UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Hydrologic response during snowmelt in three steep headwater catchments : Ringrose Slope, Slocan Valley,… Whyte, Dion M. 2004

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

Item Metadata


831-ubc_2004-0325.pdf [ 7.29MB ]
JSON: 831-1.0091595.json
JSON-LD: 831-1.0091595-ld.json
RDF/XML (Pretty): 831-1.0091595-rdf.xml
RDF/JSON: 831-1.0091595-rdf.json
Turtle: 831-1.0091595-turtle.txt
N-Triples: 831-1.0091595-rdf-ntriples.txt
Original Record: 831-1.0091595-source.json
Full Text

Full Text

H Y D R O L O G I C R E S P O N S E D U R I N G S N O W M E L T IN T H R E E S T E E P H E A D W A T E R C A T C H M E N T S : R I N G R O S E S L O P E , S L O C A N V A L L E Y , BRITISH C O L U M B I A by DION M . W H Y T E B . S c , University of Victoria, 2001 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department of Geography) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F BRITISH C O L U M B I A A P R I L 2004 © Dion M . Whyte, 2004 \ Library Authorization In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. g><p/i / J U ^ if/6/zaaq Name of Author (please print) Date (dd/mm/yyyy) Title of Thesis Degree: Department of {f)pQ0. x&p^y The I Iniworci+w /-if Rritioh f^niumhia / The University of Briu^hCoWf bra Vancouver, BC Canada A B S T R A C T Using a combination of hydrometric measurements, geochemical and isotopic tracers, and analysis of historical data, three steep headwater catchments of the Slocan Val ley , Bri t ish Columbia, Canada, were described and contrasted in terms of runoff processes and pathways during the 2002 freshet. Streamflow responsiveness was shown to be related to the degree to which connectivity between source areas of meltwater and stream channels was maintained as the snowline progressed upslope. The Thickett Creek catchment displayed the most sluggish streamflow response and was dominated by relatively deep subsurface flow throughout the freshet. The Gurn Brook catchment demonstrated the flashiest streamflow response and experienced significant channel flow relative to the other catchments, with subsurface flow maintaining connectivity between source areas and the channel network. The G u m Spring catchment demonstrated streamflow response intermediate between the other catchments; however, high unit runoff and unit peak flow relative to the other catchments suggested significant inter-catchment transfers of bedrock groundwater into the G u m Spring catchment. Whi le a soil-water reservoir within Gurn Spring contributed to streamflow during peak flow, contributions from bedrock groundwater dominated during baseflow and hydrograph recession. Attempts to perform an isotopic hydrograph separation were confounded by the complex variation in new-water isotopic concentrations in both time and space. Qualitative interpretation of isotopic results indicated the dominance of old water contributions during baseflow, with increasing new water contributions during hydrograph rise. Similar isotopic concentrations between meltwater and streamwater at the time of peak flow prevented resolution of the relative contributions from old and new-water at that time. Relative to other studies, the study catchments displayed surprisingly little diurnal streamflow response during snowmelt. Possible explanations for this are presented and discussed, as are the potential impacts of forest harvesting and road construction on hydrologic response in the three catchments. Finally, the implications of the research findings for catchment-scale modelling were examined. T A B L E O F C O N T E N T S Abstract ii Table of Contents iii Lis t of Figures V Lis t of Tables vi Acknowledgements vi i 1 Introduction 1 1.1 Background 1 1.2 Review of Runoff Generation in Temperate Forested Catchment 1 1.2.1 Runoff processes and pathways 2 1.2.2 Methods for studying runoff pathways 4 1.2.3 Distinctive characteristics of snowmelt runoff processes 6 1.3 Objectives, Questions, and Research Hypotheses 9 1.4 Structure of the Thesis 10 2 Methods 11 2.1 Study Area 11 2.1.1 Location 11 2.1.2 Climate 11 2.1.3 Geology 12 2.1.4 Vegetation 13 2.2 F ie ld Site 13 2.2.1 Location and physiography 13 2.2.2 Soils IV 2.3 Climate, Snowpack, and Streamflow Data 17 2.3.1 Meteorology 17 2.3.2 Snow surveys 18 2.3.3 Snow line mapping 18 2.3.4 Streamflow 18 2.4 Piezometric Data 19 2.5 F ie ld Observations of Surface F low 20 2.6 Water Chemistry 21 2.6.1 Streamwater 21 2.6.2 Mel t and rainwater 21 2.6.3 Piezometers 21 2.6.4 Overland flow 22 2.7 Laboratory Analysis 22 2.7.1 Oxygen-18 22 2.7.2 Hydrochemistry 23 2.7.3 Dissolved organic carbon 23 2.8 Data Analysis 23 2.8.1 Analysis of peak flow generating mechanisms 23 2.8.2 Inter-catchment comparison of diurnal streamflow response 23 2.8.3 Statistical analysis of 1 8 0 inputs 25 2.8.4 Principle components analysis of streamwater chemistry 26 3 Results 27 3.1 Overview of the Study Period 27 iii 3.2 Streamflow Variability ...31 3.2.1 Comparison of streamflow amongst the three streams 31 3.2.2 Timing and magnitude of peak flows 36 3.2.3 Diurnal streamflow variability 37 3.3 Piezometric Data.... . . . . . . . . ...39 3.3.1 Time series of piezometer water levels 39 3.3.2 Distribution of maximum recorded water levels ..42 3.4 Observations of Surface Flow. 43 3.4.1 Saturation Overland Flow (SOF).... 43 3.4.2 Overland flow along roads 46 3.5 Isotopic Concentrations 51 3.5.1 Isotopic concentrations in melt and rainwater 51 3.5.2 Isotopic concentrations in streamwater 56 3.5.3 Isotopic concentrations in subsurface waters....... 56 3.6 Water Chemistry.... .'56 3.6.1 Water chemistry in piezometers 56 3.6.2 Water chemistry in streamflow .62 3.6.3 Concentration - discharge relations 62 3.6.4 Principle components analysis of water chemistry 68 4 Discussion. • • • • 72 4.1 The Potential Role of Inter-Catchment Groundwater Flow 72 4.2 What are the relative proportions of new and old water in streamflow and how do they change over the snowmelt period? . 73 4.2.1 Challenges in applying isotope hydrograph separation 73 4.2.2 Qualitative interpretation of oxygen-18 data 74 4.2.3 Comparison of results with other studies 75 4.3 Comparison of Streamflow Response and Runoff Pathways 75 4.3.1 Comparison amongst the study catchments 75 4.3.2 Comparison of results with other studies • .78 4.4 Implications for the Hydrologic Effects of Forest Harvesting and Road Construction 79 4.4.1 Forest harvesting 80 4.4.2 Road construction 80 4.5 Implications for Catchment Scale Modelling 82 5 Conclusions 83 5.1 Summary of Findings. 83 5.2 Recommendations for Future Research 84 References • • 86 Appendix A: Freshet Hydrographs ••92 Appendix B: Piezometer Water Levels and Crest Data 95 Appendix C: Observations of Saturation Overland Flow 100 Appendix D: Isotopic Concentrations in Melt and Rainwater 103 iv "1 \ L I S T O F F I G U R E S Figure 1.1 Runoff pathways 3 Figure 2.1 Climate normals at N e w Denver 12 Figure2.2 Map o f the field site 16 Figure 2.3 Illustration of the method used to measure diumal streamflow variation 25 Figure 3.1 Snow-water equivalents 28 Figure 3.2 Overview of the field season '.' .30 Figure 3.3 Snow line position .32 Figure 3.4 Freshet hydrographs .33 Figure 3.5 Inter-annual variation in unit runoff. 34 Figure 3.6 Diumal streamflow variability.. 38 Figure 3.7 Water levels and crests measured in 3 piezometers 40 Figure 3 8 Distribution o f maximum water levels recorded in piezometers 41 Figure 3.9 Max imum subsurface water levels grouped by hydrogeomorphic position. . . . . . . . .42 Figure 3.10 Saturation overland f low. . . ........45 Figure 3.11 Overland flow along roads (Apr. 15-18, 2002). 48 Figure 3.12 Overland flow along roads (Apr. 22-24, 2002). . . . . . . ....49 Figure 3 13 Overland flow along roads (May 15-16, 2002) 50 Figure 3.14 Variation in oxygen-18 with elevation .....53 Figure 3.15 Variation in oxygen-18 with time 54 Figure 3.16 Time series oxygen-18 in streamwater and meltwater ..55 Figure 3.17 Chemical concentrations in subsurface water 59 Figure 3.18 Variation in chemical concentrations in subsurface water with elevation . . . . . . . .60 Figure 3.19 Comparison o f RNa/k in subsurface water and streamwater. 61 Figure 3.20 Time series o f chemical concentrations in streamwater 63 Figure 3.21 Concentration - discharge relations in Thickett Creek .65 Figure 3.22 Concentration - discharge relations in Gurn Brook. 66 Figure 3.23 Concentration - discharge relations in G u m Spring 67 Figure 3.24 Bivariate plot o f S i vs. R N A / K • ••• 71 v \ L I S T O F T A B L E S Table 1.1 Catchment-scale studies of snowmelt runoff processes 8 Table 3.1 Air temperature and rainfall at Ringrose Slope for several years on record.. 29 Table 3.2 Timing and magnitude of peak flows in relation to snow-water equivalents 35 Table 3.3 Comparison of diurnal streamflow response 37 Table 3.4 Observations of saturation overland flow 44 Table 3.5 Results of correlation analysis (oXygen-18 and date)..... 52 Table 3.6 Results of correlation analysis (oxygen-18 and elevation) .52 Table 3.7 Results of correlation analysis (subsurface water chemistry and elevation) 58 Table 3.8 PCA results. , ...69 vi A C K N O W L E D G E M E N T S I would like to take this opportunity to thank a number of people for their support and help along this path. M y humble thanks go to my supervisor Dr. Dan Moore for his support (both financial and academic), interest, enthusiasm, and keen intelligence throughout this project. I hope that all graduate students are so lucky as to have such a helpful and involved supervisor as he. Many thanks to my favourite pool shark, George Richards, for his invaluable assistance in the field, for English teas, and for nearly getting me into a bar-fight on the first night of the field season by beating all the locals on their own table. I would also like to thank Geoff H i l l and Leigh Herbert for helping me with my field work. Thanks to al l of you for climbing up and down that darned road with me, and for accompanying me on those oh-so-anticipated midnight sampling visits. M y thanks also to Dave Toews, Dr . Peter Jordan, David Gluns and especially Emilee Fanjoy of the Brit ish Columbia Ministry of Forests for their continued assistance and support, and for putting up with my plethora of information requests. A l s o thanks to David Hutchinson for his assistance in accessing climate data, and to Dr. A m o d Dhakal for his assistance in generating catchment boundaries and for his painstaking efforts to model hydrologic response on Ringrose Slope using D H S V M . I would also like to thank Dr . Olav Slaymaker and Dr. Hans Schreier for their insightful comments and suggestions during the formative stages of this project. Special thanks to Dr. Slaymaker for his added contributions as I was revising this work. Final ly, and most importantly, I give thanks to my beloved wife Lindsay Cole. Thank you for putting up with me, for never giving up on me, and for supporting me while I finished this. Y o u are my superstar! vii C H A P T E R 1 I N T R O D U C T I O N 1.1 B A C K G R O U N D The processes and pathways by which water moves through a catchment to its outlet exert a control on the timing, magnitude and chemistry of streamflow. These processes can be influenced by land use activity, such as the construction of logging roads (e.g. Megahan, 1972; Luce and Wemple, 2001; Wemple, 2003). An understanding of runoff processes is therefore essential for predicting the effects of land use, such as forest harvesting, on water quantity and quality. Many residents of southern interior British Columbia derive their water supply from small, weakly incised streams draining steep valley walls, locally referred to as "face unit" streams (Forest Practices Code, 1999). To address the potential effects of forest harvesting on these small streams, a paired-catchment experiment involving three streams was initiated in the late 1980s on the Ringrose Slope in the Slocan Valley in the southern interior of British Columbia. One of the catchments was to be subjected to logging and road construction, with the other two being maintained as untreated experimental controls. Two studies were initiated in 2001 to complement the paired-catchment experiment, one to study runoff processes, the other to apply a spatially distributed, physically based hydrologic model. The research presented in this thesis examined pre-harvest snowmelt runoff processes in the three catchments to provide a basis for interpreting the results of the paired-catchment study. It will also provide data for internal validation of the hydrologic simulations. The next section presents a review of the current understanding of snowmelt runoff processes in forested catchments in order to provide the scientific context for the research objectives. 1.2 R E V I E W OF RUNOFF GENERATION IN T E M P E R A T E F O R E S T E D C A T C H M E N T S 1.2.1 Runoff processes and pathways Runoff processes encompass both the sources of water, usually distinguishing between old (pre-event) and new (event) water, as well as the pathways by which water travels to the stream channel. The four major runoff pathways are: (1) Overland Flow (Q0); (2) Subsurface Storm Flow (SSF) or Throughflow (Qt); (3) Groundwater Flow (Qg); and (4) Direct Precipitation onto the Channel (Qp) (Figure 1.1; Ward and Robinson, 2000). Overland flow can be generated 1 by two processes: Hortonian Overland F l o w (HOF) and Saturation Overland F low (SOF). The former occurs when the rate o f water input exceeds the infiltration capacity o f the soil and excess rainfall or snowmelt discharges over the land surface (Horton, 1933). Consequently, the temporal source o f H O F is new rainwater and the runoff pathway is overland. Saturation Overland F low occurs when the water table rises to the ground surface, creating saturated conditions (Dunne and Black, 1970a, b). It comprises Direct Precipitation onto Saturated Areas ( D P S A ) and Return F low (RF). The former involves the deflection o f inputs o f rain or meltwater onto saturated areas overland towards the stream channel, whereas R F involves the exfiltration o f subsurface water that is then free to travel overland to the stream channel. Saturation overland flow can therefore be a mixture o f old subsurface water and new water from either rainfall or snowmelt. Subsurface Storm F l o w occurs when infiltrated water flows laterally through the upper soil horizons either as unsaturated flow or as a perched water table above the main saturated zone (Freeze, 1974). Groundwater flow is that portion o f the subsurface flow that moves laterally through the permanently saturated zone to the stream channel (Ward and Robinson, 2000). 2 Figure 1.1. Runoff pathways (from Ward and Robinson, 2000). Studies in forested environments have generally discounted HOF as a significant runoff process due to the high infiltration capacity of forest soils (Sidle et al., 2000). In snowmelt dominated systems, HOF may be significant where impermeable frost develops at the ground surface or where an impermeable ice layer forms at the base of the snowpack (Dunne and Black, 1971; Jordan, 1978; Price and Hendrie, 1983; Nyberg et al., 2001). In near-channel environments where the water table and capillary fringe are close to the ground surface, as in wetlands and riparian corridors, SOF is the dominant runoff process (Freeze, 1972; Freeze, 1974; Sidle et al., 2000). Subsurface stormflow has been suggested as a dominant runoff process on forested slopes in areas characterized by shallow soil horizons of high permeability at the surface, convex slopes feeding deeply incised channels, and high saturated hydraulic conductivities (Freeze, 1974). Rapid downslope flow through forest occurs via preferred pathways, such as macropores formed by decayed root channels (Mosley, 1979; Turton, 1992; Hutchinson and Moore, 2000; Noguchi et al., 2001). Preferential flow through fractured bedrock has also been cited as a possible means by which water may move downslope and enter the channel network (Laudon and Slaymaker, 1997). The ability of SOF and SSF to contribute to rapid streamflow response is linked to the variable source area concept, which essentially treats saturated areas as extensions of the channel network. As saturated areas grow and shrink in response to water inputs, so do the volumes of SSF, tapped arid discharged as return flow, and DPSA entering this extended channel network (Dunne and Black, 1970a, b; Dunne and Black, 1971; Freeze, 1972, Freeze, 1974). 3 Sidle et al. (2000), working in a 2.48 ha catchment at Hitachi Ohta Experimental Watershed, Japan, developed a conceptual hydrogeomorphic model to describe the spatially distributed nature of stormflow response over a period of rainfall and increasing soil moisture. Sidle et al. identified three hydrogeomorphic units, which are essentially source areas that display distinctive hydrologic behaviour: (1) narrow riparian corridors; (2) linear hillslopes; and (3) geomorphic hollows or zero-order basins. During dry conditions the majority of stormflow was generated from channel interception and SOF in narrow riparian units. As wetness increased, the lower extent of the linear hillslope units linked with riparian corridors to produce subsurface flow from the soil matrix and augment streamflow. Further wetting resulted in subsurface flow propagating over larger hillslope areas, the commencement of preferential flow, and "tapping" of geomorphic hollows, thereby linking all three hydrogeomorphic units and further augmenting streamflow. As the study basin was relatively small and water inputs (rainfall) were assumed uniform over the basin area (Sidle et al., 2000), no attempt was made to describe the interaction between spatially distributed water inputs and the spatial distribution of the resulting runoff processes. The hydrogeomorphic concept is a conceptual framework for examining temporal and spatial variation in the characteristics of catchment flowpaths and hydrologic response. This is achieved by focussing on the connectivity between hydrogeomorphic units (i.e., source areas possessing relatively homogeneous hydrologic and geomorphic characteristics) in order to generate a conceptual model of catchment response (Sidle et al, 2000). Other studies have adopted the hydrogeomorphic concept in principle, if not in name. Ladouche et al. (2001) examined the runoff response from three spatially distinct hydrologic zones within the Strengbach catchment, France. Hangen et al. (2001) documented the temporal process of stormflow generation from near-stream and hillslope areas as follows: (1) rapid delivery of water by infiltration and soil water displacement at near-stream areas supplemented by SOF; (2) rapid depletion of the near-stream soil and groundwater reservoir; and (3) delayed subsurface water contribution from hillslope areas. Spence and Woo (2003) analysed runoff processes in a subarctic Canadian Shield catchment in terms of the interactions between bedrock upland and soil-filled valley units. 1.2.2 Methods for studying runoff processes Methods employed in the study of runoff processes vary in their uses and limitations A number of isotopic tracer studies have revealed that pre-event water dominated streamflow in rainfall and snowmelt driven systems and produced rapid streamflow response (Sklash and 4 Forvolden, 1979; Pearce et al. , 1986; Sklash et al., 1986; Dewalle et al . , 1988, Moore, 1989; Buttle and Sami, 1992; Sueker et al., 2000; Ladouche et al., 2001). Geochemical tracers have been used to separate streamflow into components based on the chemistry o f specific contributing waters, or end-members (Christophersen et al., 1990; Hooper et al., 1990; Sueker et al., 2000; Burns, 2001; Katsuyama et al., 2001). End-members have been defined in terms o f chemical differences between riparian vs. hillslope waters (Hooper et al. , 1990; Hooper et al., 1998), surface vs. subsurface flowpaths (Wels et al., 1991; Sueker et al. , 2000), and between subsurface flowpaths at different depths (Christophersen et al., 1990; M c G l y n n et al., 1999). Hydrometric methods involve measurement of hillslope flow (e.g., by using interception trenches and overland flow collectors) and subsurface water potential (using tensiometers, wells and piezometers), and have been used to identify the mechanisms responsible for the typically large proportions o f pre-event water contributing to stormflow, as inferred from isotopic studies. These include groundwater ridging, saturation overland flow, translatory flow, kinematic waves, and release o f water from surface storage (Buttle, 1994). Integrated tracer studies, which combine hydrometric measurements with tracer analysis, have underscored the risks o f inferring streamflow generating mechanisms from tracer studies without supporting evidence from hydrometric measurements and observation (Buttle and Sami, 1992). Ladouche (2001) recommended that hydrologic studies employ "complementary approaches" in time and space. Tracer studies should employ both isotopic and geochemical tracers i f they are to identify both runoff pathways and temporal sources o f runoff (Ladouche, 2001). A range o f chemical species has been used to provide information on flow paths. Dissolved organic carbon ( D O C ) has been employed to identify runoff contributions from organic-rich surface and shallow subsurface waters (Ladouche et al. , 2001; Shanley et al. , 2002) and from different soil depths within the riparian zone ( M c G l y n n et al., 1999). Si l ica and the major cations Ca, M g , N a and K have been used extensively to identify runoff contributions from deeper subsurface pathways characterized by long contact times in the presence of mineral soils and/or bedrock (Christophersen et al., 1990; Hooper et al., 1990; Wels et al., 1991; Anderson et al., 1997; Buttle and Peters, 1997; Sueker et al. , 2000; Ladouche et al. , 2001). Subsurface flow contributions have also been inferred using temperature (Kobayashi, 1985; Kobayashi et al., 1999) and electrical conductivity (Kobayashi, 1986; Kobayashi et al., 1999; Laudon and Slaymaker, 1997) by employing a mass-balance equation and variations on Hal l ' s mixing models (Hall , 1971). Fluorescence properties o f dissolved organic carbon have been employed to identify runoff contributions from the transient saturated zone (Katsuyama and Ohte, 2002). 5 1.2.3 Distinctive characteristics of snowmelt runoffprocesses Most studies o f runoff processes in steep forested catchments have focused on rainfall events (e.g., DeVries and Chow, 1978; Anderson et al., 1997; Montgomery et al., 1997; Hoeg et al., 2000; K a t s u y a m a and Ohte, 2002). Snowmelt runoff processes differ from rainfall-runoff processes due to the distinctive spatial variation in the depth and attributes o f the snowpack, in meltwater generation (e.g., in relation to aspect, elevation and vegetation cover), as wel l as the effect o f internal snowpack processes influencing percolation rates. Snowmelt runoff processes can be viewed as a function of: (1) the spatial distribution o f snow-water equivalent (SWE) and the rate o f melt (controlled by meteorological variables); (2) internal snowpack processes such as metarhorphism, phase changes and percolation; and (3) the physical processes and characteristics that govern the movement o f meltwater through the basin (Ward and Robinson, 2000). The distribution o f the snowpack and the rate o f melt w i l l influence the timing, location, and magnitude o f meltwater inputs. For example, alpine catchments may experience a positive net balance of snow storage at higher elevations during years o f high snow accumulation, resulting near-normal catchment runoff despite above average snow-water-equivalents (Teti, 1979). Dunne and Black (1971) observed that the lag-time between the onset o f melt and runoff initiation was primarily controlled by the depth o f the snowpack. There was an approximately linear relation between the depth o f the snowpack and the lag between the onset o f snowmelt and runoff initiation. Early in the melt period internal snowpack processes may dominate (Colbeck, 1979), and are influenced by the physical properties o f the snowpack (Marsh and W o o , 1984, 1985). For example, dye tests by Jordan (1978) in a small alpine basin in M c G i l l i v r a y Pass, British Columbia, revealed that the flow of meltwater within the snowpack was affected by irregularities in the snowpack such as ice layers and snow layers o f different textures and hydraulic conductivities. The rate at which meltwater can be transmitted through the snowpack greatly affects the timing o f the streamflow response (Ward and Robinson, 2000). A s mentioned earlier, the presence o f impermeable ground frost can limit infdtration at the snow-soil interface (Price and Hendrie, 1983; Thunholm et al., 1989; Stahli et al., 1996); however, many studies in forested environments have found either no impermeable frost layer (Moore, 1989); or its effects on runoff response to be negligible (Roberge and Plamondon, 1987; Shanley and Chalmers, 1999). Later in the melt period, streamflow response is governed by the movement o f meltwater through the drainage basin by means o f the runoff pathways outlined above. Runoff processes in steep, snowmelt dominated catchments are further complicated by elevational changes in hydrogeomorphic setting and the upslope movement o f the snowline 6 \ \ resulting in spatially variable meltwater inputs. O f the studies focused on catchment runoff processes and pathways during snowmelt in temperate forested regions, only the study by Sueker et al. (2000) was conducted in a mountainous environment and none was conducted in steep headwater catchments (Table 1.1). Studies conducted in British Columbia have examined runoff processes and pathways in alpine catchments (Zeman and Slaymaker, 1975; Laudon and Slaymaker, 1997), and the processes of meltwater generation and the routing of water through the snowpack (Braun 1980, Jordan, 1983). However, little attention has been given to the study of runoff processes and pathways in steep, temperate, forested headwater catchments of the Pacific Northwest during snowmelt. In particular, there appear to have been few significant contributions to the study of catchment runoff processes during snowmelt from temperate forested areas of either coastal or interior British Columbia. 7 e o 03 O o 1-3 T3 O I S Oh -^ ' $ II 00 rj- lO CM ~-(N (M '—1 •s .8 3 C/3 C/} O © O o s© in ^ O 2 OO «N So ^ m m 0 0 as ON ON O ON > ^ O S os w o\ <tf —: Gfl CO a u 03 S PQ ^ ON ON ON (S «1 -43 00 .g u 2 S a 03 O O . 43 a r j - 1 >i 03 « S q « s ^ s N 03 CO in o ^ I o oo o OO o ON ON 00 ON 9 5 ^ o j> NO O o ON s » o (N 03 c 03 C 03 6 N \ 1.3 OBJECTIVES, QUESTIONS AND R E S E A R C H HYPOTHESES The objective of this study is to compare the spatial and temporal distribution of runoff processes in response to spatially and temporally variable snowmelt inputs among three face unit catchments in the southern Interior of British Columbia. The hydrogeomorphic concept will be adopted as a framework within which the following specific question can be addressed: 1. What are the dominant runoff pathways and how do they interact and evolve over the snowmelt period in response to spatial and temporal variation in water inputs? It was anticipated that a comparison of the three catchments would reveal differences in snowmelt runoff processes due to differences in their associated hydrogeomorphic characteristics. The following hypotheses have been proposed as alternatives to the null hypothesis that no difference in catchment runoff processes will be observed (these hypotheses draw upon catchment descriptions presented in Chapter 2: Methods, Section 2.2: Field Site): 1. Differences in catchment streamflow response are related to the degree to which connectivity between stream channels (i.e. areas capable of rapidly transmitting water from source downslope) and meltwater source areas is maintained as the snowline progresses upslope. 2. Thickett Creek, having a short, poorly incised permanent channel will experience an early disconnect between source areas and stream channels as the snowline progresses upslope. Runoff pathways will be dominated by subsurface flow through deeper soil layers and subject to long contact times. Streamflow response in Thickett Creek will be sluggish relative to the other streams 3. Gurn Brook, having a deeply incised stream channel extending into the upper catchment, will experience the greatest degree of connectivity between meltwater source areas and the stream channel as the snowline progresses upslope. Surface flow (e.g. channel flow or saturation overland flow) and shallow subsurface flow characterized by short contact times will constitute important runoff pathways in the Gum Brook catchment. Streamflow response in Gurn Brook will be flashy relative to the other streams. 4. Gurn Spring, having a moderately incised channel of moderate length and small catchment area, will exhibit an intermediate degree of connectivity between source areas and the stream channel relative to the Gurn Brook and Thickett Creek catchments. 9 Runoff pathways will be comprised of a mixture of deep subsurface flow and surface/shallow subsurface flow. Streamflow response will be neither so flashy as Gurn Brook nor so sluggish as Thickett Creek. 1.4 STRUCTURE OF T H E THESIS The remainder of this thesis is composed of four chapters. Chapter 2 provides an introduction to the study area and an overview methodology and data analysis. Chapter 3 presents results of the data analysis and Chapter 4 provides interpretation and discussion of results in the context of the research objectives. Chapter 5 contains a summary of key findings and makes recommendations for future research. 10 C H A P T E R 2 M E T H O D S 2.1 STUDY A R E A 2.1.1 Location The study area was located at Ringrose Slope, a steep southwest-facing side wall of the Slocan River watershed, in the West Kootenay region of British Columbia (Figure 2.2, inset). The Slocan River watershed covers an area of approximately 340,000 ha, and is located in the Selkirk Mountains, in what is referred to as the "wet interior" forest zone. Steep terrain dominates the Slocan Valley landscape. Narrow flat valley bottoms, changing abruptly to steep, broken side walls, characterize the main valley and its tributaries, with slopes often exceeding 30 degrees or 60% (Silva Forest Foundation, 1996). Ringrose Slope marks the western toe of the Kokanee Range of the Selkirk Mountains further to the east. Directly across the main valley from Ringrose Slope lies the Valhalla Range of the Selkirks. The terrain of Ringrose Slope is that of a steep (slope gradients dominantly in excess of 20°) valley side-wall accentuated by rock outcrops and small cliff bands, and dissected by small headwater streams. 2.1.2 Climate The climate of the Slocan Valley can be generally described as continental, with cold winters and hot summers. During winter and early spring, cold air travels down slopes and stream gullies effectively extending colder conditions of the upper elevations into lower elevations and along stream channels (Silva Forest Foundation, 1996). Monthly mean daily temperatures, based on 1961-1990 climate normals measured at the Atmospheric Environmental Service (AES) climate station in New Denver (570 m elevation), range from -3.3 °C in January to 18.7 °C in August. Total annual precipitation at New Denver is 866 mm, of which the majority (64%) is produced between October and April. Snowfall comprises approximately 24% of the total annual precipitation (Figure 2.1). 11 w Figure 2.1. Climate normals (1961-1990) for monthly mean daily temperature, and monthly Drecioitation at N e w Denver CSlocan Val lev . British Columbia). 2.1.3 Geology The northern portion o f the Slocan valley is dominated by intrusive rocks o f the Kuskanax Batholith, represented predominantly by leuco-quartz monzonite, minor leucosyenite, and leucogranite. The southern half is underlain by the Nelson Batholith to the east and the metamorphosed rocks o f the Valhalla Complex to the west. Rocks o f the Nelson Batholith, which underlie the Ringrose Slope field site, are represented predominantly by porphyritic granites while biotite-hornblende quartz monzonite, and granodiorite monzonite are also found. The Geological Survey o f Canada's Slocan map sheet (1932) identifies an eastward and upslope transition from phases o f gneiss, crushed granite, and partly altered pre-batholithic rocks in the valley bottom, to mainly porphyritic crushed granite on mid-slopes immediately east o f Slocan Lake and the Slocan River. The Valhalla Complex consists o f low-dipping, gneiss rocks ranging in composition from granodiorite to quartzite, marble, biotite-quartz-plagioclase paragneiss and pelitic schist. A northwest trending belt (15-40 km wide) o f metamorphosed volcanic and sedimentary rocks belonging to the Kaslo and Slocan Groups separates Nelson and Valhal la rocks 12 from Kuskanax rocks. Rocks of the Nelson Batholith display greater textural and compositional heterogeneity relative to the Kuskanax Batholith and associated stocks (Ministry of Energy, Mines and Petroleum, 1981). Surficial geology of the Slocan Valley consists of glaciofluvial deposits and alluvium in valley bottoms of the Slocan River and its major tributaries. Deep (i.e., generally >1.5 m) phases of colluvial deposits commonly accumulate at the base of slopes along the upper reaches of tributary valleys, with shallow (i.e., <1.5 m) colluvial deposits being common throughout the Slocan River watershed on steep slopes and overlying bedrock. Discontinuous basal and ablation tills occur throughout the watershed (Jungen, 1980). 2.1.4 Vegetation The Slocan Valley lies in the Interior Cedar-Hemlock biogeoclimatic zone. Forests are composed of mixed tree species at a wide range of elevations and aspects. Englemann spruce (Picea englemannii) and subalpine fir {Abies lasiocarpa) represent the dominant species at higher elevations. At middle and lower elevations on drier sites, ponderosa pine (Pinus ponderosa) and Douglas-fir (Pseudotsuga menziesii) are the dominant tree species. Western red cedar (Thuja plicatd), western hemlock (Tsuga heterophylla), western larch (Larix occidentalis), and western white pine (Pinus monticola) stands develop wherever sufficient moisture and nutrients accumulate. At middle and lower elevations and on moist aspects, mixtures of these species together with Douglas-fir are often found. Early successional tree cover includes lodgepole pine (Pinus contorta var. latifolid), paper birch (Betula papyrifera), and trembling aspen (Populus tremuloides)(Si\va Forest Foundation, 1996). 2.2 F I E L D SITE 2.2.1 Location and physiography The research site comprised three small stream catchments on Ringrose Slope: Thickett Creek, Gum Brook, and Gum Spring (Figure 2.2). Al l three catchments share the same westerly aspect. However, catchments display varying morphologies. The presence of fracture systems in bedrock underlying the study catchments was suggested by field observations of bedrock outcrops and by the nature of the local bedrock composition and structure as identified in geologic maps (Geologic Survey of Canada, 1932). Catchment areas were determined in ArcGIS from a digital elevation model (DEM) derived from a 1:5000 map with 5 m contour intervals supplied by Slocan Forest Products. 13 \ Uncertainty in measurement of catchment areas arose from two sources. First, subdued topography made it difficult to determine drainage divides. This was not an issue for the deeply incised and topographically distinct Gurn Brook catchment. Though not so defined as Gurn Brook, the Gurn Spring catchment is relatively well defined by topography and uncertainties in catchment area are likely small. The Thickett Creek catchment is characterized by poorly defined topography and likely possesses the largest uncertainty in drainage area from this source. This is confirmed by previous maps of the Thickett Creek catchment that showed an area greater than that of Gurn Brook. The second source of uncertainty relates to the potential for crOss-catchment transfers of water along forest roads. This would yield little uncertainty with regards to estimates of Gurn Spring's catchment area due to the near absence of roads in that catchment. Conversely, field observations revealed the potential for overland flow across drainage divides along road segments in the middle and upper elevations of the Gurn Brook catchment, and lower elevations of the Thickett Creek catchment, suggesting that uncertainties in catchment areas from this source were largest in these catchments. In order to address uncertainties in catchment areas, boundaries were field-checked during both low and high flow conditions. In both cases catchment boundaries generated using the D E M were judged to provide a good representation of actual drainage divides. Thickett Creek is composed of intermittent seeps, is weakly channelized, and surfaces approximately 10 m upslope of the weir. The Thickett Creek catchment has an estimated area of 13.7 ha at the weir, and ranges from approximately 725 m to 1445 m elevation. The Thickett Creek drainage could be characterized as a broad network of seeps and riparian zones in its lower two thirds with a small upslope contributing area. The Gurn Brook catchment drains a broad concave slope in its upper extent, emerging into a pair of ephemeral channels that converge and flow through a deeply incised gully to the catchment outlet. The catchment is largely devoid of any riparian zone, particularly in the lower reaches. Gurn Brook is the largest of the three watersheds (57.9 ha), and ranges in elevation from approximately 650 m to 1575 m. The Gurn Spring catchment, at 12.5 ha, is the smallest of the three. It displays a short, moderately incised channel within a broad shallow gully. A narrow riparian zone occupies the base of the shallow gully and extends the length of the stream channel. Topography of Gurn Spring's drainage area could be described as intermediate between that of the other two catchments: its channel is more incised than Thickett Creek's but not so much as Gurn Brook's, it has distinct riparian zones (though they are more limited than the broad network of seeps and 14 \ riparian areas found in Thickett Creek), and it possesses moderately steep near-stream hillslopes. The Gurn Spring catchment ranges from just under 650 m to over 1075 m in elevation. The study catchments are distinguished from other snowmelt-dominated, forested, headwater catchments in that they are small, relatively steep (slope gradients dominantly in excess of 20° or 35%), and in the case of Thickett Creek and to a lesser extent Gum Spring, are characterized by minimal channel incision and poorly distinguished topographic boundaries. In comparison, other headwater catchments that have been the subject of snowmelt-runoff studies are characterized by lesser slope gradients and better developed channel incision and topographic boundaries. These include the 40.5 ha W9 sub-basin of the Sleepers River watershed with an average slope gradient of 13% (Kendall et al., 1999; McGlynn et al., 1999; Shanley et al., 1999; Shanley et al., 2002), and sub-catchment S of the Moshiri experimental basin (128 ha) with slope gradients in the range of 20-25% (Kobayashi, 1985; Kobayashi, 1986; Kobayashi et al., 1999). Other snowmelt-runoff studies have focussed on larger meso-scale catchments in mountainous areas (e.g. Sueker et al., 2000), on alpine catchments (e.g. Zeman and Slaymaker, 1975; Laudon and Slaymaker, 1997), on sub-arctic catchments (e.g. Spence and Woo, 2002; Spence and Woo, 2003), and on headwater catchments of relatively low relief (e.g. Moore, 1989; Maule and Stein, 1990; Buttle and Sami, 1992; Hazlett et al., 2001). 15 2.2.2 Soils Soils within the field site are relatively shallow (often < 1.5 m), and soil horizons tend to be poorly defined. In general, a surficial litter layer is underlain by a thin organic horizon that rests above a heavily leached mineral horizon (the organic layer becoming deeper in near-stream riparian areas). Soils display a sandy-loam texture with significant amounts of cobbles and gravels within the soil matrix. Soil characteristics observed in the field closely resemble the Bonnington Soil Association (BG) described by Jungen (1980). Bonnington soils occur on the lower slopes of major valleys, and may extend up to 1220 m on south and west aspects. B G soils are developed in loose, shallow colluvium and/or glacial till over bedrock. B G soils are dominated by the Orthic Dystric Brunisol soil sub-group, and are characterized by rapid drainage. A transition between the lower and middle valley slopes with respect to soil associations is common in the region. The soil association found on Ringrose Slope above 1200 m is likely the Buhl Creek Association (BH), one of the most common of the Nelson map area. B H soils extend to elevations of approximately 1770 m, are typically shallow (0.5-1.5 m), and are developed in shallow, loose colluvium overlying crystalline bedrock (mainly granite and granodiorite). These soils are generally well to rapidly drained (Jungen, 1980). 2.3 C L I M A T E , SNOWPACK, AND S T R E A M F L O W D A T A Most of the data for this research were collected during the 2002 field season, which began February 11, 2002 (well before the onset of snowmelt), and proceeded until April 25, 2002 (approximately two weeks after peak flows). Additional visits were made on May 15 and 16, and June 18-20. 2.3.1 Meteorology Meteorological parameters were measured at two climate stations located within the field site (Figure 2.4). The lower climate station is located at approximately 975 m while the upper station is located at approximately 1175 m. Both stations were installed in November of 1993. Air temperature ^relative humidity (Campbell Scientific temperature and RH probe model HMP 35C), and rainfall (TE525M Texus tipping bucket raingauge) were measured hourly at both stations and recorded on Campbell Scientific CR-500 dataloggers. Daily maximum, minimum, and mean air temperatures, mean relative humidity, and total precipitation were also recorded. In August of 2001, an anemometer (Met One wind speed sensor model 014A), a 17 pyranometer (Li-cor model LI200S), and an automatic snow depth sensor (Sonic Ranging Sensor SR50) were added to the upper climate station to provide hourly measurements of wind speed, incoming solar radiation, and snow depth, respectively. Daily maximum wind speed and daily accumulated snow depth were also recorded. 2.3.2 Snow surveys Snow depth and snow water equivalent (SWE) during the spring of 2002 were measured weekly at two snow courses maintained by the Ministry of Forests (MoF) (see Figure 2.4). The upper snow course is composed of 32 snow survey points ranging in elevation from 1150 m to 1175 m. The lower snow course is composed of 30 points at an elevation of approximately 975 m. Both snow courses have been monitored by MoF since 1988, with no snow surveys conducted in 1995, 1996, or 1997. An additional snow course consisting of 6 snow survey points was established at an elevation of approximately 780 m during the February of 2002 to facilitate interpolation of snow depth and SWE over a broader range of elevations (see Figure 2.4). Snow depth and SWE were measured at all snow courses using a Mount Rose snow corer. 2.3.3 Snow line mapping Following the onset of snowmelt, the snow line at the field site was mapped on a weekly basis. For mapping purposes, the "snow line" was defined as a line above which the majority of the ground surface (>50%) was covered with snow. Mapping was conducted in the field by locating the snow line at the northern extent of the field site and following it southwards on foot. Using a contour map, altimeter, and pace and compass, points along the snow line were drawn using topographic features as reference points. A total of five snow line maps were constructed showing the upslope recession of the melting front over a five-week period. 2.3.4 Streamflow Streamflow from Thickett Creek, Gurn Brook, and Gurn Spring has been monitored since 1990 by the B C Ministry of Forests using V-notch weirs installed on each of the streams (see Figure 4). Hourly stage measurements were recorded using pressure transducers (CS400/CS405 Kellar series 169/173 as of Jan 2002, formerly Sensotec pressure transducers) attached to Campbell Scientific CR-10 dataloggers with SM4 storage modules. Weirs were equipped with stage rulers for manual stage measurements and to allow for calibration of transducer data. Stream temperature was measured hourly and recorded along with daily maximum, minimum and mean stream temperature using a 107B temperature probe connected to the dataloggers. Hourly 18 air temperature and relative humidity was also measured at the weirs and recoded along with daily maximum, minimum, and mean air temperatures and relative humidity using a Campbell Scientific temperature and RH probe model 207. At least twice each week, manual stage measurements were taken for each stream in both the morning and the afternoon. Electrical conductivity and stream temperature were also measured at these times using a portable W T W LF-340 conductivity probe. Once during baseflow conditions and weekly for five weeks during the spring freshet, manual stage, conductivity, and temperature measurements were made every 4 hours over a 24-hour period to investigate diurnal streamflow fluctuations. 2.4 PlEZOMETRIC DATA A total of 30 piezometers were installed at various locations throughout the research site (Figure 4). Of these, 7 were short (<75 cm long), made of 1.5 inch P V C perforated in the lower 6 inches and wrapped in nylon mesh to prevent blockage at the intake. The others were longer (160 cm, 3 inches in diameter), prefabricated stainless steel piezometers with drive-points. All piezometers were installed by driving an iron bar into the ground to create a hole, and then driving the piezometer into the ground until either bedrock was struck or the entire piezometer (minus a few cm) was embedded. Bentonite clay was applied around the protruding end of each piezometer to prevent infiltration of surface water. Piezometers were capped with plastic bags to prevent the capture of rain or snow. Four of the piezometers (P9D, P9S, P10, and PI 1) were located on Ringrose Slope just north of the T H catchment. At the time of installation, these piezometers were believed to be within the catchment. Reassessment of catchment boundaries following the field season revealed these to be outside the T H catchment. Transects of four piezometers were installed perpendicular to Thickett Creek (1 transect) and Gurn Spring (2 transects) near the weirs. Due to steep gully walls and limited access immediately above the weir on Gurn Brook, transects could not be established in this area and 2 piezometers were located next to the stream in the gully bottom. An attempt was made to locate the remaining piezometers at representative hydrogeomorphic positions throughout the catchments (e.g., near-stream riparian areas, near-stream hillslopes, upslope contributing areas). Wherever piezometers were installed in riparian areas, short piezometers were paired with the longer stainless steel piezometers in order to estimate a vertical hydraulic gradient, and to sample shallow and deep soil water for chemical analysis. At all other locations, the longer stainless steel piezometers were used in order to penetrate as deeply as possible. Water levels in all 19 piezometers were measured manually at least once per week. Crest stage gauges constructed from rigid transparent plastic tubes containing Styrofoam "floats" were inserted into piezometers, thereby providing a means of measuring maximum water levels between visits. Two piezometers were equipped with pressure transducers attached to Campbell Scientific CR-10X dataloggers to provide water level measurements at 10 minute intervals. 2.5 F I E L D OBSERVATIONS OF SURFACE F L O W Throughout the field season the locations and timing of surface flow were recorded whenever it was observed within the study site. Surface flow was intensively monitored immediately following peak flows and field sketch maps of surface flow locations were created over the period April 15-18. Points were plotted on field maps and flagged in the field with high visibility tape wherever surface flow either: 1. Commenced on the road due to interception from adjacent cut-banks, 2. Left the road either due to infiltration into the road-bed or as a result of diversion off of the road, or 3. Was observed in other areas not on the road (e.g. seepage observed at the base of forested hillslopes or SOF observed in riparian zones). In order to investigate the extent to which the road influenced drainage patterns, an attempt was made to estimate the proportions of surface flow continuing along the road and diverted off of the road for each point plotted along the road on field maps. Observation points were re-visited on Apr 22-24, 2002 and May 15-16, 2002. On these dates each point was listed as being either active (i.e. surface flow persisted) or inactive (i.e. no surface flow observed) and proportions of flow diverted from active locations were again estimated. Surface flow observed elsewhere was mapped in order to record the extent to which SOF occurred within the study catchments. These points were also re-visited on the dates stated above, and were listed as either active or inactive. 20 \ 2.6 W A T E R CHEMISTRY 2.6.1 Streamwater Streamflow samples were collected at least twice weekly for analysis of oxygen-18, major cations (Ca, Mg, K, Na), and silica (Si). Twice-weekly samples for analysis of dissolved organic carbon (DOC) were collected during the three weeks straddling peak flows. Once during baseflow conditions and weekly for five weeks during the spring freshet, water samples were collected for analysis of oxygen-18 and water chemistry every 4 hours over a 24-hour period to investigate diumal fluctuations in stream chemistry. Al l streamwater samples were collected manually at the weirs using polyethylene sampling bottles. Al l bottles received a minimum of two rinses prior to sample collection. Streamwater electrical conductivity (EC) and temperature were measured in the field at each sampling time. 2.6.2 Melt and rainwater Meltwater samples were collected from 8 sampling trays inserted into the base of the snowpack at various locations within the field site. Sampling trays were constructed using flat-bottomed plastic bins, approximately 0.5 m wide, 1 m long, and 0.25 m deep. The back end of each bin was cut away leaving three sides and a bottom, thereby allowing the sampling tray to be manually inserted at the base of the vertical profile of the snowpack. Sampling bottles were then placed under a drain hole located at the front of the sampling tray to collect composite samples of melt and rain water. The date and time each bottle was placed was recorded, and bottles were then observed daily and collected when at least 100-200 mL of water had accumulated (again recording date and time of sample collection). Samples were then analyzed for oxygen-18. To avoid changes in isotopic signatures due to evaporation, a thin film of oil was added to each bottle at the time of its placement, thereby sealing accumulated melt/rainwater from interaction with the atmosphere. Upon retreat of the snow-line above the sampling trays, they continued to be used as rain water collectors following the same procedure. 2.6.3 Piezometers Up to three subsurface water samples were collected from each piezometer (Feb. 11-13, May 15, and June 18). Samples were collected as follows: 1. The piezometer was purged and allowed to refill. 21 2. Using a handheld Nalgene vacuum pump connected to a length of Nalgene tubing and polyethelene collection bottle, water was extracted and used to rinse the sampling apparatus (this rinsing process was repeated twice prior to sample collection). 3. The sample was extracted into the collection bottle and transferred to 125 mL polyethelene storage bottles (each receiving a minimum of two rinses prior to filling with sample). These were then labelled and stored according to laboratory-recommended protocols. Subsurface water samples were analyzed for oxygen-18, the major Cations, and silica. Temperature and E C were measured in the field at each sampling time. 2.6.4 Overland flow Overland flow samples were collected on an opportunistic basis, primarily from surface flow along the deactivated access road. Samples were collected manually using polyethylene sampling bottles. Each bottle was rinsed at least twice prior to sample collection. Al l overland flow samples were submitted for analysis of oxygen-18, the major cations, silica, and DOC. Temperature and E C were measured in the field. 2.7 L A B O R A T O R Y ANALYSIS 2.7.1 Oxygen-18 Oxygen-18 ( 1 80) was analyzed by means of mass spectrometry at the Stable Isotopes Laboratory of the National Water Research Institute in Saskatoon, SK. The ratio of the heavier l 8 0 to the 1 6 0 isotope, referenced against its deviation from the ratio in 'standard mean ocean water' (SMOW), is used to express the abundance of stable isotopes in a water sample, and is denoted as follows: 8 I 8 0 = [(Rsample ~ RsMO\v)/RsMOw] - 1000 (1) where 8 I 8 0 = the relative difference in the ratios in units of parts per thousand (per mil), Rsampie and RSMOW = the ratios of 1 8 0 to 1 6 0 in the water sample or in SMOW, respectively (Buttle, 1994). The analytical precision of the method was ±0.2 per mil. A number of blank samples were re-submitted to the laboratory for Q A / Q C purposes. Al l 1 8 0 samples were stored in cool and dark conditions prior to analysis. 22 2.7.2 Hydro-chemistry Concentrations of the major cations (Ca, Mg, Na, and K) and silica (Si) were measured by inductively coupled plasma mass spectrometry (ICP-MS) at the University of British Columbia's Soil Resources Laboratories. Concentrations were reported in milligrams per litre (mg/L). Al l water samples were pre-filtered in the field using glass-fibre filter papers (<0.2 micrometer pore size) to remove any suspended solids. 2.7.3 Dissolved organic carbon Concentrations of non-purgeable dissolved organic carbon (DOC) were measured at the Environmental Sciences Laboratories of Okanagan University College, Kelowna, British Columbia. All DOC samples were pre-filtered in the field using Watman GF-C glass fibre filter papers. Al l filter papers were muffled before use in the field (i.e. heated at 450 °C for approximately 2 hours to remove trace amounts of carbon). 2.8 D A T A ANALYSIS 2.8. 1 Analysis ofpeak flow generating mechanisms Peak flow generating mechanisms (i.e. snowmelt, rain-on-snow, or rainfall) were determined for each stream catchment by considering the SWE at the time of peak flows and inspecting rainfall and air temperature data collected during the week immediately prior to peak flows. 2.8.2 Inter-catchment comparison of diurnal streamflow response Data from published research papers was used to compare diurnal streamflow response observed in the three study catchments with that observed in other headwater catchments. The aim was to find data for as many different snowmelt-dominated headwater catchments as possible from published material. The following criteria were used to select studies for this meta-analysis: 1. The study catchment is small (i.e., not exceeding 1.5 km 2 in area), predominantly forested, and is snowmelt-dominated or received significant water inputs as a result of snowmelt. 2. Information regarding snowpack water equivalent (SWE) prior to melt is provided. 3. Information regarding catchment area and elevation range is included. 23 4. Rainfall data are provided for the period of interest (i.e., melt season). 5. A streamflow hydrograph presenting diurnal fluctuations is presented for the period of interest (i.e., melt season). A non-dimensional index of catchment steepness (S*) known as the Melton ruggedhess index (Melton, 1965) was computed as: S* = ( H m a x - H m i n ) / A 0 5 (2) where and H™,, are the maximum and minimum elevations of the catchment, respectively, and A is the catchment area. A measure of each catchment's diurnal streamflow response that would allow for direct comparison between catchments was computed as: DR - (Qmx - Q m , „ ) / Q „ u n (3) where DR is a unitless measure of catchment diurnal streamflow response, and Qmax and Qmin are the maximum and minimum discharges recorded on a given day, respectively. DR was calculated for each catchment for the day showing the largest diurnal variation over a period of no more than 10 days before or after the date of peak flow. Days on which rainfall was recorded were excluded. For consistency, diurnal variation was measured from minimum to maximum daily discharge on either the rising or falling limb of the peak flow hydrograph (Figure 2.3). Where more than one year of data was available for a given catchment, the mean DR was used to represent the catchment's diurnal response index. In the case of data from published research papers, daily minimum and maximum discharge was measured directly from printed hydrographs. Similarly, where no value was expressly stated for maximum pre-melt SWE, it was estimated from printed time series graphs of SWE. For the three study catchments (Thickett Creek, Gurn Brook, Gurn Spring) and Upper Redfish Creek, daily minimum and maximum discharge and antecedent SWE was taken from hourly discharge data and digital snow survey data. The accuracy of the measurement technique was estimated by printing hydrographs from the digital datasets and comparing the raw data with estimates derived from printed hydrographs (as per the method described above). Estimated daily maximum and minimum discharge measured from printed hydrographs differed from the raw data by less than 1% on average. 24 300 50 j 0 +- 1 — 1 . 1 1 15/05/1996 22/05/1996 29/05/1996 05/06/1996 12/06/1996 19/06/1996 Figure 2.3. Example of a peak flow hydrograph illustrating the method used for measuring daily streamflow variation. Daily variation was measured from daily minimum to maximum discharge. A thorough search of published literature yielded 11 catchments (including the study catchments) that met the selection criteria. Two years of data were used for two of these catchments (Moshiri Experimental Basin and Redfish Creek). For the three study catchments (Thickett Creek, Gurn Brook, and Gurn Spring), four years of data were used. A single year of data was available for each of the remaining catchments (i.e., Sleepers River Watershed "W-9", Hermine Basin, and Woods Lake Watersheds "W02" and "W04"). 2.8.3 Statistical analysis of 180 inputs Spearman correlation analysis was performed to test whether systematic temporal and/or spatial patterns among the concentrations of 1 8 0 in meltwater inputs existed. To test whether a temporal pattern existed, meltwater ! 8 0 data were stratified according to the elevation of meltwater samplers (i.e. isolating all samples collected from one elevation/sampler). To test whether a spatial (i.e. elevational) pattern existed, meltwater 1 8 0 data were stratified by date (i.e. isolating all samples collected on a given date or sequence of dates. The use of Spearman's rank-25 \ • based correlation test requires no assumptions about the probability distributions underlying the data, and can also accommodate nonlinear relations. Observed patterns and/or lack thereof were displayed using scatterplots generated from the selected data and fitted with LOWESS trendlines. 2.8.4 Principal components analysis of streamwater chemistry Principal components analysis (PCA) was performed on streamwater chemistry data. Factors with eigenvalues greater than 1.0 were identified as being "significant" (Hair et al., 1995). PCA's were performed separately for each stream using the following variables: Si, DOC, the ratio (Na+K)/(Na+K+Ca+Mg) (R^-id and E C . Silica has been employed as an indicator of runoff contributions from deep subsurface waters characterized by a relatively long contact time (Maule and Stein, 1990; Buttle and Sami, 1997; Ladouche et al., 2001), and the E C of subsurface waters has been shown to increase with depth (Kobayashi, et al. 1999). Dissolved organic carbon has been used as an indicator of surface flow or flow through shallow saturated soil horizons (McGlynn et al., 1999; Ladouche et al., 2001). RNa-K is commonly employed in Piper's Diagrams to describe subsurface water chemistry (Freeze and Cheery, 1979), and was used as a means of collapsing cation concentrations into a single parameter. Spearman correlation analysis showed that RNa-K and DOC concentrations in streamwater samples collected during the field season were strongly related (p < 0.005, n = 89), suggesting that in this study R N a . K could be used as an indicator of surface/shallow subsurface flow. A separate PCA was conducted for Ringrose Slope using all streamwater, piezometer, and overland flow samples collected during the field season. The PCA was run using the same four parameters described above. Results of this PCA are intended to describe factors responsible for the observed variance in the chemistry of all waters sampled from Ringrose Slope, and particularly to explore chemical contrasts among the streams. 26 C H A P T E R 3 R E S U L T S This chapter draws upon streamflow, snowpack, and meteorological data collected over several years as well as detailed measurements for one field season. The detailed field study was conducted during the winter and spring o f 2002, beginning on February 8, 2002, wel l before the onset o f snowmelt, and continuing until A p r i l 25, 2002. Two further visits were made to the field site on M a y 15-16 and June 18-20 respectively. Sections 3.1 and 3.2 provide a context for the field season by presenting results o f snowpack, climatic and streamflow measurements collected over a number o f years and for the detailed field season. Sections 3.3 and 3.4 present results o f piezometer measurements and observations o f overland flow collected during the field season. Section 3.5 addresses the temporal and spatial variation in isotopic concentrations o f melt and rainwater. Section 3.6 contains results of analyses o f water chemistry, and 3.7 presents results o f the isotopic hydrograph separation. 3.1 O V E R V I E W OF T H E STUDY PERIOD The timing and magnitude o f peak S W E varied substantially amongst years (Figure 3.1). The peak snow accumulation at 975 m was high in 2002 compared to previous years, and the melt season began relatively late. A t 1175 m, the peak snow accumulation and timing o f melt appeared to be intermediate compared to previous years. The main snowmelt period began about March 19, when air temperatures jumped above 0°C, and streamflow peaked on A p r i l 14 for al l three streams (Figure 3.2). A i r temperatures measured during the 2002 field season were low relative to previous years, while maximum daily and total monthly rainfall were average relative to previous years (Table 3.1). 27 E in GO m 3. Year X • • 1994 1998 1999 2000 2001 2002 Figure 3.1. Snow water equivalents measured in forested areas at 1175 m and 975 m on Ringrose Slope. SWE measured during the field season is shown in black. 28 > •a c o co 03 D D 43 CO c £ <D 43 -*-» 00 c •c •§ &, _o on CO o I-I 00 c 2 . c o IT) O r- « ON P T3 OS ^ & SS S3 .s a eg ^  o O co O o3 O £ j co ^ . — i 3 * CU .5 43 1—1 T3 c 03 a) 3 03 _ O 53 _S> O H 33 I s .S «8 03 Q 03 j-i Q ° ro, cu 3 - « c o CO >-> 03 H> >> 03 CN NO © < 30 co 00 d d ON CO rj-i n ON ON 00 2002 Apr co co <N ON © i. r-<fr CO NO © NO NO •<3-NO . 00 o3 co i f-co i ON NO i n •n ON 1 od © ON 00 ON i n NO 03 NO ON oo © CN NO CN . c> ON . 00 ON iri 00 ON CO 2001 Apr 00 CN NO ON 1 © CN d CN CN ON iri CN CN © I-I 03 r -© o ON CN i vo NO oq • ON ON r--NO oo © ON 00 NO 03 CN CO NO CN ON . 00 CN © © CO co od 2000 Apr co 00 CN i oo iri © © ON .00 ON NO CN r -co -* d o NO . CN t i n co 00 d • m NO oo >, 03 co 00 CN d ON 00 CN ON CN NO i n © m CN iri CN m 1999 Apr r-; co . © © •<r iri CN © CN CN ON oo od i n co 03 © co iri ON co • © CN NO NO CN i r-; oo cn vo CN CN oo ' o? m NO 00 d NO © iri 00 ON CO CN 00 CN »—i 1996 Apr ON 00 ON d iri 00 ON *CN m "in i n t--i n NO CN © 03 co d • i n . NO co © r - NO 1 >n ON *r-CO i n od 03 m © NO i n CN © CN i n i n CO CN i n m 1995 Apr 00 © CN i n CN ON NO ON 00 i n NO CN 03 co d NO ON 00 1 ON NO °. 00 1 in c<i CI 129.5 i n od CN oo Mean X 03 S3 Mean X d S3 X 03 Total X 03 s Total ' (0.) n i e l l i © druax J i V ^ l l B Q (rata) u i gt\\ (ram) ui c/,6 C5 CM Figure 3.2. Overview of the field season. From top to bottom, time series of snow depth at 1175 m, mean daily air temperatures, daily rainfall, and streamflow from each of the three stream catchments during the spring 2002 field season. 30 The snow line began to recede above the weirs on Gurn Brook and Gurn Spring during the week of March 26, and above the weir on Thickett Creek during the week of April 4 (Figure 3.3). Snow lines drawn on April 10 and 18 bracket the time of peak flows recorded on all streams. By April 18, almost 100% of the snow had melted off Gurn Spring. On April 24, roughly 17% and 55% of the Thickett Creek and Gurn Brook catchments remained snow covered. Upon return to the field site on May 15, all snow cover had melted from the slope. Peak flows occurred on April 14 for all three streams and were preceded by rise in air temperatures and moderate rainfall onto partial snow cover. Gurn Brook and Thickett Creek experienced the highest and lowest flows, respectively (Table 3.2). At the time of peak flows, snow depth at the upper climate station was 50.1 cm. 3.2 S T R E A M F L O W VARIABILITY This section presents results of streamflow data analysis. Section 3.2 compares peak flows and unit runoffs from a number of years, while section 3.2.2 examines diurnal streamflow variability at the three study streams. A comparison between the diurnal streamflow response at the study catchments and that observed in other published studies is presented in section 3.2.3. 3.2.1 Comparison of streamflow amongst the three streams Streamflow was highly variable amongst years for all three streams (Figure 3.4). Gurn Brook consistently displayed the flashiest response of the three streams, while Thickett Creek showed the most subdued streamflow. For all years, the GS catchment produced the largest unit runoff, while the T H and G B catchments produced similar unit runoffs (Figure 3.5). Unit runoffs for all streams in 2002 were relatively low compared with those of previous years. 31 Figure 3.3. Snow line position on Ringrose Slope during the spring 2002 field season. 32 1-Jan 1-Feb 1-Mar 1-Apr 1-May 1-Jun Date Figure 3.4. Freshet hydrographs for Thickett Creek (TH), Gurn Brook (GB) and Gurn Spring (GS) for several years on record. The detailed field season is shown in black for comparison. 33 • TH • GB A GS 1994 1996 1998 2000 2002 Figure 3.5. Inter-annual variation in unit runoff (mm). Table 3.2. Timing o f peak flows in relation to snow water equivalents for selected years on record. A long both snow courses, only measurements at forested survey points were included in the average. N o snow survey data was available for the years 1995 and 1996. Year Stream Date and Time of Peak Peak S W E at S W E at Peak Flow Catchment Peak F low Flow F low 975m at 1175m at Generated (L/s) (Ls- 'ha 1 ) Time o f Peak Time o f B y : Flows (mm) Peak Flows (mm) 1994 T H Mar . 18 (20:00) 4.7 0.34 40 90 R O S G B Mar . 4 (04:00) 14.2 0.25 80 150 R O S GS Mar . 18 (01:00) 7.9 0.63 40 90 R O S 1995 T H Mar . 15 (18:00) 9.2 0.67 - - -G B Mar . 15 (14:00) 53.5 0.92 - - -GS Mar . 16 (6:00) 20.9 1.67 - - -1996 T H Apr . 12 (16:00 4.3 0.31 - - -G B Apr . 10(2:00) 41.0 0.71 - - -GS Apr . 11 (6:00) 12.6 1.01 - - -1998 T H Mar . 25 (11:00) 7.1 0.52 41 125 R O S G B Mar . 25 (22:00) 26.2 0.45 41 125 R O S GS Mar . 26 (18:00) 25.0 2 27 117 R O S 1999 T H Mar . 26 (07:00) 14.9 1.09 115 185 R O S G B Mar . 25 (19:00) 29.6 0.51 119 188 R O S GS Mar . 26 (18:00) 26.9 2.15 115 185 R O S 2000 T H Apr . 5 (18:00) 4.0 0.29 62 156 R O S G B Apr . 28 (19:00) 34.2 0.59 0 0 Rainfall GS Apr . 5 (17:00) 11.6 0.93 62 156 R O S 2002 T H Apr . 14 (03:00) 7.8 0.57 19 92 R O S G B Apr . 14 (05:00) 39.2 0.68 19 92 R O S GS Apr . 14 (11:00) 22.4 1.8 19 92 R O S (rainfall)* * See discussion below. 35 3.2.2 Timing and magnitude ofpeak flows Peak flows on Thickett Creek, Gurn Brook, and Gurn Spring occurred within 8 hours o f each other on A p r i l 14, 2002, with Thickett Creek and Gurn Spring peaking first and last, respectively (Table 3.2). A t the time o f peak flows, the snow line elevation within the T H , G B , and G S catchments was approximately 1000 m. The proportion o f snow covered area remaining within the T H , G B , and G S catchments was approximately 52%, 80%, and 5% respectively, and S W E measured in forested areas at 1175 m (where snow cover persisted) was approximately 92 mm. Gurn Brook experienced the highest absolute peak flows, followed by Gurn Spring and Thickett Creek (Table 3.2). However, Gurn Spring had the highest unit peak flows. Peak flows in 2002 were relatively high (lower only than 1995 and 1996) and occurred late in the season relative to previous years. Hydrographs generated from hourly streamflow data for al l years on record are presented in Appendix A . Peak flows for each o f the catchments generally occurred within 1-2 days o f each other in most years, with the exception of 1994 and 2000, when peak flows at Gurn Brook occurred wel l before (1994) and after (2000) peak flows for the other catchments in response to large rain events. Rain-on-snow (ROS) was judged to be the dominant mechanism for generating peak flows on all streams. The single exception occurred in 2000, when peak discharge from Gurn Brook occurred in response to a rainfall event after the snow cover had receded above the upper snow course at 1175 m. Lack o f snow survey data prevented evaluation of peak flow generating mechanisms for the years 1995 and 1996. Weekly snowline maps drawn during the 2002 field season, however, call the preceding conclusion into question. Figure 3.3 clearly shows that nearly no snow cover remained in the Gurn Spring catchment at the time of the peak flows (Apr i l 14, 2002). This is l ikely the result of: (a) Gurn Spring's lower elevation relative to the other two study catchments, and (b) the pattern of melt followed by the snow line as it receded upslope. The conclusions drawn from the preceding analysis of peak flows must be qualified. In particular, peak flows on Gurn Spring in the year 2002 appear not to have been generated as a result of R O S as originally concluded, but rather in response to rainfall. Snowline maps were not available for any year except 2002. A s a result, conclusions regarding peak flow generating mechanisms for the remaining four years could not be checked. A further limitation involved with assigning peak flow generating mechanisms without the aid o f snowline maps is that a significant proportion of the area o f the Gurn Brook catchment lies above the elevation of the upper snow course (1175 m). A s a result, it may be that an S W E o f zero measured at 1175 m does not necessarily indicate that little or no snow remains within the Gurn Brook catchment. 36 The dominant peak flow generating mechanisms assigned to Gurn Spring for the years 1994, 1998, 1999, and 2000, and to G u m Brook for the year 2000 in Table 3.2 are thus qualified. 3.2.3 Diurnal streamflow variability Hourly streamflow from each o f the streams was plotted for three 6-day dry periods immediately preceding peak flows for the years 1999, 2000, and 2002, in order to examine diurnal streamflow variation resulting from snowmelt alone (Figure 3.6). A clear diurnal response was observed in G u m Brook for 1999 and 2000, and in G u m Spring for 1999. N o clear diurnal response was seen in G u m Spring for 2000 and 2002, and the diurnal fluctuation in Gurn Brook streamflow during 2002 was marginal at best. A distinct lack o f diurnal streamflow fluctuation was observed in Thickett Creek for each year shown. Snow cover remained on all three catchments during the periods shown (Table 3.2). Snowline maps drawn during the 2002 field season show the snow line rapidly receding upslope during the period shown, indicating ripe snowpack conditions at lower elevations (Figure 3.3). Consequently, a systematic diurnal pattern was expected. When compared to other catchments reported in the literature, Thickett Creek, Gurn Brook, and G u m Spring displayed the lowest diumal streamflow responses despite being the steepest of the catchments surveyed. Spearman correlation analysis revealed a significant inverse relation between S* and D R (r, = -0.785, p < 0.005, n=12). Table 3.3. Comparison o f diumal hydrograph responses during snowmelt (S* = index of catchment steepness, D R = index o f diurnal response). Catchment Name Area (ha) S* D R Reference Moshi r i Exp. Basin "S" 128.0 0.22 0.65 Kobayashi (1985,1986) Moshi r i Exp. Basin "S" 128.0 0.22 1.06 Kobayashi (1999) Redfish Upper Basin 115.0 0.76 0.26 Unpublished data, B C M o F Sleepers River Watershed "W-9" 40.5 0.24 0.12 Shanley et al. (2002) Perch Lake Watershed "3" 3.1 0.08 0.38 Buttle and Sami(1992) Hermine Basin, 5.3 0.13 3.00 Moore (1989) Woods Lake Watershed " W 0 2 " 41.3 0.16 1.74 Burns and McDonne l l (1998) Woods Lake Watershed " W 0 4 " 61.2 0.15 1.73 Gray Creek Catchment 149.5 0.20 1.00 K i m et al., (in press) Thickett Creek 13.7 1.95 0.08 This study Gurn Brook 57.9 1.22 0.14 Gurn Spring 12.5 1.20 0.02 " 37 TH G S Figure 3.6. Diurnal streamflow variability during three 6-day dry periods preceding peak flows for the years 1999, 2000, and 2002. 38 3.3 P I E Z O M E T E R D A T A This section presents results of piezometer data collected during the field season. Section 3.3.1 presents piezometer water level variations through time, while section 3.3.2 examines the distribution of maximum recorded water levels in piezometers during the field season. 3.3.1 Time series of piezometer water levels Time series plots of water level and crest data collected from all piezometers are presented in Appendix B. Three examples, one selected from each of these hydrogeomorphic positions (near-stream riparian areas (NSRIP), near-stream hillslopes (NSHIL), and upslope contributing areas (UP) are presented in Figure 3.7. Water levels at P2 varied little throughout the field season, while P19 showed some response over time. Of the three piezometers represented, only PI7 exhibited a water level variation corresponding to streamflow variation. Time series plots of water level and crest data for the remaining piezometers generally support these patterns (Appendix B). Figure 3.7 also shows continuously recorded water level data from two piezometers, P3d and A l . P3d is located at approximately 1135 m within the Gum Brook catchment, next to an ephemeral tributary of Gum Brook in an area of convergent hillslopes. A l is located at approximately 940 m elevation within the Thickett Creek catchment. Water levels in P3D peaked approximately 19.5 hours after the end of the rainfall, while streamflow in Gurn Brook (measured at the weir) peaked another 2.5 hours after this. Water levels in A l peaked approximately 30.5 hours after the end of the rainfall, while streamflow in Thickett Creek peaked another 49.5 hours after this. 39 0 -25 -50 -75 -100 1 U P19 (680 m) E > CD 0> 0 -25 -50 -75 1 U P17(650m) O-100 C L CD Q 0 -25 -50 -75 —-100 If ID C TO * 40 _ 30 1 r P2(1270 m) h— U A1 (940 m) UP NSH4 ^25 -50 75 NSFUP N S R H f A|.J. 11 i .lll.liL j 0 -100 -100 50 40 30 20 10 0 1-Jan 1-Feb 1-Mar 1-Apr 1-May 1-Jun 1-Jul Date o WL Crest 975 m 1175 m TH G B G S Figure 3.7. Water levels and crests measured in piezometers in relation to rainfall and streamflow. Data are shown for P2 (GB), PI9 (GS) and P17 (GS), automatically recorded water levels piezometers P3D (GB) and A l (TH). Hydrogeomorphic positions for each piezometer location are shown in the top-right of each plot. Elevations are shown in parentheses. Water levels are expressed as elevations relative to the soil surface. 40 \ 3.3.2 Distribution of maximum recorded water levels Piezometers located in near-stream riparian (NSPJP) areas (e.g. P3d, P15d) experienced relatively high water levels (Figure 3.8). With the exception of P12s, a shallow piezometer located within the T H stream channel during high flows, water levels in piezometers located in NSRIP areas immediately upslope of the weir on Thickett Creek were lower than those in NSPJP areas in the Gurn Spring and Gurn Brook catchments. Shallow piezometers located in NSRIP areas experienced higher water levels than deeper piezometers located in the same areas. Piezometers located in near-stream hillslope areas (NSHIL) (e.g. P7, P21) and those located in upslope contributing areas (UP) (e.g. P2, P5) exhibited relatively low water levels (Figure 3.9). E o Q . CD Q 0) - Q CO & -100 ro -150 NSHIL NSRIP UP Hydrogeomorphic Position Figure 3.9. Box plot showing maximum water levels in piezometers grouped by hydrogeomorphic position: near-stream riparian areas (NSRIP), near-stream hillslope areas (NSHIL), and upslope contributing areas (UP). Water levels are expressed relative to the ground surface (0). 42 3.4 OBSERVATIONS OF SURFACE F L O W This section presents observations of overland flow made on Ringrose Slope during the field season. Section 3.4.1 presents observed saturation overland flow and section 3.4.2 presents observed surface flow along roads. 3.4.1 Saturation overland flow (SOF) The timing and location of saturation overland flow (SOF) were recorded throughout the field season on an opportunistic basis. Approximate locations of SOF observed between Feb 11, 2002, and May 16, 2002, are presented in Figure 3.10 and cross-referenced with brief descriptions in Table 3.4. Detailed descriptions of observed SOF are presented in Appendix C. It should be noted that the maps consider observed SOF only, and may not include all locations experiencing active SOF for a given time period. Due to the relatively large area of the field site, its steep terrain, and the fact that access was by foot only, it was impossible to make regular visits to all areas of the field site. Despite these qualifications, observations indicated that locations of SOF did not exhibit a high degree of connectivity in Thickett Creek (Figure 3.10). In Gurn Spring and Gurn Brook, observed surface flow was mainly restricted to the stream channel itself, or to small zones immediately next to the stream channel. These observations are consistent with maximum water levels measured in piezometers (Figure 3.8). 43 \ Table 3.4. Observations of saturation overland flow (for cross-reference with Figure 3.10 and Appendix C). Map Label Date of First Date of Last Comments Observation Observation 1 Feb 25,2002 Apr 24,2002 Seepage generating downslope overland flow. 2 Mar 31, 2002 Apr 24, 2002 Incipient channel feature and diffuse surface saturation. 3 Apr 9, 2002 Apr 18, 2002 2 parallel incipient channel features. 4 Apr 10, 2002 Apr 18,2002 Incipient micro-channel at base of convergent slopes. 5 Apr 10,2002 Apr 18,2002 Seepage and surface saturation. 6 Apr 15,2002 Apr 15,2002 Incipient channel feature intersecting road. 7 Apr 15,2002 Apr 18,2002 Surface saturation. 8 Apr 15,2002 May 16,2002 Multiple seeps generating downslope overland flow. 9 Apr 16,2002 Apr 16, 2002 Incipient micro-channel at base of convergent slopes. 10 Apr 16,2002 Apr 24,2002 Seepage. 11 Apr 16,2002 Apr 16,2002 Upslope extension of #3. 12 Apr 16, 2002 Apr 24, 2002 Downslope convergence of channel features in #3. 13 Apr 17,2002 Apr 24,2002 Seepage generating downslope overland flow. 14 Apr 18,2002 Apr 24,2002 Incipient channel feature intersecting road. 15 Apr 18,2002 May 16,2002 Seepage generating downslope overland flow. 16 May 15,2002 May 15,2002 Incipient channel feature intersecting road. 44 LEGEND — > s ^ Fbads — F e r m a n e n t Stream Channels — - Ephemeral Stream Channels •—^—•'' Catchment Boundaries Saturation Overland Row Figure 3.10. Saturation overland flow observed between Feb 11, 2002 and May 16, 2002. Numbers are provided for cross-reference with Table 3.4 and Appendix C. No observations were made above 1275 m, and no SOF was observed above 1200 m (the upslope extent of the above map). S C A L E 100 0 100 200 300 400 ' Symbols representing overland flow are not drawn to scale. " Locations of observed overland flow were estimated in the field. "These locations are approximations and were not accurately surveyed or otherwise georeferenced. 45 \ 3.4.2 Overland flow along roads Figures 3.1.1 to 3.13 show overland flow observed along roads during the periods Apr 15-18, 2002, Apr 22-24, 2002, and May 15-16, 2002, respectively. Note that arrows representing overland flow indicate the direction of flow, but are not intended to indicate relative volumes of flow. Locations of observed overland flow along roads were estimated in the same manner as for saturation overland flow. Some locations were surveyed using differential GPS. Overland flow along roads was greatest during the period Apr 15-18, 2002 (Figure 3.11). Flows were only observed below 1000 m, the approximate elevation of the snow line (Figure 3.3), and were higher at all recorded locations than on subsequent visits. The connectivity of surface flows along roads was highest during this period, with some sections of the road carrying water through multiple switchbacks or long distances along straight sections. The estimated proportion of water diverted off the road surface at many points of diversion (PODs) was lower during this period. This resulted from higher rates of spillage out of PODs and back onto roads, as many PODs were incapable of handling the higher volumes of flow. Subsurface flow interception along cut-banks was also at its highest during this period, with water often observed dripping or flowing freely from the soil profde. During the following week, overland flow along roads persisted in many locations, but the magnitude and continuity of flows was diminished (Figure 3.12). Between Apr 22, 2002, and Apr 24, 2002, all overland flow along roads was observed below the snow line (1175 m), with the snowline often following the road within the Gum Brook catchment. One new occurrence of overland flow was observed just below 1175 m. This followed the road for a short distance before being diverted off the road by a wooden cross-ditch. Active PODs generally diverted the majority of overland flow off roads during this period, with proportions estimated to be > 90% for 11 of 12 active PODs. Relative to the week prior, overland flow along roads tended to be more intermittent during this period with flows appearing and then disappearing into the roadbed along some sections. Although the snow cover had left the catchments entirely by May 15-16, significant overland flow persisted along the readjust below 1175 m and above and below the lower climate station between 900 and 1000 m (Figure 3.13). Flows persisted in some locations at lower elevations within the Thickett Creek catchment, but these were limited to slow trickles. During this period 12 of the 13 active PODs were estimated to be diverting > 98% of flows off road surfaces. 46 Overland flow along road surfaces was observed crossing the T H and GB catchment boundaries during each of the three observation periods. There appears to be zero potential for cross-catchment flow of water along roads in GS, and minimal potential in GB. Because of the high road density at lower elevations near the T H catchment outlet, the potential for cross-catchment flow along roads in T H appears to be relatively high. This is particularly true during the period April 15-18, 2002, when flow volumes were highest and POD's were least effective at removing water from road surfaces. 47 APRIL 1 5 - 1 8 , 2 0 0 2 Fbads Catchment Boundaries Fermanent Stream Channels Ephemeral Stream Channels Start of overland flow Continuing overland flow Row diverted by wooden cross-ditch Interception of subsurface flow at cut-banks 100 S C A L E 100 200 300 400 ' A r r o w s r e p - r e s e n t i n g o v e r l a n d f l o w a l o n g r o a d s a r e n o t d r a w n to s c a l e . ' L o c a t i o n s o f o b s e r v e d o v e r l a n d f l o w a l o n g r o a d s w e r e e s t i m a t e d i n the f i e ld . T h e s e l o c a t i o n s a r e a p p r o x i m a t i o n s a n d w e r e not a c c u r a t e l y s u r v e y e d o r o t h e r w i s e g e o r e f e r e n c e d . Figure 3.11. Overland flow observed along road surfaces between April 15, 2002 and Apr 18, 2002. Discontinuous arrows (i.e. arrows that are not connected) indicated intermittent flow while connected arrows indicate connectivity of overland flow along the road. Arrows with black-ringed dots on the end represent the start of flow along roads, and were generally associated with interception of subsurface flow at cut-banks on the upslope side of roads. Arrows leaving the road surface and pointing downslope indicate points where flow was diverted off the road surface. No observations were made above 1275 m, and no overland flow was observed above 1200 m (the upslope extent of the above map). 48 APRIL 2 2 - 2 4 , 2 0 0 2 Fbads Catchment Boundaries Fermanent Sream Channels Ephemeral Stream Channels Start of overland flow Continuing overland flow How diverted by wooden cross-ditch Interception of subsurface flow at cut-banks 100 S C A L E 100 200 300 400 ' Arrows rep-resenting overland flow along roads are not drawn to scale. ' Locations ol observed overland flow along roads were estimated in the field. These locations are approximations and were not accurately surveyedor otherwise georeferenced. Figure 3.12. Overland flow observed along road surfaces between A p r i l 22, 2002 and A p r i l 24, 2002. See caption for Figure 3.11 for further information. N o observations were made above 1275 m, however no overland flow was observed above 1200 m (the upslope extent of the above map). 49 M A Y 1 5 - 1 6 , 2 0 0 2 Fbads Catchment Boundaries Fermanent Stream Channels Ephemeral Stream Channels Start of overland flow Continuing overland flow Row diverted by wooden cross-ditch Interception of subsurface flow at cut-banks 100 SCALE 100 200 300 400 ' Arrows rep-resenting overland flow along roads are not drawn to scale. ' Locations of observed overland flow along roads were estimated in the field. These locations are approximations and were not accurately surveyedor otherwise georeferenced. Figure 3.13. Overland flow observed along road surfaces between M a y 15, 2002 and May 16, 2002. See caption for Figure 3.11 for further information. N o observations were made above approximately 1275 m, however no overland flow was observed above 1200 m (the upslope extent of the above map). 50 \ 3.5 I S O T O P I C C O N C E N T R A T I O N S 3.5.1 Isotopic concentrations in melt and rainwater In this section, the term "integrated melt" w i l l be used to refer to a sample composed o f melt and rainwater, and "melt sampler" w i l l refer to the sampling trays used to collect either integrated melt or rainwater. Isotopic concentrations o f melt and rainwater samples collected from melt samplers are presented Appendix D . This section presents results o f analysis of isotopic concentrations in integrated melt and rain samples and explores their variation through time and space. Results are presented for samples collected on four dates: March 24, 2002, and March 27, 2002, were days o f rain-on-snow above 1175 m, while A p r i l 10, 2002 followed a 5-day dry period and experienced minor rainfall (< 5 mm). Finally, a number o f rainfall samples along with integrated melt samples were collected on A p r i l 14, 2002. Spearman rank order correlation analysis revealed statistically significant relations between 5 l s O and elevation (p < 0.05) for samples collected on each date save March 24, 2002 (Table 3.5, Figure 3.14). Rainwater samples were generally less depleted in l s O relative to integrated melt samples. Though not statistically significant, samples collected M a r 24, 2002, suggested a positive relation between 8 1 8 0 and elevation. Negative relations were seen for all other time periods (Figure 3.14). Statistically significant correlations were found between 8 1 8 0 and time for integrated melt samples collected from most elevations/samplers (Table 3.6), with water inputs becoming increasingly enriched in oxygen-18 over time (Figure 3.15). 51 Table 3.5. Results of Spearman correlation analysis comparing variation in isotopic concentrations of integrated melt and rainwater samples with changes in elevation. Data are stratified by date to control for temporal variability in isotopic Concentrations of water samples. Date(s) Sample Type n r s P Mar 24, 2002 Integrated Melt 6 0.714 > 0.05 Mar 27, 2002 Integrated Melt 6 -0.943 0.05 Apr 10, 2002 Integrated Melt 7 -0.786 0.025 Apr 14,2002 Rainwater 5 -0.9 0.05 Table 3.6. Results of Spearman correlation analysis comparing variation in isotopic concentrations of integrated melt samples over time. Data are stratified by elevation to control for spatial variability in isotopic concentrations of water samples. Elevation (m) Sample Type n r, P 1275 Integrated Melt 5 0.5 > 0.05 1270 Integrated Melt 9 0.561 > 0.05 1135 Integrated Melt 10 0.952 < 0.005 1060 Integrated Melt 6 1.000 <0.01 945 Integrated Melt 6 0.886 < 0.025 935 Integrated Melt . 7 0.786 0.025 825 Integrated Melt 6 0.943 <0.01 655 Integrated Melt 4 0.4 > 0.05 52 -14 - - 2 0 ~ • Melt X Rain I I I I I I J .22 600 700 800 900 1000 1100 1200 1300 Elevation Figure 3.14. Variation in isotopic concentrations o f meltwater and rainwater with changes in elevation. Data for three periods of melt (top 3) and one period o f rain at lower elevations (bottom) are presented ( L O W E S S smoother, tension = 0.9). 53 -14 -14 • Melt X Rain Figure 3.15. Time series of isotopic concentrations in meltwater measured at four elevations. Rainwater samples are included for comparison; however, the LOWESS smoother (tension = 0.9) is fitted to the meltwater data points only. 54 m m X 3.5.2 Isotopic concentrations in streamwater Oxygen-18 concentrations were similar amongst streamwater samples collected Dec. 20, 2001 (-16.5 for T H , -16 for GB, -16.9 for GS) and Feb. 11-13, 2002 (-16.53 for T H , -16.15 for GB, -17.3 and -16.8 for GS) indicating little isotopic variation in during baseflow. The concentration of oxygen-18 tended to decrease in all streams up until about the time of peak flows (Figure 3.19), becoming similar to that of integrated melt samples collected at that time (Figure 3.16). Streams displayed differing concentrations early in the season, with concentrations becoming similar around the time of peak flows. Following peak flows, streams again displayed differing concentrations. Concentrations were lowest iii Gurn Spring both prior to and following peak flows. Both early in the season and after peak flows, Thickett Creek and Gurn Brook streamwaters were more enriched in oxygen-18 relative to Gurn Spring (with one exception being the water sampled from GB on June 18, 2002). 3.5.3 Isotopic concentrations in subsurface water Subsurface water sampled from piezometers in the T H and GS catchments contained similar concentrations of oxygen-18 (Figure 3.17). Excepting the single sample collected February 11, 2002, samples collected from GB piezometers were depleted in 8 l s O relative to samples collected from T H and GS. In T H and GS piezometers there was little variation in 8 1 8 0 among the three sampling dates. In GB piezometers, 8 I 8 0 decreased throughout the freshet period. Concentrations of oxygen-18 in all piezometer samples collected during base-flow conditions (Februaru 11, 2002) were similar. 3.6 W A T E R C H E M I S T R Y This section presents result from analyses of water chemistry. Section 3.6.1 provides a discussion of water chemistry in subsurface waters sampled from piezometers, while section 3.6.2 examines streamwater chemistry. Concentration - discharge relations are presented in section 3.6.3. Section 3.6.4 presents results from principal components analyses of water chemistry from each of the three streams and from the whole of Ringrose Slope. 3.6.1 Water chemistry in piezometers Water samples were collected from the majority of active piezometers on each of three dates (Feb 11, 2002, May 16, 2002, and Jun 18, 2002), except where water levels on one or more of these dates were insufficient for sample collection (e.g., P20) or piezometers failed to 56 refill within a short time after being pumped dry (e.g., PI2s). Only those piezometers located in near-stream riparian areas (NSRIP) contained enough water on all three sampling dates for sample collection. Exceptions were PI 1, a near-stream hillslope (NSHIL) piezometer located just north of the T H catchment from which samples were collected on all sampling dates, and P20, a NSHIL piezometer located in the GS catchment that produced a sample on Jun 18, 2002. Both PI 1 and P20 were located on short (1-2 m), moderately angled slopes defining the adjoining riparian valley bottoms. No piezometers located in upslope contributing areas (UP) contained enough water on any of the sampling dates to produce a water sample. Concentrations of Ca, Mg, Si, E C , and Na concentrations appeared higher iii water collected from GS piezometers (Figure 3.17). Water from GB piezometers contained the lowest concentrations of K, and the ratio (Na+K)/(Na+K+Ca+Mg) (RNA-K)- Within catchments, chemical concentrations vary amongst piezometers. Concentrations of most chemical parameters varied throughout the freshet period. Concentrations of all parameters appear to decrease with increasing elevation (Figure 3.18). However, concentrations varied substantially amongst piezometers at the same elevation. Spearman correlation analysis revealed significant inverse relations between elevation and concentrations of Ca, Mg, Si, and E C in piezometer samples collected February 11, 2002 (p <= 0.05, n = 8) (Table 3.7). A comparison of RNA/K values in piezometer samples and in streamwater samples revealed the following (Figure 3.19): 1. Higher values of the ratio were measured in streamwater than in subsurface waters at GB. 2. Values of the ratio in T H and GS streamwater appear to fall within the range of values measured in subsurface waters. 3. The ranges of the ratio in subsurface water and streamwater T H are almost identical. 4. The subsurface water chemistry in the GS catchment is more variable than GS streamwater chemistry. 57 w Table 3.7. Spearman correlation analysis showing relation between elevation and concentrations of chemical parameters in piezometer samples collected February 11, 2002. Chemical Parameter n r s P Ca 8 -0.908 < 0.005 K 8 -0.282 > 0 . 0 5 Mg 8 -0.786 < 0.025 Na 8 -0.503 > 0.05 Si 8 -0.921 < 0.005 5 I 8 0 8 -0.216 > 0.05 RNA-K 8 0.061 > 0.05 E C 8 -0.651 < 0.05 58 l l o Feb 11, 2002. x May 16,2002. • Jun 18,2002. 80 70 — 60 _J O) E.50 ro °40 30 20 11 10 9 £ 8 E. m 7 -r—i—i—r-r TH H - K f - h H T—i—i—i—i—r ° GS T— i—r GB 1 I I I I o B I I I I I I I I I I I - H -f - H -11 10 9 8 2 CO 7 3 6 -C 5 4 400 H300 m O o 3 H200 I I I ' i ' i r i i—|—|—i_U—I—I—I rtoo Piezometer ID 2.0 1.5 E 1.0 0.5 -16.0 -^.-16.5 1 i— 0) 3-17.0 O co to -17.5 -18.0 •J I I I T * l TH T—n—r~ GS H -x g H—r-4 - H J L I—i—r GB I 1 I I I I I I I I I x 9 o I I I I I I I ' l l ' H -J I L 4 3 co 0.11 I0.10 -2 I > 0.09 ^ 0.08 I + 0.07 S + co -m.06 Piezometer ID 0.05 Figure 3.17. Concentrations of eight chemical parameters in subsurface water collected from active piezometers on three sampling dates. Dashed vertical lines separate piezometers from the T H , GS, and GB catchments. PI 1 lies on Ringrose Slope just north of the Thickett Creek catchment. 59 80 70 ~60 _j •) E 50 o Feb 11, 2002. 6 x • • ° x 8 ° H h • f — I 4 h x May 16, 2002. • Jun 18,2002. 2.0 r 11 10 8 2 co 7 3 to 6 400 m 300 O o 3 200 100 600 700 800 900 1000 1100 1200 Elevation (m) 1.5 E 1.0 0.5 -16.0 •S • O CO -17.5 -18.0 I 1 9 1 1 • X - x X 0 X X • 1 o ° § * n - • 8 0 • 0 X • | I 1 1 X 1 I 1 1 1 1 o - -1 o _ ° x - i a • » • - • x • • g o • • 1 a * X i • 1 1 1 1 1 1 1 1 1 x o -x B o o a c ) _ ° s ' X t> • 0 X o X X x a D -1 | 1 • 1 J 1 1 1 1 X X o o X o — X o - e • X _ o X a a a — x o x — • 1 o 1 ' • 1 o + 600 700 800 900 1000 Elevation (m) 1100 1200 —'0.05 Figure 3.18. Concentrations of chemical parameters in subsurface waters plotted against elevation. Samples were collected from active piezometers on three seperate dates. Elevations of piezometers were either estimated using an altimeter and reference to topographic features or surveyed using differential GPS. 60 0.11 0.10 Figure 3.19. Comparison of R N A / K values measured in subsurface water sampled from piezometers located in the T H , GB, and GS catchments (scatterplots at left) with that measured in streamwater collected from T H , GB, and GS (box plots at right). 3.6.2 Water chemistry in streamflow Concentrations of the four major cations (Ca, K, Na, Mg) were lowest in GB at all times throughout the freshet (Figure 3.20). Cation concentrations were highest in GS streamwater except during peak flows, when they dropped to similar concentrations as in T H streamwater. The concentrations of Ca, Mg, sum of cations (Ca + Mg + Na + K), and electrical conductivity (EC) exhibited similar variation through time, dominated by a decrease associated with peak streamflow, followed by an increasing trend through the post-peak discharge recession. Spearman correlation analysis revealed that E C provides a good proxy for total cation concentrations in streamwater (rs = 0.954, P = 0.005, n = 130). Silica (Si) concentrations were highly variable, with no discernable pattern emerging over the freshet period. Concentrations were generally similar in T H and GB streamwater, and highest in GS streamwaters. Dissolved organic carbon (DOC) and RNA-K were highest in GB and lowest in T H streamwater. DOC and the ratio displayed an overall increase up to the time of peak flows for all streams. 3.6.3 Concentration — discharge relations For Thickett Creek, the relations between Ca and E C vs. discharge exhibited clockwise hysteresis. During the falling limb, Ca and E C followed a curvilinear trend with concentrations increasing rapidly as flows decreased towards original levels (Figure 3.21). Concentrations of Mg were also higher during the rising limb; however, the hysteresis observed for Ca and E C was less marked. Concentrations of all cations were more variable at a given discharge during rising stage. RNA-K increased gradually with increasing discharge, with the ratio being higher during falling stage. 8 I 8 0 decreased with increasing discharge, and showed no obvious difference between samples collected during rising and falling stages. Silica concentrations varied chaotically with discharge, with no apparent difference between samples collected during rising and falling stages. 62 4 701 i i i i i i i i i i i i i i i i i i i i i Date Date Figure 3.20. Time series of chemical parameters measured in streamwater at the weirs on the three study streams. Hydrographs are included for reference at the bottom. 63 For Gurn Brook, similar relations are observed for Ca, Mg, and E C (and to a lesser degree K): concentrations begin relatively high during rising stage at low flows, then drop steadily with increasing discharge (Figure 3.22). During falling stage, points generally plot along the same line as during rising stage; however, there is a rapid increase in concentrations below 5 L/s, resulting in higher late-season concentrations of Ca, Mg, and E C than during pre-melt base flow conditions. R N A - K increased rapidly with increasing discharge, following a roughly linear trend. No obvious distinction was observed between samples collected during the rise and fall in the hydrograph. 8 1 8 0 decreased with increasing discharge, with samples collected during falling stage generally plotting below those collected during rising stage for a given discharge. Silica concentrations show a decrease with increasing discharge; however, the relation is weak and no distinction between samples collected during rising and falling stage is seen. Concentrations of all chemical parameters were similarly variable under conditions of rising and falling stage. The concentration - discharge relations for Gurn Spring (Figure 3.23) display a number of important differences from those of GB and T H : 1. Concentrations of cations and E C do not exhibit hysteresis. 2. No clear concentration-discharge relation was observed for RNA-K-3. 8 I 8 0 generally increased with increased discharge during rising stage, but varied oppositely during falling stage. 4. Silica varied relatively chaotically with discharge. 64 Thickett Creek 50 E.40 CD O 35 30 7.0 6.5 5.5 5.0 1 V 1 1 1 1 1 1 \ X " ' '\ X O x O I s c% o o ° o i 1 1 o 1 1 o 1 - 1 1 1 1 1 1 1 1 1 -X X X X / f " ' e l * X o o X X ' 1 1 I I 1 1 1 1 1 1 1 1 1 1 1 1 - X O °x - *> x '• '• o>x X x.-O o o o * o ax>c? o * o X X X ° >x *x O * X o -X 1 1 1 1 I I 1 | I 1 1 1 1 1 1 1 1 -* X X*<XX V . o x x 6 _ • X/ 1 * C9 1 I I o ° 1 1 . O 0 1 1 x Rising Stage 1.2 1.1 g-1.0 ~& E. * 0.9 0.8 0.7 CO 5 ? 260 250 240 m O 230 220 o 3 210 200 190 180 0 1 2 3 4 5 6 Q(L/s) 7 8 9 -16.0 "g -16.5 -17.0 -17.5 o Falling Stage n — i — i — i — i — i — i i c9 I I I I I 1 1 I o o B % I I I O X I I I I I I I I I I I I I I L_ 2.6 2.4 2.2 z 2.0? <e. 1.8" 1.6 1.4 0.09 0.08 | + 0.07 S 0.06-5 + 0.05 cf 0 1 2 3 4 5 6 7 8 Q(L/s) 0.04 Figure 3.21. Concentration-discharge relations for Thickett Creek. Samples collected on the rising falling limbs of the hydrograph are represented by the black circles and grey crosses, respectively. 65 Gurn Brook 40 ?3 0 t "5) E, C O ° 2 0 10 E, 55 o T 1 1 8 r o -o x> I f $ a x 5)-V 1 I I X - o 1 1 1 -8 o -tg % -° O y / * 1 1 1 o 1 1 1 ~ 0 -o * x » -A ? o " " o o O 0*;0 0 o -o o o x 1 I I o 1 1 1 • J* _ - ° o X 1 1 1 10 20 30 Q(L/s) 200 m 150 O -M00 50 40 Rising Stage 1.2 1.1 1.0 - 0 . 9 | 0.8 * 0.7 0.6 0.5 0.4 o Falling Stage -15.5 -16.0 I -16.5 a. „ ° -17.0 -17.5 -18.0 t 8 ° ? 8 o o \ 9 € H + * ° * x x o o e 10 20 30 Q(U/s) 2.5 2.0 z CD 3" 1.5 1.0 0.12 -0.11 z > 0.10* 0.09 0.08 0.07 + + O CO + CO 0.06 40 Figure 3.22. Concentration-discharge relations for Gurn Brook. Samples collected on the rising and falling limbs of the hydrograph are represented by the black circles and grey crosses, respectively. 66 Gurn Spring 70 60 • E 30 20 E. 55 9x ° o 8*8 * 5 10 15 Q(L/s) 10 9 8 7 3 CO 350 300 m o 250 S" 3 H200 20 150 Rising Stage 1.3 1.2 1.1 1H.0 0.9 0.8 0.7 -16.4 -16.8 CD S-17 .2 O 00 -17.6 -18.0 o x 6 o Falling Stage _ j ! °6 oa, * 10 Q (L/s) 15 10.09 > J0.08 * 2 0) + + 10.07 8 J0.06 20 Figure 3.23. Concentration-discharge relations for Gurn Spring. Samples collected on the rising and falling limbs of the hydrograph are represented by the black circles and grey crosses, respectively. 67 3.6.4 Principal components analysis of water chemistry Principal components analysis (PCA) of streamwater chemistry was conducted for each of the three study catchments using the following variables: Si, DOC, RNA-K, and EC. Two separate PCA's was performed for Ringrose Slope water chemistry. The first was conducted using all streamwater samples from T H , GB, and GS, while the second was conducted without DOC as a variable but with all streamwater and piezometer samples included. Results are presented in Table 3.8. For T H streamwater, Factor 1 is composed of E C varying in strong opposition with DOC and RNA-K- The second factor is composed almost entirely of Si. For GB streamwater, Factors 1 and 2 are somewhat different, composed of E C varying in strong opposition with RN 3 -K (but weakly related with DOC) and Si varying with DOC, respectively. For GS streamwater, Factor 1 is composed of E C varying opposition with DOC and weakly in opposition with Si, while Factor 2 shows R N a . K and D O C varying together, both in weak opposition with Si. The first PCA of Ringrose Slope water chemistry showed E C varying in strong Opposition with both DOC and RNA-K (Factor 1), and Si varying independently of the other components (Factor 2). The second Ringrose Slope PCA yielded similar results, with E C varying in strong opposition with RNA-K (Factor 1), and Si varying independently (Factor 2). 68 Table 3.8. Results of PCA of streamwater chemistry from each of the three study catchments. PCA's for Ringrose Slope were conducted using all streamwater samples and Si, DOC, RN 3 -K and E C as variables (Slope [1]), and using all streamwater and piezometer samples but without DOC as a variable (Slope [2]). n Component Loadings Eigenvalue % Variance Explained Si D O C RNA/K E C T H 30 Factor 1 -0.096 0.910 0.795 -0.942 2.357 58.9 Factor 2 -0.994 0.017 0.016 0.132 1.009 25.2 GB 32 Factor 1 -0.295 0.340 0.883 -0.937 1.860 46.5 Factor 2 0.801 0.674 0.126 0.111 1.124 28.1 GS 27 Factor 1 -0.469 -0.668 0.365 0.896 1.601 40.0 Factor 2 0.466 -0.667 -0.806 0.075 1.317 32.9 Slope 89 Factor 1 -0.206 0.943 0.933 -0.973 2.747 68.7 0 ) Factor 2 0.978 0.110 0.099 -0.005 0.979 24.5 Slope 158 Factor 1 0.514 -0.817 0.913 1.767 58.9 (2) Factor 2 0.843 0.458 -0.064 0.924 30.8 69 A bivariate plot o f Si against RNA-K measured in streamwater and piezometer samples for each o f the three study catchments reveals a distinct separation between catchments (Figure 3.24). The data points form a roughly triangular space (outlined in dashed lines) resembling that observed in many diagrams describing a three-component mixing model o f water chemistry, such as that employed in many end-member mixing models ( E M M A ) (Christophersen et al., 1990; Hooper et al., 1990). Potential end members include deep groundwater (high in Si), shallow groundwater (low in Si and RNA-K), and surface water (low in Si , high in RNA-K)-Thickett Creek streamwater (red crosses) and piezometer samples (blue crosses) plotted in the same general area, being relatively low in both parameters. G u m Brook streamwater plotted towards the far right side of the graph, having the highest values o f RNA-K and relatively low S i concentrations. Piezometer samples collected within the G B catchment plotted separately from streamwater samples, being among the lowest in both parameters. Gurn Spring streamwater plotted towards the middle of the graph, having the highest Si concentrations and moderate ratio values. Piezometer samples collected from within the G S catchment plotted in the same general area, however, somewhat above points representing streamwater from G S . Piezometer samples were higher and more variable both for the ratio and concentrations o f S i than were streamwater samples from the G S catchment. 70 11 p 10-9 -8 -7 -6 -5 -A / ^ / ( / I I I I I I I O O \ / / / L -\ 0.025 0.050 0.075 0.100 0.125 (Na+K)/(Na+K+Ca+Mg) GB Streamwater • GB Piezometers GS Streamwater GS Piezometers TH Streamwater i TH Piezometers Figure 3.24. Bivariate plot of Si l ica vs. RNA-K- Samples are grouped by catchment and streamflow samples are distinguished from piezometer samples. The dashed triangle represents a possible interpretation of the plotted data as potential end-members for runoff from Ringrose Slope. 71 \ C H A P T E R 4 D I S C U S S I O N This chapter provides an interpretation of results in the context of the specific research question and hypotheses stated in chapter 1. Section 4.1 addresses the validity of a fifth hypothesis: The Gurn Spring catchment receives water from outside the catchment by means of subsurface flow through bedrock fractures. This hypothesis was generated due to results from analysis of historical unit peak flows (Table 3.2) and unit runoff (Figure 3.5), along with analyses of water chemistry data collected during the field season. Section 4.2 addresses interpretation of temporal sources of catchment runoff, and section 4.3 addresses research question 2 and hypotheses 2 - 4 (Chapter 1, Section 1.3). Section 4.4 discusses implications of the research findings for the hydrological effects of forestry operations, with a focus on forest roads. This section was included to help address the long-term goals of the broader Ringrose Slope study (of which this thesis is a part). Section 4.5 discusses implications of the research finding for catchment scale modelling. 4.1 T H E P O T E N T I A L R O L E O F I N T E R - C A T C H M E N T G R O U N D W A T E R F L O W The GS catchment consistently generated two to three times more unit runoff than both T H and GB despite its relatively low elevation (Table 3.2, Figure 3.5), suggesting inter-catchment groundwater flow into GS through bedrock fractures. This inference is consistent with field observations of fractured rock outcrops on Ringrose Slope and with the deformed nature of the local bedrock as shown in geologic maps (Geological Survey of Canada, 1932). Although there is some uncertainty with respect to the catchment area of T H (see section 2.2.1), GB and GS catchment areas are considered to be relatively accurate and should yield meaningful comparisons between unit runoffs. Comparison of unit runoff amongst these catchments suggests Gurn Spring is receiving water from outside its catchment boundaries (Figure 3.5). The timing of peak flow at Gum Spring supports the hypothesis of inter-catchment bedrock flow, particularly through large fractures allowing relatively rapid flow. Due to its small size and relatively low elevation, flows were expected to peak on Gurn Spring before those on Thickett Creek and Gurn Brook. However, peak flows in 2002 were synchronous amongst streams despite the lack of snow cover in the GS catchment by that time (Table 3.2, Figure 3.3). Peak flows on Gum Spring recorded in the majority of previous years were also relatively well synchronized with those on Thickett Creek and Gurn Brook (Table 3.2), suggesting continued meltwater inputs after the disappearance 72 w of the snow cover from GS. Alternatively, the observed synchrony of peak flows amongst catchments may be the result of forcing by rainfall/rain-on-snow events. Streamwater and subsurface water in GS had high concentrations of Si and cations relative to T H and GB, suggesting runoff pathways characterized by long contact times (Figure 3.16, Figure 3.19). A possible explanation is that there exists a soil-water reservoir within GS characterized by a long contact time, and that can be mobilized during periods of high flow. Alternatively, there may be a reservoir of water within the bedrock fracture system that possesses a long contact time, and can be mobilized during periods of high flow. This interpretation supports the inferred existence of inter-catchment transfers of water into Gurn Spring. Together, these results provide strong evidence of inter-catchment groundwater flow on Ringrose Slope. Bedrock flow has been proposed as the dominant streamflow component in at least one other steep, small headwater catchment, although bedrock in that catchment consisted of sedimentary deposits, particularly sandstones. (Anderson et al., 1997). While Gurn Spring appears to receive inputs of water from outside its boundaries, the source of this water is difficult to specify. It is possible that the water originated in the upper catchment of Gurn Brook, and flowed through fractures across the northeastern divide of the GS catchment. Despite its high elevations, Gurn Brook produced unit runoff and unit peak flows similar to, and sometimes lower than, those of Thickett Creek (Figure 3.5), consistent with the notion that GB lost water. 4.2 W H A T A R E T H E R E L A T I V E P R O P O R T I O N S O F N E W A N D O L D W A T E R IN S T R E A M F L O W A N D H O W D O T H E Y C H A N G E O V E R T H E S N O W M E L T P E R I O D ? 4.2.1 Challenges in applying isotopic hydrograph separation Isotopic concentrations in snowmelt and rainwater vary in both time and space, which generates uncertainty in the "event" water signal and confounds the application of isotopic hydrograph separation (IHS). Most studies have applied IHS to rainfall-dominated catchments (e.g. Buttle and Peters, 1997; Hoeg et al., 2000; Ladouche et al., 2001) or snowmelt dominated catchments of low relief (e.g. Moore, 1989; Maule and Stein, 1990; Buttle and Sard, 1992), where both the isotopic concentrations and volumes of event waters are relatively uniformly distributed over the catchment, thereby limiting this confounding effect. The relations between oxygen-18 concentrations in meltwater samples and elevation varied through time, reflecting elevational trends in the initial oxygen-18 concentration of the snowpack, as well as the differences in the timing and rates of snowmelt, which would influence 73 the fractionation process (Table 3.5, Figure 3.14). Oxygen-18 concentrations in meltwater samples exhibited statistically significant increases with time at some sites (Table 3.6, Figure 3.15), consistent with the effects of fractionation during snowmelt. However, the lack of significant relations at three of the eight sampling locations indicates that processes other than fractionation complicated temporal variation in snowmelt isotopic concentrations (e.g. differential mixing between snow-water and rainwater or between distinct layers within the snowpack). This spatial and temporal variability of the event-water 8 l s O concentrations, coupled with spatially variable input rates, makes it difficult to apply HIS with any reasonable level of confidence. Temporal variability in event water isotopic signatures has been addressed in previous studies using the current meltwater 8 1 8 0 sampled from snow lysimeters at each sampling time step during the snowmelt period (Maule and Stein, 1990; Wels et al., 1991; Shanley et al., 2002). However, this method assumes that each meltwater contribution is only stored within the soil until the following sampling occasion (Laudon et al., 2002). Laudon et al. (2002) introduced the runoff-corrected event water approach (runCE) as an alternative means of accommodating temporal variability in the isotopic signature of snowmelt inputs. This method accounts for both the timing and volume of meltwater entering the catchment and the discharge of previously melted and stored water at each time step during the snowmelt period. However, to the best of the author's knowledge, no methods have yet been developed to accommodate the combined spatial and temporal variability in the snowmelt isotopic signal. 4.2.2 Qualitative interpretation of oxygen-18 data Reasons for the discrepancy between the second streamflow samples collected from T H and GS (Figure 3.19) relative to the first and third are unclear. These samples were submitted to the laboratory at different times and analysed in separate batches. However, the test method employed for all samples was identical and is likely not responsible for the observed discrepancy. This conclusion is supported by the lack of a similar discrepancy in GB streamwater. Streamwater samples collected during pre-melt base flow were enriched in oxygen-18 relative to early meltwater samples (Figure 3.16). Oxygen-18 concentrations in Gurn Spring pre-melt baseflow were depleted relative to T H and GB. In Thickett Creek and Gurn Brook, isotopic concentrations remained relatively steady throughout the field season (typically ranging from about -16.5 to -17 per mil) despite inputs of significantly more depleted new water early in the melt period, suggesting domination by old water during the early hydrograph rise. Approaching peak flow, the concentration of oxygen-18 in meltwater becomes similar to that in streamwater, thereby limiting further resolution of relative contributions from new/old waters (Figure 3.16). 74 In Gurn Spring, oxygen-18 increased between the third and fourth samples to levels similar to those in Thickett and GB, then declined in a similar fashion to T H and GB (Figure 3.19). Following peak flow, oxygen-18 concentrations in GS dropped to values similar to those of pre-melt baseflow. It may be that during snowmelt there is mobilization of a reservoir of water having a higher oxygen-18 concentration relative to pre-melt baseflow. One hypothesis is that pre-melt baseflow is supplied by bedrock groundwater having a relatively depleted isotopic signature. During snowmelt, inputs of new water mobilize stored soil-water having a relatively enriched oxygen-18 signature. Following the main melt period, relatively depleted bedrock groundwater again dominates. This two-reservoir model is likely complicated by significant mixing between bedrock groundwater and resident soil water prior to the discharge of bedrock water into the GS stream channel. 4.2.3 Comparison of Results with Other Studies Concentrations of oxygen-18 in T H , GB, and GS streamwater became increasingly similar to meltwater concentrations sampled from snowmelt lysimeters. On April 14-15, 2002, oxygen-18 concentrations in streamwater and meltwater were nearly identical (Figure 3.19, Appendix D). Other studies employing isotopic hydrograph separation to identify streamflow contributions from old vs. new-water during snowmelt have found considerable differences between concentrations of oxygen-18 in streamwater and meltwater at the time of peak flows, suggesting relatively large old-water contributions at this time (e.g. Moore, 1989). The similar isotopic signatures in streamwater and meltwater during peak flow may suggest high new-water contributions in the three study catchments relative to other snowmelt-dominated catchments described in the literature. However, it could also indicate that the difference between old and new-water isotopic concentrations during peak flow is not sufficiently great to resolve relative peak flow contributions from old and new-water. 4.3 C O M P A R I S O N O F S T R E A M F L O W R E S P O N S E A N D R U N O F F P A T H W A Y S 4.3.1 Comparison amongst the study catchments Gurn Brook displayed the flashiest streamflow response of the three catchments, followed by GS and T H , suggesting larger streamflow contributions from deeper subsurface runoff pathways in T H and GS relative to GB. As the snowline progressed upslope, a greater de-coupling of source areas of meltwater from the stream channel is suggested for T H and GS relative to GB. This inference is supported by the shape of streamflow hydrographs (Figure 3.2), 75 comparison of diurnal streamflow response (Figure 3.6), and observed lags between water inputs and streamflow response (Figure 3.7). Snowpack depth has been shown to affect diurnal snowmelt signals and resulting streamflow variations. The role of deep snowpacks in attenuating diurnal melt signals has been well documented (Jordan, 1983). However, peak snow-water-equivalents in the study catchments were low (~ 0.2 m) relative to those measured by Jordan (1.25 m). Because the snowline represents the lowest snow covered area and is characterized by the thinnest snowpack and thus the least snowpack-related attenuation, the role of the snowpack in attenuating snowmelt signals is unlikely to influence the interpretations provided below regarding connectivity between the upslope-moving snowline and stream channels. The T H channel was the shortest of the three streams, except during the two week period around the time of peak flow when a discontinuous extended channel network coupled with small areas of surface saturation facilitated connectivity between rising source areas of meltwater and the stream channel (Figure 3.10). As the snowline rose above 1000 m (approximately half way up the catchment), the extended channel network receded and source areas became increasingly disconnected from the channel. Though overland flow was observed along roads within T H , this was primarily at lower elevations, suggesting that overland flow along roads did not facilitate connectivity between upslope areas and the channel (Figures 3.11 - 3.13). At all times during the freshet, connectivity between upslope source areas and channel units was maintained primarily through subsurface flowpaths. This resulted in sluggish streamflow response relative to the other streams as noted above, thereby supporting hypothesis #2. The dominance of subsurface runoff pathways in the T H catchment throughout the freshet is supported by examination of T H groundwater and streamwater chemistry (Figure 3.18, Figure 3.19). Concentrations of Ca, Mg, and E C were elevated during rising stage in T H relative to falling stage at given discharge (Figure 3.20). Other studies have associated clockwise hysteresis amongst rising and falling limbs on plots of groundwater levels vs. streamflow with a threshold in relative streamflow contributions from riparian vs. hillslope areas (Kendall et al., 1999). Kobayashi et al. (1999) observed counter-clockwise hysteresis amongst E C concentrations during a diurnal rise and fall in streamflow in response to a rainfall event in the Moshiri experimental basin and associated this with increasing discharge from shallow flowpaths. Hall (1971) attributed a clockwise hysteresis amongst E C concentrations during rising and falling streamflow in a subcatchment of the Sleepers River watershed to non-random trends in the storage-volume - discharge relation. It is hypothesised that the clockwise hysteresis observed in T H suggests a release of stored subsurface water to the stream channel during rising stage by a process of displacement by relatively dilute snowmelt and rainwater. This is similar to the 76 process of translatory flow originally described by Hewlett and Hibbert (1967), and used to account for the observed proportions of old/new water in several isotopic tracer studies (see review by Buttle, 1994). As the catchment wetted further, the saturated zone thickened, causing saturation overland flow in discrete areas and elevated streamflow contributions from surface/shallow subsurface flowpaths, associated with slightly higher concentrations of DOC, lower E C and cation concentrations, and higher R N A _ K . These chemical changes were more evident in GB and GS, probably reflecting the lesser importance of SOF in T H and the greater dominance of deeper subsurface runoff pathways that served to link areas of shallow subsurface flow, surface saturation, and ephemeral channel reaches during periods of high runoff. As streamflow declined following peak flows, the saturated zone thinned and contributions from surface/shallow subsurface runoff pathways dropped, consistent with rising E C and cation concentrations and decreases DOC and RN 3-K-The deeply incised channel of Gurn Brook was the longest of the three catchments and facilitated rapid movement of water between meltwater source areas and the catchment outlet early in the melt period. Though minimal SOF was observed within the catchment, the Gurn Brook channel network expanded both upslope and laterally as the snowline rose, thereby maintaining connectivity between source areas and the stream (Figure 3.10). The extended channel network persisted for some time following peak flows, ensuring that connectivity was maintained within the GB catchment, even when the snowline rose to the upper reaches of the catchment. Throughout the freshet, surface flowpaths (including overland flow along roads) constituted important linkages between upslope source areas and the catchment outlet (Figures 3.11 — 3.13).. This resulted in the Gurn Brook displaying the most responsive streamflow of the three catchments throughout the snowmelt period, thereby supporting hypothesis #3. Contributions from surface and/or shallow subsurface flowpaths appeared to be greater in GB compared to T H and GS, as suggested by the greater difference between streamwater and groundwater chemistry (Figure 3.18) and the greater relative changes in streamwater chemistry associated with peak flows (Figure 3.19). Channel flowpaths (including overland flow along roads) are suggested as the primary surface runoff pathways in Gurn Brook. The lack of observed SOF in GB suggests that subsurface flow is responsible for connecting source areas and the channel network (Figure 3.10). However, streamwater and subsurface water chemistry in the GB catchment suggest that subsurface flowpaths were shorter in that catchment relative to T H and GS (Figures 3.16 and 3.19). In GS, a higher water table developed in the near-stream riparian corridor relative to near-stream riparian areas in T H (Figure 3.8). Saturation overland flow directly connected to the 77 \ stream channel was observed in the near-stream riparian area upslope of the weir on Gum Spring during the three weeks bracketing peak flow (Table 3.4, Figure 3.10). This suggests that streamflow during this period received contributions from SOF linked to the Gum Spring channel by surface and shallow subsurface flowpaths. The Gum Spring channel itself was relatively short, although longer than the permanent channel on Thickett Creek, and experienced no observable expansion during the freshet period. Early in the melt period, as the snowline began to rise, the GS channel maintained connectivity between source areas of meltwater and the catchment outlet. As the snowline rose beyond the channel head (i.e. above 725 m), the lack of channel expansion coupled with the near total absence of overland flow along roads resulted in a progressive reduction in connectivity (Figure 3.11- 3.13). The specific timing and nature of runoff processes in Gurn Spring remain unclear. Streamwater chemistry suggests that during baseflow conditions, runoff consisted primarily of deep subsurface water having a relatively long contact time (Figure 3.19), such as bedrock groundwater. Soil water, having a relatively enriched isotopic signature, was mobilized as a result of new melt and rainwater inputs and became the dominant streamflow contributor in GS during freshet. This inference is supported by the rise in GS piezometer water levels, particularly in those piezometers located in near-stream riparian positions (Figure 3.7, Figure 3.8, Appendix B) and by time series of oxygen-18 concentrations (Figure 3.19). Water levels in the riparian corridor rose near to the ground surface at this time, producing areas of SOF tightly coupled with the stream channel (Figure 3.8, Figure 3.10). This resulted in greater streamflow contributions from surface and shallow subsurface flowpaths as indicated by increases in streamwater D O C and R N a . K at this time, reaching their maximum at the time of peak discharge (Figure 3.19). The high unit peak flows in GS relative to T H and GB indicate that transfers of bedrock groundwater into the GS catchment also contributed to streamflow during freshet. Following peak flow, concentrations of all chemical parameters (save perhaps Si) in streamwater returned to near baseflow levels, suggesting that bedrock groundwater again dominated streamflow contributions (Figure 3.19). This two-reservoir model of streamflow generation in GS is consistent with qualitative interpretations of isotopic results. This model is likely complicated by mixing between reservoirs due to bedrock water entering catchment soils prior to being discharged to the stream channel. 4.3.2 Comparison of Results with Other Studies The study catchments Thickett Creek, Gurn Brook, and Gum Spring displayed unusually little diumal streamflow response compared with other snowmelt dominated headwater 78 catchments surveyed from the literature (Kobayashi, 1985; Kobayashi, 1986; Moore, 1989; Buttle and Sami, 1992; Burns and McDonnell, 1998; Shanley et al., 2002). Even the Redfish Upper catchment, located less than 50 km away, displayed significantly greater diurnal streamflow response, despite being forced by similar climatic conditions (Unpublished Data, B C MoF). The study catchments were the steepest of all the catchments surveyed (Table 3.3), and the strong inverse relation between diurnal response and catchment steepness (p = 0.005, n = 10) suggests this may be the cause of the study catchments' relatively sluggish diurnal streamflow response, although this notion is counter-intuitive. It may be that relatively flat catchments experience high diurnal streamflow response as a result of one, or both, of the following: (1) synchronized generation of meltwater over a relatively large proportion of the catchment due to uniform energy inputs, and/or (2) increased rate of transfer of water to the stream channel due to the development of widespread saturation overland flow. In a study by Kim et al., (in press) in the upper Gray Creek catchment, British Columbia, comparisons were made between diurnal variations in streamflow and water table elevations (measured at 5 min intervals). Results showed that throughflow can be sufficiently rapid to produce diurnal streamflow response during snowmelt, particularly if the water table rises into more permeable shallow soil layers. This process of transmissibity feedback was observed by Kendall et al. (1999) in the W9 sub-catchment of the Sleepers River watershed and was associated with an abrupt increase in streamflow with a rise of the water table into near-surface zones of high transmissivity. Flatter catchments are more likely to experience widespread elevation of the saturated zone into more conductive soil layers and an associated throughflow-induced diurnal response. Steep catchments such as those on the Ringrose Slope may have such limited areas capable of producing saturation overland flow and rapid throughflow that response is dominated by relatively more sluggish subsurface flow paths, particularly when the meltwater source becomes disconnected from the channel network. 4.4 I M P L I C A T I O N S F O R T H E H Y D R O L O G I C A L E F F E C T S O F F O R E S T H A R V E S T I N G A N D R O A D C O N S T R U C T I O N This section briefly discusses the implications of the research findings for the hydrological effects of forest harvesting (section 4.4.1) and road construction (section 4.4.2). The hydrological effects of forest harvesting and road construction constitute a field of research in forest hydrology in and of themselves, and a detailed evaluation of these effects is beyond the scope of this thesis. 79 4.4.1 Forest harvesting The specific impacts of forest harvesting on the T H , GB, and GS catchments would depend largely on the type of harvesting prescribed. However, previous studies indicate that increases in the magnitude and timing of peak flows would result in all catchments, as would increases in annual water yield and summer low flows (Harr et al., 1982; Troendle and King, 1985; Hetherington, 1988; Van Haveren, 1988). The magnitude of these increases would depend on the area harvested. In Thickett Creek and Gum Brook, the channel network was observed to expand in response to increasing water inputs, suggesting that increases in annual water yield and the magnitude of peak flows could lead to greater expansion of the channel network. This could result in longer and more continuous extended channel networks during peak flow in the T H and GB catchments, respectively. In both catchments, this would increase connectivity between source areas and the catchment outlet producing flashier streamflow response. Increased low flows in Gum Brook might lead to measurable flows at the weir throughout the summer months. In Gurn Spring, no expansion of the channel network was observed in response to increasing water inputs. This suggests that increases in annual water yield and the magnitude of peak flows would not lead to an expansion of the channel network. Consequently, no change in the responsiveness of Gum Spring streamflow would result. If the hypothesis that GS receives inter-catchment transfers of bedrock groundwater from GB is true, then the effect of forest harvesting in Gurn Brook's upper catchment may not show up fully as a change in GB streamflow measured at the weir: some of the increased runoff could end up discharging via Gum Spring. This could complicate interpretation of the results of a paired-catchment analysis of harvesting in the GB catchment. 4.4.2 Road construction The hydrologic effects of forest roads and road construction have received significant attention in the literature (e.g., Megahan, 1972; Beschta, 1978; Wemple et al., 1996). The impact of road construction at Ringrose Slope would depend on the density of roads and compacted areas within the study catchments, as well as the spacing of drainage-relief culverts. Road construction would likely result in increased interception of subsurface waters and the generation of rapid overland flow along road segments. This would increase streamflow contributions from surface flowpaths and connectivity between source areas and channel units, thereby reducing lag times and increasing catchment responsiveness. This would likely be most pronounced in Thickett 80 Creek, where deep subsurface flowpaths dominate throughout the catchment. The effect would be least pronounced in Gurn Brook at lower and middle elevations, where deep subsurface flowpaths constitute a less important source of streamflow generation relative to channel flow. Construction of forest roads that cross drainage boundaries could further facilitate the inter-catchment transfers of water along surface flowpaths, thereby complicating the processes currently responsible for the movement of water between catchments. This would also result in some catchments losing water and thus experiencing reduced annual water yield, with water yield increases in receiving catchments. In an extreme scenario, roading could result in sufficient transfers of water out of a catchment to render it dry during low-flow periods. Culverts and cross-ditches constructed during road construction would result in lateral concentrations of flow along slopes. This would increase soil erosion at discrete locations thereby increasing the potential for mass wasting and landslide events. This is particularly true where slope gradients are highest, such as at lower and middle elevations and near channel units within the GB catchment. Increased rates of sedimentation might also result following mass wasting events or during road construction itself. From a management perspective, these impacts might be considered insignificant or manageable. However, water supply licenses have been issued for all three of these streams, with Thickett Creek and Gurn Spring currently being used as primary supplies of drinking water. Changes in their flow regimes could have significant impacts on local residents who currently depend on these streams for water supply. Assuming no significant alteration of the channel networks as a result of harvesting activities and road building, nothing in the preceding discussion suggests that forest harvesting would have adverse effects on water supply in these streams. If anything, increased low flows during summer months in Gurn Brook, currently the only one of the streams not used as a domestic water supply, could make it more suitable for this purpose. However, the potential for Cross-drainage transfers between catchments along road segments is high, as evidenced by field observations of overland flow along existing roads crossing the T H and GB catchment boundaries. However, the potential impacts of forest roads on cross-catchment drainage are minimal relative to the significant inter-catchment transfers of bedrock groundwater into Gurn Spring. The above discussion focuses exclusively on impacts to the magnitude and timing of water supply, and does not consider the potentially adverse effects of forest harvesting and road construction on water quality. 81 4.5 I M P L I C A T I O N S F O R C A T C H M E N T S C A L E M O D E L L I N G Hydrologic models such as T O P M O D E L and the distributed hydrology-soil-vegetation model (DHSVM) treat the soil-bedrock interface as an impermeable boundary, confining runoff processes to the overlying soils and surface environment. Such models are inherently incapable of representing runoff processes and hydrologic response in catchments that experience significant flow through the bedrock itself, particularly if this leads to inter-catchment transfers of water. The existence of inter-catchment subsurface flow through bedrock fractures on Ringrose Slope has been suggested in the above discussion. This calls into question the appropriateness of existing catchment-scale models for representing hydrologic response in similar catchments and areas underlain by fractured parent material. It further highlights the need for models that allow parameterization of the underlying bedrock (e.g. considering it as an extension of the soil profile) and are capable of simulating transfers of water across drainage boundaries. 82 \ C H A P T E R 5 C O N C L U S I O N S 5.1 S U M M A R Y O F FINDINGS This thesis set out to identify: (1) the relative proportions of pre-event and event water in T H , GB, and GS streamflow and how these proportions varied throughout a freshet period, and (2) the dominant runoff pathways in each catchment and how they varied throughout the freshet. It was hypothesised that differences in streamflow response would be related to the degree to which connectivity between the channel and meltwater source areas was maintained as the snowline progressed upslope. It was further hypothesised that this would result in Thickett Creek displaying the least responsive streamflow, followed by Gurn Spring, with Gurn Brook displaying the flashiest response of the three streams. Application of traditional IHS to identify old vs. new water contributions was unsuccessful due to an inability to adequately characterize the volumes and isotopic signature of the new-water component. This was due largely to significant temporal and spatial variation in meltwater isotopic concentrations, particularly in relation to elevation. However, qualitative interpretation of oxygen-18 data suggest relatively large old-water contributions in all catchments, at least during the early portion of the seasonal hydrograph rise. The similarity between streamwater and meltwater isotopic concentrations at the peak of the freshet is not consistent with results from other isotopic tracer studies in snowmelt dominated catchments, and makes it impossible to resolve the relative contributions of old and new water. In Gurn Spring, oxygen-18 concentrations actually shifted away from the new-water concentrations in the early stage of freshet. It is hypothesized that this shift resulted from the mobilization of a reservoir of relatively enriched water, likely stored within the soil. Deep subsurface runoff pathways dominated the T H catchment throughout the 2002 freshet, with minimal surface/shallow subsurface flow contributions during peak flow. In contrast, Gum Brook was dominated by channel flow throughout the snowmelt period, with relatively short subsurface flowpaths maintaining linkages between upslope source areas and the stream channel. Runoff processes in Gum Spring likely reflect a system composed of at least two reservoirs: bedrock groundwater, and soil water. Bedrock groundwater dominated streamflow contributions during baseflow, with isotopically enriched soil water dominating during freshet. Bedrock groundwater again dominated streamflow in GS during hydrograph recession. The 83 suggested model is likely complicated by mixing between reservoirs due to bedrock water entering catchment soils prior to being discharged to the stream channel. Results support the four hypotheses originally stated. Streamflow in T H was least responsive of the catchments, followed by GS and GB. In T H , this resulted from the disconnection between meltwater source areas and the channel as the snowline moved upslope. This de-coupling was less pronounced in GS, resulting in moderately responsive streamflow relative to T H and GB. In GB, expansion of the channel network ensured that connectivity between source areas and the channel was maintained to a greater degree than in the other catchments, resulting in the flashiest streamflow response. Differences in catchment streamflow response are clearly related to the degree to which connectivity between channels and source areas is maintained as the snowline rises. The potential hydrologic impacts of forest harvesting within the study catchments are likely of minimal concern from a management perspective. The main risk is associated with cross-catchment water transfers as a result of subsurface flow interception at road cuts and transfer across catchment divides within the road drainage system. However, care must be taken with respect to any future harvesting and roading operations to avoid impacting drinking water supplies to local residents. 5.2 R E C O M M E N D A T I O N S F O R F U T U R E R E S E A R C H The hydrogeomorphic concept is a promising framework for the investigation of catchment streamflow response as a function of temporal and spatial variation in the attributes of catchment runoff processes and pathways, and in connectivity between hydrogeomorphic units. Previous studies have mainly employed the hydrogeomorphic concept to investigate hydrologic response from temporally variable rainwater inputs, safely assuming little or no spatial variation in rainwater inputs (Sidle et al., 2000; Hangen et al., 2001; Ladouche et al, 2001). This study has benefited from the unique opportunity to investigate the effect of contrasting catchment geomorphologies on streamflow response during snowmelt. Further, it illustrates how streamflow response in steep snowmelt-dominated catchments is affected by spatial variation in water inputs (i.e. from the effect of elevation on snow accumulation, snowmelt, and rainfall). Studies employing the hydrogeomorphic concept to examine streamflow response in such catchments should seek to address this issue. As the concept focuses on linkages between hydrogeomorphic units, studies should endeavour to determine the degree to which connectivity between upslope areas and stream channels is maintained as the snowline moves upslope. This 84 \ study employed a combination of snowline mapping, piezometer measurements, and field observations of surface flow towards this end. This study highlights the need for improved models for conducting IHS for catchments in which the rates and isotopic concentrations of event water vary significantly both spatially and temporally. Further research should examine the role of groundwater flow in bedrock fractures in the context of headwater catchments, particularly the potential for contributing to peak flows as well as base flow. Few studies of headwater catchments have examined fractured bedrock flow in detail; Anderson et al. (1997) is a notable exception. In particular, the study calls into question the appropriateness of existing catchment-scale models for representing hydrologic response in catchments underlain by fractured bedrock, where significant flow might occur through the fracture network, given that most models assume the bedrock to be impermeable. There is need for research focused on the development of parameterizations of the underlying bedrock (e.g. considering it as an extension of the soil profile) that are capable of simulating transfers of water across drainage boundaries. Given that hydrologic models are likely to be used with increasing frequency in the near future as a tool for managing the hydrologic effects of land use (e.g., forestry), predictions of such models will only be credible if they can accurately represent all of the dominant processes. 85 R E F E R E N C E S Anderson, S.P., Dietrich, W.E. , Torres, R., Montgomery, D.R., and Lbague, K. 1997. Concentration-discharge relationships in runoff from a steep unchanneled catchment. Water Resources Research. 33: 211-225. Beschta, L.R. 1978. Long-term patterns of sediment production following road construction and logging in the Oregon Coast Range. Water Resources Research. 14: 1011-1016. Braun, L . N . 1980. Scale dependence of factors controlling the release of water from snow and ice storage. M.Sc. thesis, University of British Columbia: 142. . Burns, D.A., and McDonnell, J.J. 1998. Effects of a beaver pond on runoff processes: comparison of two headwater catchments. Journal of Hydrology. 205: 248-264. Burns, D.A., McDonnell, J.J., Hooper, R.P., Peters, N.E., Freer, J.E., Kendall, C , and Beven, K. 2001. Quantifying contributions to storm runoff through end-member mixing analysis and hydrologic measurements at the Panola Mountain Research Watershed (Georgia, USA). Hydrological Processes. 15: 1903-1924. Buttle, J.M., and Sami, K. 1992. Testing the groundwater ridging hypothesis of streamflow generation during snowmelt in a forested catchment. Journal of Hydrology. 135: 53-72. Buttle, J.M. 1994. Isotope hydrograph separations and rapid delivery of pre-event water from drainage basins. Progress in Physical Geography. 18(1): 16-41. Buttle, J.M., and Peters, D.L. 1997. Inferring hydrological processes in a temperate basin using isotopic and geochemical hydrograph separation: a re-evaluation. Hydrological Processes. 11: 557-573. Christophersen, N. , Neal, C , Hooper, R.P., Vogt, R.D., and Andersen, S. 1990. Modelling streamwater chemistry as a mixture of soilwater end-members: a step towards second-generation acidification models. Journal of Hydrology. 116: 307-320. Colbeck, S.C. 1979. Water flow through heterogeneous snow. Journal of Cold Regions Science and Technology. 1: 37-45. De Vries, J., Chow, T.L. 1978. Hydrologic behavior of a forested mountain soil in Coastal British Columbia. Water Resources Research. 14(5); 935-942. Dewalle, D.R., Swistoch, B.R., and Sharpe, W.E. 1988. Three-component tracer model for stormflow on a small Appalachian forested catchment. Journal of Hydrology. 104: 301-310. Dunne, T., and Black, R.D. 1970a. An experimental investigation of runoff production in permeable soils. Water Resources Research. 6(2): 478-490. Dunne, T., and Black, R.D. 1970b. Partial area contributions to storm runoff in a small New England watershed. Water Resources Research. 6(5): 1296-1311. 86 \ Dunne, T., and Black, R.D. 1971. Runoff processes during snowmelt. Water Resources Research. 7(5): 1160-1172. Forest Practices Code. 1999. Appendix 1: Identification of stream orders, sub-basins, and the point of interest. Watershed Assessment Procedures Guidebook, 2nd Edition. British Columbia Ministry of Forests. Forest Sciences Department, B C Ministry of Forests. Unpublished data for Upper Redfish Creek. Supplied by Dr. P. Jordan, Nelson Forest Region, February, 2003. Freeze, R A . 1972. Role of subsurface flow in generating surface runoff 2: Upslope contributing areas. Water Resources Research. 8(5): 1272-1283. Freeze, R A . 1974. Streamflow generation. Reviews of Geophysics and Space Physics. 12(4): 627-647. Geological Survey of Canada, Department of Mines. 1932. Slocan Sheet, Kootenay District, British Columbia [Map 272 A]. W.H. Collins, Director. Scale 1:63360. Hair, J.F. Jr, Anderson, R.E., Tatham, R.L. and Black, W.C. 1995. Multivariate Data Analysis, Fourth Edition. Prentice Hall, Englewood Cliffs, 745 pp. Hall, F.R. 1971. Dissolved solids-discharge relationships 2: Applications to field data. Water Resources Research. 7:591-601. Hangen, E. , Lindenlaub, M . , Leibundgut, Ch., and von Wilpert, K. 2001. Investigating mechanisms of stormflow generation by natural tracers and hyrometric data: a small catchment study in the Black Forest, Germany. Hydrological Processes. 15: 183-199. Harr, R.D., and McCorison, F .M. 1979. Initial effects of clearcut logging on size and timing of peak flows in a small watershed in western Oregon. Water Resources Research. 15: 90-94. Harr, R.D., Levno, A. , and Mersereau, R. 1982. Streamflow changes after logging 130-year-old Douglas Fir in two small watersheds. Water Resources Research. 18: 637-644. Hazlett, P.W., Semkin, R.G., and Beall, F.D. 2001. Hydrologic pathways during snowmelt in first-order stream basins at the Turkey Lakes Watershed. Ecosystems. 4: 527-535. Hetherington, E.D. 1988. Hydrology and logging in the Carnation Creek Watershed - what have we learned? Applying 15 Years of Carnation Creek Results: Proceedings of the Workship. Chamberlin, T. W. ed. Pacific Biological Station, Nanaimo, British Columbia: 11-15. Hewlett, J.D., and Hibbert, A.R. 1967. Factors affecting the response of small watersheds to precipitation in humid areas. Forest Hydrology. Sopper, W.E. , and Lull, H.W., eds., New York, Pergamon Press: 275-290. Hoeg, S., Uhlenbrook, S., and Leibundgut, Ch. 2000. Hydrograph separation in a mountainous catchment - combining hydrochemical and isotopic tracers. Hydrological Processes. 14: 1199-1216. 87 Hooper, R.P., Christophersen, N., and Peters, N.E. 1990. Modelling streamwater chemistry as a mixture of soilwater end-members: an application to the Panola Mountain catchment, Georgia, U.S.A. Journal of Hydrology. 116: 321-343. Hooper, R.P., Aulenbach, B.T;, Burns, D.A., McDonnell, J., Freer, J., Kendall, C , Beven, K. 1998. Riparian control of stream-water chemistry: implications for hydrochemical basin models. Hydrology, Water Resources and Ecology in Headwaters (Proceedings of the HeadWater'98 Conference, Meran/Merano, Italy, April 1998), IAHS Publ. No. 248: 451-458. Hutchinson, D.G., and Moore, R.D. 2000. Throughflow variability on a forested hillslope underlain by compacted glacial till. Hydrological Processes. 14: 1751-1766. Jordan, R.P. 1978. The snowmelt hydrology of a small alpine watershed. M.Sc. thesis, University of British Columbia: 201pp. Jordan, R.P. 1983. Meltwater movement in a deep snowpack 1: Field observations. Water Resources Research. 19: 971-978. Jordan, R.P. 1983. Meltwater movement in a deep snowpack 2: Simulation model. Water Resources Research. 19: 979-985. Jungen, J.R. 1980. Soil Resources of the Nelson Map Area (82F). British Columbia Soil Survey Report No. 28, British Columbia Ministry of Environment. Kelowna. 1980. Kane, D.L. , and Stein, J. 1983. Water movement into seasonally frozen soils. Water Resources Research. 19: 1547-1557. Katsuyama, M . , Ohte, N. , and Kobashi, S. 2001. A three-component end-member analysis of streamwater hydrochemistry in a small Japanese forested headwater catchment. Hydrological Processes. 15: 249-260. Katsuyama, M . , and Ohte, N. 2002. Determining the sources of stormflow from the fluorescence properties of dissolved organic carbon in a forested headwater catchment. Journal of Hydrology. 268:192-202. Kendall, K.A. , Shanley, J.B., and McDonnell, J.J. 1999. A hydrometric and geochemical approach to test the transmissivity feedback hypothesis during snowmelt. Journal of Hydrology. 219: 188-205. Kim, H.J., Sidle, R.C., Moore, R.D., and Hudson, R. Throughflow variability during snowmelt in a forested mountain catchment, coastal British Columbia, Canada. Hydrological Processes. In press. Kobayashi, D. 1985. Separation of the snowmelt hydrograph by stream temperatures. Journal of Hydrology. 76: 155-162. Kobayashi, D. 1986. Separation of a snowmelt hydrograph by stream conductance. Journal of Hydrology. 84: 157-165. Kobayashi, D., Ishii, Y . , and kodama, Y. 1999. Stream temperature, specific conductance and runoff processes in mountain watersheds. Hydrological Processes. 13: 865-876. 88 Ladouche, B., Probst, A. , Viville, D., Idir, S., Baque, D., Loubet, M . , Probst, J . -L . , and Bariac, T. 2001. Hydrograph separation using isotopic, chemical and hydrological approaches (Strengbach catchment, France). Journal of Hydrology. 242: 255-274. Laudon, H. , Slaymaker, O. 1997. Hydrograph separation using stable isotopes, silica and electival conductivity: an alpine example. Journal of Hydrology. 201: 82-101. Laudon, H. , Hemond, H.F., Krouse, R., and Bishop, K. 2002. Oxygen 18 fractionation during snowmelt: Implications for spring flood hydrograph separation. Water Resources Research. 38 (11): 1-10. Luce, C.H. , and Wemple, B.C. 2001. Introduction to special issue on hydrological and geomorphic effects of forest roads. Earth Surface Processes and Landforms. 26: 111-113. Marquis, J.P., and Golding, D.L. 1983. Hydrograph separation using natural isotope and conductance methodologies in the West Kootenay areas of British Columbia. Report to the British Columbia Ministry of Forests [contract 65-3607 between British Columbia Ministry of Forests and the University of British Columbia. Dr. D.L. Golding supervisor, Faculty of Forestry]. Marsh, P., and Woo, M.K. 1984. Wetting front advance and freezing of meltwater within a snow cover, 1: Observations in the Canadian Arctic. Water Resources Research. 20: 1853-1864. Marsh, P., and Woo, M.K. (1985). Meltwater movement in natural heterogeneous snow covers. Water Resources Research. 21:1710-1716. Maule, C P . , and Stein, J. 1990. Hydrologic flow path definition and partitioning of spring meltwater. Water Resources Research. 26(12): 2959-2970. Megahan, W.F. 1972. Subsurface flow interception by a logging road in mountains of central Idaho. Proceedings of the National Symposium on Watersheds in Transition. Csallany, S.C., McLaughlin, T .G. , and W.D. Striffler eds. Fort Codings, CO. June 19-22, 1972: 350-356 Melton, M A . 1965. The geomorphic and paleoclimatic significance of alluvial deposits in southern Arizona. Journal of Geology. 73: 1-38. Ministry of Enenrgy, Mines and Petroleum, 1981. Mineral Resources of the Slocan: Background Report for the Slocan Valley Planning Program. Resource Data and Analysis, Geology Branch. Prepared for the Kootenay Resource Management Committee and the Regional District of Central Kootenay. 55 pp. [+maps]. Montgomery, D.R., Dietrich, W.E., Torres, R., Anderson, S.P., Heffner, J.T., and Loague, K. 1997. Hydrologic response of a steep, unchanneled valley to natural and applied rainfall. Water Resources Research. 33 (1): 91-109. Moore, R.D. 1989. Tracing runoff sources with Deuterium and Oxygen-18 durning spring melt in a headwater catchment, southern Laurentians, Quebec. Journal of Hydrology. 112: 135-148. Mosley, M.P. 1979. Streamflow generation in a forested watershed, New Zealand. Water Resources Research. 15(4): 795-806. 89 McGlynn, B.L. , McDonnell, J.J., Shanley, J.B., Kendall, C. 1999. Riparian zone flowpath dynamics during snowmelt in a small headwater catchment. Journal of Hydrology. 222: 75-92. Noguchi, S., Tsuboyama, Y. , Sidle, R.C., Hosoda, I. 2001. Subsurface runoff characteristics from a forest hillslope soil profde including maeropores, Hitachi Ohta, Japan. Hydrological Processes. 15:2131-2149. Nyberg, L . , Stahli, M . , Mellander, P-E., and Bishop, K. H. 2001. Soil frost effects on soil water and runoff dynamics along a boreal forest transect: 1. Field investigations. Hydrological Processes. 15: 909-926. Pearce, A.J. , Stewart, M.K. , and Sklash, M . G . 1986. Storm runoff generation in humid headwater catchments 1. Where does the water come from? Water Resources Research. 22(8): 1263-1272. Price, A . G . , and Hendrie, L .K. 1983. Water Motion in a deciduous forest during snowmelt. Journal of Hydrology. 64: 339-356. Roberge, J., Plamondon, A.P. 1987. Snowmelt runoff pathways in a boreal forest hillslope, the role of pipe throughflow. Journal of Hydrology. 95: 39-54. Shanley, J.B., Kendall, C , Smith, T.E. , Wolock, D .M. , and McDonnell, J.J. 2002. Controls on old and new water contributions to stream flow at some nested catchments in Vermont, USA. Hydrological Processes. 16: 589-609. Shanley, J.B., and Chalmers, A. 1999. The effect of frozen soil on snowmelt runoff at Steepers River, Vermont. Hydrological Processes. 13: 1843-1857. Sidle, R.C., Tsuboyama, Y. , Noguchi, S., Hosoda, I., Fujieda, M . , Shimizu, T. 2000. Stormflow generation in steep forested headwaters: a linked hydrogeomorphic paradigm. Hydrological Processes. 14: 369-385. Silva Forest Foundation. 1996. Ecosystem-based Landscape Plan for the Slocan River Watershed. June 1996. Sklash, M . G , and Farvolden, R.N. 1979. The role of groundwater in storm runoff. Journal of Hydrology. 43: 45-65. Sklash, M . G , Stewart, M.K. , and Pearce, A.J. 1986. Storm runoff generation in humid headwater catchments 2. A case study of hillslope and low-order stream response. Water Resources Research. 22(8): 1273-1282. Spence, C , and Woo, M.K. 2003. Hydrology of subarctic Canadian shield: Soil-filled valleys. Journal of Hydrology. 279: 151-166. Stahli, M . , Jansson, P-E, and L - C Lundin. 1996. Preferential water flow in a frozen soil - a two-domain model approach. Hydrological Processes. 10: 1305-1316. 90 Sueker, J.K., Ryan, J.N., Kendall, C , and Jarrett, R.D. 2000. Determination of hydrologic pathways during snowmelt for alpine/subalpine basins, Rocky Mountain National Park, Colorado. Water Resources Research. 36(1): 63-75. Thunholm, B., Lundin, L - C , and S. Lindell. 1989. Infiltration into a frozen heavy clay soil. Nordic Hydrology. 20: 153-166. Troendle, C.A., and King, R.M. 1985. The effect of timber harvest on the Fool Creek watershed, 30 years later. Water Resources Research. 21:1915-1922. Troendle, C.A., and Stednick, J.D. 1999. "Effects of basin scale timber harvest on water yield and peak streamflow," by Timothy A. Burton [review]. Journal of the American Water Resources Association. 35: 177-181. Turton, D.J., Haan, C.T., and Miller, E L . 1992. Subsurface flow responses of a small forested catchment in the Ouachita Mountains. Hydrological Processes. 6: 111-125. Van Haveren, B.P. 1988. A re-evaluation of the Wagon Wheel Gap forest watershed experiment. Forest Science. 14: 208-214. Ward, R.C., and Robinson, M . 2000. Principles of Hydrology: 4th Edition. McGraw Hill Publishing Company, London: 450 pp. Wels, C , Cornett, R.J., and Lazerete, B.D. 1991. Hydrograph separation: comparison of geochemical and isotopic tracers. Journal of Hydrology. 122: 253-274. Wemple, B.C., Jones, J.A., and Grant, G.E. 1996. Channel network extension by logging roads in two basins, western Cascades, Oregon. Water Resources Bulletin. 32: 1195-1207. Zeman, L.J. , and Slaymaker, O. 1975. Hydrochemical analysis to discriminate variable runoff source areas in an alpine basin. Arctic and Alpine Research. 7: 341-351. 91 A P P E N D I X A F R E S H E T H Y D R O G R A P H S 1 \ A P P E N D I X B P I E Z O M E T E R W A T E R L E V E L S A N D C R E S T D A T A E 75 > CD » CO O Q . CD Q 10 -30 --70 --110 -150 L P 2 (1270 m) 10 -30 -70 -110 -150 U P3S(1135m ) a 10 -30 -70 -110 -150 10 -30 -70 -110 i r P1 (1275 m) + <*• 8 6 Q 6 P 3 D (1135 m) o o L P 5 (1060 m) • -h-P 6 D (935 P 6 S (935 m) - 1 5 0 " — 1-Jan _L &=-e— 8 -_L UP 4 UP -J Ho 30 70 -110 1-150 NSRIP -110 -150 UP -\ N S R I f l 10 -30 -70 -110 -150 N S R I f l 1-Feb 1-Mar 1-Apr 1-May 1-Jun 1-Jul O WL Crest 96 10 -30 -70 -110 -150 U P9D (825 m) 10 -30 -70 -110 -150 U P9S (825 m) \- P.10 (825 m) 10 -30 -70 -110 -150 -70 --110 --150 1 r P8D (945 m) + + U P11 (825m) P12D(825 m) P12S (735 m) _L _l_ _L 1-Jan 1-Feb 1-Mar 1-Apr 1-May UP A _L NSRIB NSRIR NSRIW N S H l U 10 NSRI?30 NSRI 10 -30 70 110 -150 10 -30 -70 110 -150 -70 -110 -150 1-Jun 1-Jul O . A WL Crest 97 U P14 (735 m) U P15D(650m) r- P15S(650m ~ 1 T P13(735m) -4r " 6- 6 + U P17(650m) h " |- P18D(680m) P18S(680m) _L 1-Jan 1-Feb 1-Mar 1-Apr 1-May T" NSRIP-J NSRIP-4 _L 10 30 -70 110 -150 NSRlff RW. BIFM NSRW-30 NSRIFt 10 30 -70 110 150 10 -70 -110 -150 1-Jun 1-Jul O A WL Crest 98 \ E > (0 CO a. Q 10 -30 -70 -110 -150 1 " U P20 (680 m) 10 -30 -70 -110 -150 U P21 (650 m) 10 -30 -70 -110 1 r P19(680m) L. P22 (650 m) + PA1 (940 m) -150 1— 1-Jan T NSHIL-J N S H l U NSHIL-NSHIL-4 NSRIP-J 10 30 70 -110 -150 10 -30 -70 -110 -150 O WL A Crest 1-Feb 1-Mar 1-Apr 1-May 1-Jun 1-Jul 99 \ APPENDIX C OBSERVATIONS OF SATURATION OVERLAND FLOW What follows is a detailed description of overland flow observed throughout the field season. Occurrences of SOF are numbered according to the order in which they were observed and can be cross referenced with both Figure 3.10 and Table 3.4. 1. On a broad forested hillslope where overland flow began as a seep in the cavity left by the roots of a wind-thrown tree and continued downslope for a number of meters. It was observed on Feb 25, 2002 and its location was estimated to be just below 900 m and immediately north of the Thickett Creek catchment1. 2. In the near-stream riparian area surrounding the lower transect of piezometers on Gurn Spring (P15s, P15d, P16, and P17) on Mar 31, 2002. This was primarily due to the formation of a short, intermittent channel on the flat valley bottom that roughly paralleled the primary channel of Gurn Spring. Commencing as heavy seepage from a large macropore in the ground upslope of P17, it ran downslope for 10-20 meters, passing within roughly 2 m of PI 7 before rejoining the main channel downstream of the piezometers. Diffuse surface saturation and overland flow were observed between this new secondary channel and the main channel This was traced upslope to an area of localized spillage at a bend in the main channel. In addition to the flows observed prior to the beginning of Apri l (#1-2, above), SOF was observed between Apr 9-10, 2002 in the following locations: 3. On Apr 9, 2002 in two small valley features running parallel to each other immediately off a switchback in the road at 825 m and just north of the Thickett Creek catchment. Channelized flows were observed in each of these micro-valley features, both generated by seepage immediately upslope. One of these micro-channels passed adjacent to P9s and P9d so that those piezometers were within centimetres of the surface water. 1 Locations of observed SOF were not surveyed precisely and are derived from field approximations using local topographic features, roads, and instrument locations (e.g. surveyed piezometers and climate stations). 100 4. On the broad forested hillslope within the Thickett Creek catchment at approximately 900 m where a micro-channel had formed by seepage in an area of short (1-2 m) convergent hillslopes. SOF was first observed in this location on Apr 10, 2002. 5. Also on the broad forested hillslope within the Thickett Creek catchment at approximately 925 m where a seep occurred approximately 10-25 m downslope of the piezometer A l causing downslope surface saturation. SOF was first observed here on Apr 10, 2002. The most widespread occurrence of SOF was observed immediately following peak streamflows during the period Apr 15-18, 2002. During this period, SOF persisted in each of the locations discussed above. New occurrences of SOF were observed in the following locations: 6. On Apr 15, 2002 intersecting the readjust below 950 m where a small channel of overland flow emerged from a shallow gully feature vegetated by herbaceous shrubs and brush. 7. On Apr 15, 2002 above the readjust below 775 m where a saturated surface conditions had developed upslope of the road-cut. This was judged to be located just within the TH catchment. 8. On Apr 15, 2002 below the road and almost directly across from #7 (above) and upslope of the weir on Thickett Creek where large quantities of SOF emerged from multiple seeps. From these, overland flow continued in a series of intermittent seeps and saturated areas down to the permanent channel of Thickett Creek. 9. On Apr 16, 2002 at approximately 910 m and upslope of #4. SOF occurred here as an intermittent micro-channel in an area of short (1-2 m), low angled convergent hillslopes. SOF observed between 900 m and 925m appeared to be connected by subsurface flows linking the observed seeps and micro-channels (i.e. linking #4, 5, and 8). 10. On Apr 16, 2002 in the forest at approximately 875 m where seepage emerged and continued a short ways downslope before disappearing. 11. On Apr 16, 2002 immediately adjacent to the road switchback at 825 m, where flows from the two micro-channels observed on Apr 9-10, 2002 (#3) had extended 10-20 m further upslope. It is possible that the SOF observed on Feb 25, 2002 actually occurred within the Thickett Creek catchment. 101 12. On Apr 16, 2002 just below #3 where the two micro-channels converged downslope ofP9sandP9d. 13. On Apr 17, 2002 at approximately 1175 m within the Gurn Brook catchment upslope of P3s and P3d where seepage commenced on a narrow terrace feature located along a forested north-facing hillslope. From here seepage continued downslope along the terrace for approximately 10-25 m. 14. On Apr 18, 2002 within the T H catchment and intersecting the readjust below 825 m where a small channelized feature emerged from the forest. The micro-channel was traced 10-20 m into the forest to its source at an area of small convergent hillslopes producing localized seepage. 15. On Apr 18, 2002 emerging as a large seep from directly beneath the base of a tree just below the road at approximately 800 m within the T H catchment. Flow continued into the forest for 10-25 m before ponding somewhere upslope of the road just above 825 m. Though not as widespread as the week prior, SOF persisted in a number of locations during the period Apr 22-24, 2002. These were: #1,2, 10, 12, 13, and 14 (above). In addition, SOF was observed in significantly diminished quantities at #8 and 15 (above). SOF was mapped a final time on May 15-16, 2002, at which time overland flow persisted only in diminished quantities at #8 and 15 (above). New SOF was observed in the following location: 16. On May 15, 2002 in the GB catchment and emerging from a forested gully with steep sidewalls upslope of the readjust above 1125 m. Flows were observed crossing the road and continuing into the forest below. 102 APPENDIX D ISOTOPIC CONCENTRATIONS IN M E L T AND RAINWATER \ Q CN a. < CL. < < Q. a. D . a. a. < 0 • CN k-CN ca s so CN cs 2. <n CN C3 CN o oo C3 2 > W _ . 00 e n i n s© SO i • • i n SO t-; t-; '—i oo so so 00 00 m O CN Os OO OS SO 00 Os S — CN OS OS d CN SO od SO OS CN •—' i n OS O od od od 00 i • ,i 1 i r~ m r~ cn r- CN od OS sd i—< m oo m P~ Os OS O Os CN >o CN d CN d CN Os d CN 00 Os o CN o CN i n © CN CN m C l -^4 Os 00 O i n Os d CN CN CN CN CN m o m o r- r» c> SO CN CN 1—< o 1—1 *-" o CN 00 CN CN Os m Os CN m oo so i—1 m m SO OO Os — J J . J J 1) 6 cu ct> C-t—4 I o c cn 6 o ti—i T3 (D • •*-» o "o o I-I <u •4-* 03 'J3 T3 CO >-a B 00 I CO * t o CN e 3 CD •*-» C3 Q 3 < > O C O > s CN OS m Os os CN m o r - so CN O m m Os m v-i m T T CN m Os 00 SO >n so oo as —' _ l J _1 _) J >—1 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items