UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Study of the stormflow hydrology of small forested watersheds in the Coast Mountains of Southwestern… Cheng, Jie-Dar 1976

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1976_A1 C49_4.pdf [ 10.33MB ]
Metadata
JSON: 831-1.0093910.json
JSON-LD: 831-1.0093910-ld.json
RDF/XML (Pretty): 831-1.0093910-rdf.xml
RDF/JSON: 831-1.0093910-rdf.json
Turtle: 831-1.0093910-turtle.txt
N-Triples: 831-1.0093910-rdf-ntriples.txt
Original Record: 831-1.0093910-source.json
Full Text
831-1.0093910-fulltext.txt
Citation
831-1.0093910.ris

Full Text

A STUDY OF THE STORMFLOW HYDROLOGY OF SMALL FORESTED WATERSHEDS IN THE COAST MOUNTAINS OF SOUTHWESTERN BRITISH COLUMBIA BY JIE-DAR CHENG B.Sc, National Chung Hsing University of Taiwan, 1969 M.Sc, University of British Columbia, 1972 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY (H Y D R OB$?ERBJ]§C I'PtE IMRI DS3$® IBS M E N T) (HYDROLOGY AND WATERSHED MANAGEMENT) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA November, 1975 c l Jie-Dar Cheng, 1976 In p r e s e n t i n g t h i s t h e s i s in p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . It i s u n d e r s t o o d tha t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department of I n t e r d i s c i p l i n a r y Studies (Hydrology) The U n i v e r s i t y o f B r i t i s h C o l u m b i a 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date A p r i l 2 , 1 9 7 6 ABSTRACT This thesis is comprised of four self contained chapters that report the results of a study on the stormflow hydrology of small forested watersheds in the Coast Mountains of southwestern British Columbia. The chapters discuss the general characteristics of the study watersheds and their instrumentation, the generation of stormflows from small forested watersheds, the stormflow (channel-phase) characteristics of one study watershed with steep topography, and the evaluation of i n i t i a l changes in peak stormflow following logging of another study watershed. Chapter I. The characteristics of the study watersheds with respect to regional climate, physiography, soil hydrologic characteristics and forest cover were evaluated and summarized from available information. Emphasis is placed on the hydrologic characteristics of the watershed soils. The instrumentation of the study watersheds pertinent to the present study is also described. Due to the highly permeable nature of the watershed s o i l s , the physical setting of the study watersheds favor a Rapid response of streamflow to rainstorms. On one study water-shed this rapid response characteristic is reinforced by i t s steep topo-graphy and high drainage density. Chapter II. The problem of stormflow generation from small forested watersheds is dealt with by analyzing results from studies completed by the author and other workers in Jamieson Creek watershed and vicinity and by making f i e l d examinations in the same study area. i i i i i A review is made of stormflow generation models, followed by analyses of rainfall intensity, saturated soil hydraulic conductivity and depression storage of the study area. These analyses revealed that overland flow rarely, i f ever, occurs on coastal watersheds with hydro-logic environments similar to that of the study area. Instead, rain water takes alternate subsurface pathways through the soil to the stream channel. Observations made by the author in the study area and in other watersheds in this coastal region confirmed the existence of these alter-nate routes of water flow. Two types of subsurface stormflow pathways have been identified by earlier workers: (1) the matrix of forest floor and mineral soil beneath and (2) channels within or passing through the mineral s o i l . In the study area most soil channels were developed from dead or decaying roots. After passing through these two types of pathways, subsurface stormflows feed the expanding stream channel system laterally while rainfall is feeding the system from above. Subsurface stormflows are mainly in the form of saturated return flow from the ground and seepage flow through saturated stream banks. The stream channel system expan-sion during, and contraction after, a storm was measured in a small sub-watershed in the study area. It was found that the rate of storm-flow from a watershed was closely related to the rate at which the stream channel expanded in response to the storm. From theestudy i t is concluded that the model of subsurface stormflow from a variable source area is more appropriate than the other two models in describing stormflow generation in this coastal region. iv Chapter III. Stormflow characteristics of Jamieson Creek watershed, a small, steep, and forested watershed in the Coast Mountains of south-western British Columbia, were evaluated by the analysis of 41 storm hydrographs from 1970-1974. During the study period, the rainfall amount per storm event varied from 5 to 330 mm, with the majority of the storm durations ranging from 20 to 60 hours. On the average, the fraction of storm rainf a l l that appeared as stormflow was 44 percent, varying from 2.5 to 81 percent. A significant number of major storms produced stormflow that accounted for more than 60 percent of the storm r a i n f a l l . Instantaneous peak flows varied considerably with storms, ranging from about 10 to 1,370 - 1 - 2 1 s km and appeared to be mainly affected by the rainfall amount and distribution before the occurrence of the peak flow. Rising time (time to the peak) was short, usually within 30 hours, depending upon the rainfall distribution before the occurrence of the peak flow. Lag time was found to be relatively constant and short, ranging from 5 to 15 hours with an average of 8.5 hours. It is suggested that to derive lag time from characteristics of small watersheds, soil hydrologic properties should also be included with those parameters that are gen-erally used. Stormflow amount was highly correlated with rainf a l l amount with 92 percent of its variance being accounted for. Antecedent base-flow rate was proposed as an index of watershed soil water storage prior to the storm hydrograph rise. One set of data from Jamieson Creek water-shed and four additional data sets from two small steep watersheds in V the Coweeta Hydrologic Laboratory were used to assess, through multiple regression analysis, the usefulness of antecedent baseflow rate in improving stormflow-rainfal1 relations. For al l data sets, the inclu-sion of antecedent baseflow as a second independent variable s i g n i f i -cantly improved the stormflow estimate in comparison to that when rain-f a l l amount was the only independent variable. Recession limbs of storm hydrographs varied with individual storms, depending on the degree of recharge to the watershed storage by the storm and the spatial distribution of such storage over the water-shed. The stormflow characteristics of Jamieson Creek watershed re-flect the influence of not only climatic conditions but also watershed characteristics: (1 ) shallow but highly permeable s o i l s , (2) steep watershed slopes and stream channels, and (3) high drariinage density. The stormflow characteristics can be interpreted in terms of the gener-ation of stormflow from a variable source area of the watershed. A comparison of the stormflow characteristics of Jamieson Creek watershed and the adjacent Elbow Creek watershed indicated that storm-flow from the former usually has a sharper peak, higher peak flow ratio and steeper recession than stormflow from the latter, but both have very similar rising times. Differences in the streamflow response of the two watersheds could be caused by their differences in some topographical features. However, these differences also suggest that leakage from Elbow Creek, revealed in a preliminary f i e l d investigation, may deserve more detailed study. vi Chapter IV. This chapter provides the f i r s t quantitative Canadian in-formation with respect to the impact of logging on peak stormflow. The paired-watershed technique was used to evaluate the i n i t i a l changes in peak streamflow during storm periods following logging of a small watershed in the U.B.C. Research Forest, near Haney, B.C. Contrary to the majority of similar studies elsewhere, the analysis indicates that significant peak flow changes after logging occurred as follows: ( 1 ) an increase in the time to the peak, and (2) a decrease in the magnitude of the peak. The changes can be explained by ( 1 ) the degree of ground sur-face disturbance associated with the logging and (2) the stormflow gener-ation mechanisms of the study area. Visual examination after the logging indicated that ground surface disturbance did not reduce the soil in-f i l t r a t i o n capacity to the extent that overland flow resulted. Workers in an earlier study speculated that forest floor disturbance could result in closure of some of the entrances to soil channels, thus increasing temporary water storage in the soil matrix. This, they further specu-lated, would result in reduced subsurface stormflow and, consequently, lower peak flow. The results of the present study tend to support the speculations, that the closure of some soil channel entrances is respon-sible for lower peak flow after logging. However, this study indicated that peak flow magnitude decreased mainly because of the flattening out of the hydrograph as a result of increased time to the peak (delayed peak rather than earlier hydrograph rise,). It is suggested that a lower rate of stormflow transmission through the soil matrix caused this in-creased time to the peak and, consequently, lower peak flow magnitude. vi i Implications of this study for better water management are sug-gested. TABLE OF CONTENTS Page ABSTRACT . . . . . . . . i i LIST OF TABLES x i i i LIST OF FIGURES xiv ACKNOWLEDGEMENTS .\ . x v i i i INTRODUCTION . . . . . . . . . 1 CHAPTER I - THE STUDY WATERSHEDS AND THEIR INSTRUMENTATION. . . . 4 A. INTRODUCTION . 5 B. REGIONAL CLIMATE . . . . . . . . . '. . . . g 1. General Description g 2. Precipitation Characteristics as Indicated by Long-term Records . . 9 a. Seymour River Basin . 9 b. The U.B.C. Research Forest 10 C. WATERSHED PHYSIOGRAPHIC CHARACTERISTICS . . . . . . . . . 13 1. Jamieson Creek and Elbow Creek Watersheds . 13 a. General Features 13 b. Soils . 19 2. Watersheds 1 and 2, U.B.C. Research Forest 26 a. General Features 26 b. Soils 30 D. WATERSHED FOREST COVER 35 1. Jamieson Creek and Elbow Creek Watersheds 35 2. Watersheds 1 and 2, U.B.C. Research Forest 35 vi i i ix Page E. WATERSHED INSTRUMENTATION 36 1. Precipitation . . . . 36 a. Jamieson Creek and Elbow Creek Watersheds . . . . . 36 b. Watersheds 1 and 2, U.B.C. Research Forest 38 2. Streamf 1 ow ' 38 a. Jamieson Creek and Elbow Creek Watersheds 38 b. Watersheds 1 and 2, U.B.C. Research Forest 39 F. LITERATURE CITED '. . 40 CHAPTER II - THE GENERATION OF STORMFLOW FROM SMALL FORESTED WATERSHEDS IN THE COAST MOUNTAINS OF SOUTHWESTERN BRITISH COLUMBIA 43 A. INTRODUCTION . . 44 B. SUMMARY AND REVIEW OF MODELS OF STORMFLOW GENERATION . . . 45 1. Model of Overland Flow from the Entire Watershed Resulting from Rainfall Exceeding Infiltration Rate . 45 2. Model of Subsurface Stormflow from a Variable SSurce Area in the Watershed . 46 ouur 3. Model of Overland Flow after Soil Saturation of a , Partial Area of the Watershed , 47 C. HYDROLOGIC INSIGNIFICANCE OF OVERLAND FLOW IN WATERSHEDS OF THE COAST MOUNTAINS 48 D. PATHWAYS OF WATER MOVEMENT IN FOREST SOIL ' 49 1. Water Flow Through the Forest Floor-andl St>tiT."Matrix • . 5 1 2. Water Flow Through the Soil Channels 51 E. STUDIES OF THE HYDROLOGIC RESPONSE OF FOREST SOILS TO A RAINSTORM 52 X Page F. THE RESPONSE OF THE STREAM CHANNEL SYSTEM TO A RAINSTORM . 54 G. CONCLUSION . ' 57 H. LITERATURE CITED . . . 57 CHAPTER III - THE STORMFLOW CHARACTERISTICS OF A SMALL, STEEP AND FORESTED WATERSHED IN THE COAST MOUNTAINS OF SOUTHWESTERN BRITISH COLUMBIA . . . 60 A. INTRODUCTION 61 B. DATA USED . 63 C. ANALYTICAL METHODS' 64 1. General Streamflow Characteristics 64 2. Hydrograph Analysis . . . . . . . . 64 a. The Selection of Stormflow Events 64 b. Hydrograph Separation 65 c. Stormflow Parameters Evaluated from Hydrograph Analysis 68 d. Recession Analysis . . . . . 69 e. Rainfall-Stormflow Relationships 71 i) Graphical comparison of rainfall input and the response in the streamflow hydrograph . . 71 i i ) Regression analysis of stormflow amount against rainfall amount 71 3. The Role of Antecedent Baseflow.in Ra.infall-Stormflow Relationships 72 a. Rationale 72 b. Statistical Assessment 74 4. Comparison of the Stormflow Characteristics of Jamieson Creek and Elbow Creek Watersheds 75 xi Page D. RESULTS AND DISCUSSION 77 1. General Streamflow Characteristics 77 2. Stormflow Characteristics as Indicated by Stormflow Parameters Evaluated from Hydrograph Analysis . . . . 81 3. Recession Characteristics of Storm Hydrographs . . . . 89 4. Rainfall-Stormflow Relationships 96 a. Graphical Comparison . 96 b. Statistical Relationships between Stormflow Amount and Rainfall Amount . 102 5. The Significance of Antecedent Baseflow Rate in Improving the Statistical Relationship between Stormflow and Rainfall Amount 103 6. Relationships between the Stormflow Characteristics of Jamieson Creek and Elbow Creek Watersheds . . . . 108 a. Hydrograph Comparison . . . 108 b. Statistical Relationships 112 E. SUMMARY AND CONCLUSION . . . . ' 117 F. LITERATURE CITED . : 119 CHAPTER IV— THE EVALUATION OF INITIAL CHANGES IN PEAK STORMFLOW FOLLOWING LOGGING OF A WATERSHED ON THE WEST COAST OF CANADA 124 A. INTRODUCTION 125 B. [LITERATURE REVIEW . . . . 126 C. THE STUDY AREA 127 D. THE EXPERIMENT ' 130 E. ANALYSIS AND RESULTS 133 F. DISCUSSION 135 x i i Page G. CONCLUSION . . . . . 144 H. LITERATURE CITED . . . . . . . . . . . . . . . . . . . . . . . . 145 APPENDIX 1 - PHOTOGRAPHS OF THE INSTRUMENTATION OF THE STUDY WATERSHEDS . . . . . . . . . . . 149 APPENDIX 2 - A LISTING OF STORMFLOW PARAMETERS FROM STORM HYDRO-GRAPHS OF JAMIESON CREEK WATERSHED 155 APPENDIX 3 - THE RELATIONSHIP OF PEAK FLOW MAGNITUDE AND STORM RAINFALL BEFORE PEAK FLOW OCCURRENCE FOR JAMIESON CREEK WATERSHED . . . . . . . . . . . 158 LIST OF TABLES Table Page CHAPTER I 1 Values of Saturated Hydraulic Conductivity for Soils of Jamieson Creek Watershed and Vicinity . . . . 25 CHAPTER II 1 Maximum Expected Short Duration Rainfall Inten-sities of Different Return Periods as Determined from Measurements at Seymour Falls 50 2 Measured or Estimated Saturated Hydraulic Con-ductivity for Soil and Forest Floor of Jamieson Creek Watershed and Vicinity . . . . . . . . . . . . . 50 CHAPTER III 1 The Means and Ranges of Stormflow Parameters for Jamieson Creek Watershed (1970-1974) 82 2 Comparison of Stormflow Parameters of Jamieson Creek Watersheds and Watersheds 35 and 37, Coweeta Hydrologic Laboratory, North Carolina, U.S.A. . . . . 87 3 Values of the Constants in the Double Exponential Equation (4) Determined for Recession Limbs of Selected Storm Hydrographs of Jamieson Creek Water-shed 94 4 Regression Equations Showing the Relationships of Stormflow Amount to Rainfall Amount and Antecedent Baseflow Rate . . ' 106 x i i i CHAPTER IV 1 The Means and Ranges of Peak Flow Parameters for Watershed 1 (W.l) and Watershed 2 (W.2) 132 xi i i LIST OF FIGURES Figures CHAPTER I 1 2 3 4 5 6 7 8 9 10 11 12 Page A map of southern British Columbia, showing the region ( • ) within which the study watersheds are located . . . . . . 6 A map of Seymour River and Capilano River Basins, showing the location of Jamieson Creek and Elbow Creek watersheds . . . . . . 7 A map 6ifr watershed 1 and 2 in the U.B.C. Research Forest . . . . . . . . 8 Monthly distribution of precipitation as averaged from measurements at Seymour Falls, Seymour River Basin (1941-1970) 11 Monthly distribution of precipitation as averaged from measurements at the U.B.C. Research Forest Administration Building (1958-1974) . . . 12 A topographic map of Jamieson Creek watershed (A) and Elbow Creek watershed (B) in the upper Seymour River Basin. Scale, 1:50,000, Contour Interval: 30.5 m (100 feet) 14 Area-elevation curve for Jamieson Creek watershed . '. 15 Landslope distribution curve for Jamieson Creek watershed 16 Longitudinal channel profile for the main stem of Jamieson Creek . 18 Longitudinal profile of Elbow Creek 20 Partial water retention characteristics of disturbed samples from the 0-40 cm, 40-80 cm and 80-100 cm soil layers of Strachan gravelly sandy loam (after Cheng, 1972) 21 Partial hydraulic conductivity characteristics of the 70 cm depth of Strachan gravelly sandy loam; "(Measurements were made in situ, (.after Cheng et al'., 1975)) 22 xiv XV Figure Page 13 Partial water retention characteristics of undisturbed samples from the forest floor of Jamieson Creek water-shed (1 bar = 1 ,022 cm H90) (after Plamondon,. et a l . , 1972) L 23 14 Hydraulic conductivity characteristics of undistunbed samples from the forest floor of Jamieson Creek water-shed (after Plamondon et a l . , 1972) '. . 24 15 A topographic map for watersheds 1 (A) and 2 (B), U.B.C. Research Forest. Scale: 1 to 250,000; Contour interval: 7.6 m (25 feet) 27 16 Area-elevation curve for watershed 1,.U.B.C. Research Forest . . . . . 28 17 Longitudinal streammchannel profile for watershed 1, U.B.C. Research Forest . 29 18 Area-elevation curve for watershed 2, U.B.C.. Research Forest . ' 31 19 Longitudinal stream channel profile for watershed 2, U.B.C. Research Forest . . . . . . . . 32 20 Partial water retention characteristics of disturbed sample from the 0-60 cm and 60-90 cm soil layers of Capilano gravelly sandy loam (after Willington, 1971) 33 21 Partial hydraulic conductivity characteristics of the 68 cm depth of Capilano gravelly sandyloam.((Measured ments were made in situf{after Cheng et a l . , 1975) . . 34 22 A map showing the instrumentation in Jamieson Creek and Elbow Creek watersheds 37 CHAPTER II 1 A stormflow hydrograph of Jamieson Creek and the measured changes in the stream channel network of a sub-watershed in response to a rainfa l l event in October 1973 . 56 xvi Figure Page CHAPTER III 1 1 A diagram illustrating the hydrograph separation method and stormflow parameters used in the study (after Hewlett and Hibbert, 1967) . . . . 67 2 Mean monthly distribution of streamflow from Jamieson Creek watershed (1970-1974) . . . . . . . . . 78 3 Hydrograph of Jamieson Creek watershed for the water year of 1970-1971 . 79 4 Mean monthly distribution of precipitation for Jamieson Creek watershed (1970-1974) . . . . . . . . . 80 5 Distribution of peak flow magnitudes with respect to the percentage of a l l occurrences for Jamieson Creek watershed (1970-1974) . . . . . . . . . . . .'. 84 6 Recession limbs of selected storm hydrographs in 1970, 1972 and 1973 for Jamieson Creek watershed . . . 90 7 Recession limbs of selected storm hydrographs in 1971 for Jamieson Creek watershed . 91 8 Recession limbs of selected storm hydrographs in 1974 for Jamieson Creek watershed . . . 92 9 The response of streamflow from Jamieson Creek watershed to a storm in July, 1972 . 97 10 The response of streamflow from Jamieson Creek watershed to two storms in October, 1973 98 11 A comparison between hourly rainfall and streamflow of Jamieson Creek watershed for a storm in July, 1972 . 99 12 The relationships between rainfall amount and stormflow amount of Jamieson Creek watershed (1970-1974) . ' . . 104 13 A comparison of the response of streamflows in Jamieson Creek and Elbow Creek to a storm in July, 1972 . . , 109 14 A comparison of the response of streamflows in Jamieson Creek and Elbow Creek to a storm in November, 1974 . . . . . . . . . . . . . . . 110 xvi i Figure Page 15 A comparison of the response of streamflows in Jamieson Creek and Elbow Creek to another storm in November, 1974 I l l 16 The relation of time to the peak of Jamieson Creek watershed to that of Elbow Creek watershed (1971-1974) . . . . . . . . . . . . . . . 113 17 The relation of peak flow magnitude of Jamieson Creek watershed to that of Elbow Creek watershed (1971-1974) . . . . . . . . . . . . . 115 CHAPTER IV 1 A map of watersheds 1 and 2 in the U.B.C. Research Forest, showing the instrumentation and areas of clearcut , \ 128 2 The relation of time to the peak of watershed 1 to that of watershed 2 136 3 The relation of peak flow magnitude of watershed 1 to that of watershed 2 . 137 4 Total water potential (\b) plotted as a function of depth, during a wetting cycle,before forest floor disturbance. The number against each line is the time in minutes since the beginning of water appli-cation. The reference height at which the gnavi-tational potential equals zero is 45 cm above the soil surface (after de Vniies and Chow, 1973) 5 Total water potential (^ ) plotted as a function of depth in the same plot as figure 4 after forest floor disturbance. The number against each line is the time in minutes since the beginning of water application. The reference height at which the gravitational potential equals zero is 45 cm above the soil surface (after de Vries and Chow, 1973) . . . ACKNOWLEDGEMENTS The assistance, advice, encouragement and guidance of Dr. R. P. Willington, throughout this study and my whole graduate studies is deeply appreciated. To Dr. T. A. Black, I extend my very sincere thanks for his patient help and guidance throughout the analytical and writing phases of this study. To Dr. J. de Vries, I wish to express my deep appreciation for his advice and assistance to this study and for his continuous encourage-ment throughout my graduate career. Special thanks are due Professor S. 0 . Russell and Dr. A. Kozak. They served on my research committee and provided invaluable help during the writing phase of this study. The review of this thesis by Messrs. H. Hunter, H; Coulson and D. Reksten of the Hydrology Division, B.C. Water Resources Service, is highly appreciated. I would like to thank Mr. Dave Helvey, Hydrologist, Pacific Northwest Forest and Range Experimental Station, Wenatchee, Washington for his constructive criticism of this thesis. I wish to extend my thanks to Mr. Kuo-Chi Rai, who was of great assistance in the completion of this study program. To Mrs. Kathleen Hejjas, I extend my very special thanks for her help in the computation of the considerable quanity of data. x v i i i The help provided by Dr. J. P. Kimmins, Dr. M. C. Feller and Mr. Min Tse is appreciated with gratitude. The f i e l d assistance provided by Director J. Walters of the U.B.C. Research Forest and his staff and by the Greater Vancouver Water Board is highly appreciated. This study was supported by the National Research Council of Canada through a post-graduate scholarship to the author (NRC:1560) and a research grant (NRC:67-6123). Additional funding was provided by the Greater Vancouver Water Board. DEDICATION TO DR. B. C. GOODELL AND MY PARENTS & *.as to® m&~n±mife%MU:m XX INTRODUCTION The heavily forested Coast Mountains of British Columbia are characterized by steep slopes, shallow but extremely permeable soils and high annual precipitation ranging from 750 to well over 2,500 mm. Flooding of small streams and rivers by peak stormflow resulting from long duration rains or a combination of rain and snowmelt have caused extreme damage locally and in rare cases over large areas. Clearcutting of the forest has been traditionally a common logging practice in this region. The stormflow characteristics of small streams and their poten-t i a l changes following logging of the watersheds are of major concern to resource managers as well as to the public. For example, the B.C. Forest Service Coastal Logging Guidelines, recognizing that adequately sized, culverts at stream crossingsaareae'ssefiitfa 1 fforS-tt ^ m i 'protection and maintenance of good road conditions, specify that culverts be de-signed ibo handle the 25-year storm. However,, a central problem for foresters and engineers to implement this specification is the i n s u f f i -cient measurement and study of streamflow of the countless small streams in the coastal region. Fishery managers are also interested in the impact of peak stormflow and the change caused by logging on the environ-ment of these small streams where a large proportion of coho, pink and chum salmon, and steel head and cutthroat trout are produced. Such small streams, without lakes in their watersheds to provide some flow con-t r o l , are more susceptible to damage by changes in peak stormflow caused 1 by logging than the larger streams and rivers. It is also considered that the understanding of stormflow hydrology of small watersheds is useful in solving flood problems of both small and larger streams which are of particular interest to the water resources managers. The public, on the other hand, is also increasingly demanding the proper management of these renewable resources and the protection of the environment against damages caused by flooding of the streams and rivers. A reliable base of knowledge of the interactions between forests and peak stormflow is needed in developing guidelines for the proper management of our forest, water and fishery resources in the Coast Moun-tains of B.C.; however, there is no such quantitative information avail-able from studies conducted in the area. This dissertation reports the results from a stormflow hydrology study conducted on four small forested watersheds in the Coast Mountains in southwestern B.C. This dissertation is comprised of the following four chapters: Chapter I. The Study Watersheds and Their Instrumentation Chapter II. The Generation of Stormflow from Small Forested Water-sheds in the Coast Mountains of Southwestern British Columbia. Chapter III. The Stormflow Characteristics of a Small, Steep, and Forested Watershed in the Coast Mountains of Southwestern Br British Columbia Chapter IV. The Evaluation of Initial Changes in Peak Stormflow Following Logging of a Watershed on the West Coast of Canada 3 Chapter II of this dissertation was presented as a paper, under the senior authorship of the candidate, at the Canadian Hydrology Sym-posium in Winnipeg, August 11-14, 1975. The material contained in Chapter IV, with only minor changes, has been accepted as a paper, also under the senior authorship of the candidate, for presentation at the International Hydrology Symposium in Tokyo, December 1-8, 1975. The present study is the f i r s t of its kind conducted in the coastal region of British Columbia. It is hoped that i t will contri-bute to the understanding of the stormflow behavior of small streams in this region and elucidate some of the natural and man made factors which are involved in causing flood problems as a result of peak storm-flows. It is also hoped that this study will help stimulate further studies for which the methodology illustrated in this dissertation may serve as an useful example. CHAPTER I THE STUDY WATERSHEDS AND THEIR INSTRUMENTATION 4 CHAPTER I THE STUDY WATERSHEDS AND THEIR INSTRUMENTATION A. INTRODUCTION The objective of this chapter istto provide background infor-mation about the study watersheds for the interpretation of the results in the subsequent chapters and, possibly, for use in future hydrologic studies which may be conducted on these study watersheds. Four small watersheds in the southwest fringe of the Coast Mountains of British Columbia were selected for this study. Two of them, the watersheds of Jamieson Creek and Elbow Creek, are in the upper Seymour River basin, near Vancouver, B.C., while the other two, water-sheds 1 and 2 are in the University of British Columbia (U.B.C.) Research Forest, near Haney, B.C. These two pairs of watersheds are situated in the mountain and foothill regions of the Coast Mountains respectively (Figures 1 to 3). The present stormflow hydrology study of the watersheds of Jamieson Creek and Elbow Creek is a part of a broader research project entitled "The Hydrology of Municipal Water Supply Watersheds of Greater Vancouver Water District," which is currently conducted by the Faculty of Forestry, University of British Columbia. This broader research project which covers the basins of Seymour River and Capilano River 5 6 Figure 1 . A map of southern British Columbia, showing the region ( • ) within which the study watersheds are located. 7 Figure 2. A map of Seymour River and Capilano River Basins, showing the location of Jamieson Creek and Elbow Creek watersheds. 8 Figure 3. A map f r o f watersheds 1 and 2 in the U.B.C. Research Forest. 9 has the three objectives of (1) adding to the existing, meagre knowledge of the hydrologic phenomena of the Coast Mountains of British Columbia, (2) evaluating the hydrologic effects of forestry practices, and (3) determining how forestry practices can be and should be modified to minimize influences detrimental to the water resources and, where possible, maximize influences that are beneficial (Goodell, 1972). The watersheds 1 and 2 in the U.B.C. Research Forest are also the study sites of a research project on the impacts of logging and related operations on the chemistry of stream water (Feller, 1974). B. REGIONAL CLIMATE 1. General Description Influenced by the North Pacific Ocean to the west, the region containing the study watersheds shows a typical maritime climate charac-terized by wet, relatively mild winters and dry, moderately warm summers, and a relatively small annual range of temperature. This climate is modified to a great extent by the rapid northward increase in elevation of 1,500 m (5,000 feet) within 35 km from the Fraser delta to the North Shore mountains. The climate of each pair of study watersheds is also modified by local topographic features. 2. Precipitation Characteristics as Indicated by Long-term Records a. Seymour River Basin The characteristic orographic l i f t i n g of the moist air masses is reinforced by the narrow, steep-walled valley of the Seymour River. 10 The average monthly distribution of annual precipitation based on long-term records from Seymour Falls, about 12 km south of the mouth of Jamieson Creek is given in Figure 4. The mean annual total precipitation is 3 .,7:3.9 mm (147 inches) with October, November and December being the wettest.months, while June, July and August are the driest. Rainfall accounts for 3,490 mm (138 inches) which is more than 93 percent of the mean annual total. The water equivalent of mean annual snowfall, mainly occurring from mid-November to mid-March is 2266mnm (8.9 inches) in depth. Since the watersheds of Jamieson Creek and Elbow Creek are located at higher elevations than the Seymour Falls station, higher annual precipitation and a greater proportion of snowfall are expected. b. The U.B.C. Research Forest The monthly distribution of the mean annual total precipitation based on the long-term record at the weather station at the U.B.C. Re-search Forest Administration Building, about 1 km from the mouth of watershed 1, is given in Figure 5. Most of the mean annual precipi-tation of 2,185 mm (87 inches) f a l l s as rain during the relatively warm and moist winter months. November, December and January are the wettest months and June, July and August are the driest. This type of monthly distribution of precipitation is typically a maritime one. Snow usually comes and goes during the winter months and seldom remains on the ground for more than one to two weeks. Rainfall intensities are generally low, rarely exceeding 15 mm per hour; but long-duration steady rains may total 125 to 150 mm in a day or over 500 mm in the wettest month. 60 (H .500-1 2 z o o UJ o. 4 0 0 H 3 0 0 H 200 A 100 H 0 i » i i i i I 1 1 1 1 u J F M A M - j ' j A S O N D M O N T H Figure 4. Monthly distribution of precipitation as averaged from measurements at Seymour Falls, Seymour River Basin (1941Tl970). F M A M J J A S O N D M O N T H Monthly distribution of precipitation as averaged from measurements at the U.B.C. Research Forest Administration Building (1958-1974). 13 The precipitation amounts in watersheds 1 and 2 are expected to be some-what higher since the watersheds are at a higher elevation than the weather station at the Administration Building. C. WATERSHED PHYSIOGRAPHIC CHARACTERISTICS 1. Jamieson Creek and Elbow Creek Watersheds a. General Features A topographic map with a contour interval of 30.5 m (100 feet) for the watersheds of Jamieson Creek and Elbow Creek is given in Figure 6. The topographic boundary of Jamieson Creek watershed is well defined and the watershed area is 299 ha (740 acres). Jamieson Creek is a tributary of the Seymour River in the southwest fringe of the Coast Mountains, about 30 km north of Vancouver (Figures 1 and 2). The water-shed is underlain by mostly granitic bedrock and is considered to be water-tight. The elevation of Jamieson Creek watershed ranges from 305 m (1,000 feet) at the streamflow gauging station to 1,310 m (4,300 feet) at the highest point of the divide with more than 30 percent of the watershed area lying above 900 m (3,000 feet). The area-elevation curve of this watershed is shown in Figure 7. Watershed land slopes are generally steep (Figure 8) with a considerable area of shallow soils and occasional rock outcrops. The main channel of Jamieson Creek has a general orientation to the southeast, resulting in a majority of the watershed land slopes facing northeast and southwest. The stream 14 Figure 6. A topographic map of Jamieson Creek watershed (A) and Elbow Creek watershed (B) in the upper Seymour River Basin. Scale, 1:50,000. Contour interval-30.5 m (100 feet). 15 I35CH 150 H 0 1 1 1 1 1 " — 20 40 .60 80 100 AREA A B O V E INDICATED ELEVATION(%) Figure 7. Area-elevation curve for Jamieson Creek watershed. 120 I O O H AREA WITH SLOPE GREATER THAN INDICATED VALUE 8. Land slope distribution curve for Jamieson Creek watershed. 17 channels of Jamieson Creek are characterized by the presence of boulders, occasional bedrock exposures and numerous dams of debris from the over-mature forest. The longitudinal stream channel profile for Jamieson Creek is given in Figure 9. The gradient of the stream channel averages about 20 percent; however, some short reaches, particularly in the upper portion of the watershed, have gradients greater than 100 percent. Elbow Creek watershed which was selected as a control for the detection of hydrologic changes in Jamieson Creek watershed, as a result of planned logging in 1977, is topographically ill-defined'with an appro-ximate area of 120 ha (300 acres) (Figure 4). According to Goodell (1972), the choice of Elbow Creek as a stream comparable to Jamieson Creek was based on: (1) a channel reach suitable for accurate stream-flow gauging, (2) a stable, vegetative watershed cover, (3) accessibility, (4) proximity to Jamieson Creek, and (5) non-existence of another stream having near equivalent in desiderata. The poorly defined watershed boundaries prevent quantitative descriptions of physiographic characteristics relevant to watershed area. Generally Elbow Creek watershed has an orientation to the east, but a good portion in the upper part of the watershed, tends to be northeasterly. The land slopes of the watershed are generally steep, as is the case in Jamieson Creek watershed. However, the degree of contour curvature, as can be seen fnom Figure 6, is much less than that of Jamieson Creek watershed. The elevation of the watershed ranges from 275 m (900 feet) at the streamflow gauging station to approximately 1,065 m (3,500 feet). The main channel of the Elbow Creek runs easterly in the major part of 1200 19 the watershed, and southeasterly when approaching the streamflow gauging station. The channel profile of Elbow Creek is shown in Figure 10. Channel gradients are as high as 100 percent in some portions of the creek. Below the 305 m (1,000 feet) contour the channel gradient de-creases to about 1 percent. b. Soils The soils of Jamieson Creek watershed are mostly coarse-textured sands and gravelly sandy loams derived primarily from granodiorite and quartz diorite. Bedrock is generally covered by a layer of glaciated t i l l of extremely variable thickness. The two major types of soils are: steep mountain soils and valley bottom soils. Steep mountain soils are mainly ablation t i l l and colluvium. These soils are characteris-t i c a l l y shallow and very permeable, resulting in a very low water storage capacity. Valley bottom soils are primarily glacio-alluvial soils and lacustrine s o i l s . While glacio-alluvial type soils are very permeable, the lacustrine soils are less permeable due to their varved nature. A more detailed description of the soils and geology of Jamieson Creek watershed can be found in 0'Loughlin (1972) and Zeman (1974). The water retention and unsaturated hydraulic conductivity characteristic curves for the Strachan gravelly sandy loam, which covers most of the watershed's middle and lower slopes, and for a forest floor sample from Jamieson creek watershed are shown in Figures 11 to 14. The saturated hydraulic characteristics for the soils of Jamieson Creek watershed and vicinity have been measured by 0' Loughlin (1972) and estimated by Chamberlin (1972). The values of saturated hydraulic con-ductivity from these two references are summarized in Table 1. 20 21 0-6 o ro 5 o z LU 2 O O CC LU I O CE LU _J O > 10 20 30 40 50 60 70 80 TENSION (CM H 2 0 ) 110 120 Figure 11. Partial water retention characteristics of disturbed samples from the 0-40 cm, 40-80 cm and 80-100 cm soil layers of Strachan gravelly sandy loam (after Cheng, 1972). 22 10 100 TENSI0N(CM H 2 0) 1000 Figure 12. Partial hydraulic conductivity characteristics of the 70 cm depth of Strachan gravelly sandy loam. Measurements were made in si£u";(after Cheng et a l . 1975). 23 FOREST FLOOR DEPTH • • 2 enO • » 6 cm J F HORIZON \ VOLUMETRIC WATER CONTENT (ci Figure 13. Partial water retention characteristics of undisturbed samples from the forest floor of Jamieson Creek watershed (1 bar = 1,022 cm H2O) (after Plamondon et a l , 1972). 24 TENSION (bars) Figure 14. Hydraulic conductivity characteristics of undisturbed samples from the forest floor of Jamieson Creek watershed (after Plamondon et a l . , 1972). 25 Table 1. Values of Saturated Hydraulic Conductivity for Soils of Jamieson Creek Watershed and Vicinity Saturated Hydraulic Conductivity -Inches per Soils Cm per hour hour A. Soils containing no channels^ Sample 1 (parallel to slope) 2 (parallel to slope) 3 (parallel to slope) 4 (parallel to slope) 5 (parallel to slope) 6 (parallel to slope) > 7 (vertical) 8 (vertical) 9 (vertical, unweathered, compacted t i l 1) B. Field soils containing channels^ 20.88 20.16 187167 19.80 2.52 3.60 1.80 10.80 0.07 &7'35o 0,000 8.22 7.94 7.23 7.80 0.99 1.42 0.71 4.25 0.03 ,13.80 o From measurements by 0'Loughlin (1972). From estimates by Chamber!in (1972). 26 The values of hydraulic conductivity in Table 1 indicate that the highly permeable nature of the watershed soils is in large part due to the existence of soil channels which are usually in the form of old root holes (Chamber!in, 1972). The high permeability of the watershed s o i l , the thick but permeable forest floor (Plamondon, 1972), the low intensities of rainfall and negligible presence of frozen soil ensure the entry of rainfall and snowmelt water into the mineral s o i l . Thus there is no observed Horton type overland flow (Horton, 1933). The mechanisms of stormflow generation from small forested watersheds will be discussed in more detail in chapter II. The soils of Elbow Creek watershed are considered to be generally similar to those of Jamieson Creek watershed except that they are some-what shallower. 2. Watersheds 1 and 2, U.B.C. Research Forest a. General Features A topographic map for watersheds 1 and 2 with a contour inter-val of 7.6 m (25 feet) is presented in Figure 15. Watershed 1 has a drainage area of'23.1 ha (57.1 acres) with an elevation ranging from 145 m (475 feet) to 310 m (1,017 feet). The area-elevation curve for watershed 1 is presented in Figure 16. Water-shed .land slopes average 12 percent fior the lower part and 29 percent for the upper part of the watershed. The longitudinal stream channel profile for watershed 1 is shown in Figure 17. The watershed has a general orientation to the south. Figure 15. A topographic map for watersheds 1 (A) and 2 (B) in the U.B.C. Research Forest. Scale: 1 to 25,000; Contour Interval: 7.6 m (25 feet). 360H I2CH 60H 0 1 1 i . r 1 — 20 40 60 80 100 AREA ABOVE INDICATED ELE VATION(%) Figure 16. Area-elevation curve for watershed 1, U.B.C. Research Forest. 29 60 300 600 900 DISTANCE FROM GAUGING STATION(M) Figure 17. Longitudinal stream channel profile for watershed 1, U.B.C. Research Forest. 30 Watershed 2 has a drainage area of 68 ha (168 acres) above its streamflow gauging station. The elevation of watershed 2 ranges from 285 m (950 feet) to 450 m (1,485 feet). The area-elevation curve for watershed 2 is given in Figure 18. The watershed land slopes average about 12 percent. The watershed has a southwesterly to southerly aspect. The longitudinal stream channel profile for watershed 2 is given in Figure 19. The bedrock of the watersheds 1 and 2 consists mostly of quartz diorite, granodiorite or diorite (Roddick, 1965). Outcrops and expo-sures of the quartz diorite display a smooth, superficially weathered surface relatively free of open joints (Feller, 1974), suggesting that the bedrock is generally impermeable. This would prevent significant deep seepage losses from the watersheds. b. Soils The soils are generally sandy loam, overlying unweathered, com-pact basal t i l l and/or bedrock being the most common pattern of the terrain. The so i l s , which contain many interconnected channels, are generally very permeable (Feller, 1974).. Soil frost rarely forms under the protection of thick forest floor. As a result, i n f i l t r a t i o n capa-citi e s of the soil usually remain higher than expected maximum rainfall intensities throughout the year. Overland flow seldom, i f ever, occurs. The water retention and hydraulic conductivity characteristic curves for the Capilano gravelly sandy loam, which covers parts of these two watersheds, are given in Figures 20 and 21. "io" 40 60 80 100 A R E A A B O V E INDICATED E L E V A T I O N ( % ) F i g u r e 18 . A r e a - e l e v a t i o n c u r v e f o r w a t e r s h e d 2 , U.B.C. R e s e a r c h F o r e s t . 32 Figure 19. Longitudinal stream channel profile for.watershed 2, U.B.C. Research Forest. 33 0, (o j ' 1 1 < 1 1 ' 1 1 r o O 2 0 -40 6 0 8 0 (OO 120 140 l&O 180 2 0 0 T E N S I O N ( C M K 2C>) Figure 20. Partial water retention characteristics of disturbed samples from the 0-60 cm and 60-90 cm soil layers of Capilano gravelly sandy loam (after Willington, 1971). 34 = I I I IIIU| I I I 11111J—I I I 1114] TENSION (CM H20) Figure 21. Partial hydraulic conductivity characteristics of the 68 cm depth of Capilano gravelly sandy loam. Measurements were made in situ (after Chenq et a l . , 1 1975). 35 D. WATERSHED FOREST COVER 1. Jamieson Creek and Elbow Creek Watersheds With the exception of the most precipitous bluff and bedrock areas, Jamieson Creek watershed is covered by mature and over-mature coniferous forest. Western red cedar (Thuja plicata Donn), western hemlock (Tsuga heterophylla (Rafn.) Sarg.) and Douglas f i r (Pseudotsuga  menziezii (Mirb.) Franco) are the dominant species on the lower slopes. At approximately 900 m (3,000 feet), the occurrence of mountain hemlock (Tsuga mertensiana (Bong.) Carr), yellow cedar (Chaemacyparis nootkatensis D. Don) Spach), amabilis f i r (Abies amabilis (Dougl.) Ford) is increas-ingly noticeable; while that of western red cedar and western hemlock is decreasing. At this elevation Douglas f i r is almost absent. This marks the lower limits of the Mountain Hemlock Zone, Wet Subzone as described by Krajina (1965). Elbow Creek watershed has a pristine and stable but inaccessible coniferous forest cover (Gbodell, 1972). Both watersheds are within the Southern Pacific Coast Section (c-2) of the Coast Forest Region as described by Rowe (1959). 2. Watersheds 1 and 2, U.B.C. Research Forest These two watersheds are mostly covered by coniferous species of western hemlock, western red cedar and Douglas f i r with some red alder (Alnus rubra Bong.), black cottonwood (Populus trichocarpa Torr. and Gray), big-leaf maple (Acer macrophyllum Pursh) and western white birch (Betula papyrifera Marsh) in the occasional opening or wet site. 36 The relative distribution of the major tree species in these two water-sheds is discussed by Feller, ( 1 9 7 4 ). The distribution of forest cover type for these two watersheds is also given by Feller. Both watersheds are within the dry subzone of the Coastal Western Hemlock Biogeoclimatic Zone of British Columbia (Krajina, 1 9 6 5 ) . Both watersheds are also within the Southern Pacific Coast Section (c - 2 ) of the Coast Forest Region as described by Rowe ( 1 9 5 9 ) . E. WATERSHED INSTRUMENTATION 1 . Precipitation a. Jamieson Creek and Elbow Creek Watersheds Twelve Sacramento-type precipitation storage gauges were estab-lished along specially constructed contour t r a i l s to measure the spatial distribution of monthly and annual total precipitation in Jamieson Creek watershed. Only one gauge was used in Elbow Creek watershed and this was located near i t s streamflow gauging station (Figure 2 2 ) . These gauges were serviced at the end of each month, with those at higher elevations requiring helicopter access during winter months when acces-s i b i l i t y was hindered by heavy snowfalls. In addition to precipitation storage gauges, five recording rain gauges (Belfort weighing-type precipitation gauge) were installed along the contour t r a i l s of Jamieson Creek watershed and one recording rain gauge was installed near the streamflow gauging station of Elbow Creek (Figure 2 2 ) . These gauges are operated during the period from Legend • Snow course • Precipi tat ion storage gauges O Recording rain gauges v Streamflow gauging stations Figure 22. A map showing the instrumentation in Jamieson Creek and Elbow Creek watersheds. The shaded area is the sub-watershed used for the invest i -gation of stream channel expansion discussed in Chapter I I . 38 April to November as snow permits. Weekly service for these gauges is required. Wpttortogtf&pho'&ftlfh'esabbve^mentiion'eddg:aug[e2sisa®fooWrP'fh Appen-.diixerldix 1. b. Watersheds 1 and 2, U.B.C. Research Forest There are weather stations located near each of these two water-sheds. Both stations were instrumented by the U.B.C. Research Forest as part of its meteorological network. Four recording rain gauges, two on each watershed, were installed especially for the present study to provide data on rainfall intensities and occurrence periods. These gauges were operated throughout the year as weather permitted. Weekly service of these rain gauges is required. Precipitation input to the watersheds in the form of snow was measured at the nearby weather station at the Administration Building of the U.B.C. Research Forest. 2. Streamflow a. Jamieson Creek and Elbow Creek Watersheds A 120° v-notch weir and water stage recorder provide an accurate and continuous record of water discharge from Jamieson Creek watershed (Appendix 1). The weir, located in a narrow canyon of Jamieson Creek., was completed by November, 1970. It consists of a 7.3 m (24 feet) wide structure containing a 2.1 m (7 feet) high v-notch. The rectangular spill-over is 1.2 m (4 feet) high. The designed maximum capacity of the weir, 11.3 m 3s _ 1 (400 f t 3 s _ 1 ) , is the estimated 1,000 year storm event. The weir pond stage is monitored by a Leupold and Stevens water 39 level recorder with a manometer attached to the recorder. A nitrogen purge system with servo-manometer and servo-control is ut i l i z e d . An Ott current meter was installed on a bridge spanning the weir pond about 4.6 m (15 feet) upstream from the weir. Water stage is also checked with a hook gauge (point gauge) once every week or two. Water velo-city is measured during high flows with the Ott current meter. The 120° v-notch weir of Elbow Creek is 1.1 m (3.5,tfeet) high with an enclosed pond of 7.3 m (24 feet) long and 6.5 m (21.2 feet) wide. This weir began operating in November, 1971. The weir pond stage is continuously measured by a Leupold and Stevens water stage recorder using a float system (Appendix 1). Servicing of this weir is similar to that for the Jamieson Creek weir. b. Watershed 1 and 2, U.B.C. Research Forest A 120° v-notch weir was constructed at the mouth of watershed 1 in 1971. The height of this v-notch weir is 1.2 m (4 feet). A rec-tangular weir was constructed at the mouth of watershed 2 in 1972. Another 120° v-notoh weir was also installed at about 600 m downstream from the rectangular weir of watershed 2 in 1971 (Figure 1 of Chapter IV). The streamflow data from this second 120° v-notch weir is not analysed in the present study but was used for another study on slash burning and streamflow chemistry (Feller, 1974). All three weirs are equipped with instrument shelters housing s t i l l i n g wells, and Richards-type water stage recorders (Appendix 1). All water stage recorders with spring-wound clocks and 7-day charts were regularly calibrated (approximately once every two weeks) and ad-justed as required. The water stage-discharge relationships of these three weirs were calibrated for low and intermediate flows by using a bucket and stop-watch technique. High flows through the two' v-notch weirs were determined directly from the theoretical curves for the 120° v-notch weir by assuming that the close similarity between the measured and theoretical values for low and intermediate flows would also be the case for the high flows (Feller, 1974). For high flows through the rectangular weir, am Ott universal current meter was used to determine discharge. F. LITERATURE CITED Chamber!in, T. W., 1972. Interflow in the mountainous soils of Coastal British Columbia. In "Mountain Geomorphology," B.C. Geo. Series, No. 14, pp. 121-127. Cheng, J. D., 1972. Evaluating soil water drainage of a humid moun-tain forest site in southwestern British Columbia by two f i e l d techniques. Unpublished M.Sc. Thesis, University of British Columbia, Vancouver, B.C. Cheng, J. D., Black, T. A. and Wellington, R. P., 1975. A technique for the f i e l d determination of the hydraulic conductivity of forest so i l s . Can. J. Soil Sci. 55: 79-82. Feller, M. C , 1974. Initial effects of clearcutting on the flow of chemicals through a forest-watershed ecosystem in southwestern British Columbia. Unpublished Ph.D. Thesis, University of British 41 Columbia, Vancouver, B.C. Goodell, B.C., 1972. Water quantity and flow regime--!*nfluences of land use, especially forestry. In "Mountain Geomorphology," B.C. Geo. Series, No. 14, pp. 197-206. Horton, R. E., 1933. The role of i n f i l t r a t i o n in the hydrologic cycle. Amer. Geophys. Union Trans. 14: 446-460. Krajina, V. J., 1965. Ecology of western North America. University of British Columbia, Dept. Botany, Vol. 1, 112p O'Loughlin, C. L., 1972. An investigation of the stability of the steep land soils in the Coast Mountains, southwestern British Columbia. Unpublished Ph.D. Thesis, University of British Columbia, Vancouver, B.C. Plamondon, A., 1972. Hydrologic properties and water balance of the forest floor of a Canadian west coast watershed. Unpublished Ph.D. Thesis, University of British Columbia, Vancouver, B.C. Plamondon, A., Black, T. A. and Goodell, B. C , 1972. The role of hydro-logic properties of forest floor in watershed hydrology. In "Water-sheds in Transition," pp. 341-348, American Water Resources 'Asso-ciation and Colorado State University. Roddick, J. A., 1965. Vancouver, North Coquitlam and Pitt Lake map areas, British Columbia: with special emphasis on the evolution of the plutonic rocks. Geol. Surv. Can. Mem.'335, 276 p. 42 Rowe, J. S., 1959. ' Forest Regions of Canada. Bull. 123. Forestry Branch, Canada Dept. Northern Affairs and Natural Resources, Ottawa, 71 p. Willington, R. P., 1971. Development and application of a technique for evaluating root zone drainage. Unpublished Ph.D. Thesis, University of British Columbia, Vancouver, B.C. Zeman, L. J., 1973. Chemistry of tropospheric fallout and streamflow near Vancouver, British Columbia. Unpublished Ph.D. Thesis, Univer-sity of British Columbia, Vancouver, B.C. CHAPTER II THE GENERATION OF STORMFLOW FROM SMALL FORESTED WATERSHEDS IN THE COAST MOUNTAINS OF SOUTHWESTERN BRITISH COLUMBIA 43 CHAPTER II THE GENERATION OF STORMFLOW FROM SMALL FORESTED WATERSHEDS IN THE COAST MOUNTAINS OF SOUTHWESTERN BRITISH COLUMBIA A. INTRODUCTION In the heavily forested coast mountains of southwestern British Columbia (B.C.), peak stormflows, caused by prolonged rainfall or a combination of rain and snowmelt, frequently result in flooding. L i t t l e is known with respect to the generation of such stormflows from small forested watersheds in this region. This is dealt with by analyzing results from studies completed by the author and other work'ersni.h:.Jamieson Creek watershed and vicinity within the Seymour River basin of the Greater Vancouver Water District,and by making f i e l d examinations in the same study area. Only stormflow events resulting from rainfall are dealt with in this paper. The problem of stormflow events from snowmelt water will not be considered, as there are more factors in-volved in controlling the rate of snowmelt. It i s , however, probably reasonable to assume that after snowmelt water reaches the surface of unfrozen forest s o i l s , the pathways i t takes and the flow mechanisms involved are similar to those of rainwater. 44 45 B. SUMMARY AND REVIEW OF MODELS OF STORMFLOW GENERATION A survey of literature reveals that proposed stormflow mechanisms during rainfall events may be conveniently classified under the following three models. 1. Model of Overland Flow from the Entire Watershed Resulting  from Rainfall Exceeding Infiltration Rate This model is based on Hofcton's theories of i n f i l t r a t i o n f i r s t presented in the early 1930's (Horton, 1933). Horton proposed that overland flow is generated whenever the rainfall rate exceeds the in-f i l t r a t i o n rate of the surface s o i l . The i n f i l t r a t i o n rate decreases with the time during a storm until a capacity value, a property of the surface s o i l , is reached. Implicit in Horton's theories are: (1) over-land flow is common and spatially widespread over the watershed during most rainfall events, and (2) stormflow measured at the watershed out-let is generated by overland flow processes. Since these processes are considered to occur uniformly over the watershed, the source area is considered equal to that of the entire watershed. The overland flow model may appropriately describe the production of stormflow from some types of terrain with limited i n f i l t r a t i o n capacity such as in arid and semi-arid areas. On the other hand, overland flow may be a rare occurrence in vegetated, humid environments, where i n f i l t r a t i o n capa-cities are usually greater than expected rainfall intensities. 46 2. Model of Subsurface Stormflow from a Variable Source  Area in the Watershed The concept of subsurface flow as the primary source of storm-flow from forest lands of humid areas was also presented in the 1930's (Hursh, 1936). Basically i t assumes that due to the highly porous nature of the surface s o i l , i n f i l t r a t i o n is unlimited in the vegetated humid areas. Consequently, stormflow mainly results from flow through subsurface pathways usually occurring in the soil above a less permeable layer of soil or parent material. As the velocity of subsurface flow is normally considered to be slower than overland flow, the following questions arise. How can subsurface flow deliver sufficient water to provide significant contributions to stormflow volume and, can subsurface flow supply water quickly enough to account for stormflow? These ques-tions can be answered partly by the facttthat soil channels (macropores) exist in most forest soils which provide pathways of low resistance to subsurface stormflow (Whipkey, 1967; Aubertin, 1971). The variable source area concept developed from the subsurface stormflow model by Hewlett and his co-workers helps answer these questions more precisely (Hewlett and Hibbert, 1967; Hewlett and Nutter, 1970). This concept suggests the stream channel system expands into and shrinks from inter-mittent and ephemeral source areas during and following r a i n f a l l . As a result the subsurface flow paths that join surface water systems are shortening, while the cross-sectional area through which subsurface flow can pass is expanding. .The variable source area concept also recognizes the importance of areas of low storage capacity along stream channels in contributing to stormflow. 47 3. Model of Overland Flow after Soil Saturation of a Partial  Area of the Watershed This model emphasizes that stormflow results from overland flow generated from a small partial area of the watershed that quickly becomes saturated during a rainfall event (Betson, 1964; Ragan,1968; Dunne and Black, 1970). Saturation occurs from below because of rising shallow water tables fed by vertical i n f i l t r a t i o n of rainwater from above. Although Dunne and Black (1970) refer to stormflow so generated as "saturated overland flow," i t is rather different from Horton's " i n f i l -tration rate exceeded overland flow." On the other hand, the "partial area concept" is very similar to the "variable source area concept" dis-cussed earlier. The term "variable source area" is preferred because i t also implies the dynamic nature of the source areas. Dunne and Black (1970), however, contend that the input of downward percolating water in the near-channel area is sufficient to produce a rapid response of the water table, and that downslope drainage contributions from the mountain slope, as suggested by Hewlett and his co-workers, are negli-gible. To some extent, this minor disagreement could be attributed to differences in the physical environments of watersheds. Freeze (1972) presented the results of his mathematical simu-lations of stormflow generation to provide theoretical support for the latter model. Without considering the expanding and shrinking nature of stream channels inherent in the variable source area model by Hew-lett and his co-workers, Freeze f e l t that the conditions necessary for subsurface stormflow to be a significant contributor to stormflow are quite stringent. It should be noted that, unlike Whipkey (1967) and 48 Aubertin (1971), Freeze did not recognize the role of soil channels, which are present in most forest s o i l s , in conducting quick subsurface stormflow. Hewlett (1974) questioned whether the saturated conductivity values used by Freeze were adequate to characterize surface soils and stream bank materials in forested watersheds. Hewlett (1974) then specu-' lated that soils of perennial, intermittent, and ephemeral stream banks will tend to have saturated conductivity values near the highest limits set by Freeze. For forest soils which usually contain many soil channels, Hewlett's speculation is probably very reasonable. C. HYDROLOGIC INSIGNIFICANCE OF OVERLAND FLOW IN WATERSHEDS OF THE COAST MOUNTAINS The shallow forest soils (forest floor and mineral soil) of the slopes of the B.C. Coastal Mountains which support mature and old growth trees are generally very porous. Field observations indicated that soil channels in the form of old root holes, structural channels or cracks are "evident in the soil profiles of the slopes. Overland flow in the classical Horton sense rarely takes place except when rain f a l l s on saturated soil and forest floor or on bedrock. The following analysis will serve to explain this view point. The ponding of rain water on the soil surface is the i n i t i a l stage for the occurrence of overland flow. As pointed out by Rubin (1966), the necessary conditions for ponding are a rainfall rate greater than the saturated hydraulic conductivity of the surface soil and a rainfall duration greaterrthan the time [required for the surface soil to become saturated. As there 49 is no information regarding the time required for saturation of the study area, analysis will be carried out only on the basis of the f i r s t of Rubin's c r i t e r i a . Tables 1 and 2 compare the reported values of saturated hydraulic conductivity for the soils in the vi c i n i t y of Jamieson Creek watershed and values of rainfall intensity of different duration and return periods as determined by measurements at Seymour Falls some eight miles downstream from the mouth of Jamieson Creek. Although the rainfall intensify values for Jamieson Creek watershed are expected to be higher than that at Seymour Falls due to elevation difference, i t seems reasonable to say that the majority of rainfalls on most areas of the Jamieson Creek watershed and vicinity cannot lead to overland flow in light of the f i r s t of Rubin's c r i t e r i a . This con-clusion is consistent with f i e l d observations made in this study and that of Plamondon et a l . (1972). Even after surface ponding does occur, soil surface depression storages are able to store additional quantities of water. Depression storage estimated from two sample micro-relief measurements made by the author between a 1.5 m contour interval for Jamieson Creek watershed is about 10 mm. D. PATHWAYS OF WATER MOVEMENT IN FOREST SOIL In the forest soil of mountain slopes of Coastal B.C., two types of stormflow pathways have been identified: (1) the matrix of the forest floor and mineral soil and (2) interconnected soil channel, passing through the miinerralfsochl ((id-e iVrriesaandoChowCdfl978i)es and Ciicw. T A B L E . I M A X I M U M E X P E C T E D S H O R T DURATION RAINFALL INTENSITIES OF D I F F E R E N T RETURN PERIODS A S DETERMINED FROM M E A S U R E M E N T S A T SEYMOUR F A L L S * RAINFALL INTENSITY ( mm hr -1 ) RETURN PERIOD ( Y e o r ) 5 10 25 DURATION ( M i n u t e ) 5 | ^ 74 91 112 15 48 61 74 30 38 46 56 60 29 36 43 120 21 24 29 I i T A B L E . 2 | M E A S U R E D OR E S T I M A T E D S A T U R A T E D HYDRAULIC j CONDUCTIVITY FOR SOIL AND F O R E S T FLOOR O F ; J A M I E S O N C R E E K W A T E R S H E D AND, VICINITY. SATURATED HYDRAULIC CONDUCTIVITY (mm hr"') 1. FOREST FLOOR SAMPLE 2 0 0 , SOIL SAMPLE CONTAINING NO 'CHANNELS. 120 , FIELD SOIL CONTAINING CHANNELS • 350 1. ESTIMATED FROM PLAMONDON et ol (I 972) 2. MEASURED BY O'LOUGHLIN ( 1 9 7 2 ) 3. ESTIMATED BY CHAMBERLIN (19.72) *DATA SOURCE: ATMOSPHERIC ENVIRONMENT SERVICE OF CANADA. 51 1. Water Flow Through the Forest Floor, and"1 Soil Matrix Water flow through homogeneous soil can be described by Darcy's law which states that the flow rate is propor-tional to the product of the hydraulic potential gradient and soil hydraulic conductivity. The water which inf i l t r a t e s mountain slope through the soil matrix system normally moves vertically downward until i t encounters a less permeable layer. Some water penetrates slowly into the lower less permeable layer, but as this occurs at a much slower rate, there is a tendency for a saturated condition to develop above the interface and the surplus water to begin moving laterally downslope. Where such impermeable soil layers are not present, downward movement of water continues until a relatively impermeable parent material is reached. Some water which i n f i l t r a t e s the soil during the storm periods may be stored to sustain lateral flow which subsequently becomes baseflow or to provide a source of water for evapotranspiration after the storm has ceased. 2. Water Flow Through the Soil Channels Water is conducted into channels that are present in forest soils and moves downward in response to the force of gravity. Soil channels begin to f i l l when water ponds on the soil surface in local depressions or when the surrounding soil matrix becomes saturated. Whipkey (1967) suggested that water tends to move through soil channels by mechanisms of turbulent or non-Darcian flow, in response to the gravi-tational potential gradient. Whether or not flow is turbulent or 52 laminar depends on the pore size and the velocity of the flow. In lami-nar flow conditions, Poiseuille's equation can be used to give some indication of the importance of pore size to the flow rate. In a slope which contains soil channels, water will move rapidly down to the im-permeable layer, bedrock, slope toe, water table or stream channel. The soil matrix and soil channels may serve either as separated or integrated pathways for stormflow concentration. The relative impor-tance of each of these two pathways in conducting water downward depends on the following factors: 1) the depth and hydrologic properties of the soil matrix, 2) the size and distribution of soil channels, 3) the ante-cedent wetness of the forest s o i l , and 4) the intensity and duration of the rainstorm. E. STUDIES OF THE HYDROLOGIC RESPONSE OF FOREST SOILS TO A RAINSTORM The hydrologic response of forest soils containing intercon-nected soil channels to water inputs has been observed in the study area by several workers. Based on a study in an isolated 2.5 m by 3.5 m experimental plot, located at the 300 m elevation in Jamieson Creek watershed, de Vries and Chow (1973) reported that during wetting, the non-steady state total water potential f i e l d displayed a relatively high degree of complexity. For example, tensiometer readings indicated that water reached the 57.0 cm depth before i t reached 31.3 cm depth. Working on a plot located at the 700 m elevation in the same watershed, Chamberlin (1972) reported a similar response of forest soils to water 53 inputs. Both studies suggested that the complex behavior of the soil under conditions of non-steady state is related to the presence of soil channels and the large difference in the resistance to flow in the soil channel and soil matrix systems. It is likely that especially dead-end soil channels, which have a tendency to f i l l up with water, con-tribute to wetting of the soil from below. Aubertin (1971) proposed that the rapid entry of water into soil channels was because after the water moved through porous forest floor material, i t was directed into soil channels in depressions as a result of the microrelief of the mineral s o i l , de Vries and Chow (1973), on the other hand, suggested that the opening to the soil chan-nels is located wear tlhie surface-of H horizon in the lower part of forest floor. As the water content of the soil matrix gradually in-creases with water supplied from the soil surface through i n f i l t r a t i o n and from the wetted perimeters of the soil channels, the soil matrix resistance to water flow decreases. This reduces the difference in flow resistance between soil channels and soil matrix, de Vries and Chow observed l i t t l e heterogeneity in the flow during the drainage cycles and that as drainage proceeded the heterogeneity decreased. Water which is in soil channels at the end of a rainfall event will drain rapidly from the s o i l . Once the soil channels are empty, they no longer con-tribute to the drainage process, which from that time on is confined to the soil matrix. 54 F. THE RESPONSE OF THE STREAM CHANNEL SYSTEM TO A RAINSTORM Sustained by flow through the soil channels and/or soil matrix, saturated subsurface flow may become saturated return flow, wherever environmental conditions favor i t s occurrence. Such return flow is by no means overland flow in the classical sense of Horton's theory. Rather i t is the water that has infiltrated at positions upslope and has reappeared downslope. The saturated return flow may enter the ground again or become part of an expanding stream channel system (sur-face water system), depending on topographic features, soil i n f i l t r a t i o n capacity along i t s path and the subsequent rainfall pattern. Sources which may contribute to the expansion of the surface water system are: 1) direct rai n f a l l on flowing streams and adjacent saturated areas, 2) vertical i n f i l t r a t i o n of rainwater in areas close to the stream channels, 3) seepage through saturated stream banks as a result of up-slope subsurface flow and 4) saturated return flow. During a storm, the stream channel system will expand and shrink in response to rain-f a l l characteristics and will shrink to perennial stream channels after the cessation of a storm. This is the crucial feature of the variable source area concept conceived by Hewlett and his co-workers. The occurrence of saturated return flow and its contribution to an expanding and shrinking stream channel system during a storm has been observed in the study area. In conjunction with a slope st a b i l i t y study, a network of simple piezometers established by O'Loughlin (1972) in steep drainage depressions in the vici n i t y of Jamieson Creek watershed 55 revealed that the rising of saturation levels within the soil rapidly responded to ra i n f a l l or snowmelt water. For example, within one hour of the beginning of a heavy rainfall- in September 1971, the piezometer level rose at a rate of approximately 60 mm per hour. Relationships between piezometer levels and 24-hour rainf a l l amounts indicated that 24-hour ra i n f a l l amounts exceeding approximately 120 mm can cause com-plete saturation of most depressions. O'Loughlin (1972) speculated that, during heavy r a i n f a l l , water passes readily through the soil and moves laterally over the impermeable t i l l or bedrock in a shallow, saturated layer at the base of B horizon. Flow is then directed towards the drain-age depressions from adjacent slopes. The concentration of subsurface water in a depression can cause a rapid increase in the thickness of the saturation zone and a concomitant increase in the soi l ' s a b i l i t y to conduct water downslope. Consequently saturated return flow can contribute to the expanding stream channel system during a storm. A sub-watershed^in the lower portion of Jamieson Creek watershed was used for the investigation of changes of the stream channel system in response to a rainstorm. Figure 1 shows time lapse diagrams of the expanding and shrinking nature of the stream channel network observed during a storm in October 1973. Also.shown is the stormflow hydrograph of this storm. As can be seen from the diagrams, the extent to which temporary streams expanded into the sub-watershed is closely related to the rate of stormflow as measured at the outlet of the main watershed. Sources of these temporary streams were observed to be mostly saturated return flow appearing preferentially where soils were more readily See Figure 22 of Chapter 1 for location. 56 V h - 7 0 - 6 0 CM I — E - 5 0 _ i CO •40 W £ - 3 0 9 - 2 0 g H CO M O 12 ' . 13 1 14 O C T O B E R 1 9 7 3 Figure 1. A stormflow hydrograph of Jamieson Creek and the s measured changes in the stream channel network of a sub-watershed in response to a rainfall event in October 1973. 57 saturated. These locations were at the toes of long steep slopes, in small drainage ways between microridges, in areas of thin and less per-meable s o i l s , in areas of flow concentration produced by contour and slope concavity and in areas adjacent to the flowing streams where soils tended to be wettest. Although this investigation was limited to the lower portion of Jamieson Creek watershed, i t does indicate that storm-flow generation mechanisms on mountain slopes of Jamieson Creek water-shed are similar to the model of subsurface stormflow from a variable source area of the watershed. G. CONCLUSION For B.C. coastal watersheds with hydrologic environments simi-lar to that of the study area, Horton type overland flow rarely i f ever occurs during most rainstorms. The model of subsurface stormflow from a variable source area is considered more appropriate in describing stormflow generation in this coastal region. H. LITERATURE CITED Aubertin, G. M., 1971. Nature and extent of macropores in forest soils and their influence on subsurface water movement. U.S. Forest Serv. Res. Paper NE-191. Betson, R. P., 1964. What is watershed runoff? J. Geophys. Res. 69(8), 1541-1551. 58 Chamberlin, T. W., 1972. Interflow in the mountainous forest soils of coastal British Columbia. In "Mountain Geomorphology," B.C. Geog. Series, No. 14, 121-127. de Vries, J. and Chow, L., 1973. Hydrology of mountain slopes, Environ-ment Canada, Final 1973 Report. Dept. Soil Science, University of B.C., 265 p. Dunne, T., and Black, R. D., 1970. Partial area contribution to storm runoff in a New England watershed. Water Resour. "Res., 6(5), 1296-1311. Freeze, R. A., 1972. Role of subsurface flow in generating surface runoff. 2. Upstream source areas. Water Resour. Res., 8(3), 609-623. Hewlett, J. D., 1974. Comments on letters relating to 'Role of subsur-face flow in generating surface runoff by R. A. Freeze. Water Resour. Res., 10(3), 605-608. Hewlett, J. D., and Hibbert, A. R., 1967. Factors affecting the response of small watersheds in humid areas. In Forest Hydrology, 275-290. Pergamon Press, New York. Hewlett, J. D., and Nutter, W. L . , 1970. The varying source area of stormflow from upland basins.. Paper presented at the symposium on Interdisplinary Aspects of Watershed Management, Mont. State Univer., Bozeman. 59 Horton, R. E., 1933. The role of i n f i l t r a t i o n in the hydrologic cycle. Amer. Geophy. Union Trans. 14: 446-460. Hursh, C. R., 1936. Reports, hydrology. Amer. Geophy. Union Trans. 15: 301-302. O'Loughlin, C. L., 1972. An investigation of the sta b i l i t y of the steep-land soils in Coast Mountains, southwest British Columbia, Ph.D. Thesis, University of B.C., 147 p. Plamondon, A., Black, T. A., and Goodell, B.C., 1972. The role of hydro-logic properties of the forest floor in watershed hydrology. In "Watershed in Transition," 341-348. American Water Resources Asso-ciation and Colorado State University. Ragan, R. M., 1968. An experimental investigation of partial-area con-tributions. Publ. 76, pp. 241-249, Int. Ass. Sci. Hydro!., Berne. Rubin, J., 1966. Theory of rainfall uptake by soils i n i t i a l l y drier than their f i e l d capacity and its applications. Water Resour. Res., 2(4), 739-749. Whipkey, R. Z., 1967. Storm runoff from forest catchments by subsur-face routes. Int. Ass. Sci. Hydro!., Proc. of Leningrad Symposium, 774-779. CHAPTER III THE STORMFLOW CHARACTERISTICS OF A SMALL, STEEP AND FORESTED WATERSHED IN THE COAST MOUNTAINS OF SOUTHWESTERN BRITISH COLUMBIA 60 CHAPTER III THE STORMFLOW CHARACTERISTICS OF A SMALL, STEEP AND FORESTED WATERSHED IN THE COAST MOUNTAINS OF SOUTHWESTERN BRITISH COLUMBIA A. INTRODUCTION The last chapter dealt mainly with the mechanisms of stormflow generation on mountain slopes of the watersheds. This chapter reports the results of an investigation of stormflow characteristics as deter-mined primarily from streamflow measurements made at the gauging station of Jamieson Creek watershed--a small, steep, and forested watershed in the Coast Mountains of southwestern British Columbia. This chapter deals mainly with the channel-phase stormflow hydrology. A hydrograph is a plot of the water stage or discharge rate of a stream versus time. In response to a rainfa l l input to i t s watershed, a stream will rise and then recede until the occurrence of next rain-f a l l event. This is recorded as a series of rising and f a l l i n g limbs in the streamflow hydrograph. The reaction of the hydrograph to a rain-f a l l event can be assumed to characterize the overall behavior of the watershed (Beth!amy, 1963). With respect to the response of the stream-61 62 flow hydrograph to a rainfall event, Chow (1964) wrote: "The stream-flow hydrograph is an integral expression of the physiographic and c l i -matic characteristics that govern the relationship between rainfall and runoff of a particular basin. It shows the time distribution of runoff at the point of measurement, defining the complexities of the basin characteristics by a single curve." Hydrograph analysis i s , there-fore, the f i r s t step in the investigation of the response in streamflow to rainfall input. An important aspect of stormflow hydrology is the evaluation of the time response of a stream to rainfall input. Therefore, hydrograph analysis is a prerequisite to this type of stormflow hydro-logy study. Other hydrologic analyses which are also useful in a storm-flow hydrology study, such as comparing the reactions of two watersheds to the same storm, will aiksollbesiiiseded later. As at the time of preparing this chapter, there was no published quantitative information with respect to the stormflow characteristics of the streams that drain the small, steep, and forested watersheds in the Coast Mountains of southwestern British Columbia. The response of streams in this coastal region to rainfall input, however, has been described as very flashy in some unpublished reports among which are those iiin the f i l e s of the Hydrology Division, British Columbia Water Resources Service, Victoria, B.C. The objectives of this chapter are as follows: 1. to investigate, through analysis of individual storm hydro-graphs, the characteristics of stormflow from Jamieson Creek watershed; 62 63 2. to evaluate the stormflow characteristics of Jamieson Creek watershed in relation to other hydrologic variables; and 3. to compare the stormflow characteristics of Jamieson Creek watershed and its control watershed, Elbow Creek watershed. The efforts of this study are restricted to stormflow events caused only by rainstorms. B. DATA USED Streamflow records from the gauging station located at the mouth of Jamieson Creek (1970-1974) form the basic data used for the analyses of this chapter. Relevant ra i n f a l l records for the Jamieson Creek watershed (1970-1974) and concurrent streamflow records from Elbow Creek (1971-1974) were used as supplementary data for the analyses. Streamflow data were originally available in the form of con-tinuous traces of water stage versus time. Water stages were read and converted to streamflow rate by using the-stage-discharge relationships established for the streamflow gauging stations. The total r a i n f a l l amount and variations in r a i n f a l l intensity for every storm event were averaged from available data from recording rain gauges located in Jamieson Creek watershed and v i c i n i t y , as pre-viously discussed in Chapter I. Field examinations were frequently carried out to make sure that storm events with possible snow melt effects were not included in the data analyzed (see Appendix 2). 64 C. ANALYTICAL METHODS 1. General Streamflow Characteristics Before proceeding with the specific analyses of stormflow characteristics of Jamieson Creek watershed, a general description of the general streamflow characteristics of the watershed will be made. This will focus on the monthly distribution of streamflow and precipitation for the watershed. 2. Hydrograph Analysis a. The Selection of Stormflow Events For the purpose of this study, individual stormflow events are the basic units for hydrograph analysis and other analyses related to stormflow characteristics. These stormflow events were selected by examining both r a i n f a l l and s t r e a m f l o w records. Generally every apparent r i s e i n the s t r e a m f l o w hydrograph was c o n s i d e r e d as an e v e n t . The only criterion was that the hydrograph rise be attributable to a R a i n f a l l event occurring on the watershed. Any r i s e which was only a small fluc-tuation of the hydrograph was excluded as an event. A stormflow event selected in such a way tended to include a combination of small, moderate and large storms. However, i t should be pointed out that the number of stormflow events used in each of the analyses could vary, depending on the avai l a b i l i t y of the required data for a specific analysis. The frequency of reading water stage for conversion to stream-flow rate for a selected stormflow event depends on the variations of 65 the water stage with time. In other words, more readings were taken from a portion of the hydrograph that displayed a high degree of variation in water stage with time than from another portion that only displayed a gradual change in water stage. b. Hydrograph Separation In hydrograph analysis, the streamflow hydrograph for a storm-flow event has often been processed by the application of a more or less arbitrary method of separating baseflow from stormflow (or direct runoff). This arbitrary separation of streamflow hydrograph is nece-sary because of the d i f f i c u l t y in dealing with long recessions of indi-vidual stormflow hydrographs. The complexity of the relation between rainfall and the response' in streamflow,and the predominant concern with floods as a result of extreme peak stormflows in the Coast Moun-tains of southwestern British Columbia,make i t appropriate in this study to use the stormflow component of the hydrograph. The relationship between this component and rainfall is more recognizable than that of total streamflow and r a i n f a l l . Hydrograph separation procedures (Linsly et a l . , 1949) are essen-t i a l l y empirical since there is not yet a universally applicable separa-tion technique which has a sound physical basis. In this study, no attempt has been made to partition the total streamflow into surface flow (overland flow), interflow and groundwater flow. Rather, a total streamflow hydrograph during the stormflow period is considered to be made up of stormflow and baseflow. To separate the total streamflow hydrograph into stormflow and baseflow, a time-based separation technique proposed by Hewlett and Hibbert (1967) was adopted for this> study after a preliminary test of this technique on hydrographs of 20 stormflow events of Jamieson Creek watershed. By this method, a total streamflow hydrograph for a storm event is divided into what Hewlett and Hibbert referred to as "quick flow" and "delayed flow." Quick flow and delayed flow are considered quantitative expressions of stormflow and baseflow respectively. The hydrograph separation is made by drawing a line upward from the point -1 -2-1 of i n i t i a l hydrograph rise at a slope of 0.55 1 s ikm h (0.05 3-1 -2 -1 f t s mi h ) until the line intersects the f a l l i n g limb of the hydro-graph (Figure 1). This slope of the separation line was selected be-cause when projected from the i n i t i a l rise of the hydrograph, i t ter-minates quick flow within a few hours after a short and intense storm, and within a few days after prolonged heavy storms (Hibbert and Cunning-ham, 1967). This time-based hydrograph separation method could not be claimed to be any less arbitrary than other methods, but i t has the important advantage of being more objective, i.e., the duration of the stormflow event is determined mainly by the data i t s e l f . It is also f e l t that since the separation slope used was constant throughout the study, i t would not greatly affect the accuracy of the stormflow hydro-graphs for comparison purposes or interpretations. This method is also suitable for direct computer processing of the streamflow data. For this study a computer program based on the technique suggested by 67 Figure 1. A diagram illustrating the hydrograph separation method and stormflow parameters used in the study (after Hewlett and Hibbert, 1967). 68 Hewlett and Hibbert (1967) was developed and applied to water stage data to evaluate relative proportions of quick and delayed flow for a total streamflow hydrograph during a stormflow event. c. Stormflow Parameters Evaluated from Hydrograph Analysis Several stormflow parameters are usually determined from hydro-graph analysis. Those stormflow parameters evaluated in this study are listed and defined as follows. Many of these are also shown in Figure 1. i) Antecedent baseflow (q^) is the streamflow rate prior to the hydrograph rise caused by a rainstorm. i i ) Peak flow magnitude (q ) is defined as the maximum instanv: taneous rate of streamflow produced during the period of each individual storm less the antecedent baseflow rate i i i ) Time to the peak ( t p ) is the time from the i n i t i a l hydro-graph rise to the occurrence of the streamflow peak. iv) Stormflow period or duration ( t d ) is the time period during which stormflow occurred, defined here as the time from the i n i t i a l hydrograph rise until when the hydrograph separation line intersects the f a l l i n g limb of the hydrograph. v) Stormflow amount (v) is the streamflow volume that li e s above the hydrograph separation line, expressed as depth over the watershed. vi) Rainfall amount (p) is the volume of rainfall during a storm that is considered to cause the stormflow event, also expressed as depth over the watershed. 69 v i i ) Lag time (t^) is the time between the center of mass of rainfall and that of stormflow (quick flow), v i i i ) Peak flow ratio (r ^ ) is defined as the ratio of peak flow magnitude to i t s antecedent baseflow (q /q.)-It should be noted that the definitions for some of these para-meters are, to some extent, dependent on the hydrograph separation tech-nique used. d. Recession Analysis In this study, graphical comparison is made using several selected recession limbs of individual storm hydrographs. Because of the arbitrary nature of hydrograph separation, total streamflow rather than stormflow was used for recession analysis. In order to eliminate the effects of channel precipitation, the beginning oftthe period of recession was taken as two hours after the peak or two hours following the cessation of storm r a i n f a l l , whichever was later. Two hours are considered sufficient time for the latest channel precipitation to reach the streamflow gauging station. Horton (1933), Barnes (1939) and others have shown that the entire recession curve, or a segment of i t , can be fitt e d by the simple exponential equation of the form q t - q o e - " (1) or <t = q o k t (2) 7 0 where q^ . is the streamflow rate at time t, q Q is the i n i t i a l stream-flow rate, e. is the base of the natural logarithm, a is a constant and k is a another constant equal to e . If a recession curve conforms to the simple exponential equation,then since l o g 1 Q q t = t l o g 1 Q k k + i 0 g 1 Q q Q i (3) i t can be seen that a plot of streamflow rate (log scale) against time (linear scale) results in a straight line of slope log k. The recession curves can then be described and compared using the values of the con-stants. It is for this reason that the selected recession limbs of storm hydrographs are plotted on semi-log graph paper. Due to the range of the streamflow rate during a recession period, one additional advan-tage of plotting on semi-log paper is that i t will display the recession limbs more conveniently. If the simple exponential equation is not appropriate in de-scribing the recession limbs of selected storm hydrographs from Jamieson Creek watershed, a double exponential equation is used to f i t a number of selected recession limbs by the method of least squares. The double exponential equation was f i r s t suggested by Horton (1933) as an improve-ment to the simple exponential equation (1) for a basin containing many sub-basins of different characteristics and can be written as follows: q t = q o e - a t " (4) where n is another constant. 71 This equation can be written in double logarithmic form as: q l o g 1 0 Elog10 (q^)]f n 1 o 9 1 0 1 + l o 9 i o a - ° - 3 6 2 2 2 - ( 5) The methods of least squares then can be used to derive the values of the constants (Johnson and Dils, 1956). It should be noted that equation (4) makes the constant k of equation (2) a function of time as follows: k = e" a t (6) e. Rainfall-Stormflow Relationships i) Graphical comparison of rainfall input and the response in the streamflow hydrograph The chronological variability of the response in streamflow hydrographs is most readily visible through graphical comparison. For each storm, the hyetograph of hourly rainfall intensities was plotted on the same time scale as the storm hydrograph to examine the reactions of streamflow to rainfall intensity fluctuations. The rainfall inten-sity values were averaged from available data of the recording rain gauges located in the watershed. i i ) Regression analysis of stormflow amount against rainfall amount Regression analysis used in this study involves the analy-sis of relationships between numerically defined stormflow parameters and other pertinent hydrologic parameters. 72 In a study of the relationships of stormflow and precipitation (rainfall) parameters, the storm events used in the analysis should not be limited to those producing large amounts of stormflow. In order to obtain sound regression relations both large and small storms must be included in the analysis. In the regression analysis in this section, stormflow amount, the dependent variable, was related to rainfall amount, the independent variable. 3. The Role of Antecedent Baseflow in Rainfall-Stormflow Relationships In light of some of the conclusions of the last chapter on storm-flow generation mechanisms, the soil moisture storage of a watershed before a storm is an important factor in influencing the extent of stormflow source area and consequently the stormflow amount generated during a given storm. Because of the d i f f i c u l t y in intensive sampling of soil moisture on a watershed scale by available techniques, indirect estimation by some relevant indices are often used. Among these indices are the number of days since the last major rainstorm, antecedent pre-cipitation for an arbitrarily specified period and antecedent baseflow just prior to the rise of streamflow hydrograph. These indices are used because, in most cases, the quantities of rainfall and streamflow are the only two direct measurements available. a. Rationale Antecedent baseflow rate has been successfully used as an index of groundwater storage for larger watersheds in improving stormflow 73 prediction (Lee and Bray, 1969). In small, steep watersheds such as Jamieson Creek watershed and the two experimental watersheds in western North Carolina reported on by Hewlett and Helvey (1970), a large pro-portion of baseflow is considered to result from soil water drainage. Since the rate of soil water drainage depends on soil water storage, i t would seem possible to use antecedent baseflow as an index of average watershed soil water storage before a storm. The relationship between soil water drainage and storage, when evapotranspiration is prevented or considered to be negligible, was stu-died by Ogata and Richards (1957). They found an empirical relation-ship between soil water content and time after saturation, as follows: M = a t " b (7) where M is the soil water content, t is the time after saturation and a and b are constants. The soil water drainage rate is obtained by differentiating (7) (Bawer et a l . , 1972), -abt-< b t l>. (8) Substituting (7i) into (8) gives: This relates the rate of soil water drainage (dM/dt) to soil water content, M. This equation explains why the antecedent baseflow rate before a storm can often be considered as an index of average watershed 74 soil water storage, when evapotranspiration from the watershed is assumed negligible in comparison to streamflow from the watershed. Since very few of the selected stormflow events occurred during the summer months in Jamieson Creek watershed, the assumption that evapo-transpiration is negligible may be reasonable. The contribution of soil water movement to baseflow has been observed by Nixon and Lawless (1960), Sartz (1964) and Weyman (1970) for soils in situ and by Hewlett (1961) for soils in a sloping "model" in the f i e l d . Cheng (1972) also found that soil water drainage showed a close relationship to streamflow from Jamieson Creek watershed. The purpose of this section is to assess, through multiple regression analy-s i s , the usefulness of antecedent baseflow as an index of average water-shed soil water storage by examining i t s role in modifying r a i n f a l l -stormflow relationships. b. Statistical Assessment In multiple regression analysis, one may tend to include vari-ables which in fact have very 1 i t t l e effect on the stormflow-rainfall relations. A measure of significance could be obtained by considering the analysis with and without a suspect variable and considering whether the results obtained "with" were significantly better than that obtained "without." For this purpose, a stepwise multiple regression technique was used to relate stormflow amount not only to rainfall amount but also to antecedent baseflow rate. The variance ratio (F) test was used to detect whether the inclusion of the antecedent baseflow rate in the 75 analysis would significantly improve the regression. A computer pro-gram1 based on this idea was used for the stepwise multiple regression analysis. This computer program accepts an additional independent vari-able into the regression equation i f the ratio of the decrease in the residual sum of squares (SS ) to the new residual mean squares (MS r e s) exceeds a certain acceptable F level. For the present study, an F level of 5 percent or higher was considered s t a t i s t i c a l l y significant. In the procedure of the stepwise multiple regression analysis, a simple regression equation was f i r s t derived using a radnifialclv. amount as the independent variable (X-j). Antecedent baseflow rate was then added to the regression analysis as another independent variable (X2) and yielded a multiple regression equation. The values of variance  ratio (F), standard error of estimate (SEE) and coefficient of deter- mination (R ) were calculated for both simple and multiple regression equations. The F value contributed by each independent variable was also obtained for each equation. These F values are necessary in deter-mining whether or not the inclusion of the second independent variable is s t a t i s t i c a l l y significant. 4. Comparison of Stormflow Characteristics of Jamieson Creek and  Elbow Creek Watersheds This comparison is made by ( 1 ) examination of concurrent storm-flow hydrographs of the two watersheds and ( 2 ) regression analysis of selected stormflow parameters of the two watersheds. 1 This computer program was developed by Dr. A. Kozak, Faculty of Forestry, University of British Columbia, Vancouver, B.C. 76 As the exact watershed area of Elbow Creek is s t i l l unknown because of its poorly defined watershed boundary, the streamflow values used in this section are not on the basis of flow per unit watershed - 1 - 2 -1 area, i.e., 1 s km , but rather in units of 1 s . T h e hourly r a i n f a l l distribution of the storm that caused the hydrograph rise will be shown together with the hydrographs. The rainf a l l inputs to the two watersheds are assumed to be the same. 'Since these two adjacent watersheds have similar aspect and elevation, i t is f e l t that this assumption would not greatly affect the accuracy of.comparison and interpretation. The forestin Jamieson Creek watershed was originally scheduled to be logged at the end of the 1975 water year. This regression analysis was intended to test the su i t a b i l i t y of the calibration of Jamieson Creek watershed using individual stormflow events. Basically this c a l i -bration approach is to establish the regression relationships between the stormflow characteristics of the two adjacent watersheds during the pre-logged period in order to make future comparisons with that of the post-logging period. However, since the logging of Jamieson Creek water-shed has been postponed until 1973,1 more available data than those used in the present analysis should provide a more precise calibration of the pre-logging stormflow characteristics. The calibration of watersheds based on individual stormflow events was suggested by Bethlamy (1963) and Ursic and Popham (1967) ^ Personal communication with Dr. R. P. Willington, Assistant Professor of Forest Hydrology, Faculty of Forestry, University of British Columbia, Vancouver, B.C. 77 and used by Hewlett and Helvey (1970) and Hombeck (1973). The method adopted in the present study is similar to that used by these authors. The method was to relate the stormflow parameter of Jamieson Creek to that of its control, Elbow Creek. In the analysis, the storm-flow parameters of Jamieson Creek were used as the dependent variables since its watershed will be logged after the calibration is completed. For this analysis, only two stormflow parameters were used, i.e., time to peak (t ) and peak flow magnitude (q n). This is due to the lack of a suitable hydrograph separation technique for the streamflow data of Elbow Creek for which the exact watershed area is not yet known. D. RESULTS AND DISCUSSION 1. General Streamflow Characteristics The streamflow in Jamieson Creek is rapid. This is due to the fact that the channel gradient averages about 20 percent and some short reaches have gradients greater than 100 percent (see Figure 9 of Chapter I). As is demonstrated in the mean monthly streamflow distribution graph in Figure 2 and the hydrograph for Jamieson Creek in Figure 3, there is a pronounced seasonal variation in the streamflows. Maximum streamflows occurred at two times of the year, in the spring and in the late f a l l or winter; while summer streamflows were usually low in response to light and infrequent summer rainfa l l (Figure 4). Maximum spring streamflow occurs as a result of meteorological conditions con-ducive to snowmelt, sometimes supplemented by warm spring rains. 78 70 CH 60 0-J 500-2 2 ^ 400-2 u . 2 5 300-cc 200 4 100 4 J F M A M J J A S O N O M O N T H Figure 2. Mean monthly distribution of streamflow from Jamieson Creek watershed (1970-1974). 700-, 600 500 i 5 400 2 z o 300 O UJ £ 200 100 1 i r-J F M A M J J A S O N D MONTH Figure 4. Mean monthly distribution of precipitation for Jamieson Creek watershed (1970-1974). 81 Major streamflow increases in the f a l l and winter are caused by major rain storms or rain-on-snow events in these two seasons. These stream-flow characteristics are generally similar to those of Seymour River of which Jamieson Creek is a tributary. The characteristics of storm-flow as a result of rainstorms for Jamieson Creek watershed will be discussed in detail in later sections. 2. Stormflow Characteristics as Indicatedbby Stormflow Parameters  Evaluated from Hydrograph Analysis A total of 41 rainstorm hydrographs during the study period (1970-1974) were separated and analyzed by the methods previously out-lined. The means and ranges of the stormflow parameters evaluated from hydrograph analysis are summarized in Table 1. A more detailed l i s t of the values of these parameters for a l l storm events is given in Appendix 2. The amount of storm rainfall during the study period varied from 5.1 mm (0.2 inch)tto nearly 320 mm (12.95 inches) with the majority of storm durations ranging from 20 hours to approximately 60 hours. Those storms with a duration shorter than 20 hours were usually small -1 -2 storms which produced low peak flows (less than 100 1 s km ) and a small amount of stormflow (less than 25 mm). Some storm events which were made up of a series of continuous periods of rain had durations longer than 60 hours. An event of this nature usually caused a storm hydrograph with multiple peaks. The prolonged rainstorms often exhi-bited considerable variation in rainfa l l intensity, thus causing 82 Table 1. The Means and Ranges of Stormflow Parameters for Jamieson Creek Watershed (1970-1974) Range Number of Observa-Stormflow Parameter Mean Min. Max. tions (n) Antecedent Baseflow (q.), 1 s _ 1km" 2 32.51 3.889 98.38 41 Peak flow magnitude 388.41 10.30 1,370.00 41 (q p), 1 s-W 2 Time to the peak 21.50 3.50 82.50 41 (t ), hour Stormflow duration 57.22 10.00 243.5 41 ( t d ) , hour Stormflow amount 32.04 0.152 179.30 41 (v), mm Rainfall amount 73.22 5.080 328.9 41 (p)» mm Lag time 8.50 5.00 15.00 33 ( t ^ ) , hour Peak flow ratio 16.56 0.80 97.75 41 83 streamflow to fluctuate considerably. The majority of the storm events during the study period occurred in the months of October and November. The fraction of storm rainfall that appeared as stormflow (quick flow) varied from storm to storm with a minimum of 2.5 percent to a maximum of 81 percent. The results of a regression analysis showing the relationships between the stormflow and rainfall amounts of indi-vidual storm events is presented in a later section. The magnitudes of peak flows during the stormflow periods varied -1 -2 considerably, ranging from a l i t t l e more than 10 1 s km to a maximum -1 - 2 3 -1-1 of 1 ,370 1 s km (125.7 f t sftmi ;.). The accumulative peak flow magni-tude distribution curve is presented in Figure 5. As can be seen from this diagram, approximately 30 percent of the stormflow events have a -1 -2 3 -1 -peak flow magnitude greater than 550 1 s km , equivalent to 50 f t s mi The latter figure was used by Harris and Williams (1971) as an arbitrary lower limit of high flows for small forested watersheds in coastal Oregon. During the study period two stormflow events had peak flow magnitudes -1 -2 3 -1 -2 greater than 1,100 1 s km , equivalent to 100 f t s mi . Rothacher (1973) considered this peak flow magnitude to be sufficiently high to be important to downstream flooding for small forested watersheds on the west slopes of the central Cascade Range in Oregon. A regional storm, which caused a 165 mm rainfall in Jamieson Creek watershed on -1 -2 July 11 and 12, 1972 resulted in the largest peak flow, 1,370 1 s km , recorded during the study period. This storm which resulted in rainfall amounts varying from 150 to 200 mm over the lower Fraser River Valley has been described as a record breaking storm for July in the region ^ PERCENTAGE OF ALL OCCURRENCES GREATER THAN Z INDICATED VALUE - h O O - J . -S to c + O - s 0 ) -". 3 O -- • • cz r + o o O O -5 -+i ro ro T 3 7r ro CD Qi r + - h ro — • - s o t/i s zy ro 3 Q . SD ' — 3 - a i i. r n 3=> ^ 1 C 7 ^ O Q -i ro - n — ' t o t o o • ^ i S. s : 4=. -•• c + 3 " £ D - s ro i—i t o — 1 - a ro o m r i -c h o — ' c + t o 3 " 1 ro — J T 3 3 ro | - s r>o o ro 3 r + 0 ) I Q ro O -h (D — 1 o o o e - s -5 ro 3 O ro t o o o 0) o 00 o o o o o o o O) o o 00 o o o o o o o o o fr8 85 by Shaeffer (1973). The magnitude of peak flow is largely governed by the size and intensity of the rainstorm and the watershed charac-t e r i s t i c s . In this study of Jamieson Creek watershed, peak flow magni-tude appeared to be dependent on the rainf a l l distribution over the period from the beginning of rainf a l l to the time streamflow reached its peak.(see Appendix 3). With the exception of a few very prolonged storms, generally the peak of the streamflow during the storm period was reached in a relatively short time, often in 5 to 30 hours after the i n i t i a l rise of the hydrograph. In other words, the majority of the rising times (times to the peak) were less than 30 hours. The rising times seemed to depend to a large extent on the rainstorm characteristics, particu-l a r l y the time distribution of ra i n f a l l over the period before the occurrence of the streamflow peak. Lag time was computed for 33 stormflow events, excluding those events that produced a storm hydrograph with multiple peaks. Lag times were found to be less varied than the rising times. The majority of the stormflow events examined had lag times ranging from 5 to 10 hours. The longest lag time was 15 hours. Askew (1970) studied lag times of 2 watersheds with area varying from 0.4 to 90 km in New South Wales, Australia. He found that (1) the differences in the values of lag time between watersheds were highly correlated with watershed characteristics such as area and slope, and (2) the variation in values for individual stormflow events was strongly related to the magnitude of mean storm-flow. Lag times for Jamieson Creek watershed were found to be relatively 86 constant, with small variations of random nature. However, i t is f e l t that to obtain formulae for estimating lag time using watershed charac-t e r i s t i c s , soil hydrologic properties that govern the stormflow trans-mission to the stream should also be included. Antecedent baseflow during the study period ranged from 4 to -1 -2 100 1 s km , indicating variable watershed soil moisture conditions prior to the rise of hydrograph. The results of a regression analysis on the role of antecedent baseflow in modifying rainfal1-stormfal1 relationships is presented in a later section. The peak flow ratio was found to vary from storm to storm from a minimum of 1 to a.maximum of about 100. Some stormflow parameters evaluated with the same hydrograph separation technique as that used in the present study, for watersheds 36 and 37 of the Coweeta Hydrologic Laboratory in North Carolina, were available from Hewlett and Helvey (1970). These stormflow parameters were the amount of storm rainfall (p), the amount of quick flow (storm-flow amount, v), antecedent baseflow rate (q^), time to the peak ( t p ) , peak flow magnitude (q p) and stormflow duration ( t d ) . Table 2 summarizes the results of a comparison of stormflow parameters for Jamieson Creek watershed and the two watersheds of Coweeta Hydrologic Laboratory. Since the reported minimum rainfa l l was 48 mm (1.9 inches) for the two watersheds in Coweeta, only those rainfall events in Jamieson Creek watershed that were equal to or greater than this value were used in this comparison. Despite their differences in geographic location, the stormflow parameters for the three watersheds are similar as indicated by their 87 Table 2. Comparison of Stormflow Parameters of Jamieson Creek Watershed and Watersheds 36 and 37, Coweeta Hydrologic Laboratory, North Carolina, U.S.A. Jamieson Coweeta Hydrologic Laboratory, Creek North Carolina, U.S. .A. Watershed, B.C., Canada Watershed 36 Watershed I 373 1970-1974 1943-1963 1964-1967 1943-1963 1964-1967 p A1 114 138 110 139 112 (mm) B 2 448*329 63.5-315 48.3-297 ":66-318 48.3-300 V A 54 48.5 34.3 55.1 44.7 (mm) B i i 11*179 r 6591-3.5.9 C6S9W69 ?7'.4-,171 8.9-191 <ln A 602 707 564 1.058 900 [ I s 'km d) B 209-1374 186-4329 153-2427 295-5346 240-3236 q i i o A 43.8 49.7 43.6 44.3 40 (1 s"'km^) B 66G-6S96&6 10.9-97.3 77X759299 88283825 9' r6.6-96.6 tp A 31.3 30.9 25.0 30.4 23.2 r (hour) B 88553-8255 1M2132 c- 94-90 1 M-134 " 4-91 A 79.4 109 87 n o 90.5 (hour) B 30.5-244 444^221 37373230 44436-204 • 45-^ 225 A = Mean B = Range Watershed 37 was logged in 1963. 88 means and the ranges (Table 2). This could be due to the similarity , of their hydrologic environments. All. three of these watersheds are forested, located in a humid area, and steep in their topography. More importantly, a significant portion of their stormflow moves through subsurface pathways from a variable source area in the watershed (Hewlett and Hibbert, 1967; Cheng et a l . , 1975). Jamieson Creek watershed has a slightly higher percentage of ra i n f a l l appearing as stormflow, indi-cating its lower watershed storage capacity. This would l i k e l y be the result of shallower soils in Jamieson Creek watershed than in i t s counter-parts in CoweetaJ - 1 - 2 The peak flow magnitudes ( I s km ) of Jamieson Creek watershed were usually lower than the Coweeta watersheds. This is possibly due to the differences in ra i n f a l l characteristics rather than total rain-f a l l amounts in which regard a l l three watersheds were similar. Although no information is available from Hewlett and Helvey (1970) with respect to r a i n f a l l distribution for any storms in the two Coweeta watersheds, some indication of the nature of these storms was obtained by calcu-lating the ratio of rising time to that of stormflow duration. From these calculations, i t was found ' that the majority of storms in - Coweeta that produced very high peak flows were usually those with a high total ra i n f a l l amount and a very low ratio of rising time to storm-flow duration. This indicates that a much larger proportion of the total r a i n f a l l amount occurred before the occurrence of the streamflow peak. In comparison, this type of storm occurred less frequently in Jamieson Creek watershed. ^ Personal communication with Mr. J. D. Helvey, formerly Hydro-logist with the Coweeta Hydrologic Laboratory, now with the Forest 89 Furthermore, the fact that the watershed area of Jamieson Creek watershed is about six times larger than the Coweeta watersheds might account for the lower peak flows. This is because the larger the water-shed area, the longer i t takes its storm waters to reach the lower part of the main stream, thus resulting in lower peak flows. 3. Recession Characteristics of Storm Hydrographs The recession limbs of a number of selected hydrographs plotted on semi-log graph paper are shown in Figure 6 to 8. All selected hydro-graphs exhibited well defined recession limbs which, although steep during early stages, gradually flatten out. However, i t was found that the recessions did not lend themselves to an obvious breakdown into two or three straight lines on which traditional hydrograph separation can be based. Instead, they were s t i l l continuously curvilinear overall. The steeper slopes in the upper portion of the recession curves are probably due to the rapidly decreasing importance of the soil channels in serving as the pathway of subsurface stormflow and a quick contrac-tion of the stream channel system as a result of reduced saturated flow through the s o i l . The continuously decreasing recession rates can possibly be related to the gradual change from dominantly saturated to dominantly unsaturated flow in the watershed s o i l . Figures 6 to 8 also indicate that only the lowest portion of the recession limbs can possibly be plotted as an approximately straight line on semi-log graph paper as in the case of groundwater recession curves. However, i t is f e l t Hydrology Laboratory of the Pacific Northwest Forest & Range Experiment Station in Wenatchee, Washington, U.S.A. 9 0 Figure 6. Recession limbs of selected storm hydrographs in 1970, 1972 and 1973 for Jamieson Creek watershed . 91 3 10 20 30 40 50 60 T I M E , h o u r > Figure 7. Recession limbs of selected storm hydrographs in 1971 for Jamieson Creek watershed. Figure 8. Recession limbs of selected storm hydrographs in 1974 for Jamieson Creek watershed . 93 that baseflow from Jamieson Creek watershed as well as from many other small, steep watersheds is mainly sustained by slow soil water drainage due to the apparent lack of a significant groundwater reservoir (Hewlett, 1961). In this view, the narrow saturated zone along stream channels is not a source but rather a conduit through which slow soil water drain-age passes to enter the stream (Hewlett, 1961). The lack of a s i g n i f i -cant groundwater reservoir is partly due to the steep watershed topo-graphy and the close proximity of impermeable bedrock to the ground surface. Since i t is clear that the simple exponential equation is not appropriate in describing the recession limbs of the storm hydrographs of Jamieson Creek watershed, six recession limbs (four of them are shown in Figures 16) and 171) were selected to f i t the double exponential equa-tion (4) by the least square methods. The recession limbs show a good f i t to the double exponential equation. This is not surprising since an extra constant has been included in the double exponential equation. The values of the constants determined for these six recession limbs are presented in Table 3. Any storm hydrograph is a short term event and its recession varies from storm to storm because of variations in water storage in the watershed. As can be seen from Table 3 and Figures 6 to 8, the shape of the recession curves varied with storms (even over a small range of streamflow values) which is reflected by the different values of con-stants obtained for different storm hydrographs. The differences in streamflow recessions between storms can be explained in terms of the Table 3. Values of the Constants in the Double Exponential Equation (4) Determined for Recession Limbs of Selected Storm Hydrographs of Jamieson Creek Watershed Values of Constants Periods of Recession a n September 11-12, 1971 0.0919 0.8149 October 04-07, 1971 0.1275 0.6602 October 13-17, 1971 0.2606 0.5227 October 22-24, 1971 0.0660 0.7886 September 21-23, 1973 0.0814 0.8606 October 05-07, 1973 0.2618 0.5404 95 contribution to streamflow from variable source areas. The pattern of the recession of a given storm hydrograph depends on the degree of re-charge of the watershed storage by the storm and the spatial d i s t r i -bution of such storage over the watershed. Storms with relatively in-tense rainfalls may cause large amounts of quick flow without either greatly increasing the overall watershed storage or increasing the areas of high soil moisture content adjacent to the stream channels. On the other hand, a prolonged storm of large rainfall amount will result in a larger contributing area, or larger areas with higher soil moisture content. The shape of the storm hydrographs, including the recession limbs, for these two types of storms would be different because of the degree of recharge of water storage and i t s spatial extension into the watershed. Furthermore, i t will be primarily the characteristics of the contributing areas that influence the shape of the storm hydrographs, including the recession limbs. In studying the shape of hydrographs and recession patterns, the watershed characteristics must be viewed as dynamic in nature. If streamflow is generated from a small part of a watershed, the characteristics of streamflows from this area can be quite different from those of streamflows from a much larger area of the watershed. However, i t should be noted that the overall water-shed characteristics do exert a direct influence on the portion of the watershed that does contribute to streamflow. 96 4. Rainfal1-Stormf1ow Relationships a. Graphical Comparison The reaction of streamflow to the within-storm variations of rainfa l l intensity is illustrated in Figures 9 iandlQO. As can be seen from these diagrams, the hydrographs usually start to rise very shortly after the beginning of a rainstorm. This rapidity in the start of hydro-graph rise could be attributable to the rainfall that f e l l directly into the stream channel and other adjacent areas which were saturated or nearly saturated before the storm. These two diagrams also indicate that during the storm period, the rate of streamflow normally is related closely to that of the rainfall input. However, this correlation is not caused by overland flow as a result of the rainfall exceeding the in f i l t r a t i o n rate. Rather, this is more likely because of subsurface stormflow, including saturated return flow, from a variable source area of the watershed. To further demonstrate the relationship between the input (rain-f a l l ) and output (streamflow), the hourly rainfall and streamflow rate for the stormflow event of July 11 to 12, 1972, were plotted in Figure 11 on the same scale and in the same units (mm h~^). The total rainfall amount of this storm was 165 mm and the storm duration was 36 hours. The stormflow produced by this storm was 108 mm and i f we assume that a l l this stormflow was generated by the process of overland flow as in Horton's theory, we would get an average i n f i l t r a t i o n rate of less than 1.5 mm h~^. This is not lik e l y to occur in Jamieson Creek watershed since the saturated hydraulic conductivity is usually greater than 100 97 i I I I I 10 II 12 13 14 JULY 1972 Figure 9. The response of streamflow from Jamieson Creek watershed to a storm in July, 1972. 98 II 12 13 14 15 16 17 OCTOBER 1973 Figure 10. The response of streamflow from Jamieson Creek watershed to two storms in October, 1973. 99 II 12 13 J U L Y 1 9 7 2 Figure 11. A comparison of hourly rainfall and streamflow of Jamieson Creek watershed for a storm in July, 1972. 100 mm h~^ for soils containing no soil channels in this watershed (see for example Table 2 of Chapter II). As indicated in Chapter II, there are many interconnected soil channels in the soils of Jamieson Creek water-shed mainly as a result of dead and decaying roots. Consequently, the overall permeability for a soil mass containing channels would be much higher than 100 mm h"'! (see Table 2 of Chapter II). As also pointed out in Chapter II, several f i e l d observations, during the storm periods in Jamieson Creek watershed and other forested watersheds in this coastal region, have provided evidence of the rare occurrence of overland flow in the coastal B.C. environment. The following paragraphs will briefly discuss the response of streamflow to rainfall in terms of subsurface stormflow, including saturated return flow, from a variable source area of the watershed. During the period between rainstorms, water slowly moves down-ward through the soil mantle of the mountain slopes, feeding baseflow in the'stream channels and producing a relatively high concentration of moisture at locations closer to the base of the slope. That soil water content decreases with height or distance from stream channels had, in fact, been measured in the Southern Appalachians (Hewlett, 1961; Helvey, 1971). The concentration of water at the slope base usually creates a small but permanent zone of saturated or nearly saturated conditions near the stream channels. An important point in this con-text is that the water in this zone of soils is immediately available to be dispatched into the stream in the event of further rainfall input. According to Dunne (1970), the capillary fringe zone, a zone in which 101 the soil is essentially saturated but under low tension, may intersect the soil surface close to the stream. Adding a small amount of water to this zone will reduce the tension and easily convert the capillary fringe into a saturation zone. This will accordingly increase the slope of the saturation zone toward the stream to produce subsurface storm-flow. This mechanism and the direct channel precipitation may be responsible for the rapid rising of the streamflow following the start of a rainstorm as indicated in Figure 11. During the storm, the saturation zone originating from the area close to the stream channel will expand vertically up the soil profile and laterally up the slope as water is added by flow through soil matrix and soil channels. Figure 11 also indicates that at the early stage of a rainstorm, a large portion of infiltrated water could be retained temporarily, increasing soil moisture and resulting in a higher soil hydraulic conductivity. This will in turn increase the contribution of soil water drainage to the increase in saturation zone at the slope base. If the saturation zone were to reach ithe soil surface in certain areas of favorable conditions, a l l further rain f a l l input to these areas will become stormflow. These completely saturated areas will be inte-grated into an expanding stream channel system. The expansion of the stream channel system was observed during a stormflow event from October 12 to 14, 1973. The result of this observation was discussed in Chapter II and the hydrograph of this storm is also shown in Figure 10. An expanding stream channel system during a storm also implies that the infiltrated water will enter the stream channel system at locations much closer to the point of i n f i l t r a t i o n than i f i t had to travel to the original permanent stream channels. 102 Streamflow from the watershed increases as the stream channel system expands and the hydraulic gradient of the saturation zone, through which water enters the stream channel, increases. Streamflow rate reaches its maximum when the combined contribution of flow from these two pro-cesses reaches a maximum. Unless the contribution from these two pro-cesses increases continuously, the streamflow from the watershed will decrease accordingly. The characteristics of Jamieson Creek watershed that contribute to the rapid response of streamflow to a rainstorm are summarized as follows: 1. shallow s o i l ; 2. extremely high overall permeability of the soil mass; 3. steep stream channels and watershed slopes; and 4. a high degree of contour curvature, resulting in a high density of permanent, intermittent and ephemeral stream channels. The climatic environment of the watershed may also contribute to the rapid response of streamflow to major storms. As a result of the fre-quent occurrence of small storms, the soils in the most part of the watershed remain above " f i e l d capacity" throughout the f a l l and winter seasons, thus creating very favorable conditions for the rapid response of streamflow to major storms. b. Statistical Relationships between Stormflow Amount and  Rainfall Amount The regression equation derived from 41 storm events, with storm-flow amount as the dependent variable and rainfall amount as the 103 independent variable, is presented in Figure 12. Pediction limits^for individual observations at the 95 percent level are also shown in the 2 diagram. The r value indicates that 92 percent of the variation in stormflow amount is accounted for by the variation in rainfall amount. However, there is s t i l l a scatter of points along the regression line, suggesting that there may be additional variables that also influence the amount of stormflow produced by a r a i n f a l l . The highly significant relationship between the stormflow and rainfall may be used to predict the stormflow amount from this water-shed i f some stormflow records are missing. However, such a prediction for storm events outside the range of the data that were used to deter-mine the regression equation could result in a large error. Further-more, i t should be understood that the regression equation applies only to the watershed studied, with no claims made as to i t s applicability to other watersheds. Nevertheless, this high correlation does indicate that a close relationship could also exist between stormflow and rain-f a l l amount, with a similar regression coefficient, for small forested watersheds having a similar hydrologic environment to that of Jamieson Creek watershed. 5. The Significance of Antecedent Baseflow Rate, in Improving the  Statistical Relationship between Stormflow Ml.Rainfall Amount To make a more general assessment of the role of antecedent baseflow rate in the relationship between stormflow and rainf a l l amount, one set of stormflow data from Jamieson Creek watershed and four addi- « tional sets of data from watersheds 36 and 37 of the Coweeta Hydrologic 1 Confidence limits. 104 Figure 12. The relationships between rainfall amount and stormflow amount of Jamieson Creek watershed (1970-1974). 105 Laboratory in North Carolina (Hewlett and Helvey, 1970) were analyzed. The results of a stepwise regression analysis are summarized in Table 4. When using rainfall amount as the only independent variable, the explained variance (r ) of stormflow amount ranged from 74 to 92 percent with Jamieson Creek watershed having the highest value. The inclusion of antecedent baseflow rate as an additional independent vari-able only slightly improved the accuracy of stormflow prediction with the explained variance (R ) ranging from 79 to 94 percent. However, an examination of F values for both independent variables indicates that the improvements were significant at the five percent or higher level of significance for every data set. This indicates that antece-dent baseflow rate may be a useful index of watershed soil water storage. The reductions in the standard error of estimate (SEE) for stormflow amount also suggests that the use of antecedent baseflow rate as an indicator of average watershed soil water storage improves the accuracy of the stormflow estimate. The standard error of estimate for the regression equation using two independent variables ranged from 10 to 16 mm of stormflow. One might question whether the improvement in stormflow estimate by an increase in the explained variance of 2 to 5 percent has any prac-tical value. In fact, this improvement might not be too promising from a practical point of view. However, the main purpose in this section was to test the validity of a physicalTybibasiedhjnypbithesdsr.raithertthan to advocate the practical usefulness of such an index. Direct measurements i f feasible or estimation of soil water storage on a watershed scale before storm occurrence should improve the 106 Table 4. Regression Equations Showing the Relationships of Stormflow Amount to Rainfall Amount and Antecedent Baseflow Rate 2 Regression Equations F F l 2^ R S E E Coweeta Hydrologic Lab. Watershed 36 1) 1943-1963 (n=77) Y,C=-18.24+0.4834X1 212.19 212.19 I^.S 0.739 14.81 36 1 Yo,.=-37.38+0.5087X1+0.3143Xo 140^78 212.19 18.85* 0.792 13.31 36 1 2 2) 1964-1967 (n=30) Yoc=-20.95+0.5021X, 100.13 100.13 0.782 17.31 36 I Yoc=-37.00+0.5223Xn+0.3163Xo 63.18 125.94 6.51 0.824 15.83 36 1 L Watershed 37 1) 1943-1963 (n=77) Y37=-18.38+0.5265X1 258.57 258.57 0.775 14.76 Y37=-31.7387 149.58 299.12 9.90* 0.801 13.95 2) 1964-1967 (n=30) Y37=-24.06+0.6 70X] 188.16 188.16 0.869 16.11 "kit i%^s-.mm*oz®zsm*w:wx0 106.96 213.22 4.50 0.888 15.19 37 1 1 iL 2 Jamieson Creek Watershed (n=41) Y. =-7.991+0.5468Xn 451.48 451.48 0.921 11.35 jc 1 Y. =-12.46+0.522TX1+0.1928X, 280.42 450.28 9.62* 0.937 10.27 JC 1 2 continued 107 Table 4 - continued * Significant at 1 percent level ** Significant at 5 percent level Y, stormflow amount (mm) X-j, rainfall amount (mm) -1 -2 X2, antecedent baseflow rate ( I s km ) F, variance ratio value for whole equation F.| and F 2, variance ratio values for variable X-j and X 2 respec-tively 2 R , coefficient of determination SEE, standard error of estimate for Y (mm) 108 stormflow estimate to a certain extent. However, the remaining variation in stormflow amount may be mainly due to an inadequate sampling of rain-f a l l amount over the watershed and to differences in rainf a l l intensity between storms. 6. Relationships between the Stormflow Characteristics of Jamieson  Creek and Elbow Creek Watersheds a. Hydrograph Comparison The comparison of concurrent storm hydrographs, as shown in Figures 13 to 15 reveals that the storm hydrograph of Jamieson Creek watershed displays the characteristics of a sharper peak, higher peak flow ratio, and steeper recessions than the storm hydrograph of the adjacent Elbow Creek watershed. However, the storm hydrographs of both watersheds have almost identical rising times with the differences of approximately 1 to 2 hours at the most. The recession limbs of the storm hydrographs for Jamieson Creek watershed are usually steeper than those of Elbow Creek watershed. As pointed out in Chapter I, the s o i l s , geology, and the gra-dients of watershed landslopes and stream channels of these two adjacent watersheds are considered generally similar. The differences in storm-flow characteristics could possibly be accounted for by some differences in other topographical features of the two watersheds. Field obser-vation indicated that Jamieson Creek watershed has a higher degree of contour curvature, resulting in a higher drainage density than Elbow Creek watershed in which there is very few clearly defined valley floors, Figure 13. A comparison of the response of the streamflows in Jamieson Creek and Elbow Creek to a storm in July, 1972. no 5 6 7 8 NOVEMBER 1974 Figure 14. A comparison of the response of streamflows in Jamieson Creek and Elbow Creek to a storm in November, 1974. I l l Figure 15. A comparison of the response of streamflows in Jamieson Creek and Elbow Creek to a storm in November, 1974". 112 stream channels or ravines. The differences in topographical features between these two watersheds can also be seen in part in the topographic map shown in Figure 6 of Chapter I. Higher drainage density has been related quantitatively to higher peak flows (Carlston, 1963; Rothacher et a l . , 1967) and lower baseflows (Carlston, 1963; Trainer, 1969). The differences in the response of streamflow to rainfall input may indicate that the two watersheds differ in some other fundamental aspect and this will be discussed later. b. Statistical Relationships The values of time to the peak ( t p ) in hours and peak flow magni-tude (q p) in 1 s - 1 from these two watersheds were evaluated. Regression equations, calculated for both stormflow parameters, are as follows: t p ] = -1.4892 + 0.9496 t p 2 (n = 23, r 2 = 0.9830) q p l = 498.20 + 5.0390 q p 2 (n = 23, r 2 = 0.5552) where 1 and 2 denotes Jamieson Creek and Elbow Creek watershed respec-tively. The very high coefficient of determination for the f i r s t equa-tion suggests that the time to the peak for Jamieson Creek watershed could be adequately predicted from the corresponding values from Elbow Creek watershed. This equation and i t s prediction limits for individual observations are shown in Figure 16. As pointed out previously additional data collected before the scheduled logging in 1978 will be used to improve the calibration of Jamieson Creek watershed. This calibration 113 T I M E TO T H E P E A K (hr.) E lbow C r e e k W a t e r s h e d Figure 16. The relation of time to the peak of Jamieson Creek watershed to that of Elbow Creek watershed (1971-1974). 114 will be used to detect any change in the time to the peak for Jamieson Creek watershed as a result of forest removal. Judging from the stable relationship between the times to the peak of these two watersheds, i t does not seem very likely that the data from future storm events will significantly change the regression line. Based on this preliminary calibration, a change in the time to the peak greater than 23 percent of the mean value of the calibration period is needed in order to be detected. As more stormflow data will be used in the final calibration, a decrease in the standard error of estimate is expected. This will accordingly allow the pre-logging regression equation to detect a smaller change in the time to the peak. In contrast, the second equation for the peak flow magnitude does not indicate that a very close relationship exists between these two watersheds (see Figure 17). Since such a regression equation would not be adequate to detect any foreseeable changes in peak flow magni-tude as a result of logging, no attempt was made to calculate the prediction limits for this equation. As pointed out by Bethlamy (1963), a <l>ow correlation coeffi-cient would indicate that either the two watersheds were not subject to the same climatic events or that the watersheds diff e r in some funda-mental respect. After examination of the rainfall records, i t is con-sidered that the f i r s t of these possible reasons is not very likely to be the major factor responsible for the wide scatter of the data points. Evidence supporting the second reason is given below. As pointed out in Chapter I, the selection of Elbow Creek water-shed as a control to Jamieson Creek watershed was based on several P E A K FLOW M A G N I T U D E ( I s - 1 ) , Elbow C r e e k W a t e r s h e d Figure 17. The relation of peak flow magnitude of Jamieson Creek water-shed to that of Elbow Creek watershed (1971-1974). 116 f a c t o r s , o f which one was the l a c k o f a b e t t e r a l t e r n a t i v e watershed near t h e study a r e a . At t h e time o f making the s e l e c t i o n , i t was r e a l i z e d t h a t t h e watershed boundary o f Elbow Creek i s t o p o g r a p h i c a l l y i l l - d e f i n e d . I t was assumed t h a t t h e watershed a r e a i s c o n s t a n t , a l t h o u g h unknown, and t h a t no s e r i o u s problem o f v a r i a b l e l e a k a g e from the watershed e x i s t s . However, a more r e c e n t i n v e s t i g a t i o n made by a team o f g r a d u a t e s t u d e n t s from the F a c u l t y o f F o r e s t r y and t h e Department o f Geology o f the U n i v e r s i t y o f B r i t i s h Columbia i n d i c a t e s t h a t t h e above assumption might not be v a l i d J T h i s i n v e s t i g a t i o n r e v e a l e d t h a t a snow a v a l a n c h e had t r a v e l l e d the l e n g t h o f Elbow Creek t o a p o i n t a t t h e bend i n t h e c r e e k l e a d i n g t o t h e s t r e a m f l o w gauging s t a t i o n (see F i g u r e 6 o f Chapter I ) . I t was found t h a t the upper a r e a o f Elbow Creek i s a permanent a v a l a n c h e t r a c k . I t s s l o p e i s g r e a t e r than 1G0 p e r c e n t , concave and exposed t o bedrock. T h i s f i e l d i n v e s t i g a t i o n a l s o i n d i c a t e d t h a t s t r e a m f l o w was g r e a t e r a t h i g h e r e l e v a t i o n s on Elbow Creek than near i t s e x i t i n t o the Seymour R i v e r . I t was h y p o t h e s i z e d t h a t an e x t r e m e l y permeable zone o f m a t e r i a l l i e s a t t h e base o f t h e a v a l a n c h e t r a c k (an a l l u v i a l f a n ) which p e r m i t s l e a k a g e t o o c c u r . T h i s a l l u v i a l f a n i s g r a v e l l y and a p p r o x i m a t e l y 30 f e e t deep a t the t u r n i n g p o i n t o f the c r e e k . A s o i l p i t dug c l o s e t o the main r o a d o f Seymour R i v e r b a s i n r e v e a l e d s a t u r a t e d s u b s u r f a c e f l o w t h e r e b y p r o v i d i n g e v i d e n c e f o r the p r e s e n c e o f t h i s seepage zone. However, a g e o p h y s i c a l t e c h n i q u e would have t o be used t o d e t e r m i n e 1 P e r s o n a l communication w i t h Dr. R. P. W e l l i n g t o n , A s s i s t a n t P r o f e s s o r o f F o r e s t r y H y d r o l o g y , F a c u l t y o f F o r e s t r y , U n i v e r s i t y o f B r i t i s h Columbia, Vancouver, B.C. 117 the exact structure of the seepage zone. A network of piezometers to measure the response of the saturation zone to variable storm events could be used to confirm that seepage occurs and that i t varies with the storm size. If the latter occurs, then the measured peak flow of Elbow Creek in response to a storm would not be proportional to that of Jamieson Creek. E. SUMMARY AND CONCLUSION An understanding of rainfall-stormflow relationships has been the focal point of many investigations for engineering purposes or scien-t i f i c interest. The reasons that this study deals with the short term rather than long term rainfal1-streamflow relationships are its impor-tance in the coastal region of southwestern British Columbia and the lack of this type of quantitative hydrologic information for small, steep, and forested watersheds in this coastal region. In this study the stormflow characteristics of Jamieson Creek watershed have been evaluated using hydrograph analysis. Forty-one storm events, ranging from 5 to 330 mm, occurred during the snow-free season from 1970 to 1974. Stormflow amount accounted for 44 percent of the storm rainfall on the average, varying from 2.5 to 80 percent. A significant number of storms had more than 60 percent of their total rai n f a l l appear as stormflow. Instantaneous peak flows varied considerably with individual storms and appeared to be mainly governed by the amount and distribution of rainfall prior to the 118 occurrence of the peak flows. Stormflow rise usually occurred very shortly after the start of a rainfall event; however, the rising times were dependent on rainfall distribution before the peak flow occurred, usually being less than 30 hours. Lag time was found to be relatively constant and short with an average of 8 hours. It is suggested that to derive lag time from characteristics of small watersheds, soil hydro-logic properties should also be included with those generally used para-meters . Stormflow amount was highly correlated with rainfall amount with 92 percent of its variance being explained by the variations in rainfall amount. Antecedent baseflow rate was proposed as an index of watershed soil water storage prior to the storm hydrograph rise. One set of data from Jamieson Creek watershed and four additional data sets from two small steep watersheds in the Coweeta Hydrologic Labora-tory were used to assess, through multiple regression analysis, the usefulness of antecedent baseflow rate in improving stormflow-rainfall relations. For a l l data sets, the inclusion of antecedent baseflow as a second independent variable significantly improved the stormflow e s t i -mate in comparison to that when rainfa l l amount was the only independent variable. Recession limbs of the storm hydrographs were found to vary with individual storms, depending on the degree of recharge of the water-shed storage by the storm and the spatial distribution of such storage over the watershed. The evaluation of the 41 storm hydrographs of Jamieson Creek watershed indicates an overall flashy nature of stormflow. This flashy 119 nature is typified by an immediate hydrograph rise, sharp and high peak flow, high proportion of rainf a l l appearing as stormflow, and steep but variable recession limbs. These stormflow characteristics mainly re-fl e c t the influence of not only climatic conditions but also three water-shed characteristics: (1) soils with high permeability and low water storage capacity, (2) steep watershed slopes and stream channels, and (3) high drainage density. These stormflow characteristics can be generally interpreted in terms of the generation of stormflow from a variable source area of the watershed. A comparison of the stormflow characteristics of Jamieson Creek watershed and the adjacent Elbow Creek watershed was also made. This comparison indicated that stormflow from Jamieson Creek watershed has a sharper peak, higher peak flow ratio and steeper recession than storm-flow from Elbow Creek watershed, but both have very similar rising times. Differences in the streamflow response of the two watersheds could indi-cate differences in some of their topographical features. However, i t also suggests that the possibility of leakage from Elbow Creek, re-vealed in a preliminary f i e l d investigation, may deserve more detailed study. F. LITERATURE CITED Askew, A. J., 1970. Derivation of formulae for variable lag time. Jour, of Hydrology 10: 225-242. Barnes, B. S. 1939. The structure of discharge-recession curves. Trans. Amer. Geophys. Union 20: Pt. IV, 721-724. 120 Bayer, [_. D., Gardner, W. H. and Gardner, W. R., 1972. Soil Physics. John Wiley and Sons, Inc., New York. Bethlamy, N., 1963. Rapid calibration of watersheds for hydrologic studies. Int. Assoc. Sci. Hydrl. Bull. 8 (3): 38-42 Carlston, B. S., 1963. Drainage density and streamflow. U.S. Geol. Survey Prof. Paper 422-c, 8 p. Cheng, J. D., 1972. Evaluating soil water drainage of a humid forest site in southwestern British Columbia. Unpublished M.Sc. Thesis, University of British Columbia, Vancouver, B.C. Cheng, J. D., Black, T. A. and Willington, R. P., 1975. The generation of stormflows from small forested watersheds in the Coast Moun-tains of southwestern British Columbia. Paper presented at the Canadian Hydrology Symposium, August 11-14, 1975, Winnipeg, 7 p. Chow, V. T., 1964. Runoff. In "Handbook of Applied Hydrology." McGraw-H i l l Book Co., New York, Section 14. Dunne, T., 1970. Runoff production in a humid area. ARS 41-160, Agricultural Research Service, U.S. Dept. of Agriculture. Harris, D. D. and Williams, R. C , 1971. Streamflow, sediment-transport, and water temperature characteristics of three small watersheds in the Alsea River basin, Oregon." U.S. Geol. Survey Crircular 642, 21 p. 121 Helvey, J. D., 1971. Predicting soil moisture in the Southern Appala-chians. Unpublished M.Sc. Thesis, School of Forest Resources, University of Georgia, Athens, Georgia. Hewlett, J. D., 1961. Soil moisture as a source of baseflow from steep mountain watersheds. U.S. Forest Serv. Southeast Forest Exp. Sta., Station Paper 132, 11 p. Hewlett, J. D. and Helvey, J. D., 1970. Effects of forest clear-cutting on the storm hydrograph. Water Resour. Res. 6(3): 768-782. Hewlett, J. D. and Hibbert, A. R., 1967. Factors affecting the response of small watersheds to precipitation in humdid areas. In "Inter-national Symposium on Forest Hydrology," Pergamon Press, New York, pp. 275-290. Hibbert, A. R. and Cunningham, G. B., 1967. Streamflow data processing opportunities and application. In "International Symposium on Forest Hydrology," Pergamon Press, New York, pp. 725-736. Hornbeck, J. M., 1973. Stormflow from hardwood-forested and cleared watersheds in New Hampshire. Water Resour. Res. 9(2): 346-354. Horton, R. E., 1933. The role of i n f i l t r a t i o n in the Ihydrologic feycle. Amer. Geophys. Union Trans. 14: 446-460. Johnson, E. A. and Dils, R. E., 1956. Outline for compiling precipi-t a t i o n , runoff, and groundwater data from small watersheds. South-122 east Forest Exp. Station, U.S. Forest Serv. Station Paper No. 68, 41 p. Lee, J. Y. and Bray, D. I., 1969. The estimation of runoff from rain-f a l l for New Brunswick watersheds. Jour, of Hydrology, 9: 427-437. Linsley, R. K., Kohler, M. A. and Paulhus, J. L., 1949. Applied Hydro-logy. McGraw-Hill Book Co., New York, 689 p. Nixon, 0. R. and Lawless, G. P., 1960. Translocation of moisture with time in the unsaturated soil profiles. Jour. Geophys. Res. 65: 655-661. Ogata, G. and Richards, L. A., .1957. Water content changes following irrigation of bare f i e l d soil that is protected from evaporation. Soil Sci. Amer. Proc. 21: 355-356. Rothacher, J., Dyrness, C. T. and Fredriksen, R. L., 1967. Hydrologic and related characteristics of three small watersheds in the Oregon Cascades. USDA For. Serv. Pacific Northwest Forest and Range Exp. Station, 54 p. Portland, Oregon. Rothacher, J., 1973. Does harvest in west slope Douglas-fir increase peak flow in small forest streams? U.S. Forest Serv. Res. Paper PNW-163, 13 p. Shaeffer, D. G., 1973. A record breaking summer storm over the Lower Fraser Valley. Atmospheric Environment Service, Environment Canada, Technical Memoranda, TEC 787. 123 Sartz, R. S., 1964. Duration of percolation from a loess s o i l . U.S. Forest Serv. Lake States Forest Exp. Station, Res. Paper LS-40. Trainer, F. W., 1969. Drainage density as an indicator of baseflow in part of Potomac River Basin. U.S. Geol. Survey, Prof. Paper 650-c, pp. c!77-cl83. Ursic, S. J. and Popham, T. W., 1967. Using runoff events to calibrate forested catchment. Proc. Int. Forestry Res. Organization, pp. 319-324. Weyman, D. R., 197J0. Throughflow on hil 1 slopes and i t s relations to streamflow hydrograph. Int. Assoc. Sci. Hydro1. Bull. 15(2): 25-33. CHAPTER IV THE EVALUATION OF INITIAL CHANGES IN PEAK STORMFLOW FOLLOWING LOGGING OF A WATERSHED ON THE WEST COAST OF CANADA 124 CHAPTER IV THE EVALUATION OF INITIAL CHANGES IN PEAK STORMFLOW FOLLOWING LOGGING OF A WATERSHED ON THE WEST COAST OF CANADA A. INTRODUCTION The impact of logging activities on the peak flows of a stream is of major concern to the public and the managers of water an'd fish resources in the heavily forested Coast Mountains of southwestern British Columbia (B.C.), on the west coast of Canada, where many of the watersheds may be subject to logging in the near future. In this area, peak flows, caused by prolonged rainfall or a combination of rain and snowmelt, frequently result in flooding. Important as is the need, quantitative information with respect to the impacts of logging on peak stormflows relevant to the above area is not yet available in the literature. This fact promoted this study where the i n i t i a l changes in peak stormflows during storm periods as a result of clearcut logging were determined for a small watershed in southwestern B.C. by use of the paired-watershed technique. 125 126 B. LITERATURE REVIEW A survey of literature reveals that the changes in peak flow following clearcut logging of a watershed are highly variable, depending on the climate, s o i l , vegetation and topography of the watershed and the degree of ground surface disturbance associated with the logging. Generally in North America, forests have been observed to delay snow-melt and the complete clearing of a watershed can be expected to cause streamflow peaks of snowmelt origin to be higher. As snowmelt is not a major contributor to most peak flow events observed in the study area, the following review will be restricted to studies in areas where rainstorms are responsible for most peak flows. At the Coweeta Hydrologic Laboratory in western North Carolina, clearcutting a 13 ha (33-acre) area of hardwoods, leaving trees where they f e l l , increased annual water yiel d , but did not increase peak flows during storm periods (Hoover, 1945; Hewlett and Hibbert, 1961). More recently at Coweeta, clearcutting a mature hardwood forest on a 44 ha (108-acre), high elevation watershed increased the peak flows by an average of 9 percent and increased stormflow volumes of major storms by an average of 11 percent. No overland flow was observed (Hewlett and Helvey, 1970). In a West Virginia hardwood-forested watershed where a l l timber of commercial value was removed, Reinhart (1964) reported that instantaneous peaks (all far below flood magnitude) during the growing season were increased on the average by 21 percent and during the dormant season they were apparently 127 reduced by 4 percent. Somewhat similar results were obtained in a more recent study in New Hamshire by Hornbeck (1973). He found that both peak flow and stormflow volume were increased during the growing season as a result of forest clearing by cutting and subsequent applications of herbicides; but the changes during the dormant season were negligible. Rothacher (1973) reported that logging in Douglas-f i r forests with deep porous soils on the coast mountains of Oregon had only minor effects on major peak streamflows which occurred after the soils had been thoroughly wetted by f a l l rainstorms. The results from the aforementioned studies indicate that peak flow increases that result' from logging are restricted to rain-storms that occur during the growing season when a period of drying has established a soil water storage difference between clearcut and forest areas. On the other hand, during the dormant season when soil moisture contents are similar, no increases are to be expected from clearcut areas. C. THE STUDY AREA The two adjacent watersheds, watershed 1 and watershed 2, used in this study are in the Research Forest of the University of British Columbia (U.B.C.) near Haney, B.C., about 50 km east of Vancouver (Figure 1). These two watersheds, 23.1 ha and 44 ha respectively, are located in the foothill region of the Coast Mountains. The two watersheds extend from 145 to 455 m above mean sea level and 128 GAUGES V STREAMFLOW GAUGING STATIONS 13 AREA OF C L E A R C U T IN S E P T . 1973 Figure 1. A map of watersheds 1 and 2 in the U.B.C. Research Forest, showing the instrumentation and areas of clearcut. 129 have southerly aspects, and average slopes of 20 and 12 percent respectively. Strongly influenced by the Pacific Ocean 200 km to the west, the climate of the study area is typically maritime. About 90 percent of its 2,285 mm (90 inches) of annual precipitation f a l l s as rain during relative warm and moist winter months. Snow usually comes and goes during the winter months and seldom remains on the ground for more than a few days. Rainfall intensities are usually low, rarely exceeding 15 mm per hour; but long duration, steady rains may total 100 to 150 mm a day over storm periods of a few days. Peak stream-flows are usually caused by this kind of prolonged rainstorm, or a combination of rainfall and snowmelt in the f a l l and winter months. The bedrock of the region that contains the two watersheds consists of mostly quartz, granodiorite or diorite (Roddick, 1965). Outcrops and exposures of the quartz display smooth, superficially weathered surfaces free of open joints (Feller, 1974), suggesting that the bedrock is generally impermeable. This prevents significant deep seepage losses from the watersheds. The soils are mostly gravelly sandy loams derived from glacial t i l l . The watersheds have a thick and porous forest floor (Plamondon, 1972), and a highly permeable mineral s o i l . The high permeability is partly due to a great number of channels passing through the s o i l . These soil characteristics, together with low rainfall intensities and almost complete absence of frozen s o i l , ensure the entry of rain and snowmelt water into the forest floor 130 and mineral s o i l . As a result, overland flow as defined by Horton (1933) seldom, i f ever, occurs. Essentially rain and snowmelt water are transformed into stormflow through subsurface pathways (Cheng et a l . , 1975). Before treatment, both study watersheds were covered by conifer-ous tree species of western hemlock (Tsuga heterophylla (Rafn.) Sarg.), western red cedar (Thuja plicata Donn) and Douglas-fir (Pseudotsuga  menziezii (Mayr.) Franco). A more detailed description of these two watersheds has been presented in Chapter I. D. THE EXPERIMENT • The rapid calibration method based on individual peak flow events as suggested by Bethlamy (1963) was adopted in this study. The treatment watershed (watershed 1) was calibrated against the control watershed (watershed 2) during a period of 16 months, May 1972 to August 1973, during which twenty-one peak flow events occurred. In September 1973, the mature forest area of watershed 1, which occupied 61 percent of the watershed area, was clearcut (Figure 1). Twenty-two peak flow events during the post logging period from October 1973 to January 1975 were compared with peak flow events of the pre-logging (calibration) period. Calibration of watersheds over a.period longer than that used this study would provide a more precise regression analysis. It 131 would also provide a wider range of peak flow events. However, the higher cost inherent in a longer calibration period and the benefits to land management of early research results frequently require that calibration be ended after the shortest practicable period (Hornbeck, 1973). This is particularly true in the Coast Mountains of British Columbia, where quantitative data with respect to logging and streamflow are lacking. Another reason for the adoption of rapid calibration is that these two watersheds were also used in a companion study which evaluate the i n i t i a l changes in streamflow chemistry as a result of clearcut logging (Feller, 1974). Since a wide range of peak flow events are included in the calibration period (Table 1) and clearcut logging is a drastic treatment, the 16-month calibration period was considered to be adequate. The logging road constructed for timber removal from water-shed 1 was located along the watershed boundary and well away from the stream. Several skid roads were used, especially in the upper portion of the watershed. These skid roads, however, were generally parallel to land contours and not very steep. Most timber was yarded by a high lead spar system. Although yarding across the streams occurred, landings were located on the ridge tops so that nearly a l l yarding was uphill (Feller, 1974). Logging activities did disturb a considerable area of the watershed. Skid roads took up about 20 percent of the clearcut area, but the total area disturbed--including the roads, skid roads, landing sites and t r a i l s over which logs were skidded--was about 50 percent of the clearcut area. Although logging TABLE I. THE MEANS AND RANGES OF PEAK FLOW PARAMETERS FOR WATERSHED I (W.I) AND WATERSHED 2 (W2) TIME TO THE PEAK PEAK FLOW MAGNITUDE hr m 3 s 1 k m ' 2 f t 3 s m i - 2 W. 1 W. 2 W. 1 W.2. W. 1 W.2 BEFORE LOGGING a 12.3 14.5 0.414 0.335. 37 9 30.6 b 4.0-28.0 4.3 - 36.0 0.089-1.542 0.085-1.130 8.2 - 141 7. 8 - 103.4 AFTER LOGGING a 20.2 17.3 0.392 0.398 35.9 36.4 b 8.0 - 45.0 4.0 - 4 1 0.124-1.015 0.127-1.008 11.3 - 92.8 11.6 - 92.2 a = MEAN b=RANGE 1 activities disturbed a large proportion of the ground surface down to the mineral s o i l , f i e l d observation indicated that they did not cause overland flow and increased concentration downslope. E. ANALYSIS AND RESULTS In the present study, two easily defined peak flow parameters obtained by analysis of individual stormflow hydrographs were evaluated for the pre-logging and post-logging periods. These two parameters are: (1) time to the peak, defined as the time from the beginning of the increase in streamflow, in response to a rainstorm, to the occurrence of the streamflow peak and (2) peak flow magnitude, defined as the maximum rate of streamflow produced during the period of each individual storm less the i n i t i a l baseflow. Data for corresponding time to the peak ( t p ) in hours and 3-1 -2 peak flow magnitude (q p) in m s km of the two watersheds were tabulated for the 16 months preceding and the 16 months following logging. Regression equations, calculated for each period and parameter, are as follows: Before logging = 0.9569 + 0.7808 t p (n = 21, r 2 = 0.9648) q„ =-0.0412 + 1.3593 q (n = 21, r 2 = 0.9940) P l p2> 134 After logging t = 2.8181 + 1.0066t (n = 22, r 2 = 0.9553) p l p2 q n =-0.0125 + 1 . 0 1 9 4 q n (n = 22, r 2 = 0.9540) p l ^2, where 1 and 2 denotes watershed 1 and watershed 2 respectively. The slopes of the regression equations between the two periods for each parameter are compared as follows: Regression Slope Time to Peak Peak Flow Magnitude (Regression Coefficient) a. Before logging 0.7808 1.3593 b. After logging 1.0066 1.0194 c. Difference 0.2258 0.3399 Calculated t value 3.5779 6.2106 Level of significance 1 percent 1 percent Since the regression slopes are s t a t i s t i c a l l y different, i t is not necessary to evaluate changes in intercepts and i t is concluded that there are significant changes in both peak flow parameters follow-ing logging of watershed 1. After logging the average time to the peak was 5.77 hours (about 40 percent) more than that predicted from the pre-logging regression. The average peak flow magnitude for water-3 -1 -2 shed 1 after logging was 0.108 m s km (about 22 percent) less than that predicted from the pre-logging .regression. 135 F. DISCUSSION The results of the analysis indicate that there were s i g n i f i -cant changes in peak flows following clearcut logging of watershed 1. The changes, an increase in time to the peak and a decrease in peak flow magnitude, are contrary to those found in the majority of similar studies elsewhere. The peak flow changes showed some variation for different storms, as indicated by the scatter of points in Figures 2 and 3. This variation may be associated with the time of the year, antecedent precipitation and soil moisture conditions, presence or absence of snow and other factors. Insufficient peak flow events occurred during the study period to permit stratification according to these factors. It should be pointed out that no exceptionally large storms occurred during the post-logging period. The peak flow of a stream, in response to a storm, occurs when the maximum amount of water arrives simultaneously via different pathways from different parts of the watershed at the point of flow measurement. Its magnitude is governed by characteristics of the watershed and stream system. Included among these characteristics are water storage capacity of the watershed, permeability and depth of watershed mantle, the steepness of land and stream and the nature of forest cover in the watershed. The forest's influence on peak flows is a result of the interception and transpiration of trees and the hydrologic characteris-tics of forest soil and forest floor. If forestry practices modify 136 TIME TO THE PEAK ( hr) , WATERSHED 2 Figure 2. The relation of time to the peak of watershed 1 to that of watershed 2. 137 0 0.2 0 4 0,6 0.8 1.0 1.2 1.4 PEAK FLOW MAGNITUDE (m3 s"1 km2), WATERSHED 2 Figure 3. The relation of peak flow magnitude of watershed 1 to that of watershed 2. 138 water losses from a watershed by evapotranspiration or the storage and movement of water in the watershed and stream system, changes in the time to the peak and the magnitude of the peak flow may be expected. As pointed out by Lull and Reinhart (1972), interception of rainfall in the mature forest is not a major influence on peak flows. Any reduction in interception by forest clearcutting would have l i t t l e influence on peak flows. This is particularly true because a large volume of slash (tops and branches) l e f t in watershed 1 continued to intercept r a i n f a l l . Since most peak flow events in the study water-sheds occurred in the dormant season, the influence of the reduction in transpiration due to clearcut logging on peak flows could also be considered insignificant. Snowmelt was not a major contributor to the peak flows observed during the study. Thus forest clearing effects on snowmelt may be ignored. Infiltration capacity of the soil and forest floor might have been reduced by ground surface disturbance associated with the logging a c t i v i t i e s . However, observation on watershed 1 during the storm periods after logging indicated that ground surface disturbance, although occurring extensively over the logged area, did not cause the occurrence of overland flow. An experiment carried out by Willington (1968) in an adjacent area indicated that the i n f i l t r a t i o n capacities of skid-trail and slash-burning areas are significantly less than that of the control forest area, but s t i l l greater than the maximum short duration rainfall intensity recorded for the study area. 139 The above discussion suggests that the conditions in watershed 1 following logging were not favorable to cause decreased times to the peaks and increased peak flow magnitudes. The following discussion will focus on reasons for the increased times to the streamflow peaks and decreased peak flow magnitudes observed in this study. The implication of such changes will also be briefly discussed. As pointed out previously, on the study watersheds, rainfall is transformed into stormflow as a result of flow through subsurface pathways. Two types of subsurface stormflow pathways normally exist in the forest s o i l s : (1) the soil matrix of the forest floor and mineral s o i l , (2) the interconnected soil channels, passing through the mineral s o i l . The resistance to storm flow in the soil channels is much less than that of the soil matrix. Thus soil channels can serve as very important pathways of stormflow concentration (de Vries and Chow, 1973; Cheng et a l . , 1975). Field examination indicated that ground surface disturbance did not result in the occurrence of overland flow; however, this disturbance might result in changes in subsurface stormflow pathways. Based on the results of a plot study in a nearby watershed, de Vries and Chow (1973) speculated that ground surface disturbance would close some entrances to soil channels, resulting in an increase in temporary water storage in the soil matrix. They suggested that this would reduce subsurface stormflow and, conse-quently reduce the peak flow. 140 In their study, de Vries and Chow instrumented a 2.5 m by 3.5 m plot with tensiometers located at several depths to compare the response of soil to rainfall (natural and a r t i f i c i a l ) before and after forest floor disturbance. Pertinent observations by these authors are pre-sented here as Figures 4 and 5. The non-steady state total water potential profiles of Figure 4 show that in undisturbed s o i l , water reached the B horizon before i t reached horizons closer to the soil surface, indicating non-uniform wetting or wetting of the soil from below, de Vries and Chow attributed this phenomenon of wetting from below to the conductance of water through soil channels and suggested that stormflow transmission in the undisturbed forest soils of the Coast Mountains is dominated by the presence of soil channels by means of which water tends to reach stream channels under minimal influence of the soil matrix. In contrast, in the non-steady state total water potential profiles of Figure 5 the disturbed soil portrays the normal wetting from the surface down. The change in wetting pattern follow-ing forest floor disturbance down to the mineral soil was interpreted by de Vries and Chow to be the result of the closure of some of the soil channels, rendering them inoperative as pathways of subsurface stormflow The results of the present study tend to support the specula-tions by de Vries and Chow that the closure of some soil channel entrances as a result of ground surface disturbance was responsible for the lower peak flow after logging. However, i t should be pointed TOTAL WATER POTENTIAL (cm of water) Figure 4. Total water potential plotted as a function of depth, during a wetting cycle before forest floor disturbance. The number against each line is the time in minutes since the beginning of water application. The reference height at: which the gravitational potential equals zero is 45 cm above the soil surface (after de Vries and Chow, 1973). - 5 0 - 3 0 Figure 5. TOTAL WATER POTENTIAL (cm of water) Total water potential (\b) plotted as a function of depth, during a wetting cycle, in the same plot as Figure 4 after forest floor disturbance. The number against each, line is the time in minutes since the beginning of water application. The reference height at which the gravitational potential equals zero is 45 cm above the soil surface (after de Vries and Chow, 1973). 143 out that the decreased peak flow magnitudes observed in this study were mainly due to the flattening out of the hydrograph rather than the reduced stormflow volume. This is apparent in an examination of the pre-logging and post-logging storm hydrographs of the. two . watersheds. This examination indicated that the increased time to the peak of watershed 1 after logging resulting from a delayed peak, rather than earlier hydrograph rise. It is suggested that lower rate of stormflow transmission through the soil matrix caused this increased time to the peak and, consequently, lower peak flow magnitude. One of the factors that promoted the present study is the frequent occurrence of peak flow flooding in the British Columbia coastal region. Such peak flow flooding appears to often take place even when the soil water storage volume of the watershed is only partially f i l l e d . It has been suggested that i f i t is true that lower peak flow magnitudes following logging are mainly related to the closure of soil channels caused by ground surface disturbance associated with logging, then some ground surface disturbance technique may be used to delay quick subsurface stormflow in the watershed and reduce peak streamflow magnitude (de Vries and Chow, 1973). Besides ground surface disturbance, a technique suggested by Whipkey (1967), based on a similar principle, might also be used to delay stormflow. Whipkey found that the total subsurface flow from water sprinkled plots was reduced: by one-half by cutting 30 cm (1 foot) wide trenches across the plot to a depth of 1.2 m (4 feet) 144 and backfilling with a heterogeneous mixture of displaced soil material. Such a technique could take f u l l advantage of detention storage in a soil where subsurface stormflow is otherwise so rapid that this storage is not f u l l y used even during a major storm. As yet no study has been made on the application of such techniques on a watershed scale. The best possibility of reducing peak flow magnitude by delay-ing subsurface stormflow may be the application of the above techniques to the important source areas of the watershed. A design for systematic selection of the most suitable area in the watershed based on a peak flow routing technique may be desirable. When applying such techniques, efforts must be made to keep the occurrence of overland flow to a minimum because i t would cause soil erosion and would not reduce the streamflow peak. G. CONCLUSION Following clearcut logging of watershed 1, the time to the streamflow peak increased and peak flow magnitude decreased,both significantly. Such changes are lik e l y related to the stormflow genera-tion mechanisms of the study area and the ground surface disturbance associated with logging. As there may be some other factors influencing the changes in these two peak flow parameters, the explanation given in this paper requires further investigation. For example, an increased channel roughness resulting from a large increase of debris in the stream channel following logging may account for some portion of the 145 reduced stormflow transmission rate. Although the results of this single experiment should not be interpreted too generally because they reflect the climatic, s o i l , and vegetation factors unique to the study area, they are indicative of potential changes in these two peak flow parameters for other watersheds with a similar hydrologic environment. This study also indicates that carefully controlled investigations on a watershed scale of the hydrologic impact of logging are essential in the testing of conclusions based on plot studies. Since timber harvesting in the forests of the West Coast of Canada is being extended into rugged mountainous terrain with steeper slopes, the impact of clearcut logging in these areas may be different from that observed on the relatively gentle slopes of the area of this study. H. LITERATURE CITED Bethlamy, N., 1963. Rapid calibration of watersheds for hydrologic studies. Int. Assoc. Sci. Hydrl. Bull. 8(3): 38-42. Cheng, J.D., Black, T.A. and Willington, R.P., 1975. The generation of stormflows from small forested watersheds in the Coast Mountains of southwestern British Columbia. Paper presented at the Canadian Hydrology Symposium, August 11-14, 1975, Winnipeg, 7 p. 146 de Vries, J. and Chow, L., 1973. Hydrology of mountain slopes. Environment Canada, Final 1973 report, Dept. Soil Sci. University of British Columbia, 265 p. Feller, M.C., 1974. Initial effects of clearcutting on the flow of chemical through a forest-watershed ecosystem in south-western British Columbia. Unpublished Ph.D. Thesis, University of British Columbia. Hewlett, J.D. and Helvey, J.D., 1970. Effects of forest clear-f e l l i n g on the storm hydrograph. Water Resources Res 6(3): 768-782. Hewlett, J.D. and Hibbert, A.R., 1961. Increase in water yield after several types of forest cutting. Int. Assoc. Hydrol. Oub. 6: 5-7. Hoover, M.D., 1945. Effect of removal of forest vegetation upon water yield. Amer. Geophys. Union Trans. 6: 969-977. Hornbeck, J.M., 1973. Stormflow from hardwood-forested and cleared watersheds in New Hamshire. Water Resource Res. 9(2): 346-354. Hornbeck, J.M., 1973. The problem of extreme events in paired-watershed studies. U.S. Forest Serv. Res. Note Ne-175, 4 p. 147 Horton, R.E., 1933. The role of i n f i l t r a t i o n in the hydrologic cycle. Eos Trans. AGU, 14: 446-460. L u l l , H.W. and Reinhart, K.G., 1972. Forest and floods in the eastern United States. U.S. Forest. Serv. Res. Paper NE-226. 94 p. Plamondon, A., 1972. Hydrologic properties and water balance of a Canadian west coast watershed. Unpublished Ph.D. Thesis, University of British Columbia. Reinhart, K.G., 1964. Effect of a commercial clearcutting in West Virginia on overland flow and storm runoff. J. Forestry 62: 167-171. Roddick, J.A., 1965. Vancouver, North Coquitlam and Pitt Lake map areas, British Columbia: with special emphasis on the evolution of the plutonic rocks. Geol. Surv. Can. Mem 445, 276 p. Rothacher, J., 1973. Does harvest in west slope Douglas-fir increase peak flow in small forest stream? U.S. Forest Serv. Res Paper PNW-163, 13 p. Whipkey, R.Z., 1967. Storm runoff from forested catchments by subsurface routes. Int. Assoc. Sci. Proc. Leningrad Symposium: 773-779. 148 Willing-ton, R.P., 1968. Some effects of slashburning, clearcutting, and skidroads on physical-hydrologic properties of coarse glacial soils in coastal British Columbia. Unpublished M.Sc. Thesis, University of British Columbia, 149 p. APPENDIX 1 PHOTOGRAPHS OF THE INSTRUMENTATION OF THE STUDY WATERSHEDS 149 150 1. Streamflow gauging station at the mouth of Jamieson Creek 151 2. Streamflow gauging station at the outlet of Elbow Creek 152 3. Streamflow gauging station of watershed 1, U.B.C. Research Forest 153 6. The weighing-type recording rain gauge used in this study. The one shown is located at the 270 m elevation of watershed 1, U.B.C. Research Forest APPENDIX 2 A LISTING OF STORMFLOW PARAMETERS EVALUATED FROM STORM HYDROGRAPHS OF JAMIESON CREEK WATERSHED (1970-1974). 155 A LISTING OF STORMFLOW PARAMETERS EVALUATED FROM STORM HYDROGRAPHS OF JAMIESON CREEK WATERSHED (1970-1974) Stormflow Parameters Year " ° ™ Date- " l . \ *P *d " P (1 s km ) hour hour mm mm hour 1970 1971 1972 1973 1 11/06 55.58 739.2 30.75 81.00 - 70.61 146.00 10. 00 13.30 2 11/10 40.29 262.1 20.00 41.00 11.73 50.80 13. 00 6.50 3 11/15 49.06 1261.2 9,75 90.00 87.12 195.60 12. 25 25.70 4 09/04 30.60 297.9 26.00 46.00 13.93 38.10 6. 50 9.74 5 09/08 28.91 59.4 6.00 26.00 2.26 17.78 9. 00 2.06 6 09/10 31.03 607.2 o 10.50 46.00 25.57 41.91 5. 50 19.57 7 09/27 13.42 273.9 8.50 58.00 22.95 66.04 15, 00 20.41 8 10/03 47.37 831.8 12.50 54.00 48.37 78.74 7. 00 19.56 9 10/12 19.29 455.5 10.25 48.00 16.89 40.64 7. .00 23.62 10 10/18 22.00 716.3 82.50 121.00 59.83 149.90 8. ,50 32.56 11 10/24 71.53 853.4 20.50 52.00 42.21 76.20 6. .50 11.93 12 . 11/07 93.10 568.8 54.00 167.00 100.10 152.40 6.10 13 07/07 30.45 212.9 22.00 57.00 10.91 48.51 6. ,00 6.99 14 07/10 96.72 1374.0 • 47.25 88.00 118.60 1.65.10 3. ,50 14.20 15 09/18 6.04 192.4 6.00 39.00 5.52 37.59 7. 80 31.87 16 09/20 23.44 422.9 16.00 86.00 52.33 165.90 8. 00 18.04 17 10/09 8.77 13.9 3.50 10.00 0.18 6.35 1.59 18 10/25 6.88 45.1 5.00 21.00 1.07 15.24 6.55 1'9. 10/27 10.04 24.6 32.00 36.50 0.82 33.02 11.93 20 11/01 11.42 709.7 69.00 243.50 179.30 328.90 6.10 21 08/16 4.97 78.60 6.50 38.00 2.38 21.75 9. ,50 15.82 22 09/06 5.26 23.20 6.50 22.00 0.37 8.26 4. ,75 4.41 23 09/18 3.89 12.30 6.50 16.00 0.21 12.70 5. ,00 3.16 24 09/19 6.54 222.20 25.00 67.00 17.74 69.21 10. ,50 34.00 25 09/26 12.39 10.30 10.50 25.00 0.15 5.08 13. ,00 0.80 26 09/30 9.61 14.60 7.00 14.00 0.21 7.62 5. ,50 1.50 *the date on which the storm event began. 157 A Listing of Stormflow Parameters Evaluated From Storm Hydrographs of Jamieson Creek Watershed (1970-1974) (continued) Stormflow Parameters Year Storm No. Date* q i qp (1 s'W 2 ) P hour ud hour v mm P mm °L hour pf 1973 1974 27 10/05 8.77 94.90 4.00 22.00 1.76 12.70 5.75 10.80 28 10/11 13.68 540.40 38.00 79.00 32.30 99.06 9.00 39.50 29 10/15 40.43 108.90 7.00 . 24.00 3.98 13.97 2.09 30 10/17 33.69 220.30 20.00 49.00 15.44 31.75 14.50 6.50 31 10/20 55.58 218.70 8.50 65.00 24.69 60.96 14.50 3.94 32 10/23 59.34 357.30 18.00 39.50 32.66 102.90 9.00 6.03 33 10/26 78.90 208.40 27.00 62.00 20.18 60.96 5.00 2.64 34 10/29 98.38 131.80 12.00 25.50 5.64 15.24 8.00 1.34 35 10/20 5.20 36.80 4.50 15.50 0.72 14.86 5.00 7.07 36 10/27 5.50 236.30 8.50 32.00 7.56 35.56 6.50 42,96 37 11/04 7.96 ' 777.70 32.25 74.75 32.33 80.01 5.50 97.75 38 11/08 37.46 383.50 12.50 50.00 17.62 49.02 7.00 10.20 39 11/11 47.37 726.30 20.00 62.00 51.31 81.28 5.00 15.30 40 11/17 31.46 860.70 81.50 117.00 68.47 148.60 27.35 41 11/23 70.81 739.20 33.50 69,00 107.80 205.70 10.40 *the date on which the storm event began. APPENDIX 3 THE RELATIONSHIP OF PEAK FLOW MAGNITUDE AND STORM RAINFALL BEFORE PEAK FLOW OCCURRENCE FOR JAMIESON CREEK WATERSHED 158 159 THE RELATIONSHIP OF PEAK FLOW MAGNITUDE AND STORM RAINFALL BEFORE PEAK FLOW OCCURRENCE FOR JAMIESON CREEK WATERSHED 1. q = -0.4754 + 10.OOlOPi (n=26 r2=0.62) 2. q = -384.8 + 38.8651P6 (n=26 r2=0.68) 3. q p = -352.5 + 31.1449P6 + 5.5390 (Pi - Pe) (n=26 r2=0.79) where q p is the peak flow magnitude for an individual storm, P-j is the total rainfall amount before the occurrence of streamflow peak, and P6 is the total rainfall amount during the 6-hour period proceding the streamflow peak. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0093910/manifest

Comment

Related Items