Bow River, Southern Alberta, Canada
THE UNIVERISTY OF CALGARY
Controls on the chernistry of the Bow River, southern Alberta, Canada
by Stephen E. Grasby
A DISSERTATION SUBMlTTED TO THE FACULTY OF GRADUATE STUDIES IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF =OR OF PHlLOSOPHY
Department of Geology and Geophysics
Calgary, Alberta
March, 1997
O Stephen E. Grasby 1997 Acquisitions and Acquisitions et Bibliographie SemMces- se~kesbibiiograptiiques
The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distniute or sell reprod&, prêter, distri'buer ou copies of Merthesis by any means vendre des copies de sa thése & and in any fonn or format, malQng quelque detea sous quelque this thesis avaiiable to interested forme qye ce soit pour mettre des persons. exemplaires de ce~tethèse à la disposition des personnes intéressées.
The author retains ownership of the L'auteur caoserve la propriété du copyright in Merthesis. Neither droit d'auteur qui protège sa thèse. Ni the thesis nor substantial extracts la théSe ni des extraits substantiels de fkom it may be printed or otherwise ceile-ci ne doivent être imprimés ou reproduced with the author's autrement reproduits sans son permission. ABSTRACT
Integrated chernical and stable isotope analyses were used to defhe the controls on the dissolved inorganic load in the Bow River, and thereby eiucidate the chernical and hydrologie dynamics of the nver. The dominant sources of ions in the nver are amiospheric deposition and rock weathering. The input by weathe~gis largely controiied by dissolution of carbonate and evaporite minerais. Calgary is the most significant point source input dong the river-
Effluent fiom the sewage treatment plants loads Na, K, and Cl to the nver. Cation activity ratios are strongiy controiied by exchange on smectite. Smectite is absent in the nver, suggesting that activity ratios are an iaherited signature of ground water. Stable isotope data indicate that discharge in the fail and winter is fed by groundwater. The high discharge event in the spring is a mixture of snowmelt and displaced groundwater. Summer discharge is fed by rainfall. Despite seasonal variations in the TDS load, element ratios are constant, suggesting that the chemistry of snowmelt and raid" are altered by the same processes controiling groundwater chemistry. This suggests snowmelt and rab faU must pass through the ground before becoming discharge. S1*O,, data indicates dissolved sulfate undergoes a complex redox history before reaching the river, implying that the water transporting the sulfate passes through the anoxic zone before becoming discharge. Therefore, the Bow River is largely fed by ground water.
The chernical denudation rate for the Bow River at Banff is 678 kg/ha/y. The denudation rate for the basin as a whole is 340 kg/ha&. Loading fiom Calgary accounts for 8 to 92 of the mass flux out of the basin in the spring and fdand 25% of the mass flux in the summer. 1 am gratefiil to my supe~sorIan Hutcheon, and to Roy Krouse, who were both encouraging and supportive of my work. They were never too busy to sit down and talk about my research. Funding for this pmject was provided by research grants to 1. Hutcheon and H.R. Krouse. This work was assisted by the cooperation of several agencies. Chernical data for the Bow River at Lake Louise and Ban€f was supplied by the Water QwLty Branch of the Water Survey of Canada. The Wakr Survey of Canada and Trans-Alta Utilities provided discharge data for the Bow River and tributaries. Aiberta Environment provided precipitation chemistry. Parks Canada provided groundwater data for Banff National Park. Several people at the University of Calgary helped me complete this project. 1 am gratehil to Maria Miehailescue, Jesusa Pontoy-Overend, and Nenita Lozano for teaching me how to nui the mass spectrometers, as weii as for feeding me. Maurice Shevaiier not only helped in the lab, but also solved ali my computer problems. Pat Michad ran cation andysis. The 'Pudes" provided many hours of stimulahg arguments about geochemistry. matched by many hours of playing Doorn. Marian Johuson assisted in collecting water samples, as well as distracthg me with ski trips. Once again, thanks to Teresa for her support and patience. TABLE OF CONTENTS
ABSTRAa ACENOWLEDGEMENTS TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES
CHAPTER 1 Introduction
BACKGROUND PROJECT DEFINITION OBJEcmms FIELD AND LABORATORY METHODS
CEIAPTER 2 Overview of the Bow River Basin
GEOGRAPHY CLIMATE GEOLOGY ANTaROPOGENIC WATER USE
CHAPTER 3 Meteonc verms groundwater inputs to the Bow River
INTRODUCTION O AND H ISOTOPE COMPOSI'MONS OF PRECIPITATION AND GROUNDWATER O AND H ISOTOPE COMPOSITIONS OF WATERS IN THE BOW mRBASIN VARIATIONIN THE SCABLE ISCJIDPE COMPOSlTlON OFTRlBUTARlES - THE '-" CONïTNENTALEFFmT CATIONS FOR PALEO-CWMATE STUDIES 6180AND 6D OF THE BoW RIVER - THE SOURCE OF DWHARCX ~WNSIREAMVARIATION IN 6D CONCLUSIONS
CHAP'iZR 4 Chemical weathering of the Rocky Mountains
INTRODUCTION METHODS CONTROLS ON THE CHEMISTRY OF THE BOW RIVER CORRKIWG IQRNON-WEATHEWG COWNENTS WEATHERINGREACIIONS CO~OLLINGRIVER CHEMISIRY THERMODYNAMICCONIROLSONFWERCHEMISTRY -CAL DENUDATiON RATE CONCLUSIONS CEAPTER 5 Chernical dynamics of the Bow River
INTRODUCTION CHEMICAL CaARACTERISTICS OF POINT SOURCE INPUTS CHEMICAL CHARACTERISTICS OF TEIE BOW RIVER CONTROLS ON TBE MAJOR ION CHEMISTRY OF TBE BOW RIVER CHLmcDE SODIUM SULFUR CALCIUMAND MAGNESIUM BICARBONATE CHEMICAL DENUDATION RATE SUMMARY IMPLICATIONS FOR BASIN HYDROLOGY
CaAPFER 6 Tracing anomalous TDS in Nose Creek
INTRODUCTION TBE NOSE CREEK BASIN RESULTS AND DBCUSSION SOURCE OF NOSECREEK WA'IER INORGANECHEM~~~RY OF NOSE CREEK WATER OXUGEN AND HYDROGEN SOTO OPE conmsmo~sOF NOSECREM WATER SOURCES OF SULFATE CONCLUSIONS
CEiAPTER 7 Conclusions
REFERENCES APPENDIX 1 Sâmple Locations APPENDIX 2 Chernical and stable isotope data for the Bow River APPENDLX 3 Chcmical data for springs and shallow ground water APPENDIX 4 Chernical and stable isotope data for Nose Creek Table 2-1 Average annual flow for the Bow River and major aibutaries. Table 2.2 Licensed and actual water use for the Bow River for 1991. Table 3.1 Isotope &ta for spcings in the Bow Basin. Table 4-1 Chioride normaliseci eqyivalents ratios for precipitation. Table 4.2 Mean mual wet and dry deposition rates for the Bow Basin. Table 4.3 Dominant weathering reactiom anticipated for the Bow Basin. Table 4.4 Ca and Mg bearing minerals that may occur in sedimentary rocks Table 4.5 Calculated beidellite activities for smectite in equilibrium with Bow River water at Banff and Lake Louise. Table 4.6 Long term denudation rate for the Bow Basin. Table 4.7 Chernical deaudation rates for world rivers, and world average rate. Table 5.1 Chemisüy of storm sewer discharge, sewage effluent, and irrigation return flow.
Table 5.2 Calculated beidellite activities for mectite in equiliirium Bow River water at Banff. Table 5.3 Total monthly discharge and flux of TDS for the Bow River at Banff, Carseland, and Hays. Table 6.1 Flow data for Nose Creek. LIST OF FIGURES
Figure 1.1 Location of study area and sampling locations. Figure 1.2 Discharge for the Bow River at Calgary, showing sampling periods. Figure 2.1 ~ajorweather systems for the Bow River Basin. Figure 2.2 Physiographic Regions of the Bow River Basin. Figure 2.3 Mean annual precipitation and evapotmnspiration for the Bow River Basin. Figure 2.4 Composite monthly discharge for the Bow River at Batlff. Figure 2.5 Monthly average mual precipitation for Calgary. Figure 2.6 Geology of the Bow River Basin. Figure 2.7 Dowmtream variation in discharge of the Bow River.
Figure 3.1 Variation in the stable isotope composition of precipitation from the five main weather systems that bring moisture to the Calgary area Figure 3.2 m> and 6"O of tributaries versus the distance of their confluence dong Bow River. Figure 3.3 6D and Sf8Oof tributaries versus distance of the headwaters of the tributaries hmthe Great Divide. Figure 3.4 Schematic illustration of mixing relationship of westerly and easterly winds over the Rocky Mountains. Figure 3.5 Plot of SD versus 6% for ail samples coilected dong the Bow River. Figure 3.6 Plot of 6D versus 618~. Figure 3.7 Plot of m> versus 6180best fit iines for each sample set from the Bow River. Figure 3.8 Discharge versus temperature, measured at Banff. Figure 3.9 Plot of 6D versus 6180for the tributaries dong the Bow River. Figure 3.10 Downstream variation in 6D.
Figure 4.1 Geology in the headwaters of the Bow River basin. Figure 4.2 Discharge venus TDS for the Bow River at Banff, 1978 - 1995. VlIl Figure 4.3 Composite plot of monthly TDS measurements at Banff, 1978 - 38 1995. Figure 4.4 Temqdiagram showing major cation and mion composition of the Bow River at Lake Louise and Bd. Figure 4.5 Cumulative plot of mon- Cl concentrations at Banff and Lake Louise. Figure 4.6 Na+K vs. Cl for Banff. Figure 4.7 Major ion composition of the Bow River and rivers draining a variety of Lithologies throughout the world. Figure 4.8 Ca+Mg vs. HCO,. Figure 4.9 Ca+Mg vs. HCO, + SO,. Figure 4-10 Discharge vs. SO,/total anions. Figure 4.1 1 Plot of log a~ala(~)~versus log a~g/a(~)~for Bow River water. Figure 4.1 2 Plot of log aCa/a(W2 venus log aMg/a(H)' for Bow River water with calculated phases boundaries superimposed. Figure 4.13 Plot of log aNa/H versus log aKM- Figure 4.14 Plot of a) log a~a/a(~)~versus log a~g/a(~)~,and b) log aNfl versus log aK/H for groundwater. Figure 4.15 Instantaneous Myflux versus a) discharge and b)cumulative instantaneous flux, measuted at Banff. Figure 5.1 Temary diagram, in equivalents, showing major ion composition of tributaries, storm sewer discharge and effluent from waste water treatment plants. Figure 5.2 TDS of tributaries to the Bow River. Figure 5.3 Total dissolved solids load versus distance dong the Bow River. Figure 5.4 Temary diagram of major cations and anions of Bow River water. Figure 5.5 Dowstream variation in CI concentrations for the Bow River. Figure 5.6 Dowstream variation in Na concentrations for the Bow River. Figure 5.7 Downstream variation in the Na+KICa+Mg equivalence ratio. Figure 5.8 Na + K versus Cl for the Bow River and tributaries. Figure 5.9 Na-K activity plot for the Bow River. Figure 5-10 63S of potential inputs of S to the Bow River.
Figure 5.1 1 Downstream variation in the concentration of SO, and PS, for the Bow River. Figure 5.12 6YS versus 1/S04. Figure 5.13 Dowllstream variation in 6% aud 8180in SO,. Figure 5.14
Figure 5.15 Downstream variation in a) Ca and b) Mg for each season. Figure 5.1 6 Ca + Mg versus totai ak for the Bow River and tributaries. Figure 5.17 Ca + Mg versus alk + SO, for the Bow River and tributaries. Figure 5.18 Ca-Mg activity plot, a) fd,b) winter, c) spring, and dl mmmer- Figure 5.19 Sources of DIC and their associated 6I3cvalues (after Pawellek and Veizer, 1992). Figure 5.20 Variation in pC0, dong the length of the Bow River. Figure 5.2 1 613cof DIC vernis calculated pCO, for the Bow River. Figure 6.1 Nose Creek basin, showing sample locations Figure 6.2 Ternary plot of major ions for Nose Creek Figure 6.3 Variation in major dissoived ions dong the length of Nose Creek. Figure 6.4 Plot of m> versus 6180for Nose Creek. Figure 6.5 kf80in sulfate verses ~"0of water for Nose Creek.
Figure 6.6 6% venus suUate concentration for Nose Creek. When the well's dry, we know the worth of water Benjamin Franklin CHAPTER 1 Introduction
BACKGROUND
Each year 40,000 km3 of water from the continents to the oceans. Of this approximately 20,000 km3are available for human (Clarke, 1993). This representr the global resource of renewable fiesh water. This resource is becoming increasingly stressed due to population growth; our per capita renewable water mpply has decreased by one third since 1970, due to the addition of 1.8 billion people to the Earth since then (Postel, 1993).
As water supplies becorne stressed, maintahhg water quality becomes increasingly important. In order to niaintain water quality, it is necessary to understand contmls on the chemistry of fiesh water systems. River systems are perhaps the most dEcuit fresh water system to examine due to their dynamic nature. Chemical analyses alone are often not sunicient to elucidate the sources of, and controls on, the inorganic chemistry of nver water. Chemical analyses are pdcularly limited when tracing non point-source inputs to river systems. Stable isotope analyses can enhance our understanding of the geochemistry of nver systems as they offer the potential to distinguish sources of dissolved inorganics in surface water, particularly sulfate and catbon. Stable isotopes of sulfate have ken used to irace both naturai and antbropogenic inputs of suifur into the environment (e.g. Krouse and Grinenko, 199 1; Hendry et al., 1986; Norman and Krouse, 1992; kmeu et al., 1994; van Donkelaar et al., 1995; Eden, 1996). These studies have focused on tracing industriai emissions into the atmosphere, lakes, groundwaier, soils, and vegetation. A number of workea have also used sulfiu isotopes to trace non-point source inputs into river systems (e.g. Hitchon and Krouse, 1972; Rabinvich and Grinenko, 1979; Ivanov, 1983; Longinelli and Edmond, 1983; Trembaczowski and Halas, 1993). These investigations served to identify naairal sources of sulfate, such as dûsolution of evaporites or oxidation of suifide minerals. Stable isotopes of carbon have also ken used to examine controls on the carbon cycle in river systems. The riverine carbon cycle is complex due to the various sources and interactions dong the fiow path. Stable isotopes of carbon have been used to examine the role of organic matter in the generation of CO, overpressure in riven (e-g. Richey et al. 1988; Buhl et al., 1991), and atmospheric buffering of pCO, in river systems with long residence times (Yang et al., 1996; Pawellek and Veizer, 1994). The role of algd photosynthetic activity in controliing pC4 was also examined by Pawellek and Veizer (1994). Hydrograph separation is another important appiication of stable isotopes analysis in river snidies (e.g. Fritz et al., 1976; Sklash and Farvolden, 1979; Mdec, et al., 1974).
By exarnining the stable isotope composition of H and 0, it is possible to test models of runoff generation (e.g Horton, 1933; Hewlett and Hibbert 1967; Dunne and Black, 1970). The basic problem these models try to address is how do snowmelt and rainfd reach a
Stream. By understanding the pathway water follows to the river, we can better understand the factors that innuence the chemimy of that water.
PROJECT DEFINITION
This study examines controls on the chemistry of the Bow River in southem Alberta. The Bow River is quickly reaching it's sustainable Limit as a water resome for local inhabitants. Currently, 88% of the average annual fiow of the Bow River is licensed for use and up to 39% is consurned (Table 2.2). The Bow River is the largest of three rivers that feed the South Saskatchewan River. Future increase in consumption of the Bow River is Limited by interprovincial agreements to maintain 5090 of the average annual discharge of the South Saskatchewan River. Despite heavy use, there are ody a few water quality saidies of the Bow River; the most complete data sets are in the pristine headwaters. Block md W(1988) and Block et al. (1993a,b) examined inorganic, organic, and trace elernent data of the Bow River in Banff National Park. Cross et al. (1986) and Charlton et al. (1986) examined nutrient levels down stmrn of Calgary. The Bow River Water Council (1994) provides a good summary of the state of the Bow River, cumnt water use, and an overview of the natural and anthropogenic environment of the basin. They note a general decrease in water quality dong the length of the river. Unforninately, the down Stream areas, with the highest potential for anthropogenic impacts, have the lean background data. The most noticeable detenoration of water quality is high levels of nutrients, fecal coliCorms and benthic aigae, ail of which increase downstream of Calgary.
Objectives
The primary goal of this study is to use the integrated anaiysis of major ions and stable isotopes to identify the source and controls on dissolved inorganic loads, and thereby contnie to the cmnt understanding of the chemical and hydrologic dynamics of river systems. To achieve this goal, the objectives of this work were to: 1) Characterise the hydrologic cycle of the Bow River Basin in tenns of relative inputs of ground and meteonc water. 2) Mode1 controls on the river chemistry in ternis of weathering and equilibrium exchange reactions. 3) Detedethe chemical denudation rate of the Bow River in the pristine headwaten, and the basin as a whole. 4) Distinguish natural from anthropogenic sources of sulfur and iheû relative inputs to the Bow River-
5) Examine the riverine carbon cycle.
Nose Creek, a minor tributary to the Bow River, was also examined to &temine if the anomalously high total dissolved solids (TDS) load is the resd.t of naaual phenornena or anthropogenic activity in the Nose Creek basin.
FIELD AND LABORATORY METHODS
There are two primary sources of data in this study. Discharge data dong the Bow River and triiutaries, and long term data sets of chemical analyses in the headwaters of the River, were obtained nom Environment Canada. In order to examine the chemical dynafnics along the length of the river, samples were collected along the Bow River and at the mouths of the main tributaries. These samples were analysed for major ion composition, and stable isotope measurements of SD and PO in &O, 6°C in HC03, and 6% and 8''O in SO,.
Four sample sets, fiom spring of 1993 to fall of 1994 were coiiected along the Bow River and its major tributaries (Fig. 1.1, AppendU. 1). These sample sets represent a range fkorn baseflow to highflow conditions (Fig 1.2). Each sample set was cokted over a pend of 4 to 7 days. To avoid sampling storm events, sp~gand summer samples were coilected at least 7 days after the last rainfd. A sample set includes 15 samples nom dong the Bow River, and 8 samples nom tributaries. The mean annuai discharge at the mouth of the Bow River (from 1965 to 1990) is 2.8 1 km3 (Environment Canada, 1990). The 1993 and 1994 discharge was 44% greater than and 23% less than average respectively. The 1993 and 1994 tlow does not appear as extreme when compared to the longer term flow data at Calgary (191 1 to 1990). Here the 1993 and 1994 discharge was 3% greater than and 12% las than average respectively. Data at Calgary are a more diable mesure of the variabilïty in the river flow as there are no irrigation developments upstream of the city.
Figure 1.1 Location of study area and samphg locations. BR = Bow River, BT = Bow tributary. Sample locations are given in Appendix 1.
Sample collection and bottle cleaning foilowed procedures defhed by Envirocment Canada (1983). Grab samples were collected in mid-stream, either fiom the upstream side of bridges, or by wadiag. Tnbutary samples were collected near their confluence with the Bow River. Unstable parameters (pH and temperature) were meanired on-site. For the remaining analyses, the water was passed through a 0.45 pm fdter. Samples for cation analysis were acidïfied to pH Q with ultrapure HNO,. Samples for anion analysis were untreated and stored at 4 OC. For stable isotope analysis, dissolved sulfate was precipitated in the field as barium sulfate, by the addition of BaCl,. The water was then acidified to pH 6
42 in order to dissolve any BaC03. For measurement of 613c in HCOi, water was extracteci hmthe sample bottie by syringe and then passed through a syringe nIter into a htglas boctle. The sample was preserved by poisoning with HgC4, and storage at 4
OC. In the lab, CO, was exsolved by addition of phosphoric acid in vacuo. Chernical and stable isotope analyses were conducted at the University of Calgary. Alkalinity was determined within 24 hours of sample collection using an Onon 960 auto-titrator. Anions were measured by ion liquid chromatography and cation concentrations were measured by atomic absorption. Analytical emr was estimated to be less than 2%. Samples with charge balance greater than 5% were rem.
~[FIOW Data at Calgary
Figure 1.2 Discharge data for the Bow River at Calgary, showing periods of sample coiIection. 34 Stable isotope compositions ('0, D, S, and I3c) are expressed using the usual 6 notation:
8 @O) = [(R sample - &tandard) 1 ktandard x 105 (1-1) were R is-the abundance ratio of heavy to Light isotopes. The standards used are VSMOW for oxygen and hydrogen, Caiion Diablo Troilite (VCDT) for sulfur. and V-PDB for carbon. For 6% analyses, S02was prepared using the rnethod of Yanapisawa and Sakai
(1983); SI~O,,, was measured with CO2 prepared by graphite reduction of BaSO, (Shakur. 1982); was rneasured on CO, isotopicaIIy equilibrated with (Epstein and Mayeda, 1953);and m) was measured using H, produced by the Zn-reduction method of Coleman et al. (1982). Combined sampling and analyticd errors for isotope data were 18 18 estimated to be f0.2%0for 6 O (H201and fi,and fl%o for 6D and 6 O,,, Precision of isotope measwements of carbon is high. f0.3%0,however erron induced by sarnpling and filtering increase erroa to tl%. CHAPTER 2 Overview of the Bow River Basin
GEOGRAPKY
The Bow River fiows wt hmthe Canadian Rocky Mountains into the interior plains of western Canada (Figs. 2.1.2.2). The river originates at Bow Sumrnit, 2063 m a.s.1.. and flows 619 km to join the Oldman River 1260 mm lower than i origin. to form the South Saskatchewan River that evenrualiy discharges into Hudson Bay. The river drains an area of 25,300 krd, and has an average muai discharge near its mouth of 2.8 1 km3
(Environment Canada, 1990). Variation in flow dong the river. and the relative inputs of major triiitaries, are given in Table 2.1. The river reaches it peak discharge at Calgary. Water Loss downstream of Calgary is related to extensive irrigation developments.
Figure 2.1 Map of western Canada showing the Bow River Basin aad schematicaliy illustrating the two major weather systems that bting the majonty of precipitation to the basin. The high pressure sytern in Idaho is marked "W. Table 2.1 Average annual flow data for the Bow River and major tributaries, arranged ftom the headwaters to the confïuence with the Oldman River. Tniutaries are placed in the relative position of theV confluence with the Bow. Locations are hdicated in Figure 2.2. Discharge data nom Enviromnent Canada (1990).
- - Location Tributary Bow River Index discharge discharqe Bow River Tributaries (Fig. 2-21 x 106 m3 x 106 m
Upstceam Lake Louise Pipestone Brewster Redearth Baker Banff spray Cascade Kananaskis Seebee Ghost Ghost Reservoir Jumping Pound Bearspaw Cdgary Elbow Nose Highwood Carseland Bassano Confluence with Oldman River
The Bow River Basin can be divided into three main physiographic units: 1) the Rocky Mountains. 2) the foothills, and 3) the prairies (Fig. 2.2). The Rocky Mountains are characterised by vaiieys covered by coniferous forests that are surrounded by steep mountains with exposed bedrock. The foothills are characterised by rolling hills covered in mked deciduous and coniferous forests that open eastward into grasslands. The eastem ~uo-thirdsof the basin is covered by praine grasslands that have been converted withia the 1st 100 years to agricuitural use (mostiy cereal grains and canola crops). The major
tniutaries to the Bow River all have kir headwaters in the &ont and main ranges of the Roclcy Mountains, whereas some minor tributaries have theû headwaters in the prairies
/ Bow River Basin
Figure 2.2 Physiographic Regions of the Bow River Basin and sample locations (BR = Bow River, BT = Bow Tributary). Numbered locations are discharge measurement stations cable 2.1).
CLIMATE The Bow River Basin has a cold temperate climate. Average annual temperanires reported by Environment Canada Vary hm-0.4 OC in the momtains (at Banff) to 5.5 OC in the prairies (at Calgary). Average precipitation exceeds evapompiration in the 11 mountains (600 and 250 mm/yr respectively), whereas in the prairies evapotranspiration exceeds precipitation (450 and 300 dyrrespectively) (Fig. 2.3 a and b). In parts of the eastem basin, potential evapotranspiration cm exceed 600mm/y. On average, 50% of the precipitation in the mountains is snowfd. In the prairies this drops to 25%, most of which is evaporated by Chinook winds. The dominant hydroogic event in the basin is the melting of the winter snow pack. This lasts about 40 days, beginning in May, with the highest discharge occuning in June (Fig. 2.4).
There are two main weather systems that b~gthe majority of precipitation to the Bow River Basin (Reinalt, 1970) (Fig. 2.1). The dominant source of precipitation is westerly winds that bring precipitation to the basin in the faii and winter. In spring and summer precipitation is denved largely fkom winds that swing no& around a high pressure system in Idaho. Although both systems derive precipitation £kom the Pacinc Ocean, the moisture in the westerlies onginates nom higher latitudes. In the winter months the Rockies have a strong min shadow effect on the westerly flow, creating progressively drier conditions fÎom West to east. In the spring and nunmer this affect is reversed, as air systems from Idaho rise orographicaily against the eastem dopes of the Rocky Mountains, creating progressively dner conditions from east to West (Rehalt, 1970). Moisture derived from Idaho can account for up to 45% of the annual precipitation in the eastem Rockies. Overall, the most precipitation fds during the summer months (Fig. 2.5), with the majority of precipitation falling in the Rocky Mountains. Based on Figure 2.3, the average annual raidail for the Bow River Basin is calculated to be 13 km3. In coatmt the average annual flow of the river at its mouth is approximately 2.8 km3. Using an estimated additional 1 km3consumed by irrigation, the nuioff ratio for the Bow River is 0.3 1, less than the world average of 0.46 (Berner and Berner, 1987), and consistent with the dry nature of the basin. precip. z evap \ avap > precip.
Figure 2.3 Mean annual a) precipitation, and b) evapotranspiration for the Bow River Basin in mm.(after Atlas of Alberta, 1969). I FMAM J J AS OND
Figure 2.4 Composite monthly discharge for the Bow River at Banff for 1978 to 1995, from Environment Canada (1990).
Figure 2.5 Monthly average anoual precipitation for Calgary (from Klivokiotis and Thomson, 1986) 14 GEOLOGY
Figure 2.6 iuustrates the geology of the Bow River Bd. The basin is underlain entkely by sedimntary rocks. For simplicity, the basin can be divided into 4 main LithoIogic groups. The western most part of the basin, upstnam of Lake Louise, is dominated by clastic rocks of the Miette and Gog groups. The Miette Group is dominantiy low-grade peiites and grits (feldspathic pebble conglomerates). The Gog Group is dominantly quartzite. The area between Lake Louise and Bdis underlain dorninantly by Cambrian carbonates and interùedded cdcareous shdes. From Banff to Morely, the basin is dominated by Devonian carbonates and calcareous shales. The Cambrian sequence was deposited in an open maxine setting, whereas Devonian carbonates were deposited in a resmcted sea (Mossop and Shetsen, 1994). As a consequence, the Devonian carbonates have a higher evaporite mineral content. The eastem two thirds of the basin are underlain by Cretaceous to Paleocene saadstones and shales. Wkù the exception of the exposed cWs in the Rocky Mountains, and dong the river valIey, the entire Bow Basin is overlain by tin and glacial lake deposits. The till can be divided into a western and eastem unit based on lithology (Moran, 1986). The western till is derived fiom the Cordillerian ice sheet, and consequently is dominated by carbonate clasts. The eastem till was deposited by the continental ice sheet and is characterîsed by igneous and metamorphic rocks derived fiom the Canadian shield. The boundary between the two tills runs through the eastem edge of Calgary (Fig. 2.6).
ANTHROPOGENIC WATER USE
Table 2.2 details the water use and consumption in the Bow Basin for 1991. Domestic use accounts for ody 11 % of water use in the basin. in cornparison, imgation accounts for 8595, or 46% of the average annual flow. Figure 2.7 illustrates variations in discharge dong the length of the river. Heavy irrigation use in the summer reduces the river discharge at Bassano below levels in the head waters at Lake Louise. Total iicensed withdrawal hmthe river is 88% of the long term annuai flow. with total consumption estimated at 39%. The comption estimate is high as t includes loss to ground water, which wodd evennially be returned to the river. There is one instream storage reservoir on the Bow River, the Ghost Reservoir upstream of Calgary (Fig. 2.7). The net effect of operating this resewoir decreases the naturai fiow of the Bow River during the summer and inmases it in the fdl and winter. In addition to the Ghost Reservoir, there is a large diversion development at Bassano downstream of
Calgary. Here the river Ievel is raised hmApril to October to allow diversion to a series of off-stream storage reservoirs. Through the fa and winter the river flows free. In addition to storage reservoirs, there are several "nui of the river" hydropower developments and weirs that alter the fiow on a daily basis.
Table 2.2 Licensed and actual water use for the Bow River for 199 1 (after Bow River Water Quality Council. 1994)
Licensed Withdrawa i Withdrawn Consumed User Group Amount % Amount % Amount % m3x i06 total m3x 106 total m3x 106 total 490 19.8 166.6 11.0 10.2 0.9 Irrigation 1894 76.7 1296.6 85.4 1050.9 95.7 Agriculture 8.9 0.4 8-5 0.6 8.5 0.8 Industrial 60.8 30.6 2.0 1.2 ! 2.5 13.6 Others 16 0.6 15.3 1.O 15.3 1.4 Total 2469.7 100.0 1517.6 100.0 1098.5 100.0 % of total annual flow* (inStream1 *based on Environment Canada (1990) value for long term average fiow of 28 10 m3 x 106 rather than the Bow River Water Council's nported value of 4lûû m3x 106
Lake Louise Ghost Dam. Carseland Weir Banff Bearspaw Dam 8assano1 Dam 600 4 0 fall ,*--,, O winter .. ---, -9 'Oo - 00 . --9 O spring I 0 -9- 8. A summer . 400 &**..O---+y'
distance (km) from confluence with the Oldman River
Figure 2.7 Dowostream variation in Bow River discharge for fd,winter, spring and summer (after Environment Canada, 1990). Assessing seasonal variation in meteoric versus groundwater inputs to the Bow River using 6% and 6D
INTRODUCTION
Ultimately, most moff originates fiom precipitation. However, the pathways precipitation follows to reach streams are often unclear. By understandhg thes pathways, we can gain a clearer understanding of the hydrology of the basin, and the factors that Muence the chemistry of surface water. The stable isotope composition of the Bow River was examined to discern the relative inputs of meteoric and ground water to the river system. and how they Vary seasondy. By gaining a better understanding of the hydraulic cycle of the Bow River, it will be easier to examine the controls on the chemishy of the Bow River.
In order to quantify the isotope composition of meteoric input, it is necessary to examine the "reversecl" continental effect observed by Yonge et al. (1989). Most models of the isotope composition of precipitation are based on Rayleigh htionation mechanisms; as precipitation is progressively ~leasedfrom clouds the remaining vapour becomes progressively depleted in the heavy isotopes D and "O. This OCCUIS as weather systems move inland, and is often referred to as the continental effkct (Dansgaard, 1964). The continental effect is often expressed as a progressive inland depletion in the heavy isotopes of surface water. B is more pronounced over mountainous areas where orographie uplifc enhances precipitation and also superimposes a minor bctionation due to the drop in temperature related to lifhng of the air mas. In a reconnaissance study of western Canada, Yonge et al. (1989) sarnpled surface water across southwestern Canada, fiom the Pacific
Coast, across the Great Divide, and into the Canadian Prairies. They note that from the Pacinc Coast to the Great Divide the 6D and 6'*0 of water bodies foiiow a typicd Rayleigh distiüation pattern. However, ihis trend abmptiy changes east of the divide to a trend of increasing 6D inland. Standard modek of evapotranspiration (e.g. Salati et al., 1979; Rozanski et al., 1982; Ingraham and Taylor, 1986) do not explain such a trend. Yonge et al. (1989) suggest, but do not demonstrate, that the "reversed" continental effect observed east of the Great Divide is related to either recycling of moisture fiom evapotranspiration or mixllrg of precipitation hmdinerent sources.
O AND H ISOTOPE COMPOSITIONS OF PRECIPITATION AND GROUNDWATER in order to examine relative inputs of ground water and meteonc water to the river it is necessary to characterise the isotope composition of these two sources. The isotope composition of precipitation in Calgary varies widely, with extremes in 6''0 rangkg fiOm
-5.0480 to -44.W00 and m) fiom -48.5 to -333960 (C.Yonge, unpublished &ra; Fig. 3.1). The heaviest precipitation is associated with spring and summer raios denved nom Idaho, whereas the iightest is associated with weather systems derived fiom Arctic fronts. The average isotope composition of precipitation derived hmIdaho is &"O = -15%0, 6D = - 112%0,whereas westerly winds cany precipitation with lighter isotope compositions, 6180 = -20qO0. 6D = -155%~~Calgary precipitation has a weighted average local meteoric water iine (LMWL) expressed by the equation (C. Yonge, unpublished data):
similar to the global MWL defiued by Craig (196 1). 0 Idaho 1 O
TOC
Figure 3.1 6"0 versus temperature, showing the variation in the stable isotope composition of precipitation from the five main weather systems that bring moisture to the Calgary ana (after Yonge, unpublished data).
Results hmstable isotope analyses of nine springs dong the Bow River are presented in Table 3.1. Ground water samp1es fdon a best fit line defined by:
Since both ground water and precipitation have well dehed relationships between 6% and 6D we can assess their relative contri'bution to the river. Table 3.1 Isotope data for springs in the Bow Basin
Location 6~ % amo%O Vermilian Lake Canmore Creek Many Springs Mount Yarrianuska Big Hiil Springs Silver Springs A Silver Springs B HW 1A
O AND H ISOTOPE COMPOSITIONS OF WATERS IN THE BOW RIVER BASIN
Variation in the stable isotope composition of tributaries - The ccreversed" continental effect
Results from stable isotope analysis are presented in Appendur 2. Figure 3.2 shows the variation in m) and S180of tn'butaries; 6D and 6180are plotted vs. distance dong the river where the tributaries enter the river. For each season there is a trend of increasing m) and 6'6 with distance. This trend is similar to that observed by Yonge et al. (1989) and is contrary to a simple single-source Rayleigh hctionation mode1 that pcedicts a progressive inland depletion in the heavy isotopes. Nomally, when waters becom enriched in D and 'b,it is interpreted to be due to evaporation. However, two feahires suggest that although evaporation occurs, it is not the primary control on the observed e~chment. Fust, the spring samples show the greatest increase in 8D and 6I8O with distance. However, this is when evapotranspiration would be relatively low. Second, the Elbow (ER) and Highwood (HR) nven have 6D and 8'0 values that are very negative compared to the trend of e~chmentin heavy isotopes east of the Great Divide, as defmed by the other tributaries. If the and 6180of the tributaries are plotted against the distance of their headwaten fiom the Great Divide (Fig. 3.3), then the Elbow and Highwood rivers fit the observed trend of inland e~chmentin heavy isotopes. This suggests that the stable isotope composition of
Uibutaries is a function of theïr source area rather than evaporation. As discussed in Chapter 2, the Bow River Basin receives the mjority of precipitation nom two sources, westerlies and weather systems originating in Idaho. The stable isotope composition of precipitation derived hmIdaho is more enriched in heavy isotopes than precipitation derived hmthe westetly winds (Fg. 3.1). Due to the reveaing min shadow efSect over the Rocky Mountains, precipitation hmthese two weather systems mix, with moisnire derived nom Idaho fonning a larger proportion of the total precipitation eastward in the Bow River Basin (Fig. 3.4). The rnixing effect appears greatest in the spring, when the of triiutaries is observed to increase at a rate of 3mJ100 km away fiom the Great Divide (Fig. 3.3). This is coincident with the time when weather from Idaho is a significant source of precipitation in the eastem ranges of the Rocky Mountains, and the western Ranges are adding significant fluxes of snow melt (denved hmwesterly flows) to the triiutaries drainhg the Rwky Mountains. In the fall, mixing appears less significant as the observed increase in m) of tniutaries is ody 6%d100 km. This is coincident with the season when westerlies are the dominant source of precipitation in the basin. Thus, the fact that the trend of increasing 6D eastward of the Great Divide is greatest when input fkom the two different weather systems is greatest, and potential evapotranspiration is low, indicates that this trend is related to the rnixing of weather fronts. rather than evaporation effects. a faIl
O winter
O spring
a) distance (km) along Bow River b) distance (km) along Bow River of the tributaries confluence of the tributaries confluence
Figure 3.2 Plot of a) 6D and b) 8'0 of tniutaries versus the distance of their confluence along Bow River. PR = Pipestone River, SR = Spray River, KR = Kananaskis River, GR = Ghost River, JC = Jumpingpound Creek, ER = Elbow River, NC = Nose Creek, HR = Highwd River.
a) distance (km) of heaûwaters from the Great Diide
Figure 3.3 Plot of a) 6D and b) 8'0 of tn'butaries vernis distance of the headwaten of the tn'butaries from the Gnat Divide. Letter codes are the same as in Figure 3.2. distance hm Great DMde
Figure 3.4 Schematic diagram illustrating mixing relatioaship over the Rocky Mountains of weather systems derived fiom westerly winds and nom high pressure systems in Idaho, and the resdtant surface water composition.
Implications for paleo-climate studks The above data indicate caution should be used in paleoclimate studies of continental interion. As stated by von Grafenstein et al. (1996) "The temporal GL80-temperature relatiomhip is the foundation of many attempts to reconstruct paleo temperatures ..." However, as they indicate, this relationship is generally only calibrated for the modem day. Although von Grafenstein et al. (1996) show that #'O in pmipitation is related to temperature changes for the 1st 200 years in southern Gemiany, other worken have demonstrated that historical variations in 6'b are relateci to changes in atmospheric circulation patterns (Amundson et al., 1996). In a seaing Lüre the Bow River basin, a shift in circulation pattems could cause a sipnincant shift in the PO composition of net precipitation. For example, if easterlies bmught less precipitation to the Bow Basin. the net 8'0 at any particular point east of the Great Divide would decrease, whereas if they brought more precipitation the net 6180 would increase. Ushg the temperature effect of Dansgaard (1964), a decrease in 6I8O would be wrongly interpreted as a net decease in mean temperature, where as an increase in 8'0 would be interpreted as a net inc~a~e.
Although variations in weather pattern do represent ciimatic changes, the isotope record can not always be used as a simple proxy for temperature: east of the Rocky Mountains. However, if properly calibrated (e.g. Amundson et al., 1996), historïcal variations in VO could be used to examine temporal changes in circulation pattems in this part of North America
6"O and 6D of the Bow River - The source of dikcharge A key question hydrologists seek to answer is "how is rainfd or snowmelt over a catchment aansfonned into stream moff?" Rodhe and Killingtveit (1997). Stable isotope data can be used to distinguish whether melting snow enters stream networks directly by overland flow, or indirectly by seepage through the ground (e.g. Martinec et al. 1974;
Krouse et al., 1978; Sklash and Famolden, 1979 ). If snow melt was translated directly to the river by overland flow, the river would show a meteoric water signature coincident with the rise in stage. If seepage displaces exis~ggmundwater into streams, the rise in stage will not be accompanied by a significant shift in the isotope composition of the stream water. Dincer, et al. (1970) show that in a mountainous carbonate terrain, similar to the ROC@ Mountains, only one-third of the snow melt becomes direct surface moff, the other two-thirds displaces groundwater into surface streams. The 6'*0 and 6D isotope composition of the Bow River was examhed in order to better define runoff generation in the Bow River Basin. The 6% and 6D values of the Bow River samples are plotted in Figure 3.5. The best fit Iùle is defined by: The low r value shows there is no simple relatiombip in the isotope composition of the nver through the course of the year. However, when samples hmeach season are plotted separately, seasonal variations in the isotope composition arr evident (Fig. 3.6).
Figure 3.5 Plot of 6D versus SmOfor ail samples collected dong the Bow River.
Figure 3.7 illustrates the best fit lines for each season, as well as the LMWL and the best fit line for groumi water (GWL). The best fit line for samples coilected in the fd is coiocident with the groundwater he, suggesting groundwater is the primary source of river water during this season. The fdand winter represent low stage flow of the nver when groundwater input is maximum (020ray and Bames, 1977). This is evident in Figure 1.2 as the background 50 km3/sdischarge, and Figure 4.2b where TDS of the nver is highest during low stage flow. Presurnably the winter samples, when the ground surface and most of the nver is fiozen, would exhibit the greatest groundwater input of any of the simple sets. However, this sample set is biased by upstream samples which may explain the best fit Line having a dope pater than the GWL.
1 d) summer h
Figure 3.6 Plot of m> versus 6% for a) fa,b) winter, c) spring, and d) summer. Figure 3.7 Best fit lines for each sample set, with the LMWL and GWL, and the average isotope composition of precipitation derived nom westerlies and the Idaho High.
Spring high discharge (Fig. 1.2) is related to the melt of the winter snow pack, however, the scattered data in Figure 3.7 suggests that water feeding the river is fiom a variety of sources. The low slope of the best fit iine defined by the stable isotope data (Fig.
3.7) suggests there is sipnincant groundwater input to the river at this tirne. Varying mixtures of gromdwater and melt water would explain the observed scatter of data. Temperature profiles of the river support this. If snow melt reached the river as overland fiow, then the temperattue of the water entering the river would pnsumable be close to zen, untii the snow pack was melted. However, the initiation of sp~ghigh discharge is associated with a jump in water temperatures, hmnear zero, to 5 to 10 'C as measured at Banff (Fig. 3.8). Thus, similar to the study by Dincer, et al. (1970), the spriag mlt in the Bow Basin is interpreted to displace Large amounts of groundwater into the river. The summer sarnples plot on a best fit line that has a slope close to the LMWX.,. This indicates üiat the dominant source of the summer discharge is meteoric water. This is consistent with the summer king the weaest part of the year (Fig. 2.4), with a large meteoric water input hmweather systems derived hmIdaho.
discharge m3/s
Figure 3.8 Temperature versus discharge, measured at Bd(Environment Canada, 1990).
Timing of groundwater recharge can be estimated based on stable isotope data presented above. The GWL,and the best fit line for the fall sample set, intercept the LMWL close to the average value for precipitation derived fkom westerly winds. Whereas the best fit Line for summer discharge has a less negative intercept with the LMWL, closer to average composition of precipitation from Idaho. This suggests that groundwater is dominantly recharged by precipitation nom the westerly air masses. Westedies are dominant in the winter months, suggesting that spring mit of the winter snow pack is the domhant source of groundwater recharge in the basin. This is consistent with the above interpretation that spring wltdisplaces pre-existing groundwater into the river. The 6180 and 6D values for tniutaries have a sîmiIar relation as the Bow River (Fig. 3.9). Samples hmthe spring and nunmer plot close to the LMWL, whereas the samples fiom fall and winter define slopes of 3.8 and 4.6 mpectively. This indicates that, like the Bow River, ground water is the major source of tn'butary water in the fd and winter, whereas meteonc water is the primary source in spring and sumrner. As well, the intercepts of the best fit lines with the UifWL show seasonal variations. The sp~ganci summer samples have a less negative intercept with the LMWL, relating to pmcipitation derived fiom Idaho. In cornparison, the fall and winter sampies have intercepts close to the average values for precipitation denved fiom westeriies. Again, this suggests that the dominant source of groundwater recharge is spring melt of the winter snow pack. Given the above data, the question as to "how is tainfidl or snowmelt over a catchment transformed into Stream runoff?" can be addressed to some extent for the Bow River. If we make the simple assumption that 50 m3/s is the average groundwater discharge into the Bow River (based on falywinter discharge data in Figure 1.2 and isotope data present above), then groundwater accounts for a maximum of 40% of the average annual discharge of the Bow River. However, the grotmdwater contriibution is likdy greater because the spring melt appears to have a significant component of displaced groundwater.. In order to quantq this input, and thus better define gmund water contriution to the river, it would be necessary to cokt detailed tirne-series samples of surface, ground, and snow melt waters during the spring melt. The implications these data have on the hydrologie cycle of the Bow Basin are discussed mer,in relation to chernical data, in Chapter 5. Figure 3.9 Plot of 6D versus 6180for the trioutaries dong the Bow River, with best fit hes for each sample set and the LMWL and GWL, and average isotope composition of precipitation deriveci 6rom westerlies and the Idaho High.
Downstream variation in 6D
The variation in 6D dong the length of the Bow River is plotted in Figure 3.10. Each season has a general trend of increasing 6D fkom the headwaters to the confluence with the Highwood River. This is largely a function of addition of tri'butiuy water progressively enriched in D as discussed above; there are no major tnibutaries downstream of the Highwood River confluence. In the fd samples, the isotope composition of the Bow River is relatively constant dowmtream of the Highwood confiuence. This suggests there is littie input of "new" water, and that there are no insiream processes affecting the isotope composition (e-g. instream evaporation). Due to ice cover, samples downstream of Calgary were not collected in the winter. The 6D values of spring samples show enatic variations. This is Wrely related to mUing of multiple sources of water along the river. Sp~grepresents the highest discharge in the river due to melt of the mountain snowpack, as weiI as relatively hi@ rain fail. To accommodate melt, resemoirs along the river are lowered, thereby adding stored water hmthe previous year to the sp~gflow. Thus the isotope composition of the river water will be a fiinction of muiag of these various sources along the river.
Summer samples show a progressive increase in m) downstream of the Highwood confluence. This observed enrichment in D is unWrely related to evaporation, as evaporated water tends to plot to the right of the LMWL. Table 2.1 and Figure 2.7 show an increase in discharge of the Bow River, unaccounted for by major tributaries, downstream of Bassano. This added water is iikely from ephemeral tributaries and irrigation remflow dong this portion of the river. -la- @ïiEëmcI -145 a O g -150- tO -156 - -1 60 d", *, Figure 3.10 Downstream variation in - -1 ôD, for a) fd, b) winter, c) spring, -135 - and d) summer. -140 - s a -145- ce -150 -
-156 -
-16Q , i
distance (km) CONCLUSIONS
The anomalous trend of increashg 6D inland, observed by Yonge et al. (1989), is ~lated to the rain shadow effect east of the Great Divide, and the dting east-west mixing of precipitation fiom the two predominant weather systems (westerly flow, and weather systems derived nDm Idaho). Thus, caution shouid be used in applying Rayleigh type fhctionation modeis to interpreting trends in the stable isotope composition of surface waters. Lt is important to quanti.@ whether more than one weather system is bringing moisturc! to a basin, and how mixing of Merent weather systems can create apparent hctionation trends. More importantly, these data indicate caution should be used in paleoclimatic studies of continental intenors as a shifi in circulation patterns could cause a significant shift in the 6"O composition of net precipitation. If tùis is the case, variations in the stable isotope composition of a historical record would be wrongly interpreted as changes in mean temperature. Although variations in weather patterns do represent climatic changes, the isotope record can not always be used as a simple proroxy for mean temperatme east of the Rocky Mountains. However it may be possible to use bistoncal variations in 6180to examine temporal changes in climate patterns.
Stable isotope data indicate that ground water is an important input to the Bow River, accounting for at least 40% of total mual discharge. Groundwater feeds the river in fd and winter, whereas rain water feeds the river in the summer. The spring discharge is a combination of grouudwater displaced by snowmelt and snowmelt itself. Seasonal pattems in the stable isotope composition of the Bow River suggest that spring melt of the winter snow pack recharges gromdwater systems in the basin. while at the same tirne displaciag older pundwater into the river system. CHAPTER 4 Chernical Weathering of the Rocky Mountains
INTRODUCTION
Environment Canada has monitored the chemistry of the Bow River at Lake Louise and the BaPark gate (Fig. 4.1) since 1978 (Block et ai., 1992). In addition, discharge is
monitored daily at the toms of Banff and Lake Louise. These data offer an opportunity to examine weathering rates, and controls on river chemistry, in a near pristine, cold- temperate climate carbonate-basin. These long tenn data sets are necessary for weathering rate calcdations, as they reduce the influence of short-term climatic and biotic fluctuations
that may cause yearly variations in river chemistry (Bluth and Kump, 1994). The two Environment Canada stations represent a 35 km reach of the river, nom Bow
Lake to Lake Louise, and a 80 km reach fiom Lake Louise to the Banff Park gates. Upstream of Lake Louise, the basin is an undeveloped wilderness area. Between Lake Louise and Banff the basin is largely undeveloped except for the two tomsites. The basin
has weU defined geology (Rice and Mountjoy. 1972% 19724 1978, 1979). Upstream of
Lake Louise, the basin is dominated by clastic rocks of the Mietîe and Gog groups, and Cambrian carbonates (Fig. 4.1). The Miette Group is dominantly low-grade pelites and
grits (feldspathic pebble conglornerates), and the Gog Group is dominantiy quartzite. Between Lake Louise and Banff, the basin is underlain by Paleozoic carbonates with interbedded units of calcareous shales. The vaiiey bottoms are covered by giaciolacustrian deposits. Lake 2 Louise âiiarge & chernistry Banff
Figure 4.1 Geology in the headwaters of the Bow River basin. Sarnple coilection sites and discharge measurement stations are indicated.
METHODS
The data malysed in this chapter were collectecl by the Water Survey of Canada. Cheniical data hm 1978 to 199 1 are published in Block et al. (1993a,b). Daily average discharge data for the Bow River and tributaries, hm 1901 to 1990, are published by Environment
Canada (1990). Additional unpublished cbemicd and discharge &ta for 1990 through to
December 1995 were provided by the Water Survey of Canada. At Lake Louise, discharge is measured at the same location samples are coilected for chemicai analyses. However, for
Banff, discharge is measured 15 km upstceam fiom the sampling site (Fig. 4.1). The data sets were culled for samples with an inorgaaic charge balance greater than 5%. Trend analyses by Block et al. (1993a) do not show any long tem antbmpogenic impact on this portion of the Bow. CONTROLS ON THE CHEMISTRY OF THE BOW RIVER
Typical of most rivers, discharge is the major control on the chemistry of die Bow River (Fig. 4.2). nie TDS load is highest during winter base flow conditions, vaqing between 95 to LOS mg/i at Lake Louise and 175 to 200 mgll at Banff (Fig. 4.3). During spring melt of the snow pack, there is a significant drop in TDS (approximately 50% of winter maximum), foiiowed by a steady increase through nimmer and faii as discharge returns to base flow conditions. Aithough the TDS load is largely related to runoff, the chemistry of most unpolluted rivers is controlled by weathering of bedrock mever, 1988; Meybeck, 1987). Despite large seasonal variations in TDS of the Bow River, the aonnalised composition of the river is nearly constant and distinct fkom that of precipitation (Fig. 4.4). Element ratios of the Bow River are characteriseci by a fixed WMg ratio, low Na, a slightly variable AWSO, ratio, and low Cl. The consistent nature of the normalised composition of the water, both seasonaily and for the 18 years of measurement, suggests there are fundamenial control(s) on element ratios. Processes controlling the major element ratios must be operating within the basin as the chemistry of precipitation is altered before the water entea the river (Fig. 4.4). As a first approach to determùiing controls on the river chemistry, the abundance and relative proportions of dissolved ions in surface water can be modeled usiag a mass balance approach in terms of chemical weathering of the common rock forming xninerals in the basin (e-g. Garrels and MacKenie, 1967). However, in order to examuie the role weathering reactions play in coatrolling river chemistry, it is necessary to correct the measured water chemistry for the non-weathering components (e.g. Stailard and Edmond.
198 1; White and Blum, 1995). 1 b) Banff b
discharge m3/s discharge m3/s
Figure 4.2 Discharge versus TDS for the Bow River at a) Lake Louise, and b) Banff, 1978 - 1995.
1 1 b) Banff
J FMAM J JASON0 FMAM
Figure 4.3 Composite plot of monthly TDS measurements ai a) Lake Louise, and b) Banff, 1978 - 1995. Louise
rn Na+K Alk
Figure 4.4 Temary diagram, in equivalents, for major cation and anions for Lake Louise and Banff.
Correcting for non-weathering components
The chernical dynamics of a river system can be summarised using the relation given by White and Blum (1995): Where F- represents the annual flux of material fiom weathering of bedrock and till, F,, F,, is the measured river Bux, F,, and F,, represent atmosphenc loading, F,, ,,, represents the net flux of dissolved solids hmion exchange sites in clay miaerals, and F,, represents the net flux of dissolved solids due to changes in biomass. Equation 4.1 assumes that there is no significant net ground water flux into the basin. In order to quantify this relationship it is necessary to define each component of the flux of dissolved inorganics to the river.
1) Atmospheric loading h order to examine the role of chernid weathe~g,it is important to fkst correct for atmospheric loading, both wet and dry fd, of major dissolved ions (e.g. Hoiland, 1978; Stiùlard and Edmond, 1981). Typically, corrections for atmosphenc loading are doue using methods outlined by Stallard and Edmond ( 198 1), where they separate atmospheric flux into a marine and terrestriai component. They use only the marine component to correct for atmosphenc loading, as the terresaial component is assumed to be derived fiom within the basin. This is a reasonable assumption for a basin where weather systems move inland in an upstream direction. However, the geographic settkg of the Bow River basin is signxcantly dinerent. Here the dominant weather systems move 800 km overland before reaching the headwaters of the Bow basin. Thus atmospheric loading in the basin should have a signincant tenestriai component tbat is derived from outside of the basin. This is readily observed on the westem margin of the basin, where precipitation is heavily enriched in Ca, Mg, and SO, relative to costal min (Table 4.1). As the majority of Bow River water onginates in the western part of the basin, the most reasonable way to correct 41 for cyclic sait input in the Bow River is to use the average annual loadhg rates for wet and dry fdin the headwaters of the basin. Annuai wet and dry deposition rates were taken fiom Legge's (1988) region 9. This region represeats the integrated deposition hmthe western margin of the basin to Calgary. Tbe contribution of the atmosphenc load to the river can be calculated by averaping the annual wet and dry deposition over the yearly discharge. Table 4.2 illustrates that atmosphenc loading is a signif'ant source of some ions in the river (50% of K, 17% of SO, and 16% of Cl). In contrast, dissolved NO, in the river accounts for only 4% of atmosphenc NO,. Nitrate is generdy thought to be lost from basins by denitrification reactious in the soi1 zone (Schlesinger, 199 1).
Table 4.1 Chloride normalised equivalent ratios for average concentratiom in precipitation. hmWest (Kananaskis) to east (Suffield), from Myric (1992). and world average coastal, from Berner and Berner (1987).
element Kananaskis Calgary Suffeld Coastal Rain
2) Ion exchange and biologic uptake Ion exchange reactions are much more rapid than chernical weathering reactions (Cresser and Edwards, 1987) so the net flux fiom ion exchanges processes should be minimal for basins where there hasn' t ken any sipnincant change in input, or the basin environment (e.g. acid rain, or anthropogenic activity). Similarly, if the net biomass in the basin has remained relatively constant through the, then the flux related to biomass should be minimal. Since the headwaten are in a protected park, anthropogenic impact has been 42 minimal. There have been no major fkes in BMNational Park or Kananaskis since 1940 due to fire suppression (Whh. 1985). leading to the deveiopment of a stable biomass. As weîl, the pH of precipitation has remauied relatively constant at 5.5 over the 1st 18 years (Lau and Dass, 1985; Myric, 1992). Given the above, changes in both ion exchange and biologic uptake should k minimal during the pend analysed. ln addition, carbonate weathering rates are bigh enough that the weathering flux wouid overwhelm any flux hm ion exchange and biologic upstake @river, 1997).
Table 4.2 Mean annual wet and dry deposition rate (kg/ha/y) averaged over the head waters of the Bow River (fkom the Great Divide to Calgary), from Legge (1988), and calculated atmospheric component of major ions in river water at Banff (annual atmospheric loading averaged over annual discharge).
Bow River atmospheric Species wet rate dry rate total component 5% kg/hd~ ~WYkg/hd~ mgfl mg/l atmospheric
3) Anthropogenic Input Visitor use in Banff National Park appears to have minimal impact on the water chemistry of the Bow River. The two most notable effects on the inorganic chemisüy are elevated phosphorus levels related to sewage input (Block and ZaU, 1988), and a 0.2 mu increase in minimum Na concentrations during the 1980's, attributed to road salting and water sofieners (Block et al., 1993a). Weathering reactions conîroIting river chemistry
Given the above, equation 4.1 can be simplified by dropphg the terms F,, Fion-hs, and F-=, reducing it to:
The two basic implications of equation 4.2 are th& 1) the measured river chemistry only
needs to be comcted for atmosphenc input, and 2) despite the various inorganic and organic systems operating in the basin, the dissolved load of the river can be modeled in terms of primary weathering reactions anticipated for the geohgy of the basin (Table 4.3).
Table 4.3 Dominant weathering reactions anticipated for the Bow River basin, based on relative weathering rates of doahant minerds (Lasaga et al., 1994).
- -
Ca CO, + &CO3 * Ca + 2HC0, (1) CaMg(CO3, + 2H$03 - Ca + Mg + 4HC03 (2) CaSOp Ca + SO, (3)
NaCl --. Na+ Cl (4)
2FeS, + 4qO + 60, FqO, + 8H + 40, (5)
&SO, + 2CaC03* 2Ca + SO, + 2HC0, (6)
2H,CO, + 9H20 + 2NaAiSi,O, iUS~05(OH),+ 2Na + 2HC0, + 4H,Si04 (7) 44 Of these reactions, the dissolution of NaCl is the easiest to quantify becaux Cl ôehaves consematively in surface water. Sewage effluent at Ban£€ has less than 1 mg/l Cl (Block and Zall, 1988). thus al1 Ci after correction for atmsopheric hput can be assumed to be denved from dissolution of evaporite rninerals. Cumulative Cl concentrations. hm1978 to 1995. are plotted for Lake Louise and Ba& in Figure 4.5. Overail. Cl concentrations are lower at Lake Louise than downstream at Banff. For Banff, the Cl concentrations are rdatively high in winter and close to zero during spring and summer. This cm be related to groundwater hput king diiuted during spring and summer by snow melt and precipitation. An interesting feature at Lake Louise is that there is a slight increase in Q in early spring, before the majority of the snowpack has begun to melt. A possible explination could be initiai melting of the snow pack at low elevations. where snow almg the Icefields Parkway (HW 93) would contriiute Cl fiom winter road salting. If Na and K were only denved fiom dissolution of evaporite minerais, then (Na+K) should balance Cl. A plot of (NatK) and Cl (Fig. 4.6) illustrates that (Na+K) is in excess. On average, 90% of (Na+K) is accounted for by dissolution of halite. One potentiai source of excess Na is weathe~gof detntal albite in the Miette Group (reaction 7 in Table 4.3). The abundant carbonate rock in the basin, and the rapid weathering rate of carbonate mllierals, suggest that dissolution of carbonate miaerals wül add signincant amounts of Ca and Mg to the river. As would be expected, the Bow River plots among rivers draining carbonate terrains, aad well away Born rivers draining silicate rocks (Fig. 4.7). The most common weathering reaction for carbonates is simple dissolution (Drever, l988), reactions 1 and 2 in Table 4.3, givhg a (Ca + Mg):HCO, equiiivance ratio of 1:l. For Banff, (Ca+Mg) balances alkalinity at low concentrations (Fig. 4.8), however, (Ca+Mg) progressively deviates hmthe 1: 1Lure as alkalinity increases. 33a)- buse
8 J FMAMJ JASON0 J FMAMJ J ASOND
Figure 4.5 Cumulative plot of monthly Cl concentrations at a) Lake Louise and b) Banff, corrected for atmosphenc input.
Figure 4.6 Na+K vs. Cl for a) Lake Louise, and b) Banff. 46 Two possible sources of additional Ca+Mg are oxidation of sulphides (cornbined reactions 5 and 6). or dissolution of gypsum and anhydrite (miction 3). These two sources can not be disthguished solely based on stochiometry, since both reactiom require tbat (Ca+Mg) be baianceci by (HCO, + SOJ, flabIe 4.3, Figure 4.9). The stable isotope composition of SO, can be used to distinguish SO, derived from evaporite minerals and oxidized sulphide minerals. Stable isotope data presented in Chapter 5 indicate that the major@ of SO, measured at Banff is derived nom dissolution of evaporites. men plotted against discharge (Fig. 4-10), the SO,:(SO, + Alk) ratio at Banff is highest during low flow, when groundwater discharge is the dominant source of the river. This could be a function of evaponte minerals being leached out in the near surface, andor grouadwater having a longer residence tirne dowing more dissolution.
The deviation from the 1: 1 Ca+Mg:Ak is not sipificant at Lake Louise. As well, there is no seasonal variation in the S04:(S04 + AUc) ratio as seen in Banff, suggesting evaporites are not as common. This is consistent with the local geology; evaporites are scarce in the Cambrian carbonates upstream of Lake Louise, where as they are more common in the Devonian carbonates exposed between Banff and Lake Louise (Mossop and Shetsen, 1994). In su-, the major elemnt chemistry of the Bow River, after correction for atmospheric input, is dominated by dissolution of carbonate minerals. Additionai Ca and Mg, as weil as SO, is added to the river downstream of Lake Louise by dissolution of gypsum and anhydrite. This is more signiticant during base flow when ground water is the major source of water in the river. Na and K are largely derived fiom dissolution of evaponte minerals. Again, this is more sipnincaut downstream of Lake Louise. Excess (Na+K) is likely derived fkom weathe~gof feldspars. At Lake Louise, the rise of Cl in
April may be related to inital melting of snow near the Icefields Padovay. O carbonate O shale O sanelstone Igneiss A granite A volcanic + Bow River
Figure 4.7 Temary diagram of major ion composition of the Bow River and rives draining a varieîy of lithologies throughout the world (after Meybeck and Helmer, 1989).
Figure 4.8 Ca+Mg vs. HCO, for a) Lake Louise. and b) Banff. Figure 4.9 Ca+Mg vs. HCO, + SO, for a) Lake Louise, and b)Banff.
La)Lake Louise
O 20 40 60 80 O 100 200 300
discharge m3/ s discharge ma/ s
Figure 4.10 Discharge vs. SO,/totai anions for a) Lake Louise, and b) Banff. Examuiing the major ion stoichiomeûy provides information on weathering reactions controiiing the input of ions to the river system, in contrast, thermodynamic analyses on help illustrate equiliirium rractions controlling the water chemisw There are limited snidies that examine equiltirium feactions of river water. Norton (1974), examining the Rio Tama River in Puerto Rico, used activity ratios to suggest that weathe~g
stoicâiometry was the primary control of river chemistry. However Drever (1988) reinterpreted Norton's (1974) data, and suggested tbat cation exchaage between ma- smectite and Wg-smectite is the dominant control. Similarly, in examining numerous rivers draining basaltic terrains, Bluth and Kump (1994) note that cation activity ratios always plot on trends paralieling smectite exchange boundaries. In their words, "if river chemistry depended only on bedrock composition and weathe~gstoichiometry, we would expect a much pater variation amoag the cation activity ratios". Bluth and Kump (1994) conclude that the major cation ratios are likely controiIed by smectite exchange reactions. The role equili'brium reactions play in controlling the chemisay of the Bow River is investigated by coastructing activity-activity diagrams. Activities of major ions are calculated using the geochemical modeling package SOLMINEQ.88 PUSHEU. (Wiwchar, et al., 1988). Ca and Mg activity ratios plotted in Figure 4.1 1 &fine a iine with a dope of 1 (9 = 1). The strong correlation of measwed activities to a slope of 1 suggests a 1: 1 Ca- Mg exchange reaction is contmlling the element ratios in the river. In order to identify potentiai exchange reactions, minerai stability boundaries were calculated at 5 OC by the program PTA (Brown, et al., 1988), for the system: Ca, Mg, Al, Si, C, 0,and H. The range of possible actions was resûicted by only considering common minerals in sediment.rocks (Table 4.4). Of over 7000 stable and meta-stable reactions possible, only 2 have a 1: 1Ca-Mg exchange boudary in the range of activities measured in the Bow River: Calcite + Mg = Dolomite + Ca
Recipitation of dolomite is rare in modem environxnents, iadicating the only potentiai reaction controiling the major ion ratio of the Bow River is equation 4.3.
Table 4.4 Ca and Mg bearing mine& potentiaily found in sedirnentary rocks.
Anorthite Brucite Clinochlore Gibbsite Kaolinite Wairakite Ca-beideIIite Mg-beidellite Heuiandite Calcite Dolomite
The CaMg exchange reaction in equation 4.3 is investigated fiutber using methods outlined in Abercrombie (1989). Smectite is rnodeiied as a mixtue of Ca-, Mg-, Na-. and K-beideihte end member components. The activities of these components are unknown, but can be caicuiated for an individual water sample ifequil'briwn is assumed. The sum of activites of the individual smectite components would equal 1: Three indepentant cation exchange reactiom can be written, with their equiliirium constants defked as:
Equüirium constants for these reactions are cdcdated using EQCALB, a version of the program EQCALC (Flowers, 1986) modified to use thermodynamic propeaies of minerals tabulated by Bernian et al. (1985). Thus, Equations 4.4 through 4.7 can be solved simdataneously to obtain the activites of the four beidellite components in equili'brium with a given water sample. Activities of the beidellite components were calculated for samples 80/04/23 fkom Banff and 79/06/06 from Lake Louise flable 4.5). The calculateci activties iodicate that smectite in equilibirum with the river water would be dominantiy Ca- and Mg- beideilite, with subordhant Na- and K-beidellite. Using the calculated activity of the smectite components, miueral stability boundaries are caicuiated by the program FTA (Brown, et al., 1988). Because samples 79/06/06 for Lake Louise, and 80/04/23 for Banff, were used to calculate the activities of the sxnectite components, they must faIl on the caiculated exchange-reaction boundary. However, the location of the other river samples are independent of the calcuiated boundaries. If the activities nom water analyses cluster dong a phase boundary, it suggests that the water 52 may be in equiliirium with the exchange reaction (Hutcheon, 1989). Water samples are plotted on the calcuiated phase diagrams in Figure 4.12. The measured activy ratios of Ca and Mg show a stmng comIation with the calulated exchange boundary, arguing that exchange between Ca- and Mg-beideilite is contmiling the Ca/Mg activity do..
Table 4.5 Caicuiated beideIlîte activities for smectite in equiliïrium with samples 79/06/06 for Lake Louise and 80/04/23 for Banff.
Smectite Lake Louise Banff component (79~06106) (8W04/23)
%-kidci~ia 0.4955 0.4955
For Na and K. the activity &ta are more scattered, but they do cluster around the caiculated Na-K beidellite exchange boundary (Fig. 4.13). Again, this suggests equiliiexchange between Na- and K-beiâellite is controUing the NaK activity ratio in the river. l ai Lake Louise
12 13 14 log a MgM2 log a Mg/H2
Figure 4.11 Plot of log aCa/a(H)2 versus log aMg!a(~)~for Lake Louise and BauK
l5 5a) Lake Louise
log a Mg/H*
Figare 4.12 Plot of log aCa/a(H)' versus log aMg/a(H)2 as in Figure 4.1 1. with
calculated reaction boundaries at t bar and 5 OC. La)- Louise a O
Figure4.13 Plot of log aNa/H versus log aK/H for a) Lake Louise, and b) Banff-
Reaction boudaries are calculated at I bar and 5 OC.
10 11 12 13 14 15 log a MgMa
Figure 4.14 Plot of a) log aCa/a(~)' venus log aMg/a(H)', and b) log aNaM venus log aKRI for ground water (closed circle). Open circles represent high and low values for the Bow River data at Banff. Reaction bomdaries are for Banff,
calculated at 1 bar and 5 OC, 55 An important issue to address is: are equiiiiwn exchange reactions occwiag within the river, or are we obsenring an inherited signature of groundwater chemisûy? For the reactions to be occuring within the river, smectite wouid have to be present as suspended sediment. Given the cationexchange capacity of smectite (80 - 150 mq1100g; Drever, 1988). and an average Ca concentration of 19 meqll. 1.3 - 2.4 g/i of suspended smectite is required. This is significantly higher than the average total suspended load of 5.5 mg. In addition. suspended and bottom sediments hmthe river were analysed by XRD, and smectite is only a trace component This suggests that activity ratios observed in the river water represent hhented signatures of equilibrium reactions occurrjng in the soii and ground water zones. This is supported by the fact that activity ratios calculated for shallow ground water samples collected by Park Canada (Appendix 3), upstrearn of the Banff town site, have the same range of activity ratios as river water (Fig. 4.14). If we are observing an inherited signature of groundwater chemistry, then smectite must be present in the basin. Smectite is commonly associted with volcanic ash layers in Crrtacous shales in the basin. There are few studies on the clay minedogy of older units in the basin, however, ash layers have been observed in the Banff and Exshaw formations (G. Davies, pen. communication, 1997). The presence of smctite bearing units supports the interpretation that the activity ratios observed in the river are an inherited gromdwater signature. The implications of river chemistry being an inherited signahue of groundwater chemistry is discussed in relation to basin hydrology in Chapter 5.
Chernical denudation mte
The river chemistry discussed above only provides infomtion on weathering processes, not on weathering rates and mass tramfer fiom the basin. By combining chexnical data with discharge information, chemicai denudation rates in the Rocky Mountains can be 56 caicuiated Annuai denudation rates are caiculateA by nomaiking the weighted average
annual flux (C, x Q ) over the basin area (421 kmt above Lake Louise and 22 10 kmZ above Banff Padc gate). Where weighted average aunual concenaaiion (C,) is defieci as the average of the monthly samples (Zemm, 1978; White and Blum, 1995):
where C, is the concentration of the individual Stream sample, is the discharge during
the sampling interval j to j-1, and Q is the total mual discharge. htantaneous flux is
defmed as (Cax Q), where Q is the daily discharge. Figure 4.15 Uidicates instantaneous flux is a function of discharge, consistant with most world rivers (Berner and Bemer, 1987). Figure 4.15 also shows that the majonty of the dissolved ioad, approximately 75%, is transported during the summer months, when concentrations are lowest. The weighted average chemical &nu&tion rate (CDR) for the Roclues upstream of Banff, after conecting for atmosphenc loading, is 997 kghaly. If reaction I in Table 4.3 is the dominant meam of carbonate weatherins, then only one half the HCO, in the river is from rock weathering. This reduces the CDR to 678 kgha& or 1.5 x ld kg of rock removed as dissolved load each year flable 4.6). This falls in the typical range for carbonate basins (Table 4.7). The denudation rate above Lake Louise is ody slightly lower
at 804 kgmaly, or 516 kgha& after correction for HCO,. The lower denudation rate upstream of Lake Louise is iikely related to the large exposures of silisiclastic rocks in that part of the basin. The Bow River is unusual in that the suspended load is much less than the dissolved load. The non-nlterable residue (NFR) measund by Environement Canada at
Banff averages 2 mgll, w hich translates to an additional 11 kgmaly , or 1.1% of the to ta1 57 load IDconûast, the North Amencan average suspended Ioad comprises 70% of the total load (Schiesinger, 1991), and carbonate basins in the Himalayas have ~pendedloads that are 90% of the total Ioad (Sarin et al., 1989; Berner and Bemer, 1987). The iow suspended solids load may be reiated to the fact the vaüey flwrs of the basin are heavily forested, Ibducing the amount of loose sediment that cm mach the river.
0000000 J FMAMJJ ASOND V)oV)oV)o r - CU CU m discharge m3/s mont hs
Figure 4.15 Iostantaneous daily mix versus a) discharge and b) cummunulative instantaneous flux, measured at Banff. Table 4.6 Weighted average long term denudation rate (kg/ha/y) for the Bow River at Banff and Lake Louise (total flux - atmosphenc input).
K -5 .9 so4 33 106 HCO, 575 638 SiO, 15 18 cl 2 4 CDR 804 997 CDR - corrected 5 16 678 for HCO,
Table 4.7 Chemical denudation rates (CDR)for world nvers, and world average rate. CDR River Basin Type kdw~ Re ference Yamunu Carbonate 1430 Sarh et al. (1989) Indus Yagtze 1 Ming-hui et al. (1982) Hwangho 1 Brahmaputra 1 Edmond (1982) Meykong 1 Zaire 1 Amau>n 3 World 1 Berner and Berner (1987) 59 CONCLUSIONS
Discharge is the dominant control on the TDS load of the Bow River. Spring melt and summer raias inmase discharge and düute groundwater input to the river. Although discharge is the dominant contrd on concentration, the source of ions in the river is controlled by atmospheric deposition and water/rock interaction. Atmospheric Ioadbg oui be a signincant source of some ions in the pristine headwaters of the river (e.g. 50% of JC, 17% of SO,, 16% of Cl). In terms of water/mk interaction, the input of ions to the river is largely controiled by dissolution of carbonate and evaponie miner&. Cation exchange reactions exert a strong control on element ratios in the river. The WMg activity ratio is strongiy controlied by exchange between Ca- and Mg- beidellite. NaK activity ratios are controUed to a lesser &pe by exchange between Na- and K- beidellite. These activity ratios appear to be ïnherited signatures of ground water. The fixed element ratios in the river suggest that that both snowmelt and rainfall mut pass through the ground before reaching the river.
The chernical denudation rate for the Bow River at Banff is 678 kg/ha/y, or 1.5 x 108 kg of rock that is removed as dissolved load each year, consistant with CDR of other carbonate basins. Up to an additional 11 kg/ha/y are removed as nispended load. Ushg a density for ibnestone of 2.75 @cm3 (Daly, et al. 1961). this weathe~grate represents a rock volume, carried by the Bow River each year, of 54,545 m3. CHAPTER 5 Chernical dynamics of the Bow River
INTRODUCTION Chapter 4 examined geochemical coatrols on water chemistry in the pristine headwaters of the Bow River. This chapter examines variations in the chemical and stable isotope composition dong the length of the nver, eom the headwatea above Lake Louise to the confiuence with the Oldman River. Characterishg controls on the chemistry of the river is more complex than in Chapter 4 because the basin is less homogenous, and there is more anthropogenic activity downstrearn of Banff National Park. An integrated chemical and stable isotope approach is used to trace point and non-point source inputs to the river-
Four sample sets, representing fail, winter, spring, and summer, were collected to characterise the chemicai variation along the length of the river. Main tributaries to the nver were also sampled. Sample locations are shown in Figure 1.1. The basin can be subdivided into four main segments. The Bow River, ftom Lake Louise to Morely (sample stations BR1 to BR4, Fig. 1.1) drains the near pristine Rocky Mountains. The foothills area between the Rocky Mountains and Calgary (BR4 to BR6) is largely used for ranching, but is otherwise undeveloped. Calgary, the only large urban centre in the basin. lies within stations BR6 to BRIO. The city has numerous storm sewer out falls along the river, and two outlets for treated domestic sewage fiom the Bonny Brook and Fish Creek treatment plants. Effluent fiom the plants enters the river between sample stations B R9 and BR10 (Fig. 1.1, Appendix 1). Downstream of Calgary, agriculture is the dominant land use (mostly cereal grains and canola crops). Before examinhg the chemical dynamics of the Bow River itself, it is necessary to characterise the chemistry of major point source inputs dong the river. 61 CHEMICAL CHARACTERISTICS OF POINT SOURCE INPUTS The main point source inputs to the Bow River are tniutaries, Storm sewer outlets, treated domestic sewage, and irrigation retum ffow dong the river. The main tributaries to the Bow River ali have headwaters in largely undeveloped basins in the Rocky Mountains (Fig. 1.1). The aibutaries have Ca-Mg-HCO,SO, type waters, simiiar to the Bow River in its headwaters (Chapter 4). With the exception of Nose Creek, the cation ratios of tributary waters are characterised by a fixed CalMg ratio and low Na (Fig. 5.1). Cation ratios of triiutary waten are similar to each other and the headwaters of the Bow River (Chapter 4). This is consistent with the fact that tributaries are drainhg essentially the same rock units as the Bow River in the headwaters. Anion ratios of tributary waten are characterised by low CI, and S0,fA.k ratios that have similar seasonal variations as the Bow River at Banff (Chapter 4). With the exception of Nose Creek, the total dissolved solids (TDS)loads of tributaries are also in the same range as the Bow River (Fig. 5.2). Because tributaries have the same chemical makeup and variability as the Bow River, the addition of tributary water wiU not greatly affect the chemical characteristics of the river, although tributaries will affect mass balance calculations (discussed below). Nose Creek has anomalously high TDS, and a significantly difterent chemicai makeup compared to the Bow River and other tributaries (Fig. 5.1). However the volume of Nose Creek is too small to alter the Bow River composition (Table 2.1). Factors controlling the chemistry of Nose Creek are addressed separately in Chapter 6. 0 storm sewer 0 effuient @ irrigation return 0 Bow River aver.
B-. aroundwater Na+K Alk [ O spring water
Figure 5.1 Temary diagram, in equivalents, showing major ion composition of tributaries to the Bow River, as weil as storm sewer discharge and effluent fiom waste water treatment plants in Calgary. Tniutary samples emïched in Na+K and Cl are from Nose Creek (NC).
The major ion composition of storm sewer discharge in Calgary is similas to Bow River water (Table 5.1, Fig. 5.1). and has a simiiar range in TDS (160 mgll in the summer to 300 mgA in the winter). Thus, like the tributaries, storm sewer discharge wili not greatly affect the major ion chemistry of the river, although they may dilute the river in the summer. In contrast, effluent from water treatment plants in Calgary has significantly higher TDS (575 to 875 mgll) and is enriched in Na, Cl and SO, compared to the Bow River (Table 5.1, Fig. 5.1). Irrigation return flows have TDS levels within the range of the Bow River (280 mg/l) and are e~chedin Na compared to the river. Return flows generdy discharge water to the river during summer irrigation. Range of TûS for the Bow River
Figure 52 TDS of tributaries to the Bow River.
Table 5.1 Chemistry of storm sewer discharge (this work), effluent from the Bomy Brook and Fish Creek sewage matment plants, and average composition of Mgation retum flow (fkom Sosiak, 1996).
storm sewer Bonny Brook Fish Creek irrigation ion average effluent* effluent* retumflow* (mm mgm (mp;li) (md) Ca 44 55 70 35 Mg 13 25 33 17 Na 2.7 82 118 15 K 0.8 16 10.5 2.0 HCO, 148 185 325 153 SO4 28 142 165 54 Cl 2.3 68.4 133 5.7 *hSosiak (1996) 64 CHEMICAL CHARACTERISTICS OF THE BOW RlVER Waters nom the Bow River are dominantly Ca-Mg-Alk-SO,. The TDS load of the river varies seasonally, as weU as dong the river. The highest TDS values are observed duriog baseflow conditions in the id and winter, where as the lowest TDS levels are observed during the summer rainy season Fig. 5.3). Intermediate levels are observed during peak discharge related-to spring snowmelt. The following dowmtream variations in TDS are observed in all four sample sets (Fig. 5.3): 1) the lowea TDS values are observed in the headwatea, 2) there is a general trend of increasing TDS as the river fiows through the foothills to Calgary, 3) TDS drops as the river flows through the City of Calgary, particularly in spring and summer, and 4) downstream of the sewage treatment plants in Calgary there is a noticeable increase in TDS. In the fd, TDS remains constant through the agriculnual areas downstream of Calgary, whereas in spring and summer TDS
A summer 50 I I I I I O 8 8in 8 n8 8CI( 8 Distance from the Oldman canfiuenœ (km)
Figure 5.3 Total dissolved solids load versus distance dong the Bow River for faii, winter, spring, and summer. 65 The variations in TDS as the Bow River fiows through Calgary are consistent with point source loading hmStorm sewers and efnuent fkom treatment plants. Storm sewer discharge, with relatively low TDS, would dilute the river as it flows through the city. This is most noticeable during the rainy season in spriag and summer Fig. 5.3). In contrast, loading hmwaste water treatment plants increases the TDS levels of the nver above those upstream of Calgary. Despite variations in TDS dong the length of the river, element ratios are relatively fixed (Figure 5.4). The Ca/Mg ratio of Bow River water is nearly constant. However, waters from upstream and downstream of Calgary plot in iwo distinct clusters, where waters downstream of Calgary are more enriched in Na+K. In order to examine controls on the observed seasonal and downstream variations in the nver chemistry, individual ion chemistry is discussed below.
Mg Na+K Alk
Figure 5.4 Temary diagram, in equivaleats, of major cations and anions of Bow River water for di four seasons. CONTROLS ON THE MAJOR ION CHEMISTRY OF THE BOW RIVER
Chloride is the easiest ion to account for because of its conservative nature and limited sources. The main sources of Cl in rivers are sea sdt (fiom tain fall), weathering of halite, and pollution (Berner and Berner, 1987). The downstream variations of Cl concentrations in the Bow River are plotted for each season in Figure 5.5 (distances are measured fiom the confluence with the Oldman River, the end of the Bow River). Cl concentrations are Iow (around 0.5 meq/l) and slightly variable upstream of Calgary. Concentrations are within the range observed at Banff in Chapter 4, and thus represent cornbined atmospheric loading and evaporite dissolution. There is no loading of Cl between Banff and Calgary. In contrast, the concentration of Cl increases nearly 4x as the river flows through Calgary (the Calgary city limit starts at Bearspaw reservoù in Figure 5.5). Cl is a common pollutant from municipalities, denved fiom domestic sewage and road salt. The most significant increase in Calgary occurs between the hua sampling stations that bracket the outlets for the Bomy Brook and Fish Creek sewage treatment plants. Effluent waters are significantly e~chedin Cl compared to the Bow River (Table 5.1, Fig. S. 1). As the river flows through agricultural land downstream of Calgary, there is no additionai loading of CI. There are two spikes in Cl concentrations downstream of
Calgary in the spring and summer. It would be difocult for Cl concentration to decrease as rapidly as observed without significant amounts of dilute water king added to the river. What these spikes likely represent is effluent from the sewage treatment plant that has not been Mymixed with the river. Cross et al. (1986) estimate that the mixing zone for the effluent is over 20 km long. distance fmOldman cmfiueme (km)
Figure 5.5 Dowstream variation in Cl concentratiom for the Bow River.
Sodium
The downstream variations of Na concentrations in the Bow River are ploned for each season in Figure 5.6. Like Cl, Na concentrations are low upstream of Calgary, and increase signincaatly as the river passes the outfall for the Bomy Brook and Fish Creek sewage treatment plants. The most common sources of Na fiom municipalities are NaCl, N%C03, NaSO,, Na-borate, and other Na salts used in industry (Bemer and Berner, 1987). As weii, Na is commonly derived fiom Na,C03 and Na-zeolite used as domestic water softeners. diitance from Oldman duence (km)
Figure 5.6 Dowstream variation in Na concentrations for the Bow River.
Loading of Na fkom Calgary causes a significant shift in the (Na+K):(Ca+Mg) ratio of the river (Fig. 5.7), resuiting in the two data clusters observed in Figure 5.4. The (Na+K):(Ca+Mg) ratio continues to rise as the river flows through agricultural areas downstream of Calgary, particularly in the sp~gand summet* This increase is solely due to loading of Na+K, as Ca and Mg concentrations remain relatively constant downstrram of Calgary (Appendix 2). Although both Na and the (Na+K):(Ca+Mg) ratio increase progressively through the agricultural areas downstream of Calgary, Cl concentrations nmain constant (Figs. 5.5.5.8). This desout road salt or other fonns of NaCl as the source of Na. The progressive increase in Na concentrations is coincident with a progressive e~chmentin 6D observed in Figure 3.10. As was indicated in Chapter 3, this shift in 6D is related to addition of water relatively e~chedin D to the Bow River. The coincident increase in Na implies that the added water must also be eariched in Na relative to the Bow River. The most obvious source of this water is irrigation return flow. This is consistent with the observations that 1) Na is king loaded progressively dong the river, and 2) Na is oaly king loaded in the spring and Nmmer when irrigation retum canals discharge to the river.
distance from Oldman confluence (km)
Figure 5.7 Dowmtream variation in the Na+K/Ca+Mg equivaience ratio. Figure 5.8 Na + K versus Cl for the Bow River (open circles) and tributaries (closed circles) for a) fd,b) winter, c) spring, and d) summer.
The role smectite exchange reactions play in controllhg Na and K ratios dong the length of the river is investigated hen using methodology outlined in Chapter 4. The activities of individual beidellite components were calculated assuming equilibrium with the Banff sample for each season. Smectite in equilibrium with river water at Banff 71 wodd be dominantly Ca- and Mg- beidellite. with subordinate Na- and K-beideiiite (Table 5.2). Because the Banff samples were used to calculate activities of the beideilite components they must faon the calculated exchange reaction boundary (Abercrombie, 1988). However, the location of other water samples are independent of the cdculated boundaries. Water samples nom the river and main tributaries cluster around the calculated phase boundary (Fig. 5.9), suggesting that the Na-K activity ratio in the river is hxed by smectite exchange reactions.
Table 5.2 Caicdated beideiiite ac tivities for smectite in equilibrium with samples colIected at Banff for each season.
Smectite component fall winter spri. summer
%kidcilie 0.4929 0.4939 0.49 19 0.4956 -/b) winter
NaBd + K = KBd + Na
2 3 4 5 6 7 2 3 4 5 6 7 log a WH log a WH
Figure 5.9 Na-K activity plot for samples dong the Iength of the Bow River (open circles), and main tributaries (closed circles), for: a) fa, b) winter, c) spring, and d) summer.
Sulfate in river systems can be derived from a variety of sources such as: the dissolution of sulfate minerais, oxidation of pyrite or other reduced forms of sulfur, and anthropogenic input fiom fenüizer, industrial emissions fkom sour gas processing, or municipal effluent. If any of these potential sources have an unique ratio of 34 SI32 S, then 73 it is possible to use the 6% vaiue of dissolved datein the river to trace S input. Figure 5.10 illustrates the 634~values of potentid S sources in the Bow River Basin. Background values for precipitation in the basin are +5 to +IO%, but nach + 23%0 near Sour gas plants (Norman, 1991). However, SO, in precipitation is not a sipnincant source of SO, in the river, it comprises at most 17% of the S flux (Chapter 4). Sulfate hmtills in the basin is largely derived nom reduced S, and thus has negative 634~values in the range of -8 to -12 %O (Hendry et al, 1986; 1989; Ferne& 1994). The 6% of S-based fertilizer sold in the basin was measured to be +14%0, S-based fertiIizer is not commonly used as soils in the ana generally have sufficient S for most crop plants (Aiberta Wheat Pool - personal communication). Sour gas processing plants emit S with approximately the same range of 6% values as evaporite minerais in Devonian carbonate rocks in the basin. Springs discharging from Devonian carbonates in the Rocky Mountains (the dominant source of evaporite minerals) bave 6MS, values ranging bm+17 to +25%0.
till - effluent from sewage - treatrnent plants
evaporite minerals 1-1 precipitation -spnngs 1-1
Figure 5.10 6% of potentid inputs of S to the Bow River. Downstream variations in sulfate concentration and 6Y~,,, are plotted for each season in Figure 5.1 1. Sulfate concentrations vary nom 2 mgn in the headwaters, to 50 mgA near the confiuence with the Oldman River at Hays. 6M~,, varies inversely, decreasing fiom +2(%0 in the headwatea to 5%0 at Hays. In the winter, both sulfate concentrations and the 6%~~remain relatively constant along the river. Loading fkom the sewage trcatment plants at Calgary is ody significant in the faand winter, when discharge in the river is low. Sulfate concentrations remain relatively constant through the agricultural areas downstream of Calgary in the fd, but increase in spring and surnmer. The inverse relationship between SO, and 6Y~so4along the length of the river indicates that two or more sources of SO, with different SYsS, values are mixing dong the length of the river. By plotthg the inverse of SO, concentration veaus YS,,, it is possible to determine the 6US,value of S04king added to the river if simple mixing of two or three end rnembers is taking place. For the fd, sp~gand summer sample sets, a minimum three component mixing is indicated (Fig. 5.12). Samples fiom Lake Louise have Ps, values fiom +17 to +2M~,in the typical range for evaporite minerais. There is a significant increase in SO, concentration from Lake Louise to Banff, with a concurrent shift in the 6%&,. From Banff to Calgary, concentrations of SO, and the 6%,, remain relatively constant. The best fit lines, from Lake Louise through the data cluster representing Banff to Calgary, give intercepts of +8 to +1û%0. If simple mixing is occuring, +8 to +1û% is the 6% of SO, king loaded to the river. A value of +lm0 is consistent with SO, derived from soils. However, Mayer et al. (1995) show that forest soils tend to be a suhsink rather than a sulfur source. The more iikely origin of the SO, is a mixture of the two dominant S sources in the bedrock, SO, denved from evaporites and oxidized pyrite. Taking +19%0, the 6WSso4value at Lake Louise (Appendix 21, to represent the average composition of evaporites, and -10% as the average value for pyrite (Hendry et al, 1986; 1989; Femeu, 1994), then oxidized pyrite comprises approximately 3 1% of the SO, king loaded. A second mixing relationship is apparent downstream of Calgary. The 6%~~ progressively decreases as the concentration of SO, increases. Best fit lines drawn for the data downstrearn of Calgary have intercepts that indicate a FS, value of -10% for SO, king loaded to the river in the fail and spring, and -2%0 in the summer. A value of -lM~ is consistent with the range of values observed for SO, in tills in the basin (Hendry et al, 1986; 1989; Ferneil, 1994). The more positive vdue (-2%0) for the summer is not consistent with any primary source of sulfùr in the basin. This Uely represents sulfate fiom till that has mixed with a source more e~chedin "S. Sulfur derived fkom sour gas operations is unlikely as the more positive 6U~,,4values are ody observed in the summer. Given that the summer represents the active fanning season in the basin, soi1 sulfate is a possible source. Unlike stablised forest soils that are sulfur sinks (Mayer et al., 1995), tilled soils in the prairies may be a sulfur source due to oxidation of organic sulfur. Taking -Ima as the tiü component, and +6%0as the soi1 component, a -2% value would represent a 1: 1 ratio of SO, denved £rom soils and till. The winter samples suggest a single SO, source is king loaded to the river with a 6%,, value of +12%0. As interpreted above, this Likely represents the combined signature of S from evaporites and oxidized pyrite.
8'0 in sulfate The 6'8~in sulfate provides additional insight into the sulfur cycle in the basin. The dowwtream variation in 6'8~,, and 6% is plotted in Figure 5.13. With the exception of the winter samples, the PO, values are lowest at Lake Louise, and increase through to Calgary. Downstream of Calgary the 6180,, values remain relatively constant. In the winter, the 6'b,, values remain constant dong the river, similar to 6%. Figure 5.11 Downstream variation in the concentration of SO, and 69, for the Bow River in a) fall, b)winter, c) spring, and d) summer. P * - dow nstrearn
Banff to Calgary
10
Figure 5.12 PSversus l/SO,for a) fall, b) winter, c) spring, and s) summer. distance hmOldman confluence (km) dinœfrom Oldman confluence (km)
Figure 5.13 Downstream variation in 8% and 6180 in SO, for a) fall, b) winter, c) spring, and d) summer.
The 6'b, is plotted versus 6'b,,, in Figure 5.14- 6'b, has a narrow range, between +6 and -6%0,and varies independently of S"O,,,. The O in SO, is depleted in 180compared to the typical range for evaporite minerals (6f80, = +10 to +i6%0, Claypool et ai., 1980). Exchange reactions between water and sulfate cm not be invoked 18 to explain the depleted 6 O,,, because the SO, - $0 exchange is extremely slow at temperatme and pH conditions comparable to the Bow River (Lloyd, 1968). As well, the 6'8~, values are not consistent with oxidation of reduced sulfur, with the exception of samples from Lake Louis, as waters fiom the Bow River plot outside of the theoretical sulfide oxidation field of Van Stempvoort and Krouse (1993) (Fig. 5.14). The lower boundary of the sulfide oxidation field represents 1006 conmiution of oxygen in SO, from H20, The upper bounâary is derived from num&ous experiments (e.g. Taylor et al., 1984) of sulfide oxidation and represents the minimm contribution of oxygen from H,O (approxirnately 62% of O in SOJ. Thus the sulfide oxidation field dehes the range of 6'8~,, possible when suifide is oxidized in the presence of water with a given S'b,,. There are two possible explanations for the observed 6'8~,,, in the Bow River L) a mixture of sulfate fiom evaporite rninerals and oxidized sulfides, and 2) bacterial moderated sulfate reduction which preferentially reduces 32 S16 O (Vao Stempvoort and Krouse, 1993), leaving the remaining 6'*0,,) e~chedin "O. A mixture of sulfate nom evaporite minerals and oxidized sulndes does not secm possible given the downstream variation in 6''0[,, (Fig 5.13). At Lake Louise, the 634Sindicates SO, is derived from dissolution of evaponte minerals, whereas the 6180, plots in the suEde oxidatioa field (Fig. 5.14). This implies that most of the sulfate derived fiom dissolution of evaporite minerals is reduced to sulfide, and then reoxidized in the presence of groundwater before reaching the river. As well, as discussed above, the 6% indicates that SO, king loaded to the river downstream of Calgary is largely from a reduced sulfur source (oxidized pyrite and organic matter in tus). When sulfides are oxidized to SO,, the 6"0,, should plot in the suKde oxidation field. Although there is gypsum preseat in the eastern tiil, the 634~(-13%~ Fenneu, 1994) indicates it is derived from oxidized pyrite, and thus can 80
not be the source of the high values. The fact the St80, plots welî above the sulfide oxidation field suggests that the 6%, of the newly formed SO, has been
e~chedby sulfate ~duction.Similar e~chmentsin the 6'9, relative to the suitide oxidation field have been observed in tills in Saskatchewan (Van Stempvoort et al., 1994) and streams in Italy (Schwarn and Cortecci, 1974).
0 spring rn0" .-= O
Figure 5.14 #'O in SO, versus S1'O in H,O.
The basic implication of these data is that SO, in the river goes through a complex history of redox conditions before entering the river. This in tum supports previous arguments in Chapter 3 and 4 about the flow path of water feeding the river. Sulfate derived nom precipitation contributes at most 17% of sulfate in the river. Thus the vast majority of SO, in the river is denved nom some form of rock weathering. The 6% value of sulfate in the head waters of the river indicate that dissolution of marine evaporites is the dominant source of this sulfate. The 8180,,,, indicates that once dissolved in the water, this sulfate must be reduced and then reoxidized before entering the river. As well, S king Ioaded in the prairie regions must onginate as oxidued sudes, pass through the anoxic zone, and finally be added to the river. This implies that the water transporthg this S must move fkom the surface, through the anoxic zone, and then to the river, indicating that the river is largely gmundwater fed
Cakium and Magnesiuin
In most riven calcium and magnesium are derived dominantiy from rock weathe~g; they are not a common poiiutant (Berner and Berner, 1987). This is evident in the Bow River, where uniike the other major ions, calcium and magnesium are not king loaded by CaIgary (Fig. 5.15). The loading of Ca and Mg to the river is consistent with the local geology. Ca and Mg behave similarly, with concentrations increasing from the headwatea to Calgary. This trend occurs in the western part of the basin that is largely underlain by carbonate rock, and covered by the carbonate dominated Cordilleran till.
Where as downstream of Calgary, where the boundary between the Cordilleran till and the igneous and metamorphic dorninated eastem tili occurs (Fig. 2.6), Ca and Mg concentration remain relatively constant.
As stated above, calcium and magnesium are dominantly derived from rock weathering. The most common weathering reaction for weathering of carbonates is simple dissolution (Drever, 1988). reaction 1 in Table 4.3, giving a (Ca + Mg):HCO, equivalence ratio of 1: 1. Figure 5.16 iliustrates that (Ca+Mg) is in excess of alkaluüty by up to 25%. where as it is balanced by (alk + SOJ (Fig. 5.17). As discwed in Chapter 4, excess (Ca+Mg) cmbe denved nom oxidation of sulfides or dissolution of gypsum and anhydrite. Both processes require Ca+Mg to be balanced by Alk + SO,. Based on stable 82 isotope evidence discussed above, at least 7046 of the SO, king loaded upstream of Calgary is derïved from evaponte dissolution.
- - -- 0 fall Q winter O spnng A surnmer
distance (km) from the confluence with the Oiârnan
Figure 5.15 Downstream variation in a) Ca and b) Mg for each season. 1 2 3 4 5 1 2 3 4 5 alk meq/l alk meqll
Figure 5.16 Ca + Mg versus total ak for the Bow River (open circles) and tributaries (closed circles) for a) fail, b) winter, c) spring, and d) summer. 2 3 4 2 3 4 alk + m4meqll alk + 93, meqll
Figure 5.17 Ca + Mg versus a1.k + SO, for the Bow River (open circles) and tributaries (closed circles) for a) fd, b) winter, c) spring, and d) Sumner. Phase diagrams were used to examine the role smectite exchange reactions play in controlling the Ca:Mg activity ratio of the river. As above, phase boundaries were calculated assuming equilibrium with the nver water at Banff. Bow River waters show a strong correlation with the calculated exchange boundary, suggesting that Ca-Mg exchange reactions exert a strong coatrol on the river chemistry (Fig. 5.18). As in Chapter 4, it is important to examine whether Ca-Mg exchange reactions, and those controliing the Na:K activity ratios, are instream processes or inherited groundwater signatures. A number of factors argue that these equilibrium exchange reactions are an inhented ground water signature: 1) As calculated in Chapter 4. 1.3 to 2.4 g/l of suspended smectite is required to exert the observed conaols on activity ratios in the river. However, non-filterable residue, a measure of total suspended solids, is too Low along the length of the river (5 to 20 mgll) for there to be sunicieut smectite, 2) XRD analyses of suspended sediment along the river indicates that smectite is a trace component. 3) Activity ratios observed in the river are similar to those measured for discharge nom springs dong the river (Fig. 4.14). The basic implication of this is that the chemistry of the majority of water entering the river is controiled by waterlrock interaction. In hm, this indicates that, as previously discussed, the majority of water in the river must have passed ihrough the groundwater zone before reaching the nver. log a MgM2 log a MgH2
Figure 5.18 Ca-Mg activity plot, a) fd, b) winter, c) spring, and d) summer.
Bicarbonate
The carbon cycle in riven is complex due of the various sources and interactions dong the flow path. Most rivers in the world are characterised by large overpressures in pCO,, 10 to 15x equilibrium with atmospheric CO, (Pawellek and Veizer, 1994; Stallard and Edmond, 198 1, 1983; Buhl et al. 199 1). Carbon isotope studies of these riven attribute overpressures of CO, to oxidation of organic matter, generaliy attributed to enhanced organic productivity related to nutrient poilution (Richey et al. 1988; Buhl et al., 1991). In contrast. river systems that have natural interfluvial lakes (e.g. the St. Lawrence) or extensive impoundments (e.g. the Danube) tend to have pC02 values near, or below, equilibrium with ahwspheric CO, ( Yang et ai., 1996, Paweliek and Veizer, 1994). In these cases the long residence time dows pCO, to equilibrate with annosphecic pressures, and this in tum ailows partial isotopic equilibration with atmosphenc CO,. In the case of the Danube River, intense algal photosynthetic activity causes a significant &op in pC4 during the summer with a concurrent enrichment in 6°C due to preferential withdrawal of "C by algae. These shldies demonstrate that by examining the stable isotope composition of DIC, it is possible to elucidate the controls on the riverine carbon cycle. In order to do this, it is necessary to first dehe the 6I3cvalues of potential carbon sources. Pawellek and Veizer (1994) sllmmarise the various sources of dissolved inorganic carbon @IC) and their expected S13C values (Fig. 5.19). DIC can be derived from extemal sources via uptake of atmospheric CO,(-~XG),or carbon derived from dissolution of carbonate minerais (-5 to +2%0). DIC may ais0 be generated within the river by oxidation of organic matter (-24 to -31%) and respiration of aquatic plants. It is important to note that the isotope composition of these potential sources of DIC are altered on entering solution. Under equiiibrium conditions. the enrichment factor between DIC and atmospheric CO, is expressed as:
where T is in OK (Mook et al., 1974). Based on observed water temperahues, and 613cof atmospheric CO, of -7%0.the 6I3C of DIC in equilibrium with atmospheric CO, would be +3%0. The 8°C of DIC derived from CO2 from oxidation of organic matter wïU be in the range -14 to -24%. Carbonate rock is typically weathered by carbonic acid produced in the soil zone (Drever, 1988; reaction 1 in Table 4.3). The resultaat composition of CO, hmthe soil zone and rock weathering would be -7 to -12%~.DIC may be lost nom the river by degassing to the atmosphere or photosyuthetic activity.
(PBD) r I I I 1 I 1 I 1 -35 -25 -1 5 -5 +5 -soi1 organic matter CO2 in soi1 water - carbonate rock DIC derived from bicarbonate acid dissolution of carbonate rocks - = atmospheric CO2 dissolved in water atrnospheric CO2 I
Figure5.19 Sources of DIC and their associated 613c values (after Pawellek and Veizer, 1992).
Variation Ut pCO, of the Bow River The pCO, for Bow River waters were calculated using Solmin88 (Wiwchar. et al., 1988) and are presented in Fig. 5.20. Overall, the pCO, of the Bow River is relatively low compared to larger river systems. This may be due to the turbulent nature of the river. particularly in the upper reaches, ailowing excess CO2 to escape to the atmosphere. in the summer and fd pCO, values are generaliy the same order of magnitude as expected for equilibrium with the atmosphere (350 ppm). The winter samples have erratic variations in pC4dong the river, varying nom highly overpressured to near equilibrium values. The erratic variations in pC4are likely related to ice cover on the river. During the winter, the river is mainly fed by groundwater (Chapter 3) which genetally has high pCO, compared to atmospheric values. Due to intermittent ice cover dong its length, the river is oniy able to release this excess CO, in ice fm portions. The variations in pCO, dong the river would then refîect the degree tbat water has been able to degas. In the spring pCO, is near equilibrium with atmospheric pressures fiom the headwaters to Calgary, and steadily increases dowostream of Calgary. The observed high pCO, values in the river downsaarn of Calgary may be related to either instream production or the introduction of groundwater with high pCO, to the river.
Laice Louise Banff Bearspaw ûonny Bmok Bassano Dam
O fall o winter O spring A sumrner
distance from Oldman confluence (km)
Fip* 5.20 Variation in pCO, dong the length of the Bow River.
Figure 5.2 1 illustrates the variation of pCO, with 6°C. The most noticeable trend is in the spring, where low pCO, values tend to be associated with higher 6"~values, 90 whereas high pCO, vaiues trend towards a S13C of -8%0 (withui the range expected for carbonate minerais). The river itseif wodd aot be able to generate the overpressures in CO, by carbonate weathering, implying that high pCO, values in the river are associated with groundwater discharge. In the spring, lower pCO, values trend towards Waa, suggesting that as pCO, reaches equiliiwith atmospheric pressures there is also a partial isotope equilibrium of DIC with atmosphenc CO2. This is similar to relations observed in the Danube by PawelIeck and Veizer (1994). The high pCO, value observed downstream of the Bassano Dam in the spring (Fig. 5.20) is associated with a much lower 613cvalue than typicd (Fig. 5.21). This concurrent increase in pC4 and decrease in 8l3cmust be related to oxidation of organic matter. The high spring discharge may be fiushing out organic material fiom the Bassano resewoir that is then oxidized in the water
downstream O Eo of Bassano O fall O winter e spring A summer I
pCO2 (PP~)
Figure 5.21 613Cof DIC versus calcuiated pCO, for the Bow River. 91 CEEMICAL DENUDAION RAIE
Foilowing methods ouilioed in Chapter 4, the mass tlux and chemical denudation rate for the Bow River Basin cmbe estimated. Mass flux is calculated at three stations dong the river: Banff (BR2). Caneland (BR1 l), and Hays (BRIS). These stations represent the headwaters of the river (Banff), a midpoint downstream of Calgary (Carseland), and the end of the river at the conîiuence with the Oldman River (Hays). Mass flux can be calculated using a modified version of Equation 4.2, where the term Fmk,,, is included.
The majority of water that feeds the Bow River onginates within the Rocky Mountains (Table 2. l), therefore atmospheric loading rates used in Chapter 4 (Table 4.2) can be used for the temis Fm,, and FM* The only anthropogenic input of major ions
EnthmPogenic) identifïed is the city of Calgary. Given the average TDS of effluent (Table 5.1) and the average discharge at Bomy Brook and Fish Creek (5 and 2 m3/s respectively), then the mass flux fiom Calgary is approximately 4 xld kglday . Using total discharge for the months samples were collected, mass flux for the Bow River was calculated for each station (Table 5.3). There is a large increase in mass flux nom Banff to Carseland. mainly due to the addition of tributary water. Mass flux from the basin is highest during the spring and lowest during the summer. The low rnass flux during the summer is related to heavy use of the river for irrigation. This is readiiy observable in Table 5.3, where the flux at Hays is oniy 40% of the flux upstream at Carseland. Outside the irrigation season, the mass flux fiom Carseland to Hays remains relatively constant. Loading from Calgary accounts for 8 to 9% of the mass flux out of the basin in the spring and fd. In contrast Calgary accounts for 25% of the mass flux in the summer, when both concentrations and water volumes of the Bow River are relatively low. Averaged over the year, the flux at the mouth of the Bow River (near Hays) is 1,112 x 1o6 kg. After comcting for non-weathering components this gives a chemicai denudation rate of 340 kghdy, close to the world average (Table 4.6).
Table 5.3 Total monthiy discharge (x 1o6 m') and flux (x 106 kg) of TDS for the Bow River at Banff, Carseland, and Hays for the four seasons sampled.
Location 10/93 OU94 06/94 08/94
~OW ~UX ~OW ~UX ~OW flux ~OW ~UX x10 m3 x106kg x10m3 d06kg x10 m3 x106 kg x10 mJ xlo6 kg Banff 69.6 13.6 26.4 7.0 282.0 47.5 145.0 22.9 Carseland 325.0 99.2 n.a. n.a 553.0 148.0 216.0 47.8 Hays 320.0 110.2 157.0 n.a. 481.0 148.9 78.1 18.8
SUMMARY Seasonaüy, the main variation in the chemistry of the Bow River is in the TDS load; TDS is mainiy a hinction of discharge. Element ratios remain relatively constant through the year. Upstream of Calgary, then are no point source inputs that affect the major ion chemistry of the Bow River. The chernical makeup of the river is identical to that of the Bow River in its pristine headwaters. The source of ions in this section of the river are interpreted to be atmospheric loading and waterhck interaction. Sulfate is dominantly derived fiom dissolution of evaporite minerais, with up to 30% denved from oxidation of sulfides. Once dissolved, sulfate under goes a complex redox history before reaching the river, implying that the water transport@ sulfate to the river passes through the anoxic zone before becoming discharge. Calgary is the most significant point source input dong the river. Effluent fiom the sewage treatment plants loads a significant amount of Na, K, and CL to the river, and minor arnounts of SO,. Downstream of Calgary, Na and SO, is loaded to the nver by irrigation nuioff. Sulfate fkom this part of the basin is largely derived from oxidized sulfides in the local till. In the summer, sulfate derived fiom soils can be a signü'icant component of sulfate being added to the nver (up to 50%). As observed upstream of
Calgary, sulfate has a complex redox history before reaching the river. 6180 data suggests that after sulfides are oxidued, they are partialiy reduced before reachuig the nver. Again, this indicates that water transporthg this sulfate must pass through the ground before becomuig discharge. DLC in the river is largely derived from the weathering of carbonate rock by soil CO, The pC4 of the Bow River is generaliy near equilibrium with atmospheric pressures, particularly in the turbulent headwaters. High pC4 values are associated with groundwater discharge. In the sp~gthere is a partial isotope equilibrium of DIC with atmospheric CO,, as pCO, reaches equiliirium with atmospheric values. Cation exchange reactions exert a strong control on cation activity ratios in the river.
The CalMg and NaK activity ratios are controlled by exchange reactions with smectite. These activity ratios appear to be inherited signatures of ground water as smectite is absent in the suspended and bottom load. Mass flux from the Bow River Basin is highest during the spring and lowest during the summer. The low mass flux during the summer is related to heavy irrigation use of the river. Loading from Calgary accounts for 8 to 98of the mass flux out of the basin in the spring and faii and 25% of the mass flux in the summer. The chernical denudation rate of the Bow River, 340 kg/ha/y,is close to the world average. 94 IMPLICATIONS FOR BASIN HYDROLOGY
In Chapter 3, the question was posed: "how is rainfail or snowmelt over a catchent transformed into stream runoff?" One of the earliest attempt to address this was by Horton (1933). In Horton's model (Hortoaian overland flow) stream mnoff is generated when raiafall intensity exceeds the infiltration capacity of soil. Water in excess of the infiltration capacity reaches the stream by overland flow. In this model, runoff is the sum of slowly changiog groundwater discharge and rapidly changing direct nuioff. Later models are based on the observation that rainfall intensity does not normaIly exceed soil infiitration capacity. Still, these models differ in the relative importance of overland and subsurface flow (e.g. Dunne and Black, 1970; Hewlett and Hibbert, 1967). Recent studies using stable isotope techniques have tried to test these models (e.g. Maainec et al.
1974; Krouse et al., 1978; Skiash and Famolden, 1979, Dincer, et ai., 1970) What these studies show is that in high discharge events, whether generated by snow melt or rainfaii, the 'new* water being added to the basin displaces pre-existing groundwater into the river. So that high discharge in comprised of a mixture of displaced groundwater, and 'new* water. Although the stable isotope signature can be used to quantify the relative proportions of groundwater and 'new' water in stream discharge (hydrograph separation) it stül does not provide any information of the pathway that the 'new* water takes to reach the stream (i.e. as overland flow or subsurface flow). Some recent studies have suggested that the chemistry of snowmelt is altered by reactions in soil and groundwater before it becomes stream discharge (e.g. Williams et al., 1993; Campbell et al., 1995; Williams et al.. 1995), implying that the 'new' water must foiiow a subsurface path to the river. Several lines of evidence fiom this study provide further insight into the processes of runoff generation. with particular reference to the Bow River basin; they are surnmarised below : 1) In Chapter 3, discharge in the fail and winter is shown to be fed by groundwater. In contrast, the high discharge event in the spring is a mixture of 'new' snowmelt and groundwater displaced hto the river. Summer discharge is fed by summer rainfall. These data indicate that at least 40% of the river is fed by groundwater (Le. water that has had sufficieut residence time for isotope exchange reactions with rock fonning minerals to occur). The remainder of the discharge is comprïsed of 'new' water (water with a short residence time). Although these data belp determine the relative proportion of 'old' water (Le groundwater) and 'new' water feeding the river, they do not provide information on the flowpath of 'new' water to the river.
2) Chernical data in Chapter 4 show that the TDS of the river varies with discharge. The spring and summer high discharge events are more dilute than groundwater fed baseflow in the fa11 and winter. If snowmelt and rainfall were diluthg groundwater, a simple mixing relationship between the groundwater and precipitation end members would be expected. This can be tested using a mass balance calculation. Taking the average TDS of precipitation (7.4 mg/l), winter discharge (190 mgA) and summer discharge (100 mg/l) (Chapter 4), and the average discharge in winter (10m3/s)and summer (1 10 m3/s), then summer discharge caries four time the mass bat would be expected for simple dilution of groundwater. This indicates that precipitation generating the high spring and summer discharge undergoes some degree of water/mck interaction. The observation that summer discharge has nearly identicai chernical composition as baseflow, and quite distinct from that of precipitation (Figs. 4.2 and 4.4)' suggests that precipitation undergoes similar water/rock interaction as groundwater. This is supported by themodynamic models that show equilibrium exchange reactions on smectites are controlling the activity ratios of major cations in the river water, although smectite is absent from the river. The interpretatîon that this is an inberited groundwater signahm is supported by the similarity of activity ratios for river water and groundwater in the basin. Therefore, contrary to common opinion, the relationship of discharge to TDS is not one of dilution, but rather, the controlling variable is the relative residence tirne of the water in the ground. During high discharge events, related to snowmelt and heavy rainfall, there would be a large flux of water through the ground. This would reduce the time available for water-rock interaction dong the flow path, and thus the TDS of water.
For this mode1 to work, snow melt must be able to enter the ground during a period when it would presumably be frozen (Le. during the spring melt). Work by Harris (1986) indicates that the relatively high snowfall in the headwatea of the Bow Basin thermally insulates the ground, making the pedost iine anomalously high compared to areas north and south. As well, mass balance calculations (S. Harris, persona1 commun., 1997) indicate that melt water hmthe winter snow pack in the Bow Basin is transferred to the groundwater zone as early as mid-February. In addition, the headwaters of the Bow River basin are domhated by carbonate rock, which tends to have a high permeability.
3) The 6'8~, indicates that after evaporite minerals are dissolved, sulfate goes through a complex redox history &fore entering the river. Similarly, after oxidation sulfides show evidence of partial reduction before entering the river. This complex history of redox conditions implies that the water transporthg this S must move fiom the surface, through the anoxic zone, and then to the river.
In summary, the above data all support the conclusion that the Bow River is almost entirely fed by groundwater input. Snowmelt and rainfail flushes groundwater into the river. The seasonal variations in TDS of the river is not a function of dilution, but rather 97 hydrology is a dynamic system, where the majority of water feeding the river flushes through the groundwater system seasonaiiy. Care should be taken in applying these results to runoff generation models. The basin geology, and anomaiously high permafrost line, ailow snowmelt to enter the ground, reducing surface runoff. It is likely that riven draining dinerent other lithologies may not be dominated by groundwater discharge. However, a survey of basins with Merent lithologies, using a combined stable isotope and geochemicai approach, would help resolve the process of runoff generation. CHAPTER 6 Tradng anomalous TDS in Nose Creek
INTRODUCTION
Nose Creek was examined as part of the Bow River Basin study (Fig. 1.1). It was chosen for a more deiailed study because the total dissolved solids load (TDS) is significantly higher than the Bow River or its other tniutaries. These high concentrations must be the result of either natural phenomena or anthropogenic activity in the basin. This study combines both chernical and stable isotope anaiysis to determine the origin of the high dissolved load in Nose Creek. Samples nom Nose Creek representing two flow regimes, fa base flow conditions and spring/summer high discharge, were couected. The fdsamples were coiIected dong Nose Creek over a two day period (October 30 to 31, 1993). Spring samples were coiiected on June 5, 1996. For each set, seven samples were collected dong Nose Creek, and one from West Nose Creek (Fig. 6.1). s
THE NOSE CREEK BASIN
Nose Creek is located in southem Alberta, Canada (Fig. 6.1), in an area dominated by prairie grassland. The basin has a dry-subhumid ciimate, with annuai precipitation of 400 mm, and potential evapotranspiration of 530 mm (Ozoray and Bames, 1977). Precipitation is concentrated in the spring and summer, with a dry fall and winter (Fig. 6.2). Nose Creek extends 45 km f~omits confluence with the Bow River to its head waters north of the town of Crossfield (Fig. 6.1). The creek drains a 986 km2 area, and has an average annual fiow of 0.73 m3/s (Environment Canada, 1990). The creek has one main tributary (West Nose Creek) and several ephemeral tributaries. The basin lies 99 outside of any of the Bow River irrigation districts, so groundwater and precipitation are tbe only sources of water in the creek. Discharge in the creek is relatively high during spring and early summer, and low in the fdand wintet.
The Nose Creek basin is underlain by up to 15 m of Balzac Till, a silty to sandy till with abundant blocks of limestone and quartzite, and rare blocks of granite and gneiss (Moran, 1986). The Porcupine Hills Formation, a non-marine, fme graine4 calcareous, cherty sandstone, forms the bedrock of the basin (Green. 1972). The dominant land use in the north part of the basin is agricultural (mostly cereal grains and canola crops). In addition, there are two natural gas processing facilities that extract sulfur, one at Crossfield and one at Balzac. The southernmost part of the basin lies within the northeast section of the City of Calgary. This section of the city is dominated by residential housing, the Calgary International Airport, and some light industry.
Figure 6.1 Nose Creek basin, showing sample locations and the Bearspaw Reservoir. Sample stations dong Nose Creek have the prehx NC and those on the Bow River, the prefix BR (Fig. 6.1, Appendur 2,4). An important feature to note is that although station NC5 Iies within the City of Calgary, the surroundhg area is undeveloped agricultural land. Stations NC6 and 7 lie withïn the developed part of the city.
RESULTS AND DISCUSSION
Source of Nose Creek Woirr
For the fall samples, precipitation and surface mnoff can be eliminated as major contributors to Nose Creek, as only 9 mm of precipitation were recorded in the 2 months preceding sampling. The ody water that municipalities pipe directly into Nose Creek is storm water runoff, and aLl of the storm sewers observed during sampling were dry.
Therefore, groundwater must be the primary source of water in Nose Creek in the fa. This is consistent with Ozoray and Bames (1977) who indicate that in the fail, groundwater is the principal source of surface water in large parts of southem Alberta. In contrast, during spring sampling storm sewea draining paved streets were observed discharging into Nose Creek. Environment Canada recorded 78 mm of precipitation in the 2 months preceding the spring sampling.
Inorganic chemistry of Nose Creek water
Chernical data for Nose Creek are presented in Appendix 4. Two features distinguish Nose Creek water from the other tributaries: 1) concentrations of inorganic ions are anomalously high, and 2) the dominant ions in Nose Creek are Na-SO, -HC4 whereas aIi other tributaries to the Bow River are Ca-HC03 waters (Fig. 6.2). Precipitation can not be a major source of the dissolved load because it is relatively dilute (north Calgary precipitation has an average TDS of 4 ma).Therefore, the high TDS load of Nose 101 Creek can only be explained by either natumi controls (e-g. weathe~gof till andlor bedrock, evaporation, etc.). or anthropogenic activity in the basin.
Figure 6.2 Ternary plot of major ions for Nose Creek (open circles) and other tributaries to the Bow River (closed circles).
Figure 6.3 illustrates the variation in major ion concentrations dong the length of Nose Creek for spring and fd. For the spring samples, the most important features to note are that: 1) upstream of Airdrie, in the headwaters of Nose Creek, concentrations are high cornparrd to other tributaries of the Bow River, and 2) Nose Creek is diluted where the creek flows through the cities of Airdrie and Calgary. For the fa11 samples, concentrations of Na and SO, are generally higher. the cities of Airdrie and Calgary cause more significant dilution, and concentrations of major ions (particularly Na and SO,) increase steadily as the creek flows through agriculhiral land between Aircirie and Calgary. These spatial relations imply that municipalities act as point source inputs of relatively dilute water, and that in the fdl either a dispersed source, or a physicaVchemica1 process, significantly increases the TDS of Nose Cree k within the 102 agricultural areas of the basin. In the spring, the major ion chemistry is relatively consistent dong the creek. althwgh the concentration of SO, does increase through the
Airdrie Airdrie . Calgary 1 12.5
10
7 -5
5
2 -5
I
distance fmm Bow River b) distance from Bow River confience ()an) confluence (km)
Figure 6.3 Variation in major dissolved ions dong the length of Nose Creek for a) spring, and b) fall runoff. Note that the Land between the Calgary City Limits and 'tuban Calgary" is undeveloped agricultural land.
Of the dissolved solids, chloride is the easiest to account for because of its conservative nature and Limited numbers of potential sources. There are no naturai sources of Cl in the Nose Creek Basin, the concentration of Cl in precipitation is Low (0.165 mgil), and neither the bedrock nor the till in the basin contain evaporites. This 103 implies that Cl is derived fiom an antbropogenic source. The most Likely source of the Ci is road salt used to melt winter ice on Highway (HW)2 and other secondary roads (Fig. 6.1). Road salt used by Aiberta Transportatioa is typically 95% NaCl, with the remainder king K, Ca, and Mgchlorides. Accepting that road salt is the major source of Cl, it only accounts for 10 - 20 % of the cations in Nose Creek, and does not account for the significant Levels of sulfate. The steady increase in major ion concentrations dong the length of Nose Creek between Airdrie and Calgary, pdcularly in the fd(Fig. 6.3), makes it dïffkult to define a source based on chemûtry aione. It is unlikely that in- stream evaporation could account for the observed increase. Thus, there must be addition fiom non point-source(s) dong the length of the creek. Stable isotopes of water and sulfate were used to trace the source(s).
Oxygen and hydrogen isotope compositio~~~of Nose Creek water Stable isotope data for Nose Creek are presented in Appendix 4. On a plot of 6~ vs PO
(Fig. 6.4), the data for the sp~gsamples plot in a relatively tight cluster, with 6D of -120 to 125960 and 6% of -15%0. The data falis just off the local meteoric water Line
(LMWL),suggesting precipitation is the dominant source of creek water. This is consistent with the buik of precipitation occurring during the spring (Fig. 2.5). If groundwater is king added to the'creek, the consistency of the isotope composition dong the creek suggests that the proportion of groundwater to rain water is nearly constant. In the fali, Nose Creek water becornes relatively depleted in the heavy isotopes of O and H in the downstream direction, 6"O decreases nom -12.5 to -19.0470. On a plot of 6D vs 6"O (Fig. 6.4), the faii samples dehe a best fit line with a dope of 5.4 (8- .92), as compared to a slope of 8 for the local meteonc water line (Lm).Data defining on a dope lower than the LMWL represents either evaporation or mixing. In a small creek, evaporation dong the flow path would be unlikely. If it did occur, the 6D and 6"0 104 values would progressively increase downstream, opposite to what is obsewed. The progressive dowustream decrease in 6D and 6"0 in the fall sample set can only be explained by addition of water relatively depleted in D and '*O.
a Bow Rhr a Nose Creek -1 00
between Airclrie
within Calgary 4-1 Bow River downstream of Bearspaw -1 60
Figure 6.4 Plot of 6D versw S"O for a) spring and b) fd, aad the Bow River below the Bearspaw Reservoir (square). The local meteoric waterline @D = 8PO + 6) is given for reference.
The isotope data for the fasamples in Figure 6.4 plot in three clusters, with each cluster representing a separate reach of Nose Creek: 1) upstream of Airdne, 2) between Calgary and Aiidrie, and 3) within Calgary. The higher #'O values upstream of Airdrie (relative to the LMWL) are consistent Mth groundwater king the principal source of the head waters of Nose Creek (Ozoray and Bames, 1977). The best fit hedefined by Nose Creek water passes through Bow River water sampled downstream of the Bearspaw Reservoir (Figs. 6.1,6.4; Bowness sample in Appendix 2). Normally, such a relationship would be interpreted as miung between two end members (i.e. the head waters of Nose 105 Creek and the Bow River). Because Nose Creek is a tributary of the Bow River, it seems impossible that Nose Creek water is mixing with Bow River water dong the length of Nose Creek. However, it is important to note that the stable isotope composition of Nose Creek water only changes as it flows through the two cities (Airdrie and Calgary). This suggests that the cities are adding signincant amounts of water to the creek, consistent with the observed dilution (Fig. 6.3). Cities cm add water to a creek directly via storm sewer discharge and sewage outlets, or indirectly from leaking pipes via groundwater (curent estimates by the City of Calgary put water loss from Calgary pipes at 15%). Although lawn watering can add additional water, it is minor relative to water loss from pipes. Tliere was no storm sewer discharge observed during sampling, and residential or industrial sewage is not discharged into Nose Creek. Therefore, the water king added to the creek by Airdrie and Calgary must be derived ftom groundwater. Both Airdrie and north Calgary draw their water supply fkom the same source, the Bearspaw Reservoir on the Bow River, consistent with the mUing relationship observed in Figure 6.4. Thus, the observed dilution and increased flow of Nose Creek is a result of addition of municipal water, denved from the Bearspaw Reservoir, and added to Nose Creek via groundwater infiltration. Assuming the relationship in Figure 6.4 represents perfect mixiag, an isotope balance approach can be used to caiculate the amount of Bearspaw Reservoir water king added to Nose Creek at Airdrie and Calgary in the fd. These calculations suggest that the fiow of Nose Creek increases by a factor of 1.6 as it fîows through Airdrie and 3.8 as it fiows through Calgary. There are iimited Stream flow data to compare with these results. Environment Canada operated two gauging stations on Nose Creek fiom 1980 to 1986, one below the confluence with West Nose Creek, just south of station NC5, and one at the same location as station NC7 flable 6.1). Excluding the high flow year of 1986, the flow between the two stations increases an average of 4.5 times (n=4), in good agreement 106 with the 3.8 fold increase in flow calculated nom the isotope data. Given the average flow for October of 0.33 m3/s,a 3.8 fold increase in flow implies Calgary is adding approximately 22 MVday to Nose Creek, or 10% of the 220 MV &y that are piped into north Calgary hmthe Bearspaw reservoir. This is reasonable when compared to the estimated 15% water loss through lealry pipes Ïn Calgary.
Table 6.1 - Environment Canada (1990) flow data for Nose Creek flow data south of NCS' flow data at NO Increase in flow
Envimument Canada stations 'OSBH003 and '05~~901
In surnmary, m) and 6"O data indicatc that in a dry climate, municipalities can add signifcant amounts of water to local water bodies. During base flow conditions, water nom the head waters of Nose Creek is mïxed with two pulses of "Bow River" water added via leaky pipes in the cities of AUdrie and Calgary, increasing the creek discharge approximately 4 fold. It should be noted that the town of Baizac apparentiy does not alter the isotope composition of Nose Creek water. The smail population of Baizac (250), as compared to Calgary (700,000) and Aircirie (25,000), would not introduce as much municipal water into the basin. The volume of water added to the creek by spring rains
(3 times bwflow conditions) appears to overwhelm any isotope signature of municipal water king added to the creek. Municipal water use in the city is relatively constant, summer consumption is 10-15% greater than winter (Engineering and Environmentai
SeMces, The City of Calgary). Thus, we can assume the addition of muncipal water 107 calcuiated for October (22MVday), is consistent through the year. Taking the Enviroment Canada (1990) average discharge for Nose Creek of 11,500 Ml, fiom May to October, municipal water accounts for 35% of discharge during this time. Whereas during low fiow in October, municipal water accounts for 77% of discharge. Flow data is not available for wintet months, however flow conditions would be similar to, if not less
Sources of su@te
Evaporation can not account for the observed increases in major ion concentrations between Airdrie and Calgary. This impiies that a dispened source(s) adds inorganic solutes dong this reach. The S and O stable isotope compositions of sulfate cm serve to identify these source(s). Dissolved sulfate can be derived fiom the dissolution of sulfate minerals, the oxidation of pyrite and other forms of reduced su,and anthropogeaic inputs (e.g. fertiluer, industrial emissions from natural gas facilities that process suh, etc.). Given the typical concentration of sulfate in Calgary rain (1.5 mg/l), atmosphenc sulfate can not be a major contributor of sulfate to Nose Creek. Soils in southem Alberta generally have sufticient suifur content for rnost crop plants, so that S-based fertilïzer is not commonly used (Hendry et ai., 1986). Even where used in parts of the eastern Bow Basin, Hendry et al. (1986) noted that the application of S-based fertiher did not affect the groundwater chemistry. The only plausible origin(s) of sulfate in Nose Creek are 1) weathering of till andlor bedrock, or 2) anthropogenic input. The oxygen isotope composition of sulfate was used to determine if the sulfate was denved fiom a primary source (e-g. sulfate minerals), or the oxidation of reduced sulhu. The 6"0 of sulfate and water are plotted in Figure 6.5. The fall samples have a relatively wide distribution with 6180,,,,, ranging from +7 to -7460, and a positive correlation with 8*0,,,,.In contrast the spring samples only have a 5%~variation in 108
6'80,, and 6"Om, is constant. Overaii, the 6180, values of ail samples are depleted cornparrd to evaporite minerais; the 6'80, compositions of evaporites range nom +10 to +16%0 (Claypool et al. 1980). Exchange reactions between water and sulfate can not be invoked to explain the depkted 6%- because the SO, - H,O exchange is extremely slow at temperature and pH conditions comparable to Nose Creek (Lloyd, 1968). However, when reduced S is oxidized, 50 to LOO% of the oxygen is derived from water (Van Stempvoort and Krouse, 1994). Therefore, the positive correlation between 6180,s,, and 6'80,,, in the fail samples suggests SO, is derived nom oxidation of a reduced sulfur source. The unifonn 81800,,,, in the spring samples makes a correlation impossible. The sp~grains would Wtely be washing out soluble salts built up in the soil, so a correlation between 6'8~~0,,,,,and 6'8~,H20)would not even be expected. However, as in the fail, the depleted 6"0,, values suggest at least some of the sulfate was derived fkom a reduced sulfur source*
Figure 6.5 8'0 in sulfate verses 6"O of water for spring (open cirlces) and fd (closed circles). The stable isotope composition of sulfur in sulfate was examined in an attempt to identify the sources(s) of sulfur. The 6% valws are plotted against sulfate concentrations in Figure 6.6. For the fklI samples, these data plot in clusters as in Figure 6.4, representing the three reaches of Nose Creek: 1) upstream of Airdrie, 2) between Airdrie and Calgary, and 3) withh Calgary. Unlike Figure 6.4, these data do not defme a muciog he, but rather, each mach of the creek has a different source of sulfate. There is ody one data point upstream of Airdne so it is difficdt to Say if it represents a unique source. However, the 6% composition of +l% is weli above values typical of surface water in southem Alberta, but consistent with S4emissions during processing of Sour gas. Emissioas hmthe Crossfield gas plant have g4s values near +25%0 (Norman and Krouse, 1992). Between Airdrie and urban Calgary, the sulfate concentration increases 34 almost Linearly with distance, however the 6 S,, remains relatively constant it +5%.
This indicates that a single. but uniformly dispersed source of sulfate, is being added between the two cities during the fd. This source is likely related to the dominant land use, agriculture. Fertilizer application can be dedout because it is not common, and the 6% valw of S-based fertilizer sold in this area was determined to be +14%0,higher than that observed. One possible source is oxidation of organic S in soils. Samples were collected in the last part of October, near the end of the local harvest. During this time soils are loose and uncovered. making them susceptible to wind erosion. The high suspended solid load and brown colour. of Nose Creek are consistent with large amounts of particdates king added. Fennell(1994) examined an area 25 km NW of the Nose Creek basin that is underlain by the Balzac Till. In this area A-horizon soils typically have S'~Svalues in the range of +3 to + 10%~~The SUs of the sulfate added to Nose
Creek between Airdrie and Calgary (+ 5%) fds within this range. between Airdrie snd Qlgmy a a e. t a between Aiidrie and Caigary
Figure 6.6 6*~versus sulfate concentration for a) spring and 6) fd.
In the fd, the 6% value of sulfate in Nose Creek is -7.7% in Calgary. Using the calculated increase in volume of Nose Creek as it fiows through Calgary, a combined mass and isotope balance can be used to calculate the 6% composition of the sulfate source within Calgary needed to cause the observed change in the 6%- of Nose Creek. These calculations yield a 6% value of -14.6460 and a sulfate concentration of 4.3 meqA for water king added from Calgary. The calculated concentration of SO, is significantly higher than that of the Bow River (Appendix 2). The excess sulfate must be denved fiom the till and/or bedrock that the groundwater fiows through. The calculated value of the sulfate (-14%0)is consistent with the average 634~value of -12.5%0 @=IO) for total S
(pyrite + organic S) in the Balzac till (Fenneli, 1994). This implies that the large flux of water king added by the City of Caigary is oxiduing reduced forms of sulfur in the till, which is then mobilized as SO, and transported into Nose Creek via groundwater flow. This is also observed by Hendry et ai. (1986, 1989) in the in the eastern Bow River B asin. 111 where groundwater in oxidized tüls has anomalously high suifate concentrations relative to unoxidized tus, and has 634~values of -9.2560. Hendry et al. (1986, 1989)
demonstrated that in this area, the high sulfate concentrations an related to oxidation of organic matter in the till. The sp~gsamples show a simpler relationship between sulfate concentration and 6% The sample for FUS, upstnam of Air& was lost, so only the trend dom stream of Airdrie cm be analysed. There is a progressive decrease in the S4svalue of dissolved sulfate between Airdrie and Calgary, with a concurrent increase in sulfate concentration (Fig. 6.6). This appears to be a simple mixing relation between a source with 634~> +
12960 (possibly emissions nom the Crossfield naniral gas processing facility) and a relatively depleted source. The 634s value of ciissolved sulfate approaches +5% as the creek reaches Calgary, similar to the fdl samples. This may reflect soluble sulfate king Ieached fiom soils by spring rains. The fs,, drops 5%~where Nose Creek enters urban Calgary. This observed çhift in the 634~is likely the result of r.nixing of three sources: 1) sulfate in the creek before it reaches the city, 2) sulfate in storm water, and 3) oxidized sulhir fiom tills in the Calgary maas observed in the fidl samples. The $s,,, measured from storm sewer discharge is +13%0. There was no stomi water discharge in the fall, so the addition in the spring of sulfate in storm sewer discharge, with a relatively 34 enriched isotope composition, would explain why the &op in 6 S, at Calgary is less signiûcant than in the fd.
CONCLUSIONS
This study shows that in a dry climate. municipalities can add significant amounts of water to local aquifers. During base flow conditions two pulses of Bow River water (the municipal water supply) are added to Nose Creek via leaky pipes in the cities of Airdrie and Calgary. This water increases discharge in the creek 4 fold during base flow, diluting 112 the dissolved inorganics, and thus enhancing water quality in Nose Creek. Munifipal water accounts for 35% of spring and summer discharge, and up to 7796 of fd and winter discharge in Nose Creek. In terms of basin scale water budgets, water fiom leaky pipes has kenrecorded as lost 6rom the river system, however this study illustrates that at least two thirds of this "lost watei' is eventually renirned to the Bow River via Nose Creek. The more positive 6% values above Airdrie suggests that the processing facility that removes sulfur nom natural gas near Crossfield may be a major source of dissolved sulfate in the headwaters of the creek. Significant loading of inorganic constituents occurs in the agricultural area between Airdrie and Calgary. Stable isotope evidence suggests that oxidation of orgaaic matter in soüs is the primary source of sulfate. Sulfate relatively depleted in %S is added within Calgary through oxidation of reduced forms of sulfur (pyrite + organic-S) in tills, by the anthropogenicaliy increased groundwater recharge. This study iilustrates how a combined chernical and stable isotope study cm help elucidate processes controiiing surface water chemistry. However, this work wodd have been merenhanced by measuring discharge at sample sites, thus aiiowing for more accurate mass balance cdculations. Conclusions
Controis on the chernistry of the Bow River
Although TDS is relateci to discharge, the source of ions in the headwaten of the nver is controlled by atmospheric deposition and waterirock interaction. Amiosphenc loaduig can be a signincant source of some ions in the headwaters of the river (e-g. 50% of K, 17% of SO,, 1696 of Cl). In terms of water/rock interaction, the input of ions to the river is largely controiled by dissolution of carbonate and evaporite minerals. Calgary is the most significant point source input dong the nver. Effluent fiom the sewage treatment plants loads a signincant amount of Na, K, and Cl to the nver, and minor amounts of SO,. Dowastream of Calgary, Na and SO, is loaded to the river by ûrigation runoff.
In the headwaters of the basin, sulfate is dominantly denved kom dissolution of evaporite minerals, with up to 30% derived nom oxidation of suifides. Once dissolved, sulfate under goes a complex redox history before reaching the river, implying that the water transporthg sulfate to the river passes through the anoxic zone before becoming discharge. Downstream of Calgary, sulfate is largely derived hmoxidized sulfides in the local till. In the summer, sulfate derived from soils cm be a signincant component of sulfate king added to the river (up to 50%). As in the headwaters, sulfate undergoes a complex redox history before reaching the nver. SL80data suggests that after sulfides are oxidized, they are partially reduced before reactiing the river. This indicates that water transporthg this datemust pas through the anoxic zone before becoming discharge. The pCO, of the Bow River is generally eear equiliirium with atmospheric pressures, particularly in the turbulent headwaters. High pC4 values are associated with groundwater discharge. DIC in the river is mainly denved from the weathering of carbonate rock by soil CO,. As pC0, reaches equiliirium with atmospheric values there is a partiai isotope equiiibrium of DIC with atmospheric CO,. Cation exchange reactions exert a strong control on eiement ratios in the river. The WMg activity ratio is strongly conmIIed by exchange between Ca- and Mg- beidellite. NaK activity ratios are connolled to a lesser degree by exchange between Na- and K- beidellite. These activity ratios appear to be inherited signatures of ground water. The fked element ratios in the river suggest that that both snowmelt and rainfall must pass through the ground before reaching the river. The chernical denudation rate for the Bow River at Banff is 678 kghaly. The denudation rate for the basin as a whole is 340 kgma/y. Loaduig fiom Calgaiy accounts for 8 to 9% of the mass flux out of the basin in the spring and fall and 25% of the mass flux in the summer.
Hydrology of the Bow River Basin
Discharge in the fdand winter is fed by groundwater. In contrast, the high discharge event in the spring is a mixture of 'new' snowmelt and 'old' groundwater displaced into the river. Summer discharge is fed by sumrner rainfall. Combined geochernical and stable isotope data indicate that snowmelt and &&ilMiow a subsurface path to the river, flushing pre-existing groundwater into the river system. The groundwater residence time durhg the spring and summer must be short, the matter of &YS or weeks.
Source of anomalous TDS in Nose Creek
This study shows that in a dry climate, municipalities cm add significant amounts of water to local aquifers. hiring base flow conditions two pulses of Bow River water (the municipal water supply) are added to Nose Creek via leaky pipes in the cities of Aircirie and Calgary. This water increases discharge in the creek 4 fold during base flow. diluthg the dissolved inorganics, and thus entisncing water quality in Nose Creek. Municipal water accounts for 35% of spring and summer discharge, and up to 77% of fd and winter discharge in Nose Creek. Significant loadiag of inorganic constituents occurs in the agriculaurit ma between Airdrie and Calgary. Stable isotope evidence suggests thaî oxidation of organic matter in soils is the primary source of sulfate. Sulfate relatively depleted in 34~is added within Calgary through oWdati011 of reduced forms of sulfur (pyrite + organic-9) in tas, by the anthropogenically increased groundwater recharge.
FUTURE RESEARCH
Several avenues of future research have ken identified in this study. Stable isotope data in Chapter 3 indicate that the discharge of the Bow River is a mixture of '018 and 'new' water. By collecting detailed time-series samples of surface, ground, and snowmelt waters, it would be possible to quant@ the seasonai variations in their relative contributions to the river. Chapter 3 also indicates that the stable isotope composition of surface water is a fwiction of rnixiog of two weather systems. By conducting detailed sampling of surface water and precipitation, east of the Great Divide, it may be possible to calibrate this relationship. This would make it possible to use temporal variations in 6180 records to examine historical variations in weather patterns in this part of North Amenca. Chapter 4 examined chernical controls on the Bow River and weathering rates in the headwaters of the basin. Several data sets, similar CO those used in this snidy, are available for river basins north of the Bow River. As these basins are underlain by similar lithologies, it is possible to do comparative studies, examining how colder ciimates mer north affect weathering rates. Several Lines of evidence indicate that the Bow River is almost entirely fed by groundwater discharge. The application of these results to modeis of runoff generation need to be tested by conducthg a survey of basins with difterent Lithologies. This would indicate if the Bow River is a special case due to the dominance of carbonate rock in the headwaters. REFERENCES
Abercrombie, H. J. (1989) Water-rock interaction during diagenesis and thermal recovery, Cold Lake, Aiberta. W.D. thesis, Uaiv. of Calgary, Calgary, Alberta, Canada, 183 pages. Amundson, R., O. Chadwick, C. Kendall, Y. Wang, M. DeNiro, Isotopic evidence for shifts in atmosphenc circulation patterns during the late Quaternary in mid-North Amenca, Geolugy ,24,23-26,1996. Atlas of Alberta (1969) University of Alberta Press, Edmonton Alberta, 158 pages. Berman, R.G.; Brown, T.H. and Greenwd, H.J. (1985) A thermodynamic data base for minerals in the system Na$-50-Cao-MgO-Fe0-Fe203-AI 203-Si022Ti02-H ,MO2.
Atomic Energy Canada Ltd., Technical Report TR-377,70pages.
Berner, E.K. and Bemer, R.A. (1987) The global water cycle. Prentice-Hall, New Jersey, 397 pages. Block, H.; Crosley, R.; Shaw, P. and Donald, D. (1993a) Water quality monitoring in Banff and Jasper National parks (1973 to 1991), Part 1: Interpretive Report. Conservation and Rotection, Water Quality Branch, Environment Canada, 74 pages. Block, H.; Crosley, R. and Shaw, P. (1993b) Water quality monitoring in Banff and Jasper National parks (1973 to 1991), Part 2: Data Report. Conservation and Protection, Water Quality Branc h, Envimnment Canada, 370 pages. Block, H. and Zaal, F. (1988) An assessrnent of the Banff sewage lagoon and its effects on the Bow River during February 1988. Conservation and Protection, Water Quality Branch, Enviromnent Canada, 26 pages. Bluth, G.J.S. and Klump, L.R. (1994) Lithologic and climatologie controls of river chemistry. Geochimica et Cosmochhica Acta, 58,234 1-2359. 118 Bow River Water Quality Council(1994) Preserving our Iifeline: A report on the state of the Bow River. Calgary, Alberta, October 1994,98 pages.
Brown, G.W.; Berman, R.G. and Perkins, E.H. (1988) GEO-CALC: A software package for rapid calculation of stable pressure-tempature-activity phase diagrams using a [BM or compatible personal cornputer. Cornputers and Geosciences, 14,279-289. Buhl, D.; Neuser. R.D.; Richter, D.K; Riedel, D.; Roberts, B.; Strauss, H. and Veizer, J. (1991) Nature and Nurture: Environmental Story of the River Rhine. Naturwissenschaften ,78,3 37-346. Campbell, D.H.; Clow, D.W.;Ingersoll, GR;Mast, M.A.; Spahr, NE. and Turk, J.T.
(1995) Processes controlling the chemistry of two snowmelt-dominated streams in the Roc@ Mountains. Water Resources Research, 31,28 11-282 1. Chariton, S.E.D.;Hamilton, H.R. and Cross, PM. (1986) The limological characteristics of the Bow, Oldman, and South Saskatchewan nvers (1979-82). Part II, The Primary
Producers. Alberta Environment, Water Quaiity Control Branch, Edmonton, Alberta, 110 pages. Clark, R. (1993) Water: the international crisis. MlT Press, Cambridge, 193 pages.
Claypool, G.E., Holser, W.T., Kaplan, I.R, Sakai, M. and Zak, 1. (1980) The age curves
of sulfûr and oxygen isotopes in marine sulfate and their mutuai interpretation.
Chernical Geology,28,199-260.
Coleman, M.L.; Sheperd, T.I.; Durham, J.J.; Rouse, J.D. and Moore, G.R. (1982) Reduction of water with zinc for hydrogen isotopes analysis. Analytical Chemistry, 54,993-995. Craig, H. (1961) Isotopic variations in meteoric waters. Science, 133, 1702- 1703. Cresser, M. and Edwards, A. (1987) Acidification of freshwaters. Cambridge Univ. Ress. Cross, PM; Hamilton, HB.;and Charlton, S.E.D. (1986) The limological characteristics
of the Bow, Oldman, and South Saskatchewan nvers (1979-82). Part 1, Nutrient and water chemistry. Alberta Environment, Water Quality Control Branch, Edmonton, Alberta, 182 pages. Daly, R.A, Manger, G.E. and Clark, S.P. (1961) Density of rocks. in Handbook of
physicai constants, edited by S.P. Clark, The Geologicd Society of America Mernoir 97,19-26. Dansgaard, W. (1964) Stable isotopes in precipitation. Tellus, 16,436-468. Drever, I. 1. (1988) The geochemistry of natural waters, second edition. Prentice Hall, New Jersey, 437 pages. Drever, J.I. (1997) Catchent mass balance. in GeochemicaI processes, weathering and groundwater recharge in catchments, edited by O.M. Saether and P. Caritat, A.A. Balkema, Rotterdam, 400 pages.
Dincer, T., Payne, BK;FLorkowski, T.; Martinet, I. and Tongiargi, E. (1970) Snowmelt runoff from measurements of tritium and oxygen-18. Water Resources Research , 6, 110-124. Dunne, T. and Black, R.D. (1970) Partial area contribution to storm runoff in a srnail New England watershed. Water Resources Research, 6(5), 1296131 1. Eden, G.F. (1996) Sources of sulpbur in pst-indusmal sediments of Waterton Reservoir, S.W. Alberta. M.Sc. thesis, University of Calgary, Canada, 147 pages. Edmond, J.M. (1982) Major ion chemistry of some large chinese rivers. Nature, 298, p. 550-553. Environment Canada (1990) Historical Streamflow Summary, Alberta. Inland Waters Directorate, Water Resources Branch, Water Survey of Canada, Ottawa, Canada, 629 pages. 120 Environment Canada (1983) Sampling for water quality. Inland waters directorate, Water resources branch, Water survey of Canada, Ottawa, Canada, 55 pages. Epstein, S. and Mayeda, T.K. (1953) Variation of "0 content of waters fiom namal sources. Geochimica Cosmochimica Acta, 4,2 13-224. Ferne& J.W. (1994) Source and distribution of sulfate and associated organics at a sour gas plant in southem Alberta. M.Sc. thesis, Univ. of Calgary, Calgary, Alberta, Canada, 192 pages. Fenneil, J.W.; Bentley, L.R.; Lyness, L.S. and Anderson, P. (1994) Source and distn'bution of sulphate in groundwater near a sour gas plant in southem Aiberta 4th Aanual Conference on groundwater and soil remediation, Sept. 21-23, Calgary, 529- 543.
Flowers, G.C. (1986) Computation of thermai dynamic propdes of reactions involving rninerals and aqueous solutions with the aid of a personal cornputor. Cornputors and Geosciences, 12,36 1-380. Fritz, P.; Cherry, I.A.; Weyer, K.U. and Sklash, M. (1976) Storm runoff malyses ushg environmental isotopes and major ions. in Interpretation of Environmental Isotope and Hydrochemical Data in Ground Water Hydrology. Vienna: International Atornic Energy Agency, Viema, 11 1 - 13 1. Garrels, R.M. and MacKenzie, F.T. (1967) Ongin of the chexnical composition of some springs and Iakes. in Equilibrium concepts in natural water systems. American Chemical Society, Advances in chemistry series, 67,344~~. Green, R. (1972) Geological map of Alberta. scaie 1:267,000, Alberta Geological Survey, Nanual Resources Division.
Harris, S.A. (1986) Permafrost distribution, zonation and stability dong the eastem
ranges of the cordiliera of North Amenca. Arctic, 39,29-38. 121 Hendry, ML; Krouse, H.R. and Shah, M.A. (1989) Interpretation of oxygen and sulfur isotopes from dissolved sulphates in tas of southem Alberta, Canada. Water Resources Research, 25(3), 567472. Hendry, M.J.; Cherry, J.A. and Wallick, El. (1986) Origin and distriion of sulfate in
a fiactured tili in southem Alberta, Canada. Water Resources Research, 22(1), 45-6 1.
Hewlett, J.D. and Hibbert, A.R. (1967) Factors affécting the response of small watersheds to precipitation in humid areas. in International Symposium on Forest Hydrology 1965 (W.E. Sopper and H.W. Ld,eds.), Pe~sylvaniaState University, Pergamon Press, New York, 275-290. Hitchon, B. and Krouse, H.R. (1972) Hydrogeochemistry of the surface waters of the Mackenzie River drainage basin, Canada - III, Stable isotopes of oxygen, carbon and sulphur. Geocliimica et Cosmuchimica Acta, 36,1337 - 1357. Holland, H.D. (1978) The chemistry of the atmosphere and oceans. Wiley, New York. Horton, R.E. (1933) The role of infiltration in the hydrologie cycle. Transactions American Geophysical Union, 14,446460. Hutcheon, LE. (1989) Application of chernical and isotopic analyses of fluids to problems in sandstone diagenesis. in Short course in burial diagenesis handbook (LE. Hutcheon, ed.), Mineralogicd Association of Canada, 15,279-306. Ingraham, NL. and Taylor, BE. (1986) Hydrogen isotope study of large scale meteoric water transport in northem Caüfornia and Nevada. Journal of Hydrology, 85, 183- 197. Ivanov, M.V. (1983) The sulphur cycle in continental reservoirs. in The GLobal Biogeochemical Sulphur Cycle, SCOPE Report 19, edited by M.V. Ivanov and J.R. Freney, John Wiley, New York, 33 1-356. 122 KLivokiotis, P. and Thomson, R.B. (1986) The climate of Calgary. Climatological Studies No. 38, The climate of Canadian cities No. 3, Environment Canada, Atmosphenc Environment Service, Ottawa, Canada, 52 pp.. Krouse, H.R. and Grinenko, V.A. (1991) Stable Isotopes: Natural and Anthropogeoic Suiphur in the Environment. SCOPE Report 43, John Wey and Sons, New York, 440 P%=- Krouse, HP.;Holecek, G. and Steppuhn, H. (1978) Detecting snowmelt in streamfhw by D/Hand 160abundances in two western Canadian basins. in Stable isotopes in the Earth Sciences (ed. B.W. Robinson), DSIR Bulletin 220, New Zealand Department of Scientific and Indusrrial Research, New Zealand. Lasaga, A.C.; Soler, J.M.; Ganor, J.; Burch, T.E. and Nagy, K.L. (1994) Chernical weathering rate laws and global geochemical cycles. Geochimica et Cosmochimica Acta ,SB, 2361-2386- Lau, Y.K. and Das, N.C. (1985) Precipitation quality monitoring in Alberta. PNWIS- APCA 1985 &muai meeting, Calgary, Alberta, November 13- 15. Legge, A.H. (1988) The present and potentiai effects of acidic and acidifjhg air poilutants on Alberta's EnWoument. Crtical Point 1, 1988. Final Report. Prep. for the Acid Deposition Research Program by the Biophysics Research Group, Kananaskis Cenm for Enviroimental Research, The University of Calgary, Calgary, Alberta, 664 pages. Lloyd, R.M. (1968) Oxygen isotope behavior in the sulfate water system. Journal of Geophysical Research ,73,6099-6 110. Longbelli, A. and Edmond, LM. (1983) Isotope geochemistry of the Amazon Basin: A recomaissance, Journai of Geophysical Research, 88,3707-37 17. 123 Martinec, I.; Siengenthaier, U.; Oescbger, H. and Tongiorgi, E. (1974) New insights into
nuisff mechanism by environmental isotopes. in Isotope techniques in groundwater hydrology, V1, Vienna: International Atomic Energy Agency, Vienna, 129-143.
Mayer, B.; Fritz, P.; Prietzel, J. and Krouse, H.R. (1995) The use of stable sulfur and oxygen isotope ratios for uiterprethg the mobility of sulfate in aerobic forest soils. Applied Geochemistry ,IO, 16 1- 173. Meybeck, M. (1987) Global chemical weathering of surficial rocks estimated hmriver dissolved loads. American Journal of Science, 401-428. Meybeck, M. and Helmer, R (1989) The Quaiity of Rivers: fkom pristhe stage to global pollution. Palueogeography, Palaeoclimatology, Palaeoecology, 75,283-309. Ming-hui, H; Siallard, R.F. and Edmond, J.M. (1982) Major ion chemistry of some large Chinese rivers, Natute, 298,550-553. Mook, W.G.; Bommerson, J.C. and Staverman, W.H. (1974) Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide. Eurth Planet. Sci. Lett., 22, 169- 176. Mossop, G.D. and Shetsen, 1. (compilers) (1994) Geological atlas of the Western Canada Sedimentary Basin. Canadian Society of Peaoluem Geologis ts and Alberta Research Council, 5 10 pages. Moran, S.R. (1986) Su15cial geology of the Calgary urbao area. Alberta Research Council, Terrain Sciences Department. Myric, R.H. (1992) Summary of air quality monito~gin Alberta 1992. Air assessrnent section, Environmental Qualit. Monitoring Branch, Environement Assessrnent Division, Alberta Environment Protection, 25 pages. Norman, A. and Krouse, H.R. (1992) Stable isotope studies of atmospheric sulphur: cornparison of Calgary, Canada and Bermuda. in Critical issues in the global
environment, vol. 2, Atmospheric Chemistry, Air and waste management association, 124 Pittsburgh, 16 pages. Norton, D. (1974) Chernical mass transfer in the Rio Tanama system, west-central
Puerto-Rico. Geochmtica et CornochimicaActa, 38,267-277 Ozoray, G.F. and Barnes, R. (1977) Hydroiogy of the Calgary-Golden are4 Alberta. Aiberta Researçh Council Report 77-2,38 pages.
Pawellek, F. and Veizer, 1. (1994) Catbon cycle in the upper Danube and its tributaries:
6I3C,,, constraints. Isr. Journui of Earth Science, 43, 187- 194.
Postel, S. (1993) Facing water scarcity. in State of the world 1993, edited by L. Starke, W. W. Norton and Company, New York, 22-4 1. Pnce, R.A. and Mountjoy, E.W. (1972a) Banff (est and West halves), 150,000 scale, Geological Sweyof Canada, Maps No. l294A and 1295A.
Price, R.A. and Mountjoy, E.W. (1972b) Mount Eisenhower (east and West halves), 150,000 scale, Geological Survey of Canada, Maps No. 1296A and 1297A. Price, R.A. and Mountjoy, E.W. (1978) Hector Lake (east and West halves), 150,000 scale, Geological Survey of Canada, Maps No. 1463A and 1464A. Price, R.A. and Mountjoy, E.W. (1979) Lake Louise (east and West halves), 150,000
scale, Geological Swey of Canada, Maps No. 1476A and 1477A- Rabinovich, A.L. and Grinenko, V.A. (1979) Sulfate sulfur isotope ratios for USSR river water. Geokhimja, No. 3,441-454. Reinalt, E.R. (1970) On the role of orography in the precipitation regime of Alberta. Albertan Geogmpher, 6,4548. Richey, J.E.; Devol, A.H.; Wofsy, SC; Victoria, R. and Riberio, M.N.G. (1988) Biogenic gases and the oxidation and reduction of carbon in Amazon River floodplain water, LimoL Oceanographer., 33,55 1-56 1. Riley and Chester ( 197 1) Introduction to marine chemistry. Academica Press. 125 Rodhe, A. and Kilüngtveit, A (1997) Catchment hydolgy. in Geochemical processes, weathe~gand groundwater recharge in catchments, edited by OM. Saether and P. Caritat, A.A. Badkema, Rotterdam, 400 pages. Rozanski, K.; So~tag,C. and Munnich, K.O. (1982) Factors controlling the stable isotope composition of Eumpean precipitation. Tellus, 34, 142-150. Salati, E.; Dail"Olio, A.; Matsui, E. and Gat, J.R. (1979) Recycling of water in the Amazon Basin: An isotopic study. Water Resources Research, 15, 1250- 1258. Sarh, MM.; Krishnaswami, S.; Dilli, K.; Somayajuia, B.L.K. and Moore, W.S. (1989) Major ion chemistry of the Ganga-Brahmaputra river system: Weathering processes and fluxes to the Bay of Bengal. Geochimica et Cosmochmtica Acta, 53,997-1009. Schlesinger, W.H. (199 1) Biogeochemistry - Aaalysis of global change. Academic Press, New York, 443 pages. Schwarcz, H.P. and Cortecci, G. (1974) Isotopic analyses of spring and Stream water sulfate from the Italian Aipes and Apenoines. Chernical Geology, 13,285-294.
Shakur, A. (1982) 34~and "O variations in terrestriai suiphates. Ph.D. thesis, Univ. of Caigary, Calgary. Aiberta, Canada, 229 pages. Sklash, M.G. and Farvolden, RN. (1979) The role of groundwater in storm runoff. Joumai of Hydrology, 43,4565 Sosiak, A. (1996) Bow River synoptic surveys 1994-1995, preliminary evaluation of resuits. Water Management Division, Alberta Environmentai Rotection, 103 pages. Staiiard, R.F. and Edmund, JM. (1981) Geochemistry of the Amazon 1. Precipitation chemistry and the marine contribution to the dissolved load at the time of peak discharge. Journal of Geophysical Research, 86, No. C 10,9844-9858. Staiiard, R.F., and Edmund, JM. (1983) Geochemistry of the Arnazon 2. The influence of geology and weathering envkonemnt on the dissolved load. Journal of Geophysical Reseurch ,88, No. C 14,9671-9688. 126 Taylor, BE.; Wheeler, M.C. and Nordstrom, D.K. (1984) Stable isotope geochemistry of acid mine drainage: Experimental oxidation of pyrite. Geochemica et Cosmochimica AmT48,2669-2678. Trembaczowski, A. and Halas, S. (1993) Sulphur isotopes in sulphates in natural waters:
( 1) surface waters of relatively unpolluted terraios. Isotopenpr~~~'s,28,2 15-228. van Donkelaar, C.; Hutcheon, I.E. and Krouse, HR(1995) 6%. 8180, 8 D in shallow
groundwater: tracing anthropogenic sulfate and accompanyhg groundwaterlrock interactions. Wuter,Air and Soi1 Pollution, 79,279-298. Van Stempvoort, D.R.;Hendry, M.J.; Schoenau, J.J. and Krouse, H.R. (1994) Sources and dynamics of sulfur in weathered till, western glaciated plains of North Arnerica. Chernical Geology ,111,35-56.
Van Stempvoort, D.R.; and Krouse, H.R. (1994) Controls of 6180in sulfate: Review of
experimental data and application to specific environrnents. in Environmental geochemistry of sulfide oxidation, edited by C.N. Alpes and D.W. Biowes, ACS Symposium Series 550,446480. von Grafenstein, U.; Erlenkeuser, H.; Müller, I.; Trimborn, P. and Alefs, J. (1996) A 200 year mid-European air temperature record preserved in lake sediments: An extension
of the 6180p-airtemperature relation into the past. Geochemica et Cornochimica Acta,
21,4025-4036. White, C. (1985) Wildland fires in Banff National Park 1880- 1890. Occasional paper No. 3, National Parks Branch, Parks Canada, Environement Canada, 10 pages. White, A.F., and Blum, A.E. (1995) Effects of chnate on chernical weathering in watersheds. Geochemica et Cosmochimica Acta, 59, 172%1747. Williams, M.W.; Brown, A.D. and Melack, LM. (1993) Geochemical and hydrologie controls on the composition of surface water in a high-elevation basin, Sierra Nevada, 127 California. Limnol. Oceanographer, 38,775-797.
Williams, M.W.;Yang, D.; Liu, F.; Tudc, I. and Melack, J.M. (1995) Controls on the major ion chemistry of the Üriimqui River, Tian Shan, People's Repubiic of China. Journal of Hydrdogy, 172,209-229. Wiwchar, B.; Perkins, E.H. and Gunter, W.D. (1988) SOLMlNEQ.88 PCISHELL manual. Alberta Research Council, Oil sands and hydrocarbon recovery department. Yanagisawa, F. and Sakai, H. (1983) Precipitation of SO, for sulphur isotope ratio measurements by thermal decomposition of BaS0,-V,O,SiO, mixtures. Anufytical Chemistry, 55,9859987. Yang, C.; Telmer, K. and Veizer, 3. (1996) Chemicai dynamics of the 'St. Lawrence"
nverine system: m), ,6'6,, 6'3Ç,1c,6% &. and dissolved ''~r/~~r.Geochemica
et Cosmachimica Acta, 60, 85 1-866. Yonge, C.J.; Goldberg, L. and Krouse, H.R. (1989) An isotope study of water bodies dong a traverse of southwestern Canada. Cunadian J. Hydrol., 106,245 - 255. Zeman, L.I. (1978) Mass balance mode1 for calculation of ionic input loads in atmospheric fdout and dishcarge from a mountainous basin. Hydrol. Sci Bull., 23, 103-1 17, Sample locations
Distance Station Station Name Location hm Access Number Comfluence (km) Bow Rivet
BR1 Lake Louise Trans Canada Highway, West 575 wading of Lake Louis BR2 Banff Banff townsite 510 bridge BR3 Carmore Canmore bridge 486 bridge BR4 Morely Morely bridge 435 bridge BR5 Cochrane Highway 22 405 bridge BR6 Bowness Nose Hill Drive, Calgary 378 bridge BR7 Edworthy Pedestnan bridge at Edworthy 367 bridge Park, Calgary BR8 LRT Pedestrian bridge below the 365 bridge noahwest LRT liae, Calgary BR9 Bonny Ogden Road, Calgary 355 bridge BR10 HW22X Highway 22X bidge 339 bridge BR1 1 Carseland Highway 24 287 bridge BR12 Cluny Highway 842 214 bridge BR13 Bow City Highway 539 110 bridge BR14 KW36 Highway 36 77 bridge BR15 Hays Highway 875 41 bridge
Tributaries BT1 Pipestone River Lake Louise townsite 574 B Spray River Banff golf course 509 wading BT3 Kananaskis River Highway 1 459 bridge Distance hm Access Comfluence
BT4 Ghost River Highway 940 425 wading BTS Jumping Pounnd Highway 1 407 wading Creek BT6 Elbow River 9th Street bridge, Calgary 362 bridge BT7 Nose Creek Calgary Zoo parking lot 360 wading BT8 Highwd River Highway 552 3 15 bridge Appendix 2 Chernical and stable isotope data for the Bow River (BR) and tributaries (BT)
Lake Louise 4.0 Banff 4.3 Canmore 6.6 More1y 8.3 Cochrane 1 1*4 Bowness 9.1 Edworthy 10.7 LRT 9.2 Bonny 8.1 HW22X 7.5 Carsland 6.9 Cluny 10.2 Bow City 9.9 HW36 0.0 Hays 10.1 Pipestone River no sample Spray River 4.6 Kananaskis 9.7 Ghost River 9.1 Jumping Pound 3.6 Elbow River 11.1 Nose Creek 8.1 Highwood River 6.6 Appendix 2 (continued) Chernical and stable isotope data for the Bow River (BR) and tributaries (BT) FaII 1993 Appendix 2 (continwd) Chernical and stable isotope data for the Bow River (BR) and tributaries (BT) Winter 1994
Lake Louise 0.0 4.09 Banff 0.0 3.17 Canmore 2.2 4.65 More1y no sample Cochrane 0.5 3.98 Bowness -0.1 5.18 Edworthy O. 1 N.A. LRT 0.5 3.62 Bonny 0.6 3.81 HW22X 0.2 3.50 Carsland 0.6 3.77 Cluny no sample Bow City no sample HW36 no sample Hays no sample Pipestone River no sample Spray River 0.0 8.64 Kananaskis 2.1 2.53 Ghost River no sample Jumping Pound no sampie Elbow River 0.4 NvA, Nose Creek no sample Highwood River no sample Appendix 2 (continueà) Chernical and stable isotope data for the Bow River (BR)and tributaries (BT) Winter 1994
BR01-1 0193 13.2 BR02-10193 15.3 BR03-10193 10.5 BR04-10193 no sarnple BR05-10193 14.8 BR06-10193 14.7 BR07-10193 14.6 BROS-10193 14.9 BR09-10193 14.9 BR10-1 0193 15.4 BR1 1-10193 14.5 BR12-1 0193 no sample BR1MOI93 no sample BR14-1 0193 no sarnple BR15-1 OB3 no sampie BTOI -10193 no sample BT02 - 10193 18.7 BT03 -10193 t 0.5 BT04 - 10193 no sarnple BT05 -10193 no sample BT06 - 10193 18.5 BT07 - 1OB3 no sample BTOB - t 0193 no sample
Appendix 2 (continued) Chemical and stable isotope data for the Bow River (BR) and tributaries (BT) Spring 1994 Appendix 2 (continuai) Chernical and stable isotope data for the Bow River (BR) and tributaries (BT) Summer 1994
Lake Louise -20.22 -3.80 19.92 -4.81 1.56 Banff -17.48 -4.03 13.13 -1 .O4 2.22 Canmore -19.83 -7.88 13.74 5.26 2.67 Morel y -19.90 -3.65 13.41 3.07 IOS~ Cochrane Bowness Edworthy LRT Bonny HW22X Carsland Cluny Bow City HW36 Hay s Pipestone River Spray River Kananaskis Ghost River Jumping Pound Elbow River Nose Creek Highwood River
APPENDIX 3 Chernical data for springs and ground water
T Ca Mg Na K HCO, S04 Cl Location pH OC m@ mJI mg/l mfl m@i mpli mgll
------Springs
Vermilian Lk 8.4 20.1 74.3 23.5 27.5 2.3 157 136.2 55.2 Canmore Ck 8.4 11.1 117 41.8 6.2 .9 252.4 213.6 2.1
Many Springs 8.6 1 1.4 77.4 24 1.3 -5 169.5 114.4 2.1
Yamanuska 6 67.4 15.9 13.9 1.2 292.1 7.2 5 -2
Big Hill 8.4 7.3 71.4 32-6 7.9 2-9 369 a.a. 7-2
SiIverSprings 7.8 n.a. 78.6 49.2 45.4 3.1 469 82.2 15.3
Shallow groundwater in Banff National Park (from Parks Canada) Appendix C Chernical and stable isotope data for Nose Creek (NC)
north Airdrie Airdrie North crossing South Crossing Country Hills 32nd Ave. NW CWY Centre Street
Spring 19%
north Airdrie Airdrie North crossing South Crossing Country Hills 32nd Ave. NW Cakary Centre Street ~~~g~wX~ omorrn~cu- CU rrrrrr