UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Creek temperatures in shaded reaches downstream of forestry activities, Central British Columbia Story, Anthony Charles 2002

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

Item Metadata

Download

Media
831-ubc_2002-0571.pdf [ 13.65MB ]
Metadata
JSON: 831-1.0090484.json
JSON-LD: 831-1.0090484-ld.json
RDF/XML (Pretty): 831-1.0090484-rdf.xml
RDF/JSON: 831-1.0090484-rdf.json
Turtle: 831-1.0090484-turtle.txt
N-Triples: 831-1.0090484-rdf-ntriples.txt
Original Record: 831-1.0090484-source.json
Full Text
831-1.0090484-fulltext.txt
Citation
831-1.0090484.ris

Full Text

Creek Temperatures in Shaded Reaches Downstream of Forestry Activities, Central British Columbia by Anthony Charles Story B.Sc, Trent University, 1999  A THESIS SUBMITTED IN P A R T I A L F U L F I L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF SCIENCE  in THE F A C U L T Y OF G R A D U A T E STUDIES Department of Geography  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH C O L U M B I A September, 2002 ©Anthony Charles Story, 2002  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  Department of Geography The University of British Columbia Vancouver, Canada  ABSTRACT Removal of streamside vegetation often causes daytime increases in stream temperature, due to greater warming by solar radiation. Several studies have reported creeks subsequently cooling as they flow back into shade, but none have established the causal processes. Fisheries and Oceans Canada observed downstream cooling during several summers in two small streams (catchments of 0.5 - 1.5 km ) that were monitored as part of a 2  multi-year study in the central interior of British Columbia. The purpose of this thesis was to use an energy balance framework to evaluate the physical processes responsible for the observed temperature patterns. Field investigations in July-August, 2000, showed that both streams continued to gain energy from the overlying atmosphere/vegetation on most afternoons as they flowed through the -200 m forested reaches. This effect only consistently resulted in measurable downstream warming in the upper section of the first reach, where groundwater did not contact the channel. Net inflow of groundwater in the lower section of the reach (-0.02 L s"  1  m" ) sometimes caused rapid downstream cooling, depending on the magnitude of 1  streamflow input at the head of the reach. The greatest downstream cooling (up to 4 °C) occurred when the streamflow input was <5 L s" , because infiltration in the upper 150 m of 1  the reach consumed the warmer streamwater from the upstream catchment. Modest inputs of groundwater (-0.002 L s" m" ) occurred throughout the second 1  1  study reach, causing an estimated 0.5 °C cooling in the average daily creek temperature. Two-way exchanges of energy between the creek and its subsurface, driven by conduction, and possibly hyporheic exchange, were important influences on the daily temperature extremes. Heat appeared to be stored within the riparian zone during daytime and released to the stream at nighttime. Heat storage also affected the mean daily temperatures, apparently depending on the relative strengths of the daytime heat sink and nighttime heat source effects. This study adds to a growing recognition of the high spatial variability in environmental responses to forest management, even within small watersheds. A better understanding of the hydrology-temperature linkage is required for effective management of the thermal effects of forestry.  ii  \  T A B L E OF CONTENTS ABSTRACT ." LIST OF FIGURES LIST OF T A B L E S LIST OF APPENDICES ACKNOWLEDGEMENTS  ii v ix x xi  CHAPTER 1 INTRODUCTION 1.1 Problem Statement 1.2 Previous Studies ". 1.3 Physical Controls on Stream Temperatures 1.4 Study Objectives & Thesis Organization  1 1 2 6 11  CHAPTER 2 S T U D Y SITE & M E T H O D S 2.1 Study Site 2.1.1 Location 2.1.2 Physiography & Bedrock Geology 2.1.3 Regolith 2.1.4 Vegetation 2.1.5 Regional Climate 2.1.6 Local Climate 2.1.7 Study Reaches 2.2 Methods 2.2.1 Creek Temperature Monitoring 2.2.2 Modelling Downstream Temperature Change 2.2.3 Stream-Groundwater Interactions 2.2.4 Meteorological Measurements & Calculations 2.2.5 Bed Heat Conduction 2.2.6 Hyporheic Exchange  12 12 12 12 17 19 19 20 20 24 24 24 26 32 36 37  CHAPTER 3 RESULTS 3.1 Overview of the Study Period 3.2 Stream Temperature Data Quality 3.3 Results for B5 3.3.1 Stream Temperature Patterns 3.3.2 Temporal Variations in Streamflow 3.3.3 Spatial Variations in Streamflow 3.3.4 Stream-Subsurface Interactions 3.3.5 Electrical Conductivity & Temperature Patterns 3.3.6 Temperature Patterns in Relation to Hydrology 3.3.7 Energy Exchanges across the Water Surface 3.3.8 Bed Heat Conduction  41 41 47 49 49 52 53 55 60 63 66 69  iii  3.3.9 Energy Exchanges Driven by Groundwater Inflow & Hyporheic Exchange 3.3.10 Stream Temperature Variations in Relation to Energy Exchanges 3.4 Results for B3 3.4.1 Overview of Stream Temperature Patterns 3.4.2 Variations in Streamflow 3.4.3 Stream-Subsurface Interactions 3.4.4 Tracer Tests 3.4.5 Electrical Conductivity & Temperature Patterns 3.4.6 Energy Exchanges across the Water Surface 3.4.7 Bed Heat Conduction 3.4.8 Energy Exchanges Driven by Groundwater Inflow & Hyporheic Exchange 3.4.9 Effects of Downstream Transport Processes on Temperature Patterns 3.4.10 Energy Balance Analysis of Downstream Cooling 3.4.11 Temperature Modelling Based on Groundwater Inflow 3.5 Summary of B5 and B3 Results  73 76 82 82 84 85 87 89 91 93 96 98 100 100 108  CHAPTER 4 DISCUSSION 4.1 Energy Balance Considerations 4.1.1 Energy Exchanges across the Water Surface 4.1.2 Bed Heat Conduction 4.1.3 Groundwater Inflow 4.1.4 Hyporheic Exchange 4.2 Catchment-Scale Interactions between Hydrologic & Thermal Regimes 4.2.1 Influence of Catchment Hydrology on Stream Temperature 4.2.2 Effects of Stream Temperature on Reach-scale Hydrology 4.3 Reach-Scale Linkages between Hydrologic & Thermal Regimes 4.3.1 Hydrologic Contrasts Between the Two Study Reaches 4.3.2 Temporal Linkages Between Hydrologic & Thermal Regimes 4.3.3 Linking Spatial Variations in Stream & Subsurface Thermal Regimes 4.4 Management Implications  109 109 109 111 115 118 122 122 123 125 125 126 128 131  CHAPTER 5 CONCLUSIONS 5.1 Summary of Key Findings 5.2 Suggestions for Future Research  135 135 136  REFERENCES APPENDICES  137 146  iv  LIST OF FIGURES Figure 1.1 Changes in summer-time afternoon creek temperatures as creeks flowed into forest, after being heated in clearings  3  Figure 1.2 Maximum daily stream temperatures in 1999 at three sites along three Baptiste Creek tributaries  5  Figure 1.3 Schematic of the three main terms in a traditional stream energy balance  6  Figure 1.4 Four stream types classified by their interactions with groundwater  8  Figure 1.5 Schematic of vertical and lateral hyporheic flow between a stream and its bed and banks  9  Figure 2.1 Map of western Canada, showing Fraser River Basin and major cities  13  Figure 2.2 Map of Stuart-Takla drainage basin  14  Figure 2.3 Map of Baptiste Creek study site  15  Figure 2.4 Cutblocks in the Baptiste Creek watershed  16  Figure 2.5 Photographs of road-cuts  18  Figure 2.6 Average monthly rainfall and snowfall recorded at the Takla Landing station from 1962-1990  19  Figure 2.7 Maps of study reaches  21  Figure 2.8 Long profiles of the two study reaches  22  Figure 2.9 Step-pool sequence at 60 m downstream along the B5 study reach  23  Figure 2.10 Mariotte bottle on upstream end of B5 culvert at head of study reach  27  Figure 2.11 Generalized cross-sectional schematic of the subsurface thermal and hydrologic monitoring transects used in the B5 study reach Figure 3.1 (a) Daily precipitation measured at the Middle River DFO camp. (b) Daily precipitation measured at the Baptiste Creek Open site. (c) Daily temperature extremes during the study period measured at the Open site  v  30  43  Figure 3.2 Incident solar radiation and precipitation measured at the Open site during three segments of the study period  45  Figure 3.3 Air temperature at the Open and Forest meteorological sites during i  the study period  46  Figure 3.4 Quality control of data-logged creek temperature observations  48  Figure 3.5 Creek temperatures from the three DFO data loggers in stream B5  50  Figure 3.6 Daily precipitation measured at the Open site and instantaneous estimates of streamflow at the B5 flume Figure 3.7 Streamflow measurements at the B5 culvert vs. estimates of  52  streamflow at the B5 flume  53  Figure 3.8 Streamflow along B5, measured using constant-rate salt injection  53  Figure 3.9 Spatial extent of streamflow in B5 from July 17 to August 30  54  Figure 3.10 Bed hydraulic gradients measured along B5  55  Figure 3.11 Hydraulic head at cross-stream transects in upper B5 and lower B5, measured July 27, 2000 Figure 3.12 (a) Daily precipitation measured at the Open site, (b) Changes in creek depth relative to depths on July 25, and (c) bed hydraulic gradients measured at 46 and 59 m downstream in B5. (d) Water table elevations recorded at selected wells in lower B5 during study period  56  58  Figure 3.13 Contour plots of phreatic surface through lower B5 on July 27 and September 4, 2000  59  Figure 3.14 Spatial patterns of electrical conductivity and temperature along B5 Figure 3.15 Relations between creek temperature and electrical conductivity measured along B5 on July 18 and July 29 Figure 3.16 (a) Creek temperatures at B5BR and B5LO. (b) Estimates of streamflow at the B5 flume, (c) Rainfall measured at the Open site  61  62 65  Figure 3.17 (a) Measured net radiation and modelled net longwave radiation at the Forest site over B5. (b) Air and creek temperatures measured at the Forest site, (c) Net radiation modelling results  67  Figure 3.18 Energy exchanges across the water surface in B5  68  vi  Figure 3.19 (a) Creek and bed temperatures measured at 40 m and (b) 183 m downstream, (c) Estimated bed heat conductivefluxesat 183 m downstream, and (d) creek water depths at 183 m downstream  70  Figure 3.20 (a) Creek and bed temperatures and (b) estimated bed heat conductive fluxes through B5 at -13:30 PST on July 28, 2000  71  Figure 3.21 (a) Estimated bed heat conduction, and (b) creek and bed temperatures at the 165 m pool site in B5  72  Figure 3.22 Downstream creek temperature change predicted for lower B5 with varying streamflow and groundwater conditions  73  Figure 3.23 (a) Creek water depths and (b) bed temperatures at 199 m pool. (c) Creek water depths and (d) bed temperatures at 199 m riffle  75  Figure 3.24 Downstream changes in mean daily creek temperature in upper B5, as a function of modelled daily Q*  77  Figure 3.25 Downstream changes in daily creek temperature extremes in upper B5, as a function of average modelled daily Q* during the 2 hour period preceding the first occurrence of the maximum or minimum at the downstream site  77  Figure 3.26 (a) Creek temperature observations from lower B5, and (b) streamflow estimates at the B5 flume  80  Figure 3.27 Creek temperature data from near the upstream and the downstream ends of the B5 study reach from August 10-18  81  Figure 3.28 Creek temperature time series from the three DFO data loggers in stream B3  83  Figure 3.29 Temporal variations in streamflow at the B3 culvert, measured using constant-rate salt injection  84  Figure 3.30 Spatial variations in streamflow along the B3 study reach  84  Figure 3.31 Bed hydraulic gradients measured along B3  85  Figure 3.32 Hydraulic head at cross-stream transects in B3  86  Figure 3.33 Tracer test injection results  87  Figure 3.34 Fitted parameters from OTIS-P analysis of tracer tests in B3  88  vii  Figure 3.35 (a) Temperature patterns in B3 on the afternoons of August 15,16. (b) Electrical conductivity patterns on August 14, 15, 16. (c) Streamflow on August 16  90  Figure 3.36 Relations between creek temperature and electrical conductivity measured along B3 on August 14 and August 15  91  Figure 3.37 Energy exchanges across the water surface at the B3 Forest site  92  Figure 3.38 Comparison of modelled and measured net radiation at the B3 Forest site Figure 3.39 Estimated conductive bed heat fluxes along B3 at -16:00 on August  93  15  94  Figure 3.40 Creek and bed temperatures measured in B3 Figure 3.41 Contour plot of predicted average daily creek temperatures at B3LO for the July 17-August 31 period Figure 3.42 Creek temperatures observed at three sites along B3, and temperatures simulated for the furthest downstream site using the OTIS advection-dispersion model Figure 3.43 Creek temperatures modelled for B3LO during July 17 to August 31 using a simple groundwater inflow model Figure 3.44  Figure 3.45  Analysis of residuals from B3LO mean daily creek temperatures simulated using groundwater inflow model (Fig. 3.43a) Statistical analysis of model residuals from Fig. 3.43  viii  95  97  99  102  104 106  LIST OF TABLES Table 1.1 Details of studies plotted in Fig. 1.1  4  Table 3.1 Peak snowpack water equivalent and snowmelt streamflows observed in the Baptiste watershed  42  Table 3.2 Summaries of air temperature and precipitation records from the Middle River camp  42  Table 3.3 Climatic characteristics of the three segments of the study period, and comparison of precipitation measured at the Open site to that measured at the Middle River Camp  44  Table 3.4 Summaries of mean daily temperature extremes at three sites along stream B5 from July 17 to August 31, 2000  51  Table 3.5 Comparison of modelled and observed afternoon creek warming in upper B5  78  Table 3.6 Summaries of mean daily creek temperature extremes at three sites along stream B3 from July 17 to August 31, 2000  82  Table 3.7 Best fit values of hydrologic parameters for four tracer test segments along B3, obtained using OTIS-P, and various subsequently calculated values  89  Table 3.8 Summary of influences on B3LO creek temperature daily maximum on August 15, 2000, relative to B3BR daily maximum  100  Table 3.9 Best simple (bivariate) and multiple (trivariate) linear regressions between daily model residuals from B3 groundwater inflow modelling, daily air temperature variables measured at the Open site, and total daily incoming solar radiation measured at the Open site  105  ix  LIST OF APPENDICES Appendix 1. Description of the error estimation for the terms in Eqs. 2.3, 2.4  146  Appendix 2. (a) Creek temperatures observed in upper B5 using Campbell Scientific data-loggers, and (b) modelled net radiation at 40 m downstream from July 18-29  147  Appendix 3. (a) Creek temperatures observed in upper B5 using Tidbit dataloggers, and (b) modelled net radiation at 20 m downstream from August 4-13  148  Appendix 4. Downstream temperatures changes (B3BR-B3LO) modelled during July 17 to August 31  149  Appendix 5. Groundwater inflow modelling for B3 including daily variations in groundwater temperature  150  Appendix 6. Matrix of coefficients of determination (r ) for simple linear regressions between daily model residuals from B3 groundwater inflow modelling ("RD"), daily air temperature ("T") variables measured at the Open site, and total daily incoming solar radiation ( " K i " ) measured at the Open site  151  Appendix 7. Checks of energy balance estimates reported in the literature: (a) estimation of Brown's (1969; 1972) thermal conductivity value for gravel stream beds, and (b) examination of Comer and Grenney's (1977) stream energy balance fluxes  152  Appendix 8. Longitudinal dispersion coefficients ("D") estimated in six studies using tracer tests, as a function of streamflow: (a) linear plot, (b) log-log plot, (c) Log-log plot of D as a function of mean flow velocity ("u"). (d) and (e) are linear and log-log plots, respectively, of u as a function of streamflow  153  2  x  ACKNOWLEDGEMENTS While only my name appears on the title page of this thesis, the list of people who have contributed in some way would not fit on one page. From a "group effort" perspective, most of the Canadian population contributed unknowingly, in the form of a Natural Sciences and Engineering Research Council of Canada postgraduate scholarship awarded to me. In terms of time and effort expended by an individual, Prof. Dan Moore proved an outstanding supervisor, by providing enough guidance to allow successful completion of a valuable final product, while allowing me to make enough of my own mistakes to ensure a rich learning experience. His patience with my often glacial progress, and his fair-minded approach set examples for me. Prof. Olav Slaymaker gave invaluable advice from the outset. Thanks to Ragnar Kaltenbach for providing excellent assistance in the field, and Eric Mellina for help early in the field season and continued interest in my work. Fisheries and Oceans Canada (DFO) provided logistic support, data, and, most importantly, interesting field sites. In particular, I thank Herb Herunter and Steve Macdonald of DFO. Thanks to Profs. Tim O k e ' and Andy Black for lending valuable instruments despite the imminent threat of "bear damage" in the field. While too numerous to list by name, I thank all of my fellow students in the Department of Geography and residents of St. John's College who offered camaraderie and encouragement during my time at U B C . It is impossible to overstate my gratitude to friends and family in northeastern Ontario, who have provided many kinds of support during both my time at U B C and during the past year when I retreated to the comforts of home. Renee Duval put up with my periods of introspective "mental data analysis," and even feigned interest in obscure topics like the expression of riparian heat storage in model residuals. M y sister Maria and her family welcomed me into their home at a busy time in their lives. Although coming home in September 2001 doubtless delayed my completion of the thesis, spending this time with Maria et al. and my dad made the delay worthwhile (at least in my mind). Finally, I thank my mom. Without her love, I doubt that I would have graduated from high school, let alone entered graduate school.  xi  CHAPTER 1 INTRODUCTION 1.1 Problem Statement Water temperature affects a wide array of chemical, physical, and biological processes in the aquatic environment. Temperature not only directly limits the abundance of dissolved gases, but also indirectly controls the abundance of certain compounds within streams by regulating the rates of biochemical reactions (Butturini and Sabater 1998). The temperature of liquid water determines its viscosity, affecting physical variables including the hydraulic conductivity of sediments through which the water travels. Because the hydraulic conductivity of saturated sediment at 25 °C is twice as high as at 0 °C, creek temperature can affect rates of water loss from a channel by infiltration through the bed, and thus the amount of flow in the channel (Constantz et al. 1994). The removal of streamside vegetation has long been implicated as a cause of declines in North American cold-water fish populations that followed European settlement. Titcomb (1926) stated that the clearing of riparian forests increased daytime creek temperatures by exposing the stream "to the direct action of the sun." He believed that this phenomenon was the main factor responsible for the disappearance of Atlantic salmon and trout from many streams in New England. Since Titcomb's early paper, water temperature has indeed been shown to be a key determinant of the distribution of fish species (Barton et al. 1985; Taniguchi et al. 1998), and his general theory about the processes linking deforestation and increased stream temperatures has been confirmed numerous times (e.g., Brown 1969; Beschta and Taylor 1988). However, there is an emerging recognition that our understanding of the mechanisms controlling creek temperatures is inadequate (Webb and Zhang 1999; Gaffield 2000; Johnson and Jones 2000a). For example, although Johnson and Jones (2000b) found that increased stream temperatures following forest disturbance could be attributed to more solar radiation reaching the water surface, they also observed high thermal variability along the length of one channel that appeared to be related to variations in substrate type. Hypothesized downstream changes in diurnal temperature fluctuations were an important element in the development of the influential River Continuum Concept describing variations in biological  1  communities within river systems (Vannote et al. 1980). Nevertheless, few observational studies have examined the thermal effects of forestry at the watershed scale. The present study was motivated by a specific question raised by fisheries biologists and foresters: what processes control the thermal characteristics of creek reaches located downstream of forestry activities (i.e., cutblocks and roads)? Previous research has revealed that creeks sometimes cool as they flow back under intact forest canopy during the daytime, ameliorating the effects of warming in cutblocks, but no studies have focused on identifying the mechanisms responsible for such cooling patterns where they occur. Some streams in British Columbia have recently been designated "temperature sensitive," due to concerns about fish species that require cold water habitat. For example, the Bull Trout (Salvelinus confluentus) requires water of 2-7 °C for spawning and <10 °C for rearing (Teti 2000). While it is relatively obvious that streamside forest harvesting should not occur in areas adjacent to Bull Trout habitat, the indirect impacts associated with changes in upstream shade cover are unclear. Environmental managers would like to understand, and ideally predict, how forestry activities along non-fish-bearing streams affect downstream fish habitat. The following section reviews previous studies of thermal behaviour downstream of cutblocks. Section 1.3 then reviews the physical controls on stream temperatures to provide the background for the research objectives, presented in Section 1.4.  1.2 Previous Studies Studies in North America indicate a wide range in temperature changes downstream of forest openings (Fig. 1.1, Table 1.1). It is possible to speculate on the causes of the scatter in the data. For example, the lack of measurable change observed by Brown et al. (1971) might be related to the greater thermal inertia of that relatively large stream (Table 1.1). There is also evidence that the magnitude of downstream cooling is positively related to the maximum creek temperature. However, this relation ceases to be significant when the downstream rate of temperature change (°C m" ) is considered. The greatest amount of 1  cooling occurred through the two shortest reaches (Fig. 1.1, Table 1.1), perhaps suggesting non-linearity in downstream cooling rates. These studies demonstrate that the phenomenon is controlled by processes which act at disparate rates in different locations.  2  2  0 +  5  B4  (cloudy.days) -  B3  B5 -2  O o H <  -4  5  50  (sunny days)  70  90  110  130  150  170  190  210  230  Length of Shaded Reach ( m )  Figure 1.1 C h a n g e s in s u m m e r - t i m e a f t e r n o o n c r e e k t e m p e r a t u r e s ( m a x i m u m daily w h e r e available) a s c r e e k s f l o w e d into forest, after b e i n g h e a t e d in c l e a r i n g s ( A T = f o r e s t - c l e a r i n g ) . S e e T a b l e 1.1 for details of e a c h study. D a t a f o r B 3 - B 5 a r e f r o m J u l y - A u g u s t 1 9 9 9 ( m e a n ± maximum/minimum).  In the Baptiste creek drainage of central British Columbia, Fisheries and Oceans Canada (DFO) monitored temperatures along three small tributaries affected by forestry activities during the summers of 1997-1999 (Moore et al. in press). The range of changes in maximum daily temperatures downstream of the clearings observed during July and August of 1999 spans almost the entire range of observations from other areas of North America (Fig. 1.1), indicating that the rates of the cooling processes vary substantially even within small watersheds. Creek B4 continued to warm downstream of the forest disturbances. In contrast, like most previous studies, temperatures in the other creeks (B3 and B5) almost always declined substantially downstream of cutblocks and logging roads (Fig. 1.1). Continuous temperature records from several sites within the Baptiste Creek drainage allow a more detailed examination of the spatial and temporal variability suggested by points B3-B5 in Fig. 1.1. The relatively undisturbed creek (B4) always warmed slightly in the  3  E P  ro  CD CO  zs •—'  0  •&T5 CD 0 0)  E  H  |  §  o o  £ ~ o 0 co o. c o 0) L5 o £ 2 > ro co  si I-  s** CO  £ ro  I!  CP CU JQ  5 o  ro ro >  E ro 2  O  =  0  to,  VI X <D CD -~ 3 CD "co  C L  co CD  Q  CO  ^  . CO CD SZ  o o ro  CM CD  3  m-Q E .E CO "  CD CD  2  CD  CO 0  S §I  ro  CD c  0  0  ro  a>  E ^  CO  o  o  0 O)  i l l  O D roC o  ro  <  0  CD  CD  5 =5 o SZ o « >ii _ro  I  0  15 c co  •5  i-  T3 C L 0 C O  o  CO •£ J3> O  I  _ro  LL  ro to - ro ro x: c CO  CD  (5 CD  2  CM  2  0  t  > CD O C D CO E -i = 0 <f> cn C cz CO O  .«  co _  -  >-  o O O co CD  CJ)  3  x:  co  O o  0 1  CL CZ  O !Z 0  0 ro ro c 0 Q. 0 0 ; =  ^ .52 CO C L  CO  .!2  CO  co CO  ro E 0  2 I £ V 5 => -D t; XJ CD O  CD  a*? 0  0  4  S  0  ro  CO CO  _l LL  1 sa  J= 3 CD c -= 0  CO CO  f> r- ^, 13  5  C  O co  o  8£  O o  ^1  CD 0  00  c  CD  CO m CD  S, E m ro  0  cn cn cz cn co ro  TD  0 .£ > sz  o ro  C CO O  0  C O CO °> 5  o  E  ro in 0 co  K_  o  C XJ  CO  .  0  v§  0 >  3  o  m ro  CD ,<Uro  CO D) x: x: c H o b CC £ax: TD 0 a. c >co • C O o P "c: ? CD o t~ 0  0 >  S|  0  cororo c c to o ra JQ  -Q o  CM C D0 0 1 70 T cz o3  o — co —3 ro ro r c < o o cz O— £ 0 o J m 2 1C c E D — c: O»o ro - O "a C D) CD 0 0 CL q o ^ co Z 5 Oiu C L  TD O  V, ro  O  ~  ro  o x: sz o ™*a co x: D) ro c CD 5  to  S  -<8 •->.  O  a  co i  CM I  cz  IS  c o o co ro CO S .2 £" 0 .55 "55 o C D to to £ C O ll. 0 O CD TD £ 0 ro a> CD -o C O  00 CD  g  o~-  h-  *w —8  c ro  CO6  xi ro 'ro > ro o cz  o  ro CO J= 2 x: ro  .E ro  CO >>i- £  E  —•  CO  CM  O  TD  E ro  a>  g?  0  CD i  T3  O 5  • l l  O  c: 0  •= 0 — — ^> ro 2 0 o  ro 0  0 T3 >> x: TD  o  CD§ 21 O CD o -t-» C o CD cCoD x: E _ CO CO TDx: T3 E CD  CD  1  c  TD O  m °  0 LL ro •g TD  0  o  7  CD  to  c  CD CD C  E •EX ._  CM  *" CD 0 > 0-0)0 _ CO XI CM 0 ro  0  5=  ~  i_  CD O £ O  0  3 •=)  <  x: co x 8.9. i= •D r c ® O "O  O  O) -> CD 3  lo*  0  0  CD  CM  X5  ro  O CD  CM  cz  h-  i  .E  x» ro ro > ca o  o  sz sz  CM  CNJ  ay  c  *i_  ro#  X2  H  CO E  i : 3 (D  ro  CD  m'  iS rs  .55=8  x> I— co E i= 3 CD c: T r 0 m  O  1-  o O O  O 1 - co o o O O c 0 0  CO  1  0  CL  \— CL CZ  CZ  0 0  to CO  downstream direction (Fig. 1.2a). Creeks B3 and B5 also warmed as they traveled through their more heavily disturbed catchments, but then generally cooled by 1 to 5 °C after flowing back into undisturbed forest (Fig. 1.2b,c). Thermal patterns along stream B3 were consistent, with temperatures always higher at both sites downstream of the disturbances than upstream, but some cooling invariably occurred as the creek flowed back into intact forest (BR>LO>HI in Fig. 1.2b). Thermal patterns along stream B5 were more variable in time. Temperatures in the downstream forest were sometimes lower than those observed upstream of the cutblocks and logging roads, while at other times no cooling was observed as the creek flowed into the forest (Fig. 1.2c). Fig. 1.2c is particularly interesting from a management perspective, since the heating that occurred through the cutblock and logging roads appeared to dissipate completely after the creek had flowed less than 200 m into intact forest. This rapid cooling suggests that  6 1-Jul  16-Jul  31-Jul  15-Aug  30-Aug  Figure 1.2 M a x i m u m daily s t r e a m t e m p e r a t u r e s in 1 9 9 9 at t h r e e sites a l o n g t h r e e B a p t i s t e C r e e k tributaries. T h e t e m p e r a t u r e l o g g e r s a r e located a b o v e t h e logging r o a d s a n d c u t b l o c k s ( H I ) , w h e r e t h e s t r e a m f l o w s b a c k into u n d i s t u r b e d f o r e s t ( B R ) , a n d a p p r o x i m a t e l y 140 to 2 0 0 m into t h e f o r e s t ( L O ) . S e e T a b l e 1.1 f o r details o n t h e nature of f o r e s t r y activities in e a c h tributary d r a i n a g e . Briefly, B 4 w a s a f f e c t e d only by logging r o a d s , w h e r e a s B 3 a n d B 5 f l o w e d t h r o u g h c u t b l o c k s . 5  harvesting along headwater streams may sometimes have negligible thermal impacts on downstream waters. However, the pattern was not persistent, and it is not clear what processes caused the stream to cool so rapidly. Without an understanding of the controlling mechanisms, there is no way to predict how the thermal characteristics of particular stream reaches will respond to upstream forestry activities.  1.3 Physical Controls on Stream Temperatures A traditional steady-state energy balance of small streams (e.g., Brown 1969) devoid of tributary inflows can be simplified to: AQs = Qu + Qc +  (1.1)  Q G W  where AQs is the net change in heat stored within the streamwater, Qu is the net heat exchange across the water surface, Qc is the conduction of heat between the water and the bed of the stream, and Q G W is the heat flux caused by groundwater inflow (Fig. 1.3). It should be noted that cool groundwater inflow causes stream temperature decreases because the concentration of heat is diluted, not because heat energy is lost from the stream channel (Poole and Berman 2001).  6  This energy balance approach defines the channel of a particular stream reach as the "control volume" (Fig. 1.3), and the net heat exchange across the water surface can be expanded to: Qu = Q* +  Q H +  (1.2)  Q E  where Q* is the sum of net solar radiation (K*) and net terrestrial radiation (L*), Q H is the sensible heat flux associated with turbulent air movement, and Q E is the latent heat flux (evaporation/condensation).  i  Other studies have considered the heat flux associated with precipitation falling on the water surface, but this is consistently found to be a negligible term (e.g., Webb and Zhang 1997; Evans et al. 1998). Webb and Zhang (1997) and Evans et al. (1998) considered the heat generated by friction along their study reaches. They estimated that friction was sometimes a significant source of energy, but since this term can only warm streams it cannot help explain the observed daytime cooling patterns in the Baptiste Creek tributaries. Longitudinal dispersion can affect temperatures over long distances of stream, but the error caused by ignoring longitudinal dispersion is typically less than the errors associated with calculating the fluxes of energy into and out of the stream (Polehn and Kinsel 2000). Of the results shown in Fig. 1.1, only Brown et al. (1971) used an energy balance approach to identify the physical mechanisms responsible for the downstream temperature changes in the forested reaches. For one clear July day, they found that the sum of Qu and Qc was positive throughout mid-day, although the forest canopy substantially reduced inputs of solar radiation. The latent and conductive heat fluxes were the only cooling (negative) terms, but these were offset by the warming influences of Q* and Q . Brown et al. (1971) H  concluded that downstream temperature reductions are usually caused by groundwater inflow. Because groundwater inflow is most frequently cited as a potential explanation for creek cooling downstream of forestry activities (Brown et al. 1971; Beschta et al. 1987; McGurk 1989; Keith et al. 1998), it is worth considering the fundamental controls on streamgroundwater interactions in more detail.  (  7  Where streamflow increases downstream as a result of groundwater discharge to the channel, the stream is said to be "gaining" (Fig. 1.4a). In contrast, a "losing" stream is characterized by decreasing streamflow in the downstream direction. Streamflow losses can occur when the stream is perched above the water table (Fig. 1.4b), in which case the stream water must travel through an unsaturated zone. Alternatively, losing streams may be connected to the water table (Fig. 1.4c). Finally, a "flow-through" stream simultaneously loses flow through one of its banks while gaining water from the other (Fig. 1.4d), in which case streamflow may increase, decrease, or remain constant downstream.  F i g u r e 1.4 Four s t r e a m types classified by their interactions with g r o u n d w a t e r ( D i n g m a n 1 9 9 4 ) : (a) g a i n i n g , ( b ) losing a n d p e r c h e d , ( c ) losing a n d c o n n e c t e d , ( d ) f l o w - t h r o u g h . T h e w a t e r table is indicated by t h e d a s h e d line, with a r r o w s r e p r e s e n t i n g s u b s u r f a c e flow.  8  Recently, hydrologists and ecologists have recognized that water movement between streams and the subsurface is more intricate than described in Fig. 1.4. Rather than flowing in one direction from groundwater to stream (or vice-versa), streamwater travels in and out of the bed (Thibodeaux and Boyle 1987) and banks (Harvey and Bencala 1993) as it moves through what is sometimes termed the "hyporheic" zone (Findlay 1995). Harvey and Wagner (2000) indicated that hyporheic flow occurs across a wide range of time (10 s - 100 day) and space (1 cm - 100 m) scales, including both the lateral and vertical dimensions (Fig. 1.5). Hyporheic flowpaths leave and return to the stream channel many times within the length of a single study reach, whereas groundwater flowpaths enter or exit the stream reach only once (Harvey and Wagner 2000).  Cross-section  Figure 1.5 Schematic of vertical and lateral hyporheic flow (dashed lines) between a stream and its bed and banks (Findlay 1995). While flow through riffles (longitudinal section) and meanders (plan view) is common, other features (e.g., steps and woody debris) may also induce hyporheic exchange.  9  The distinction between stream-groundwater exchange and hyporheic exchange can be better understood within the context of a water balance at the stream reach scale: dqldx-= q" - q ^ + q™ - q° " - q  (1.3)  0  w  YP  H  YP  E  where q is streamflow, x is distance downstream, q"  GW  is groundwater inflow to the channel  (L s" m" ), q™' is groundwater outflow (L s" m" ), g ^ - i s hyporheic inflow (L s" m" ), q™" 1  1  1  1  1  1  w  YP  is hyporheic outflow (L s~ m" ), and q is effective outflow due to evaporation (or inflow due l  1  E  to condensation) on the water surface (L s" m" ). At the stream reach scale q™ is assumed 1  1  YP  to balance q™ (Harvey et al. 1996), so that non-zero dq I dx values are caused either by net YP  (  exchanges in q  GW  or q . E  Much research has considered the role of hyporheic exchange in altering streamwater chemistry (e.g., Findlay et al. 1993; Mulholland et al. 1997; Harvey and Fuller 1998) since the chemically reactive minerals and active microbial life of the subsurface transform and retain dissolved constituents (Harvey and Wagner 2000). White et al. (1987) used river bed temperatures to assess hyporheic flow and several other studies have since considered its influence on subsurface temperatures (e.g., Wondzell and Swanson 1996; Evans and Petts 1997; Malard et al. 2001). Castro and Hornberger (1991) interpreted White et al.'s (1987) findings as indicating that hyporheic exchange causes stream temperature variations. However, no published studies have attempted to quantify this process. Bilby (1984) identified the causes of "cool water areas" along a fifth-order stream in western Washington affected by forestry. He attributed thirty of the thirty-nine areas along the 3.5 km reach to groundwater inflow (lateral or pool bottom seeps), four were located at confluences with cooler tributaries, and five of the cool water areas were caused by what Bilby termed "flow through the bed." These latter areas, linked to what we now might label hyporheic flow, accounted for 57% of the total area of cool water along the reach despite being relatively uncommon in number (Bilby 1984). More recently, Evans et al. (1998) recognized the potential significance of heat fluxes due to water flow through the bed, but then argued that they might not need to be quantified because the conductive bed heat flux term "probably incorporates a proportion of the other heat flux components." The empirical evidence of Bilby (1984) suggests that hyporheic flow can be an important component of  10  stream thermal budgets, and so the steady-state energy balance (Eq. 1.1) is expanded here to include a hyporheic term  (QHYP)  on the right-hand side.  1.4 Study Objectives & Thesis Organization The objectives of this thesis are: 1. To document the thermal behaviour along shaded reaches downstream of forestry activity, particularly the occurrence of downstream cooling, and 2. To apply an energy balance framework to evaluate the physical processes responsible for the observed temperature patterns.  The thesis is divided into five chapters. Chapter 2 describes the study site and methods used to achieve the objectives. The thesis results are outlined in Chapter 3, with separate sections focusing on results from each of the two study reaches. Sub-sections of the chapter describe the hydrologic environments of the reaches, evaluate the magnitudes of energy fluxes within the reaches and relate the observed stream temperature patterns to the energy fluxes. Chapter 4 discusses the relative importance of the various energy fluxes at the two reaches, describes the underlying influences on their thermal behaviours, and evaluates the management implications of the thesis results. Finally, Chapter 5 summarizes the key findings of the thesis, and outlines suggestions for future research.  11  CHAPTER 2 STUDY SITE & METHODS 2.1 Study Site 2.1.1 Location The study was conducted at the Baptiste Small Stream Study Area of the Stuart-Takla Fish-Forestry Interaction Project. The Stuart-Takla Project is a multi-disciplinary, multiagency study initiated by DFO in 1990 to address the lack of knowledge about the effects of forestry practices on aquatic ecology in the interior of British Columbia (Macdonald et al. 1992). The project initially focused on four creek drainages with areas of 36-75 km , since 2  their downstream reaches are used as spawning habitat by the economically important sockeye salmon (Oncorhynchus nerka). The Baptiste Creek tributary catchments (areas of 0.5-1.5 km ) were added to the 2  Stuart-Takla Project in 1996 because of concerns about the effects of forestry on smaller streams and species such as rainbow trout (Oncorhynchus mykiss) that depend on them. The Baptiste Creek study area is located at 54°51' N , 125°20' W within the Stuart-Takla drainage basin of central British Columbia, which is the most northerly sub-basin of the Fraser River (Figs. 2.1, 2.2). Harvesting of forests within the B3 and B5 catchments was carried out in January, 1997 (Figs. 2.3, 2.4).  2.1.2 Physiography & Bedrock Geology The terrain of the Baptiste Creek watershed is of moderate relief, lying between the Hogem Ranges of the Omineca Mountains to the north, and the Fraser Basin plateau to the east (Collett and Ryder 1997). The study streams drain the northeast-trending ridge that forms the southern boundary of the Baptiste Creek watershed (Fig. 2.2). The B5 catchment is moderately steep, with the channel flowing in a mainly northerly direction (Fig. 2.3) at a mean gradient of - 7 % (Beaudry 1998). The B3 channel flows toward the northwest (Fig. 2.3) and is substantially steeper (mean gradient -26%). Maximum elevations at the drainage divides range from 1160 m a.s.l. for B5 to 1340 m for B3. The channels drain to a swamp and lake complex at an elevation of-930 m, before subsequently flowing into the Baptiste Creek proper -3 km downstream. Baptiste Creek eventually drains into the Middle River at 695 m a.s.l. 12  Figure 2.1 Map of western Canada, showing Fraser River Basin, and major cities. The StuartTakla watershed is located within the rectangle. Prince of Wales Island, Alaska, was the location of the studies of Keith et al. (1998), and Hetrick et al. (1998).  13  Figure 2.2 Map of Stuart-Takla drainage basin, northwest of the town of Fort St. James, showing Baptiste Creek study site, Middle River DFO camp, and Takla Landing.  14  Figure 2.3 Map of Baptiste Creek study site, modified from Christie (1998). "H" represents the HI creek temperature monitoring sites, "B" the BR sites, and "L" the LO sites.  15  Figure 2.4 C u t b l o c k s in t h e B a p t i s t e C r e e k w a t e r s h e d , (a) L o o k i n g t o t h e northwest a c r o s s t h e B 5 c u t b l o c k f r o m a b o v e t h e u p p e r r o a d ( F i g . 2.3), w i t h Mt. S i d n e y W i l l i a m s ( 1 9 8 6 m a.s.l.) in t h e d i s t a n c e T h e riparian a r e a o f t h e B 5 c r e e k is v i s i b l e in t h e l o w e r left t o l o w e r m i d d l e o f p h o t o , d i s t i n g u i s h e d by t h e s t a n d i n g t r e e s o f t h e " a g g r e s s i v e b u f f e r . ' (b) T h e B 3 c a t c h m e n t in t h e f o r e g r o u n d w i t h o n e o f t h e un-named Baptiste Lakes visible to the northwest, taken f r o m the upper edge of the cutblock. Note t h a t a) w a s t a k e n in A u g u s t , 2 0 0 0 a n d b) in J u n e , 2 0 0 0 . 16  The bedrock underlying the study catchments consists of ultra-basic intrusives including amphibolite, peridotite, their serpentinized products, and serpentine itself (Collett and Ryder 1997). Collett and Ryder (1997) suggested that the terrain of the study area is generally bedrock-controlled, although they also stated that the landforms and surficial materials have been heavily influenced by glaciation. At the peak of the last glacial advance, ~15 000 years ago, ice buried the entire study site and probably flowed in a northeasterly to easterly direction (Collett and Ryder 1997),  2.1.3 Regolith The bedrock of the study catchments is mainly covered by basal till, which varies in depth from <1 m to greater than several meters thick (Collett and Ryder 1997). Collett and Ryder (1997) mapped surficial materials within the Baptiste Creek watershed using observations at road cuts, tree-throw hollows, debris slide headscarps, and in soil pits. Their map does include inter-polygon variations in till depth but it is difficult to identify the B3 study stream on the map and the substantial uncertainties associated with the estimated till depths at sites other than road cuts prohibit a reliable assessment of till depth variability. However, field observations during the summer of 2000 indicated shallower till depths near the outlet of the B5 catchment than near the outlet of the B3 catchment (i.e., just upstream of the study reaches - Fig. 2.5). Textural analyses performed by Collett and Ryder (1997) indicate that the basal till consists of a silty sand or clayey silty sand matrix containing ~30% clasts (mainly pebbles and cobbles). The upper 1-2 m of the till has weathered to produce a looser, more permeable material than the deeper, highly cohesive, unweathered till (Collett and Ryder 1997). Study area soils are dominated by grey luvisols, although humo-ferric podzols are also found (Valentine et al. 1986).  17  Figure 2.5 Photographs of road cuts at (a) right-hand bank of B5 creek, upstream of study reach, (b) left-hand bank of B5 creek, upstream of study reach, (c) B3 catchment. Note that the road in a) and b) did not exist when Collett and Ryder (1997) made their till depth observations.  18  2.1.4 Vegetation The study catchments lie within the Sub-boreal Spruce biogeoclirnatic zone (B.C. Ministry of Forests 1995). Dominant tree species include Englemann spruce (Picea engelmannii), Alpine fir (Abies lasiocarpa), and Lodgepole pine (Pinus contorta). Mountain ash and alder, juniper, wild rose, soapberry, stinging nettles, and devil's club are examples of common plants in the forest understory (Christie 1998).  2.1.5 Regional Climate The climate of this portion of British Columbia's central interior can be described as cool and moist. Mean annual air temperature at Fort St. James (Fig. 2.2) is 3 °C, with mean monthly air temperatures of 15 °C in July and -12 °C in January (Environment Canada 2002). Mean annual precipitation is 520 mm at Takla Landing (location in Fig. 2.2) and 480 mm at Fort St. James, with 60% rain and 40% snow at both sites (Environment Canada 2002). Mean monthly air temperatures peak at 14 °C in July, and mean air temperature in January is -10.5 °C. Most precipitation falls as snow from November to March, with a monthly precipitation minimum in April, and almost constant average monthly rainfall from June to October (Fig. 2.6). Regional annual evapotranspiration is on the order of 300 mm (den Hartog and Ferguson 1978).  2,  60 -.  Figure 2.6 Average monthly rainfall and snowfall recorded at the Takla Landing station from 1962-1990.  19  2.1.6 Local Climate Beaudry's (2001) snowcourse data show that peak snow water equivalent averaged 60% greater at the study site during the 1996-2000 period than mean annual snowfall recorded at the Takla Landing station from 1962-1990 (340 mm vs. 210 mm). It is possible that the differing measurement periods and/or local site factors such as drifting snow caused this discrepancy. However, orographic enhancement of snowfall at the -400 m higher Baptiste site is a more likely explanation. Similar evidence of orographic enhancement of rainfall (section 3.1) suggests that mean annual precipitation in the Baptiste Creek tributary .catchments is probably closer to 800-900 mm than the 520 mm measured at Takla Landing. Combined with the annual evapotranspiration estimate of den Hartog and Ferguson (1978), this suggests that annual runoff from the Baptiste Creek tributary basins is 500-600 mm, neglecting other possible water balance terms such as recharge of deep groundwater (i.e., precipitation = evapotranspiration + runoff, on an annual basis).  2.1.7 Study Reaches Each reach was shaded by the forest canopy and generally dense undergrowth, except for a section immediately downstream of the culverts where the creeks flowed through the road "right-of-way," (upstream of BR sites in Fig. 2.7). The B3 study reach includes the confluence with stream B2 (Figs. 2.3, 2.7b), but no surface flow occurred from B2 during much of the study period and for the sake of simplicity that study reach will be referred to simply as "B3." Surveys of the study reaches conducted using a level and compass during the summer of 2000 showed that mean channel gradients within the study reaches (7% in B3, 3% in B5) were substantially lower than in the upstream catchments (26% for B3, 7% for B5). Gradients also declined with distance downstream in the study reaches, particularly in the case ofB3 (Fig. 2.8). Bed materials of both reaches were dominated by gravel (2-64 mm). Data from Christie (1998) indicate that the size fractions >2 mm accounted for 68% of sample weight at B3 and 76% at B5. Trends of downstream fining suggested by Christie's data, collected in both reaches during 1996 and 1997, were supported by field observations during the study  20  a  199m-^' ^183 m 165m-»f^  •-LO  S  -  _f131 m V-Forest  108 m-A  7 5 m - J — ' " V -58  -  40  m-J  1  0 m  \  I  20 m  m  \  BR-A  40 m  Culvert Outlet J [  -  road I  1  1  1  1—  1  1  1  1  1  1  1  1  1  1  1  b k  218 m  LO(224m 201  m^N \*-170 m Multiple Channels  -  \  confluence with B2  ^135 m  / A-Forest  90 m-A  Multiple __Sv*  Channels  5  4  ' V  C U | V E R T  \  r 0 m  20 m  m  Q  ^ A  BR  40 m  Figure 2.7 Maps of study reaches: (a) B5 study reach, with downstream distances of bed temperature monitoring sites, locations of DFO stream temperature loggers, and Forest meteorological site, (b) B3 study reach, showing downstream distances of bed temperature monitoring sites, locations of DFO stream temperature loggers, the Forest meteorological site, and unconstrained sections of multiple channel. Both streams flow in an approximately northerly direction. 21  period. Downstream fining of the bed materials within the study reaches is consistent with longitudinal sorting associated with the observed trends in channel gradients (Fig. 2.8). During installation of subsurface monitoring equipment, bank materials were observed to consist predominantly of silt overlying gravels. At some locations in B5, buried gravel lenses provided preferential pathways for subsurface water flow between the stream channel and the riparian zone. Measurements during summer 2000 suggested that average bankfull (distance between vegetated banks) widths of the two reaches are similar at 1.3-1.4 m despite the fact that the B3 drainage area upstream of the study reach is less than 1/3 the size of B5 (42 vs. 150 ha). This surprising observation appears to be due to the presence of two unconstrained sections of B3, in which multiple low flow channels meander across a total channel width of up to 3 m (Fig. 2.7b). Neither reach can be described as highly constrained since longitudinal topographic gradients along the weakly incised channels were generally greater than lateral gradients across the riparian zones. However, streamflow in B5 was always contained within one channel.  Figure 2.8 L o n g profiles of t h e t w o s t u d y r e a c h e s . N o t e that w h i l e t h e local e l e v a t i o n c h a n g e s a s s o c i a t e d with s t e p - p o o l s e q u e n c e s w e r e i n d e e d m o r e p r o n o u n c e d in B5 t h a n B3, a g r e a t e r effort w a s also m a d e to s u r v e y that s c a l e of variability in B5. H e n c e , t h e plots a r e not c o m p a r a b l e with r e s p e c t to s m a l l - s c a l e variability a s s o c i a t e d with s t e p - p o o l s e q u e n c e s  22  Step-pool morphology (e.g., Fig. 2.9) was the dominant bed configuration within the study reaches, although riffle-pool sequences also occurred (particularly in the lower half of each reach, where channel gradients declined). Steps develop around woody debris, with gravel and cobbles accumulating behind them, and pools form through scouring immediately downstream The elevation of the steps in the B5 streambed was commonly 40-50 cm higher than the deepest areas of downstream pools. Local elevation changes associated with steps were not as pronounced in B3 as in B5, with maximum pool depths of only 10-20 cm in B3.  Figure 2.9  Step-pool sequence at 60 m downstream along the B5 study reach.  23  2 . 2 Methods The study initially focused on the B5 reach, but much of that channel dried up in early August, 2000, so the study was expanded to include the B3 reach.  2.2.1 Creek Temperature Monitoring The two DFO data loggers (BR and L O - section 1.2) in the B5 reach were located approximately 16 and 158 m downstream of the culvert outlet, respectively (Fig. 2.7a). The DFO records consist of hourly sampled temperature data. Ten-minute resolution creek temperature data were collected along with bed temperatures at 40 m and 183 m downstream, and also with the microclimatic data at the Forest site (section 2.2.4.1) at 114 m (Fig. 2.7a). On August 4, four Stowaway® Tidbit temperature data-loggers were distributed along the B5 reach. The Tidbit data loggers were programmed to store measurements every 10 minutes, and were initially placed on the stream bed at four locations along B5: near the DFO logger at 16 m, at 111 m, near the DFO logger at 158 m, and at 209 m. On August 13, the Tidbit loggers initially placed at 111 m and near the DFO logger at 158 m were moved to the B3 reach. These were each placed near the two DFO temperature data loggers at the B3 reach (36 and 224 m downstream of the culvert outlet - Fig. 2.7b). 2.2.2 Modelling Downstream Temperature Change A model for predicting downstream changes in water temperature is derived as follows. The effects of different processes on downstream changes in temperature can be estimated by considering the energy and mass balances of a stream segment of length "Z.". Meteorological conditions are assumed to be relatively constant during the time it takes water to flow through the segment, which should be a reasonable assumption for the short (-200 m) reaches examined in this study. Given this assumption of steady state conditions, the energy budget for the reach can be expressed as: Heat in = Heat out where "heat in" includes advective input from upstream, vertical energy exchange across the water and bed surfaces, advective input by groundwater and hyporheic exchange. These terms are expressed below:  24  Advective input from upstream = C-qus'Tus where C is the heat capacity of water (4.18 x 10 J L" ° C ' ) , qus is streamflow at the 3  1  upstream end of the reach (L s") and Tus is water temperature at the upstream end of the 1  reach (°C).  Inputs across water and bed surfaces =LW -(Qu + Qc) m  where L is the reach length (m), and W is the mean surface width of the stream (m). m  Advective input from groundwater = C-Lqaw'TGw where qcw is groundwater inflow per unit length of stream (L s" rrf') and T G W is 1  groundwater temperature (°C). Hyporheic exchange = C-Lqmp'iTmp-  T) w  where qnyp is hyporheic exchange per unit length of stream (L s" rrf ) which is equal to q"^ 1  and  \q™y \  in Eq. 1.3,  P  T YP H  1  is the temperature of upwelling hyporheic water (°C) and T is w  local water temperature (°C).  Heat out =C-qos'TDS where q s and T D  are the streamflow (L s") and temperature at the downstream end of the 1  D S  reach (°C).  Solution for computing downstream temperature Setting heat in = heat out and solving for T T  D S  yields the following expression:  qusTys + /-[qGwTGw + qHYp(THYP~ Tus) + BQp + BQc] V^' ) 1  1DS -  n  where B = W / C . Eq. 2.1 assumes that longitudinal dispersion is negligible. The four terms m  in the square brackets represent the effects of groundwater, hyporheic exchange, energy exchange across the upper water surface and heat conduction across the streambed,  25  respectively. The effect of each term on downstream temperature (e.g., ATy) can be "partialled out" by comparing the computed T D S with and without the term. Details on the measurements and calculations required to quantify the terms are provided in the following sections.  2.2.3 Stream-Groundwater Interactions Three distinct methods were used to assess groundwater inflow and its effects on stream temperatures. First, measurements of streamflow at intervals of -40 m along the study reaches provided a direct indication of whether the stream was gaining water from the subsurface. Measurements of sediment temperatures in the riparian zone were used to quantify groundwater temperatures (T w in Eq. 2.1). These two pieces of information could G  be combined with stream temperatures to calculate the thermal impacts of groundwater inflow. Second, comparison of longitudinal patterns of creek temperature and electrical conductivity provided insight into the thermal effects of groundwater inflow. Finally, measurements of hydraulic gradients within the streambeds and banks were used to examine spatial variability in the stream-groundwater interactions.  2.2.3.1 Streamflow Measurements Streamflow was measured using constant-rate salt injection. A solution of NaCl was injected into the creek at a constant rate using a Mariotte bottle set up on a surveying tripod, except when streamflow was gauged at the culvert, in which case the bottle was placed on the upstream end of the culvert (Fig. 2.10). Streamwater electrical conductivity (EC) was monitored about 5-10 m downstream, using a WTW™ conductivity meter. Streamflow measurement was limited to sites with appropriate channel characteristics to facilitate thorough, and rapid, mixing of injection solution with the streamwater (e.g., multiple riffles and few, small pools). The DIN 19 266 nonlinear calibration was used by the W T W to correct automatically the electrical conductivity values to a standard temperature of 25 °C.  26  Figure 2.10 Mariotte bottle on upstream end of B 5 culvert at head of study reach. Cobbles were placed at the bed near the entrance of the culvert to facilitate rapid mixing of injection solution into the streamwater during this low-flow period.  27  Streamwater E C was monitored for several minutes prior to injection to ensure a stable background value. The salt solution was injected into the stream until a constant (plateau) value of E C was observed for several minutes and then the injection was stopped. Calibration was performed in the field by repeatedly mixing known volumes of the injection solution into a known volume of creek water, so that the solution gradually became more concentrated. By monitoring the increase in E C during five-to-twelve additions of 1-10 mL of injection solution, streamflow could be calculated by comparing the degree of dilution observed in the stream to that in the calibration: ^  q=  b(EC  (2.2) -EC )  P  B  where: q is streamflow (L s" ), qj is injection rate (L s" ), ECp is plateau electrical 1  1  conductivity (u.S cm" ), E C is background electrical conductivity (u,S cm" ), and b is the 1  1  B  slope of the calibration curve (cm u.S"'). The potential error (5) associated with the streamflow data was calculated as:  5q_  2  =  \9, J  +  '»Y ( S(EC T  P  — EC ) B  +  (2.3)  where: S(EC  P  -EC ) B  = yl(SEC )  2  P  +(SEC )  2  B  (2.4)  A detailed description of the estimation of individual errors associated with each of these terms is included in Appendix 1. Total uncertainty associated with the streamflow measurements averaged 4%, with a range of ±2-12%. Streamflow at the B5 flume (Fig. 2.3) was also estimated on each day that a visit was made to the field site, by measuring the depth of water adjacent to the stilling well of the 22.9 cm Parshall flume and applying the flume's rating equation:  q=[3.07(S-3.281) -0.0283]-1000  (2.5)  153  (Beaudry 1998)  where S is stage height adjacent to the stilling well (m). A n error of ±2 mm was assumed in measuring stage height.  28  2.2.3.2 Riparian Zone Sediment Temperatures Temperatures of the sediments within the riparian zone were monitored with copperconstantan thermocouples buried to various depths in the banks and beds of the creeks. After installing the wells (section 2.2.3.4), holes were hand-augered into the banks and the thermocouple sensors were placed at specific depths (30-100 cm) below the ground surface. Augered holes were back-filled with kaolinite clay and native sediments. No sensors were installed above the water table at the time of installation, and measurements of water levels in the wells allowed the sensor depth to be compared to the depth of the water table throughout the study period. Each thermocouple wire was cut to length so that the indicator end emerged ~10-20 cm above the ground surface. Thermocouple indicators were fitted with "mini-plugs" which allowed them to be monitored periodically using a digital thermometer (Omega™ HH-25TC), with a resolution of 0.1 °C. Temperatures within the creek banks were monitored along transects which ran roughly perpendicular to the stream. At B5, the three transects (163 m, 192 m, and 208 m downstream) consisted of two augered holes on each side of the creek: one ~1 m from the creek and one ~3-4 m from the creek (Fig. 2.11). Generally, two thermocouples were installed in each auger hole, at approximate depths of 0.5 m and 1 m. The soil temperature transects were similar at B3, except that thermocouples were only installed at one augered hole in each of the creek banks (~ 1 m from the stream). In both reaches, sampling depths were influenced by soil conditions. Where the banks consisted predominantly of silt, the wells and thermocouples could be readily installed to the desired depth. However, augering holes into gravel and cobbles proved difficult, hence installation depths were often shallower than desired where those deposits were encountered. Creek bed temperatures were monitored with thermocouples fixed to narrow (1 cm ) z  wooden stakes with duct tape. Three thermocouples were attached to each stake, and the stakes were inserted into the bed sediments using a small shovel so that bed temperatures were measured at depths of 5, 10, and 20 cm below the streambed surface (Fig. 2.11). Bed temperature gradients at B5 were sampled using three approaches. Temporal patterns of bed and stream temperatures were monitored continuously at two sites (40 and 183 m) in B5 using Campbell Scientific CRI OX electronic data-loggers equipped with reference thermistors. The temperatures of all four thermocouples were measured every 10 s and  29  averaged over ten-minute periods. The data-logger used to collect the 40 m data was disconnected and moved to 199 m on August 7 since it was felt that the bed temperatures at 40 m did not need to be monitored continuously beyond that date.  1m  2 m  unsaturated zone saturated zone creek  well or bed piezometer x thermocouple sensor  F i g u r e 2.11 Generalized cross-sectional schematic of the subsurface thermal and hydrologic monitoring transects used in the B5 study reach. None of the transects contained all of these components.  Spatial patterns of bed temperature along the B5 reach were examined with measurements at six additional sites (Fig. 2.7a). At each site, one thermocouple stake was placed in the upstream step or riffle, while a second thermocouple stake was placed in the downstream pool. Spatial patterns of bed temperature were assessed in a more detailed manner at some sites by including a greater number of thermocouple stakes (e.g., spread laterally across the channel). These bed temperature data were recorded using the digital thermometer, up to four times per day. Bed temperature data at B3 were collected using seven thermocouple stakes spread along the reach (Fig. 2.7b). A l l but one of these thermocouples were placed in riffles, in order to examine the influence of hyporheic flow through riffles. Bed temperatures were never measured more often than once per day at B3, and the thermocouples were not installed until August 15.  30  2.2.3.3 Longitudinal Electrical Conductivity and Temperature Sampling Electrical conductivity is an aggregate measure of dissolved solids, so that changes in EC along a stream indicate changes in water chemistry caused by groundwater inflow, where groundwater and streamwater differ chemically. Longitudinal patterns of stream EC and temperature were measured to determine whether downstream changes in temperature corresponded with changes in E C due to groundwater inflow. Where creek temperature is controlled mainly by groundwater inflow, downstream changes in EC and temperature should be well correlated since stream and ground waters seldom have the same chemical signature. Streamwater EC and temperature were measured with the WTW™ conductivity probe and meter. Measurements of temperature and E C were made about every 20 m along the two reaches, beginning at the culvert outlets. Each set of measurements required about 30 minutes to complete.  2.2.3.4 Stream-subsurface Hydraulic Gradients Vertical hydraulic gradients within the streambeds were measured with single piezometers driven to depths of 20-30 cm. The -60 cm long piezometers were constructed from 1.5 cm internal diameter P V C pipe with a 10 cm slot zone (four 5 mm diameter holes were drilled in the bottom 10 cm of the pipes). The slot zone was screened with nylon mesh and the piezometers were installed in the bed using a small shovel. Water levels in the piezometers were detected by observing a change in electrical resistance when water was contacted by a circuit attached to a graduated rod. Measurements of hydraulic head within the piezometers were reproducible to within <±5 mm. The graduated rod was also used to measure stream hydraulic head relative to the top of each piezometer. Vertical hydraulic gradients within the streambed (BHG) were then calculated as:  B H G = (D -Di)/D 0  (2.6)  d  where D is the measured distance from the top of the piezometer to the stream surface (cm), 0  Dj is the measured distance from the top of the piezometer to the water in the piezometer (cm), and Dd is the distance from the bed surface to the middle of the slot zone. Positive  31  values of B H G from Eq. 2.6 indicate upwelling hyporheic flow or groundwater, while negative values indicate downwelling. Hydraulic head distributions within the stream banks were measured using transects of wells positioned perpendicular to the creeks. Wells consisted of 1.5 cm internal diameter P V C pipe with 5 mm diameter holes drilled in rows at 10 cm intervals along the entire length, with holes spaced by 90° separation around the pipe. The wells were screened with nylon mesh and inserted into holes drilled by hand auger. The auger holes were back-filled around the wells using gravel and native sediments. Lateral hydraulic gradients between the creeks and their riparian zones were then evaluated by comparing the hydraulic heads measured within the banks to water levels in the stream. Streamwater levels were measured either on the bed piezometers or on wooden staff gauges driven into the streambed. Horizontal positions and vertical elevations of bed piezometers, wells and staff gauges were surveyed using a transit level. At the two creek reaches, a total of 13 bed piezometers, wells, and staff gauges were broken or removed by bears during the study period. Such equipment was only replaced i f the item was damaged early in the study period.  2.2.4 Meteorological Measurements & Calculations 2.2.4.1 Instrumentation Weather conditions were monitored continuously at the Open site in the B5 cutblock (location in Fig. 2.3) from 09:00 on July 16 to 14:00 on September 5. Air temperature was measured at a height of 1.6 m, using a temperature and relative humidity probe (Campbell Scientific model HMP45C), equipped with a standard radiation shield. Incident solar radiation was measured using a pyranometer. A tipping-bucket rain gauge was used to monitor precipitation (tipping depth = 0.32 mm). A l l data were recorded on an electronic data logger (Campbell Scientific CRI OX) once every 10 s and averaged (or summed in the case of the rain gauge) every 10 minutes. Measurements of the creek microclimates were made at one heavily shaded site (hereafter Forest) along the B5 and B3 study reaches. The instruments were initially set up at the B5 Forest site (Fig. 2.7a), and were then moved to the B3 Forest site (Fig. 2.7b) on  32  August 15, 2000. Air temperature and vapour pressure were measured 60 cm above the water surface, using a Vaisala temperature and relative humidity probe. A Met One 3 cup anemometer was used to measure wind speed at a height of 80 cm. Creek temperature was, monitored at the Forest sites using a copper-constantan thermocouple. A Middleton net radiometer was used to monitor net radiation approximately 10 cm above the creek surface. A l l data were recorded on an electronic data logger (Campbell Scientific CRI OX) once every 10 s and averaged every 10 minutes.  2.2.4.2 Calculations of Energy Fluxes across the Water Surface A l l energy flux densities, and gradients driving these fluxes, are labelled positive i f they are a source of energy to the creek water (and vice-versa). Note that this contrasts with the sign convention commonly used in micrometeorology, where non-radiative fluxes directed away from the creek are positive when focusing on energy balance closure (e.g., Oke 1987).  Radiative Fluxes For purposes including the estimation of missing data, it was necessary to develop a model of net radiation over the stream. Modelling was facilitated by separating Q* into its short- and long-wave components as follows:  Q* = K i - K t + L i - L t = K * + L *  (2.7)  where K i and K t are the incoming and reflected shortwave radiative fluxes over the stream (W rrf ), L i and L t are the incoming and outgoing longwave fluxes (W m" ), and K * and L * 2  2  are the net short- and long-wave exchanges (W m" ). The net shortwave and longwave 2  exchanges were modelled using the following expressions:  K*= f K i  (2.8)  L * = L i - L t = . -a(T . ) - s -a(T ) 4  ea  v  a  v  (2.9)  4  w  w  where (j) is a factor accounting for transmission of shortwave radiation through the canopy, K i is incoming solar radiation measured at the Open site (W m" ), £ . is the effective 2  a  33  v  emissivity of the atmosphere and vegetation overlying the creek surface, a is the StefanBoltzmann constant (5.67 x 10" W m" K" ), T . is the temperature of the atmosphere8  2  4  a  v  vegetation (K), s is the emissivity of the creek water, and T is the creek water temperature w  w  (K). The first step of the modelling exercise was simulating nighttime Q* when Q*=L*. Black et al. ( 1 9 9 1 ) showed that L l beneath a Douglas-fir canopy on Vancouver Island could be accurately approximated by assuming that T . was equal to air temperature. Therefore, a  v  the only parameters to be varied in Eq. 2.9 were s - and e , with observed ranges for' a  v  w  vegetation and water of 0 . 9 0 - 0 . 9 9 and 0 . 9 2 - 0 . 9 7 , respectively (Oke 1987). The emissivity parameters were fitted by comparing scatter plots of model results versus observations to a line of perfect fit, and time series of the modelled and observed values. These analyses showed that s . and e had to be nearly equal in order to achieve a good fit to the observed a  v  c  values. The model results were less sensitive to variations in the magnitude of the two emissivities, although values of - 0 . 9 5 produced the best results. Negative L * values modelled with higher emissivities were excessively negative in comparison to observed nighttime Q* values. Consequently, the L * model was calibrated by holding 8 constant at W  0.950  and varying the 8 . value slightly between the two locations of the Forest site to a  v  account for differences in canopy characteristics. After optimizing the 8 . values for the two sites, values of <j) were selected for the two a  v  Forest sites by comparing scatter plots of modelled daytime Q* values (sum of Eqs. 2.8 and 2.9) versus observations to a line of perfect fit. A 3 0 minute averaging period was used to minimize the effects of clouds passing over the Open and Forest sites at different times (Rowland and Moore 1992), while maintaining relatively fine temporal resolution.  Convective Fluxes The latent heat flux (Q ) and sensible heat flux (Q ) were estimated using massE  H  transfer and Bowen ratio approaches. Two sets of equations, applied by Brown ( 1 9 6 9 ) and Webb and Zhang ( 1 9 9 7 ) , respectively, were initially evaluated for use in this study. Brown's (1969)  formulations were chosen because they gave estimates of Q E and Q that were H  generally half as large as those of Webb and Zhang. This decision was based upon two considerations. First, the anemometer frequently overestimated windspeed in the Forest due  34  to its relatively high stall speed (~ 0.45 m s" ). The windspeed time series consequently contain long stretches of 0.45 m s" values when the actual windspeed was probably lower. 1  Moreover, these 0.45 m s" values were included in many ten-minute averaged values, so that 1  even when the anemometer was often turning, the ten-minute average wind speed values were probably overestimates. Second, as will be discussed in sections 3.3.7 and 3.4.6, stable conditions above the creeks often inhibited free convection. Therefore, conservative estimates of the convective fluxes were desired. The latent heat flux, Q , was calculated as: E  Q = 76.6-U(e -e ) E  a  (2.10)  w  where U is the wind speed (m s" ), e is the vapour pressure of the atmosphere over the 1  a  stream (kPa) and e is the saturation vapour pressure at the creek water temperature (kPa), w  computed as: e • exp 0  1  1  5R.  (2.11)  w J  after Stull (1995), where e = 0.611 kPa and T = 273 K are constants, U =2.5 x 10 J kg" is 6  0  1  0  the latent heat of vaporization over a liquid water surface, and 5R = 461 J K" kg" is the gas 1  1  V  constant for water vapour. The sensible heat flux, QH, was then computed as: QH = PQE  (2.12)  where [3 is the Bowen Ratio, calculated as: (3 = 0.61-(P/1000)-[(T -T )/(e -e )] w  a  w  (2.13)  a  Air pressure (P) was assumed constant at 90 kPa, given that monthly average P at the Prince George Airport is constant at 93 kPa (Environment Canada 2002), and assuming that P decreases by 1 kPa per 100 m increase in altitude (Ahrens 1994). The location of the city of Prince George is shown in Fig. 2.1. The elevation of the Prince George airport (676 m a.s.l.) is approximately 300 m lower than the Baptiste Creek study site.  35  2.2.5 Bed Heat Conduction The streambed conductive heat flux was calculated as:  Qc = K ( d T / d Z )  (2.14)  c  where K c is the thermal conductivity of the streambed sediments (W m" K" ) and dT/dZ is 1  1  the temperature gradient with depth in the streambed calculated using the difference between the water temperature and the bed temperature at a depth of 5 cm. Lapham (1989) presented a graphical summary of the positive relation between the thermal conductivity of saturated sediments and their dry bulk densities. Since dry bulk density depends primarily on the porosity of mineral sediments (Freeze and Cherry 1979), a range of K values was calculated based on a range of porosities for the sediments in the c  study reaches (a particle mass density of 2.65 g cm" was assumed - Freeze and Cherry 3  1979). A reasonable range of porosities for alluvial gravel is 0.25-0.40 (Freeze and Cherry 1979; Dingman 1994). This range of porosities corresponds to a range of Kc values of 2.82.1 W m" K" , according to Lapham's (1989) relation. 1  1  Since Evans et al. (1998) found that spatial variations in calculated thermal conductivity exerted a minor influence on Qc compared to spatial variations in bed temperature gradients, no)effort was made to estimate spatial variations in streambed thermal conductivity along the study reaches. Instead, a Kc value of 2.6 W m" K* was used, based 1  1  on a porosity of 0.30 typical of relatively coarse gravel. This thermal conductivity value is relatively high compared to those used in other similar studies, although several studies did not report the K values that they used in Eq. 2.14 (e.g., Webb and Zhang 1997, 1999), while c  others were not dimensionally homogeneous (e.g., Brown 1969). Silliman et al. (1995) estimated a Kc value of 1.0 W m" K" for a heterogeneous creek bed which included 1  1  substantial organic matter (Silliman and Booth 1993), while the highest value was 2.3 W m" K" used by Ronan et al. (1998) for the fine sand at their study site. Most K c 1  1  values have been estimated with similar methods to those used here, but Land and Paull (2001) measured the thermal conductivity of estuarine sediments in situ. The composition of their sediments at 19 sites ranged from fine-grained mud to sand and K c values ranged from 0.7 to 2.0 W m" K" . The range of K values they observed provide some support for the use 1  1  c  36  of Lapham's curve, since their highest value corresponds to that predicted for an "average sand" with porosity 0.4 (Dingman 1994), while the lowest values are similar to the 0.8 W m"  1  K" predicted for a fine-grained sediment with a porosity of ~0.6. 1  2.2.6 Hyporheic Exchange At B5, the areal extent of the hyporheic zone was evaluated by examining temperature patterns within the riparian zone. The temperature of deep groundwater is relatively constant in time at a value similar to, or slightly warmer (1 °C higher) than, mean annual air temperature (Domenico and Schwartz 1998), suggesting a temperature of 3-4 °C at the Baptiste Creek tributaries. Because the streamwater was generally warmer (7-14 °C) than this groundwater temperature during the study period, the dimensions of the hyporheic zone could be estimated by considering the degree to which riparian zone temperatures were controlled by stream temperatures. The hyporheic zone is commonly identified as that portion of the subsurface containing > 10% streamwater (e.g., Harvey and Fuller 1998; Hill and Lymburner 1998), based on concentrations of a conservative tracer. Although the temperature of water changes as it flows through sediments of different temperature, (i.e., temperature is not a "conservative" property of water), and subsurface temperatures are not determined exclusively by advection (conduction may also be important), temperature has previously been used to quantify the extent of the hyporheic zone (e.g., White et al. 1987; Wondzell and Swanson 1996). At B3, an attempt was made to quantify the energy exchange associated with hyporheic flow (QHYP)- First, the physical processes controlling hyporheic exchange were quantified using the transient storage model OTIS-P. Runkel (1998) describes transient storage as "the temporary detainment" of stream water within areas that are moving slowly relative to the "faster moving waters near the center of the channel." These areas consist of two components: stagnant surface water storage zones such as pool margins, and the subsurface hyporheic zone (section 1.3). The areal extent of the transient storage zone, and the rate at which water moved between it and the stream channel, were evaluated by injecting a dilute salt solution into the creek and measuring the change in streamwater EC through time at a downstream point.  37  A coupled set of differential equations describes the changing concentrations of a conservative solute in the main channel and the storage zone: a C __q_d£ dt riC  ^  A dx  +  j__a_  f  + ^(C A  -C) + a(C  L  A dx  -C)  s  A  =a -  (2.15a)  (2.15b)  (C-C ) s  where: A As C CL Cs D q q t x a L  is main channel cross-sectional area (m ) is storage zone cross-sectional area (m ) is main channel solute concentration (uS cm" ) is lateral inflow solute concentration (uS cm" ) is storage zone solute concentration (uS cm" ) is the longitudinal dispersion coefficient (m s" ) is streamflow (m s" ) is lateral inflow rate (m s" m" ) is time (s) is distance (m) is the storage zone exchange coefficient (s" ). 1  1  1  2  3  1  1  3  1  1  1  The OTIS-P model described by Runkel (1998) numerically solves finite difference approximations to Eq. 2.15 with the Crank-Nicolson Method, and uses an adaptive nonlinear least squares method to determine an optimal set of values for the parameters A , As, D, and a. The statistical optimization used by OTIS-P minimizes investigator bias and provides an assessment of parameter uncertainty that was unavailable using the trial-and-error method of parameter optimization in the original OTIS model (e.g., see Stream Solute Workshop 1990). Salt injection tracer tests were carried out at three -20 m sub-reaches of B3 on August 30. A fourth tracer test was conducted at a 17m segment that included the culvert at the head of the reach and a pool at its outlet. This test provided some measure of the amount of transient storage that might be expected to occur in individual pools, although this pool was probably better-mixed than most of those in the natural channel. The methods used for these trials were similar to those used for streamflow measurement using salt injection (section 2.2.3.1), except that downstream EC values were monitored as the plateau value was  38  approached and during the period of time required for EC to return to background levels after the injection was stopped. When significant pools are absent from a stream reach, it is reasonable to assume that the transient storage area parameter (As) estimated by OTIS-P is dominated by the hyporheic zone (e.g., Harvey and Fuller 1998). The large pools along B5 precluded the use of tracer tests to characterize its hyporheic zone, but no large pools were present along the three tracer test sub-reaches of B3. Therefore, the average flux of water through the hyporheic zones of the B3 sub-reaches (L s" m" ) was calculated as: 1  1  q YP = a-A  (2.16)  H  after Harvey and Wagner (2000). This estimate of the hyporheic flux was used to calculate the potential thermal effects of hyporheic exchange, by assuming that qnyp is cooled to the average temperature of the hyporheic zone. The average temperature of the hyporheic zone (THYP) was calculated by combining the bed temperature data with an estimate of the average depth of the hyporheic zone. Harvey and Wagner (2000) suggested the average depth of the hyporheic zone can be approximated as: d HYP  ~  (2.17)  W-n  where W is stream width and n is streambed porosity. Harvey and Wagner (2000) suggest that Eq. 2.17 is valid only when the storage-zone cross-sectional area is considerably smaller than the cross-sectional area of the stream (As<A), and when the stream is at least twenty times as wide as it is deep. Both of these conditions were met in the B3 tracer test subreaches. Because bed temperature measurements were not made in all three sub-reaches, a reach-averaged bed temperature gradient (dT/dz as in Eq. 2.14) was used to calculate THYP values for each of the three sub-reaches:  T, HYP  T + 0.5 • d w  HYP  • dTjdz  (2.18)  39  It should be emphasized that the  T YP H  value calculated for each sub-reach using Eq. 2.18  only varied according to differences in estimated d y p values, because of the use of a single H  reach-averaged bed temperature gradient (i.e., deeper hyporheic flow was assumed to be cooler than shallower flow). Finally, the effect of the hyporheic flux on stream temperature [(dT/dx) Yp] was H  calculated using a simple model:  d J HYP x  —  (THTF-TJ  (2-19)  where (dT/dx)nYP is expressed in °C m" , and T is the temperature of streamwater. Eq. 2.19 1  w  was applied separately to each of the three sub-reaches, so that the longitudinal variability of the estimated hyporheic effect could be assessed. The cumulative reach-scale thermal impact of the hyporheic term ( A T Y P ) was calculated by averaging the results for the three subH  reaches, and multiplying by 200 m.  40  CHAPTER 3 RESULTS This chapter presents results of the field campaign during summer 2000. The initial intent was to study stream B5 in detail, since a pilot study in 1999 revealed intriguing behaviour suggestive of cooling driven by hyporheic exchange (Moore et al. in press). However, as will be described below, stream B5 dried up part way through the study period, so research was expanded to include stream B3. The chapter begins with an overview of the study period and discussion of the quality of stream temperature data (sections 3.1 and 3.2). Results are presented separately for streams B5 and B3 in sections 3.3 and 3.4, respectively. The chapter ends with a summary of the key findings.  3.1 Overview of the Study Period Beaudry (2001) presented snowpack and snowmelt hydrograph data from the Baptiste Creek watershed for the years 1996-2000 (Table 3.1). Five or six snow surveys were conducted each year, at intervals of 1-4 weeks. Peak snowpack water equivalent occurred in March and averaged 340 mm during the five year period (Table 3.1). The second-smallest snowpack was observed in 2000, with a peak value of 300 mm. Maximum instantaneous (15-minute resolution) streamflow values occurred in May during all five years, in association with peak snowmelt inputs. As a result of the relatively small snowpack and a prolonged melt period, peak snowmelt streamflows in 2000 were the lowest observed during the five years (Table 3.1). On May 17, 2000, streamflow reached 23 L s" at the B3 flume 1  and 150 L s" at the B5 flume (locations of flumes shown in Fig. 2.3). Specific peak 1  streamflows at B5 were roughly twice those at B3 in all years (Table 3.1). During the study period, July mean air temperature was close to the ten year average recorded at the Middle River DFO camp, while the August mean was 0.9 °C lower than average (Table 3.2). Total precipitation for the months of June-August was 8% below average, due to July being the second driest of the ten year record (Table 3.2). The Middle River camp is 25 km northwest of the Baptiste study site, and is -300 m lower in elevation (Fig. 2.2). There is poor correspondence between daily precipitation recorded at the study site and the Middle River camp (Fig. 3.1a,b), but for longer periods there appears to be some  41  correlation between the two sites (Table 3.3). The greater precipitation at the study site is consistent with orographic processes.  Table 3.1  P e a k s n o w p a c k w a t e r e q u i v a l e n t a n d s n o w m e l t s t r e a m f l o w s o b s e r v e d in t h e B a p t i s t e w a t e r s h e d . D a t a a r e f r o m f i g u r e s in B e a u d r y ( 2 0 0 1 ) . N o t e that 1 9 9 6 c o n s t i t u t e d t h e p r e - h a r v e s t p e r i o d . Significant l e a k a g e o c c u r r e d at t h e B 5 f l u m e in 1 9 9 7 ( B e a u d r y 1998). B3 (0.42 k m ) Peak Specific Peak 2  Peak Snow W a t e r  B5 ( 1 . 5 k m ) Peak 2  Streamflow  Streamflow  Specific Peak Streamflow  Equivalent  Streamflow  (mm)  (L's )  420 370  25  59  190  127  1997  50  118  -350  233  1998  275  30  71  200  133  1999  340  32  75  200  133  2000 mean  300 341  23 32  54 75  150 185  100 145  1996  (L s" k m " ) 1  1  2  (L s" k m " )  (L s )  1  1  2  Table 3.2 S u m m a r i e s o f air t e m p e r a t u r e a n d precipitation r e c o r d s f r o m t h e Middle River c a m p for t h e s p r i n g - s u m m e r period (station is not active in late fall or w i n t e r ) . V a l u e s f r o m t h e s u m m e r of 2 0 0 0 (including t h e s t u d y period) a r e in bold font. D a t a p r o v i d e d by D F O . - Monthly Precipitation ( m m )  Monthly Mean Air T (°C) 1991  J-J-A S u m  June  July  August  (mm)  14.4  20  33  49  102  9  14  39  July  August  13.5  May  1992  15.0  14.1  25  16  1993  14.3  13.5  65  99  76  56  231  1994  15.7  15.3  27  64  29  71  164  1995  14.6  11.1  17  33  58  59  151  1996  13.1  12.5  33  49  65  49  163  1997  14.3  14.7  74  50  24  149  1998  16.4  14.2  46  36  45  127  1999  13.6  14.0  49  36  42  126  2000  14.4  12.9  59  22  45  127  14.5  13.8  51  42  45  138  10 y e a r average  25  6 year average  42  25 ^ 20 E ^  15  c o  10  3 g .  o  8> o_  5 0  —i—i—i—T=i—i—|—i—i—t - !—i 1  18-Jul  1  26-Jul  i i - ]—i—i—i—i—i—I—i—|—i—i—i—I—I  r T '|' 'i i i i i TT  3-Aug  19-Aug  1  1  11-Aug  pi 1 -  n—i  27-Aug  i i | i  4-Sep  25 20 E £ c o  I Q. O  £  15 10 5  CL  0  J  -|—I—I—I—m—I—i  |  i  I  •^—Segment 1  I—I—I  —  • I  i—|—1  I—I  1  I  I  . I  |  I  I  .1 I  I  I  I  I  |  •ll.li. I, I  1 r  •Segment 3-  — S e g m e n t 2=  30  18-Jul  26-Jul  3-Aug  11-Aug  19-Aug  27-Aug  4-Sep  Date (year 2000) Figure 3.1 (a)  Daily p r e c i p i t a t i o n m e a s u r e d at t h e M i d d l e R i v e r D F O c a m p , (b) Daily p r e c i p i t a t i o n m e a s u r e d at t h e B a p t i s t e C r e e k O p e n site, (c) Daily t e m p e r a t u r e e x t r e m e s d u r i n g t h e study period m e a s u r e d at t h e O p e n site.  43  T a b l e 3.3 Climatic c h a r a c t e r i s t i c s of t h e t h r e e s e g m e n t s of t h e s t u d y p e r i o d , a n d c o m p a r i s o n of precipitation m e a s u r e d at t h e O p e n site to that m e a s u r e d at t h e M i d d l e River C a m p .  Baptiste C. Open site Air Temp. (°C) P (mm) Segment 1 (July 17-Aug. 1) Segment 2 (Aug. 2-17) Segment 3 (Aug. 18- Sept. 4) Study Period Totals  14.1 12.7 8.5  50.3 9.4 98.8 158  Middle River Baptiste PI P (mm) Middle R. P 17.4 2.8 61.2 81  2.9 3.3 1.6 1.9  The study period can be broken into three ~16 day segments.; each characterized by distinctive weather (Figs. 3.1, 3.2, Table 3.3). The first segment was warm, with occasional rain events (Table 3.3, Fig. 3.2a). The second segment was dry (Table 3.3, Fig. 3.2b) and sometimes hot (Fig. 3.1). The third segment of 18 days was dominated by cool (Fig. 3.1, Table 3.3), cloudy, and wet conditions (Fig. 3.2c), including a hailstorm on August 18 and a snowfall on September 1. Diurnal air temperature fluctuations at the Open site were often greater than at the Forest site (Fig. 3.3). Daily air temperature minima were similar in the two settings (Fig. 3.3), although it should be noted that comparisons before and after the field season showed that the probe used at the Forest site consistently overestimated air temperature by 0.6-0.8 °C relative to the probe used at the Open site. Daily air temperature maxima were up to 8 °C higher at Open than at Forest (Fig. 3.3).  44  1200  ~i—i  0  r  2 800 6  400 H  0 1 17-Jul  *~j  1  |  p-T  19-Jul  I  'j '  I  '| '  I  LU '|  1  8 1  I  |  I  |  I  |  I  |  I  |  I  |  I  |  I  [  10  21-Jul  23-Jul  25-Jul  27-Jul  29-Jul  31-Jul  2-Aug  6-Aug  8-Aug  10-Aug  12-Aug  14-Aug  16-Aug  18-Aug  1-Sep  3-Sep  E 1200 c o .2 T3  800  CO  rr  CO  o 400 -I co CO  c E  8 -  I  2-Aug  I  I  I  I  4-Aug  1200  800  400  1  '| '  1  'l'  1  '| '  1  'l '  1  '| '  1  'l'  1  |  1  l  18-Aug 20-Aug 22-Aug 24-Aug 26-Aug Figure 3.2  ' '| '  1  'I '  28-Aug  1  '| '  1  'l  30-Aug  Incident solar r a d i a t i o n a n d p r e c i p i t a t i o n m e a s u r e d at t h e O p e n site d u r i n g t h r e e s e g m e n t s (a,b,c) of t h e study p e r i o d . T h e study p e r i o d w a s s u b - d i v i d e d m a i n l y o n t h e basis of v a r i a t i o n s in t h e f r e q u e n c y o f p r e c i p i t a t i o n e v e n t s . D a t a f r o m t h e final t w o d a y s o f t h e study p e r i o d a r e not plotted (a total o f 2.5 m m o f p r e c i p i t a t i o n w a s r e c o r d e d o n S e p t e m b e r 3-4, a n d solar r a d i a t i o n p a t t e r n s w e r e similar to other segment 3 days).  45  30  -i  20 H i  i t ! ' 1*  10  0 17-Jul  18-Aug  1  1  20-Aug  i  |  21-Jul  19-Jul  '  1  22-Aug  i  '  1  24-Aug  T  |  23-Jul  27-Jul  25-Jul  '  1  26-Aug  '  |  28-Aug  29-Jul  1  I  30-Aug  !  1-Sep  *  "1  31-Jul  1  \  VF  2-Aug  1  Figure 3.3 A i r t e m p e r a t u r e at t h e O p e n a n d Forest m e t e o r o l o g i c a l sites d u r i n g t h e study p e r i o d . T e m p o r a l s u b - d i v i s i o n s (a,b,c) w e r e m a d e o n t h e basis of d a t a p r e s e n t e d in Fig. 3.2. T h e Forest d a t a - l o g g e r lost p o w e r f r o m July 2 9 t o A u g u s t 4 .  46  1  3-Sep  3.2 Stream Temperature Data Quality The quality of the various creek temperature measurements was assessed by comparing temperatures from the electrical conductivity probe to values recorded by nearby data-loggers at the time of the probe measurement. A linear regression was performed on all the available comparison data except those from B5 Forest, which suggested a systematic error of-0.19 °C on the part of the temperature probe (Fig. 3.4). No attempt was made to include the DFO data-loggers in the initial quality control exercise since the timing of conductivity probe samples seldom coincided with the one-hour sampling interval of the DFO loggers. However, since there was no obvious systematic error associated with the Tidbit loggers (Fig. 3.4), the DFO loggers were compared to re-sampled ten-minute resolution data from nearby Tidbit loggers. Results were satisfactory, with the least-squares linear regression between all three possible comparisons (at B5BR, B3BR, and B3LO) suggesting a possible systematic error of only +0.04 °C. The data are relatively uniformly scattered around the regression line on Fig. 3.4 and most (89%) of them fall within 0.19 °C of the expected values (bounded by the 1:1 line and Y = X-0.38 in Fig. 3.4). The data that fall outside of these bounds are from three sites: 117 m, 183 m and 209 m downstream along B5. Groundwater discharge was relatively high downstream of 150 m in B5, suggesting that the latter two data-loggers may have been influenced by cool groundwater which was not always detected by the conductivity probe. The B5 Forest site (117 m) sometimes underestimated creek temperature by up to ~2 °C (Fig. 3.4). This underestimation occurred because the B5 Forest site initially used the data-logger internal temperature as the thermocouple reference temperature, which is known to be an unreliable method. The erroneously low values occurred between 11:50 and 13:50 PST, suggesting that the data-logger wiring panel warmed up more than the internal temperature due to solar heating during that period. The increasing voltage signal from the thermocouple would then have been attributed to decreasing water temperature, rather than increasing reference temperature. Stream temperature data from the B5 Forest site between 11:50-13:50 PST will not be analysed. Instead, those data were reconstructed by linear interpolation across the two-hour period.  47  Conductivity Probe Temperature (°C) Figure 3.4 Quality control of data-logged creek temperature observations.  48  3.3 Results for B5 Section 3.3.1 identifies distinct temperature patterns based on data from the DFO loggers located at the HI, BR, and L O sites. Sections 3.3.2 to 3.3.6 examine the hydrologic context of the stream and its relation to the temperature patterns identified in 3.3.1. Sections 3.3.7 to 3.3.10 examine the energy exchanges between the stream and its environment and their influence on stream temperature patterns in time and space.  3.3.1 Stream Temperature Patterns Four daily temperature patterns can be identified in Fig. 3.5: 1. On pattern 1 days, temperatures at L O were virtually identical to those at BR: both series plot above the HI series, with significant diurnal fluctuations at all three sites (e.g., July 17). 2. On pattern 2 days, temperatures at L O were similar to those at BR, except that the daily maxima at L O were slightly (<~1.0 °C) lower than those at B R (e.g., July 18). 3. On pattern 3 days, temperature maxima and minima at L O were lower than those at BR, L O temperatures were lower than those at HI, and only slight diurnal fluctuations were measured at L O (e.g., July 23). 4. On pattern 4 days, temperature maxima at L O were similar to those at BR, but L O minima were substantially lower than both B R and HI minima, and wide diurnal fluctuations were apparent at L O (e.g., August 10).  Creek temperature patterns 1, 2 and 3 were all apparent during the first segment of the study period. Patterns 3 and 4 dominated the second segment of the study period. The first week of the third segment of the study period was complex, with most days not falling neatly into one of the categories, although brief periods of all four patterns were apparent (Fig. 3.5). Temperature patterns 1 and 2 were observed from August 25 to the end of the study period.  49  4H 17-Jul  19-Jul  21-Jul  23-Jul  25-Jul  27-Jul  2-Aug  4-Aug  6-Aug  8-Aug  10-Aug 12-Aug  29-Jul  31-Jul  2-Aug  14-Aug  16-Aug  18-Aug  1-Sep  3-Sep  18-Aug 20-Aug 22-Aug 24-Aug 26-Aug 28-Aug 30-Aug Figure 3.5 C r e e k t e m p e r a t u r e s f r o m t h e t h r e e D F O d a t a l o g g e r s in s t r e a m (a,b,c) o f t h e study p e r i o d a r e t h e s a m e a s u s e d in Fig. 3.2.  B5.  Segments  Diurnal fluctuations in temperature at all three sites averaged ~2 °C between July 17 and August 31 (Table 3.4). Maximum daily temperatures at B5LO averaged 0.9 °C lower than at B5BR, but were slightly warmer than at the HI site. Minimum daily temperatures at B5LO were lower and more variable than at either of the upstream sites (Table 3.4). The average time of maximum daily temperature increased slightly in the downstream direction, but the substantial variability in these times suggests that the trend was insignificant (Table 3.4). The timing of the daily maxima was least variable at the B R site, with a standard deviation of 1.0 hours (Table 3.4). The timing of the daily maxima was most variable at B5LO.  Table 3.4  S u m m a r i e s of m e a n daily t e m p e r a t u r e e x t r e m e s at t h r e e sites a l o n g s t r e a m B 5 f r o m J u l y 17 to A u g u s t 3 1 , 2 0 0 0 (prior to t h e o n s e t of cooler t e m p e r a t u r e s in S e p t e m b e r ) . " T " is t e m p e r a t u r e , a n d "t" is t i m e ( P S T ) . W h e n t h e daily m a x i m u m t e m p e r a t u r e w a s r e c o r d e d m o r e t h a n o n c e , t h e a v e r a g e t i m e o f all o c c u r r e n c e s w a s c a l c u l a t e d . C a l c u l a t i o n s o f t h e m e a n a n d s t a n d a r d deviation ("St. Dev.") of t h e t i m e of daily m a x i m u m did not include t i m e s less t h a n 6:00 P S T , s i n c e t h e s e f e w d a y s w o u l d h a v e s k e w e d t h e m e a n . T h e n u m b e r of s u c h e x c l u d e d t i m e s at e a c h site is n o t e d in t h e f u r t h e s t r i g h t - h a n d c o l u m n . M e a n Daily M i n i m u m T  M e a n Daily M a x i m u m T St. D e v . (h)  t<6.0 h  B5HI  8.0  0.8  9.6  1.0  14.5  2.3  0  B5BR  8.7  1.0  10.7  1.1  15.1  1.0  0  B5LO  7.6  1.8  9.8  1.0  15.3  2.9  5  Site  M e a n T (°C) St. D e v . (°C)  M e a n T (°C)  51  St. D e v . (°C) M e a n t (h)  3.3.2 Temporal Variations in Streamflow The three distinct segments of the study period (Table 3.3) are also apparent in streamflow records from the B5 flume (Fig. 3.6). The first segment was characterized by moderate variations in streamflow within the range of -3-10 L s" , in response to occasional 1  rainstorms. The second segment began August 2, and was dominated by streamflow recession except for the small rainfall event on August 9. Streamflow during July 21-25 and August 4-9 decreased linearly with time (Fig. 3.6), at a rate of -0.5 L s" day" . This 1  1  behaviour supports the use of linear interpolation to estimate streamflow during dry periods, at least for the range 1-5 L s" . 1  The 17 mm of rain on August 18 was sufficient to increase streamflow to values similar to those observed at the beginning of the second segment. Streamflow was maintained in a relatively constant range (3-4 L s" ) by frequent small rain events during most 1  of the final segment. The rain and snowfall event of August 31 and September 1 produced the highest flows observed during the study period.  Segment 1  25  _  20  •b E E  15  i= o  16  + H-  '•+-»  Segment 3  Segment 2 •  % 10 o  Q)  ++ + + +  12  8  eg 05 CD  CO  CL  ,++  + - H -  0 18-Jul  26-Jul  3-Aug  11-Aug  19-Aug  27-Aug . 4-Sep  Figure 3.6 Daily precipitation measured at the Open site (bars) and instantaneous estimates of streamflow at the B5 flume (crosses). Grey lines represent streamflow recession segments. 52  3.3.3 Spatial Variations in Streamflow The B5 flume was located 80 m upstream of the B5 culvert but streamflow estimates at the flume compare reasonably well to salt-dilution measurements of streamflow at the culvert (Fig. 3.7). The tendency to lower flows at the culvert could reflect the loss of some flow to seepage through the road bed, or overestimation of streamflow at the flume. Streamflow generally decreased through the study reach in the early part of the study period (Fig. 3.8). Dry sections of channel were first observed on July 23, upstream of the L O logger, when 4.5 L s" of streamflow was estimated at the B5 flume (Fig. 3.9). 1  0  50 100 150 Distance Downstream of B5 Culvert (m)  200  Figure 3.8 S t r e a m f l o w a l o n g B 5 , m e a s u r e d u s i n g c o n s t a n t - r a t e salt injection. L o c a t i o n s of t h e B R a n d L O sites a r e s h o w n o n t h e x - a x i s . Inferred pattern f o r J u l y 19 w a s c a l c u l a t e d by e x t r a p o l a t i n g m e a s u r e d s t r e a m f l o w l o s s e s to 155 m (the a p p a r e n t transition f r o m losing to g a i n i n g hydraulic g r a d i e n t s b a s e d o n Figs. 3.10, 3 . 1 1 , 3.13). 53  The maximum extent of dry channel was observed on August 16 when streamflow at the B5 culvert was measured at 0.7 L s" (Fig. 3.9). 1  30-Aug  3.7 3.9  28-Aug  4.2  26-Aug 24-Aug  2.6 2.9 3.7 2.6 3.2  22-Aug 20-Aug  14-Aug  1.5 1.1 1.0 (0.7) 1.2 1.2 1.5  12-Aug  1.6 (1.2)  18-Aug 16-Aug  1.7  10-Aug  1.3(1.4) 1.6 (1.4) 2.1 2.6  8-Aug 6-Aug  3.4  4-Aug  5.8 7.8 4.2 4.2  2-Aug 31-Jul  6.0 7.6 7.0 3.2 3.3 3.7 4.5 4.8  29-Jul 27-Jul 25-Jul 23-Jul  5.1  21-Jul •  5.0 (5.0)  19-Jul  5.0 (7.6)  17-Jul BR 40mBedT | 2 0 m Tb |  81 m Tb |  110 m Tb | Forest (117 m)  LO f  50 100 150 Distance Downstream of B5 Culvert (m)  183 m BedT 209 m Tb f  f  200  Figure 3.9 Spatial e x t e n t of s t r e a m f l o w in B5 f r o m J u l y 17 t o A u g u s t 3 0 . Horizontal lines r e p r e s e n t s e c t i o n s with c o n t i n u o u s flow, w h i l e d a s h e d lines r e p r e s e n t d i s c o n t i n u o u s f l o w (i.e., isolated pools). S t r e a m f l o w e s t i m a t e s at t h e B5 f l u m e a r e r e p o r t e d at t h e d o w n s t r e a m e n d s of t h e lines ( s t r e a m f l o w m e a s u r e m e n t s at t h e culvert a r e in b r a c k e t s ) . T h e locations of all c r e e k t e m p e r a t u r e d a t a - l o g g e r sites r e p o r t e d in t h e text a r e also s h o w n . T b is a Tidbit data-logger, B e d T is a b e d - t e m p e r a t u r e d a t a - l o g g e r .  54  3.3.4 Stream-Subsurface Interactions Dominantly negative bed hydraulic gradients upstream of 155 m (Fig. 3.10) indicate infiltration into the streambed, supporting the measured streamflow losses (Fig. 3.8). However, downstream of 155 m, positive gradients were as common as negative gradients (Fig. 3.10). Except for the step/pool sequence at 59 m, gradients in steps or riffles were more negative than those in pools (Fig. 3.10). A l l lateral hydraulic gradients upstream of 155 m were directed outwards from the stream, whereas groundwater levels were higher than the stream level in at least one well at each transect downstream of 155 m (Fig. 3.11). Water movement in the sub-reach upstream of 155 m (hereafter "upper B5") was dominantly from channel to subsurface, while there appeared to be areas of both up- and down-welling water in the sub-reach downstream of 155 m (hereafter "lower B5"). Groundwater inflow was not observed in upper B5, and the potential for local re-emergence of water lost to the subsurface appears to have been low, given the widespread negative hydraulic gradients. In lower B5, both groundwater inflow and hyporheic exchange may have been important processes.  0.2 _  I  ¥  0  JCDL  -0-  ©  -0.2  T  ©  -0.4  m  TJ  ro -0.6 CD £ -0.8 CO  -5 >>  -1  X  -o -1.2 m -1.4 -1.6  •  Pools Steps/Riffles  O  Neither  50 100 ~ 150 Distance Downstream of B5 Culvert (m) 23 6  200  F i g u r e 3.10 B e d hydraulic g r a d i e n t s m e a s u r e d a l o n g B5. Vertical r a n g e s r e p r e s e n t m a x i m u m a n d m i n i m u m v a l u e s o b s e r v e d , w h i l e s y m b o l s indicate s t u d y period m e a n s .  55  59 m  -4  -2 0 2 Distance From Stream Stage Measurement (m)  4  -8  -4 0 4 Distance From Stream Stage Measurement (m)  8  F i g u r e 3.11 Hydraulic h e a d at c r o s s - s t r e a m t r a n s e c t s in (a) u p p e r B 5 a n d ( b ) lower B5, m e a s u r e d J u l y 27, 2 0 0 0 . Black s y m b o l s r e p r e s e n t s t r e a m s t a g e levels, g r e y s y m b o l s r e p r e s e n t well levels, a n d o p e n s y m b o l s r e p r e s e n t t h e b o t t o m s of dry wells. L a b e l s a r e d i s t a n c e s d o w n s t r e a m . Positive x v a l u e s indicate l o c a t i o n s on t h e right b a n k (looking d o w n s t r e a m ) , a n d v i c e - v e r s a . N o t e t h e differing s c a l e s b e t w e e n a) a n d b). T h e elevation d a t u m is t h e c h a n n e l b e d at t h e d o w n s t r e a m e n d of t h e s t u d y r e a c h (209 m).  56  Bed piezometers in upper B5 showed a tendency to increasingly negative hydraulic gradients through the study period, although considerable spatial variability was evident (Fig. 3.12c). The water table in lower B5 declined steadily from the end of July to August 18, with many wells drying out and maximum water level decreases of 30-40 cm recorded in those wells which did not dry out (Fig. 3.12d). However, by early September, groundwater in lower B5 had risen to levels similar to those measured in late July and early August (Fig. 3.12d). Water levels rose so much in the left bank at 162 m that a shallow pool formed, from which surface flow to the creek occurred. This phenomenon was also observed during the earliest portion of the study period, before the wells were installed. The complex, spatially variable stream-subsurface relations evident in lower B5 (e.g., Fig. 3.1 lb) can be better understood by considering the positions of the individual transects within broader-scale groundwater flow patterns in that sub-reach. Although based on a relatively sparse sampling pattern, Fig. 3.13 suggests that the LO temperature logger was situated just downstream of a transition from losing to gaining conditions. Near the upstream limit of Fig. 3.13, flow was directed outward from the creek, while at the downstream end, subsurface water was directed into the creek. Between those two extremes, subsurface water generally flowed down the topographic gradient of the valley and was not obviously linked to the creek. Hydraulic gradients at several transects were influenced by the orientation of the stream to the water table contours (Fig. 3.13). For example, at 192 m, the creek channel was nearly parallel to the water table contours of July 27, and the water table configuration was similar on September 4. Consequently, the hydraulic gradient at that transect was directed across the stream, suggesting that the creek channel may have been gaining subsurface water from the left bank but losing water to the right bank. Thermal patterns in the riparian zone of the 192 m transect support this interpretation, since soil temperatures were ~2 °C warmer in the right bank than in the left bank on days with warm stream temperatures (Fig. 3.13).  57  25 T3  E 20 E  15  c o  '  10  Q. O CU  5  S  0  Q_  -0.8  ~i—II—i—II  n  i  i  i  i  i  i  i  i  i  i  r  l—I—I—i—I—I—I—i—I—i—i—i—i—I—i  -B-  0.8  |  m m 199 m 192 m  i  208  LW2^  208  RW1  i  r  ~i—i—i—i—i—i—r  LW1 LW2  0.6  03 Q CD  o 0.4 CO  c o 'rc 0.2 >  ja-B^g^B—BE  J  0)  LU ~i  18-Jul  i  i  i  i  26-Jul  r  T  3-Aug 11-Aug 19-Aug Date (year 2 0 0 0 )  ~i—i—j—i—i—i—I—I—I—I—|—I  27-Aug  4-Sep  F i g u r e 3.12 (a) Daily precipitation m e a s u r e d at t h e O p e n site, ( b ) C h a n g e s in c r e e k d e p t h relative to d e p t h s o n July 2 5 at 4 6 a n d 5 9 m d o w n s t r e a m in B 5 . (c) B e d hydraulic g r a d i e n t s m e a s u r e d at 4 6 m d o w n s t r e a m a n d 5 9 m d o w n s t r e a m (pool site in Fig. 3.10). (d) W a t e r t a b l e elevations r e c o r d e d at s e l e c t e d wells in lower B 5 d u r i n g t h e s t u d y period. Labels a r e d i s t a n c e s d o w n s t r e a m a n d positions a l o n g transect looking d o w n s t r e a m (e.g., L W 1 is left well closest to s t r e a m ) . T h e d a t u m in d) is t h e c h a n n e l bed at t h e d o w n s t r e a m e n d of t h e s t u d y r e a c h ( 2 0 9 m ) . 58  •  1  1  r^-  I  —5 m—  _  —5 i n -  j u r y 27  September 4  F i g u r e 3.13 C o n t o u r plots of phreatic s u r f a c e t h r o u g h lower B5 on July 2 7 a n d S e p t e m b e r 4 , 2 0 0 0 . S t r e a m f l o w s t o w a r d t h e top of t h e p a g e . I r r e g u l a r l y - s p a c e d well d a t a w e r e c o n v e r t e d to a regularlys p a c e d grid (0.10 m s p a c i n g ) u s i n g kriging (linear s e m i v a r i o g r a m ) in S U R F E R . C o n t o u r interval is 0.10 m, with e l e v a t i o n s r e f e r e n c e d to t h e c h a n n e l b e d at 2 0 9 m d o w n s t r e a m . W e l l s are r e p r e s e n t e d as solid circles, d i a m o n d s r e p r e s e n t s t r e a m s t a g e m e a s u r e m e n t sites, a n d c r o s s e s indicate soil t e m p e r a t u r e t h e r m o c o u p l e s . L a b e l l e d d a s h e d lines s h o w t h e locations of t h e t r a n s e c t s p r e s e n t e d in Fig. 3 . 1 1 . S m a l l bold n u m b e r s are soil t e m p e r a t u r e s o n A u g u s t 1, a n d S e p t e m b e r 4 ( w h e r e c r o s s e s are not s h o w n , t h e r m o c o u p l e s w e r e located within - 1 0 c m of t h e w e l l s ) . Soil t e m p e r a t u r e s w e r e m e a s u r e d at a d e p t h of 1 m, e x c e p t for t h o s e m a r k e d with a n asterisk (*), w h i c h a r e f r o m a d e p t h of 0.7 m.  59  3.3.5 Electrical Conductivity & Temperature Patterns Longitudinal patterns of streamwater EC and temperature along B5 varied between days. On pattern 1 days, electrical conductivity was relatively steady throughout the reach, with only slight declines or increases in the downstream direction (Fig. 3.14a,b). Downstream temperature changes on those days were also relatively modest, with decreases of less than 0.3 °C between the BR and L O sites, and decreases of less than 0.5 °C until 208 m. Fig. 3.14a demonstrates that it is important to consider temporal variations in water quality at the upstream end of the reach while the longitudinal sampling was completed. The measurements were made in the upstream direction, beginning at 208 m, but on some days these were supplemented with a measurement taken at the culvert outlet prior to the start of the longitudinal sampling. Thus, the second set of data at 0 m in Fig. 3.14a can be used to distinguish the effects of temporal variability from those of spatial variability. The downstream increases in E C on July 27 (Fig. 3.14a) might have been related to the decrease in EC at the culvert through time, rather than being caused by the inflow of higher EC subsurface water in lower B5. Similarly, the abrupt decrease in stream temperature below 150 m may indicate that the warming streamwater from upstream had not yet flowed beyond that point. The fact that marked, though slight, changes in both stream chemistry and temperature occurred at ~150 m in Fig. 3.14a may not be coincidental. A series of large pools (~2 m wide by up to 0.5 m deep) in the creek channel from -145 to 155 m could have delayed the arrival of streamwater from upstream. Pattern 2 days displayed greater variation in both EC and temperature along the reach (Fig. 3.14c,d) than profiles from pattern 1 days. Electrical conductivity was relatively constant in upper B5, but dropped by 5-8 uS cm" below 150 m. Creek temperatures were 1  not obviously related to variations in EC in the upper sub-reach, while in lower B5 decreases in creek temperature were strongly linked to decreases in electrical conductivity. Linear regressions using all data from distances > 150 m indicated highly significant (P<0.01) positive relations between E C and creek temperature on both days (Fig. 3.15). Combined with evidence of positive hydraulic gradients downstream of 150 m (Figs. 3.10, 3.11), the strong relations between EC and temperature (Fig. 3.15) suggest that creek temperatures in lower B5 were dominantly controlled by the inflow of cool, low E C subsurface water on pattern 2 days. The creek experienced a net loss of water as it flowed  60  EC  -e-  Temperature  0  11  1 7 0 $ _ _ _ ^ _ g _ ^ .  July 27 8:30-9:30 (Pattern 1)  ^140 co 180  8  150  7  ^140  12 11  o  10 160  9  0 150 ro  July 18 13:10-13:45 (Pattern 2)  •|140  8 7  T  180. <s—e-e-  170  E  CD CD  O  T  ."l^o'© | 160 c 0 150 ro  1 I  O OO  July 29 11:30-12:00 (Pattern 2)  140  180 o-  Q_,Q_.  Cft- (D  170 ,road right-of-way  9 8  150  7 | i | i 50 100 150 200 Distance Downstream (m)  140  - i  Aug 1 12:40-13:00 (Pattern 1)  co 180  160  July 26 10:00-10:30 (Pattern 3- day 5)  0  CD Q.  10  160  140  CD i_ *—< CD  11  +-++  150  9  12  -e-eo  ^ C ^ O ^ O ^  160  9  150  1  170 t^l-  10  160  a  180  12  180  r  August 4 14:20-14:46 (Pattern 3- day 2)  0  50 100 150 200 Distance Downstream (m)  F i g u r e 3.14 Spatial p a t t e r n s o f electrical c o n d u c t i v i t y a n d t e m p e r a t u r e a l o n g B5 d u r i n g c r e e k t e m p e r a t u r e pattern 1 ( a , b ) , pattern 2 ( c , d ) , a n d pattern 3 (e,f) d a y s . T h e low s h a d e a s s o c i a t e d w i t h t h e r o a d r i g h t - o f - w a y e x t e n d s f r o m 0 t o - 1 5 m. T h e a r r o w s in a ) i n d i c a t e t h e relative t i m i n g o f t h e m e a s u r e m e n t s . N o t e t h e different r a n g e of t h e t e m p e r a t u r e s c a l e in f ) , c o m p a r e d to t h e o t h e r f i v e panels.  61  from 120 to 195 m (Fig. 3.8), but the low spatial resolution of the sequential streamflow measurements may have obscured groundwater inflow in the lower sub-reach. Fig. 3.8 shows a more plausible streamflow pattern for July, 19 inferred from the E C and hydraulic gradient data.  11.5  11 o o o  10.5  y = 0.13x-11.19 r = 0.97  o  Jul-18, <150 m  •  Jul-18, >150 m  •  Jul-29, <150 m  •  Jul-29, >150 m  2  10  £  • ED  9.5  91  Linear (Jul-29, >150 m) Linear (Jul-18, >150 m)  y = 0.11x-8.57 r = 0.93 2  8.5 174  172  170  168  166  164  162  160  Electrical Conductivity (^S cm" ) 1  Figure 3.15 Relations b e t w e e n c r e e k t e m p e r a t u r e a n d electrical c o n d u c t i v i t y m e a s u r e d a l o n g B 5 o n July 18 a n d J u l y 2 9 (both c r e e k t e m p e r a t u r e pattern 2 d a y s ) . N o t e that t h e x - a x i s is r e v e r s e d to reflect t h e spatial patterns of Fig. 3.14.  The two pattern 3 days (Fig. 3.14e,f) support the inferred streamflow pattern in Fig. 3.8 since infiltration consumed the 3.2-3.4 L s"' of water entering the head of the reach, but the stream channel still contained water below -155 m. Consequently, a section of the channel approximately 35 m long was entirely dry on both days (Fig. 3.14e,f). The reach  62  comprised two distinct sections, with warm, high EC water in the upper portion, and cool, low EC water in the lower portion. Differences between water quality in the upper and lower sub-reaches were more marked on these pattern 3 days than observed during either of the other patterns (Fig. 3.14), with water temperatures 3 °C or more cooler, and E C values 1025 pS cm" less in the lower reach. In the upper sub-reach, variations in EC were negligible 1  and temperatures increased slightly in the downstream direction, particularly within the road right-of-way (Figs. 3.14e,f). Temperature and EC varied erratically in lower B5 on July 26 (Fig. 3.14e) because water was isolated to individual pools and no flow occurred between them (Fig. 3.9). The chemical signature of each pool on July 26 probably reflected variations in the origins of inflowing groundwater. Flow from pool to pool was more significant on August 4, since the lower section of the reach had not been disconnected from incoming streamflow for as long as on July 26 (Fig. 3.9). Consequently, E C patterns were more spatially coherent in lower B5 in Fig. 3.14f, although temperature fluctuated erratically. The longitudinal temperature data of Fig. 3.14 were not all sampled during periods of peak daily temperature. However, the differences between temperatures near the BR and L O sites in Fig. 3.14e,f of -3.4 and -4.0 °C compare favourably to the values of -3.4 and -3.6 °C indicated as the differences between peak afternoon stream temperatures by the DFO loggers on July 26 and August 4. This suggests that pattern 3 days identified in Fig. 3.5 were caused by the streamflow discontinuity between the sub-reaches.  3.3.6 Temperature Patterns in Relation to Hydrology The previous section demonstrated that the temperature patterns identified in section 3.3.1 were linked to temporal variations in longitudinal patterns of both creek temperature and chemistry. These longitudinal patterns, in turn, were broadly related to variations in stream-subsurface interactions and streamflow along the reach. This section links temporal variations in streamflow at the upstream end of the study reach to the temperature patterns identified in section 3.3.1.  63  The timing of unusual fluctuations in creek temperature on specific days coincided with changes in the magnitude of streamflow at the upstream end of the reach. For example, on the afternoon of July 26, temperatures at the L O site were relatively steady at -7.7 °C (-3 °C cooler than at the B R site). At about 23:00 PST, the L O creek temperature abruptly warmed to 10.2 °C, approximately the same temperature as at the B R site (Fig. 3.16a). The 10 mm of rain on the evening of July 26 (Fig. 3.16c) increased the input of streamflow at the upper end of the reach so that it was not entirely lost to infiltration before reaching the lower sub-reach. Hence, warmer streamwater flowed downstream to the L O logger where the measured temperature had previously been controlled by locally emergent subsurface water. Fig. 3.14a confirms that the reach was re-connected by 9:00 on July 27, in contrast to the disconnected profile of EC and temperature observed the previous morning (Fig. 3.14e). Fig. 3.16a,b suggests that when streamflow exceeded 5 L s" at the flume, creek 1  temperature patterns 1 or 2 were consistently observed. When streamflow dropped below 4.5-5.0 L s'\ creek temperature pattern 3 often occurred, indicating that 4.5-5 L s" of 1  streamflow at the flume was required to maintain continuous streamflow along the reach. A threshold value of 4.5 L s" for the B5 disconnection is in good agreement with the inferred 1  pattern of streamflow in Fig. 3.8. On July 19, inferred streamflow at 155 m was 0.6 L s"  1  when streamflow at the culvert was 5.0 L s" , suggesting that 4.4 L s" of streamflow at the 1  1  culvert would have been entirely consumed by infiltration before reaching the L O logger. However, while the relation between streamflow and creek temperature identified in Fig. 3.16 was generally supported through segments 1 and 2 of the study period, streamflow disconnection was not observed as frequently in segment 3 as would be predicted by streamflow variations (Fig. 3.9).  64  26-Jul  28-Jul  30-Jul  1-Aug  3-Aug  5-Aug  F i g u r e 3.16 (a) C r e e k t e m p e r a t u r e s at B 5 B R a n d B 5 L 0 . ( b ) E s t i m a t e s of s t r e a m f l o w at t h e B5 f l u m e , ( c ) Rainfall m e a s u r e d at t h e O p e n site.  65  3.3.7 Energy Exchanges across the Water Surface Nighttime net radiation at the Forest site was successfully modelled using Eq. 2.9 with an s . value of 0.953 (Fig. 3.17a). The direction of the net radiation flux corresponded a v  closely to the difference between air and stream temperature (Fig. 3.17a,b). During daylight hours, net radiation exceeded the net longwave flux (L*), indicating that a proportion of solar radiation penetrated the forest canopy. Half-hourly averages of net radiation were modelled by summing the net longwave flux and 1% of the solar radiation measured at the Open site (i.e., <> j = 0.01 in Eq. 2.8). The model was reasonably accurate, although substantial underestimation occurred at some times (Fig. 3.17c). In particular, average net radiation for the half-hour ending at 14:00 PST was under-predicted by up to 166 W m" . Estimates of (j) at 2  that time of day ranged from 0.02 to 0.22, with a six-day mean of 0.08. The net radiation model of Fig. 3.17c and the temperature data of Fig. 3.18a were used to generate continuous time series that span periods of missing net radiation data (e.g., July 18-22) and creek drying at the Forest site (Fig. 3.18b). The daytime modelled values should be considered a lower bound, since no effort was made to account for the temporal variations in <) suggested by Fig. 3.17. Modelled net radiation near the upstream end of the study reach was positive during the hours 8:00 to 20:00 on all days except for the coldest night, when negative values occurred from 18:00 on August 9 to 9:00 on August 10 (Fig. 3.18b). The sum of the estimated convective fluxes (Q +QH) at the Forest site exhibited E  similar temporal patterns to net radiation, with positive values from 8:00 to 21:00 on most days (Fig. 3.18c). During the day, air temperature was usually higher than water temperature (Fig. 3.18a), producing a positive sensible heat flux (Fig. 3.18c). The air above the creek was often nearly saturated with water vapour, generating vapour gradients conducive to condensation onto the stream surface (Fig. 3.18d). Occasional increases in windspeed were associated with decreased vapour pressures and potential for evaporation from the creek, likely caused by down-mixing of drier air from above the canopy (Fig. 3.18d,e). For example, on the afternoon of July 28, negative latent heat fluxes offset positive sensible heat fluxes, resulting in slightly negative  QE+QH  values as early as 16:10 (Fig. 3.18c). However,  the insensitivity of the anemometer to the low wind speeds observed above the stream (Fig. 3.18e) creates uncertainty in the estimated magnitudes of the convective fluxes.  66  400.8  160 Q*  120  L* ( m o d e l l e d )  x =  80  CD >  40  03  A  •o  CO  or  ~i  1  1—i—i  1  T  r  AirT  20  Creek T  9  Interpolated C r e e k T  16  I !  <D  I  \  V 1 «  1  1  1  1  12  0  o_ E CD  „  8  T  24  23  -20  25  T  i  26 27 Date (July, 2 0 0 0 )  • i  28  i  i  i  i  i  i  i  ~l  r  29  5 August  20 40 60 M e a s u r e d Q* (WITT )  80  100  2  Figure 3.17 o v e r B 5 . (b) c r e e k drying (circled d a t a  (a) M e a s u r e d net radiation a n d m o d e l l e d net l o n g w a v e radiation at t h e Forest site Air a n d c r e e k t e m p e r a t u r e s m e a s u r e d at t h e Forest site. D a t a g a p s resulted f r o m or d a t a - l o g g e r m a l f u n c t i o n d u e to p o w e r loss, (c) Net radiation m o d e l l i n g results w e r e m e a s u r e d / m o d e l l e d for t h e half-hour e n d i n g at 14:00 P S T ) .  67  17-Jul  23-Jul  29-Jul  5-Aug  11-Aug  Figure 3.18 E n e r g y e x c h a n g e s a c r o s s t h e w a t e r s u r f a c e in B5 (a) air t e m p e r a t u r e m e a s u r e d at the Forest site, a n d c r e e k t e m p e r a t u r e s m e a s u r e d near the u p s t r e a m e n d of t h e B5 s t u d y r e a c h , (b) m o d e l l e d net radiation, u s i n g t h e m o d e l s h o w n in Fig. 3.17c a n d t h e t e m p e r a t u r e d a t a s h o w n in panel a), (c) e s t i m a t e d c o n v e c t i v e f l u x e s at Forest site, (d) air v a p o u r p r e s s u r e , s a t u r a t i o n v a p o u r p r e s s u r e ( S V P ) a n d c r e e k v a p o u r p r e s s u r e s , (e) w i n d s p e e d . T h e a n e m o m e t e r stall s p e e d w a s 0.447 m s'\ h e n c e t h e a p p a r e n t c o n s t a n t w i n d s p e e d s d u r i n g m u c h of t h e period. N o t e that t h e c r e e k b e d w a s dry at t h e Forest site f r o m A u g . 5-14.  68  There is also doubt about the reliability of the equations used to calculate the fluxes under the dominantly stable conditions that occurred when the overlying air was warmer than the creek (Fig. 3.18a). Although the magnitudes of the convective fluxes are uncertain, their directions can certainly be inferred based on differences between air and stream temperature, and air and stream vapour pressure. Since these differences were generally positive during afternoon and early evening (i.e., air values exceeded water values), the convective fluxes could not have contributed to downstream cooling in maximum daily temperatures. Evaporative fluxes from the water surface explain a negligible amount of the streamflow losses observed in upper B5. The most negative QE values estimated at the B5 Forest site were about -10 W m" , which translates to a downstream flow loss of 0.0006 L s" 2  1  in upper B5, using a latent heat of vaporization of 2.48 M J kg" and a stream surface area of 1  150 m . Arbitrarily increasing Q by an order of magnitude results in streamflow losses of 2  E  0.006 L s" in upper B5, or less than 0.2% of the observed losses (~4 L s" ). 1  1  3.3.8 Bed Heat Conduction Bed temperature gradients and heat conduction varied with location along B5. At 40 m downstream of the culvert, stream and bed temperatures were visually indistinguishable from one another at all times (Fig. 3.19a), indicating that bed heat conduction was negligible at that location. The ten-minute average values varied by only hundredths of a degree Celsius between the four samples. This lack of significant bed temperature gradients suggests rapid infiltration of creekwater, consistent with the observed streamflow losses in upper B5 (Fig. 3.8). In contrast, bed temperatures at 183 m downstream were usually lower than stream temperature by up to 2 °C (Fig. 3.19b), indicating heat conduction into the bed. Peak conductive losses of heat from the stream at 183 m were estimated at -160 W m" , but more typically ranged from -40 to -120 W m" at the time of maximum daily stream 2  temperature (Fig. 3.19c). Only temperatures at a depth of 5 cm in the creek bed will be considered for the remainder of this section since the temperature gradient between the creek and that depth was used to calculate the conductive fluxes. The distinction between bed temperature patterns in the continuous, data-logged, records from 40 and 183 m is also generally apparent in manual  69  0  ST  E  £  -80  O  -160  E 4 <D  >2.5 c m <2.5 c m  r-  >- Q. ^ O 0J . Q 1  0  1  19  21  23  i  1  i  1  r  25 27 29 31 Date (July-August, 2000)  Figure 3.19 (a) C r e e k a n d b e d t e m p e r a t u r e s m e a s u r e d at 4 0 m a n d (b) 183 m d o w n s t r e a m , (c) E s t i m a t e d b e d heat c o n d u c t i v e f l u x e s at 183 m d o w n s t r e a m , a n d (d) c r e e k w a t e r d e p t h s at 183 m d o w n s t r e a m . T h e s t r e a m b e d dried at 183 m d o w n s t r e a m o n A u g u s t 6, a n d t h e d a t a - l o g g e r w a s r e m o v e d f r o m 4 0 m d o w n s t r e a m o n A u g u s t 7.  70  measurements from the upper and lower sub-reaches (Fig. 3.20a). For example, on July 28 at 2  2  -13:00, bed heat conduction averaged -2 W m" upstream of 150 m, and -59 W m" downstream of that point (Fig. 3.20b). This heat flux in lower B5 would have caused a downstream temperature change of 0.4 °C in 50 m (i.e., dT/dx = 0.007 °C m" ), assuming a 1  c  mean stream width of 1 m and mean streamflow of 2 L s" . 1  However, the mean values for each sub-reach obscure considerable spatial heterogeneity. Pool bed temperatures were often cooler than bed temperatures observed within steps or riffles, which led to greater heat losses via conduction at those sites (Fig. 3.20). In the pool at 165 m, the steep temperature gradient in the top 5 cm of the bed caused conductive heat losses which often exceeded -200 W m" and occasionally approached 2  -300 W m " (Fig. 3.21a). 2  11 10  O  1  +  ffi-  9  o (D i_  3  8  CL  7  '  6  00 i_ CD  E  CD  5 c m depth bed T: pool  •  -  +  •  ^  5 c m depth bed T: step/riffle  • +  5 c m depth bed T: data-logged creek T: data-logged  5. 50 0 ^ -50  X  \  ^  ©-  -Q7  "O"  o  data-logged  t  -100 co CD  X  data-logged  -150  " S -200 co  O  -250  — i  40  1  60  1  1  80  i  j  i  |  i  j  100 120 140 Distance Downstream (m)  160  180  200  Figure 3.20 (a) C r e e k a n d b e d t e m p e r a t u r e s a n d (b) e s t i m a t e d b e d h e a t c o n d u c t i v e f l u x e s t h r o u g h B5 at - 1 3 : 3 0 P S T o n July 2 8 , 2 0 0 0 . F o r g r a p h i c a l p u r p o s e s , t h e c o m p o n e n t s of t h e step/riffle-pool s e q u e n c e s h a v e b e e n arbitrarily s e p a r a t e d by 2 m, a l t h o u g h t h e y w e r e c l o s e r t h a n that.  71  Bed temperature measurements across the pool revealed that conductive losses of heat at the margins of the pool at 165 m were only about half those in the deepest area, indicating that the -200 W m " fluxes were isolated to a small area. While conductive fluxes in lower B5 were mainly negative, positive values were also estimated (e.g., riffle at 197 m in Fig. 3.20) suggesting that solar radiation heated the bed in some locations.  *T E  -100  It x  -150  LT -•—»  -200  I  -250  DO.  -300  X X  X  CO  X X  X X  T3 CU  X  LO Creek T -5 cm Bed T  13 12  o o CD ZJ  CO  11  -  10  ;  ••••+-••  -10 cm Bed T  -Eh  -20 cm Bed T  i  9 -  1  \ I  -  w  8  1  CD Q.  E  cu  7 6 5  89"  4 3 22 Jul  i 24-Jul  >  i 26-Jul  •  i  1  28-Jul  i 30-Jul  1  i 1-Aug  1  i  1  3-Aug  i  i  1  5-Aug  7-Aug  Figure 3.21 (a) E s t i m a t e d b e d heat c o n d u c t i o n , a n d (b) c r e e k a n d b e d t e m p e r a t u r e s at t h e 165 m pool site in B 5 .  72  3.3.9 Energy Exchanges Driven by Groundwater Inflow & Hyporheic Exchange Infiltration, not groundwater inflow, was the dominant water exchange in upper B5. This was indicated by the close correspondence between stream and bed temperatures (Figs. 3.19, 3.20) and hydrometric evidence (sections 3.3.3, 3.3.4). The weak bed temperature gradients also suggest that hyporheic exchange was not an important thermal process in upper B5 (i.e., Eqs. 2.18, 2.19). Further, the positive hydraulic gradients required to facilitate local upwelling of infiltrated water into the channel were not apparent (Fig. 3.10). Several lines of evidence indicate that groundwater inflow influenced stream temperature in lower B5. Soil temperatures at a depth of ~1 m were as low as 4.6 °C around the 163 m transect (Fig. 3.13), indicating the presence of cold groundwater. Discharge of this cold groundwater into the pool at 165 m was suggested by mean bed temperatures of 5.0 °C and 5.4 °C at depths of 5 and 20 cm, respectively, until August 5 (Fig. 3.21b). Mean bed hydraulic gradients of 0.05 m m" in the pool at 165 m were higher than at any other site 1  (Fig. 3.10). At the 183 m site, relatively cool bed temperatures (Fig. 3.19b) and mean bed hydraulic gradients of +0.02 m rrf also indicated a groundwater influence. The thermal 1  effect of groundwater inflow in lower B5 can be predicted using Eq. 2.1, the inferred streamflow pattern from Fig. 3.8, and a range of groundwater temperatures observed in the streambed. Predicted downstream cooling ranged from 2-3 °C at relatively low streamflow to ~1 °C at relatively high streamflow conditions (Fig. 3.22).  O o  Approximate Streamflow at the B5 Culvert (L s ) 1  11  5.5  6.5  0  7.5  -0.02  10  -  TO h-  o  -  CD  Groundwater = 6°C  CD i—  O  £  E -0.04 o  Groundwater = 7°C  8  Groundwater = 8°C  O  -0.06 X T3  -0.08  h73  o  TD CD  CL  r  -1  |  ,  !  ,  !  -0.1  1 2 3 Streamflow at 155 m Downstream (L s ) 1  Figure 3.22 Downstream creek temperature change predicted for lower B5 with varying streamflow and groundwater conditions. Initial creek temperature at 155 m was held constant at 11 °C, with streamflow varied from 0.5 to 3.5 L s'\ Streamflow at the B5 flume roughly corresponding to these values is shown on the upper x-axis. Temperatures of groundwater inflow varied from 6 to 8 °C .  73  I  Vertical hyporheic exchange may also have influenced stream temperatures in lower B5. Downwelling of streamwater into shallow depths of the bed at 183 m was suggested by the rapid, but sustained, warming of the 5 cm depth on July 27 and August 1 (Fig. 3.19b), in association with rain events and increased streamflow (Fig. 3.16). This pattern suggests that hyporheic exchange was limited to the upper 5 cm of the bed at this groundwater-influenced site, and that it only occurred when creek depth exceeded -2.5 cm (Fig. 3.19d). Smaller magnitude, higher frequency, thermal fluctuations occurred at all three depths on many days, with temperatures declining by 0.2-0.4 °C in 30-40 minutes (Fig. 3.19b). These fluctuations often coincided with rapid decreases in stream temperature, but their magnitude increased with depth in the bed (Fig. 3.19b), suggesting a mechanism other than hyporheic exchange. A deeper hyporheic zone was apparent at the riffle-pool sequence at 199 m. On August 7, the channel was essentially dry at that site. The water table declined from near the bed surface to a depth of -20 cm from August 7-18 (Fig. 3.23a,c). Until August 20, bed temperatures at a depth of 20 cm in the pool (Fig. 3.23b) rarely exceeded 6 °C, suggesting a groundwater influence. Similarly, bed temperatures at 20 cm in the riffle fluctuated at - 7 °C from August 10-19 (Fig. 3.23d). The return of streamflow to the channel on August 20 modified thermal patterns at all depths (Fig. 3.23). Rapid temperature fluctuations in the bed and relatively close correspondence between stream and bed temperatures suggest that downwelling of streamwater was the dominant control on bed temperatures at this site following August 20. Mean bed hydraulic gradients of -0.03 m m" in the pool and -0.05 m m" in the 1  1  riffle facilitated streamwater penetration into the bed. Fig. 3.23 suggests that the hyporheic zone was at least 20 cm deep in the 199 m riffle, and at least 10 cm deep in the 199 m pool. This indicates that relatively deep hyporheic exchange can occur even in groundwaterinfluenced reaches, potentially causing downstream cooling on warm summer afternoons. Vertical hyporheic flow might have affected stream temperatures in lower B5, but lateral hyporheic exchange may also have played a role. The water table configuration in the riparian zone of lower B5 indicated that the creek might have been gaining groundwater from the left bank at 192 m while simultaneously losing flow to the right bank (Figs. 3.11 and 3.13). The thermal data in Fig. 3.13 support this hypothesis, since the higher temperatures measured on the right bank at 192 m suggest that the aquifer had been recharged by warm streamwater at that location. The phreatic surface contours and the subsurface thermal  74  I  20  8  10  12  14  16  18  20  22  24  26  28  30  1  3  5  28  30  1  3  5  August / September 2000  8  10  12  14  16  18  20  22  24  26  August / September 2000 Figure 3.23 (a) C r e e k w a t e r d e p t h s a n d (b) b e d t e m p e r a t u r e s a t 1 9 9 m p o o l , (c) C r e e k w a t e r d e p t h s a n d (d) b e d t e m p e r a t u r e s at 199 m riffle. D a t a - l o g g e d c r e e k t e m p e r a t u r e s f r o m 183 m are s h o w n in b ) a n d d). O n A u g u s t 7, the t h e r m o c o u p l e s at 199 m w e r e c o n n e c t e d to t h e d a t a - l o g g e r that w a s previously situated at 4 0 m d o w n s t r e a m  75  patterns in Fig. 3.13 suggest that a portion of the streamwater which flowed into the right bank at 192 m may have flowed back into the creek near the downstream end of the study reach, thereby forming a lateral hyporheic flowpath. The travel time of this flowpath would have been on the order of 10 days, assuming a relatively permeable silt with a hydraulic conductivity of 2-4 x 10" m s" and a porosity of 0.4. 6  3.3.10  1  Stream Temperature Variations in Relation to Energy Exchanges Sections 3.3.5 and 3.3.6 examined the controls on temperature patterns at the scale of  the entire B5 reach. Temperature patterns within upper B5 and lower B5 differed significantly, in response to the distinct hydrologic environments of the two sub-reaches (Section 3.3.4). Downstream warming occurred during the daytime in the upper sub-reach (e.g., Figs. 3.14b,c,f), where streamwater rapidly infiltrated to the subsurface. In contrast, downstream cooling occurred in the lower sub-reach, where groundwater influenced the channel. This section focuses on linking the various energy balance terms (sections 3.3.73.3.9) to observed creek temperatures in the two sub-reaches of B5, beginning with upper B5 and proceeding to lower B5. In upper B5, downstream changes in both daily average and extreme temperatures were positively related to modelled net radiation (Figs. 3.24, 3.25). The relations were weak to moderate. However, the intercepts of the relations were within 0.1 °C of zero (Figs. 3.24, 3.25), indicating that the stream tended to continue to warm in the downstream direction when modelled Q* was positive and that the stream tended to cool when Q* was negative. Scatter about the regression line in Fig. 3.24 was not random: all July data fell above the line, while all August data fell below the line. One conceivable cause of this pattern is that average streamflow at the culvert was higher on the July days (5.3 L s" ) than on the August 1  days (2.5 Ls"'). This difference, however, cannot explain why average temperatures decreased in the downstream direction on two days with positive modelled Q* values (August 10/11 and 11/12 in Fig. 3.24). This suggests either that Q* was not the only important energy flux in the upper B5 sub-reach, or that the modelled values are not accurate. To test whether the magnitude of observed temperature changes was consistent with modelled Q* values, Eq. 2.1 was used to simulate afternoon warming through the upper sub-  76  July 1 8 - 2 8 ( 4 0 m - 1 1 7 m j ^ August 4-5  (20 m - 1 1 0 m)  A u g u s t 7 - 1 2 ( 2 0 m - 81 m )  0  •• •  0  o o  H  (11:0011:00)  0  < E  7/8 Y = 0.28X - 0.03 r = 0.38  0  CO CD i  to  8/9  2  -0  c  $ o Q  -0  11/12 • 10/11  9/10  -0.6 -1  2  0 1 M o d e l l e d Daily Q * ( M J rrr ) 2  Figure 3.24 D o w n s t r e a m c h a n g e s in m e a n daily c r e e k t e m p e r a t u r e in u p p e r B 5 , a s a f u n c t i o n of m o d e l l e d daily Q*. A v e r a g e s a n d s u m s a r e b a s e d o n 2 4 h o u r p e r i o d s f r o m 1 2 : 0 0 P S T ( e x c e p t f o r A u g u s t 4 - 5 ) , in o r d e r t o m a x i m i z e t h e n u m b e r of d a t a points. L a b e l s b e l o w points a r e d a t e s in A u g u s t . T i m e s e r i e s a r e s h o w n in A p p e n d i c e s 2 a n d 3.  July 1 8 - 2 8  August 4-5  40 m - 1 1 7 m  20 m - 1 1 0 m  A u g u s t 7-12 20 m - 8 1 m  o  Daily M i n .  A  Daily M i n .  •  Daily M i n .  •  Daily M a x .  A  Daily M a x .  B  Daily M a x .  0.6 ®  0.4 Q  o  E  CO CD i  tf) o Q  o  O 0  9 11  %  •  • 11 • 9  • 10  12  -0.6  ®  •  • H  -0.4  A  8DO  °  -0.2  m •  ®  0.2  O  •  9  All d a t a Y = 0.014X-0.13 r = 0.36 2  -0.8  •  -1  10  X f  . A u g . 12 m a x . r e m o v e d Y = 0.014X-0.10 r = 0.55 2  2  -1.2 -30  -20  -10  0 10 20 M o d e l l e d Q * (Wnrr )  30  40  50  2  Figure 3.25 D o w n s t r e a m c h a n g e s in daily c r e e k t e m p e r a t u r e e x t r e m e s in u p p e r B 5 , a s a f u n c t i o n of a v e r a g e m o d e l l e d daily Q * d u r i n g t h e 2 hour period p r e c e d i n g t h e first o c c u r r e n c e o f t h e m a x i m u m o r m i n i m u m at t h e d o w n s t r e a m site. L a b e l s b e l o w points a r e d a t e s in A u g u s t . T h e daily m a x i m u m o n A u g . 12 m a y h a v e b e e n a f f e c t e d b y s t r e a m drying (Fig. 3.9). T i m e s e r i e s a r e s h o w n in A p p e n d i c e s 2 a n d 3.  77  reach during July 18-28. The average predicted warming for all nine days is only about onehalf the observed warming (Table 3.5). There are at least two explanations for this underprediction. First, under-prediction would result if mean stream surface area was greater than 100 m . Detailed stream width data are not available, but average bankfull width upstream of 2  the Forest site was 1.2 m, which suggests that the ~1.3 m estimate used for low flow is likely an overestimate (the stream surface area was originally estimated using the total shaded distance upstream of the Forest site (-100 m), and an assumed average width of 1 m).  Table 3.5 C o m p a r i s o n of m o d e l l e d a n d o b s e r v e d a f t e r n o o n c r e e k w a r m i n g in u p p e r B 5 , b a s e d o n o b s e r v a t i o n s of m a x i m u m daily t e m p e r a t u r e s f r o m t h e 4 0 m a n d 117 m sites. M o d e l r e a c h a r e a w a s held c o n s t a n t at 100 m f o r all c a l c u l a t i o n s . S t r e a m f l o w (q) w a s c a l c u l a t e d a s t h e a v e r a g e of t h e f l u m e e s t i m a t e a n d e s t i m a t e d f l o w at t h e 1 1 7 m site, b a s e d o n e x t r a p o l a t i o n f r o m t h e s e q u e n t i a l s t r e a m f l o w m e a s u r e m e n t s of J u l y 17/19. E n e r g y f l u x e s a r e a v e r a g e s c a l c u l a t e d f o r t h e 2-hour period prior to t h e first o c c u r r e n c e of t h e daily c r e e k t e m p e r a t u r e m a x i m a at 117 m ( t h e Forest site). All o b s e r v e d d a t a are in bold font. Particularly c r u d e e s t i m a t e s of s t r e a m f l o w a r e in g r e y font. T h e " R e q u i r e d Flux" w a s c a l c u l a t e d a s t h e e n e r g y flux r e q u i r e d to s i m u l a t e t h e o b s e r v e d w a r m i n g , after s u b t r a c t i n g t h e e s t i m a t e d w a r m i n g c a u s e d by L*. 2  Mean Q* (4) = 0 . 0 1 ) ( W m" ) 2  Mean  Observed Warming (°C)  Required Flux ( > L * ) ( W m" )  ( W m" )  ( W m" )  5.2 3.8 3.5 3.1  Modelled Warming (°C) 0.22 0.20 0.12 0.20 0.24 0.28  0.27 0.49 0.53 0.20 0.34 0.48  10.9 48.5 92.8 7.5 17.3 34.0  25 23 19 11 21 11  253 488 373 712 343 704  0.23 0.39 0.28 0.36  0.2 68.2 56.2  14 13 8  46.8  16  273 187 589 436  q (L s ) 3.6 1  18-Jul 19-Jul 20-Jul 21-Jul 22-Jul 23-Jul  32.7 30.9 26.1 31.1 35.3 36.2  26-Jul 27-Jul 28-Jul  20.2 19.2 17.6  1.8 5.2 5.8  0.26 0.09 0.07  means  27.7  4.0  0.19  3.6  2  QH  +QE 2  Ki 2  Required * ' 0.04 0.10 0.25 0.01 0.05 0.05 0.00 0.36 0.10 0.11  The second explanation is that net radiation modelled for the 40 m downstream site underestimated the actual mean energy inputs through the sub-reach. The mean incoming long-wave flux ( L i ) may have been underestimated due to the particularly shaded and sheltered conditions at the Forest site. If the atmosphere/vegetation overlying the Forest site was 2 °C cooler than the sub-reach mean, Q* would have been underestimated by -10 W m . 2  Similarly, the solar radiation (K*) component of net radiation may have been under-  78  estimated in the Q* model. Adding an average of 11% of the K-l values observed at the Open site to the L * values from the Forest site fills the gap between the modelled and observed temperature increases (Table 3.5). The daily estimates of the required (j) parameter vary widely (Table 3.5). However, the mean (j) value of 0.11 is not unreasonable considering that similar values were estimated at the heavily-shaded Forest site for a half-hour period of early afternoon (section 3.3.7). The convective fluxes (QH + QE) may also have contributed to downstream warming. On four of the nine days, the estimated convective fluxes were sufficient to provide the required warming (Table 3.5). However, on average the convective fluxes were estimated to provide only one-third of the heating fluxes (Table 3.5), suggesting that inputs of solar radiation would still be required (albeit at a lower rate i.e., <j)~0.06-0.07). Given the poor constraints on these energy balance estimates, the best estimate of the (j) value for the upper B5 sub-reach is 0.1 (one significant figure). This analysis suggests that afternoon modelled Q* values were likely underestimates, but it appears that there may also have been a heat storage mechanism in upper B5. On four days, downstream decreases in temperature maxima occurred despite positive modelled Q* values (Fig. 3.25), indicating that cooling processes were not accounted for. Downstream decreases in temperature minima also occurred on those four days (Fig. 3.25), suggesting that the presence of cool water in the sub-reach during the preceding mornings may have influenced the afternoon temperatures. Thus, the B5 stream often continued to warm as it flowed through the upper sub-reach, in response to positive energy exchanges across the water surface. However, downstream cooling sometimes occurred that could not be explained by energy exchanges across the water surface. The most common temperature pattern in the lower "gaining" sub-reach of B5 was downstream cooling, although like upper B5 this trend was also not temporally constant. Creek temperatures at 183 m were consistently up to ~1 °C cooler than at 158 m on relatively cloudy, moderate streamflow days (e.g., July 18-20 in Fig. 3.26). Estimated downstream cooling due to bed heat conduction (0.007 "Cm" ) explains only 18% of the observed cooling 1  (0.04 °C m" ) on those days. In contrast, the observed cooling rate corresponds to that 1  predicted by inflow of 8 °C groundwater (Fig. 3.22). Groundwater discharge of 5 to 6 °C was observed at 165 m (3.21b), although bed temperatures at 183 m suggested a warmer groundwater influence of 7.5-8 °C (Fig. 3.19b). When streamflow at the B5 flume exceeded ^  79  7 L s" , downstream cooling was on the order of 0.5 °C or less between 158 m and 183 m 1  (e.g., July 27-28, August 1 in Fig. 3.26). This amount of decreased cooling with increased streamflow (<0.02 °C m" at 7 L s" vs. 0.04 °C m" at 5 L s" ) is consistent with predictions 1  1  1  1  (Fig. 3.22). Downstream warming often occurred below 158 m when streamflow became discontinuous and originated as locally emergent subsurface water around 158 m (i.e., when streamflow at the B5 flume was less than 4.5-5 L s" , Fig. 3.26b). Spatial variations in the 1  temperature of groundwater inflow may have been responsible for some of this warming. For example, creek temperature minima at 183 m were about 1 °C warmer on July 25-26 than on July 23-24 (Fig. 3.26a), although daily air temperature minima were similar at 9-10 °C on all four days (Fig. 3.3). The stream in lower B5 was isolated to individual pools on July 26  18-Jul  20-Jul  22-Jul  24-Jul  26-Jul  28-Jul  30-Jul  1-Aug  3-Aug  5-Aug  Figure 3.26 (a) Creek temperature observations from lower B5, and (b) streamflow estimates at the B5 flume. 80  (Fig. 3.9), suggesting that the minimum temperature at the 183 m site may have been controlled by relatively warm groundwater similar to the bed temperatures at that site (Fig. 3.19b). In contrast, the cooler groundwater inflow at 165 m (Fig. 3.21b) likely affected the daily minima at 183 m on July 23-24, when the two sites were connected by channel flow (Fig. 3.9). Rapid mid-day fluctuations in temperature at 183 m, and the fact that the daily maxima at that site often preceded those at 158 m (Fig. 3.26a), suggest that energy exchanges across the water surface were also an important influence. When streamflow in lower B5 became isolated to the furthest downstream point of the reach (209 m), creek temperatures were relatively constant at ~7.7 °C (Fig. 3.27). This suggests a relatively warm source of subsurface water, perhaps reflecting the influence of the slow lateral hyporheic flowpath identified in section 3.3.9. However, this temperature is consistent with groundwater temperatures inferred from bed temperatures at the 183 m site.  _  13  O ^  12  20 m Downstream 209 m Downstream  CD  3  11  'i i  CO  H E  £  j  10  i  9  CD CD  if  \t  8  + -  6  i_r  7 11-Aug  12-Aug  13-Aug  14-Aug  15-Aug  16-Aug  17-Aug  18-Aug  Figure 3.27 Creek temperature data from near the upstream and the downstream ends of the B5 study reach from August 10-18.  81  3.4 Results for B3 Section 3.4.1 describes temperature patterns based on data from the DFO loggers located at the HI, B R and L O sites. Sections 3.4.2 to 3.4.5 examine the hydrologic context of the stream and its relation to temperature variations. Sections 3.4.6 to 3.4.11 evaluate the processes governing energy exchanges and their expression in temperature variations along B3.  3.4.1 Overview of Stream Temperature Patterns Except for four days (August 10, 14, and September 1, 2), temperatures at B3HI were always lower than at B3BR and B3LO (Fig. 3.28). Diurnal variations at HI were also smaller than at the two lower sites (Table 3.6). Diurnal creek temperature variability was greatest at BR, with the daily maxima exceeding those at L O on all but one day (September 1). The  ^  daily maxima occurred earliest at the B R site, with the L O site maxima occurring, on average, over two hours later (Table 3.6). Variability in the timing of the daily maximum was also smallest at the B R site, with a standard deviation of 1.1 hours (Table 3.6). Daily temperature minima were generally similar at the B R and L O sites, although the daily minima at the L O site exceeded those at the B R sites on 41 of the 50 study period days. The daily minima at the L O site were usually <0.5 °C warmer than those at the B R site (Fig. 3.28). However, the L O minima were 0.5 °C or more warmer than B R on eight days of the study period and the L O minima were less than the B R minima on seven days. The former pattern was associated with "cooling limbs" or local valleys in the daily minima time series, while the latter pattern was associated with "warming limbs" or local peaks in the daily minima time series (Fig. 3.28).  Table 3.6 S u m m a r i e s of m e a n daily c r e e k t e m p e r a t u r e e x t r e m e s at t h r e e sites a l o n g s t r e a m B 3 f r o m J u l y 17 t o A u g u s t 3 1 , 2 0 0 0 (prior to t h e o n s e t of c o o l e r t e m p e r a t u r e s in S e p t e m b e r ) . " T " is t e m p e r a t u r e , a n d "t" is t i m e ( P S T ) . W h e n t h e daily m a x i m u m t e m p e r a t u r e w a s r e c o r d e d m o r e t h a n o n c e , t h e a v e r a g e t i m e of all o c c u r r e n c e s w a s c a l c u l a t e d . C a l c u l a t i o n s of t h e m e a n a n d s t a n d a r d deviation ("St. Dev.") of t h e t i m e of daily m a x i m u m did not i n c l u d e t i m e s l e s s t h a n 6:00 P S T , s i n c e t h e s e f e w d a y s w o u l d h a v e s k e w e d t h e m e a n . T h e n u m b e r of s u c h e x c l u d e d t i m e s at e a c h site is n o t e d in t h e f u r t h e s t r i g h t - h a n d c o l u m n . M e a n Daily M a x i m u m T  M e a n Daily M i n i m u m T Site  M e a n T (°C) St. D e v . (°C)  M e a n T (°C) St. D e v . ( ° C ) M e a n t (h) St. D e v . (h)  t<6.0h  B3HI  7.3  0.6  7.7  0.6  16.1  2.5  B3BR  8.4  1.1  11.7  1.5  14.2  1.1  0  B3LO  8.7  1.0  9.8  0.8  16.9  1.8  2  82  5  o 4 2-Aug  4-Aug  6-Aug  8-Aug  10-Aug  12-Aug  14-Aug  16-Aug  18-Aug  18-Aug  20-Aug  22-Aug  24-Aug  26-Aug  28-Aug  30-Aug  1-Sep  3-Sep  F i g u r e 3.28 C r e e k t e m p e r a t u r e t i m e series f r o m t h e t h r e e D F O d a t a l o g g e r s in s t r e a m B 3 . (a,b,c) of t h e study period are t h e s a m e a s u s e d in Figs. 3.2 a n d 3.5.  83  Segments  3.4.2 Variations in Streamflow Streamflow varied relatively little at the B3 culvert between the four days it was measured (Fig. 3.29). Only 3 mm of precipitation fell in the two weeks prior to August 16, while 62 mm fell in the two weeks prior to August 30, yet streamflow at the culvert only increased from 1.1 L s" to 1.3 L s" between those two dates. 1  1  Downstream increases in streamflow indicated net groundwater inflow through the study reach (Fig. 3.30). The largest streamflow increases occurred in the lower 50 m on both days, although slight increases were observed throughout the reach (Fig. 3.30).  ~  10  1.6  8  1.2  6  0.8  SS 4 a>  0.4 £ 2  0 11  15  19 23 Date (August, 2000)  27  31  Figure 3.29 Temporal variations in streamflow at the B3 culvert, measured using constant-rate salt injection. Bars indicate precipitation measured at the Open site.  0  50  100 150 Distance Downstream (m)  200  250  Figure 3.30 Spatial variations in streamflow along the B3 study reach. Locations of the BR and LO data-loggers are shown on the x-axis. Streamflow from B2 occurred on August 27, as shown. 84  3.4.3 Stream-Subsurface Interactions Hydraulic gradients in the B3 streambed were generally weak, with the exception of two negative sites (Fig. 3.31). The positive gradient in the pool at 224 m (0.007 m m" ) was 1  at the limit of confident detection. However, a gradient of 0.007 would be sufficient to provide the inflow evident in Fig. 3.30 through a substrate with saturated hydraulic conductivity of 0.1 cm s" (coarse sand or fine gravel), assuming that upwelling zones were 1  limited to an area of 75 m (~ one-half of the reach). The high frequency (6/7) of negative 2  mean gradients was likely related to the predominance of step/riffle sampling sites. Well data from six cross-stream transects (Fig. 3.32) are also not entirely consistent with the downstream increases in streamflow, as negative gradients were observed at two sites in the first 100 m of the reach. However, bank-stream hydraulic gradients were consistently positive further downstream, with particularly high water levels observed in the right bank (Fig. 3.32). Hydraulic heads varied modestly through time, with a maximum increase of 19 cm between August 15 and September 4 in the left well at 90 m downstream (Fig. 3.32). Together with observations of net groundwater inflow along the reach (Fig. 3.30), Figs. 3.31 and 3.32 indicate that groundwater interacted closely with the B3 channel, and that hydraulic gradients were sufficiently diverse to generate a variety of hyporheic flowpaths.  0.2  r  6  0  T  3  ^5  J (pool)  -0.2 -0.4 -0.6 -0.8 -1  T  T  T  50 100 150 200 Distance Downstream of B3 Culvert (m)  250  Figure 3.31 B e d hydraulic g r a d i e n t s m e a s u r e d a l o n g B 3 . Vertical r a n g e s r e p r e s e n t m a x i m u m a n d m i n i m u m v a l u e s o b s e r v e d , w h i l e horizontal lines indicate s t u d y period m e a n s ( n = 4 , e x c e p t f o r site 7 at 5 4 m, w h e r e n = 2 ) . Site n u m b e r s a r e s h o w n only to highlight t h e locations of plotted points.  85  60  54 m  40 20 0 60  90 m  +  40 20  -  0  i  i  i  1  i  i  i  i  i  1 A u g u s t 15 August 20 August 27 September 4  60 40  DFO well at 218 m  218 m  -  20 0  + I  I  I  co* i  I  I  I  !  0 D i s t a n c e from S t r e a m S t a g e M e a s u r e (m)  I  j I  5  10  Figure 3.32 Hydraulic h e a d at c r o s s - s t r e a m t r a n s e c t s in B 3 at v a r i o u s d i s t a n c e s d o w n s t r e a m . A c o m m o n d a t u m w a s not u s e d f o r all c r o s s - s e c t i o n s (i.e., t h e arbitrary d a t u m w a s different at e a c h transect).  86  3.4.4 Tracer Tests Constant-rate salt injections revealed that solute pulses traveled differently along four segments of B3 (Fig. 3.33). Time to plateau at the three -20 m sub-reaches ranged from 9-70 minutes, compared to 3.5 minutes for the 17 m long culvert-pool segment (Fig. 3.33).  430 _ 460  I  E 420 o  440  O  co  a) Culvert  420  L = 17 m q = 1.27 L s -  O  b) Sub-reach 1 A>  410  L = 20 m q = 1.65 L s  LU  1  400 4 0  400 0.2 Time (hours)  1  1  e - e — e - o  0.2  0.4 0.6 Time (hours)  0.8  1  430 xnrfflfjxiOEn  E 420 co  • •  -a—a  •  D  •J  u • '  c) Sub-reach 2  O 410  L = 23m q = 1.86 L s  LU  1  ie©oo_o_Q_,  400  ~  0  \  r  r  n  0.2 0.4 0.6 0.8  1  ~i  1.2 1.4 1.6 1.8 Time (hours)  1  2  1  r  ~i  r  2.2 2.4 2.6 2.8  430 -B-B-B-B  E 420 co O LU  B B S  d) Sub-reach 3 ?  410  , L = 20 m ]q=1.84Ls-  I  1  400 0  0.2  0.4 0.6 Time (hours)  0.8  Figure 3.33  1  T r a c e r test injection results ( E C = Electrical Conductivity). Circles indicate t h e postinjection p e r i o d . S y m b o l s indicate m e a s u r e d v a l u e s ; lines indicate s i m u l a t e d v a l u e s .  87  3  When the uncertainties associated with the four parameters fitted in OTIS-P are considered, only A and A varied significantly between the four tracer test segments (Fig. 3.34). Stream s  cross-sectional area (A) at the culvert/pool was substantially lower than values for the three sub-reaches (Fig. 3.34a). The stream cross-sectional areas of sub-reaches 1 and 3 were similar and both were substantially smaller than that of sub-reach 2 (Fig. 3.34a). Similarly, the A value for the culvert/pool was not significantly different from 0, while values for subs  reaches 1 and 3 were significantly greater than 0 but not significantly different from one another, and sub-reach 2 had the highest value by a factor of about four (Fig. 3.34c).  0.06  0.03  0.04 0.02 </>  E  E  <  0.02  Q  0.01  0 -0.02 culvert  s-r 1  s-r 2  s-r 3  culvert  s-r 1  s-r 2  s-r 3  culvert  s-r 1  s-r 2  s-r 3  0.006  0.03  0.004 0.02 &  0.002  03 <  0.01  0  • -0.002 culvert  s-r 1  s-r 2  s-r 3  Figure 3.34 Fitted parameters from OTIS-P analysis of tracer tests in B3: a) channel crosssectional area, b) longitudinal dispersion coefficient, c) storage zone cross-sectional area, d) storage zone exchange coefficient. Error bars are 95% confidence intervals. Sub-reaches are indicated by "s-r", and locations are shown on Fig. 3.35a.  88  The storage zone cross-sectional areas (A ) computed for the sub-reaches appear to be s  related to channel complexity. Mean depths of all three sub-reaches, computed as crosssectional area divided by measured wetted width, were 3-4 cm. However, sub-reach 2 displayed the most complex morphology of the three sub-reaches, with the channel separating into as many as three sub-channels spread over a total width of -3.5 m, before flowing through a few small pools associated with woody debris. The ratio As/A was highest for sub-reach 2 at 0.5, with values of less than 0.2 for the other two sub-reaches (Table 3.7).  Table 3.7 Best fit v a l u e s of hydrologic p a r a m e t e r s f o r f o u r tracer test s e g m e n t s a l o n g B3, o b t a i n e d using O T I S - P , a n d v a r i o u s s u b s e q u e n t l y calculated v a l u e s .  Derived Parameters  Fitted Parameters Segment  A(m )  D(m s" )  A (m )  « (s- )  Culvert  0.005 0.023 0.029 0.020  0.020 0.025 0.043 0.031  0.0004 0.0033 0.0153 0.0038  0.0020 0.0010 0.0006 0.0027  0.022 0.017 0.054  0.024  0.033  0.0075  0.0014  0.031  Sub-reach 1 Sub-reach 2 Sub-reach 3  Mean (Subreaches 1,2,3)  2  2  1  2  s  1  dT/dx (°C rrf )  A /A  1.3 5.5 2.5  -0.0009 -0.0025 -0.0037  0.14 0.53 0.19  3.1  -0.0024  0.29  q p(Ls" rrf ) d p (cm) 1  1  HY  HY  HYP 1  s  3.4.5 Electrical Conductivity & Temperature Patterns After warming over 1 °C through the unshaded road right-of-way, the creek temperature decreased at an approximately linear rate between ~30 and 210 m on the afternoons of August 14 and 15 (Fig. 3.35a). Downstream decreases in electrical conductivity (Fig. 3.35b) generally corresponded to downstream increases in streamflow (Fig. 3.35c). However, creek temperature also decreased in at least one section (150-175 m) where no changes in electrical conductivity, and insignificant increases in streamflow, were observed. The longitudinal temperature gradient in that section was about twice as steep as the reach-length average (Fig. 3.35a), suggesting a disproportionate cooling effect despite the lack of an obvious groundwater influence. Moderately strong relations between stream temperature and electrical conductivity (Fig. 3.36) suggest that groundwater inflow was a significant influence on B3's thermal regime, but that other factors were also important.  89  Figure 3.35 (a) T e m p e r a t u r e p a t t e r n s in B 3 o n t h e a f t e r n o o n s of A u g u s t 15,16. (b) Electrical conductivity p a t t e r n s o n A u g u s t 14, 15, 16. (c) S t r e a m f l o w o n A u g u s t 16, with c a l c u l a t e d uncertainties. T h e m i x i n g m o d e l results in a) a r e b a s e d o n a T of 7.4 + 0.6 ° C , a l t h o u g h t h e r a n g e b e t w e e n t h e points a n d the u p p e r error bar is believed to be m o s t realistic b a s e d o n t h e results of s e c t i o n s 3.4.8 a n d 3.4.11. G W  90  Figure 3.36 Relations b e t w e e n c r e e k t e m p e r a t u r e a n d electrical conductivity m e a s u r e d a l o n g B 3 o n A u g u s t 14 a n d A u g u s t 15. N o t e that t h e x-axis is r e v e r s e d to reflect t h e spatial patterns of Fig. 3 . 3 5 .  3.4.6 Energy Exchanges across the Water Surface The direction of the net radiation flux corresponded closely to the difference between air and stream temperature (Fig. 3.37a,b). In general, net radiation at the Forest site was positive from 8:00 to 18:00, except for August 26, 27, 31, and September 1. The sum of the estimated convective fluxes ( Q +QH) at the Forest site exhibited similar temporal patterns to E  net radiation, although the magnitude of the fluxes was considerably smaller, particularly during the daytime (Fig. 3.37c). The directions and magnitudes of the sensible and latent heat fluxes were similar on most days, although there were some exceptions. On the afternoons of August 15, 18, and 27, slightly negative latent heat fluxes were estimated while the sensible fluxes were positive. Relatively high windspeeds were also recorded on the first two of those afternoons (Fig. 3.37e), suggesting down-mixing of drier air.  91  r  0.52  £  0.48 0.44  I ill  15-Aug  I  ,  ~i  J.  L  i  i  i  r  21-Aug  27-Aug  2-Sep  Figure 3.37 Energy exchanges across the water surface at the B3 Forest site: (a) measured air and creek temperatures, (b) measured net radiation, (c) estimated convective fluxes, (d) measured air vapour pressures, calculated air saturation vapour pressures (SVP) and creek vapour pressures, and, (e) measured windspeed. The anemometer stall speed was 0.447 m s" , hence the apparent constant wind speeds during most of the period. 1  92  Net radiation observations at the B3 Forest site could be simulated reasonably well using a <) value of 0.035 and an e . value of 0.945 (Fig. 3.38). The modelling results exhibit a v  relatively uniform scatter around the 1:1 line, although substantial under-estimation occurred during some mid-day periods (Fig. 3.38).  -40  -20  0  20  40  60  80  Measured Q* (Wrrr ) 2  Figure 3.38 Comparison of modelled and measured net radiation at the B3 Forest site (halfhourly averages). In the model, e . = 0.945 and rj) = 0.035. The data enclosed in the elipse were measured between 11:30 and 13:30 on four days (August 15, 18, 27, 29). Data from September 1 are not plotted because snow accumulation on the radiometer likely affected the measurements. a v  3.4.7 Bed Heat Conduction No continuous time series of bed temperature were recorded in B3 and the sampling scheme was also not as spatially-intensive as at B5. On the warmest afternoon for which bed temperature data are available, conductive fluxes cooled the creek water at all seven sites as the maximum temperature pulse traveled downstream (Fig. 3.39). The spatial mean of the best estimates of Qc in Fig. 3.39 was -30 W m" , with a standard deviation of 14 W m" . The high degree of spatial variability in Fig. 3.39 indicates considerable uncertainty associated  93  with the reach-averaged conductive fluxes, which is compounded by the relative lack of data from pool sites. The conductive flux at the one pool site is about twice the average of the flux at the nearest two or three sites (Fig. 3.39), suggesting that mean Q for the reach might c  have been -50% more negative if pools had been sampled as often as riffles/steps. Estimated bed heat conductive fluxes were not significantly correlated with bed hydraulic gradients (Fig. 3.31), even with the data from 54 m downstream excluded. Measurements of bed temperature gradients during afternoon or early evening of days other than August 15 also indicate negative conductive bed heat fluxes, but bed temperature gradients on the mornings of August 18 and 27 were reversed between the 5 and 20 cm depths (Fig. 3.40). Meteorological and hydrological conditions on those two mornings  0  -20  -40  -60  pool  ~1  1  1  I  1  200 50 100 150 Distance Downstream of B3 Culvert (m) F i g u r e 3.39 E s t i m a t e d c o n d u c t i v e b e d heat f l u x e s a l o n g B 3 at - 1 6 : 0 0 o n A u g u s t 15. Vertical lines r e p r e s e n t t h e effect of u n c e r t a i n t i e s in b e d t h e r m a l conductivities, w h i l e t h e horizontal lines indicate t h e 'best e s t i m a t e s , ' b a s e d o n a b e d s e d i m e n t porosity of 0 . 3 0 .  94  250  were not unusual, suggesting that the bed temperature gradient reversals apparent in Fig. 3.40 were related to the timing of the measurements during those days (rather than their occurrence on those particular days). Less than 5 mm precipitation fell during the preceding 24 hour periods, and minimum air temperatures were 3-4 °C on those mornings, compared to the August mean of 5.4 °C. Estimated conductive heat fluxes on those two mornings ranged from negligible (August 27) to negative (-15 W m" on August 18), based on the temperature 2  gradient between the creek and 5 cm depth.  30 m Downstream Creek T 224 m Downstream Creek T  A  5 cm Depth Bed T (PM) 10 c m D e p t h B e d T ( P M )  O  15-Aug  20 cm Depth Bed T (PM)  19-Aug  23-Aug  •  5 c m Depth Bed T (AM)  •  10 c m D e p t h B e d T ( A M )  •  20 c m Depth Bed T (AM)  27-Aug  31-Aug  4-Sep  Figure 3.40 C r e e k a n d b e d t e m p e r a t u r e s m e a s u r e d in B 3 . M e a n b e d t e m p e r a t u r e s a r e b a s e d o n all s e v e n s a m p l i n g sites s h o w n in F i g . 3.39.  95  3.4.8 Energy Exchanges Driven by Groundwater Inflow & Hyporheic Exchange To estimate the thermal effect of groundwater inputs using Eq. 2.1, groundwater inflow rates were calculated from the sequential streamflow measurements (Fig. 3.30), while groundwater temperatures were inferred from riparian soil temperatures. Soil temperatures at a 50 cm depth in the banks of B3 ranged from 7.7-8.3 °C (mean 8.0 °C, n=6), while 1 m depth bank temperatures ranged from 6.3-7.4 °C (mean 6.8 °C, n=8), when measurements were conducted on August 27. Mean temperatures at these two depths declined to 6.9 °C and 6.5 °C, respectively, on September 4, following a major rain and snow event on August 31 and September 1 (-40 mm). It is not clear which of these depths provides a better estimate of the temperature of inflowing groundwater, since both depths were below the water table in the banks. The effect of different groundwater inflow temperatures was explored by modelling the average daily creek temperature at B3LO for the July 17- August 31 period, based on two main assumptions:  1. the difference between average creek temperatures at the B R and L O sites during this period was entirely due to groundwater inflow, and,  2. net groundwater inflow observed along the B3 reach on August 16 and 27 (Fig. 3.30) represented the range of hydrologic conditions [i.e., (qns-qus)/qus] at B3 from July 17August 31 (after which streamflow and groundwater inputs increased by an unknown amount).  These assumptions are not entirely realistic. Other processes, particularly seasonal heat storage in the riparian zone, may also have caused average temperature differences between the two sites. Further, there are no hydrologic data available for B3 for most of the modelling period. The modelling exercise should therefore be considered heuristic. The model results indicate that groundwater inflow could result in a cooling of 0.4 to 1.0 °C in the average creek temperature at B3LO during July 17 to August 31, using the observed mean temperature (9.8 °C) at B3BR as the upstream temperature (Fig. 3.41). This amounts to a dT/dxcw value of -0.002 to -0.005 °C m" in the daily mean creek temperature. 1  96  On the afternoons of August 14 and 15, dT/dxcw was estimated at -0.004 to -0.006 °C m"  1  (i.e., combining the stream temperature data from August 14 and 15 with the streamflow data of August 16 - Fig. 3.35). If the creek gained groundwater from one bank and simultaneously lost flow to the other bank, this through-flow would have confounded the mixing model results. Bank-stream hydraulic gradients appeared conducive to this type of stream-groundwater interaction at a few sites along B3 (Fig. 3.32). However, there were no significant differences between left and right bank temperature data at the four cross-stream ground temperature transects to suggest migration of streamwater into the banks.  0.7  co 0.6 CO. D)  c o ro  1  0.5  CD  -*—<  CO  TD C  o  (3 9.2-  7.0  7.2  7.6  7.4  7.8  8.0  Groundwater Temperature (deg. C) F i g u r e 3 . 4 1 C o n t o u r plot o f p r e d i c t e d a v e r a g e daily c r e e k t e m p e r a t u r e s at B 3 L O f o r t h e July 1 7 - A u g u s t 31 period. G r o u n d w a t e r inflow r a n g e is b a s e d o n o b s e r v a t i o n s f r o m A u g u s t 16 a n d 2 7 , with culvert s t r e a m f l o w held c o n s t a n t at 1.1 L s" ( a s o b s e r v e d o n t h e t w o d a y s ) . A v e r a g e t e m p e r a t u r e at B 3 B R w a s 9 . 8 5 ° C f r o m July 17 to A u g u s t 3 1 . 1  97  The influence of hyporheic exchange on creek temperatures can be modelled crudely by assuming that the subsurface storage zones dominated transient storage (section 2.2.6). Data from B3 indicated mean temperature gradients of 0.1 °C cm" in the top 5 cm of the bed 1  on the afternoon of August 15, suggesting that the hyporheic zone was 0.15 °C cooler than the creek water (based on the mean d^y? value shown in Table 3.7). The model results indicate that the magnitude of the effect of hyporheic exchange varied by a factor of four between the three tracer test sub-reaches, from 9 x 10" to 3.7 x 10" °C m" (Table 3.7). The 4  3  1  lowest afternoon downstream cooling rate was modelled for sub-reach 1, due to its relatively shallow hyporheic zone and moderately slow exchange rate. The deeper, cooler, hyporheic exchange at sub-reach 2 (d YP H  =  5.5 cm) was overwhelmed by the faster exchange associated  with the higher a value of sub-reach 3. The concerns expressed in section 3.4.7 about the high degree of uncertainty surrounding the bed temperature gradients in B3 also apply to the calculated hyporheic fluxes, meaning that the cooling effect of hyporheic exchange may have been -50% higher than shown in Table 3.7.  3.4.9 Effects of Downstream Transport Processes on Temperature Patterns The effect of longitudinal dispersion was estimated by applying the OTIS model to simulate the downstream propagation of the diurnal temperature wave measured at B3BR. Ten-minute resolution creek temperature data from the B R Tidbit were used as boundary conditions in the OTIS model. Temperature was treated as a conservative solute and no mass or energy exchanges were included in the simulation. Longitudinal dispersion did not significantly affect the maximum daily temperature migration over the -200 m distance between B3BR and B3LO (Fig. 3.42). The maximum daily temperature modelled for L O using D=0.04 was <0.01 °C lower than the boundary conditions, while increasing D by a full order of magnitude (to 0.4 m s" ) resulted in a maximum temperature decrease of only 2  1  0.06 °C. It is, however, interesting that the daily temperature peak appeared to travel down the reach at a much slower speed than predicted on the basis of the velocities observed in the tracer test sub-reaches, perhaps indicating a delay caused by transient storage. In Fig. 3.42,  98  the peak temperature reached the Forest site (102 m downstream) at approximately the same time as it was predicted to arrive at the L O site (224 m downstream).  Figure 3.42 C r e e k t e m p e r a t u r e s o b s e r v e d at t h r e e sites a l o n g B 3 , a n d t e m p e r a t u r e s s i m u l a t e d f o r t h e f u r t h e s t d o w n s t r e a m site u s i n g t h e OTIS a d v e c t i o n - d i s p e r s i o n m o d e l .  Sub-reach scale variability in transient storage and hydrologic retention may have influenced temperature patterns. The longitudinal temperature gradient through sub-reach 2 was about twice as steep as observed in sub-reaches 1 and 3 on both August 14 and 15 (Fig. 3.35a). The ratio of A / A was substantially higher at sub-reach 2 than at the other subs  reaches (section 3.4.4). The modelled thermal effects of hyporheic exchange in the three sub-reaches do not correspond in any obvious way to the spatial temperature patterns observed in Fig. 3.35a since the greatest rate of cooling was calculated for sub-reach 3 (Table 3.7). This suggests that either the input data were erroneous due to insufficient measurement detail (e.g., use of a reach-averaged bed temperature gradient) or that the conceptual foundation of the model is incorrect.  99  3.4.10  Energy Balance Analysis of Downstream Cooling The afternoon of August 15 is the only period for which sufficient information is  available to estimate all of the major energy balance terms at B3. The warming effect of the energy exchanges across the water surface was more than offset by the cooling effects of the other three terms (Table 3.8). Groundwater inflow was the single most influential term, causing 0.9-1.2 °C, or about one-half, of the downstream cooling over the -200 m study reach (Fig. 3.35a). However, the combined influences of "heat exchanges with riparian sediments" (conduction plus hyporheic exchange) were more important at that time. The high end of the ranges for all the cooling mechanisms must be used to reproduce the observed decrease in maximum daily temperatures between the B R and L O sites (Table 3.8).  T a b l e 3.8 S u m m a r y of influences o n B 3 L O c r e e k t e m p e r a t u r e daily m a x i m u m o n A u g u s t 15, 2 0 0 0 , relative t o B 3 B R daily m a x i m u m . E s t i m a t e s a r e b a s e d o n t h e t w o - h o u r period s p a n n i n g t h e final o c c u r r e n c e of t h e daily m a x i m u m at B 3 B R (at 15:20) a n d t h e first o c c u r r e n c e of t h e m a x i m u m t e m p e r a t u r e at B 3 L O (at 1 7 : 2 0 ) . * N o t e that AT o n l y i n c l u d e s net radiation, a n d n o u n c e r t a i n t y is c o n s i d e r e d in o r d e r to e m p h a s i z e t h e e f f e c t s of u n c e r t a i n t i e s in other t e r m s . T h e u n c e r t a i n t y in t h e w a r m i n g e s t i m a t e d by net radiation a l o n e is ~ ± 0 . 2 ° C ( d u e to uncertainties in s t r e a m s u r f a c e a r e a ) , w h i l e t h e total uncertainty a s s o c i a t e d w i t h t h e s u r f a c e e x c h a n g e t e r m is at least ± 0 . 3 ° C . U  Component A  Low estimate  V 0.7  A  T  A G  W  -0.9  T  C  -0.7  AT  H y p  -0.5  £ A  T  -1.4  Z(A < T  0 )  observed  A  T  -2.1  -2.3 High estimate  3.4.11  0.7  -1.2  -1.1  -0.7  -2.3  -3.0  Temperature Modelling Based on Groundwater Inflow This section explores B3's thermal behaviour and its causes over longer time scales.  The working null hypothesis is that groundwater inflow is the dominant process controlling downstream temperature changes (Table 3.8). Creek temperatures at B3LO are therefore simulated using a simple mixing model of groundwater inflow. The model residuals are then related to meteorological variables, and examined for autocorrelation effects, to explore the influences of energy exchanges across the upper water surface and the bed. For example,  100  underestimation of the mean daily temperatures on hot, sunny, days would suggest that the Q u term had a significant, direct, impact on temperatures at B3LO (i.e., that the downstream warming effect caused by Q u was significantly greater on hot days). The first step in the modelling exercise was identifying optimal values of T G W and fractional groundwater inflow [(qos - q u s V q u s ] to simulate the mean daily creek temperatures at B3LO. A n average creek temperature of 9.3 °C was observed at B3LO during the July 17August 31 period. On the basis of the model results in Fig. 3.41, this suggests both a relatively warm groundwater inflow temperature (7.4-8.0 °C) and that mean hydrologic conditions were closer to those observed on August 16 (0.3 L s" inflow, or 21% of 1  downstream flow) than on August 27 (0.7 L s" inflow, or 56% of downstream flow). A n 1  average groundwater inflow of 0.4 L s" at 7.7 °C provides a reasonably accurate prediction 1  of mean creek temperature at B3LO on a daily basis through the July 17-August 31 period (Fig. 3.43a). This T w corresponds to the lower end of the range of 50 cm depth soil G  temperatures, while groundwater inflow is 0.1 L s" greater than indicated by the sequential 1  streamflow data of August 16. The assumptions underlying this exercise (particularly the first as outlined in section 3.4.8) are unrealistic; hence these "optimal values" of T G W and qG\v may be biased. It should also be noted that Fig. 3.43, and in particular 3.43a, overstates the explanatory power of groundwater inflow because the variability at B3BR is essentially duplicated in the model results. Appendix 4 compares the modelled downstream temperature change in each of the daily variables to the observed change. The modelled temperatures at B3LO are presented here because they are easier to interpret, and they provide qualitatively the same information While the model explained much of the variability in daily mean temperatures at B3LO (Fig. 3.43a), it generally underestimated the daily minima (Fig. 3.43b), and substantially overestimated the daily maxima (Fig. 3.43c). Surface energy exchanges do not help explain the relatively poor modelling of the diurnal temperature extremes (Fig. 3.43b,c). The model over-predicted creek temperatures during the daytime and early evening when Q u usually warmed the creek (section 3.4.6). Similarly, the model overestimated the daily minima at B3LO, but Q u often cooled the creek during the nighttime and early morning. It thus appears that energy exchanges across the water surface within the B3 reach were relatively unimportant, as suggested by the energy balance estimate for August 15.  101  o  11  o  E =3  10 9  CD  Q -o _CD  ~a3  8 7  T3  O  6 7  8  9  10  11  M e a s u r e d Daily M i n i m u m ( ° C )  O  13  .  9  8  10  11  12  M e a s u r e d Daily M a x i m u m ( ° C ) Figure 3.43 C r e e k t e m p e r a t u r e s m o d e l l e d f o r B 3 L O d u r i n g July 17 t o A u g u s t 31 u s i n g a s i m p l e g r o u n d w a t e r inflow m o d e l : a) daily m e a n s , b) daily m i n i m a , a n d c) daily m a x i m a . M o d e l l i n g u s e d E q . 2 . 1 , w i t h _ q =1.1 L s" , q s = 1.5 L s" , T = 7.7 ° C a n d o b s e r v e d t e m p e r a t u r e s at B 3 B R a s T . D a s h e d line a n d e q u a t i o n in a ) is a s i m p l e linear r e g r e s s i o n . 1  u s  D  u s  102  G  W  In contrast, temporal variations in bed temperature gradients suggest that inclusion of the bed heat conduction and hyporheic exchange terms would reduce the model residuals in Fig. 3.43 (if sufficient data were available). The bed acted as a heat sink for the creek water on warm afternoons, when the water cooled as it flowed downstream (Fig. 3.40). While no nighttime measurements of bed temperature are available from B3, the data from 9:30 PST on August 18 and 27 suggest that the bed was likely a source of heat for the creek water at earlier times on those days (Fig. 3.40). Further, since the magnitudes of the conductive and hyporheic terms depend on the magnitude of the bed temperature gradient, the higher gradients observed during afternoons are consistent with the greater magnitude of the maximum residuals than the minimum residuals (Fig. 3.43b,c). Statistical analysis of the residuals in relation to meteorology provides further insight into heat exchanges between the B3 channel and its riparian sediments. On days forming local peaks in the maximum air temperature time series, the groundwater inflow model overpredicted mean creek temperatures, while it under-predicted "in the valleys" (Fig. 3.44a,b). The majority (54%) of the unexplained variance in the predicted mean daily temperatures at B3LO is explained by daily anomalies in maximum air temperature, relative to the seven-day centered mean of this variable (Fig. 3.44c). The centered mean helps identify the local peaks and valleys in the air temperature time series (Fig. 3.44b). The maximum air temperature anomaly accounted for twice as much of the unexplained variance as maximum daily air temperature itself (Fig. 3.44c vs. 3.44d). The explanatory strength of this variable might be due to its ability to index changes in the difference between creek and riparian zone sediment temperatures (and thus the magnitude and net effect of the heat exchanges with riparian sediments over entire days). The predictive power of maximum daily air temperature likely improves when compared to recent (and future) weather because it then provides statistical evidence of the heat already (and yet to be) stored in the riparian zone. Fluctuations in groundwater temperature, as indexed by variations in creek temperature at B3HI, do not account for this pattern (Appendix 5). Inclusion of total daily incoming solar radiation in a multiple regression resulted in a modest (8%) increase in explanation of the mean daily residuals (Table 3.9), perhaps because of its strong influence on stream temperature. The daily minimum residuals are significantly (P<0.05) related to both the preceding day's mean air temperature and local maximum air temperature anomaly, although the signs  103  0.4  O o  CD \T CD CO 1  9g  I  •O CD = ^ CD O  "S -  n  1  O co CO -0.4  2  n i  i  i i r  n—i  i  i  i  I  i  i  i r  30 O  25  CD  20  "co CD  15  CL  § 10 H §  5  Daily M i n i m a  0  I—i—i—?—i  i—i—i—|—i—i  18-Jul  i i  26-Jul  i i  3-Aug  11-Aug  19-Aug  27-Aug  Date (year 2 0 0 0 ) 0.4 c  O  Y = 0.07X-0.02 r = 0.54  o  CD CO  T3  i  o°  0  o  CD  CD  O  —I CO  Q'  :> m -0.4 -6 Max.  -2  0  2  \T  c5> p  i CO TD CD = ^ CD O  o o o o -0-  -4  2  CD CO CO Q  o  4  O  CO  S  CQ  6  Daily A i r T - 7 D a y C e n t e r e d  Y = 0.02X - 0.44 r = 0.26  o  =^ CD O "2  j  o  OcD / c9 9) / ° o o ' -o °  CD "7  9  _0.4  cP  2  -0.4  o. O o o oo .R o o, OO'  o o o°©£h q o o o o oo o, o o o ,o -p-  8  12  16  20  24  28  Maximum Daily Air Temperature ( ° C )  A v e r a g e of M a x . Daily A i r T ( ° C )  Figure 3.44 A n a l y s i s o f residuals f r o m B 3 L O m e a n daily c r e e k t e m p e r a t u r e s s i m u l a t e d u s i n g g r o u n d w a t e r inflow m o d e l ( F i g . 3 . 4 3 a ) : (a) t i m e series of m o d e l r e s i d u a l s plotted a s a bar chart, (b) daily air t e m p e r a t u r e e x t r e m e s at t h e O p e n site f r o m July 17 t o A u g u s t 3 1 , (c) residuals v s . "local a n o m a l i e s " in t h e daily m a x i m u m air t e m p e r a t u r e s e r i e s f r o m t h e O p e n site, a n d (d) residuals v s . m a x i m u m daily air t e m p e r a t u r e (linear r e g r e s s i o n s a r e r e p o r t e d in  104  of the relations differ (Appendix 6). A multiple regression model based on the two variables explains 68% of the variance in the minimum residuals (Table 3.9, Fig. 3.45a). The inverse relation between the daily minimum residuals and the previous day's mean air temperature suggests a general tendency of greater heat release on mornings following warm days, perhaps due to a warmer riparian zone enhancing the heat source effect. In contrast, the positive effect of the previous day's peak air temperature anomaly on the minimum residuals suggests a smaller heat source effect on the mornings following relatively warm days. This effect can be linked to the tendency of creek temperature minima to be relatively warm on mornings following relatively warm days. Although the creek temperature minima at B R were not significantly related to the previous day's peak air temperature anomaly (r =0.05, P>0.05), there was a significant positive relation between the "local anomaly" in the creek temperature minima and T  _p (r =0.24, P<0.01). Thus, the negative relation between 2  m a x  an  minimum residuals and the previous day's mean air temperature is attributed to the effect of subsurface temperature variations, while the positive relation with the previous day's maximum air temperature anomaly can be linked to variations in creek temperature. Statistical evidence suggests that subsurface temperature variations were a less important influence on the daily maximum creek temperatures. The daily maximum residuals are most strongly related to variations in solar radiation (Fig. 3.45b), and including  T a b l e 3.9 Best s i m p l e (bivariate) a n d multiple (trivariate) linear r e g r e s s i o n s b e t w e e n daily m o d e l r e s i d u a l s f r o m B3 g r o u n d w a t e r inflow m o d e l l i n g ("RD"), daily air t e m p e r a t u r e ("T") v a r i a b l e s m e a s u r e d at t h e O p e n site, a n d total daily i n c o m i n g solar radiation ("Ki") m e a s u r e d at t h e O p e n site. V a r i a b l e s w e r e s e l e c t e d f o r testing of predictive s i g n i f i c a n c e b a s e d o n physical r e a s o n i n g a n d relatively high initial bivariate relations with t h e m o d e l l i n g r e s i d u a l s ( A p p e n d i x 6). S u b s c r i p t s : "mean" is daily m e a n , " " is daily m a x i m u m , " " is daily m i n i m u m , " " is t h e p r e v i o u s d a y ' s v a l u e , " " is t h e local a n o m a l y (day's v a l u e - s e v e n d a y c e n t e r e d a v e r a g e of the variable). m a x  m i n  p  ari  Modelling Residual Daily M e a n  Daily Min  Daily Max  r or R  Regression Equations  2  RD  m e a n  = 0.07T  RD  m e a n  = 0.058T  RD  m i n  = -0.06T  RD  m i n  = -0.08T  RD  m a x  = 0.08K^-0.55  RD  m a x  = 0.07'K^ + 0 . 0 6 T  105  - 0.02  m a x a n  m a x a n  m e a n p  m e a n p  0.54  + 0 . 0 0 9 K ^ - 0.173  0.34  +0.28 + 0.09T  0.62  m a x a n p  + 0.57  0.68 0.81  m a x a n  - 0.45  0.87  2  0.4  2  Multiple r e g r e s s i o n (Table 3.8, R = 0.68)  Y = 0.08X-0.55 r = 0.81 2  2  o  i.6  ofb  0 O  Q  or TD CD  1.2  •  o  I  -0.4  •  0.8  o  Q  -*—'  O a>  o  a:  -0.8  o  0.4 ,6°  •o  0  o  -1.2 -1.2  -0.8  -0.4 RD  m i n  0  -0.4  0.4  fC)  0  5 10 15 20 25 Ki ( M J m- day- ) 2  30  1  CZ O "co c  'E CD  -»—*  CD  Q  M—  O CD  o it  CD O O 0  1  2  3  4 5 6 7 L a g (Days)  8  9  10  Figure 3.45 Statistical analysis o f m o d e l r e s i d u a l s f r o m F i g . 3 . 4 3 : (a) p r e d i c t e d daily m i n i m u m t e m p e r a t u r e r e s i d u a l s at B 3 L O b a s e d o n multiple r e g r e s s i o n e q u a t i o n of T a b l e 3.9 a s a f u n c t i o n of t h e actual m o d e l r e s i d u a l s of F i g . 3.43b, (b) m o d e l residuals of Fig. 3.43c a s a f u n c t i o n of daily Ki, (c) a u t o c o r r e l a t i o n in m o d e l residuals ( R D ) , a n d daily m e a n air t e m p e r a t u r e ( T ) a n d total daily solar radiation ( K i ) at t h e O p e n site (all significant v a l u e s a r e positive c o r r e l a t i o n s ) . m e a n  the daily anomaly in maximum air temperature at the Open site in a multiple regression accounted for only an additional 6% of the unexplained variance (Table 3.9). Based on the previous inference that the maximum air temperature anomaly is an index of the difference between creek and subsurface temperatures over the course of an entire day, its limited explanatory power suggests that riparian sediment temperature fluctuations had a relatively  106  minor effect on the maximum residuals. The principal determinant of the daytime heat sink effect was likely the maximum daily stream temperature at the upstream (BR) site, since this temperature appeared both to be more variable, and to be consistently warmer, than the bed temperatures (Fig. 3.40). As a result, the positive association between the maximum modelling residuals and total daily solar radiation (Fig. 3.45b) might be mainly due to collinearity between maximum daily stream temperature at the B R site and K-l (r = 0.70, 2  plot and equation not shown). A l l three of the model residual time series are significantly autocorrelated to lags of one day, with the minimum residuals significantly autocorrelated to at least the five day lag (Fig. 3.45c). The pattern of autocorrelation in the minimum residuals is similar to the autocorrelation pattern in the mean daily air temperature, while the autocorrelation patterns for the mean and maximum residuals correspond more closely to that of solar radiation (Fig. 3.45c). Since daily bed temperature minima are apparently closer to the daily stream temperature minima than is the case for the daily maxima (Fig. 3.40), the daily minimum residuals may be more significantly affected by longer-term, lower amplitude sediment temperature fluctuations. In addition, air temperature has a stronger influence on stream temperatures during nighttime than daytime, due to the greater importance of longwave radiation at night (when solar radiation = 0). The extended autocorrelation in the air temperature time series is consequently more likely to be reflected in the daily minimum residuals. The correspondence between the autocorrelation pattern in the mean daily residuals and that in total daily solar radiation (Fig. 3.45c) suggests that daytime heat transfers were the main determinant of the net role of the heat storage effect over the entire day. However, the strength of the relation between the daily mean residual and the maximum air temperature anomaly (Fig. 3.44c) also suggests that the nighttime heat transfers and subsurface temperature variations were equally important influences. Generally, the B3 riparian zone appeared to act as a heat sink during the afternoon/evening and a heat source during the morning throughout the study period, as suggested by energy balance results for one afternoon and our limited bed temperature data. However, the results of our heuristic modelling exercise suggest that the heat exchanges also fluctuated on a daily basis, with the relative magnitudes of the daytime sink and nighttime source terms determining the net role of the riparian zone through the entire day.  107  3.5 Summary of B5 and B3 Results Stream temperatures along the B5 study reach were highly variable in space and time, creating three distinct temperature patterns. The upper sub-reach was characterized by streamflow losses, which caused the L O data-logger to be periodically disconnected from the warm streamwater flowing from the cut-block. Temperatures at the L O logger were dominated by locally emergent subsurface water during the dry weather associated with these disconnection events, until the L O site dried completely during a two-week drought. Energy exchange across the water surface (mainly net radiation) was the dominant control on the thermal regime of the upper "losing" sub-reach of B5, while groundwater inflow was generally the most important control in lower B5. Bed heat conduction and hyporheic exchange may also have been significant influences in lower B5. The relatively stable temperature patterns of the B3 study reach were linked to stream-subsurface energy exchanges conditioned by stable hydrology. Net groundwater inflow was observed through the reach, which may have caused up to one-half of the downstream cooling in the daily maximum temperature. Longitudinal patterns of creek temperature and electrical conductivity were not as well correlated in B3 as on pattern 2 days in lower B5, suggesting that processes other than groundwater inflow were important at B3. Two-way heat exchanges with riparian sediments (conduction and hyporheic exchange) appear to have heavily influenced both the daily maximum and daily minimum time series. A heuristic modelling exercise of the temperatures at B3LO based only on groundwater inflow suggested that heat storage within the B3 reach also affected the mean daily temperatures, depending on the relative strengths of the daytime heat sink and nighttime heat source effects. In Chapter 4, the following key issues serve as main points for discussion: 1. the relative importance of the various energy balance terms at the two reaches, and the broader significance of these findings 2. links between the hydrology and the thermal regimes of the study reaches 3. effects of upstream forestry activities on reach-scale stream temperatures and hydrology 4. relations between surface-subsurface water exchanges and thermal regime.  108  CHAPTER 4 DISCUSSION The relevance of the results presented in Chapter 3 for streams in other settings is considered in section 4.1, within a reach-scale energy balance framework. Section 4.2 relates the reach-scale results to catchment-scale hydrologic processes. Section 4.3 describes the interactions between the hydrologic and thermal regimes within the two study reaches. Finally, section 4.4 considers the management implications of the thesis results.  4.1 Energy Balance Considerations 4.1.1 Energy Exchanges across the Water Surface Both creeks B5 and B3 typically gained energy across their air-water interfaces as they flowed back into the forest on warm afternoons. Significantly negative Q E values were rarely observed in the study reaches at these times, because air vapour pressures were consistently higher than stream vapour pressures, except on a few relatively windy afternoons. The absence of significant turbulence over the streams likely facilitated accumulation of water vapour in the boundary layer, possibly from transpiration by understory species such as Devil's club. Brown et al. (1971) estimated that evaporation was cooling the stream ( Q was —100 W m" ) at their forested site. The different results of the 2  E  two studies probably reflect the influence of site-specific conditions, particularly the often dense understory vegetation at our site. Our meteorological sites may have been less than ideal from this perspective. We selected heavily-shaded sites in an attempt to maximize the possibility of finding meteorologically-driven "cooling sites." While this was an effective strategy in that radiative fluxes were small at the two Forest sites, it may have had the undesired effect of reducing the estimates of evaporation due to the sheltered conditions lowering turbulence and increasing local vapour pressures. However, the observations during this study are consistent with the assertion of Beschta et al. (1987) that "evaporative and convective transfers of energy are typically low for forested streams, because vapor pressure and temperature gradients close to the water surface are small and wind speeds are usually low." Since the latent heat fluxes did not appear to be significant, the atmospheric equilibrium temperature of the creeks in the forested sections of the study reaches was  109  roughly approximated by the air temperature within the boundary layer above the creek. This observation is important because the Baptiste creeks were cooler than others affected by forestry, perhaps due to the buffer strips which were left in place through the cutblocks. For instance, a similar stream studied by Keith et al. (1998) in coastal Alaska (location in Fig. 2.1) warmed to 21°C after flowing through -50 m open canopy segments. That creek temperature may have exceeded air temperatures within the closed canopy segments, but it is unlikely that heat losses from the water surface could account for the rapid cooling that Keith et al. (1998) observed in closed canopy segments (-5-6 °C in -60 m). The latent heat flux of -100 W m~ observed by Brown et al. (1971) at creek temperatures of 27 °C would have 2  caused a downstream cooling of <2°C through the 60 m closed canopy segments (applying Eq. 2.1 using a stream width of 2 m, and streamflow of 1.5 L s" for Eleven Creek, as 1  reported in Keith et al. 1998). The amount of solar radiation penetrating the forest canopy is a fundamental issue, since this is the greatest source of energy available to heat creek water. Beschta et al. (1987) suggested that solar radiation within shaded reaches is "likely to be greater than energy losses from" longwave or convective cooling (or even streambed conduction). Our data do not contradict their main assertion, but at our heavily-shaded Forest sites the magnitude of net radiative inputs were generally lower than the 140 to 200 W m" values at Brown et al.'s 2  (1971) forested site. We did not measure the spatial average of incoming solar radiation along the forested reaches. However, back-calculations of Q* required to produce the warming in upper B5 indicated that -10% of the incident solar radiation measured at the Open site may have reached the creek water in that sub-reach. Using a hand held densiometer, Mellina et al. (accepted) estimated the amount of stream shading provided by the riparian vegetation along seven stream reaches (mean bankfull width - 2 m) in the same biogeoclimatic zone as the Baptiste tributaries. Percent canopy cover averaged 77%, with a standard deviation of 10%. Our <j) value of 0.1 suggests 90% canopy cover for the upper subreach of B5, which is at the upper end of the range reported by Mellina et al. However, the <j) estimation was made for late afternoon/early evening while Mellina et al.'s observations indicate the percent canopy cover within two hours of solar noon. Net radiation data from both of the Forest meteorological sites indicate increased (j) values during mid-day, qualitatively in agreement with the measurements of Mellina et al. :  /  110  If advection of warm streamflow from upstream reaches is the main determinant of the daily maximum stream temperature, the timing of the daily maxima should be weakly related to local fluxes of Qu- For two relatively poorly-shaded stream reaches of 160 m and 655 m in Colorado and New Mexico, Constantz et al. (1994) reported that the timing of daily temperature extremes was identical at upstream and downstream sites. This suggested that temporal trends in local heat exchanges were important in determining the timing of stream temperature extremes in the two reaches (Constantz et al. 1994). In contrast, the daily maxima at the downstream sites in B3 and upper B5 were delayed relative to those at the BR sites. This suggests that the timing of the daily creek temperature maxima in the forested reaches was determined primarily by the timing of the daily maxima at the B R sites and the time required for the temperature pulses to travel downstream through B3 and upper B5. Edinger et al.'s (1968) theoretical analysis suggested that daily maxima should regularly occur at ~16:00-18:00 hours, due to water temperatures lagging peak inputs of solar radiation by up to 6 hours, with the length of the time lag positively related to stream depth. The temperature maxima at our B R sites occurred 1-2 hours earlier than this range, perhaps due to the low thermal inertia of the shallow streams. The relatively low variability in the timing of the daily maxima at both B R sites (standard deviations of ~1 h) suggested regular timing of maximal inputs of heat, consistent with a local solar influence. This local solar influence was particularly apparent in the B3 road right-of-way, where afternoon creek temperatures increased by 1-1.5 °C through the first 30 m of the study reach. Brown et al. (1971) observed even faster rates of stream warming near roads (up to 7 °C in 46 m).  4.1.2 Bed Heat Conduction Streambed heat conduction is often regarded as a minor energy balance term, or has sometimes even been ignored entirely (e.g., LeBlanc et al. 1997). Results from the Baptiste Creek tributaries indicate that conduction can be an important energy balance component in forested reaches. Daytime estimates of conductive heat transfer from stream to bed ranged from about -300 to +20 W m" at points within B5. At B3, bed heat conduction ranged from 2  about -50 to -10 W m on one afternoon, while smaller positive values were inferred at 2  nighttime. These diurnal patterns at B3 are consistent with observations at other sites of  111  nighttime reversals in bed temperature gradients (e.g., Crisp 1990) and bed heat conduction fluxes (Webb and Zhang 1999). In general, the magnitude of conductive heat losses during daytime was similar to inputs of heat across the water surface in our forested reaches. Other studies (e.g., Hondzo and Stefan 1994; Evans et al. 1998; Webb and Zhang 1999) have reported similar bed heat conduction fluxes, but they were not of such high relative importance at those sites because inputs of radiant energy were generally larger. The magnitude of the conductive heat flux was related to hydrological conditions within the streambed at multiple spatial scales. Streamwater infiltrated rapidly into the bed in upper B5, causing weak or negligible bed temperature gradients and conductive fluxes. The largest conductive transfers of heat from stream to bed occurred in the groundwaterinfluenced lower B5 sub-reach, although streamwater penetration into the subsurface caused weaker bed temperature gradients at some sites. At the finest spatial scale, bed heat fluxes tended to be more negative in pools than in riffles or steps in B5. This tendency was likely related to enhanced streamwater downwelling through steps and riffles (White et al. 1987; Evans and Petts 1997), due to local increases in negative hydraulic gradients, although variations in hydraulic or thermal conductivity cannot be ruled out. Clark et al. (1999) found that bed temperature patterns associated with enhanced downwelling of water are not always observed at riffles. Bed temperature patterns at B3 were not documented as well as at B5, particularly for the finest spatial scale. However, bed heat conductive fluxes at B3 varied spatially, with a coefficient of variation of 0.52 on the afternoon of August 15. Measuring most of the bed temperature gradients in steps or riffles may have led to underestimation of average bed heat conductive fluxes at B3. These observations raise important questions about sampling schemes for reach-scale estimates of bed conduction. Shepherd et al. (1986) noted the contrast in bed thermal regimes between downwelling and upwelling zones at three coastal sites in British Columbia and Alaska, but they did not recognize the potential implications for streamwater temperatures. Evans et al. (1998) considered spatial variability in thermal gradients, but based their main conduction estimates on bed temperatures measured at the head of a riffle, where "groundwater was least likely to be encountered." It is not clear that areas of groundwater influence should be avoided unless their cumulative area makes up a negligible portion of total reach area. Spatial variations in temperature gradients within the cold pool at  112  165 m in B5 indicated that the area of anomalously high (up to 300 W m" ) conductive losses 2  was probably negligible, but groundwater influenced bed temperatures throughout the lower sub-reach. Brown's (1969) conclusion that conduction was an important energy budget term where streams have solid rock bottoms, but that "gravel bottoms seem to be insignificant energy sinks," was influenced by the low thermal conductivity value he used for gravel. Brown's (1969) thermal conductivity units (BTU/ft -inch-min-°F) were not dimensionally 2  homogeneous, but SI equivalents of his values can be estimated using the thermal conductivity reported for solid green breccia by Brown (1972). This transformation suggests that the thermal conductivity value Brown (1969) used for gravel was 0.4 W m" K" 1  1  (Appendix 7a), which is less than one-half of the lowest thermal conductivity value reported in recent studies. It is not clear how Brown (1969) estimated this value, but it appears that he used a value for dry gravel, since the thermal conductivity of a dry sandy soil with porosity of 0.4 is 0.30 W m" K" (Oke 1987, p. 44). Measurements of in-situ streambed thermal 1  1  conductivity using instrumentation such as that of Land and Paull (2001) should be made in the future, since a wide range of Kc values have been estimated in the recent literature. The nature of the specific hydrologic environments where pioneering studies of small stream energy balances were conducted also may have contributed to an underestimation of the importance of bed heat conduction in later studies (e.g., LeBlanc et al. 1997). Brown (1969) chose study reaches where groundwater inflow was negligible. Bed temperature gradients were not reported, and he attributed his findings of insignificant conduction in gravel streambeds to the small magnitude of his estimate of gravel's thermal conductivity. However, Brown (1972) described the temperature gradients observed in the top 20 cm of his gravel-bed study stream as isothermal. Maximum temperature differences between the streamwater and the 5 and 20 cm depths were 0.05 °C and 1.1 °C, respectively (Brown 1972). In the groundwater-influenced lower B5 sub-reach, the 5 cm depth temperatures were often 0.5-1 °C cooler than stream temperature (e.g. at the 183 m site or the 164 m riffle), and were greater than 5 °C cooler than stream temperature at the focused site of groundwater discharge at 165 m. Average bed temperature gradients observed in the upper 5 cm of the more weakly groundwater-influenced B3 streambed on August 15 were ten-times greater than those  113  Brown (1972) observed. Even in the strongly losing upper B5 sub-reach, temperature gradients in the streambed at certain sites were sometimes as great as Brown's. In contrast to Brown's weak bed temperature gradients, Ringler and Hall (1975) observed temperature differences of up to 3.5 °C between the stream and 5 cm depth in an entirely clear-cut catchment of the Alsea Basin in Oregon. Comer and Grenney (1977) reported bed heat conduction fluxes of -230 to -700 W m" in a groundwater-influenced reach 2  of a Utah creek, although these may be overestimates by a factor of about three (Appendix 7b). The bed temperature gradients observed in the gaining reach of Silliman and Booth 2  1  (1993) would have caused Qc fluxes of -60 to -120 W m" , assuming a Kc range of 1-2 W m " ° C (Silliman et al. (1995) assumed a value of 0.9 W m" ° C ) . These ranges are similar to 1  1  1  values observed in this study. Nevertheless, the highest bed heat fluxes observed in our study (up to -300 W m" at the focused groundwater discharge site at 165 m in B5) do not 2  necessarily represent an upper bound on the magnitude of this term. Since the stream temperatures observed in the Baptiste Creek tributaries were relatively low, conductive fluxes might cause a greater cooling effect where warmer streamwater flows into a groundwater-influenced reach. For instance, bed heat conduction could have caused the rapid downstream cooling observed by Keith et al. (1998) at Eleven Creek on sunny days. Streamflow at a temperature of 21 °C flowed into closed-canopy sections where the streambed may have been influenced by groundwater of ~10 °C, since mean annual air temperature at Prince of Wales Island, Alaska is ~7 °C ( N O A A 2002, location of Prince of Wales Island is shown in Fig. 2.1). Hence, bed heat fluxes of -300 W m" might have been more spatially extensive than 2  observed in lower B5 where streamwater temperatures of up to 11 °C flowed across a streambed influenced by 5-8 °C groundwater. Applying a flux of -300 W m" in Eq. 2.1 at 2  Eleven Creek gives a downstream cooling of 5.7 °C in 60 m, compared to the 5-6 °C observed. This estimate, and the importance of bed heat conduction in the energy balance of the 200 m B3 reach, contradict Beschta et al.'s (1987) suggestion that conduction does not need to be included in energy balances of reaches shorter than 1000 m. Bed heat conduction is probably a particularly important term in shaded reaches downstream of open environments, since warm streamwater flows across shaded sediments. The bed sediments of shallow streams in clear-cuts are likely heated by solar radiation, although theoretical  114 \  considerations suggest that most of this heat is rapidly transferred to the streamwater (Adams 1999). Solar heating of bed sediments in clearings may partly explain why bed heat conduction has sometimes been successfully ignored when modelling the downstream temperature changes of deforested creeks (e.g., Hetrick et al. 1998). Conduction may also be relatively less important in open settings where radiative fluxes are greater.  4.1.3 Groundwater Inflow The important volumetric balance between local groundwater inflow and advection of streamflow from upstream has long been understood (e.g., Brown et al. 1971), but the data from B5 provided a particularly clear example of its role in determining reach-scale thermal regimes. The highly variable nature of the hydrology in the B5 reach caused the development of three distinct stream temperature patterns (not including "pattern 4" which was associated with stream channel drying, and hence measurements of air temperature). At the "dry extreme," local inflow of subsurface water entirely controlled B5LO temperatures when the warm streamwater entering the reach was completely consumed by infiltration in upper B5 (pattern 3 days). At the "wet extreme," streamflow at the upper end of the B5 reach was high enough to prevent any obvious cooling between the BR and L O sites (temperature pattern 1 days). In contrast, the B3 reach was characterized by relatively stable hydrology and groundwater inflow consistently caused about one-half of the observed downstream cooling in the daily maxima. Reach-averaged groundwater inputs on August 16, after a two week drought, were 0.0014 L s" m" or 0.1% of downstream flow per meter. In comparison to 1  1  other sites, this rate of groundwater inflow is not particularly high, but in relative terms its contribution to streamflow is substantial. For example, Cey et al. (1998) observed groundwater inputs one order of magnitude higher (-0.01 L s" m" ) at baseflow along a 1  1  450 m reach in south-western Ontario but because downstream flow was about 8 L s"  1  groundwater inflow also contributed -0.1% flow irf'. Similarly, groundwater inflow along 11 stream reaches within the same biogeoclimatic zone as the Baptiste Creek tributaries averaged 0.004 L s" m" , or about three times higher than in B3, but this contributed only 1  1  0.03% flow m" because downstream flow averaged 14 L s" (Mellina et al. accepted). In 1  1  115  contrast, during winter low flow at six first-order forested reaches of -20 m in the Coweeta Hydrologic Laboratory of North Carolina, groundwater inflow averaged 0.016 L s" rrf or 1  1  ~1% flow rrf because streamflow was only 1 L s" (D'Angelo et al. 1993). 1  1  High rates of groundwater discharge into small streamflows can cause rapid rates of downstream cooling i f stream temperatures have been elevated above those of groundwater. For example, groundwater inflow of 0.01 L s" rrf at 10 °C would be sufficient to cause 1  1  much of the rapid cooling that Keith et al. (1998) observed at Eleven Creek, where stream temperatures exceeded 20 °C. Mixing a 0.5 L s" groundwater inflow into an initial 1 L s" 1  1  flow at 20 °C over 50 m results in downstream flow of 1.5 L s" at a temperature of 16.7 °C, 1  compared to observed downstream temperatures of ~15 °C. It should be noted that if this groundwater inflow of -0.01 L s" rrf (0.7% flow rrf ) occurred along that stream, Hetrick et 1  1  1  al. (1998) probably would not have been able to model successfully the warming observed in the three open-canopy sections without quantifying that term. A l l three open canopy sections were included in Hetrick et al.'s modelling, and it seems unlikely that substantial groundwater inflow only occurred in the two closed canopy sections where rapid cooling occurred (three open-canopy sections alternated with upstream closed-canopy sections). . This discussion has heretofore emphasized the influences of three variables on the thermal effect of groundwater inflow: streamflow magnitude, groundwater discharge, and stream temperature, all of which are known to vary in space and time (e.g., see Genereux et al. 1993; Sinokrot and Stefan 1993). It is often supposed that the fourth relevant variable, the temperature of inflowing groundwater, is relatively constant and can be accurately approximated as mean annual air temperature (+1 or 2 °C). However, results from the Baptiste creek tributaries show that this presumption will not provide accurate estimates of stream temperatures in all situations. The temperature of inflowing groundwater was approximately as low as predicted by conventional theory at only one location (-5 °C at 165 m in B5), whereas it appeared to be closer to 8 °C throughout the rest of the study reaches. Few measurements of shallow groundwater temperatures are available, but Mellina et al. (accepted) observed a similar range of temperatures at four seepages monitored from June 1 to August 31, 2001 along three streams in the nearby Nation River drainage (-50 km east of the Baptiste tributaries). One site averaged 4.6 °C, while the mean temperatures of the other three seepages ranged from 7.2 to 7.9 °C (Mellina et al.). These temperatures are slightly  116  cooler than those observed in the Baptiste tributaries during the 2000 study period, but differences in the timing of the two monitoring periods may be responsible. In 2001, Mellina et al.'s mean daily seepage temperatures fluctuated at -7.5 °C during the mid-July to end of August timeframe of our study period, after warming from - 4 °C in early June. Similarly, Gaffield (2000) observed that the temperature of groundwater discharging to his study streams in Wisconsin warmed by up to 3 °C between May and September. Underestimating the temperature of groundwater inflow by 3 °C could result in overestimation of downstream cooling. Temperature of deep groundwater (water table depth > -10 m) may be relatively constant over the year, but shallower groundwater can exhibit temporal variability at a variety of timescales, depending on aquifer characteristics (Bundschuh 1993) In cases where creek temperatures are close to those of groundwater, such as was often the case in lower B5, the direction of downstream temperature changes may depend on spatial variations in the temperature of inflowing subsurface water. For example, using only a mixing model of groundwater inflow (e.g., Brown et al. 1971), Mellina et al. were able to predict correctly downstream warming of-1 °C in daily mean temperature through a 500 m reach of a forested headwater creek (TCH C C , unharvested). This counter-intuitive result occurred because the mean upstream temperature (4.3 °C) was 2.5 °C cooler than the estimated temperature of groundwater inflow, which contributed 44% of the downstream flow (based on one set of sequential streamflow measurements and T w as an average of the G  four seepages described above). Mellina et al's model of groundwater inflow predicted downstream changes in mean daily temperature at 11 stream reaches reasonably accurately (average absolute deviation of 0.3 °C, range of observed downstream temperature changes from -+1 to -2.5 °C). However, they cautioned that the importance of other energy balance components could not be completely ruled out based on their evidence. The complex hydrology of these small streams is best characterized through the use of multiple approaches, including streamflow measurements along the reach, monitoring of hydraulic gradients in the riparian zone and tracer tests (Harvey and Wagner 2000). However, time constraints too often preclude the application of multiple methods, necessitating the identification of simple, robust approaches that provide maximal information. Comparing longitudinal profiles of streamwater electrical conductivity and temperature can be a powerful tool for assessing the effect of groundwater inflow on stream  117  temperature. A surprising finding in our study reaches was the contrast between the electrical conductivity of stream and groundwater. Groundwater is commonly observed to have higher electrical conductivity than surface water (e.g., Geist 2000), but discharge of subsurface water into both study reaches was associated with declines in creek water electrical conductivity. Comparison of streamflow and E C patterns suggests that groundwater electrical conductivity was consistently about 10% lower than that of creek water in both reaches, despite the fact that B3 electrical conductivity values were at least twice those in the B5 study reach (-400 vs. <200 uS cm" ). It thus appears that the gross 1  chemical composition of subsurface water in these small catchments varied only subtly between upstream sources and the study reaches. The shallower topographic gradients within the study reaches might cause groundwater to travel through slightly shallower and shorter flowpaths than further upstream, resulting in lower solute concentrations in subsurface groundwater within the study reaches. Alternatively, the differences may be due to spatial variations in regolith geochemistry. Investigating links between the chemical and thermal signatures of subsurface water would facilitate a deeper understanding of the hydrological influence on stream temperature.  4.1.4 Hyporheic Exchange This study did not clearly demonstrate that hyporheic flow influences stream temperature, but it did reveal a diverse range of thermal interactions between the hyporheic zone and the stream channel, and provided some evidence of how these interactions are conditioned by a stream's hydrologic environment. The B5 reach could be broadly separated into upper "losing" and lower "gaining" sub-reaches. Bed temperature patterns in the upper sub-reach of B5 indicated that streamwater rapidly infiltrated to depths of 20 cm, but there was no evidence that streamwater flowed through the subsurface and returned to the stream. Although some investigators have characterized the "hyporheic zone" based on similar thermal patterns (e.g., Malard et al. 2001), Harvey and Wagner (2000) questioned whether hyporheic zones exist in losing streams. In the lower sub-reach of B5, the hyporheic zone ranged from depths of 5 cm or less at sites with positive bed hydraulic gradients (165 m and 183 m), to up to at least 20 cm at a site (199 m) where negative hydraulic gradients were  118  observed. These depths are roughly consistent with the 0-20 cm range estimated using conservative tracer techniques in other groundwater-influenced streams (Hill and Lymburner 1998; Harvey and Fuller 1998). The estimated depth at the 199 m site may have been deeper than 20 cm if the hydrological definition of the base of the hyporheic zone (>10% channel water) had been evaluated rigorously using a chemical mixing model. This evidence from lower B5 indicates vertical migration of streamwater into the groundwater-influenced subsurface and potential thermal effects driven by hyporheic exchange. Vertical hyporheic exchange is likely an important control on stream temperatures in certain hydrologic environments, but lateral hyporheic exchange and its relation to the broader hydrologic setting must also be considered. Soil temperatures and hydraulic gradients measured in the banks of lower B5 suggested a lateral hyporheic flowpath caused by the stream flowing roughly perpendicular to the down-valley groundwater flow, as in the scenario described by Huggenberger et al. (1998). In sub-reach 2 of B3, lateral hyporheic flowpaths may also have developed due to channel meandering (e.g., Wroblicky et al. 1998) or the existence of multiple channels (e.g., Kasahara and Wondzell 2001). Data from one afternoon indicated that the observed downstream cooling in the B3 study reach could not be explained using only the traditional energy balance terms, suggesting the possible thermal influence of hyporheic exchange. A relatively high rate of downstream cooling in sub-reach 2 corresponded to high measures of transient storage in that section of the reach. The modelled depth of the hyporheic zone in B3 ranged from 1 cm in the first sub-reach to 6 cm in the second sub-reach, roughly consistent with bed temperature observations at those sites in lower B5 with positive hydraulic gradients. The ratio As/A, which has been shown to be correlated with nutrient retention and uptake in small streams (Valett et al. 1996; Mulholland et al. 1997), was about three times greater for sub-reach 2 than for the other two sub-reaches. The As/A ratio for sub-reach 2 of 0.53 is not exceptionally large relative to other studies, suggesting that greater  QHYP  effects might be  expected in other settings. The highest reported ratios have been in the range of five (e.g., Gallina Creek in Morrice et al. 1997), although most (-75%) have fallen below 0.5 (Harvey and Wagner 2000). Harvey and Wagner (2000) emphasized that the stream-tracer approach used to estimate these parameter values only detects relatively short timescales of hyporheic exchange. They also asserted that water chemistry is most likely to be influenced by  119  hyporheic exchange at the timescales detected by tracer tests, but it remains to be seen whether this is the case for stream temperature. Harvey and Wagner (2000) found that As/A covaries with a measure of channel roughness known as the channel friction factor, which is directly related to stream depth and streambed slope and indirectly related to stream velocity. Their relation should provide an effective starting point for estimating the thermal effect of transient storage in particular streams. However, the variation in As/A values from the three B3 sub-reaches indicates that Harvey and Wagner's relation cannot account for the relatively subtle variations in channel structure that influence transient storage at the within-reach scale, and may be important in determining longitudinal temperature patterns. Stream depth, bed slope and stream velocity appeared to vary only slightly between the three sub-reaches, yet As/A varied by a factor of three. This amount of unexplained scatter is typical of Harvey and Wagner's relation, in which plotted values of A / A vary over about one order of magnitude at particular values of s  channel friction factor. The A / A values observed in the B3 tracer test sub-reaches are s  predicted reasonably well by their relation, based on an estimated channel friction factor of nine for all three sub-reaches. Sub-reaches 1 and 3 plot on the low side of the relation, while sub-reach 2 falls almost exactly where a line of best fit would appear. Harvey and Wagner's (2000) relation should be useful for predicting how transient storage varies at broad spatial scales, for example from headwater to valley bottom streams. The greater morphologic complexity of sub-reach 2 suggests that visual clues such as woody debris might be used to identify sites with high hyporheic influence at finer spatial scales. The simple model used to quantify the thermal effects of hyporheic flow also suggested that the rate of exchange between the hyporheic zone and channel might be as important as the relative size of the transient storage area. The transient storage exchange coefficients (a) estimated here were at least one order of magnitude higher than estimated at baseflow in similar sized streams by Morrice et al. (1997), but similar to those estimated by D'Angelo et al. (1993) at their first order sites. The higher a value of sub-reach 3 resulted in a higher estimate of hyporheic cooling than that for sub-reach 2, where a larger transient storage zone was estimated. These results should not be over-interpreted, since several dubious assumptions were made in the modelling. For instance, no attempt was made to account for spatial variations in bed temperature gradients among the three sub-reaches. If  120  the hyporheic zone of sub-reach 2 cooled more rapidly with depth than in the other subreaches, substantially greater hyporheic cooling could have taken place therein. This scenario appears somewhat unlikely, however, since observations in B5 suggest that zones of deeper hyporheic flow are associated with smaller bed temperature gradients. This apparent contradiction highlights our poor understanding of the hyporheic influence on stream temperature. The relative thermal effects of deep or shallow, and of slow or fast, exchange are unclear. It is clear, however, that measurements of net groundwater inflow do not necessarily provide a full indication of the interaction between channel and subsurface waters. Inverse modelling of stream tracer data using OTIS-P is just one of several ways to quantify this interaction (Harvey and Wagner 2000). Future research on the thermal effects of hyporheic exchange should employ a wider range of field techniques. The stream-tracer approach could be combined with measurements of chemical tracer penetration into the subsurface to verify that the different perspectives of hyporheic storage area are mutually consistent. These comparisons are not necessarily straightforward because, as Harvey and Wagner (2000) pointed out, their simple approximations of hyporheic zone depths do not account for the mixing between stream and groundwater in the subsurface (the source of all water in the hyporheic zone is assumed to be the stream). Thus, the hyporheic zone depth delineated using the >10% streamwater criterion is not truly comparable to the modelled storage zone depth (d Yp) calculated from the OTISH  P estimate of transient storage area (e.g., Harvey and Fuller 1998). This clarification highlights the need to develop a deeper understanding of the exchanges of energy and mass that take place at the centimeter- or sub-centimeter-scale between streamwater, groundwater, and hyporheic sediments. Laboratory experimentation (e.g., Vaux 1968; Thibodeaux and Boyle 1987) could prove a valuable tool for advancing this understanding. Further application of the relatively simple field techniques used here could also help clarify the specifics of thermal "bed processes." Our understanding of the relative roles of hyporheic exchange and bed heat conduction at B3 could be refined with more spatially- and temporally-intensive bed temperature measurements. For instance, i f continuous time series of bed temperatures were recorded at B3, the relative influences of conduction and advection on the thermal properties of the hyporheic zone could be assessed using modelling approaches (e.g., Silliman et al. 1995; Land and Paull 2001).  121  4.2 Catchment-Scale Interactions between Hydrologic & Thermal Regimes Section 4.1.3 emphasized that the balance between streamflow inputs from upstream and local groundwater inflow is a fundamental control on the thermal regimes of these small streams. Influences of the hydrologic regimes of the B3 and B5 catchments upstream of the study reaches will be examined in section 4.2.1. Section 4.2.2 considers the possibility that variations in the temperature of creek water flowing from upstream catchments might significantly alter the hydrology of particular reaches.  4.2.1 Influence of Catchment Hydrology on Stream Temperature Differences in mean regolith depth between the two catchments may have influenced their hydrologic regimes. Bedrock was visible in the road cut immediately upstream of the B5 study reach, suggesting till depths of ~1 m. In contrast, no bedrock was visible at the road cut in the B3 catchment. Multiple lines of evidence suggest that this apparent variability in till depth between the two catchments caused important differences in their hydrologic regimes. First, the snowmelt hydrograph data of Beaudry (2001) indicate subdued snowmelt freshets at B3 relative to B5. Peak specific streamflow during the snowmelt period averaged about half as high at the B3 flume (75 L s" km" ) as that at the B5 flume (145 L s" km" ) 1  2  1  2  from 1996 to 2000 (Beaudry 2001). This relation was evident during the pre-harvest year (1996), as well as post-harvest years, indicating that forest harvesting was not the cause of different hydrologic responses to snowmelt in the two catchments. Second, both specific and absolute streamflow at the B3 culvert were higher than those at the B5 culvert following the two-week drought of August, 2000. Streamflow of 1.1 L s" (2.6 L s" km" ) was measured at the B3 culvert on August 16, 2000, compared to 0.7 L 1  1  2  s" (0.5 L s" km" ) at the B5 culvert. A considerable portion of snowmelt water is probably 1  1  2  stored within the deeper B3 regolith and slowly released as summer baseflow, causing the higher minimum baseflows from the B3 catchment. This hypothesis is supported by the higher electrical conductivity values of the B3 catchment (often > 400 uS cm"' during the study period) compared to the B5 catchment (always <200 uS cm" during the study period). 1  The greater concentration of solutes in B3 streamwater likely reflects the longer residence times of water within the catchment subsurface due to deeper flowpaths. This is also  122  consistent with the subdued hydrologic response of the B3 catchment to the rain events of 1  2  late August, 2000. Specific streamflow from the B3 catchment increased by 0.5 L s" km" , while that from the B5 catchment increased by 2.2 L s" km" , in response to the 62 mm of 1  2  precipitation that fell from August 16 to 28. These hydrologic characteristics have important implications for temperature patterns within the reaches downstream of the cutblocks, from the streamwater side of the streamgroundwater balance. For example, the flashier hydrologic regime of the B5 catchment plays a role in creating the discontinuous streamflow patterns in the B5 study reach. If the catchment upstream of the B5 study reach had a seasonal hydrologic regime similar to the B3 catchment, discontinuous streamflow in the B5 reach might be a less common occurrence. Conversely, the quick hydrologic response of B5 sometimes allows continuous flow to be restored in the B5 reach after small rain events (~10 mm events were sufficient in July, 2000). The high temporal variability in the thermal regime of the B5 study reach was consequently linked to the flashy hydrology of the upstream catchment. In comparison, the subdued hydrologic response of the B3 catchment was a key factor in creating the temporally consistent temperature patterns observed there. The shallow regolith of the B5 catchment could also be a partial cause of the streamflow losses in the upper section of the B5 reach. The water table within the upper B5 sub-reach may be dominantly recharged by local sources (i.e., by infiltrating streamwater and precipitation) since the upslope regolith is too shallow, and the bedrock is likely not permeable enough, to allow significant groundwater flow. The visible bedrock at the B5 road cut suggests that groundwater flow between the upstream catchment and the study reach is minimal. An abrupt transition to deeper sediments at the head of the B5 study reach would explain the rapid streamflow losses in upper B5, similar to Tijeras Arroyo, New Mexico (Constantz and Thomas 1997), Vicee Canyon, Nevada, (Ronan et al. 1998), and the "perennial transition zone" of Amargosa River, Nevada, described by Constantz et al. (2001).  4.2.2 Effects of Stream Temperature on Reach-scale Hydrology Constantz et al. (1994) demonstrated that diurnal variations in creek temperature can cause diurnal variations in streambed infiltration at losing reaches. The inverse relation  123  between water temperature and viscosity causes the hydraulic conductivity of saturated sediments to double between 0 and 25 °C. Constantz et al. (1994) measured diurnal variations in streamflow losses of 26-27% at streams in Colorado and New Mexico with diurnal temperature variations of 14-15 °C. The diurnal creek temperature variations at the B5 study reach were much smaller, rarely exceeding 3 °C. The rate of hydraulic conductivity increase with temperature is ~3% °C through the relevant range of temperatures (10-15 °C), _1  suggesting diurnal variations in streamflow losses of up to 9% in upper B5, although Constantz et al.'s results would suggest smaller diurnal variations (-5%). Significant diurnal variations in streamflow extent (15-20 m) were observed on several days in upper B5, but these fluctuations could also be due to diurnal variations in streamflow inputs from the upstream catchment caused by enhanced daytime evapotranspiration (e.g., Bren 1997). Evapotranspiration within the riparian zone of the upper B5 sub-reach might also have driven diurnal variations in streamflow (Bond et al. 2002). A more relevant issue in this case is the previously unexplored possibility that land use changes might enhance streamflow losses through increased stream temperatures. Since no formal comparison of pre- versus post-harvesting temperatures in B5 has been published, it is difficult to assess the magnitude of the potential effect. However, harvesting forests in the B5 watershed may have increased creek temperatures by ~3 °C (H. Herunter, DFO, personal communication). Consequently, this suggests increased streamflow losses of up to 9% in upper B5 due to logging, indicating that discontinuous streamflow along the reach may have become more common after harvesting. The threshold streamflow value (-4.5 L s" ) for 1  discontinuity might have been slightly lower prior to harvesting. Decreasing infiltration rates in upper B5 by 9% suggests a pre-harvesting threshold streamflow of 4.1 L s " . This lower 1  threshold would have delayed the onset of the three major streamflow disconnection events observed during the first two segments of the study period by ~1 day each time (the slope of the streamflow recessions through the range of 1 -5 L s" was -0.5 L s" day"'). In fact, 1  evidence from the 2000 study period weakly supports this relation between creek temperature and streamflow continuity. Streamflow was continuous along the B5 reach more often than would be predicted by the 4.5 L s" threshold from August 18 to 31. Mean creek 1  temperatures at B5BR during that period were 1.6 °C cooler than during the July 17- August 2 period, when the 4.5 L s" threshold was more consistent. 1  124  The spatially-averaged infiltration flux in upper B5 was calculated at 10 cm hr" based 1  on streamflow measurements. This is on the high side of the range of streamflow losses observed at other sites, but it is not exceptional. Constantz et al. (1994) measured average infiltration fluxes of about 6-8 cm hr" at a 160 m reach of St. Kevin Gulch, Colorado. At a 1  1 km study reach of the Tijeras Arroyo in New Mexico, estimated infiltration rates ranged from 6-25 cm hr" , with a spatial mean of 11 cm hr" (Constantz and Thomas 1997). Ronan et 1  1  al. (1998) measured infiltration rates of 2-14 cm hr" at Vicee Canyon, Nevada. While most 1  of the research on the effects of creek temperature variations on streamflow losses has been done in arid or semi-arid climates, losing reaches have been reported in humid climates. Huggenberger et al. (1998) referred to a study of perialpine rivers in Switzerland that reported infiltration rates of up to 4 cm hr" . Genereux et al. (1993) observed two reaches 1  with ephemeral flow situated downstream of reaches with perennial flow in the 0.6 km East Fork catchment of Walker Branch, Tennessee, where annual precipitation is 1400 mm. Gburek and Folmar (1999) estimated a streamflow decrease from 11.9 to 8.6 L s" in 500 m, 1  and measured a streamflow decrease from 7.3 to 3.2 L s" in 1000 m, in two upper reaches of 1  a 7.3 k m catchment in Pennsylvania, where annual precipitation is 1100 mm. These studies 2  demonstrate that losing conditions similar to those encountered in the upper sub-reach of B5 have been observed in other humid environments, although they may be relatively uncommon. Enhanced streamflow losses can be expected i f stream temperatures increase substantially following forest harvesting or other perturbations.  4.3 Reach-Scale Linkages between Hydrologic & Thermal Regimes 4.3.1 Hydrologic Contrasts Between the Two Study Reaches Spatial variability in stream-subsurface interactions was high along the B5 reach. Streamflow rapidly infiltrated into the streambed in the upper 150 m of B5, while cool groundwater often interacted closely with the stream channel in the lower 60 m of the reach. Although measurements were not as detailed in the B3 reach, its riparian hydrology was different than that of B5. Hydraulic gradients were directed from the B3 channel to the subsurface in some areas, but these negative gradients were not as strong as in upper B5, and downstream decreases in streamflow were not observed in any sections of B3. While  125  downstream increases in streamflow were not observed in B5, inferred rates of net groundwater inflow in lower B5 (0.02 L s" m" ) were an order of magnitude higher than 1  1  observed along B3 (0.001-0.003 L s" m" ). Lateral hydraulic gradients in some sections of 1  1  B3 were as strongly positive as any observed in lower B5, suggesting that the lower apparent rate of groundwater inflow in B3 was caused by lower hydraulic conductivity of the bank materials. Less permeable bank materials at B3 are consistent with observations of fewer buried gravel lenses in B3 than in lower B5. Hydrologic conditions exhibited less temporal variability at B3 than at B5. For example, between August 15 and September 4, the maximum well level increase observed in B3 was 19 cm, while most (15 of 17) well levels increased by less than 10 cm. In contrast, of those wells in lower B5 which did not dry out during the study period, the maximum increase between August 17 and September 4 was 45 cm, and most (8 of 10) well levels increased by more than 20 cm between those dates. These reach-scale hydrologic differences are superimposed onto, and probably related to, the different streamflow regimes of the two catchments noted in section 4.2.1.  4.3.2 Temporal Linkages Between Hydrologic & Thermal Regimes To understand the aquatic effects of forestry activities, we need to understand how the hydrologic and thermal characteristics of particular streams vary through time. Since biotic uses of habitat vary through the year (e.g., spawning, incubation) the seasonal timescale of variability is a key consideration and since different years or multi-year periods (e.g., Mantua et al. 1997) often have different climatic characteristics interannual variability is also important. With respect to the seasonal timescale, the threshold value of streamflow required to maintain continuous channel flow along the B5 reach appeared to vary during the study period, perhaps in response to changing groundwater levels. From July 17 to August 2, continuous streamflow was generally observed along the B5 channel when streamflow at the flume exceeded 4.5-5 L s" , but following August 19, the threshold for continuous flow 1  appeared to be closer to 3 L s" . This apparent decrease in streamflow losses in upper B5 1  cannot be readily linked to changes in near-stream hydraulic gradients. A weakening of negative bed hydraulic gradients and an increased water table elevation would be expected to  126  cause decreased streamflow losses. In contrast, observations indicate increasingly negative bed hydraulic gradients in upper B5, and similar or slight decreases in water table elevations in lower B5 between the July 17 to August 2 and post-August 19 periods. Alternatively, i f water table drawdown caused the stream channel to become disconnected from the aquifer in some key areas of upper B5 during late August, streamflow losses might have decreased as a result of the lower hydraulic conductivity of unsaturated sediments (i.e., Fig. 1 Ac vs. Fig. I. 4b). Variable saturation of the bed sediments may complicate the linkage between hydrology and stream temperature in B5. This sub-seasonal scale hydrologic variability also suggests the potential for interannual variability. However, Moore et al.'s (in press) observations suggest that the hydrologic behaviour of the B5 reach identified during the 2000 study period was similar in other years. They observed continuous channel flow through the B5 reach on July 8, 1999, when streamflow at the flume was measured at 7.2 L s" , while a section of dry channel was observed on August 1  II, 1999, when measured streamflow at the flume was 3.1 L s"'. This provides some support for the interannual consistency of the 4.5-5.0 L s" threshold value, although their 1  observations do not rule out the possibility that it may have varied by up to at least ±1 L s" . 1  Variations in hydrology at B3 through the study period appeared to be unimportant, but interannual variations in hydrologic conditions may have been important. Moore et al. (in press) observed evidence of net groundwater inflow in sub-reach 2 of B3 on August 12, 1999. On August 14-16, 2000, there was no definite evidence of net groundwater inflow in sub-reach 2 from either longitudinal profiles of electrical conductivity or sequential streamflow measurements. However, streamflow below the B2/B3 confluence was higher on August 12, 1999 (3.3 L s" ) than measured the following year, suggesting generally drier 1  conditions and lower potential for groundwater inflow in August 2000. Consequently, subreach 2 may have only been a site of net groundwater inflow at specific times. Discharge of cooler subsurface water into sub-reach 2 at B3 may have occurred during the drier conditions of August 2000, but could have escaped detection because of the difficulties in distinguishing gross inflow from net inflow. Along a 36 m reach of a gaining stream in the Rocky Mountains of Colorado, Harvey and Bencala (1993) estimated that the net inflow of water from subsurface to stream (0.0016 L s irf') was 40% less than the gross inflow (0.0027 L s" _l  1  m" ). Streamwater entered the subsurface upstream of steps, flowed laterally around the 1  127  steps, and re-entered the stream at the downstream end of steps (Harvey and Bencala 1993). Similarly, Hill et al. (1998) observed significant lateral hyporheic flowpaths in a creek where riffles of 2-3% slope alternated with pools of slope less than 1%. Even in the B3 reach, with its simpler hydrology, potentially complex surface-subsurface water interactions confounded our understanding of its thermal regime.  4.3.3 Linking Spatial Variations in Stream & Subsurface Thermal Regimes The nature of stream-groundwater interactions along a reach is a major influence on creek temperatures, as previously suggested by Constantz (1998), and numerous others (e.g., Titcomb 1926). The most obvious mechanism driving this influence is groundwater inflow, which is notoriously variable in space (e.g., Duff et al., 2000). The thermal regime of the shallow subsurface can be used to characterize the direction and rate of stream-groundwater flow (e.g., Silliman and Booth 1993), and hyporheic exchange and bed heat conduction are also both affected by the linkage between subsurface water flow and thermal regime. This section demonstrates that the differing temporal patterns of energy exchanges with, and heat storage in, the riparian zone sediments are fundamentally governed by stream-subsurface water interactions. Thermal heterogeneity in the B5 streambed indicates that the high degree of variability observed by Malard et al. (2001) within a glacial floodplain is not unique to that environment. Like the B5 reach, Malard et al. (2001) observed a range of bed temperature behaviours that fell between the two extremes of rapid streamwater downwelling (their K I and our 40 m downstream) and strong groundwater discharge (their K3 and our 165 m pool). While based on fewer measurements, it appears that the thermal regime of the riparian sediments in the B3 reach was intermediate between these two extremes, with active downwelling of streamwater into groundwater-influenced sediments. These three stream segments with differing hydrologic environments allow insight into the controls on the riparian sediment heat exchanges and the timescales of their thermal effects. Assuming that advection dominantly controlled the temperature of bed sediments of upper B5, the timescale of heat storage in the upper 20 cm of the bed would be <1 hour, since mean infiltration velocities were -30 cm hr" (assuming perfectly vertical infiltration and 1  porosity of 0.3). If this infiltration velocity was directly reflected in bed temperature  128  gradients, the mean temperature difference between streamwater and 5 cm depth in the bed (0.1 °C) would have been barely detectable since maximum warming limb slopes of daily thermographs from upper B5 were about 0.6 °C h" . It is likely, however, that a 1  disproportionate amount of the infiltration in upper B5 occurred in small areas of higher than average hydraulic conductivity, where bed temperature gradients were effectively negligible (e.g., 40 m downstream of the culvert). Hence, the spatially-averaged infiltration velocity reflected in bed temperature gradients might have been less than 30 cm hr" , and probably 1  was much lower than that at specific sites. For instance, at a hypothetical site with an infiltration velocity of 5 cm hr" , the timescale of heat storage in the upper 20 cm of the bed 1  sediments would be at least several hours. Bed temperature gradients at these sites dominated by streamwater infiltration are also affected by the rate of change in creek temperature. Given that peak rates of change in creek temperature in upper B5 were about 0.6 °C h" , conductive fluxes of -30 W m" are suggested for the hypothetical site with an 1  2  infiltration velocity of 5 cm hr" . Larger magnitudes of heat conduction would be expected in 1  losing streams with wider diurnal temperature fluctuations. At nearly the opposite end of the hydrologic spectrum, in lower B5 the rates of temperature change in the subsurface were moderated by upwelling groundwater of a relatively constant temperature. The relatively short bed temperature time series available from lower B5 suggest that heat storage occurred in the deeper riparian sediments at weekly time scales. For instance, at the 183 m site, mean daily temperature at a depth of 20 cm increased linearly with time (r =0.98) from 7.3 to 7.9 °C between July 19-26. This warming 2  could be attributed to the cumulative effects of conductive heat transfers from stream to bed or to an increase in groundwater temperature. In either case, temperature changes in deeper bed sediments appeared to occur slowly at sites with positive hydraulic gradients in lower B5, likely due to the dominance of conductive heat transfer. Advection of streamwater into the shallow (5 cm) subsurface did, however, decrease bed temperature gradients at one site with a positive mean hydraulic gradient (183 m), perhaps reflecting the non-Darcian, turbulent, flow that can occur in shallow depths of gravel beds (Packman and Bencala 2000). This pattern lessened the effect of bed heat conduction at the 183 m site, and would also have decreased the thermal effect of hyporheic exchange had it been calculated using Eq. 2.19. However, the subsurface energy exchanges in lower B5 tended to be unidirectional, with the  129  magnitude of the cooling fluxes largely depending on the stream temperature. When streamflow was continuous along the B5 reach (pattern 1 and 2 days), conductive fluxes generally cooled the stream all day long at sites with positive bed hydraulic gradients. In contrast, the results of the heuristic groundwater modelling exercise for B3 suggest that two-way "buffering processes" (Poole and Berman 2001) were an important influence on creek temperatures in that reach. Poole and Berman (2001) distinguished buffering processes from "drivers" of stream temperature, in that buffers operate "by storing heat already in the stream system rather than by adding or removing heat", and "by integrating variations in discharge and temperature over time." Heat was apparently stored within the B3 riparian zone as warm water flowed through the reach during the day, and a portion of this heat was subsequently released back to the stream as cool water flowed through the reach at night. Statistical analysis of the model residuals further suggested that the B3 riparian zone was a source of heat on relatively cool days, possibly reflecting the release of heat stored in the riparian zone on relatively warm days. Thus, important timescales of heat exchange with the B3 riparian sediments appear to have ranged between -12 to 48 hours. These timescales might be particularly effective in buffering stream temperatures because they integrate across two time cycles at which some of the largest contrasts in stream temperature occur (diurnal and multi-day changes in weather systems). Similarly, but with respect to a different timescale, Poole and Berman (2001) argued that recharge of alluvial aquifers during winter and spring floods can be an important cause of temperature buffering because this cold water may subsequently be discharged to the channel when stream temperatures are highest. Comparing these three creek segments reveals that variations in hydrologic setting affect the rates of riparian sediment heat exchanges, and the timescales at which they affect stream temperature. Faster energy exchange in upper B5 is caused by the dominant role of streamwater advection into the bed, compared to slower energy exchange associated with upwelling groundwater and conduction in lower B5. The smaller inputs of groundwater at B3 compared to lower B5 increased the relative importance of the two-way riparian sediment heat exchanges in B3's energy budget. However, the specific nature of the hydrologic environment at B3 (streamwater downwelling into groundwater-influenced sediments) may also be particularly conducive to two-way "buffering processes." The intermediate  130  characteristics of stream-subsurface water interactions in B3 likely facilitated a combination of conductive and advective energy exchanges that occurred at appropriate rates to buffer stream temperatures across important timescales. Similarly, Hill (2000) focused attention on the internal hydrologic flowpaths within riparian areas as a fundamental control on stream chemistry. However, he also emphasized the external links between the near-channel zone and its catchment, proposing a hydrogeological framework to explain why riparian zones have differing effects on stream chemistry. This conceptual framework is useful in elucidating the different thermal behaviours of the B3 and B5 study reaches. The B5 riparian zone probably falls into the "thin aquifer-rain dependent" type, having the water table close to the riparian zone most of the year, but with a large water table drawdown during extended droughts (Hill 2000). The summer-time thermal regime of B5 is likely erratic because a steady supply of groundwater is not supplied to the reach by an extensive upland aquifer. In contrast, the B3 riparian zone would likely be classified as an "intermediate aquifer" or "thick aquifer" type, with a continuous, relatively large stream baseflow, and minor water table drawdown during drought. The reach-scale linkages between surface-subsurface thermal regimes and hydrology need to be understood within the context of spatial variations at the catchment scale(as discussed in section 4.2), and ultimately the landscape scale.  4.4 Management Implications Managing the thermal impacts of forestry activities can only be feasible if those impacts are predictable at specific sites, which requires an understanding of the magnitudes of the governing processes and how they vary in space and time. Teti (2000) proposed guidelines for stream temperature management and asserted that "[t]he physical processes and parameters which control stream temperature are well understood." However, prior to this study, the main factors affecting creek temperature change in forested reaches downstream from cutblocks had not been documented. The two causes proposed by Keith et al. (1998) for the rapid cooling they observed in the closed canopy segments of Eleven Creek were groundwater inflow and evaporation, which are probably the two most frequently cited explanations. Physical reasoning and evidence from this study suggest that neither of those mechanisms was responsible for the bulk of the downstream cooling at Eleven Creek.  131  Instead, the most likely scenario is that hyporheic flow and bed conduction were the main causes, since it is plausible that these processes may have operated at sufficiently low rates in the open canopy segments to allow the relatively successful modelling of Hetrick et al. (1998). Thus, the results from this study contradict Teti's assertion, at least in the context of the downstream propagation of thermal impacts. Teti (2000) suggested that maximum daily temperatures "cannot be cumulative downstream" partly because longitudinal dispersion attenuates the daily temperature peaks. Analysis of the effect of longitudinal dispersion on temperature migration down the 200 m B3 reach showed that it was a negligible factor at that spatial scale. Similarly, Gaffield (2000) ignored longitudinal dispersion in his modelling work, after estimating that the Peclet numbers for his streams were greater than 1000. He noted that dispersion is likely to be more important in rivers larger than his study streams (where flows ranged from -3-400 L s" in 21  60 k m catchments). It is possible that the effects of bed heat conduction and hyporheic 2  exchange are sometimes mistakenly attributed to longitudinal dispersion, since these processes also reduce the magnitude of daily temperature extremes. Moreover, transient storage might slow the downstream passage of diurnal temperature waves in a manner similar to longitudinal dispersion. This is an important distinction, however, because longitudinal dispersion generally varies in predictable ways compared to the other "apparently dispersive" fluxes. For instance, an analysis of 41 longitudinal dispersion coefficients from five published studies and this-study showed that -80% of the variation in D could be explained by regression against streamflow magnitude (Appendix 8). In contrast, results of this study and many others have shown that the factors controlling bed heat conduction and hyporheic exchange vary within, and between, creeks of similar streamflow. The results from upper B5 indicate that streams that have been warmed slightly in clearings can continue to warm as they flow into heavy shade, meaning that forestry activities could have substantial thermal effects on downstream reaches. However, this behaviour may be limited to reaches that lose streamflow to the subsurface. In hydrologically "neutral" stream reaches, insignificant changes in temperature might occur due to cooling fluxes of bed heat conduction and hyporheic exchange off-setting the warming influence of the atmosphere. When streams gain substantial flow from groundwater inputs, downstream cooling is almost certain (although groundwater inflow itself will not necessarily  132  be the primary cooling mechanism). Thus, in all situations but those of streamflow losses, stream temperatures will either tend to remain constant or decrease in heavily shaded reaches during daylight hours of warm summer days. Combined with an assumption that gaining conditions prevail in small streams of humid climates, this generalization might explain why Meehan (1970) consistently observed downstream cooling during summer afternoons at shaded reaches in southeast Alaska. Downstream cooling was measured in all twelve of the shaded reaches, at an average rate of 0.1 °C per 20 m, while downstream warming was observed at all thirteen of the open reaches, at an only slightly higher rate (Meehan 1970). Mellina et al. (accepted) observed net groundwater inflow along all eleven of their study reaches (260-900 m long) in British Columbia's central interior, providing some support for the assumption of gaining conditions in small streams of humid climates. In most management situations, detailed hydrology-temperature studies employing multiple approaches, such as this study, are not feasible. Simpler methods for assessing the vulnerability of stream reaches to upstream temperature increases need to be identified. Teti (2000) referred to the concept of equilibrium temperature with the atmosphere, yet in shaded reaches heat exchanges across the stream-subsurface interface are probably more important than the heat exchanges across the water surface. Consequently, equilibrium temperatures of small, forested, creeks should be evaluated with respect to subsurface temperatures, rather than, or in addition to, the traditional atmospheric variables (e.g., Mohseni and Stefan 1999; Caissie et al. 2001). One method may be the careful measurement of streambed temperature gradients during hot summer afternoons. This single variable appeared to be closely related to downstream temperature patterns within our two study reaches, and it has previously been shown to be an effective tool for differentiating between losing and gaining reaches (Silliman and Booth 1993). However, guidelines and protocols would need to be developed to deal with the substantial variability in bed temperature gradients, especially the systematic differences related to channel morphology. Downstream temperature patterns need to be further studied at a wide variety of field sites before bed temperatures could be applied as a management tool. Sampling schemes might vary depending upon the ecology of the particular aquatic species of concern. For instance, the special affinity of some fish species for groundwater (e.g., Salvelinus fontinalis; Curry and Noakes 1995) means that they may tend to occupy habitats that are generally well-  133  protected from thermal perturbations. This hypothesis appears to have been supported in the Baptiste tributaries. Rainbow trout (Oncorhynchus mykiss) were observed spawning in the B3 sub-reach during a preliminary study site visit in June, 2000, and fry were prevalent in that study reach during fieldwork in August. In contrast, fish were not observed in the B5 study reach at either time. Trout were rarely observed at B5 by DFO staff in previous summers, and only in the groundwater-influenced lower sub-reach (H. Herunter, DFO, personal communication). Other variables such as channel gradient may have influenced fish distribution. However, fry were observed upstream of the B2/B3 confluence in August, 2000, where channel gradients are at least as great as in upper B5 (Fig. 2.8). Perhaps only bed temperatures of streams used by species that are particularly sensitive to increased temperatures (e.g., Salvelinus confluentus) would need to be intensively monitored, although effects on the ecology of any resident species should be considered (Torgersen et al. 1999; Power etal. 1999). Based on the limited results presented here, it is unclear whether measurements of bed temperatures would enable a quantitative prediction of post-harvesting thermal regimes, or remain a largely qualitative technique. This relatively narrow channel-based approach would also have to be supplemented by a basin-wide perspective on the impacts of forestry practices (e.g., Curry and Devito 1996). In particular, this study suggests that information on catchment-scale variability in regolith depth would be invaluable. The threshold between riparian zones that are permanently connected to upland aquifers, and those where the connection varies seasonally or inter-annually may depend on small differences in regolith depths (Devito et al. 1999; Vidon and Hill 2002). Consequently, different catchment-scale hydrogeologic responses to forest harvesting and climatic variability could confound the abilities of subsurface thermal monitoring to identify stream reaches that are thermally vulnerable to forestry activities. This study adds to a growing recognition of the high spatial variability in environmental responses to forest management, even within small watersheds (e.g., Brardinoni 2001).  134  ,\  CHAPTER 5 CONCLUSIONS - 5.1 Summary of Key Findings Field investigations in July-August, 2000, showed that both streams continued to gain energy from the overlying atmosphere/vegetation on most afternoons as they flowed through the -200 m forested reaches. This effect only consistently resulted in measurable downstream warming in upper B5, where groundwater did not contact the channel. Net inflow of groundwater in lower B5 (~0.02 L s" m" ) sometimes caused rapid downstream 1  1  cooling, although the thermal behaviour depended on the magnitude of streamflow input at the head of the reach. The greatest downstream cooling occurred when the streamflow input was <5 L s" , because infiltration in the upper 150 m of the reach consumed the warmer 1  streamwater from the upstream catchment. Modest inputs of groundwater (-0.002 L s" m" ) occurred throughout the B3 study 1  1  reach, causing an estimated 0.5 °C cooling in the average daily creek temperature. Two-way exchanges of energy between the creek and its subsurface, driven by conduction, and possibly hyporheic exchange, were important influences on the daily temperature extremes at B3. Heat appeared to be stored within the riparian zone during daytime and released to the stream at nighttime. Heat storage also affected the mean daily temperatures at B3LO, apparently depending on the relative strengths of the daytime heat sink and nighttime heat source effects. These interactions within the riparian zone are heavily conditioned by the characteristics of the surrounding catchment. Regolith depth was probably a particularly important variable in the case of these two streams. The sensitivity of a stream reach to upstream thermal perturbations appears to be closely related to its hydrologic environment. In particular, reaches that rapidly lose water to the subsurface should probably be treated cautiously. Not only are upstream temperature increases more likely to persist through such reaches, but the increased temperatures might also exacerbate streamflow losses. Increased streamflow losses might potentially lead to dry sections of channel that did not occur, or only occurred rarely, prior to harvesting. Measuring bed temperature profiles could be the easiest way to assess a stream's hydrologic environment, and thereby identify vulnerable losing reaches, although this approach needs to be tested on a wide variety of streams.  135  5.2 Suggestions for Future Research The small time and space scales addressed in this study mean that many questions remain about the thermal behaviour of small, forested streams. At the coarsest scale, we need to know more about interactions between small streams and their aquifers and how they vary between different catchments and landscapes. In what areas of catchments should we find discharge and recharge zones in different landscapes, and how do the hydrologic and thermal characteristics of these zones vary through the year and between years? Energy balance studies should be conducted in forested reaches with a wide range of hydrologic environments, to clarify the role of stream-subsurface interactions and their effects on energy exchanges. At intermediate spatial scales, we need to know more about the effects of channel features such as meanders on water flows and stream temperature. Step-pool units, and riffle-pool units in shallower gradient streams, also deserve further study from a thermal perspective, since these are known to be important generators of hyporheic flow and may also partially determine locations of groundwater discharge. At the finest spatial-scales, we will ultimately need to know more about the mechanisms of energy exchanges between streamwater and the hyporheic zone. There is currently confusion about the relative roles of bed heat conduction and hyporheic exchange, how mixing takes place between stream and subsurface waters in the hyporheic zone, and how all of these processes are inter-related. The approaches taken to resolve these issues will vary depending on the question of interest. The complexity of hydrologic interactions in small streams and their influence on stream temperature necessitates multiple research approaches, and simpler integrative methods for "prescriptions." A common challenge lies in reliably monitoring both the time and space dimensions of riparian zone thermal regimes. "Scaling-up" temporally-intensive (i.e., data-logged) measurements using more spatially-extensive manual measurements appears to be a promising approach. Physical modelling of subsurface thermal regimes would enhance our understanding of the relative roles of advection and conduction at specific sites. Tracer tests should be conducted at the reach-scale to examine further the potential for transient storage to mediate the downstream propagation of temperature pulses. Combining these latter two approaches using an advection-dispersion model (e.g., OTIS) and energy balance data will be necessary, given the dynamic nature of the heat exchanges and the need to integrate the effects of multiple energy exchanges across multiple time and space scales.  136  REFERENCES Adams, T.N. 1999. Primer on the Physics of Forest Stream Temperature (draft):, Tacoma, WA. Ahrens, C D . 1994. Meteorology Today 5 ed.. West Publishing, St. Paul, M N , 592 p. th  Barton, D.R., W.D. Taylor, and R . M . Biette. 1985. Dimensions of riparian buffer strips required to maintain trout habitat in southern Ontario streams. North American J. Fisheries Management 5: 364-378. B.C. Ministry of Forests. 1995. Biogeoclimatic zones of British Columbia. Folio map available from B.C. Ministry of Forests, Research Branch, 31 Bastion Square, Victoria, British Columbia. Beaudry, P.G. 1998. Effects of forest harvesting on streamflow and sediment concentrations of small streams in central British Columbia. Paper presented at the Canadian Water Resources 51 Annual Conference: Mountains to Sea: Human Interactions with the Hydrological Cycle, Victoria B.C. st  Beaudry, P.G. 2001. Effects of riparian management strategies on the hydrology of small streams in the Takla region of British Columbia. Final report submitted to the Science Council of British Columbia by P. Beaudry and Associates Ltd., Prince George B.C., 33 p. Beschta, R.L., and R.L. Taylor. 1988. Stream temperature increases and land use in a forested Oregon watershed. Water Resources Bulletin 24: 19-25. Beschta, R.L., R.E. Bilby, G.W. Brown, L.B. Holtby and T.D. Hofstra. 1987. Stream temperature and aquatic habitat: fisheries and forestry interactions. Chapter six in: Streamside Management: Forestry and Fishery Interactions. Eds. E.O. Salo and T.W. Cundy. Institute of Forest Resources, University of Washington, Seattle, Washington p. 191-232. Bilby, R.E. 1984. Characteristics and frequency of cool-water areas in a western Washington stream. J. Freshwater Ecology 2: 593-602. Black, T.A., J-M. Chen, X . Lee, and R . M . Sagar. 1991. Characteristics of shortwave and longwave irradiances under a Douglas-fir forest stand. Canadian J. Forest Research 21: 1020-1028. Bond, B.J., J.A. Jones, G. Moore, N . Phillips, D. Post, and J.J. McDonnell. 2002. The zone of vegetation influence on baseflow revealed by diel patterns of streamflow and vegetation water use in a headwater basin. Hydrological Processes 16: 1671-1677. Brardinoni, F. 2001. Identification of natural and logging-related landslides in the Capilano River Basin (Coastal British Columbia): A comparison between remotely sensed and 137  field survey. Unpublished MSc thesis, Department of Geography, University of British Columbia, Vancouver, British Columbia, 127 p.  1  Bren, L.J. 1997. Effects of slope vegetation removal on the diurnal variations of a small mountain stream. Water Resources Research 33: 321-331. Brown, G.W. 1969. Predicting temperatures of small streams. Water Resources Research 5: 68-75. Brown, G.W. 1972. A n improved temperature prediction model for small streams. Oregon State University, Water Resources Research Institute, WRRI-16. (Obtainable from O.S.U.'s Center for Water and Environmental Sustainability - Corvalllis, Oregon.) Brown, G.W., G.W. Swank, and J. Rothacher. 1971. Water temperature in the Steamboat Drainage. U.S. Department of Agriculture Forest Service Res. Paper PNW-119, 17 p. Bundschuh, J. 1993. Modeling annual variations of spring and groundwater temperatures associated with shallow aquifer systems. J. Hydrology 142: 427-444. Butturini, A., and F. Sabater. 1998. Ammonium and phosphate retention in a Mediterranean stream: hydrological versus temperature control. Canadian J. Fisheries and Aquatic Sciences 55: 1938-1945. Caissie, D., N . El-Jabi, and M . G . Satish. 2001. Modelling of maximum daily water temperatures in a small stream using air temperatures. / . Hydrology 251: 14-28. Castro, N . M . , and G . M . Hornberger. 1991. Surface-subsurface water interactions in an alluviated mountain stream channel. Water Resources Research 27: 1613-1621. Cey, E.E., D.L. Rudolph, G.W. Parkin, and R. Araveria. 1998. Quantifying groundwater discharge to a small perennial stream in southern Ontario, Canada. J. Hydrology 210: 21-37. Christie, T . M . 1998. Application of multi-element geochemical methods to identify sediment sources and trace the transport of sediment in small (S4) streams before and after watershed disturbance. Unpublished M A S c thesis, University of British Columbia. Clark, E., B.W. Webb, and M . Ladle. 1999. Microthermal gradients and ecological implications in Dorset rivers. Hydrological Processes 13: 423-438. Collet, A . and J.M. Ryder. 1997. Baptiste Creek Watershed: Detailed terrain and sediment source mapping with interpretations for slope stability, erosion potential and sediment transfer. Report prepared for the B.C. Ministry of Forests, Department of Fisheries and Oceans Canada and the University of British Columbia. J.M. Ryder and Associates Terrain Analysis Inc. Vancouver, B.C.  138  Comer, L.E., and W.J. Grenney. 1977. Heat transfer processes in the bed of a small stream. Water Research 11: 743-744. Constantz, J. 1998. Interaction between stream temperature, streamflow, and groundwater exchanges in alpine streams. Water Resources Research 34: 1609-1715. Constantz, J., and C L . Thomas. 1997. Stream bed temperature profiles as indicators of percolation characteristics beneath arroyos in the Middle Rio Grande Basin, U S A . Hydrological Processes 11: 1621-1634. Constantz, J., C L . Thomas, and G. Zellweger. 1994. Influence of diurnal variations in stream temperature on streamflow loss and groundwater recharge. Water Resources Research 30: 3253-3264. Constantz, J., D. Stonestrom, A . E . Stewart, R. Niswonger, and T.R. Smith. 2001. Analysis of streambed temperatures in ephemeral channels to determine streamflow frequency and duration. Water Resources Research 37: 317-328. Crisp, D.T. 1990. Water temperature in a stream gravel bed and implications for salmonid incubation. Freshwater Biology 23: 601-612. Curry, R.A., and K.J. Devito. 1996. Hydrogeology of brook trout (Salvelinus fontinalis) spawning and incubation habitats: implications for forestry and land use development. Canadian J. Forest Research 26: 767-772. Curry, R.A., and D.L.G. Noakes. 1995. Groundwater and the selection of spawning sites by brook trout (Salvelinus fontinalis). Canadian J. Fisheries and Aquatic Sciences 52: 1733-1740. D'Angelo, D.J., J.R. Webster, S.V. Gregory, and J.L. Meyer. 1993. Transient storage in Appalachian and Cascade mountain streams as related to hydraulic characteristics. J. North American Benthological Society 12: 223-235. den Hartog, G. and H.L. Ferguson. 1978. Water balance-derived precipitation and evapotranspiration. Plate 25, Hydrological Atlas of Canada, Ottawa, Department of Fisheries and Environment, Ottawa, Ontario, map. Devito, K.J., A.R. Hill, and P.J. Dillon. 1999. Episodic sulphate export from wetlands in acidified headwater catchments: Prediction at the landscape scale. Biogeochemistry 44: 187-204. Dingman, S.L. 1994. Physical Hydrology. Prentice-Hall, Inc., New Jersey, 575 p. Domenico, P.A., and F.W. Schwartz. 1998. Physical and Chemical Hydrogeology 2 John Wiley & Sons, New York, N Y , 506 p.  139  nd  ed.  Duff, J.H., B . Toner, A.P. Jackman, R.J. Avanzino, and F.J. Triska. 2000. Determination of groundwater discharge into a sand and gravel bottom river: a comparison of chloride dilution and seepage meter techniques. Verhandlungen der Internationalen Vereinigung fur Theoretische und Angewandte Limnologie 27: 406-411. Edinger, J.E. D.W. Duttweiler, J.C. Geyer. 1968. The response of water temperatures to meteorological conditions. Water Resources Research 4: 1137-1143. Environment Canada. 2002. www.msc-smc.ec.gc.ca/ climate/climatenonnals_1990/index_e.cfrn [cited July 2002]. Evans, E.C., and G.E. Petts. 1997. Hyporheic temperature patterns within riffles. Hydrological Sciences Journal 42: 199-213. Evans, E.C., G.R. McGregor, and G.E. Petts. 1998. River energy budgets with special reference to river bed processes. Hydrological Processes 12: 575-595. Findlay, S. 1995. Importance of surface-subsurface exchange in stream ecosystems: The hyporheic zone. Limnology and Oceanography 40: 159-164. Findlay, S., D. Strayer, C. Goumbala, and K . Gould. 1993. Metabolism of streamwater dissolved organic carbon in the shallow hyporheic zone. Limnology and Oceanography 38: 1493-1499. Freeze, R.A., and J.A. Cherry. 1979. Groundwater. Prentice-Hall, Englewood Cliffs, New Jersey, 604 p. Gaffield, S.J. 2000. Evaluation of the controls of summer stream temperature in the Driftless area of southwestern Wisconsin. Unpublished PhD thesis, University of Wisconsin at Madison. Gburek, W.J., and G.J. Folmar. 1999. Flow and chemical contributions to streamflow in an upland watershed: abaseflow survey. J. Hydrology 217: 1-18. Geist, D.R. 2000. Hyporheic discharge of river water into fall chinook salmon {Oncorhynchus tshawytscha) spawning areas in the Hanford Reach, Columbia River. Canadian J. Fisheries and Aquatic Sciences 57: 1647-1656. Genereux, D.P., H.F. Hemond, and P.J. Mulholland. 1993. Spatial and temporal variability in streamflow generation on the West Fork of Walker Branch Watershed. J. Hydrology 142: 137-166. Greene, G.E. 1950. Land use and trout streams. J. Soil and Water Conservation 5: 125-126.  140  Harvey, J.W., and K . E . Bencala. 1993. The effect of streambed topography on surfacesubsurface water exchange in mountain catchments. Water Resources Research 29: 89-98. Harvey, J.W., and C C . Fuller. 1998. Effect of enhanced manganese oxidation in the hyporheic zone on basin-scale geochemical mass balance. Water Resources Research 34: 623-636. Harvey, J.W., and B.J. Wagner. 2000. Quantifying hydrologic interactions between streams and their subsurface hyporheic zones. Chapter 1 in: Streams and Ground Waters, eds. J.B. Jones and P.J. Mulholland, Academic Press, San Diego, C A . Harvey, J.W., B.J. Wagner, and K . E . Bencala. 1996. Evaluating the reliability of the stream tracer approach to characterize stream-subsurface water exchange. Water Resources Research 32: 2441-2451. Hetrick, N.J., M . A . Brusven, W.R. Meehan, and T . C Bjornn. 1998. Changes in solar input, water temperature, periphyton accumulation, and allochtonous input andstorage after canopy removal along two small salmon streams in southeast Alaska. Transactions of the American Fisheries Society 127: 859-875. Hill, A.R. 2000. Stream chemistry and riparian zones. Chapter 3 in: Streams and Ground Waters, eds. J.B. Jones and P.J. Mulholland, Academic Press, San Diego, C A . Hill, A.R., and D.J. Lymburner. 1998. Hyporheic zone chemistry and stream-subsurface exchange in two groundwater-fed streams. Canadian J. Fisheries and Aquatic Sciences 55: 495-506. Hill, A.R., C.F. Labadia, and K. Sanmugadas. 1998. Hyporheic zone hydrology and nitrogen dynamics in relation to the streambed topography of a N-rich stream. Biogeochemistry 42: 285-310. Hondzo, M . , and H.G. Stefan. 1994. Riverbed heat conduction prediction. Water Resources Research 30: 1503-1513. Huggenberger, P., E. Hoehn, R. Beschta, and W. Woessner. 1998. Abiotic aspects of channels and floodplains in riparian ecology. Freshwater Biology 40: 407-425. Johnson, S.L., and J.A. Jones. 2000a. Spatial and temporal dynamics of stream temperature: Geomorphic and riparian influences. In: Spatial and Temporal Variation in Aquatic Communities II, North American Benthological Society annual meeting. Johnson, S.L., and J.A. Jones. 2000b. Stream temperature responses to forest harvest and debris flows in western Cascades, Oregon. Canadian J. Fisheries and Aquatic Sciences 57(Suppl. 2): 30-39.  141  Kasahara, T., and S.M. Wondzell. 2001. Effects of geomorphic features on hyporheic exchange flow in mountain streams. Oral presentation at the 27 annual scientific meeting of the Canadian Geophysical Union, May, 20001, Ottawa, Ontario. th  Keith, R . M . , T.C. Bjornn, W.R. Meehan, N.J. Hetrick, and M . A . Brusven. 1998. Response of juvenile salmonids to riparian and instream cover modifications in small streams flowing through second-growth forests of southeast Alaska. Transactions of the American Fisheries Society 127: 889-907. Land, L . A . , and C K . Paull. 2001. Thermal gradients as a tool for estimating groundwater advective rates in a coastal estuary: White Oak River, North Carolina, U S A . J. Hydrology 248: 198-215. Lapham, W.W. 1989. Use of temperature profiles beneath streams to determine rates of vertical ground-water flow and vertical hydraulic conductivity. U.S. Geological Survey water-supply paper 2337, 35 p. LeBlanc, R.T., R.D. Brown, and J.E. FitzGibbon. 1997. Modelling the effects of land use change on the water temperature in unregulated urban streams. J. Environmental Management 49: 445-469. Levno, A . , and J. Rothacher. 1967. Increases in maximum stream temperatures after logging in old-growth Douglas-fir watersheds. U S D A Forest Service Research Note PNW 65. Pacific Northwest Forest and Range Experimental Station, Portland, Oregon. Macdonald, J.S., J.C. Scrivener, and G. Smith. 1992. The Stuart-Takla Fisheries/Forestry Interaction Project: Study Description and Design. Can. Tech. Rep. Fish. Aquat. Sci. 1899. 39 p. Malard, F., A . Mangin, U . Uehlinger, and J.V. Ward. 2001. Thermal heterogeneity in the hyporheic zone of a glacial floodplain. Canadian J. Fisheries and Aquatic Sciences 58: 1319-1335. Mantua, N.J., S.R. Hare, Y . Zhang, J.M. Wallace, and R.C. Francis. 1997. A Pacific interdecadal climate oscillation with impacts on salmon production. Bulletin of the American Meteorological Society 78: 1069-1079. McGurk, B.J. 1989. Predicting stream temperature after riparian vegetation removal, p. 157-164 in: Proceedings of the California Riparian Systems Conference, Davis, California. U S D A Forest Service General Technical Report PSW-110. Meehan, W.R. 1970. Some effects of shade cover on stream temperature in southeast Alaska. U S D A Forest Service Research Note PNW-113. Mellina, E., R.D. Moore, S.G. Hinch, J.S. Macdonald, and G. Pearson. Stream temperature responses to clear-cut logging in the central interior of British Columbia: the  142  moderating influences of groundwater and headwater lakes. Accepted for publication in: Canadian J. Fisheries and Aquatic Sciences, cited original manuscript submitted May 21, 2002. Mohseni, O., and H.G. Stefan. 1999. Stream temperature/air temperature relationship: a physical interpretation. J. Hydrology 21%: 128-141. Moore, R.D., J.S. Macdonald, H . Herunter. In Press. Downstream thermal recovery of headwater streams below cutblocks and logging roads. A progress report on the Stuart-Takla Fish-Forest Interaction Project. Morrice, J.A., H . M . Valett, C.N. Dahm, and M . E . Campana. 1997. Alluvial characteristics, groundwater-surface water exchange and hydrological retention in headwater streams. Hydrological Processes 11: 253-267. Mulholland, P.J., E.R. Marzolf, J.R. Webster, D.R. Hart, and S.P. Hendricks. 1997. Evidence that hyporheic zones increase heterotrophic metabolism and phosphorus uptake in forest streams. Limnology and Oceanography 42: 443-451. NOAA 2002. http://lwf.ncdc.noaa.gov/img/documentlibrary/clim81 supp3/tempnormal_hires.jpg Oke, T.R. 1987. Boundary Layer Climates. Routledge, New York, 435 p. Packman, A.I., and K . E . Bencala. 2000. Modeling surface-subsurface hydrological interactions. Chapter 2 in: Streams and Ground Waters, eds. J.B. Jones and P.J. Mulholland, Academic Press, San Diego, C A . Polehn, R.A., and W.C. Kinsel. 2000. Transient temperature solution for a river with distributed inflows. Water Resources Research 36: 787-791. Poole, G.C., and C.H. Berman. 2001. A n ecological perspective on in-stream temperature: Natural heat dynamics and mechanisms of human-caused thermal degradation. Environmental Management 27: 787-802. Power, G , R.S. Brown, and J.G. Imhof. 1999. Groundwater and fish - insights from northern North America. Hydrological Processes 13: 401-422. Ringler, N.H., and J.D. Hall. 1975. Effects of logging on water temperature and dissolved oxygen in spawning beds. Transactions of the American Fisheries Society 104: 111121. Ronan, A.D., D.E. Prudic, C.E. Thodal, and J. Constantz. 1998. Field study and simulation of diurnal temperature effects on infiltration and variably saturated flow beneath an ephemeral stream. Water Resources Research 34: 2137-2153.  143  Rowland, J.D., and R.D. Moore. 1992. Modelling solar irradiance on sloping surfaces under leafless deciduous forests. Agricultural and Forest Meteorology 60: 111-132. Runkel, R.L. 1998. One-dimensional transport with inflow and storage (OTIS): A solute transport model for streams and rivers. United States Geological Survey WaterResources Investigations Report 98-4018. Shepherd, B.G., G.F. Hartman, and W.J. Wilson. 1986. Relationships between stream and intragravel temperatures in coastal drainages, and some implications for fisheries workers. Canadian J. Fisheries and Aquatic Sciences 43: 1818-1822. Silliman, S.E., and D.F. Booth. 1993. Analysis of time-series measurements of sediment temperature for identification of gaining vs. losing portions of Juday Creek, Indiana. J. Hydrology 146: 131-148. Silliman, S.E., J. Ramirez, and R.L. McCabe. 1995. Quantifying downflow through creek sediments using temperature time series: one-dimensional solution incorporating measured surface temperature. J. Hydrology 167: 99-119. Sinokrot, B.A., and H.G. Stefan. 1993. Stream temperature dynamics: measurements and modelling. Water Resources Research 29: 2299-2312. Stream Solute Workshop. 1990. Concepts and methods for assessing solute dynamics in stream ecosystems. J. North American Benthological Society 9: 95-119. Stull, R.B. 1995. Meteorology Today for Scientists and Engineers. West Publishing Co, St. Paul, M N , 385 p. Taniguchi, Y . , F.J. Rahel, D.C. Novinger, and K . G . Gerow. 1998. Temperature mediation of competitive interactions among three fish species that replace each other along longitudinal stream gradients. Canadian J. Fisheries and Aquatic Sciences 55: 18941901. Teti, P. 2000. Recommendations for managing the effects of forest practices on stream temperature in British Columbia. Prepared for Temperature Sensitive Stream Working Group, B.C. Ministry of Forests. Thibodeaux, L.J., and J.D. Boyle. 1987. Bedform-generated convective transport in bottom sediment. Nature325: 341-343. Titcomb, J.W. 1926. Forests in relation to fresh water fishes. Transactions of the American Fisheries Society 56: 122-129. Torgersen, C.E., D . M . Price, H.W. L i , and B.A. Mcintosh. 1999. Multiscale thermal refugia and stream habitat associations of chinook salmon in northeastern Oregon. Ecological Applications 9: 301-319.  144  Valentine, K.W.G., P.N. Sprout, T.E. Baker, and L . M . Lavkulich (eds.). 1986. The soil landscapes of British Columbia, Ministry of Environment, Victoria, British Columbia, 197 p. Valett, H . M . , J.A. Morrice, C.N. Dahm, and M . E . Campana. 1996. Parent lithology, surface-groundwater exchange, and nitrate retention in headwater streams. Limnology & Oceanography 41: 333-345. Vannote, R.L., G.W. Minshall, K . W . Cummins, J.R. Sedell, and C.E. Cushing. 1980. The river continuum concept. Canadian J. Fisheries and Aquatic Sciences 37: 130-137. Vaux, W.G. 1968. Intragravel flow and interchange of water in a streambed. Fishery Bulletin 66: 479-489. Vidon, Ph., and A.R. Hill. 2002. Landscape controls on stream riparian zone hydrology. Oral presentation at 51 Annual Meeting of the Canadian Association of Geographers, Toronto, Ontario, May 28- 1 June, 2002. st  Webb, B.W., and Y . Zhang. 1997. Spatial and seasonal variability in the components of the river heat budget. Hydrological Processes 11: 79-101. Webb, B.W., and Y . Zhang. 1999. Water temperatures and heat budgets in Dorset chalk water courses. Hydrological Processes 13: 309-321. White, D.S., C.H. Elzinga, and S.P. Hendricks. 1987. Temperature patterns within the hyporheic zone of a northern Michigan river. J. North American Benthological Society 6: 85-91. Wondzell, S.M., and F.J. Swanson. 1996. Seasonal and storm dynamics of the hyporheic zone of a 4 -order mountain stream: I: Hydrologic processes. J. North American Benthological Society 15: 3-19. th  Wroblicky, G.J., M . E . Campana, H . M . Valett, and C.N. Dahm. 1998. Seasonal variation in surface-subsurface water exchange and lateral hyporheic area of two stream-aquifer systems. Water Resources Research 34: 317-328.  145  APPENDICES  Appendix 1. Description of the error estimation for the terms in Eqs. 2.3, 2.4.  Inflow rates (qi) were typically 50-60 mL in ~2 minutes, where the uncertainty associated with reading the graduated cylinder was estimated at ±0.5 mL, and the uncertainty associated with measuring the time was ±2 s (1 s on either end of the timing). These estimations yielded a Sqj/q; value of ~ 2% for all measurements. Values of 5 E C and 8 E C P  B  were 0.1 ps cm" when EC values were less than 200 ps cm" , except when a range of 1  1  fluctuations around either value were observed, and 1 ps cm" when values were above 200 1  ps cm" . Finally, the standard error around the slope (b) of the least-squares regression 1  equation calculated from the calibration procedure was used as the 8b value.  I  146  Appendix 2. (a)  C r e e k t e m p e r a t u r e s o b s e r v e d in u p p e r B5 u s i n g C a m p b e l l Scientific d a t a - l o g g e r s , a n d (b) m o d e l l e d net radiation at 4 0 m d o w n s t r e a m f r o m J u l y 1 8 - 2 9 . T h e c r e e k t e m p e r a t u r e t h e r m o c o u p l e at 117 m d o w n s t r e a m w a s dry for m o s t of t h e period f r o m J u l y 2 4 to t h e m o r n i n g o f J u l y 26.  18  19  20  21  22  19  20  21  22  23  24  Z-20 H 18  147  23 24 Date (July, 2000)  27  28  29  Appendix 3. (a)  C r e e k t e m p e r a t u r e s o b s e r v e d in u p p e r B 5 u s i n g Tidbit d a t a - l o g g e r s , a n d m o d e l l e d net radiation at 2 0 m d o w n s t r e a m f r o m A u g u s t 4 - 1 3 .  (b)  A p p e n d i x 4. D o w n s t r e a m t e m p e r a t u r e c h a n g e s (B3BR-B3LO) m o d e l l e d d u r i n g J u l y 17 to A u g u s t 31: a ) daily m e a n s , b) daily m i n i m a , a n d c ) daily m a x i m a . M o d e l l i n g u s e d E q . 2.1, with q =1.1 L s" , q = 1.5 L s" , T w = 7.7 ° C a n d o b s e r v e d t e m p e r a t u r e s at B3BR a s T . D a s h e d line a n d e q u a t i o n in a ) is a linear r e g r e s s i o n . 1  u  s  D S  G  u s  0  M:1  -0.4 y = 0.70x-0.18 r2 = 0.78  -0.8  |  -1.2  TD  o -1.6  i  -1.6  i  1  -1.2  i  1  -0.8  r  1  -0.4  0  O b s e r v e d C h a n g e in M e a n ( ° C )  1.2  o  1:1  o  ~  0.8  i  0-4  cn c  CO  I• •* • •  0  o  -0.4  jD CD TD O  -0.8  -0.8  -0.4  0  0.4  0.8  1.2  O b s e r v e d C h a n g e in M i n . ( ° C )  9  0  x CO  . I • • • • • •  CD  ? -2 CO sz  O  TD  CD O  _3  1:1  TD  A  -3  -2  -1  O b s e r v e d C h a n g e in M a x . ( ° C )  149  0  Appendix 5.  G r o u n d w a t e r inflow m o d e l l i n g for B 3 including daily variations in g r o u n d w a t e r t e m p e r a t u r e : (a) t i m e s e r i e s of r e s i d u a l s f r o m m o d e l u s i n g o b s e r v e d m e a n daily c r e e k t e m p e r a t u r e s at B 3 H I a s daily v a l u e s of T w , (similar to Fig. 3.44a but different s c a l e ) , (b) daily air t e m p e r a t u r e e x t r e m e s at t h e O p e n site ( s a m e a s Fig. 3.44b), (c) m o d e l l e d v s . o b s e r v e d daily m e a n t e m p e r a t u r e s at B 3 L O , a n d (d) m e a n daily c r e e k t e m p e r a t u r e s at B 3 H I a s a f u n c t i o n of t h e daily m a x i m u m air t e m p e r a t u r e a n o m a l y at t h e O p e n site. G  0.8 O o  0.4 CD CO 00 O  Q  c  I  o  •I I  i co T3 CD = CD  2  o  -0.4  o co 2  1  CO  I I I I I I I i i \ i I i i I i  -0.8  ~|—i—i—i—I—i—i—i—|—i—I—i—r  18-Jul  26-Jul  T—|—i—i—i—n—i—1™|—i—i—I—i—i—i—i—p—i  3-Aug  11-Aug  i i i I r  19-Aug  i  27-Aug  Date (year 2000) 12  Y = 1 . 1 4 X - 1.39 r = 0.96  10  1:1, O  2  c  11  CO  CO  9 H  co Q c  0)  a>  2  3  10 "O  Y . 0 2 X + 7.55 r = = 00.01  ^co 9  CO CD  Q  ^  8  CO CD  10 11 12 8 Measured Daily Mean at B3LO (°C) 150  o  o  o  o  8 O  oo CO  o  o  0  o  ^  oo  o  -2 0 2 4 6 Max. Daily Air T - 7 Day Centered Average of Max. Daily Air T (°C)  0  C  ro o co CO  -  co  x: > CD  .<« § =  g> • • —  %E c o o  o  i2 h  -  CD  c  %B u-  —-  CD •a o  >.£ co -a  <-  OT  c  <»  O  CD  S "°  CO CO >  CO  +^  +^  +^  +^  +^  +^  +^  +^  +^  CO o d  CM d  CD O d  CO CO d  CM  o o T—  +^  +^  +^  +^  CO CM d  in o d  o o d  CO CO d  CM O d  +  +  +  o d  CO i— d  o o  +^  +^  d  CD O d  +^  +^  Tt  d  d  T— o d  +  +  +  +  +  +  +  o d  o o d  CM O d  m o d  CM d  O d  CM CM d  +^  +^  +^  +^  +^  +^  +^  m CM d  o d  o CM d  Tt  oo  o o  oo  T-  M"  00  d  CD CM d  co d  +^  +^  CD CO d  LO CD d  S«M  CD  x:  +^  =3  O CO ro cnx: 0 3  3  CO 0  <  S  co 1 ~~ o 1—  o0 =3 •=g ro  i l l  i CM  F 0  1  ro ro 0^ |  -a  2  c  CD P  E —  O XJ  CD  TJ" N T3 co !.  to • o-g CD CO  sco s»c ; 0 S  x:  o E  c  1 -  +^  CD CM d  r~m d  C O  0  CD i-  c  co .2 g  • D CO CD ^J-  CM d  CO o d  co CM d  CM d  1  Tiro d  +^  d  Tt  CO co d  in d  +^  +^  Tt  o o  o o  m d  d  d  o o  CM O d  co CD d  CO  4^  +^  4^  d  co d  m CO d  CO o d  +  +  +  +  +  CD O d  in d  m o d  +^  CM CM d  in o d  CD CO d  in Md  +^ a> CM d  CD o d  +^  +^  +^  o d  Tt  +^  +^  d  CO CO d  +^  o d  CO CM d  +^  +^  o d  CM d  o d  +^  +^  +^  LO CM d  m o d  +^  o CM d  •  +  o o d  CD CM d  +^  +  co o d  m CD d  o o d  d  +  d  d  m  cn o d  oo  Tt  •  o d  -  T—  co d  CO o d  CM T— d  T"  o o  CO d  +^ CO o d  CO m d  o d  +^  d  LO o d  +^  +^  d  CD o d  a> o d  d  +^  +^  +^  +^  +^  O d  CM o d  o d  i  +^  +^  Tt  O d  CO T— d  +^  +^  •  o o  +^  +^  •  +^  1--.  CM Md  d  St 0  * ro a> Ca> J)to 8^ •o 5 Q5 SS C T 3 C 250 «P •3 E a) « C ro^ Q. £ ro > , ° Q. O "g ro ro ro .S>  CM d  o o  Tt  d  a> o d  T-'  +^  m o d  CM d  1  •  i  i  CO o d  CM M" d  CM d  LO d  in CO d  i  +^  +^  +^  +^  +^  +^ o o T—  m d  CO CM d  CM d  i o o  Tt  d  Tiro O  o p  CO o d  CO in d  oo oo d  +^  +^  +^  o  C D co * ro to 2CD cu o£  —  -  Tt  CM CO d  ro  T3  I 0  £ >."S  +^  33o  3  £  in co d  o a  CD >s. '43  •  d  CM  co  . -o  T "  o o  +^  o d  m  .fo  CO d  m CM d  •sf lO d  2  0  Tt  t u> o  m o d  s:  co  d  MO d  CO  i2 b 0 5 '*9> _ E > o  S  o  o d  o d  T— o d  II E  x:  o o  •  d  +  0  +^ CO  d  ^Q.  co o d  m CO d  +  Q) CO t  o  d  d  0  o o d  T— LO d  +  3  cn o d  CD o d  +  O x:  d  o d  +  £=  o  LO o d  +  to  \—  cn o d  +  i r  CO  M" O  d  Eco a) _>,  X  +^  +  w  0 D)>C0  0  +^  °  x:  1  •S E (» ^ O,: CO  +^  Q - — , „ CO >  D  C " , a. (0 ^ ro E  i-  _ ro o  ,  0  co p d  a> o d  £  25  in o d  co  E '  « CD  +^  CO d  oo CO d  « X>  £ co CO CD  CO LO d  • 5  =' - -  co  3  CM CO d  -  0  O  CO CO >  o I £ =*<o * o0  to  i_ CO CD jjj CD Q.  > 0  >  ^ CD  T 6 -g o"  •E 0  ro  w  0  co  3 J? o XI  j)  <D Z r- *^ i_r O 0 i_ > 0  -~ -~ D)  E *-  3  co °  d  i  +^  w  Cf  —  o o d  o  CD  m  CD CD x: x:  o  S  c CO  O O O  T—  co  g "*  CO O .g> • -o O S£ c co m O  +^  M" M"  CM d  •  co  So  T3  +^  x: co  OT  3  00 d  =-  roo  <£L >, —  x:  0  CM O  Tt  o  Tt Tt  Tt  d  o d  +^  +^  +^  o o d  in CM d  o d  o o d  +^  +^  +^  +^  CO in d  o d  CO o d  CD CM d  CM  d  d  T— o d  CO co d  o d  +  +  +  +  +  +  +  +  +  iv CM d i  CO d •  Tt  Tt  d  d  o o d  CM d  +  +  +  +  H  I-  Tt  m  <  CO C  "D  "D  >  x:  e  E  E  Q Ka K Ql l - h r - r - r - r -  c  151  ^  A p p e n d i x 7. C h e c k s of e n e r g y b a l a n c e e s t i m a t e s r e p o r t e d in t h e literature: (a) e s t i m a t i o n of B r o w n ' s ( 1 9 6 9 ; 1 9 7 2 ) t h e r m a l conductivity v a l u e f o r g r a v e l s t r e a m b e d s , a n d ( b ) e x a m i n a t i o n o f C o m e r and Grenney's (1977) stream energy balance fluxes.  a) Brown (1972) used a K c of "8 x 10" cal/cm sec °C" for solid green breccia. This is 3  presumably equivalent to the value of "0.39 Btu/ft -inch-min-°F" given by Brown (1969) for 2  breccia, while Brown (1969) gave the thermal conductivity of the "water-gravel mixture" as "0.05 Bru/ft -inch-min-°F". 2  Converting 8 x 10" cal/cm sec ° C t o W m " K " : 3  8 x IO"  3  1  1  cal x 1J x 100 cm =3.4 W i n K" , cmsec°C 0.239 cal lm 1  1  multiplying by 0.05 Btu/ft -inch-min-°F gives Brown's gravel K c as 0.4 W m" K" . 0.39 Btu/ft -inch-min-°F 2  1  1  2  In contrast, calculating a simple weighted mean thermal conductivity for a saturated gravel with a porosity of 0.3, Brown's (1972) value for solid green breccia, and Oke's (1987) value for the K c of water gives a thermal conductivity for gravel that is close to that used in the present study (2.5 vs. 2.6 W m" K" ). 1  1  b) Comer and Grenney (1977) calculated net heat fluxes across the water surface of 4602900 W m" (40-250 cal cm" h" ), which seem excessively high. Net heat fluxes across the 2  2  1  water surface of even poorly-shaded creeks and rivers are usually less than 1000 W m" (Sinokrot and Stefan 1993; Webb and Zhang 1997). Further, fluxes in excess of the solar constant (~1400 W m" ) seem implausible. Hence it appears that Comer and Grenney (1977) 2  may have overestimated all of their heat flux terms by a factor of about three [(2900 W m" )/(1000 W m" )]. This apparent overestimation suggests that Comer and 2  2  Grenney's bed heat conduction fluxes were in the range -80 to -230 W m , rather than the 2  values of-230 to -690 W m" (-20 to -60 cal cm" h" ) presented in their 1977 paper. 2  2  1  Sample conversion of cal cm h" to W m : 2  1  20 cal x 1J x (100 cm) x 1 h = 232 W m" . cm h 0.239 cal 1m 3600 s 2  2  2  152  2  Appendix 8. Longitudinal d i s p e r s i o n coefficients ("D") e s t i m a t e d in six s t u d i e s u s i n g t r a c e r tests, a s a f u n c t i o n o f s t r e a m f l o w : (a) linear plot, (b) log-log plot. T h e linear r e g r e s s i o n in a ) d i d n o t i n c l u d e t h e t w o filled points lying o u t s i d e t h e plot, but t h e s e w e r e i n c l u d e d in all o t h e r fits, (c) L o g - l o g plot of D a s a f u n c t i o n o f m e a n f l o w velocity ("u"). Multiple r e g r e s s i o n o f D a g a i n s t q a n d u d i d not result in a n i m p r o v e d R v a l u e , (d) a n d (e) a r e linear a n d log-log plots, respectively, o f u a s a f u n c t i o n o f s t r e a m f l o w . N o t e that f o u r of t h e s t u d i e s u s e d statistical m e t h o d s o f p a r a m e t e r o p t i m i z a t i o n ( H a r v e y et al. 1 9 9 6 , H a r v e y a n d Fuller 1 9 9 8 , M u l h o l l a n d et al. 1997, a n d this study), w h i l e t w o u s e d s u b j e c t i v e "fit b y e y e " m e t h o d s ( D ' A n g e l o et al. 1993; M o r r i c e et al. 1 9 9 7 ) . Artificial s t r e a m s in D ' A n g e l o ( 1 9 9 3 ) w e r e not i n c l u d e d . 2  (120, 8.3) •  1  3n  T -Eh-B-  Y = 0.0036X + 0.091 r = 0.81 A  /  2  ^  2  10  (2000, 2.5)  /  • JDD  CO  /  CM  Q 1  O0  l  •  '  I  I  l  200 400 600 Streamflow (L s ) — i — i — i — i  11111  Y = 0.016X r 2  ^  =  0  1  0  9  1  o  Jo"v a  •  0.01  •  •  o  1 I I I III  TTTT|  •  1 I I I 11111  1 I I I I lll|  1—n  10 100 1000 Streamflow (L s )  i  •  o a  D ' A n g e l o et al. ( 1 9 9 3 ) H a r v e y e t al. ( 1 9 9 6 )  or A  A  • O  o r !  X  / A A A,  O  3 I  /  V  mil—n:  _1  1 — i — i — i 11 i  •  1  0.1  M i  7  CO  Q  i  8  0  o  800  egg  .49  2  A A  X  0.01  —  6  A&  0.1  '  m iiiMA  | 0  2  1  10  I I I I II  Y = 0.030X r = 0.78  CO  •  AAA/  0  I  Q  A  E  . TTTj  H a r v e y a n d Fuller ( 1 9 9 8 )  O  Morrice e t a l . (1997)  V  M u l h o l l a n d e t al. ( 1 9 9 7 )  O  this s t u d y  •fa "1  1 1—I M i l l  1  1 1—I  10 u (cm s )  If  100  1  100  1 i i I inn  rrm  Y = 4.07X r = 0.55  0  1 I I I IIIIA i—i -  3  i i mil  4  O  §  Or  1 0  •o  i u  • * •  2  CO  i  •  CP "TTJ  0  4 0 0 800 1200 1600 2000 Streamflow (L s ) -1  153  1  i—rniTn|  i  i  i in  (  -rnTrm |  1—rr  10 100 1000 Streamflow (L s ) 1  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090484/manifest

Comment

Related Items