THE CONTRIBUTION OF FLUVIAL PROCESSES TO THE FORMATION OF CRYOPLANATION TERRACE S : THE ROLE OF PERIGLACIAL SORTED STRIPES By Raven Jezell Mitchell A THESIS Submitted to Michigan State University i n partial fulfillment of the requirements f or the degree of Geography Master of Science 2020 A BSTRACT THE CONTRIBUTION OF FLUVIAL PROCESSES TO THE FORMATION OF CRYOPLANATION TERRACES: THE ROLE OF PERIGLACIAL SORTED STRIPES By Raven J ezell Mitchell The mechanisms of material excavat ion and remov al from large periglacial landforms known as cryoplanation terraces (CTs) have been debated for over a century. These unknowns hinder acceptance of a unified understanding of CT formation. The research reported in this thesis links sorted stripes a type of p eriglacial patterned ground frequently encountered on CT treads to the hydrologic connectivity of periglacial hillslopes, which ha s not been considered in the context of CT development. Traditional interpretations hold that the presence of sorted patterned ground indicat es geomorphic quiescence , a view that has contributed to the dismissal of these features as a factor in the creation of periglacial topography. This thesis addresses the geomorphic role of sorted stripes as fluvial features by investigating their hydrologic effectiveness in removing weathered material from CT treads. Process - focused investigations and watershed morphometric analysis were conducted on thee cryoplanation terraces in an active upland periglacial environment near Atlin, British Columbia, Canada. Results demonstrate the landscape - scale spatial organization and geomorphic effectiveness of sorted stripe network s for transporting water and suspended sediment across CT surfaces. A qualitative model of sediment production and t ransportation is presented that: 1) outlines erosional processes responsible for CT formation; and 2) defines the distinct hydrologic - geomorphic imprint imparted by sorted stripes on periglacial hillslopes. Copyright by RAVEN JEZELL MITCHELL 2 020 iv ACKNOWLEDG MENTS The thesis is the culmination of two years of hard work, exciting adventures, and new experiences, and represents my most rigorous academic pursuit to date. Without the help from many, this work would have been entire ly impossible and significantly less fulfilling. I would like to thank the members of my committee, Drs. Alan Arbogast and Ashton Shortridge for their careful and patient editing and for their willingness to lend a helping hand when needed. A special acknowledgment goes to my advisor, Dr. Fritz Nelson, who championed me throughout my degree and whose unwavering support and ad vising I will cherish for the rest of my life. To Dr. Kelsey Nyland, thank you for your patience, your mentorship, your help during my very first field season, and most of all, thank you for your continued belief in me. I also acknowledge Dr. Randall Schae tzl for the use of his laboratory facilities, Christopher Baish and Jarrod Knauff for help in the lab, and my fellow geography students for their comradery. Acknowledgment is extended to the Department of Geography, Environment, and Spatial Sciences and t o the College of Social Science for their generous funding over the years. Thank you to the Circumpolar Active Layer Monitoring (CALM) program, for affording me the opportunity to travel to some of the most beautiful and remote parts of the Arctic, for exp osing me to permafrost research, and especially for the beautiful friendships forged through the months of fieldwork. I would also like to express my gratitude to the Juneau Icefield Research Program for use of their research facilities in British Columbia , which has enabled the work of myself and others since the 1970s. Lastly, I would like to acknowledge my family. To my mother, grandmother, and little sister: thank you all for being my very own support team, offering encouragement when I v needed it most. To my partner, Ally: thank you for wearing the hats of thesis edit or, life coach, cheer team, and for always supporting my dreams. The love from you all was limitless. Countless others assisted me over the last two years in innumerable ways to them and to those named above, thank you ! vi T ABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ................................ ix LIST OF FIGURES ................................ ................................ ................................ ............................... x Chapter 1 Introduction ................................ ................................ ................................ ................... 1 1.1 Brief History of Cryoplanation Research ................................ ................................ ............... 3 1.1.1 First encounters. ................................ ................................ ................................ ............. 3 1.1.2 Introduction of periglacial geomorphology ................................ ................................ .... 4 1.1.3 Early conceptions of the nivation hypothesis ................................ ................................ 5 1.1.4 Periglacial geomorphology gains traction ................................ ................................ ...... 6 1.1.5 Definitive studies on cryoplanation terraces ................................ ................................ . 7 1.1.6 Quantitative studies ensue ................................ ................................ ............................. 8 1.2 Revisiting the Nivation Hypothesis ................................ ................................ ....................... 9 1.2.1 Scrutiny ................................ ................................ ................................ ........................... 9 ................................ ......................... 10 Chapter 2 Hydrologic Significance of Sorted Patterned Ground ................................ .................. 12 2.1 Periglacial Hydrology ................................ ................................ ................................ ........... 13 2.1.1 Increasing interest in periglacial hydrology ................................ ................................ .. 14 2.1.2 Characteristics of periglacial watersheds ................................ ................................ ..... 14 2.1.3 Sorted patterned ground ................................ ................................ .............................. 15 2.2 Sorted Stripes as Fluvial Features ................................ ................................ ....................... 15 2.2.1 Sorted patterned ground formation ................................ ................................ ............ 16 2.2.2 Overlooked features ................................ ................................ ................................ ..... 17 2.2.3 Previous work ................................ ................................ ................................ ............... 17 2.2.4 Hillslope connectivity. ................................ ................................ ................................ ... 19 2.3 Conclusion: Sorted Stripes, Nivation, and Cryoplanation Terraces ................................ .... 22 Chapter 3 Study Area ................................ ................................ ................................ .................... 24 3.1 The Greater Atlin Area ................................ ................................ ................................ ........ 26 3.1.1 Geographic description. ................................ ................................ ............................... 26 3.1. 2 Geology. ................................ ................................ ................................ ........................ 26 3.2 The Cathedral Massif ................................ ................................ ................................ ........... 26 3.2.1 Physical description. ................................ ................................ ................................ ..... 26 3.2.2 Geology. ................................ ................................ ................................ ........................ 27 3.2.3 Climate. ................................ ................................ ................................ ......................... 28 3.2.4 Glacial history. ................................ ................................ ................................ .............. 28 3.3 Frost Ridge ................................ ................................ ................................ ........................... 29 3.3.1 Physical description. ................................ ................................ ................................ ..... 29 vii 3.3.2 Climate. ................................ ................................ ................................ ......................... 29 3.3.3 Glacial history. ................................ ................................ ................................ .............. 29 3.3.4 Hydrology. ................................ ................................ ................................ ..................... 30 3.3.5 Periglacial features. ................................ ................................ ................................ ...... 30 3.4 Previous Work on Frost Ridge ................................ ................................ ............................. 31 3.5 Conclusion ................................ ................................ ................................ ........................... 32 Chapter 4 Statement of Problem and Hypothesis ................................ ................................ ........ 33 4.1 Statement of Problem ................................ ................................ ................................ ......... 33 4.2 Hypothesis ................................ ................................ ................................ ........................... 34 Chapter 5 Process Investigations on Two Incipient Cryoplanation Terraces ............................... 36 5.1 Process - focused Geomorphology ................................ ................................ ....................... 36 5.1.1 Process work in geomorphology. ................................ ................................ ................. 37 5.1.2 Suggestions for process - based investigations on cryoplan ation terraces. .................. 38 5.2 Field Investigations on Frost Ridge ................................ ................................ ..................... 39 5.2.1 Field methods. ................................ ................................ ................................ .............. 40 5.2.2 In lab calculations. ................................ ................................ ................................ ........ 43 5.2.3 Laboratory methods. ................................ ................................ ................................ .... 46 5.3 Results ................................ ................................ ................................ ................................ . 47 5.3.1 Monitoring data. ................................ ................................ ................................ ........... 48 5.3.2 Thermal diffusivity calculations. ................................ ................................ ................... 51 5.3.3 Transported water/sediment data ................................ ................................ ............... 52 5.4 Discussion ................................ ................................ ................................ ............................ 54 5.5 Conclusion ................................ ................................ ................................ ........................... 55 Chapter 6 Network analysis of a Sorted Patterned Ground Field ................................ ................ 58 6.1 Remote Sensing in Geomorphology ................................ ................................ .................... 59 6.1.1 Data acquisition. ................................ ................................ ................................ ........... 59 6.1.2 Data. ................................ ................................ ................................ .............................. 60 6.1.3 Applications of remote sensing models in periglacial geomorphology. ...................... 61 6.2 Drainage Network Analysis ................................ ................................ ................................ . 63 6.2.1 Sorted patterned ground drainage channels. ................................ .............................. 63 6.2.2 Drainage network Analysis. ................................ ................................ .......................... 64 6.2.3 Hydrologic flow models. ................................ ................................ ............................... 66 6.2.4 Applications of flow modeling in geomorphology. ................................ ...................... 68 6.3 Network Analysis of a Sorted Patterned Ground Field ................................ ....................... 69 6.3.1 Study area. ................................ ................................ ................................ .................... 70 6.3.2 Study area data acquisition. ................................ ................................ ......................... 71 6.3.3 Network analysis inputs. ................................ ................................ .............................. 74 6.3.4 Comparison of digitized and automated drainage network extract ions. .................... 77 6.3.5 Network analysis results. ................................ ................................ .............................. 81 6.4 Discussion ................................ ................................ ................................ ............................ 82 viii 6.4.1 Bifurcation ratio. ................................ ................................ ................................ ........... 82 6.4.2 Stream length ratio. ................................ ................................ ................................ ...... 84 6.4.3 Drainage frequency. ................................ ................................ ................................ ..... 84 6.4.4 Drainage density. ................................ ................................ ................................ .......... 85 6.4.5 Relief ratio. ................................ ................................ ................................ ................... 86 6.5 Conclusion ................................ ................................ ................................ ........................... 87 Chapter 7 Synthesis and Conclusions ................................ ................................ ........................... 90 7.1 Summary of Research ................................ ................................ ................................ ......... 90 7.1.1 Field - based process investigations. ................................ ................................ .............. 90 7.1.2 Sorted patterned ground network analysis. ................................ ................................ 92 7.2 Synthesis ................................ ................................ ................................ .............................. 94 7.2.1 The role of fluvial action in cryoplanation terrace formation. ................................ ..... 94 7.2.2 The conveyor system model. ................................ ................................ ........................ 94 ....................... 97 7.3 Conclusions ................................ ................................ ................................ ........................ 100 7.4 Recommendations for Future Work ................................ ................................ ................. 101 7.4.1 Long - term process investigations. ................................ ................................ .............. 101 7.4.2 Broader implications of Frost Ridge periglacial studies. ................................ ............ 102 APPENDIX ................................ ................................ ................................ ................................ .... 104 REFERENCES ................................ ................................ ................................ ................................ 107 ix LIST OF TABLES Table 5.1 Temperature logger data breakdown. ................................ ................................ .......... 51 Table 5.2 Input parameters for determination of apparent thermal diffusivity. ......................... 52 Table 5.3 Sediment trough numerical particle size analysis results. ................................ ............ 54 Table 6.1 Field site topographic information. ................................ ................................ .............. 71 Table 6.2 Ancillary drainage basin data. ................................ ................................ ....................... 81 Table 6.3 Quantitative drainage basin characterization. ................................ .............................. 82 Table A.1 Chapter 5 data logging equipment information. ................................ ........................ 105 Table A.2 Chapter 6 GIS tools information. ................................ ................................ ................ 106 x LIST OF FIGURES Figure 1.1 Examples of characteristic cryoplanation terraces at Eagle Summit, AK. ..................... 2 Figure 1.2 Diagrammatic model of nivation - driven scarp retreat of cryoplanation terrace formation. ................................ ................................ ................................ ................................ ....... 8 Fi gure 2.1 Sketch of water movement through sorted patterned ground. ................................ . 19 Figure 2.2. Organizational similarities between sorted stripes and water tracks. ....................... 21 Figure 3.1 The Juneau Icefield. ................................ ................................ ................................ ..... 25 Figure 3.2 Orientation of Frost Ridge. ................................ ................................ .......................... 27 Figure 5.1 Oblique photo of Frost Ridge sorted stripe with silt fan. ................................ ............ 40 Figure 5.2 Site arrangement on Frost Ridge. ................................ ................................ ................ 42 Figure 5.3 Sediment trough and mast installation on cryoplanation terrace. ............................. 45 Figure 5.4 Field activities. ................................ ................................ ................................ ............. 46 Figure 5.5 Equipment orientation on Frost Ridge. ................................ ................................ ....... 47 Figure 5.6 Field data logger records. ................................ ................................ ............................ 50 Figure 5.7 Particle size analysis of sediment trough data. ................................ ........................... 53 Figure 6.1 Panoramic view of Frost Ridge and inset photo of sorted patterned ground features. ................................ ................................ ................................ ................................ ....................... 71 Figure 6.2 Aerial survey data. ................................ ................................ ................................ ....... 73 Figure 6.3 Compound filter map. ................................ ................................ ................................ .. 75 Figure 6.4 Drainage network identification results. ................................ ................................ ..... 77 Figure 6.5 Stream ordering results. ................................ ................................ .............................. 80 ................................ ................................ .............................. 99 1 Chapter 1 Introduction Cryoplanation terraces (CTs) are large upland periglacial landforms resembling giant staircases, with series of gently sloping treads bounded by ascending steep risers (Figure 1.1). Although the formative mechanisms related to CT development are better understood today than they were at the onset of the 20th century when CTs were first documented, some aspects of CT formation and development remain mired in confusion (Ballantyne, 2018, 221) Recent research has confirmed the importance of CT tread expansion through parallel retreat of scarps, demonstrating the time - transgressive nature of CT formation (Matthews et al., 2019; Nyland & Nelson, 2020). The most widely accepted hypothesis of tread expansion is the nivation hypothesis, which emphasi zes the role of perennial snow cover in scarp erosion and subsequent transportation of snowbank meltwater and weathering products (Ballantyne, 2018, 221). Weathering near the snowpatch margin has been substantiated in recent years, and qualitative models o f time - transgressive scrap retreat have been proposed (e.g. Kelsey E. Nyland & Nelson, 2020a) while the mass movement/transportation component of nivation on CT slopes has received little more than cursory investigation (Queen, 2018, 41; Nyland, 2019, 59 ; Thorn & Hall, 2002a, 545). Some geomorphologists have argued that, under the nivation hypothesis of CT formation, rampart - like accumulations of sediment should exist on subjacent treads near the scarp - tread junctions, representing past scarp positions. No such accumulations have been documented on cryoplanation terraces in unglaciated Beringia (Ballantyne, 2018, and references therein) and the absence of such ramparts has been used to question the nivation hypothesis of CT formation. 2 Figure 1.1 Exampl es of characteristic cryoplanation terraces at Eagle Summit, AK . Cryoplanation terraces (top and bottom) typical of upland, unglaciated Beringia. One tread (solid black lines) and its accompanying scarp (dotted black lines) comprise a cryoplanation terrace unit, labeled. Treads range in size from 10 to several hundred meters in both length and width, with slopes between 1 - 5°; scarp height ranges between 1 - 75 meters, with slopes between 15 - 40° (Reger & Péwé, 1976; Ballantyne, 2018, 220). Here, one CT unit i s approximately 200 m in the horizontal dimension. Photos taken in July 2018 by R.J. Mitchell. Some workers have alluded to the removal of eroded material in sorted stripes a type of periglacial patterned ground (e.g., Taylor, 1955; Demek, 1969; Thorn & Hall, 2002b; Queen, 2018; Nyland, 2019;) over lateral terrace margins, but few quantitative data exist to substantiate this proposition. Competing opinions within the literature assert the presence of patterned ground as being indicative of geomorphic quiescence (e.g., Karte, 1979 , 1981; Thorn & Hall, 2002 , 545) , while others (Paquette et al., 2017a, 2018; Wilkinson & Bunting, 1975) emph asize the dynamic hydrologic and geomorphic controls implied by the presence of sorted 3 patterned ground, although not in the context of cryoplanation. Given these diametrically opposing views, it is imperative that sorted stripes, which abound on CT treads (Nyland, 2019, 57) , be made an investigative focal point about the mode of sediment transport on CT treads and side slopes. 1.1 Brief History of Cryoplanation Research Across the spectrum of periglacial environments, CTs are some of the largest and most widespread landforms , giving rise to entire upland landscapes. Assemblages of periglacial microfeatures made up of frost - fractured rock, geometrically arranged ground, and solifluction lobes, characterize cryoplanated ridges. Despite the pervasiveness of CTs throughout perigl acial regions and their distinct geomorphic imprint, workers since the opening of the 20 th century have struggled to relate specific periglacial processes to CT formation, a trend that contributes to the current contention surrounding the origin of these f eatures. 1.1.1 First e ncounters . Observational accounts of cryoplanation terraces in North America were first recorded in the early 1900s in the course of surveys by the United States and Canadian Geological Surveys ( e.g., Prindle, 1905; C airnes, 1912; Eakin, 1916; Mertie, 1937) . normal cycle of erosion , struggled to reconcile their observations with Davisian principles. conceived by creating an analogy with the life cycle of biological organisms, was based on the concept of geomorphic landscapes going through young, mature, and old stages. A cycle culminates in leveled surfaces of regional extent at low elevation, which Davis called peneplains . Following epeirogenic uplift or a drop in base level , which signals the beginning of a new geographic cycle, these peneplain 4 surfaces are dissected by fluvial processes, leaving discontinuous remnants of former peneplains at relatively high elevation. However, the flat elevated surfaces observed in northwestern North America by early surveyors displayed non - accordant summit elevations at the local scale and were criteria for identifying elevated peneplain remnants (see Thornbury 1969, 182 - 185). Davis (1909) also recognized the imprint of other climatic conditions on geomorphic landscapes, ref which he developed alternative cycles of erosion. Davis himself, however, did not develop a modification to the normal cycle that adjusted for periglacial clim atic conditions and geomorphic processes. 1.1.2 Introduction of periglacial geomorphology . Around the time that CTs were first encountered, periglacial studies gained relevance in the European literature with the introduction of foundational periglacial te rminology. Lozinski (1909) coined the term periglacial ( near glaciers ), the study of landforms formed under cold but nonglacial conditions, a concept expanding F.E. Matth es, working in the Big Horn Mountains of Wyoming, introduced the concept of nivation . The process suite invoked by the term begins when snow accumulate s within topographic hollows on north - facing slopes, persists late into the summer or , in some cases, the entire year. As these snow accumulations achieve greater densities and ablate , i ntensive freeze - thaw, chemical weathering, slopewash, and solifluction occur in the vicinity of the snowpatch. The recurring weathering and transportation processes excavate small depressions known as nivation hollows. In the context of CT formation, ridge crests occupied by perennial 5 snowpatches promote headward scarp retreat, widening nivation hollows so that the ridge profile resembles a giant staircase of ascending treads and scarps. Early work by Matthes (1900) , Andersson (1906), and Lozinski (1909) con tributed to the nascent field of climatic geomorphology , which gained traction as a continental European reaction to Davisian geomorphology (see Penck 1905; Martonne 1913; Büdel 1948a; Tricart & Cailleux 1972; Büdel 1977; Beckinsale & Chorley 1991, Chapter 12). Periglacial studies formed an important component of climatic geomorphology . E arly work by Troll (1944) and Büdel omorphology, led by Hans Poster (e.g., Hövermann & Oberbeck 1972; Poser, 1977; also see Karte, 1979, 1982). The International Geographical Union facilitated interactions between periglacial scientists of the Eastern and Western blocs through its long - stan ding Periglacial Commission, particularly through its journal Biuletyn Peryglacjalny , founded in Poland by Jan Dylik in 1954. 1.1.3 Early conceptions of the nivation hypothesis . In the early 1900s, Prindle (1913) and Cairnes (1912) made genetic inferences about CT development that were consistent with the nivation process - Cairnes emphasized widespread leveling of the ground surface, a process he called equiplanation . Under equiplana tion, erosion occurs by nivation and solifluction where the surface is ultimately lowered by freeze - thaw action and the movement of weathered ramparts to the level of the underlying permafrost table. Prindle (1913), without using the term, indirectly conne - - central Alaska. Other scientific general ist s operating in Alaska in the early 1900s interpreted CTs 6 as a consequence of geologic structure (s ee Nyland, 2019 , 25) . Following these initial observations in the early 1900s, very little work was published about CTs until the 1960s, using the term altiplanation terraces Eakin, 1916) . The literature on nivation, although more voluminous, remained d eductive and descriptive until the 1970s. 1.1. 4 Periglacial geomorphology gains traction . of erosion was presented to the Anglophone audience in The Geographic Cycle in Periglacial by Louis Peltier (1950). cycle holds that morphogenetic regions, delineated by regional prevailing climatic regimes, progress geomorphically through c limate - specific erosional processes. Erosional stages were demarcated using Davisian nomenclature youth, maturity, and old age but emphasis was placed on climatically induced landscape development. As such, climates are not viewed as as naturally occurring sets of controls over geomorphic processes. Peltier (1950) proposed the periglacial realm as one of nine morphogenetic regions (cf. Büdel 1948a). Nivation and solifluction, concepts proposed years earlier, are emphasized under Pel - shattered material contributes to material accumulation and mass movement processes solifluction, slopewash, and congeliturbation that promote scarp retreat. Cryoplanation surfaces characterize Pel cycle and are achieved as rampart accumulation s eventually rise to a level at which further denudation is limited, giving way to the formation of broad erosional surfaces. Cryoplanation, a term first coined by Bryan (1946), was adopted by Peltier (1 950) and became used widely in reference to the long - term generation of cryoplanation terraces via scarp retreat in cold regions (Demek, 1968; Ballantyne, 2018 , 220) . Peltier (1950) was successful in incorporating periglacial 7 relief generat mainstream of Anglophone geomorphology. 1.1. 5 Definitive studies on cryoplanation terraces . The first comprehensive studies concerning cryoplanation terraces were conducted by Demek (1969) and Reger (1975). These works drew on all the available information (e.g. stratigraphy, climate, sedimentology, and morphology) to characterize CTs and to map their extent across the N orthern H emisphere. One of the significant findings of these works was their documentation of CTs cutting across geologic structure, a major blow to the geological - control hypothesis of CT formation. Both authors embraced the nivat ion hypothesis of CT formation, in which scarp retreat is related to the mass - balance of large and persistent banks of snow (Figure 1.2). Both authors understood the limitations of their works, which relied on classification and mapping, and made poignant calls for process - based studies, dating of CT surfaces, and investigations on transportation of weathered material. Specifically, these authors highlighted the importance of tracing the fate of weathered material on CT slopes to explain the stepped appeara nce of CT profiles. Ultimately these works would not be highly influential, however, largely because the ir distribution was severely limited. Calls for process - based investigations would remain unaddressed for many years. 8 Figure 1.2 D iagrammatic model of nivation - driven scarp retreat of cryoplanation terrace formation . The stages of CT development are characterized by (A) the accumulation of snow within a topographic irregularity on a north - facing slope. (B) the elongation of the topographic hollow through backwall retreat which forms a visible break in slope between the CT tread and tread. The formation of sorted patterned ground from available snow melt water occurs on the CT tread. (C) The CT tread continues to lengthen via scarp retreat but mater ial accumulations representing past scarp position are not visible. Figure after Nyland (2019, Figure 18). 1.1. 6 Quantitative studies ensue . Process studies focusing on nivation would not be undertaken until the mid - 1970s following the field research cond ucted by Thorn (1976) who worked in the alpine periglacial environment of the Colorado Front Range. This work quantitatively substantiated the intensified weathering and transportation of material in the vicinities of late - lying snowpatches but ultimately, Thorn concluded that his observed erosion rates by nivation were inconsistent with the size of cirque - scale features and argued forcefully against the concept of a continuum between nivation hollows and glacial cirques. Nelson (1975) data set, confirmed that the size of CTs is significantly 9 smaller than with the dimensions of CTs, providing some support for the nivation hypothesis of CT formation. Nelson (1989, 1998) and Nelson and Nyland (2017) argued that CTs are periglacial analogs of glacial cirques; both features form as a consequence of the mass balance of bodies of snow and ice, both show poleward - facing orientation, and both follow the clima tically defined regional snowline. Thorn & Hall (1980) have raised concerns about the efficacy of the nivation process suite and took issue with its terminology. Although their concerns had been tempered by the turn of the 21 st century (Thorn & Hall 2002), their earlier objections were translated into textbook form by H.M. French, who recommended that the term nivation be abandoned altogether (French, 2017 , 233 and references therein) . By the early 2000s only indirect lines of evidence continued to attribute CT formation to niv ation. Process - based studies on the landforms remained entirely lacking. 1.2 Revisiting the Nivation Hypothesis A recent trend in periglacial research highlights the transference of water, fine clastic sediment, and weathered solutes within sorted patt erned ground and associated periglacial water tracks (e.g., Paquette et al., 2017, 2018) . In these studies, sorted patterned ground, specifically sorted stripes, were found to be responsible for the initiation of periglacial water tracks, and thus the widespread hydrologic co nnectivity of the hillslope. This research trend informs an important line of investigation that has profound implications in the context of CT formation hypotheses. 1.2.1 Scrutiny . The most widely invoked scrutiny of the nivation hypothesis of CT formati on states that moraine - like deposits, representative of past scarp positions, should 10 accumulate on CT inner treads (Ballantyne, 2018, 221) . No such accumulations have been observed in characteristically periglacial regions, a situation supposedly indicative of the absence of processes that remove weathered products from CT treads. In the absence of another widely supported hypothesis, the ni vation explanation for CT formation seems the most appropriate but rampart accumulations must be accounted for before the hypothesis can be widely embraced. link . The networks of sorted patterned ground and solifluction lobes frequently encountered on CT treads and side slopes are potentially important features for the transmission of weather ing products across CT treads. A record of work that notes the presence of running water within sorted stripes on CT t reads exists (e.g., Taylor, 1955 ; Caine, 1963; Smith, 1968; Nelson, 1979) and provides impetus for investigations into the hydrology of sorted patterned ground. However, no study has related sorted patterned ground to landscape - scale evolution, specifically cryoplanation terraces, alt hough previous work has indicated the ability of sorted stripes to exert major hydrologic controls over large extents of the hillslope (Paquette et al., 2018) . An exploration into the hydrologic and geomorphic significance of sorted stripes on CT treads is required to reveal the information necessary to explain the absence of ramparts on the inner (proximal) areas of CT treads. This study attempts to address the absence - of - ramparts issue by examining the role of sorted stripes through two avenues of research: (1) process - related studies; and (2) fluvial geomorphometric analysis of a sorted stripe network. 11 The results of this study are used to specify the nature of the missing link issue of th e absence of ramparts on CT treads, thereby a ddressing for scrutiny of the nivation hypothesis. In doing so, it has potential to remove two of the last major objections to the nivation hypothesis of cryoplanation - terrace forma tion: (1) demonstration of a mechanism for transporting large volumes of sediment over and off CT treads; and (2) growth and maintenance of the characteristic step - like CT form. 12 Chapter 2 Hydrologic Significance of Sorted Patterned Ground Prior to the 1970s, the role of water in periglacial landscape was rarely addressed in the literature . This research deficit has been overturned by the documentation of fluvial features such as gullies, rivers, and V - shaped valleys abounding in periglacial landscapes (Ballantyne, 2018, 253 ). In arid periglacial environments, hydrologic activity is largely determined by the occurrence and distribution of perennial snowpatches and ice, which are responsible for the prolonged release of moisture to slopes (Ballantyne, 2018, 253 ). Underlying ic e - bonded permafrost also exerts a hydrologic control by limiting vertical water percolation and by concentrating hydrologic activity in the seasonally thawed active layer (Slaughter & Kane, 1979; Woo, 1986) . Fluvial action, in association with periglacial hydrologic - related constituent processes, is one of the most importan t geomorphic processes in periglacial regions (Ballantyne, 2018, 253 ). A recurring theme in periglacial literature of the past several decades hints at the hydrologic controls imposed by the minor periglacial features known as patterned ground (Nelson, 1975; Wilkinson & Bunting, 1975; Paquette et al., 2017a, 2018) . Some wr iters regard implies a trivial relationship between patterned ground and the underlying topography. Further opposition is presented through the interpretation of patte rned ground as an indication of geomorphic dormancy (e.g., Thorn & Hall, 2002b, 545) , underscoring notions about the hydrologic insignificance of the features. 13 Research by Paquette et al., (2018) explicitly attributes enhanced channeling of water within areas of patterned ground to the formation of periglacial water features known as water tracks. The connectivity of this patterned ground - water track complex provides striking evidence that favors the interpretations about the geomorphic and hydrologic significance of patterned ground. This thread is exemplified in a thesis by Queen (2018), who applied the concept of feature assemblages to perigla cial landscapes, demonstrating that repeating assemblages of minor periglacial features such as patterned ground are the elemental building blocks of larger units that are recognizable and operational at the landscape scale, i.e. , cryoplanation terraces. The conception of patterned ground as being of minor or no geomorphic significance is at odds with literature that documents patterned ground hydrology. On the one hand, the reader is led to believe that patterned ground is nothing more than periglacial d écor, while on the other hand, substantial evidence demonstrates its hydrologic and geomorphic significance. An understanding of slope development in a periglacial setting warrants consideration of patterned ground as a fluvial feature. This approach to as sessing the hydrology of a periglacial setting permits the exhaustive exploration of associated geomorphic development. 2.1 Periglacial Hydrology Peltier ( 1950, 215 ) opined that mass movement and moderate to strong wind action are important generators of relief in the periglacial morphogenetic region. Under the geographic (geomorphic) cycle as it relates to periglacial environments, the impact of water is given littl e mention, conveying a sense of insignificance regarding periglacial hydrology. Major contributions to periglacial hydrology were made beginning only in the 1970s, revealing the 14 unique runoff regime in periglacial regions, which is characterized by the ces sation of runoff in winter, the dominance of water storage as ice and snow, and underlying permafrost conditions. 2.1.1 Increasing interest in periglacial hydrology . One of the first serious considerations of hydrology in the periglacial literature was pu blished by McCann et al. (1972), a paper in which the lack of studies focusing on periglacial hydrology is noted very explicitly. The dearth of information on periglacial hydrology at that time was accentuated by the fact that of the few published works on periglacial catchments mostly focused in Alaskan tundra basins most were qualitative (McCann, et al. , 1972) . The late 1970s brought about a marked change in interest about periglacial hydrology, with extensive research by Ming - k o Woo, working in Canadian watersheds in the late 1970s and 1980s, (e.g., Woo, 1976; Woo, 1986) research focused on tundra rivers and on the oddities of periglacial rivers such as the influence of snow - choked channels (Woo, 1976; Woo & Sauriol, 1980) . Research on the hydrology of periglacial regions advanced rapidly thereafter, as reflected by an in - depth review by Clark (1988), who nonetheless lamented that the subject was still in its infancy. Progress has continued unabated in the intervening years, and the subject was treated comprehensively in a recent volume by Woo (2012). Today, active areas of research in periglacial hydrology apply field - based monitoring efforts and modeling techniques to modern issues such as the response of the periglacial hydrologic regime to warming global temperatures (Woo et al., 2008) . 2.1.2 Characteristics of periglacial watersheds . Arid and semi - arid periglacial watersheds are typified by a snowfed, or nival regime (Woo, 2012 , 475 477) . Under this regime, rivers and other water features are controlled primarily by snowme lt water (Christiansen, 1996) confining most hydrologic activity to the spring, when snow begins to melt in the N orthern H emisphere. 15 Threshold - like melting is achieved in snowpatches at the onset of above - freezing temperatures, (Quinton & Marsh, 1999) . Vertical percolation of water into the substrate underlying the snowpatch is at first inhibited due to the formation of basal ice, which induces later al flow between the ground surface and the overlying snowpatch (Ballantyne, 2018, 255 ). In the early spring, limited groundwater storage due to frozen ground, rapid melting of snowpatches, and dominantly lateral flow on the hillslope all contribute to inte nse hydrologic activity in periglacial watersheds. Irregular accumulations of snow concentrate hydrologic activity and erosive power in those areas proximal to snowpatches, invoking the influence of nivation in periglacial watersheds. 2.1.3 Sorted pattern ed ground . Patterned ground is the term given symmetrical forms, such as circles, polygons, nets, steps, and stripes, that are characteristic of, (Washbur n, 1956) . A subgrouping of patterned ground, sorted patterned ground concentrated presence of coarse fragments bordering fine - textured areas is frequently encountered occupying broad areas extend ing downslope from perennial snow - patch margins (Ballantyne, 2018, 145 and references therein). Sorted patterned ground imparts a distinctive imprint on local hydrology by establishing microtopographic variation, material sorting, and channelization of water. 2.2 Sorted Stripes as Fluvia l Features Underlying permafrost conditions inhibit fluvial incision of water features into mature channels so that in some periglacial settings, smaller and less accentuated periglacial hydrologic features 16 may be the only indications of water transportation on the g round and represent substantial caches of water (W oo, 2012, 242 249) . The initial focus of periglacial hydrology on rivers contributed to some degree of dismissal of the hydrologic control of other features and processes such as rillwash and gullying (Embleton & King, 1975) . Rec ent research has linked patterned ground and other periglacial features to significant hydrologic activity (e.g., Levy et al., 2011; Paquette et al., 2017a, 2018) , marking a resurgence of research on these features. 2.2.1 Sorted patterned ground formation . Modeling efforts aimed at establishing the origins of sorted patterned ground demonstrate that differential frost heave occurring at the interface between unfrozen and frozen ground causes uplift of coarse fragments, a process known as frost sorting that leads to squeezing and confinement of stones along their margins (Hallet, 1990; Kessler & Werner, 2003) . Macrofabrics from the coarse borders of sorted circles (Nelson 1982a) and sorted stripes (Nelson 1982b) ref lect this process; tabular clasts lie with their long axes parallel to borders and are frequently on edge. Under recurrent freeze - thaw cycles within a soil medium composed of unsorted coarse and fine - textured material, particles migrate toward areas of sim ilarly sized particles, i.e., large clast s migrate toward higher concentrations of similar particles, invoking a self - organizing formation scheme related to continued patterned ground development (Kessler & Werner, 2003) . Sorted pa tterned ground is further altered under the influence of slope, where the greater the slope, the more elongate the forms; i.e., stripes (Washburn, 1956) . The self - propagation of sorted patterned ground formation ensures the maintenance of textural boundaries, which are the defining components of water movement. 17 2.2.2 Overlooked features . Sorted stripes appear on the ground surface as repeating linear bands of coarse fragments alternating with bands of fine sediment oriented parallel to slope and always occurring in groups (Washburn, 1956) . Sorted stripes occupy slopes between 4 - 11° and have been associated with efficient water transportation, which flows in the large interstitial spaces of the coarse stripes ( e.g., Nelson, 1975; Wilkinson & Bunting, 1975) . fine - textured material (Washburn, 1956) . That water flows within coarse sorted stripes is not controversial and is well - documented e.g., (Wilkinson & Bunting, 1975) . The hydrologic significance of sorted patterned ground is less well reco gnized (Paquette et al., 2017a) . Goldthwait (1976, 31) described the flow of water through a sorted patterned g round landscape: Woo (2012, 245 - 246) noted that hydrologically, sorted stripes direct runoff from snowmelt into surface drainage networks, indica ting that when present sorted stripes comprise the rudiments of an informal drainage network. Qualitative documentation of sorted stripe hydrology provides a foundation for their consideration as fluvial features , but in the absence of quantitative data, l eaves the concept open to question. 2.2.3 Previous work . Hydrologic activity within sorted stripes was tangentially outlined interpreted here as sorted stripes superimposed on terraces off the coa st of Tasmania. Taylor (1955, 136) noted that the 18 178) postulated the origin of sorted stripes in northern England and qua litatively deduced that drainage of water from a heavily saturated field site was directed through the coarse portions of sorted stripes. He observed: place through the coarse stripes or the grooves of thawed material just below them as long as Work conducted by Wilkinson and Bunting (1975), focused on the movement of water as overland flow in a periglacial settin g via subtle water features, known as rills, demonstrating that a system of rills fed by snow meltwater were hydrologically connected to other water sources on the hillslopes through effective transport of water and sediment. Nelson (1975) outlined the geo morphic and hydrologic character of sorted stripes at an alpine periglacial site, describing the preferential flow of water through coarse stripes (16). Nelson (1975), following Ule (1911), Cairns (1912), Salomon (1929), Mortenson (193 0 ), Poser (1931), Czeppe (1961), and Washburn (1969), also hypothesized that the sorted stripes originate from snow meltwater rillwash (Nelson, 1975, 20). Troll (1944) rejected this explanation on grounds that many stripes lie parallel to one another and/or do not occupy depressions in the soil matrix, arguing instead that solifluction is responsible for stripes through elongation of sorted nets or circles. This view has become a standard explanation for sorted stripes (e.g., Embleton and King 1975, 91 ). Goldthwai t (1976) also argued against a rillwork origin on the basis that rill networks do not exhibit the geometric regularity he perceived, without quantitative characterization, in patterned ground fields he investigated. The similarities between the network cha racteristics of 19 a field of sorted stripes and those of low - order fluvial networks is the subject of detailed investigation in Chapter 6 of this thesis. Nelson (1975) used two diagrams depicting water flow through a sorted stripe and sorted net landscape to summarize his observations (Figure 2.1). Hodgson & Young (2001) found the hydraulic conductivity within coarse sorted stripes to be three orders of magnitude higher than in intervening frost - mound centers; hydraulic conductivity values for coarse strip es were between 90 - 1000 m/day whereas frost mound values ranged from 0.1 - 1.0 m/day. Figure 2.1 Sketch of water movement through sorted patterned ground . (A) W ater movement from a sorted stripe to and through a solifluction lobe downslope; (B) Sorted patt erned ground features on Frost Ridge, showing pathways of running water through interstitial space in coarse portions of the patterns. The direction of water movement in both sketches is indicated by arrows. Sketches from Nelson (1975). Local late - lying sn owbanks supplied meltwater to these areas throughout the summer season. This thesis explores the hypothesis that movement of water through a sorted patterned ground environment prevents the formation of accumulative ramparts on subjacent treads. 2.2.4 Hillslope connectivity . Recent studies have suggested that poorly incised water features, known as water tracks, exert a dominant role in linking terrestrial and aquatic systems in upland periglacial regions (Levy et al., 2011; Paquette et al., 2017a, 2018) . Water tracks, 20 which appe ar as linear bands of rich vegetation that alternate with bands of sparse vegetation (Figure 2.2) , have been shown to control nutrient transportation, vegetation establishment, and water fluxes from upland slopes in hillslopes underlain by permafrost (Levy et al., 2011) . In periglacial catchments that lack formal river systems, water tracks are in some instances the only evidence of drainag e pathways (Woo, 2012, 242 243) . McNamera et al . (1998) found that water tracks occupying hillslopes in the Kuparuk River basin in Alaska were responsible not only for basin drainage, but also for significant water storage. Although water tracks do not achieve the organization of a mature drainage netw ork, their drainage and storage characteristics make them significant components of the catchment hydrologic system (McNamara, et al., 1998) . Observations by Levy et al (2011) in Antarctica demonstrate that within water tracks, which were composed of silt, sand, and pebbles, solute transport was two orders of magnitude faster than in adjacent soils, indicating the prominent transportation pathways imparted by water tracks. 21 Figure 2.2. Organizational similarities between sorted stripes and water tracks . (A) sorted stripes near Atlin, British Columbia, Canada. Here, stripes are downslope trending and extend towards the slop e margin in the background of this photo. Photo taken by C.W Queen 2017. (B) Water tracks extend from upslope in the background and converge in the foreground of the photo. Photo taken in the Arctic Foothills, AK, by A.E Klene. A series of works by Paquette et al., (2014, 2017, 2018) demonstrated the importance of patterned ground in water - track initiation. Presented at the Fourth European Conference on Permafrost in 2014, this research introduced the hypothesis of patterned - grou nd - initiated water track formation (Paquette, et al., 2014) . Preferential sorting, eluviation of fine material, 22 and channelized water flow associated with patterned ground were attributed to downslope water track formation, defining the hydrologic connectivity of the hillslope. Papers published by the same authors in 2017 and 2018 expand on the concept of hillslope - scale connectivity, establishing that the patterned ground - water track complex tended to dominate the local hillslope hydrology, and that patterned ground increased hillslope connectivity to the principal water sources, snow patches. The resonating theme of these contemporary studies lies in their compelling evide nce that patterned ground increases connectivity between upland water sources (in this case, snowpatches) and lowland water features, e.g., lakes, through the exertion of geomorphic and hydrologic control. These findings challenge the notion that patterned ground indicates geomorphic quiescence and further demonstrate that when moisture conditions permit, patterned ground is a critical component in periglacial hydrology. 2.3 Conclusion: Sorted S tripes, Nivation, and Cryoplanation T erraces It is evident that where present, sorted stripes and water tracks represent preferential flow pathways on periglacial hillslopes. The hydrologic significance of these features is especially prominent in arid and semi - arid periglacial regions, where snow me ltwater constitutes a major hydrologic input. The linkages between patterned ground, water tracks, and hydrology have been documented since the 1950s (Taylor, 1955) , but have seldom been embraced widely in the literature (Paquette et al., 2014, 2017a) . Work by Queen (2018) established cryoplanation terraces as foundational units in a periglacial setting. Inherently related to the basic CT unit is the operation of periglacial assemblages, whose repeating morphologic signatures link micro - scale periglacial features and processes to landform - and landscape - scale geomorphic activity. Patterned ground, solifluction 23 lobes, and pe rennial snow patches favor CT profiles owing to the poleward orientation of CTs (Nelson, 1998) , and the rubble mantle cover that typifies CT treads (Ballantyne, 2018, 220). The co - location of patterned ground and perennial snowpatches atop cryoplanation terraces is an arrangement that has not been explored previously in the context of CT formation. Allusion to evacuation of sediment from CT treads via sorted patterned ground (e.g., Nelson 1975; Nyland, 2019, 59; Queen, 2018, 41) , is only an initial step in recognizing the importance of low - order fluvial networ ks in the formation and maintenance of cryoplanation terraces. 24 Chapter 3 Study Area The Juneau Icefield is the fifth largest expanse of continuous upland ice in North America (Miller, 1975, 1) , covering an area of about 1, 800 km 2 in the 1950s (Field & Miller, 1950, 180) . The icefield extends from the Taku River Valley , AK in the south, to the vicinity of Skagwa y, AK in the north . The north ern and eastern flanks of the icefield extend into Canada, covering a confluence area across the Alaska, British Columbia, and Yukon Territory borders . The Juneau Icefield remained largely unexplored until the establishment of the Juneau Icefield Research Project (JIRP) , under the auspices of the American Geographical Society (Heusser , 2007). T he first research - motivated reconnaissance mission w as conducted in 1948 (Field & Miller, 1950) . A series of research camps established at the onset of icefield explorations form a northeast - southwest transect extend ing from the coastal side of the Boundary Range in Alaska to the northeast into British Columbia, Canada (Miller, 1975, 1) . Most periglacial work in the interior of the icefield has been conducted on nunataks (e.g., Hamelin, 1964; Nelson , 1979 , 101 - 102 ; Dixon et al. , 1984). The research camp transect terminates at Camp 29 near the village of Atlin, British Columbia, which is the site where the research for this thesis was conducted in the summer of 2019 (Figure 3.1) . The glacial history of this area, among other environmental conditions, has promote d the development of a wealth of both relict and active periglacial features (Nelson, 1979, 8) , making it a highly un usual site for periglacial investigations. 25 Figure 3.1 The Juneau Icefield . The extent of the Juneau Icefield with the approximate location of Camp 29, the study site for this thesis , denoted with (1979), map after C.J. Cialek. 26 3.1 The Greater Atlin Area 3.1.1 Geographic d escription . The village of Atlin is situated along the eastern shore of Atlin Lake and lies w ) Stikine Region of northwest British Columbia (Statistics Canada, 2018) . Atlin L ak e, like many of the lakes in this region, occupies a glacial valley that (Slupetzky & Krisai, 2009a, 192) . 3.1.2 Geology . The geology of northwestern British Columbia is composed of a complex mosaic of terranes fragments of crustal material that have coalesced due to tectonic rifting (Monger et al. 1972) . As such, the geology of the area invokes a multitude of geologic discon tinuities represent ing distinct geologic and physiographic boundaries (Monger et al. , 1972 , 577 579) . Atlin lies within the fault - bounded westernmost limit of the Atlin Terrane and is composed of little - metamorphosed upper Paleozoic rock s belong ing to the Cache Creek Group (Monger, 1975, 2 ; Colpron & Nelson, 2011) . 3.2 The Cathedral Massif 3.2.1 Physical description . The Cathedral Massif an isolated, rocky highland is located approximately 35 km southwest of Atlin . The highest point of the massif, Cathedral Peak, (2314 m .a.s.l. ), is flanked to the west by Mount Edward Little (1932 m .a.s.l. ) and to the east by Splinter Peak (1925 m.a.s.l. ) (Figure 3. 2 ) . T he Cathedral Glacier occupies t wo glacial cirques . This glacier, sometimes referred to as the Cathedral Massif Glacier ( e.g., Slupetzky & Krisai, 2009a) , occupies the north - facing side of the Cathedral Massif, immediately downslope from Cathedral Peak. Slupetzky & Krisai (2009, 194) report ed the terminus of Cathedral Glacier to be at 1615 m.a.s.l. (in 1999) , though the current terminus elevation is likely much higher as the 27 glacier has receded substantially in recent years (Nelson, 2018, personal communication ) . Two major flow units are derived from the Cathedral glacier : flow from the eastern cirque drains eastward into the southern reaches of Atlin Lake , while f low from the western lobe drains directly downslope, resulting in a proglacial lake at its base. Downslope from the present terminus of the Cathedral glacier , a glacial valley is occupied by a terminal and several recessional moraine s , which represe nt past glaciations that occurred during the Little Ice Age (Slupetzky & Krisai, 2009b, 196) . Figure 3. 2 Orientation of Frost Ridge . (A) Site map of Frost Ridge and relevant features. (B) Oblique view of Frost Ridge facing south. The Cathedral Glacier can be seen between Splinter Peak and Mt. Edward Little. The western lobe of the Cathedral Glacier is located between Cathedral Peak and Splinter Peak. The eastern lobe is located between Mt. Edward Little and Cathedral Peak. Photo taken July 2018 by R.J. Mitchell. 3.2.2 Geology . Jones (1975) provide d reconnaissance - level geologic data for the Cathedral massif region , which lies immediately outside the western border of the Atlin Terrane , within a narrow strip of the Stikinia Terrane ( MacIntyre et al., 2001; Colpron & Nelson, 2011) . T h e geology of the area is composed predominantly of volcanic, plutonic, and 28 sedimentary rock of Late Triassic to Early Jurassic age (Colpron & Nelson, 2011) . A d ark, heavily stressed, fine - grained gran o diorite intrusion characterizes most of the Cathedral Massif (Jones, 1975 , 22 23) . Near and above the Neoglaciation limit, Splinter Peak is composed of highly weathered metamorphic rock (Jones, 1975, 25) . 3.2.3 Climate . The climate of the Cathedral Massif region is classified as continental, semi - arid, and sub - polar (Miller, 1975, 6) ., The climate is i nfluenced by its location on the continental side of the Boundary R ange , which inhibits penetration of moist air from the Pacific Ocean to the southwest (Jones, 1975, 39). The continental climate of the Cathedral Massif and greater Atlin area is characterized by cold and dry conditions , with the Atlin area receiving 285 mm of precipitation per year (Jones 1975, 39; Nelson 1979, 12) . Owing to the climatic continental ity , much of the unglaciated parts of the Cathedral Massif are classified as periglacial ( Tallman, 1975a, 106; Johnson, 1983, 2;) . 3.2.4 Glacial h istory . The glacial history of the Cathedral Massif is reveal ed by a network of moraines that occupy the glacial valley downslope from Cathedral Peak. Rounded peaks, glacial scouring, fields of glacial till and proglacial lakes also reflect glacial activity withi n the area. Pre - Wisconsinan glaciation from the ancestral Hobo - Llewellyn glacier deposited a large terminal moraine over 1.5 km north of Cathedral peak (Tallman 1975, chap.6; Miller 1975 131 - 132, Slupe t sky & Krisai 2009). I ce flowing north and northeast ar ound the Cathedral deeply scoured the southern flanks of the massif, creating the Atlin Lake basin (Jones, 197 5b , 31; Tallman, 1975b, 32; Slupetzky & Krisai, 2009) . Late Pleistocene glaciation resulted in the deposition of an impressive lateral moraine found on the eastern flank of Frost Ridge (Slupetzky & Krisai, 2009). Later, Neoglacial - period ice deposited m inor recessional moraines that are well - 29 confined within the Cathedral glacier valley . In addition to glacial alteration of the Cathedral Massif, periglacial processes have contributed to alteration of local geomorphology (Jones, 1975; Nelson, 1979) . 3.3 Frost Ridge 3.3.1 Physical description . Frost Ridge (FR) is a northeast trending line ar feature extend ing from Splinter P eak to the southern end of Atlin Lake. FR is flanked to the east by a large lateral moraine and to the west by the unoccupied Snowdrift C irque (Jones , 1975) . Hummocky, vegetated terrain characterizes the lower reaches of FR while well - expressed patterned ground sorted nets, circles, and stripes and other periglacial features abound in its upper reaches. 3.3.2 Climate . The same continental climate described for the Cathedral Massif area prevails on Frost Ridge . Topographic irregularities and minimized solar radiation on north - facing slopes contribute to the accumulation of what little snow does fall on the ridge. Late - lying snowpatches occur on FR where snow accumulations persist into the summer months, supplyi ng prolonged moisture to the slope. 3.3.3 Glacial history . During the Wisconsinan glaciation the upper reaches of Frost Ridge remained above the glacial margins (Nyland 2018, 82). Further downslope, ice contact from the ancestral Hobo - Llewellyn glacier carved marginal drainage features on the north - facing side of FR (Nyland 2019, 84). Above these marginal drainage scars, the upper reaches are characterized by an abundance of periglacial features formed within the frost - shattered ma ntle cover since the waning stages of the Wisconsinan (Queen, 2018, 57) . 30 3.3. 4 H ydrology . Late - lying snowpatches are common on Frost Ridge and permit the relatively moist substrate conditions necessary for patterned ground formation, mechanical and chemical weathering, and overland flow. Frost Ridge constitutes an alpine nival regime (Woo 2012, 475 477) in which snow and ice are the primary suppliers of moisture to the area. Runoff from the snowpatch margin s facilitate s observable overland flow, slopewa sh , and rillwash . The only indications of appreciable water flow atop FR are solifluction lobes and sorted patterned ground, through which the preferential flow of water has been documented (Nelson, 1975) . 3.3. 5 Periglacial features . Relict and active periglacial features abound in the frost - shattered mantle of Frost Ridge. Nelson (1979) provide d a description of the well - developed patterned ground field occup ying the uppermost reaches of FR. Notable features here are well - developed coarse stripes, sorted circles, and solifluction lobes composed primarily of silty material (Nelson, 1979, 32 ). The occurrence of inactive sorted stripes has also been reported (Nelson 1979) . Inactivity of stripes is indicated by the subdued appearance of the coarse lineations relative to intervening fines, which had been penetrated by vertical plant roots (Nelson, 1979 , 43 ). Since the time of that research the periglacial features have undergone modifications (Nelson, 2018, personal communication) owing to complex relationships between snow mass balance and climate trends. For example, a snowpatch on the upper reaches o f FR used to persist until early September in the mid - 1970s, which is no longer the case. The resulting loss of moisture has resulted in widespread inactivity of the well - developed patterned ground field , over the last four decades. 31 At the landscape scale, a series of incipient cryoplanation terraces, or nivation hollows, impart a notched appearance to Frost Ridge. These terraces, which were incised into the ridge by the ancestral Hobo - Llewelyn glacier, have been modified under the influence of nivation since deglaciation of the area. Today, the incipient terraces are comparable in scale to characteristic cryoplanation terraces found throughout unglaciated Be ringia (Nyland, 2019, 81 ). Field reconnaissance during the summers from 2017 - 2019 confirm active periglacial processes operating on these incipient terraces, evidenced by the occurrence of late - lying snowpatches that persist into mid - July, and by the burg appearance of the sorted patterned ground. 3.4 Previous Work on Frost Ridge Detailed periglacial investigations on Frost Ridge were first recorded in an open file report by Nelson (1975) , in which he postulate d the origins and formation of patterned ground. Nelson (1975) present ed a rill - initiated hypothesis of patterned ground formation on Frost Ridge, emphasizing channelization by snow meltwater - fed rivulets, the obstruction of vertical percolation by the froz en substrate, and annual freezing to perpetuate ground sorting. Trenching of stripes normal to their axes by Nelson (1975) revealed t he occurrence of anchored sorted stripes on FR whose depth of sorting (~80 cm) appeared to coincide with the position of th e permafrost table , thereby fixing stripe position s on the slope. A thesis by Nelson ( 1979 ) , which focused on the distribution and internal structure of sorted patterned ground on Frost Ridge , revealed that the distribution and form of patterned ground is related to both the moisture of the substrate, clast shape, slope, and local climate conditions. Nelson (1982) also 32 addressed clast fabric within the coarse stripes, which he attributed to the squeezing of clasts along their axes oriented away from interve ning bands of fine - textured sediment . P eriglacial studies on FR lapsed for more than 40 years. Work of this nature was resumed in 2017 by Michigan State University personnel (Queen 2018; Nyland 2019) and has focused largely on the role of nivation in the development of the FR incipient terraces . This body of work has contributed an understanding of cryoplanation terrace formation rates (Nyland & Nelson , 20 20 ) and the characterization of a periglacial morphometric signature (Q ueen, 2018 ). 3.5 Conclusion The glacial history, geology, and geomorphic activity of the Cathedral Massif region are well understood . Detailed studies conducted on Frost Ridge in the 1970s , in addition to the recent resurgence of research at the site have contributed knowledge pertaining to nivation, cryoplanation terrace development , and periglacial activity. The existence here of actively developing cryoplanation terraces provide s an unusual opportunity to focus on the relationship between cryoplanation terrace development and the fluvial characteristics of sorted patterned ground . 33 Chapter 4 Statement of Problem and Hypothesis Process - based investigations on CT formation are required be fore the nivation hypothesis can be widely embraced (Nyland, 2018, 96 - 97). Research points to the importance of snow patches in the generation of the characteristic stepped profiles (e.g., Nelson, 1989; Nelson & Nyland, 2017; Nyland & Nelson, 2020) but direct lines of evidence linking snow erosion to CT tread expansion are absent. 4.1 Statement of Problem Cryoplanation terraces have been the subject of two genetic interpretations: (1) geologic structure; and (2) the nivation process suite. Geologic structure has been ruled out by some workers (e.g., Demek, 1969; Reger, 1975) because in many instances, CTs are observed to have no relation to foliation, faults, or compositional layering, and have been observed cutting across geolog ic structure. The erosive efficacy of nivation has been scrutinized in recent years (e.g., Th orn & Hall, 2002; French, 2017) , calling into question the role of nivation in land - forming processes. Recent literature has been more accepting of the nivation hypothesis (e.g., Nyland & Nelson, 2020a) , but only indirect evidence exists to link nivation with CT formation. A fau lt in the nivation concept is that it fails to explicitly account for the absence of constructional ramparts on the inner (proximal) parts of CT treads. Little effort has been made throughout the history of cryoplanation - terrace research to address the hi llslope hydrology of CT treads and side slopes. The fluvial role of sorted patterned ground, specifically sorted stripes, has been documented qualitatively (e.g., Taylor, 1955; 34 Nelson, 1975, 1979; Queen, 2018; Nyland, 2019) , and quantitatively (e.g., Hodgson & Young, 2001; Paquette et al., 2014, 2017a, 2018) . Aside from short treatments in theses by Queen (2018) and Nyland (2019) no studies have leveraged the well - documented role of fluvial activity within sorted stripes in periglacial environments to better understand landscape - scale geomorphic evolution, specifically that of cryoplanation landforms. Queen (2018) confirmed the existence of a characteristic periglacial landscape, describing the presence of periglacial la ndform assemblages, which he suggested are form suggested that these features operate as integrated, functional units. Paquette et al. (2018 , 1087 ), demonstrat ed that the hydrologic influence of sorted stripes and water tracks can extend outside of their immediate hydrologic zone. The conception of sorted stripes operating on the landform assemblage scale, as demonstrated by Queen (2018) and Paquette et al. (201 This thesis directly challenges opposition to the nivation hypothesis through a single research q uestion: - weathered material on cryoplanation terrace If an adequate answer to this question can be constructed it will be among the first stud ies to unambiguously link CT morphology with periglacial processes. 4 .2 Hypothesis The above - stated research question is summarized by a hypothesis accounting for sediment transport from the location of weathering at scarp - tread junctions across CT treads: 35 Fluvial processes operating within the coarse portions of sorted stripes provide a mechanism for removal of weathered sediment from cryoplanation terrace scarps and treads, thereby preventing development of ramparts on treads below. Solifluction assists in the removal of sediment in the intervening fine stripes. This thesis examines processes operating atop Frost Ridge through two distinct but complementary lines of investigation: (1) analysis of field data collected on two incipient cryoplanation terraces (Chapter 5); and (2) geomorphometric analysis of the network formed by sorted stripes in a large patterned ground field, based on high - resolution remotely sensed imagery (Chapter 6). Chapter 7 synthesizes this material to provide a new qualitative model of cryoplanation terrace development. This work is motivated by the apparent high degree of organization of sorted stripes on CT treads, their prevalence in these positions, the demonstrated hydrologic significance of coarse sorted stripes within the literat ure, and the failure of past work to account for the absence of weathered material accumulations on CT inner treads. 36 Chapter 5 Process Investigations on Two Incipient Cryoplanation Terraces In some areas where well - developed cryoplanati on terraces exist, e.g., throughout unglaciated Beringia, periglacial processes are more subdued than during past glacial periods, contributing to the interpretation that CT are inactive (Ballantyne, 2018, 221). Many studies conducted on cryoplanation terr aces have been focused on geomorphically dormant CTs so that the processes responsible for CT formation are deduced from form (Ballantyne, 2018, 221). Contention exists within the literature surrounding the origins and formation of CTs because no studies have been conducted that explore periglacial process that may be related to CT formation. Recent work on cryoplanation terraces has provided support for the nivation hypothesis of CT formation (Nyland & Nelson, 2020b) . Other work has emphasize d the importance of permafrost groundwater seepage in the localized weathering of CT scarp - tread junctions to maintain CT profiles (Matthews et al., 2019) . Both studies implemented chronologically based methods to confirm the time - transgressive nature of CT treads but neithe r study confirms the processes responsible for scarp retreat. 5.1 Process - focused Geomorphology Contemporary research on cryoplanation terraces has implemented relative and absolute age - determination techniques to define the origins and development of CT s (Matthews et al., 2019; Nyland & Nelson, 2020a,b; Nyland et al. 2020) . Sufficient evidence now exists to unambiguousl y recognize the time - transgressive nature of CT treads through scarp retreat, and the resulting stepped CT profiles. Rates of CT formation have also been defined using sophisticated remote sensing techniques (Nyland & Nelson, 2020) . With the integration of dating, remote sensing, 37 mapping, and cirque - analog characterization of CTs, the case for causal association between late - lying snow patches and cryoplanation terraces has been reinforced. It is still the case, however, that the processes responsible for scarp retreat have not been confirmed through field - based investigations. 5.1.1 Process work in geomorphology . Process - based work in geomorphology emphasizes explanations of operational geomorphic processes to better understand landform development (Bradshaw, 1982) . Adoption of process - based investigative strategies as replacements for qualitative theories of landscape evolution began in the late 1950s and have been attributed to the quantification of geomorphology (Sack, 1992) . The quantitative characteri zation of drainage basins presented by Horton (1945), Strahler (1952), and Schumm (1956) serve as examples of studies that were some of the first to relate geomorphic form to process. Similarly, modeling efforts rely on the relation between landform and pr ocess to predict landform development (Goudie, 1990, 15) . Nested within the process study paradigm is the systems theory approach as an explanatory model of geomorphic processes (Goudie, 1990, 6) . Under the systems approach, landscapes form through interrelated processes that result in the maintenance of a stable form as opposed to processes resulting in greate r change (Goudie, 1990, 7). Landforms are then viewed as inherent end members of models that are driven by underlying land - forming processes. Process investigations are difficult in many periglacial environments because, in regions where CTs are well - exp ressed, periglacial processes are not as dominant as they were during cold climatic intervals (Reger, 1975, 202). Research on relict CTs has confirmed important characteristics such as elevation trends, dominant aspect, and the time - transgressive nature of 38 tread development, but further work is needed to synthesize the qualitative models of CT formation that have been outlined by workers such as Demek (1969), Reger (1975), and Nyland (2019, 59). 5.1.2 Suggestions for process - based investigations on cryop lanation terraces . Calls for process studies on cryoplanation terraces have been made by Demek (1969), Reger (1975), and Thorn & Hall (1980; 2002). Assessments of erosion rates, morphological characterization, and the study of active CTs have been produced in recent years (Queen, 2019; Nyland, 2019; Nyland & Nelson 2020b). Still, data concerning 1) the hydrological regime of CTs (Demek, 1969, 70), 2) over - tread transportation of weathering products (Reger, 1975, . 202), and 3) monitoring of climate conditions (Reger, 1975) are absent. Although CTs have been extensively mapped ( Demek, 1969; Reger, 1975; Queen, 2018), the general remoteness of CTs and logistical constraints have hampered attempts to verify periglacial activity in all the locations where CTs are known to occur. As such, the ability to directly measure the processes thought to be associated with CT formation, specifically nivation, has been negatively impacted. Specific attention should be paid to the investigation of material displacement in the generation of the characteristic stepped CT profile. The installation of mass movement pillars can be used to assess rates of sediment movement via solifluction (Benedict, 1970) . Such features are known to occupy CT treads. Gerlach troughs have been utilized to estimate rates of erosion in inte r - rill and rill zones (Roels & Jonker, 1983) , which are similar in spatial distribution to the coarse and fine - textured sorted stripes in periglacial regions. Miniature temperature loggers can be programmed to collect and store climate data at varying temporal scales, and are easily deployable, and economical (Humlum, 2008, 22) . Similarly, temperature loggers with 39 external sensors can be used to collect air and ground temperature data for long - term monitoring investigations. The need for process - based studies of CT formation has been indicated over the past four decades. Past work has eliminated some previously advanced CT formation hypotheses, but work remains to be conducted that relates periglacial processes to the existing qua litative models of CT formation. The availability and advancement of field equipment and techniques employed in periglacial regions facilitates process - based investigation of CTs. Cryoplanation terraces that are in a state of active scarp retreat are ideal locations to conduct process - based work. 5.2 Field Investigations on Frost Ridge Two active incipient terraces are the sites of study for this chapter. At both sites, sorted stripes carried by running water, while the intervening fine stripes function as solifluction lobes. In early summer, running water can be heard flowing within the interstitial spaces of coarse stripes. Water is channeled through the sorted patter ned ground landscape, moving sediment across CT tread surfaces. Some coarse stripes terminate near the margins of CT treads with silty deposits reminiscent of alluvial fans (Figure 5.1). Water running as overland flow from the termini of sorted stripes has been observed. Sediment transfer is the result of preferential flow of water through the coarse stripes over the gently sloping CT treads. The source of the water is the snowpatch upslope. The hydrologic activity occurring within sorted stripes atop FR wa s Stone stripes are slightly sinuous and serve as channels for 40 Figure 5 .1 Oblique photo of Frost Ridge sorted stripe with silt fan . From left to right; Frost Ridge and an extent box outlining sorted stripes downslope from a perennial snow patch with an extent box outlining a silt fan, forming at a coarse stripe terminus. Left photo taken July 9 th , 2019, middle and right photos taken July 14 th , 2019 by R. Mitchell. The oc cupation of CT treads and scarp backwalls by active periglacial features provides an unusual opportunity to study contemporary processes related to CT formation. The methodology presented in this chapter utilizes climatic monitoring and data collection to investigate the role of water and sediment transmission via sorted patterned ground and solifluction to assess CT formation, thereby addressing decades - old calls for such investigations. Results and data gleaned from this study are used to assess qualitati ve models of CT formation. The duration of the observation period reported here was necessarily very short, owing to logistical and scheduling limitations during the summer of 2019. The automated instrumentation described below was programmed to operate fo r up to two years and the resulting records are expected to span the period from July 2019 to July 2021. 5 .2.1 Field m ethods . Field samples were collected over a week in mid - July 2019 from two incipient cryoplanation terraces (transverse nivation hollows). The i ncipient terraces, Terraces 2 and 4 , are located on Frost Ridge at approximately 1622 and 1525 m.a.s.l., respectively (F igure 5.2). T he t errace nomenclature adopted here follows terminology used by 41 Nyland (2019) and Queen (2018) to maintain consistency and to avoid confusion. To assess the microclimate - related characteristics of coarse and fine - textured patterned ground fea tures, a series of temperature loggers were installed. On Terrace 2, Station 1 was established to characterize the subaerial and ground temperature profiles of areas occupied by solifluction lobes. Station 1 was instrumented with S - TMB - M0xx 12 - bit air and ground temperature sensors (Onset Computer Corporation, Bourne, MA; see Appendi x ) , stabilized by a tripod mast. The air temperature sensor was installed within a radiation shield, 2 m above the ground surface, while the ground temperature sensors were inst alled within a solifluction lobe. Excavating a pit within the solifluction lobe, the ground temperature sensor was installed 10 cm below the ground surface within the upslope wall of the pit. S - SMC - M005 soil moisture sensors (Onset Computer Corporation, Bo urne, MA) were installed at 10 cm and 30 cm below the ground surface, accompanying the temperature sensor at Station 1. A similar instrumental array was established downslope on Terrace 4 (Station 2). Here, a mast with air and ground temperature sensors wa s accompanied by an additional ground temperature sensor placed 30 cm below the ground surface. Soil moisture sensors were placed at 10 cm and 30 cm depths. Ground temperature logging equipment on Terrace 4 was installed within a solifluction lobe, mirrori ng the instrument setup on T2. 42 Figure 5.2 Site arrangement on Frost Ridge . Field sites are indicated by dashed circles. T2 and T4 are active periglacial sites evidenced by the occurrence of late lying snowpatches occupying the incipient terrace scarp backwalls. The patterned ground field is considered inactive because snow no longer persists late into the summer. Frost Ridge faces north. Photo taken July 9 th , 2019 by R. Mitchell. Six MX2203 TidbiT® (Onset Computer Corporation, Bourne, MA) minia ture temperature data loggers were installed in two transects extending from the mast on Terrace 4 (located near the toe of the CT tread) upslope towards the CT scarp. At the time of data collection, the CT scarp backwall was occupied by a receding snowpatch. To monitor one coarse sorted stripe, coarse fragments were removed to excavate small, shallow (<10 cm) spaces within which loggers were deployed. These cavities were loosely covered with the excavated blocky material. The three loggers comprising the coarse - stripe monitoring transect were placed at approximately 1528, 1532, and 1546 m.a.s.l on the slope. The logger at the highest position on the slope was placed at the snowpatch margin on Terrace 4. To monitor fine - texture d areas 43 moving as solifluction lobes, three of the miniature temperature sensors were deployed on the ground surface directly upslope from the mast on Terrace 4 at approximately 1523, 1526, and 1529 m.a.s.l on the slope. All climate monitoring equipment o n both terrace study sites was programmed to log data at 5 - minute intervals for the duration of the field campaign. Miniature temperature sensor data were collected and stored using HOBOmobile® software. Ground, air, and soil moisture data were collected a nd stored using H21 - USB Micro Station data loggers (Onset Computer Corporation, Bourne, MA) and Hoboware® graphing and analysis software (Appendix , 10 7 ). 5.2.2 In lab calculations . Calculation of thermal diffusivity was made using temperature data from Terrace 4 to indicate characteristics of the ground thermal regime. Thermal (Williams & Smith, 1989, 94) , has been used to discern the operation of non - conductive forms of heat transfer in the substrate (Nelson et al., 1985; Rajeev & Kodikara , 2016) . In a purely conductive system, it is assumed that within a homogenous medium, heat flows only by conduction in the vertical direction (Williams & Smith, 1989, 85) . Non - conductive heat transfer within the thermal regime is introduced through variations in soil moisture and texture, and through phase changes of water within the ground profile (Williams & Smith, 1989, 106) . Apparent thermal diffusivity, , a non - laboratory method for calculating thermal diffusivity, can be used to detect phase change or other non - conductive modes of heat transfer occur. Two sinusoidal temperature records from varying depths within the material of interest can be treated with the apparent thermal diffusivity equat ion, which takes the form 44 ( 5.1 ) where Z refers to temperature sensor depths (m), A is thermal amplitude ( C o ) at depths Z 1 and Z 2, and is the angular frequency of oscillation given by (5.2 ) where P is the period of the temperature wave (s). Equation 5.1 yields a quantity expressed in units of length squared per unit time (m 2 s - 1 in SI units). An unreasonably large value would imply rapid and large changes in temperature (Williams & Smith, 1989, 104) . In the case on Frost Ridge, apparent thermal diffusivity was used to assess the impacts of water on heat transfer at the feature level. W ater and transported sediment from the termini of three sorted stripes were captured using Gerlach Trough Runoff Sediment Samplers (Rickly Hydrological Co., Inc.) (Figure 5.3). The basin of each trough collects sediment that settles to the bottom of the trough while transported water is routed to a collection receptacle downslope. The design of the sediment tr ough facilitates separate measures of transported water and sediment samples. Clear nylon piping was used to connect troughs to 19 - liter (5 - gallon) jugs placed downslope from the trough to capture transported water. Overflow holes were drilled into the jug s to accommodate excess water flow. Troughs were installed at the terminus of a coarse sorted stripe by excavating some coarse fragments, digging a pit approximately the size of the trough beneath the stripe terminus, and inserting the trough into the pit. The lips of each trough were inserted beneath some coarse fragments of the sorted stripe to direct water and sediment flow into the trough (Figure 5.4B). Troughs were anchored using large fragments from surrounding areas. Two 45 sediment troughs were instrum ented for two separate coarse stripes on T2. One trough was instrumented at the base of a coarse stripe on T4, while another trough was installed at the base of a solifluction lobe, also on T4. Owing to the rocky substrate, troughs could not be installed f lush with the coarse stripe termini. As such, some water and sediment were assumed to flow beneath and around the sides of the troughs. Volumes of transported sediment and water collected using the troughs are used as relative measures of sorted stripe hy drology. Figure 5.3 Sediment trough and mast installation on cryoplanation terrace . (Center) mast with sensors measuring soil temperature and moisture conditions in intervening fine area. (Right and left) sediment traps collecting transported water and sediment from the termini of sorted stripes. Sediment trap locations are indicated by dashed red circles. 46 Figure 5. 4 Field a ctivities . (A) Mast on T4. (B) Sediment trough installed on T4 at the terminus of a coarse sorted stripe (not pictured). (C) Location of miniature logger within a coarse stripe, outlined by a white dashed circle. (D) Ground temperature sensors within a solifluction lobe on T2. (E) Collection of transported water and sediment from the snowpatch margin in July durin g the summer 2018. (F) Organization of a sediment trough at a coarse stripe terminus and 5 - gallon collection jug downslope. All photos except for photo E (taken in July 2018 by K. Nyland), were taken during the week of July 9 th - 14 th , 2019. 5 .2. 3 Laboratory m ethods . Transported sediment from the sediment troughs were collected and textures analyzed. During the summer of 2018, transported water and sediment were collected from the base of the snowpatches that occupy the nivation hollows on T2 a nd T4 (Nyland & Nelson, 2020). Snowpatch samples were compared to material collected from sediment troughs. Sediment trough samples were analyzed using laboratory facilities at Michigan State University during the early spring following the field campaign . Wet samples were oven dried at 40°C to remove all moisture. Samples were gently disaggregated by mortar and pestle and material sieved to separate coarse (>2mm), fine, and organic fragments. Fine 47 samples were passed 3 times through a sample splitter to en sure a representative and homogeneous sample. 10mL of dispersant solution ([NaPO 3 ] 13 · Na 2 O) and ~10mL of deionized water were added to 20mL vials containing ~0.5g of fine sediment sample. Vials were shaken for at least 30 minutes. Laser diffraction using a Malvern Mastersizer 2000E unit generated a particle size analysis report containing detailed textural information for each sample. Maps of the field arrangements on both terraces are presented in Figure 5.5. Figure 5 . 5 Equipment orientation on Frost Ridge . (Left) Terrace two instrumented with a mast and two sediment troughs. (Right) Terrace 4 with six miniature temperature loggers, a mast and two sediment troughs. Small numbers next to the temperature sensor symbols indicate the miniature logger number. Larger numbers on top of the sediment trough symbols indicate trough number. Relative locations and distances between features are represented in this map. 5 .3 Results The length of the record from all monitoring equipment spans approximately 3 d ays (Figure 5.6) . Data logging began for each sensor around midday on July 11 th and ceased midday on July 48 14 th . During the logging period, a general decrease in temperature was observed in all logger records. 5 .3.1 Monitoring data . Recorded air t emperatures at the Terrace 4 station tended to be higher (8.19 °C) than those at Terrace 2 (7.09 °C), which is located a t a higher elevation on Frost Ridge than Terrace 4. Ground temperatures at both sites tended to be higher than the recorded air temperatures . For example, on Terrace 4, at 10 cm depths, the average temperature was 12.84 °C compared to the average air tempe rature of 8.19 °C. I n the fine soil , phase lags and progressively subdued thermal amplitude with depth were observed (Figure 5.6), as would be expected in a conductive heat - transfer system. Within the coarse stripe, the logger labeled TB_1 was positioned closest to the snowpatch margin on Terrace 4 (Figure 5.5). Increasing logger numbers indicate progressively lower logger positions on the slope, with logger TB_6 occupying the lowest position on Terrace 4. A numerical breakdown of the miniat ure and mast temperature data records is presented in Table 5.1. Within the miniature temperature logger data, warmer temperatures were recorded from the loggers instrumented on fine sorted stripes. Within the fine - textured stripe closest to the CT tread margin, average temperatures were recorded at 12.85 °C. Generally, h igher maximum and minimum temperatures were recorded for the fine stripes . From TB5, the miniature logger placed mid - CT tread, maximum and minimum temperatures were 19.04 °C and 5.79 °C, r espectively . The greatest fluctuation s in temperature were recorded from TB_1, located nearest the snowpatch margin , where the temperature amplitude was 14.12 °C , and from TB_6, located nearest the CT toe tread , where the temperature amplitude was 20.82 °C . Within all temperature data records, peak temperatures were achieved just after midday when 49 the sun is at its highest position in the sky and receipts of solar radiation are greatest on the hillslope. Temperatures dip sharply near the end of the day and continue t o decrease until the sun rises. Observable water flow, which occurred from the 12th to the 13 th , corresponds with recorded temperatures. On the last day of the logging interval, no observable water flow was observed on the hillslope. Higher soi l moisture conditions were recorded at the 30 cm depth versus conditions at 10 cm depths. Soil moisture conditions recorded at 10 cm depths generally follow temperature trends. Soil moisture tends to be higher at midday than later in the day. The soil mois ture sensor on Terrace 4 depicts this diurnal trend in soil moisture conditions clearly. 50 Figure 5.6 Field data logger records . (Fig. 5.6A - B) Terrace 2 and Terrace 4 station temperature sensor data display a systematic diminution of thermal amplitude with depth, and a progressive lag in thermal response with depth. The diurnal variation of temperature is also depicted. (Fig. 5.6C - D) Coarse and fine stripe temperature monitoring records depict temperature vari ations induced by position on the slope and substrate material. (Fig.5.6E - F) Soil moisture data from both terraces depict the same diurnal variances as indicated in the temperature records. Amplitude increases are shown at depth, attributable to enhanced s oil moisture conditions at depth and the leaching of surface moisture through vertical water percolation. 51 T able 5 .1 Temperature logger data breakdown . Temperature indices are shown below for temperature sensors on masts and for the miniature temperature loggers. Mast sensor names refer to the station number and the depth at which the sensors were placed. Air sensors are labeled with the station number an . Flux temperatures were calculated as the difference between the maximum and minimum temperatures for each sensor. Position Sensor Average Temperature ( ° C) Minimum Temperature (°C) Maximum Temperature (°C) T emperature Amplitude (°C) Mast Station 1 TS1_air 7.09 4.74 10.49 5.75 TS1_10 10.68 8.00 14.53 6.53 TS2_air 8.19 5.23 13.59 8.36 Station 2 TS2_10 12.84 9.53 16.63 7.10 TS2_30 12.70 11.05 14.41 3.36 Miniature D ata Loggers TB1 7.23 1.03 15.15 14.12 Coarse Stripe TB2 8.39 4.59 11.91 7.32 TB3 9.32 5.58 14.19 8.61 TB4 12.30 6.66 19.57 12.92 Fine Stripe TB5 11.29 5.79 19.04 13.26 TB6 12.85 5.77 26.58 20.82 5.3.2 Thermal diffusivity calculations . The temperature record for Station 2 on Terrace 4 full 24 - hour period of observation. Values of the input parameters are given in Table 5.2. Equations 5.1 and xpected for a moderately wet silt in which heat transfer occurs entirely by conduction (Johnston et al. 1981, 119 - 123). Although the temperature record shows decreasing amplitude and a temporal lag with dept h, the large variation in soil moisture (Figure 5.6) on this day contributed a significant nonconductive component to the ground - thermal regime. Variations in soil moisture, attributable to the arrival of snow meltwater on this day, depressed soil temperat ure rapidly. Over the course of a 52 summer, snow meltwater also acts to decrease soil strength and encourage sediment transport by solifluction. Table 5.2 Input parameters for determination of apparent thermal diffusivity . P (s * day - 1 ) Z 1 (m) Z 2 (m) A 1 ( o C) A 2 ( o C) 2 s - 1 ) 86,400 0.1 0.3 7.1 3.36 30.0 x 10 - 7 5 .3. 3 Transported water/sediment data : Installation occurred on July 11 th and material from all four sediment troughs was collected on July 14 th . Particle - size analysis was conducted on all samples in the lab the following January. Results from this analysis are presented in Figure 6.8. Trough 3_4, which was installed within a solifluction tongue on Terrace 4, collected very little material, all of which was assumed to have blown/saltated in as opposed to having flowed in via solifluction. Texture curves in Figure 5. 7 show the dominance of silt - textured material within samples c ollected from coarse stripe termini. Jugs collecting water from all three troughs installed at coarse stripe termini were found completely full after the first day following their installation. Based on observations of the overflowed jugs, more than 19 lit ers of transported water is inferred to have flowed through each coarse stripe. The transported sediment/water collection methodology implemented in this study yielded only relative volumetric measures ; however, mass - movement pillars installed within one f ine - textured stripe during the field campaign will yield data sufficient for volumetric estimates of sediment removal upon annual visits to the site. Texture size percentages are expressed numerically by sample in Table 5. 3 . Textural classifications presen ted in Table 5. 3 confirm the silty texture of sediment transported in the coarse stripes. 53 Figure 5. 7 Particle size analysis of sediment trough data . Particle size curves above are expressed as a percentage of each individual sample volume. Higher curves indicate a higher occurrence of a given soil texture within the sample. Particle sizes define the texture of the soil. Trough names indicate the number of the trough first, followed by the terrace on which the trough was instrumented. Particle size analysis was also conducted on transported sediment collected from the snowpatch margins on both Terrace s 2 and 4 by Nyland (Nyland & Nelson, 2020) . On both Terraces 2 and 4, samples collected from snowpatch margins were dominantly of sandy texture. As indic ated in Table 5.2, transported sediment textures collected at coarse sorted stripe termini are dominantly silty. 54 Table 5 . 3 Sediment trough numerical p article size analysis results . Texture classifications based on the data presented in Figure 5.7. Troughs 1, 2, and 4 were installed at the base of coarse stripe termini while Trough 3 was installed at the base of a fine stripe. 5 .4 Discussion Results from the climate and soil monitoring equipment identify strong diurnal trends in conditions on both terraces and scale - based temperature gradients. Phase lags and subdued temperature amplitudes observed in soils at mast locations is indicative of c onductive heat transfer at depth. A slope - scale temperature gradient is inferred from temperature records between Terraces 2 and 4 warmer conditions are present at lower positions on the slope. A terrace - level gradient is implied from the sorted stripe tem perature monitoring equipment where the highest temperatures were recorded from the loggers operating closest to the CT toe. Substrate albedo seemed to impact recorded temperatures t emperatures on average were higher on fine stripe material compared to tem peratures recorded within coarse material. Differences in temperature occurred between loggers that were placed next to one another on the CT tread TB_3 and TB_4 but were placed within contrasting stripe material, indicating the presence of micro - scale top oclimatic influences on the hillslope. The largest fluctuations of sorted stripe temperatures were observed at the location closest to the snowpatch margin and closest to the CT toe. Large fluctuations at the location nearest the snowpatch demonstrate Trough % Clay % Silt % Sand Texture 1_2 11.7 57.5 30.8 Silt Loam 2_2 7.9 60.4 31.7 Silt Loam 3_4 5.8 16.6 77.7 Loamy Sand 4_4 5.7 44.6 49.7 Very Fine Sandy Loam 55 non - conductive heat transfer induced by diurnal pulses of water through the coarse stripes. Pronounced fluctuations in temperatures near the snowpatch margin may also be attributed to localized cool, dense air descending from the snowpatch. Near the CT tread m argin, temperature variability may be related to contrasts in albedo between the darker solifluction lobes and the lighter - colored coarse stripes. A feature - level temperature gradient is indicated by the air, 10 cm depth, and 30 cm depth temperature senso rs, where temperature fluctuation was less pronounced. Soil moisture monitoring confirms the presence of moisture within fine stripes and displays a vertical gradient of soil moisture in which moisture increases with depth. Soil moisture conditions, especi ally near the ground surface, display a diurnal trend that follows diurnal temperature fluctuations. The presence and behavior of soil moisture within the fine stripes at both sites promotes solifluction and frost creep. Silty material collected from th e base of coarse - stripe termini and the overflowed 19 - liter collection jugs confirm the active transport of silt - sized material suspended in water via the coarse stripes. Differences in textures from material collected at sorted stripe termini and snowpatc h margins could be attributable to the fluvial sorting of material as it is transported away from the snowpatch margin towards the CT tread. Fluvial sorting as water and sediment is transported across the CT tread would help to explain the occurrence of si lt fans that extend from the termini of coarse sorted stripes, as illustrated in Figure 5.1. 5 .5 Conclusion Climate monitoring records and soil texture analysis characterize periglacial processes operating on Terraces 2 and 4. Strong diurnal temperature f luctuations were observed on both 56 terraces at three distinct spatial scales; slope level, terrace level, and feature level. At the slope level, a gradual increase in air temperature exists. The occurrence of snowpatches at both the snowpatch on the lower t errace being the larger of the two demonstrates that snow mass balance is in fluence d by the relationship between snow accumulation and topographic hollows that shield the snow from incoming solar radiation. At the terrace level, increases in temperature were observed as one moves from the snowpatch margin toward the CT tread margin , a consequence of water being warmed as it travels across the terrace. At the feature - level, marked differences were observed within coarse and fine stripes. Temper ature amplitudes achieved in the coarse stripes closest to the snow margin demonstrate the non - conductive transfer of heat attributable to pulses of water flowing through coarse stripes. Temperature amplitudes within coarse stripes at progressively downslo pe positions indicate the flow of water in the large interstitial boulder spaces. A system of sediment transportation driven by the onset of warm temperatures is indicated through the analysis of the quantitative data presented in this chapter. With the onset of warmer temperatures early in the day, snowpatches warm and release meltwater downslope, which is demonstrated clearly by large temperature fluctuations recorded by the temperature logger placed at the snow patch margin. The lag in temperature peak s within coarse stripe temperature loggers further downslope reflect the propagation of water through coarse stripes, reflecting the hydrologic connectivity that exists within coarse stripes. As water flows downslope through coarse stripes, coarser - texture d material is deposited higher on the slope while finer - textured material continues to flow within coarse stripes. Near the CT tread toe, sorted stripes terminate in large silt fans sometimes burying their lower reaches and 57 resulting in the formation of tu rf - banked solifluction lobes. Solifluction and flowing water both contribute to the efficient transport of water and sediment across CT treads. Diurnal decreases in temperature likely influence the potential of snowmelt water contributions. In this interpr etation, nivation is responsible for much of the transport of material on the hillslope, a phenomenon that can be used to infer the lack of material ramparts on subjacent CT treads. Sediment transport on the hillslope is strongly influenced by both tempera ture and moisture availability. Underlying permafrost conditions and depth of sorting within the patterned ground network limit the vertical percolation of water, providing a maximum depth of hydrologic activity attributable to CT flatness. 58 Chapter 6 Network analysis of a S orted P atterned G round F ield The patterned ground field occupying the majority of the T1 cryoplanation tread atop Frost Ridge is comprised primarily of sorted stripes, although sorted circles, sorted nets, and turf - banked solifluction lobes are also found within the field. Sparse vegetation overlying a fine - textured substrate is also present in the sorted patterned ground field. Fieldwork conducted at this site in the mid - 1970s confirmed that snowmelt water from a late lying snow patch upslope used to flow within the coarse segments of the patterned ground field throughout the summer (Nelson, 1975). Proximity to a perennial snow patch, snow meltwater, and cool climatic conditions promoted the development and maintenance of the patterned ground. Today, the patterned ground field, although seemingly geomorphically dormant owing to earlier disappe arance of the snowpatch, is still intact and in fact, bears resemblance to a drainage network. This conceptualization has not previously been explored in the context of CT formation. The methodology used in this thesis to address a network of well - develope d sorted patterned ground atop a cryoplanation tread explores: 1) the efficacy of hydrologic modeling in a remote upland periglacial environment; 2) the comparison of manual versus automated channel detection; and 3) quantification of the characteristics o f a sorted patterned ground channel network. Results of the channel network analysis will be used to infer local hydrology, an approach that has not previously been attempted in the context of sorted patterned ground (cf. Godin et al. 2019). Network analys is may provide insight into a system of features that operates to evacuate water and sediment from snowpack margins near CT scarp backwalls, 59 thereby clarifying some of the problematic aspects of the nivation hypothesis raised by Hall (1998) and others, and contributing to enhanced understanding of CT formation. 6 .1 Remote Sensing in Geomorphology Remote sensing is a standard tool in geomorphic investigations (Rhoads, 2004) . Geomorphology, a subdiscipl ine of physical geography, is the study of landforms and landscapes (Short & Blair, 1986) and is concerned primarily with surface morphology and composition (Smith & Pain, 2009) . When available, remotely sensed data provide useful information if a site is difficult to access due to physical restrictions, enabling geomorphic stud ies. Utilization of remote sensing is particularly helpful when research topics are focused in - eye - perspective of satellites and other unmanned aerial vehicles permit analysis of processes that op erate over wide areas, revealing relationships not clearly visible from the ground (Short & Blair, 1986) . Advances in computer technology, storage capacity, and the amount a nd availability of remote sensing data have added to the utility of such data in physical geography (Rhoads, 2004; Short & Blair, 1986; Smith & Pain, 2009) . When remote sensing data are integrated with field - based validation, geomorphic studies benefit from increased accuracy and testing of previously untestable hypotheses (Smith & Pain, 2009) . The state of remote sensing data acquisition and produ cts as they relate to the methodology of this chapter is presented below. 6 .1.1 Data acquisition . The term remote sensing refers to the acquisition of data from a distance (Colwell, 1966, 1; Campbell et al. , 2011, 6) . The process of remotely acquiring information about the land surface involves the physical object of study, a sensor such as a camera, and a platform on which the sensor is mounted, e.g., an unmanned aerial vehicle 60 (UAV). Sensors are responsible f or capturing and recording the electromagnetic energy being emitted from the ground surface acquired during a survey, or the period of data acquisition (B runn et al., 2004, 112) . A temporal record of aerial imagery is achieved when an area is repeatedly surveyed by a sensor at different times. The most commonly derived data from a taken from above (Smith & Pain, 2009) . Electromagnetic information (sunlight reflected from the ground surface) nverted into information that reflects the physical properties of the features of interest. For this reason, remotely sensed imagery is a model or 2004, 119 ). 6 .1.2 Data . Aerial images, photos of the ground surface taken aloft, are converted to surface that depict details as imagery and have been adjusted to a standard datum and map projection (Jensen, 1995; Brunn et al., 2004, 124). Orthoimages reflect the positional and electromagnetic characteristics of objects on the ground. Orthomosaic maps, which are particularly useful products, are produced when two or more orthoimages are stitched together, giving the impr ession of one continuous image (Fernandez, Garfinkel, & Arbiol, 1998) . Orthomosaic maps have the benefit of a seamless appearance, as well as having large spatial extent. generated from aerial imagery (Zhang & Montgomery, 1994). DEMs reflect the topography of a elevation information. A continuous elevation surface is created through interpolation methods 61 such as kriging. Together, orthoimages and DEMs offer a strong visualiza tion tool that can be used qualitatively for visual analysis and quantitatively as data inputs for modeling efforts implemented in a geographic information system (GIS) (Kamp et al., 2005; Smith & Pain, 2009). 6 .1.3 Applications of remote sensing models in periglacial geomorphology . Physical geographers and other earth scientists benefit from the spatial and temporal coverage offered by remote sensing (Short & Blair, 1986; Boyd, 2009 , 456 & 645) . In certain cases, data of i nterest may be difficult to access or difficult to measure, e.g., remotely located sites, and tree canopies. A sensor mounted on aerial craft capture images covering large swaths of land in a fraction of the time that would be required for a human to trek across the same terrain. Similarly, the spatial coverage offered by aerial surveys encompass landform - size data (Singh, 2018) scales (Short & Blair, 1986, ix) , in addition to open - source data repositories, e.g., Google Earth, provide easy access to remotely sensed information at temporally robust scales (Smith & Pain, 2009) . Remotely sensed data products have been used in periglacial regions to assess the overall spatial organization of features, to test previously untestable hypothese s, and to map hard - to - access features. A study conducted by Nyland and Nelson (2020) used high - resolution DEMs to estimate denudation rates by nivation. Volumetric comparison of marginal drainage features and incipient cryoplanation terraces was conducted to estimate rates of erosion attributable to nivation (Nyland, 2019, 86 - 87) . Results indicate that a nivation - altered hillslope had achieved the appearance and size of other cryoplanation terraces found in unglaciated Beringia since deglaciation of the area (Nyland, 2019, 87) . Working in the same study area, 62 Queen (2018) used large - scale geomorphometry to map periglacial features across Alaska and in northwestern British Columbia. Using elevation data from the Arctic DEM - Polar Geospatial Center (Porter et al., 2018) and field obser vations, Queen constructed local - scale maps of periglacial assemblages at several sites distributed across eastern Beringia, documenting the presence of charact eristic periglacial landscapes. Similar efforts to map periglacial features were conducted by Grosse et al. (2005) . Automated identification of features and manual digitizing were used and the results of the two methods compared to assess the accuracy of u sing remotely sensed imagery in feature extraction. High - resolution maps of periglacial features produced from the automated extraction of features using satellite imagery and DEMs underline the utility of such data inputs in periglacial geomorphology. Kam p et al. (2005) used elevation data from the ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) sensor to produce a map of periglacial features including solifluction lobes and patterned ground. The results of that study provide support for the use of remotely sensed elevation data to map periglacial features. Other inventories of periglacial features using air photographs, digital elevation models, or remote sensing products include Evans et al. (2017), , Mithan et a l. (2019), and Eichel et al. (2020). Results from studies using remotely sensed imagery as data input indicate that use of such data is feasible and yield useful applications in periglacial environments. The integration of remote sensing into periglacial r esearch broadens the spatial extent of the studies (e.g., Queen 2018) and enables testing of hypotheses that are otherwise difficult to evaluate because of field limitations (e.g., Nyland, 2019). 63 6.2 Drainage Network Analysis A drainage network is comprised of paths on the land surface along which water flows (Tarboton, 1989, 11) . Sorted patterned ground features often occur as an interconnected system of boulder - filled gutters on the ground surface. Such patterned ground fields resemble drainage networks but have not been quantitatively explored as such. Quantitative methods to c haracterize drainage networks are abundant in the literature and provide useful metrics to define drainage networks past general qualitative assessment. 6 .2.1 Sorted patterned ground drainage channels . All sorted patterned ground forms are comprised of bo ulder/stone gutters and areas of fine soil cells (Goldthwait, 1976) . Sorted patterned ground forms often develop as net works of interconnected circles, polygons, and stripes (Washburn 1956), and are prominent periglacial features. Boulder gutters, which are the defining boundaries of sorted forms, connect to one another, unifying the individual forms. Elongated sorted stri pes display this arrangement on slightly sloping surfaces and serve as a shallow drainage system on hillslopes (Goldthwait, 1976). Within this sorted network, water flows through the large interstitial spaces of the boulder gutters, promoting removal of fi nes from the gutters. The interconnectedness of sorted stripes and the available interstitial space offered by boulder gutters facilitate a system of water and sediment movement, a phenomenon that has been documented over the past half century (Caine, 1963; Smith, 1968; Nelson, 1975; Goldthwait, 1976; Queen, 2018, 57) . Although sorted patterned ground has been documented as effective water channels for removal of sediment, (e.g., Nelso n, 1975), no study has been conducted that assesses the geometric or topological arrangement of such networks. 64 6 .2.2 Drainage n etwork Analysis . River network characteristics were first quantified in terms of stream order by R.E. Horton in 1945 (Melton, 1957, VIII) . Previously, drainage networks had been qualitatively classified based on pattern (e.g., Thornbury, 1969). The stream - ordering methods presented by Horton were lat er revised by Strahler (1952) to produce the widely known Horton/Strahler stream ordering method (Smart, 1978) . The methodology presented by Horton involved identification of basin stream order, the most fundamental drainage basin assessment. Stream ordering gives rise to other quantitative drainage - basin parameters, such as stream length and number of streams. Horton related stream order and The bifur cation ratio (R b ) is the ratio of the number of stream lengths of a given order (N) to the number of stream lengths of the next highest order (N+1). Bifurcation ratios generally range from 3.0 - 5.0 and indicate degree of structural influence (Pareta & Pareta, 2011) . Stream length ratio (R L ) is the ratio of the stream segment lengths of the next lower order (L - 1). Stream length ratio varies at the basin and sub - basin levels, and indicates the relationship between flow discharge and the erosional stage of the basin (Hajam, Hamid, & Bhat, 2013) . Drainage density (D d ) is the ratio of the total length of all strea ms in a drainage network (L T ) to the drainage basin area (A). Drainage density is an indicator of drainage basin dissection, where a high value of D d is related to a highly dissected basin (Melton, 1957, 35; Hajam et al., 2013) . Drainage density also indicates relief steep basins tend to have a low drainage density ratio value (Melton, 1957, 37) . The above - stated morphometric calculations are given by the formulae: 65 The Strahler stream ordering method (Strahler, 1952) is used in this study. The Strahler ordering method assigns an order N = 1 to all streams that do not have tributaries. Downstream from the first - order streams, when two or more streams of the same order N meet, an order of N + 1 is assigned. If stream order N is met by a stream order less than N, no change to the downstream section occurs. All sediment and water flow out of the highest order stream lengths. A higher stream order is associated with a high er basin discharge (Hajam et al., 2013). (Smart, 1978) and are still implemented ( e.g., Pareta & Pareta, 2011; Hajam et al., 2013) . Commonly accepted analyses of drainage networks involve calculation of Horton as well as parameters suggested by other workers. Hajam et al., (2013) and Pareta & Pareta (2011) used quantitative metrics presented by Horton (1945) and others, e.g., Schumm (1956) and Strahler (1954), to quantitatively detail their drainage bas in study areas and to assess hazards such as flood risk. The work done by Horton (1945) confirmed that numeric data can be related to form. From form, processes can be inferred. Calculation of quantitative parameters Bifurcation Ratio (6.1) Stream Length Ratio (6.2 ) Drainage Frequency (6.3 ) Drainage Density (6.4 ) Relief Ratio (6.5 ) 66 can confirm field conditions on the gro und and reveal relationships that are not observable from maps. 6 .2.3 Hydrologic flow models . Numeric hydrologic modeling involves prediction of hydrologic activity through implementation of algorithms within computer software (Martz & Garbrecht, 1992) . Algorithms are usually parameterized by data derived from field measurements (Jayawardena, 2014, 16) because field measurements promote accurate prediction of hydrologic responses (Beven & Kirkby, 1979) . Models representing natural features include digital elevation models (DEMs), which are gridded representations of the Ea (Martz & Garbrecht, 1992) . Hydrologic models are useful because outputs of algorithms such as stream number and order are often difficult if not impossible to accurately measure in the field. Because drainage characterization is critical in land/water management, hydrologic modeling has become an integral component in such management strat egies (Arnold, Srinivasan, Muttiah, & Williams, 1998) . Stream modeling is commonly used as base da ta to quantify drainage basin morphology, (e.g., Hajam et al., 2013) The advent and ongoing development o f computers and geographic information systems (GIS) since the 1960s have enhanced processing ability and storage of hydrologic data, spurring advances in hydrologic modeling (Arno ld et al., 1998; Singh, 2018) . Various models have been developed and used since the 1850s, (e.g., Mulvaney 1851) to predict surface runoff, subsurface flow, and groundwater storage (Singh, 2018) . Flow models use assumptions about the influence of topography on the flow of water to digitally r epresent flow paths based on the individual cells in the DEM grid (Martz & Garbrecht, 1992) . Most flow 67 models are adapted for DEMs, which are the primary data inpu t for flow - model algorithms. Flow models have applications in channel identification and can aid in the quantitative analysis of channel networks because the digital representations produced by flow models are easily analyzed within a tabular interface A series of steps implemented with a GIS to extract or identify stream channels within a DEM. The D8 algorithm assumes that water from one cell of the DEM flows completely int o one of eight implementation can be broken down into three steps: 1) artificial pit removal and flow direction computation DEM, 2) flow accumulation calculation, a nd 3) definition of channels based a user - defined threshold of flow accumulation cells (Tarboton, 1989, 56). The result of the D8 application over a DEM is the identification of streams where stream channels represent those connected cells with the larges t flow accumulation values. The D8 flow algorithm is a useful method to map networks of channels when a DEM grid is available. Other flow models developed, e.g., the multiflow direction algorithm from Quinn et al. (1991), are based on the conceptual framew ork of the D8 model but adjust the generation of flow direction to reflect the ability of water to flow in multiple directions, or into multiple adjacent cells. The main conceptual difference between the D8 and multiflow direction algorithms is that the D8 model assumes that flow is directed only in the direction of the steepest slope, i.e., convergence while the multiflow algorithm assumes that flow is directed into all downslope directions , i.e., divergence (Wolock & McCabe, 1995) . The D8 model is implemented in this study because the effects on channel extraction using either D8 or multiflow algorithms are minimal (Wolock & 68 McCabe, 1995) . I n the case of the patterned ground field on Frost Ridge, both the D8 and multiflow channel detection methods produced similar channel ide ntification results, highlighting the prominence of convergence on the hillslope. 6 .2. 4 Applications of flow modeling in geomorphology . McNamara et al. (1999) utilized flow modeling with a DEM as data input to locate water tracks at a high - latitude site. Through the organizational characteristic s of a drainage basin underlain by permafrost. The results of network, probably due to underlying permafrost conditions that inhibit incision. They postulate that as unde rlying permafrost conditions are compromised under warming trends in climate, incision of water tracks will commence, contributing to hillslope erosion. Water tracks in Antarctica have also been identified using flow modeling infer groundwater activity (Levy et al., 2011) . The identification of water tracks in that study, in conjunction with field observations, revealed that water tracks we re efficient transporters of meltwater and rock - weathering - derived solutes, indicating their geomorphic significance. Flow model algorithms have been used to parameterize TOPMODEL (Beven & Kirkby, 1979) , a hydrologic model that has been used in a wide array of hydrologically based predictions and simulations. Flow models can be used to infer depth to water table, which is parameter in TOPMODEL (Beven & Kirkby, 1 979; Wolock & McCabe, 1995) . Flow modeling uses assumptions about water flow and elevation inputs to predict flow paths. Identification of drainage networks using flow models produce outputs, e.g., stream order, stream length, and number of streams, th at can be used as numerical input to 69 characterize the morphology of the drainage basin (e.g., Gleyzer et al., 2004; Pareta & Pareta, 2011; Hajam e t al., 2013) . Modeling of flow pathways is an important tool to predict the distribution of moisture in the subsurface, to understand the organization of water features, and to infer geomorphic processes. Understanding the distribution of such features is especially important in periglacial regions, where water features are sometimes difficult to see from ground level, e.g., water tracks, or where the local hydrology has not yet been confirmed by quantitative data, e.g., a drainage network comprised of sorted stripes. 6 .3 Network Analysis of a Sorted Patterned Ground Field Sorted stripes are characteristic periglacial features that represent subsurface flow pathways for water. Aerial imagery and elevation data can provide insight into the organizational structure of sorted stripes. Flow modeling is a useful tool to identify drainage patterns and to obtain stream parameters such as stream length and number of streams. Stream parameters derived from flow modeling, implemented on a DEM, facilitate the quant itative analysis of the drainage network. The availability of high - resolution orthophotos and a DEM facilitate the application of the D8 flow model on the sorted patterned ground field. Quantitative characterization of it as a drainage network can be conducted once flow models have identified the locations of subsurface flow on the CT tread. Two technical approaches can be used: (1) manual digitizing to identify features; and (2) use of automated flow models to identify stream pathways. Both the manually digitized pathways and those derived from th e flow model can be used to characterize the organization of the patterned ground field. The organization of the patterned 70 ground can then be used to in evaluate the geomorphic significance of sorted stripes in cryoplanation terrace formation, specifically in the context of the nivation hypothesis. 6 .3.1 Study area . A series of incipient cryoplanation terraces are incised into Frost Ridge, imparting a notched profile appearance to its north - facing flank. A well - expressed patterned ground field occupies a gently sloping CT tread located at the junction between Splinter Peak and the uppermost cryoplanation terrace tread (Figure 6.1). This field of patterned ground is dominated by sorted stripes extending from the scarp - tread junction to the toe of the CT tre ad. Other types of sorted patterned ground, frost - fractured clasts mantling slopes, solifluction lobes, needle ice creep, and nivation hollows are periglacial features also found on Frost Ridge (Queen, 2018, p. 42) . The widths of sorted stripe unit s range from 1.7 m to 5 .4 m, and are 2.75 m wide on average (Queen, unpublished data ) . There is no apparent correlation between the (1976) assertions about regular spacing. The pattern - diameter to depth - of - sorting ratio is 3.44 , within the ran ge reported by Uxa et al. (2017). The underlying geology of the site is Paleozoic sedimentary rock (Queen, 2018, 5 6 and references therein) . Additional site information is summarized in Table 6.1. 71 Figure 6 . 1 Panoramic view of Frost Ridge and inset photo of sorted patterned ground features . The dashed white line delineates Frost R idge; the black box indicates the location of the inset photo. Within the inset map, sorted stripes are oriented northward. Dark brown areas are fine - textured sediment, lighter gray, coarse - textured areas are coarse sorted stripes. Table 6 .1 Field site topographic information . *Elevation at the scarp - tread junction * * Tread length is approximated by measuring the horizontal distance from the scarp tread junction to the ridge of the CT scarp immediately downslope. 6 .3.2 Study area data acquisition . A DJI Mavic 2 Pro drone was used to survey the patterned ground field, covering a total area of 0.143 km 2 in the fall of 2018 (Figure 6.2). Using DJI mission flight planning software, a double grid flight path was flown with 80° camera angle, 85% front and 82% side overlap. No ground control points were used. The orthomosaic , DEM, and a quality report were generated using Pix4Dmapper Pro version 4.2.27 software (Pix 4D Total relief (m) 29 Average slope (°) 10 Aspect North, northeast Elevation (m) * 1690 CT tread length* * (m) 279 72 Inc., San Francisco, California). Areas with 3 - 5 plus overlapping images were deemed acceptable for analysis. The area outlined in Figure 6.2 was chosen to lim it the number of pixels that did not have sufficient overlap to be included in the network analysis. The study area used in the analysis includes the scarp backwall and most of the CT tread extending from the base of the scarp - tread junction. The patterned ground study area lies within areas considered to be of good quality. A spatial resolution of 3.27 cm/pixel was achieved for both the DEM and orthophotos. 73 Figure 6 . 2 Aerial survey data . A UAV survey flown September 4, 2018 northwest at 80 meters above the takeoff site located in the northwest portion of the map (indicated by an arrow). An orthomosaic (A) and Digital Elevation Model (B) were generated though post - survey processing using Pix4DMapper Pro software. Coarse stripes are light gray, highl y textured linear features oriented approximately parallel with the slope. Within the DEM, elevations are represented as a gradient from white (relatively higher elevations) to black (relatively lower elevations). The gradient indicated the north - northeast orientation of Frost Ridge The extent of the field study site is indicated by the solid white outline. The scarp is indicated by a dashed white line, separating it from the outward - extending CT tread. Data quality decreases with increasing distance from t he patterned ground field study area. Black areas outside the map area contain no data values. Remote sensing products courtesy of Merlin Geoscience Inc. 74 6 .3.3 Network analysis inputs . Before flow modeling was implemented in the study area, visual analysi s was used for preliminary characterization of the patterned ground field. Using the orthomosaic as a guide, manual digitizing was used to identify coarse portions of the sorted stripes thought to channel water and transported sediment across the CT tread. The polyline (Environmental Systems Research Institute, Inc., 2020) was used to delineate the coarse patterned ground from intervening fine - textured areas. The coarse portions of sorted circles and net s found near the scarp - tread junction were identified by Nelson (1975) as channels for water flow, based upon visual assessment and the sound of water running in them throughout the summer. These features grade into sorted stripes found just downslope. Sor ted nets and circles were included in the digitized channel network in addition to sorted stripes. The arrangement of the features revealed a network reminiscent of a drainage basin. The process of digitizing the network was guided by the high - resolution orthomosaic, DEM (indicative of the downslope direction), and expert field knowledge derived from fieldwork conducted during the summers of 2018 and 2019. Other filters applied to the study area DEM and orthomosaic aided in the visual characterization of the sorted patterned ground field (Figure 6.3). 75 Figure 6 . 3 Compound filter map . creates a shaded relief surface taking the source angle and shadows into consideration, highlighting subtle relief features on the surface. The hillshaded surface was made partially transparent and draped over the orthomosaic map (B) to create a compound map (C). Solifluction lobes indi cated by small black arrows on map C , linear coarse stripes light gray areas , and the scarp backwall are especially prominent in the compound map. Darker features resembling raindrop shapes constitute the fine stripes, which are fine - textured sediment mo ving as solifluction lobes. From the compound map, solifluction lobes and coarse stripes appear to extend out from the scarp - tread junction to the edge of the study area boundary to the northeast. 76 Equations for the dimensionless ratios used for quantitati ve network analysis require numeric inputs such as stream order, derived from the results of flow algorithms implemented in a GIS. The D8 flow algorithm was used to model flow in the study area, and tabular data were used later to calculate the drainage ne twork parameters. The stream identification process was conducted on the DEM within GRASS GIS version 7.8.1 (GRASS Development Team, 2019) using tools within the Hydrologic Modeling toolset. Following the methodology for stream r.fill.dir tool to remove pits within the area. Using the r.watershed tool, flow direction and flow accumulation rasters were generated. Using the raster calculator, a conditional statement, which performs conditional if/else statements set by the user on each of th e input cells of an input raster, was used to determine a flow accumulation threshold, set to a value of 100,000. The assumption of this step is that only cells having a certain amount of flow accumulation constitute flow paths on the surface. The chosen f low accumulation threshold excluded noise from the cells with small flow accumulation values that are unlikely to represent drainage channels, while preserving the overall pattern identified by the drainage direction matrix. This threshold identified the d ominant flow paths extracted by the flow modeling algorithm. In the context of fluvial geomorphic activity, the flow accumulation value represents the amount of water convergence that must occur to initiate a flow path. Conceptually, this channelization re sults in the initiation of a sorted stripe on FR. Other workers have also indicated the importance of fluvial processes in sorted patterned ground initiation/formation (e.g., Nelson, 1975; Paquette et al., 2017b) . The flow paths identified indicate a minimum estimate of the hydrologic activity atop FR, as many of the minor channels have been excluded from the 77 analysis. Similarly, the flow accumulation threshold chosen for this analysis represents the upper limit of total convergence required to initiate a sorted stripe flow path. The results of the manual digitizing and stream identification processes are shown in Figure 6.4. The goal of this methodology was to identify the coarse portions of the patterned ground network, to identify its overall pattern and orientation as verified by the digitized network, and then to derive the numeric data needed to quantitatively assess the drainage basin and channel network. Figure 6.4 Drainage network identification results . ( A) Flow pathways identified through manual digitizing e fforts. (B) Flow pathways identified through flow modeling methodology. Both results indicate an interconnected system of coarse components of the patterned ground. 6 .3.4 Comparison of digitized and automated drainage network extractions . The results of both manually digitized and automated stream extraction methods are shown in Figure 6.4. The automated map identifies the most prominent stream channels based on basin topography. Both maps identify areas of accumulation that connect with on e another, 78 constituting the stream channels in both networks. Areas of accumulation were identified in the manually digitized map based on the spectral differences between the coarse and fine - textured stripes. The automated channel detection method did not identify the coarse stripes in the same locations as they were identified via manual digitization. This result most likely reflects the microtopography on the seemingly flat CT tread: linear piles of coarse fragments constituting the coarse stripes are to pographically higher relative to intervening bands of fine - textured sediment, which function as solifluction lobes. The flow algorithm assumes water flow through topographically low points on the slope, however, water flows within topographic high lineatio ns on FR. The results of the automated channel detection method were channels located at the borders of coarse and fine stripes , as opposed to the interior of coarse stripes . For identifying terrace - level drainage basin organization, the automated methodol ogy was sufficient because the D8 flow algorithm was able to capture large - scale sorted stripe feature organization , based on visual comparison with the manual channel detection method . Both maps depict similar drainage network patterns where stream channels extend outward from the scarp - tread junction -- in the southwestern corner of the map in a northeasterly direction. Generally, streams in both maps follow the northeasterly aspect of th e CT tread but at the southeastern edges of the study area, streams assume an easterly flow path. Flow paths on the CT tread probably reflect the slightly convex topography of the tread (Nyland, 2019 , 57 58) . The predominant orientation of sorted stripe s in both maps is reminiscent of a parallel drainage network ( Hobbs, 1910; Zernitz, 1932) although the entire CT tread must be analyzed to make conclusive assertions about the organizational structure of the drainage network on Frost Ridge. The predominantly parallel na ture of the network depicted in Figure 6.4 is part of a 79 larger distributary network that reflects the slightly convex topography of the CT tread and clearly depicts flow off the side of the treads eastern margin. Visual assessment of the maps in Figure 6.4 is useful for an integrated characterization of the drainage on the entire CT slope. the network occur near the scarp - tread junction while the network mouth is located near in the CT tread toe. Both the interconnectedness of the patterned ground field and the organization of the stripes into a parallel network extending toward the CT toe indicates efficient transportation of snowmelt water and sediment a cross and over CT treads and side slopes. The stream order of the drainage basin based on flow modeling data is shown in Figure 6.5. Stream ordering was based on the Strahler method. The FR drainage basin is a third - order basin. First - order streams are m The network converges into third - order streams near the toe of the CT tread. Overall stream discharge and channelization of water is probably highest in the 3 rd order streams. Stream order of the FR drainage network confirms the direction of the flow of water and helps to locate areas of elevated discharge on the CT tread. Comparison of the automated network identification map with the manually digitized map shows that automated methods to delineat e a periglacial drainage network can be used to achieve useful flow pathway maps. Automated maps are particularly effective for obtaining drainage parameters such as stream order, whereas manual methods of stream delineation tend to be labor intensive and subject to user error, possibly resulting in biased network parameters. 80 Figure 6 . 5 Stream ordering results . The above stream ordering map was generated using the Stream Order tool (Environmental Systems Research Institute, Inc., 2020) and demonstrates an overall stream order of 3 . The Strahler stream order method (Strahler, 1952) was used. Black arrows indicate the flow direction. Light gray circles represent the main outflow points of the drainage network, or the areas where the most water is thought to be channeled out through based on the high order stream channel. 81 6 .3.5 Network analysis results . Drainage basin characteristics were quantified usi ng methods presented in Horton (1945), Strahler (1952 & 1964), and Schumm (1956). All numeric information required to calculate the parameters presented in the drainage network analysis was sourced from the DEM of the sorted patterned ground field. Multipl e metrics were chosen to capture the breadth of the drainage basin characteristics. Inputs to the chosen morphometric parameters are summarized in Table 6.2. The results of the morphometric parameter calculations are summarized in Table 6.3 . Table 6 .2 Ancillary d rainage b asin d ata . Values in this table were derived from a GIS based on the study area DEM. Some values, e.g., basin relief were measured within a GIS using ruler tools. Stream order and number values were calculated following the implement ation of the D8 flow algorithm. Values in this table were used in the quantitative drainage basin characterization. Parameter Value Source Basin Relief, R , (m) 29 * Basin Area, A , (m 2 ) 26,382 * Basin Perimeter, P , (m) 635 * Number of 1st Order Streams, N 1 50,085 ** Number of 2nd Order Streams, N 2 31,496 ** Number of 3rd Order Streams , N 3 5,249 ** Total Number of Streams, N T 86,831 ** Length of the Principal Drainage Line, L P , (m) 184 * Total Stream Length, L T , (m) 3,420 ** Mean Length of 1 st Order Stream, 1 , (m) 4.1 * Mean Length of 2 nd Order Stream, 2 , (m) 4.7 * Mean Length of 3 rd Order Stream, 3 , (m) 19.0 * Mean Length of all Streams, T , (m) 4.5 * * Data derived from ArcMap v. 10.6 software ** Data derived from GRASS GIS v. 7.8.1 software 82 Table 6 .3 Quantitative drainage basin characterization . Parameter values were calculated and are organized by parameter types. Stream order ratios are specified for the bifurcation and stream length ratio calculations. For example, the first bifurcation ratio calculation (1 st :2 nd ) indicates that the calculation was performed using the 1 st order stream data as the numerator and 2 nd order stream data was used as the denominator. A list of sources, indicated by stars, is located below the table. Both Horton (1945) and Strahler (1952) are listed as sources for stream ing method; the Strahler method was used for the classification of the drainage basin in this study. Parameter Value Source Linear Bifurcation Ratio, Rb 1 (1st:2nd) 1.6 ** Bifurcation Ratio Rb 2 (2nd:3rd) 6.0 ** Stream Length Ratio, RL 2 (2nd:1st) 1.2 ** Stream Length Ratio, RL 3 (3rd:2nd) 4.0 ** Stream Order 3.0 **, *** Areal Drainage Frequency, D f , (m - 1 ) 3.3 * Drainage Density, D d, (m - 2 ) 1.7x10 - 4 ** Relief Relief Ratio, R r 0.16 **** * Equation from Horton (1932) ** Equation from Horton (1945) *** Equation from Strahler (1952) **** Equation from Schumm (1956) 6 .4 Discussion 6 .4.1 Bifurcation ratio . Bifurcation ratio (R B ) is the ratio of the number of streams of a given order, N , to the number of streams of the next highest order, N+1 , and serves an index of the relief and dissection of a basin and reflect the branching within a drainage basin ( Horton, 1945). Bifurcation can also indicate geologic control and degree of disturbance within a drainage basin (Pareta & Pareta, 2011 and references therein). Bifurcation ratios are not constant between order, and values range from 3.0 to 5.0 (Schumm, 1956, 603) . In flat or rolling drainage basins, lower bifurcation values are typical (Horton, 1945). Bifurcation values of 83 3 and greater are typical in highly dissected or mountainous drainage basins (Horton, 1945). The bifurcation ratio value of 1.6 for 1 st to 2 nd order streams field reflects the flat topography encountered in the interior of the CT tread. Typically, low - relief bas ins exhibit low bifurcation ratios because the degree of dissection is low in flat areas relative to high - relief areas. Less dissection resulting in a lower bifurcation ratio is especially true between 1 st and 2 nd order streams in the FR basin, a fact evid enced from the stream - order map (Figure 6.5). The 6.0 bifurcation ratio of 2 nd to 3 rd order streams reflects large amounts of bifurcation of 2 nd order streams into 3 rd order streams within the drainage basin. The higher bifurcation value for the 2 nd to 3 rd order is expected because near the CT toe tread, where the 2 nd and 3 rd order streams are located, slope tends to be greater than th at near the scarp - tread junction. Increases in slope impact the amount of bifurcation due to an increase in stream power, wh ich is related to relief. The bifurcation ratio for the FR drainage accurately reflects the subtle changes in slope that occur across the CT profile. The CT tread drainage basin tends to be more highly dissected closer to the toe and less dissected near th e scarp - tread junction, based on inputs from the flow model stream results. Attention should be paid to the influence of the flow accumulation threshold used in the flow model algorithm. Due to the accumulation threshold chosen for the analysis, the flow algorithm likely underestimated the number of 1 st order streams because those streams have less accumulation than 2 nd and 3 rd order streams and thus may have not been identified. Comparison of the two maps in Figure 6.5 underline this possibility. However, the bifurcation ratios accurately reflect the subtle sloping conditions that exist on the CT tread, suggesting that 84 the bifurcation ratios calculated can be used to quantitatively characterize the FR drainage basin. 6 .4.2 Stream length ratio . Stream length ratio is the ratio of the mean stream length of a given order, L , to the mean stream length of the next lowest order, L - 1 , and reflects the relationship between surface flow and discharge and the erosion stage of the basin (Hajam et al., 2013 and references therein). Stream length ratios calculated for the FR drainage basin are 1.2 for 2nd to 1 st order streams and 4.0 for 3 rd to 2 nd order streams, falling near the range of typically observed stream length ratios, which range from 0.5 - 3 (Horton , 1945). For 2 nd and 3 rd order streams, the high stream length ratio indicates a large jump in average stream length from 2 nd to third order. Increasing length with increasing order is typical in most drainage basins. Large flow accumulation values identif ied within the 3 rd order streams indicates inflated discharge rates from higher order streams within the basin. 6 .4.3 Drainage frequency . Calculated as the ratio of the total number of all streams, N T , to the basin area, A , drainage frequency reflects the texture of the drainage network and indicates substrate permeability. The drainage frequency value of 3.3 is high relative to results reported by others e.g., (Vittala, Govindaiah, & Honne Gowda, 2004; Hajam et al., 2013; Martins & Gadiga, 2015; Rai, Mohan, Mishra, Ahmad, & Mishra, 2017) , which range from 0.3 - 3.0. The high D f calculated for the FR drainage basin reflects the large number of water features identified by the flow algorithm. High D f values have also been attributed to the permeability of the substrate, where a high value indicates low substrate permeabilit y (Reddy, Maji, & Gajbhiye, 2004) . In the case of the Frost Ridge drainage network, the existence of a frozen substrate and the depth of sorting within the sorted patterned ground both contribute to low 85 permeability. Frozen material and the depth to fine - textured material would inhibit vertical flow, encouraging the dispersion of flow in a semi - horizontal plane. The dominance of lateral water flow could enhance the formation of sorted patterned groun d, which is closely related to moisture availability. Overall, the drainage frequency of the sorted patterned ground field indicates low permeability of the substrate and the abundance of drainage features identified by the flow algorithm. 6 .4.4 Drainage density . Drainage density, D d , is the average length, , of streams per unit area A , and indicates the degree of drainage within a basin. Rainfall and relief commonly influence drainage density values within a basin but infiltration capacity the maximum rate that soil can absorb water and erosivity of the basin substrate are also importan t (Horton, 1945). A well - drained basin will tend to have a lower drainage density than a poorly drained system. Because drainage density is calculated based on the average stream lengths, drainage density is related to the degree of branching within the ba sin. Basins with high D d values rarely exceed 3.0 and in most humid basins where soil erosion is active, D d varies between 1.0 - 2.0 (H orton, 1945, 359) . The calculated D d value for the FR sorted patterned ground basin, 1.7x10 - 4 m, is low based on common values presented by Horton (1945). The low drainage density values are likely due to the abundance of short stream lengths identifie d by the flow algorithm. Although some of the higher - order streams in the basin are long, much of the basin is dominated by relatively shorter stream lengths. The abundance of small features is responsible for the removal of water across the seemingly flat CT tread that is underlain by permafrost. The abundance of sorted stripe features on the CT tread result in a well - drained basin despite basin topography and substrate impermeability. 86 Low drainage density values are typically associated with youthful drainage basins that are dominated by 1 st order streams. This is not the case for the sorted stripes atop FR, which have been developing since the waning stages of the Wisconsinan. In some cases, basins that exhibit low drainage density values may actually be at a mature stage of development if overland flow has not eroded specific areas of the drainage basin due to flat slopes, high infiltration capacity, surface resistance to erosion, or a combination of all those factors (Horton, 1945). In the c ase of FR, relatively flat slopes and frozen substrate probably contribute to the propagation of the 1 st order streams and the low drainage density value. Because of the flatness and resistance to erosion due to frozen ground in the early spring when snowm elt water is abundant, water is unable to erode into well - defined channels. Water is then channeled through the coarse portions of the sorted patterned ground network , which are small forms relative to the total area of the CT tread. The low drainage densi ty of the FR basin indicates that the area is well - drained and when physical characteristics such as slope are considered, the drainage basin may also be characterized as mature. The results of the drainage density calculation suggest that sorted stripes c lassified as first order dominate the basin and contribute to the drainage of water from the CT tread. 6 .4.5 Relief ratio . The relief ratio of a basin is calculated as the ratio of basin relief, R , to the length of the principal drainage line, L p . Here, t he principal drainage line was determined by identifying the longest 3 rd order stream and measuring a line parallel to the flow path. A positive relationship between relief ratio and sediment loss was shown in Schumm (1956), suggesting that relief ratio can be used to predict sediment transportation within a drainage basin. Th e relief ratio of the Frost Ridge patterned ground network is 0.16, which is high based 87 on the values presented by Schumm (1956), which range from 0.009 in low - relief basins to 0.15 in high - relief basins. The high relief ratio in this case is not indicativ e of a high relief basin, which contradicts the trends found in Schumm (1956). However, the high relief ratio could have been influenced by the small spatial extent that was analyzed in this study compared to the regional/landscape - scale studies from which relief - ratio values have been reported. The high value calculated does provide some support for the occurrence of large amounts of sediment transportation on the slope , following research from Schumm (1956) . 6. 5 Conclusion The organization of the sorted patterned ground field on Frost Ridge was evaluated in this chapter through network analysis. Visual assessment of manually digitized and flow modeled drainage maps revealed that, although flow modeling can lead to underestimation of 1 st order strea ms in a basin, both drainage identification methods identified drainage patterns within the network that are reminiscent of a parallel drainage network, with ancillary distributary characteristics near lateral tread margins. The network of soil pipes analy zed by Bernatek (2015) has similar characteristics. Quantitative analysis of drainage basin characteristics based on flow algorithm inputs confirm trends in microtopography and slope on the CT tread. Calculated bifurcation ratios confirm that relief and sl development and sorted stripe bifurcation. High stream - length ratios within the drainage basin in addition to the location of 3 rd order streams indicate areas of greater stream discharge to e xist at the terminus of 3 rd order sorted stripes, which occur near the CT tread toe. Drainage frequency and drainage density values conform with the manually digitized map, which shows a 88 high density of sorted stripe features in accordance with the gentle topography shown in the DEM. Drainage frequency and drainage density values indicate that flow is dominantly lateral as vertical percolation is inhibited , owing to the frozen substrate and the depth of patterned ground sorting. The high relief ratio of the FR drainage basin indicates that sorted stripes are related to effective sediment transportation. Three main points are derived from the network analysis of the sorted patterned ground field: 1) the abundance of sorted stripes results in a highly dissecte d drainage network; 2) the flow of water within the sorted stripe network is likely dominantly horizontal; and 3) the greatest amount of discharge is probably located at the terminus of 3 rd order streams, which occur near the CT toe tread. The agreement of quantitative drainage parameters with the manually digitized map and field observations indicate that flow modeling of a sorted patterned ground network is an appropriate analytic methodology. Quantitative basin characteristics also indicate the operati on of phenomena, such as sediment transportation, that are not known based on visual assessment of remotely sensed data. The visual and quantitative assessment of the sorted patterned ground field lend support to the hypothesis that the network of sorted p atterned ground atop FR is conducive to the efficient transportation of water and sediment despite the gently sloping conditions on the CT tread. Quantitative parameters help to define the hypothesis that the highly dissected network directs water laterall y though the sorted landscape toward the CT toe tread, facilitating the transportation of water and sediment across the tread. Based on these results, the lack of material accumulations on flat CT treads may be attributed to the flow of water within a sort ed patterned ground field. The flow of water through the sorted landscape would probably promote further development of sorted stripes 89 while helping to maintain the flatness of the tread through the removal of fines over low permeability substrates. Furthe r analysis of this drainage basin could be improved by confirming the locations of first - order stripes , underlying permafrost conditions , and comparison of morphometric parameters that have been calculated for other sorted stripe drainage basins. 90 Chapter 7 Synthesis and Conclusions 7.1 Summary of Research Competing hypotheses of CT formation lie primarily within two distinct groupings. The first group attributes CT form to geologic structure (e.g., French, 2016) whereas the second, and more widely accepted hypothesis, attributes CT formation to climatic influences, particularly from the microclimate in and around late - lying snowpatches. The second formation hypothesis, known as the nivation hypothesis, requires investigation of the capability of periglacial processes to initiate CT formation, (e.g., Lauriol et al., 2006) , the existence of active periglacial processes that maintain or further modify CTs (e.g., French, 2016) , and the lack of material rampart accumul ations o n inner CT treads (Ballantyne, 2018). Recent research lends support to the time - transgressive nature of CT tread growth (Nyland, 2019; Matthews et al., 2019; Nyland & Nelson, 2020) the poleward orientation of CTs (Nelson, 1998) , the widespread occurrence of CTs in periglacial regions across the globe (Demek, 1969) , and repeating assemblages o f periglacial features on CTs (Queen, 2018), all pointing to a strong association between CTs and periglacial conditions. However, previous work fails to account for the absence of ramparts of weathered material on CT inner treads, contributing to skeptici sm about the nivation hypothesis. The work presented in this study focuses on relating periglacial processes to CT formation and also challenges the notion of geomorphic quiescence implied by the occurrence of sorted patterned ground (e.g., Karte 1979, 81; Thorn & Hall, 2002b, 545) . 7.1.1 Field - based process investigations . Installation of temperature and hydrologic monitoring equipment on two incipient cryoplanation terraces was undertaken to fulfill the need for process - focused investigations on CTs (e.g., Demek, 1969) that link nivation to CT 91 formation. The short temperature monitoring data records available from Frost Ridge clearly display cyclic diurnal temperature behavio r, at distinct scales. At the slope level, a temperature gradient exists between upper and lower positions on the slope, confirming the expected air temperature decreases with increasing elevation. At the terrace level, a temperature gradient extending fro m the scarp to the CT tread toe is also apparent. At the feature level, distinctive diurnal temperature behaviors were observed in the coarse and fine stripes. Within the coarse stripes, the temperature records demonstrate the impacts of surface albedo, pr oximity to the snowpatch margin, and non - conductive thermal processes on the thermal regime. Pulses of snowpatch meltwater moving through the coarse stripes, coinciding with the onset of increasing air temperatures, constitute non - conductive heat - transfer processes operating within coarse stripes. These pulses and the large interstitial spaces within coarse stripes contribute to the spiky temperature record observed. Within the fine - textured stripes, temperature records displayed less extreme diurnal amplit udes, except near the toe of the CT tread, indicating the dominance of conductive heat transfer within fine - textured stripes where surface temperatures are largely impacted by larger - scale (terrace - scale) temperature gradients and the dark, low - albedo surf aces. Phase lags and progressively subdued diurnal amplitudes with depth in the vertical temperature profile of a fine stripe confirm that heat transfer in the fine stripes was primarily conductive during the period of observation, although variations in s oil moisture at the diurnal scale significantly impact the ground thermal regime. Sediment troughs and overland flow - collection jugs indicate key outflow points that occur at coarse stripe termini. The duration of the field campaign provides a rough enve lope of fill time for the overland flow jugs, which imply flow volumes in excess of 19 liters per day from 92 a single sorted stripe, coinciding with summertime temperatures, upslope snowpatch conditions, and the thermal regime of the coarse stripes. The comp arison of coarse stripe termini and snowpatch - margin transported sediment via particle size analysis is strong evidence for sediment transport by fluvial processes, in which suspended particles derived from weathering processes at the snowpatch margin are sorted by texture en route to the CT toe tread. Deposition points of transported material are evidenced by silt fans and turf - banked solifluction lobes. 7.1.2 Sorted patterned ground network analysis . The sorted patterned ground field on the uppermost te rrace of Frost Ridge is characterized by alternating coarse boulder stripes and fine - textured solifluction lobes that extend from the scarp - tread junction out toward the CT tread toe and over CT side slopes. Motivated by the well - documented hydrologic acti vity within coarse stripes (Nelson, 1975) , the coarse stripes were manually digitized to outline flow paths on the CT tread. Automated flow modeling confirmed the organization of the manually digitized flow map and provided input data for calculation of quantit ative morphometric drainage basin parameters. Results from the morphometric parameters provide convincing evidence for strong hydrologic influence over the organizational structure in the patterned - ground field and quantitative support for the characterist ics inferred from the process - focused investigations. The relief ratio, bifurcation ratio, drainage frequency, and stream - order values calculated indicate that basin morphometry is influenced by underlying impermeable conditions, which direct flow along do wnslope paths toward major outflow points at the toe and sides of the CT tread. 93 Unusual values calculated for some morphometric parameters are explained through consideration of the periglacial characteristics of the drainage basin. Specifically, the numbe r of streams is influenced both by underlying permafrost conditions, which are related to the depth of the permafrost table, and the consequent depth of particle sorting. The agreement of morphometric analyses with field observations highlights the advanta ges of applying automated flow modeling in a sorted patterned ground landscape. Based on the morphometric analyses presented here, the organization of the sorted patterned ground field is interpreted as a well - integrated system of flow paths that occupy a gently sloping tread underlain by permafrost. The network of patterned ground thereby connects the snowpatch margin hydrologically to outflow points located near the toe tread. Ancillary drainage paths remove sediment from CT treads along their margins. Re sults from the automated and manual flow - detection methods presented here both demonstrate that a pattern akin to a parallel drainage network exists over most of the sorted patterned - ground field. Areas on the periphery of the study area were not all inclu ded owing to decreased quality of input data, so the quantitative network analysis in this study largely covers the interior of the patterned ground field. Consideration of areas near the periphery of the present study area in the morphometric characteriza tion would clearly reveal feature organization reminiscent of a distributary pattern, which would agree with the slightly convex topographic profile of the CT tread (Nyland, 2019, 57 58) . The southwest part of the area with adequate coverage shows draina ge pathways leading to the lateral margin of the CT tread. 94 7.2 Synthesis 7.2.1 The role of fluvial action in cryoplanation terrace formation . Sorted patterned ground features have long been associated with cryoplanation landforms but have rarely been considered in the context of CT formation. On Frost Ridge, the climatic regime, periglacial processes, and fluvial influences combine to impart a distinct hydrologic signature on CT treads that exerts an important control over CT formation. Whe re snowpatches persist into the summer, the coarse and fine stripes on CT treads constitute dynamic surfaces of transportation over which sediment and water produced at snowbank sites are removed from terrace treads. Contrasting substrate material influenc es spatial divergences in the ground thermal regime, attributable to diurnal water pulses within coarse stripes, while long - term frost creep and slow flow of saturated soil occurs within fine - textured stripes. 7.2.2 The conveyor system model . Considering CT treads as efficient surfaces of transportation gives rise to a conceptualization of the system as a conveyor - like device for transporting weathered material between the snowbank and tread margins. Rapid transport of water and fine sediment occurs as diu rnal pulses of snow meltwater within coarse stripes, which form an interconnected network of features terminating near or extending over the CT tread toe and lateral margins. The intervening fine stripes function as solifluction lobes that move large volum es of sediment at much slower rates across CT treads and side slopes. The underlying frozen ground functions in both the fine and coarse stripes as an impermeable substrate that governs the depth of saturation and sorting , promotes laminar flow in the lobe s and turbulent flow in the coarse gutters, and anchors the coarse stripes. Together, these mechanisms work to prevent the construction of rampart - like features at the bases of 95 subjacent scarps. Material removed from the scarp - tread junction is transported over tread toes and lateral margins and moves further downslope. If transported over the distal edge it is recycled to the subjacent CT scarp - tread junction, downslope of which another conveyo r transport system operates. If transported over tread margins and onto side slopes, material is rapidly removed from the ridge entirely or incorporated in solifluction lobes (Brunnschweiler , 1965). Cryoplanation terrace side slopes in interior and western Alaska are festoo ned with solifluction lobes more comprehensively than on Frost Ridge, where Neoglacial moraines dominate. Under the conveyor - system model of sediment excavation on CT treads, localized weathering near snowpatches is accentuated as material derived from the snowpatch vicinity is subsequently transported away during snowmelt. Fluvial action within sorted patterned ground is a key component of scarp retreat and tread expansion. The coarse stripes function as a the tread surface, confines, and directs , acts as channels for rapid transport by flowing water. Other factors, particularly such microclimatic parameters as topographic shadowing and slope aspect, are important for preserving the conveyor system because they promote the maintenance of underlying permafrost and the late - lying snow cover. The erosive power of this system is demonstrated by the time - transgressive back - wasting of scarps through localiz ed weathering and efficient removal of sediment by processes enabled by snow meltwater (Nyland and Nelson 2020; Nyland et al. 2020a, b). The hydrologic significance of sorted patterned ground has been described in other studies that highlighted the ease w ith which water flows in boulder gutters (e.g., Caine, 1963; 96 Nelson, 1975; Wilkinson & Bunting, 1975) . The conveyor - system model underlines the importance of periglaci al hydrology in the context of geomorphic activity and aligns with the results of previous research on patterned - ground hydrology. The results of this study refute the notion that patterned ground is indicative of geomorphic dormancy and highlight the hydr ologic imprint of patterned ground on CT treads. The occurrence of patterned ground is directly related to CT form, i.e., flat treads, and helps to explain formation process attributable to CTs, i.e., prevention of weathered rampart accumulations. The res ults presented in this study provide substantial evidence for the hydrologic significance of periglacial assemblages existing as parts of CT treads, and demonstrate the efficacy of periglacial processes in landscape formation, a phenomenon that has been indicated in other periglacial regions, (e.g., Paquette et al., 2014) . The hydrologic significance of other fluvial features in permafrost environments, known as water tracks, has also been noted in recent years (Levy et al., 2011; Paquette et al., 2017b, 2018) . These sm aller periglacial water - related features can impart significant controls on periglacial hillslope hydrology. By considering cryoplanation landforms as integral, functional parts of the periglacial assemblage, an integrated perspective on the upland perigla (Brunnschweiler , 1965; Reger , 1975). Minor periglacial landforms, including sorted patterned ground, blockfields, and solifluction lobes, occupy distinct functional niches in the creation of cryoplanated terrain. actively contribute to sculpting it , imparting both a characteristic geomorphometric signature (Queen, 2018) and a distinctively periglacial appearance to the landscape. 97 system . The uppermost patterned ground field on Frost Ridge exemplifies the conveyor model and highlights the dynamic nature of upland periglacial assemblages (Figure 7.1). The micro - and meso - scale features are integral components of the nivation process suite, the various components of which are represented by photos around the periphery. The tread is a tessella ted surface composed of interconnected periglacial microforms functioning as an integrated system for moving weathering products from the snowbank vicinity, across the CT tread, and on to lower elevations. The conveyor - system model of sediment trans port on CT treads presents a new and compelling argument in favor of the nivation hypothesis of CT formation. As noted above, m any elements of the periglacial literature describe periglacial microforms as having been superimposed atop pre - existing topogra phy (e.g., Berthling & Etzelmüller , 2001). From a process perspective, however, it bears repeating that this term fails to convey what is perhaps the most important consideration for the long - term evolution of upland periglacial geomorphic landscapes: upla nd periglacial assemblages function as systems cryoplanation terrace on Frost Ridge: From a vantage point midway up Splinter Peak (Figure 7.1) the observer gains a vivid impression of the periglacial processes and landforms that constitute a terminology presented to describe systems of smaller periglacial forms that comprise landscape - scale features (Brunnschweiler, 1965 ) . The fundamental unit of 98 the a ltiplanorium, cryoplanation terraces, are maintained under the influence of minor periglacial features , which provide evidence for the dynamic system of weathered material transport: lobate solifluction tongues, alluvial fan - like silt deposit s that extend from coarse stripe termini, audible flowing of water through coarse stripes, the appearance of needle ice, and accumulations of angular fragments on the upslope sides of large immobile boulders. The periglacial features o n Frost Ridge form a continuous tessellat ion with a geometric pattern reminiscent of a distributary drainage system , an organization that operates to remove weathered material from the scarp - tread junction and transport it over the tread. W ater flows rapid ly through the interstitial spaces of coarse gutters, or slowly as pulses of moisture through fine - textured soil stripes. During periods of diurnal freeze - thaw action, the phase change of water results in the incremental downslope movement of soil via fros t creep and the growth and ablation of needle ice. The sorted stripe features on Frost Ridge and associated periglacial processes comprise a system of transport in which material is excavated from the vicinity of the snowpatch margin, exposing new material to the influence of nivation, leading to the retreat of CT scarps and tread extension. As noted in C hapter 3 of this thesis, the large patterned - ground field at the confluence of Splinter Peak and Frost Ridge had a far less dynamic appearance in the late 2010s than it did in the mid - 1970s. This may be attributable to the fact that the snowbank at this lo cation A critical experiment will form the next phase of research on Frost Ridge . 99 Figure 7. 1 System . (S een from mid - slope on Splinter Peak ) (a) Movement pegs displaced two years after emplacement in a T1 fine (solifluction) stripe by V. Jones. Pegs were installed in summer 1974 and show 1 - 2 cm downslop e movement. Photo by C. Cialek, August 1976. 100 . (b) Sorted stripes on T1. Sorting extends to the depth of 80 cm, below which it is inhibited by the presence of permafrost. Water could be heard running in these coarse . (c) Sorted circle near crest of Frost Ridge on flattest part of T1 tread surface. Despite the low gradient, water flowed through the coarse segment of this feature. JIRP Director Maynard M. Miller standing just beyond the sorted circle. See Figure 2.1. Photo by C. Cialek, August 1976. (d) silt fan formed by sediment exiting sorted stripe, near distal edge of T2. Photo by F. Nelson, July 2019. atop T1. Note stone - banked front and concentric parabolic furrows pointing downslope, indicative of movement. Photo by C. Cialek, August 1976. (f) Large turf - banked solifluction lobe immediately downsl ope of distal edge of T1 tread. Robert L. Nichols near center of photo. Photo by F. Nelson, August 1976. (g) Splaying solifluction lobes on T1 tread, formed where the fine stripes encounter a decrease in slope angle. Note that splaying lobes have buried th e bounding coarse stripes. Photo by F. Nelson, August 1975. (h) needle ice with superincumbent load formed atop fine stripe on T1 terrace. Photo by F. Nelson, September 1975. (i) Large boulder, possibly rooted in permafrost, in vicinity of T3. Note accumul ation of clastic material on upslope side of boulder, probably emplaced by frost creep. (j) Snowpatch at scarp - tread junction of T1. Photo by C. Cialek, August 1976 . 7.3 Conclusions This study is among the first to relate quantitative field - based and remotely sensed data to cryoplanation terrace formation. One of the primary contributions of this work is the proposition, supported by process - focused investigations, that periglacial patterned groun d plays a significant role in water/sediment transportation on CT treads, thereby providing an important dimension to the nivation hypothesis of cryoplanation terrace formation. From th is analysis of hillslope hydrolog y , the main conclusions of this study are derived: The presence of sorted patterned ground does not indicate geomorphic quiescence. Well - developed patterned ground networks facilitate geomorphic activity through hydrologic processes. 101 Sorted patterned ground constitutes a transportation netwo rk that promotes the toe - ward and lateral transmission of water and sediment, despite the seemingly flat microtopography of the CT tread. stabiliz ing the patterned ground f ield, and facilitating rapid transmission of water and sediment through boulder gutters. Ancillary sediment evacuation occurs as solifluction and sidewalls. Sorted patterne d ground, solifluction lobes, and other periglacial microforms are not - scale terrain. These features perform important, interconnected functional roles in sculpting the distinctively periglacial topography typified by cry oplanation terraces. The arrangement of sorted stripes on Frost Ridge resembles a fluvial network characterized by interconnected features that extend from the snowpatch to terrace margins. The apparent organization of the sorted - stripe networks underlines their propensity for efficient water transportation and points to a fluvial origin for sorted stripes. 7.4 Recommendations for Future Work 7.4.1 Long - term process investigations . Process investigations conducted for this study can be extended and enhanced through continued monitoring efforts and through installation of additional loggers on Terraces 2 and 4. The short - term temperature and soil moisture records established in this study would be complemented by continued data collection at 102 similar temporal resolutions, i.e., 5 - minute logging intervals, to capture both long - term patterns and short - term pulses in the fine and coarse stripe records. The resulting highly detailed records would contribute substantially to future exploration of the ground thermal regime and for establishing data for long - term monitoring studies. Temperature sensors should be installed at additional depths on both Terrace 2 and Terrace 4 to characterize the thermal processes operating in fine and coarse stripes. Based on the initial results derived from this study, continued monitoring of patterned ground on active CT terraces through repeat UAV survey and process measurements would complement efforts to define more precisely the role of patterned ground in periglacial hydrology, an association that has been indicated in previous research, but has yet to be widely embraced (Paquette, 2017). Mapping of the permafrost table via ground - penetrating radar (GPR) at key locations on Frost Ridge would confirm underlying permafrost conditions, complementing the morphometric characterization of the patterned ground network. 7.4. 2 Broader implications of Frost Ridge periglacial studies . T he work reported in this study has potential to assist in the resolution of some outstanding questions in Qua ternary studies. Continued monitoring efforts of the periglacial activity on Frost Ridge, such as those established in this study, are critical in developing an extensive body of quantitative data to achieve a definitive theory of cryoplanation terrace for mation. Both relict cryoplanation terraces and sorted patterned ground have been reported widely in the Appalachian Mountains south of the glacial border (e.g., Braun , 1989; Clark & Ciolkosz, 1988; Clark & Hedges, 1992; Clark & Schmidlin 1992), but the periglacial affinities of some of these features has been 103 challenged (e.g., Kerwa n, 2001). Application of the fluvial morphometric methods employed in Chapter 6 to fields of relict sorted stripes could help to resolve such differe nces. Future mapping and monitoring efforts focused on the permafrost of Frost Ridge would also contribute to efforts related to investigating the response of areas affected by permafrost to warming global temperatures, and contributing to a growing area o f research in periglacial geomorphology (Ballantyne, 2018, 2). The response of underlying permafrost conditions and snow accumulation to warming temperatures has the capacity to impact the formation and distribution of sorted stripes, which have been shown to influence the distribution of vegetation, nutrient transport, and hillslope - scale hydrologic connectivity (Paquette et al., 2017; 2018). Continued monitoring of sorted stripe morphology and underlying permafrost are therefore critical to assess changes in CT response to warming temperatures and will provide input for climate and geomorphic modeling that result in improved assessments of the effects of environmental change. 104 APPENDIX 105 Table A.1 Chapter 5 d ata l ogging e quipment i nformation . Tool Purpose/descriptio n Specifications URL Hobo® TidbiT® MX Temp 400 (MX2203) Measure temperatures in water bodies and soil environments Temperature r ange: - 20° to 70°C Accuracy: ± 0.25 °C from - 20° to 0°C; ±0.2°C from 0° to 70°C Resolution: 0.01°C Dimensions : 4.45 x 7.32 x 3.58 cm https://www.onsetco mp.com/products/da ta - loggers/mx2203/ Soil Moisture Smart Sensor (S - SMx - M005 ) Measure soil water content Measurement range: 0 to 0.550 m 3 /m 3 ; Temperature range 0° to 50°C Accuracy: ±0.031 m 3 /m 3 Resolution: 0.0007 m 3 /m 3 Probe dimensions: 89 x 15 x 1.5 mm https://www.onsetco mp.com/products/se nsors/s - smd - m005/ HOBO® Micro Station (H21 - USB) Store data recorded from temperature and soil moisture sensors Operating Range: - 20° to 50°C Time accuracy: ±5 seconds per week at 25°C https://www.onsetco mp.com/products/da ta - loggers/h21 - usb/ 12 - Bit Temperature Smart Sensor 9s - TMB - M0xx) Temperature monitoring Temperature r ange: - 40° to 100°C Accuracy: to 50°C 0° to 50°C Probe Dimensions: 5.1 x 33 mm https://www.onsetco mp.com/products/se nsors/s - tmb - m0xx/ 106 Table A.2 Chapter 6 GIS t ools i nformation . Tool Source Information/operation URL Stream O rder ArcMap Stream ordering - assign a numeric order to links in a stream network using the Strahler (1957) method which assigns an order of 1 to all links without any tributaries https://desktop.arcgis.com/en/a rcmap/10.3/tools/spatial - analyst - toolbox/how - stream - order - works.htm#ESRI_SECTION1_332 E8909620C461B9B991A7FC1A5 E843 Polyline ArcMap Digitize s coarse patterned grou nd constriction template tool implemented within the Create Features window https://desktop.arcgis.com/en/a rcmap/10.3/manage - data/editing/what - is - editing - .htm#ESRI_SECTION1_8E30256B C6 C5408C832128CF45C36C5B Hillshade ArcMap Reveal s minor periglacial features - a grayscale representation of the surface position into account to shade the image. Default azimuth; 315°, and sun height; 45°, were used. https://desktop.arcgis.com/en/a rcmap/10.3/manage - data/raster - and - images/hillshade - function.htm r.fill.dir GRASS Fill DEM generates a depressionless elevation map, and flow direction map. All depression s within the input elevation raster are filled, subsequent filling occurs if the flow direction algorithm identifies additional pits. https://grass.osgeo.org/grass78 /manuals/r.fill.di r.html r.watershed GRASS Flow direction (using D8method) and flow accumulation https://grass.osgeo.org/grass78 /manuals/r.watershed.html#des cription r.mapca lc GRASS Define s flow accumulation threshold - raster map calculator that performs user - defined arithmetic expressions https://grass.osgeo.org/grass78 /manuals/r.mapcalc.html 107 REFERENCES 108 REFERENCES Arnold, J. G., Srinivasan, R., Muttiah, R. S., & Williams, J. R. (1998). Large area hydrologic modeling and assessment part I: Model development. JAWRA Journal of the American Water Resources Association , 34 (1), 73 89. Andersson, J. G. (1906). Solifluction , a component of subaërial denudation. The Journal of Geology , 14 (2), 91 112. Ballantyne, C. K. (2018). Periglacial Geomorphology (1st ed.). Hoboken, NJ: John Wiley & Sons. Beckinsale, R.D. & Chorley, R.J. (1991). The History of the Study of Landforms or t he Development of Geomorphology, Volume 3: Historical and Regional Geomorphology 1890 - 1950. London and New York: Routledge, 496 pp. Benedict, J. B. (1970). Downslope soil movement in a Colorado alpine region: Rates, processes, and climatic significance. Ar ctic and Alpine Research , 2 (3), 165 226. Berthling, I., & Etzelmüller, B. (2011). The concept of cryo - conditioning in landscape evolution. Quaternary Research , 75 (2), 378 384. Bernatek, A. (2015). Visualizing morphometric changes in a piping system using D EM and GIS analysis: the Bieszczady Mts., Poland. In: Hengl, T. (eds.). Geomorphometry for Geosciences Poznan, Poland: International Society for Geomorphometry, 157 - 160. Beven, K. J., & Kirkby, M. J. (1979). A Physically Based, Variable Contributing Area Model of Basin Hydrology. Hydrological Sciences Journal , 24 (1), 43 69. Taylor & Francis. Boyd, D. S. (2 009). Remote sensing in physical geography: A twenty - first - century perspective. Progress in Physical Geography , 33 (4), 451 456. Bradshaw, M. (1982). Process, time and the physical landscape: Geomorphology today. Geography , 67 (1), 15 28. Braun, D.D. (1989). Glacial and periglacial erosion of the Appalachians. Geomorphology 2(1): 233 - 256. Brunn, S. D., Cutter, S. L., & Harrington Jr, J. (2004). Geography and Technology (1st ed.). New York: Springer Science & Business Media. Bryan, K. (1946). Cryopedology, the study of frozen ground and intensive frost - action, with suggestions on nomenclature. American Journal of Science , 244 (9), 622 642. 109 Büdel, J. (1948) Das System der klimatischen Geomorphology. Verhandl. Deutscher Geographie 27: 65 - 100. English translation : The climatic geomorphic system. In: Derbyshire, E. (ed.), Climatic Geomorphology . London, The MacMillan Press, 104 - 130. Büdel, J. (1948b). Die klima - morphologischen Zonen der Polarländer. Erdkunde 2(1/3): 22 - 53 Büdel, J. (1977). Klima - Geomorphologie . Ber lin: Gebrüder Borntraeger, 304 pp. English translation: Climatic Geomorphology. Princeton, NJ: Princeton University Press, 443 pp. Brunnschweiler, D. (1965). The morphology of altiplanation in interior Alaska. Unpublished report to US Army Cold Regions Res earch and Engineering Laboratory, Hanover NH. 48 pp. Caine, T. N. (1963). The Origin of Sorted Stripes in the Lake District. Northern England. Geografiska Annaler , 45 (2/3), 172 179. Cairnes, D. D. (1912). Differential erosion and equiplanation in portions of Yukon and Alaska. Bulletin of the Geological Society of America , 23 (1), 333 348. Campbell, J. B., & Wynne, R. H. (2011). Introduction to Remote Sensing, Fifth Edition (5th ed.). New York: Guilford Press. Christiansen, H. H. (1996). Effects of nivation on periglacial landscape evolution in western Jutland, Denmark. Permafrost and periglacial processes , 7 (2), 111 138. Clark, G.M. and Schmidlin, T.W. 1992. Alpine periglacial landforms of eastern North America: A review. Permafrost and Periglacial Process es , 3: 225 - 230. Clark, G.M. and Ciolkosz, E.J. (1988). Periglacial geomorphology of the Appalachian Highlands and interior highlands south of the glacial border - a review. Geomorphology , 1: 191 - 220. Clark, G.M. and Hedges, J. (1992). Origin of certain hi gh - elevation local broad uplands in the central Appalachians south of the glacial border, U.S.A. - A peleoperiglacial hypothesis. In: Dixon, J.C. and Abrahams, A.D. (eds.) , Periglacial Geomorphology . John Wiley & Sons, Chichester, New York, Brisbane, Toront o, Singapore, 31 - 61. Colpron, M., & Nelson, J. (2011). A Digital Atlas of Terranes for the Northern Cordillera, Yukon Geological Survey. Retrieved May 23, 2020, from http://yukon.maps.arcgis.com/apps/webappviewer/index.html?id=b7fa3a4a0ad94ed19 412bbad5ec6ab72 Colwell, R. N. (1966). Uses and limitations of multispectral remote sensing. Proceedings of the 4th International Symposium on Remote Sensing . University of Michigan, Ann Arbor, MI. 110 Czeppe, Z. (1961) Participation of flowing water in creation of the periglacial microrelief. In : International Association for Quaternary Research (INQUA), 6th Congress, Warsaw, Poland, August - September 1961: 8g. Davis, W.M (1909). Geographical Essays , Bos ton: Ginn and Company. Demek, J. (1968). Cryoplanation terraces in Yakutia. Biuletyn Peryglacjalny 17: 91 - 16. Demek, J. (1969). Cryoplanation terraces, their geographical distribution, genesis and development. Rozpravy Ceskoslovenski Akademi ved, Rada, Mat ematickych a Prirodnich , 79 (4), 80. Dixon, J.C., Thorn, C.E., and Darmody, R.G. (1984). Chemical weathering processes on the Vantage Peak nunatak, Juneau Icefield, southern Alaska. Physical Geography 5(2): 111 - 131. Eakin, H. M. (1916). The Yukon - Koyukuk r egion, Alaska. U.S. Geological Survey Bulletin 631, 88 pp. Eichel, J., Draebing, D., Katenborn, T., Senn, J.A., Klingleil, L., Wiel, M., and Heinz, E. (2020). Unmanned aerial vehicle - based mapping of turf - banked solifluction lobe movement and its relation to material, geomorphometric, thermal and vegetation properties. Permafrost and Periglacial Processes 31:97 - 109. DOI: 10.1002/ppp.2036 Embleton, C., & King, C. A. M. (1975). Periglacial Geomorphology . New York: John Wiley & Sons. Environmental Systems Res earch Institute, Inc. (2020). ArcGIS . Redlands, CA: Environmental Systems Research Institute Inc. Evans, D.J.A., Kalyan, R., & Orton, C. (2017). Periglacial geomorphology of summit tors on Bodmin Moor, Cornwall, SW England. Journal of Maps . 13(2): 342 - 349. DOI: 10. 1080/17445647.2017.1308283 Fernandez, E., Garfinkel, R., & Arbiol, R. (1998). Mosaicking of aerial photographic maps via seams defined by bottleneck shortest paths. Operations Research , 46 (3), 293 304. Field, W. O., & Miller, M. M. (1950). The J uneau Ice Field Research Project. Geographical Review , 40 (2), 179. French, H. M. (2016). Do periglacial landscapes exist? A discussion of the upland landscapes of northern interior Yukon, Canada. Permafrost and Periglacial Processes , 27 (2), 219 228. French , H. M. (2017). The Periglacial Environment (4th ed.). Hoboken, NJ: John Wiley & Sons. Gerloch Trough Runoff Sediment Sampler. (2018). Rickly Hydrological Co., Inc. Retrieved April 10, 2020, from https://rickly.com/gerloch - trough - runoff - sediment - sampler/ 111 Gleyzer, A., Denisyuk, M., Rimmer, A., & Salingar, Y. (2004). A Fast Recursive Gis Algorithm for Computing Strahler Stream Order in Braided and Nonbraided Networks1. Journal of the A merican Water Resources Association , 40 (4), 937 946. Godin, E., Osinski, G.R., Harrison, T.N., Pontefract, A., and Zanetti, M. (2019). Geomorphology of gullies at Thomas Lee Inlet, Devon Island, Canadian high Arctic. Permafrost and Periglacial Processes 30: 19 - 34. DOI: 10.1002/ppp. 1992. Goldthwait, R. P. (1 976). Frost sorted patterned ground: a review. Quaternary Research , 6 (1), 27 35. Goudie, A. (1990). Geomorphological Techniques (2nd ed.). London, UK: Unwin Hyman Ltd. GRASS Development Team. (2019). GRASS GIS . Open Source Geospatial Foundation. Retrieved from https://grass.osgeo.org Hamelin, L. - E. (1964). Le périglaciare du Massif Juneau en Alaska. Biuletyn Peryglacjalny , 13: 5 - 14. Hajam, R. A., Hamid, A., & Bhat, S. (2013). Application of morphometric analysis for geo - hydrological studies using geo - spatia l technology a case study of Vishav Drainage Basin. Hydrology Current Research , 4 (3), 1 12. Hallet, B. (1990). Spatial self - organization in geomorphology: from periodic bedforms and patterned ground to scale - invariant topography. Earth - Science Reviews , 29 ( 1 4), 57 75. Heusser, C. (2007). Juneau Icefield Research Project (1949 - 1958) . Oxford, UK: Elsevier. Hobbs, W.H. (1910). Soil stripes in cold humid regions, and a kindred phenomenon. 12 th Report, Michigan Academy of Science , 51 - 53. Hodgson, R., & Young, K. L. (2001). Preferential groundwater flow through a sorted net landscape, Arctic Canada. Earth Surface Processes and Landforms , 26 (3), 319 328. Horton, R. E. (1945). Erosional development of streams and their drainage basins; hydrophysical approach to quan titative morphology. Geological society of America bulletin , 56 (3), 275 370. Hövermann, J. and Oberbeck, G. (eds., 1972). Hans - Poser - Festschrift. Göttinger Geographicsche Abhandlungen , 60, 571 pp. + maps, diagrams. Humlum, O. (2008). Alpine and polar periglacial processes: The current state of knowledge (p. 8). Presented at the Ninth International Conference on Permafrost, University of Alaska Fairbanks. 112 Jayawardena, A. W. (2014). Environmental and Hydrological Systems Modelling . Boca Raton, FL: CRC Pr ess. Jensen, J. R. (1995). Issues involving the creation of digital elevation models and terrain Geocarto International , 10 (1), 5 21. Taylor & Francis. Johnson, R. F. (1983). Ice movement and structura l characteristics of the Cathedral Glacier system, Atlin Provincial Park, British Columbia (M.S. Thesis). Department of Earth Sciences, Montana State University, Bozeman, Montana. Johnston, G.H., Ladanyi, B., Morgenstern, N.R., and Penner, E. (1981). Engin eering characteristics of frozen and thawing soils. In: Johnston, G.H. (ed.), Permafrost: Engineering Design and Construction . Toronto: John Wiley & Sons, 73 - 147. Jones, V. K. (1975). Contributions to the Geomorphology and Neoglacial Chronology of the Cath edral Glacier System, Atlin Wilderness Park, British Columbia. (M.S. Thesis), Department of Geology, Michigan State University, East Lansing, MI. Kamp, U., Bolch, T., & Olsenholler, J. (2005). Geomorphometry of Cerro Sillajhuay (Andes, Chile/Bolivia): comp arison of digital elevation models (DEMs) from ASTER remote sensing data and contour maps. Geocarto International , 20 (1), 23 33. Taylor & Francis. Karte, J. (1979). Räumliche Abgrenzung und regionale Differenzierung des Periglaziärs. Bochumer Geographische Arbieten , 35, 211 pp. Karte J. (1982). Development and present state of German periglacial research in Arctic and alpine environments. Biuletyn Pe ryglacjalny , 29, 183 - 201. Kerwan, M.L. (2001). Origin and evolution of High Elevation Southern Appalachian Plateaus . (M.S. Thesis), Department of Geology, The College of William and Mary. Kessler, M. A. (2003). Self - organization of sorted patterned gr ound. Science , 299 (5605), 380 383. Kessler, M.A., Murray, A., Werner, B., & Hallet, B. (2001). A model for sorted circles as Journal of Geophysical Research: Solid Earth , 106 (B7), 13287 13306. , Z., Treml, V., & Traczyk, A. (2019). Patterned ground above the alpine timberline in the high Sudetes, central Europe. Journal of Maps 15(2): 563 - 569. DOI: 1080/17445647.2019.1636890 Lauriol, B., Lamirande, I., & Lalonde, A. E. (2006). The giant steps of Bug Creek, Richardson Mountains, NWT, Canada. Permafrost and Periglacial Processes , 17 (3), 267 275. Wiley Online Library. 113 Levy, J. S., Fountain, A. G., Gooseff, M. N., Welch, K. A., & Lyons, W. B. (2011). Water tracks and permafrost in Taylor Valley, Anta rctica: Extensive and shallow groundwater connectivity in a cold desert ecosystem. GSA Bulletin , 123 (11 12), 2295 2311. von Lozinski, W. (1909). Über die Mechanische Verwitterung der Sandsteine im Gemässigten Klima . Acad Sci, Cracovie Bull. Internat. cl. s ci. math et naturelles. 1, 1 - 25 (English Translation: On the Mechanical Weathering of Sandstones in Temperate Climates. MacIntyre, D., Villeneuve, M., & Schiarizza, P. (2001). Timing and tectonic setting of Stikine Terrane magmatism, Babine - Takla lakes area, central British Columbia. Canadian Journal of Earth Sciences , 38 (4), 579 601. NRC Research Press. Martins, A. K., & Gadiga, B. L. (2015). Hydrological and morphometric analysis of upper Yedzaram catchment of Mubi in Adamawa state, Nigeria usingg eogr aphic information system (GIS). World Environment , 5 (2), 63 69. Martonne, E. de (1913). Le climat - facteur du relief. Scientia 1913: 339 - 355. English translation: Climate: Factor of relief. In: Derbyshire, E. (ed.), Climatic Geomorphology . London, The MacMi llan Press, 61 - 75. Martz, L. W., & Garbrecht, J. (1992). Numerical definition of drainage network and subcatchment areas from digital elevation models. Computers & Geosciences , 18 (6), 747 761. Matthes, F. E. (1900). Glacial sculpture of the Bighorn Mountai ns, Wyoming. US Geological Survey Twenty - first Annual Report Part II , 167 190. Matthews, J., Wilson, P., Winkler, S., Mourne, R., Hill, J., Owen, G., Hiemstra, J., et al. (2019). Age and development of active cryoplanation terraces in the alpine permafrost zone at Svartkampan, Jotunheimen, southern Norway. Quaternary Research , 92 , 641 664. McCann, S., Howarth, P., & Cogley, J. (1972). Fluvial processes in a periglacial environment: Queen Elizabeth Islands, NWT, Canada. Transactions of the Institute of Briti sh Geographers , 69 82. JSTOR. McNamara, J., Kane, D., & Hinzman, L. (1998). An analysis of streamflow hydrology in the Kuparuk River basin, Arctic Alaska: A nested watershed approach. Journal of Hydrology , 206 (1 2), 39 57. McNamara, J., Kane, D., & Hinzman, L. (1999). An analysis of an arctic channel network using a digital elevation model. GEOMORPHOLOGY , 29 (3 4), 339 353. Melton, M. A. (1957). An Analysis of the Relations among Elements of Climate, Surface Properties, and G eomorphology (Technical Report No. 11) (pp. 100). Office of Naval Research, Department of Geology, Columbia University, New York, NY. 114 Mertie, J. B. (1937). The Yukon - Tanana Region, Alaska (p. 276). U.S. Geological Survey Bulletin 872. Miller, M. M. (1975). Mountain and Glacier Terrain Study and Related Investigations in the Juneau Icefield Region, Alaska - Canada (No. AD - A019 703) (p. 265). Seattle, Washington: Foundation for Glacier and En vironmental Research. Mithan, H.T., Hales, T.C., & Cleall, P.J. (2019). Supervised classification of landforms in Arctic mountains. Permafrost and Periglacial Processes 30: 131 - 145. DOI: 10.1002/ppp2015 Monger, J. (1977). Upper Paleozoic rocks of the Atlin Terrane, northwestern British Columbia and South - Central Yukon. (No. 77 1A) (pp. 255 262). Geological Survey of Canada. Monger, J., Souther, J., & Gabrielse, H. (1972). Evolution of the Canadian Cordillera; a plate - tectonic model. American Journal of Scie nce , 272 (7), 577 602. American Journal of Science. Mortensen, H. (1930). Einige Oberflächenformen in Chile und auf Spitsbergen im Rahmen einer vergleichenden Morphologie der Klimazonen. Petermanns Geographische MittelungenErgänzungscheft 209: 147 - 156. Mulv aney, T. J. (1851). On the use of self - registering rain and flood gauges in making observations of the relations of rainfall and flood discharges in a given catchment. Proceedings of the institution of Civil Engineers of Ireland , 4 , 19 31. Nelson, F. E. (1 975). Periglacial Features on the Cathedral Massif, northwestern British Columbia: A Preliminary Investigation (Open File No. 75). Seattle, Washington: Juneau Icefield Research Program and The Foundation for Glacier and Environmental Research, 36 pp. Nelso n, F. E. (1979). Patterned Ground in the Juneau Icefield Region, Alaska - British Columbia (M.S. Thesis). Department of Geography, Environment, and Spatial Sciences, Michigan State University, East Lansing, MI. Nelson, F. (1982a). Sorted - circle macrofabrics. Polarforschung 52(1/2): 43 - 53. Nelson, F. (1982b). Sorted - stripe macrofabrics. Geografiska Annaler 64A(1 - 2): 25 - 33. Nelson, F. E. (1989). Cryoplanation terraces: Periglacial cirque analogs. Geografiska Annaler: Series A, Physical Geography , 71 (1 2), 31 41. Nelson, F. E. (1998). Cryoplanation terrace orientation in Alaska. Geografiska Annaler: Series A, Physical Geograph y , 80 (2), 135 151. Nelson, F. E., & Nyland, K. E. (2017). Periglacial cirque analogs: Elevation trends of cryoplanation terraces in eastern Beringia. Geomorphology , 293 , 305 317. 115 Nelson, F., Outcalt, S., Goodwin, C., & Hinkel, K. (1985). Diurnal thermal re gime in a peat - covered palsa, Toolik Lake, Alaska. Arctic , 310 315. JSTOR. Earth Surface Processes and Landforms , 45 (3), 526 5 34. Wiley Online Library. Nyland K.E., & Nelson, F. E. (2020). Long - term nivation rates, Cathedral Massif, northwestern British Columbia. Canadian Journal of Earth Sciences . DOI: 10.1139/cjes - 2019 - 0176 Nyland, K. E. (2019). Stairways to Heaven: Origins and Development of Cryoplanation Terraces (Ph.D Thesis). Department of Geography, Environment, and Spatial Sciences, Michigan State University, East Lansing, MI. elevat ion data. Computer vision, graphics, and image processing , 28 (3), 323 344. Onset HOBO and InTemp Data Loggers . 470 MacArthur Blvd. Bourne, MA 02532: Onset Computer Corporation. Paquette, M., Fortier, D., & Vincent, W. F. (2014). The geomorphic and hydrolog ic relation between water tracks development and patterned ground on a High Arctic slope. Presented at the Fourth European Conference on Permafrost, Evora, Portugal. Retrieved from https://www.researchgate.net/publication/271645193_The_geomorphic_and_hydro log ic_relation_between_water_track_development_and_patterned_ground_on_a_High_A rctic_slope/comments Paquette, M., Fortier, D., & Vincent, W. F. (2017). Water tracks in the High Arctic: A hydrological network dominated by rapid subsurface flow through patte rned ground. Arctic Science , 3 (2, SI), 334 353. Paquette, M., Fortier, D., & Vincent, W. F. (2018). Hillslope water tracks in the High Arctic: Seasonal flow dynamics with changing water sources in preferential flow paths. Hyrological Processes , 32 (8), 1077 1089. Pareta, K., & Pareta, U. (2011). Quantitative morphometric analysis of a watershed of Yamuna basin, India using ASTER (DEM) data and GIS. International Journal of Geomatics and Geosciences , 2 (1), 248 269. Penck, A. (1905). Climatic features in the l and surface. American Journal of Science (4 th Series), 109: 165 - 174. Porter, C., Morin, P., Howat, I., Noh, M. - J., Bates, B., Peterman, K., Keesey, S., et al. (2018). ArcticDEM. Harvard Dataverse. Retrieved May 19, 2020, from https://doi.org/10.7910/DVN/OHHUKH 116 Poser, H. (1931). Beiträge zur Kenntnis der Arktischen Bodenformen. Geologische Rundschau 22: 200 - 231. Poser H. (ed., 1977). Formen, Formengesellschaften, und Untergrenzen in den heutigen Periglazialen Höhenstufen der Hochgebirge Europas und Africas zwischen Arktis und Äquator: Bericht über eine Symposium. Abhandlungen der Akademie der Wissenschaften in G ö t tingen, Mathematische - physikalische Klasse , Dritte Folge, 31: 354 pp. Prindle, L. M. (1905). The gold placers of the Fortymile, Birch Creek, and Fairbanks regions, Alaska . U.S. Geological Survey Bulletin, 251, 89 pp. Prindle, L.M. (1913). A geologic reconn aissance of the Circle Quadrangle, Alaska. U.S. Geological Survey Bulletin 538 , 82 pp. Queen, C. W. (2018). Large - Scale Mapping and Geomorphometry of Upland Periglacial Landscapes in Eastern Beringia (M.S. Thesis). Department of Geography, Environment, and Spatial Sciences, Michigan State University, East Lansing, MI. Quinn, P., Beven, K., Chevallier, P., & Planchon, O. (1991). The prediction of hillslope flow paths for distributed hydrological modelling using digital terrain models. Hydrological processes , 5 (1), 59 79. Quinton, W. L., & Marsh, P. (1999). Image analysis and water tracing methods for examining runoff pathways, soil properties and residence times in the continuous permafrost zone. IUGG 99 Symposium HS4 (pp. 257 264). Presented at the Integrate d Methods in Catchment Hydrology -- Tracer, Remote Sensing and New Hydrometric Techniques, Birmingham: IAHS PRESS - INTERN ASSOC HYDROLOGICAL SCIENCE. Rai, P. K., Mohan, K., Mishra, S., Ahmad, A., & Mishra, V. N. (2017). A GIS - based approach in drainage morpho metric analysis of Kanhar River Basin, India. Applied Water Science , 7 (1), 217 232. Springer. Rajeev, P. & Kodikara, J. (2016). Estimating apparent thermal diffusivity of soil using field temperature time series. Geomechanics and Geoengineering 11(1): 28 - 4 6. DOI: 10.1080/17486025.1006266 Reddy, G. P. O., Maji, A. K., & Gajbhiye, K. S. (2004). Drainage morphometry and its influence on landform characteristics in a basaltic terrain, Central India a remote sensing and GIS approach. International Journal of App lied Earth Observation and Geoinformation , 6 (1), 1 16. Elsevier. Reger, R. D. (1975). Cryoplanation Terraces of Interior and Western Alaska (Ph.D. Thesis). Arizona State University, Tempe, Arizona. 117 Reger, R. D., & Péwé, T. L. (1976). Cryoplanation terrace s: Indicators of a permafrost environment. Quaternary Research , 6 (1), 99 109. Rhoads, B. L. (2004). Whither physical geography? Annals of the Association of American Geographers , 94 (4), 748 755. Taylor & Francis. Roels, J. M., & Jonker, P. J. (1983). Proba bility sampling techniques for estimating soil erosion. Soil Science Society of America Journal , 47 (6), 1224 1228. Sack, D. (1992). New wine in old bottles: The historiography of a paradigm change. Geomorphology , 5 (3 5), 251 263. Salomon, W. (1929). Arktisch Bodenformen in den Alpen. Sitzber. Heidelberg Akademie Wissenschaften, Math. - Naturw Klasse , 5: 1 - 31. Schumm, S. A. (1956). Evolution of drainage systems and slopes in badlands at Perth Amboy, New Jersey. Geological society of A merica bulletin , 67 (5), 597 646. Geological Society of America. Short, N. M., & Blair, R. W. (1986). Geomorphology from space: A global overview of regional landforms . Washington, DC: Scientific and Technical Information Branch, National Aeronautics and Sp ace Administration. Singh, V. P. (2018). Hydrologic modeling: Progress and future directions. Geoscience Letters , 5 (1: 1 - 18). Slupetzky, H., & Krisai, R. (2009a). Indications of late glacial to Holocene fluctuations of Cathedral Massif Glacier, Coast Range (Northern British Columbia, Canada). Zeitschrift für Gletscherkunde und Glazialgeologie , 43 (44), 187 212. Smart, J. (1978). The analysis of drainage network composition. Earth Surface Processes , 3 (2), 129 170. Wiley Online Library. Smith, H.T.U. (1968). " Piping" in relation to periglacial boulder concentrations. Biuletyn Peryglacjalny 17: 204. Smith, M., & Pain, C. (2009). Applications of remote sensing in geomorphology. Progress in Physical Geography , 33 (4), 568 582. Sage Publications Sage UK: London, Eng land. Statistics Canada. (2018, May). Stikine Region. British Columbia Statistics. Retrieved from https://www2.gov.bc.ca/assets/gov/british - columbians - our - governments/local - governments/governance - powers/stikine_region_map.pdf Strahler, A. N. (1952). Dynami c basis of geomorphology. Geological Society of America Bulletin , 63 (9), 923 938. 118 Tallman, A. M. (1975). The Glacial and Periglacial Geomorphology of the Fourth Of July Creek Valley, Atlin Region, Cassiar District, northwestern British Columbia (Ph.D. The sis). Department of Geology, Michigan State University, East Lansing, MI. Tarboton, D. (1989). The Analysis of River Basins and Channel Networks using Digital Terrain Data (Ph.D Thesis). Massachusetts Institute of Technology, Cambridge, MA. Taylor, B. (195 5). Terrace formation on Macquarie Island. The Journal of Ecology , 133 137. JSTOR. Thorn, C. E. (1976). Quantitative evaluation of nivation in the Colorado Front Range. Geological Society of America Bulletin , 87 (8), 1169 1178. Thorn, C. E., & Hall, K. (1980). Nivation: An Arctic - alpine comparison and reappraisal. Journal of Glaciology , 25 (91), 109 124. Thorn, C. E., & Hall, K. (2002). Nivation and cryoplanation: The case for scrutiny and integration. Progress in Physical Geograp hy , 26 (4), 533 550. Thornbury, W.D. (1969). Principles of Geomorphology . New York: John Wiley & Sons, Inc., 594 pp. Tricart, J. & Cailleux, A. (1972). Introduction à la Géomorphologie Climatique . Paris: Société ranslation: Introduction to Climatic Geomorphology Troll C. (1944). Strukturböden, Solifluction und Frostklimate der Erde. Geologische Rundschau 34: 545 - 694. English translation: Structure Soils, Solifluction, and Fr ost Climates of the Earth. U.S. Army Snow Ice and Permafrost Research Establishment, Wilmette, IL., 1958, 121 pp. Ule, W. (1911). Glazialer Karree -- eder Polygonenboden. Zeitschrift Ges. Erkunde , 253 - 262. limate on morphology and development of sorted circles and polygons. Permafrost and Periglacial Processes 28: 663 - 674. DOI: 10.1002/ppp.1949 Vittala, S., Govindaiah, S., & Honne Gowda, H. (2004). Morphometric analysis of sub - watersheds in the pavagada area of Tumkur district, South India using remote sensing and GIS techniques. Journal of the Indian Society of Remote Sensing , 32 (4), 351 362. Washburn, A. L. (1956). Classification of p atterned g round and r eview of s uggested o rigins. Geological Society of America Bulletin , 67 (7), 823. Washburn, A. L. (1969). Weathering, frost action, and patterned ground in the Mesters Vig district, Northeast Greenland . CA Reitzels. 119 Wilkinson, T. J., & Bunting, B. T. (1975). Overland transport of sediment by rill water in a periglacial environment in the Canadian high Arctic. Geografiska Annaler. Series A, Physical Geography , 57 (1/2), 105 116. Williams, P. J., & Smith, M. W. (1989). The Frozen Earth: Fundamentals of Geocryology . Cambridge, UK: Cambridge University Press. Wol ock, D. M., & McCabe, G. J. (1995). Comparison of single and multiple flow direction algorithms for computing topographic parameters in TOPMODEL. Water Resources Research , 31 (5), 1315 1324. Woo, M. - k. (1986). Permafrost hydrology in North America. Atmosphe re - Ocean , 24 (3), 201 234. Woo, M. - k. (2012). Permafrost hydrology . Berlin Heidelberg: Springer - Verlag. Woo, M. - k., Kane, D. L., Carey, S. K., & Yang, D. (2008). Progress in permafrost hydrology in the new millennium. Permafrost and Periglacial Processes , 19 (2), 237 254. Woo, M. - k., & Sauriol, J. (1980). Channel development in snow - filled valleys, Resolute, NWT, Canada. Geografiska Annaler: Series A, Physical Geography , 62 (1 2), 37 56. Taylor & Francis. Woo, M. - k. (1976). Hydrology of a small Canadian High Arctic basin during the snowmelt period. Catena , 3 (2), 155 168. Elsevier. Woo, M. - k. (2012). Permafrost Hydrology . Berlin Heidelberg: Springer - Verlag. Zernitz, E. R. (1932). Drainage patterns and their significance. Journal of Geology , 40 (6), 498 521. J STOR. Zhang, W., & Montgomery, D. R. (1994). Digital elevation model grid size, landscape representation, and hydrologic simulations. Water Resources Research , 30 (4), 1019 1028. Wiley Online Library.