Tectonic Geomorphology of Mountains: A New Approach to Paleoseismology William B. Bull äjk Blackwell \§ß Publishing Tectonic Geomorphology of Mountains Tectonic Geomorphology of Mountains: A New Approach to Paleoseismology William B. Bull äjk Blackwell \§ß Publishing © 2007 William B. Bull BLACKWELL PUBLISHING 350 Main Street, Maiden, MA 02148-5020, USA 9600 Garsington Road, Oxford OX4 2DQ, UK 550 Swanston Street, Carlton, Victoria 3053, Australia The right of William B. Bull to be identified as the Author of this Work has been asserted in accordance with the UK Copyright, Designs, and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs, and Patents Act 1988, without the prior permission of the publisher. First published 2007 by Blackwell Publishing Ltd 1 2007 Library of Congress Cataloging-in-Publication Data Bull, William B., 1930- Tectonic geomorphology of mountains : a new approach to paleoseismology / William B. Bull, p. cm. Includes bibliographical references and index. ISBN-13: 978-1-4051-5479-6 (hardback: alk. paper) ISBN-10: 1-4051-5479-9 (hardback : alk. paper) 1. Morphotectonics. 2. Paleoseismology. I. Title. QE511.44.B85 2007 551.43'2-dc22 2006100890 A catalogue record for this title is available from the British Library. Set in 10.74/llpt AGaramond by SPi Publisher Services, Pondicherry, India Printed and bound in Singapore by C.O.S Printers Pte Ltd The publisher's policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. For further information on Blackwell Publishing, visit our website: www.blackwellpublishing.com Contents Preface................................... viii 1 Scrunch and Stretch Bedrock Uplift 1.1 Introduction.............................. 3 1.2 Pure Uplift, Stretch and Scrunch Bedrock Uplift................. 6 1.2.1 Isostatic and Tectonic Uplift..................... 6 1.2.2 Stretch and Scrunch Tectonics..................... 12 1.3 Landscape Responses to Regional Uplift.................... 23 2 Concepts for Studies of Rising Mountains 2.1 Themes and Topics........................... 27 2.2 The Fundamental Control of Base Level.................... 28 2.2.1 Base Level............................ 28 2.2.2 Base-Level Change......................... 28 2.2.3 The Base Level of Erosion...................... 31 2.2.4 The Changing Level of the Sea.................... 33 2.2.5 Spatial Decay of the Effects of Local Base-Level Changes........... 37 2.3 Threshold of Critical Power in Streams.................... 39 2.3.1 Relative Strengths of Stream Power and Resisting Power............ 41 2.3.2 Threshold-Intersection Points..................... 42 2.4 Equilibrium in Streams.......................... 42 2.4.1 Classification of Stream Terraces.................... 42 2.4.2 Feedback Mechanisms........................ 45 2.4.3 Dynamic and Static Equilibrium.................... 46 2.5 Time Lags of Response.......................... 49 2.5.1 Responses to Pulses of Uplift..................... 50 2.5.2 Perturbations that Limit Continuity of Fluvial Systems............ 51 2.5.3 Lithologic and Climatic Controls of Relaxation Times............ 54 2.5.4 Time Spans Needed to Erode Landforms................. 57 2.6 Tectonically-Induced Downcutting...................... 58 2.6.1 Straths, Stream-Gradient Indices, and Strath Terraces............. 58 2.6.2 Modulation of Stream-Terrace Formation by Pleistocene—Holocene Climatic Changes . 65 2.7 Nontectonic Base-Level Fall and Strath Terrace Formation............. 66 2.8 Hydraulic Coordinates.......................... 69 3 Mountain Fronts 3.1 Introduction.............................. 75 3.2 Tectonically Active Escarpments....................... 79 3.2.1 Faceted Spur Ridges........................ 79 3.2.2 Mountain—Piedmont Junctions.................... 83 3.2.3 Piedmont Forelands........................ 86 vi Contents 3.3 Fault Segmentation of Mountain Fronts.................... 97 3.3.1 Different Ways to Study Active Faults.................. 97 3.3.2 Segmentation Concepts and Classification................ 104 3.3.3 Fault-Segment Boundaries...................... 105 3.3.4 Normal Fault Surface Ruptures.................... 106 3.3.5 Strike-Slip Fault Surface Ruptures................... 113 3.4 Summary.............................. 115 4 Tectonic Activity Classes of Mountain Fronts 4.1 Tectonic Setting of the North America-Pacific Plate Boundary............ 117 4.2 Appraisal of Regional Mountain Front Tectonic Activity.............. 119 4.2.1 Geomorphic Tools For Describing Relative Uplift Rates........... 119 4.2.1.1 Mountain-Front Sinuosity.................. 122 4.2.1.2 Widths of Valleys...................... 124 4.2.1.3 Triangular Facets...................... 127 4.2.2 Diagnostic Landscape Classes of Relative Tectonic Activity.......... 128 4.2.3 Regional Assessments of Relative Tectonic Activity............. 141 4.2.3.1 Response Time Complications and Strike-Slip Faulting........ 141 4.2.3.2 Maps of Relative Uplift................... 145 4.3 Summary.............................. 164 5 Fault Scarps 5.1 General Features............................ 165 5.2 Scarp Morphology Changes with Time.................... 172 5.2.1 Changes in Scarp Height...................... 173 5.2.2 Decreases in Maximum Scarp Slope.................. 174 5.2.3 Diffusion-Equation Modeling..................... 175 5.3 Climatic Controls of Fault-Scarp Morphology................. 181 5.4 Lithologic Controls of Fault-Scarp Morphology................. 184 5.4.1 Fault Rupture of Different Materials.................. 185 5.4.2 Lithologic Controls on an 1887 Fault Scarp................ 187 5.4.2.1 Geomorphic Processes.................... 190 5.4.2.2 Scarp Materials...................... 193 5.4.2.3 Scarp Morphology..................... 194 5.5 Laser Swath Digital Elevation Models.................... 196 5.6 Dating Fault Scarps with Terrestrial Cosmogenic Nuclides............. 201 5.6.1 Alluvium............................ 201 5.6.2 Bedrock............................ 204 5.7 Summary.............................. 207 6 Analyses of Prehistorical Seismic Shaking 6.1 Paleoseismology Goals.......................... 209 6.2 Earthquake-Generated Regional Rockfall Events................. 212 Contents vii 6.2.1 New Zealand Earthquakes...................... 212 6.2.1.1 Tectonic Setting...................... 212 6.2.1.2 Background and Procedures.................. 215 6.2.1.3 Diagnostic Lichen-Size Peaks.................. 225 6.2.1.4 Tree-Ring Analyses..................... 227 6.2.1.5 Alpine Fault Earthquakes................... 241 6.2.1.6 Recent Marlborough Earthquakes................ 246 6.2.2 California Earthquakes....................... 255 6.2.2.1 Calibration of Lichen Growth Rates............... 257 6.2.2.2 Recent Cliff Collapse..................... 258 6.2.2.3 Rockfall Processes in Glaciated Valleys............... 262 6.2.2.4 San Andreas Fault Earthquakes................. 265 6.2.2.5 Lichenometry and Precise Radiocarbon Dating Methods........ 270 6.3 Summary............................... 273 References Cited................................ 275 Index.................................... 305 Preface Uplift by mountain-building forces changes fluvial landscapes. Pulsatory tectonic activity on a range-bounding fault increases relief, changes rates of geo-morphic processes, and modifies the shapes of hills and streams. Landscape responses to uplift occupy a critical time frame for studies of past earthquakes between the brevity of instrumental seismic data and long-term geologic crustal shifts. The appealing challenge for us is to determine how and when nearby and distant parts of the landscape change in consecutive reaches upstream from a tectonically active range front. Each climatic and lithologic setting has a characteristic style and rate of erosion, which adds spice to the scientific challenge. Landscape analyses include the geomorphic consequences of seismic shaking and surface rupture and their associated hazards to humankind. Tectonic geomorphology is essential for complete paleoseismology investigations. Locations, sizes, times, and patterns of seismic shaking by prehistorical earthquakes can be described and surface rupture and seismic-shaking hazards evaluated. This book explores tectonic geomorphology of mountain fronts on many temporal and spatial scales to encourage expansion of paleoseismology inquiries from the present emphasis on stratigraphic investigations in trench exposures. Evaluating earthquake hazards is in part a study of mountain-front segments. Cumulative displacements over late Quaternary time spans create landscape assemblages with distinctive signatures that are functions of uplift rate, rock mass strength, and the geomorphic processes of erosion and deposition. Such interactions define classes of relative uplift. Tectonic activity class maps define tectonically inactive regions as well as fronts of slow to rapidly rising mountains. Fault scarps focus our attention on recent surface ruptures and propagation of active faults. Dating and describing the characteristics of single prehistoric surface-ruptures is important. But now we can link sequences of events and depict sequences of prehistorical earthquakes along complex plate boundary fault zones. Examples here include the Alpine fault in New Zealand and the northern Basin and Range Province in the United States. This book applies a variety of geomorphic concepts to tectonics and paleoseismology. Don't expect landscape summaries for all major mountain ranges. Repetitive descriptions would dilute explanation and application of basic principles. Do expect essential concepts that should help you better understand the landscape evolution of your favorite mountains. Mountain front tectonic geomorphology studies can determine: 1) Which faults are active [Holocene ruptures], 2) Fault slip rates for short time spans [offset landforms] and long time spans [landscape evolution], 3) Time of most recent surface rupture and degree of irregularity of earthquake recurrence interval, and 4) Intensity and extent of seismic shaking. The amount of related literature cited borders on being unwieldy because of topic diversity of and the rapidly increasing interest of earth scientists in these subjects. I had to pick and choose so as to not overwhelm the content with citations of relevant literature. My citations are merely a gateway to related literature. Dating times of prehistoric earthquakes and estimating rates of tectonic and geomorphic processes continue to be of paramount importance. Study methods are changing, and precision and accuracy are improving. Diffusion-equation modeling of fault scarps and stratigraphic radiocarbon dates on pre- or post earthquake material collected from trenches have long been bastions for approximate age estimates. Sykes and Nishenko made a plea in 1984 for better ways of dating frequent earthquakes along plate boundary fault zones whose earthquake recurrence intervals may be shorter than the intervals defined by groups of overlapping radiocarbon age estimates. The rapid development of terrestrial cosmogenic nuclides broadens dating perspectives by estimating ages beyond the reach of radiocarbon analyses and by making surface-exposure dating a cornerstone for studies of geomorphic processes. Tree-ring analyses and lichenometry have potential for dating prehistorical earthquakes with a precision of ± 5 years. Prefai Both methods are used here in a study of Alpine fault, New Zealand, earthquake history. The subjects of the six chapters are wideranging. Acknowledging the scrunch and stretch horizontal components of bedrock uplift is assessed from a geomorphic standpoint in Chapter 1. Diverse, essential conceptual models and methods for fluvial tectonic geomorphology are presented in Chapter 2. Contrasting tectonic land-forms and landscape evolution associated with thrust and normal faults are the focus of Chapter 3. Uplift, stream-channel downcutting, and piedmont aggradation are interrelated base-level processes that are used to define relative classes of mountain-front tectonic activity in Chapter 4. The fault scarps of Chapter 5 are incipient mountain fronts with surface-rupture recurrence intervals ranging from 200 years to 200,000 years. Chapter 6 considers how mountains crumble from seismic shaking. It uses coseismic rock-falls and tree-ring analyses for precise, accurate dating of earthquakes of the past 1,000 years and for mapping the intensity of seismic shaking of these prehistorical events. Readers should know basic geologic principles as these essays are written for earth scientists and students of geomorphic processes, landscape evolution, and earthquake studies. This book is appropriate for upper division and graduate-level courses in active tectonics, geologic hazards, tectonic geomorphology, physical geography and geomorphology, engineering geology, and paleoseismology. This project began in 1975 when Luna Leopold encouraged me to embark on selected in-depth geomorphic syntheses using book manuscripts as a career development tool. Global climate change and tectonic deformation are major factors influencing the behavior of fluvial geomorphic systems. Book goals determined my study emphases in a series of projects. "Geomorphic Responses to Climatic Change" (Bull, 1991) revealed pervasive impacts on geomorphic processes of arid and humid regions. This second book examines tectonic geomorphology of mountain ranges in a paleoseismology context. Of course the varied content of this book is indeed a team effort by the earth-science community. Students in the Geosciences Department at the University of Arizona played essential roles in every chapter. Peter Knuepfer, Larry Mayer, Les McFadden, Dorothy Merritts, and Janet Slate were among the many who tested the conceptual models of Chapter 2 with field-based studies. The first true positive test of the fault segmentation model (Schwartz and Coppersmith, 1984) in Chapter 3 is the work of Kirk Vincent. Les McFadden and Chris Menges broke new ground with me for the Chapter 4 elucidation of tectonic activity classes of mountain fronts of the Mojave Desert and Transverse Ranges of southern California. Susanna Calvo, Oliver Chadwick, Karen Demsey, Julia Fonseca, Susan Hecker, Phil Pearthree, and Kirk Vincent helped define the essential aspects for studies of normal-fault scarps of the Basin and Range Province in a vast region stretching from Idaho into Mexico. Andrew Wells kindly provided fascinating details about the sensitivity of New Zealand coastal and fluvial landscapes to seismic shaking. The integration of geomorphic and structural features shown in the Figure 1.12 map is the work of Jarg Pettinga. Kurt Frankel and Mike Oskin shared results and concepts of work in progress and Figures 5.35-5.40. The book project expanded in scope during a decade when a new lichenometry method was developed to date and describe how seismic shaking influences rockfalls and other landslides. Lichenometry projects included expeditions into the Southern Alps and Sierra Nevada with Fanchen Kong, Tom Moutoux, and Bill Phillips. Their careful fieldwork and willingness to express divergent opinions were essential ingredients for this paleoseismology breakthrough. I appreciate the assistance of John King in sampling and crossdating the annual growth rings of trees in Yosemite, and of Jim Brune's help in measuring lichen sizes near the Honey Lake fault zone. Jonathan Palmer introduced me to Oroko Swamp in New Zealand, which turned out to be a key dendroseismology site. Images are essential for landscape analysis and portrayal. Tom Farr of the Jet Propulsion Laboratory of the California Institute of Technology always seemed to have time to help find the essential NASA and JPL images used here. The banner photo for Chapter 2 and Figure 4.14 are the artistry of Peter Kresan. I thank x Preface Frank Pazagglia for Figure 2.4, Malcolm Clark for the a pleasure to work w Chapter 4 banner photo, Tom Rockwell for the Figure 5.28 image, Greg Berghoff for Figure 5.34, Scott Miller for Figure 6.2 and Eric Frost for Figure 6.9A. Formal reviews of the entire book manuscript by Lewis Owen and Philip Owens provided numerous suggestions that greatly improved book organization and content. I am especially indebted to Wendy Langford for her meticulous proofreading and to Rosie Hayden for editorial suggestions. Their thoroughness improved format and uniformity of expression. It was th the efficient production staff at Blackwell Publishing including Ian Francis, Rosie Hayden, and Delia Sandford. Essential financial and logistical support for this work was supplied by the U.S. National Science Foundation, National Earthquake Hazards Reduction Program of the U.S. Geological Survey, National Geographic Society, University of Canterbury in New Zealand, Hebrew University of Jerusalem, Royal Swedish Academy of Sciences, and Cambridge University in the United Kingdom. Earthquakes! Active Tectonics! Evolution of Mountainous Landscapes! Landscapes have a fascinating story to tell us. Tectonic geomorphology intrigues laypersons needing practical information as well as scientists curious about Earths history How fast are the mountains rising? When will the next large earthquake occur? Will the seismic shaking disrupt the infrastructures that we depend on? How do the landscapes surrounding us record mountain-building forces within the Earths crust, and how does long-term erosion influence crustal processes? Humans are intrigued by tectonic geomorphology on scales that include origins of continents, grandeur of their favorite mountain range, and the active fault near their homes. Let us expand on the purpose and scope summarized in the Preface by elaborating on the structure of this book. I introduce, describe and use geomorphic concepts to solve problems in tectonics and paleoseismology. The intended geographical focus is global application of examples from southwestern North America and New Zealand. A fluvial emphasis excludes glaciers, sand seas, and active volcanoes. I present data and analyses from diverse tectonic, climatic, and lithologic settings so you can resolve similar problems in other geographical settings. This book emphasizes responses of fluvial systems to uplift, or more specifically the adjustments of geomorphic processes to base-level fall. Uplift terminology usage continues to change since the hallmark paper by Molnar and England (1990). Geomorphologists may use uplift terms in a different context than structural geologists. So Chapter 1 is a brief review of terminology and types of base-level change induced by tectonic deformation in extensional and contractional settings. Such crustal stretching and scrunching is nicely recorded by landforms ranging in size from mountain ranges to fault scarps. A variety of useful geomorphic concepts are assembled in Chapter 2 instead of being scattered. Get familiar with these principles. This broad base of essential concepts lets you evaluate and explore new and diverse approaches in tectonic geomorphology. These include a sensitive erosional-depositional threshold, time lags of response to perturbations (changes in variables of a system), types of equilibrium (graded) conditions in stream Photograph of 59,000 and 96,000 marine terraces (Ota et al., 1996) and 330,000 year old mountains (Bull, 1984, 1985) rising out of the sea at Kaikoura, New Zealand Chapter 1 systems, local and ultimate base levels, and the process of tectonically induced downcutting to the base level of erosion. These guidelines are a foundation for understanding interrelations between tectonics and topography in the next three chapters. Chapter 3 compares the landscape evolution and useful tectonic landforms for mountain ranges being raised by slip on active thrust and normal faults. These fluvial systems are affected differently by the two styles of tectonic base-level fall. Strike-slip faulting tends to tear drainage basins apart: a much different subject that is not emphasized here. Some tectonic landforms, like triangular facets, are rather similar in different tectonic settings. But piedmont landforms are much different in thrust-and normal-fault landscapes. Comparable contrasts should be expected elsewhere, such as the countries bordering the Mediterranean Sea, and Mongolia. The next three chapters discuss tectonic geo-morphology for three distinct time spans (Fig. 1.1) of about 2,000,000, 12,000, and 1,000 years. The tec-tonic-geomorphology theme continues to be applications for paleoseismology. The landscape tectonic activity classes of Chapter 4 are based on universal geomorphic responses to different rates of base-level fall during the Quaternary time span. The resulting diagnostic landscape assemblages are defined and mapped for diverse tectonic and structural settings in California. This model could have been created, and applied, just as easily for suites of mountain fronts in Japan, China, Mongolia, and Russia. Fault scarps are the focus of Chapter 5, with an emphasis on the Holocene time span. Choosing to discuss recent surface ruptures in southwestern North America was done in part to hold variations of several controlling factors to a limited range. These include climate and alluvium mass strength. Such studies of incipient mountain fronts can be made just as easily in the Tibetan Plateau, the Middle East, and Africa. New approaches are overdue to decipher the sequences of frequent earthquakes that characterize plate-boundary fault zones. Chapter 6 develops a new geomorphic way to precisely date earthquakes in New Zealand and to describe their seismic shaking. It then tests the model in California. This geomor- Stretch and scrunch bedrock uplift Conceptual models for fluvial tectonic geomorphology \ 1 Thrust-and normal- Classes of rising Late Quaternary Surface-rupture and faulted mountains landscapes 'Fault scarps seismic-shaking events Pleistocene ^~ Holocene ^^^^ - The past 1,000 years T Figure 1.1 Major topics of this book and their application to paleoseismology. Scrunch and Stretch Bedrock Uplift phic approach to paleoseismology provides essential information about the frequency and magnitude of recurrent tectonic perturbations such as surface ruptures and seismic shaking. Other plate-boundary settings, such as the Andes of South America, Anatolian fault zone of Turkey, and the Himalayas may be even better suited for this way to study earthquakes than my main study areas. This book uses two primary, diverse study regions to develop concepts in tectonic geomorphol-ogy for fluvial systems in a global sense. Principal sites in New Zealand are shown in Figure 1.2 and southwestern North America sites in Figure 1.3 together with the links to their chapter section numbers. 1.1 Introduction Continental landscapes of planet earth are formed in large part by interactions of tectonic and fluvial processes, which are modulated by the pervasive influence of late Quaternary climate changes. Tectonics is the study of crustal deformation: the evolution of Figure 1.2 Locations of Southern Alps study sites in the South Island of New Zealand discussed in Sections 1.2, 2.4, 2.5, 2.6, and 6.2.1. This is a grayscale version of Shuttle Radar Topography Mission image PIA06662 furnished courtesy of NASA and JFL. 4 Chapter 1 Scrunch and Stretch Bedrock Uplift 5 geologic structures ranging from broad transition zones between crustal plates to small faults and folds. Geomorphology is the study of landscapes and the processes that shape them. The influences of vertical and horizontal earth deformation on fluvial, coastal, and glacial processes and the resulting landscapes comprise the domain of tectonic geomorphology. The main emphasis here is on fluvial system responses to tectonic deformation. The challenge for all of us is to more fully recognize and use tectonic signals in the landscapes around us. The consequences of earth deformation by specific geologic structures profoundly affect geomorphic processes and landscape evolution. Conversely evolution of landscape assemblages can be used to decipher the kinematics of faults and folds. Changes in style, rate, and locations of faulting and folding change the landscape too. An example is the Hope fault of New Zealand where Eusden et al. (2000) describe a 13 km long and 1.3 km wide transpressional duplex structure (adjacent areas of rise and fall) that has migrated northeast along a range-bounding oblique-slip fault that is as active as the San Andreas fault of California, USA. This leading portion of the duplex structure is rising on thrust faults. In the trailing southwest portion, formerly active duplex structures are now collapsing, undergoing a reversal of slip style to become normal faults. Rising geomorphic base levels become falling base levels with dramatic consequences for hills and streams of upstream watersheds. Another example is drainage nets that change as tips of faults propagate (Jackson et al., 1996). Structural geologists need to recognize how tectonic deformation affects erosion, deposition, and landforms. Tectonic geomorphology aids tectonic inquiries on many temporal and spatial scales. Some of us seek to understand how horizontal, as well as vertical, earth deformation affects the shapes of hills and streams in a quest to better understand long-term partitioning of strain along plate boundary-fault systems (Lettis and Hanson, 1991). Others study landslides in order to determine earthquake recurrence intervals and to make maps depicting patterns of seismic shaking caused by prehistorical earthquakes (Chapter 6). Tectonic geomorphology seismology and paleoseismology are cornerstone disciplines for studies of active tectonics (neotectonics). Seismology—historical instrumental studies of earthquakes - contributes much to our understanding of crustal structure and tectonics by 1) defining earthquake hypocenters (location and depth of initial rupture along a fault plane), 2) describing earthquake focal mechanisms (strike-slip, normal, and reverse styles of displacement), 3) evaluating the frequency magnitude, and spatial distributions of present-day earthquakes, and 4) modeling how yesterdays earthquake changes the distribution of crustal stresses that will cause future earthquakes. Paleoseismology — the study of prehistorical earthquakes - utilizes many earth science disciplines including dendrochronology geochronology, geodesy geomorphology seismology soils genesis, stratigraphy and structural geology Tectonic geomorphology is indispensable for complete paleoseis- Figure 1.3 Locations of study sites in the western United States and northern Mexico and their book section numbers [5.5]. F3, Pleistocene Lake Bonneville [5.2.3]; F3L, F3ig Lost River [5.3]; F3R Borah Peak and the Lost River Range [3.3.4]; CP, Curry Draw [2.2.3]; CR Colorado Plateau; PR, Piablo Range [4.2.3.2]; PV Death Valley, Panamint Range, and Saline Valley; FR, Front Range [1.3]; GC, Grand Canyon [2.5.2]; GY, Great Plains [1.3]; HL, Hebgen Lake [5.6.2]; KK, Kings River [6.2.2.2]; L, Pleistocene Lake Lahontan [5.2.3]; LS, Laguna Salada [2.1]; MC, McCoy Mountains [3.2.2]; MP, Mojave Pesert [4.2.3.2]; ML, Mount Lassen and the southern end of the subduction related Cascade volcanoes [4.1]; MR, Mogollon Rim [4.2.2]; NPR, North Platte River [1.3]; OM, Olympic Mountains [1.2.2, 5.5]; PIT, Pitaycachi fault [2.2.5, 5.4.2]; PR, Panamint Range, Peath Valley, and Saline Valley [4.1, 4.2.2, 4.2.3]; PS, Puget Sound [5.5]; KG, Rio Grande River and extensional rift valley [1.3]; RM, Rocky Mountains [1.3]; S, Socorro [6.6.1]; SGM, San Gabriel Mountains [3.2, 3.3.1, 4.2.3.2]; SJV San Joaquin Valley [4.1, 4.2.3.2]; SM, Sheep Mountain [4.2.2]; SN, Sierra Nevada microplate [4.1, 6.2.2]; ST, Salton Trough [4.2.3.2]; TR, Tobin Range, Pleasant Valley, Pixie Valley, and the Stillwater Range [3.2.1, 4.2.3.2, 5.1, 5.6.2]; WC, Wallace Creek [2.5.2]; WL, Walker Lake [5.2.3]; WR, Wasatch Range [3.3.3]; YO, Yosemite National Park [6.2.2.1, 6.2.2.3]. Pigital topography courtesy of Richard J. Pike, US Geological Survey. Chapter 1 mology investigations. For example, stream-channel downcutting and diffusion-equation modeling of scarp erosion to complement stratigraphic information gleaned from trenches across the fault scarp. Quaternary temporal terms (Table 1.1) have been assigned conventional ages*. The 12-ka age assignment for the beginning of the Holocene is arbitrary and is preceded by the transition between full-glacial and interglacial climatic conditions. Unless specifically noted, radiocarbon ages are conventional (using the old 5,568 year half-life allows comparison with dates in the older literature) and have been corrected for isotope fractionation. The term "calendric radiocarbon age" means that the correct 5,730 year half-life is used and that variations in atmospheric 14C have been accounted for, using the techniques of Stuiver et al. (1998). Calibration of radiocarbon ages (Bard et al., 1990) shows that the peak of full-glacial conditions may be as old as 22 ka instead of the conventional radiocarbon age estimate of 18 ka. The 125 and 790-ka ages are radiometric and paleomagnetic ages that have been fine-tuned using the astronomical clock (Johnson, 1982; Edwards et al., 1987a, b). The 1,650-ka age is near the top of the Olduvai reversed polarity event (Berggren et al., 1995). Landscape evolution studies accommodate many time spans. Topics such as the consequences of rapid mountain-range erosion on crustal processes involve time spans of more than 1 My. Examinations Age Ka Holocene Late 0-4 Middle 4-8 Early 8-12 Pleistocene Latest 12-22 Late 12-125 Middle 125-790 Early 790-1650 Table 1.1 Assigned ages of Quaternary temporal terms, in thousands of years before present (ka). *1 ky = 1000 years; 1 ka = 1 ky before present. 1 My = 1 million years; 1 Ma = 1 My before present. of how Quaternary climate changes modulate fluvial system behavior generally emphasize the most recent 50 ky. Understanding the behavior of fault zones concentrates on events of the past 10 ky. The first concepts discussion about examines several processes that raise and lower the land surfaces of Chapter 1 study sites. Streams respond to uplift by eroding mountain ranges into drainage basins. So Chapter 2 then examines how far streams can cut down into bedrock - their base level limit. We also explore the behavior of fluvial systems to lithologic and climatic controls in different tectonic settings in the context of response times, the threshold of critical power, and tectonically induced downcutting. These concepts will give you a foundation for perceiving tectonic nuances of mountain fronts and hillslopes. 1.2 Pure Uplift, Stretch and Scrunch Bedrock Uplift 1.2.1 Isostatic and Tectonic Uplift My approach to tectonic geomorphology examines some of the myriad ways that uplift may influence fluvial landscapes. Many new methods and models alter our perceptions of tectonics and topography as we seek to better understand everything from landscapes and prehistorical earthquakes to crustal dynamics. So we begin this chapter by examining the intriguing and occasionally puzzling meanings of the term "uplift". My emphasis is on how subsurface processes affect altitudes of all points in a landscape. Read England and Molnar's 1990 article and you will come away with a fascinating perspective about several components that influence uplift of points on the surface of a large mountain range. The key to using their breakthrough is to recognize the factors influencing uplift of bedrock, not only at the land surface but also at many positions in the Earth's crust. I introduce additional parameters that also influence rock uplift. Both tectonic and geomorphic processes influence bedrock uplift (Fig. 1.4). S.I. Hayakawa's semantics philosophy (1949) certainly rings true here; "The word is not the object, the map is not the territory". Not only will each of us have different (and changing) impressions of uplift terminology, but also my attempts to neatly organize key variables are hindered by substantial overlap between categories. Fault displacements do more than raise and lower bedrock (pure uplift), because Scrunch and Stretch Bedrock Uplift Pure Uplift Tectonics Crustal changes >105 km2, >105 ky >10s km2, >10s ky Isostatic uplift Scrunch and Stretch Tectonics Tectonic denudation or buria >105 km2, >105 km2, >105 ky >105 ky Tectonic displacements Local 10'to >103 km2, 10"1to>105ky Crusta 103to >104km2, 103to >104ky Bedrock uplift Geomorphic processes Surficia 10zto >103 km2 10° to >103 ky Subsurface 102to >104km2 10° to >104ky Net change of land-surface altitude Tectonic uplift Figure 1.4 Links between tectonic, isostatic, and nontectonic variables affecting landscape altitudes and bedrock uplift. Feedback mechanisms to isostatic and tectonic uplift are shown with dashed lines. earth deformation usually entails tectonic shortening (scrunch) and extension (stretch) processes. So, rocks move both vertically and horizontally (Willett, Slingerland, and Hovius, 2001). Do not expect crisp black-and-white definitions, because a model of fuzzy overlap is closer to the truth in the world of scrunch-and-stretch tectonics. Erosion of mountains is like rain falling on a marine iceberg; the height of both results from buoyant support. Rainfall can never melt enough ice to lower the surface of an iceberg to the water line. This is because ice melted above the waterline is largely replaced by "uplift" of submerged ice. Sea level is a handy reference datum for uplift of ice or mountain ranges. Uplifted materials may be above or below that worldwide waterline. Altitude is the specific term for height above present sea level, whereas the engineering term "elevation" can have several geologic connotations, including uplift. Isostatic uplift occurs because ice is only 90 % as dense as seawater. If 100 tons is melted from the exposed surface of an iceberg, it is compensated by 90 tons of ice raised by isostatic uplift. This is pure uplift because it is not complicated by shearing or tensional failure of ice. Similarly, isostatic uplift of mountain ranges continues despite eons of surficial erosion because continental crust "floats" on the denser rocks of the Earth's mantle. Continental crust with a density of about 2,700 kg/m3 is in effect floating on mantle with a density of about 3,300 kg/m3- a density contrast of roughly 82% (90% contrast for oceanic crust with a density of 3,000 kg/m3). The iceberg analogy is appropriate because materials deep in the earth behave as viscous fluids over geologic time spans (Jackson, 2002). Fluvial and glacial denudation of 1,000 m only seems to significantly lower a mountain range because it is largely compensated by 820 m of concurrent isostatic rebound. Neither ice nor rock landscapes remain the same, unless erosional lowering is the same for all points in a landscape. Relief and altitudes of peaks increase if melt of ice, or erosion of rocks, is mainly along valley floors. Removal of mass above our sea-level datum causes pure isostatic uplift of all parts of the landscape. The average altitude of both the iceberg and the mountain range decreases with time because buoyancy-driven isostasy can never fully compensate for the mass lost by erosion. Chapter 1 A substantial proportion of mountain-range uplift is the result of these crustal isostatic adjustments (Molnar and England, 1990; Gilchrist et al, 1994; Montgomery, 1994; Montgomery and Greenberg, 2000). Isostatic uplift is both regional and continuous (Gilchrist and Summerfield, 1991), and generally does not cause pulses of renewed mountain building. This is done by scrunch and stretch tectonics. A major difference between icebergs and mountain ranges is that mountains do not float in a Newtonian fluid such as water, which has no shear strength. Continental rock masses float on hot litho-spheric materials whose rigidity provides some support. Rocks at shallower depths are stronger (cooler) and respond to changes in load by flexing in an elastic manner. Small, local changes in rock mass will not cause the lithosphere to flex because it has enough strength to support minor changes in load. But beveling of a 10,000 km2 mountain range will indeed influence crustal dynamics. Prolonged erosion has resulted in substantial cumulative isostatic rebound of the Appalachian Mountains of the eastern United States for more than 100 My Tectonic geomorphologists would prefer to discern how different uplift rates influence land-forms and geomorphic processes, but reality is not that simple. Mountain-building forces may continue long after tectonic quiescence seems to have begun, as revealed by strath terraces (a tectonic landform discussed in Sections 2.4.1 and 2.6) in pretty dormant places like Australia (Bierman and Turner, 1995). Space and time frameworks of references vary greatly for the Figure 1.4 surface-uplift variables. Generally, they are large and long for pure uplift, tectonic denudation, or burial, and small and short for tectonic displacements and geomorphic processes. The predicament is that uplift has two components - tectonic and isostatic. Tectonic mountain-building forces may cease but the resulting isostatic adjustments will continue as long as streams transfer mass from mountains to sea. The best we can do at present is to observe landscape responses to the algebraic sum of tectonic and isostatic uplift. Bedrock uplift = Tectonic uplift + Isostatic uplift (1.1) This seems simple, until we attempt to quantify the Figure 1.4 variables that influence tectonic uplift and isostatic uplift. The term bedrock is used here in a tectonic instead of a lithologic context. Bedrock is any earth material that is being raised, with no regard as to the degree of lithification or age. We should note the "fuzziness" of this definition. Three exceptions are acknowledged; these occur when the nontectonic surficial process of deposition raises a landscape. The most obvious and dramatic is volcanic eruption, which raises landscape altitudes by depositing lava and tephra. Of course volcanic eruptions may also be associated with tectonic shortening and extension. Tectonic geomorphologists are interested in how climate change affects the behavior of streams in humid and arid regions. Mountain valleys and piedmonts undergo aggradation events as a result of major climate changes (Bull, 1991) that change the discharge of water and sediment. We do not class such stream alluvium as bedrock because its deposition is the result of a nontectonic process that raises valley-floor altitudes. Alluvium laid down before the particular time span that we are interested in would be treated like other earth materials, as bedrock. Studies of Pleistocene uplift would treat Miocene fluvial sand and gravel as bedrock. Thirdly, nontectonic deposition includes eolian processes such as the creation of sand dunes. Least obvious, but far more widespread, is deposition of loessial dust. In New Zealand windblown dust is derived largely from riverbeds after floods and the loessial blanket that covers much of the stable parts of the landscape may contain layers of volcanic ash, such as the 26.5 ka Kawakawa tephra (Roering et al., 2002, 2004). Hillslopes where this ash has been buried by 0.5 to 5 m of loess are landscapes where deposition has slowly raised the altitudes of points on the land surface during the 26 ky time span at average rates of <0.02 to >0.1 m/ky. Deposition - by volcanic ejecta, inability of a stream to convey all bedload supplied from hillslopes, and dust fall - is just one of several nontectonic geomorphic processes that change altitudes of points in a landscape (Fig. 1.4). I prefer to emphasize bedload transport rates in this book because bedload governs stream-channel responses to bedrock uplift. Rivers transport mainly suspended load to the oceans and deposit silty sand and clay on floodplains. Dissolved load is bedrock conveyed in solution. Both require little stream power, but the unit stream power required to mobilize and transport bedload reduces the energy available for tectonically induced downcutting of stream channels. Saltating cobbles and boulders are tools for abrasion of bedrock. With suspended load being Scrunch and Stretch Bedrock Uplift flushed downstream it is bedload that is deposited as fill stream terraces and many alluvial fans. Such land-forms are used to analyze responses of fluvial systems to bedrock uplift and to changes in late Quaternary climate too. Three classes of deposition in Figure 1.5 illustrate the care needed in defining sand and gravel as bedrock. The active range-bounding fault controls the behavior of this fluvial system. Stream channel processes normally switch abruptly from net erosion to net deposition after crossing the fault zone. Sandy alluvial-fan deposits of Miocene age have been elevated and now underlie watersheds in this hypothetical mountainous landscape. Most of us would agree that the uplifted Miocene fan deposits, although unconsolidated, should be classed as bedrock in a geomorphic sense. They are mountainous terrain into which drainage basins are carved. The gravelly fill terraces in the valley upstream from the mountain front are the result of climate-change perturbations. Without perturbations the watersheds of tectonically active mountain ranges in the Mojave Desert would have undergone uninter- rupted long-term degradation of their valley floors. But late Quaternary climatic fluctuations significantly affected sediment yield and stream discharge. Climate-change perturbations in arid and humid watersheds can temporarily reverse the tendency for stream-channel downcutting, even in rapidly rising mountain ranges. Climate-change perturbations are dominant because they quickly affect geomorphic processes throughout a drainage basin, whereas uplift on a fault zone is local and the resulting increase in relief progresses upstream relatively slowly. Climate-change induced aggradation events in the Mojave Desert raised valley floors <5 to >50 m. The range is largely due to lithologic controls on weathering and erosion. Aggradation was the result of insufficient stream power to convey bedload supplied from hillslopes whose vegetation changed drastically when the climate changed. Major aggradation events at about 125 ka and 10 ka were times of widespread stripping of hillslope sediment reservoirs that were no longer protected by dense growth of plants. A climatic perturbation at about 60 ka also coincides with a global sea-level highstand and caused an aggra- Age of aggradation event valley fill 125 ka 60 ka 10 ka soil Figure 1.5 Summary of late Quaternary deposition for a typical Mojave Pesert, California fluvial system where times and locations of aggradation are controlled by climatic perturbations that overwhelmed the effects of uplift along active fault zones. Hachures show soil profiles that postdate the ends of aggradation events and record brief intervals of nondeposi-tion on the fanhead. 10 Chapter 1 dation event of smaller magnitude. So the sedimen-tology and thickness of each late Quaternary fill-terrace, and the concurrent increments of alluvial-fan deposition, were different (Table 1.2). Vertical separations between the beveled bedrock beneath several valley fills record stream-channel down cutting induced by uplift along the range-bounding fault zone in the intervals between climate-change induced aggradation events. The potent 10 ka aggradation event might have buried the equally strong 125 ka aggradation event at cross section A—A if there had been no tectonically-induced lowering of the valley floor. More tectonically induced degradation has occurred at cross section B-B' than at A-A because it is closer to the active fault zone. Depositional elevation of the stream terrace tread is a clear-cut example of nontectonic elevation of landscape elements. Such deposits should not be classed as bedrock. How should we regard the area of active alluvial-fan deposition downstream from the range-bounding fault? Surface ruptures on the normal fault create the space for continuing accumulation of basin fill. Such fans are tectonic landforms because nearly constant deposition would not have occurred without continuing uplift. Differential uplift along the fault has been sufficiently rapid to maintain late Quaternary aggradation adjacent to the mountain front (Section 4.2.2). Major Late Quaternary climatic changes caused the rate of fan aggradation to vary and influenced the locations of fan deposition. Minor, brief climatic fluctuations are superimposed on the long-term climatic controls. They caused brief episodes of stream-channel downcutting in the mountains and temporary entrenchment of the fan apex. Brief local cessation of depositional processes allowed incipient soil-profile development on the fan surfaces adjacent at cross section C-C. Each aggradation event was strong enough to backfill the fanhead trench, thus allowing fan deposition to continue to radiate out from an apex at the mountain front. It is debatable as to whether such fan deposits should be regarded as bedrock. Perhaps they should be classed as bedrock because the locations of fan deposition are tectonically controlled. Deposition of a thick fan would not occur here in the absence of active faulting. Alternatively, one might argue that rates of sedimentation vary with late Quaternary climates. Deposition merely tends to partially offset tectonic lowering of basin altitudes in an extensional terrain. Such fans should not be classed as bedrock. Lithospheric rigidity interjects the important element of scale into our perception of what constitutes uplift. Tectonic-uplift variables behave differently at the local scale of a single hillside or small watershed as compared to large chunks of the Earth's crust. For each point in a landscape, tectonic deformation caused by different styles of faulting and folding is superimposed on regional uplift (or subsidence) caused by broad warping of the lithosphere. This is a matter of different wavelengths for different earth-deformation processes. Alluvial geomorphic surface Aggradation age, ka Basis for age estimate Q4 Active washes, riparian trees, no rock varnish on cobbles Q3b ~S> 14C dating of plant fossils, lake stratigraphy, rock varnish Q3a -12 14C dating of plant fossils, lake stratigraphy rock varnish Q2c -60 230Th/234U ages of pedogenic carbonate, uranium-trend date, calibrated fault slip age estimate, cosmogenic 10f3e age estimate Q2b -125 230Th/234U ages of pedogenic carbonate Q2a 240-730 K/A dating of tuff, basalt flow, normal paleomagnetic polarities Q1 >1,200 K/A dating of basaltic sources dissected into ridges and ravines Table 1.2 Pulses of climate-change Induced alluviation in the Mojave Desert of California. Summarized from Tables 2.13 and 2.15 of Bull (1991). Scrunch and Stretch Bedrock Uplift 11 Tectonic uplift = Local uplift + Crustal uplift (1.2) The background regional crustal warping may be slow or fast, but it affects erosion rates of local land-forms as well as those of entire mountain ranges. Local faulting creates topographic anomalies such as rising mountain fronts that attract tectonic geo-morphologists (Chapter 3). We analyze landforms to separate local tectonic deformation from background regional uplift. But separating tectonic from isostatic uplift can be difficult at the watershed spatial scale because not all earth deformation is purely vertical. Scrunch and stretch tectonics plays an important role in deformation of Earths crust. For example, plate-boundary subduction is a tectonic process, but how much of the resulting bedrock uplift is the result of isostatic uplift caused by thickening of the crust? How much is the result of scrunch induced by concurrent folding and thrust faulting? Conversely, in extensional terrains how much of a decrease in altitude is offset by isostatic adjustment resulting from concurrent erosion of mountain ranges? How much of lowering induced by stretch tectonics is offset by aggradation (Fig. 1.5) in basins that receive the deposits? Let's begin with brief summaries of the contents of the "Pure Uplift" and "Geomorphic Processes" boxes of Figure 1.4 to gain background before delving into "Stretch and Scrunch" box. Many factors affect magnitudes and response times for isostatic uplift. Important slow changes in the crust include accretion, or thinning, of light, buoyant crustal materials. Temperature increase or decrease changes the density of crustal rocks, thus changing their buoyancy. Phase changes in minerals that reflect changing pressures or temperatures alter buoyancy contrasts with adjacent rocks. Change to denser minerals decreases rock volume, which also tends to directly lower land-surface altitudes. Pure strike-slip faulting does not raise or lower a landscape, but major horizontal shifts of mountain ranges and crustal blocks may alter regional distributions of isostatic forces. Many plate-boundary strike-slip faults have cumulative displacements of more than 50 km, so this style of tectonic deformation may change the crustal loads on opposite sides of a fault sufficiently to cause isostatic re-adjustments. This important aspect of strike-slip faulting deserves its separate box within pure uplift tectonics in Figure 1.4. Changes in altitude that occur at bends and sidesteps of strike-slip faults, are included in the local- tectonic-displacement box because they are classified as scrunch and stretch tectonics. Transpressional or transtensional components of most plate-boundary fault zones also are best considered as part of scrunch and stretch tectonics. Tectonic processes and isostatic uplift may increase land-surface altitudes, but landscape altitudes also change because of several geomorphic processes. We have already mentioned the surficial processes of fluvial and volcanic deposition. Another is fluvial erosion, which tends to lower hills and streams. Both sets of processes affect crustal weight, and when sufficient may cause isostatic adjustments. Diagenesis of recently deposited basin fill tends to lower land-surface altitudes. Compaction of saturated clayey, silty beds in a sedimentary basin is analogous to crustal changes that produce denser minerals. It is pure vertical subsidence. Bulk density increases as water is gradually expelled from sediments by the weight of the overlying stratigraphic section, plus several hydrodynamic forces. The resulting decrease in bed thickness lowers the overlying strata and the land surface. Ground water derived from infiltrating rain and snowmelt dissolves minerals. Solution is a greatly different geomorphic process than landsliding because it is not visually conspicuous. It occurs below the land surface and the resulting ions are invisible in emerging clear springs that nourish streamflow. But substantial mass is removed over Quaternary time spans at depths that range from surficial soil profiles to more than 1 km. The net surface uplift resulting from all Figure 1.4 processes is an algebraic sum. Surface uplift = Rock uplift + Geomorphic Processes (1.3) The sum of geomorphic processes has feedback loops to isostatic uplift and tectonic deformation. Stretching and scrunching are important tectonic processes that lower or raise landscape altitudes. Most importantly, they (not isostatic uplift) initiate the creation of mountain ranges. Let us think of these as being tectonic denudation (Fig. 1.6) and tectonic burial (Fig. 1.8). Both are common, and operate at a variety of spatial scales. I'll focus mainly on scrunch processes because local uplift may appear anomalous when it is ten times the expected regional uplift. Also, it seems that tectonic denudation processes are already nicely discussed in the literature of the past two decades. 12 Chapter 1 Figure 1.6 diagrammatic sketch of extension associated with normal faulting that causes tectonic denudation and crustal thinning. Rollover folds form where gravitational collapse progressively increases closer to the normal fault. Frictional resistance during displacement of the hanging-wall block generates the shear couple responsible for drag folds next to the fault. 1.2.2 Stretch and Scrunch Tectonics Tectonic stretching (Fig. 1.4) is important. The resulting tectonic denudation is widely recognized, and generally is thought of as normal faulting that thins the crust (Armstrong, 1972; Davis and Coney, 1979; Shackelford, 1980; Spencer, 1984; Coney and Harms, 1984; Pain, 1985; Wernicke, 1992; Dickinson and Wernicke, 1997; Burbank and Anderson, 2001, p. 149-151). Normal faulting also occurs locally in compressional settings (Molnar and Lyon-Caen, 1988; Gammond, 1994; Eusden et al., 2005a). England and Molnar (1990) combined tectonic denudation and surficial erosion into a single process called "exhumation". Low angle detachment faulting (Lister et al., 1986; Bradshaw and Zoback, 1988; Lee and Lister, 1992; Dokka and Ross; 1995; Bennett et al., 1999) can efficiently remove large amounts of bedrock, thereby promoting isostatic rebound (Wernicke and Axen, 1988). Normal faulting in the Basin and Range Province of the western United States has resulted in extension of more than 250 km (Wernicke and Snow, 1998), with a crust that has thinned to about 30 km (Jones et al., 1992). The lower crust of the Basin and Range province should behave as a viscous fluid (Bird, 1991; McCarthy and Parsons, 1994), tending to fill voids created by tectonic extension. Stretch tectonics has distinctive features and resulting landforms (Fig. 1.6). The footwall block typically has minimal secondary faulting, but ten- Scrunch and Stretch Bedrock Uplift 13 sional forces create a myriad of antithetic and synthetic faults in the hanging-wall block. These result mainly from removal of vertical support. Complex structures are induced in hanging-wall blocks where normal fault dip becomes less with depth below the surface to create listric faults. This promotes fault-bend folding. Gravitational collapse is greatest near master detachment faults to create rollover folds (Hamblin, 1963, 1965). These tectonic processes lower surface altitudes. Local vertical displacements, such as range-bounding faults, create space that allows deposition of alluvial fans and other basin fill. Such aggradation raises surficial altitudes so deposition of basin fill is a process that partially offsets tectonic lowering, perhaps by a factor of half. Upper crust thinning enhances the potential for upwelling and isostatic uplift. Crustal rigidity extends this iso-static rebound into the footwall block at the left side of Figure 1.6, in an exponentially decreasing manner with increasing distance from the range-bounding fault. Spatially variable isostatic rebound tilts the land surface. Tectonic denudation caused by a variety of stretch processes thins the upper crust. These reduce crustal loading, and together with an increase in geothermal gradient and lithospheric upwelling promote isostatic uplift that partially offsets the stretch-induced subsidence (Bird, 1991). This self-arresting feedback mechanism is opposite of that caused by tectonic scrunching. The style of normal faulting affects the behavior of fluvial systems. The example used here examines stretch-tectonics controls on the thickness of piedmont alluvial fans. Continuing lowering of a valley and/or uplift of the adjacent mountains creates the space for new increments of piedmont deposition. The resulting alluvial fans reflect the style and rate of tectonic deformation. Prolonged displacement on a range-bounding normal fault can result in fan deposits more than 1,000 m thick. Alluvial-fan deposits are thickest where basins quickly drop away from the mountains, such as the high-angle normal faults of the Basin and Range Province of the western USA. Fan deposits are much thinner where tectonic displacements occur on low-angle faults. Examples include where thrust-faulted mountain fronts are shoved up and over adjacent basins along low-angle faults that dip back into the mountains (Section 3.2.3). I use many examples from the Death Valley region of southeastern California in subsequent chapters, so introduce an interesting example of stretch tectonics here. The locale is the western flank of the Panamint Range. Low-angle normal faults have played an important role in both tectonic extension and landscape evolution of the Death Valley region. Style of alluvial-fan deposition varies with type of fault. Debate continues as to how important such detachment faults are as compared to normal faults that dip steeply at 45° to 65° (Wernicke, 1981, 1995; Walker et al., 2005). Cichanski (2000) made a detailed study of the cur-viplanar low-angle normal faults on the west flank of the Panamint Range that were first noted by Noble (1926) and Maxon (1950). As a geomorphologist, I have no doubt that normal faults that dip only 15° to 35° had substantial slip during the late Cenozoic. My premise is based on the idea that changes in the kinematics of faulting change the landscape. The evidence is the contrasting styles of alluvial-fan deposition. One would expect different types of alluvial fans resulting from low-angle and high-angle normal faulting. Adjustments of fluvial systems to movements on 60° and 25° normal faults are much different (Fig. 1.7). Slip on either steep or gentle fault surfaces causes fluvial systems to cut down into the footwall block and to deposit a new increment of detritus on the hanging-wall block. Part of the newly exposed fault plane is subject to the initial stages of dissection by water flowing in rills, and part is quickly buried by the newest increment of alluvial-fan deposition. Fan slope is also a function of magnitude and type of streamflow events, and the amount and particle-size distribution of the entrained sediment (Bull, 1962; Hooke, 1967). Although many alluvial fans in the Basin and Range Province slope less than 10°, steeper fans are common. Most fans along the Lost River fault zone near Borah Peak in Idaho have fanhead slopes of more than 20° (Section 3.3.4). An assumed fan slope of 20° for the ancestral fans along the western flank of the Panamint Range seems reasonable for this discussion. Thicknesses of tectonic alluvial fans are a function of fault dip and fan slope. The combination of a 60° normal fault and a 20° fan surface provides ample space for thick deposits to accumulate adjacent to the footwall block. Fan thickness in the Figure 1.7B example is 40 m, and would be the maximum of 50 m if the range-bounding fault were vertical. Steep faults are sites of thick fans of small areal extent. Extension on high-angle normal faults also favors incision of deep valleys in the footwall block. 14 Chapter 1 Lx-z=220 m Lx-z=57 m Figure 1.7 Diagrammatic sketches showing how change from low-angle to high-angle normal faulting changes landscape characteristics. H is slope fall, and L is slope length horizontal distance. Vertical tectonic displacements, Hx-z, total 100 m in both cases as the footwall block slips from X to Z. Horizontal tectonic displacements, Lx-z, of 57 and 220 m are a function of normal fault dip. F3 is the present threshold-intersection point where erosion changes to deposition, assuming that the increase of relief of the footwall block is distributed evenly between alluvial-fan deposition and valley deepening. A. 25° normal-fault dip and a 20° fan slope. F3. 60° normal-fault dip and a 20° fan slope. In contrast, only thin veneers of deposits accumulate on a 20° sloping fanhead in response to movements on a 25° low-angle normal fault. Fan thickness in the Figure 1.7A example is only 12 m, but the width of the newest increment of onlapping fan deposits is 110 m - four times that of the high-angle fault example. Such low-angle faults are sites of thin fans of large areal extent. An emphasis on horizontal instead of vertical displacement also inhibits erosion of deep valleys in the footwall block. These shallow valleys are part of a diagnostic landscape assemblage suggestive of low-angle normal faulting, as are the smooth sloping hillsides that resemble the carapace of a turtle, the "turtlebacks" of Wright et al. (1974). The thinnest deposits near the intersection point (where erosion changes to deposition) are readily removed by fluvial erosion after deposition ceases. Such erosion may have occurred along the west flank of the Panamint Range, and elsewhere in the Death Valley region. Initiation of steep range-bounding faults in the Pleistocene that cut the now inactive low-angle faults (Cichanski, 2000) would stop deposition of the ancestral fans and begin the process of eroding them. The combination of incremental exposure of the plane of a low-angle fault while it is active and subsequent partial stripping of a thin mantle of fan deposits results in spectacular rilled fault planes (Fig. 2.19A). Scrunch deformation is everywhere in hanging-wall blocks of thrust faults. In addition to synthetic and antithetic faulting, scrunch processes include folding, flexural-slip faulting along bedding planes, and shoving of wedges of crumpled, brittle rocks up gently inclined fault planes. Scrunch style tectonics may dominate locally to the extent of raising surface altitudes an order of magnitude faster than regional uplift rates. It makes for pretty messy earth deformation, but adds much variety to rock uplift (Fig. 1.8). The belt of former piedmont terrain between the two thrust-fault zones is called a piedmont foreland, the topic of Section 3.2.3. Bedrock uplift resulting from scrunch tectonic processes increases landscape altitudes and relief of mountains, thus accelerating erosion that partially offsets regional uplift. Scrunch processes may promote lithospheric downwelling opposite in style to the mantle upwell-ing described for tectonic stretching. Deposition in tectonic basins raises altitudes. Scrunching and Scrunch and Stretch Bedrock Uplift 15 Flexural slip folds and faults with displacements along bedding-plane faults Abandoned streamcourse is now a wind gap Deflected stream flows through a water Figure 1.8> Diagrammatic sketch of types of contractional faulting and folding associated with tectonic shortening that causes burial and crustal thickening. Overturned strata may suggest displacement by a normal fault. Displacements along bedding-plane faults occur where planes between beds are relatively weak; note rock flowage into fold axes. Thrust faulting buries the apex of the piedmont alluvial fan, and a younger fault folds the fan surface. deposition thicken the crust, thereby promoting iso-static subsidence that partially offsets concurrent rock uplift. Geothermal gradients become cooler where the crust is thickened from the surface down, and the relatively cooler rocks have a lesser potential for iso-static uplift. Tectonic burial has not received as much attention in the literature as tectonic exhumation, so I use Figures 1.9-1.16 to illustrate the diversity and importance of scrunching. Creation of a fault zone causes more than just uplift, because thrust faults are not vertical. Horizontal rock displacement is a major consequence of scrunching. The hanging wall block is raised as it is shoved up the incline of a gently dipping thrust fault (Fig. 1.9A). The horizontal component of displacement increases local crustal thickness. Amounts of horizontal displacement are a tangent function of fault-plane dip: 100 m of vertical dis- placement is accompanied by only 27 m of horizontal displacement for a 75° dipping fault. This increases to 100 m for 45° and to 373 m for a fault with a 15° dip. Mass is added to the footwall-block terrain by tectonic conveyance and deposition of sediments eroded from the newly raised block as the fault trace advances in an incremental manner. In the best of all worlds, tectonic geomor-phologists would use planar or conical landforms as time lines passing through tectonically deforming landscapes. Dating of faulted alluvial geomorphic surfaces can provide valuable information about late Quaternary uplift rates. However, estimation of tectonic displacement rates of faulted stream terraces probably is more reliable for stretch than for scrunch tectonics. The fan surface upslope from the scarp crest in Figure 1.9B is no longer linear. Its undulations sug- 16 Chapter 1 Footwall 0 400 m Figure 1.9 Tectonic uplift and burial induced by thrust faulting. A. Piagram showing components of uplift and burial created by movement along a thrust fault. Both processes thicken the crust and are functions of fault-plane dip. gest complicated tectonic deformation. A common first impression is that scarp height is indicative of the magnitude of tectonic throw, but scarp height exceeds true displacement where sloping alluvial surfaces are ruptured. A closer approximation can be obtained by noting the vertical separation of projections of the tectonically undeformed fan surfaces upslope and downslope from the fault zone. But this Cucamonga Canyon alluvial fan has a slope that decreases down-fan resulting in lack of parallelism of the projected surfaces. A mean apparent throw of 9.3 m based on maximum and minimum displacements is triple the deformation attributable to scrunching. These apparent displacements need to be corrected for the dip of the faults, which is unknown. A complete discussion is deferred until Section 3.3.4, which describes how to estimate throw for normal-fault scarps on alluvial fans. The interpretation shown in Figure 1.9B is that several synthetic thrust faults ruptured the surface, during several Holocene earthquakes (Morton and Matti, 1987). Another possibility is that the hum-mocky terrain is nothing more than piles of debris near the fault tip that have been bulldozed by thrust faulting along a single thrust fault. Third, compression may have folded the surficial materials. Most likely, the scrunched material resulted from several processes. Holocene bedrock uplift varies from point to point, but approximates the sum of the vertical 720 -P 11.5 m 13-0 m 4 3.3 m 700 21.S m 1 zs < 6&0 660 Apparent vertical tectonic displacements (throw) 1 5carp height 2 Scrunching uplift component 3 Minimum fan surface offset 4 Maximum fan surface offset 5 Mean rock uplift 100 200 Distance in meters Figure 1.9 Tectonic uplift and burial induced by thrust faulting. F3. Inferred thrust faults along cross section based on topographic profile. All estimates of displacements are apparent, and except for scarp height are based on projections of adjacent undeformed alluvial-fan surfaces upslope and downslope from the fault zone. Cucamonga alluvial fan, San Gabriel Mountains, southern California. Figure 1.10 Deformation of the stream terraces of the Waimangarara River caused by recent surface ruptures of the range-bounding Hope thrust fault, Seaward Kaikoura Range, New Zealand. The young T1 stream terrace is strongly backtilted, and is anomalously high when compared with estimates of late Quaternary uplift rates for this mountain front. component of thrust-fault displacement, folding, and other scrunching that results from compressional deformation of the wedge of material above the thrust fault. The magnitude of horizontal displacement determines the amount of tectonic burial. The algebraic sum of these processes equals the changes of surficial altitudes because this young alluvial fan is virtually uneroded. Isostatic adjustments are not likely at this small scale. Tilted stream terraces are sure to catch the attention of the tectonic geomorphologist, especially when alluvium deposited with a 3° downvalley dip now slopes 2° to 5° upvalley (Fig. 1.10). A splay of the Hope fault that bounds the Seaward Kaikoura Range of New Zealand ruptured the Waimangarara River stream terraces. The two oldest, late-Holocene, stream terraces, Tl andT2, have the same backtilt, so the tectonic deformation is younger than the T2 terrace-tread age. Terrace Tl is 5 m above T2. Terrace tread age was estimated with weathering rind analyses, a surface-exposure dating method (Whitehouse and McSaveney, 1983; Whitehouse et al., 1986; Knuepfer, 1988). Analysis of boulders on the T2 tread implies a late Holocene age (Fig. 1.11). This tuffaceous greywacke sandstone does not have nice, sharp weathering rinds, and rind thickness ranges from 1 to 4 mm. I used the McSaveney (1992) procedure. A peak at -2.5 mm dates as 2,200 ± 300 years before present. Even a 4 mm peak would date to only -4,700 years B. P. Terrace T3 is not backtilted but has a fourfold decrease in slope as it approaches the deformed older stream terraces (Fig. 1.10). So it appears that the range-bounding fault ruptured between T2 and T3 time, and again since T3 time. 12 3 4 Kind thickness, mm Figure 1.11 Distribution of Waimangarara River T2 stream terrace tread weathering rinds in cobbles of greywacke sandstone deposited before the older of two recent surface-rupture events. Normal distribution curve has been added. 0.25 mm class interval, n = 40. 18 Chapter 1 Figure 1.12 Geologic structures, landforms, and tecto nica I ly deformed stream terraces at the mouth of the Waimangarara River, Seaward Kaikoura Range, New Zealand as mapped by Jarg Fettinga, University of Canterbury. Scarp height is an impressive 18 m. Vertical offset of 7 m is a minimum value because Tl and T2 have been buried by an alluvial fan downstream from the fault scarp. The prominent graben at the folded scarp crest (Fig. 1.12) can be used to postulate locations of antithetic and synthetic faults above the master thrust fault, which is presumed to dip less than 50° (Van Dissen, 1989). I suspect that neither the large scarp height nor upvalley stream-terrace tilt is indicative of slip rates on this segment of the Hope fault. TheTl fault scarp on the other side of the river is only about 3 to 4 m high, which is a more reasonable offset for two surface-rupture events. Adjacent segments of this mountain front lack high fault scarps that date to the most recent event. Dip and style of faulting may change within short distances, and subsurface exploration techniques are needed here to fully appraise two possible scenarios. The geologic map (Fig. 1.12) portrays a zone of deformation that tapers towards the southwest, seems to be diffuse on the north side, and is abruptly terminated by the range-bounding thrust fault on the south side. A model of imbricate thrust faulting (Fig. 1.13A) can account for the width of the deformation zone, and synthetic and antithetic faults could produce grabens. If one uses the critically tapered wedge model of Davis and Namson, (1994), the scrunching shown in Figure 1.13B reflects a fault-kinematic equilibrium. Wedge shape would influence dip of the basal detachment surface, synthetic and antithetic fault movements, and thickness of scrunched rock and alluvium. So, much of the rock uplift here may be the result of tectonically induced scrunching processes of folding and bulldozing. Brittle fractured greywacke sandstone under low confining pressures may behave like loose boulders. More coherent bedrock slabs may Scrunch and Stretch Bedrock Uplift 19 220 1 200 \ < \b0 . P. • .».»»VI 100 200 300 400 Horizontal distance, m 500 600 160 100 200 300 400 500 Horizontal distance, m 600 Figure 1.13 Models for deformation of the Waimangarara River terraces. A. Fault steepening towards the surface rotates the stream terraces, creating the backtilting of T1 and T2. Grabens at scarp crest record antithetic and synthetic faulting. fail by rupture along secondary faults. The systematic deformation shown in Figures 1.12 and 1.13B could result from folding instead of haphazard bulldozing processes. Scarp-crest grabens would result from tensional stresses at the crest of an anticline in a folding-dominant model. The Waimangarara River has frequent large flow events that deposit bouldery alluvium on the adjacent piedmont. Erosional widening of the bedrock valley floor in the mountain-front reach may have thinned the slab above the range-bounding thrust fault prior to the recent surface rupture events. Reduction of rock mass strength below a critical-tapered-wedge threshold would have favored tectonic scrunching processes in the broad valley floor upstream from the fault trace, but not along the adjacent parts of this steep mountain front. Rock uplift, r , at location 1 in Figure 1.13B is mainly a function of magnitude of slip along the fault plane, D, and dip of the thrust fault, a. ru = sinaD (1.4) Rock uplift at location 2 in Figure 1.13B could be largely bulldozed materials above the plane of the thrust fault where scrunch rock uplift, sr , has occurred at several fault splays. ru = sinaD + sri:(1,2,3,4) (1.5) The longitudinal profile of the Waimangarara River reflects several possible tectonic inputs. The Figure 1.13 Models for deformation of the Waimangarara River terraces. F3. Critical wedge model in which movement along gently dipping thrust fault has bulldozed and/or folded the fractured greywacke sandstone. stream changes its vertical position in the landscape in response to bedrock uplift. However, fluvial adjustments to rock uplift in the longitudinal profile do not distinguish between regional isostatic uplift, slip on thrust faults, folding, and local bulldozer scrunching of fractured greywacke sandstone. I conclude that the Waimangarara River stream terraces are not ideal time lines passing through a tectonically deforming landscape. The deformed stream terrace treads are good reference surfaces for describing the complicated total bedrock uplift, but should not be used for estimating fault slip rates. Thrust-fault displacement has a vertical component, but secondary folding and crushing is largely a function of horizontal displacement. Both contribute to rock uplift. This local increase in crustal loading due to scrunching is too small to overcome lithospheric rigidity, so let us examine an example that is sufficiently weighty to influence isostatic processes. Erosion becomes ever more important with increase in spatial extent and steepness of a landscape, longer time spans, and decrease of rock mass strength. Erosion rates increase exponentially with hillslope steepness (Ahnert, 1970), so relief that is increased by scrunching accelerates the denudational processes that tend to lower a mountain range that is being created by tectonic forces. One impressive example of large scale thrust faulting, and rapid erosion during the past 1.5 My, is the Salt Range in Pakistan (Burbank and Anderson, 2001). Potential tectonic burial by a tectonically translocated mountain-range size block that is 3 km high and 18 km long (Fig. 1.14) never transpired because erosion occurred as rapidly as scrunching raised poorly consolidated fluvial sediments of the Siwalik Formation 20 Chapter 1 North 20 km South Volume of foreland fluvial strata that has been eroded away as fast as soft fluvial sediments have been thrust up the footwall ramp Basin fill deposited as foreland strata are eroded Basin fill alluvial-fan conglomerate Siwalik Formation fluvial molasse deposited in Himalayan foreland basin Strata of Paleozoic to Eocene age Salt of Paleozoic to Eocene age Pre-Cambrian basement rocks Figure 1.14 Potential large scale tectonic burial that has been offset by erosion of the Salt Range, Pakistan. From Burbank and Anderson, 2001, Figure 7.5, and Burbank and Beck. 1991. up the ramp of the footwall block. Note the lack of a tapered wedge of mountain range relief away from the top of the ramp. The Salt Range tectonic setting appears to represent a case where regional fluvial erosion balances the tendency for uplift to increase relief. Estimated rates of erosion are 2 m/ky over the large area of 1500 km2. Tectonic loading has been largely offset by concurrent erosion, which increases crustal thickness elsewhere in depositional basins. Active folding provides extreme examples of scrunch-induced bedrock uplift even where rates of regional uplift are modest. Spatial contrasts in uplift rates generally are gradual for active folds and abrupt for active faults. Horizontal strata under a constant rate of tectonic shortening are folded upward, but the crest of the resulting anticline does not rise at a uniform rate. Rockwell et al. (1988) show that uplift for a single, simple fold quickly accelerates to a maximum, and then slows to zero despite unabated com- pression (Fig. 1.15). About 36% of potential uplift has occurred after only 4% shortening of a horizontal bed, a situation where a slow rate of horizontal displacement causes remarkably rapid uplift of the fold hinge. But there is a limit to how much uplift can be produced by contraction (shortening) of a single fold. Continued scrunching creates faults and new folds. Anticlines in fold and thrust belts commonly have thrust faults in their cores (Fig. 1.8), which further complicates assessment of bedrock uplift. Folding may be largely replaced by tilting after a thrust fault propagates through to the land surface. We should expect the landforms and geomorphic responses to tectonic deformation to vary along the trace of fold created by a propagating thrust fault (Fig. 1.8). Spatial variations of local tectonic deformation should reflect the cumulative displacements of individual earthquakes. Level-line surveys of recent historical earthquakes nicely show the contrast in Fold hinge 40% shortening 16% shortening 4% shortening 0% shortening Figure 1.15 Tectonic uplift and burial induced by contractional folding. Deceleration of rates of folding Induced uplift, using a model of uniform rates of tectonic shortening (from Rockwell et a I., 1 E 0.3 c -40 km -20 0 20 Distance from fault Figure 1.16 Spatial variations in deformation caused by two magnitude M 7.3 earthquakes. The thrust-fault example is the 1952 Kern County, California earthquake. The normal fault example is the 19S>3 Borah Peak, Idaho earthquake. Note that both subsidence and uplift occur with extensional and contractional earthquakes {from Stein, etal., 19&&). styles of folding and faulting associated with normal and thrust faulting (Stein et al., 1988). Such work reveals that some local uplift occurs during a normal-fault surface rupture and some subsidence occurs with a thrust-fault rupture event (Fig. 1.16). Rebound uplift of the footwall block was about 20% that of the hanging-wall block tectonic subsidence result- Scrunch and Stretch Bedrock Uplift 21 ing from the Borah Peak earthquake. Minor faulting and folding is concentrated in the hanging-wall block of normal faults, but the footwall and hanging-wall blocks of an active thrust fault may have similar secondary deformation. Not all tectonic deformation occurs at the moment of an earthquake due to the response time needed for subsequent mantle upwelling. Fault creep and gradual folding may also deform the land surface. Modeling done by Freed and Lin (2002) links tectonic deformation to post-seismic relaxation of viscous lower crust and/or upper mantle - a process that continues for decades. Although some folding or warping during a particular earthquake event occurs as post-seismic deformation, few studies have the data to assess both pre- and post-seismic folding rates over time spans of centuries. One such investigation uses stream terraces formed as a result of downcutting induced by folding and faulting of an anticline. Streams flowing across rising mountains incise bedrock and tectoni-cally induced downcutting is proportional to bedrock uplift rates. Nicol and Campbell (2001) estimated uplift rates for an anticline by measuring the heights of terrace treads above the active channel, and by using weathering-rind and radiocarbon methods to date abandoned fioodplain remnants. The scene is a young fold-and-thrust belt in the foothills of the Southern Alps of New Zealand. The Waipara River has a watershed area of 950 km2 where it cuts though Doctor's Anticline. The Karetu thrust fault in the core of the anticline has broken through to the surface. Regional tectonically induced downcutting has been subtracted from the total tectonic displacement measured in the anticline reach to produce the graph of Figure 1.17. The highest terrace has been raised 23 m relative to the active channel but is only about 600 years old. Downcutting curves for two reaches of the stream channel show that accelerated downcutting occurred between 0.6 and 0.2 ka. The mean local bedrock uplift rate was an astonishing 52 m/ky! This example of extreme scrunching is 50 times the estimated Holocene uplift rate for this tectonic province. Nicol and Campbell also use the terrace ages and heights (Fig. 1.17) to assess the temporal distribution of uplift before and after an earthquake on the Karetu thrust fault that occurred 350 ± 50 years ago. Maximum uplift occurred in that century-long time span. The steep sections of the tectonically induced downcutting plot between 0.6 ka and 0.35 ka and 40 km 22 Chapter 1 Figure 1.17 Variable rates of tectonically Induced stream channel downcutting caused by a late Holocene folding event on a thrust-cored anticline in the foothills of the Southern Alps, New Zealand. The magnitude of tectonic deformation shown here is a minimum because stream terraces are not present at the crest of the anticline. From Figure 12F3 of Nicol and Campbell (2001). Marine terraces provide an opportunity to assess horizontal as well as vertical movements of rocks. Pazzaglia and Brandon (2001) provide an elegant example in their discussion of the tectonic land-forms of the Cascadia forarc high (Fig. 1.18). They examine coastline landforms where the Juan de Fuca plate is converging with the North American plate at 3.6 m/ky at a bearing of 54 °. Shore platform-sea cliff landform couplets generally are created only at times of prolonged sea-level highstands, such as the present. The horizontal and vertical distances between modern and ancient inner edges of marine terraces differ as a function of the horizontal and vertical rates of tectonic displacement. Sea level has also varied in a eustatic (world Distance parallel to plate convergence direction (m) Figure 1.1S> Horizontal and vertical displacement of marine-terrace landforms on a subducting plate boundary. Schematic cross section across the modern shore platform, sea cliff, 122 ka Sangamon shoreline, and a partially buried sea cliff. Mouth of the Queets River, Olympic Mountains of northwestern Washington, USA. Figure 17 3 of Pazzaglia and Brandon (2001). 0 0.2 OA 0.6 03 1.0 ka Terrace age between 0.35 ka and 0.2 ka are inferred to be the result of aseismic folding. These several examples provide interesting food for thought about scrunching that results in vertical tectonic displacements but tell us very little about magnitudes of horizontal earth deformation. Scrunch and Stretch Bedrock Uplift 23 £ 50r Time Before Present, ky Figure 1.19 Fluctuations of global sea level since 330 ka (from Chappell and Shackleton, 1956). wide) sense (Fig. 1.19) due to changing volumes of glacial ice and ocean temperatures (Shackleton, 1987). The Sangamon sea-level highstand at about 124 ka was about 5 m above the present ocean level. This means that even under tectonically inactive conditions the Sangamon shoreline for this gently sloping coast should be 5 m higher and quite far inland. Marine terraces were created at times of globally synchronous sea-level highstands (Chappell, 1983; Chappell and Shackleton, 1986; Lambeck and Chappell, 2001). The -124 ka terrace has been uranium-series disequilibrium dated using coral from New Guinea, New Hebrides, Barbados, Haiti, the Mediterranean Sea, Hawaii, Japan, and California. Bloom et al. (1974) estimated the altitudes of many sea-level highstands. These were brief time spans of unchanging terminal base levels for rivers, much like the past 6 ky. Remnants of shore platforms created at a variety of sea levels indeed are useful time lines passing through tectonically active landscapes. Rapid sea-level changes between the highstands raised and lowered the mouths of streams but this is not the same type of base-level fall as faulting of a streambed (Sections 2.2.4, 2.6). The Figure 1.18 analysis assumes similar geo-morphic processes and alluvium mass strength (gravels) at Sangamon time as compared to the present. Because of the higher eustatic sea level the Sangamon highstand sea cliff would have formed at 945 ± 145 m inland of the modern sea cliffs. Instead it is located an additional 505 ± 150 m farther inland. This suggests a horizontal tectonic displacement of about 450 m during the past 122 ka or a mean horizontal tectonic velocity of 3.7 ± 1.1 m/ky The eustatic component is much smaller than the tectonic component for uplift during the same time span and is simply the change in world-wide sea level. 1.3 Landscape Responses to Regional Uplift Streams incise ever deeper as bedrock is raised into the powerful buzz saw of stream-channel downcutting. Amounts and rates of tectonically induced downcutting are functions of vertical tectonic displacement rates, excess unit stream power, and resistance of earth materials to degradation. Downcutting by small streams flowing over resistant welded tuff may be unable to match a bedrock-uplift rate of 0.1 m/ky; such reaches erode continuously Downcutting by perennial rivers flowing over soft rock easily keeps pace with bedrock uplift of 5 m/ky. But stream-channel downcutting occurs only during appropriate climatic and tectonic conditions. The tendency of streams to cut down to the minimum gradient needed to transport their sediment load has been a long standing fundamental concept in fluvial geomorphology (Powell, 1875; Mackin, 1948; Leopold, Wolman, and Miller, 1964; Leopold and Bull, 1979; Bull, 1991). Headwater reaches of streams in rising mountains tend to stay on the degradational side of the threshold of critical power, but downstream reaches, with their greater unit stream power, are more likely to attain the base level of erosion through the process of tectonically induced downcutting. Gregory and Chase (1994) minimize the influence of base level in a diametrically opposite model. They conclude that Cenozoic canyon cutting in the Front Range of the Rocky Mountains in Colorado resulted entirely from climatic changes that increased stream power. Molnar and England (1990) also favor dominance of climate-change causes of stream-channel downcutting in this same region. The resulting isostatic uplift would promote further canyon down-cutting. Zaprowski et al., (2005) prefer a model where climatic changes would increase the concavity of the longitudinal profiles of rivers crossing the western Great Plains. Greater concavity would require more intense rainfalls and larger, more frequent flood events in the Quaternary than during the Pliocene. However, analysis of gradient changes of rivers flowing east from the Rocky Mountains (Figs. 1.20A, B) reveals that flexural isostatic rebound of the lithosphere due to Cenozoic erosional unloading accounts for only 20% of the concurrent increase of relief (McMillan et al., 2002). Therefore, tectonic uplift is necessary in order to explain the Front Range canyon cutting and the concurrent deepening of val- 24 Chapter 1 leys across the adjacent Great Plains. McMillan et al. conclude that post-depositional changes in slope of the stream channels in the western Great Plains of Wyoming and Nebraska since 18 Ma are the result of broad-wavelength tectonic uplift centered under the Rocky Mountains. Uplift began during deposition of braided-stream gravels of the Miocene Ogallala Formation. Tectonically induced downcutting has continued to lower the active stream channels relative to the strath beneath the basal Ogallala fluvial gravels. It is not easy to discern uplift in landscapes that lack obvious Quaternary faulting and folding. So this book emphasizes tectonic influences on the landscapes of individual watersheds, preferably where tectonic controls are obvious such as active range-bounding faults. McMillan et al. were able to estimate regional tectonic influences on landscape evolution with a combination of paleohydrologic, strati-graphic, and geophysical analyses involving a spatial scale of 250 km and a time span of -15 My. This challenging project produced some interesting results. The post Ogallala time span coincides with the gradual northward extension of the Rio Grande rift from New Mexico; and it seems reasonable that the accompanying regional tectonic uplift also decreased towards the north. Tectonic rock uplift was followed by an episode of erosion-induced isostatic uplift that began when the rivers of the region ceased deposition and began 5 My of fluvial degradation. Leonard (2002) analyzed the larger valleys draining the eastern flank of the Rocky Mountains. Uplift caused tectonically induced downcutting, which promoted isostatic uplift. He assumes that the base of the Ogallala formation was planar and tilted eastward. The Arkansas River valley in southeastern Colorado was eroded to deeper levels than the valley of the North Platte River in southeastern Wyoming. Maximum warping of the Colorado piedmont occurred near the Arkansas River. Leonard's modeling suggests that the isostatic component of rock uplift (Fig. 1.4) accounts for 50% of the total rock uplift with the remainder being tectonic uplift. About 540 m more uplift occurred along the Figure 1.20C transect at the Arkansas River valley than at the valley of the South Platte River. Leonard's results approximate the lesser uplift amounts suggested by McMillan etal., (2002) along the valley of the North Platte River. These thoughts about the diverse character of uplift will be used when we explore uplift of specific mountainous landscapes. Chapters 3 and 4 go into more detail regarding the complications that arise when one attempts to determine how fast the mountains are rising. Figure 1.4 is a rudimentary summary. It hits the main points, but the influences of many of the variables are not easily constrained to the tidy boxes of this simple model. Deep seated crustal flow (Zandt, 2003) is largely ignored. The isostatic component of rock uplift is the algebraic sum of many processes. It is a function of crustal temperature or mineralogy changes, crustal subduc-tion, spreading and flexing, strike-slip fault loading changes, tectonic denudation or burial, faulting and Figure 1.20 Late Cenozoic tectonic and isostatic uplift of the Colorado piedmont east of the Rocky Mountains. A. Trunk channels of the major rivers flowing eastward across the western Great Plains. Rio Grande is in a rift valley that has propagated northward during the Cenozoic. Scrunch and Stretch Bedrock Uplift 25 50 100 150 200 Kilometers east of 105° IV longitude 250 as » in ü as as t o in -§ « zs as < O 2,500 2,000 1,500 1,000 South North .-*** .............. North Canadian Arkansas South Platte Platte Rij/er , , River , .River, .River, 1,000 500 0 U as m Q Q -200 -100 0 100 200 300 400 500 600 Figure 1.20 Late Cenozolc tectonic and isostatic uplift of the Colorado piedmont east of the Rocky Mountains. F3. Post-depositional changes in slope of fluvial gravels in the western Great Plains of Wyoming and Nebraska, USA. The "present slope" is the strath at the base of the 15 Ma Ogallala Group fluvial gravels. The estimate of the original depositional slope - paleosurface slope - during gravel deposition was calculated using an equation (Paola and Mohrig, 1996) relating size of gravel in a braided stream to depth and slope of flow. Cashed line shows the modeled uplift of the Ogallala strath caused by isostatic rebound due to erosion. Change in relief is relative to fixed hinge point at eastern edge of study area. Flexural rigidity used in model is 1024 l\Lm. From Figure 4 of McMillan et al. (2002). C. Polynomial fit of basal Ogallala Formation surface (heavy solid line) based on reference points (solid circles) projected into transect line of Figure 1.20A. Modeled flexure due to erosional unloading (dashed line) used a flexural rigidity of 1024 N-m. Topographic profile shown by thin solid line. Transect line is along 103° 50' W meridian. From Figures 2 and 3 of Leonard (2002). 26 Chapter 1 folding that changes crustal thickness, and the geo-morphic processes of erosion, deposition, solution, and compaction. The bedrock-uplift concept modernizes the ways in which we study tectonics of mountain ranges on active or passive plate margins. Geomorphology now plays a major role in studies of earth history because of the need to understand landscape responses to uplift caused by either tectonic or isostatic uplift. Temporal and spatial scales of study dictate research objectives and procedures. Evaluation of sediment flux from continental landmasses to ocean basins uses different time spans, areas, and geomorphic processes than local erosion and deposition associated with a single-rupture event fault scarp. Conceptual geomorphic models that seem ideal for their formative study area and dataset may become rather tenuous when applied to different spatial, tectonic, and climatic settings. Fortunately, individual drainage basins are the basic components of mountain-range landscapes and even for rivers that flow across continents. This book appraises tectonics and topography at the scale of moderately small watersheds with an emphasis on processes that shape hills and streams. Response times for drainage nets and their adjacent hillslopes are best studied when tectonic perturbations are nearby. Different rock types and climates add spice to landscape studies. These important variables need to be added to the recipe if you want to more fully understand the tectonic geomorphology of mountains. All these aspects are accommodated by the basic theme of fluvial-system behavior. Now that we have outlined the essence of bedrock uplift, let us examine constraints on the limits of how fast and far streams can cut down into landscapes at the watershed scale. The next chapter is about specific processes and landforms that help us understand how active faulting and folding shapes the hills and streams of fluvial systems. I Concepts for Studies of [Rising Mountains Apulse of uplift along a range-bounding fault is transmitted to all parts of a fluvial landscape. How does this occur? Streams are the connecting link between the different parts of watersheds that we treat here as fluvial systems. Is this connecting link equally strong in all humid and arid watersheds? How long does it take for a tectonically steepened reach to migrate upstream to the headwaters? Surely large rivers respond more quickly than small streams. Hillslope-erosion processes don't even start to feel the effects of increased relief until the upstream migrating steeper stream reach arrives at their footslopes. Response times to a surface-rupture event on a range-bounding fault vary greatly with drainage-basin area, climate, and rock type. In contrast, the impacts of a seismic-shaking event are felt quickly throughout modest-size watersheds as changes in sediment yield and mass-movement processes. Let us discuss standard and new ways to study fluvial-system behavior with an emphasis on responses to tectonic perturbations. Chapter 2 concepts will help you evaluate and explore new and diverse approaches in tectonic geomorphology. They are my foundation for understanding interrelations between tectonics and topography Aerial view of Laguna Salada, Mexico. An active normal fault separates landscapes characterized by erosion and deposition. Photograph by Peter L. Kresan ©. 2.1 Themes and Topics Chapter 2 concepts focus on rates and styles of geomorphic processes in diverse tectonic, climatic, and lithologic settings. Continental landscapes of planet Earth are formed in large part by interactions of tectonic and fluvial processes, which are modulated by pervasive late Quaternary climate changes. Tectonics is the study of crustal deformation: the evolution of geologic structures ranging from broad transition zones between crustal plates to small faults and folds. Geomorphology is the study of landscapes and the processes that shape them. The influences of vertical and horizontal earth deformation on fluvial, coastal, and glacial processes and the resulting landscapes comprise the domain of tectonic geomorphology. Many processes shape the surface of planet Earth, but the action of running water is responsible for most subaerial landscapes. This book is about fluvial systems—hilly to mountainous source areas that supply water and sediment to streams, which convey their load to depositional basins. Sustained uplift along active faults and folds may create mountain-front escarpments. Tectonically active mountain fronts appeal to tectonic geomorphologists, because uplift steepens stream gradients, which accelerates watershed (synonymous with the term drainage basin) erosion by making hillslopes steeper. Climatic changes may create landforms that approximate time lines passing through tectonically 28 Chapter 2 deforming landscapes. Examples include shoreline marine terraces from polar to tropical realms, and abandoned flood plains rising like flights of stairs above rivers. Both tectonic landforms are vital sources of information about local uplift history and regional isostatic adjustments. Tectonic geomorphology has two facets - basic research to better understand landscape evolution and practical applications. Basic research is applied to define potential hazards posed by active tectonics processes (Hecker, 1993) and to diminish risk to people and engineering structures. Engineers and planners need knowledge gained from landscape studies (Fig. 2.1) in order to predict earthquake hazards (frequency and magnitude of surface ruptures, seismic shaking, and coseismic tsunamis, landsliding ,and flooding) to minimize earthquake risk (loss of life and property). Chapter 1 discussed several processes that raise and lower land surfaces. This chapter examines how far streams can cut down into bedrock - their base level limit as they erode mountain ranges into drainage basins. We also explore fluvial-system behavior to lithologic and climatic controls in different tectonic settings in the context of response times, the threshold of critical power, and tectonically induced down-cutting. These concepts will give you the necessary foundation for perceiving the nuances of how hills and streams respond to mountain-building forces. 2.2 The Fundamental Control of Base Level 2.2.1 Base Level Studies of tectonics and topography use base level as a reference datum for rivers. John Wesley Powell (1875, p. 203-204) introduced the term base level as the altitude below which a stream cannot down-cut. The ocean is regarded as a general, or ultimate, base level even though Quaternary sea levels have fluctuated 130 m (Chappell and Shackleton, 1986: Chappell, 2001). Fluvial processes cease where rivers flow into lakes or the ocean, because the hydraulic gradient is reduced to zero and potential energy is not further transformed into kinetic energy. The concave longitudinal stream profile that typically develops upstream from a base level reflects adjustments between hydraulic factors that G.K. Gilbert referred to as "an equilibrium of action" (1877). Anomalously steep and narrow reaches may reflect lithologic controls on a longitudinal profile (Kirby et al., 2003). Many local base levels occur between the headwaters of a stream and the terminus in an ocean, lake, or basin of internal drainage (Fig. 2.2). Resistant outcrops are considered local base levels because a stream is unable to lower its bed as easily as through relatively softer materials in adjacent upstream and downstream reaches. Downcutting promoted by uplift is reduced or delayed (sometimes greatly; see Sections 2.5.1, 2.5.2) where resistant outcrops create relatively stable reaches. Such local base levels are temporary compared to the relative permanence of the oceans. Mean sea level may change slowly but short-term fluctuations of sea level are as much as 130 m in 15 ky. Downstream parts of fluvial systems at 20 ka now are deep under the sea. Alluvial reaches of streams may be regarded as an even more temporary category of local base levels. Indeed, each point along a stream, be it underlain by rock or alluvium, is part of a continuum of streambed altitudes. Each short reach of a stream exerts a base-level control on adjacent reaches that partly determines the longitudinal profile and stream-channel patterns such as meandering and braided. Base levels for adjacent reaches of a stream can be raised or lowered. The base-level processes of aggradation and degradation may be caused by either tectonic or climatic perturbations, or they can result from internal adjustments initiated by changes in the hills and streams of a fluvial system. The exciting challenge for the tectonic geomorphologist is to recognize and interpret key features of tectonically controlled aggradation and degradation within a fluvial system that provide clues about styles and rates of earth deformation. 2.2.2 Base-Level Change The spatial consequences of base-level change emanate both upstream and downstream from a base-level perturbation. Base-level fall, such as tectonic lowering of a streambed downstream from a fault or fold axis, is readily transmitted upstream by creation of a short reach of increased gradient. This local increase in stream power tends to initiate degradation. The degrading reach propagates upstream as headcuts, waterfalls, and rapids that become smaller as the perturbation migrates away from its tectonic origin. An important consequence of accelerated stream-channel downcutting is the increase of sediment yield caused by undercutting that steepens adjacent hillslopes. The result is an increase in bedload transport rate Concepts for Studies of Rising Mountains Controlling Variables Climate and types of climatic change Lithology and structure Time Vertical earth deformation Process - Response Modele Hillslopes Streams Piedmonts Evaluation of Tectonic Landforme 'Fault scarps Cross-valley profiles Triangular facets Mountain-piedmont junctions Longitudinal-valley profiles Stream terraces Alluvial fans Pediments Tectonic Inferences Spatial patterns of earth deformation Amounts of Quaternary tectonic deformation Horizontal slip rates Vertical slip rates Seismic moment and moment rate Earthquake recurrence Interval Practical Applications Regional planning Building and zoning codes Local land uses Hazard maps for surface rupture and seismic shaking Earthquake hazards reduction Figure 2.1 Flow-chart checklist for tectonic-geomorphology studies. 30 Chapter 2 Terrace tread is a stable base level for adjacent hillslope Degrading stream channel is a ^t^a'^a-.g lling base level pill k Piedmont reach Mountains reach Local base leve^ Stream-channel downcutting is a base-level fal □I Bedrock Figure 2.2 Reaches of stable, falling, and rising base level along a hypothetical fluvial system in an arid region. Mountains are being raised relative to the basin on a normal fault so the stream is degrading. A. Cross-valley topographic profile. F3. Longitudinal stream topographic profile. Waterfall knickpoint is a local base level that separates stream-channel reaches and hillslopes with different characteristics. that tends to offset the effects of an anomaly created by a tectonically steepened gradient. Base-level fall affects reaches that are downstream from a tectonic perturbation much differently than upstream reaches. Active range-bounding normal faults typically separate eroding mountains from aggrading basins. An example from Mexico is shown in the photo on the title page for this chapter. A base-level fall that accelerates degradation upstream from an active fault also creates the space for accumulation of thick basin fill downstream from the fault. The effects of base-level rise seem to be minimal in a variety of settings. An aggrading playa lake bed affects only short terminal reaches of streams. The effects of base-level rise are not transmitted nearly as far upstream as those of a base-level fall. Construction of a dam across a stream is a local base-level rise that is propagated upstream from the reservoir (Leopold and Bull, 1979; Leopold, 1992; Gellis et al., 2005). The longitudinal profile of the newly created valley fill would be the same as before the base-level rise if equal thicknesses of new alluvium extended to the headwaters. This does not occur. Instead, a wedge of alluvium is deposited that extends only as far upstream as needed to maintain the slope at 50 to 70% of its original value. Other variables, such as hydraulic roughness, change concurrently as gradient is decreased. The affected reach reestablishes a new unchanging (equilibrium) configuration through a different set of interactions between variables. The sediment wedge does not migrate farther upstream. So, reaches upstream from the small wedge of new valley fill are not affected by the base-level rise. One important conclusion is that aggradation of river valleys over long reaches is primarily the consequence of climatic-change impacts over the hillslopes of an entire watershed. Climatic perturbations influence bedload supply from hills and transport capacity of streams in ways that maintain the process of deposition for the duration of an aggradation event. Propagation of the effects of a local base-level rise in downstream reaches is minor. Local aggradation along a mountain stream creates a patch of alluvium consisting of a gentler upstream reach and steeper downstream reach (Fig. 2.3). The locations where the stream changes its mode of operation from degradation to aggradation (or vice-versa) are Figure 2.3 Longitudinal profile of a valley floor showing adjacent alluvial reaches that are moregentle (X) and steeper (Y) than the slope of the bedrock channel prior to temporary deposition of an alluvial channel fan. Threshold-intersection points are shown by TIF! Concepts for Studies threshold-intersection points. The steepened reach is inherently unstable. Increase of stream power favors entrenchment, leading to the development of an incised channel that concentrates flow, thereby initiating a self-enhancing feedback mechanism that favors removal of the patch of alluvium. Patches of alluvium deposited in active channels tend to be temporary. Localized brief episodes of accelerated deposition also occur in depositional environments such as deltas and alluvial fans where sediment tends to subsequently be redistributed. Aggradation upstream from lakes and the ocean is different because it occurs in the terminal reach of a stream. 2.2.3 The Base Level of Erosion I now define and use a valuable equilibrium concept that describes adjustments between the hierarchy of adjacent reaches in a drainage net. The base level of erosion is the equilibrium longitudinal profile below which a stream is unable to degrade and at which neither net vertical erosion nor deposition occurs (Powell, 1875, p. 203-204; Barrell, 1917). A stream cannot permanently degrade below its base level of erosion and maintain the gradient needed to transport its bedload. A reach of stream at the base level of erosion has achieved a time-independent configuration of its longitudinal profile that is maintained as long as the controlling variables do not change in an average sense over suitably long time spans. This notion integrates the system, equilibrium, and base-level concepts. It considers the longitudinal profile spatially as being an infinite sequence of adjacent base levels (Gilbert, 1877), and temporally as capable of being reestablished at multiple positions within a rising landscape. Changes in any of the variables affecting stream power or resisting power (terms are defined in Figure 2.13) may change the base level of erosion. The importance of base level in studies of landscape evolution can be illustrated by the simple example of a stream flowing into a lake, or the ocean. The terminal reach of the stream cannot erode below the lake level because it will no longer have sufficient gradient (stream power) to transport the bedload supplied to it by upstream reaches. So the reach can only cut down to the minimum gradient needed to convey its sediment load with the prevailing stream discharge. Of course, both of these dependent variables of fluvial systems changed whenever late Quaternary climate changed. of Rising Mountains 31 An upstream migration of the base level of erosion begins at the lake base level. Assume that next reach upstream from the terminal reach is steeper than is needed to convey the bedload and that resistance of rocks to erosion remains constant. The excess stream power will degrade the streambed until it too reaches the base level of erosion. Unit stream power (discharge per unit width of streamflow width) becomes progressively less in consecutive upstream reaches, so more time is needed to downcut to the base level of erosion. This attainment of equilibrium conditions progresses spatially at exponentially slower rates because stream discharge decreases upstream in an exponential manner. Attainment of the base level of erosion in a sequence of upstream reaches of a stream initiates spatially consecutive pulses of landscape change that migrate up the adjacent valley sides. Bedrock reaches in the headwaters of a watershed never achieve the base level of erosion. Matmon et al. (2003) came to the same conclusion for the southern Appalachians where they evaluated bedrock-erosion rates based on 10Be analyses of river sediment samples. "... it appears that Hacks dynamic equilibrium might never be achieved at the scale of headwater streams" of <50 km2. This is because equilibrium longitudinal profiles are never attained in the headwater extremities of the drainage basin. The time needed to achieve the base level of erosion stretches to infinity for the mere trickles of headwater streamflow. Rock mass strength becomes progressively more important upstream. Eaton and Church (2004) used a stream table to examine the relation between equilibrium stream channel morphology and discharge, bedload supply, and valley slope. Streambed and channel morphology changes were minor, except for channel slope. "... the system tends to move toward the minimum slope capable of transporting the sediment supply". Attainment of the stream channel base level of erosion affects hillslopes too. A valley floor at the base level of erosion is a stable base level for the adjacent hillslopes in much the same way as a lake is the base level for the terminal reach of the stream. Hillsides adjacent to stable valley floors typically have three segments. The footslope is the concave base of the hillslope. It is a surface of detrital conveyance where geomorphic processes cannot further lower a hillslope graded to broad valley floor. The crestslope is the convex erosional topographic profile descending from the ridgecrest. The midslope 32 Chap is where hillslope topography changes from convex to concave and may be long and straight. Concave footslopes are most common in downstream reaches of a watershed and midslopes and footslopes may not be present in upstream reaches if rapid stream-channel downcutting generates hillslopes that are convex from ridgecrest to valley floor. The base level of erosion concept is important because it defines equilibrium conditions that favor beveling of broad bedrock valley floors (Hancock and Anderson, 2002). Previous valley floors may be preserved above an active channel as flights of stream terraces when valley-floor downcutting by the stream is renewed at times of excess stream power. Several types of stream terraces are defined in Section 2.4.1. The point here is that straths are surfaces beveled by streams while at the base level of erosion. Return to similar climatic conditions promotes development of parallel strath terraces. Vertical separations between successive straths are a measure of stream-channel downcutting induced by bedrock uplift. Each longitudinal terrace profile represents a similar base level of erosion that has been raised tectonically relative to the active channel. A reach of a stream at the base level of erosion is easily identified in the field as a surface of detrital sediment transport. Bedload is moved across the beveled bedrock surface leaving a veneer of gravel on top of a roughly planar bedrock surface. Repeated aggradation events during the Pleistocene were followed by stream-channel down- ter2 cutting. The past 10 to 15 ky of net degradation has allowed even small streams of humid regions to catch up with their base level of erosion. This is why the present is a time of strath formation for many streams, for example the Ventura River (Rockwell et al., 1984) of southern California and the Reno River of Italy (Fig. 2.4). This position remains the same in tectonically stable landscapes (Fig. 2.5), but will be at progressively lower positions in landscapes of rising mountain ranges, including both of the above examples. The time needed to achieve the base level of erosion is a function of 1) available unit stream power in excess of that needed to transport sediment and overcome hydraulic roughness, and 2) the resistance to erosion of materials beneath the streambed. Duvall et al. (2004) analyze the mutual influence of uplift rate and rock type on longitudinal profiles of streams in the California Coast Ranges. Streams downcut and then bevel straths quickly in reaches underlain by soft materials such as unconsolidated mudstone. Streams of tectonically inactive regions also achieve the base level of erosion but flights of strath terraces are not the typical landform of these fluvial systems. Valley-floors may be buried instead of being raised above the active channel and preserved as strath terraces. Note the similar depths of incision shown in Figure 2.5. Even the depth of incision of the late Pleistocene valley was at the same base level of erosion in this tectonically inactive setting, despite changing streamflow characteristics as a result Figure 2.4 Strath forming under the active channel of the Reno River at Marzabotto, northeastern Italy. The 4,597 km2 watershed is characterized by frequent large floods. Large unit stream power allowed the river to remain at the base level of erosion during the late Holo-cene. Lack of gravel in the active channel may in part be the result of prolonged mining of the streambed for concrete aggregate. Photo provided courtesy of Frank Pazzaglia. Concepts for Studies of Rising Mountains 33 13 ka 6 ka Aka 13 ka >20 ka Figure 2.5 Cross section at Curry Praw, southeastern Arizona showing three Holocene episodes of arroyo cutting and backfilling since deposition of the Coro Marl at 13 ka. Vertical hachures indicate the relative ages of soil-profiles and the numbered arrows date times of maximum channel down-cutting for three prehistoric valley-floor degradation events. Note the similar 4 to 4.6 m depths of the modern, 4 ka, and 6 ka arroyos. Similar levels of backfilling were attained by the intervening aggradation events. Stratigraphic cross section is from Haynes (19&7). Figure 44 of Bull (1997). of Pleistocene-Holocene climatic change. Stream behavior changed after the Pleistocene-Holocene transition, and many streams in the American southwest underwent repeated episodes of arroyo cutting (Bull, 1997). This unchanging base level of erosion clearly reveals lack of significant uplift during the late Quaternary Neither isostatic nor tectonic uplift occurred here so former valley floors were not isolated by the process of tectonically induced downcutting. The base level of erosion helps us understand why stream channels, although occupying only a miniscule percentage of mountain-range area, are so important in controlling bedrock erosion rates. The base level of erosion for each consecutive reach restrains degradation of the adjacent hillslopes. Consequently the regional drainage net controls the tectonic geomorphology of mountainous landscapes. 2.2.4 The Changing Level of the Sea We use sea level as the ultimate base level for specific rivers and as the reference framework for altitudi-nal positions within fluvial landscapes. Rotation of the Earth causes the radius at the poles to be about 21 km shorter than at the equator, and gravity varies on the surface of this oblate spheroid because of spatial variations in the mass of Earths crust and mantle. Earths mean sea level, although perfectly level at each point, has minor bumps and hollows as large as 106 m (Heirtzler and Frawley, 1994). So the geoidis that particular equipotential surface of the Earths gravity field which best fits, in a least squares sense, global mean sea level. In contrast to minimal rates of non-climatic change of this level surface, the altitude of a river mouth varies greatly over short time spans as sea level rises and falls. A key question concerning the importance of the ocean as a base level is whether or not entrenchment of coastal streams occurred during times of declining sea level as continental glaciers expanded. The answer lies in the slope of the new reach of a river that is created as sea-level decline exposes part of the continental shelf. Assume that no changes occur in discharge of water and sediment from the upstream reach. The continental shelf for the first 100 m below present sea level commonly is gentler than the slope of the adjacent upstream reach (Fig. 2.6A). This gentler new terminal reach would tend to aggrade because of a decrease in stream power (lesser slope). The opposite situation would occur where 100 m of sea-level decline dropped a river into the head of a submarine canyon, or onto the steeper foreset part of a delta. The new terminal reach would be steeper than the adjacent upstream reach and would tend to degrade (Fig. 2.6B). In contrast to the aggradation case, such channel incision is a base-level fall that can be propagated far upstream. The depth to the edge of the continental shelf and its distance offshore play important roles in the responses of rivers to decline in sea level during the late Reach 1 Reach 2 3 Present sea level Resell 1 Reach 2 200:1 vertical twajgs ration Range of sea level since 140 ka 80 120 Distance, km Figure 2.6 Changes in gradients of terminal reaches of coastal rivers caused by sea-level fall or rise. A. diagrammatic sketch of 100 m of sea-level fall that decreases fluvial hydraulic gradient, decrease of gradient in reach 2 favors aggradation. This model of effective base-level rise also applies to basins of internal drainage where lakes become playas when the climate changes. F3. diagrammatic sketch of 100 m of sea-level fall that increases fluvial hydraulic gradient. Increase of gradient in reach 2 favors stream-channel entrenchment of both reaches when sea level falls. C. Topographic and bathymetric profile of the Colorado River of Texas where it enters the Gulf of Mexico. From Figure 1 of Tailing (199S>). See Figure 1.19 for sea-level curve since 140 ka. Chapter 2 Quaternary. Tailing (1998, Fig. 1) shows topographic and bathymetric profiles for river mouths in a variety of plate-tectonic settings. An example from a passive margin setting illustrates some key points (Fig. 2.6C). About 40 m of sea-level decline would extend the mouth of the Colorado River off the present coastline of Texas by almost 100 km but the gradient would remain essentially unchanged. Aggradation or degradation of this reach would therefore be in response to changes in discharge of water and/or sediment load. The major sea-level decline of the latest Quaternary represented a lowering of base level of about 130 m below the present level of the sea. This resulted in sufficient extension of the mouth of the Colorado River so that it dropped over the outer edge of the continental shelf. The result was a major base-level fall that propagated far upstream (Blum and Valastro, 1994). The Mississippi River also drains into the Gulf of Mexico but underwent a different response to the same sea-level fall. Tailing (1998) notes the 200 times greater rate of sediment delivery to the mouth of the Mississippi River as compared to the Colorado river. The result is that the Mississippi River delta has been built out to the edge of the continental shelf, so even moderate sea-level fall substantially steepens the terminal reach of the river. But stream-channel entrenchment does not necessarily result from this base-level fall. Times of maximum sea-level decline also were times of large increases in sediment load in the drainage basins of many mid-latitude rivers as a result of increased glacial erosion and periglacial processes on hillslopes. Thus it is common for fluvial aggradation of some river valleys to coincide with times of low sea levels (Bull, 1991, Section 2.1.3). The great influx of bedload may exceed the concurrent increase in stream power where a falling sea level steepens the terminal reach of a river. Deposition would result. Postglacial alluviation resulting from Pleistocene glacier retreat is one of several paraglacial processes (Ballantyne, 2002 a,b). Large rivers may entrench and backfill their terminal reaches. In studies of the lower Mississippi River, Fisk described 100 m of late Quaternary channel entrenchment (Fisk, 1944, 1947; Saucier, 1994, 1996; Albertson and Patrick, 1996). Degradation was in part due to increased stream power as sea-level decline exposed the foreset beds of the delta and created a steep terminal reach. Entrenchment propagated upstream, but was followed by aggradation when influx of glacio-fluvial detritus overwhelmed 35 Figure 2.7 Shore platform at Kaikoura Peninsula, New Zealand beveled during the past 6 ky of stable sea level. Pocks are folded and faulted mid-Tertiary mudstone and limestone. Waves keep the platform swept clean of littoral deposits. Two persons near base of sea cliff for scale. the effects of coastal base-level fall and backfilled the entrenched channel. The terminal reach of the river became progressively gentler during the ensuing sea-level rise, which promoted further deposition that continues presently. Streams do not permanently degrade below their ocean base level. It is unlikely that a river will scour 80 m below sea level even temporarily during major floods if a terminal reach has an altitude of only 20 m. However changing the total relief from 10 to 120 m, by 100 m of sea-level decline, might allow 80 m of stream-channel entrenchment because of diminished constraint of the ocean as a "permanent" base level. Such entrenchment could occur only when discharges of sediment and water favored a degradational mode of operation - for example, during an increase in water discharge and a decrease in bedload transport rate. Two conclusions emerge. First, sea-level position constrains the amounts of possible channel downcutting. Second, changes in sea level may not be as important as changes in streamflow characteristics in determining aggradation and degradation modes of stream operation in terminal reaches of rivers. Marine base-level controls also include lateral erosion into sea cliffs during sea-level highstands to create shore platforms. The sea has remained at about the same level during the past 6 ka (Lajoie, 1986), which has favored retreat of rocky coastlines (Fig. 2.7) that would not undergo much erosion if only briefly exposed to the force of waves. Consider the interactions of a stable sea-level highstand and rivers along two steep coastlines with hills of soft rock; one coastline is tectonically stable (Fig. 2.8A) and the other is rising (Fig. 2.8B). With no uplift and a stable sea level for 6 ky, shore platforms become wider as sea cliffs retreat. The rate of widening of the shore platform decreases with time, because wider platforms are more effective in dissipating wave energy. Shore-platform widening causes a lateral erosion induced base-level fall where streams enter the ocean. This base-level fall can be large for steep streams. The river cuts down to a new equilibrium profile, similar to the former longitudinal profile, as the consecutive base-level falls migrate upstream. The new profile is offset horizontally by a distance equal to the amount of sea-cliff retreat, and is lower than the profile before the sea-level rise by an amount equal to the lateral erosion induced base-level fall (Fig. 2.8A). Much different interactions prevail for a rising coastline and a stable sea level (Fig. 2.8B). Sea-cliff retreat could have occurred during all of the past 6 ka for the stable coastline (Fig. 2.8A), but only for about 1 to 3 ky for a coastline rising at more than about 1 ml ky. Furthermore, much of this brief period of coastal erosion would have occurred prior to attainment of the 6 ka sea-level highstand. Optimal conditions for shore-platform development occur when the rates of sea-level rise and uplift are similar. Assume that the shore platform began to form at 8 ka in the Figure 2.8B example. Coastline retreat, by 5 ka, would be 36 Chap much less than for the tectonically inactive coast, but other processes create wide shore platforms. Rock that is raised into the surf zone is beveled too. The inner edge of the shore platform coincided with the river mouth at 8 ka in the hypothetical situation portrayed in Figure 2.8B. Then 8 ky of uplift raises the longitudinal profile of the river to a higher position. The river continues to cut down in response to a base-level fall that is the sum of the uplift and lateral erosion induced base-level fall. Erosiveness by a variety of shore platform processes (Kirk, 1977: Stephenson and Kirk, 1998, 2000a, 2000b, 2001) and erodibility of the rock type largely determine whether or not shore-platform degradation equals bedrock-uplift rate. The platform enlarges mainly in a seaward direction as rock is raised into the surf zone after initial sea-cliff retreat at the beginning of a sea-level highstand. Wave abra- ter 2 sion may be more important on the outer edge than on the inner edge where salt weathering and biota are key factors of bedrock erosion. Shore-platform degradation has kept pace with bedrock uplift along parts of the New Zealand coast (Fig. 2.9). Landscape evolution at the mouths of rivers also is a function of available stream power. Small streams may be unable to achieve the base level of erosion. Downcutting in response to uplift of their coastal reaches steepens their longitudinal profiles. Such streams undergo a net increase in altitude while continuing to maintain a concave longitudinal profile. Powerful rivers generally have rates of down-cutting that equal bedrock-uplift rates. They tend to remain at the same position, relative to sea level, and have similar longitudinal profiles throughout a sea-level highstand. The position of the longitudinal profile in the landscape is progressively lower 40 m wide shore platform 100 m wide shore platform Figure 2.5 Influences of a 6 ky of stable sea level on the longitudinal profiles of powerful rivers. A. Lateral erosion induced base-level fall along a tectonically stable coast. F3. Elevation of a 6 ka stream profile along a rapidly rising coast. Concepts for Studies of Rising Mountains 37 Figure 2.9 Shore platform at Kaikoura Peninsula, New Zealand that has not been widened by trimming of the sea cliff since a mid-Holocene coseismic uplift event. Late-Holocene increase of shore-platform width is the result of ~0.5-1.0 m/ky of rock uplift that has raised soft rocks into the surf zone. than before the sea-level rise by an amount equal to the sum of the lateral erosion induced base-level fall, plus the amount of bedrock uplift. Coastal erosion base-level fall creates mountain fronts with characteristics that are similar to faulted mountain fronts (Fig. 2.10). Lateral erosion induced base-level fall affects hillslopes as much as river valleys. Larger streams of the Seaward Kaikoura Range in New Zealand deliver so much sediment to the coast that deltas are deposited. Such depositional base-level rise keeps the waves away from the mountain front. Apparent lateral erosion can also be the result of tectonic horizontal displacements. Pazzaglia and Figure 2.10 Truncated coastline where the Seaward Kaikoura Range, New Zealand has an abrupt, straight escarpment, deeply incised V-shaped canyons, and triangular facets. Brandon (2001) describe how a shoreline that was tectonically translated towards the ocean results in apparent uplift of marine terraces. 2.2.5 Spatial Decay of the Effects of Local Base-Level Changes Exponential equations may be used to describe either the curving longitudinal profiles of streams, or the decreasing influence of a pulse of uplift generated by a range-bounding fault with increasing distance upstream. Hack (1957) defined the exponential nature of longitudinal profiles of streams. Morisawa (1968) described the relation between decrease in altitude, H, and distance in the downstream direction, L, as H = be- (2.1) where b and m are constants. The effects of vertical tectonic displacement of a streambed decrease exponentially upstream and downstream from the surface rupture. If PQ is the magnitude of the base-level fall (Fig. 2.11), and P is the magnitude of change in streambed altitude at distance L from the fault, then: P=P eKb, L O (2.2) 38 Chapter 2 Longitudinal profile of stream before base-level fall P, P .- Longitudinal profiles of stream after initial base-level fall ._ ^ Alluvial fan Figure 2.11 Decreasing effects of a local base-level fall upstream and downstream from a hypothetical lateral erosion induced base-level fall. Po is the magnitude of the perturbation, and P is the effect of the base-level perturbation atan upstream distance L Equation 2.2 describes spatial adjustments by the processes of tectonically induced downcutting or aggradation. The base-level-reaction constant Kbl is InP -InP, Kb= —^ (2.3) /C is also a function of discharge, structure and lithology, and climate and vegetation-all of which influence the rate of tectonically induced downcutting. Values of Kbl are small because the numerator of equation 2.3 is divided by the horizontal distance from the tectonic perturbation. Strath terrace longitudinal profiles commonly are parallel. Values of Kbl approach 0.0 for pairs of terraces where minimal net change in the longitudinal profile has occurred after re-establishment of equilibrium conditions. Maximum values of Kbj occur where vertical tectonic displacement is large relative to distance from the active fault zone, and where stream power or time has been insufficient for the stream to return to a longitudinal profile similar to that present before vertical tectonic deformation. When used in this context, /C describes the degree of departure, from pre-base-level fall conditions. Both erosion and deposition in fluvial systems respond to elevation by mountain-building forces. In 1887 A.D., a small ephemeral stream crossing the Pitaycachi fault in northeastern Sonora, Mexico was displaced 3 m vertically by a Magnitude 7 earthquake (Bull and Pearthree, 1988). A major event, this is the longest known surface rupture by a normal fault. This large tectonic perturbation greatly affected fluvial processes in the reaches upstream and downstream from the surface rupture. The soil profile on a pre- 1887 alluvial fan downstream from the fault zone has characteristics that correlate with the soil profile on an alluvial fill terrace upstream from the fault zone. Both were part of a continuous alluvial surface (Fig. 2.12A) deposited during a late Pleistocene aggradation event. Minor entrenchment of the stream channel terminated the aggradation event and allowed the distinctive late Pleistocene soil to form on the stable terrace tread. A 3 m offset in 1887 of the active channel and the Pleistocene fill terrace caused accelerated stream-channel entrenchment upstream and alluvial-fan deposition downstream from the fault. Deposition of a new alluvial fan began, not adjacent to the fault, but 30 m downstream. This base-level rise caused partial backfilling of the stream channel entrenched into the fill terrace. In 1987, the first 16 m downstream from the fault trace was still entrenched, but the mode of operation had switched to aggradation in the next 14 m of the partially backfilled fanhead trench. Equation 2.2 was used to describe the spatial changes in channel depth below the late Pleistocene fill terrace tread in a 12.4 m long reach upstream from the fault. The rupture plane is exposed in a 0.3 m high dry waterfall in weathered quartz monzonite. The amounts of tectonically induced downcutting decrease in an exponential manner with increasing distance upstream from the fault. The value of Kbj for this reach is 0.106. A regression analysis of the effects of the 1887 surface rupture (Fig. 2.12B) predicts a stream-channel depth at the fault of 2.82 m; the actual value is 2.74 m. Tectonically induced aggradation of the downstream reach of the fanhead trench also can be described by equation 2.2 (Fig. 2.12C, D). The Concepts for Studies of Rising Mountains 39 '••/|v Post 1ÖÖ7 alluvial far 10 0.1 3 y = 2.ßx10-a050x : 1 R = 0.99 1 5 10 Distance upstream from fault, m 15 5 10 15 Distance upstream from fan apex, m 5 10 15 Distance downstream from fault, m Figure 2.12 Spatial decay of an uplift perturbation, for a 100-year time span, along a stream that was displaced 3 m vertically in 1557 by rupture along the Pitay-cachi normal fault, Sonora, Mexico. A. Diagrammatic sketch of a small Pitay-cachi fault study area. B. Decreasing amounts of tectonically induced downcutting in the reach upstream from the fault. C. Decreasing amounts of tectonically induced aggradation in the reach upstream from the apex of an alluvial fan that is 30 m downstream from the fault because of pre-1557 stream-channel downcutting. D. Decreasing amounts of tectonically induced downcutting in reach downstream from the fault. 1887 uplift event increased the relief and sediment yield of the small watershed. Stream power increased because of the steeper gradient, but resisting power increased even more (terms are defined next in Figure 2.13). Aggradation in the fanhead trench reach is the combined result of increased watershed sediment yield and base-level rise caused by alluvial-fan deposition. The resulting post-1887 tectonically induced aggradation was described by measuring thicknesses of modern gray alluvium above reddish brown weathered Pleistocene deposits along the 14 m of the partially backfilled entrenched stream channel. Spatial changes in thickness of aggradation are systematic (Fig. 2.12C). The Kbl value is 0.124. Channel depth decreases systematically in the entrenched stream channel (Fig. 2.12D); Ku is 0.224. We continue by introducing the topics of stream power, resisting power, and thresholds in geo-morphic processes. These concepts provide a base to examine topographic profiles of stream channels that have attained equilibrium configurations. 2.3 Threshold of Critical Power in Streams Altitude and topographic relief increase where bedrock uplift exceeds surficial erosion (equation 1.3). Bedrock uplift increases valley-floor slope, and thereby the stream power available to transport bed- 40 Chapter 2 load and to erode valley floors. Bedrock uplift is an independent variable of fluvial systems where it is not affected by other geomorphic variables. Slopes of hillsides and streambeds are dependent variables because they are affected not only by tectonic processes, but also by independent variables such as climate and rock type, and by dependent variables such as soils, plants, and erosional processes. Perhaps we should use the less restrictive term "controlling" instead of "independent" for time spans of more than 1 My. For example, uplift of the Himalayas during the past 35 My has controlled climate on a global scale (Raymo and Ruddiman, 1992). Lithologic control of erosional landforms is associated with spatial variations in structure, fabric, and mineralogy of rocks (Weissel, and Seidl, 1997). The time span being considered influences how we classify variables (Schumm and Lichty, 1965). Climate should be considered as an independent variable for short time spans, but over long time spans orographic influences of bedrock uplift make watershed climate dependent on bedrock uplift. Climatic control of fluvial systems is so pervasive that a companion book is devoted to the fascinating subject of Geomorphic Responses to Climatic Change (Bull, 1991). Changes to plant communities are central to climate-change impacts (Vandenberghe, 2003; Flenley and Bush, 2006). Late Quaternary climatic changes commonly overwhelmed the effects of concurrent bedrock uplift by abruptly changing the amounts of water and sediment supplied to streams. As a tectonic geomorphologist, you need to separate tectonic from climatic influences on landscape evolution. Hemisphere-scale climate variations affect both styles and rates of erosion of mountain ranges, such as the Andes (Montgomery, 2002). INCREASING RESISTING POWER INCREASING STREAM POWER Figure 2.13 Schematic balance between modes of aggradation and degradation in streams; zero is the threshold of critical power. Increases or decreases of one or more variables may cause the mode of stream operation to depart markedly from a threshold condition. (Originally from notes of E.W. Lane; modified from Chorley et al., 19S>4 and Bull, 1991.) Concepts for Studies Streamflow behavior may be regarded as a delicate balance between several controlling factors -stream slope and discharge, sediment size and amount, and hydraulic roughness (Fig. 2.13). The threshold of critical power separates two disequilibrium modes of operation in streams, degradation and aggradation. This threshold is defined as a ratio (Equation 2.4) where the numerator consists of those variables that if increased tilt the balance in favor of degradation, the mode of operation that lowers the altitude of a reach of a stream by fluvial erosion. The denominator consists of those variables that if increased tilt the balance in favor of aggradation. Sustained aggradation raises the altitude of a reach of an active stream channel by selective deposition of bedload. of Rising Mountains 41 Stream Power (driving factors) Resisting Power (prohibiting factors) = 1.0 (2.4) The available stream power to transport bedload (Ferguson, 2005) as defined by Bagnold (1977) is Q = yQS (2.5) where Q is the total kinetic power, or in terms of power per unit area of streambed, yQS d)=-L^- = yQSv = icq (2.6) where J is the specific weight of sediment-water fluid, Q is stream discharge, S is the energy slope, W is streambed width, d is streamflow depth, V is mean flow velocity, and tco is the shear stress exerted on the streambed (Baker and Costa, 1987). Unit stream power is a measure of the power available to do work in a reach of stream. The denominator of equation 2.4, resisting power, becomes greater with increases in hydraulic roughness, and the amount and size of bedload. When resisting power exceeds stream power, aggradation of bedload occurs as the suspended load continues to be washed downstream. The deposits of most stream terraces and many small alluvial fans consist largely of gravelly bedload materials, as compared to the silty suspended-load overbank deposits of large rivers. Both bedrock uplift and climatic change profoundly affect the components of the threshold of critical power (Starkel, 2003). Bedrock uplift increases watershed sediment yields by increasing landscape relief and steepening valley floors. Climatic changes during the Pleistocene and Holocene changed discharges of both water and sediment. Climate change may increase bedload production so much that aggradation prevails along nearly all streams of a region, including reaches with tectonically active faults and folds. Watersheds in each climatic setting have different styles and magnitudes of geomorphic responses to changes from glacial to interglacial climates (Bull, 1991). Aggradation events in New Zealand occurred during full-glacial times. In the Mojave Desert of California they occurred during the transition to interglacial climates. Use of the threshold of critical power model emphasizes rates of change of the variables affecting a geomorphic process, thus encouraging you to consider the relative importance of many interrelated factors. 2.3.1 Relative Strengths of Stream Power and Resisting Power Relative strengths of stream power and resisting power vary spatially in drainage basins (Graf, 1982, 1983). Consider the hypothetical longitudinal profile for the tectonically active arid watershed depicted in Figure 2.14. Stream power continually exceeds resisting power in the headwater reaches. Slope remains excessively steep, especially where resistant rock types and small streamflows near drainage divides favor slow degradation of valley floors. Bedrock uplift of the mountain block occurs as displacements along the range-bounding normal fault zone. This tectonic perturbation to the entire fluvial system occurs in a narrow linear zone crossing the canyon mouth. Episodic bedrock uplift increases the slope component of stream power. Increases of orographically-induced precipitation also occur in rising mountains over time spans of 100 ky to 10 my, so local climate and daily weather patterns commonly are linked to long-term tectonic controls. In marked contrast, late Quaternary climatic changes over time spans of 0.1 to 3 ky were relatively abrupt. They greatly influenced both stream power and resisting power and simultaneously affected all the watersheds of a mountain range. Numerical modeling suggests vastly longer watershed response times to surface ruptures (Allen and Densmore, 2000). The situation depicted in the aggradational reach of Figure 2.14 is opposite to that of the headwaters. Deposition has been continuous because resist- Mountains Piedmont Figure 2.14 Relative strengths of stream power (5F) and resisting power (RP) along a hypothetical fluvial system in an arid closed basin. Tectonic perturbations are initiated by ruptures on a normal fault in reach T and climatic perturbations are initiated in reach C (all the drainage-basin hill-slopes). TIP is threshold-intersection point. ing power has persistently exceeded stream power. Such reaches are particularly obvious in basins of internal drainage where accumulation of playa and associated alluvial-fan deposits constitute a base-level rise that gradually tends to decrease overall piedmont slope and thereby stream power. Tectonic subsidence of a basin - stretch tectonics - tends to offset the base-level rise caused by aggradation. Examples of depositional basins include Great Salt Lake in Utah, the San Joaquin and Death Valleys in California, and lake basins in North Africa and the Middle East. The mountain-front reach of Figure 2.14 will be the most likely to change its mode of operation as a consequence of climatically induced changes in hill-slope water and sediment yield. This is the most sensitive part of a fluvial system to climatic and tectonic perturbations because it is where stream power may exceed, equal, or be less than resisting power for a given streamflow event. Relative discharges of water and bedload are especially important because moderate changes in resisting power may initiate episodes of aggradation or degradation. 2.3.2 Threshold-Intersection Points Threshold-intersection points are spatial crossings of the threshold of critical power along longitudi- :er 2 nal profiles of valleys (Fig 2.3). Aggradation in a mountain-front reach that formerly was downcut-ting into bedrock moves the threshold-intersection point upstream. Renewed stream-channel downcut-ting through the new valley fill and fan deposits then shifts the threshold-intersection point downstream leaving the recently elevated valley floor as the tread of a fill terrace (Figure 2.14). Threshold-intersection points may shift in a great variety of time scales. More than 2 ky may be needed for a shift such as that depicted in Figure 2.14. In contrast, sediment discharge commonly peaks before water discharge during a flood event. This causes the threshold-intersection point to migrate upstream and then downstream as the ratio of stream power to resisting power changes during the flow event. Threshold-intersection points shift more rapidly along the valleys of streams in humid regions than in arid regions, because wet climates provide greater annual stream power to do the work of transporting the sediment load imposed by weathering and erosion of hillslopes. 2.4 Equilibrium in Streams Those who seek tectonic information from landscapes need to know how tectonic perturbations change the behavior of fluvial systems. How do landscapes evolve when perturbed by vertical and horizontal earth deformation? Do they tend towards unchanging configurations? Do some landforms achieve such a steady state (equilibrium) sooner than others? Vital skills include being able to assess how far removed a landform is from a steady-state condition, and how quickly part of a fluvial system can achieve a new equilibrium. Depositional landforms, such as growing deltas and alluvial fans, do not even tend toward steady-state conditions (Bull, 1977a, 1991, Section 1.6.2). 2.4.1 Classification of Stream Terraces Some former valley floors are equilibrium reference surfaces that record tectonic deformation. Let us classify types of stream terraces, noting their suitability for tectonic studies. Two types of equilibrium, and the crossing of the threshold of critical power, can be recognized by the presence of distinctive types of stream terraces. We are particularly interested in stream terraces resulting from tectonic and climatic perturbations, but acknowledge that common minor Concepts for Studies of Rising Mountains 43 terraces (called autocyclic) form because of responses ceding degradation event when tectonically induced to fluctuations of geomorphic processes within a flu- downcutting lowers the active channel below the for- vial system. mer valley floors represented by stream terraces. First Alternating aggradation and degradation let us get acquainted with standard stream-terrace events are common in fluvial systems that are sen- terminology (Fig. 2.15A). sitive to climatic changes. Termination of aggrada- Terraces may be paired or unpaired. Unpaired tion and degradation events creates key landforms. A terraces (Davis, 1902) typically occur on the insides of degradation event is an interval of net lowering of a meander bends along a stream that is steadily down-valley floor by fluvial erosion. Minor stream terraces cutting, but sometimes occur where a stream tempo-form during pauses in the downcutting process. A rarily erodes laterally into a bedrock hill. A paired degradation event commonly is terminated by a pro- terrace consists of remnants of a former floodplain longed episode of lateral beveling when the stream that when connected describe a former longitudinal stays at the same low position in the landscape. profile of the stream. They occur on both sides of a Different climatic controls promote an aggra- valley where not removed by erosion. An important dation event, which tends to backfill valleys (Barnard characteristic of a paired terrace is that it has continu-et al., 2006) and commonly is terminated by deposi- ity along a valley. tion of a broad alluvial surface. Aggradation tends to Flights of stream terraces can be likened to bury fluvial landforms created by the preceding deg- flights of stairs; both consist of a sequence of alternat- Figure 2.15 Tectonic, climatic, and internal-adjustment terraces of the Charwell River, New Zealand. A. Sketch of basic definitions. A strath beveled at the lowest possible position in a landscape is the tectonic reference landform. A climate-change induced aggradation event buries the tectonic strath. The tread of the resulting fill terrace records a time of change from aggradation to renewed degradation. Pauses in the subsequent stream-channel downcutting that generate stream terraces may be due mainly to adjustments caused by variable hillslope water and sediment yield. 3.Multiple strath and aggradation surfaces. Geomorphic responses to late Quaternary climatic changes and tectonic perturbations, and internal adjustments of the fluvial system, are recorded as a flight of stream terraces in this rising landscape. The Flax Hills and Stone Jug aggradation surfaces are fill-terrace treads that record terminations of climate-change induced aggradation events. Internal adjustments are preserved as minor strath, fill-cut, and fill terraces that record pauses in degradation from the aggradation surface to the younger broad tectonic strath terrace at the base level of erosion. radation event. Such burial is continuous as shown by lack of buried soil profiles in the stratigraphic section of an aggradation event. Fortunately, net stream-channel downcutting occurs in tectonically active landscapes. Watersheds with infrequent aggradation events are less likely to conceal landforms of the pre- A Broad surface that marks the end of an aggradation event is _^\the fundamental climate-change reference surface of streams / \ Internal-adjustment t| ez__: Broad strath beveled at the base level of erosion is — fundamental tectonic reference surface of streams ^- Climatic aggradation surfaces nternal-adjustment t terraces / Internal-adjustment fill terrace -y Tectonic straths -; cut in bedrock Internal adjustment minor strath terrace Charwell River at the base level of erosion V 44 Chap ing treads and risers. A terrace tread is remnant of a former valley floor that has been abandoned by the stream as a result of stream-channel incision. The corresponding terrace riser is a scarp created by fluvial lateral erosion above the tread. Former levels of streams may be classed as fill, fill-cut and strath terraces (Leopold & Miller, 1954; Howard, 1959; Bull, 1990). A fill terrace is formed by aggradation of a valley floor and subsequent channel incision into the alluvium that leaves remnants of the former active channel as the tread of a paired fill terrace. The tread represents a time of crossing of the critical power threshold when the mode of operation switches from aggradation to degradation. Longitudinal profiles of fill terraces generally do not represent attainment of equilibrium, because this threshold crossing tends to be abrupt and may occur at different times along a valley A strath terrace is genetically the same as a fill-cut terrace: straths are surfaces beveled in bedrock and fill-cut surfaces are beveled in alluvium. This generally brief equilibrium (pauses in a degradation event) is followed by renewed stream-channel downcutting that preserves remnants of the beveled surface, and gravel being transported at the time of formation, as the terrace tread. In this case, the riser (or streambank) was formed shortly before initiation of renewed stream-channel downcutting. Strath and fill-cut terraces differ from fill terraces in that only thin layers of stream gravel cap their erosion surfaces. These thin deposits may be regarded as lag deposits or gravelly cutting tools that most likely would have been entrained and redeposited by the next large flood. Erosional lowering of a valley floor does not proceed at a constant rate. Stream-channel downcutting rates vary with unit stream power and bedload transport rate. Pauses in valley-floor degradation allow fill-cut and strath terraces to form. Brief reversals in degradation of valleys caused by deposition of bedload result in minor fill terraces. Typically, both of these brief, minor, types of events create internal-adjustment terraces. In contrast to tectonic and climatic stream terraces, they result from temporal variations of dependent variables in a fluvial system. Such stream terraces have been studied in experimental (sandbox) models. They are called "complex responses" (Schumm, 1973; Parker, 1977; Schumm et al., 1987) and are referred to as "auto-cyclic" landforms by Hasbargen and Paola (2000). These authors studied internal-adjustment terraces ter2 in sandbox models by initiating upstream migration of knickpoints that locally increased sediment yield from hillslopes. Although internal-adjustment terraces are important for understanding adjustments in fluvial systems, they do not record the start or end of climate-change induced aggradation, nor can they be used to assess rates of valley floor downcutting that are related to bedrock uplift. Many other processes cause perturbations of water and sediment discharge that favor formation of internal-adjustment terraces; fires and landslides are examples. A kinematic wave of bedload increase sweeps down trunk stream channels only to be followed by stream-channel incision to produce an internal-adjustment fill terrace. Aggradation may last only a few years or decades. Dating flights of internal-adjustment terraces in adjacent drainage basins usually reveals a lack of synchroneity (Bull, 1991). Each watershed behaves differently. In contrast, tectonic and climatic stream terraces tend to be synchronous throughout much of a mountain range. Examples of synchronous terraces range from a brief episode of valley backfilling caused by an earthquake-induced sediment-yield increase to regional aggradation of longer duration caused by stripping of the hillslope sediment reservoir during the Pleistocene—Holocene climatic change. Important terraces form at the end of major aggradation and degradation events. Thick alluvial fill deposited in response to regional climatic change raises valley floors and alluvial-fan surfaces. The amount and rate that valley floors are raised by such deposition are mainly a function of the time span of the climatic perturbation and magnitude of departure from the threshold of critical power. A fill-terrace tread represents the maximum altitude attained by backfilling a valley floor, and records the end of persistent aggradation that commonly lasts 1 to 15 ky. This type of aggradation surface is the fundamental climatic stream-terrace landform. The tread is the only landform created that has temporal significance if the riser above the tread is merely a valley side. Aggradation event surfaces in adjacent watersheds are synchronous where response times to climatic perturbations are similar. There are limits to the depth of erosion of any valley floor, and the termination of a degradation event sets the stage for creation of an important type of strath terrace. Lateral erosion that bevels bedrock and widens an active channel at the end of a degradation event Concepts for Studies may persist for 1 to 10 ky. It creates a strath that is the fundamental tectonic stream-terrace landform (Fig. 2.4). This alluvium-bedrock interface records the lowest possible longitudinal profile for a particular tectonic setting. Strath age is defined as the end of an interval of strath formation. This occurs when renewed, persistent deposition raises the streambed and buries the beveled bedrock surface. Termination usually requires a climatic change. Alternatively, valley floor bedrock beveling can be terminated by base-level fall induced by a large surface rupture in a downstream reach. This increases stream power sufficiently that the threshold of critical power is crossed to create a tectonic strath terrace landform. Multiple late Quaternary tectonic strath terraces record important times of attainment of equilibrium along the streams of rising mountains. In contrast to the generally minor internal-adjustment strath terraces, tectonic straths tend to be broader, and are more likely to be correlated between adjacent reaches. Reaches in adjacent drainage basins with similar bedrock resistance, discharge, and bedload transport rates generally complete a degradation event at about the same time. If so, time spans of tectonic strath formation may be synchronous for a mountain range. Adjustments of reaches of streams that achieve such equilibrium conditions are discussed in the next two sections. 2.4.2 Feedback Mechanisms The threshold approach to geomorphic investigations emphasizes how far removed a system is from stable conditions. Thresholds are essential for studying interactions between geomorphic variables that do not tend towards an unchanging "steady-state" condition. So, we need to consider such relations in a bit more detail. Interactions between geomorphic variables may tend to create a landform, or landscape, that does not change with the passage of time. Alternatively, such interactions may result in an opposite tendency away from equilibrium conditions. Key to understanding both modes of landscape evolution are feedback mechanisms between dependent variables that drive the system toward or away from steady-state conditions. Studies of both self-enhancing and self-arresting feedback mechanisms are encouraged by the threshold approach to geomorphology whereas studies only of self-arresting feedback mechanisms are encouraged by the equi- of Rising Mountains 45 librium approach to geomorphology The behavior of streams is one of the better examples of how self-arresting feedback mechanisms promote equilibrium. Landslides represent the opposite situation where self-enhancing feedback mechanisms drive the process towards its culmination rather than toward a balance between variables. Progressive accumulation or erosion of hill-slope colluvium are examples of self-enhancing mechanisms. Colluvium deposited on bare rock increases infiltration capacity thereby providing both water and soil to support vegetation. The vegetation traps additional colluvial materials from upslope sources, thereby furthering the tendency for accumulation of more colluvium. For a reversal in the mode of operation to occur, a threshold must be passed that separates tendencies for progressive accumulation from those of progressive erosion of colluvium. Climate-change perturbations cause this system to alternate between two modes of operation and both landforms and landscape evolution cannot tend towards unchanging conditions. Such nonequilibrium modes of operation may involve time spans of decades or millions years. Progressive increase in areas of massive granitic outcrops in the Sierra Nevada of California over several million years may be considered an irreversible change. The self-enhancing feedback mechanism of rapid runoff from bare rock continues to erode the soil produced by weathering of granitic rocks (gruss) at the margins of hillslope outcrops. Clyde Wahrhaftig (1965) made a classic study of such a landscape evolution. He sought to learn more about the formation of stepped topography developed in the massive granitic rocks of the western slopes of the Sierra Nevada. Bedrock exposed by streams expanded in area to become the dominant hillslope landform. Such topographic inversion — valley floors eventually become ridgecrests - requires that the drainage net, as well as the hillslopes, undergo progressive change. Thus Wahrhaftig's study provides a nice example of the importance of self-enhancing feedback mechanisms. The steady-state model is inappropriate because a key independent variable - erodibility of surficial materials - changes in both time and space. A different example of self-enhancing feedback mechanisms is the progressive development and ultimate collapse of a slump rotation block in massive sandstone cliffs (Schumm and Chorley 1964). A seemingly insignificant tension crack at the top of the cliff signals initiation of a process that may require 46 Chapter 2 thousands of years. Insufficient lateral support for the cliff face allows the minute crack to gradually widen at an exponentially faster rate, thereby enhancing several processes that culminate in failure of the sandstone monolith when it collapses into a pile of rock-fall blocks. Water and ice accumulate in the ever widening crack, rotation shifts the center of gravity of the monolith, and block rotation fractures the lower portion of the sandstone pillar thereby causing a progressive decrease in rock mass strength. Now, let us examine the opposite type of feedback mechanism. The adjustment of a stream to an increase of gravel eroded from the hillslopes of its fluvial system is an example of self-regulation by self-arresting feedback mechanisms. Large increases of bedload change all hydraulic variables in meandering stream channels, which in turn results in a decrease in channel sinuosity, so that more bedload can be conveyed. In braided channels, increases in bedload may cause aggradation of the valley floor with maximum alluviation in the upstream reaches. This increase in streamflow gradient is the principal way stream power is increased. In both meandering and braided streams, adjustments of the hydraulic variables will continue as long as the stream is supplied an excess of bedload (Paola and Mohrig, 1996). Changes in flow characteristics or stream-channel characteristics that are greater than needed to transport the bedload will result in counterbalancing adjustments that will decrease the transporting capacity and competence of the stream. A key tectonic landform is the longitudinal profile of a stream channel (Merritts et al., 1994). We consider present and past longitudinal profiles as potential reference datums passing through tectonically deforming landscapes. So, it is worth our while to consider several types of equilibrium in such streams, and the landscape characteristics associated with each type. 2.4.3 Dynamic and Static Equilibrium The base level of erosion concept describes reaches of streams that have achieved one of two types of equilibrium-static equilibrium and dynamic equilibrium (which can be further classified as type 1 or type 2). We define and discuss these terms here in order to clarify how climatic controls and internal adjustments in fluvial systems lead to static equilibrium, and how tectonic controls create situations of dynamic equilibrium in landscape evolution. Unless classed as static or dynamic, the general term "equilibrium" refers to reaches that have attained the base level of erosion. Static equilibrium is attained briefly when bedload derived from the watershed hillslopes is transported through a reach with neither aggradation nor degradation of the streambed (Leopold and Bull, 1979). One example is a gradual, instead of abrupt, switch from aggradation to degradation at the conclusion of an aggradation event. Bull (1991, Section 5.4.4.1) describes field identification of an aggradation surface that records a period of static equilibrium. A stream that pauses briefly during a degradation event creates minor fill-cut or strath terraces that also record temporary static equilibrium. John Hack used the ideas of G.K. Gilbert when he applied the concept of dynamic equilibrium (steady state) to open fluvial systems of mountains. Hack believed that a steady-state landscape configuration would develop between processes that tend to raise and denude mountains. All elements of a landscape in dynamic equilibrium are mutually adjusted to each other. The resulting landscape assemblage downwastes at a uniform rate, and its configuration does not change (Gilbert, 1877; Hack, I960, 1965; Carson, 1971). Thus, dynamic equilibrium is independent of time and acknowledges that climatic and tectonic energy is continually entering and leaving the system. Adjustments to perturbations restore a dynamic steady state without significant delays. The steady-state model of landscape evolution continues to be highly regarded by many geomorphologists. The Gilbert-Hack model is heuristic, because we have yet to obtain field evidence for attainment of true steady state for an entire landscape (Bull, 1977a). Important tectonic and climatic perturbations occur frequently during time spans of 0.1 to 100 ky and affect denudation rates that vary with both time and space within drainage basins. In my opinion, response times of geomorphic processes are sufficiently long for hillslopes to prevent attainment of steady state for either watersheds or mountain ranges. Models using non-steady state assumptions are preferable for studies of landscape evolution. Solyom and Tucker (2004) nicely describe obvious examples of why steady state is inappropriate for modeling short-term behavior of entire watersheds by emphasizing interactions between the variables of storm duration, basin shape, and peak discharge. Central to their argument is that precipitation-runoff inputs are not uniformly distributed in Concepts for Studies a spatial sense, creating disparities that increase with drainage-basin area and with increasing watershed aridity. Nonuniform inputs of precipitation typically create situations of partial-area contributions of runoff from a watershed (Dunne and Black, 1970; Yair et al., 1978), which lead to lesser concavity (rate at which a longitudinal profile becomes flatter downstream). Lack of a vigorous proof for dynamic equilibrium for entire watersheds does not detract from the value of the base level of erosion concept, and the tendency of specific landscape assemblages toward time-independent shapes. One nice application of the dynamic equilibrium model is the way in which powerful streams respond to bedrock uplift. Two categories of dynamic equilibrium in streams may be defined in terms of length of attainment of the base level of erosion. Hack (1973, 1982) defined steady-state adjustment between variables as those longitudinal profiles that plot as straight lines on semi-logarithmic graphs because stream discharge increases logarithmically downstream (Wolman and Gerson, 1978). Two types of dynamic equilibrium can be recognized in the field. Diagnostic landforms for attainment of type 1 dynamic equilibrium include a strath beneath the active channel and a valley floor that is sufficiently wide for preservation of strath terraces. Type 1 dynamic equilibrium is present when stream down-cutting and bedrock uplift rates are the same. An analogy would be the constant shape and position of a rotating circular saw (the stream) as a log (the mountains) is raised into it. The resulting sequence of longitudinal profiles represents an infinite number of tectonic base levels of erosion as the stream reestablishes equilibrium conditions after each tectonic perturbation in the rising landscape. Climatic perturbations in real world fluvial systems do not allow perpetual uniform downcutting in response to bedrock uplift. Instead, climatic factors modulate the discharge of water and bedload so as to vary the rate of valley-floor downcutting. The result is a series of closely spaced minor strath terraces, each created briefly at an appropriate time as dictated by changing climatic constraints. Type 2 dynamic equilibrium is present in streams with a strong tendency toward, but lack of sustained attainment of, the base level of erosion. Diagnostic landforms include narrow valley floors whose longitudinal profiles (like those of type 1 stream reaches) plot as concave lines on arithme- of Rising Mountains 47 tic graphs and as straight lines on semi-logarithmic graphs. Straths and remnants of strath terraces generally are not present. Recent studies of bedrock stream-channel longitudinal profiles and stream-channel characteristics are a worthwhile departure from the previous emphasis on less confined, easily eroded alluvial streams (Baker and Kochel, 1988; Seidl and Dietrich, 1992: Wohl, 1994; Montgomery et al., 1996; Baker and Kale, 1998; Hancock et al., 1998; Pazzaglia et al., 1998; Sklar and Dietrich 1998; Tinkler and Wohl, 1998a, b; Massong and Montgomery, 2000; Wohl and Merritt, 2001, Whipple, 2004). Irregular longitudinal profiles are characteristic of disequilibrium streams upstream from most type 2 equilibrium reaches where interactions between variables have yet to approximate the base level of erosion. Landforms in disequilibrium reaches include convex valley sideslopes plunging into V-shaped canyons with waterfalls and rapids, and convex longitudinal profiles, even when plotted on logarithmic graphs. Stream-channel widths for degrading streams tend to be narrower than for streams in equilibrium. Consider the case where renewed degradation converts an equilibrium reach characterized by strath cutting (active-channel width is not confined) to confined streamflow that is typical of downcutting reaches. Streams degrading into bedrock may establish a self-enhancing feedback mechanism that accelerates downcutting until equilibrium conditions are approached. Narrowing of streamflow tends to increase unit stream power (equation 2.6). Millions of years may be needed for ephemeral streams flowing over resistant rocks to approach their base levels of erosion. Some response time is needed for large rivers to adjust to a pulse of uplift even in humid mountain ranges. The time spans needed for landscape adjustments typically are longer than intervals between uplift. Upstream reaches and adjacent hillslopes will have longer response times. So a model where interacting landscape elements continue to adjust makes more sense than proclaiming still another example of steady-state conditions. One alternative is the allometric change conceptual model (Bull, 1975a; 1991, Section 1.9) that describes orderly interactions between dependent variables in a changing landscape before or after the base level of erosion is attained. Allometry in biology describes relative systematic changes in different parts of growing organ- 48 Chap isms (Vacher, 1999). Allometric change in geomor-phology describes orderly behavior in nonsteady state fluvial systems. Hydraulic geometry of stream channels (Leopold and Maddock, 1953) is a good example. Increase of flow width with increasing discharge at a gauging station is static (at a single location in a watershed) allometric change. Increase of flow width with downstream increase of discharge (a sequence of locations) is dynamic allometric change. The allometric change model emphasizes the degree of inter-connectivity of different geomorphic processes and landscape characteristics. Allometric change emphasizes system behavior instead of attainment of steady state and is mentioned here to alert readers that most regressions between geomorphic variables in this book imply nonsteady-state assumptions. Perennial streams of humid regions flowing over soft or highly fractured rock are better suited for studies of responses of streams to bedrock uplift during the past 40 ky. One such stream is the Charwell River, whose main branch drains 30 km2 of the Seaward Kaikoura Range of New Zealand. Times of aggradation events were estimated by radiocarbon dating of wood in the fill-terrace deposits of the region. The Charwell River is the type area for the Stone Jug, Flax Hills, and Dillondale climate-change induced aggradation events, which have synchronous counterparts elsewhere in New Zealand (Barrell et al., 2005; Litchfield and Berryman, 2005). Terrace tread ages were estimated with weathering rind analyses (Knuepfer, 1988). Strath ages were estimated by radiocarbon dating of tree trunks, in growth position, just above the straths. These age estimates indicate termination of intervals of major strath beveling at about 40 and 29 ka. Tectonic strath terraces are those that record times of attainment of the base level of erosion (type 1 dynamic equilibrium) in rivers that are being tec-tonically elevated. The Charwell River has both internal-adjustment and tectonic strath terraces. The piedmont reach extends 12 km downstream from the highly active range-bounding Hope fault, has local faulting and folding, and has experienced episodes of climate-change induced aggradation. Bedrock uplift ranges from 0.2 to 1.3 m/ky. Successive tectonic strath terraces are at lower positions in the landscape because the Charwell River lowers its channel in response to continued bedrock uplift (Fig. 2.15B). Each broad tectonic strath was created at a time of interstadial or interglacial climatic conditions, such as the present Holocene climate. Each episode of tee- ter 2 tonic strath formation ended at the time of initiation of an aggradation event. The long-term tendency of rivers to cut down into a rising landmass is temporarily reversed during aggradation events, which in humid, mesic to frigid climates coincide with times of full-glacial climates (Bull, 1991, Chapter 5; Wegmann and Pazzaglia, 2002). Rivers then switch to degradation and may catch up to, and re-establish, type 1 dynamic equilibrium during times of interglacial climates. This is when streams have excess stream power, relative to resisting power. Rapid stream-channel downcutting (Bull, 1991, Fig. 5.18) began at 14 to 16 ka after termination of the latest Pleistocene aggradation event, and ended at about 4 ka. By then, the Charwell River had attained its tectonic base level of erosion downstream from the range-bounding Hope fault. Bedrock uplift has continued to raise the stream during the past 4 ky, and the active channel has cut down into soft piedmont-reach lithologies at a rate that matches the local bedrock-uplift rate. Fluvial erosion switched from largely vertical to mainly horizontal and the stream beveled a strath as wide as 200 m (discussed and shown later in Figure 2.26). Degradation stream terraces also record times of attainment of the base level of erosion. The flight of Charwell River stream terraces provides examples of how self-arresting feedback mechanisms result in temporary static equilibrium. Latest Pleistocene aggradation buried the 29 ka tectonic strath and raised the streambed altitude 40 m. Then rapid degradation of the bouldery valley fill occurred. The stream paused a dozen times as it cut down before it caught up to its long-term (tectonic) base level of erosion (Fig. 2.16). Tectonic and climatic controls on other Marlborough streams follow a similar pattern but magnitudes of change in positions of valley floors are functions of the amounts of latest Quaternary aggradation and uplift. Mean rates of post-glacial incision for the Saxton River on the Awatere fault are 1.4 m/ky compared to 5.1 m/ky for the Charwell River (Fig. 2.16). The shapes of the two incision curves are the same, and both streams have returned recently to the base level of erosion where the rate of stream-channel downcutting matches the uplift rate. Each pause was a time of attainment of static equilibrium that was recorded by a minor terrace. Stream-channel entrenchment into poorly sorted sandy gravel promoted selective entrainment and transport of the finer portion of the valley fill. Accumulation of residual boulders on the streambed Concepts for Studies Types of stream terraces Tectonic I Internal adjustment ICIimaticI /ium of the Stone aggradation event Bedrock Age, ka 12 16 Figure 2.16 Tectonic, climatic, and internal-adjustment terraces of the Charwell River: variations of stream-channel downcutting rate since 16 ka. Uncertainty cross estimates are for terrace ages and distances below the aggradation surface reference level. From Figure 6 of Bull and Knuepfer (19S>7). Gray plot is for Saxton River on the Awatere fault. From Figure 13 of Mason et al. (2006). created a lag deposit that armored and protected the streambed from further degradation by 1) increasing the shear stresses needed to initiate movement of streambed gravel, and 2) by increasing the hydraulic roughness. Riparian vegetation grows beside streams with armored beds, further increasing hydraulic roughness. Neither net aggradation nor net degradation occurs, even though the active channel is still far above its tectonic base level of erosion. These brief episodes of static equilibrium end abruptly when a 1-ky flood event entrains and smashes streambed-armor boulders, and destroys riparian vegetation. Both changes reduce hydraulic roughness and the shear stresses needed to entrain bedload in the normal range of stream discharges. Renewed streambed degradation and winnowing then create a new lag gravel at a lower altitude. Remnants of the former streambed, with the characteristic stream-bed armor preserved as a capping layer of boulders, remain as treads of fill-cut or minor strath terraces. Thus without invoking either climate-change or tectonic perturbations, a self-arresting feedback of Rising Mountains 49 mechanism can occur repeatedly during the formation of a flight of degradation terraces. These are regarded as internal-adjustment terraces because they are not the result of changes in the independent variables of this fluvial system. Each records a pause in valley-floor degradation caused by temporary increases of resisting power. Different geomorphic responses to bedrock uplift in the reach upstream from the Hope fault have resulted in narrow V-shaped canyons (Bull, 1991, Fig. 5.7), convex footslopes, and waterfalls. The fractured greywacke sandstone is more resistant to erosion than the soft Cenozoic sediments downstream from the fault. Bedrock-uplift rate is about 3.8 ± 0.2 m/ky based on altitudinal spacing analysis of uplifted marine-terrace remnants. Degradation by small headwater streams is unable to keep pace with bedrock uplift, but some mid-basin reaches of the Charwell River have sufficient unit stream power to approximate type 2 dynamic equilibrium. The reach just upstream from the Hope fault has remnants of two fill terraces, perhaps reflecting repeated episodes of base-level rise caused by rapid deposition of 30-50 m of latest Pleistocene valley fill in the adjacent reach downstream from the Hope fault. Many possible interactions between variables can produce equilibrium in streams. Increased slope and reduced stream-channel width may be important in achieving equilibrium after vertical offset of a streambed by an earthquake surface rupture. Changes in hydraulic roughness and stream-channel pattern may occur after a landslide increases bed-load. Powerful streams may maintain equilibrium by adjusting interactions between variables, within a limited range, as short-term climatic change influences unit stream power and bedload transport rate. Perturbations that force a reach of a stream beyond its capacity to maintain equilibrium conditions initiate aggradation or degradation events. Sections 2.6 and 2.7 provide several examples of how strath terraces can be used by tectonic geo-morphologists. But first, let us outline the characteristics of response times of fluvial systems to climatic and tectonic perturbations. 2.5 Time Lags of Response Several concepts, aggradation and degradation events, attainment of the base level of erosion, and time lags of response can be summarized with a simple diagram (Fig. 2.17) Reaction time is the delay before Chapter 2 Figure 2.17 Anatomy of an aggradation-degradation event illustrating the concepts of reaction time, relaxation time, and response time after a climate-change perturbation causes an episode of alluviation of a valley floor. streambed aggradation begins: for example, the interval when hillslope plant cover decreases thus increasing sediment yield sufficiently that the stream can no longer maintain equilibrium conditions. Relaxation time is the time span needed to complete the aggradation event. Response time is shown as the total elapsed time from the time of a climatic perturbation to end of the aggradation event. It is the sum of reaction and relaxation times (Allen, 1974; Thornes and Brunsden, 1977; Brunsden and Thornes, 1979; Brunsden, 1980). The ensuing degradation event returns the stream to the same base level of erosion in this case. This is indicative of a lack of uplift during the time span represented by the aggradation-degradation event. Fluvial systems also respond to being elevated. Persistence time is the time span during which fluvial system behavior is constant, which here is a condition of type 1 dynamic equilibrium. 2.5.1 Responses to Pulses of Uplift The concept of response time in a fluvial system to tectonic inputs can be illustrated with a threshold-equilibrium plot (Fig. 2.18). Reactions of a stream to two hypothetical perturbations are illustrated for a reach that is 2 km upstream from the mouth of a watershed and initially in type 1 dynamic equilibrium. The first perturbation is a 2 m surface rupture on the range-bounding normal fault. Tectonic lowering of the reach immediately downstream from the fault, relative to the mountain block, creates a knickpoint. This short tectonically steepened reach time Time->- Figure 2.18> Hypothetical threshold-equilibrium plot showing the components of response time, R, for a reach 2 km upstream from a normal fault. Delayed responses are depicted for two perturbations; a 2 m surface rupture on the range-bounding fault, and a landslide upstream from the study reach. Response time is the sum of the reaction time, R , a and relaxation time, R . P is the time of x s persistence of new equilibrium condition, and T and E are the times of threshold and equilibrium conditions respectively. Concepts for Studies migrates upstream. A knickpoint may be a waterfall initially but commonly becomes rapids - a knick-zone - where rock mass strength is low Processes of upstream migration are different for these two types of stream-channel perturbation. Knickpoints retreat mainly by undermining and collapse and knickzones by streambed abrasion and plucking by fast moving saltating bedload. Both may be significant local departures from the typically concave longitudinal profile indicative of a reach in equilibrium. As such they can inhibit continuity of fluvial systems (Section 2.5.2) to the extent that waterfalls separate relict landscapes in a dramatic fashion (Crosby and Whipple, 2006). Rates of incision into bedrock are considered to be proportional to shear stress exerted by streamflow (Howard and Kerby 1983; Tucker and Slingerland, 1994; Snyder et al., 2000, 2003). Some geomorphologists prefer to emphasize total stream power (Seidl and Dietrich 1992), bedload transport rates relative to the threshold of critical power (Kooi and Beaumont, 1994), bedload abrasion potential (Sklar and Dietrich, 1998), or unit stream power (Whipple and Tucker, 1999). The interval is quite short for the first persistence time of Figure 2.17. Migration of the hypothetical knickzone (Figure 2.17) still further upstream undercuts a hillside and triggers a landslide - the second perturbation. The result is an increase in bedload input to the stream that causes minor aggradation in the study reach, followed by a new equilibrium condition. Note that equilibrium is regarded as an interval with no change of streambed altitude, whereas the threshold is depicted as a point in time when the system reversed modes of operation (for example, from degradation to aggradation). The first two equilibrium intervals are examples of type 1 and type 2 dynamic equilibrium. The third is an example of static equilibrium, because the stream is above its long-term (tectonic) base level of erosion. Reaction time is a measure of the sensitivity of a fluvial system to a perturbation, and relaxation time is a measure of how efficiently a geomorphic system adjusts to a perturbation. The reach immediately upstream from a surface rupture reacts quickly as waterfalls and rapids are created. Hillslopes in the headwaters of the same stream react slowly and flat summits and plateaus have such exceedingly long reaction times that they are essentially isolated from downstream perturbations. Resistance of rocks may be the same throughout a drainage basin, but stream power decreases exponentially in the upstream direc- of Rising Mountains 51 tion. Spatial variations in response time inhibit attainment of steady-state conditions for entire drainage basins. Outputs of fluvial systems, such as sediment yield, integrate response times from all parts of a drainage basin. Renewed degradation of a valley floor in response to a range-front uplift event provides a new source of sediment as streamflow downcuts into reaches previously at equilibrium. Stream-channel entrenchment steepens adjacent hillsides, which promotes mass movements; both processes increase sediment yield. Tectonically induced increases of watershed sediment yield increase resisting power and tend to accelerate deposition in reaches downstream from the active fault. Chapter 4 has many examples of the abrupt initiation of deposition of alluvial fans that coincide nicely with this tectonic base-level fall. An aggradation-rate increase on the alluvial fan reach of a fluvial system alluvial fan occurs as quickly as upstream channel downcutting is initiated, but maximum rate increase may be delayed substantially until tributary streams and their large areas of hillslopes are affected by an episode of range-front faulting. So reaction time is short for the depositional reach, but relaxation time may be extended until the effects of a tectonic perturbation have spread throughout much of the drainage basin. The response-time model described above applies only to those fluvial systems with continuity Both erosional and depositional reaches of some streams can have characteristics that inhibit continuity of fluvial-system behavior. 2.5.2 Perturbations that Limit Continuity of Fluvial Systems Waterfalls are dramatic landforms that typically isolate upstream reaches from the effects of surface rupture on downstream fault zones. These streambed cliffs effectively decouple upstream from downstream reaches. Decoupling isolates the reaches in the headwaters of a watershed, which may show minimal consequences of incremental increases in watershed relief emanating from an active range-bounding fault. Decoupling inhibits stream-channel downcutting in a trunk valley from migrating up tributary valleys. Instead, such upstream reaches are graded to the top of a waterfall, or to a rapid with a drop sufficient to dissipate energy in a hydraulic jump. Rapids and falls are local base levels (Fig. 2.2B). 52 Chap Geomorphic responses to climatic changes may control the magnitude, time of formation, and position in a watershed of a waterfall (Sections 2.6.2, 2.7). Episodes of uplift along a range-bounding fault increase the height of a bedrock fault scarp even when deep beneath the gravels of a climate-change induced aggradation event. Age controls for the Charwell River landscape evolution allow estimation of process rates (Bull and Knuepfer, 1987). The period between about 26 and 14 ka ago was a time of valley floor backfilling. Episodic vertical movements along the Hope Fault during this time span continued to displace both bedrock and valley fill. Streamflows would quickly eradicate scarps formed in unconsolidated fill, but a sub-alluvial bedrock fault scarp would have become progressively larger until Holocene degradation was sufficient to expose bedrock once again at about 9 ka. The Hope Fault separates the mountain and piedmont reaches; both are rising and relative uplift is about 2.5 m/ky. Thus, late Quaternary cli- Figure 2.19 Decoupled hillslopes, piedmonts, and stream channels. A. Slowly eroding upland in the Kelly Range, Southern Alps, New Zealand changes abruptly to cliffy headwater reaches of rivers that are downcutting rapidly in response to bedrock uplift of 5 m/ky (Bull and Cooper, 1956). Uplands are shore platform(s) of uplifted marine terraces (SF). Degraded former sea cliff at SC. The foreground cliffs decouple the upland terrain from the adjacent downstream reaches of these fluvial systems. Hut at H is 110 m below shore platform at SR ter2 matic perturbation modulated tectonic processes to create and then exhume a prominent bedrock waterfall about 40 m high. After exposure it would retreat upstream as a knickpoint or steeper reach. The present anomalously steep reach 1300 m upstream from the Hope Fault departs from a smooth profile by about 40 m. Cliffs prevent continuity of geomorphic processes on hillslopes too. One example is the free face on a young fault scarp. This is why we can't do diffusion-equation modeling for a fault scarp that still has a free face. Another common example of lack of fluvial system continuity is the precipitous drop below a tabletop upland or mesa. Flux of sediment and water through cliffy landscapes is interrupted in a system that otherwise would behave systematically enough to be described by a single set of equations. Summit uplands are subject to such different processes that they may be eroding at a miniscule rate relative to the valley floors and hillslopes downvalley from the cliffy terrain that decouples the headwaters from the rest of the fluvial system. Remnants of shore platform-sea cliff landscapes of raised marine terraces (Fig. 2.19A) are another example of decoupled terrain. Reaches of active deposition may decouple fluvial systems as efficiently as waterfalls. Consider the case of a downcutting stream that emerges from a rising mountain range, crosses an aggrading alluvial fan, and joins a trunk river. A base-level fall on the trunk river will migrate upstream as a headcut but cannot be transmitted through the alluvial-fan reach as long as it remains on the aggradational side of the threshold of critical power. Fans decouple fluvial systems by not allowing base-level falls further downstream to influence stream channels and hillslopes in their source areas. The Carrizo Plains adjacent to the Temblor Range of California have ephemeral streams that respond to tectonic and climatic perturbations (Ouchi, 2005). Sieh and Wallace (1987) note that the fanhead trench of Wallace Creek has been incised for about 3.7 ky and records 128 m of right-lateral displacement by the San Andreas fault zone during this time span (Fig. 2.19B). Alluvial-fan deposition has subsequently occurred on the fanhead of the unnamed creek, whose young surface has undergone minimal stream-channel offset. Both streams have undergone arroyo cutting (Bull, 1964a, b) during the past 200 years. Erosional knickpoints during this climate-change induced stream-channel entrenchment can easily move up the continuous channel of Concepts for Studies of Rising Mountains 53 WMltce Crank Unnamsd ensdk 5 tOM tomour mterv»l Wallace Creek. The 400-m long aggradational reach stops knickpoint migrations in the unnamed creek. The fluvial system of the unnamed creek will remain decoupled until all threshold intersection points are eliminated so all reaches are on the erosional side of the threshold of critical power. Even big rivers can be decoupled. The Grand Canyon reach of the powerful Colorado River illustrates a third style of fluvial decoupling. This reach lacks the smooth concave longitudinal profile indicative of attainment of equilibrium conditions that one would expect for a big river. Instead, the profile is convex with numerous abrupt steps (Fig. 2.19C). Steps coincide with tributaries from cliffy 900 500 zs < 700 600 500 Nankoweap „Kwagunt , Unkar Creek n,/ Hance ^ Sockdolager horn Creek Salt Creek Crystal Walthenberg ./ Elves Chasm Kanab -.02 -.01 - 0 '50 100 150 200 250 Distance downstream from Lees Ferry (km^ Figure 2.19 Decoupled hillslopes, piedmonts, and stream channels. F3. Topographic map of Wallace Creek crossing the San Andreas fault that shows piedmont stream channels that are decoupled from their source areas by intervening reaches of alluvial-fan deposition. Wallace Creek is the only stream channel with continuity that permits upstream migration of knickpoints. Trace of the San Andreas fault is at base of the escarpment. From Figure 3 of Sieh and Wallace (1957). watersheds whose frequent debris flows deliver 1 to >3 m boulders to small fans (Fig. 2.19D). The river quickly removes the sand fraction from such debris fans, and can shift some of the smaller boulders a short distance downstream during flood discharges (Webb et al., 1999, Figure 41). Repeat photography (Webb et al., 1999, Figure 14) shows the same large boulders in the Lava Falls rapids after a moderate flow of 6,200 m3/s (220,000 ft3/s). Encroachment by debris fans controls the longitudinal profile of the river and the hydraulic behavior of the rapids (Dolan et al., 1978; Howard and Dolan, 1981; Kieffer, 1985; Webb et al., 1989; Melis and Webb, 1993; Webb, 1996; Griffiths et al., 1996, 2004). The Grand Canyon reach does not have the typical pool-riffle sequence characteristics of bedload streams (Webb et al., 1996, p. 152) because of the size and flux rate of boulders from tributaries. The result is constricted, steep reaches with near-critical streamflow characteristics and upstream migrating hydraulic jumps. Half the drop in altitude in the Grand Canyon reach occurs in short rapids (Leopold, 1969). Each bouldery obstruc- Figure2.19 Decoupled hillslopes, piedmonts, and stream channels. C. Longitudinal profile and gradients of the Grand Canyon reach of the Colorado River, Arizona between 50 and 250 km downstream from Lees Ferry. Steep reaches coincide with rapids (only a few are named here) resulting from bouldery debris flows derived from cliff/ tributary streams. 1927 data set supplied courtesy of Robert H. Webb, U. S. Geological Survey. sz 54 Figure 2.19 Decoupled hillslopes, piedmonts, and stream channels. D. diagrammatic Grand Canyon debris fan and rapids. 1 is debris-flow fan from tributary watershed. 2 is constricted river flow plunging down a rapid with large immobile boulders. 3 is debris bar of cobbles and small boulders derived from debris fan. 4 is secondary rapid caused by debris bar. From Griffiths et al., 1996 as modified from Hamblin and Rigby, 1965. Chapter 2 As elsewhere in the American Southwest (Bull 1991, Chapter 2) the change to Holocene climates resulted in much more frequent debris flows. Summer monsoon-type rainstorms and winter cyclonic storms, such as cutoff lows off the coast of Southern California and Mexico, that are sufficiently warm to produce thunderstorms were scarce or absent during times of cooler full-glacial climates. So, the influence of debris flows on the Colorado River longitudinal profile may not have been important before the change to a Holocene monsoonal climate. Pazzaglia (2004, p. 268) concludes "There L remains no good single explanation for why dramatically steep slopes on the Great Escarpments of passive margins erode at slow rates approaching 5 m/My TTl whereas equally steep slopes in tectonically active settings may erode at rates three orders of magnitude faster approaching 5000 m/My" The concept of impediments to continuity in fluvial systems may explain this paradox. Disconnected landscape elements cannot transmit base-level falls in a manner resembling watersheds where the trunk stream channel is the connecting link between different parts of an integrated fluvial system. Numerical analyses that fail to recognize the presence of two separate, adjacent, landscape systems might use a process-response model that does not represent the real world. Conceptually, the ancient Great Escarpments of South Africa resemble the sandstone spires of Monument Valley, Arizona. Both are spectacular weathering-limited cliffs whose rates of denudation have little relation to base-level changes in the streambeds of nearby fluvial systems. tion acts as a local base level for the upstream pool reach in much the same manner as first described by John Wesley Powell (1875, p. 203-204). Continued influx of debris-flow boulders is sufficient to maintain rapids as significant local base levels. Lava that flowed into the canyon was a local bedrock base level that was removed by the river. Bouldery debris fans are a different type of perturbation. They are renewable base-level controls that define the character of the Grand Canyon reach. Human impacts are increasing the importance of this perturbation. Heights of some rapids are increasing since dam construction eliminated large annual floods (Graf, 1980; Melis and Webb, 1993; Webb et al, 1996). 2.5.3 Lithologic and Climatic Controls of Relaxation Times The independent variables of lithology and structure, and of climate, largely determine the types of processes that create hillslope landforms, and the time needed for streams to achieve equilibrium conditions. Joints and fractures and petrologic fabric beneath a watershed change little with time, whereas climate change is ubiquitous. Climate also varies between north- and south-facing slopes, and with altitude. Lithology also varies with space, and even monolithologic watersheds - those drainage basins underlain by a single rock type such as quartz mon-zonite or greywacke sandstone - typically have a highly variable density of joints, fractures, and shears. 6781823517058112369312 Concepts for Studies of Rising Mountains 55 Precipitation Mean Annual (mm) Temperature Mean Annual (°C) Extremely arid Arid Semiarid Semihumid Humid Extremely humid <50 50 -250 250 - 500 500 - 1,000 1,000 - 2,000 > 2,000 Pergelic Frigid Mesic Thermic Hyperthermic > 0 0-8 8-15 15-22 >22 Seasonality Index (Sp)* Seasonality Index (St) * Nonseasonal Weakly seasonal Moderately seasonal Strongly seasonal <2 2-5 5-15 > 15 1-1.6 Nonseasonal 1.6-2.5 Weakly seasonal 2.5-10 Moderately seasonal > 10 Strongly seasonal * Precipitation seasonality index (Sp) is the ratio of the average total precipitation for the three wettest consecutive months (Pw) divided by the average total precipitation for the three consecutive driest months (Pd). Sp = Pw/Pd * Temperature seasonality index (St) is the mean temperature of the hottest month (Th) minus the mean temperature of the coldest month (Tc) in °C. St=Th- Tc Table 2.1 Classification of climates. Relaxation times after fault ruptures of streambeds are short for large rivers flowing on soft rocks, and long for ephemeral streams flowing on hard rocks. Locally massive, hard rocks greatly increase the lifespan of a waterfall that decouples upstream from downstream reaches. Multiple surface-ruptures produce knickpoints that may migrate upstream only to increase height and permanence of waterfalls. Table 2.1 defines the climatic terms used in this book. Each category, including extremely humid and extremely arid, is characterized by major differences in geomorphic and pedogenic processes. The temperature terms are from Soil Taxonomy (Soil Survey Staff, 1975). Mean annual air temperature at a site approximates soil temperature at a soil depth of 50 cm. Climatic constraints affect the time needed for fluvial processes to shape a given landform by at least an order of magnitude (a ten-fold variation). Consider the triangular facets shown in Figure 2.20A. Weak to moderately resistant rocks and an arid, thermic, strongly seasonal climate are responsible for Saline Valley triangular facets with minimal dissection. The lower portion of the rilled facet approximates the plane of an exhumed 35° to 40° range-bounding normal fault such as those described by Cichanski (2000). One might conclude that bedrock uplift must be rapid to form such dramatic triangular facets. The bedrock-uplift rate probably is typical of other rapidly rising mountain fronts in the Basin and Range Province, most likely being 0.3 to 1.0 m/ky. The minimal degradation seen here is in large part the result of the arid, thermic to hyperthermic climatic setting. A much different climate influences the triangular facets of Figure 2.20B. The northwest front of the Southern Alps of New Zealand is being raised along the oblique-reverse range-bounding Alpine fault that dips under the range. Quartz-biotite schist offers little resistance to erosion after being weathered in the extremely humid, mesic, weakly seasonal climate. Deep valleys dissect the triangular facets. One might erroneously conclude that this landscape is indicative of a slow bedrock-uplift rate. Instead, this is one of the fastest rising major mountain fronts in the world - rock uplift and ridgecrest uplift is about 5 to 8 m/ky Chapter 2 swan Figure 2.20 Lithologic and climatic control of tectonic landforms illustrated by a comparison of triangular facets. A. Mountain front along the southwest side of arid Saline Valley in southeastern California. The mountain-piedmont junction coincides with a normal fault. The slightly rilled lower surface, just above the mountain-piedmont junction, has a homogeneous appearance because it is fault gouge. Contrasting litholo-gies are obvious higher on the slope where the thin layer of gouge has been removed by erosion. Local patches of colluvium and alluvium cling to the fault plane such as at the top of the waterfall at the left side of the view. (Bull and Cooper, 1986; Yetton and Nobes, 1998). Valley-floor surface uplift is < 1.0 m/ky because these big rivers with large annual stream power have impressive stream-channel downcutting rates. Increases of bedload size and amount may have contributed to modest long-term increases in stream-channel gradient in reaches upstream from the range-bounding Alpine fault. The main divide of the Southern Alps is high partly because sustained rapid erosion of deep valleys promotes isostatic compensation that further increases the altitudes of peaks and ridgecrests. The factors that influence surface uplift (Fig. 1.4) can be used to elaborate on the usefulness of tectonic landforms such as triangular facets. Surface uplift more closely approximates bedrock uplift in Saline Valley because erosion is mini- F3. Mountain front along the northwest side of extremely humid southern Alps of New Zealand. The dense rain forest provides little protection against rapid erosion of schist. The mountain-piedmont junction coincides with the oblique-reverse Alpine fault. mal. This favors preservation of triangular facets as a tectonic landform. However, rapid erosion of weathered rocks creates tectonic landforms suggestive of relatively less bedrock uplift in the Southern Alps. The magnitude of climate-controlled erosion is large enough to affect styles of crustal faulting (Koons, 1989; Norris and Cooper, 1995). Average surface uplift for Southern Alps watersheds surely is reduced by rapid erosion of the landscape, but still exceeds that of less tectonically active Saline Valley. The importance of climatic control on landscape evolution demonstrated by this comparison underscores the difficulty of using landforms for quantitative estimates of bedrock or surface uplift rates. Alternatively, one can rank qualitative classes of surface uplift based on assemblages of tectonic land-forms within a given climatic province (Chapter 4). The concept of relaxation time also applies to the consequences of Pleistocene-Holocene climatic change. Reaction times typically are brief when protective plant cover is changed and hillslope soils undergo net erosion instead of net accumulation. The pulse of valley-floor alluviation caused Concepts for Studies by stripping of the hillslope sediment reservoir has a relaxation time of only a brief 1 to 3 ky in hot deserts (Bull, 1991), but is much longer for vegetated hillslopes of humid regions. Density of hill-slope plant cover is not changed as much, and the volume of soil and colluvium is an order of magnitude greater. The relaxation time of Japanese watersheds to the Pleistocene-Holocene climatic change exceeds 10 ky and may be a factor in the present high watershed sediment yields (Oguchi, 1996). 2.5.4 Time Spans Needed to Erode Landforms Tectonic geomorphology studies focus mainly on the past 10 to 100 ky in areas of accelerated landscape evolution (rapid bedrock uplift, soft rocks, and extremely humid climate) and on more than 10 My in slowly changing pedimented landscapes of some arid regions. Hills and streams continue to change after tectonic uplift of mountains has virtually ceased. The time needed for erosion to create landforms indicative of tectonic stages of landscape evolution ranges from less than 1 ky to more than 1,000 ky. The time span needed for each landform noted on the left side of Figure 2.21 is a function of uplift, of Rising Mountains 57 rock resistance, and volume of material to be eroded after cessation of uplift. Only a short time is needed for the concentrated power of a stream to remove a small volume of unconsolidated alluvium to create a fanhead trench. Immense time spans (>10 My) are needed to consume the last vestiges of an uplifted planar surface. Such escarpment retreat is accomplished by gradual weathering of bedrock and slow erosion of hillslopes, and the volume of rock to be removed is huge. Isotopic ages allow rough estimates of the times needed to erode landforms in the Mojave Desert and the Coast Ranges of California. Potassium-argon ages for volcanic materials in mountains and basin fill range from 0.5 to more than 5 Ma. Granitic and metamorphic rocks predominate in the arid Mojave Desert, and soft mudstone and sandstone predominate in the semiarid to subhumid Coast Ranges. Erosion rates vary with climatic setting by at least two orders of magnitude (Fig. 2.20). The Mojave Desert and Coast Range plots on Figure 2.21 are separated by approximately an order of magnitude. The sheared granitic and metamorphic rocks of the semiarid to subhumid Transverse Ranges occupy an intermediate position. Estimates of denudation rates based on amounts of sediment trapped in 450 Eliminate planar uplands, form circular drainage basins Form pediment with inselbergs Eliminate triangular facets Embay mountain front Erode U-shaped valley Entrench alluvial fan 104 105 106 107 Sequence of erosional stages Years since cessation of uplift Figure 2.21 Piagram comparing estimated times needed for changes in landforms after cessation of active uplift for different climates and rock types in 10 km2 fluvial systems. The Coast Range and Mojave Pesert stages are spaced on the ordinate so that most points approximate straight lines. Plots without control points have less dating control. A. Sheared and fractured greywacke sandstone in humid New Zealand. 3. Soft mudstone and shale in the semiarid central Coast ranges of central California. C. Sheared and altered granitic and metamorphic rocks of the subhumid San Gabriel Mountains of southern California. Quartz monzonite in the arid Mojave Pesert of California. E. Gneissic and granitic rocks in the extremely arid Sinai Peninsula of Egypt. After Bull (1955). 58 Chapter 2 Tectonically induced downcutting Figure 2.22 Sketch of two longitudinal stream profiles graded to similar sea-level highstands. Rivers erode down into rising mountains and then widen their valley floors by beveling strath surfaces when they are not able to downcut further. In this case tectonically Induced downcutting between times A and 3 has left the downstream reaches of the longitudinal profile as a strath terrace passing through the rising landscape. debris basins (Scott and Williams, 1978; Brown and Taylor, 1982) suggest rapid denudation of the San Gabriel Mountains at about 1.5 m/ky Extremes of rates of landscape evolution are represented by the easily eroded fractured greywacke sandstone and schist of the extremely humid Southern Alps of New Zealand and by the extremely arid Sinai Peninsula. A wide range of spatial and temporal scales of investigation is needed for the overdue incorporation of landscape analyses as an integral component of the plate tectonic paradigm. Process-oriented studies emphasize small spaces and time spans as short as the elapsed time since a recent earthquake (Arrowsmith and Rhodes, 1994). At the other extreme, spaces can be as large as mountain ranges, or entire tectonic provinces, and time spans may exceed 10 My (Davies and Williams, 1978; Oilier, 1982). 2.6 Tectonically Induced Downcutting Streams incise ever deeper as bedrock is raised into the powerful buzz saw of stream-channel downcutting. Amounts and rates of tectonically induced downcutting are functions of vertical tectonic displacement rates, excess unit stream power (equation 2.6), and resistance of earth materials to degradation. Downcutting by small ephemeral streams flowing over resistant welded tuff may be unable to match a bedrock-uplift rate of 0.1 m/ky; such reaches remain on the erosional side of the threshold of critical power. Downcutting by perennial rivers flowing over soft rock easily keeps pace with bedrock uplift of 5 m/ky. But stream-channel downcutting occurs only during appropriate climatic and tectonic conditions. The tendency of streams to cut down to the minimum gradient needed to transport their sediment load has been a long standing fundamental concept in fluvial geomorphology (Powell, 1875; Mackin, 1948; Leopold, Wolman, and Miller, 1964; Leopold and Bull, 1979; Bull, 1991). Headwater reaches of streams in rising mountains tend to stay on the degradational side of the threshold of critical power, but downstream reaches, with their greater unit stream power, are more likely to attain the base level of erosion through the process of tectonically induced downcutting (Fig. 2.22). 2.6.1 Straths, Stream-Gradient Indices, and Strath Terraces Many streams return to the base level of erosion after tectonically induced downcutting is interrupted by aggradation events that temporarily raise the streambed. The Charwell River, New Zealand fluvial system (Figs. 2.23, 2.24) is sensitive to both tectonic and climatic perturbations; it has frequent climate-change induced aggradation events, numerous internal-adjustment terraces, and occasional times when the stream bevels its valley floor to create a tectonic landform - a major strath. Prior piedmont valleys, with their flights of Pleistocene stream terraces, have been preserved. Their rich history of landscape evolution has been set to one side as a result of rapid right-lateral displacement of the watershed by the Hope fault, which is at the mountain front-piedmont boundary. The following discussion focuses on the present-day valley and its flight of terraces, whose creation was modulated by several late Quaternary global climatic changes. The mere presence of either marine or strath terraces has tectonic significance. Only one sea-level highstand was higher than the present high stand during the past 350 ka. It occurred at about 125 ± 5ka Concepts for Studies Figure 2.23 Prainage networks of the Main and Right Branches of the Charwell River, New Zealand, k is knickpoint migration from Hope fault since ~9 ka. K is where several knickpoints have accumulated to create a large step in the streambed. Numbers are for Strahler (1952, 1964) stream ordere for a third order tributary. From Infomap 260 031, New Zealand Pepart-ment of Survey and Land Information. when the oceans were about 5 to 6 m above present sea level (Chappell, 1983, 2001; Chappell et al., 1996; Israelson and Wohlfarth, 1999). So just the presence of coastal shore platforms higher than 6 m shows that the land is rising relative to the sea-level datum. Similarly, the presence of flights of paired strath terraces shows that the terrain is rising, relative to the long-term base level of erosion of the stream. Dating the times of formation, and measuring heights, of either strath or marine terraces provides estimates of bedrock-uplift rates. Reaches of the Charwell River upstream and downstream from the range-bounding Hope fault have different styles of response to uplift. This part of the Seaward Kaikoura Range is rising three of Rising Mountains 59 times faster than the adjacent piedmont reach. Rock mass strength also is much greater in the mountains, where unit-stream power becomes progressively less farther upstream. The longitudinal profile in the mountains has the characteristics of a disequilibrium stream, whereas the river flowing down the piedmont easily achieves type 1 dynamic equilibrium. Tectonic strath terraces of the Charwell-River reach downstream from the front of the Seaward Kaikoura Range illustrate the importance of this landform to tectonic geomorphologists. The longitudinal profile is much more concave upstream from the range-bounding Hope fault and the average gradient is fivefold less downstream from the fault (Figs. 2.23, 2.24). The marked change in concavity of the two reaches mainly reflects rapid uplift of more resistant mountain bedrock, and pronounced overall widening of active-channel streamflow in the piedmont reach. Downstream increases of discharge and decrease in size of bedload are only moderate in this short distance, so may be less important than litho-logic and tectonic controls. Slower uplift and softer rocks in the piedmont reach favor prolonged attainment of the base level of erosion at the conclusion of degradation events that follow pulses of aggradation. Valley-floor portions of fault zones were buried beneath thick alluvium during aggradation events. The highly irregular longitudinal profile upstream from the Hope fault in part reflects cumulative surface ruptures as much as 40 m that were not able to migrate upstream until bedrock beneath episodic valley fill was exposed to erosion. We need ways to quantify both the irregularities in the longitudinal profile upstream from the Hope fault and degree of smoothness downstream from the fault zone. Stream-gradient indices are introduced as a valuable concept here. John Hack used characteristics of large rivers in the humid Appalachian Mountains of the eastern United States to define a stream-gradient index that describes influences of many variables that influence the longitudinal profiles of stream channels (Hack, 1957, 1973, 1982). Equilibrium adjustments, termed hydraulic geometry, assume orderly interactions between streamflow variables. Hydraulic geometry of stream channels is based on stream-gauging data, and typically has an order of magnitude scatter of data when used in logarithmic regressions of discharge and streamflow characteristics (Leopold and Maddock, 1953; Leopold, Wolman, and Miller, 1964). It defines statistical relationships between streamflow Stream length, km Figure 2.24 longitudinal profile of the Charwell River from the headwaters to the junction with the Conway River, South Island, Hew Zealand. SL is stream-gradient index, k is a knickpointthat has migrated upstream from the Hope fault where it originated as a fault scarp beneath alluvium between 26 and 9 ka. K is larger multiple-event knickpoint. From Figure 4 of Bull and Knuepfer (19&7). parameters and channel morphologies. Using the approximate relations provided by hydraulic geometry dispenses with having to measure streamflows in virtually inaccessible localities. Discharge (Q) from a watershed increases as a power function with drainage-basin area {AJ: Q = cAf (2.7) Many studies have compared length down a stream channel, L, from the main divide with drainage basin area, Ad, and have found that L increases at least as rapidly as Ad, (exponent is >0.5), L = bA°-6 (2.8) where the units for L are miles and for Ad are square miles. This exponential function is now revered as "Hack's law" and has been the subject of many re-evaluations (Smart and Surkan 1967; Mueller, 1972; Seidl and Dietrich, 1992; Montgomery and Dietrich, 1992; Pvigon et al., 1996). Experimental watershed studies by Lague et al. (2005) found that power functions of mainstream length increase almost linearly with drainage area, and that Hack's law is not significantly dependent on uplift rate. The systematic decrease in slope as described by concave longitudinal profiles of stream channels is nicely described by a power function between the slope of a reach of a stream, S , and A.. Ss=kAf (2.9) Hack used length, L, as a reasonable proxy for discharge, Q. He tested whether streams had achieved equilibrium by analyzing spatial variations in the product of slope of a reach, AH/ALr (change of altitude/length of reach) and the horizontal length to the midpoint of the reach from the watershed divide, Lsc. This is the "stream gradient index", or SL index, where SL is defined as: SL = 4-?L (2-10) A L sc r Verification of Hack's SL model was achieved when he showed that reaches of Appalachian rivers had fairly constant values of SL for consecutive reaches. The Appomattox River has remarkably constant SL values over a distance of 150 km (Hack, 1982). See the Figure 2.30 example discussed in Concepts for Studies Section 2.7. This implies that 1) larger "bankfull" flow events were cumulative responses of the entire watershed, 2) streambed hydraulic roughness is constant downstream, and 3) bedload size and transport rate remained about the same downstream. Analyses of stream-gradient index should be classed as narrow or inclusive. Narrow pertains to an anomalous stream-gradient index. It typically is only a short reach that describes only one or several contour intervals. It is useful for describing magnitudes of longitudinal profile abnormalities caused by locally high rock mass strength and/or knickzones that have migrated upstream from a source of base-level fall perturbations. Inclusive pertains to long reaches of a longitudinal profile that have a constant rate of longitudinal-profile decrease in gradient associated with progressively larger streamflows from a headwater divide source area. Adjacent reaches with dissimilar inclusive gradient-indices describe variations in longitudinal profile caused by factors such as adjustments to spatially variable uplift rates (Keller and Rockwell, 1984), changes in rate of downstream increase of stream discharge, change in the direction of a valley, and change in median particle size of gravelly stream-beds. Narrow gradient-indices describe disequilibrium reaches of streams. Inclusive gradient-indices can be used to describe situations of type 1 or type 2 dynamic equilibrium. Narrow SI values of 1300, 97, and 1900 for adjacent reaches upstream from the Hope fault (Fig. 2.24) record the inability of the Charwell River to smooth out some irregularities in the longitudinal profile. These anomalies result from frequent large tectonic displacements of the streambed and variable rock mass strength of greywacke sandstone. The heights and present positions of these knickpoints are also a function of late Quaternary climate change. Stream-gradient index analyses may not apply equally well to all streams. I suspect that this index describing the behavior of an erosional fluvial system should not be used where streams are aggrading. Of course, a brief pulse of deposition that has uniform thickness would not change analysis results because erosional processes prior to deposition created the form of the longitudinal profile. Application to ephemeral streams should proceed with caution, especially where most convective-storm rainfalls generate flash floods over only part of a watershed. Infiltration of streamflow into a dry streambed results in progressive decrease of Rising Mountains 61 in discharge, which is opposite to the trend of the large perennial rivers where Hack defined the stream-gradient index. Hack's model assumes that longitudinal-profile concavity results from ever-increasing stream power in the downstream direction. Concentration of sediment load concurrent with decreasing discharge of water can move a degrading ephemeral streamflow closer to, or across to the depositional side of, the threshold of critical power. For such reasons, ephemeral streams typically have longitudinal profiles that are much less concave than for humid region watersheds that generate bankfull discharges of similar size (Wolman and Gerson, 1978). We can expect stream-gradient indices to increase downstream, even in equilibrium reaches, where ephemeral streamflow behavior has constant or even decreasing stream power in consecutive downstream reaches. Hack also noted that long equilibrium reaches of perennial rivers plotted as a straight line on semilogarithmic regressions of altitude H and In of L . Each long reach can be numerically described by an inclusive gradient-index. Such linear relations are described by H = C-k(lnL) (2.11) C is a constant and k is the inclusive gradient-index (slope of the regression). The derivative of equation 2.10 with respect to L is streambed slope, S: c, _dhj d(kinL) k dL- dL ~ L (2.12) The inclusive gradient-index can be estimated by regression analysis or by using data points from the longitudinal profile: (H-H) k= ' 1 llnL.-lnL) (2.13) where H and H. are the altitude and distance from the watershed divide for an upstream point on the stream channel, and H and L are for a downstream point on the longitudinal profile. Examples from the Charwell River are introduced here. The Right Branch of the Charwell River is presently beveling a strath as it flows from the Hope fault to its junction with the Main Branch. 62 Chapter 2 500 450 - < 400 350 ..65 S.75 &.&5 &.95 9.05 In (Horizontal distance (L) along valley floor from watershed divide) 450 350 < 250 i i i i i i i i i i i i i i i i i i i i i i i i i i i Junction with Right Branch Crossover point sed for both plots 50 I__................_■ ■ ■ ■ &.& 9 9.2 9A 9.6 93 10 10.2 In (Horizontal distance (L) along valley floor from watershed divide) _L Figure 2.25 Inclusive stream-gradient indices for piedmont reaches of the Charwell River. A. Right Branch from Hope fault to junction with the Main Branch. B. Main Branch from Hope fault to the junction with the Conway River. The crossover point is used in both regressions. A semi-log regression of altitude and distance from the headwater divide (Fig. 2.25A) indeed plots as a straight line. The perfect correlation coefficient is in part due to soft mudstone bedrock beneath the stream channel, and the lack of topographic obstacles. Of course correlation coefficients tend to be high where cumulative altitude is regressed against cumulative distance. The inclusive gradient-index for this fairly small stream is 260. A similar analysis for the Main Branch also demonstrates attainment of equilibrium conditions. The complete dataset has a correlation coefficient of 0.994 and an inclusive gradient-index of 214, just what one would expect for a stream whose watershed is three times larger than that of the Right Branch. Alternatively, the Main Branch can be modeled as two reaches with different characteristics. This improves the correlation coefficients slightly. The trend of consecutive points for the reach upstream from the crossover point does not reveal where the Right Branch enters the Main Branch or where the Main Branch narrows where it flows through a gorge cut in massive sandstone. The upstream reach flows straight down the piedmont at a bearing of 170°. The river impinges on Flax Hills at the location of the crossover point. This topographic obstacle deflects the course of the river by 50°, changing the direction of the valley to a bearing of 230°. The steeper regression trend for the downstream reach has an inclusive gradient-index of Concepts for Studies 245. The contrast between gradient indices of 184 and 245 supports treating these as two datasets, thus counteracting the initial impression. Evaluation of the equilibrium stream channels of the Right Branch and Main Branch for reaches just downstream from the Hope fault is more appropriate for comparison of gradient-indices for these humid region streams, 184 for the Main Branch and 260 for the Right Branch. Straths are beveled along reaches of streams at equilibrium - where the inclusive gradient-index remains constant from reach to reach. It is useful to view strath formation in a context of a strath-formation threshold defined as the effective stream power needed to mobilize streambed materials, and above which the stream can do the work of beveling a strath. This threshold is reached more often in downstream reaches of a stream as is suggested by the general lack of straths in the upstream half of most watersheds. Climatic and lithologic variables play critical roles in determining the wide range of conditions affecting the strath-formation threshold for a specific drainage basin. Rainfall-runoff magnitudes and rates greatly affect peaks and durations of streamflows and amount and size range of suspended and saltating sediment. Such interactions between the fluctuating variables over the long term affect the numerical values for the inclusive gradient-index. A broad modern strath in the reach immediately downstream from the Hope fault has been beveled across soft sedimentary rocks (Fig. 2.26) and is indicative of prolonged attainment of type 1 dynamic equilibrium. This late Holocene valley floor is a nice example of a steady-state landform. Peter Knuepfer (1988) did the weathering-rind dating of exposed of Rising Mountains 63 boulders on the treads of the flight of degradation terraces (Fig. 2.16). The lower strath terraces of the flight are shown here. The scarp of the oblique right-lateral Hope fault is the bushy riser at the far end of the pasture. This is a nice example of attainment of equilibrium for a specific reach of the fluvial system (Fig. 2.25B). The Charwell River quickly re-established the base level of erosion here many times after brief departures during the past 4 ka. These minor variations in streambed altitude are merely the product of the normal fluctuations in the spectrum of discharge of water and sediment. Strath terrace heights in the Flax Hills reach, 7 to 8 km downstream from the Hope fault, were surveyed using the modern strath as a reference level (Bull, 1991, Fig. 5.19). It is 2.5 ± 0.5 m below the active channel of the Charwell River. Seven radiocarbon ages on fossil wood collected from basal aggradation gravels just above several older straths reveal that tectonically induced downcutting in this reach is 0.37 ± 0.03 m/ky. Assuming that this bedrock-uplift rate was uniform during the past 200 ka, one can estimate strath ages by dividing strath height by tectonically induced downcutting rate. For example, the strath presently at 30.3 ± 0.5 m above the modern strath is estimated to have formed at about 82 ka. 30.3 m 30.3 m/ky' :82+8ka (2.14) Times of tectonic strath formation occurred at approximately 0, 40, 54, 62, 72, 82, and 114 ka in the Flax Hills reach. Times of strath terrace forma- Figure 2.26 Reach of the Charwell River downstream from the Hope fault has been at the base level of erosion. The resulting steady-state landform is a 400 m wide strath cut in soft Ceno-zoic basin fill that is capped with a veneer of stream gravel. This surface of detrital transport since 4 ka continues to be lowered at a rate equal to the rock-uplift rate. 64 Chapter 2 Strath height, m Marine terrace age, ka Inferred strath age, ka 0 0 0 * 29 # 14.9 40 40 44 # 20.1 53 54 23.1 59 62 26.6 72 72 30.3 81 82 96 # 100 # 42.1 118 114 124 53.7 No match for 54 m strath 145 62.0 176 168 Table 2.2 Relations between technically induced valley-floor down-cutting and inferred ages of major straths of terraces along the Flax Hills reach of the Charwell River, South Island, Hew Zealand. Marine-terrace ages are from Chappell and Shackleton (1936); and Shackleton (1937). * River cut a strath near the mountains, but only incised part way through the Flax Hills aggradation-event alluvium in the study reach. # Tectonic strath of this age is present in another reach of the River tion at about 29, 44, 96, and 100 ka were observed in other reaches, but not in the Flax Hills reach because 1) erosion has removed the strath, 2) the strath was not exposed at the time of my survey, 3) or insufficient vertical separation to distinguish between adjacent straths because of the locally low bedrock-uplift rate. Strath terrace age estimates for the Charwell River (Table 2.2) coincide with the ages of dated global marine highstands of sea level (Chappell and Shackleton, 1986; Gallup et al., 1994). The coincidence between the inferred ages of Charwell River straths and the isotopic ages of global marine terraces is the result of similar timing of climate-change modulation of marine coastal and fluvial geomorphic processes during the late Quaternary Both the Charwell fluvial system and the coastal marine system are controlled by global climatic changes that fluctuate between full-glacial and interglacial extremes. Times of rapid aggradation of New Zealand valleys occurred at times of maximum accumulation of ice on the continents and lowstands of glacio-eustatic sea levels. Downstream reaches of these streams easily attain type 1 dynamic equilibrium during interglacial climates at times that coincide with the times of maximum melting of continental ice masses and attainment of sea-level highstand. Hillslope plant cover and geomorphic processes in the Charwell River watershed were greatly different for these two regimes (Bull, 1991, Chapter 5). The 29, 40, and 54 ka tectonic straths of the Charwell and nearby rivers can be identified by radio- carbon and luminescence dating of the adjacent overlying deposits. Together with the modern (0 ka) strath, they provide readily accessible time lines to assess local rates of tectonically induced downcutting of the valley floor. For example, the 0 and 29 ka straths can be used to estimate the bedrock-uplift rate for a reach of the Charwell River that is 1 km downstream from the Hope fault. Analyses of weathering-rind thicknesses on surficial greywacke cobbles by Knuepfer (1984, 1988) were the basis of the 10.8 ± 1.9 ka age estimate for the tread of the fill-cut degradation terrace shown in Figure 2.27. The lack of paleosols or beds of loess in the basal 23 m of uniformly massive sandy gravels is suggestive of a single pulse of aggradation of Stone Jug gravels. The process of returning to the Charwell River base level of erosion involves stream-channel downcutting through the aggradation event gravels and then through a thickness of bedrock equal to the total rock uplift since the time of the pre-aggradation event strath. Holocene degradation of an additional 39 m below the 29 ka strath occurred in this reach. The rate of tectonically induced downcutting is 39m 29ky = 1.3 ± O.lm/ky (2.15) Two additional points are illustrated by Figures 2.26 and 2.27. Even the small 30 km2 main branch of the Charwell River may easily attain the base level of erosion in downstream reaches for sufficiently long time spans to bevel tectonic straths (type Concepts for Studies of Rising Mountains 65 climate change in much of the South Island of New Zealand are nicely synchronous because of fairly similar topography lithology and humid climate. Brief regional aggradation events can also have nontectonic origins such as regional coseismic landslides (Hancox et al., 2005). Termination of periods of strath formation is much more likely to be synchronous. Studies of the Greenland ice cores give us a better appreciation of the rapid onset of major climate changes (Alley 2000; Peteet, 2000). The age uncertainties for isoto-pically dated times of global marine terraces are less than the ±5 to ±10 ky uncertainties for strath ages that are a function of surveying and uplift-rate-calibration errors (see equation 2.14 for an example). 2.6.2 Modulation of Stream-Terrace Formation Figure 2.27 View of 11 ka Charwell River fill-cut terrace. About 23 m of gravel lie on a tectonic strath that formed at about 29 ka. The 39 m between the buried strath and the present tectonic strath reflects the amount of tectonically Induced river downcutting since 29 ka: the basis for estimating an uplift rate of 1.3 ±0.1 m/ky. 1 dynamic equilibrium). Second, it is not necessary to preserve a complete section of aggradation gravels in order to identify the aggradation event that buries a tectonic strath. The excellent agreement between ages of global marine terraces and local tectonic stream terraces ties the Charwell River terrace chronology to a global climatic chronology. Global climatic changes result from variations in the Earth's orbital parameters - the astronomical clock - (Berger, 1988). The similarity of the Table 2.2 pairs of ages may permit assignment of ages for straths with less radiocarbon dating control. Potential dating uncertainties for straths older than 40 to 50 ka include violation of the assumption that both systems have similar response times to global climatic perturbations. This can happen for aggradation events because watershed characteristics are sufficiently variable to result in crudely synchronous, or even diachronous, aggradation surfaces for a suite of adjacent watersheds. Watershed responses to by Pleistocene—Holocene Climatic Changes Times of formation of tectonic landforms commonly reflect other important variables of fluvial systems such as annual unit stream power (a measure of a stream's capacity to do work). Times of tectonic strath formation along the Charwell River were largely controlled by the rather overwhelming influence of late Quaternary climatic changes. Climatic-change impacts of watershed geomorphic processes raise and lower the streambed at rates faster than the concurrent bedrock uplift caused by the sum of tectonic forces and isostatic adjustments (Fig. 2.28). The piedmont reach of the Charwell River was either aggrading or was catching up to new base levels of erosion. This reach was raised by a combination of uniform rapid bedrock uplift and intermittent valley-floor backfilling of 30-60 m. Aggradation events were the dominant process during the Pleistocene, whereas the Holocene has been characterized by degradation. In order to occasionally catch up to a new tectonic base level of erosion the stream had to degrade through the most recently deposited valley fill, and then through a thickness of bedrock equal to the amount of bedrock uplift since the last time the stream attained type 1 dynamic equilibrium. The Charwell River barely had enough time to bevel a new tectonic strath after attaining the base level of erosion, before the onset of the next aggradation event. These brief episodes of attainment of equilibrium allow comparison between the times of strath cutting with the times of solar insolation maxima and sea-level rise. The agreement 66 Chapter 2 Degradation \owere the valley floor Stream cute down 'and remains at new base 'of erosion —I-1-1- Valley floor raised by climate-change induced aggradation Times of tectonic strath terrace formation^ 120 Q Q Time, ka would not be nearly as nice (Table 2.2 and Figure 2.28) for larger streams that may remain at the base level of erosion for 60-90% of the time. One reason for distinguishing between type 1 and type 2 dynamic equilibrium is that tectoni-cally induced downcutting can be used to estimate uplift rates only when comparing situations of type 1 dynamic equilibrium. These streams have parallel longitudinal profiles of stream terraces that indicate return to similar combinations of variables for reaches where channel width is less than valley-floor width. This is not the case for type 2 dynamic equilibrium streams incising into bedrock. Longitudinal profiles may be concave and exponential, but unfortunately we can only examine the present assemblages of land-forms because type 2 streams do not create suitable landforms that are preserved. Consider the Grand Canyon reach of the Colorado River in northern Arizona. Active normal faulting at the western end of the Canyon during the Quaternary caused tectonically induced downcutting of roughly 0.4 m/ky but this decreased to 0.2 m/ky at 100 km upstream in the eastern reach of the Canyon (where strath terraces are more likely). It appears that this 100 km long reach has been steepened 400 m in the past 2 My. Steeper gradient and narrower channel width are the obvious consequences, but changes in hydraulic roughness may be just as profound. These several alterations do not let us use changes in the altitude of the longitudinal profile to estimate uplift rate. Furthermore, influx of large boulders from cliffy tributary streams does not allow the river to behave Figure 2.28> Changes in the streambed altitude of the Charwell River, New Zealand reflect the combined influence of tectonic and climatic controls during the past 45 ka. Tectonic strath terraces are created only during brief time spans that follow climate-change modulated episodes of tectonically induced down-cutting. Simplified from Figure 5.24 of Bull (1991). as a system of interrelated reaches (Figs. 2.19C, D). Waterfalls in the Charwell River upstream from the Hope fault also limit use of fluvial landforms to estimate uplift rate. 2.7 Nontectonic Base-Level Fall and Strath Terrace Formation Not all strath terraces represent time lines in tectonically deforming landscapes. So let us clarify other aspects of this valuable landform with examples of the few exceptions to what might have seemed a general rule in the preceding discussions. The most obvious nontectonic strath is an unpaired terrace resulting from local lateral migration of a stream into a bedrock hillslope. Such a nontectonic strath could even form while a valley floor is being slowly raised during the terminal stages of an aggradation event. Pauses in a degradation event may temporarily allow a stream to bevel either fill-cut surfaces in alluvium or strath surfaces in bedrock before the stream has downcut sufficiently to return to a new base level of erosion. These common erosion surfaces are nontectonic internal-adjustment terraces (Charwell River at - 14 to 4 ka (Fig. 2.16) for example). Base-level falls can be induced by climatic perturbations to fluvial systems as well as by uplift. A good example is from the piedmont along the mid-Atlantic coast of the eastern United States. Isostatic uplift continues at a very slow rate in response to Concepts for Studies gradual erosion of the Appalachians and tectonic uplift generally is so minor as to be trivial. So, this would seem to be an improbable region to observe large amounts of stream-channel downcutting below prominent strath levels. But, beautiful, prominent strath surfaces occur along the lower reaches of large rivers that have cut spectacular bedrock gorges just above their terminal tidewater reaches. The Great Falls are in the terminal reach of the Potomac River west of Washington, DC (Fig. 2.29). The prominent strath surface was beveled across the highly resistant late Proterozoic sandstone and schist of the Mather Gorge Formation. Cosmogenic 10Be and 27. River (basin-position coordinates of 0.71 to 1.00). It is presently at its base level of erosion and beveling a tectonic strath since 4 ka. Reach D has achieved type 1 dynamic equilibrium as described by the Figure 2.25 equation. Reach B is still actively downcut-ting, but plots as a straight line on this semi-log plot, even though the stream flows through steep, rugged landscape. The central section of reach B, between altitudes of 800 and 600 m, has an inclusive gradient-index of 250 (same as reach D). So, lacking straths, reach B is a nice example of type 2 dynamic equilibrium between basin-position coordinates of about 0.27 and 0.52. Convex reach C (0.52 to 0.71) is in disequilibrium because of 40 m of uplift along the Hope fault. Post-26 ka reaction time (-17 ky) to surface ruptures is abnormally long. The stream could not degrade into bedrock as long as latest Pleistocene climate-change induced stripping of the hillslope sediment reservoir kept the system strongly to the aggradational side of the threshold of critical power. Exhumation of the sub-alluvial fault scarp that started at about 9 ka initiated knickzone retreat, which has progressed only a few hundred meters upstream. Long-term rock uplift favors maintenance of type 2 dynamic equilibrium conditions in reach B. Response times may be so long that the next climate-change aggradation event may occur before the 40 m of cumulative Hope fault displacements, that occurred between 26 to 9 ka, extends upstream to a basin-position coordinate of 30. Lack of fill terrace remnants in reach B suggests that it was sufficiently steep to remain on the degradational side of the threshold of critical power. Continued stream-chan- of Rising Mountains 73 nel downcutting might reduce or eliminate convex reach C. Climate-change induced aggradation seems to occur only in reach D, a base-level rise that spreads upstream far enough to bury the Hope fault for the duration of an aggradation event. Headwater reach A (basin-position coordinates of 0.00 to -0.27) is persistently degrading and unable to achieve equilibrium because of low annual unit stream power relative to the rock mass strength of materials beneath the trunk stream channel. Surprisingly the Right Branch fluvial system is neither sensitive (long reaction time) nor efficient (long relaxation time) in its response to a large tectonic perturbation emanating from the mountain front. System adjustments here are strongly modulated by climatic and lithologic controls. Chapter 2 tectonic concepts should be applied in the context of watershed climatic and lithologic controls on geomorphic processes. This helps us better understand the significance of external factors such as late Quaternary global climate change, sea-level fluctuations, and vertical tectonic deformation. The base level of erosion is the reference datum for studies of tectonics and fluvial topography. The threshold of critical power separates degradation and aggradation modes of operation of fluvial systems. It is purposely defined as a multi-variable ratio to remind us not to overemphasize a single variable, such as streamflow gradient, when trying to comprehend fluvial-system behavior. Time lags of response help us focus on the frequencies and magnitudes of tectonic and climatic perturbations, their locations within a fluvial system, and the magnitudes and time spans of departures from equilibrium conditions that such perturbations usually cause. Bedrock uplift has a major influence on geomorphic processes and landscape evolution. Relief orographically controls precipitation and temperature, and defines potential energy of flowing water even where tectonic elevation of mountains ceased long ago. Increases of fluvial-landscape relief emanate from active geologic structures through the process of tectonically induced downcutting. Streams act as connecting links that transmit tectonic perturbations to upstream reaches. Active faults and folds separate fluvial reaches with vastly different processes and landforms. Degradation changes to piedmont aggradation where a stream crosses an active range-bounding fault. Let us explore how bedrock uplift affects mountain fronts in the next two chapters. Mountain Fronts 3.1 Introduction Mountain-front escarpments have caught the attention of humankind for centuries. Settlers of the American west viewed distant mountains as a change of scene and as an impending challenge. Geologists wonder if active faults and folds separate mountains from lowlands. Active geologic structures in topographic escarpments are zones of concentrated tectonic base-level fall for fluvial systems. One challenge is to discern which mountain fronts have fault zones that are sufficiently active to generate damaging earthquakes. Progressive urban encroachment onto mountainous escarpments occurs after gently sloping land is occupied or when homeowners seek impressive views from their residences. Residential construction on tectonically active escarpments, such as in Los Angeles, California, Salt Lake City, Utah, and Wellington, New Zealand, increases earthquake and landslide risks. Steep, high mountain fronts can be menacing but the next surface rupture may occur along the more subtle low fronts and scarps. In order to assess potential earthquake-related hazards the paleoseis-mologist needs to identify and date key tectonic land-forms and apply her or his knowledge of landscape evolution. Tectonic geomorphologists are faced with diverse questions when viewing a sea of suburban development that laps onto foothills of lofty mountains (Fig. 3.1). How old is the escarpment, and what are the past and present rates of uplift? How seismogenic are the pulses of mountain-building uplift? Has a steady-state balance been achieved between uplift and denudation of the mountain slopes? (Lave and Bur-bank, 2004), or is a model of continuously changing landscape more appropriate? Earth scientists, engineers, and planners benefit from geomorphic tectonic assessment of whether or not range-bounding fault zones are active or inactive. How long has it been since the most recent surface rupture, and when will the next one occur? What advice should be given to those seeking to bulldoze low piedmont fault scarps in order to build new housing subdivisions, or to those already admiring Image showing the proximity of the Los Angeles metropolitan area (lower-right coastal plain) of southern California to the imposing mountain front of the rugged San Gabriel Mountains which are rising >2 m/ky. Pacific Ocean in foreground. Mt. Baden Powell at the right side rises to 2,866 m. Shuttle Radar Topographic Mission perspective view with Landsat overlay; image PIA02779 courtesy of the Jet Propulsion Laboratory and NASA. Chapter 3 Figure 3.1 Urban development encroaching onto a thrust-faulted mountainous escarpment east of Cucamonga Canyon, San Gabriel Mountains, southern California. Both the high mountains and the lower structural bench are being raised along thrust faults. Less obvious active faults rupture the urbanized piedmont alluvial fans. their views from homes built on the crests of high fault scarps? Clearly there is a need to do more than merely describe the locations and types of faults present. With each passing year Quaternary earth scientists are better able to define the locations and magnitudes of future surface ruptures, and to estimate the rates of uplift along faults associated with low and high escarpments. The San Gabriel Mountains and the thrust faults along its south flank are associated with a bend in the strike-slip San Andreas fault. Right-lateral movements along this restraining bend cause local crustal shortening, so the thrust and strike-slip styles of faulting are intimately related. It is logical for pa-leoseismologists to ask "do synchronous surface ruptures of the San Andreas and thrust faults occur as a single mega-earthquake event?" Alternatively, thrust-fault earthquakes occur independently. The hazard implications for the Los Angeles metropolitan region are profound. The San Fernando (U.S. Geological Survey, 1971), Whittier Narrows (Hauksson et al., 1988), and Northridge (Hudnut et al., 1996) earthquakes demonstrate the seismically active nature of the mountain front and adjacent basin (Dolan et al., 1995). We can expect more damaging earthquakes especially if the major range-bounding Sierra Madre-Cucamonga fault (Fig. 3.3) ruptures. Will 40-90 km of this fault rupture synchronously with the next surface rupture of the San Andreas fault? The result would be a Mw magnitude >8.0 earthquake. Mw is earthquake moment magnitude (Hanks and Kanamori, 1979). Thrust faults in the Elkhorn Hills (Sieh, 1978a), and perhaps elsewhere, ruptured during the 300 km surface rupture of the San Andreas fault in 1857. An appraisal of the Mw magnitude 8.3 Gobi-Altay, Mongolia 1957 earthquake (Bayarsayhan et al., 1996; Kurushin et al. 1997) serves as a useful prototype. It's surface-rupture length was 250 km. The spatial arrangement of thrust and strike-slip faults is remarkably similar to the Cucamonga and San Andreas faults. They conclude that the probability of such an event is speculative, but "the similarities are too great for the possibility of such an event to be ignored". Although the rapid contractional strain rate between the coast and the San Gabriel Mountains dictates big thrust-fault earthquakes (Dolan et al., 1995), it is rather unlikely that they will occur concurrently with the next San Andreas strike-slip earthquake (Hough, 1996). Mountain Fronts 77 Figure 3.2 Image showing the tectonic setting of the San Gabriel Mountains, which appear in the lower right part of the view as a lens of raised terrain caught between the San Andreas fault (prominent diagonal slash at right side) and the range-bounding thrust faults on the Los Angeles side of the mountain range. The Garlock fault at the top right of the view bounds the north side of a wedge of technically quiet terrain in the western Mojave Desert. Shuttle RadarTopographic Mission perspective view with Landsat overlay; image PIA03376 courtesy of the Jet Propulsion Laboratory and NASA. The purpose of Chapter 3 is to review several ways to assess the hazard potential of tectonically active mountain-front landscapes. Mountain fronts are created by diverse styles of faulting and folding, so the overall theme is landscapes that respond to tectonic base-level fall. I apply the conceptual models of Chapter 2 and use fluvial landforms to better resolve several specific problems. Stream channels, terraces, and faulted alluvial fans are used to determine which thrust faults are capable of producing the next earthquake, and to measure the true throw of normal-fault surface ruptures. Passive margin escarpments fall outside of the paleoseismology emphasis of this book. Erosion-in- Los Angeles Pomona Mountain front Major fault zone Stream I Watershed Figure 3.3 Location map for place names and illustrations pertaining to the San Gabriel Mountains. 78 Chap duced isostatic uplift of tectonically inactive escarpments is fascinating. Examples include Drakensburg in South Africa (Gilchrist et al., 1994; Brown et al., 2002; van der Beek, 2002), Blue ridge in the eastern United States (Spotila et al., 2004), and Western Gnats bordering the western side of the Deccan Plateau in India (Ramasamy, 1989). The term "mountain front" pertains to more than the topographic junction between the mountains and the adjacent piedmont. A mountain front is a topographic transition zone between mountains and plains. This landscape assemblage includes the escarpment, the streams that dissect it, and the adjacent piedmont landforms. Our discussion starts with the diagnostic land-forms of triangular facets, mountain-piedmont junctions, and piedmont forelands of an active mountain range bordering the Los Angeles metropolitan area in southern California (Figs. 3.2, 3.3). Triangular facets evolve during a million years of erosion and episodic uplift of resistant rocks of arid regions (Fig. 2.20A). Formation of mountain-piedmont junctions may be likened to a contest between the relative strengths of uplift along a range-bounding fault zone and fluvial dissection. Locations of active thrust faulting typically shift from mountains into adjacent basins. Piedmontforelands are newly raised and deformed blocks between the new and old thrust faults. These low piedmont scarps are easy to overlook, but may eventually rise to become impressive escarpments with the passage of geologic time. The Gurvan Bogd mountains of the Gobi Altay, Mongolia were formed by a system of strike-slip faults with a reverse component. The magnitude Mw-8.0 earthquake of 1957 has attracted paleoseis-mologists from around the world to study the marvelous scrunch tectonics of this remote arid region. Low ridges rise through broad piedmonts of >3,000 m high mountains and roughly parallel the older mountain fronts. Florensov and Solonenko (1963) used the term "foreberg" for these hills created by the complexities of scrunch thrust faulting (Kurushin et al., 1997). Such folds, antithetic and synthetic faults, and elongated backtilted ridges result from the shortening component on a broad active intracontinental fault zone. They have a common function, which is to broaden the deforming zone by creating new structures that accommodate both strike-slip and scrunch shortening components of tectonic deformation. The landscape further suggests that these new structures evolve by lateral propagation, increase in 3 Figure 3.4 Scrunch tectonics of piedmont forelands and forebergs. Illustration and caption are from Figure 5 of Bayasgalan et al. (1999). A Cartoon of a transpressional "flower structure" adapted from Sylvester (19&8>). B. Cartoon of the internal deformation within a foreberg, based on observations at Gurvan Bulag, Mongolia. Note the flattening of the underlying thrust at very shallow depths, which is probably responsible for the collapse of the thrust "nose" by normal faulting (Kurushin et al., 1997) and the left-stepping back-thrusts and right-stepping normal faults, which suggest a component of left-lateral slip. C. Cartoon showing migration of active faulting away from the main range front, leaving uplifted and dissected fans in the hanging wall of the newfaultand older, abandoned faults and shear fabrics within the uplifting mountain range. Mountain Fronts 79 amplitude, and may eventually merge and form new fault zones of considerable length. The evolution described here is thus peculiar to strike-slip faults with a reverse component, and can form many of the features of the "flower structures" that are often described in such oblique-shortening zones (Fig. 3.4A). The interplay between the rates of sedimentation and erosion allows some elevated fans between the foreberg ridges and the mountains to be much less dissected than would otherwise be expected, because the rising foreberg is a base-level rise (Fig. 3.4B). Few arid-region streams have sufficient annual unit stream power to accomplish the tectoni-cally induced downcutting needed to cross the rising landform (Owen et al., 1997). The streams of the highly seasonal semiarid San Gabriel Mountains dissect such piedmont forelands easily. Range-bounding faults of the lofty mountain range become much less active as tectonic deformation is transferred to the newest outermost fault zone (Fig. 3.3C). Piedmont foreland and foreberg shapes result from changes in thrust-fault dips of the underlying thrust faults in the uppermost 200 m and also at depths or more than 1 km. Add features like fault-bend and fault-propagation folds and it is easy to see why each structural geologist devises a different tectonic scenario for a study region. I use the model of Ikeda and Yonekura (1979) and Ikeda (1983) for the San Gabriel Mountains, where characteristic suites of fluvial landforms document shifts in the locations of scrunch tectonics. Chapter 3 also explores normal-fault landscapes. I evaluate a conceptual model for segmented surface-rupture behavior of active faults, and then apply the fault segmentation model to a normal fault in Idaho. The best way to test the characteristic earthquake model is to make measurements of historical and prehistorical surface ruptures with sufficient precision to define surface rupture behavior in the boundaries between fault segments. Such responses of hills and streams to episodic surface ruptures are then used in Chapter 4 to identify active range-bounding faults and to discern regional patterns of uplift rates of mountain fronts. 3.2 Tectonically Active Escarpments Hills record long-term interactions between uplift and landscapes. Compared to streams, hills respond slowly to the cumulative effects of many small increments of uplift along active geologic structures. Long response times to uplift are due primarily to huge volumes of rock that have to be weathered into erod-ible-size fragments before tectonic landforms such as triangular facets can be created (Menges, 1987, 1990a, b). Mountainous topography is the consequence of fluvial erosion initiated by the first pulse of uplift. Mountains continue to evolve for millions of years after tectonic uplift has ceased. However, isostatic uplift (Fig. 1.4) continues in response to denudational unloading. The resulting landscape assemblages record the rates and magnitudes of rock uplift and concurrent fluvial erosion; both processes increase relief. The topic of mountainous escarpments is part of the much broader subject of hillslope development whose erosion is initiated by base-level fall. Many of the early papers are worth reading. Hill-slope processes and forms are reviewed by Young (1972), Carson and Kirkby (1972), and Cooke and Warren (1973); important papers about hillslopes include those by Gilbert (1877), Horton (1945), Strahler (1950,1957), Schumm (1956), Leopold and Langbein (1962), Hack (1965), Abrahams (1994), and Anderson and Brooks (1996). Significant papers about the effects of uplift on mountainous escarpments include those by Davis (1903), Louderback (1904), Blackwelder (1934), Gilbert (1928), King (1942, 1968), and Wallace (1977, 1978). 3.2.1 Faceted Spur Ridges The splendid triangular facets of the Wasatch Range escarpment in north-central Utah have been a classic example of a tectonic landform since the time of William Morris Davis (1903). Blackwelder (1934), Hamblin (1976), and Wallace (1978) describe triangular facets as being fault planes that have been modified by erosion, an explanation that seems appropriate for mountains bounded by normal faults. Triangular facets result from base-level fall, and occur in a variety of tectonic settings. Erosion of facets at the truncated ends of spur ridges may be associated with normal faults (Fig. 3.5A), anticlines (Fig. 3.5B), thrust faults (Fig. 3.5C), and even along escarpments formed by erosional base-level fall (Fig. 3.5D). The overall similarity of the facets shown in Figure 3.5 is suggestive of a more general relation than erosional notching of normal-fault planes. The faceted ends of the spur ridges are steep hillslopes that reflect recent cumulative range-front uplift. Sharp-crested spur ridges divide an escarp- 80 Chapter 3 Figure 3.5 Triangular facets of different tectonic environments. A. Spur ridges truncated by a normal fault on the east side of the Toiyabe Range, central Nevada. merit into drainage basins. Each spur ridge terminates at the range front in a characteristic triangular outline. Initial development of faceted spurs is similar, even where uplift is along a reverse fault that dips into the mountains (Fig. 3.5C). An early stage consists of crudely planar 20° to 40° hillslopes. Keller and Pinter (2002, p. 10) nicely depict the key topographic and stratigraphic features for many normal faults (Fig. 3.6). Uplift of range fronts in west-central Nevada proceeds as 1 to 3 m surface ruptures every 5 to 10 ky (Wallace, 1978). These may seem small and infrequent, but the resulting topography is spectacular and the landforms are truly indicative of the relative rates of rock uplift and erosion. Wallace's block diagrams depict the evolution of a fault-generated mountainous escarpment (Fig. 3.7) that reflects the long-term (>10 My) interactions between uplift and denudation. Initial faulting (stage A) creates a linear scarp crest that migrates away from the range boundary A range crest is created by merging of scarp crests that migrate from the range-bounding faults on opposite sides of the rising block. Valley floors notched into the rising block (stage B) are zones of most rapid tectonically induced downcut-ting. The intervening spur ridges and the range crest are gently sloping. These landscape elements have the slowest rates of tectonically induced degradation (stages C and D). The mountain-piedmont junction continues to be straight and the valley floors narrow during continuing rapid uplift, even where rocks are Figure 3.5 Triangular facets of different tectonic environments. Triangular facets on the north side of the Wheeler Ridge anticline, south edge of the San Joaquin Valley, California. The topographic benches may be the result of mass movement processes (Bielecki and Mueller, 2002). soft. An aerial view of the Tobin Range (Fig. 3.8) reveals the simplicity of the terrain from which Wallace developed his concepts of landscape evolution. Rugged faceted spur ridges owe their substantial heights to sustained tectonic base-level fall at the range-front landscape boundary, and to the profound initial difference between valley-floor and ridgecrest rates of denudation. Increase of hillslope steepness and relief also increases the rate of hillslope erosion. Landslide processes become more important as the ever steeper valley side slopes become progressively more unstable (Pain and Bowler, 1973; Pearce and Watson, 1986; Keefer, 1994; Hovius et al., 2000; Dadson et al., 2003, 2004). The style of landscape evolution reverses after cessation of rapid uplift, but hillslopes never attain a steady-state condition. Hillslope denudation rates exceed uplift rates, so the mountain-piedmont junction becomes sinuous as it retreats from the position of the range-bounding fault to create a pediment (stage E of Fig. 3.7). Non-steady state denudation brings ridgecrests closer to the valley floors, which remain at type 1 dynamic equilibrium with valley floors that become progressively wider. Dissection of range front triangular facets proceeds independently of the stream-channel down-cutting of the adjacent trunk valley floors. Consequent drainage nets on young triangular facets initially consist of parallel rills. Capture of flow from adjacent rills occurs. Small elongate watersheds form on these planar surfaces and become more circu- Mountain Fronts 81 C2 / 3 N^v. / 2 ^\ ^\ \ DB J \ ___J- -''" /---"' Houses Figure 3.5 Triangular facets of different tectonic environments. C1. Aerial view of a set of triangular facets that terminates the spur ridge of a thrust-faulted mountain front of the San Gabriel Mountains, Southern California. C2. Facet-dissection stages as described in Table 3.1. Younger facets 1 and 2 are nested inside older higher stage 3 and stage 4 facets that are deeply incised by small stream channels. DB, basins to catch debris swept off recently burned steep hillslopes during winter storms. lar through the processes first described by Horton (1945). Development of progressively larger drainage nets at the ends of spur ridges concentrates available stream power, promoting efficient erosional destruction of the triangular facets. Planar facets with numerous closely spaced parallel rills (Fig. 2.20A) are eventually transformed into a deeply dissected ridge-and-ravine topography (Fig. 2.20B) in which the triangular shape of the facet is less obvious (Fig. 3.9). Erosional dissection of faceted spurs can be described as six stages (Table 3.1). The stages are easy to distinguish in the field, on aerial photographs, or on detailed topographic maps. The time needed to achieve a given stage is a function of two compensating processes. Uplift increases facet height, and fluvial erosion deepens valleys. With the passage of time, faceted spurs adjacent to an active fault become higher and more dissected. The lowest, most recently created, part of a facet is less dissected (Fig. 3.5C) because it has been exposed to erosion for time. These qualitative descriptions of facet dissection are used to help define mountain front tectonic activity classes in Chapter 4. Figure 3.5 Triangular facets of different tectonic environments. U. Triangular facets along the edge of a fanhead embayment at Cucamonga Canyon, San Gabriel Mountains, southern California. These nontectonic facets were created by lateral erosion induced base-level fall caused by streamflow. 82 Chapter 3 V-shaped canyon and remnants of stream terraces record tectonically induced downcutting Figure 3.6 Diagrammatic sketch of the topographic expression of an active normal--fault system. Uplift on a range-bounding normal fault creates a base-level fall that causes deep valleys to be eroded in the mountain block. This tectonic displacement favors accumulation of thick alluvial-fan deposits downstream from the normal-fault zone. From Keller and Pinter (2002, Fig. 1.7C). Figure 3.7 Block diagrams showing the sequential development of a fault-generated mountainous escarpment. A. Initial faulting creates a linear scarp. F3. Scarp crest migrates away from the rising range boundary to form range crest. C. Valleys are notched into the rising block; their floors are the locations of rapid tectonically induced downcutting by streams. The crests of spur ridges are the locations of slow tectonically induced degradation. D. 'Episodic displacement along the range-bounding fault maintains a steep, straight mountain-piedmont junction. Main and spur ridge divides continue to rise faster than degradation can lower them. E. The mountain-piedmont junction becomes sinuous and the valley floors become wider after cessation of uplift. Relief becomes less as degradation lowers the ridgecrests. Figure 4 of Wallace (1977). Mountain Fronts 83 Figure 33 View of the west side of the normal-faulted terrain of the Tobin Range, west-central Nevada showing the basic landscape elements of a tectonically rising landscape as described by Wallace (1977). These include the range crest, spur ridges extending to triangular facets at a straight range front, and a deep valley with a narrow valley-floor width. The irregular dark line at the mountain front is the surface rupture of the 1915 earthquake. The planimetric trace of this topographic transition between mountains and piedmont is useful for assessing whether or not the mountain front coincides with an active range-bounding fault zone. The sinuosity of the mountain-piedmont junction represents a balance between 1) the tendency of uplift to maintain a sinuosity as low as that of the range-bounding fault or fold, and 2) the tendency of streams to erode an irregular junction between the mountains and the plains. Straight mountain-piedmont junctions generally indicate the presence of an active fault. Embayed, pedimented mountain-piedmont junctions suggest tectonic quiescence. Downstream Facet class Erosional landforms 1 Planar surface with only rills. Includes scarps that have yet to be carved into facets by streams flowing across the scarp. 2 Planar surface with shallow valleys extending a short distance into the facet. 5 Valleys extend more than 0.7 the horizontal distance between the base and top of the facet. 4 Peep valleys extend more than 0.7 the horizontal distance. 5 Greatly dissected but the general form is still obvious. 6 So dissected that the general form of a facet is not obvious. 7 Triangular facets are not present because they have been removed by erosion, or they never existed. Table 3.1 Stages of dissection of triangular facets. 3.2.2 Mountain—piedmont junctions Transitions between mountainous escarpments and adjacent basins typically are abrupt. Steep hills give way to gentle piedmont slopes in both tectonically active and tectonically inactive landscapes. Piedmonts may consist of either the depositional environment of coalescing alluvial fans, or the erosional environment of pedimented terrain. Fans and pediments may be smooth in the arid realm. They tend to be dissected and less obvious in humid regions where floodplains are a common piedmont landform and forests may cloak subtle features of the landscape. 84 Chap Figure 3.9 Aerial view of triangular facets near La Canada, San Gabriel Mountains, Southern California. Stage 4 of Table 3.1. increase in stream power maximizes the potential for downcutting and lateral erosion where streams leave the mountains. The result is a highly sinuous mountain-piedmont junction, even in homogenous rocks, but only under tectonically inactive conditions. Small structures, such as joints, foliation, and bedding planes, also influence sinuosity of mountain-piedmont junctions. Tectonically inactive mountain fronts with structures that parallel the range front may have an anomalously straight mountain-piedmont junction and well defined triangular facets. Sinuosity of the mountain-piedmont junction also is a function of the width of a mountain range (Parsons and Abrahams, 1984; Mayer, 1986). Wide mountain ranges have large drainage basins that are more likely to have sufficient stream power to quickly attain the base level of erosion and create pediment embayments after uplift has ceased. Range width decreases with erosional retreat of range fronts. Drainage-basin size, and stream power, become less. Mountain-piedmont junction sinuosity may become lower as the mountain landscapes are progressively replaced by the beveled bedrock of pediments. The constraint of drainage-basin size on sinuosity of the mountain-piedmont junction is illustrated by the McCoy Mountains of southeastern California. Geophysical studies by Rotstein et al. (1976) suggest that the faults initially bounding the McCoy Mountains structural block now are 1 to 2 km from the present range front. The mountains are only half of their original width. Average drainage-basin length 3 Figure 3.10 Tectonically inactive mountain fronts of the McCoy Mountains in southeastern California. A. Map of the mountain-piedmont junction and watershed divides. L and H are low and high sinuosity mountain fronts. 1-5 are study watersheds. for fronts at L (Fig. 3.10A) is only about 1 km and the sinuosity of the mountain-piedmont junction is moderate. Mountain fronts are highly sinuous at H where drainage basins are twice as large. Granitic rocks weather slowly and streamfiow is ephemeral in this arid, hyperthermic, moderately seasonal climate (Table 2.1). Unit stream power is large during flash floods during infrequent incursions of tropical depressions into the southeastern Mojave Desert or during some wintertime cyclonic storms. The stream channels of the McCoy Mountains degrade by abrasion and plucking during floods, but long-term weathering of the granitic rocks into small particles plays a much larger role than in humid settings such as the Potomac River (Fig. 2.29). McCoy Mountain Fronts 85 Figure 3.10 Tectonically inactive mountain fronts of the McCoy Mountains, California. F3. Topographic map of watershed 4. Contour interval is 50 m. Gray is mountain bedrock. Pattern is piedmont alluvium whose extent defines the present mountain-piedmont junction. Heavy dashed line is present trace of ends of spur ridges. J ( near 3) is stream junction. 1, 2, 3, 4 are defined in Figure 3.10C. fluvial systems have had >10 My to greatly modify this landscape. There has been ample time for erosion of bedrock stream channels to achieve equilibrium relationships, in marked contrast to a rapidly rising mountain front such as the Charwell River (Figs. 2.16, 2.24-2.28). Planimetric and longitudinal-profile aspects of fluvial-system equilibrium in a tectonically inactive watershed are shown in Figures 3.10 B, C. The 2 km wide strath at the mountain front is a pediment embayment indicative of ample time and stream power to achieve prolonged type 1 dynamic equilibrium. Diminishing watershed area resulting from 2 km of mountain front retreat to its present position is offset by the prolonged time span for the ephemeral stream to do this work. Pediments are not a special landform when viewed in terms of processes. They form where stream(s) at the base level of erosion bevel straths that coalesce to form the beveled bed- rock piedmont landform that we call a pediment. It takes millions of years for this process to remove spur ridges between adjacent drainage basins. The smooth slightly concave longitudinal profile in the reach between locations 3 and 4 records attainment of type 1 dynamic equilibrium conditions. The upstream narrowing of the pediment embayment (Fig. 3.1 OB) reflects the importance of concomitant spatial decrease of unit stream power and the relative increase of the importance of rock mass strength of materials beneath the stream channel. Stream power prevails at the mountain front but eventually a threshold is crossed where unit stream power is insufficient to overcome rock mass strength. So the upstream end of the pediment embayment coincides with the junction (J) of the two largest streams in this drainage basin. Stream power upstream from this junction is insufficient to bevel broad valley floors in this rock type. * . « » in». 'I in». « 'I in». « '' 1--r 2 3 Distance, km Figure 3.10 Tectonically inactive mountain fronts of the McCoy Mountains. C. Longitudinal profile of trunk stream channel of watershed 4 of figure 3.10A. 1 is watershed divide. 1-2 is disequilbrium reach. 2-3 is type 2 dynamic equilibrium reach. 3-4 is type 1 dynamic equilibrium reach. Vertical exaggeration is 4.0. 86 Chapter 3 Valley-floor widening of the strath extends only a short distance further to location 3 (Figs. 3.10B and 3.IOC). The longitudinal profile becomes steeper upstream from 3, but still has a form indicative of attainment of equilibrium. But, in contrast with the downstream reach (inclusive gradient-index is 129), the valley floor is narrow (inclusive gradient-index is 200). So, reach 2-3 is best regarded as being type 2 dynamic equilibrium. Disequilibrium conditions prevail in the headwater's reach, 1-2. Unit stream power is minis-cule relative to rock mass strength. Rates of stream-channel downcutting are so slow that disequilibrium prevails even after 10 My of tectonic quiescence. The watershed area needed to generate sufficient stream power to achieve type 1 dynamic equilibrium in the McCoy Mountains is partly a function of drainage net configuration (Fig. 3.10D). Headwaters reaches, such as 1-2 or 1-3, occupy only a small portion of total watershed area in large drainage basins, so equilibrium conditions are achieved at a basin-position coordinate (Section 2.8) of only 0.25. An example is watershed 1 of Figure 3.10A. The location at which type 1 dynamic equilibrium is attained is farther downstream where headwater's source areas are smaller. 3.2.3 Piedmont Forelands Normal, reverse, and strike-slip faults associated with tectonically active landscapes may be classed geomor- 6 as «4 ws in i_