Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The downstream effects of salt application on Horstman Glacier, Whistler, British Columbia Teichrob, Nicolas Daniel 2010

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

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

Item Metadata


24-ubc_2010_spring_teichrob_nicolas.pdf [ 5.27MB ]
JSON: 24-1.0068938.json
JSON-LD: 24-1.0068938-ld.json
RDF/XML (Pretty): 24-1.0068938-rdf.xml
RDF/JSON: 24-1.0068938-rdf.json
Turtle: 24-1.0068938-turtle.txt
N-Triples: 24-1.0068938-rdf-ntriples.txt
Original Record: 24-1.0068938-source.json
Full Text

Full Text

The Downstream Effects of Salt Application on Horstman Glacer, Whistler, British Columbia by Nicolas Daniel Teichrob B.Sc. Hons, The University of British Columbia, 2006 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in The Faculty of Graduate Studies (Geography) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) January 2010 c© Nicolas Daniel Teichrob 2009 Abstract Skiing and snowboarding occur on the seasonal snow of Blackcomb Moun- tain, BC, which includes two glaciers: Blackcomb Glacier and Horstman Glacier. Ski operations involve the application of salt (NaCl) to the glacier and redistribution of snow on Horstman Glacier to allow for additional skiing and snowboarding during the months of June and July. Although European studies have documented that salting ski pistes can have negative effects on the environment, no research has been conducted to study the continuous application of salt onto a glacier and the effects on the downstream aquatic environment. In this study we examined the temporal changes in chloride concentration in Horstman Creek. Using neighbouring Blackcomb Glacier and Blackcomb Creek as a reference system, automatic samplers were used to collect water samples for two periods of sixteen days each during the sum- mer of 2008. The concentration of chloride in Horstman Creek was found to be greatly elevated compared to regional background values and those of Blackcomb Creek during both sampling periods. The pattern of chloride concentration suggests that some of the NaCl applied to the glacier makes its way downstream as initial snowmelt runoff and some is stored within the glacier and released at a later time. However, despite the fact that 90 020 kg of NaCl was applied during summer 2008, the elevated concentrations did not reach a level of environmental concern during the study period. ii Table of Contents Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . vii 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Use of salt in summer ski operations . . . . . . . . . . . . . . 2 1.2 Hydrology and solute dynamics in glacierized catchments . . 3 1.2.1 Glacial and proglacial flow paths . . . . . . . . . . . . 3 1.2.2 Proglacial solute dynamics in the absence of salting . 4 1.3 Objectives and hypotheses . . . . . . . . . . . . . . . . . . . 6 1.4 Organization of the thesis . . . . . . . . . . . . . . . . . . . . 8 2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.1 Regional context . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2 Study sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.3 Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.3.1 Continuous stream monitoring . . . . . . . . . . . . . 12 2.3.2 Streamflow measurement . . . . . . . . . . . . . . . . 14 2.3.3 Water sampling . . . . . . . . . . . . . . . . . . . . . 14 2.3.4 Snowpack information . . . . . . . . . . . . . . . . . . 15 2.3.5 Bedrock and till mineralogy . . . . . . . . . . . . . . 17 2.3.6 Climate data . . . . . . . . . . . . . . . . . . . . . . . 17 2.3.7 Salt composition and loading . . . . . . . . . . . . . . 17 2.4 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.5 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 iii Table of Contents 3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.1 Overview of the study period . . . . . . . . . . . . . . . . . . 21 3.2 Bedrock and till mineralogy . . . . . . . . . . . . . . . . . . 22 3.3 Stream temperature and electrical conductivity profiles down- stream of Horstman Glacier . . . . . . . . . . . . . . . . . . . 24 3.4 Chloride variations in grab samples . . . . . . . . . . . . . . 26 3.5 Chloride concentrations in streamwater . . . . . . . . . . . . 26 3.6 Relations between chloride and discharge . . . . . . . . . . . 31 3.7 Statistical modeling of chloride concentrations . . . . . . . . 32 4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 4.1 Sources and patterns of chloride in streamwater . . . . . . . 35 4.2 Implications for downstream water quality and summer ski operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 5.1 Summary of key findings . . . . . . . . . . . . . . . . . . . . 39 5.2 Recommendations for future study . . . . . . . . . . . . . . . 40 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 iv List of Tables 1.1 Background chloride concentrations in the Coast Mountains . 5 2.1 Characteristics of Horstman and Blackcomb glaciers . . . . . 12 2.2 Characteristics of Horstman Creek and Blackcomb Creek at the gauging stations . . . . . . . . . . . . . . . . . . . . . . . 12 2.3 Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.4 Instrumentation installed on Blackcomb Creek and Horstman Creek stream stations. . . . . . . . . . . . . . . . . . . . . . . 13 2.5 Chemical composition of the salt used by WB . . . . . . . . . 18 3.1 Whole rock mineralogy of bedrock and till samples . . . . . . 24 3.2 Chloride concentrations for the grab samples of glacier and bedrock water . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.3 ”Best” statistical models for Cl concentrations . . . . . . . . 33 3.4 ”Best” statistical models for logCl concentrations . . . . . . . 34 4.1 British Columbia environmental regulations . . . . . . . . . . 38 v List of Figures 1.1 Conceptual diagram . . . . . . . . . . . . . . . . . . . . . . . 6 1.2 Horstman Creek chloride flow chart . . . . . . . . . . . . . . . 8 2.1 Site location map. . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2 Climate normals for Whistler Mountain. . . . . . . . . . . . . 11 2.3 Automated camera’s view of Horstman Glacier . . . . . . . . 16 3.1 Study period climatic conditions . . . . . . . . . . . . . . . . 22 3.2 Time series plots of Horstman and Blackcomb discharge, snow cover, air temperature and precipitation. . . . . . . . . . . . . 23 3.3 Horstman Creek electrical conductivity and temperature dis- tance downstream of Horstman Glacier . . . . . . . . . . . . . 25 3.4 Horstman Creek and Blackcomb Creek chloride concentra- tions for the entire study period . . . . . . . . . . . . . . . . . 27 3.5 Horstman Creek and Blackcomb Creek chloride concentra- tions for July 23, 2008 to August 8, 2008. . . . . . . . . . . . 28 3.6 Horstman Creek and Blackcomb Creek chloride concentra- tions for August 14, 2008 to August 30, 2008. . . . . . . . . . 29 3.7 Chloride, discharge and cumulative discharge for Horstman Creek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.8 Blackcomb Creek chloride and discharge separated into the two sampling periods. . . . . . . . . . . . . . . . . . . . . . . 31 3.9 Horstman Creek chloride and discharge separated into the two sampling periods. . . . . . . . . . . . . . . . . . . . . . . 32 vi Acknowledgements The work presented in this thesis could not have been done without the help of others. I was able to focus solely on my research in thanks to NSERC which awarded me with a PGS–M scholarship. The Department of Geogra- phy also helped me along with a couple of scholarships and funding from Dr. Markus Weiler. Dr. Weiler believed in this project when I first presented the idea to him and I am thankful for his participation and guidance of my work in its early stages. I would like to acknowledge the supervision in the later stages of my research given by Dr. Dan Moore. Dr. Moore provided me with countless ideas to further push the analysis of the data and helped make this project as scientifically sound as possible. Thank you to Dr. Mark Johnson for providing me with the equipment and knowledge to perform the water sample analysis in his laboratory. I also owe a big thanks to Chris Loewen who planted the seed in my head to do a Master’s degree on this subject back in 2005. I had many research assistants help dig holes, survey glaciers, gauge water, and freeze our hands installing stilling wells. David Teichrob, Ryan Loewen and Marcio Penner helped ski and survey the glaciers, Andy Orr provided geologic guidance and helped keep us safe in the field, Mike Van Der Laan and Natalie Stafl helped dig holes in the snow and collect streamflow data, John Richards taught me how to do salt dilution gauging with a high level of accuracy and precision in mind, and Kory Dumas scored some pre-season glacier powder. Arthur DeJong, of Whistler Blackcomb, allowed us to access our field sites with ease and offered his knowledge on the current state of the Horstman Glacier. I very grateful of the support given by my wife Heather. Heather has sup- ported me throughout this degree and she came out for some late night stream gauging missions. Without her love and encouragement, I would not have been able to develop my photographic career while finishing this thesis. Finally, I want to thank all of my friends and family for keeping me smiling and having fun throughout the last two years. All of the powder days, surf trips and random adventures were crucial to maintaining a balanced life. vii Chapter 1 Introduction Skiing and snowboarding occur on the seasonal snow of Blackcomb Moun- tain, which includes two glaciers: Blackcomb Glacier and Horstman Glacier. Ski operations involve the application of salt (NaCl) and redistribution of snow on Horstman Glacier, allowing for additional skiing and snowboarding during the months of June and July. From this point forward, the word ’skiing’ will refer to both skiing and snowboarding and ’salt’ will refer to the NaCl applied to the glacier snowpack. Similar summer ski operations occur on a glacier at Mount Hood, Oregon, and are also planned for the proposed Jumbo Glacier Resort, BC (Pheidias Project Management Corp., 2007)). This proposed resort is highly controversial, with one point of dispute being the environmental effects of salting (Greenwood, 2004). There do not appear to have been any published studies of the environ- mental effects of salting glaciers for summer skiing in North America. The Swiss Federal Institute for Snow and Avalanche Research (SLF) studied the chemical preparation of ski pistes in Switzerland (Schneebeli et al., 2007). They found that water in streams fed by salted glaciers had much higher concentrations of salts than background levels measured in neighbouring streams. Elevated salt levels were detected more than 15 years after salting ended. In addition, the salting of ski pistes has been studied in elsewhere in Europe, but these studies were conducted privately and the results have not been published in an accessible form. One major difference between the European salting practices and those of Whistler–Blackcomb is that in Europe they used salts other than NaCl, such as nitrates and phosphates. These types of salts can be more detrimental to the natural environment than NaCl as plants and animals are more sensitive to low levels of nitrates and phosphates than chloride in the environment. As a result, the practice of salting glaciers is now restricted in Europe, as many European ski resorts have agreed to use salt in the preparation of ski pistes only when necessary for building World Cup ski race courses. This is still a grey area in practice as different resorts have different ideas of what is ”necessary.” The glaciers of British Columbia have been in an overall recession since the Little Ice Age maximum, which occurred sometime in the late 1800s 1 1.1. Use of salt in summer ski operations (Moore and Demuth, 2001). It has been suggested that Blackcomb Glacier will be at half of its present mass by 2040 and be completely gone by 2090 (Shea, 2006). With neighbouring Horstman Glacier in a similar state, the skiing and snowboarding in the summer months may become more and more limited on Blackcomb mountain. The economic benefits of operating a ski area in June and July are high, and if this is to be maintained, a scien- tific understanding of how the present mountain operations are affecting the glacier must be developed. Furthermore, Whistler–Blackcomb seeks to maintain environmental sustainability and the findings from this research may have implications for the present glacier salting and monitoring prac- tices. The remainder of this chapter will describe how and why salt is used for summer ski operations on glaciers and how solute concentrations vary in proglacial streams. The chapter will end with a statement of research objectives and proposed hypotheses regarding the anticipated chloride con- centrations in Horstman Creek downstream of Horstman Glacier. 1.1 Use of salt in summer ski operations The goal of the mountain operations is to provide a firm surface for skiers to enjoy. For purposes of ski racing training and terrain park jumps and fea- tures, the firmer the snow, the better. During summer evenings on Horstman Glacier, cooling by emission of longwave radiation decreases the surface tem- perature to near 0◦ C. The dissolution of salt is endothermic, and thus acts in concert with noctural energy losses to promote cooling and freezing of the snow–ice–water mixture when applied to the surface of a glacier or snow- pack. The addition of salt also increases the specific heat of the snow. As a result, during the morning, the snow warms back up to the melting point more slowly, prolonging the period of firm snow suitable for skiing. By mid–afternoon, the liquid–to–solid ratio is often too high and the snow too soft for skiing. The process of applying salt to the surface in the evenings and skiing during the mornings and early afternoons continues for the du- ration of the summer glacier operations, which is usually between six and eight weeks. 2 1.2. Hydrology and solute dynamics in glacierized catchments 1.2 Hydrology and solute dynamics in glacierized catchments Proglacial solute concentrations vary in response to the shifting contribu- tions of streamflow originating from different sources, such as glacier dis- charge, lateral flow from hillslopes and discharge of deeper groundwater (Anderson, 2007). This section first provides a review of the pathways by which rain and meltwater flow over or through a glacier, followed by a consideration of processes influencing hydrology and solute concentrations downstream of the glacier outlet. Proglacial solute dynamics in the absence of salting are then reviewed, with a specific focus on background concentra- tions of chloride in the southern Coast Mountains. The information from this review provides the basis for the formulation of hypotheses regarding the influence of salt application for summer skiing on downstream chloride concentrations (Section 1.3). 1.2.1 Glacial and proglacial flow paths In a snow–covered glacial system, water and solute movement can occur via flow through the snowpack, along the snowpack–glacier interface, englacially within the ice and sub–glacially at the interface between the ice and glacier bed. Solute concentrations in a glacier–fed basin vary in response to contin- uing changes in snow heterogeneity, water routing, storage in the snowpack and glacier, summer precipitation and glacier mass balance (Theakstone and Knudsen, 1996; Singh et al., 1997; Moore and Demuth, 2001). Early in the melt season the snowpack has the most significant effect on meltwater composition (Sharp et al., 1995; Marsh, 1999; Harper and Bradford, 2003; Campbell et al., 2006). Liquid water exiting a snowpack–glacier system can originate either from snowmelt or rainfall. As water flows through snow, it alters the structure and stability of the snowpack, thus increasing the spatial variability and heterogeneity of the snowpack (Waldner et al., 2004). Ice layers within the snowpack can change the course of water flow by causing ponding, diversion of flow along sloping ice layers, detention of flowing water within the ice, and slowing the downward movement of water (Furbish, 1988; Singh et al., 1997; Campbell et al., 2006). As the snowpack evolves throughout the melt season, the permeability of the snow changes, driving changes in the percolation rate of water through the snow (Colbeck and Davidson, 1972). In addition, long wave radiation cooling causes refreezing of the snowpack during cold and clear nights and thus can delay the release of meltwater from the snowpack 3 1.2. Hydrology and solute dynamics in glacierized catchments the following day (Marsh, 1999). The supraglacial drainage system changes throughout the ablation season as the volume of the snowpack decreases and air temperature increases. If water makes its way through the entire snowpack, it will be intercepted by the snowpack–glacial ice interface. Water that reaches this surface will either flow downslope along the surface of the ice or it will enter the glacier via moulins, crevasses, or micro–structures in the ice (Hooke, 1989; Harper and Humphrey, 1995; Schuler, 2002). These englacial and subglacial flow structures, fed by the overlying snowpack, store and transmit water and solutes. Flow paths and residence times for englacial and subglacial flow vary through the melt season in response to changes in the glacial drainage system (Sharp et al., 1995). In addition to inputs of water from above, the subglacial system can receive groundwater discharge via springs. The subglacial water is then exported along the interface between ice and bedrock, through a perme- able basal layer of till, or through channels eroded upward into the basal ice (Hooke, 1989). In the proglacial stream, glacier discharge can be aug- mented by lateral inflow from the surrounding hillslopes or by discharge of groundwater from a deeper aquifer. In addition, water can temporarily exit a stream system by infiltration into the stream bed and banks and then re–enter the stream at some point downstream. This process is known as hy- porheic exchange. In steep mountain streams with a step-pool morphology, water infiltrates the bed in the steps and discharges back into the stream in the pools (Tonina and Buffington, 2009a,b). 1.2.2 Proglacial solute dynamics in the absence of salting In the absence of salt application, snow and ice melt are typically dilute. Solutes in the snowpack are flushed at the beginning of the melt season and by mid–summer, the meltwater is mostly solute free (Molles and Gosz, 1980; Fountain, 1996). Zeman and Slaymaker (1975) suggested that the influence of atmospheric sources of chloride is paramount in determining the ionic composition of snowpack meltwater. In their study, the chloride concentra- tion of strictly snowmelt in the Coast Mountains was 0.15 to 0.30 mg L−1 while glacier melt contained an average chloride concentration of 0.20 mg L−1 (Zeman and Slaymaker, 1975). Most solutes exported by glacier dis- charge are acquired by water following sub–glacial flow paths via dissolution of glacial till and/or bedrock (Collins, 1979; Tranter et al., 1996; Anderson et al., 1997; Brown, 2002). The geochemistry of the bedrock under a glacier determines what ions enter the meltwater, and variations in bedrock compo- sition can result in distinct chemistries (Schneebeli et al., 2007). In addition 4 1.2. Hydrology and solute dynamics in glacierized catchments to the transient influence of the flushing of solutes in the early period of snowmelt, solute concentrations in glacier discharge exhibit a hysteretic re- lation with discharge (Collins, 1979) due to the changes in the sources of water exiting the glacial system. During the afternoons and evenings, flow levels are high and streamwater is dilute, as most of the water is derived from supra–glacial flow. As the dominant water source changes to the sub–glacial system, solute content and EC rise due to enrichment from the bedrock/till, and flow levels decrease. In addition, as glacier discharge decreases, there is an increase in the relative contributions of lateral inflow and groundwater discharge to proglacial streamflow. As these waters typically have higher solute concentrations than glacier outflow, ionic concentrations in proglacial streams tend to be inversely related to streamflow (e.g., Richards and Moore, 2003). Hyporheic exchange provides a mechanism for transient storage of solutes, extending their residence time within a channel reach. The ionic composition of streamwater in an alpine basin of the Coast Mountains of British Columbia was reported by Zeman and Slaymaker (1975). Their study location was approximately 30 km north of Whistler and the basins are composed of the same geology as that found on Black- comb Mountain: granitic and granodioritic bedrock. They collected samples from the snowpack, the glacier meltwater at the snout and at four points downstream of the snout of the glacier and analyzed these for chloride and other constituents. All of the samples contained a chloride concentration between 0.20 and 0.60 mg L−1, regardless of catchment size, discharge or presence/absence of glaciers (Table 1.1). Considering the similarity in basin geology, it is believed that the background chloride levels on Blackcomb Mountain should fall in or near the range listed in Table 1.1. Table 1.1: Background chloride concentrations for two streams in the Coast Mountains of British Columbia (data from Zeman and Slaymaker (1975)). Miller Creek is glacier–fed, while Central Creek is not. Stream Discharge Chloride Drainage Area Site Elevation Miller Creek 1.9 - 4.8 m3s−1 0.20 - 0.50 mg L−1 21.6 km2 1370 m.a.s.l. Central Creek 0.05 - 0.60 m3s−1 0.20 - 0.60 mg L−1 2.35 km2 1140 m.a.s.l. 5 1.3. Objectives and hypotheses 1.3 Objectives and hypotheses The objective of this research is to understand how the salt applied to the Horstman Glacier affects the chloride concentrations in the streamwater of Horstman Creek. Figure 1.1 illustrates the three main sources of water and chloride to Horstman Creek: 1) the glacier–snowpack (Qg, Cg); 2) lateral sources of water and chloride to the proglacial stream, which should be most influential during rain events (Qlat, Clat); and 3) hyporheic exchange of water and chloride out of the stream into the bed and banks and recharge back into the system some distance downstream (Qhyp, Chyp). The amount of chloride in Horstman Creek (Qt, Ct) is thus a function of the interaction between all three major inputs and outputs upstream of the gauging site. Figure 1.1: Conceptual diagram of the main sources of chloride and water in a glacierized basin. 6 1.3. Objectives and hypotheses It is anticipated that chloride variations in proglacial streamwater will exhibit distinctive patterns among the following periods: (1) the period of active salting, (2) the period following the cessation of salting, but while there is still snow covering (at least partially) the glacier ice, and (3) the period following depletion of all of the snowpack that had been salted. Dur- ing period (1), chloride variations in glacier outflow should be controlled by the flow paths followed by meltwater, with some chloride following rapid flow paths and exiting the glacier within a day, while some will be trans- ported along slower flow paths and be stored for one or more days (e.g., in the glacier, snowpack, or in sub–glacial till) before being discharged. During period (2), chloride stored in the glacier snowpack will be gradually depleted, leading to a negative trend in streamwater chloride concentration. Ice melt below the glacier snowline would tend to dilute the chloride released from the snowpack. However, it is possible that increasing ice melt could result in a re–organization of the sub–glacial drainage system, producing a flushing of chloride stored at the glacier bed (e.g., in cavities). During period (3), chlo- ride discharging from the glacier would be derived from longer–term storage in the englacial and sub–glacial reservoirs. Rain events would likely result in downstream dilution due to increases in relatively low–chloride lateral inflow. Particularly during periods (1) and (2), it is likely that chloride will be transported into the hyporheic zone as a result of the diurnal increases in discharge and stage, which would generate hydraulic gradients conducive to infiltration of water across the bed and banks (Loheide and Lundquist, 2009). This water would be released back to the stream particularly during periods of low flow, when the hydraulic gradient across the stream bed and banks would tend to favour discharge from the hyporheic zone. Depending on the volume of hyporheic storage and the rate of transfer, this reservoir could maintain elevated concentrations throughout period (3). Figure 1.2 outlines hypotheses regarding the timing of peak chloride concentrations in relation to streamflow sources. Starting at the uppermost level of organization (1a, 1b) alternative hypotheses relate to whether the glacier and hyporheic contributions of chloride (Qg, Cg + Qhyp, Chyp) are greater or less than the lateral sources of chloride (Qlat, Clat). The second level (2a, 2b) presents alternatives based on whether or not the glacier and hyporheic components of chloride are greater than background levels during the entire ablation season, or just during specific times. Finally, alternative hypotheses at the most detailed level specify the timing of elevated chloride levels in relation to low or high flows (3a, 3b and 4a, 4b). 7 1.4. Organization of the thesis Figure 1.2: Flow chart of chloride in Horstman Creek. 1.4 Organization of the thesis The remainder of this thesis comprises four chapters. Chapter 2 provides a description of the study site and the methods used to collect and analyze the field data. Chapter 3 presents the results of the analyses, while Chapter 4 discusses the results in relation to water quality concerns and the governing processes. Chapter 5 provides a summary of the key findings along with recommendations for further study. 8 Chapter 2 Methods 2.1 Regional context Both Horstman Glacier and Blackcomb Glacier are located on Blackcomb Mountain in Whistler, British Columbia, Canada, approximately 123 km north of Vancouver. Horstman Glacier has heavy winter and summer skiing traffic, while neighbouring Blackcomb Glacier is only skied on during the winter months. Figure 2.1 illustrates the location of Whistler, along with an aerial view of Horstman and Blackcomb glaciers relative to the town of Whistler. Whistler has a maritime climate and receives an average of 1230 mm of precipitation, with snowfall accounting for 380 mm of snow–water equivalent annually at an elevation of 640 m a.s.l. Based on the 1971–2000 climate normals, monthly mean air temperature ranges from -3.2◦C in December to 16.1◦C in August (Figure 2.2). Air temperatures at the elevations of the glaciers are approximately 6 to 13◦C lower than those at the 640 m climate station based on dry and saturated adiabatic lapse rates. Consequently, a greater proportion of annual precipitation falls as snow on the glaciers than at the valley bottom site. Horstman and Blackcomb glaciers are temperate, meaning their ice is near the pressure melting point throughout the glacier, except for the sur- face layer which is cooled during the winter months. The coastal alpine lo- cation of these glaciers yields a large positive mass balance during the winter months, but currently there is an even larger negative mass balance during the ablation season. Both glaciers on Blackcomb Mountain are in states of negative net mass balance (Shea, 2006), as are other nearby glaciers such as Helm and Place glaciers (Moore and Demuth, 2001). 2.2 Study sites Horstman Glacier is a north facing glacier with a surface area of 0.36 km2 and an elevation range from 1980 to 2330 m a.s.l. This glacier is skied on 9 2.2. Study sites Figure 2.1: Location map for Horstman and Blackcomb glaciers, Whistler, British Columbia (Google Earth, 2009). every summer during June and July, during which time it is heavily salted. There are two T–bar ski lifts located on Horstman Glacier, which carry skiers up the glacier from December until the end of July. Blackcomb Glacier is 10 2.2. Study sites Figure 2.2: Precipitation and temperature normals for Whistler village at an elevation of 640 m a.s.l. located just south of Horstman Glacier with an arete separating the two. It has a northeast aspect and a surface area of 0.63 km2. In addition to being larger than Horstman Glacier, Blackcomb Glacier has a greater elevation range, from 1890 to 2360 m a.s.l. (Table 2.1). Blackcomb Glacier has no lifts located on it and is not salted. It is therefore used in this study as a reference system to assist with interpreting the effects of salting. Stream gauging stations were installed on both Horstman Creek and Blackcomb Creek. Both streams have a cascade–pool morphology, a bank- full width of approximately 4 m, similar channel gradients and elevations that differ by 52 m (Table 2.2). However, Blackcomb Creek’s station has a drainage area of 3.62 km2, more than triple the drainage area at the Horstman Creek station. Despite this difference, the glaciers should be the major source of late–summer streamflow for both creeks. 11 2.3. Data collection Table 2.1: Characteristics of Horstman Glacier and Blackcomb Glacier CHARACTERISTIC HORSTMAN GLACIER BLACKCOMB GLACIER Surface Area 0.36 km2 0.63 km2 Elevation Range 1980 - 2330 m.a.s.l. 1890 - 2360 m.a.s.l. Salted in Summer? Yes No Table 2.2: Characteristics of Horstman Creek and Blackcomb Creek at the gauging stations CHARACTERISTIC HORSTMAN CREEK BLACKCOMB CREEK Elevation 1702 m.a.s.l. 1650 m a.s.l. Slope 18◦ 17◦ Morphology Cascade–pool Cascade–pool Bankfull Width 4 m 4 m Drainage Area at Site 1.16 km2 3.62 km2 Glacier Area 0.36 km2 0.63 km2 2.3 Data collection Data collection occurred during summer and autumn of 2008. Both contin- uous and temporally discrete data were collected, as summarized in Table 2.3. In addition, downstream variations in electrical conductivity and tem- perature were measured on one date. Further details are provided below. 2.3.1 Continuous stream monitoring Gauging station locations on each creek were chosen during low flows in September 2007, based on the criteria that there was a well-mixed reach upstream, to allow accurate measurement of streamflow using salt dilution, and a relatively calm pool was present where the well would be located. Each stream gauging station consisted of two stilling wells that were attached to a piece of angle iron which was bolted to the bedrock. The Blackcomb Creek wells were installed in October 2007. Due to early winter snowfalls that halted field access, Horstman Creek wells were not installed until May 2008. The stilling wells housed the following equipment on each stream: Odyssey water level probe, pressure transducer, ion–specific chloride probe, a thermocouple, and an electrical conductivity probe. The instrument data were recorded on Campbell Scientific dataloggers that were powered by solar 12 2.3. Data collection Table 2.3: Data collection LOCATION PARAMETERS DATA TYPE Horstman Creek Temperature Continuous Water level Continuous Electrical conductivity Continuous Chloride concentration Continuous Water samples Instantaneous Blackcomb Creek Temperature Continuous Water level Continuous Electrical conductivity Continuous Chloride concentration Continuous Water samples Instantaneous Horstman Glacier Surface elevations in May and September Instantaneous Percentage of glacier covered in snow Continuous Blackcomb Glacier Surface elevations in May and September Instantaneous panels. The specific equipment housed at each station is listed in Table 2.4, including the make and model of each instrument and the parameter be- ing measured. Each of the listed parameters in Table 2.4 was recorded at ten–second intervals and averaged every five minutes from June 6, 2008, until October 23, 2008. Unfortunately, the only data that yielded enough information to be useful were the water levels recorded by the Odyssey Wa- ter Level Logger. Due to various problems described in Section 2.5, the other data collected were no complete enough to include in the analysis. Table 2.4: Instrumentation installed on Blackcomb Creek and Horstman Creek stream stations. PARAMETER INSTRUMENTATION EST. of ACCURACY Water level Water level logger - Odyssey 0.2 cm Water level Pressure transducer - Sensor Technics PS900 0.2 cm Electrical conductivity Electrical conductivity probe - custom made 0.5 µS cm−1 Chloride Cl− specific probe - TempHion 0.1 mg L−1 Temperature Thermocouple - custom made 0.1 ◦C Dataloggers CR10X, CR1000, CR10 - Campbell Scientific 13 2.3. Data collection 2.3.2 Streamflow measurement The water levels recorded by the Odyssey water level probes were converted to streamflow by developing a rating curve for each creek. Streamflow was measured using salt dilution due to the steep, irregular nature of the chan- nels (Moore, 2004). The method of dry salt slug injection was chosen for convenience (Hudson and Fraser, 2005). Reaches were chosen on each stream that included at least three constrictions to ensure mixing, minimal inputs or outputs of water, no pools, and a mostly constrained channel. During each measurement, 1.00 kg of salt was injected into the stream. Salt was injected into a high velocity zone at the center of each stream to promote rapid disso- lution and mixing. A datalogger downstream recorded the salt wave passing through the gauging reach, which was completed once the electrical conduc- tivity returned to background values. Calibrations were made with solutions of 1.00 g salt in 1.00 L of streamwater after each dry injection wave passed through the reach. Gauging was conducted 11 times on Horstman Creek and 12 times on Blackcomb Creek during low flow periods and high flow events throughout July, August and September, in an attempt to achieve the best possible rating curves. Rating curves were developed using an R script, which fits an exponential curve to the data. 2.3.3 Water sampling To complement the continuous data, ISCO auto–samplers were installed on Horstman Creek and Blackcomb Creek to collect water for chemical analysis. The auto–samplers include a programmable computer, a peristaltic pump, and sample bottles. Water samples were collected on each creek via two samplers. One sampler was programmed with an offset date allowing for sampler one to collect for eight days and sampler two to collect for the following eight days. Each set of auto–samplers were powered by one truck battery and they were programmed to collect 400 mL of water every eight hours. There were two periods of sample collection. The first period extended from July 23 to August 8, 2008. The first set of samples was collected from the samplers on August 14, after which the samplers were re–programmed to begin the second collection period. Again collecting one sample every eight hours, the second period of water sampling finished on August 30, 2008. The sampling interval of eight hours was chosen in an effort to balance the frequency of collection with the timing of each sample. The samples were taken at 04:00, 12:00 and 20:00 for each collection period. These times were 14 2.3. Data collection chosen based on the literature (e.g. Moore and Demuth, 2001), in order to capture samples during high flows (04:00 early in the melt season and 20:00 later in the summer) and during low flows, which occurred closest to the 12:00 sampling time during most of the collection period. In addition to the samples from Horstman and Blackcomb creeks, grab samples of surface melt were taken from Horstman Glacier on August 18, 2008. These samples were taken from two different surface channels on a section of exposed glacial ice. These channels were approximately 50 m from the snout of the glacier. At the time of the sampling, the snow cover on the glacier was at 50%, resulting in sampled meltwater that was a combination of snowmelt from higher up on the glacier mixed with glacier ice melt from lower down. On September 26, 2008, grab samples of water were collected at four locations where water was in contact with the bedrock. Two of these samples were from stagnant pools in the bedrock, while the other two were taken at points of seepage: one from the main bedrock body and one from a mafic intrusive body. These six samples of glacier and bedrock water were analyzed at the same time and using the same techniques as the stream samples. The purpose of the streamwater samples was to provide additional data that could be correlated to the continuous data. The concentration of chlo- ride was measured in all of the water samples taken from both streams, Horstman Glacier and the bedrock pools. The author performed the analy- ses in the Land and Food Sciences lab at the University of British Columbia. Assistance was provided by Dr. Mark Johnson during the initial phases of the analysis. The samples were analyzed using inductively coupled plasma mass spectrometry (ICP–MS). This method provided chloride concentra- tions for each sample with a precision of 0.00001 mg L−1. 2.3.4 Snowpack information An automated wide angle digital camera was placed above the Horstman Glacier housed in a Lexan case and mounted on the Horstman Hut just above the glacier. The camera was a Nikon Coolpix P50 powered by an outlet on the outside of the Horstman hut. It was programmed to take one photo per hour with the intention of documenting changes in glacier snow cover over the entire ablation season. Figure 2.3 shows the view from the camera during an evening in mid–July 2008. One of the most variable components of the snow and ice melt is how snow is added to the glacier by the mountain operations. During May and June, the Blackcomb Mountain operations use snow–cats to push snow onto the Horstman Glacier from the 15 2.3. Data collection surrounding areas. The purpose of this operation is to stockpile snow and to build jumps and other features with the snow on the glacier. These snow–cat built features can be seen in Figure 2.3 in the bottom left and right parts of the frame. Figure 2.3: Horstman Glacier in July, as seen from the automated camera located on the Horstman Hut. Dye tracer studies of water flow through the snowpack were conducted between August 14 and August 20, 2008 on Horstman Glacier. Dye was applied to three 2 m by 2 m square patches of snow, with the intention of digging snow pits two to three days later to investigate preferential flow within the snowpack as well as any direct routing to the glacier. Unfortu- nately, upon return to the field sites, all of the zones where dye had been applied had melted down to glacier ice. The high rate of melt was not ex- pected as 60 to 90 cm of snowpack had melted during the three day period, where as average melt rates were around 5 cm per day. This early depletion of the remainder of the snow on the glacier was not expected based on the snow depth trends of the previous years. As a result, no information was 16 2.3. Data collection gathered about the flowpaths within the snowpack. 2.3.5 Bedrock and till mineralogy Two samples of till and two samples of bedrock were taken from areas ad- jacent to the glacier, at approximately 10 m from the margin. These were analyzed for whole rock mineralogy and chloride by ALS Chemex in North Vancouver. 2.3.6 Climate data Climate data for this research were acquired from Environment Canada. The data are from a climate station located on Whistler Mountain at 1680 m elevation, which is part of Environment Canada’s 2010 Winter Olympic Autostation Network1. Data included hourly precipitation, divided into snowfall (SWE) and rainfall, and air temperature from May to October, 2008. 2.3.7 Salt composition and loading The salt used on Horstman Glacier is sodium chloride (NaCl). The salt is composed of at least 99.6% NaCl with other elements making up the re- mainder (Cargill, 2007). Whistler Blackcomb uses salt from Cargill and it is classified as ”Kiln Dried Solar Salt–Medium.” Table 2.5 outlines the dif- ferent chemical constituents in the Cargill salt used by Whistler Blackcomb. Due to the fact that the salt is 99.6 % or greater NaCl, we can ignore the minor components. The salt comes in plastic bags, each with a mass of 22.7 kg. The salt in these bags is then applied by hand by Whistler Blackcomb employees. From the onset of salting during the warmest days of May 2008 until the last day of summer skiing on July 27, 2008, Whistler Blackcomb applied a total of 3969 bags of salt, producing a total load of 90 020 kg (DeJong, 2009). Although this number varies from year to year, the volume of salt applied to the Horstman Glacier has been steadily increasing during the past two decades (DeJong, 2009). 1Station name: Whistler Mt. High Level VOA. 17 2.4. Data analysis Table 2.5: Chemical composition of the salt used by Whistler Blackcomb (Cargill, 2007) COMPONENT UNITS (%) TYPICAL SPECIFICATION Sodium Chloride (dry) 1 % 99.83 99.6 min. Calcium & Magnesium (as Ca) % 0.05 0.08 max. Sulfate (as SO4) % 0.08 - Surface moisture2 % 0.05 0.2 max. Water insolubles % 0.01 0.15 max. aBy difference of impurities b110◦C for 2 hours 2.4 Data analysis In addition to plotting time series of the untransformed concentrations and scatterplots with stream discharge, the logarithm of Cl was plotted against the cumulative discharge. This plot is based on the notion that, if the Horstman Glacier and its proglacial stream behave like a continuously stirred tank reactor (CSTR), then log(Clt) should decline linearly with cumulative discharge upon cessation of salt application (Eriksson, 1971). The absolute value of the slope is the reciprocal of the volume of water stored in the system. Following visual inspection of graphs, statistical models were fitted to each of the three periods hypothesized in Section 1.3. The models had the following form: Clt = bo + b1Qt + b2Qcum + b3P24 + I + e (2.1) logClt = bo + b1Qt + b2Qcum + b3P24 + I + e (2.2) where Clt is the chloride concentration in the streamwater at time t (mg L−1), Qt is the discharge at time t (m3 s−1), Qcum is cumulative discharge, which accounts for depletion of chloride from glacier storage (m3), P24 is cumulative rainfall over the previous 24 hours (mm), and I is an interaction term that includes all of the interactions between the three predictor variables (I = b4Qt·Qcum + b5Qt·P24 + b6Qcum·P24). 18 2.5. Challenges Models were initially fitted using ordinary least squares (OLS) regres- sion, using the lm( ) function in R (v. 2.7.2) within the stats package. Linear models were created using every possible combination of the candi- date predictor variables as presented in Equations 2.1 and 2.2. Residuals were examined to assess the validity of the assumptions underlying OLS re- gression. In particular, the Durbin–Watson statistic was used to check for autocorrelation. Because of the high degree of residual autocorrelation, sub- sequent model fitting employed generalized least squares (GLS) regression, which can explicitly account for autocorrelation (Gomi et al., 2006). The best overall model was chosen based on the Akaike Information Criterion (AIC). In principle, the model with the lowest AIC is the model with the best fit for the data. However, an underlying assumption is that the residu- als follow a normal distribution. As a result, the normality of the residuals was tested using the Shapiro–Wilk test (Shapiro and Wilk, 1965) and was used as an additional criterion for choosing the ”best” model. 2.5 Challenges Despite having programmed the data loggers to collect EC, chloride con- centration, water level, and temperature, these data were not usable. The first problem arose when the power supply was being drained quicker than was expected based on the specifications of the chloride-specific probe. This problem was quickly remedied with the installation of solar panels at each site. Having reset the programs and with more than sufficient power supply, the loggers were set to record. The following field visit four days later yielded good data. After two weeks of operation, a field inspection revealed that the chloride-specific probe on Horstman Creek had failed. This probe was sent off for repair and a new one ordered. Unfortunately the new probe took three weeks to arrive. By this time it was late July and the two months of salting the glacier had come and gone without any consistent chloride data. As a result, the water samplers were deployed earlier than planned and they began to run on July 23, 2008. By mid–August, one of the solar panel voltage regulators had seized and was no longer charging the battery. As a result, the power supply for the data logger was drained. At this time it was decided to collect another set of water samples. The automated camera above the Horstman Glacier contained a 2 GB memory card, which was more than enough to record images between field visits. The camera was powered by an extension cord connected to the 19 2.5. Challenges Horstman Hut (operated by Whistler Blaccomb). The images were down- loaded and the camera re–set every two weeks. Although the hut manager was aware of the camera and the connected power cord, maintenance workers had unplugged it on July 9, 2008. During the following trip to download the images the power was re–set and the camera resumed photographing. The final error with the camera occurred at the completion of the summer ski and snowboard season. When the mountain operations crews were cleaning up and shutting everything down on the mountain, the extension cord was yet again unplugged and left like that until mid–August. During this period without images, the greatest change in snow cover on the glacier occurred and it is unfortunate that we do not have the information provided by the time–lapse imagery. The ISCO auto–samplers operated flawlessly during both sampling in- tervals on Horstman Creek. All of the sample bottles contained the desired amount of streamwater, 400 mL, and the pump heads remained free of sedi- ment. Some trouble arose on the Blackcomb Creek site during the first sam- pling period. One of the older auto–samplers had problems with the inter- nal arm that moves from bottle to bottle and delivers the pre–programmed amount of streamwater via the peristaltic pump. Due to possible slight alignment errors in the bottle placement and an undulating ground surface, 11 of the 48 sample bottles were not filled. The same auto-samplers were used on Blackcomb Creek during the second period and all 48 bottles con- tained water during round two. Some bottles had less than the desired 400 mL of water, but each bottle did contain sufficient water in order to run the samples on the ICP–MS. 20 Chapter 3 Results 3.1 Overview of the study period The Horstman Glacier was open for summer skiing from June 13 to July 27, 2008. During this time, along with the month of May, when the mountain operations crews were preparing the glacier, 90 020 kg of salt was applied. Whistler Blackcomb has noted an increase in salt use over the past 15 years, largely due to warmer summers and greater variability in snowpack depth from year to year. Snowcats were used from May to July 2008 to relocate snow on the glacier as well as to take snow from areas adjacent to the glacier and put it on the glacier. Precipitation in 2008 was generally lower than normal except during August, when there was one major rain event, while air temperature was close to normal (Figure 3.1). The cumulative snowfall in the Coast Moun- tains in the 2007–2008 winter was approximately 25% less than during the 2006–2007 winter. This was noted by Whistler Blackcomb (DeJong, 2009), and is consistent with data from the Tenquille Lake snow pillow station, located approximately 78 km northeast of Whistler (British Columbia Min- istry of Environment, 2009). The lower than average snowpack for 2008, the large rain–on–snow event on August 24, 2008, and the many days of higher–than–average air temperature contributed to a glacier surface that was nearly devoid of snow by the end of August. The combination of these effects allowed the auto–samplers to sample chloride concentrations for Horstman Creek during both high flows and baseflow within a 35 day period. Figure 3.2 illustrates Horstman and Blackcomb creek streamflow, Horstman Glacier snow cover, air temperature and precipitation from July 10 to Sept 10, 2008. Horstman and Blackcomb creeks have similar patterns of streamflow, al- though Blackcomb discharge is almost an order of magnitude higher. Note that the highest gauged flow at Blackcomb Creek is 4.0 m3 s−1. Higher flows were estimated by extending the rating curve and are considered unreliable because during the high flows there was some water flow that deviated away from the main channel as it flowed through the bouldery banks of Blackcomb 21 3.2. Bedrock and till mineralogy Figure 3.1: Mean monthly precipitation (as water equivalent) and temper- ature for the 2008 study period, compared to 2005, 2006 and the 1970–2000 normals (2007 data were not available). Creek. In addition, the high flows are only considered unreliable for Black- comb Creek because there is much less lateral spreading of water during high flows in Horstman Creek than in Blackcomb Creek. Notable changes in all data presented in Figure 3.2 occurred around August 15, 2008. During this time, the temperature remained high for a period of four days which was then followed by an intense rain event. At the same time, the rate of change of snow cover was -6% per day, which was the highest rate of snow cover loss of the summer. Following the peak rain event during this period of peak melt, both Horstman and Blackcomb creeks rose to their highest water levels of 2008. 3.2 Bedrock and till mineralogy Table 3.1 outlines the results of the laboratory analysis on each sample using the technique of X–ray fluorescence (XRF). The two samples of bedrock were basalt and basaltic andesite, respectively, according to Le Maitre (2002). The loss on ignition (LOI) values for both rock samples were low, meaning there was not a lot of carbonate and most of it was in the form of veins (Galle and Runnels, 1960). By nature, till is a sampling of the bedrock and 22 3.2. Bedrock and till mineralogy Figure 3.2: Time series of Horstman and Blackcomb Creek discharge, Horstman Glacier snow cover, air temperature and precipitation during the period of record. The Blackcomb Creek streamflow estimates that plot above the dotted line are not reliable. surficial geology under and around a glacier. As a result, the till samples were collected in a way to best represent the overall types of rock in the till. The two till samples were granodiorite (e.g., dacite extrusive) source and granite (e.g., rhyolite extrusive) source, respectively. It was estimated, by visual inspection, that approximately 90% of the till was of the dacite extrusive source with 10% of the rhyolite extrusive source. In addition, the low amount of mafic minerals also points toward a granitoid rich source. However, this could also be due to weathering and subsequent removal of mafic minerals. 23 3.3. Stream temperature and electrical conductivity profiles downstream of Horstman Glacier Table 3.1: Whole rock mineralogy of bedrock and till samples MINERAL ABUNDANCIES (weight %) Rock Names Granodiorite source Granitic source Basalt Basaltic Andesite Mineral Till #1 Till #2 Bedrock #1 Bedrock #2 SiO2 60.33 66.91 51.11 58.34 Al2O3 16.39 15.85 18.23 17.59 Fe2O3 6.38 3.50 8.08 5.84 CaO 9.01 3.53 8.40 6.25 MgO 1.43 1.45 5.51 2.54 Na2O 2.33 4.28 2.95 3.96 K2O 0.65 1.28 0.10 0.64 Cr2O3 < 0.01 < 0.01 0.01 < 0.01 TiO2 0.57 0.34 0.80 0.82 MnO 0.12 0.10 0.17 0.09 P2O5 0.165 0.155 0.298 0.233 SrO 0.12 0.08 0.09 0.10 BaO 0.06 0.13 < 0.01 0.06 LOI 1.74 1.27 3.71 2.49 Total 99.3 98.88 99.45 98.96 Cl 0.012 0.035 0.011 0.014 3.3 Stream temperature and electrical conductivity profiles downstream of Horstman Glacier Beginning at the snout of Horstman Glacier, measurements of electrical conductivity (EC) and temperature were taken approximately every 70 to 100 m on August 14, 2008, from 14:00 to 15:30 (Figure 3.3). The changes in EC indicate that lateral inflow to Horstman Creek does occur below the glacier. Increases in EC typically indicate inflow of relatively solute- rich water, while decreases indicate gains of relatively solute-poor water. However, it cannot be inferred from EC alone whether lateral inputs of water cause an overall increase or decrease of chloride in Horstman Creek. 24 3.3. Stream temperature and electrical conductivity profiles downstream of Horstman Glacier Figure 3.3: Electrical conductivity and temperature values are plotted against their position downstream from the snout of Horstman Glacier. 25 3.4. Chloride variations in grab samples Table 3.2: Chloride concentrations for the grab samples of glacier and bedrock water SAMPLE ID LOCATION CHLORIDE (mg L−1) DATE SAMPLED NOTES HSTM GL1 Glacier surface channel 1 0.1650 Aug 18, 2008 Exposed glacier ice melt HSTM GL2 Glacier surface channel 2 0.2690 Aug 18, 2008 Exposed glacier ice melt HSTM RW1 Bedrock seep 0.0000 Sept 26, 2008 Main bedrock body HSTM PW1 Bedrock pool 0.0000 Sept 26, 2008 HSTM PW2 Bedrock pool 0.1280 Sept 26, 2008 Algae present 3.4 Chloride variations in grab samples The two samples of glacial meltwater collected on August 18, 2008, con- tained 0.165 mg L−1 and 0.269 mg L−1 of chloride (Table 3.2). At the time of sampling, the glacier was over 50% exposed blue ice with snow cover re- stricted to an upper section of glacier, with depths varying from 20 to 50 cm. These grab samples were collected from supra–glacial channels incised into the surface of Horstman Glacier, which was actively melting at the time of sampling. Therefore, it is likely that a large portion of the meltwater at the time of sampling was being derived from glacier ice melt rather than snowmelt. The bedrock pool samples collected on September 26 had chloride concentrations of 0.000 mg L−1 and 0.128 mg L−1, respectively (Table 3.2). All of the grab samples are not significantly different than the background levels of chloride in Blackcomb Creek, and thus there are not considered a significant source of chloride to Horstman Creek. 3.5 Chloride concentrations in streamwater The two periods of water sampling for which chloride data were available extend from July 23, 2008 at 20:00 until August 8, 2008 at 12:00 and from August 14, 2008 at 20:00 until August 30, 2008 at 12:00. During Period 1, the snow cover on the glacier ranged from 100% to around 60–70%. In contrast, during the second period of sampling, snow cover decreased from approximately 50% in mid–August to less than 10% at the end of the sam- pling period. Chloride concentrations in Blackcomb and Horstman creeks are pre- sented in Figure 3.4. The chloride in Horstman Creek is clearly detectable during the entire period of record and, notably, it is over one order of mag- nitude greater than chloride levels in Blackcomb Creek at all times. Fur- 26 3.5. Chloride concentrations in streamwater thermore, the sample with the highest chloride concentration for Blackcomb Creek is suspect. At the time of sample collection, twelve sample bottles from one of the auto samplers contained less than 5 mL of water. Of these twelve, only three had enough water to run the analysis and one of these had the highest chloride concentration recorded for Blackcomb Creek (6.595 mg L−1). Figure 3.4: Chloride concentrations for Horstman Creek and Blackcomb Creek for the entire period of stream sampling. Figure 3.5 illustrates the variations in chloride concentration during the first sampling period. At Horstman Creek, chloride concentrations averaged 57.2 mg L−1 during the period of active skiing and salting. Chloride peaked at 64.3 mg L−1 on July 27, then declined to about 10 mg L−1 by August 8. 27 3.5. Chloride concentrations in streamwater In contrast, chloride concentrations at Blackcomb Creek were mostly below 1 mg L−1 with the exception of three samples, at least one of which is likely erroneous (see comments above, and Section 2.5). Figure 3.5: Chloride concentrations for the first auto-sampled period, for Horstman Creek and Blackcomb Creek. Note the difference in scale of the y-axis. The second sampling period, August 14 to 30, 2008, represents a differ- ent regime than the first sampling period as there was less than 50% snow coverage on Horstman Glacier and the snow was no longer being salted. Fig- ure 3.6 illustrates the chloride concentrations in Blackcomb and Horstman creeks during this period. The chloride concentrations in Horstman Creek (averaging about 6 mg L−1) were one order of magnitude higher than in 28 3.5. Chloride concentrations in streamwater Blackcomb Creek, which averaged < 1 mg L−1. Figure 3.6: Chloride concentrations for the second auto-sampled period, for Horstman Creek and Blackcomb Creek. Note the diffference in scale of the y-axis Figure 3.7 shows that, for the period following the end of salting, log(Cl) initially follows a negative linear relation with cumulative discharge, consis- tent with the behaviour of a CSTR. However, the chloride concentrations eventually shift to a roughly constant value with fluctuations on both diurnal and multi-day time scales (see also Figure 3.4). 29 3.5. Chloride concentrations in streamwater Figure 3.7: Chloride, log(chloride) and discharge plotted against cumula- tive discharge for Horstman Creek. The dotted lines represent the breaks between periods 1, 2 and 3. 30 3.6. Relations between chloride and discharge 3.6 Relations between chloride and discharge The chloride–discharge relation for Blackcomb Creek did not vary markedly between the two sampling periods with the exception of two values greater than 1 mg L−1, which occurred during low flows in the first period (Figure 3.8). These two values are substantially higher than those reported by Ze- man and Slaymaker (1975) for streams draining similar geological units and appear to be outliers, although it is not possible to ascertain whether they represent erroneous values. Figure 3.8: Chloride concentrations and discharge for Blackcomb Creek during the two sampling periods. Chloride concentrations in Horstman Creek varied from about 10 to over 60 mg L−1 during the first sampling period, but with no clear correlation with discharge, which was generally below 0.15 m3 s−1 (Figure 3.9). By 31 3.7. Statistical modeling of chloride concentrations contrast, the second sampling period was characterized by highly variable flows, ranging from 0.032 m3 s−1 to 0.37 m3 s−1, and relatively uniform chloride concentrations, ranging from 2.69 mg L−1 to 10.3 mg L−1, again showing no clear correlation with discharge. Figure 3.9: Chloride concentrations and discharge for Horstman Creek during the two sampling periods. 3.7 Statistical modeling of chloride concentrations Examination of Figures 3.4 and 3.7 suggests that the study period encom- passes three distinct regimes of chloride concentrations in Horstman Creek: (1) the period of active salting, (2) a period of declining concentrations fol- lowing the cessation of salting, with an approximately exponential decline 32 3.7. Statistical modeling of chloride concentrations in Cl, and (3) a period with roughly constant Cl concentrations that exceed background values based on Blackcomb Creek and those reported by Zeman and Slaymaker (1975). The best fit models for each of the three periods using both Equations 2.1 and 2.2 are summarized in Table 3.3 and 3.4. In all cases, the Shapiro-Wilk test suggests that the residuals do not deviate significantly from normal distributions (p–values substantially greater than 0.05). For periods 1 and 3, the ”best” models for both Cl and log(Cl) include Qt as the only predictor variable, with negative coefficients in all cases (i.e., higher discharge associated with lower concentration). However, only one of these coefficients (-23.1 for Cl: period 1 ) is significantly different from 0. In all cases, and for both Cl and log(Cl), the residuals are positively autocorrelated. For period 2, the best-fit models for both Cl and log(Cl) in- clude statistically significant negative coefficients for cumulative discharge, reflecting the dominant negative trend in chloride concentration. Examina- tion of Figures 3.4 and 3.7 indicate that the model for log(Cl) is the more appropriate model, as the relation between Cl and cumulative discharge is clearly nonlinear. Table 3.3: ”Best” statistical models for Cl concentrations. Blank cells should be interpreted as variables that were not included in the ”best” model for that respective period. PERIOD 1 PERIOD 2 PERIOD 3 Coefficient Coefficient Coefficient (p–value) (p–value) (p–value) Constant 60.0 77.4 5.35 Qcum -0.00055 (0.0003) P24 Qt -23.1 -0.088 (0.50) (0.98) Residual Autocorrelation 0.62 1 0.77 Shapiro–Wilk p–value 0.73 0.98 0.16 33 3.7. Statistical modeling of chloride concentrations Table 3.4: ”Best” statistical models for log(Cl) concentrations. Blank cells should be interpreted as variables that were not included in the ”best” model for that respective period. PERIOD 1 PERIOD 2 PERIOD 3 Coefficient Coefficient Coefficient (p–value) (p–value) (p–value) Constant 1.78 2.13 0.716 Qcum -0.00000087 (0.0000) P24 Qt -0.17 -0.091 (0.52) (0.80) Residual Autocorrelation 0.24 0.49 0.75 Shapiro–Wilk p–value 0.77 0.71 0.79 34 Chapter 4 Discussion 4.1 Sources and patterns of chloride in streamwater Variations in chloride concentrations at Horstman Creek clearly reflect the effects of salt application. The highest values were observed during the period of active salting, followed by a period of declining and then stabilizing concentrations following cessation of salting. Even during the final period, chloride concentrations remained an order of magnitude above background levels indicated by observations in Blackcomb Creek and other studies in the southern Coast Mountains (Zeman and Slaymaker, 1975). The patterns of chloride concentrations in Horstman Creek reflect the timing and quantities of salt application and the processes by which salt is transported through and stored within the glacier and proglacial stream system. Salting typically begins in May each year, with daily application be- ginning in mid–June. The amount of salt applied typically increases through the summer ski season in response to increasing melt rates associated with higher air temperatures and lower snow albedo. The highest Cl concen- tration observed in Horstman Creek occurred at the end of the period of salting, likely reflecting the pattern of increasing salt application through time. Although not statistically significant, the regression models for period 1 suggested that chloride concentrations (both raw and log–transformed) were negatively related to discharge during the period of salting, possibly reflecting a dilution effect. The presence of positive autocorrelation in the residuals from the statistical model is typical for systems involving mass storage, and could reflect changes in the flow paths through time as the glacier and its snowpack evolved through the melt season. Period 2 was dominated by a declining trend in chloride concentrations, associated with the depletion of snow cover and the resulting flushing of salt. Most of the salt that was stored in the snowpack appears to have been flushed into or through the glacier in this period. This inference is based on two lines of evidence. First, the supraglacial meltwater samples collected on 35 4.1. Sources and patterns of chloride in streamwater August 18 had chloride concentrations consistent with background values (Table 3.2). Second, chloride concentrations in Horstman Creek dropped to relatively steady values by mid–August (Figure 3.4), at which point the glacier still had about 50% snow cover. Neither antecedent precipitation nor current discharge were included in the best-fit statistical models for period 2, suggesting that dilution by rainfall runoff or melt of glacier ice was not a dominant process. During period 3, snow cover on the glacier, as determined via the pho- tographs from the automated camera, was less than 50% and the chloride present in Horstman Creek was largely derived from storage in englacial, subglacial and proglacial reservoirs. The best–fit model for this period in- cluded statistically insignificant negative relations with Qt, providing weak evidence for dilution effects either due to rainfall runoff or glacier melt. Chloride concentrations exhibited multi-day fluctuations during this period, which could result from re–organization of the glacial drainage system and associated release of chloride from isolated portions of the drainage system, or changes in proglacial hydrology, such as changes in hillslope runoff or groundwater discharge. While hyporheic exchange represents a potential mechanism for transient storage and prolonged release of chloride between the glacier and the stream sampling point, the data collected in this study do not provide any basis for quantifying the effect. A complete mass balance of chloride could not be completed due to the lack of data prior to July 23, 2008. A simple estimation of chloride flux during the melt season, suggests that only 15 % of the chloride applied to the glacier gets flushed out with the melting of the snowpack. As a result, a large amount of chloride is held in storage for a possible slow release over time. It is likely that the bedrock is not a significant contributor of chloride into Horstman Creek, as illustrated by the bedrock mineralogy results. The bedrock pool–water grab samples also suggest that the bedrock is not re- leasing chloride above background values. The two–phase pattern of chloride concentrations following the cessation of salting suggests that the Horstman system could be modelled using at least two distinct reservoirs: (1) a relatively active reservoir that behaves approximately like a continuously stirred tank reactor, which dominated the period from the cessation of salting to mid–August (period 2 ), and (2) a reservoir that releases salt slowly, contributing to the relatively constant chloride levels during the last half of August (period 3 ). Further research should focus on identifying the physical processes associated with these two reservoirs and their variation through time. Such studies would contribute 36 4.2. Implications for downstream water quality and summer ski operations to the development of integrated hydrochemical models for application in alpine catchments. 4.2 Implications for downstream water quality and summer ski operations British Columbia environmental regulations for drinking water indicate a maximum chloride limit of 250 mg L−1. During the study period, the highest recorded chloride concentration in Horstman Creek was 64.3 mg L−1. This is significantly less than the drinking water standard as well as government standards for aquatic flora and fauna, as outlined in Table 4.1. Furthermore, these concentrations would become increasingly diluted in downstream sec- tions of Horstman Creek by inflow of uncontaminated hillslope runoff and groundwater. One caveat is that our study only sampled the latter part of the period of active salting, and samples were only taken every 8 hours. It is possible that higher chloride concentrations occurred but were not sam- pled. Such an event would likely have been recorded by the chloride and electrical conductivity probes, so it is unfortunate that those data had to be discarded. Climate warming could result in higher streamwater chloride concentra- tions in the future. Warming would likely increase the amount of salt needed to maintain a firm surface on the glacier. Furthermore, as current climate projections suggest that winter snow accumulation should decrease (Bar- nett et al., 2005), and that the seasonal snowpack should disappear earlier in the season, there may be less hillslope runoff or groundwater discharge downstream of the glacier to dilute the salt discharging from the glacier. In addition, if the glacier retreats, there would be less meltwater generated and thus less dilution of applied salt. Given the caveats and concerns raised above, it is recommended that Whistler Blackcomb initiate a monitoring program to provide more detailed information on chloride concentrations and their variability both during and following the period of ski operations, and to develop strategies for mitigat- ing downstream impacts of salting should concentrations approach levels indicated in Table 4.1. Results of this study have relevance not only to ski operations at Whistler Blackcomb, but also to Mt. Hood, Oregon, which is the only other ski resort in North America that currently applies salt to glacier snow to support summer skiing. They are also relevant to the proposed Jumbo Glacier ski resort in the Kootenay Mountains of B.C. 37 4.2. Implications for downstream water quality and summer ski operations Table 4.1: British Columbia environmental regulations PARAMETER CHLORIDE LIMIT (mg L−1) Drinking Water 250 Freshwater Aquatic Life 600 (max instantaneous) 150 (30 day average) Wildlife 600 38 Chapter 5 Conclusion This study has generated unique observations related to the fate of salt applied to glacier snow, with practical implications for managing summer ski operations and fundamental implications for our understanding of alpine hydrochemistry. This chapter provides a summary of key findings, followed by suggestions for future research. 5.1 Summary of key findings Chloride concentrations in Blackcomb Creek, which is not influenced by salting, were consistent with those found in a previous study in the southern Coast Mountains of B.C. (Zeman and Slaymaker, 1975), and were used as a reference against which to assess the influence of salting on concentrations in Horstman Creek. Chloride concentrations in Horstman Creek were at least an order of magnitude higher than background values during the entire ablation season, clearly reflecting the effects of salt applied to the glacier for summer skiing. There were three distinct patterns of chloride variations during the ablation season: (1) the period of active salting, with the highest chloride concentrations, (2) a period of declining chloride concentrations, associated with the depletion of the glacier snowpack, and (3) a period of persistent elevated chloride concentrations from mid- to late August. Statistical modelling of chloride concentrations provided weak, but sta- tistically insignificant, evidence of short–term dilution effects during periods 1 and 3. In period 2, the best–fit statistical models included cumulative dis- charge as the sole predictor variable, representing the progressive depletion of chloride from storage. In all three periods, the residuals exhibited posi- tive autocorrelation and multi–day fluctuations, likely reflecting progressive changes in the glacial drainage system and proglacial streamflow contribu- tions through the study period. The maximum observed chloride concentration was 64.3 mg L−1, which is substantially lower than standards for drinking water or aquatic organ- isms. However, it is possible that higher concentrations occurred but were not sampled. In addition, climate warming could result in higher chloride 39 5.2. Recommendations for future study concentrations in Horstman Creek due to the need to apply more salt, as well as decreases in late–summer streamflow due to an earlier disappear- ance of seasonal snow and glacier retreat. In addition, the town of Whistler receives some of its drinking water from Horstman Creek (after it merges with Blackcomb Creek downstream of the gauging sites) Therefore, it is rec- ommended that Whistler Blackcomb undertake more detailed monitoring to ensure that glacier salting does not have negative downstream impacts. 5.2 Recommendations for future study This study provided some insight into the processes controlling the fate of chloride in the meltwater downstream of a salted glacier. The field studies for this project utilized some effective and other not so effective techniques of data collection and future studies would benefit by learning from the mistakes we made. I propose the following three areas where future study on salt application on glaciers for the purpose of recreation could focus on to help fill the void of knowledge on this subject: • Snowpack analysis and surface flow drainage • Englacial and subglacial glacier–specific drainage • Hyporheic exchange and lateral inputs downstream of a glacier Snowpack studies during and after the salt application would provide a better understanding of how the surface meltwater and salt are transported through the snowpack and glacier and downstream. Such studies should include direct sampling of the snowpack and analysis of choride content throughout the ablation season. In addition, snow pits dug throughout the summer skiing period would provide some understanding of how layers in the snowpack evolve throughout the summer. Flow paths of higher conductivity can quickly route water through the snowpack and the glacier, possibly expediting the releases of chloride. It would also be useful to map the salt application as the spatial variability of the salting process could alter the timing of chloride release from the snowpack. Glacier ice coring would provide information about the structures within the ice and could possiby detect brine pockets or small storage reservoirs, if they exist. Sampling of discharge and chloride concentration immediately below the glacier outlet would provide valuable information on the patterns of chloride released from glacial storage. Higher frequency sampling than 40 5.2. Recommendations for future study conducted in this study, as well as continuous monitoring of electrical con- ductivity and/or chloride, should be conducted to ensure that short–term spikes are not missed. Further research should focus on the hydrology of the proglacial stream to clarify the influences of hillslope runoff, groundwater discharge and hy- porheic exchange. Such a study should involve monitoring stream discharge and chloride concentration near the glacier outlet as well as at a downstream station. Chloride concentrations should be analyzed for water samples col- lected from wells or piezometers in the stream bed and riparian zone, as well as samples of water draining from the hillslope. These data would allow both water and chloride budgets to be estimated for the proglacial stream reach, providing direct information on the effects of hyporheic exchange and dilution effects. As there has been a large amount of year–to–year variability in the sum- mer temperature, precipitation, snowpack depth, degree of packing by snow- cats, and amount of salt applied to Horstman Glacier in the past few years, a multi–year study would be beneficial to understanding the chloride con- centrations in Horstman Creek throughout the entire ablation season. The construction of at least two stream gauging stations on each of Horstman and Blackcomb creeks, intense sampling of the snowpack throughout the sum- mer months, repeated sampling of lateral sources of water and tight record keeping on the spatial distribution of salt on the glacier would provide a more in depth understanding of how salts are transported downstream of a salted glacier. 41 References Anderson, S. P., 2007. Biogeochemistry of glacierl landscape systems. Annu. Rev. Earth. Planet. Sci. 35, 375–99. Anderson, S. P., Drever, J. I., Humphrey, N. F., 1997. Chemical weathering in glacial environments. Geology 25 (5), 399–402. Barnett, T. P., Adam, J. C., Lettenmaier, D. P., 2005. Potential impacts of a warming climate on water availability in snow-dominated regions. Nature 438, 303–309. British Columbia Ministry of Environment, 2009. Automated snow pillow reporting. URL Brown, G. H., 2002. Glacier meltwater hydrochemistry. Applied Geochem- istry 17 (7), 855–883. Campbell, F. M. A., Nienow, P. W., Purves, R. S., 2006. Role of the supraglacial snowpack in mediating meltwater delivery to the glacier sys- tem as inferred from dye tracer investigations. Hydrological Processes 20 (4), 969–985. Cargill, August 2007. Technical information: Kiln dried solar salt- medium. Specifications sheet. Colbeck, S. C., Davidson, G., 1972. Water percolation through homogenous snow. Tech. rep., USGS Water Resources Division. Collins, D. N., 1979. Hydrochemistry of meltwaters draining from an alpine glacier. Arctic and Alpine Research 11 (3), 307–324. DeJong, A., 2009. Total salt use on the Horstman Glacier during 2008: Whistler Blackcomb. Email correspondance. Eriksson, E., 1971. Compartment models and reservoir theory. Annu. Rev. Ecol. Syst. 2, 67–84. 42 References Fountain, A. G., 1996. Effect of snow and firn hydrology on the physical and chemical characteristics of glacial runoff. Hydrological Processes 10 (4), 509–521. Furbish, D. J., 1988. The influence of ice layers on the travel time of meltwa- ter flow through a snowpack. Arctic and Alpine Research 20 (3), 265–272. Galle, O. K., Runnels, R. T., 1960. Determination of CO2 in carbonate rocks by controlled loss on ignition. Journal of Sedimentary Research 30 (4), 613–618. Gomi, T., Moore, R. D., Dhakal, A., 2006. Headwater stream temperature response to forest harvesting with different riparian treatments, coastal British Columbia, Canada. Water Resources Research 42 (W08437). Google Earth, 2009. Google Earth image screenshot. Greenwood, J., 2004. Investors wait since 1991 on ski resort plan–Regulatory faceoff. National Post. Harper, J. T., Bradford, J. H., 2003. Snow stratigraphy over a uniform depo- sitional surface: spatial variability and measurement tools. Cold Regions Science and Technology 37 (3), 289–298. Harper, J. T., Humphrey, N. F., 1995. Borehole video analysis of a temperate glacier’s englacial and subglacial structure: Implications for glacier flow models. Geology 23 (10), 901–904. Hooke, R. L., 1989. Englacial and subglacial hydrology: A qualitative review. Arctic and Alpine Research 21 (3), 221–233. Hudson, R., Fraser, J., 2005. Introduction to salt dilution gauging for streamflow measurement. Part IV: The mass balance (or dry injection) method. Streamline Watershed Management Bulletin 9 (1), 6–12. Le Maitre, R. W. (Ed.), 2002. Igneous rocks: A classification and glossary of terms, Recommendations of the International Union of Geogological Sciences. Subcomission of the Systematics of Igneous Rocks, Cambridge University Press. Loheide, S. P., Lundquist, J. D., 2009. Snowmelt-induced diel fluxes through the hyporheic zone. Water Resources Research 45, W07404, doi:10.1029/2008WR007329. 43 References Marsh, P., 1999. Snowcover formation and melt: recent advances and future prospects. Hydrological Processes 13 (14-15), 2117–2134. Molles, M. C., Gosz, J. R., 1980. Effects of a ski area on the water quality and invertebrates of a mountain stream. Water, Air, & Soil Pollution 14 (1), 187–205. Moore, R. D., 2004. Introduction to salt dilution gauging for streamflow measurement: Part 1. Streamline Watershed Management Bulletin 7 (4), 20–23. Moore, R. D., Demuth, M. N., 2001. Mass balance and streamflow variabil- ity at Place Glacier, Canada, in relation to recent climate fluctuations. Hydrological Processes 15 (18), 3473–3486. Pheidias Project Management Corp., 2007. Jumbo Glacier resort master plan: Appendix 2-C: Climate, water and glaciers fact sheets. URL Richards, G., Moore, R. D., 2003. Suspended sediment dynamics in a steep, glacier fed mountain stream, Place Creek, Canada. Hydrological Processes 17, 1733–1753. Schneebeli, M., Schworer, C., Rhyner, H., Rixen, C., 2007. Chemische pis- tenpraparation - grundlagenbericht. Tech. rep., Eidg. Institut fur Schnee- und Lawinenforschung SLF. Schuler, T., 2002. Investigation of water drainage through an alpine glacier by tracer experiments and numerical modeling. Tech. rep., Swiss Federal Institute of Technology Zurich. Shapiro, S. S., Wilk, M. B., 1965. An analysis of variance test for normality (complete samples). Biometrika 52 (3-4), 591–611. Sharp, M., Tranter, M., Brown, G. H., Skidmore, M., 1995. Rates of chemical denudation and CO2 drawdown in a glacier-covered alpine catchment. Geology 23 (1), 61–64. Shea, J. M., 2006. Blackcomb Glacier: Status and recommendations, un- published, UBC. Singh, P., Spitzbart, G., Hubl, H., Weinmeister, H. W., 1997. Hydrological response of snowpack under rain-on-snow events: a field study. Journal of Hydrology 202 (1-4), 1–20. 44 References Theakstone, W. H., Knudsen, N. T., 1996. Isotopic and ionic variations in glacier river water during three contrasting ablation seasons. Hydrological Processes 10 (4), 523–539. Tonina, D., Buffington, J. M., 2009a. Hyporheic exchange in mountain rivers I: Mechanics and environmental effects. Geography Compass 3, 1063– 1086. Tonina, D., Buffington, J. M., 2009b. Hyporheic exchange in mountain rivers II: Effects of channel morphology on mechanics, scales, and rates of ex- change. Geography Compass 3, 1038–1062. Tranter, M., Brown, G. H., Hodson, A. J., Gurnell, A. M., 1996. Hydrochem- istry as an indicator of subglacial drainage system structure: A compari- son of alpine and sub-polar environments. Hydrological Processes 10 (4), 541–556. Waldner, P. A., Scheenbeli, M., Fluhler, U. S.-Z. H., 2004. Effect of snow structure on water flow and solute transport. Hydrological Processes 18 (7), 1271–1290. Zeman, L. J., Slaymaker, H. O., 1975. Hydrochemical analysis to discrim- inate variable runoff source areas in an alpine basin. Arctic and Alpine Research 7 (4), 341–351. 45


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