedict, 2005). The valley is eroded into Precambrian igneous and metamorphic bedrock, where glacial and periglacial processes have molded its topography (Benedict, 1981). Bounded to the north and west by the Continental Divide, the 4th of July Valley includes Mt. Neva (3906 m) and South Arapaho Peak (4803 m). Between these two peaks lies Arapaho Pass, a bedrock and tundra covered saddle that in prehistoric times was the most heavily-utilized east-west travel route for Native Americans in an 18 km section of the Continental Divide (Benedict, 1981). Within the 4th of July Valley, moraines, rock glaciers, and outwash terraces, as well as manmade artifacts, in the timberline ecotone were mapped and dated in order to correlate the presence of human activity with the advance of several rock glaciers (Fig. 27) (Benedict, 1981). Most of the course of the valley has a deeply cut “U”-shape in profile, with high valley-side benches. Erratic boulders and lateral-moraine remnants record maximum ice thicknesses of 325-250 m. Longitudinally, the 4th of July Valley is a series of broad steps, separated by steeply-sloping risers. The study area and major pre­historic campsites are located on the second-highest step (elevation of 3460 m) in the forest ecotone. Small-scale periglacial landforms are common within the 4th of July Valley. Late lying snowfields are covered with close-packed boulders, heaved to the surface by frost during spring freeze-thaw cycles. Sorted nets are visible where boulder pavements are intruded from beneath plugs of fine-textured sediment, and gelifluction lobes are present on saturated slopes (Benedict, 1970; Benedict, 1981). Although the sorted nets and gelifluction lobes were present within 4th of July Valley during the 1970s, the lack of active periglacial activity on Niwot Ridge suggests that the sorted nets and gelifluction lobes are inactive today. Figure 27. Late winter view of the Fourth of July Valley looking eastward towards four lobate rock glaciers (image from Benedict, 1981, Figure 77, p. 95) Gordon Gulch West of the knickzone along Boulder Creek is a rolling surface of low-relief that extends from 2800 to 2300 m over a 10-20 km stretch (Dethier et al., 2003). The surface exemplifies deep weathering, with an average regolith thickness of 3.3 m and denudation rates of ~20 µm/year (Anderson et al., 2006). The erosion rate locally exceeds the rock weathering rate in the gulch and exposes isolated bedrock outcrops as soil is removed. Gordon Gulch, a 2.74 km2 region ranging in elevation from 2737 to 2446 m, exemplifies these high denudation rates on its prominent north and south facing slopes (Fig. 28). Geologic Background Bedrock To understand and interpret the periglacial features on Niwot Ridge and in GLV, an understanding of the local geology of the region must be established. A brief review of the rocks and landforms of the Estes Park and Denver West 30’x60’ quadrangles allows better comprehension of the processes that create each of the presently exposed periglacial features. These two quadrangles contain an exceptionally complete record of geologic history in the northern Front Range of Colorado (Cole and Braddock, 2009). The Laramide Orogeny (~64 to 45 Ma) was the most recent major tectonic event affecting the Colorado Front Range, and resulted in exhumation of the Precambrian crystalline bedrock in the core of the range (Befus et al., 2011). This mountain building event uplifted the Rocky Mountains, allowing differential erosional processes to produce the Front Range geomorphology observed today (Anderson et al., 2006). Niwot Ridge is primarily composed of cordierite-and magnetite-bearing sillimanite-biotite gneiss (Xgns, Xgnc), biotite-bearing schists and gneisses (Xb) that formed before 1,700 m.y. ago, and the intrusive Long’s Peak monzogranite (Ysp) that intruded 1,440 m.y. ago (Fig. 29) (Gable and Madole, 1976). The dominantly metasedimentary gneisses are quite variable in composition, dark-grey to strongly banded black-and-white, and contain fine-to medium-grained layers of gneiss and schist. The layers are generally 1-30 mm thick with magnetite content averaging 4% in the gneiss, around 8% in weaker schist layers, but as high as 15% in adjacent Tertiary intrusive rocks. The Long’s Peak monzogranite is exposed above Green Lakes 5 to the Continental Divide. It is light to pinkish gray, medium-to coarse-grained equigranular to moderately porphyritic rock characterized by aligned tabular potassium feldspar grains. This monzogranite occurs as both discordant and small lenticular bodies, ranging from dikes and sills to stocks and batholiths. Its composition is granite, with microcline, quartz, and plagioclase as well as small amounts of biotite and muscovite. The whole-rock Rb-Sr and U/Pb isochrons of the Long’s Peak granite are about 1400-1440 m.y. (Hills and Dickinson, 1982; Gable and Madole, 1976; Williams, 2006). In addition to the gneisses and Long’s Peak Granite, Niwot Ridge is also underlain by bodies of Tertiary granodioritic biotite-quartz porphyry (Tgd), quartz monzonite (Tqm), and syenite (Ts). The granodiorite is fine-to medium-grained grayish brown material, which becomes stained by iron oxide due to weathering. The quartz monzonite resembles the lighter phases of the Long’s Peak monzogranite and forms as dikes containing prominent plagioclase phenocrysts throughout the quadrangle. Syenite within the region is light-grey to pinkish-gray, fine-grained to very coarse grained rock that forms large bodies and dikes on the eastern flanks of the Rocky Mountains. Quartz-bearing syenite accounts for over one-quarter of the Ward 7.5 minute Quadrangle, where Niwot Ridge is located (Fig. 30) (Gable and Madole, 1976). Green Lakes Valley is primarily underlain by the Precambrian cordierite-and magnetite-bearing sillimanite-biotite gneiss, with layers of cordierite-bearing garnet­sillimanite-biotite gneiss and Long’s Peak Granite. Small outwash gravel, block fields, and nivation scree slopes cover the rest of Green Lakes Valley. A large area of gray, medium-grained monzonite underlies the region immediately north and east of Lake Albion, while Pinedale till covers most of the valley below that. Surficial Deposits Most of Niwot Ridge is composed of low, rounded summits and shallow saddles covered by a thick mantle of periglacial deposits, which reflect the influence of prolonged mass wasting upon structurally complex igneous and metamorphic basement (Benedict, 1970). A large percentage of the ridge is covered by periglacial deposits, including gelifluction lobes (Qs), turf-banked lobes, terraces, hummocks, and patterned ground, all of which were active from the Quaternary (Fig. 29) (Gable and Madole, 1976). The cobbles and pebbles rest within a poorly sorted fine-grained matrix consisting of rubble, sand, and fines crushed from bedrock. These features are derived from weathering of the underlying bedrock described above and contain material that was transported only about one hundred meters during the Holocene (Gable and Madole, 1976). The upper 10-20 cm typically consists of roots and an organic A horizon, covering a thick layer of cobbles and pebbles embedded in a poorly sorted fine-grained matrix of sand and fines derived from bedrock and dustfall (Fahey, 1975). Geophysical methods and drilling data show that depth to bedrock ranges from 4 to over 10 m on Niwot Ridge, and that deposit thickness is not simply related to the local slope (Leopold, 2008). Climate Understanding the past and present climate at a particular location allows future climate models to be interpreted in order to predict biological and geomorphic variations. Details of paleoclimate can be understood by studying the present day climate, which can yield reasonable estimates of climatic variation over the past 20,000 years. Data from the last century’s climate reveal information about the paleoclimate which are important to understanding periglacial activity since the LGM. Throughout the Holocene, the annual precipitation, temperature, and mean solar radiation have varied considerable, causing severe changes in the elevation of periglacial activity and the ELA (Greenland, 1989). Studying the variation in these same variables due to global warming can provide clues as to how permafrost and periglacial features reacted over the past 20,000 years. Historic Climate Temperature and precipitation values vary greatly throughout the Boulder Creek catchment due to the 2500 m gradient between the glaciated Continental Divide and the alluvial Colorado Piedmont. The continental climate regime exerts strong diurnal and seasonal variations in temperature and allows for a steep gradient of vegetation to grow within the catchment (Fig. 31). Niwot Ridge has one of the most accurately studied climates of any research area in the world dating back to governmental funding from the 1950s. The CZO provides a perfect location to examine both the magnitude and variation in temperature and precipitation during the 20th and 21st centuries (Greenland, 1989). Measurement of the temperature and precipitation along Niwot Ridge and within Green Lakes Valley for over half a century provide a data set for the Boulder Creek CZO that is unparalleled in both quality and quantity with other CZOs from around the world. Figure 31. Vegetation zones and climatic data for a transect of the Front Range from the Continental Divide to the Colorado Piedmont near Boulder, CO, showing the areas investigated in this study. Elevation ranges are typical for the vegetation zones shown, where gaps between elevation ranges represent ecotones between the zones (image modified from Birkeland et al., 2003, Figure 5, p. 85) Climate on Niwot Ridge and Adjacent Areas The MAT near the Tundra Research station (3528 m) near the Fahey Site is -2.13 °C whereas MAT near the metrological station D1 (3739 m) is -3.7 °C, with a local lapse rate of approximately 7.1 °C/km (Greenland, 1989). At the Fahey site, MAP ranges from 1000-1100 mm/year, while at D1 it averages 1226 mm/year, of which 80% falls as snow (Leopold et al., 2010; Barry, 1973; Williams et al., 1996; Caine, 1996). A map of the D­1 Meteorological Station, Saddle Research Station, and the Albion Meteorological Station is displayed below for reference (Fig. 32). 1 Meteorological Station, Saddle Research Station, and the Albion Meteorological Station Niwot Ridge, Green Lakes Valley, and 4th of July Valley have high amounts of wind and precipitation, with moderate to heavy winter accumulation of wind-drifted snow. Many of these snowbanks persist throughout the summer and fall, supplying meltwater to tiny streams which are important to the ecologic diversity of this region. Wind velocities are predominantly strong from the west or northwest, with an average annual velocity of 10.3 m/s on Niwot Ridge (Barry, 1973). Winter snow builds dunes and corniced drifts which cover almost the entire valley, with the exception of treetops and the tops of rock glaciers (Benedict, 1981). Mid-continental locations such as the study sites have historically recorded large average temperature and precipitation variations between summer and winter, especially at high elevations (Greenland, 1989). Summer air temperatures measured at the D1 station on Niwot Ridge have shown nearly a 2 °C rise since 1950 with decreased variability in winter temperatures (Fig. 33) (Leopold et al., 2010). Since the 1950s, annual precipitation has increased by ~ 300 mm while the daily mean solar radiation has decreased by almost 100 W m-2 (Fig. 34) (Williams et al., 1996). Although these fluctuations are small compared with those on timespans of hundreds to thousands of years, the rate at which temperature and atmospheric gas concentrations are changing is unprecedented. Mean Annual Precipitation (mm) 1800 1600 1400 1200 1000 800 600 Year Figure 34. Timeline of the recorded (solid line) and reconstructed (dashed) mean annual precipitation at D1 station from 1965 to 2010. Recorded data obtained from culter.colorado.edu. Estimated precipitation from Greenland, 1989 Rising summer temperatures (Fig. 33) have contributed to an increase in late-summer discharge of high-elevation catchments and suggest that permafrost is melting from underneath talus slopes and rock glaciers (Janke, 2012). Although the effects of rising summer air temperatures due to global warming are likely constrained on north facing slopes, melting permafrost within the Boulder Creek Watershed has influenced surface geomorphology, water chemistry, and biogeochemistry over the past decade (Caine, 2010). For example, Leopold et al. (2010) documented complete melting of an ice lens within a gelifluction lobe on Niwot Ridge at the Fahey Site, which had been thought to contain permafrost for nearly 50 years. Caine (2010) notes that more than half of the apparent increase in annual discharge and almost all of the increase in late season streamflow from the alpine region of the Green Lakes Valley in the past 30 years may have been due to the melting of permafrost and ice trapped beneath the GL5 rock glacier. Meltwater from permafrost and rock glaciers can alter the biogeochemical signature because the wet and warm summer climate facilitates the mobility of solutes from weak areas, which provide material and transport for mass wasting (Gooseff, 2009). Williams et al. (2006) presented data on a disturbance in the sensitive hydrochemistry of meltwater from RG5. He observed 2+2+.. extremely high values of Mg, Ca, and ... during the late fall, resulting in a 4-5 fold increase in the magnitude of solute release from the snowpack in the form of an ionic pulse and causing episodic acidification of the meltwater in headwater catchments (Williams et al., 1996). Climate in 4th of July and Green Lakes Valley Climate in the 4th of July and Green Lakes Valley is similar to that on Niwot Ridge due to their proximity and similar characteristics. By extrapolating data from Niwot Ridge, I infer that 4th of July Valley has a MAT of -3.3 °C and MAP between 1000 and 1200 mm at an elevation of 3745 m (Marr, 1967). Benedict (1981) describes 4th of July Valley as “cool, moist, and windy,” (p. 7) although wind velocities are not as high as those on Niwot Ridge due to the orientation of the valley’s topography as well as protection from large vegetation. Windblown precipitation creates moderate to heavy winter snow drifts covering the valley floor and all but the tops of trees, rock glaciers, and the highest bedrock outcrops. Many of these snow drifts persist through the summer and into the fall, providing cold meltwater underground and providing ample insulation to keep the ground cool enough for permafrost to exist in the subsurface (Benedict, 1981). Climate in Gordon Gulch The Gordon Gulch climate contrasts with that recorded on Niwot Ridge although they are located only 12 km apart. Situated within the upper montane climatic zone, MAP at Gordon Gulch is ~519 mm, most of which falls as rain (Barry, 1973; Hinckley et al., 2012). The steep north-facing slope of Gordon Gulch maintains a MAT of 5.1 °C, while the south-facing slope may be slightly warmer due to solar radiation. With much warmer temperatures than Niwot Ridge, Green Lakes Valley, and the 4th of July Valley, most of Gordon Gulch is dominated by lodgepole pine (especially on the north-facing slopes) as well as ponderosa pine and Douglas fir (which dominate the south-facing slopes). Toeslope areas support aspen groves and moist meadows at the bottom of the gulch as well as a seasonal stream that carries regolith away from the center of the gulch during large precipitation events. Paleoclimate The history of glaciation and periglacial deposits throughout the Colorado Front Range is Range is still a highly debated topic among paleoclimatologists. I present a of the most current summary of climate information through the last 20,000 years ( Table 5). Accurate regional precipitation and temperature data are needed to help reconstruct permafrost and periglacial activity during these periods. Table 5. Estimated altitude of the equilibrium line in Green Lakes Valley during the past 20ka Calendar Years BP (ka) Inferred ELA (m) Source of Information 0 4000 Ward et al., 2009 0.2 3980 Ward et al., 2009 0.4 3980 Dühnforth and Anderson, 2011 0.38 4020 Vierling, 1998 0.9 4020 Elias, 2001 1 4050 Mears, 1981 1.2 4100 Benedict, 1970 1.6 3900 Madole, 2012 1.7 3900 Mears, 1981 1.8 4000 Lyle et al., 2012 2.4 4000 Dühnforth and Anderson, 2011 3.77 3990 Markgraf and Scott, 1981 5.1 4100 Ward et al., 2009 5.16 4100 Dühnforth and Anderson, 2011 7.42 4200 Vierling, 1998 12.7 3700 Marker, 1990 14.7 3575 Marcott, 2012 20 3300 Ward et al., 2009 The paleoclimate community uses small alpine glaciers as sensitive indicators of climate variations, allowing glacial geologists to help reconstruct a Holocene and Pleistocene climate history in the Front Range (Dühnforth and Anderson, 2011). Most data coincide with rapid climate events from the North Atlantic oscillation, emphasizing the importance of local-to-global climate variability that contributes to glacier mass balance changes (Marcott et al., 2012). Although dates and ELA values are still being discussed within the paleoclimatic scientific community, it seems that many events in the Colorado Front Range were synchronous with other glaciers across western North America. Latest Pleistocene Ward et al. (2009) used cosmogenic radionuclide exposure ages from polished and striated bedrock to constrain numerical simulation of deglaciation through all of Middle Boulder Creek Valley during the Pleistocene. Using absolute dating of moraine boulders by cosmogenic 10Be exposure dating in collaboration with a set of 1-D and 2-D numerical glacial models, Dühnforth and Anderson (2011) created a model of the past 20,000 years of glacial history specifically within Green Lakes Valley (Fig. 35). Although the timing and magnitude of early glacial advances are still not precisely understood, the early Pinedale glaciation was most likely more extensive than the late Pinedale in the Boulder Creek Watershed (Elias, 2001). During the LGM, Pinedale glaciers flowed from the mountain cirques at the continental divide 14 to 15 km downslope to elevations between 2400 and 2470 m (Elias, 2001). Pinedale glaciers in the Colorado Front Range were relatively thin and small compared with those in Wyoming and Canada, preventing them from coalescing and forming larger glaciers or ice sheets. The first two Pinedale glacial advances occurred during the Wisconsin stage of the Pleistocene glaciation and occurred in valleys both to the north and south of Niwot Ridge. Although neither Pinedale glaciation covered the top of Niwot Ridge, associated temperatures likely coincided with permafrost and periglacial activity. Both Ward et al. (2009) and Dühnforth and Anderson (2011) suggest the initiation of deglaciation around 18-20 ka and completion by 12-13 ka, with a slight re-advance between 16 and 14 ka. Because of their size and thickness, Pinedale glaciers in Green Lakes and 4th of July Valley melted rapidly, left behind great piles of rubble and fine-grained sediments, and released torrents of sediment-choked water into Boulder Creek. After the main period of retreat, Dühnforth and Anderson (2011) hypothesize that an additional moraine formed between 15-17 ka, which could represent a short re-advance during deglaciation. According to Ward et al. (2009), glaciers extended far into many valleys in the Front Range during the Pleistocene as a result of the equilibrium line altitude at approximately 3300 m. This cold period lowered the MAT at Gordon Gulch below 0 °C, likely creating marginal periglacial conditions for much of the late Pleistocene. Sub-zero temperatures within Gordon Gulch would have caused frost-cracking and regolith transport that rapidly damaged the parent bedrock and moved large amounts of material downslope. Several studies suggest that periglacial activity was formerly more extensive and tree line must have been at a lower elevation through most of the Pleistocene (Marker, 1990). Since recent glacial advances were confined to narrow zones along the north-facing walls within GLV and the 4th of July Valley, many paleoclimatologists have referred to the Pinedale glaciation as the last true valley glaciation to affect the Front Range (Elias, 2001). Colder weather could have provided suitable climate conditions for forming periglacial features before temperatures rose to their present levels. Some paleoclimatic reconstructions from the entire Rocky Mountain region indicate that the Colorado Front Range received less moisture than the Yellowstone region to the north and the San Juan Mountains to the south during the late Pinedale environment (Elias, 2001). Mears (1981) studied ice wedge-forming conditions in Wyoming’s intermontane basins and concluded that during the late Pleistocene, the North American periglacial zone extended up to 650 km south of the continental glaciers in the high basins of Wyoming. Periglacial conditions persisted on the high mountain-rimmed plains down to 40 °N due to the precipitation shadow from the cordillera. Mears (1981) estimates that the former Wyoming temperatures must have been 10-13 °C colder during the LGM than today, compared with only 6-8 °C colder in Colorado (Dühnforth and Anderson, 2011). Holocene Although it is not ideal, paleoclimatic information for the Colorado Front Range through the Holocene is much more accurate than that for the Pleistocene. By investigating details and changes of the climate, glacier equilibria, and vegetation, we can reconstruct a model of temperatures throughout Niwot Ridge that indicate the presence or lack of permafrost and periglacial activity. Climate fluctuated within the Colorado Front Range during the Holocene, indicating warmer-than-present summer temperatures and colder-than-present winter temperatures (Elias, 2001). On Niwot Ridge, the combination of low winter temperatures, high effective precipitation, and large snow-free expanses of tundra created an environment during much of the Holocene time that favored permafrost and periglacial processes with their associated deposits (Benedict, 1970). The soils that currently mantle the stable areas of the Colorado alpine zone must have developed during the Holocene as a result of various glacial and periglacial processes (Elias, 2001). The most recent glaciation is represented in the study area by numerous patches of till and vast expanses of striated bedrock (Benedict, 1981). Thermal insolation at the top of the atmosphere in the northern hemisphere was probably higher during the early Holocene (11,700-8000 cal yr BP) than at present, causing temperatures to quickly rise (Fig. 36). Periods of increased precipitation falling as rain, decreased snowpack, and an earlier onset of snowmelt runoff were common during this time period. During the Middle Holocene (8000-4000 cal yr BP) timberline in the northern Front Range reached a maximum elevation around 7420 cal yr BP and began to recede around 3770 cal yr BP due to warmer temperatuers with a wide fluctuation of precipitation (Madole, 2012). °N to glacial events in the Colorado Front Range (data from Benedict, 1970 and 1981). Also shown are the glacial periods throughout the past 30ka (image from Muhs and Benedict, 2006, Figure 10, p. 127) Vegetation provides a sensitive record of past environmental conditions, specifically for lower timberline elevation and ELA. Markgraf and Scott (1981) observed vegetation changes resulting from elevation shifts of the forest zones and used dating of radiocarbon from pollen to obtain detailed paleoclimatic information. They conclude that observed pollen changes result directly from elevation shifts of the forest zones, which in turn depend on temperature and precipitation. The summer monsoonal boundary is a major climatic margin which runs through central Colorado and deflects increased summer precipitation to central and northwestern Colorado. When this boundary is shifted to higher latitudes, it increases the amount of summer moisture in the Colorado Front Range. The southern border of Pacific storm tracks presently lies just north of Colorado, but Markgraf and Scott (1981) suggest that the shift in temperatures and precipitation around 10,000 yr B.P resulted from shifts of the summer monsoonal boundary. In addition, Lyle et al. (2012) concludes between 12,000 and 9,000 years BP an 800 km southward shift of the 8 °C isotherm due to the persistence of the glacial age ice bodies in North America caused higher precipitation (~140%) at lower latitudes (Vierling, 1998; Lyle et al., 2012). Strengthening of this summer monsoon pattern between 9000 and 4000 yr BP most likely included higher summer temperatures and an increase in the amount of precipitation falling as rain. These warm, wet summers probably raised the ELA throughout much of the Rocky Mountains. Vierling (1998) estimates that during this time period, there was 8-11 cm more precipitation during the year than today, causing glaciers to shrink past the present equilibrium. During the past 4,500 years, three smaller ice advances have caused small glaciers to form and deposit material on the slopes of many Front Range cirques. Little Ice Age moraines are present throughout the Colorado Front Range, but this short-term glaciation may not have produced many geomorphic or periglacial structures. Vierling (1998) suggests a shift towards a modern climatic regime circa 1800 yr BP, while studies of pollen suggests that vegetation types have remained constant during the last 2000 yr. METHODS This study uses many different field and lab techniques to ascertain the location and extent of permafrost, learn about the subsurface within a rock glacier, measure the temperature of surface water, and describe the periglacial deposits and landforms present today in the Colorado Front Range. Permafrost is difficult to locate in the field since it occurs completely within the subsurface. To determine the location of active permafrost and subsurface ice lenses, I worked with Dr. Mathias Leopold from the University of Western Australia and collected both ERT and GPR geophysical lines. The location of each ERT line was chosen to be representative of its surrounding area in order to ascertain subsurface measurements from a diverse region. After the data were collected, digital post-processing was performed at Williams College using several types of computer software. Both water and soil temperature data were also collected in the field to determine the depth of permafrost in several areas. Field ERT Working with Dr. Mathias Leopold from the University of Western Australia, I collected a total of 16 Electric Resistivity Tomography lines throughout Niwot Ridge (7 line), Green Lakes Valley (4 lines), 4th of July Valley (3 lines) and Gordon Gulch (2 lines) using the multi electrode system “4punktlight hp” from Lippmann Geophysikalishe Messgeräte (see www.l-gm.de). We used line consisting of either 25 or 50 electrodes to calculate two-dimensional DC resistivity tomography profiles. At each location we measured a 25 or 50 meter profile and hammered 30 cm conductive metal stakes into the ground at equal intervals (usually every 50 cm or 1 m) using a rubber mallet (Fig. 37). After noting GPS points at each terminus of the line using a Garmin eTrex, we measured the inclination every five meters along the line using a TruPulse™ 360° laser rangefinder manufactured by Laser Technology, Inc. Detailed surveying of each line using a tape measure, compass, and GPS included thorough descriptions including surface rock sizes, areas with or without vegetation, surface water, and surface periglacial activity. For ERT-4 on the southern wall of Green Lakes Valley, the 36° average slope made multiple measurements both impractical and unsafe. Instead, we used the average slope since it was relatively constant throughout the entire 30 meter line. After hammering stakes into the ground, we unraveled chains containing 25 electrodes and placed an electrode on each stake. After connecting the chains to the field computer and an electrical ground, we computed tests for contact resistance to determine if the stakes could conduct enough electricity through the soil. Contact resistances were generally between 0.15-0.6 k.m, indicating excellent coupling of the electrode to the ground. At ERT Lines 5, 6, 10, and 11 (Fig. 24), we had to squirt water on specific electrodes to lower the contact resistivity and increase the accuracy of the measurement in particularly rocky areas. After testing the equipment to double check for inaccuracies, we tested both Wenner and dipole-dipole arrays to increase penetration depth and resistivity accuracy. Using the program GeoTest© (see www.geotestinc.com/) on the field computer, I performed three resistivity measurements for each point. If these values were within a 3% failure limit, I averaged them for that resistivity point; if not we averaged eight data values. I used frequencies of 5 Hz and 0.1-10 mA for my measurements and recorded resistivities on the order of 50 to 2000 .m. The tests resulted in observation depths between 1.5 and 18 m with a resolution of ±10-50 cm. The results of Line 11 on Niwot Ridge suggest permafrost in the laterally continuous high resistivity layer (70 k.m) at a depth below 75 cm. To field check this value, we excavated a 150 cm deep pit to search for frozen ground on July 24 (Julian calendar day 206). In this pit, we measured soil temperatures at depths of 0, 20, 40, 60, 80, 100, 120, 140, and 150 cm to predict the depth of permafrost. Upon finding open­work gravel instead of permafrost within the pit, we performed a resistivity check to directly test the resistivity of the gravel. We spaced five stakes in a 50 cm line and tested the resistance between four of the stakes using the fifth as a ground (Fig. 38). For an accurate measurement, I averaged 150 measurements with a 0.02% failure limit. For several ERT lines on the 4th of July Rock Glacier, throughout Gordon Gulch, and on Niwot Ridge, we used a roll-along method to extend the measured distance at high resolution. At these locations, we measured subsurface resistivities between 0 to 49 m along the profile, then from 25 to 74 m after moving half of the electrodes to collect accurate data for the entire 75 meter line. To avoid having to invert and interpret two separate overlapping files, we then stitched the data together into one file using GeoTest©. GPR M. Leopold collected GPR data along two lines (Fig. 39) on Niwot Ridge using a portable RAMAC CU II GPR system (MALÅ Geosystems) following methods from Leopold et al. (2008). He used a combination of 25, 100, and 200 MHz antennae to obtain 2D-profiles along the lines. During data collection, the antennae were spaced at 1 m intervals, parallel to each other and perpendicular to the direction of the survey line. M. Leopold collected data every 0.25 m, and each trace of the line was stacked 16 times to insure maximum quality. To measure the local electromagnetic wave velocity through the Quaternary diamict, Leopold executed common midpoint surveys near the site of the two GPR lines. The different frequency antennae yielded different quality results; in this paper those with the 100 MHz antennae are presented. Temperature Measuring Measurements of soil and ground water temperature indicate the temperature of underlying regolith and bedrock, providing useful clues when searching for the locations of permafrost in alpine environments. Since soil moisture highly affects the accuracy of electrical resistivity, I measured temperatures of standing water at 153 locations between Cable Gate and D1 on Niwot Ridge. I measured the temperature of standing water and small streams along several transects, including one following the road from the Tundra Lab to D1, a southern traverse through Martinelli Snow Field following the contour up to D1, a northern traverse from between D1 and Lake Isabelle towards the north side of the West Knoll, and along the gelifluction lobes to the east of the Tundra Lab. At each location I recorded the GPS coordinates (Fig. 40a) using a Garmin eTrex with accuracy of ±3 m and measured soil temperatures (Fig. 40b). We collected water temperature data using a VWR Lollipop-type thermometer (H-B Instrument Company) with an accuracy of ± 0.5 °C after allowing three minutes to equilibrate before recording the temperature. At several locations I took and averaged multiple readings because of the temperature variability. Temperature Modeling Predictions regarding the depth at which permafrost will, or will not, occur are possible from models of subsurface temperatures. Annual oscillations of the surface temperatures drive subsurface temperature changes, although the effects are substantially damped deep into the subsurface. The latent heat associated with the freezing and thawing of water is ignored, since the effect is orders of magnitude smaller than other variables, and the surface temperature fluctuations are treated as sinusoidal waves with a period of one year. For this first-order approximation, advective energy transfer is ignored because the temperature difference between adjacent soil particles remains relatively constant over the course of several days. Radiation fluctuation is assumed to average out over the course of weeks, and affects the extent of Niwot Ridge uniformly, although in reality steep north-facing slopes receive much less radiation than south-facing slopes (Fig. 41). Taking the above considerations into account, average subsurface temperatures can be calculated over long periods of time. The temperature at depth varies as a damped and lagged sinusoidally varying function depending on surface temperature, since the ground averages out temperature fluctuations at depth and because it takes time for the temperature information to penetrate deep into the ground. The temperature can be found by: . . ... . ........................... . (Eqn. 4) . .. ... Here, the first two terms are the same as in Eqn. 3, and the last term describes the damped lagged sinusoid which decreases in the magnitude of oscillations with depth and increases in lag with depth. The quantity .... is defined as the magnitude of the seasonal oscillation from mean summer high to mean winter high at the surface, .. as the .. depth scale at which the temperature is .of the surface temperature, and P as the .. period (usually taken to be one year or 365 days). This fundamental equation describing the thermal behavior of periglacial landscapes can be used to predict temperatures at depths for regions across the world given the thermal diffusivity of the regolith (assumed -6 2 -1 -6 2 -1 to be approximately 1.3*10m sec for granite and 0.65*10m sec for diamictite) and the magnitude of the temperature oscillations (Lauchenbruch et al., 1986; Kim and Koo, 2007; Oladunjoye and Sanuade, 2012). Laboratory ERT After exporting the data from GeoTest as an inversion file (*.inv), I ran the data through an inversion process using the program RES2DINV© from Geotomo Software (see www.geotomosoft.com/ ) to account for topography and surface interaction and to calculate DOI indices. To compile specific resistivity models for each line, I computed least-squares inversion techniques according to the formulae (from Hilbich, 2009): ............................. .... (Eqn. 5) .......................... (Eqn. 6) where ... is the change in model parameters for the iteration k, mk-1 is the model parameter vector at the iteration k-1, m0 is a homogeneous half-space reference model, and dk is the discrepancy vector. The variables Wx, Wz, and Ws as well as Cx, Cy and Cz are the weighting and smoothing matrices, respectively, in the x, y and x directions and .k determines the relatively importance given to the model roughness. By definition, ax, ay and az are the relative weights given to the smoothness filters in the x, y and z directions and as is the relative weight for the damping factor, which minimizes deviation of the model from a reference model. The variable Jk is the Jacobian matrix of partial derivatives, which is recalculated after each iteration of the inversion process (Hilbich, 2009). RES2DINV calculated at least five iterations of the inversion process to reach the convergence limit of 3.0% using a damping factor of .......and ended with .... . ..... Root Mean Squared (RMS) errors were between 1.4 and 15.8 with an average of 6.6% for the 16 ERT lines in this study. For each line interpreted, I calculated the DOI using methods from Oldenburg and Li (1999) to determine which areas of the model were sensitive to the measured physical properties and which areas were not accurate enough to use with confidence. After completing the inversion process, I compared resistivity values with ground-truth measurements to correlate the resistivities of specific deposits. Combining these values with Leopold et al. (2013) allowed us to assign specific resistivity ranges to fine-and coarse-grained slope deposits, weathered and fractured bedrock, saprolite, ice lenses, and permafrost (Fig. 5). I was then able to display the results using Adobe Illustrator© to convey the physical properties of the subsurface in a graphical manner. Along ERT lines where soil pits were dug to saprolite for stratigraphy, I indicated the locations in the interpretation images to portray corresponding depths. GPR Interpretation of GPR data was based upon visual inspection of the reflection pattern following Neal (2004) and Leopold et al. (2008). I also considered factors such as nearby drill cores, SSR and ERT results, as well as data collected in previous studies (Leopold et al., 2008; Leopold et al., 2011). I used REFLEXW© ver. 7.0 by K.J. Sandmeier to interpret GPR results using the following procedure. I employed similar techniques for the interpretation of both GPR lines following Leopold et al., 2008: 1. Subtracted the running mean from the central point of each trace using the subtract-mean (dewow) 1D-filter with a time window of 10 ns. This eliminates a possible low frequency signal which can obstruct the true signal 2. Applied a bandpassbutterworth filter to remove any signal noise outside the main amplitude, removing any data outside of 36.5 to 137.0 MHz 3. Applied static correction with time window 28 ns to compensate for the time delay of the first return so the same time baseline can be used for all traces. 4. Changed plot options for a 1D velocity of .....m s -1 for better resolution of the returns from the upper few meters of subsurface 5. Applied manual y-gain between 40-120 ns to interactively define a digitized cain curve along the time axis to increase the resolution of the light grey layers and allow subsurface features to be seen easier 6. Applied topography with static correction -3D topography to compensate for surface topography, allowing a more accurate interpretation of the subsurface model 7. Removed air and ground traces using Photoshop© to better display the subsurface RESULTS Surface Deposits Surface deposits covering Precambrian-Paleogene granitic bedrock along Niwot Ridge and within Green Lakes Valley include periglacial terraces and lobes, stone banked terrace deposits, talus and glacial deposits, periglacial deposits, diamict, and gelifluction lobes (Fig. 42). Below the LGM glacial extent, exposed areas of bedrock have been polished smooth by glaciers. Above the maximum glacial extent, the few areas of exposed bedrock are very steep talus fields, since everything else has been covered with Quaternary diamict and periglacial deposits derived from the diamict and bedrock. The interesting pattern of periglacial terraces and lobe crests follows topography along Niwot Ridge from high elevations in the west to lower elevations near the Tundra Research Station. Temperature modeling The expected subsurface temperature structure throughout the year can be calculated from Eqn. 4 by varying the depth, z, at which the equation is calculated (Fig. 43). It is interesting to note how the amplitude of oscillation about the MAT decreases as the depth increases, and how the maximum temperature occurs later in the season at greater depths. °C and ideal temperature oscillations with an amplitude of 15 °C. This plot shows temperatures over a two year time period at one meter intervals from 0-7 m Given a typical geothermal gradient on the order of 25-30 °C/km, the temperature varies less than 0.1°C over the near subsurface where permafrost is found. It is now possible to solve for the depth of the active layer of permafrost, zactive, by setting the sine term equal to one (since it fluctuates between -1 and 1), yielding Eqn. 7. The depth scale, .., can also be calculated from the thermal diffusivity of the regolith, ., and the period of oscillations, P, taken to be one year (Eqn. 8). . .. . (Eqn. 7) ....... ........ .... ...... (Eqn. 8) . I have ignored the latent heat associated with the phase change of water throughout this derivation. While ignoring this variable captures the order of magnitude approximation of the problem, it is still worth mentioning here. Freezing water introduces a heat source, while melting water causes a heat sink in the thermal problem discussed above. The energy required to heat or cool 1 g of water by 1 °C is 4.2 J while the latent heat released or absorbed during melting and freezing is 334 kJ/kg, two orders of magnitude larger than the change in its temperature. The absorption of heat to melt subsurface frozen water will decrease the depth to which a thaw front is able to penetrate, causing Eqn. 3 to overestimate the active penetration depth by a factor proportional to the water content. Also important to mention is the lateral advection of cold water, either from melting snow or seasonal ice lenses, through the subsurface, which may be significant in the amount of heat convected to the surface. While this parameter is difficult to calculate, it is possible to estimate by measuring surface water temperatures throughout the region. Modeling Hopkins Memorial Forest Subsurface Temperatures It is helpful to examine subsurface temperature distributions in a region with no permafrost to ascertain what patterns on Niwot Ridge would look like if permafrost was not present. The method of calculating subsurface isotherms on a yearly basis can be applied if daily temperature measurements are made. Hopkins Memorial Forest is owned and operated by Williams College, located in Williamstown, MA (Fig. 44), providing mean daily surface temperature, pressure, wind, solar radiation, humidity, and precipitation measurements in addition to soil temperatures at 10 cm and 80 cm, and water temperatures 5.15 m below the surface (web.williams.edu/weather/archives.php). The water temperature is assumed to be in equilibrium with the soil at that depth due to the low permeability of the regolith and slow rate at which thermal energy can transfer horizontally. The upper three meters of Hopkins Memorial Forest are compose of unconsolidated glacial Lake Bascom clay sediments with interlayered lacustrine silt and sand varves, while the material below this depth is primarily densely packed glacial till that conducts thermal energy efficiently (Dahlberg, 1960). A plot of daily temperature measurements (Fig. 45) shows the yearly sinusoidal trend, with increased variability and amplitude towards the surface. An average surface temperature of 9.1 °C is far above the threshold for permafrost cited in the literature, corresponding with the lack of frozen ground in Williamstown, MA. Best fit sinusoidal functions where t is measured in years can be fit to the data to predict what temperatures will be on certain days during the following year (Fig. 46). The following equations are obtained with R2 values of 0.9203, 0.9886, 0.9975, and 0.9997 respectively: ... Surface Air Temperature: ......‰ .......".........‰ .. (Eqn. 9) ... ... Soil (10 cm): .......‰ ......."........‰ .. (Eqn. 10) ... ... Soil (80 cm): .......‰ .......".........‰ .. (Eqn. 11) ... ... Water (5.15 m): ......‰ .................‰ .. (Eqn. 12) ... It is interesting to note that the accuracy of each equation increases as the depth of the thermometer increases, due to the decrease in noise caused by weather fluctuations. The temporal lag of each equation can be calculated by comparing the date of the maximum temperature; the lag also increases with depth, as seen in Table 6. Based on the four equations calculated from these data (Eqn. 9-Eqn. 12), it is .. easy to calculate the missing parameters ........... ...and ................. .. . from Eqn. 4. By treating the sine term as equal to one, the ratio can be obtained by . dividing the difference between the offset of the surface temperature and the offset of the temperatures at depth with the depth of those measurements (Eqn. 13). The value .. can be calculated by solving (Eqn. 14) for each depth. A summary of these calculations . shows the ....and temporal lag constants derived from data collected in Hopkins . . Memorial Forest (Table 6). Although in theory the ....and temporal lag parameters . should remain constant with depth, they are all functions of the material in which the temperatures are calculated. ...... . .(Eqn. 13) .. . ... (Eqn. 14) ................... Table 6. List of the .......and temporal lag constants summarizing calculated parameters from Hopkins Memorial Forest, MA using Eqn. 4 Depth (cm) . . (m K -1) .. (m) Temporal Lag Behind Surface Temp (days) 10 1.91 0.593 10.5 80 1.59 1.838 26.1 515 0.448 2.91 99.1 Soil Pit Temperature Measurements on Niwot Ridge, CO A plot of the daily averages of maxima and minima temperatures at D1 station shows a sinusoidal oscillation of the temperature throughout the year, where the MAT is below the permafrost threshold of -1 °C for nearly every year studied (Ives, 1973) (Fig. 47). Temperature measurements near the Tundra Lab are similar to those at D1 due to the stations’ proximity, although the Tundra Research Station is 0.3 °C warmer due to the adiabatic lapse rate (Greenland, 1989). Temperature measurements from 2012 at the Tundra Research Station are not available due to technical difficulties. Temperature measurements taken at depth from the excavated pit near ERT Line 11 are showed below (Fig. 48a). A plot of temperature vs. depth with a best fit exponential curve according to Eqn. 4 (with !.......) is also displayed (Fig. 48b). The best fit model is Eqn. 15, where T is the temperature (°C) and z is the depth (cm), which yielded parameters . . ..... . . and .. ..... m. Both of these factors agree well with parameters provided in Material Density, . (kg/m3) Specific heat, c (kJ/kg °C) Thermal conductivity, k Latent heat of evaporation (W/m K) (kJ/kg) Quartz 2650 .732 8.4 43 Many soil minerals * 2650 .732 2.9 43 Soil organic matter * 1300 1.925 0.25 15 Water 1000 4.184 0.6 1.42 Air 1.2 1.004 .026 0.21 Table 1 and Table 6. ................ ......... (Eqn. 15) . Taking the parameters and .. calculated above, the Julian Day in which these . temperatures were calculated, a mean annual surface temperature of -3.7 °C, and amplitude of fluctuation of 15 °C, I computed a temperature-depth plot for the entire year (Fig. 49). The calculation demonstrates that a seasonally frozen layer is able to develop seasonal ice lenses and promote freeze-thaw motion which can propagate material downslope, but the inactive layer does not experience sub-zero temperatures and, thus, does not contribute to the movement of material through gelifluction. Figure 49. Annual temperature-depth profile near ERT Line 11 interpolated from average temperature measurements from 1952-2012 (Eqn. 4). Black dots are temperature measurements from the soil pit on July 24 and the red line is their calculated temperature profile. Thin black lines are calculated from running this model at 10 day intervals to create an outline for the daily temperature profile to oscillate through Using Eqn. 9 that describes the air temperature variations at Hopkins Memorial Forest, I can produce an annual temperature-depth profile using similar techniques. Using Eqn. 4 to vary the depth and day of measurement, we can create a plot of temperature depth profiles every 10 days (Fig. 50). Using Eqn. 9, Eqn. 10, Eqn. 11, and Eqn. 12 to describe the temperature at the surface, 10 cm, 80 cm, and 515 cm depth respectively, we can make a similar graph which shows measured temperature variations throughout the calendar year 2010-2011. Water Temperatures Map I conducted 153 water temperature measurements in all bodies of water that I was able to find on Niwot Ridge, as well as transects through both Saddle Stream and Como Creek (Fig. 52). Note how the cold temperatures at head of Saddle Stream and Como Creek quickly warm up as water begins to flow downhill (see HFLUX Modeling section). From surface water temperatures, we can map the location of the summer standing water in relation to the front of periglacial terrace crests (Fig. 53). The snow drift heights (D. Dethier, unpublished data, 2013) are calculated from subtracting summer and winter LiDAR data. The location of many standing bodies of water indicate that the water is trapped just above the crest of the gelifluction lobes, most likely supported by a layer of clay. HFLUX Modeling Using a stream temperature modeling program called HFLUX, developed by Gloss (2013), I was able to model the temperature in Como Creek throughout a 24 hour period along a 392 m transect based on 24 water temperatures we measured on August 4, 2012 (Appendix D) (Fig. 54). I approximate the temperature on this day as a sine wave with mean air temperature of 11.7 °C and amplitude of 7 °C, based off data from nearly meteorological stations, and the incoming shortwave radiation as a sine wave with mean of 1070 W/m2 with amplitude 37 W/m2 (Greenland, 1989; Williams et al., 1996). Clearly these approximations are only accurate to an order of magnitude since fluctuations from cloud coverage and pressure changes effect daily temperatures by several degrees. Although the HFLUX model is not very sensitive to discharge values, I estimated water discharge based on values from I. Nesbitt’s (2013) research on nearby Saddle and Martinelli Streams, which has nearly a constant width of 50 cm and area of .02 m2 along a gravel bed. Provided that the input water temperature at the top of the transect remained near 3 °C, and that all measurements were calculated at 5 cm depth, I estimated that the ground water warmed linearly between 3 and 4.5 °C along the transect due to ground’s lapse rate. I was able to measure the distance between each water temperature measurement and the source of the creek using ArcGIS© and calibrated HFLUX to calculate water temperatures between 10:00 am on August 4 and August 5, provided a constant wind speed of 0 m/s, relative humidity of 20%, and estimated zero cloud cover and shade along the creek. For August, 4, 2012, my model predicts a maximum temperature of 11.5 °C at 1:00 pm and a minimum temperature of 2.9 °C which extends roughly from 9:00 pm to 5:00 am. Stream temperature measurements from Martinelli Stream reveal maximum temperatures between 12.25 °C at approximately 2:00 pm with a minimum of 4.04 °C at 6:00 am. The model accurately predicts that the stream loses heat at first, but fails to calculate the manner in which the creek retains heat during the early morning. In both the modeled and measured temperatures, the stream heats up quickly once exposed to solar radiation: 7.5 °C in 4.5 hours (modeled) compared with 10.2 °C over the course of 8 hours (measured), and stays warm for most of the day while exposed to solar radiation. -4 3-1 Since discharge in the stream is low (approximately 10m s ) and water is shallow and turbulent for most of the transect, water temperatures are extremely sensitive to the ambient air temperature, incoming solar radiation, and latent heat flux. The streambed conduction, longwave radiation, and sensible heat flux do not affect the temperature of the water at night, but cannot be completely ignored in the calculation, especially during the day (Fig. 56). program showing solar (shortwave) radiation, longwave radiation, latent heat flux, streambed conduction, sensible heat flux, and the total heat flux in the system over a 24 hour time period Between August 4, 2013 and August 6, 2013 the minimum temperature was 4.04 °C and the maximum temperature was 10.26 °C, with an average of 7.36°C. During this time period, the subsurface ground temperature follows a sinusoidal pattern (Fig. 57) based off calculations from Eqn. 4. Water that reaches the surface must have come from a depth of at least 145 cm, where the ground temperature is at or below 4.04°C. During the early hours of the morning, high humidity and low latent heat flux allow the water temperature to fall to the average ground temperature at depth. ERT This study recorded a total of 16 ERT lines, 2 GPR lines, and 153 temperature measurements along Niwot Ridge, Gordon Gulch, and 4th of July Valley (Fig. 24). Each of the ERT lines was conducted using both a Wenner and dipole-dipole array, with additional arrays used as needed in the field to determine the subsurface composition. Many of my ERT lines show various periglacial deposits and subsurface compositions that are difficult to detect from surface data collection. I compiled a complete table of inversion images (Appendix C) showing all 16 ERT lines computed in my study (Fig. 24). Of these 16 lines, I chose four representative lines that I display and discuss in detail (Fig. 58). ERT Line 3 is located at an elevation between 3694 and 3705 m on the north facing slope of Green Lakes Valley, 500 m southwest of RG5, and is representative of a steep shaded scree slope (Fig. 60a). Large 10-20 cm cobbles and sand cover the surface of this line (Fig. 59) underlain by an extremely high resistivity layer that extends horizontally in both directions. This line maintained an average slope of 36° during its length, providing an easy surface for small scale landslides and debris flows when small amounts of weight are applied (Fig. 60a). steep scree face in upper Green Lakes Valley ERT Line 4 is located at an elevation of 3635 in Green Lakes Valley adjacent to Green Lake 5 and Rock Glacier 5 (Fig. 62). The upper 200 cm of this line are colluvial saturated silt and sand underlain by an extremely high resistivity layer that extends horizontally in both directions (Fig. 61). ERT Line 10 is located at an elevation of 3500 m on Niwot Ridge, 200 m southwest of the Tundra Research Station, and is representative of an area containing periglacial and terrace crests (Fig. 64). A small spring emanates from the hillside at 1.5 °C a few meters south of the line, eventually feeding Saddle Stream. A low resistivity organic layer underlies the upper 50 cm of this line (Fig. 63), underlain by an extremely high resistivity layer that extends horizontally in both directions. We excavated two 100 and 150 cm deep soil pits at the west end of this profile to determine the origin of the high resistivity layer and measure soil temperatures along the profile wall. vegetation above the head of Martinelli and Saddle Stream ERT Line 15 is located on a steep north-facing slope in Gordon Gulch at an elevation of 2500 m (Fig. 66). We excavated five pits along this line (black lines in Fig. 65a), which reveal the composition of the subsurface and that depth to bedrock varies much more than previously imagined. Results of the ERT line indicate that high resistivity bedrock underlies most of the line and crops out at the surface near the 40 meter mark. Interpretation of the ERT line (Fig. 65b) shows a pocket of lower resistivity saprolite surrounded by more solid weathered bedrock near the center of the line and a toeslope deposit of low resistivity material at the bottom of the slope adjacent to the creek. (>2500 Om) and low resistivity (<50 Om). Black lines represent soil pits along the line. b. Interpretation of ERT Line 15 showing surface boulders, toeslope deposits, saprolite, weathered bedrock, and bedrock GPR M. Leopold ran two GPR lines during August, 2009, located 400 m NW of the Tundra Research Station (Fig. 39). These two nearly perpendicular lines were chosen to analyze cross sections through and across what was thought to be an active gelifluction lobe. Each line is 25 m long and the two lines cross 12 m from the western terminus of line 1 and 12 m from the northern terminus of line 2. Unsurprisingly, GPR 1 is more interesting than GPR 2 because it crosses through the gelifluction lobe as opposed to remaining near the edge of it (Fig. 67). Although the GPR results do not tell us much we do not already know about these gelifluction lobes, it is useful to verify and cross-check measurements from ERT lines nearby. We can clearly see both parallel and subparallel layers below the surface of the line in the GPR image (Fig. 68a). The parallel layers reflect the inherent structure of the lobe, whereas subparallel layers imply previous gelifluction lobe surfaces that have since been buried by the main lobe present today. GPR Line 2 shows parallel subsurface layers of gelifluction deposits (Fig. 69). This line is relatively uninteresting compared with GPR Line 1 because of the simple structure underlying the gelifluction lobe. When we place both these lines in a 3D environment according to their proper elevations and angle of intersection, the inactive gelifluction lobe surfaces are visible (Fig. 70). DISCUSSION Understanding the processes that produce permafrost and permanent ice lenses is an important step in studying how the Critical Zone evolves with climatic changes over time. Since the distribution of frozen ground is controlled by minute changes in temperatures, precipitation, and solar radiation, studying inactive permafrost terrain can reveal details about the history of periglacial processes. Geophysical investigations of Niwot Ridge, Green Lakes Valley, 4th of July Valley, and Gordon Gulch reveal specific details about both present and Holocene Critical Zone processes. In this study I: (1) examine and map surface deposits; (2) compute temperature models based on both air and water temperature measurements; (3) predict subsurface temperatures based on temperatures measured in excavated pits; (4) compute both ERT and GPR geophysical lines; and (5) and interpret how paleoclimate influenced periglacial activity on Niwot Ridge and in Gordon Gulch. Results from ERT lines do not indicate the presence of permafrost by revealing that the subsurface of Niwot Ridge and Green Lakes Valley is devoid of permafrost, except at locations adjacent to rock glaciers or on steep, shaded, north-facing slopes. Soil temperature measurements from soil pits and heat modeling of the subsurface thermodynamics demonstrate that the temperature at depth is too high to support permanent ice lenses or permafrost. Surface Deposits Understanding the regolith and periglacial features on Niwot Ridge and in Green Lakes Valley is fundamental for developing a theory behind the location and thickness of permafrost and permanent ice lenses. Using a combination of field measurements, ArcGIS©, geologic maps, previous studies, and orthophotos, I compiled a map showing the location of periglacial terrace and lobe crests, stone banked terraces, talus and glacial deposits, periglacial deposits, diamict, gelifluction lobes, and granitic bedrock. Gelifluction lobes transport portions of hill slopes and locally erode the subsurface as they slowly move downhill. Associated frost-cracking and freeze-thaw action may increase the regolith transport rate past equilibrium, resulting in toeslope deposits at the bottom of many slopes (Fig. 65). The weathered bedrock pattern and toeslope deposits within Gordon Gulch provide evidence that periglacial conditions were maintained throughout the late Pleistocene at an elevation of 2500 m. Although there is no periglacial activity today in Gordon Gulch, and little on Niwot Ridge, mapping inactive periglacial deposits provides evidence that these locations are capable of maintaining permafrost supporting climates. Climate models from the LGM show that temperatures were 6-8 °C colder in the Colorado Front Range, providing conditions that could have supported both permafrost and gelifluction lobes at an elevation of 2500 m (Dühnforth and Anderson, 2011). Temperature Modeling One important consequence of conductive heat flow is that the minimum and maximum temperatures in the subsurface lag behind surface temperatures. The maximum temperature at a depth of three meters is 55 days behind the maximum surface temperature, and at a depth of ten meters the minimum temperature occurs during the maximum surface temperature, and vice versa. This phenomenon could allow permafrost at depth to survive the warm summer months and maintain much colder temperatures than would be predicted from extrapolating summer air temperatures down into the subsurface. The values for each parameter in Table 6 should be constant along a depth profile for a homogeneous medium exposed to daily and seasonally varying solar radiation and surface temperatures (Anderson and Anderson, 2010). In Hopkins Memorial Forest, or anywhere in the real world, constant subsurface thermal properties are obviously not the . case. By varying and .. over an order of magnitude, we increase the error when . predicting temperatures at depth based upon surface temperature measurements. Temperature modeling based on thermal energy equations functions relatively well for estimating temperatures over kilometer scale depth profiles, but appears to lack the accuracy required when searching for permafrost in the subsurface based upon surface temperature measurements. We do not see the same yearlong temperature-depth profile from parameters derived in the field (Fig. 49) as we do in the model according to only thermal energy equations (Fig. 9). Since the temperature below two meters never crosses the 0 °C isotherm, the ground is not able to freeze at this depth. According to the model, the ground above two meters depth periodically freezes and thaws throughout the year (in addition to diurnal cycles), causing frost-heave and other processes that combine to transport mass downhill. Based on observed values from regolith and water temperature measurements, it is likely that only the top 40-50 cm experiences frost-heave. Based on the temperature model from Anderson and Anderson (2010), it is evident that the subsurface below two meters does not remain frozen throughout the year, eliminating the possibility of permafrost at this depth. Regolith less than two meters below the surface may freeze for several months during the spring, but melts again once the surface temperature warms substantially. This freeze-thaw action provides the necessary conditions for slow movement of gelifluction lobes seen on Niwot Ridge (Fig. 42). While holding all other parameters constant, we can vary the MAT on Niwot Ridge to replicate conditions of the early 20th century based on data from meteorological stations along Niwot Ridge, around 2 °C lower than today (Leopold et al., 2010). This model produces a thin, permanently-frozen subsurface layer that provides conditions in which permafrost could form. Modeling of Gordon Gulch using temperatures estimated from the LGM suggests that the area may have also supported ice lenses and permafrost. Temperatures at Niwot Ridge are much colder and more intense than those in Hopkins Memorial Forest due to the low elevation of Williamstown, MA (approximately 300 m). Besides a shift of the temperature-depth profile to warmer temperatures, we see steeper thermal gradients at Niwot Ridge because of the regolith’s composition. Since the thermal conductivity of glacial till is much higher than that of varved glacial lake sediments, temperatures within the near-subsurface at Niwot Ridge are able to remain closer to the surface temperature than those at Hopkins Memorial Forest. Below 3 m depth at Hopkins Memorial Forest, temperature gradients reflect those near the surface of Niwot Ridge because both locations are composed of glacial till and diamict. Within Hopkins Memorial Forest, a shallow ~50 cm freeze-thaw zone creates seasonal frost heave and needle ice, although there is presently zero periglacial activity in Massachusetts. Temperatures at the surface, 10 cm, 80 cm, and 515 cm depth fit very well to sinusoidally varying functions, with R2 values all above 0.9 (Fig. 45 and Fig. 46). Modeling temperature-depth profiles at Hopkins Memorial Forest using equations generated from best fit curves produces models that closely resemble temperature structures witnessed on Niwot Ridge (Fig. 49 and Fig. 50). Due to the relatively mild winter of 2011-2012, and a quick snow melt period in the early spring, ground temperatures at 10 cm depth were not given sufficient time to equilibrate and reach their minimum temperature. Temperatures at 515 cm were measured in the water column, which convects thermal energy away from the thermometer much more efficiently than glacial till would. The difference in material allows temperatures to change quickly and follow the surface temperature oscillations, therefore, the amplitude of oscillations at depth is much higher than it would be if temperatures were measured in regolith. We expect the temperature below 4 m depth to remain constant within 1 °C during an entire year, but the convection and advection through water allows temperatures to change easily. The ground subsurface in Hopkins Memorial Forest is not homogenous, but rather consists of glacial lake sediments for the upper three meters with glacial till below. This explains the variance in parameters seen in Table 6, as well as the slightly wider variance of temperatures at depth. Water temperature measurements on Niwot Ridge indicate that the shallow ground water is mainly above 1.5 °C until at least late July and early August, even before it has been exposed to solar radiation. Temperatures this warm provide difficult conditions for ice lenses to persist through the summer. The few cold springs that emanate from the ground occur at nearly the same locations as the snow drifts (Fig. 53). From temperature models of excavated pits, we estimate that the coldest the temperature reaches at depth is 1.1 °C, too warm for the development of permafrost. Sufficient advection of thermal energy from snowmelt can maintain low temperatures in small pockets of ground, which provide a conduit for water to seep through without losing large quantities of heat. The three cold water temperature points along the ridge from left to right have values of 2.7, 1.1, and 2.5 °C (Fig. 53). According to Eqn. 15, this water must be coming from depths of at least 3.6 m for the 2.7 and 2.5 °C points. According to my conduction-based model, it does not seem possible for meteoric water to reach the ground surface at a temperature of 1.1 °C on August 5, regardless of what depth it came from. The three cold water measurements are located at sites where winter snow drifts have depths of 3.3, 3.6, and 3.1 m, respectively (Fig. 53). Water melting from these drifts remains at or near 0 °C for lengthy periods of time, since there is very little thermal energy in the subsurface to heat it up. Snow drifts at these three locations, as well as other areas covered by thick winter snow drifts, insulate the ground long into the summer, keeping subsurface water at much lower temperatures than areas not insulated by snow banks. It is because of these snow drifts that water at these three locations, as well as at the head of Saddle Stream, Martinelli Stream, and Como Creek, is able to stay near freezing for most of the summer. HFLUX Modeling Using HFLUX, I was able to determine how stream water responds to solar and longwave radiation, latent heat flux, streambed conduction, and sensible heat flux along a 392 m transect of Como Creek. Knowledge of the thermal energy within alpine streams on Niwot Ridge provides clues regarding the temperature structure of the subsurface and reveals that the source of the water is not melting permafrost, as previously believed. Martinelli and Saddle Streams are located in catchments to the east of Como Creek at approximately the same elevation, have similar discharge and streambed compositions, and provide excellent parallels for temperature modeling for each stream. Saddle Stream is 900 m to the east of Como Creek while Martinelli Stream is 1.2 km to the east. The Boulder Creek CZO monitors both these streams with thermometers that record the average temperature, pH, and water height over a period of 10 minutes. The temperature-time profile for Saddle Stream is less accurate due to its vulnerability to precipitation and solar radiation changes from cloud coverage, while the Martinelli Stream temperature sensor is more robust. Because of the low discharge values and lack of solar radiation at night, water temperatures in the early morning drop to nearly 4 °C. These low temperatures reflect the thermal energy of the regolith in which spring water seeps through. Based on both measured and modeled water temperature data, the ground surface in Como, Martinelli, and Saddle Stream catchments must be above 0 °C, too warm to support the development of permafrost. Results of this modeling are important because they refute evidence of permafrost or permanent ice lenses on Niwot Ridge, establish a near surface water table of 3-4 °C, and explain how water penetrates into and seeps out of the subsurface. By comparing modeled temperature values in Como Creek (Fig. 54) with measured temperature values in Martinelli Stream (Fig. 55), we verify the accuracy of the HFLUX model and can confidently estimate the thermal energy of surface water over a large transect. ERT The results of ERT lines on several aspects of Niwot Ridge reveal that permafrost is absent beneath periglacial terrace and lobe crests on Niwot Ridge and within Green Lakes Valley. These results agree with those from Leopold (2008, 2010, 2013) in that permafrost is absent on Niwot Ridge in locations where it existed in the 1970s. On the other hand, ERT lines within Green Lakes Valley reveal that patches of permafrost persist throughout the summer on steep shaded slopes or at the foot of rock glaciers. To determine how accurately we could interpret our ERT inversion results, we computed DOI measurements for ERT Lines 3, 4, 10, and 15 (Fig. 71). According to Oldenburg and Li (1999) we considered DOI index values higher than 0.2 as unreliable regions of the ERT model. In the figures below, only the light, medium, and dark shades of blue can be analyzed in detail, whereas resistivity values in regions beneath any other color cannot be interpreted as accurately. The four lines discussed in this paper have exceptional DOI indices, allowing us to discuss ERT inversion results without risk of inaccurately analyzing the subsurface. At an elevation of 3700 m, ERT Line 3 is the highest of the four lines interpreted in this study. This line extends northward up a steep slope which remains shaded nearly the entire day. The high elevation causes a low MAT and the shade provides limited solar radiation, which allows ice lenses to persist throughout an entire year and create permafrost below two meters depth. Most winters through late spring, the south side of ERT Line 4 is covered with a thick layer of insulating snow, whereas the north side has been blown free of snow and remains exposed to cold winter air temperatures. Variable depths of snow accumulate in front of RG5 during the winter, ranging in depths of up to 4 m. Thermal energy escapes from depths of several meters, allowing cold temperatures to freeze the rock and interstitial water to create a thick layer of permafrost at the foot of the rock glacier. Most likely frozen conditions exist throughout the profile, but the ice is more concentrated to the northwest of the line. Solid unfrozen bedrock in the area has a resistivity between 3-6 kOm whereas cracked unfrozen bedrock has a slightly lower resistivity (M. Leopold, personal communication, April 11, 2013). Cracks within the bedrock can fill with ice and drastically increase the resistivity without requiring different material. The bedrock resistivity in this survey is between 50-150 kOm. These high resistivities correlate well with frozen bedrock with varying ice concentrations filling small cracks in the bedrock. Published studies (e.g. Ives and Fahey, 1971), MATs below -1 °C, sufficient precipitation, and the initial interpretation of ERT Line 10 (Fig. 63) suggest the possible presence of permafrost in the high (>70 kOm) layer underlying the surface. Field observations of excavated pits (Fig. 48) showed that the high resistivity material is open­work gravel, which conducts electricity extremely poorly due to the interstitial air between large clasts. Wind swept snow does not accumulate above this line, allowing cold temperatures to penetrate deep into the subsurface, but the exposure to solar radiation quickly warms up the ground and melts seasonal ice lenses. We expected to not find permafrost under ERT Line 15 at Gordon Gulch (Fig. 66) because of the warm MAT (5.1 °C) but we did not anticipate discovering inactive permafrost features hidden in the subsurface of the north facing slope of Gordon Gulch. The pocket of low resistivity saprolite (800-1200 Om) records more alteration than the surrounding saprolite and bedrock (1800-2500 Om). This pocket was most likely derived from preferential freeze-thaw cracking within the bedrock during periglacial conditions within the LGM. Once freeze-thaw cracking had sufficiently split the bedrock, frost-heave and gelifluction most likely removed some of the material and allowed the cracks to expand. Once temperatures began to rise, these large cracks became inactive and were filled with material that lowered the overall resistivity of this pocket. GPR Using GPR as well as ERT techniques allows us to verify the absence of permafrost on Niwot Ridge. The buried gelifluction lobes examined in this study (Fig. 68) match the description of “topographically subdued” older lobes that have been subsequently buried by additional gelifluction lobes (Benedict, 1970 p. 216). Benedict (1970) hypothesizes that these older lobes and terraces developed during the interval separating the Middle Pinedale glacial maximum and the beginning of the Altithermal, a period filled with cool-wet conditions ideal for frost-creep and gelifluction. Radiocarbon dates from similar buried gelifluction lobes on and near Niwot Ridge provide a minimum age of formation between 5,300 and 5,800 BP. Benedict (1970) describes a similar lobe overtaking an older buried lobe in the Indian Peaks region and has advanced a distance of 4.45 m during the past 2340 ± 130 years. These buried gelifluction lobes provide paleoclimatic evidence of periglacial conditions while protecting dateable organic matter from erosion. Paleoclimatic Conditions Inactive periglacial features and glaciated deposits reveal clues about the paleoclimate of Niwot Ridge and Gordon Gulch. By examining periglacial deposits, 10Be measurements, buried organic material, and climate oscillations today, we can infer temperature and precipitation conditions for the Colorado Front Range throughout the past 20,000 years. Regolith within Gordon Gulch does not appear to be moving due to frost-heave or gelifluction today, although there is overwhelming evidence for periglacial activity during the mid-Holocene. Evidence that rock glaciers began advancing between 4000 and 2000 years ago within the 4th of July Valley provides climatic and temporal evidence for periglacial activity at lower elevations during that time period. Benedict (1970) concludes that 2000 years ago rock glaciers within the 4th of July Valley, and presumably within Green Lakes Valley, were still moving from ample precipitation and cold temperatures. As a result of excellent Depth of Investigation indices (Fig. 71), we can confidently conclude that evidence of periglacial activity exists on Niwot Ridge and within Gordon Gulch. For freezing temperatures to occur at an elevation of 3500 m on Niwot Ridge, a drop in the ELA of only 100 m is required, corresponding to lowering the MAT by 0.1-0.5 °C. Climate change since 1950 has already increased the MAT on Niwot Ridge by 2 °C, enough to completely melt permafrost that was prominent only a century ago (Fig. 33). For freezing temperatures to occur at an elevation of 2500 m within Gordon Gulch, a drop in the ELA of at least 1000 m is required. Although this appears to be a large change in regional climate, lowering the MAT by 1.5-2 °C and increasing the precipitation by 50-100 mm provides the conditions necessary to form gelifluction lobes within Gordon Gulch. Conditions during the first half of the 20th century were cold enough to develop and maintain periglacial formations on Niwot Ridge, while the climate during the LGM provided temperature and precipitation values that supported periglacial conditions within Gordon Gulch. By studying the regions climate history, an accurate model of temperatures that show the location and extent of permafrost can be created. Future Today there is practically zero permafrost present within Green Lakes Valley or on Niwot Ridge, except in locations nearly completely shielded from solar radiation (Fig. 41). If current global warming trends continue (Fig. 33) it is possible that the melting of permafrost on Niwot Ridge and throughout most of the Colorado Front Range will continue as predicted by Williams et al. (1996). Although it is very difficult to quantify the magnitude of rising temperatures and changing precipitation on Niwot Ridge, the small amount of permafrost remaining is very susceptible to any increase and may seize to exist completely within the next 10-15 years. If global warming trends reverse within the next century, it is possible that lowering the MAT by 0.1 to 1 °C could once again establish the role of permafrost, permanent ice lenses, and gelifluction lobes within the Boulder Creek Critical Zone. CONCLUSIONS Slopes on Niwot Ridge cover elevations between 3450 and 3800 m, with varying MATs and exposure to solar radiation. While in the 1970’s permafrost was pervasive on Niwot Ridge above 3500 m, periglacial landforms (Fig. 42) are moving slowly, if at all (Benedict, 1970; Ives and Fahey, 1971). Today, Niwot Ridge and Green Lakes Valley are located at the threshold of a permafrost environment, where many sites have completely lost permafrost by global warming during the last 40 years. A combination of geophysical, field, and meteorological measurements, in addition to heat-flow modeling, indicate that the present subsurface on Niwot Ridge is too warm to contain permafrost. By using non-invasive geophysical techniques along Niwot Ridge and Gordon Gulch, this study examines the subsurface of recently active and inactive periglacial features in the Front Range of Colorado. Niwot Ridge appears to be at the threshold of becoming a completely permafrost-free location within the next decade, based on resistivity interpretations, water and soil temperature measurements, and field observations. Using climate records as the main driving mechanism for gelifluction forming locations, one can infer that freeze-thaw conditions were present in Gordon Gulch during the LGM and contributed to shaping the present geomorphology of the region. REFERENCES Anderson, R.S., Riihimaki, C.A., Safran, E.B., and MacGregor, K., 2006, Facing reality: Late Cenozoic evolution of smooth peaks, glacially ornamented valleys, and deep river gorges of Colorado’s Front Range: Geological Society of America, v. SPECIAL PA, no. 398, p. 397. Anderson, R.S., and Anderson, S., 2010, Geomorphology: the mechanics and chemistry of Landscapes: Cambridge University Press, Cambridge. Anderson, S., Von Blanckenburg, F., and White, A.F., 2007, Physical and chemical controls on the Critical Zone: Elements, v. 3, no. 5, p. 315–319, doi: 10.2113/gselements.3.5.315. Anderson, S., Bales, R.C., and Duffy, C.J., 2008, Critical Zone Observatories: Building a network to advance interdisciplinary study of Earth surface processes: Mineralogical Magazine, v. 72, no. 1, p. 7–10, doi: 10.1180/minmag.2008.072.1.7. Anderson, S., Qinghua, G., and Parrish, E.G., 2012, Snow-on and snow-off LiDAR point cloud data and digital elevation models for study of topography, snow, ecosystems and environmental change at Boulder Creek Critical Zone Observatory, Colorado. Andersson, J., 1906, Solifluction, a component of subaerial denudation: The Journal of Geology, v. 14, no. 2, p. 91–112. Barry, R., 1973, A climatological transect on the east slope of the Front Range, Colorado: Arctic and Alpine Research, v. 5, no. 2, p. 89–110. Befus, K.M., Sheehan, A.F., Leopold, M., Anderson, S., and Anderson, R.S., 2011, Seismic constraints on Critical Zone architecture, Boulder Creek Watershed, Front Range, Colorado: Vadose Zone Journal, v. 10, no. 3, p. 915, doi: 10.2136/vzj2010.0108. Benedict, J.B., 1966, Radiocarbon dates from a stone-banked terrace in the Colorado Rocky Mountains, USA: Geografiska Annaler. Series A, Physical Geography, v. 48, no. 1, p. 24–31. Benedict, J.B., 1970, Downslope soil movement in a Colorado alpine region: Rates, Processes, and Climatic Significance: Arctic and Alpine Research, v. 2, no. 3, p. 165–226. Benedict, J.B., 1976, Frost creep and gelifluction features: A review: Quaternary Research, v. 6, no. 1, p. 55–76, doi: 10.1016/0033-5894(76)90040-5. Benedict, J.B., 1981, The Fourth of July Valley: Glacial geology and archeology of the timberline ecotone: Center for Mountain Archeology. Benedict, J.B., 2005, Rethinking the Fourth of July Valley site: A study in glacial and periglacial geoarchaeology: Geoarchaeology, v. 20, no. 8, p. 797–836, doi: 10.1002/gea.20084. Berg, N.H., 1986, Blowing snow at a Colorado alpine site: measurements and implications: Arctic and Alpine Research, v. 18, no. 2, p. 147–161. Bierman, P., and Montgomery, D., 2013, Glacial and periglacial geomorphology, in Key concepts in geomorphology. W. H. Freeman & Company. New York, NY. Brantley, S.L., 2008, Understanding soil time: Science, v. 321, no. 5895, p. 1454–5, doi: 10.1126/science.1161132. Bruemmer, F., 1987, Life upon a permafrost: Natural History, v. 96, no. 4, p. 30. Burger, R.H., Sheehan, A.F., and Jones, C.H., 2006, Introduction to applied geophysics: W.W. Norton & Company, New York & London. Caine, N., 2010, Recent hydrologic change in a Colorado alpine basin: an indicator of permafrost thaw?: Annals Of Glaciology, v. 51, no. 56, p. 130–134. Cardarelli, E., and Fischanger, F., 2006, 2D data modelling by electrical resistivity tomography for complex subsurface geology: Geophysical Prospecting, v. 54, no. 2, p. 121–133, doi: 10.1111/j.1365-2478.2006.00522.x. Cho, W.J., Kwon, S., and Choi, J.W., 2009, The thermal conductivity for granite with various water contents: Engineering Geology, v. 107, no. 3-4, p. 167–171, doi: 10.1016/j.enggeo.2009.05.012. Cole, J.W., and Braddock, W., 2009, Geologic Map of the Estes Park 30’ X 60’ Quadrangle, North-Central Colorado: USGS. Cross borehole electrical resistivity tomography measurements, 2003, Advanced Geosciences, Inc., p. 1–9. CZO about us -Critical Zone Observatory National, 2012, Critical Zone Observatory National Program. Web. Dahlberg, C., 1960, Geologic map of the Hopkins Forest: Undergraduate thesis, Williams College, Williamstown, MA. DeGroot-Hedlin, C., and Constable, S., 1990, Occam’s inversion to generate smooth, two-dimensional models from magnetotelluric data: Geophysics, v. 55, no. 12, p. 1613–1624. Deidda, G., Bonomi, E., and Manzi, C., 2003, Inversion of electrical conductivity data with Tikhonov regularization approach: some considerations: Annals of Geophysics, v. 46, no. June. Dethier, D., Birkeland, P.W., and Shroba, R.R., 2003, Quaternary stratigraphy, geomorphology, soils, and alpine archaeology in an alpine-to-plains transect, Colorado Front Range, in Easterbrook, D.J. ed,. Quaternary Geology of the United States, INQUA 2003 Field Guide Volume, Desert Research Institute, Reno, p. 81– 104. Dethier, D., and Racela, J., 2013, Weather archive: Weather in Hopkins Memorial Forest. Dühnforth, M., Anderson, R.S., Ward, D.J., and Stock, G.M., 2010, Bedrock fracture control of glacial erosion processes and rates: Geology, v. 38, no. 5, p. 423–426, doi: 10.1130/G30576.1. Dühnforth, M., and Anderson, R.S., 2011, Reconstructing the glacial history of Green Lakes Valley, North Boulder Creek, Colorado Front Range: Arctic, Antarctic, and Alpine Research, v. 43, no. 4, p. 527–542, doi: 10.1657/1938-4246-43.4.527. Electrical Resistance Tomography for Subsurface Imaging, 2000, Innovative Technology Summary Report. Elias, S.A., 2001, Paleoecology and Late Quaternary environments of the Colorado Rockies, in Bowman, W.D. and Seastedt, T.R. eds., Structure and Function of an Alpine Ecosystem, p. 285–303. Embleton, C., and King, C.A.M., 1975, Periglacial geomorphology: St. Martin’s Press, London: Edward Arnold. Fahey, B.D., 1973, An analysis of diurnal freeze-thaw and frost heave cycles in the Indian Peak region of the Colorado Front Range: Arctic and Alpine Research. Fahey, B.D., 1975, Nonsorted circle development in a Colorado alpine location: Geografiska Annaler. Series A. Physical Geography, v. 57, no. 3, p. 153–164. Fall, P.L., 1997, Timberline fluctuations and late Quaternary paleoclimates in the Southern Rocky Mountains, Colorado: GSA Bulletin, doi: 10.1130/0016­7606(1997)109<1306. Foster, M., Anderson, R.S., Dühnforth, M., and Kelly, P., 2011, Constraining regolith production on a hillslope over long timescales: interpreting in situ 10Be concentrations on an evolving landscape. Abstracts with Programs, Geologic Society of America, v. 46-8. French, H., 2007, The periglacial environment: John Wiley & Sons. Gable, D., and Madole, R.F., 1976, Geologic map of the 7.5 minute Ward quadrangle, Boulder County, Colorado: USGS. Giardino, J.R., Vitek, J.D., and DeMorett, J.L., 1992, A model of water movement in rock glaciers and associated water characteristics. Glose, A., 2013, Stream heat budget modeling with HFLUX to predict system response to change: Abstracts with Programs, Geological Society of America Northeastern Section, v. 48-6. Gooseff, M., Balser, A., Bowden, W., and Jones, J., 2009, Effects of hillslope thermokarst in northern Alaska: Eos, v. 90, no. 4, doi: 10.1029/2007JG000470.Hassol. Greenland, D., 1989, The climate of Niwot Ridge, Front Range, Colorado, USA: Arctic and Alpine Research, v. 21, no. 4, p. 380–391. Greenstein, L.A., 1983, Alpine permafrost on Niwot Ridge based on indirect locational techniques: University of Colorado. Hamann, H., 1998, Snowcover controls on alpine soil surface temperature patterns, Niwot Ridge, Colorado, USA: University of Colorado (Masters Thesis). Harris, C., Arenson, L.U., Christiansen, H.H., Etzelmüller, B., Frauenfelder, R., Gruber, S., Haeberli, W., Hauck, C., Hölzle, M., Humlum, O., Isaksen, K., Kääb, A., Kern-Lütschg, M. a., Lehning, M., et al., 2009, Permafrost and climate in Europe: monitoring and modelling thermal, geomorphological and geotechnical responses: Earth-Science Reviews, v. 92, no. 3-4, p. 117–171, doi: 10.1016/j.earscirev.2008.12.002. Hilbich, C., Marescot, L., Hauck, C., Loke, M.H., and Mausbacher, R., 2009, Applicability of electrical resistivity tomography monitoring to coarse blocky and Ice-Rich permafrost landforms: Permafrost and Periglacial Processes, v. 284, no. December 2008, p. 269–284, doi: 10.1002/ppp. Hills, F., and Dickinson, K., 1982, Silver Plume Granite, possible source of uranium in sandstone uranium deposits, Tallahassee Creek and High Park areas, Fremont and Teller Counties, U.S.G.S. Hinckley, E.-L.S., Ebel, B.A., Barnes, R.T., Anderson, R.S., Williams, M.W., and Anderson, S., 2012, Aspect control of water movement on hillslopes near the rain-snow transition of the Colorado Front Range: Hydrological Processes, p. n/a–n/a, doi: 10.1002/hyp. Hinzman, L., Goering, D., and Kane, D., 1998, A distributed thermal model for calculating soil temperature profiles and depth of thaw in permafrost regions: Journal of Geophysical Research, v. 103, no. 98. Ives, J.D., and Fahey, B.D., 1971, Permafrost occurrence in the Front Range, Colorado Rocky Mountains, USA: Journal of Glaciology, v. 10, no. 58, p. 105–111. Ives, J.D., 1973, Permafrost and its relationship to other environmental parameters in a midlatitude, high-altitude setting, Front Range, Colorado Rocky Mountains, in Proceedings of Permafrost: The North American Contribution to the Second International Conference, p. 121–125. Janke, J.R., 2005, The occurrence of alpine permafrost in the Front Range of Colorado: Geomorphology, v. 67, no. 3-4, p. 375–389, doi: 10.1016/j.geomorph.2004.11.005. Janke, J.R., Williams, M.W., and Evans, A., 2012, A comparison of permafrost prediction models along a section of Trail Ridge Road, Rocky Mountain National Park, Colorado, USA: Geomorphology, v. 138, no. 1, p. 111–120, doi: 10.1016/j.geomorph.2011.08.029. Kellogg, K., Shroba, R.R., Bryant, B., and Premo, W., 2008, Geologic Map of the Denver West 30 X 60 Quadrangle, North-central Colorado: USGS, p. 80225. Kim, J., Lee, Y., and Koo, M., 2007, Thermal properties of granite from Korea, in AGU Fall Meeting Abstracts, p. 576. Lachenbruch, A.H., 1957, Three-dimensional heat conduction in permafrost beneath heated buildings: Experimental and Theoretical Geophysics. Lachenbruch, A.H., and Marshall, B.V., 1986, Changing climate: geothermal evidence from permafrost in the Alaskan arctic: Science, v. 234. Leopold, M., Dethier, D., and Raab, T., 2008, Using geophysical methods to study the shallow subsurface of a sensitive alpine environment, Niwot Ridge, Colorado Front Range, USA: Arctic, Antarctic, and Alpine Research, v. 40, no. 3, p. 519–530, doi: 10.1657/1523-0430(06-124). Leopold, M., Dethier, D., Völkel, J., and Raab, T., 2008, Combining sediment analysis and seismic refraction to describe the structure, thickness and distribution of periglacial slope deposits at Niwot Ridge, Rocky Mountains Front Range, Colorado, USA: Zeitschrift für Geomorphologie, Supplementary Issues, v. 52, no. 2, p. 77–94, doi: 10.1127/0372-8854/2008/0052S2-0077. Leopold, M., Völkel, J., Dethier, D., Williams, M.W., and Caine, N., 2010, Mountain permafrost – a valid archive to study climate change? Examples from the Rocky Mountains Front Range of Colorado, USA: Nova Acta Leopoldina, v. 289, no. 384, p. 281–289. Leopold, M., Williams, M.W., Caine, N., Völkel, J., and Dethier, D., 2011, Internal structure of the Green Lake 5 rock glacier, Colorado Front Range, USA: Permafrost and Periglacial Processes, v. 22, no. 2, p. 107–119, doi: 10.1002/ppp.706. Leopold, M., Völkel, J., Dethier, D., and Williams, M.W., 2013, Changing mountain permafrost from the 1970s to today – comparing two examples from Niwot Ridge, Colorado Front Range, USA: Zeitschrift für Geomorphologie, , no. March, p. 1–21, doi: 10.1127/0372-8854. Li, X., Cheng, G., Jin, H., Kang, E., Che, T., Jin, R., Wu, L., Nan, Z., Wang, J., and Shen, Y., 2008, Cryospheric change in China: Global and Planetary Change, v. 62, no. 3, p. 210–218. Lin, H., Hopmans, J.W., and Richter, D., 2011, Interdisciplinary sciences in a global network of Critical Zone observatories: Vadose Zone Journal. Loke, M.H., 2012, Tutorial.: 2-D and 3-D electrical imaging surveys. Lyle, M., Heusser, L., Ravelo, C., Yamamoto, M., Barron, J., Diffenbaugh, N.S., Herbert, T., and Andreasen, D., 2012, Out of the tropics: the Pacific, Great Basin lakes, and late Pleistocene water cycle in the western United States.: Science, v. 337, no. 6102, p. 1629–33, doi: 10.1126/science.1218390. Madole, R.F., 2012, Holocene alluvial stratigraphy and response to climate change in the Roaring River valley, Front Range, Colorado, USA: Quaternary Research, v. 78, no. 2, p. 197–208, doi: 10.1016/j.yqres.2012.05.005. Marcott, S.A., Clark, P.U., Shakun, J.D., Hostetler, S.W., Brook, E.J., Alder, J.R., Bartlein, P.J., Novak, A.M., Caffee, M.W., and Davis, P.T., 2012, 10Be-Based Chronologies Of Late Pleistocene And Holocene Cirque Glaciation Across Western North America: GSA Annual Convention. Marker, M.E., 1990, Minimum altitudes for former periglacial landforms adjacent to longitude 106°W in Colorado and New Mexico, USA, between latitudes 33°N and 40°N: Arctic, Antarctic, and Alpine Research, v. 22, no. 4, p. 336–374. Markgraf, V., and Scott, L., 1981, Lower timberline in central Colorado during the past 15,000 yr: Geology, p. 0–4, doi: 10.1130/0091-7613(1981)9<231. Marr, J.W., 1967, Data on mountain environments: Front range, Colorado, sixteen sites, 1952 -1953: University of Colorado Press. Mears, B., 1981, Periglacial wedges and the Late Pleistocene environment wyoming’s intermontane Basins: Quaternary Research, v. 198, p. 171–198. Millar, S.W.S., and Nelson, F.E., 2001, Sampling-surface orientation and clast macrofabric in periglacial colluvium: Earth Surface Processes and Landforms, v. 26(5), p. 523–529. Mühll, D.V., Hauck, C., and Gubler, H., 2002, Mapping of mountain permafrost using geophysical methods: Progress in Physical Geography, v. 26, no. 4, p. 643–660, doi: 10.1191/0309133302pp356ra. Muhs, D.R., and Benedict, J.B., 2006, Eolian additions to Late Quaternary alpine soils, Indian Peaks Wilderness Area, Colorado Front Range: Arctic, Antarctic, and Alpine Research, v. 38, no. 1, p. 120–130, doi: 10.1657/1523­0430(2006)038[0120:EATLQA]2.0.CO;2. Nesbitt, I.M., and Dethier, D., 2013, A comparative study of snowmelt-driven water budgets in adjacent alpine basins, Niwot Ridge, Colorado Front Range: Abstracts With Programs, Geological Society of America Northeastern Section, v. 48-6. Noetzli, J., and Mühll, D.V., 2010, Permafrost in Switzerland 2006/2007 and 2007/2008: Glaciological Report Permafrost, no. 8/9. Van Nostrand, R., and Cook, K.L., 1966, Interpretation of resistivity data. Oladunjoye, M., and Sanuade, O., 2012, Thermal diffusivity, thermal effusivity and specific heat of soils in Olorunsogo powerplant, Southwestern Nigeria: arpapress.com, v. 13, no. November, p. 502–521. Oldenburg, D.W., and Li, Y., 1999, Estimating depth of investigation in dc resistivity and IP surveys: Geophysics, v. 64, no. 2, p. 403, doi: 10.1190/1.1444545. Péwé, T., 1983, Alpine permafrost in the contiguous United States: A review: Arctic and Alpine Research, v. 15, no. 2, p. 145–156. Putkonen, J., 1998, Soil thermal properties and heat transfer processes near Ny-Alesund, northwestern Spitsbergen, Svalbard: Polar Research, v. 17, no. 2, p. 165–179, doi: 10.1111/j.1751-8369.1998.tb00270.x. Putnam, S.N., and Chapman, D.S., 1996, A geothermal climate change observatory.: First year results from Emigrant Pass in northwest Utah: Journal of Geophysical Research, v. 101, no. B10. Ray, R., Krantz, W., Caine, N., and Gunn, R., 1983, A model for sorted patterned-ground regularity: Journal of Glaciology, v. 29, no. 102, p. 317–337. Roman, I. Electrical Resistivity of Snow and Ice. Romm, J., 2012, IPCC’s planned obsolescence: fifth assessment report will ignore crucial permafrost carbon feedback!: ThinkProgress. Sandmeier, K.J., 2011, REFLEXW Manual Ver. 6.0, no. 49, p. 532. Schrott, L., and Sass, O., 2008, Application of field geophysics in geomorphology: advances and limitations exemplified by case studies: Geomorphology, v. 93, no. 1­2, p. 55–73, doi: 10.1016/j.geomorph.2006.12.024. Smith, M.W., and Riseborough, D.W., 2002, Climate and the limits of permafrost: a zonal analysis: Permafrost and Periglacial Processes, v. 13, no. 1, p. 1–15. Software, G., 2011, RES2DINVx32 ver. 3.71. Stephenson, D.G., 1987, A procedure for determining the thermal diffusivity of materials: Journal of Building Physics, v. 10, no. 4, p. 236–242, doi: 10.1177/109719638701000404. Vierling, L., 1998, Palynological evidence for late-and postglacial environmental change in central Colorado: Quaternary research, v. 49, no. 2, p. 222–232, doi: 10.1006/qres.1997.1960. Ward, D.J., Anderson, R.S., Guido, Z.S., and Briner, J.P., 2009, Numerical modeling of cosmogenic deglaciation records, Front Range and San Juan Mountains, Colorado: Journal of Geophysical Research, v. 114, no. F1, p. 1–21, doi: 10.1029/2008JF001057. Washburn, A.L., 1967, Instrumental observations of mass wasting in the Mesters Vig District, northeast Greenland: Meddelelser om Gronland, , no. 166, p. 1–297. Washburn, A.L., 1970, Instrumental observations of mass-wasting in an arctic climate: Zeitschrift fur Geomorphologie. Washburn, A.L., 1973, Periglacial processes and environments: Edward Arnold, London. White, S.E., 1971, Rock glacier studies in the Colorado Front Range, 1961 to 1968: Arctic and Alpine Research, v. 3, no. 1, p. 43–64. Wilkinson, B.H., and McElroy, B.J., 2007, The impact of humans on continental erosion and sedimentation: Geological Society of America Bulletin, v. 119, no. 1-2, p. 140– 156, doi: 10.1130/B25899.1. Williams, M.W., Losleben, M., Caine, N., and Greenland, D., 1996, Changes in climate and hydrochemical responses in a high-elevation catchment in the Rocky Mountains, USA: Limnology and Oceanography, v. 41, no. 5, p. 939–946, doi: 10.4319/lo.1996.41.5.0939. Wilson, G.W., Clifton, A.W., Barbour, S.L., and Canada, N.R.C. of, 1999, The emergence of unsaturated soil mechanics: NRC Research Press National Research Council of Canada. Williams, M.W., Knauf, M., Caine, N., Liu, F., and Verplanck, P.L., 2006, Geochemistry and source waters of rock glacier outflow, Colorado Front Range: Permafrost and Periglacial Processes, v. 17, no. 1, p. 13–33, doi: 10.1002/ppp.535. APPENDICES Appendix A Basic Physics Review Since electrical resistivity tomography measures potential differences at points on the Earth’s surface which are produced by directing current to flow through the subsurface, a brief review of the basic electrical concepts seems necessary for this study. Once understanding of these concepts is complete, the determination of resistivity distribution in the subsurface and an interpretation of those results can be attained (Burger et al., 2006). A basic electrical circuit consists of a battery, connecting wires, and a resistor as well as an ammeter to measure current and a voltmeter to measure voltage. A battery maintains a potential voltage difference between its positive and negative terminal, functioning as a power source in moving charge through the circuit. To accomplish this electron flow, the battery moves positive charges from a low potential at its negative terminal to a high potential at its positive terminal. The force which achieves this work is the electromotive force with units of volts. Current is defined as the movement of charges across a given cross-sectional area in a particular time, denoted by the unit amps. Objects in the circuit such as light bulbs, heating lamps, and computer chips act as resistors which limit the ability of an electric current to pass through the circuit. Ohms Law was discovered by the German physicist Georg Simon Ohm, which states that the voltage V is directly proportional to the product of the current I and the resistance R: .... (Eqn. 16) Since various geologic materials such as sand, silt, bedrock, and ice have different resistances to current flow, it is possible to measure current and voltage to calculate resistance and determine the material of the subsurface. Many complications arise, however, since resistance depends on both the material and its dimensions, which are not visible below the surface. For a perfect resistor, with resistance R, length l, and cross- sectional surface area A, it is possible to measure its resistivity, denoted by ., following the equation . ...(Eqn. 17) . Point Current Source Electrical resistivity methods consist of applying current at one location (denoted by C1) and measuring potential at other locations (denoted as P2, P3 etc.). Assuming a uniform material resistivity of ., the current flows radially outward in all directions, appearing to form a hemispherical surface (Fig. 72). Since air’s resistivity (roughly 2 x 1016 .m at 20 °C) is orders of magnitude higher than materials in the earth’s surface (ranging from 1-5*106 .m), electricity is unable to flow upwards. Current distribution is the same everywhere in the earth’s surface a distance r from the current electrode C1, so the potential is also equal at this distance. Using Ohm’s Law and Eqn. 17, the potential difference across a shell with thickness dr can be defined as . .. ............ . (Eqn. 18) . .... single point source of current C1. Equipotential surfaces are shown; the outer two are separated by a distance dr The potential at the point P1 must compared with the potential at a point infinitely far away to calculate its voltage. The easiest way to do this is to integrate Equation 9 over its distance D to infinity: . .. ... .. ...... . . (Eqn. 19) . .. ... ... This is the fundamental equation used in electrical prospecting discussions, which will be expanded and developed for more practical relationships. It is already possible to notice that current flow must be perpendicular to the equipotential surfaces, so it is possible to visualize the direction of current flow from these simple diagrams. Two Current Electrodes After deriving the equation for potential in the earth’s surface given a single point current source, the next step towards understanding the geophysics of ERT is to determine the current flow in a homogenous, isotropic Earth when there are two current electrodes. Here, the current flows from the positive current electrode (the source) to the negative current electrode (the sink) along paths which are not as simple as in the previous discussion. For simplicity, I will derive the case when the both current electrodes lie on the same plane (Figure 73). Equation 10 can be used to determine the potential at point P1, taking into account the effects of both the source at C1 (positive) and the sink at C2 (negative). The potential can then be expressed as: .. .. .... ... . (Eqn. 20) .... .... We can rewrite Eqn. 20 in terms of the x-z coordinate system as: ... . .... . ... . .... (Eqn. 21) .. . . ...... .... ...... .... .. Now it is necessary to determine ...at many points in the x-z plane, draw contours through points of equal potential, and finally draw current flow lines perpendicular to the equipotential surfaces. Although subsurface current distribution is fairly complicated, it results in a simple equation which provides current distribution as a fraction of the total current (Van Nostrand and Cook, 1996). Along a vertical plane halfway between two current electrodes, the fraction of the total current I which penetrates to a depth z for an electrode separation D is given by . .. ..#...... (Eqn. 22) .. As a result, fifty percent of the current is confined above a horizontal plane with a depth of one-half of the current electrode separation. Thus, the greater the electrode separation, the greater the depth to which a given percentage of the current penetrates, but the pooer the resistivity resolution is. Although this equation is derived for a homogenous, isotropic surface, the concepts can still apply in realistic scenarios. Two Potential Electrodes In electrical resistivity surveys, the goal is to measure the potential difference between two points, just as in simple electrical circuits, in order to calculate the apparent resistivity of the ground subsurface. In addition to applying current at multiple points, the physical equations for measuring potential at different points on the ground surface are derived below. Two potential electrodes, P1 and P2, are located on the surface between the current electrodes, C1 and C2, as seen in a Wenner array (Figure 73). Refer to Two Current Electrodes, above, for the derivation of an equation used to determine the potential at a point due to both a source and sink. Here, the potential difference is obtained by determining the potential at P1 and subtracting it from the potential at P2. Using Eqn. 20, it is clear that .... .... .... . and .... . (Eqn. 23 and Eqn. 24) .... .... .... .... Therefore, the potential difference .V equals .... .... .. . ............ . ... . .. .. . . .. (Eqn. 25) .... .... .... .... .. ... .. .. .. .. These derivations and equations show that if there exists a current at a specific amperage with measured spacing between electrodes and some pre-determined potential difference, it is possible to calculate the resistivity of subsurface materials. A Single Horizontal Interface In reality, the subsurface is not a homogenous-isotropic surface, but rather consists of parallel strata (i.e. sedimentary layers) each having their own resistances. It is possible to start by deriving equations for a single horizontal interface, and work towards the case with many horizontal layers. To begin understanding the equations which describe how current is distributed through the subsurface, an equation yielding the fraction of the current which penetrates below the interface is employed. This current fraction can be found from the current fraction, I, electrode spacing, a, depth of interface, z, and resistivities of the materials above and below the interface .. and ..: ... . ........ . .. ........ .......... (Eqn. 26) ... ... ... where ..... ..(Eqn. 27) ..... (Van Nostrand and Cook, 1996). For the case in which ....., such as ice underlying saturated regolith, the material below the interface will have higher resistances to current flow so k-values will be positive. Substantially less current penetrates below the level of the interface compared with the homogeneous case, where k-values are 0, so current flow will avoid poor conductors in favor of good conductors. Similar to the discussion of current flow in a homogeneous, isotropic subsurface, the percentage of current flowing at depth is increased as current electrode spacing is increased. However, the percentage of current penetrating below an interface between two layers with different resistivities is controlled by the relative magnitudes of .. and .. as well as by electrode spacing. By varying the electrode spacing in the field, we can guarantee depths of penetration down to at least the depth we expect permafrost to exist. Appendix B The temperature of each body of water, elevation, easting, and northing (in UTM 18N) of each waypoint measured along Niwot Ridge, in Green Lakes Valley and Fourth of July Valley, and in Gordon Gulch Label Geophysics Line Number Elevation (m) Temperature (°C) Easting (m) Northing (m) 1 3433 10 449181 4433546 2 3427 10.5 449190 4433512 3 3449 3 449092 4433573 4 3448 5 449074 4433576 5 3418 8 449415 4433288 6 3415 13 449450 4433318 7 3417 11 449456 4433342 8 3420 10 449476 4433367 9 3436 10 449492 4433404 10 3446 10 449505 4433441 11 3446 9.5 449520 4433497 12 3460 13.5 449542 4433535 13 3467 14.5 449561 4433586 14 3464 13 449567 4433635 15 3472 12.5 449563 4433677 16 3477 9 449557 4433716 17 3489 9.5 449574 4433718 18 3476 11.5 449563 4433691 19 3488 4 449557 4433734 20 3488 5.5 449577 4433731 21 3486 3.5 449572 4433732 22 3491 1.75 449569 4433751 23 3509 2 449574 4433823 24 3511 3.25 449623 4433830 25 3059 453548 4433110 26 3092 453499 4433191 27 3137 454023 4434080 28 3135 454026 4434074 29 Line 1 Start 3790 445412 4433745 30 Line 1 End 3800 445371 4433776 31 Line 2 Start 3790 445401 4433635 32 Line 2 End 3790 445440 4433605 34 Line 3 Start 3694 446129 4433412 35 Line 3 End 3705 446129 4433383 36 Line 4 End 3629 446535 4433776 37 Line 4 Start 3638 446575 4433746 38 Line 5 Start 3514 449568 4433845 39 Line 5 End 3515 449544 4433840 40 Line 6 Start 3527 449133 4434691 41 Line 6 End 3519 449105 4434731 42 Line 7 Start 3501 449886 4433708 43 Line 7 End 3506 449863 4433664 44 Line 8 Start 3862 446291 4434270 45 Line 8 End 3857 446308 4434281 46 3811 446643 4434390 47 Line 9 Start 3698 447939 4434373 48 Line 9 End 3697 447974 4434409 49 3703 7.3 447939 4434369 50 Line 10 Start Line 14 End 3520 449582 4433843 51 Line 10 End 3519 449557 4433839 52 Line 11 Start 3463 442749 4428682 53 3449 442761 4428728 54 3456 442753 4428698 55 Line 11 End 3446 442765 4428743 56 Line 12 Start 3446 442776 4428741 57 Line 12 End 3449 442731 4428720 58 Line 13 Start 3463 442759 4428683 59 Line 13 End 3467 442795 4428678 60 3428 10.8 442644 4428681 61 3428 0 442685 4428729 62 3429 1.4 442778 4428764 63 3429 2.5 442782 4428767 64 3433 5.6 442807 4428764 65 3436 4.5 442814 4428764 66 Line 14 Start 3525 449656 4433855 67 Line 15 Start 2512 460134 4429197 13TIM Line 15 End 2486 460161 4429263 68 Line 16 Start 2531 460185 4429172 69 Line 16 End 2510 460218 4429240 70 3195 446781 4433717 71 3639 446351 4433723 72 3376 10.3 450405 4433300 73 3378 10 450399 4433316 74 3381 9.7 450397 4433329 75 3384 10 450400 4433340 76 3382 10.6 450399 4433350 77 3383 10.4 450393 4433358 78 3388 9.5 450379 4433376 79 3392 9.1 450390 4433391 80 3392 9.4 450393 4433411 81 3400 9.2 450388 4433422 82 3403 8.6 450384 4433434 83 3410 8.1 450367 4433447 84 3416 7.8 450360 4433466 85 3417 7.4 450351 4433485 86 3419 7.4 450342 4433501 87 3429 7.4 450331 4433519 88 3431 7.7 450329 4433549 89 3435 7.6 450325 4433577 90 3436 7.3 450310 4433589 91 3446 7.4 450309 4433613 92 3444 8.2 450309 4433614 93 3448 3.2 450286 4433623 94 3448 3 450298 4433632 95 3448 5.4 450305 4433631 96 3474 450230 4433754 97 3488 10.2 450175 4433816 98 3492 9 450186 4433841 99 3493 13.2 450208 4433864 100 3494 9.1 450218 4433867 101 3497 9 450198 4433890 102 3506 7.9 450179 4433910 103 3503 6.6 450181 4433915 104 3511 10.6 450167 4433971 105 3517 10.7 450040 4433856 106 3513 9.9 449911 4433762 107 3512 3.4 449781 4433832 108 3539 11.2 449614 4433985 109 3650 8.1 448416 4434131 110 3680 9.2 448163 4434221 111 3681 7.2 448153 4434240 112 3679 11.2 448121 4434241 113 3691 8 447984 4434316 114 3701 9.1 447931 4434374 115 3701 10.2 447932 4434404 116 3588 6.8 448175 4434681 117 3539 3.6 448362 4434690 118 3531 3.5 448472 4434687 119 3532 5 448854 4434637 120 3533 4 448875 4434636 121 3535 4.1 448875 4434633 122 3558 3.7 449191 4434579 123 3473 16.3 450211 4433706 124 3506 8.2 449894 4433742 125 3503 8 449881 4433708 126 3495 9.8 449869 4433677 127 3492 3.6 449748 4433663 128 3485 8.5 449742 4433598 129 3476 8.8 449746 4433552 130 3481 8.1 449842 4433497 131 3481 7.3 449721 4433593 132 3491 8.7 449677 4433699 133 3508 3.1 449608 4433826 134 3516 1.8 449562 4433817 135 3510 4.3 449333 4433769 136 3499 2.7 449021 4433708 137 3502 4 449010 4433693 138 3540 2.8 448927 4433759 139 3543 2 448920 4433756 140 3664 10.7 448048 4434023 141 3660 6.4 448047 4434022 142 3666 7.8 448010 4434056 143 3669 8 447971 4434091 144 3662 7.1 447887 4434070 145 3663 8.4 447862 4434062 146 3665 7.4 447852 4434075 147 3666 7.7 447820 4434094 148 3662 7.2 447815 4434111 149 3667 12.2 447786 4434142 150 3666 2.7 447758 4434135 151 3665 7.2 447729 4434115 152 3661 4.4 447652 4434173 153 3661 5.8 447648 4434170 154 3658 7 447647 4434168 155 3716 7.8 447690 4434432 156 3716 10.1 447653 4434448 157 3724 7.7 447632 4434455 158 3743 6.8 447487 4434525 159 3740 7.8 447472 4434531 160 3741 7.6 447443 4434548 161 3698 5.4 447923 4434455 162 3686 10.5 448089 4434409 163 3678 11 448102 4434403 164 3676 13.6 448113 4434401 165 3674 10.2 448192 4434373 166 3669 11.6 448201 4434377 167 3667 8.8 448222 4434374 168 3662 11.7 448260 4434337 169 3660 1.1 448261 4434312 170 3659 6.6 448380 4434249 171 3656 13.5 448387 4434262 172 3647 17.2 448400 4434252 173 3638 2.5 448472 4434276 174 3631 9.1 448476 4434311 175 3636 18.4 448527 4434296 176 3619 12.4 448767 4434177 177 3536 17.2 449616 4433985 178 3526 7.1 449622 4433896 179 3331 16.5 451047 4433040 180 3315 14.4 451149 4432965 181 2421 1.5 449569 4433751 182 2320 2 449580 4433735 183 2208 5 449532 4433736 184 2135 11 449118 4434713 185 2133 10 449113 4434720 186 2122 10.6 446301 4434276 187 2118 10.4 447946 4434379 188 2139 11.9 447953 4434387 189 2136 10.4 447960 4434394 190 2148 11 447967 4434402 191 2152 12 447974 4434409 192 2171 8.4 447981 4434416 193 2164 6 449476 4433367 194 3414 2.7 449206 4433487 195 GPR Line 1 Start 3564 449455 4434213 196 GPR Line 1 End 3558 449481 4434213 197 GPR Line 2 Start 3563 449459 4434202 198 GPR Line 2 End 3560 449475 4434222 Appendix C © th th Figure 74. RES2DINVinversion image showing a. 4of July 1-2 Wenner array. b. 4of July 1-2 dipole-dipole array 153 © th th 154 © th th th 155 156 Figure 78. RES2DINV© inversion image showing a. Gordon Gulch 2 Wenner array. b. Gordon Gulch 2 dipole-dipole array 157 158 Figure 80. RES2DINV© inversion image showing a. Green Lakes Moraine Wenner-half array. b. Green Lake Moraine Schlumberger array 159 Figure 81. RES2DINV© inversion image showing a. Green Lakes North Slope Schlumberger array. b. Green Lakes North Slope dipole-dipole array 160 Figure 82. RES2DINV© inversion image showing a Green Lakes South Slope Schlumberger array. b. Green Lakes South Slope Wenner array 161 162 163 164 165 166 167 168 Appendix D Input parameters used during HFLUX modelling of water temperatures in Como Creek Distance from source (m) Measured water Temperature (°C) Discharge (m3/s) Ground Water Temperature (°C) 0 3 0.00010001 3 15 5.4 0.00025003 3.06 22.07 3.2 0.00032075 3.08828 39.53 8.2 0.00049542 3.15812 40.51 7.4 0.00050523 3.16204 64.53 7.3 0.00074558 3.25812 83.74 7.6 0.00093784 3.33496 112.02 7.7 0.00122094 3.44808 142.09 7.4 0.00152206 3.56836 163.18 7.4 0.0017333 3.65272 181.53 7.4 0.00191714 3.72612 202.56 7.8 0.00212786 3.81024 222.81 8.1 0.00233081 3.89124 244.21 8.6 0.00254533 3.97684 256.86 9.2 0.00267216 4.02744 268.94 9.4 0.00279329 4.07576 289.17 9.1 0.00299618 4.15668 307.76 9.5 0.00318265 4.23104 330.57 10.4 0.00341151 4.32228 340.57 10.6 0.00351185 4.36228 350.62 10 0.00361271 4.40248 362.02 9.7 0.00372713 4.44808 375.17 10 0.00385913 4.50068 392.26 10.3 0.0040307 4.56904 Appendix E Calculating equations of temperatures at depth for Hopkins Memorial Forest . . .... Using: #.. #.............. ........ .... . ... Where: #.: Calculated Temperature at Depth z and Time t #..: Mean Annual Surface Temperature Q: Local Geothermal Heat Flow k: Thermal Conductivity of Local Rock Tamp: Amplitude of Thermal Oscillation . . * .. z : Depth at which .. . Surface air temperatures (*input data as grids called AirTemp, Soil10, Soil80, WaterTemp*) (*fit sinusoidal model to data*) model3=a3+b3 Sin[2p/365 t+d3]; fit3=FindFit[AirTemp,model3,{a3,b3,d3},t]; func3=Function[{t},Evaluate[model3/.fit3]]; (*plot the curve which was calculated as the best fit above*) Plot[func3[t],{t,0,365},Epilog.Map[Point,AirTemp],PlotRange.{-10,30}] a=Plot[func3[t],{t,0,365},PlotRange.{-10,30},PlotStyle­>Thickness[.005],AxesLabel.{"Julian Day","Temperature °C"}]; (*calculate statistical significance of regression curve*) nlm3=NonlinearModelFit[AirTemp,model3,{a3,b3,d3},t]; nlm3[{"RSquared","ParameterTable"}] {0.902931,{ {, Estimate, Standard Error, t-Statistic, P-Value}, {a3, 9.14422, 0.209696, 43.607, 2.25144×10-146}, {b3, -11.4241, 0.296206, -38.5682, 1.89922×10-130}, {d3, 1.2885, 0.0259891, 49.5783, 1.06179×10-163} }} Temperatures at 10 cm depth model2=a2+b2 Sin[2p/365 t+d2]; fit2=FindFit[Soil10Temp,model2,{a2,b2,d2},t]; func2=Function[{t},Evaluate[model2/.fit2]]; ß=Plot[func2[t],{t,0,365},PlotRange.{­10,30},PlotStyle.{Red,Thickness[.005]},AxesLabel.{"Julian Day","Temperature °C"}]; nlm2=NonlinearModelFit[Soil10Temp,model2,{a2,b2,d2},t]; nlm2[{"RSquared","ParameterTable"}] {0.988589,{ {, Estimate, Standard Error, t-Statistic, P-Value}, {a2, 11.0588, 0.0731866, 151.104, 8.463223207955×10-330}, {b2, -9.65359, 0.103415, -93.348, 7.73132×10-256}, {d2, 1.0979, 0.0107305, 102.316, 9.03994×10-270} }} Temperatures at 80 cm depth model1=a1+b1 Sin[2p/365 t+d1]; fit1=FindFit[Soil80Temp,model1,{a1,b1,d1},t]; func1=Function[{t},Evaluate[model1/.fit1]]; Plot[func1[t],{t,0,365},Epilog.Map[Point,Soil80Temp],PlotRange.{-10,20}] .=Plot[func1[t],{t,0,365},PlotRange.{­10,20},PlotStyle.{Orange,Thickness[.005]},AxesLabel.{"Julian Day","Temperature °C"}]; nlm1=NonlinearModelFit[Soil80Temp,model1,{a1,b1,d1},t]; nlm1[{"RSquared","ParameterTable"}] {0.99754,{ {, Estimate, Standard Error, t-Statistic, P-Value}, {a1, 10.7367, 0.0310892, 345.353, 4.134574394972×10-459}, {b1, -7.39143, 0.0439592, -168.143, 2.091332723705×10-346}, {d1, 0.831452, 0.00594936, 139.755, 1.0690024739944×10-317} }} Temperatures at 515 cm depth model4=a4+b4 Sin[2p/365 t+d4]; fit4=FindFit[WaterTemp,model4,{a4,b4,d4},t]; func4=Function[{t},Evaluate[model4/.fit4]]; Plot[func4[t],{t,0,365},Epilog.Map[Point,WaterTemp],PlotRange.{­10,30},AxesLabel.{"Julian Day","Temperature °C"}] ß=Plot[func4[t],{t,0,365},PlotRange.{­10,30},PlotStyle.{Green,Thickness[.005]},AxesLabel.{"Julian Day","Temperature °C"}] nlm4=NonlinearModelFit[WaterTemp,model4,{a4,b4,d4},t]; nlm4[{"RSquared","ParameterTable"}] {0.999771,{ {, Estimate, Standard Error, t-Statistic, P-Value}, {a4, 9.59209, 0.00973501, 985.318, 9.02045137081288×10-501}, {b4, 1.95895, 0.0118459, 165.369, 8.53787×10-283}, {d4, 2.72153, 0.00770901, 353.033, 3.466455534848509×10-375} }} Using best fit curves, I found the following four equations AirTemp:9.14422-11.4241 Sin[1.2885+(2pt)/365] with R2: 0.902931 Soil10:11.0588-9.65359 Sin[1.0979+(2pt)/365] with R2:0.988589 Soil80: 10.7367-7.39143 Sin[0.831452+(2pt)/365] with R2:0.99754 Water:9.59209+1.95895 Sin[2.72153+(2pt)/365] with R2:0.999771 Displaying data with and without best fit curves Show[AirTempPlot,Soil10Plot,Soil80Plot,WaterTempPlot] Show[a,ß,.,ß,AirTempPlot,Soil10Plot,Soil80Plot,WaterTempPlot] Using temperature equations to generate measured profile animation (*plotting the background net of thin black lines based on the air temperature, as well as dashed blue lines at 40 cm depth*) Graph=ParametricPlot[{{Table[{9.1442+.001316z +11.414Exp[-z /250]Sin[(2p t)/365-z /250],-z},{t,0,365,10}]},{z,-40},{-z,-40}},{z,0,550},AspectRatio.1/GoldenRatio, PlotRange.{{-5,22},{0,­550}},PlotStyle.{{Black,Thin},{Blue,Dashed},{Blue,Dashed},{Red,Thick},{Red,Thic k},{Red,Thick}},ImageSize.{1000,600},FrameLabel.{Text[Style["Temperature (°C)",FontSize.30]],Text[Style["Depth (cm)",FontSize.30]]},AxesOrigin.{0,0},Frame.{True,True,True,True},FrameStyle .{{Directive[Black,20],Black},{Directive[Black,20],Black}}]; (*define functions for temperatures at the surface, 10 cm, 80 cm, and 515 cm depth*) xSurface[t_]=9.1442_-11.424 Sin[1.2885_+(2p t)/365]; ySurface[t_]=0; xTen[t_]=11.0588_-9.6536 Sin[1.098_+(2p t)/365]; yTen[t_]=-10; xEighty[t_]=10.7367_-7.3914 Sin[0.8314_+(2p t)/365]; yEighty[t_]=-80; xFFT[t_]=9.5920_+1.9589 Sin[2.7215_+(2p t)/365]; yFFT[t_]=-515; (*join the known temperature values together with a curve*) TempCurve[t_]:=JoinedCurve[{{xSurface[t],ySurface[t]},{xTen[t],yTen[t]}}] (*plot the surface temperature points*) surfacetrajplotSurface=ParametricPlot[{xSurface[t],ySurface[t]},{t,0,365},Prolog.Inset [Point,{0,0},Automatic,0.7],Axes.False,Background.White,PlotStyle.Yellow,PlotR ange.{{-10,30},{-520,0}},AspectRatio.GoldenRatio]; (*plot the temperature points at 10 cm depth*) trajplotTen=ParametricPlot[{xTen[t],yTen[t]},{t,0,365}, Prolog.Inset[Point,{0,­10},Automatic,0.7],Axes.False,Background.White,PlotStyle.Yellow]; (*plot the temperature points at 80 cm*) trajplotEighty=ParametricPlot[{xEighty[t],yEighty[t]},{t,0,365}, Prolog.Inset[Point,{0,­80},Automatic,0.7],Axes.False,Background.White,PlotStyle.Yellow]; (*plot the temperature points at 515 cm depth*) trajplotFFT=ParametricPlot[{xFFT[t],yFFT[t]},{t,0,365},Prolog.Inset[Point,{0,­515},Automatic,0.7],Axes.False,Background.White,PlotStyle.Yellow]; (*create an animation showing the temperature points oscillating with time*) v=Manipulate[Show[Graph,trajplotSurface, Graphics[{Black,PointSize[.02],Point[{xSurface[t],ySurface[t]}]}], Graphics[{Black,PointSize[.02],Point[{xTen[t],yTen[t]}]}], Graphics[{Black,PointSize[.02],Point[{xEighty[t],yEighty[t]}]}], Graphics[{Black,PointSize[.02],Point[{xFFT[t],yFFT[t]}]}], Graphics[{Red,Thickness[.005],JoinedCurve[Line[{{xSurface[t],ySurface[t]},{xTen[t],y Ten[t]}}],CurveClosed.True]}], Graphics[{Red,Thickness[.005],JoinedCurve[Line[{{xTen[t],yTen[t]},{xEighty[t],yEigh ty[t]}}],CurveClosed.True]}], Graphics[{Red,Thickness[.005],JoinedCurve[Line[{{xEighty[t],yEighty[t]},{xFFT[t],yF FT[t]}}],CurveClosed.True]}]], Dynamic[StringJoin[{"Julian Day = ",ToString[Evaluate[N[t,1]]]}] ] ,{t,0,365/2}] (*export the animation*) Export["Hopkins Temps.mov",v,"FrameRate".60] Using temperature equations to generate measured temperature-depth profile (*create lines at 50 and 515 cm depth*) Graph2=ParametricPlot[{{-z,40}},{z,0,515},AspectRatio.1/GoldenRatio, PlotRange.{{-5,22},{0,-515}}, PlotStyle.{{Red,Thick}}, ImageSize.{1000,600}, FrameLabel.{Text[Style[ "Temperature (°C)",FontSize.30]],Text[Style["Depth (cm)", FontSize.30]]},AxesOrigin.{0,0},Frame.{True,True,True,True}, FrameStyle.{ {Directive[Black,20],Black}, {Directive[Black,20],Black}}]; (*create a table of the background graphics, and joined curves connecting all temperature points*) Show[Table[Show[Graph2,trajplotSurface, Graphics[{Black,Thin,JoinedCurve[Line[{{xSurface[t],ySurface[t]},{xTen[t],yTen[t]}}], CurveClosed.True]}], Graphics[{Black,Thin,JoinedCurve[Line[{{xTen[t],yTen[t]},{xEighty[t],yEighty[t]}}], CurveClosed.True]}], Graphics[{Black,Thin,JoinedCurve[Line[{{xEighty[t],yEighty[t]},{xFFT[t],yFFT[t]}}], CurveClosed.True]}]] ,{t,0,365,10}]] Calculating temperature-depth profile throughout the year at Hopkins Memorial Forest (*create a background graphic showing lines from the surface temperature readings*) Graph=ParametricPlot[{{Table[{9.1442+.001316z +11.414Exp[-z /250]Sin[(2p t)/365-z /250],-z}, {t,0,365,10}]},{z,-54},{-z,-54},{(z-32.44)/2002.55, -10}, {(z+9674.44)/3007.85,-80}, {(z+43955)/4300.85,-515}}, {z,0,550},AspectRatio.1/GoldenRatio, PlotRange.{{-5,22},{0,-550}}, PlotStyle.{{Black,Thin},{Blue,Dashed},{Blue,Dashed},{Red,Thick},{Red,Thick}, {Red,Thick}},ImageSize.{1000,600},FrameLabel.{Text[Style["Temperature (°C)", FontSize.30]],Text[Style["Depth (cm)",FontSize.30]]}, AxesOrigin.{0,0},Frame.{True,True,True,True}, FrameStyle.{ {Directive[Black,20],Black}, {Directive[Black,20],Black}}]; Show[Graph] Calculating temperature-time graphs for various depths on Niwot Ridge (*inputting temperature equations for different depths based on the surface temperature measurements calculated above*) a1[t_]:=-3.7-Exp[-(1000*1)/z]21.4Cos[2p t-(1000*1)/z]; a2[t_]:=-3.7-Exp[-(1000*2)/z]21.4Cos[2p t-(1000*2)/z]; a3[t_]:=-3.7-Exp[-(1000*3)/z]21.4Cos[2p t-(1000*3)/z]; a4[t_]:=-3.7-Exp[-(1000*4)/z]21.4Cos[2p t-(1000*4)/z]; a5[t_]:=-3.7-Exp[-(1000*5)/z]21.4Cos[2p t-(1000*5)/z]; a6[t_]:=-3.7-Exp[-(1000*6)/z]21.4Cos[2p t-(1000*6)/z]; a9[t_]:=-3.7-Exp[-(1000*9)/z]21.4Cos[2p t-(1000*9)/z]; a16[t_]:=-3.7-Exp[-(1000*16)/z]21.4Cos[2p t-(1000*16)/z]; (*plot the different temperatures at depth as a function of time*) Needs["PlotLegends`"] Plot[{2,-3.7-21.4Cos[2p t],a1[t], a3[t],a9[t],a16[t]},{t,0,2}, PlotStyle.{{Thickness[.007],Red}, {Thickness[.007],Orange}, {Thickness[.007],Yellow}, {Thickness[.007],Green}, {Thickness[.007],Blue}}, PlotRange.{-30,20},Ticks.{{0,.5,1,1.5,2},{-30,-25,-20,-15,-10,-5,0,5,10,15,20}}, AxesLabel.{Style["Time (Years)",Large], Style[ "Temperature °C",Large]}, LabelStyle.Large,ImageSize.{1000,1000},AxesStyle.Thickness[.007]] Fitting pit temperature measurements on Niwot Ridge to model (*inputting data*) Data={{0,19.2},{20,13},{40,14.8},{60,12.7},{80,12.6},{100,7.6},{120,7},{140,6.5},{1 50,6.8}}; (*creating a list plot of the data*) LPlot=ListPlot[Data,PlotRange->{{0,400},{0,20}}]; (*fitting data to a curve*) model=Qk z+19.2Exp[-z/zs]; fit=FindFit[Data,model,{Qk,zs},z]; (*calculating statistical parameters based on fit*) lm=NonlinearModelFit[Data,Qk z+19.2Exp[-z/zs],{Qk,zs},z]; lm["ParameterTable"]; lm["RSquared"] modelf=Function[{z},Evaluate[model/.fit]] (*plotting temperatures and best fit curve*) Temps=Show[ListPlot[Data,PlotRange­>{{0,400},{0,20}},PlotStyle.{Black,Directive[PointSize[Medium]]}],Plot[modelf[z],{ z,0,400}],AxesLabel.{"Depth(cm)","Temperature (°C)"}] Generating temperature-depth profile throughout the year (*animating the temperature-depth profile throughout the year*) Manipulate[ParametricPlot[{{z,-208.5},{-z,-208.5},{2.1+.00391311 z +15Exp[-z .0080903] Sin[(2p t)/365-z .0080903],-z}, Table[{2.1+.00391311 z +15Exp[-z .0080903]Sin[(2p t)/365-z .0080903],-z}, {t,0,365,10}]}, {z,0,500},AspectRatio.1/GoldenRatio, PlotRange.{{-15,20},{0,-500}}, PlotStyle.{{Blue,Dashed},{Blue,Dashed},{Red,Thickness[.007]},{Black,Thin}}, ImageSize.{1000,600},FrameLabel.{Text[Style["Temperature (°C)",FontSize.30]], Text[Style["Depth (cm)", FontSize.30]]},AxesOrigin.{0,0}, Frame.{True,True,True,True}, FrameStyle.{{Directive[Black,20],Black}, {Directive[Black,20],Black}}],{t,0,365,1}]