UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Hyporheic exchange processes in a coastal headwater stream Scordo, Elisa Branson 2007

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

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

Item Metadata

Download

Media
831-ubc_2007-0585.pdf [ 8.34MB ]
Metadata
JSON: 831-1.0101232.json
JSON-LD: 831-1.0101232-ld.json
RDF/XML (Pretty): 831-1.0101232-rdf.xml
RDF/JSON: 831-1.0101232-rdf.json
Turtle: 831-1.0101232-turtle.txt
N-Triples: 831-1.0101232-rdf-ntriples.txt
Original Record: 831-1.0101232-source.json
Full Text
831-1.0101232-fulltext.txt
Citation
831-1.0101232.ris

Full Text

H Y P O R H E I C E X C H A N G E P R O C E S S E S I N A C O A S T A L H E A D W A T E R S T R E A M b y E L I S A B R A N S O N S C O R D O B . N R S . , T h o m p s o n R i v e r s Univers i ty , 2003 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E i n T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Geography) U N I V E R S I T Y O F B R I T I S H C O L U M B I A October 2007 © E l i s a Branson Scordo, 2007 ABSTRACT H y p o r h e i c exchange f l o w involves the two-way movement o f water between the stream channel and the bed and banks. These exchange f lows create dist inct ive habitats and influence b iogeochemica l processes and water temperatures. Th i s study focused on the characterizat ion o f the spatial dis t r ibut ion o f subsurface f low pathways and associated travel t imes through the hyporhe ic zone w i t h i n a low-order , high-gradient headwater stream located i n the U B C M a l c o l m K n a p p Research Forest, approximately 60 k m east o f Vancouve r , B r i t i s h C o l u m b i a . H y p o r h e i c zone processes were examined M a y to October 2006, at three spatial scales i n East Creek: point, channel-unit and reach. Hydrome t r i c data col lected from piezometers instal led w i t h i n the stream channel , a long w i t h solute injection tracer experiments, were used to characterize subsurface f low pathways w i t h i n a 100 m stream section. Stream tracer breakthrough curves were used to mode l the processes o f advect ion, dispersion, lateral in f low and transient storage w i t h i n the hyporhe ic zone us ing the numerica l mode l O T I S - P . Tracer injections at i nd iv idua l step-pool units were used to identify locations o f hyporheic discharge, as w e l l as to estimate separate travel t imes for hyporheic and surface-water transient storage zones. Solute transport process var ied w i t h discharge at the reach scale. Transient storage area (As) increased w i t h discharge, w h i l e transient exchange coefficient (a) remained fa i r ly constant. A t the scale o f ind iv idua l pools , transient storage area and residence times were higher than the reach scale estimate, suggesting that pools and back eddies do contribute to transient storage i n headwater streams. Water fluxes calculated w i t h D a r c y ' s L a w i n one channel-unit d id not "scale-up" to the reach scale estimate o f hyporhe ic exchange (a), and was two orders smaller than the reach scale. Di rec t measurements o f water fluxes into the streambed, i nc lud ing ver t ical h y d r a u l i c gradients and inf i l t ra t ion rates, d id not vary systematical ly w i t h discharge. H y d r a u l i c gradients var ied s ignif icant ly w i t h scaled loca t ion w i t h i n the channel-unit, but not w i t h the downstream step height. H y d r a u l i c conduct iv i ty var ied w i t h site condit ions (upwel l ing , d o w n w e l l i n g and neutral sites), suggesting that channel geometry and hydrau l ic conduct iv i ty control exchange f low. T h i s mu l t ip l e scale approach highl ights the considerable spatial and temporal var iab i l i ty and complex i ty o f hyporheic exchange processes w i t h i n step-pool streams. TABLE OF CONTENTS A B S T R A C T . . - i i T A B L E O F C O N T E N T S i i i L I S T O F T A B L E S v L I S T O F F I G U R E S . . . . v i i A C K N O W L E D G M E N T S x D E D I C A T I O N '. x i C H A P T E R O N E : I N T R O D U C T I O N 1 1.1. H y p o r h e i c exchange f low i n smal l streams 1 1.2. Phys ica l controls on exchange f low 3 1.2.1. Bed fo rm scale 4 1.2.2. Channel-uni t scale , 5 1.2.3. Reach scale - Transient storage processes 8 1.2.4. Study objective and scale-dependent questions 12 C H A P T E R T W O : M E T H O D S 14 2.1. Study locat ion .". 14 2.2. Study design 19 2.3. Stream discharge and geometry 21 2.3.1. Discharge measurements 21 2.3.2. Discharge calculations 22 2.3.3. Character iz ing lateral 'exchanges 24 2.3.4. Cross-sect ion measurements 24 2.4. Stream tracer experiments - Reach scale 24 2.4.1. M e t h o d o f injection J ....24 2.4.2. Quant i fy ing poo l storage and residence t ime 25 2.4.3. H y d r a u l i c parameters ( O T I S - P ) 25 2.4.4. Eva lua t ion o f parameter uncertainty 27 2.5. Stream tracer experiments - Channel-uni t scale 28 2.5.1. Qual i ta t ive observations o f hyporhe ic discharge 28 2.5.2. Quant i fy ing residence times 29 2.5.3. M o d e l l i n g mean residence t ime 29 2.6. Subsurface f l ow measurements - Po in t scale 30 2.6.1. Piezometer design and instal lat ion 30 2.6.2. H y d r a u l i c head 31 2.6.3. H y d r a u l i c conduct iv i ty : 32 2.6.4. Re la t ing discharge and recharge zones to stream geometry 34 2.6.5. Stream bed infiltrometers 35 2.6.6. Subsurface relative connect ivi ty 37 2.7. Statistical A n a l y s i s 38 C H A P T E R T H R E E : R E S U L T S 39 3.1. S tudy per iod condi t ions 39 3.2. Da ta qual i ty 41 3.3. Solute transport mode l analysis - Reach scale 42 3.3.1. Summary o f O T I S - P simulations 42 3.3.2. V a r i a b i l i t y o f fitted parameters 44 3.3.3. Parameter uncertainty 45 3.3.4. D e r i v e d quantities 47 3.3.5. Latera l in f low rates 49 3.3.6. Quan t i fy ing p o o l storage and residence t imes. .....50 3.3.7. Subsurface relative connect iv i ty .' : 53 3.4. Solute injection experiments - Channel-uni t scale 54 3.4.1. Qual i ta t ive observations o f hyporhe ic discharge 54 3.4.2. Quant i fy ing residence times 56 3.5. Subsurface f l ow - Po in t scale 58 3.5.1. H y d r a u l i c gradients 58 3.5.2. V H G and scaled locat ion w i t h i n channel units 63 3.5.3. H y d r a u l i c conduct iv i ty 65 3.5.4. Streambed infi l t rat ion rates 67 3.5.5. Streambed water fluxes computed f rom D a r c y ' s l aw 68 C H A P T E R F O U R : D I S C U S S I O N 70 4.1. Reach scale ; 70 4.1.1. M o d e l l e d parameter uncertainty , 70 4.1.2. Solute transport parameters and discharge 72 4.1.3. Residence times and retention 74 4.1.4. In-channel transient storage 75 4.2. Channel-uni t scale 75 4.2.1. V a r i a b i l i t y i n exchange f low pathways 75 4.2.2. Transient storage mode l l i ng 77 4.3. Poin t scale '. 77 4.3.1. Interpretation o f f low pathways 77 4.3.2. Interactions between hyporhe ic f low and lateral in f low : 79 4.3.3. Water fluxes and discharge 80 4.3.4. Sca l ing streambed water fluxes 81 C H A P T E R F I V E : C O N C L U S I O N S 83 5.1. Summary o f m a i n results 83 5.2. Areas for future research 85 R E F E R E N C E S •. 87 A P P E N D I X A : M O D E L S I M U L A T I O N S 93 i v LIST OF TABLES Tab le 3.1: C o m p a r i s o n o f 2006 mean da i ly temperature and month ly precipi tat ion to 30 year c l imate normal (1961-1990) as measured at the H a n e y - U B C Research Forest A d m i n cl imate station (Environment Canada) 39 Tab le 3.2. S u m m a r y o f streamflow (Q) measurements conducted dur ing lower reach ( L R ) , upper reach ( U R ) stream tracer injections over the study per iod M a y 31 to October 20, 2006. Included are the rates o f injection (Rj), slope o f the cal ibra t ion regression (k), and standard error o f k (SE) , electrical conduct iv i ty at background ( E C D k g , ) and plateau (ECp| a t) and the probable error i n stream f low measurement 41 Tab le 3.3: S u m m a r y o f best fit mode l parameters for solute releases i nc lud ing stream discharge (Q) , dispersion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone ( A s ) , storage zone exchange coefficient (a), net lateral i n f low (QL) , the D a m k o h l e r number (Da l ) : 42 Tab le 3.4. S u m m a r y o f uncertainty ratios for the parameter estimates o f dispersion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , and the storage zone exchange coefficient (a) 46 Tab le 3.5. S u m m a r y o f der ived quantities i nc lud ing stream ve loc i ty (w), hydrau l ic residence t ime for the stream ( T s l r ) and storage zone ( T s l o r ) , hydrau l i c uptake length (Si i y d) , hydrau l ic retention factor ( R n ) and the standardized storage zone coefficient ( A s / A ) 47 Tab le 3.6. S u m m a r y o f the simulated parameter estimates o f dispersion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , and the storage zone exchange coefficient (a), stream ve loc i ty (u), the standardized storage zone coefficient ( A s / A ) and the D a m k o h l e r number ( D a l ) 51 Tab le 3.7. M e a n residence times for hyporheic zone (step) and p o o l storage zones i n P o o l 4. Ve r t i c a l hydrau l ic gradients ( V H G ) measured us ing piezometer 61 .56 Tab le 3.8. Spearman correlation coefficient (r s ), associated p-values, number o f observations (n) and vert ical hydrau l ic gradients ( V H G ) for each piezometer ind ica t ing a significant correlat ion w i t h discharge 62 Tab le 3.9. A n a l y s i s o f variance table compar ing three l inear mixed-effects models for ver t ical hydrau l ic gradients i nc lud ing a base mode l w i t h o n l y channel-unit , a second m o d e l w i t h channel-unit and step height ( S H ) , and a third mode l w i t h channel-unit , step-height and channel pos i t ion ( X / L ) . A chi-square (%2) statistic was used to test for s ignif icance 63 v Table 3.10. A n a l y s i s o f variance table compar ing two linear mixed-effects models for hydrau l ic conduct iv i ty ( log transformed), i nc lud ing a base mode l w i t h on ly reach as a factor and a second mode l w i t h reach and site condi t ion (upwel l ing , neutral and downwel l i ng ) 65 Tab le 3.11. Spearman con-elation coefficient (r s), associated p-values and number o f observations (n) for inf i l t rat ion rates versus discharge at each infil trometer locat ion 68 Table 3.12. Water fluxes w i th in one step-pool unit (Poo l 1) a long w i t h scaled-up and reach-scale estimates o f hyporheic exchange (s"1) : 69 Table 4.1. Range o f parameter values reported for high-gradient streams (Wagner and H a r v e y 1997) compared to mode l l ed parameter values i n East Creek 71 v i LIST OF FIGURES Figure 1.0.1. D i a g r a m showing a typ ica l subsurface f low pathway from the stream channel to the hyporheic zone 6 F igure 2 . 1 . Site map showing locat ion o f U B C Research Forest 1 5 F igure 2 . 2 . Site map showing East Creek drainage basin 1 6 F igure 2 . 3 . East Creek study reach wi th upper and lower sub-reaches and stream tracer experiment inject ion and sampl ing locations 1 7 F igure 2 . 4 . E l eva t ion gradient for East Creek study reach downstream from a culvert cross ing at R o a d M '. 1 9 F igure 2 . 5 . Downs t ream v i e w o f l o g steps and piezometers i n upper reach o f East Creek 2 0 Figure 2 . 6 . Ups t ream v i e w o f lower reach o f East Creek showing steps and poo l complexes 2 0 F igure 2 . 7 . M a p o f sampl ing locations, morpho logy and thalweg prof i le for both reaches 3 3 Figure 2 . 8 : Streambed infiltrometer. Adap ted from M a r t i n ( 1 9 9 6 ) 3 5 F igure 3 . 1 : D a i l y precipi tat ion, m a x i m u m and m i n i m u m da i ly temperatures, measured discharge and net lateral i n f low from tracer injections conducted dur ing the study per iod o f M a y to October 2 0 0 6 . Discharge values represent streamflow measured at the l ower reach boundary. Lateral i n f low measured as the difference between upstream and downstream streamflow measurements. N o t e l o g scale for Q and Q L . C l ima te data recorded at the H a n e y - U B C Research Forest A d m i n cl imate station (Envi ronment Canada) 4 0 F igure 3 .2 . M o d e l s imulat ions us ing O T I S - P for June 1 9 for the upper reach (a) and the lower reach (b) .....43 F igure 3 . 3 . S imula ted mode l parameters for solute releases i n the upper and lower stream reach. Di spe r s ion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , storage zone exchange coefficient (a) versus stream discharge ( Q ) . E r ro r bars represent ± 1 standard deviat ion 4 4 F igure 3 .4 . Exper imenta l D a m k o h l e r number ( D a l ) versus stream discharge ( Q ) 4 5 v n Figure 3 .5 . Uncer ta in ty ratio ( U R ) for the s imulated mode l parameters o f dispersion (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , storage zone exchange coefficient (a) versus stream discharge (Q) 46 F igure 3 .6 . H y d r a u l i c residence t ime o f solutes in the stream ( A ) and storage zone (B) versus stream discharge ( Q ) for the upper and lower reach. 4 8 F igure 3 .7 . Stream ve loc i ty ( A ) , hydrau l ic uptake length ( B ) , hydrau l ic retention factor (C) and the standardized storage zone coefficient (D) versus stream discharge ( Q ) for the upper and lower reach 4 9 F igure 3 .8 . Ne t lateral i n f low rates (QL) versus discharge ( Q ) for a l l solute releases 5 0 Figure 3 .9 . M o d e l s imulat ions us ing O T I S - P for September 3 0 . Results are from one poo l locat ion i n the lower reach : 51 Figure 3 . 1 0 . S imulated mode l parameters for solute releases i n the upper and lower stream reach dur ing September 2 9 and 3 0 . Poo l s 1 and 2 were located i n the upper reach. P o o l 3 was located i n the lower reach. Dispe r s ion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , storage zone exchange coefficient (a) versus locat ion. Er ro r bars represent ± 1 standard deviat ion 5 2 Figure 3.1 l . H y d r a u l i c residence t ime o f solutes i n the stream and storage zone for solute releases i n the upper and lower stream reach dur ing September 2 9 and 3 0 . Pools 1 and 2 were located i n the upper reach. P o o l 3 was located i n the lower reach 5 3 Figure 3 . 1 2 . Rela t ive connect iv i ty ( R C ) as measured us ing a non-dimensional m i x i n g ratio (%) for piezometers sampled dur ing tracer injection experiments i n the upper and lower reach 5 4 Figure 3 . 1 3 . Channel-uni t observations o f f l ow pathways i n P o o l 4 , i nc lud ing side v i e w and aerial v i e w 5 6 Figure 3 . 1 4 . Step-pool residence t ime experiment conducted on September 2 5 , 2 0 0 6 5 7 Figure 3 . 1 5 . Step-pool residence t ime experiment conducted on October 5, 2 0 0 6 . 5 7 Figure 3 . 1 6 . Ve r t i c a l hydrau l ic gradients measured i n the upper ( A ) and lower (B) reaches. Symbo l s indicate study per iod means 5 9 Figure 3 . 1 7 . Ve r t i c a l hydrau l ic gradients measured over the study per iod i n piezometers 1-6 i n the upper reach 6 0 Figure 3 . 1 8 . Ve r t i c a l hydrau l ic gradients measured over the study per iod i n piezometers 1 0 , 1 1 , 1 3 - 1 5 i n the upper reach 6 0 V l l l Figure 3.19. V e r t i c a l hydrau l ic gradients measured over the study period^in the piezometers 21-23 and piezometers 24-26 i n the upper reach 61 F igure 3.20. V e r t i c a l hydrau l ic gradients measured over the study per iod i n the piezometers 60, 61 , 62 and 67 to 68 i n the lower reach 62 F igure 3.21. V e r t i c a l hydrau l ic gradient (cm/cm) versus scaled locat ion w i t h i n the channel-unit. H y d r a u l i c gradients are averaged over the entire study per iod. 64 F igure 3.22. V e r t i c a l hydrau l ic gradient (cm/cm) versus step height (m) as a function o f scaled loca t ion w i t h i n the channel-unit ( X / L ) . H y d r a u l i c gradients are averaged over the entire study per iod 64 F igure 3.23. H y d r a u l i c conduct iv i ty ( K ) for d o w n w e l l i n g (D) , neutral (N) and u p w e l l i n g (U) sites located i n the lower (n = 24) and upper reach (n = 17). N o t e l o g scale 66 F igure 3.24. H y d r a u l i c conduct iv i ty ( K ) w i t h depth o f piezometer instal lat ion for the upper and lower reaches. N o t e l og scale 66 Figure 3.25. Infiltration rates over the study per iod. Er ror bars represent probable errors based on Equa t ion 2.25 67 Figure 3.26. H y d r a u l i c conduct iv i ty calculated us ing infi l t rat ion rates and slug-tests for f ive locations. V a l u e s represent the geometric mean ± standard error 68 Figure 3.27. Water fluxes calculated us ing D a r c y ' s L a w for each X / L category w i t h i n one step-pool channel-unit dur ing l o w f low (Q = 1.1 L / s ) and h igh f low (Q = 15.4 L/s). . . . . . . .69 Figure A . l . M o d e l s imulat ions us ing O T I S - P for M a y 31 for the lower reach 93 Figure A . 2 . M o d e l s imulat ions us ing O T I S - P fo r June 27 for the lower reach 93 F igure A . 3 . M o d e l s imulat ions us ing O T I S - P for Sept 21 for the upper reach (a) and the lower reach (b) '. 94 Figure A . 4 . M o d e l s imulat ions us ing O T I S - P for Sept 29 for the upper reach 95 F igure A . 5 . M o d e l s imulat ions us ing O T I S - P for Sept 30 for the lower reach .95 Figure A . 6 . M o d e l s imulat ions us ing O T I S - P for October 20 for the upper reach (a) and the lower reach (b) 96 Figure A . 7 . M o d e l s imulat ions us ing O T I S - P for September 29. Results are from two pools located i n the upper reach 97 i x ACKNOWLEDGMENTS R.D. Moore M. Weiler H. Schreier E. Morgan T. Lagemaat S. Guenther J. Leach A. Bier A. Zimmermann J. Caulkins J. Phillips A. Home Perkins A. See ton UBC Geography/ Forestry National Science and Engineering Research Council of Canada (NSERC-CGM) Forest Investment Account (FIA) La mia famiglia DEDICATION For Claire and Kees, the Adventurers CHAPTER ONE: INTRODUCTION 1.1. Hyporheic exchange flow in small streams H y p o r h e i c exchange f low involves the two-way movement o f water between the active stream channel and subsurface sediments in the stream bed and banks. These exchange f lows play an important role i n the funct ioning o f stream ecosystems. Frequent hyporheic exchange keeps stream water i n close contact w i t h chemica l ly or b i o l o g i c a l l y active stream bed sediments, w h i c h increases the opportunities for b iogeochemica l processing. Interactions between the stream channel and subsurface influence water qual i ty b y generating gradients o f nutrients and d isso lved gases (Bou l ton et a l . 1998) and regulat ing water temperatures ( M o o r e et al . 2005a). The hyporheic zone plays an important role i n ecosystem funct ioning inc lud ing stream metabol i sm ( M u l h o l l a n d et a l . 1997), nutrient retention and c y c l i n g (Tr i ska et a l . 1989, W o n d z e l l and Swanson 1996b), habitat for benthic invertebrates (Stanford and W a r d 1988), and general ecosystem stabil i ty (Valett et al . 1994). Spatial heterogeneity in exchange f low pathways results i n a hyd ro log i ca l l y l i nked region beneath and adjacent to streams and rivers where surface water and subsurface ground water m i x (Tr i ska et a l . 1989, Stanford and W a r d 1993, F i n d l a y 1995). T h i s m i x i n g zone is defined as the "hyporhe ic zone ," i n w h i c h the water chemistry reflects a mixture o f streamwater and groundwater. Exchange f lows through the hyporheic zone l i n k aquatic and terrestrial components o f the riparian ecosystem ( W o n d z e l l and Swanson 1996b). E a r l y attempts to define the phys ica l boundaries o f the hyporheic zone were based on the distributions o f aquatic invertebrates, i n c l u d i n g hypogeon (groundwater or ig in) and epigeon (channel or igin) . T r i s k a et a l . (1989) used solute patterns to operat ional ly define the boundaries o f the hyporheic zone as the depth to w h i c h greater than 1 0 % advected channel water and less than 9 0 % groundwater is present. Beneath the hyporhe ic zone is the groundwater zone where the water chemist ry is not inf luenced b y stream water. Hyporhe i c zones are l i nked to a nested series o f f low paths that can travel both 1 laterally and ver t ica l ly through subsurface f low paths, rather than entering the stream bed i n one locat ion as does groundwater (Harvey and Benca l a 1993, H a r v e y et al . 1996, H a r v e y and Wagne r 2000, Kasahara and W o n d z e l l 2003). H y p o r h e i c exchange creates distinct zones o f aquifer discharge (upwe l l i ng or outwel l ing) f rom the sediments into the stream channel, and recharge (downwel l ing) from the stream channel into the saturated sediments. U p w e l l i n g water can supply stream organisms w i t h the nutrients w h i c h influence p r imary production, w h i l e d o w n w e l l i n g water can provide d issolved oxygen and organic matter to benthic invertebrates l i v i n g i n the sediments (Bou l ton et a l . 1998). D o w n w e l l i n g water also provides oxygen to f ish eggs i n the subsurface (Baxter and Hauer 2000). S m a l l streams (<10 m width) w i t h active exchange between surface and subsurface waters (hyporheic exchange) are thought to facilitate ni t rogen-removal and reduce the export o f nitrate ( N C v ) downstream (Tr i ska et al . 1989, Jones and H o l m e s 1996, D u f f and T r i s k a 2000). Ni t ra te is considered a major pollutant o f aquatic systems in m u c h o f the northern hemisphere ( N R C 2000). T h e compet ing processes o f nutrient retention and hydro log ica l export are expressed in the nutrient spi ra l ing concept (Webster and Patten 1979, N e w b o l d et a l . 1981). Vale t t et al . (1996) proposed a conceptual mode l suggesting that the solute retention is a product o f chemical transformation rates and surface-subsurface interactions w h i c h increase residence times. A d d i t i o n a l studies support the theory that the extent o f subsurface interactions influences solute retention ( M u l h o l l a n d et a l . 1997, H i l l and L y m b u r n e r 1998). Solute retention and residence t ime i n the subsurface depend p r imar i ly on hydrau l ic gradients and hydrau l ic conduct iv i ty . The hyporhe ic zone increases the residence t ime for water w i t h i n the stream ecosystem and enhances transient storage (Benca la 1984). Transient storage refers to the temporary detainment o f solutes i n s low m o v i n g areas such as side pools or back eddies relative to the faster f l o w i n g areas i n the m a i n channel. Solutes, such as nitrogen and phosphorus, m a y also enter the permeable substrate surrounding the stream (i.e. hyporheic zone) and travel at a s lower ve loc i ty than that o f the m a i n channel . Solutes detained w i t h i n transient storage zones are eventually re-released back into the m a i n channel, but at a s lower rate relative to solutes t ravel ing at the advection rate i n the m a i n channel (Jones and M u l h o l l a n d 2000). In the m a i n channel, solutes are transported 2 through the hydro log ica l processes o f advect ion and dispersion (Stream Solute W o r k s h o p 1990). A d v e c t i o n refers to the downstream transport o f solute mass at the mean ve loc i ty o f the streamwater. D i spe r s ion is the spreading o f the solute mass due to shear stress and molecular diffusion in the downstream direct ion. Interactions between the stream channel and subsurface have been a focus o f research for over two decades (Jones and M u l h o l l a n d 2000). Natura l r i f f le -pool and step-p o o l units have been c o m m o n l y studied as channel morpho logy that exerts a strong control on hyporheic exchange. Current knowledge o f hyporhe ic zone processes is based largely on studies conducted w i t h i n smal l (<10 m width) , a l luv ia l , l o w to m i d order gravel-bed, headwater streams w i t h r i f f le-pool (e.g. T r i s k a et a l . 1989, H a r v e y et a l . 1996, H i l l et a l . 1998, Wagner and Bretschko 2002) and step-pool morphologies (e.g., H a r v e y and B e n c a l a 1993, W o n d z e l l and Swanson 1996a, Vale t t et a l . 1996, Hagger ty et a l . 2002, Kasahara and W o n d z e l l 2003, Storey et a l . 2003, A n d e r s o n et a l . 2005), as w e l l as w i t h i n high-order, l o w gradient, braided to meandering, r iver systems (e.g. Stanford and W a r d 1988, Boano et a l . 2006). T h i s thesis is concerned w i t h the characterization o f the spatial dis t r ibut ion o f subsurface f l ow pathways and associated residence t imes through the hyporhe ic zone w i t h i n a low-order , high-gradient headwater stream. Sect ion 1.2 reviews the current literature on the phys ica l controls o f hyporheic exchange to provide the background for the specific research objectives presented i n Sec t ion 1.3. 1.2. Physical controls on exchange flow Surface-subsurface interactions, or hyporhe ic exchange, are dr iven b y variat ions i n hydrau l ic head gradients as a result o f instream structural complex i ty created from large w o o d y debris (e.g. l o g j a m s ) and geomorphic features such as step-pool sequences, or breaks i n topography (Harvey and B e n c a l a 1993). H y p o r h e i c f low varies considerably over space scales (1 c m - 100 m) and t ime scales (10 s - 100 days) and at var ious rates through different types o f substrate (Harvey et a l . 1996). Depend ing on loca l geology and channel morpho logy , the extent o f the hyporheic zone can range i n length f rom centimeters to hundreds o f meters (Stanford and W a r d 1993). L o c a l channel features 3 i nc lud ing sediment compos i t ion , permeabi l i ty (hydraul ic conduct ivi ty) and the bed topography control the lateral extent o f the hyporhe ic zone be low the saturated stream channel (Tr i ska et a l . 1993). Exchange f lows occur at different spatial scales i nc lud ing those o f (1) ind iv idua l bedforms, (2) channel units and (3) reaches. The f o l l o w i n g sections w i l l further examine the exchange f low processes occurr ing at the different spatial scales. 1.2.1. Bedform scale Topograph ic features k n o w n as bedforms (riffles and dunes) develop due to streamflow over a loose sediment bed. Obstructions or irregularities i n the streambed such as sand riffles (Johnson 1980) or even f ish redds (Tonina and Buf fmgton 2005) create a high-pressure zone upstream o f the obstruction and a l o w pressure region downstream. F l u m e studies conducted at this scale have shown that f low is induced b y pressure imbalances generated from gradients o f temperature, density and hydrostatic head (Thibodeaux and B o y l e 1987, E l l i o t t and B r o o k s 1997). The process o f solutes and water f l o w i n g between high-pressure and low-pressure zones i n the bed is referred to as "advect ive pumping exchange" (Savant et a l . 1987, Thibodeaux and B o y l e 1987). Wate r also moves through the sediments through the process o f "turnover". Turnover occurs as m o v i n g bedforms trap and release interstitial f lu id . Studies in laboratory f lumes have indicated that hyporheic exchange rates increase w i t h discharge, spec i f ica l ly w i t h stream f low ve loc i ty , due to an increase i n the pressure difference between h igh and l o w pressure regions (Thibodeaux and B o y l e 1987, E l l i o t t and B r o o k s 1997). W o n d z e l l (2005) speculated that interactions between streamflow and channel bedforms must dr ive exchange f low i n headwater streams; unfortunately field studies have not been able to incorporate finer scale effects (as studied i n flumes) into a cohesive framework for exchange i n headwater streams. A s a result, the influence o f ind iv idua l bedforms on hyporhe ic exchange f low has not been investigated w i t h i n a field setting. 4 1.2.2. Channel-unit scale A t the scale o f ind iv idua l channel units, morpholog ica l features i n the stream channel, such as large w o o d y debris, create head gradients that dr ive advect ion o f stream water through the hyporheic zone. Several authors have shown that longi tudina l gradients i n step-pool and r i f f le-pool sequences dr ive smal l scale exchange f low both ver t ica l ly and lateral ly (Harvey and B e n c a l a 1993, H i l l et a l . 1998, Storey et al . 2003, A n d e r s o n et a l . 2005, G o o s e f f e t a l . 2006). H y p o r h e i c exchange has been described conceptual ly as short pathways that enter the subsurface and return to the stream channel at mul t ip le locations (Harvey and B e n c a l a 1993). Stream water f l o w i n g through wel l -def ined f low pathways i n the a l l u v i u m m a y enter a streambed at the top o f a r iff le or step and then return to the stream a short distance downstream i n the bot tom o f p o o l (Figure 1.1). F l o w begins when the total head i n the surface channel is greater than that i n the subsurface, resul t ing i n a negative vert ical hydraul ic gradient ( V H G ) , w h i c h drives stream water d o w n into the subsurface sediments. Surface water m a y m i x w i t h or displace groundwater and eventually return to the surface where the total head i n the subsurface is greater than the stream channel . Hyporhe i c exchange creates dist inct zones o f aquifer discharge (upwel l ing or outwel l ing) from the sediments into the stream channel , and recharge (downwel l ing) from the stream channel into the saturated sediments. F l o w paths m a y travel ver t ica l ly between the channel to the subsurface or laterally from the adjacent r iparian zone (Whi te 1993, F i n d l a y 1995). W o n d z e l l and Swanson (1996) a long w i t h Kasahara and W o n d z e l l (2003) further identif ied channel-unit features w h i c h dr ive hyporheic exchange, such as side channels, meander bends, gravel bars and boulder or log-steps. Kasahara and W o n d z e l l (2003) found that steps accounted for approximately 5 0 % o f the exchange f lows i n second and fifth-order stream reaches based on the results o f a sensi t ivi ty analysis us ing groundwater f l ow models . Channe l morpho logy has also been documented as a significant control for lateral hyporhe ic exchanges (Verv ie r et a l . 1993, M o r r i c e et a l . 1997, Storey et a l . 2003). These studies suggest that channel-unit form and geometry are significant controls o n hyporhe ic exchange f low. 5 Figure 1.0.1. D i a g r a m s h o w i n g a typ ica l subsurface f low pathway from the stream channel to the hyporheic zone. Studies examin ing hyporheic exchange f low at the channel-unit scale t yp i ca l ly employ a hydrometr ic approach. T h i s approach requires an extensive network o f piezometers and/or we l l s to measure hydrau l ic gradients and hydraul ic conduct iv i ty i n order to characterize and map exchange f lows. Studies are therefore l imi ted to a smal l spatial area (e.g. H a r v e y and B e n c a l a 1993), and inter-reach comparisons are chal lenging. A s a result, B e n c a l a (2000) expressed the need to identify the phys ica l and hydrometr ic properties o f the stream system that contribute to solute transport w i t h i n the hyporheic zone, and that can be rout inely measured or map across spatial scales. Recent studies have begun to examine channel-unit spacing i n stream longi tudinal profiles to predict the spacing between zones o f u p w e l l i n g and d o w n w e l l i n g i n step-pool and p o o l -riffle morphologies us ing a hydrometr ic (Ander son et a l . 2005) or mode l l i ng approach ( G o o s e f f e t a l . 2006). In both these studies, channel-unit spacing, s ize and sequence were considered important controls i n determining exchange patterns o f u p w e l l i n g and d o w n w e l l i n g . These results suggest that a sca l ing relationship to identify zones o f u p w e l l i n g and d o w n w e l l i n g based on channel-unit geometry w o u l d be a useful tool for character izing 6 and predict ing exchange f low i n step-pool streams. A n objective o f this thesis is to develop a geometric sca l ing relat ionship relating vert ical hydrau l ic gradients to channel-unit geometry i nc lud ing p o o l length and downstream step height, in order to determine i f hyporheic discharge and recharge zones are a function o f stream channel locat ion. W h i l e it is k n o w n that structural complex i ty from large w o o d y debris and geomorphic features such as step-pool sequences dr ive exchange f low, there are s t i l l uncertainties regarding the spatial patterns o f hyporheic f low, and the locations where hyporheic water discharges back into the stream. Recent studies o f hyporheic exchange i n steep, headwater streams i n the L o o k o u t Creek basin (Oregon, U S A ) have general ly not observed coherent u p w e l l i n g o f hyporheic water be low steps, as described b y the typ ica l f low pathway, despite predict ions from groundwater f low models that u p w e l l i n g should occur (Anderson et a l . 2005 , G o o s e f f et al . 2005, W o n d z e l l 2005). These results suggest that hyporheic discharge occurs under different mechanisms, i nc lud ing lateral in f low or ou twe l l ing f rom other locations w i t h i n the step-poo l . F o r example, M o o r e et a l . (2005b) observed u p w e l l i n g sites w i t h i n a concentrated zone o f lateral in f low that was consistent w i t h the convergent topography and h i l l s lope o f a section o f headwater stream i n coastal B r i t i s h C o l u m b i a . Solute tracer tests indicated that these u p w e l l i n g sites underwent litt le to no m i x i n g w i t h water f rom the stream channel , suggesting that u p w e l l i n g sites are a result o f lateral i n f low. In this scenario it is hypothesized that f low pathways cou ld include an interaction between groundwater or lateral i n f low and hyporhe ic exchange pathways. A n objective o f this thesis is to determine w h i c h mode l best conceptualizes exchange f low w i t h i n steep, step-pool streams. Subsurface f l ow pathways that create zones o f discharge (i.e. u p w e l l i n g or outwel l ing) and recharge (i.e. downwe l l i ng ) w i t h i n step-pool units can be described us ing three different conceptual models : 1. M o d e l l a - Represents a typica l f low pathway i n w h i c h d o w n w e l l i n g o f water occurs at the top o f a r iff le or step and returns to the stream channel a short distance downstream i n the bot tom o f poo l . F l o w is a l igned w i t h the channel creating u p w e l l i n g i n the poo l . M o d e l s l b and 2 represent possible alternative hypotheses to the typ ica l f l ow pathway and cou ld exp la in the lack o f observed u p w e l l i n g in previous studies. 7 2. Mode l l b - H y p o r h e i c f l ow is al igned w i t h the channel and dr iven by ver t ical hydrau l ic gradients as per M o d e l l a , but f l ow creates u p w e l l i n g or ou twe l l ing sites at another locat ion wi th in the step-pool, such as d i rect ly be low the step. M o d e l s l a and l b do not inc lude a lateral i n f low component, w h i c h is represented by M o d e l 2. 3. Mode l 2 - Hyporhe i c exchange f low includes a lateral i n f low component i n w h i c h zones o f u p w e l l i n g are a result o f lateral in f low from the r iparian zone and adjacent h i l l s lope due to convergent topography. A s w e l l , f l ow pathways cou ld include an interaction between groundwater or lateral i n f low and hyporheic exchange pathways. F o r example , hyporheic water cou ld f l ow lateral ly into the r iparian zone after infi l t rat ing i n a step, then f low laterally into the channel . 1.2.3. Reach scale - Transient storage processes H y p o r h e i c exchange processes are typ ica l ly studied at the reach scale us ing a transient storage mode l ( T S M ) consis t ing o f a one-dimensional advect ion-dispersion equation wi th an addit ional term for transient storage (Benca la and Wal ters 1983, R u n k e l 1998). T h e T S M provides reach-scale estimates o f the solute transport processes o f advection, dispersion, transient storage and lateral i n f low, by s imula t ing the breakthrough curves ( B T C ) generated f rom stream tracer injections (e.g. D ' A n g e l o et a l . 1993, H a r v e y and Benca l a 1993, M o r r i c e et a l . 1997, M u l h o l l a n d et a l . 1997, G o o s e f f et al . 2003). A d d i t i o n a l l y , T S M s have been w i d e l y applied to l i n k hydro log ica l transport parameters to b io log ica l processes (e.g. Vale t t et al , 1996, M u l h o l l a n d et a l . 1997, Goose f f et a l . 2004). Transient storage includes (1) in-channel storage i n "dead zones" (i.e. side pools , back eddies) and (2) storage i n the hyporhe ic zone. Current T S M s , such as the One D i m e n s i o n a l Transport w i t h Inf low and Storage ( O T I S - P ; R u n k e l 1998) mode l , l u m p both transient storage zones together into a single model -der ived estimate o f the cross-sectional area o f the transient storage zone (As) . Exchange between the m a i n channel and the transient storage zone is control led b y a transient exchange coefficient (a), w h i c h is considered an estimate o f hyporheic exchange. T y p i c a l l y , the model -der ived transient storage area (As) parameter is assumed to represent storage i n the streambed, or the hyporhe ic zone (Benca la et a l . 1993), due to the inab i l i ty to separate in-channel storage from storage i n the hyporheic zone. However , previous studies have suggested that both transient storage mechanisms occur at the reach scale (Harvey and B e n c a l a 1993, G o o s e f f et a l . 2003, W o n d z e l l 2005). Harvey and Benca l a (1993) observed a difference i n the transient storage residence-times estimated from stream tracer experiments and hydrometr ic w e l l breakthrough dynamics , and concluded that at the reach-scale solute transport processes are sensitive to both transient storage mechanisms. Attempts to dis t inguish between in-channel storage and hyporheic transient storage have resulted i n the development o f T S M s wi th two storage or mul t ip le exponential residence t ime distributions (Castro and Hornberger 1991, C h o i et a l . 2000, G o o s e f f et a l . 2004). However , these models are often diff icult to correctly parameterize us ing current field techniques (Runke l 2002). A s a result, there is a need to better understand processes at the point and channel-unit scale for interpretation o f reach scale tracer tests, par t icular ly i n terms o f the relative roles o f hyporheic exchange and transient storage i n pools . Solutes detained w i t h i n dead zones are eventually re-released back into the ma in channel , but at a s lower rate than the ma in solute pulse (Jones and M u l h o l l a n d 2000). T h i s s low release is manifested b y a long tail i n the breakthrough curves (Benca la and Wal ters 1983, Stream Solute W o r k s h o p 1990, R u n k e l 1998, Chapra and R u n k e l 1999, C h o i et a l . 2000). T h e O T I S - P transient storage mode l assumes that the late-time residence times o f solutes w i t h i n these dead zones are exponent ia l ly distributed (Runke l 1998). H o w e v e r , recent workers have shown residence t ime distributions that are better characterized us ing a lognormal dis tr ibut ion ( W o r m a n et a l . 2002), or d isp layed scale-invariant behavior , w i t h the under ly ing residence t ime f o l l o w i n g a power l a w dis tr ibut ion (Haggerty et a l . 2002). G o o s e f f et a l . (2003) characterized exchange processes i n three reaches at L o o k o u t Creek basin us ing two solute transport models w i t h different residence t ime distr ibutions (exponential and power- law) . Transient storage w i t h i n a bedrock reach fo l l owed an exponential d is t r ibut ion w i t h a mean residence t ime o f three hours. In contrast, transient storage i n an a l luv ia l reach fo l l owed a power l aw dis t r ibut ion w i t h a 9 mean residence t ime o f > 100 hours. A l t h o u g h this study makes the assumption that the a l luv ia l reach represents both in-stream and transient storage processes, whereas the bedrock reach represents on ly in-channel transient storage due to the absence o f a hyporheic zone, these results show that in-ehannel features such as pools and back eddies do contribute to transient storage in smal l , headwater streams. Harvey et al . (1996) assumed that the residence-time for in-channel transient storage to be very short and is therefore accounted i n the dispersion coefficient rather than the transient storage coefficient. Howeve r , G o o s e f f et a l . (2003) highl ights the var iab i l i ty i n residence t imes and late-time distributions that can occur depending on the dominant transient storage mechan i sm at the reach scale. A l t h o u g h it is sometimes possible to el iminate one o f the transient storage mechanisms based on morpho logy (i.e. a l luv ia l versus bedrock reach), studies have not been able to isolate the relative contr ibut ion from each transient storage zone i n stream reaches where both processes occur together. A s w e l l , the contr ibut ion from in-channel storage has on ly been examined at the reach scale, and not at the channel-unit scale w i t h i n i nd iv idua l pools us ing a stream tracer approach. In stream reaches where pools are located, no previous study has speci f ica l ly treated an ind iv idua l pool , as a "stream reach" to test the assumptions o f the transient storage mode l . Th i s appl icat ion o f the T S M m a y help to characterize the spatial extent and residence times o f transient storage as measured from stream tracer experiments. A key objective o f this thesis is to quantify transient storage and estimate separate travel times for hyporheic and surface-water transient storage zones. Several studies have sought to identify the factors that control transient storage and hyporheic exchange, inc lud ing channel morpho logy (Harvey and Benca l a 1993, D ' A n g e l o et a l . 1993, Kasahara and W o n d z e l l 2003, G o o s e f f et a l . 2003, W o n d z e l l 2005), discharge (Harvey et a l . 1996, W o n d z e l l 2005, Zarnetske et a l . 2007), parent material ( M o r r i c e et a l . 1997), groundwater in f lows (Harvey and B e n c a l a 1993, W r o b l i c k y et a l . 1998) and stream complex i ty (Patschke 1999, G o o s e f f et a l . 2007). D ' A n g e l o et a l . (1993) used the stream tracer approach to investigate the influence o f geomorphologica l constraint on transient storage w i t h i n a fifth-order stream i n the L o o k o u t Creek basin. Transient storage was lower i n constrained stream reaches (n = 5) as compared to reaches that were unconstrained (n = 2). T h i s study also reported a 10 decrease i n transient storage w i t h stream order (first-order to fifth-order), indica t ing that transient storage is var iable w i t h spatial scale. M o r r i c e et a l . (1997) compared three headwater streams w i t h different parent l i thologies and showed that the extent o f hyporheic exchange increased w i t h increased hydrau l ic conduct iv i ty . T h i s study also suggested that conduct iv i ty and infi l t rat ion capacity influences the amount o f water that infiltrates the stream bed and enters the hyporheic zone. W o n d z e l l (2005) conducted stream tracer experiments i n two smal l , steep-mountain streams i n the H . J . A n d r e w s Exper imenta l Forest (Oregon, U S A ) at l o w basef low and h i g h basef low condit ions. T h i s study found that transient storage var ied w i t h channel morpho logy and to a lesser extent w i t h discharge. A n increase i n stream discharge is hypothesized to result in a greater wetted area, w h i c h provides for enhanced hyporheic exchange. In headwater streams, the influence o f discharge on hyporhe ic exchange and spatial extent w i l l vary w i t h catchment wetness, h i l l s lope connect iv i ty and the strength o f hydrau l ic gradients to the stream channel ( W o n d z e l l and Swanson 1996a). W r o b l i c k y et a l . (1998) applied a 2 - D groundwater f l ow mode l to study the surface-subsurface interactions w i t h i n two first-order headwater streams i n N e w M e x i c o ( U S A ) , and found that the spatial extent o f the hyporhe ic zone decreased b y approximate ly 5 0 % dur ing higher f lows. A sensit ivi ty analysis indicated that exchange f lows were governed b y the interplay o f extr insic (e.g. va r iab i l i ty i n precipi tat ion and recharge) and intr insic geomorphologica l and hydro log ica l factors (e.g. hydrau l ic conduct iv i ty) . Stream tracer experiments have conf i rmed that solute transport processes vary w i t h discharge at the reach scale ( L e g r a n d - M a r c q and Laudelout 1985, D ' A n g e l o et a l . 1993, H a r v e y et a l . 1996, M o r r i c e et a l . 1997, Hart et al . 1999, Patschke 1999, W o n d z e l l 2005); al though conf l i c t ing responses have been documented. M o r r i c e et a l . (1997) compared transport processes dur ing 4 tracer releases w i t h i n one headwater stream reach, and observed a decrease i n transient storage (As) w h i l e transient exchange (a) increased w i t h increasing discharge. S i m i l a r l y , D ' A n g e l o et a l . (1993) found that the transient exchange coefficient increased w i t h discharge and that the ratio o f storage zone to cross-sectional area ( A s / A ) decreased w i t h discharge. A d d i t i o n a l studies have reported s imi la r trends (Harvey et al . 1996, Hart et a l . 1999, W o n d z e l l 11 2005), however trends that are i n contradict ion to those studies have also been documented (Legrand-Marcq and Laudelout 1985, Hart et a l . 1999, Patschke 1999). Patschke (1999) observed an increase i n A s w i t h discharge i n reaches wi th a greater degree o f stream complexi ty . Hart et a l . (1999) reported that transient storage remained constant w i t h discharge, but transient exchange increased as discharge increased. Tracer tests by Leg rand -Marcq and Laudelout (1985) showed a rapid decrease i n A s as discharge increased to a threshold value o f 2 L / s , after w h i c h transient storage remained constant to a value o f 12 L / s . L e g r a n d - M a r c q and Laudelout (1985) and Patschke (1999) observed that the transient exchange coefficient remained constant w i t h discharge. These conf l ic t ing results i n the literature h ighl ight the uncertainty surrounding the response o f transient storage area and exchange parameters to discharge i n smal l headwaters streams. Part o f this uncertainty is attributed to the T S M , w h i c h is sensitive to experimental design (Wagner and H a r v e y 1997) and parameter f i t t ing routines (Runke l 1998). H a r v e y et a l . (1996) explained that because stream ve loc i ty increases w i t h discharge, a majori ty o f the observed differences i n parameters for intra-reach or inter-reach comparisons w i l l result from a swi tch i n dominance o f advect ive f l ow over transient storage, and not from a change i n phys ica l processes. F o r example, W o n d z e l l (2005) found that the results o f T S M simulat ions were at odds to hydrometr ic data, w h i c h showed li t t le change i n the locat ion and extent o f the hyporhe ic zone despite a four-fold increase i n discharge (1.0 - 11.5 L /s ) . These results are s l ight ly confounded b y h igh uncertainty i n parameter values as indicated by D a m k o h l e r values w h i c h were >2 (ranged f rom 2.4 to 21.1) for 5 o f the 9 simulated tracer experiments. H o w e v e r , W o n d z e l l (2005) cautioned that T S M comparisons should be restricted to stream tracer experiments performed w i t h i n different reaches o f a s ingle stream under comparable f l ow condit ions. 1.2.4. Study objective and scale-dependent questions The broad objective o f this study is to examine h o w hyporhe ic exchange processes vary spat ial ly and temporal ly w i t h i n a low-order , high-gradient headwater stream under a range o f f l ow condit ions. T h i s study focused on the characterization o f the 12 spatial d is t r ibut ion o f subsurface f low pathways and associated travel times through the hyporhe ic and.surface-water transient storage zones us ing both a hydrometr ic and stream tracer approach. A significant challenge o f hyporheic zone research has been to scale up smal l scale phys i ca l measurements to the results o f reach-scale stream tracer injections. A s a result, specif ic questions are appl ied to the spatial scale o f interest i nc lud ing the reach scale, the channel-unit scale and the loca l or point scale. A mul t ip le scale approach to examin ing hyporhe ic exchange processes has not been exp l i c i t l y appl ied i n previous research. It is expected that the var iab i l i ty i n solute transport processes observed at the reach scale, spec i f ica l ly transient storage area (As) and the transient exchange coefficient (a), can be expla ined us ing observations made at smaller spatial scales. The f o l l o w i n g research questions w i l l be addressed i n this thesis: 1. H o w do solute transport processes at the reach-scale, speci f ica l ly transient storage area and transient exchange, vary over space and t ime (i.e. w i th discharge)? O f specific interest is the contr ibution o f hyporheic and surface-water transient storage zones to transient storage area (As) . 2. , C a n separate residence t imes for hyporheic and surface-water transient storage zones be quantified us ing a stream tracer approach at the channel-unit scale? 3. Where are zones o f hyporheic discharge and recharge located wi th in the channel-unit , and do exchange f lows vary w i t h pos i t ion i n the channel-unit and downstream step height? 4. H o w do water fluxes at the point scale vary, and can these be quantified and "scaled u p " to reach scale estimates o f hyporheic exchange? 4' The f o l l o w i n g four chapters examine the research question i n detail . Chapter 2 describes the study area, f ie ld and laboratory methods, and data analysis. Results are presented i n Chapter 3 and are discussed i n Chapter 4. The m a i n conclusions o f the study and recommendat ions for future research are summar ized i n Chapter 5. 13 CHAPTER TWO: METHODS T h i s chapter provides a detailed descript ion o f the study locat ion and study reaches. A n overv iew o f the study design, a long w i t h the methods used for f ie ld sampl ing , laboratory analysis and data analysis are also described. 2.1. Study location T h i s study was conducted i n the Un ive r s i t y o f B r i t i s h C o l u m b i a ( U B C ) M a l c o l m K n a p p Research Forest (49° 16' N , 122° 34 ' W ) , located i n the temperate Fraser V a l l e y foothi l ls o f the Coast Mounta ins , approximately 60 k m east o f Vancouve r , B r i t i s h C o l u m b i a (Figure 2.1). The research forest is located wi th in the Coasta l Wes tern H e m l o c k (Tsuga heterophylla) b iogeocl imat ic zone (Coupe et al . 1991). T h i s zone is characteristic o f the mar i t ime cl imate w i t h wet, m i l d winters and warm, dry summers. M e a n annual precipi tat ion at the U B C Research Forest headquarters (147 m elevation) is approximately 2184 m m (Environment Canada 1993), o f w h i c h 7 0 % falls p r i m a r i l y as ra in between October and A p r i l due to Pac i f ic frontal systems. Snowfa l l comprises o n l y 5% o f the total annual precipi tat ion at the headquarters. T h e higher elevat ion areas are snow covered for approximately four months o f the year. St reamfiow is t yp i ca l ly very " f lashy" i n response to rainfal l . Periods o f l o w f low dominate dur ing the summer months (base f low <1 L / s ) . Per iods o f peak f low correspond to fal l and winter storms. R u n o f f generation processes are dominated b y subsurface f l ow and saturation over land f low (Hutch inson and M o o r e 2000). So i l s w i t h i n the watershed are h i g h l y permeable, and are dominan t ly sha l low • podzols (averaging about 1 m deep) formed from glac ia l t i l l ! S o i l parent material is g lac io f luv ia l i n or ig in ( K l i n k a and Kra j i na 1986). So i l s are typ i ca l ly under la in b y basal t i l l over granitic bedrock ( K l i n k a and K r a j i n a 1986). In some locations, the soils direct ly over l ie bedrock. G e o l o g i c a l condit ions have contributed to the observed l o w background concentrations o f nitrogen, phosphorus and l o w conduct iv i ty i n surface water. 14 C h e m i c a l l y the study stream is nearly pristine wi th l o w concentrations o f major ions (Fel ler and K i m m i n s 1984). Vegetat ive cover a long East Creek is dominated b y coniferous forest consis t ing o f Douglas- f i r (Pseudotsuga menziesii) and western cedar {Thuja plicata), that are approximately 130 years o ld . R e d alder (Alnus rubra) occurs a long the stream banks in patches. Select ive forest harvesting occurred adjacent to the stream i n the 1970's , but instream large w o o d y debris was retained in the stream. The study was conducted w i t h i n the East Creek watershed, a 4 k m long second-order stream that drains approximate ly 100 ha at the study reach (F igure 2.2). East Creek is located upstream from the confluence o f Spr ing Creek. Research was conducted from M a y to October, 2006. Hyporhe i c exchange processes were examined w i t h i n a 100 m long section o f East Creek , w h i c h extends downstream from a culvert c ross ing at road M to the stream's confluence at Spr ing Creek (Figure 2.3). T h i s stream was selected due to its h igh structural c o m p l e x i t y and variety o f geomorphic features, w h i c h inc lude approximate ly 7-10 pool-cascade sub-units w i th in the reach (Patschke 1999). Structural complex i ty f rom in-stream large w o o d y debris and heterogeneous substrate deposi t ion has contributed to dist inct pool-cascade sequences w i t h both boulder and l o g steps. F igure 2.1. S i t e m a p s h o w i n g locat ion o f U B C Research Forest 15 I I I I I I I 0 0.1250.25 0.5 Kilometers Projection: UTM Zone 10 Datum: NAD 1983 1:15,500 Legend ==== Roads Streams Lakes Cutblocks/Thinning Study Location Figure 2.2. Site map s h o w i n g East Creek drainage bas in 100 Meters Projection: UTM Zone 10 Datum: NAD 1983 1:2,500 Legend = Roads Streams Contour Lines (5m) it Reach Upper Boundary (EC sample) • Reach Lower Boundary (EC sample) A Tracer Injection Site Figure 2.3. East Creek study reach with upper and lower sub-reaches and stream tracer experiment injection and sampling locations The reach has an elevation drop o f approximately 20 m (Figure 2.4), and varies i n morpho logy and substrate a long the study length (Figure 2.5, F igure 2.6). T h e channel w i d t h for the reach is 2.5 - 3 m wide . A large sediment dam exists at m i d reach, w h i c h divides the reach into two sections. A b o v e the sediment dam, the top 45 m o f the reach has a gradient o f approximately 4%. Stream morphology consists o f l o w gradient riffle-run and pool sequences wi th in natural l og steps, and the substrate is composed o f fine to coarse gravel and m e d i u m size cobbles. D u r i n g periods o f l o w f low, the stream goes subsurface through preferential f low pathways i n the sediment dam (at approximate ly 42 m downstream from the culvert at M road). D i r e c t l y be low the l o g j a m , the stream is braided for approximate ly 5-10 m . A t the beg inn ing o f the study per iod, seepage f rom a cutbank was observed at this locat ion. The bot tom 45 m section has a gradient o f approximately 12% and contains p o o l -cascade sequences w i t h boulder and l o g steps. F i n e to m e d i u m gravels i n deposi t ion areas behind steps and large anchor cobbles and boulders contribute to a spat ial ly heterogeneous substrate. D u r i n g periods o f h igh f low (> 100 L / s ) , a secondary channel at approximate ly 55 m downstream from the culvert becomes active. Sediment transport processes i n East Creek vary seasonally w i t h discharge. Since 2003, approximate ly 10 to 15 sediment m o b i l i z i n g events per year have been observed, o f w h i c h 3 to 4 peak rainfal l events dur ing the fa l l and winter months o f each year move the majori ty o f sediment (J. C a u l k i n s , pers. comm.) . These events result in the greatest geomorphic changes to the channel . 18 0 20 40 60 80 100 D i s t a n c e d o w n s t r e a m (m) Figure 2.4. E leva t ion gradient for East Creek study reach downstream from a culvert crossing at R o a d M 2.2. Study design T h i s study employed a combina t ion o f hydrometr ic data and solute injection experiments to characterize subsurface f low pathways. Tracer experiments were designed: (1) to examine the extent o f stream-subsurface interactions at the reach scale (Stream Solute W o r k s h o p 1990), (2) to measure the relative connect ivi ty o f the subsurface to the stream channel, (3) to quantify storage i n i nd iv idua l pools dur ing reach scale tracer injections, and (4) to provide reach scale estimates o f the processes o f advection, dispersion, lateral exchange and transient storage. Stream tracer experiments were conducted at a range o f f low condit ions to examine h o w transport processes vary w i t h discharge. Tracer injections at i nd iv idua l step-pool units were used to identify locations o f hyporhe ic discharge, as w e l l as to estimate separate travel times for hyporhe ic and surface-water transient storage zones. Hydromet r i c data inc luded streamflow, hydrau l ic head, hydrau l ic conduct iv i ty measured us ing piezometers, a long w i t h direct measurements o f inf i l t rat ion into the stream bed. 19 Figure 2 .6. Ups t ream v i e w o f lower reach o f East Creek showing steps and p o o l complexes 2.3. Stream discharge and geometry 2.3.1. Discharge measurements Streamflow, or discharge, was measured us ing the constant-rate salt injection method ( M o o r e 2004a). Th i s method involves injecting a conservative solute tracer solut ion o f k n o w n concentration into a stream at a constant rate and then measur ing the stream water electrical conduct iv i ty ( E C ) as it becomes un i fo rmly m i x e d across the stream some distance be low the injection point. Downs t ream m i x i n g was moni tored us ing a conduct iv i ty probe to ver i fy lateral and ver t ical m i x i n g at a downstream locat ion. Tracer injections were performed accord ing to procedures out l ined b y M o o r e (2004a). S o d i u m chlor ide ( N a C l ) or c o m m o n table salt was used as a tracer because it is inexpensive, readi ly avai lable and envi ronmenta l ly benign for the concentrations and durations invo lved i n discharge measurement ( M o o r e 2004a). Ch lo r ide is a conservative i o n that occurs i n l o w background concentrations (0.6-1.2 m g / L ) in East Creek (Fel ler and K i m m i n s 1984). S o d i u m chlor ide was injected into East Creek to raise the concentration above background unt i l the concentrat ion reached plateau (approximately 2-3 m g / L ) . Solut ions o f N a C l and water were m i x e d i n the laboratory us ing pre-weighed bags o f salt. In the f ie ld , solutions were v igorous ly shaken to ensure that the added salt was complete ly d issolved. Solut ions were prepared at concentrations be low the saturation point o f 238 g / L (Webster and E h r m a n 1996). F o r injection trials pr ior to Augus t 11, solutions were added at a constant rate us ing a Mar io t t e bottle (Story et al . 2003, M o o r e 2004b) . The Mar io t te bottle consisted o f a 30 L carboy sealed at the top w i t h a rubber stopper. The release rate was control led us ing a pipette connected to a spigot at the base o f the Mar io t t e bottle. Release rates were measured for each tracer experiment both pre-inject ion and post-injection since changes i n barometric pressure can affect the release rates (Webster and E h r m a n 1996). B e g i n n i n g on Augus t 11, a battery-operated Sol ins t M o d e l 410 peristaltic pump connected to a 30 L d rum was used to conduct longer duration injections dur ing base f low condit ions. A voltage regulator attached to the 12 vol t battery was used to prevent fluctuations i n the release rate due to changes i n the battery voltage. 21 Elec t r i ca l conduct iv i ty was measured as a surrogate for tracer concentration dur ing stream tracer tests. E C increases l inear ly w i t h salt concentration and is considered an economica l alternative to direct measurements o f chlor ide (Goose f f and M c G l y n n 2005, W o n d z e l l 2005). E C was measured at the downstream end o f the stream reach dur ing discharge measurements us ing either a W T W ™ L F 3 4 0 , 340i or 350i conduct iv i ty meter. A non-l inear cal ibrat ion bui l t into the W T W ™ meter was used to correct automatical ly the E C values to a standard temperature o f 2 5 ° C . Downs t ream changes i n conduct iv i ty were recorded every 30 seconds unt i l a steady-state plateau was reached (Stream Solute W o r k s h o p 1990). T h e L F 3 4 0 and 340i W T W ™ conduct ivi ty meters were attached to a C a m p b e l l Scient i f ic C R 5 1 0 data logger to record measurements. E C values measured us ing the W T W ™ 350i meter were recorded to the meter 's internal memory . Once steady-state was achieved, as determined as a constant conductivity reading for more than 10 minutes at the farthest downstream location, the injection was stopped, and conduct iv i ty recordings continued unt i l the stream water returned to background levels to characterize the " f a l l i n g l i m b " o f the B T C . 2.3.2. Discharge calculations Discharge was calculated as: Q = (2.1) k • (ECss - ECBG) where Q is stream discharge (L/s ) , q is the injection rate o f tracer solut ion (L/s ) , k is the cal ibrat ion coefficient, E C s s is steady state conduct iv i ty (u.S/cm), and E C B G is the background or pre-release conduct iv i ty (uS/cm) . The cal ibrat ion coefficient (k) represents the slope o f the relat ion between relative concentration ( L / L ) and electrical conduct iv i ty (f iS/cm). E lec t r i ca l conduct iv i ty is l inear ly related to relative concentration ( R C ) for di lute solutions. A s a result, the relative concentration at steady state (RCss ) can be determined f rom E C measurements: RCss = (ECss - ECBG) -k (2.2) 22 Secondary and cal ibrat ion solutions were created i n the f ie ld for each tracer experiment i n order to determine k. T o perform the cal ibrat ion, two 1-L vo lumes o f stream water were first measured us ing a glass vo lumet r ic flask and poured into sample bottles. The secondary solut ion consisted o f 10 m L o f the p r imary inject ion solut ion, measured us ing a glass pipette, m i x e d wi th 1 L o f stream water. The secondary solu t ion was then added i n 10 m L increments into a separate 1 L o f stream water to create the cal ibrat ion solution. E lec t r ica l conduct iv i ty was recorded w i t h each addi t ion o f secondary solut ion into the cal ibrat ion solution. Re la t ive concentration was calculated as: RC = S, RC VT (2.3) where S A is the vo lume o f secondary solut ion added ( m L ) , R C s e c is the relat ive concentration o f the secondary solut ion (10 m L / 1 0 1 0 m L ) , and V T is the total v o l u m e o f the cal ibrat ion solution. I f assumptions o f the constant rate injection method (i.e complete m i x i n g , steady-state plateau) are met, the method can measure f lows w i t h i n ± 5% uncertainty (Johnstone 1988). Howeve r , the potential error associated w i t h the discharge measurements was calculated us ing an error analysis (Story 2002): Q + ^Sk^2 + 5(ECss - ECBG) ECss - ECBG (2.4) where: 8(ECss - ECBG) = ^{5ECssf +(5ECBG)2 SQi Qi a KK J + a Kt J where sources o f error inc luded injection rates (Q;),.electrical conduct iv i ty measurements and the slope o f the regression (k) relat ing electr ical conduct iv i ty to the relat ive concentration. The error associated w i t h the injection rates (8Qj/Qj) was estimated based on the error associated w i t h measuring the injection rate (r) and t ime (t).The uncertainty w i t h reading the graduated cyl inder (8r) was estimated as ± 0.5 m L ; the uncertainty w i t h measur ing t ime (8t) was ± 2 s (1 s on either o f the t iming) . T h e average injection rate 23 was 74 m L / m i n (n = 14). Er ro r associated w i t h the measurement o f the electr ical conduct iv i ty at plateau (ECss ) and background ( E C B G ) were 0.1 uS / c m for E C s s <200 LiS /cm and 1 LIS / c m for E C s s >200 LiS /cm. Standard error o f the slope (k) f rom the regression was used to calculate the error i n k. 2.3.3. Characterizing lateral exchanges Constant-rate injection tracer experiments also provided estimates o f the rate o f lateral in f low. Lateral i n f low rates were estimated as the difference i n discharge measurements (Equat ion 2.1) as measured at upstream and downstream E C sampl ing locations w i t h i n the reach (Figure 2.3) d iv ided by the reach length: QL=Qtb~Q,a (2.5) L w h e r e QL is the net lateral i n f l o w rate (Ls" 'm" ' ) , Qds an&Qus are streamflow (L/s ) measured at the downstream and upstream locations w i t h i n the reach respectively, and L is the reach length (m). 2.3.4. Cross-section measurements Cross sections were measured at 10 m intervals a long the stream reach. A t each locat ion, the depth, measured at 20 c m intervals across the stream, and the wetted channel w i d t h were recorded. Measurements o f depth and channel w i d t h were used to calculate the stream cross-sectional area. T h e reach averaged cross-sectional area was used as an input var iable for hyporhe ic zone m o d e l l i n g at the reach scale. 2.4. Stream tracer experiments - Reach scale 2.4.1. Method of injection Stream tracer tests were conducted us ing the constant-rate salt inject ion method i n order to examine the extent o f surface-subsurface interactions at the reach scale. Wagne r 24 and H a r v e y (1997) determined that the constant injection method w i t h sampl ing through the concentration rise, plateau and fal l provides more rel iable parameter estimates than other sampl ing designs (e.g. s lug injections). Tracer injection points and sample locations were the same for a l l f l ow condi t ions (Figure 2.3). Tracer injections were in i t i a l ly conducted b y separating the study stream into two approximately 50 m sections (i.e. lower and upper reach) and inject ing above the upper boundary for each reach. Tracer experiments were conducted on the same day or w i t h i n 1-2 days depending on condit ions. F u l l reach tracer experiments were also conducted for ease o f f ie ld work . Stream tracer was injected above the upper boundary o f the upper reach (approximately 1 0 m downstream from the culvert at M road) and the breakthrough curves were measured at the lower boundaries o f the reaches. D u r i n g base f low condit ions (Q < 2 L / s ) , the reach was again separated into two sub-reaches to conduct longer duration tracer experiments. 2.4.2. Quantifying pool storage and residence time D u r i n g two reach-scale tracer injections, electrical conduct iv i ty meters were placed at the i n f low and outf low locat ion o f three p o o l sub-units i n order to quantify poo l storage and residence time. Breakthrough curves from ind iv idua l pools were s imulated us ing O T I S - P . S imula t ions were conducted for two pools located i n the upper reach (September 29) and one poo l located i n the lower reach (September 30). E a c h p o o l was s imulated as a distinct reach. 2.4.3. Hydraulic parameters (OTIS-P) Transport and transient storage mechanisms i n the stream reach were analyzed b y fi t t ing a numer ica l mode l to breakthrough curves generated from stream tracer experiments. T h e processes o f advect ion, dispersion, lateral i n f l o w and transient storage w i t h i n the hyporhe ic zone were mode led us ing the O T I S - P code, w h i c h was developed b y the U S G e o l o g i c a l Survey (Runke l 1998). O T I S - P , a mod i f i ed vers ion o f O T I S (One-d imens iona l Transport w i t h Inf low and Storage), numer ica l ly solves finite difference 25 approximations to Equation 2.6 with the Crank-Nicolson Method, and uses a nonlinear least squares method to optimize parameters by minimizing the sum of squared errors. The model uses a modification of a one-dimensional advection-dispersion model with additional terms for groundwater inputs and transient storage (Bencala and Walters 1983, Stream Solute Workshop 1990). The following set of differential equations (Runkel and Broshears 1991) is used by OTIS-P to solve for solute mass balance between the main channel (Equation 2.6) and the transient storage zones (Equation 2.7): dC QdC I d , , „ 3 C . aim . • „ „ — = -—— + -—(AD—) + ^ —(CL - C) + a(Cs - C) (2.6) dt A ax A dx dx A — = -a — (Cs-Cs) (2.7) dt As where A is main channel cross-sectional area (L 2 ); As is storage zone cross-sectional area (L 2); C is main channel solute concentration (M/L 3 ) ; C L is lateral inflow solute concentration (M/L ); Cs is storage zone solute concentration (M/L ); D is dispersion 2 3 3 coefficient (L /T); Q is volumetric flow rate (L IT); quN is lateral inflow rate (L /T-L) ; t is time (T) and x is distance (L); q is storage zone exchange coefficient (1/T). Cross-sectional area (A, m2) as entered into the model is calculated as mean of the summed products of mean wetted depth and channel width for each cross-section. The model estimates a reach-averaged A to match observed solute concentration data. Average stream velocity (u, m/s) is calculated as Q/A, using the model-estimated A value. 2 2 Dispersion (D, m Is), reach-averaged transient storage zone cross-sectional area (A s , m ), and transient storage exchange coefficient (a, s"1) are estimated using the model. Transient storage refers to any zone within the channel, such as an eddy, pool, or hyporheic zone, where some water is temporarily detained relative to the faster moving water near the center of the channel. Model parameters solved using OTIS-P were used to derive the following quantities: (1) hydraulic uptake length (Mulholland et al. 1994), (2) hydraulic residence time (or contact time) in the storage zone and stream (Mulholland et al. 1994), (3) the hydraulic retention factor (Morrice et al. 1997), and (4) the standardized storage zone area (Stream Solute Workshop 1990, D'Angelo et al. 1993). Derived quantities are often 26 used to compare solute transport processes between stream reaches (Stream Solute W o r k s h o p 1990). T h e hydraul ic residence t ime, T s l o r (s), in the storage zone can be calculated as: Tstor = — (2.8) Aa whereas residence t ime i n the stream ( T s t r y i s calculated as: Tsir = - (2.9) a T h e hydraul ic uptake length (Shyd) is the average distance a water molecu le travels downstream before entering the storage zone: SM = -Q- (2.10) Aa The hydraul ic retention factor ( R H ) is a measure o f the storage zone residence t ime per unit o f stream reach traveled: R„ = — (2.11) Shyd T h e standardized storage zone area ( A s / A ) is the ratio o f storage cross-sectional area to stream cross-sectional area and is the mathematical equivalent o f storage zone residence t ime to stream residence t ime. 2.4.4. Evaluation of parameter uncertainty Uncer ta in ty surrounding the parameter estimates was examined us ing two indices: the experimental D a m k o h l e r number (Da l ) and the uncertainty ratio. The D a l number was calculated as: ^ T a(l + AI As)L Dal = ——= — (2.12) 27 where L is length o f the stream reach over (m). Wagne r and H a r v e y (1997) showed that when the relationship between the exchange rate and advection deviates f rom 1.0, the uncertainty i n the mode l led parameters increases. Parameter uncertainty is m i n i m i z e d when D a l is close to 1.0. H i g h values may occur because exchange wi th the streambed is re la t ively fast compared to the water ve loc i ty or the reach length m a y be too long. A D a l less than 1.0 indicates that o n l y a smal l amount o f the stream tracer is exchanging w i t h the storage zone. S m a l l D a l numbers (<0.1) cou ld result f rom: (1) h igh stream veloc i ty , f (2) long exchange t ime scale as indicated b y a l o w a and A s / A ratio, and/or (3) short reach length. T h e uncertainty ratio for each estimated parameter ( D , A , A s , a) is equal to the parameter estimate d iv ided by its standard deviat ion. A l o w ratio indicates that the parameter estimate is h igh ly uncertain.-\ 2.5. Stream tracer experiments - Channel-unit scale > 2.5.1. Qualitative observations of hyporheic discharge Tracer injections o f the dye tracers B r i l l i a n t B l u e F C F (C. I . 42090) and Rhodamine W T ( R W T ) were conducted at three locations ( I I , 14 and 15; F igure 2.8) us ing infiltrometers instal led w i t h i n the stream bed as described i n Sec t ion 2.5.6. Injections were used to identify qual i tat ively the locat ion o f hyporhe ic discharge w i t h i n ind iv idua l channel units. A p p r o x i m a t e l y 1-5 g o f R W T was injected, and less than 50 m L o f br i l l iant blue was injected at one t ime. Exper iments us ing R W T were conducted over s ix dates (July 11, Augus t 14, Augus t 21 , September 27/28, October 6) i n the lower reach at two locations (1-4,1-5). Tracer inject ion experiments were also conducted in the upper reach at one locat ion (1-1) o n four dates (Ju ly 12, Augus t 23, and September 28, October 6). 28 2.5.2. Quantifying residence times S o d i u m chlor ide was also used as a tracer to quantify solute residence t ime w i t h i n one boulder step-pool unit at the lower reach. A solut ion o f 1 L de ionized water and 20 g o f salt was added di rect ly to the infil trometer 15 (Figure 2.8) installed w i t h i n the . streambed. E lec t r i ca l conduct iv i ty was measured at the base o f the step (or p o o l in f low) and at the p o o l out f low locat ion to quantify residence t ime w i t h i n the hyporhe ic zone and pools respectively. Cal ibra t ions were conducted us ing the same procedure as described i n section 2.3.2. T w o channel-unit experiments were conducted over the study per iod (September 25 and October 5, 2006). 2.5.3. Modelling mean residence time A s imple m o d e l i n g approach was used to mode l the mean residence t ime o f water w i t h i n the transient storage zones. The system was modeled us ing l inear reservoir theory assuming that the poo l unit behaves as a continuously stirred tank reactor ( C S T R ) (Chapra 1997). Comple te m i x i n g o f the solute and steady-state water f low for the duration o f the tracer inject ion experiment was assumed. A mass balance for the p o o l can be expressed as: ^ = « J 0 - ^ ( 0 (2-13) dt where M(t) is the mass o f tracer i n poo l (kg), qin(t) is the input o f tracer (kg/s) and qout(0 is the output (kg/s). M a s s can be related to tracer concentration us ing the equations: M(t) = V-c(t) (2.14) qo,At) = Q-c0HI{t) (2.15) where Vis the v o l u m e o f water i n poo l (m 3 ) , c(t) is the concentration o f tracer i n the p o o l 3 3 at t ime t (kg/m ), Q is the water discharge at the outlet o f the p o o l (m /s), and cout(t) is the tracer concentration at the outlet. U s i n g C S T R theory, the mean residence t ime ( M R T ) o f solutes w i t h i n the system can be expressed as: 29 V 1 MRT = — = - (2.16) Q k where A: is a first order exchange coefficient ( s 1 ) . T o w a r d the end o f the experiment, tracer discharge to the p o o l w i l l become smal l , and the relat ionship between concentration and time can be mode l led as a first-order reaction: ^ - = -k-c(t) " (2.17) dt Integrating Equa t ion 2.17, subject to the in i t ia l condi t ion c(tG) = c 0 , y ie lds: c(t) = c 0 e _ * ( ' - ' o ) ' • (2.18) where Co is the concentration at t ime = to, w h i c h is an arbitrari ly selected t ime. T h i s equation specifies an exponential deplet ion o f the tracer concentration over t ime. Equa t ion 2.18 can be transformed b y tak ing logari thms o f both sides to y ie ld : ln[c(0] = l n [ c 0 ] - M < - O • (2-19) I f a plot o f the logar i thm o f concentration against t ime yields a straight l ine , Equa t ion 2.15 holds true, and k can be calculated as the slope o f a straight l ine fitted to the l inear port ion o f the log-transformed breakthrough curve. 2.6. Subsurface flow measurements - Point scale 2.6.1. Piezometer design and installation Piezometers were instal led i n a dense network near the center o f the channel i n s ix step-pools to provide a h i g h spatial resolut ion o f subsurface f low pathways (Figure 2.7). A total o f 68 piezometers were instal led i n the stream dur ing early summer ( M a y to June, 2006). T w o types o f piezometer designs were used: (1) a l u m i n u m piezometers for taking water qual i ty samples (n = 28), and (2) plast ic piezometers to measure hydrau l ic gradients and measure hydrau l ic conduct iv i ty (n = 40). 30 A l u m i n u m piezometers were constructed us ing an approximately 60 c m length o f 1 c m internal diameter a l u m i n u m tube w i t h a slot zone o f 5 c m . A l u m i n u m piezometers were dr iven into the streambed us ing a sledgehammer. H y d r a u l i c gradients were also measured for the a l u m i n u m piezometers. Plas t ic piezometers were approximately 60 c m in length and constructed us ing 0.7 c m internal diameter p o l y v i n y l chlor ide ( P V C ) pipe wi th a 5 c m slot zone. T h e slot zones were screened w i t h n y l o n mesh to prevent c logg ing w i t h fine sediments. Plas t ic piezometers were instal led into the streambed b y first d r iv ing i n a steel rod, contained w i t h i n a metal sleeve. T h e metal rod was then removed from the sleeve and the plast ic piezometer was placed inside. Once the metal sleeve and rod were removed, the piezometer was left imbedded i n the subsurface. T h e piezometers were installed at mul t ip le depths (0-15, 15-30, 30-50 cm) both longi tudinal and perpendicular to the wetted stream channel . T h e locat ion and elevation o f a l l piezometers were surveyed us ing a L e i c a G e o s y s t e m s © total station. 2.6.2. H y d r a u l i c h e a d Ver t i c a l hydrau l i c gradients ( V H G ) i n the streambed were used to locate areas o f u p w e l l i n g and d o w n w e l l i n g w i t h i n the hyporheic zone. V H G was calculated as: Ah V H G = — (2.20) Al where Ah is the elevat ion o f the water i n the piezometer minus the elevat ion o f the stream water surface (cm), and Al is the distance between the surface o f the stream bed and the m i d d l e o f the slot zone (cm). Pos i t ive V H G indicates u p w e l l i n g hyporheic or groundwater f l ow (f low potential f rom the bed towards the channel); negative V H G indicates d o w n w e l l i n g f l ow ( f low potential is directed from the channel into the bed). Reversa l o f exchange potential was defined as a shift between posi t ive and negative V H G . Neut ra l piezometers were defined as hav ing hydrau l ic gradients between -0.05 to 0.05 c m / c m , w h i c h l ie w i t h i n the uncertainty o f V H G measurements (Guenther 2007). H y d r a u l i c heads were measured us ing a water l eve l sensor. The sensor consisted o f two electrical wires attached to a rod and connected to a battery and buzzer . The rod was 31 lowered into the piezometer. When the wires contact water, the electrical circuit is closed and a buzzer sounds. The water depth was measured using a measuring tape, resulting in an accuracy of ± 0.05 cm. 2.6.3. Hydraulic conductivity Piezometers were also used to measure saturated hydraulic conductivity of the bed sediments using a Hvorslev test (also known as falling head slug test; Freeze and Cherry 1979). Water was added to the piezometer using a syringe connected to a tube and inserted to the top of the piezometer. The tube was disconnected, and the time required for the water to return to a specified level on the piezometer was measured using a stopwatch. Hydraulic conductivity (K) was computed based on the empirical equation of Hvorslev (1951) as modified by Baxter et al. (2003) for closed-bottom perforated piezometers: K^SfM"'1^ '• (2.21) (t2-t\) where: SF = (L) \ 2 d2 - In + f + where Hi and H2 are the head in the piezometer (cm) at time t/ and (s), SF is the shape factor for the piezometer intake (m), L is the length of the perforations (m), D is the diameter of the perforated intake (m) and d is the inside diameter of the piezometer (m). Hydraulic conductivity measurements of stream sediments are typically reported to be positively skewed or not normally distributed (Ryan and Boufadel 2007). As a result, the geometric mean (rather than arithmetic mean) was calculated. The mean was computed from of all falling head tests taken at each site (n =4). 32 24 P o o l ! ^ Upper Reach i i i i i i i 0 1.25 2.5 5 Meters Projection: UTM Zone 10 Datum: NAD 1983 ^ J 2 / Infiltrometer (35) Piezometer Thalweg Channel Morphology Pool Step Figure 2.7. M a p o f s ampl ing locations, morphology and tha lweg prof i le for both reaches 33 2.6.4. Relating discharge and recharge zones to stream geometry Calculated vertical hydraulic gradients (Equation 2.20) were used to map zones of hyporheic discharge (upwelling) and recharge (downwelling) along the stream profile in ArcGIS version 9.1. A geometric scaling relationship was used to describe vertical hydraulic gradients as a function of location in the stream channel based on mapping results. The general form of the relation is: C-^ = f£,SH) (2.22) dl L where X is the distance from the upstream end of the pool to the piezometer (m), L is the distance from the upstream end of the pool to the edge of the step downstream of the pool (m), and SH is the height of the downstream step (m), as defined by Zimmermann and Church (2001). It is hypothesized that dh/dl should be negatively related to X / L , with positive values for values near 0, and increasingly negative values as X / L approaches 1. Furthermore, hydraulic gradients in the downwelling zone should be negatively associated with SH. That is, higher steps should exhibit stronger downwelling gradients. A total of 7channel-units were used for the analysis. The average water flux for one step-pool unit (Pool 1) was calculated under baseflow (September 29) and high flow conditions (June 19) using Darcy's law: dh q = -K^A . (2.23) dl dh where q is the volume of water (L/s), K is the average hydraulic conductivity.(m/s), — dl is the average hydraulic gradient (cm/cm), and A is the area of stream channel (m ) as determined by the X / L category. The area above and below the step was segmented into spatial units based on the X / L category. The average flux for each segment was computed and totaled for each channel sub-unit. The total channel-unit flux was divided by the wetted channel volume (L) of the reach in order to obtain a scaled-up estimate of hyporheic exchange (s"1). This estimate was compared to the reach-scale estimates of 34 transient exchange obtained by fi t t ing the O T I S - P mode l to tracer inject ion breakthrough curves. 2.6.5. S t r e a m b e d i n f i l t r ome te r s Direct measurements o f stream bed infi l tration rates were conducted using a constant-head stream bed infi l trometer (Figure 2.8). Infiltrometers were constructed us ing an approximate ly 20-30 c m long cy l ind r i ca l , open-ended, P V C pipe w i t h a 7.5 c m internal diameter. A hole d r i l l ed into the P V C pipe at m i d length was used to connect the pipe to a Mar io t t e cy l inde r us ing a piece o f tygon tubing. T h e P V C pipe was installed into the stream bed at a depth o f 5-10 c m such that the mid- length opening to the P V C pipe was l eve l w i t h the stream bed. T h e Mariot te tube was instal led ver t i ca l ly i n the stream us ing a piece o f rebar hammered into the stream bed. T h e Mar io t t e tube was f i l l ed w i th water pr ior to conduc t ing measurements. The basic premise for the Mar io t te tube is that when water w i t h i n the P V C pipe "infil trates" into the bed, the Mar io t t e tube supplies addit ional water to main ta in a constant water leve l . The Mar io t te tube consisted o f a 30 c m length o f plastic tube w i t h both a plastic stopper at both ends. T h e bot tom end o f the tube attached to the P V C pipe installed i n the stream, w h i l e the stopper at the top was used to h o l d a smaller diameter tube in place. 20 cm Figure 2.8: Streambed infil trometer. Adap ted from M a r t i n (1996). 35 Infiltrometers were installed into the stream bed i n five d o w n w e l l i n g locat ions w i t h i n the reach, spec i f ica l ly above the step w i t h i n a step-pool unit (Figure 2.7) to measure the amount o f water direct ly infil trating into the hyporheic zone. Infi l trat ion measurements were taken per iod ica l ly during the study per iod at a range o f f lows. The infi l t rat ion rate (IR) was calculated as: IR Ah n{r{ - r 2 2) A ' K(r2o) (2.24) where Ah is the change in water level i n the Mar io t te tube (cm) over an interval , At (s), ri, rj and rp are the inside radius o f the Mar io t te reservoir (cm), the outside radius o f the bubbler tube i n the Mar io t te reservoir (cm), and the inside radius o f the pan (cm), respectively. The uncertainty associated w i t h the infi l t rat ion measurements was calculated us ing an error analysis: I S(Ah) Y , (S(At))2 , (2(r, + r 2 • * 2 ) Y + f 2 r ~ S r ^ Ah + At + 2 \ 1/2 J V (2.25) where b(Ah) is the error associated w i t h measuring the change in water l eve l i n the Mar io t te tube, estimated at ± 0.2 c m ; 8(At) is the uncertainty w i t h measur ing t ime, estimated at ± 2 s ( l s on either o f the t iming) ; &>/ , or 2 and orp are the uncertainty i n measuring the inside radius o f the Mariot te reservoir, the outside radius o f the bubbler tube i n the Mar io t t e reservoir, and the inside radius o f the evaporat ion pan, respectively. Infiltrometers were installed adjacent to piezometers (approximately 10 cm) i n order to back-calculate hydrau l ic conduct iv i ty (Ki): IR VGH (2.26) where IR is the measured infi l t rat ion rate (cm/s) and V G H is the vert ical hydrau l ic gradient (cm/cm) measured us ing Equat ion 2.20. Infiltrometer measurements o f hydrau l ic conduct iv i ty were then compared to measurements o f conduct iv i ty as measured i n the piezometers us ing the fal l ing-head test. 36 2.6.6. Subsurface relative connectivity Subsurface water samples were collected f rom the a l u m i n u m piezometers to measure the relative connect iv i ty o f the subsurface to the stream channel dur ing reach scale tracer injections. Piezometers were in i t i a l ly purged and a l lowed to ref i l l pr ior to sampl ing. A polypropylene syringe (60 m L ) attached to tygon tubing (3 -mm inside diameter) was used to wi thdraw water samples from the piezometers. S m a l l vo lumes o f water (approximately 15-30 m L ) were removed for purging. A vo lume o f approximate ly 25-30 m L was removed to analyze for electrical conduct iv i ty . Samples were removed i n smal l quantities to m i n i m i z e the influence on the subsurface f low system. The syringe, tubing and sample bottles were de ionized . Water samples were taken pr ior to tracer release and again once the stream concentration reached plateau. The electrical conduct iv i ty o f the water samples was measured to calculate relative connect ivi ty . A two-component m i x i n g equation us ing the electrical conduct iv i ty o f stream water and piezometer porewater at stream tracer ini t ia t ion (t = 0) and steady state were used to calculate the fraction o f stream water present in the subsurface (Goose f f and M c G l y n n 2005). The percent stream water i n the hyporheic zone was calculated as: where E C P ( t ) , EC P (o), EC s ( t ) and E C s ( o ) refer to the conduct iv i ty i n the piezometers and stream at steady state (t) and background (time 0) respectively. T h e m i x i n g ratio w i t h i n the hyporheic zone was calculated for a l l piezometers. It is assumed that no stream water has exchanged w i t h the subsurface when % is equal to 0; when % is equal to 1, complete replacement o f the hyporhe ic zone water w i t h tracer labeled stream water has occurred. ECp(D — ECp(0) • 100% (2.27) X = ECs(t) — ECs(0) 37 2.7. Statistical Analysis V a r i a b i l i t y o f exchange f low was assessed b y compar ing changes i n V H G and infi l t rat ion rates over the range o f f low condit ions observed. Spearman's rank correlat ion analysis was used to explore the relationships between discharge, ver t ical hydraul ic gradient and infi l t rat ion at each point locat ion. In addi t ion, V H G as a funct ion o f the downstream step height was also tested us ing this approach. A value o f 1 or -1 indicates a strong posi t ive (or strong negative) correlat ion between two nonparametric variables (Kutner et a l 2004). A l inear mixed-effects mode l ( M a i n d o n a l d and B r a u n 2007) was used to examine the in-site va r iab i l i ty i n hydrau l ic gradients and hydrau l ic conduct iv i ty us ing the L M E R -function i n R 2.5.1 for unbalanced experimental designs (R Deve lopment Core T e a m 2007). T o mode l the var iab i l i ty i n hydraul ic gradients, channel-unit (n = 7) was considered a random effect wh i l e downstream step height (7 levels) and pos i t ion w i t h i n the channel (as defined b y X / L ; 5 levels) were f ixed effects. Three separate l inear models were created i nc lud ing (1) a base mode l w i t h on ly channel-unit, (2) channel-unit w i t h step height and (3) channel-unit , step height and X L factors. A sequential analysis o f variance was used to determine the signif icance o f the f ixed effects on hydrau l i c gradients b y compar ing a l l three models us ing the A N O V A function i n R 2.5.1. Da ta were weighted b y the inverse o f X / L to account for the heteroscedasticity o f the residuals (Kutner et a l . 1994). T h i s statistical method was also used to examine the var iab i l i ty i n hydrau l ic conduct iv i ty due to site condi t ion (i.e. upwe l l i ng , d o w n w e l l i n g or neutral) and reach loca t ion (i.e upper or lower reach). Reach (n = 2) was considered a random effect w h i l e site condi t ion was a f ixed effect. Da ta were l o g transformed to account for the heteroscedasticity o f the residuals (Kutner et a l . 1994). A s ignif icance leve l o f 0.05 was used for a l l analyses. 38 CHAPTER THREE: RESULTS T h i s chapter presents the results o f f ie ld observations i n East Creek between M a y and October, 2006. The chapter starts w i th an o v e r v i e w o f the study period condit ions (section 3.1) and discuss ion o f the data qual i ty (section 3.2). Results are then presented for each scale o f invest igat ion beginning w i t h the reach scale (Sect ion 3.3), fo l lowed b y the channel-unit scale (section 3.4), and the i nd iv idua l point scale (section 3.5). The chapter concludes w i t h a summary o f the key f indings (section 3.6). 3.1. Study period conditions Based on data from the H a n e y - U B C Research Forest A d m i n cl imate station (Envi ronment Canada), located w i t h i n the U B C Research Forest (49° 1 6 . 2 ' N , 122° 3 4 . 2 ' W ) , summer temperature and precipi ta t ion condit ions were warmer and drier than the 30 year no rm (Table 3.1). F igure 3.1 provides an overv iew o f f low condit ions in East Creek a long w i t h da i ly precipi ta t ion and temperature i n the Research Forest. Stream tracer experiments were conducted f rom M a y 31 to October 20, 2006, over a range o f f low condit ions to examine hyporheic exchange processes over t ime. A majori ty o f the tracer injections were conducted dur ing basef low condit ions (Q <5 Ls" 1 ) when the contr ibut ion o f the hyporheic zone is hypothet ica l ly m a x i m i z e d . Streamflows ranged f rom 0.21 L / s on September 13 to 30.6 L / s on M a y 31 over the study per iod. Discharge levels were too h igh b y early N o v e m b e r ( Q >200 L / s ) to continue the study. Table 3.1: C o m p a r i s o n o f 2006 mean da i ly temperature and month ly precipi tat ion to 30 year cl imate normal (1961-1990) as measured at the H a n e y - U B C Research Forest A d m i n cl imate station (Envi ronment Canada). M e a n da i ly temperature (°C) Tota l precipitat ion (mm) M o n t h N o r m a l s 2006 Difference N o r m a l s - 2006 % o f N o r m a l M a y 11.8 12.7 +0.9 114 121.8 106.8 Jun 14.6 16.1 +1.5 93.1 55.2 59.3 Ju l 16.8 18.7 +1.9 80.9 27.2 33.6 A u g 17.0 17.8 +0.8 74.3 17.0 22.9 Sep 14.5 15.7 +1.2 119.7 90.8 75.9 Oct 9.9 11.6 • +1.7 223.8 32.6 14.6 39 o \ May 01 Jun 01 Jul 01 . Aug 01 Sep 01 Oct 01 Nov 01 Figure 3 .1 : D a i l y precipi tat ion, m a x i m u m and m i n i m u m da i ly temperatures, measured discharge and net lateral in f low from tracer injections conducted dur ing the study per iod o f M a y to October 2006. Discharge values represent streamflow measured at the lower reach boundary. Lateral i n f low measured as the difference between upstream and downstream streamflow measurements. No te l og scale for Q and Q l , C l ima te data recorded at the H a n e y - U B C Research Forest A d m i n cl imate station (Envi ronment Canada). 40 3.2. Data quality Error rates var ied between 3-7 % for a l l f l ow condit ions (Table 3.2). These error rates are roughly consistent w i th the cited uncertainty o f ± 5% (Johnstone 1988). D u r i n g three tracer injections conducted dur ing l o w f low condit ions (Ju ly 20, Augus t 13 and 31), the lower reach boundary d id not achieve plateau. Streamflow at the lower reach boundary was then estimated as a percentage o f streamflow as measured at the upper reach boundary. T h i s was based on the relationship between the measured discharge at the upper and lower reach boundaries for ful l reach tracer injections that reached plateau (n = 4). The increase i n discharge was estimated at 2 0 . 5 % o f the upstream discharge. M o d e l s imulat ions were o n l y conducted for tracer injections where the upper and l ower reach boundary concentrations reached plateau (n = 10). Table 3.2. S u m m a r y o f streamflow (Q) measurements conducted dur ing lower reach ( L R ) , upper reach ( U R ) stream tracer injections over the study per iod M a y 31 to October 20, 2006. Included are the rates o f injection (R;), s lope o f the cal ibrat ion regression (k), and standard error o f k (SE) , electrical conduct iv i ty at background (ECbkg,) and plateau (ECpi at) and the probable error i n stream flow measurement. Date Q R k S E ECbkg ECp| a i Probable Reach (L/s) (mm/mi) ( io- 6 ) (kx 10"9) (jj,S/cm) (u.S/cm) Error (%) M a y 31* 30.6 71 2.8 3.0 20.0 34.0 3.6 L R June 7 24.9 34 5.4 6.8 20.4 24.6 5.0 L R June 19* 15.4 33 1.7 14 22.6 43.2 3.8 U R June 19* 17.2 23 3.4 1.7 22.3 27.8 4.8 L R June 27* 5.5 59 3.3 13 23.8 79.0 3.5 U R / L R July 20 2.1 87 3.3 13 29.0 270.0 3.4 U R / L R A u g 13 0.53 40 9.2 12 29.6 174.0 3.6 U R A u g 31 0.33 116 130 3.0 38.8 84:0 4.0 L R Sept 13 0.21 70 52 3300 35.4 143.0 7.1 L R Sept 21* 9.9 109 6.6 8.3 31.1 58.9 3.4 U R / L R Sept 29* 1.1 92 21 60 33.0 102.0 3.7 U R Sept 30* 1.9 92 21 60 32.4 71.9 3.4 L R Oct 20* 11.9 104 4.5 140 30.2 74.6 4.5 U R / L R Tracer injections simulated b y O T I S - P 41 3.3. Solute transport model analysis - Reach scale 3.3.1. Summary of OTIS-P simulations Best fit mode l parameters for s ix tracer experiments conducted i n the lower reach and four tracer experiments conducted i n the upper reach are summarized i n Table 3.3. A l l tracer experiments doubled or quadrupled the background electrical conduct iv i ty o f stream water to a plateau concentration. F igure 3.2 shows an example o f s imulated and observed tracer concentrations for the June 19 tracer injection conducted i n the upper and lower reach. A d d i t i o n a l breakthrough curves and mode l s imulat ions are p rov ided i n A p p e n d i x A . Tab le 3.3: S u m m a r y o f best fit mode l parameters for solute releases i nc lud ing stream discharge (Q) , dispersion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , storage zone exchange coefficient (a), net lateral i n f low (QL) , the D a m k o h l e r number (Da l ) . Parameters Reach Q (10"3 mV) D (10"2 mV1) A (10"2m2) A s (10"2m2) a ( 1 0 V ) QL ( l O ^ m V ' n r 1 ) D a l Upper June 19 15.9 31.6 11.6 5.83 2.1 26.8 0.12 Sept 21 9.8 18.3 6.73 1.45 2.7 8.5 0.22 Sept 29 1.0 2.9 2.82 0.6 3.0 3.3 0.50 Oct 20 9.7 4.0 9.28 1.46 3.5 10.5 0.38 Lower M a y 31 30.6 48.6 21.7 13.5 1.4 31.8 0.08 June 19 17.2 29.5 19.2 5.9 1.9 20.4 0.14 June 27 5.5 1.9 13.7 4.1 2.7 5.6 0.44 Sept 21 9.9 8.4 14.3 2.8 2.1 6.9 0.36 Sept 30 1.9 4.6 6.5 1.3 0.8 6.9 0.17 Oct 20 11.9 5.0 19.2 2.0 1.6 9.4 0.28 42 o c o O CO O o O CN CD A . A A A AA"y A A A A A A A AAA -TL_ • • • • • • • A Upper boundary ° Lower boundary Simulated Time (hour) 13 C g I c 8 c 8 .1 LO CM O CM LO O -4 0.0 0.5 A A A A A r F * 1 3 ^ A A ^ A Upper boundary D Lower boundary Simulated ' a A ^ A A A A l 1 A A A 7 A 1 A 1 A A ( B ) ^H^-^St"-a—a... 1.0 1.5 Time (hour) 2.0 2.5 Figure 3.2. M o d e l s imulat ions us ing O T I S - P for June 19 for the upper reach (a) and the lower reach (b) 43 3.3.2. Variability of fitted parameters Solute transport processes var ied both temporal ly (wi th variations i n discharge) and to a lesser extent, spatial ly (i.e. between reaches). Dispers ion rates, transient storage area and stream cross-sectional area increased w i t h discharge (Figure 3.3). T h i s general trend is observed for both the upper and lower reaches; however, for a l l tracer s imulat ions the stream cross-sectional area o f the lower reach is greater than the upper reach. Increased channel complex i ty and poo l storage i n the lower reach m a y expla in this difference. Transient exchange rates between the m a i n channel and the storage zone d id not vary systematical ly w i t h discharge or between reaches (Figure 3.3). 5 0.6 LO -o ^ o 11 i -^ ' m E o _ T o o I a" " -o _ • U p p e r o o A Lower o -1 i i '•I • U p p e r A L o w e r < <° o o Q ( L s " ' ) Q ( L s " ' ) Figure 3.3. S imula ted mode l parameters for solute releases i n the upper and lower stream reach. D i spe r s ion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , storage zone exchange coefficient (a) versus stream discharge (Q) . Er ro r bars represent ± 1 standard deviat ion. 44 3.3.3. Parameter uncertainty A general decreasing trend i n D a l numbers was observed as discharge increased (Figure.3.4) . D a m k o h l e r numbers ranged from 0.08 on M a y 31 (30.6 L /s ) to 0.5 on September 29 (1.04 L / s ) . The tracer injection wi th the best parameter estimates, as indicated b y a D a l number closest to one, was on September 29. A A A A A A • Upper ^ Lower 1 I I I I I I 0 5 10 15 20 25 30 Q(Ls~ 1 ) Figure 3.4. Experimental Damkohler number (Dal) versus stream discharge (Q) The uncertainty ratio for each estimated parameter (D, A , A s , a), calculated as the ' parameter estimate divided by its standard deviation, varied over the range o f flow conditions (Table 3.4). N o trend between discharge and the uncertainty ratio was observed (Figure 3.5). D 45 Table 3.4. S u m m a r y o f uncertainty ratios for the parameter estimates o f dispersion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , and the storage zone exchange coefficient (a). R e a c h D A A s A U p p e r June 19 1.9 211.1 8.1 20.7 Sept 21 5.1 69.5 2.7 8.2 Sept 29 5.5 56.5 14.2 5.2 Oct 20 1.2 35.2 6.2 2.1 L o w e r M a y 31 9.8 .76.7 3.1 4.8 June 19. 7.6 60.2 0.9 .1.4 ' June 27 4.2 78.7 25.8 8.6 Sept 21 7.9 87.1 11.9 5.2 Sept 30 13.7 99.1 8.7 9.3 Oct 20 5.6 51.4 6.0 2.5 M e a n Upper 3.4 93.1 7.8 9.1 L o w e r 8.1 75.5 9.4 • 5.3 • Upper A . Lower D • Upper A Lower DC " i 1 1 1 1 1 r 0 5 10 15 20 25 30 Q (Ls" 1 ) CN i i — n r 10 15 20 25 30 Q (Ls~ 1) Figure 3.5. Uncer ta in ty ratio ( U R ) for the s imulated mode l parameters o f dispersion (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone ( A s ) , storage zone exchange coefficient (a) versus stream discharge (Q) 46 3.3.4. D e r i v e d quan t i t i e s D e r i v e d quantities i nc lud ing hydrau l ic residence t ime and retention and uptake lengths are summarized i n Table 3.5. H y d r a u l i c residence t ime i n the stream was higher than i n the storage zone for a l l tracer injections and d id not appear to vary w i t h discharge for either reach (Figure 3.6). The average stream residence t ime was higher i n the lower reach (109 min) compared to the upper reach (49.8 m i n , F igure 3 . 6 A ) . Storage zone residence t ime was also higher i n the lower reach (31.0 min) compared to the upper reach (10.2 m i n , F igure 3 .6B) . Tab le 3.5. S u m m a r y o f der ived quantities i nc lud ing stream ve loc i ty (u), hydrau l ic residence t ime for the stream ( T s t l ) and storage zone ( T s t 0 1 ) , hydrau l ic uptake length (Shyd), hydrau l i c retention factor (Rh) and the standardized storage zone coefficient ( A s / A ) . R e a c h Q (10"3 m V ) u (ms"1) T s t r (min) T 1 s t o r (min) Shyd (m) R h (sm"1) A s / A U p p e r June 19 15.9 0.13 80.7 40.4 648 3.7 0.50 Sept 21 9.8 0.15 61.8 13.3 542 1.5 0.22 Sept 29 1.0 0.04 • 55.2 11.6 122 • 5.7 0.21 Oct 20 9.7 0.10 48.3 7.6 302 1.5 0.16 L o w e r M a y 31 30.6 0.14 121.5 75.7 1029 4.4 0.62 June 19 1.7 0.09 86.2 26.7 465 3.4 0.31 June 27 5.5 0.04 61.0 18.1 146 7.4 0.30 Sept 21 9.9 0.07 80.2 16.0 334 2.9 0.20 Sept 30 1.9 0.03 196.5 38.9 343 6.8 0.20 Oct 20 11.9 0.06 106.3 11.0 397 ' 1.7 0.10 Stream ve loc i ty increased wi th discharge (Figure 3 . 7 A ) . Genera l ly , ve loc i ty was higher i n the upper reach than the lower reach. T h e hydraul ic uptake length increased as discharge increased for both the upper and lower reach (Figure 3 .7B) . V a l u e s ranged from 122 m (1.04 L / s ) i n the upper reach to 1028 m (30.6 L/s) i n the lower reach. O n average the uptake length was higher i n the upper reach (318 m) than the l ower reach (305 m) . Howeve r , dur ing two tracer injections (June 19, September 21), uptake length was greater i n the upper reach than the lower reach, and less than the lower reach dur ing two addit ional injections (September 29, October 20). 47 The hydraul ic retention factor, a measure o f the storage zone residence t ime per unit o f stream reach traveled, showed no clear trend wi th discharge (Figure 3 .7C) . D u r i n g l o w f low condit ions, retention was highest, ranging from 5.7 s/m at 1.0 L / s to 7.4 s/m at 5.5 L / s . A t m i d to h igh stream f low rates (10 L / s to 30 L/s ) a slight increasing trend in retention was observed; however , retention factors d id not reach l o w f low values. The standardized storage zone area also d i d not show a clear trend w i t h discharge (Figure 3 .7D). A majori ty o f the values ranged from approximate ly 0.10 to 0.30, w i th the except ion o f the M a y 31 injection at a value o f 0.62. o -CM A A O U") -X— 001. A A A o • A • • o -• A Upper Lower o o - B o co A o CD o A • A . CM O -• A • A Upper Lower 1 0 1 5 1 10 I I 15 20 I 25 i 30 Q(Ls" 1 ) Figure 3.6. H y d r a u l i c residence t ime o f solutes in the stream ( A ) and storage zone (B) versus stream discharge (Q) for the upper and lower reach. 48 • Upper A Lower < o ID O Q(Ls ) Q(Ls") Figure 3.7. Stream velocity (A), hydraulic uptake length ( B ) , hydraulic retention factor (C) and the standardized storage zone coefficient (D) versus stream discharge (Q) for the upper and lower reach . 3.3.5. Lateral inflow rates Lateral inflow rates increased with discharge (Figure 3.8). Inflow rates ranged 6 3 1 1 5 3 1 1 from 3.26x10" m s" m" at low flow conditions to 3.18 x 10" m s" m" at high flow conditions. Net inflow represents less than 1% gain of streamflow. Lateral inflow rates did not appear to vary significantly between reaches. 49 • Upper A Lower H 0 5 10 15 20 25 30 Q(Ls" 1) Figure 3.8. Ne t lateral i n f low rates ( Q l ) versus discharge (Q) for a l l solute releases 3.3.6. Quantifying pool storage and residence times Breakthrough curves f rom two pools located in the upper reach and one poo l located i n the lower reach were s imulated us ing O T I S - P . Results are presented for one injection on September 30 (Figure 3.9). A d d i t i o n a l mode l s imulat ions are presented i n A p p e n d i x A . F o r al l poo l s imulat ions, the transient storage area was greater than channel area, result ing i n a larger A s / A ratio than the reach s imula t ion (Table 3.6). H i g h uncertainty values were observed for parameter estimates associated w i t h dispersion (D) and storage area (As) for two p o o l s imulat ions (Figure 3.10). T h e s imulat ions for poo l ' 2 and poo l 3 had the greatest uncertainty based on the D a m k o h l e r number. The D a l numbers were 0.09 and 2.8, respectively, compared to a D a l number o f 0.8 i n poo l 1. Storage zone residence t imes were higher than for the entire reach for both pools located i n the upper reach (Figure 3.11). Res idence times were also greater than the reach scale estimated values. H o w e v e r , this pattern was not supported i n the poo l located in the lower reach (Poo l 3). 50 Table 3.6. S u m m a r y o f the simulated parameter estimates o f clispersion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone ( A s ) , and the storage zone exchange coefficient (a), stream ve loc i ty (ii), the standardized storage zone coefficient ( A s / A ) and the D a m k o h l e r number (Da l ) . Parameters R e a c h Q (10- 3mV) D (10- 2mV) A ( l o W ) A s (l(r2m2) a ( l O V ) U (ms-1) A s / A D a l Sept 29 Reach 1 1.0 2.9 2.8 0.6 3.0 0.04 0.21 0.50 Poo l 1 2.5 • 6.3 60.7 3.1 0.02 9.61 0.80 Poo l 2 1.8 0.8 6.1 3.4 0.13 7.40 0.09 Sept 30 Reach 2 1.9 4.6 6.5 1.3 0.85 0.03 0.20 0.17 Poo l 3 8.4 2.5 5.9 145 0.07 2.33 2.60 F igure 3.9. M o d e l s imulat ions us ing O T I S - P for September 30. Results are f rom one poo l locat ion i n the lower reach. 51 E. Q E, < E, < Pool 3 Upper Pool 3 Figure 3.10. S imula ted mode l parameters for solute releases i n the upper and lower stream reach dur ing September 29 and 30. Poo l s 1 and 2 were located i n the upper reach. P o o l 3 was located i n the lower reach. Dispe r s ion coefficient (D) , stream cross-sectional area ( A ) , cross-sectional area o f storage zone (As ) , storage zone exchange coefficient (a) versus locat ion. Er ro r bars r e p r e s e n t ' ± 1 standard deviat ion. 52 o o -CO • Stream A Storage o A CO o o -A o CO o o -CM • O O -• • • A -o - A * ^ 1 1 1 r-^ Upper Pool 1 Pool 2 Lower Pool 3 Location Figure 3.11. H y d r a u l i c residence t ime o f solutes in the stream and storage zone for solute releases i n the upper and l ower stream reach dur ing September 29 and 30. Poo l s 1 and 2 were located i n the upper reach. P o o l 3 was located i n the lower reach. 3.3.7. Subsurface relative connectivity A total o f 5 to 8 observations o f m i x i n g ratios (Equat ion 2.27) were made for each piezometer, except for piezometers P56 , P I 8, and P 4 8 , where on ly three observations were conducted due to the d ry ing out o f the stream channel over the study per iod. T h e relative connect iv i ty var ied between reaches (Figure 3.12). T h e upper reach appeared to have a higher percentage o f piezometers where the tracer concentration i n the piezometer decreased dur ing the tracer injection, as indicated by a negative value o f % (e.g. P I 6, P 2 0 , P 2 2 , and P21) . A negative m i x i n g ratio cou ld be a result o f stratification o f water chemistry i n the streambed. In general, the tracer concentration for piezometers located w i t h i n the lower reach increased dur ing tracer injections as indicated b y a posi t ive value o f x- T h i s suggests that piezometers located w i t h i n the lower reach have a higher degree o f connect iv i ty to water i n the stream channel . There was no clear dis t inct ion between u p w e l l i n g and d o w n w e l l i n g zones i n terms o f relative connect ivi ty . In the upper reach, 53 hydraul ic gradients were not measured in piezometers 16, 18 and 19. A l l three sites located at the upstream end o f a riffle habitat. o o Upper o 0 0.5 , O o i o 1 1 d m d ~ — • — i 0 o p o o 0 • • Downwelling Neutral o f I P12 P16 i P18 I P19 1 P20 1 P21 I P22 i P23 1 P27 1 P7 i P8 Lower —j— o .0 0.5 - T - 1 1 .0 0.5 ° o o i i ' i — 1 4 .0 0.5 \W*.....-±:....mmm n i l • 1 1 - j- L o 1 1 0 a o 0 Q •1.0 -0.5 • • •! Downwelling Upwelling Neutral I I I P31 P33 P35 i P36 1 1 P38 P47 P48 I P50 i P51 I P52 1 1 P53 P54 1 P56 1 1 P64 P67 i P68 Figure 3.12. Rela t ive connect iv i ty ( R C ) as measured us ing a non-d imens iona l m i x i n g ratio (x) for piezometers sampled dur ing tracer injection experiments i n the upper and lower reach 3.4. Solute injection experiments - Channel-unit scale 3.4.1. Qualitative observations of hyporheic discharge T w o different f low pathways were observed w i t h i n one channel-unit i n the lower reach (Figure 3.13). T h e f o l l o w i n g observations were made dur ing experiments at the infi l trometer 1-5 loca t ion: • July 11 - R W T tracer water infiltrated the bed, traveled lateral ly around a large anchor boulder in the riparian zone and returned to the stream at the base o f the step wi th in 5-10 minutes post-injection. 54 • A u g u s t 14 - R W T fo l lowed the same f low pathway as J u l y 11. R W T was v i s ib l e at the start o f the poo l w i t h i n four minutes post-injection and was v i s ib le w i t h i n the poo l un t i l approximately 18 minutes post-injection. • S e p t e m b e r 27 - D i scha rg ing hyporheic water was again observed at this locat ion; however , dur ing this experiment R W T infiltrated the streambed, t ravelled ver t ica l ly through the step and returned to the stream at the start o f the poo l . R W T was v i s ib le i n the poo l w i t h i n 9 minutes, and was st i l l s l ight ly v i s ib le w i t h i n the p o o l after one hour. S i m i l a r observations were made at infi l trometer 1-4: • A u g u s t 21 - N o R W T was v i s ib le w i t h i n the poo l after 4 hours o f observation. H o w e v e r , an increase i n fluorescence (as measured us ing a fluorometer) was observed. Fluorescence returned to background i n approximately 24 hours, suggesting the presence o f hyporheic f l ow at this locat ion. • S e p t e m b e r 28 - R W T infiltrated the streambed, traveled ver t ica l ly through the step and returned to the stream at the start o f the poo l . • O c t o b e r 6 - D i s c h a r g i n g hyporheic water was observed at the base o f the p o o l ( s imi lar locat ion as the September 27 and 28 injections). R W T was observed 13 minutes after injection and was s t i l l v i s ib le i n the poo l after 1 hr 24 m i n . N o tracer was observed at 1 hr 48 m i n after injection. • Injections o f B r i l l i a n t B l u e F C F were also conducted; however , d ischarging dye was not observed at any locat ions w i t h i n the lower reach. Observat ions were also made at infi l trometer 1-1 instal led i n a log-step i n the upper reach; however , d ischarging o f R W T or B r i l l i a n t B l u e F C F was not observed dur ing any tracer inject ion experiments at this loca t ion (July 12, A u g u s t 23, and September 28, October 6), except for an in i t ia l t r ial on June 27. Hyporhe i c discharge o f B r i l l i a n t B l u e F C F was observed w i t h i n the log-step channel-unit dur ing the trial on June 27. Tracer infil trated the stream bed and was observed u p w e l l i n g from the sediments at the upstream end o f the p o o l . 55 Figure 3.13. Channel -uni t observations o f f low pathways i n P o o l 4, i n c l u d i n g side v i e w and aerial v i e w 3.4.2. Quantifying residence times The poo l behaved l ike a cont inuously stirred tank reactor ( C S T R ) , as indicated by a straightening o f the late t ime period o f the breakthrough curve when concentrat ion is plotted on a logar i thmic scale (Figure 3 .14B, Figure 3 .15B) . M R T var ied over the experiments poss ib ly due to var iabi l i ty in streamflow and hydrau l ic gradients (Table 3.7). E lec t r ica l conduct iv i ty increased rapidly to a peak concentrat ion w i t h i n the step and pool sub-units. D u r i n g both experiments, the m a x i m u m electrical conduc t iv i ty was greater i n the step than p o o l . A d d i t i o n a l minor peaks i n E C are v i s i b l e o n the r i s i ng and recession l i m b o f the breakthrough curves for each experiment. These peaks could indicate separate flow pathways wi th different residence t imes w i t h i n the channel-unit . Tab le 3.7. M e a n residence t imes for hyporheic zone (step) and p o o l storage zones i n P o o l 4. Ve r t i c a l hydrau l ic gradients ( V H G ) measured us ing piezometer 61 . M e a n Residence T i m e (min) Date Q ( L s 1 ) V H G (cm/cm) Step P o o l Sept 25 2.4 -0.96 23.8 26.4 Oct 5 1.4 -0.89 68.8 71.2 5 6 o LU o LU "i 1 1 1 1 r 00:00 00:12 00:24 00:36 00:48 01:00 01:12 Time (hr:min) ~\ r 00:00 00:12 00:24 00:36 00:48 01:00 01:12 Time (hr:min) Figure 3.14. Step-pool residence t ime experiment conducted on September 25, 2006 CO O LU 01:00 03:00 Time (hnmin) O o tu 05:00 Figure 3.15. Step-pool residence t ime experiment conducted on October 5, 2006 57 3.5. Subsurface flow - Point scale 3.5.1. Hydraulic gradients Considerable spatial and temporal var iabi l i ty i n hydrau l ic gradients was observed at a l l locations. Strong negative hydrau l ic gradients typ ica l ly occurred at the start o f a step or riffle ind ica t ing infi l t rat ion into the stream bed (Figure 3.16). Consistent negative gradients were observed i n piezometers 1 to 6 w i t h i n the log-step channel-unit , al though gradients var ied tempora l ly (Figure 3.17). In general, gradients tended to get weaker 'as stream f low decreased dur ing l o w f low condit ions i n m i d to late summer. Pos i t ive hydraul ic gradients were consistently observed downstream from obstructions i n the stream channel , such as the l og step located in the upper reach 13-15 m downstream. Piezometers 10, 11, 13, 14 and 15 showed consistent posi t ive hydrau l ic gradients over the study per iod (Figure 3.18); however, hydraul ic gradients var ied over short spatial scales. F o r example, piezometers 10 and 13 were located approximately 20 c m apart and were \ installed at s imi la r depths; however, the average gradient in piezometer 13 was 1.06 c m / c m over the study per iod compared to 0.03 c m / c m i n piezometer 10. A s imi la r pattern o f hydraul ic gradients was observed i n a boulder-step channel-unit located i n the upper reach 20 m downstream from the culvert crossing. Piezometers 21 to 23 showed a consistent strong d o w n w e l l i n g response (Figure 3 .19A) . T h e magnitude o f response increased as the distance to the step increased w i t h piezometer 23 showing the strongest response (average = -0.6 cm/cm) . Piezometers 24 to 26 a l l showed consistent u p w e l l i n g and were located near the head o f the p o o l (Figure 3 .19B) . N 58 o x > o o p c\i . . ^ A ^ . . . # A A . A — r ~ 13 14 - A . • Step/Riffle A Pool 15 16 17 ' T -IS — i r~ 19 20 o E O x > o o o c\i Distance downstream(m) Figure 3.16. V e r t i c a l hydrau l ic gradients measured i n the upper ( A ) and lower (B ) reaches. Symbol s indicate study per iod means. LO o o d -K-.T.^T. ,> ... :-• A A ' -I > o P 1 A p2 A P 3 • P 4 • P 5 + P 6 1 J u l 01 1 A u g 01 1 Sep 01 1 O c t 01 Figure 3.17. V e r t i c a l hydrau l ic gradients measured over the study per iod i n piezometers 1 -6 i n the upper reach 0.2 0.3 i i A . . A 1 0 A A • 4 P10 P11 P13 P14 P15 o ~ i A A. A., . ,.o 0... . 0 . o */ ••••fl--\v;;::!i:;:;: A d A A •"" O "" 6' CN d _ 9'' T JinOI JJ01 AteOI SepOl Oct 01 Figure 3.18. V e r t i c a l hydrau l ic gradients measured over the study per iod i n piezometers 10,11,13-15 i n the upper reach 60 + P25 • P24 x P26 1 1 1 3 01 Sep 01 Oct 01 Figure 3.19. Ve r t i c a l hydrau l ic gradients measured over the study per iod i n the piezometers 21-23 and piezometers 24-26 i n the upper reach U p w e l l i n g sites were also observed w i t h i n the lower reach, al though hydraul ic gradients were not as strong as i n the upper reach. Va lues for u p w e l l i n g sites tended to be w i t h i n measurement error (± 0.05 cm/cm) . Reversa l o f hydrau l ic head gradients was also more c o m m o n w i t h i n the lower reach, w i t h piezometer response swi tch ing between neutral, d o w n w e l l i n g and u p w e l l i n g over the course o f the study per iod. F o r example, i n one boulder step channel-unit , piezometers installed at the downstream end o f a large step (60, 61 and 62) showed consistent d o w n w e l l i n g (average -1.3, -0.88, -1.1 c m / c m respectively), whereas the gradients measured i n piezometers located w i t h i n the poo l (67 and 68) fluctuated from neutral to slight u p w e l l i n g and back to d o w n w e l l i n g dur ing the late summer (Figure 3.20). Pos i t ive V H G ' s observed i n piezometers 67 and 68 dur ing m i d to late September correspond to qualitative observations made at the channel-unit, where R W T injected into the streambed traveled ver t ica l ly through the step and returned to the poo l at the start o f the poo l dur ing an inject ion on September 28. P r io r to this, exchange f low was on ly observed f l o w i n g laterally around a large anchor boulder located wi th in the channel-unit . (A) I > P21 P22 P23 (B) £ I > Jul 01 Aug 01 Sep 01 Oct 01 Aut 61 * t A o • • P60 P61 P62 P67 P68 -"t— A, ... A '•A " " ^ < ' * " •---.-W -vp........ . • ET' .S3'"" . . . - O - - . . 'o Jul 01 ' Aug 01 Sep 01 Oct 01 Figure 3.20. V e r t i c a l hydrau l ic gradients measured over the study per iod i n the piezometers 60, 61 , 62 and 67 to 68 in the lower reach O f the 66 piezometers tested, on ly four locations had correlations between discharge and V H G us ing a non-parametric Spearman's rank correlat ion that are significant at a = 0.05 (Table 3.8). A n u p w e l l i n g (P25) and d o w n w e l l i n g site (P2) had negative correlations w i t h discharge, w h i l e two d o w n w e l l i n g sites had pos i t ive correlations (P27, P47) w i t h discharge. These results suggest that stream discharge does not control ver t ical hydrau l ic gradients i n East Creek. Tab le 3.8. Spearman correlat ion coefficient (r s ) , associated p-values, number o f observations (n) and vert ical hydraul ic gradients ( V H G ) for each piezometer ind ica t ing a significant correlat ion w i t h discharge. Piezometer P-value N V H G (cm/cm) 2 -0.87 0.002 9 -0.01 25 -0.72 0.02 10 0.03 27 0.84 0.02 7 -0.12 47 0.62 0.04 11 ' -0.03 62 3.5.2. V H G and scaled location wi th in channel units Zones o f hyporheic discharge and recharge appear to be a function o f the scaled locat ion i n the channel unit (Figure 3.21). Zones o f hyporheic discharge, or u p w e l l i n g (as indicated b y a posi t ive V H G ) , were generally confined to the upper por t ion o f the channel units ( X / L = 0.0 to 0.4). Fo r X / L > 0.2, there is a trend to increas ingly negative hydrau l ic gradients w i t h increasing distance from the head o f the channel-unit . The height o f the downstream step does not appear to control hydrau l ic gradients. A Spearman's rank correlat ion analysis indicated that hydrau l ic gradients were not correlated wi th step height ( r s = -0.186, p = 0.17). In addit ion, a sequential analysis o f variance us ing three linear mixed-effects models conf i rmed that pos i t ion w i t h i n the stream channel (as defined by a category o f X / L ) s ignif icant ly contributed to the observed var iab i l i ty i n hydraul ic gradients (x2 = 40.9, p < 0.001, Tab le 3.9). T h e addi t ion o f downstream step-height to the base model ' (i.e. channel unit as a random effect) d i d not s ignif icant ly contribute to the observed var iab i l i ty i n hydrau l ic gradients (%~ = 4.6, p = -009, Table 3.9). Table 3.9. A n a l y s i s o f variance table compar ing three l inear mixed-effects models for vert ical hydrau l ic gradients inc lud ing a base mode l w i t h on ly channel-unit , a second mode l w i t h channel-unit and step height ( S H ) , and a third mode l w i t h channel-unit, step-height and channel pos i t ion ( X / L ) . A chi-square (%2) statistic was used to test for significance. Mode l D F 2 X D F X2 p-value Un i t 2 U n i t + S H 4 4.6 2 0.09 Uni t + S H + X L 8 40.9 4 <0.001 63 o o E O x > °...<&s..°.....rA..s> a.. ° - o o o o ° 0 1 1 1 1 r 0.0 0.2 0.4 0.6 0.8 • 1.0 X/L Figure 3.21. Vertical hydraulic gradient (cm/cm) versus scaled location within the channel-unit. Hydraulic gradients are averaged over the entire study period. CD X > 1 ^ 0.0 & 0.2 + 4-4-x A/ 4-X/L<0.2 0.2<X/L<0.4 0.4<X/L<0.6 0.6<X/L<0.8 0.8<X/L< 1.0 0.4 0.6 0.8 1.0 Step height (m) Figure 3.22. Vertical hydraulic gradient (cm/cm) versus step height (m) as a function of scaled location within the channel-unit (X/L). Hydraulic gradients are averaged over the entire study period. 64 3.5.3. H y d r a u l i c c o n d u c t i v i t y T h e geometric means o f conductivi t ies for the lower and upper reaches were 2.54 x 10" 4 (n = 24) and 2.37 x 10" 4 m/s (n = 16) respectively. Conduct iv i t ies appeared to be higher at neutral and u p w e l l i n g sites than at d o w n w e l l i n g sites i n the lower reach (Figure 3.23); however , on ly three sites i n the analysis were considered u p w e l l i n g sites, compared to d o w n w e l l i n g (n = 21) and neutral sites (n = 17). A sequential analysis o f variance us ing two l inear mixed-effects models conf i rmed that site condi t ion s ignif icant ly contributed to the observed var iab i l i ty i n hydrau l ic conduct iv i ty (%2 = 6.7, p = 0.01, Tab le 3.10). H y d r a u l i c conduct iv i ty also d i d not appear to vary systematical ly w i t h instal lat ion depth (Figure 3.24). Tab le 3.10. A n a l y s i s o f variance table compar ing two linear mixed-effects models for hydrau l ic conduct iv i ty ( log transformed), i nc lud ing a base mode l w i th on ly reach as a factor and a second mode l w i th reach and site condi t ion (upwel l ing , neutral and downwel l ing ) . M o d e l D F " I1 D F X 2 p-value Reach 2 Reach + Site C o n d i t i o n 3 6.7 1 0.01 65 Figure 3.23. H y d r a u l i c conduct iv i ty ( K ) for d o w n w e l l i n g (D) , neutral (N) and u p w e l l i n g (U) sites located i n the lower (n = 24) and upper reach (n = 17). No te l o g scale. o CN • Upper 0 Lower E N 10 15 " T T 20 25 Depth (cm) 30 Figure 3.24. H y d r a u l i c conduct iv i ty ( K ) w i t h depth o f piezometer instal lat ion for the upper and lower reaches. N o t e l o g scale. 66 3.5.4. Streambed infiltration rates Temporal variation in infiltration rates was observed over the study period (Figure 3.25) . Rates were difficult to measure with the constant head infiltrometer during low flow conditions, resulting in a lack of observations during base flow conditions (mid to late August). The probable errors for infiltration measurements were almost ± 60% of the measured value. The relationship between discharge and infiltration rates was tested using a non-parametric Spearman's rank correlation. In two locations, infiltration rates were significantly correlated with discharge (Table 3.11), including the sediment-step (1-3) and boulder step (1-5). The boulder step (1-5) location also had the strongest mean V H G (-1.1 cm/cm). Hydraulic conductivity estimates based on streambed infiltrometers (Equation 2.26) were higher than estimates from falling head tests (Figure 3.26). This result suggests that bed infiltration computed from piezometer data alone may underestimate actual infiltration rates. or o L O o Cxi L O ci o d Jul Aug Sep Oct Figure 3.25. Infiltration rates over the study period. Error bars represent probable errors based on Equation 2.25. 67 Table 3.11. Spearman correlation coefficient (r s), associated p-values and number o f observations (n) for inf i l t rat ion rates versus discharge at each infil trometer locat ion. Locat ion r s p-value n L o g step (1-1) 0.37 0.21 13 Sediment step (1-3) 0.92 0.01 ' 5 Boulder step (1-4) 0.51 0.25 6 Boulder step (1-5) 0.83 0.01 7 12 10 • Infiltrometer k El Slug k 11 i n Log step (1-1) Boulder step (I-2) Sediment step (I-3) Boulder step (I-4) Boulder step (I-5) Figure 3.26. H y d r a u l i c conduct iv i ty calculated us ing infi l tration rates and slug-tests for f ive locations. Va lues represent the geometric mean ± standard error. 3.5.5. Streambed water fluxes computed from Darcy's law Streambed water fluxes w i t h i n one step-pool unit i n the upper reach (Poo l 1) var ied w i t h f low condit ions. T h e total computed f lux into the bed was larger dur ing h i g h f low condit ions on June 19 ( Q = 15.4 L / s ) than dur ing l o w f low condit ions on September 29 (Q = 1.1 L / s ) , as calculated us ing Equa t ion 2.23. The fluxes into the bed were also larger than fluxes out o f the bed as summar ized i n Table 3.12. The average fluxes out o f 68 the bed, as categorized where X / L < 0.4, d id not change s ignif icant ly w i t h f low (Figure 3.27). Howeve r , Jhe f luxes into the area above the step, where X / L > 0.6, d id increase s l igh t ly w i t h f low, spec i f ica l ly where X / L ~ 0 . 8 . T h e reach scale estimate o f hyporheic exchange was two-orders o f magnitude greater than the scaled-up estimate o f exchange for both f low condit ions (Table 3.12). o o o o o CO o o Low flow High flow I I I I T~ 0.2 0.4 0.6 0.8 r.O X / L Figure 3.27. Water fluxes calculated us ing D a r c y ' s L a w for each X / L category w i t h i n one step-pool channel-unit dur ing l o w f low (Q =1.1 L / s ) and h igh flow ( Q = 15.4 L / s ) Table 3.12. Water fluxes w i t h i n one step-pool unit (Poo l 1) a long w i t h scaled-up and reach-scale estimates o f hyporheic exchange (s _ l ) . Water F luxes (L/s) D a t e Q (Ls" 1 ) qin qout Scaled-up (10 'V) R e a c h scale a(10"4 s _ l ) June 19 15.4 0.78 -0.07 4.1 2.1 September 29 1.1 0.46 -0.02 2.6 3.0 69 CHAPTER FOUR: DISCUSSION T h i s chapter discusses the results presented i n Chapter 3. Sections 4.1 to 4.3 discuss the research objectives out l ined in Chapter 1. The last section integrates observations from the different scales o f interest. 4.1. Reach scale 4.1.1. Modelled parameter uncertainty Wagner and H a r v e y (1997) expla ined that parameter uncertainty is m i n i m i z e d when D a l = 1.0. Parameter uncertainty, especia l ly w i t h respect to the estimates o f transient storage and exchange, increases i n experiments w i t h very h igh or l o w D a l values. In East Creek, D a l values were less than 1.0 for a l l experiments and ranged f rom 0.1 to 0.5, w h i c h is consistent w i t h Wagne r and H a r v e y ' s (1997) conc lus ion that " w e l l estimated" parameters are l i k e l y to be obtained when the D a m k o h l e r number is on the order o f 0.1-1.0. M o d e l l e d parameter estimates i n East Creek were also fai r ly consistent w i t h the typical range o f parameter values reported i n the literature for h igh gradient streams (1-15%) before 1997 (Wagner and H a r v e y 1997, Table 4.1). Unde r condit ions when D a l < 1.0, parameter uncertainty is h igh because o n l y a smal l amount o f tracer interacts w i t h the storage zone (Wagner and H a r v e y 1997). T h i s may occur because: (1) stream ve loc i ty is h igh , (2). exchange timescales are short, as indicated b y l o w values o f a or (3) the reach length is short. Parameter uncertainty was greatest dur ing periods o f higher f lows , poss ib ly due to the s low rates o f transient exchange relative to the stream water ve loc i ty . T h e transient exchange coefficient (a) was estimated at 0.00014 s"1 under h igh f low condi t ions on M a y 31 ( Q = 30.6 L / s , u = 0.14 m/s), w h i c h was h a l f the average a for a l l s imulat ions (0.00022 s~'). H a r v e y et a l . (1996) concluded that the stream tracer method does not re l iab ly characterize exchange at higher f lows, w h i c h complicates the efforts o f studies examin ing the influence o f discharge on transient storage processes over a range o f f low condit ions (e.g. Har t et al . 1999). 70 The Damkohler number was highest in pool 3 (Dal = 2.8) compared to values of 0.8 and 0.09 in pools 1 and 2 respectively, indicating higher uncertainty in the parameter -estimates associated with transient storage. However, parameter uncertainty was lower in pool 1 than at the reach scale (Dal = 0.5). The low Dal numbers within pool 1 and 2 may be attributed to the length of the pool (4 m) compared to 50 m at the reach-scale. Wagner and Harvey (1997) indicated that the length of the study reach should be adjusted to maintain a balance between advective transport and transient storage. This suggests that-reducing the "reach-length" to the scale of individual pools may have contributed to parameter uncertainty. The high Dal number in pool 3 indicates that exchange rates were fast relative to the stream water velocity and that all the solute was exchanged with the storage zone over the reach length (Wagner and Harvey 1997). The uncertainty ratio was also used to quantify parameter uncertainty for reach scale and channel-unit scale simulations. Ratios were lowest for the reach-scale dispersion and transient exchange parameters (Table 3.4), which indicates a higher degree of uncertainty. Low uncertainty ratios were also associated with the dispersion and transient storage parameters for two of the three pool simulations. Uncertainty ratios were within the ranges reported by a previous study examining the broad heterogeneity of hyporheic zone processes across seven small streams in western Washington (Reidy 2004). However, that study encompassed a wider range of flow conditions (0.7 to 216 10 3m3/s) and channel morphologies than studied in East Creek. Table 4.1. Range of parameter values reported for high-gradient streams (Wagner and Harvey 1997) compared to modelled parameter values in East Creek. Range Parameter Wagner and Harvey (1997) East Creek Q (m3/s) 0.005 - 0.2 0.002 - 0.03 Q L (m3/s/m) 0-0.0001 0 - 0.00003 A(m 2 ) 0.02 - 0.6 0.06-0.2 D (m2/s) 0.025 - 0.8 0.02-0.5 A s (m2) 0.01 -2.0 0.006 - 0.1 a(s-') 5.0 x 10"6-0.001 8.0 x 10"5 - 0.0001 71 4.1.2. Solute transport parameters and discharge Solute transport parameters var ied w i t h discharge at the reach scale. The results o f the O T I S - P simulat ions were i n partial agreement w i t h previous studies w h i c h reported an increase i n dispersion (D) and the channel cross-sectional area ( A ) w i t h discharge (Legrand-Marcq and Laudelout 1985, D ' A n g e l o et a l . 1993, H a r v e y et a l . 1996, M o r r i c e et a l . 1997, Hart et al . 1999, W o n d z e l l 2005). However , those studies reported that transient storage area (As) decreased w i t h discharge w h i l e transient exchange (a) increased w i t h increasing discharge, w i t h the except ion o f L e g r a n d - M a r c q and Laudelout (1985) who reported that transient exchange remained constant w i t h discharge. In East Creek, on the other hand, exchange rates remained fa i r ly constant w i t h discharge w h i l e the storage area increased. A previous study i n the upper and lower reaches o f East Creek observed s imi lar trends i n the response o f a and A s (Patschke 1999), as d i d Hart et a l . (1999). A recent study i n a fourth-order stream i n central M i c h i g a n ( U S A ) found that the size o f the transient storage zone increased w i t h discharge from 1.9 m at baseflow (2.5 m 3 / s ) to 7.3 m 2 at a f low o f 19.1 m 3 / s (Phanikumar et a l . 2007). Studies have attributed a constr ict ing o f the transient storage area dur ing periods o f higher discharge to increased groundwater discharge and catchment wetness (Bou l ton et a l . 1998, W h i t e 1993). A s the catchment wetness increases, hydrau l ic gradients to the stream from the riparian zone are stronger and can ove rwhe lm the influence o f channel morpho logy (Harvey and Benca l a 1993, W o n d z e l l and Swanson 1996a), result ing i n a decrease i n the extent o f the hyporheic zone. Channel complex i ty is cited as a possible explanat ion for the observed differences i n the response o f the transient storage area ( A s ) to discharge i n East Creek (Patschke 1999). Storage area increased w i t h discharge i n two reaches associated w i t h a re la t ively h igh degree o f stream complex i ty (i.e. upper and l ower reach) and remained constant w i t h discharge i n two less complex reaches located upstream from the current study locat ion (Patschke 1999). T h i s study attributed the difference to var iab i l i ty i n stream complex i ty and suggested that as discharge increased, the contr ibut ion f rom storage i n pools and back eddies m a y have contributed to the increase i n storage area. Story (2002) 72 also attributed the var iab i l i ty in storage zone cross-sectional area w i t h i n three sub-reaches to differences i n channel complex i ty . The ratio A s / A increased w i t h discharge i n both reaches. T h i s trend is i n contradict ion to results presented b y H a r v e y et a l . (1996) and M o r r i c e et a l . (1997) in w h i c h A s / A decreased w i t h discharge. Patschke (1999) reported a s imi la r trend for the lower reach; however , a decreasing trend i n A s / A was observed for the upper reach. D ' A n g e l o et al . (1993) and M o r r i c e et al . (1997) reported transient storage zones that were as large as or larger than the surface water area. M u l h o l l a n d et al . (1997) determined that metabol ic rates and nutrient uptake were s ignif icant ly greater i n streams w i t h a larger transient storage zone relative to the channel cross-sectional area ( A s > A ) . The transient storage area ( A s ) i n East Creek was consistently smaller (on average 70%) than the surface water area ( A ) . Estimates o f A s / A ranged from 0.1 to 0.6 i n both reaches. A t higher discharges, l o w - l y i n g areas adjacent to the ma in channel i n East Creek were f i l l ed w i t h stagnant water w h i c h m a y have p rov ided addit ional surface storage. A s w e l l , at higher f lows, a side channel at mid- reach (approximately 45 - 50 m) downstream from the culvert at M road was activated. W o n d z e l l and Swanson (1996a) observed significant exchange f low between p r imary and secondary stream channels. A t higher discharges more wetted channel area m a y also be avai lable for hyporheic exchange to occur. The transient exchange coefficient remained fa i r ly constant w i t h discharge. Some evidence o f a threshold response cou ld be suggested, as transient exchange plateau to a value o f 2.7 x 10~4 s"' at 5.5 L / s and then decl ined wi th discharge i n the lower reach. M o r r i c e et a l . (1997) also reported a threshold response i n w h i c h a increased then decreased w i t h discharge i n a first-order stream. A d d i t i o n a l studies have reported either a steady increase ( D ' A n g e l o et al . 1993, Harvey.et a l . 1996, Hart et a l . 1999, W o n d z e l l 2005) or a decrease i n transient exchange w i t h discharge ( L e g r a n d - M a r c q and Laudelout 1985, Hart et a l . 1999, Patschke 1999). 73 4.1.3. i. Residence times and retention Residence times were higher i n the stream channel than i n the transient storage zone for a l l tracer injections. Residence times also var ied spat ial ly (i.e. between reaches), but d id not vary temporal ly w i t h discharge (Figure 3.6). Be tween reaches, stream channel and transient storage residence times were consistently higher i n the lower reach. Th i s var ia t ion m a y be attributed to stream complex i ty . W o n d z e l l (2005) observed that the storage zone residence t imes and the hydraul ic retention factor (R|,) were greater i n reaches w i t h l o g j a m s than i n a companion pool-step reach. The Rh factor was also consistently greater in unconstrained reaches than constrained reaches. Retent ion factors were fa i r ly comparable between the two reaches i n East Creek (Table 3.5), despite the upper reach hav ing a greater number o f l o g steps and a higher degree o f inc i s ion . The lower reach was generally less conf ined than the upper reach, w h i c h m a y exp la in the longer residence times observed for transient storage. Reach-scale estimates o f residence times and retention are not consistent w i t h estimates for ind iv idua l pools dur ing two tracer injections. F o r both pools located i n the upper reach (Pools 1 and 2), storage zone residence times were greater than i n the stream channel. Storage zone residence times were also greater than reach scale estimated values. H a r v e y et al . (1996) assumed that the residence t ime for the in-channel transient storage is very short and is therefore accounted i n the dispersion coefficient rather than the transient storage coefficient. Results f rom T S M show that residence times w i t h i n the p o o l are higher than the m a i n channel . T h i s suggests that it m a y be i n v a l i d to assume that the residence times w i t h i n the in-channel storage zones (i.e. pools or back eddies) are negl igible . Howeve r , this pattern was not comple te ly supported i n the p o o l located i n the lower reach (Poo l 3). Residence t imes i n both the stream channel and storage zones were between 1 to 2 m i n and were lower than the reach-scale estimates o f 196 and 39 m i n , respectively. 74 4.1.4. In-channel transient storage A major c r i t i c i sm o f current transient storage models is the inabi l i ty to separate in-channel transient storage from storage wi th in the hyporheic zone (Harvey et a l . 1996). Based on the assumption that the dominant transient storage processes w i t h i n the pools are from in-stream transient storage, T S M simulat ions conducted at the scale o f ind iv idua l pools were used to separate in-channel storage from hyporheic exchange. Transient storage area ( A s ) was generally higher w i th in the pools , resulting i n a higher A s / A ratio than at the reach scale (Table 3.6). T h e transient storage area parameter (As ) is assumed to incorporate both storage processes despite the inab i l i ty to separate in-channel transient storage from hyporhe ic zone storage (Harvey et al . 1996). T h e results f rom the ind iv idua l poo l s imulat ions suggest that this m a y be a v a l i d assumption. 4.2. Channel-unit scale 4.2.1. Variability in exchange flow pathways Qual i ta t ive and quantitative observations made at the channel-unit scale h ighl ight the temporal var iab i l i ty i n exchange f lows. W i t h i n one channel-unit (Poo l 4), two separate f low pathways were observed (Figure 3.11). O n J u l y 11 and Augus t 14, R W T tracer water infiltrated the bed, traveled laterally due to the deflect ion from a large anchor boulder i n the r iparian zone and returned to the stream at the base o f the step. T h i s confirms that exchange f l o w pathways can include a lateral f low component, as described by mode l 2 i n Sect ion 1.2.2. H o w e v e r , i n this channel-unit, lateral f low was a result o f the step morphology , and therefore does not conf i rm whether zones o f u p w e l l i n g are the result o f lateral in f low or return f low from the r iparian zone and adjacent h i l l s lope . O n September 27, R W T infiltrated the streambed, traveled ver t ica l ly through the step and returned to the stream at the start o f the poo l . T h i s is described as mode l 1 a, or the typ ica l f low pathway, i n Sect ion 1.2.2. A d d i t i o n a l injections i n a different locat ion wi th in the channel-unit ( locat ion 1-4) and i n the upper reach ( locat ion 1-1) conf i rmed M o d e l l a . D i scha rg ing B r i l l i a n t B l u e F C F 75 tracer water was on ly observed on one occas ion w i t h i n the upper reach (June 27). E v e n though R W T tracer was injected i n a comparable quantity at this loca t ion ( locat ion 1-1), d ischarging tracer water was not observed w i t h the log-step i n the upper reach. T h i s observation m a y be due to Rhodamine W T absorbing to the sediments or organics i n the streambed. Rhodamine has been observed to absorb to sediments i n laboratory experiments, especial ly finer sediments ( M u n n and M e y e r 1988). Exchange f lows are induced b y the hydraul ics o f streamflow over an irregular streambed i n the process k n o w n as "advect ive p u m p i n g exchange" as described b y Savant et a l . (1987) and Thibodeaux and B o y l e (1987). These exchange f lows are control led by the spatial and temporal var iab i l i ty i n hydraul ic heads a long the stream boundary, result ing i n a distr ibution o f travel times through the hyporheic zone. Observat ions made at the channel-unit scale h ighl ight the var iab i l i ty i n residence times associated wi th transient storage. Solute residence times w i t h i n an ind iv idua l p o o l were consistently greater than those i n the step (or hyporheic zone) based on m o d e l l i n g the mean residence t ime o f solutes w i t h i n both storage zones us ing l inear reservoir theory. M e a n residence times var ied over between experiments poss ib ly due to var iab i l i ty i n streamflow and hydraul ic gradients (Table 3.7). Previous studies have reported a w i d e range o f residence times i n the hyporheic zone; for example, W o r m a n et a l . (2002) reported residence times i n the range o f 160 to 800 minutes E x a m i n i n g the breakthrough curves from these stream tracer experiments also highl ights the var iabi l i ty in exchange f low pathways observed w i t h i n a smal l spatial area (Figure 3.14, 3.15). E lec t r i ca l conduct iv i ty increased rap id ly to a peak concentration w i t h i n the step and poo l sub-units. The m a x i m u m E C was also greater i n the step than poo l . T h i s cou ld be an effect o f dispersion or poss ib ly indicate a d i lu t ion i n concentration w i t h i n the poo l . However , addit ional minor peaks i n E C are v i s ib l e on the rising and recessional l i m b o f the breakthrough curves for each experiment, w h i c h cou ld indicate separate f low pathways w i t h different residence times w i t h i n the channel-unit . 76 4.2.2. * Transient storage modelling M o d e l l i n g residence t imes w i t h i n the poo l us ing linear reservoir theory indicated that the poo l behaved l i k e a cont inuously stirred tank reactor ( C S T R ) . Residence t imes w i t h i n the poo l were higher than the hyporhe ic zone. Th i s suggests that it m a y be an i n v a l i d assumption to assume that the residence times wi th in pools are negl ig ib le (Harvey et al . 1996). Results f rom O T I S P - P simulat ions i n P o o l 1 and 2 also support this hypothesis. Residence times w i t h i n both transient storage zones (i.e. step and pool sub-units) fo l lowed an exponential dis tr ibut ion. These results support the use o f an exponential residence t ime dis tr ibut ion to mode l transient storage. Late- t ime solute residence times or " t a i l i ng" o f solute concentration after the m a i n solute pulse, are represented as an exponential probabi l i ty density i n current T S M ' s (Benca la and Walters 1983). A d d i t i o n a l studies have suggested alternative approaches to represent the t imescale o f transient exchange inc lud ing T S M ' s w i t h mul t ip le storage zones ( C h o i et al . 2000). Howeve r , the results from this study suggest that it m a y be v a l i d to use one transient storage zone to represent residences times us ing an exponential distr ibution. 4.3. Point scale 4.3.1. Interpretation of flow pathways Repeated observations o f vert ical hydrau l ic gradients measured f rom piezometers installed w i t h i n the stream channel conf i rmed the currently accepted conceptual mode l o f exchange f low w i t h i n step-pool streams, w h i c h invo lves inf i l t rat ion into the bed upstream o f a step and subsequent discharge a short distance downstream i n the bot tom o f p o o l as described by H a r v e y and B e n c a l a (1993). In East Creek, d o w n w e l l i n g f low was observed upstream o f obstructions i n the channel (i.e. steps and logs) w i t h u p w e l l i n g occur r ing at the start o f the p o o l be low a step (Figure 3.5). Patschke (1999) s imi l a r ly observed strong d o w n w e l l i n g i n piezometers located upstream from steps i n East Creek. In contrast to the observations o f u p w e l l i n g 77 observed i n East Creek and by M o o r e et a l . (2005a) at another stream i n M a l c o l m K n a p p Research Forest, W o n d z e l l (2005) failed to locate areas o f u p w e l l i n g w i t h i n a step-pool reach, despite predictions from groundwater f low models that u p w e l l i n g should occur. T h i s may be par t ia l ly explained by that author's d i f f icul ty w i th ins ta l l ing piezometers i n the streambed without s ignif icant ly dis turbing the bed sediments surrounding the piezometer (S. W o n d z e l l , pers. comm.) . A d d i t i o n a l studies i n the L o o k o u t Creek basin (Oregon, U S A ) , have also not observed coherent hyporheic discharge b e l o w steps, despite predict ions from groundwater f l ow models that u p w e l l i n g should occur (Anderson et al . 2005, G o o s e f f et a l . 2005). A l t h o u g h u p w e l l i n g was observed i n locations where it was expected i n East Creek, the hydrau l ic gradients were substantially weaker than i n the downstream sections o f the step-pool units (i.e., X / L > 0.6). G i v e n that hyporhe ic exchange should be approximately at a steady state dur ing extended periods o f baseflow, u p w e l l i n g and d o w n w e l l i n g fluxes should balance. T h e fact that d o w n w e l l i n g gradients were general ly stronger and more spat ial ly widespread than u p w e l l i n g gradients cou ld be expla ined b y one or more o f the f o l l o w i n g : (1) greater hydrau l ic conduct ivi t ies i n u p w e l l i n g zones, (2). discharge be ing concentrated a long preferred pathways, and (3) the presence o f lateral f l ow paths. A l l three are possibi l i t ies at East Creek. The presence o f higher conduct ivi t ies i n u p w e l l i n g zones w o u l d be consistent w i t h the observation that penetration o f fine sediments into the bed (resulting i n c logg ing o f larger pore spaces) is greater i n d o w n w e l l i n g areas (Scha lch l i 1992, P a c k m a n and M a c K a y 2003). Furthermore, there is weak empir ica l evidence supporting the contrast i n hydrau l ic conduct ivi t ies between u p w e l l i n g and d o w n w e l l i n g zones (Figure 3.23). There is some support for the presence o f preferred pathways from the unit-scale tracer experiments, al though it is diff icul t to quantify h o w m u c h discharge occurs v i a these preferred pathways (section 3.4). T h e presence o f a lateral component to hyporheic exchange is possible, par t icular ly at some areas a long the reach, but the study design d id not a l l ow the assessment o f lateral hyporheic exchange because hor izontal gradients were not measured. Zones o f hyporheic discharge and recharge var ied w i t h pos i t ion i n the stream channel (as defined b y X / L , Tab le 3.9). Genera l ly , ' for X / L > 0.2, negative hydrau l ic gradients increased wi th distance to the step; however , the actual step height does not 78 appear to control V G H (Table 3.9). Where X / L > 0.8, there appears to be a slight • negative trend w i t h increasing step height (Figure 3.22); however, this relationship was not significant. These results suggest that the relat ion between V H G and X / L cou ld be a useful tool for character iz ing and predic t ing hydrau l ic gradients in step-pool streams, a need expressed by B e n c a l a (2000). Channel-uni t spacing, size and sequence are also documented as significant controls on exchange f low i n previous studies (Anderson et al . 2005, G o o s e f f et al . 2006). Howeve r , the influence o f step-height on exchange f low does not appear to have been direct ly quantified i n previous studies. 4.3.2. Interactions between hyporheic flow and lateral inflow Increased lateral or groundwater in f lows from adjacent .hillslopes can reverse head gradients a long the stream marg in , leading to a reduct ion i n hyporheic zone extent and the degree o f hyporhe ic exchange (Harvey and B e n c a l a 1993, W r o b l i c k y et a l . 1998, Storey et a l . 2003, W o n d z e l l 2005). H y d r a u l i c head reversals (fluctuations between upwe l l i ng , d o w n w e l l i n g and neutral) were c o m m o n in the lower reach (Figure 3.7). Groundwater discharge f rom the stream banks cou ld have contributed to the observed reversal o f hydrau l ic gradients. T h i s observation appears to be consistent w i th documented c l imat ic condi t ions, as hydrau l ic head reversals were observed f o l l o w i n g precipi tat ion events dur ing September and October. T h i s m a y indicate that exchange f lows are control led b y seasonal var iab i l i ty i n lateral i n f low from the h i l l s lope , further suggesting that areas o f u p w e l l i n g cou ld be lateral i n f low from the r iparian zone. Lateral in f low steadily decl ined dur ing the early part o f the study per iod ( M a y to Augus t ) due to the atypical dry summer condit ions and general increased w i t h increas ing, discharge. St reamflow measurements also conf i rmed that East Creek is a ga in ing reach; however , net lateral i n f l o w remained less than 1% o f the total streamflow. W o n d z e l l and Swanson (1996a) found that the strength o f hydrau l ic gradients towards the stream varied w i t h season. D u r i n g the summer months, hydrau l ic gradients towards the stream channel were weaker than dur ing the winter months. A reduct ion i n groundwater discharge dur ing the summer months due to less ra infal l and drier so i l condit ions m a y have contributed to the var iab i l i ty i n gradients observed i n this study. 79 M o o r e et a l . (2005b) suggested that the u p w e l l i n g sites m a y be connected to lateral in f lows from the h i l l s lope , par t icular ly those wi th convergent topography based, on observations that u p w e l l i n g sites underwent litt le to no m i x i n g (% = 0) w i t h stream water dur ing tracer injections in one coastal headwater stream. In the lower reach, u p w e l l i n g and d o w n w e l l i n g sites appeared to be h i g h l y connected to water i n the stream channel , based on consistently posi t ive m i x i n g ratios dur ing reach scale tracer injections. These ratios indicate a partial replacement o f hyporheic zone water w i t h tracer-labeled stream water. In the upper reach, negative m i x i n g ratios were observed and the median m i x i n g ratio for most sites was near 0. These findings cou ld indicate that water was be ing drawn from different depths o f the hyporheic zone and thus represent water w i t h a different chemica l signature or residence t ime. 4.3.3. Water fluxes and discharge Observations o f vert ical hydraul ic gradients from piezometers instal led w i t h i n the streambed (n = 66) showed considerable spatial and temporal var iabi l i ty , suggesting that hyporheic exchange or water fluxes into the stream also vary. The h igh within-s i te var iab i l i ty demonstrated is not u n c o m m o n (Thibodeaux and B o y l e 1987, Bax te r and Hauer 2000). P h y s i c a l processes, such as discharge or loca l bedform characteristics, are considered possible controls on temporal and spatial patterns o f exchange f low. Considerable temporal variat ion i n water fluxes inc lud ing inf i l t ra t ion rates and hydrau l ic gradients were observed over the study per iod. H o w e v e r , stream discharge was not a significant phys ica l control on water fluxes into the bed, based on the Spearman's rank correlat ion analysis relating hydrau l ic gradients and infi l t rat ion rates to discharge (Tables 3.3, 3.4). These findings suggest that addit ional mechanisms contribute to the observed temporal var iab i l i ty i n water fluxes, i nc lud ing lateral i n f l o w from the r iparian. H y d r a u l i c conduct iv i ty was also hypothesized to be a significant phys ica l control on the spatial var iab i l i ty o f water fluxes into the streambed. T h e results o f a sequential analysis o f variance determined that hydraul ic conduct iv i ty var ied w i t h site condi t ion (i.e. u p w e l l i n g , d o w n w e l l i n g and neutral gradients), but not between reaches (Table 3.10). H y d r a u l i c conduct iv i ty d i d not appear to vary wi th depth o f instal lat ion i n the subsurface \ . -80 (Figure 3.24). However , hydraul ic conduct iv i ty as calculated f rom streambed infiltrometers (depth - 1 0 cm) , was higher than estimates from fa l l ing head tests (Figure 3.10), suggesting that bed infi l t rat ion computed from piezometer data alone m a y underestimate actual infi l t rat ion rates. These results are i n contradict ion w i t h previous studies, w h i c h observed an increase i n conduct iv i ty w i t h depth ( L a r k i n and Sharp 1992, Conrad and B e l j i n 1996). Those studies attributed the difference i n conduct iv i ty at the streambed to the settling o f silt , c lay and organic materials on the surface i n the process referred to as co lmat ion (Brunke and Gonser 1997). T h i s process can also reduce the degree o f hyporheic exchange. In contrast, supplementary studies examin ing the conduct iv i ty o f porous streambed sediments support the results observed at East Creek (Landon et a l . 2001, Song et a l . 2007). The streambed conduct iv i ty was typ ica l ly greater than the sediments direct ly be low this layer (~ 30 cm) for both studies. Song et al . (2007) hypothesized that hyporhe ic exchange f lows formed loca l i zed pathways w h i c h increased the sediment pore size resul t ing i n an increase i n hydrau l ic conduct ivi ty . The observed spatial var ia t ion i n conduct iv i ty cou ld be caused b y the imprec i s ion o f each measurement method. T h e probable error associated w i t h inf i l t rat ion measurements was almost ± 6 0 % o f the measured value. In addit ion, the higher conduct iv i ty value at the streambed interface cou ld be a result o f the larger sampl ing area o f the infil trometer (diameter ~ 6 cm) compared to the piezometer (diameter ~ 1 cm) , such that the infil trometers are more l i k e l y to capture the effects o f inf i l t rat ion v i a preferred pathways, consistent w i t h the results o f Song et a l . (2007). Saturated hydrau l ic conduct iv i ty has been w i d e l y documented to increase w i t h the v o l u m e o f porous m e d i u m under consideration (Freeze and Cher ry 1979). 4.3.4. Scaling streambed water fluxes The streambed water f luxes computed from D a r c y ' s L a w w i t h i n one channel-unit i n the upper reach (Poo l 1) were s ignif icant ly greater for inf i l t ra t ion (where X / L > 0.6) than for discharge (where X / L < 0.4). T h i s indicates that hyporheic discharge was a lower proport ion o f the total f lux. F luxes into the bed also increased w i t h discharge, as observed on June 19 (Q = 15.4 L / s ) , compared to l o w f low condit ions on September 29 (Q = 1.1. 81 L/s ) . T h e area o f the channel-unit that contributed to the highest proport ion to the total f lux was where X / L > 0.8, w i t h a rate o f over 0.5 m L / s on June 19. Previous studies have suggested that at higher discharges, the hydraul ic potential for d o w n w e l l i n g into the bed increases, thus increasing the potential for hyporheic exchange f low (e.g. W o n d z e l l 2005). A l t h o u g h water fluxes w i t h i n this channel-unit appeared to vary w i t h discharge, hydrau l ic gradients and infi l t rat ion rates measured at the point scale d id not va ry s ignif icant ly w i t h discharge. A s a result, it is difficult to conclude whether discharge was a first order control on water fluxes w i t h i n East Creek. At tempts to "scale-up" the total f lux to the reach scale estimate o f hyporhe ic exchange (s"') indicated that the reach-scale exchange coefficient was two orders larger than the scaled-up estimate o f hyporheic exchange (Table 3.12). Several processes m a y expla in this result. F i r s t ly , us ing D a r c y ' s L a w to calculate water fluxes m a y underestimate the amount o f exchange, due to the tremendous spatial var ia t ion i n hydrau l ic conduct iv i ty . A t least part o f this bias may result f rom underest imation o f K by the s lug tests. I f the infi l trometer measurements are accurate, they suggest that the estimated values o f K m a y be an order o f magnitude too l o w . Secondly , lateral f luxes or hor izonta l exchange f low m a y have contributed to a por t ion o f the hyporhe ic exchange f low that was not quantified at the channel-unit scale. T h i r d l y , the T S M m a y overestimate the amount o f exchange, poss ib ly due to transient exchange i n pools . H o w e v e r , this analysis was on ly conducted wi th in one channel-unit , and should be extended to addit ional channel-units. 82 CHAPTER FIVE: CONCLUSIONS The f inal chapter summarizes the m a i n results o f the thesis research and concludes w i t h areas o f future research. 5.1. Summary of main results H y p o r h e i c zone processes were examined at three spatial scales dur ing the per iod o f M a y to October 2006 i n East Creek: reach scale, channel-unit scale and point scale. A t the reach scale, the breakthrough curves f rom a total o f 10 stream tracer injection experiments were simulated us ing O T I S - P . Solute transport processes varied both temporal ly (wi th variat ions i n discharge) and to a lesser extent, spatial ly (i.e. between reaches). D i spe r s ion rates (D) , channel area ( A ) and transient storage area (As) showed an increasing trend w i t h discharge, w h i l e the transient exchange coefficient (a) remained fa i r ly constant w i t h discharge i n both reaches. T h e ratio A s / A increased w i t h discharge. Stream and storage zone hydrau l ic residence times d id not vary w i t h discharge at the reach scale. Retent ion was highest dur ing l o w f low condit ions. M o d e l parameter uncertainty was greatest dur ing periods o f h igh f lows , poss ib ly confounding the abi l i ty to examine transient storage processes over a range o f f low condit ions. D u r i n g two tracer injections (September 29 and 30), breakthrough curves from ind iv idua l pools, were s imulated i n order to quantify poo l storage and residence times. Residence times w i t h i n the transient storage zone o f the p o o l (assumed to be in-channel storage) were higher than the residence t ime i n for the entire reach. T h e transient storage area (As) was also generally higher w i t h i n the pools , resul t ing i n a higher A s / A ratio than at the reach scale. These results suggest that it m a y be v a l i d to assume that the transient storage area adequately incorporates both storage zone processes at the reach scale. T w o different f l ow pathways were observed dur ing stream tracer experiments conducted at the channel-unit scale. One f low pathway was al igned w i t h the stream channel , w h i l e a second f low pathway inc luded a lateral component w i t h i n the r iparian zone. T h i s second pathway was associated w i t h f l o w around a large boulder. 83 Observations made at this spatial scale highl ight the temporal and spatial var iab i l i ty in exchange f lows. A t the channel-unit scale, stream tracer experiments were used to determine the mean residence t ime o f solutes i n both transient storage zones, speci f ica l ly the hyporheic zone versus in-channel storage i n pools . U s i n g cont inuously stirred tank reactor theory, , mean residence times were found to be greater w i t h i n the poo l than the step (i.e. hyporhe ic zone) dur ing two stream tracer experiments. These results, a long w i t h the results f rom O T I S - P s imulat ions w i t h i n two ind iv idua l pools , suggest that it m a y be i n v a l i d to assume that residence times wi th in pools are negl igible . M o d e l results also conf i rm that the residence t ime distributions wi th in the step and poo l sub-units fo l l ow an exponential dis t r ibut ion, suggesting that current T S M (e.g. O T I S - P ) do accurately represent the late-time solute residence times us ing an exponential p robabi l i ty density . w i t h one transient storage zone. A t the point scale, direct measurements o f water fluxes into the stream bed, i nc lud ing ver t ical hydrau l ic gradients and infi l trat ion rates, showed considerable temporal and spatial var iabi l i ty . H o w e v e r , these water fluxes d id not vary statistically w i t h discharge, suggesting that other processes contribute to the observed var iabi l i ty . V e r t i c a l hydrau l ic gradients var ied systematical ly wi th the scaled locat ion wi th in the channel-unit , ind ica t ing that stream geometry is a significant control on water fluxes. Repeated observations o f V H G , as measured i n piezometers instal led w i t h i n the streambed, indicated a strong d o w n w e l l i n g o f water upstream from obstructions in the streambed such as boulders and logs, corresponding to where X / L > 0.8. Zones o f u p w e l l i n g occurred downstream at the base o f pools , and corresponded to areas i n the stream channel where X / L < 0.2. Step height was not a significant control on hydrau l ic gradients. U p w e l l i n g sites were located wi th in both reaches, al though gradients were stronger w i t h i n the upper reach. Reversa l o f hydraul ic gradients was also more c o m m o n w i t h i n the lower reach, poss ib ly due to h i l l s lope discharge. H y d r a u l i c conduct iv i ty measurements were spatial ly heterogeneous, but were w i t h i n the same order o f magnitude (10" 4 m/s) i n both reaches. However , hydrau l ic conduct iv i ty estimates based on streambed infiltrometers were higher than estimates from fa l l ing head tests. T h i s result suggests that bed inf i l t ra t ion computed from piezometer 84 data alone m a y underestimate actual inf i l t rat ion rates. A sequential analysis o f variance indicated that hydraul ic conduct iv i ty var ied w i t h site condi t ion (i.e. upwe l l i ng , d o w n w e l l i n g and neutral sites), suggesting that conduct iv i ty is an addit ional control on exchange f lows. T h i s f inding is also consistent w i t h the not ion that d o w n w e l l i n g zones should be more inf luenced by the c logg ing o f pore space b y infi l t rat ion o f fine sediment. Channel-uni t water fluxes calculated w i t h D a r c y ' s L a w d i d not "scale-up" to the reach scale estimate o f hyporheic exchange (a), w h i c h were two orders o f magnitude lower than the reach. A d d i t i o n a l processes such as lateral i n f low or transient storage i n pools cou ld have resulted i n the observed differences, i n addi t ion to bias i n the hydrau l ic conduct iv i ty measurements. 5.2. Areas for future research T h i s research contributes to a body o f w o r k examin ing the phys ica l properties o f the stream channel that influence solute transport and retention i n smal l , headwater streams. A hydrometr ic and stream tracer approach was used to characterize the spatial dis t r ibut ion and the associated residence t imes through the hyporheic and surface-water transient storage zone at three spatial scales o f interest i nc lud ing the reach, channel-unit and local or point scale. A mul t ip le scale approach to examine hyporheic exchange has not been exp l i c i t l y appl ied i n previous research, and highl ights the considerable spatial and temporal var iab i l i ty and complex i ty o f hyporhe ic exchange processes w i th in step-poo l streams. Th i s study shows that channel-unit spacing is a dominant control on the magnitude o f exchange f low and the extent o f the hyporheic zone. In addi t ion, this research suggests that a scal ing relationship based on the channel-unit geometry could be a useful and pract ical tool for characterizing and predic t ing exchange f low i n step-pool streams. Cont inued research should focus on examin ing the longi tudinal patterns i n f luv ia l geomorphology, such as channel unit spacing, i n order to characterize and apply the hyporheic exchange f low processes across a broad range o f stream size and scales. Studies attempting to l ink b io log ica l and geochemica l processes to phys ica l characteristics o f the hyporheic zone have h ighl ighted the importance o f the hyporheic zone and in-channel "dead" zones for nutrient uptake and temporary retention o f surface 85 water nutrients. T h e results f rom this study show that in-channel features such as pools and back eddies do contribute to transient storage i n headwater streams. Observat ions made at the channel-unit scale demonstrate that separate residence times for hyporheic and surface-water transient storage zones can be quantified us ing a stream tracer approach. These observations provided insight into the residence t ime dis tr ibut ion o f water i n the hyporheic zone. Howeve r , spatial and temporal repl icat ion was l imi ted i n this study, and future studies should apply the approach to mul t ip le channel-units w i t h i n a single reach, over a range o f f low condit ions. In addit ion, a metr ic relat ing in-channel residence t imes to poo l geometry w o u l d be a valuable contr ibution towards understanding the interplay between channel morpho logy and solute residence times. A cont inuing challenge i n the area o f hyporheic zone reach w i l l be to "scale-up" smal l scale phys ica l measurements to reach-scale observations o f solute transport processes. Cont inued efforts to quantify among-channel-unit var iab i l i ty i n smal l scale water fluxes w o u l d be beneficial i n order to compare hydrometr ic estimates o f exchange rates to the results o f the transient storage mode l values. 86 REFERENCES Ander son , J . K . , S . M . W o n d z e l l , M . N . Goose f f and R . Haggerty . 2005. Patterns o f stream longi tudinal profiles and impl ica t ions for hyporheic exchange f low at the H . J . A n d r e w s Exper imenta l Forest, Oregon, U S . H y d r o l o g i c a l Processes 19: 2931-2949. Baxter , C . V . , and F . R . Hauer. 2000. Geomorpho logy , hyporhe ic exchange, and selection o f spawning habitat by bu l l trout (Salvelinus confluentus). Canadian Journal o f Fisheries and A q u a t i c Sciences 57: 170-181. Baxter , C , F . R . Hauer and W . W . Woessner . 2003. M e a s u r i n g groundwater-stream water exchange: new techniques for ins ta l l ing minipiezometers and estimating hydrau l ic conduct ivi ty . Transactions o f the A m e r i c a n Fisheries Socie ty 132: 493-502. Benca la , K . E . 1984. Interactions o f solutes and streambed sediment 2. A dynamic analysis o f coupled hydro log ic and chemica l processes that determine solute transport. Water Resources Research 20: 1804-1814. Benca la , K . E . 2000. H y p o r h e i c zone hydro log ica l processes. H y d r o l o g i c a l Processes 14: 2797-2798. Benca la , K . E . , and R . A . Walters . 1983. S imula t ion o f solute transport i n a mounta in pool-and-riff le stream: a transient storage mode l . Wate r Resources Research 19: 718-724 . Benca la , K . E . , J . H . Duff , J . W . Harvey , A . P . Jackman and F . J . T r i ska . 1993. M o d e l l i n g w i t h i n the stream-catchment cont inuum. In: A . J . Jakeman, M . B . B e c k and M . J . M c A l e e r (Eds.) . M o d e l l i n g Change i n Env i ronmenta l Systems, W i l e y , N e w Y o r k , pp. 163-187. Boano , F . , C . Camporeale , R . R e v e l l i and L . R i d o l f i . 2006. S inuos i ty -dr iven hyporheic exchange i n meandering rivers. Geophys i ca l Research Letters 33: 1-4. Bou l ton , A . J . , S. F ind lay , P . M a r m o n i e r , E . H . Stanley and H . M . Vale t t . 1998. The functional s ignif icance o f the hyporheic zone i n streams and rivers. A n n u a l R e v i e w o f E c o l o g y and Systematics 29: 59-81. Brunke , M . , and T . Gonser . 1997. The ecologica l s ignif icance o f exchange processes between rivers and groundwater. Freshwater B i o l o g y 37: 1-33. Castro, N . M . , and G . M . Hornberger. 1991. Surface-subsurface water interactions i n an al luviated mounta in stream channel. Water Resources Research 27: 1613-1621. Chapra , S . C . 1997. Surface Water Qua l i t y M o d e l l i n g . M c G r a w - H i l l , N e w Y o r k , U S A . Pp . 46-85. Chapra , S . C , and R . L . R u n k e l . 1999. M o d e l i n g impact o f storage zones on stream disso lved oxygen. Journal o f Env i ronmenta l Eng inee r ing 125: 415-419. C h o i , J . , J . W . H a r v e y and M . H . C o n k l i n . ' 2 0 0 0 . Charac ter iz ing mul t ip le timescales o f stream and storage zone interaction that affect solute fate and transport i n streams. Water Resources Research 36 :1511-1518 . Conrad , L . P . , and M . S . B e l j i n . 1996. Eva lua t ion o f an induced infi l t rat ion mode l as appl ied to g lac ia l aquifer systems. Water Resources B u l l e t i n 32: 1209-1220. Coupe , R . , A . C . Stewart and B . M . W i l k e e m . 1991. Enge lmann Spruce - Subalpine F i r Zone . In D . M e i d i n g e r and J . Pojar (Eds.) . Ecosys tems o f B r i t i s h C o l u m b i a . C r o w n Publ icat ions Inc., V i c t o r i a , B C . 330 pp. 87 D ' A n g e l o , D . J . , J .R. Webster, S .V . Gregory and J .L . Meyer . 1993. Transient storage i n Appalachian and Cascade mountain streams as related to hydraulic characteristics. Journal o f the Nor th Amer ican Bentholological Society 12: 223-235. ^ Duff , J . H . , and F . J . T r i ska . 2000. N i t rogen biogeochemistry and surface-subsurface exchange i n streams. In: Jones J . B . , and P . J . M u l h o l l a n d (Eds). Streams and-Groundwaters . A c a d e m i c Press, San D i e g o . Pp . 197-220. E l l io t t , A . H . , and N . H . B r o o k s . 1997. Transfer o f nonsorbing solutes to a streambed w i t h bedforms: theory. Water Resources Research 33: 123-136. ' Env i ronment Canada. 1993. Canadian C l ima te N o r m a l s 1961-1990. A tmospher i c Env i ronment Service, Ottawa, Canada. Fel ler , M . C . , and J .P . K i m m i n s . 1984. Effects o f clearcutting and slash burning on streamwater chemist ry and watershed nutrient budgets i n southwestern B r i t i s h C o l u m b i a . Water Resources Research 20: 29-40. F ind l ay , S. 1995. Importance o f surface-subsurface exchange i n stream ecosystems: the hyporheic zone. L i m n o l o g y and Oceanography 40: 159-164.' Freeze, R . A . , and J . A . Cher ry . 1979. Groundwater . Prentice H a l l , E n g l e w o o d C l i f f s , N J . Gooseff , M . N . , S . M . W o n d z e l l , R . Hagger ty and J . Anderson . 2003. C o m p a r i n g transient storage m o d e l i n g and residence t ime dis tr ibut ion ( R T D ) analysis i n geomorphica l ly var ied reaches i n the L o o k o u t Creek basin, Oregon, U S A . Advances in Water Resources 26: 925 -937 . Gooseff, M . N . , D . M . M c K n i g h t , R . L . R u n k e l and J . H . Duff . 2004. Deni t r i f ica t ion and hydro log ic transient storage i n a g lac ia l meltwater stream, M c M u r d o D r y V a l l e y s , Antarc t ica . L i m n o l o g y and Oceanography 49: 1884-1895 . Gooseff , M . N . , and B . L . M c G l y n n . 2005. A stream tracer technique emp loy ing ion ic tracers and specif ic conductance data appl ied to the M a i m a i catchment, N e w Zealand. H y d r o l o g i c a l Processes 19: 2491-2506. Gooseff, M . N . , K . E . Benca la , D . T . Scott, R . L . R u n k e l and D . M . M c K n i g h t . 2005. Sens i t iv i ty analysis o f conservative and reactive stream transient storage models appl ied to f ie ld data f rom mult iple-reach experiments. Advances i n Water Resources 28: 4 7 9 - 4 9 2 . Gooseff , M . N . , J . K . Ande r son , S . M . W o n d z e l l , J . L a N i e r a n d R . Haggerty. 2006. A m o d e l l i n g study o f hyporheic exchange pattern and the sequence, size, and spacing o f stream bedforms i n mounta in stream networks, Oregon, U S A . H y d r o l o g i c a l Processes 20: 2443 -2457 Gooseff , M . N . , R . O . H a l l Jr., and J . L . Tank. 2007. Re la t ing transient storage to channel complex i ty i n streams o f va ry ing land use i n Jackson H o l e , W y o m i n g . Water Resources Research 43: W 0 1 4 1 7 , do i : 10 .1029 /2005WR004626 . Guenther, S. 2007. Headwater stream temperature response to alternative r iparian management: an experimental heat budget approach. M . S c . thesis, U n i v e r s i t y o f B r i t i s h C o l u m b i a , Vancouve r , B r i t i s h C o l u m b i a . Haggerty, R . , S . M . W o n d z e l l and M . A . Johnson. 2002. Power - l aw residence t ime dis tr ibut ion i n the hyporheic zone o f a 2nd-order mounta in stream. Geophys ica l Research Letters 29: 1640. do i :10 .1029 /2002 /GL014743 . Hart , B . T . , B . M a h e r and I. Lawrence . 1999. N e w generation water qual i ty guidelines for ecosystem protection. Freshwater B i o l o g y 41 : 347-359. 88 Harvey , J . W . , and B . J . Wagner . 2000. Quant i fy ing hydro log ica l interactions between streams and their subsurface hyporheic zones. In: J . B . Jones, and P . J . M u l h o l l a n d (Eds) . Streams and G r o u n d Waters. A c a d e m i c Press, San D iego , C A . Harvey , J . W . , and K . E . Benca la . 1993. The effect o f streambed topography on surface-subsurface water exchange in mountain catchments. Water Resources Research 29: 89-98. Harvey , J . W . , B . J . Wagner and K . E . Benca la . 1996. Eva lua t ing the re l iabi l i ty o f the stream water tracer approach to characterize stream subsurface water exchange. Water Resources Research 32: 2441-2451. H i l l , A . R . , and D . J . Lymburne r 1998. Hyporhe ic zone chemistry and stream-sub surface exchange i n two groundwater-fed streams. Canad ian Journal o f Fisheries and A q u a t i c Sciences 55: 495-506. H i l l , A . R . , C F . Labad ia and K . Sanmugadas. 1998. H y p o r h e i c zone hyd ro logy and , nitrogen dynamics i n relation to the streambed topography o f a N - r i c h stream. B i o g e o c h e m i s t r y 4 2 : 285-310. H v o r s l e v , M . J . 195.1. T i m e lag and soi l permeabi l i ty i n groundwater observations. U . S . A r m y Corps o f Engineers , Waterways Exper iment Station, 50 pp. Hutch inson , D . G . , and R . D . M o o r e . 2000. Throughf low var iab i l i ty on a forested h i l l s lope underlain by compacted g lac ia l t i l l . H y d r o l o g i c a l Processes 14: 1751-1766. Johnson, R . A . 1980. O x y g e n transport i n sa lmon spawning gravels. Canadian Journal o f Fisheries and A q u a t i c Sciences 55: 495-506. Johnstone, D . E . 1988. Some recent developments o f constant injection salt d i lu t ion gauging i n rivers. Journal o f H y d r o l o g y ( N . Z . ) 27: 128-153. Jones, J . B . , and P . J . M u l h o l l a n d . 2000. Streams and Groundwaters . A c a d e m i c Press, San D i e g o . Jones, J .J . , and R . M . Ho lmes . 1996. Surface-subsurface interactions i n stream ecology. Trends i n E c o l o g y and E v o l u t i o n 11: 239-242. Kasahara , T . , and S . M . W o n d z e l l . 2003. Geomorph ic controls on hyporheic exchange f low in mounta in streams. Water Resources Research 39: 1005. do i :1010-1029 /2002WR001386 . K l i n k a , K , and V . J . Kra j ina . 1986. Ecosystems o f the U n i v e r s i t y o f B r i t i s h C o l u m b i a Research Forest, Haney , B . C . Facu l ty o f Forestry, The Un ive r s i t y o f B r i t i s h C o l u m b i a , Vancouve r B . C . Kutner , M . H . , C . J . Nachtshe im, J . Neter and W . L i . 2004. A p p l i e d l inear statistical models ( 5 t h E d ) . M c G r a w - H i l l I rwin , N e w Y o r k , U S A . L a n d o n , M . K . , D . L . Rus and F . E . Harvey . 2001. C o m p a r i s o n o f instream methods for measuring hydraul ic conduct iv i ty i n sandy streambeds. Groundwater 39: 870-885. L a r k i n , R . G . , and J . M . Sharp. 1992. O n the relat ionship between r iver-basin geomorphology, aquifer hydraul ics , and ground-water f low direct ion i n a l l uv i a l aquifers. G e o l o g i c a l Socie ty o f A m e r i c a B u l l e t i n 104: 1608-1620. L e g r a n d - M a r c q , C , and H . Laudelout . 1985. Long i tud ina l dispersion i n a forested stream. Journal o f H y d r o l o g y 78: 317-324. M a i n d o n a l d , J . , and W . J . Braun . 2007. Data A n a l y s i s and Graphics us ing R - an Example -Based A p p r o a c h . 2 n d E d . Cambr idge U n i v e r s i t y Press, Cambr idge U K . M a r t i n , J . E . 1996. H y d r o l o g y and pore-water chemistry o f a t idal marsh, Fraser R i v e r estuary. M . S c . Thesis . S i m o n Fraser Un ive r s i ty , Vancouver , B r i t i s h C o l u m b i a . 89 M o o r e , R . D . 2004a. Introduction to salt d i lu t ion gauging for streamflow measurements: Part 1. Streamline Watershed Management B u l l e t i n 7: 20-24. M o o r e , R . D . 2004b. Const ruct ion o f a Mar io t te bottle for constant-rate tracer injection into smal l streams. Streamline Watershed Management B u l l e t i n 8: 15-16. M o o r e , R . D . , P . Sutherland, T . G o m i and A . D h a k a l . 2005a. Thermal regime o f a headwater stream w i t h i n a clearcut, coastal B r i t i s h C o l u m b i a , Canada. H y d r o l o g i c a l Processes 19: 2591-2608. M o o r e , R . D . , T . G o m i , , A . Story and E . M e l l i n a . 2005b. Charac ter iz ing surface-subsurface interactions i n headwater streams. Poster. A G U F a l l Mee t ing . San Francisco. 12/2005. M o r r i c e , J . A . , H . M . Valet t , C . N . D a h m and M . E . Campana . 1997. A l l u v i a l characteristics, groundwater-surface water exchange and hydro log ic retention i n headwater streams. H y d r o l o g i c a l Processes 1: 253-267. M u l h o l l a n d , P . J . , E . R . M a r z o l f , J .R . Webster, D . R . Hart , and S.P. Hendr icks . 1997. Evidence that hyporheic zones increase heterotrophic metabol i sm and phosphorus uptake i n forest streams. L i m n o l o g y and Oceanography 42: 443-451. M u l h o l l a n d , P'. J . , A . D . Steinman, E . R . M a r z o l f , D . R . Hart and D . L . D e A n g e l i s . 1994. Effect o f per iphyton biomass on hydraul ic characteristics and nutrient c y c l i n g i n streams. Oeco log ia 98: 40-47. M u n n , N . L . , and J . L . M e y e r . 1988. R a p i d f l ow through the sediments o f a headwater stream i n the southern Appa lach ians . Freshwater B i o l o g y 20: 235-240. N e w b o l d , J .D . , J . W . E l w o o d , R . V . O ' N e i l l and W . V a n W i n k l e . 1981. M e a s u r i n g nutrient spi ra l ing i n streams. Canadian Journal o f Fisheries and A q u a t i c Sciences 3 8 : 8 6 0 - 9 0 N R C 2000. C l ean coastal waters: Unders tanding and reducing the effects o f nutrient po l lu t ion . Na t iona l A c a d e m y Press, Wash ing ton , D . C . , U S A . Packman , A . I . , and J .S. M a c K a y . 2003. Interplay o f stream-subsurface exchange, c l ay particle deposit ion, and streambed evolut ion. Water Resources Research 39: E S G 4-1 to 4-9. do i :10 .1029 /2002WR001432 . Patschke, S . N . 1999. H y p o r h e i c exchange i n a forested headwater stream. M . S c . thesis, S i m o n Fraser Univers i ty , S i m o n Fraser Unive r s i ty , Vancouver , B r i t i s h C o l u m b i a . Phanikumar , M . S . , I. A s l a m , C . Shen, D . T . L o n g and T . C . V o i c e . 2007. Separating surface storage from hyporheic retention i n natural streams us ing wavelet decomposi t ion o f acoustic Dopp le r current profi les. Water Resources Research . 43: W 0 5 4 0 6 , do i : 10 .1029 /2006 /WR005104 . R Development Core Team. 2007. R : A language and environment for statistical computing. R Foundat ion for Statistical Compu t ing , V i e n n a , Aus t r i a . I S B N 3-900051-07-0, U R L ht tp: / /www.R-project .org. R e i d y , C A . 2004. V a r i a b i l i t y o f hyporhe ic zones in Puget Sound l o w l a n d streams. M . S c . thesis. Un ive r s i t y o f Wash ing ton , Washington , U S A . R u n k e l , R . L . 1998. One -Dimens iona l Transport w i t h Inf low and Storage ( O T I S ) : a Solute Transport M o d e l for Streams and Rive r s . Water-Resources Investigations . Report 98-4018. U . S . G e o l o g i c a l Survey. 43 pp. R u n k e l , R . L . 2002. A new metr ic for determining the importance o f transient storage. Journal o f the N o r t h A m e r i c a n Ben tho log ica l Socie ty 21 : 529 -543 . 90 R u n k e l , R . L . , and R . E . Broshears. 1991. One-d imens ional transport w i th in f low and storage ( O T I S ) : a solute transport mode l for smal l streams. Bou lde r , Co lo rado , U n i v e r s i t y o f Co lo rado , C A D W E S Techn ica l Report 91-01, 85 pp. Ryan , R . J . , and M . C . Boufadel. 2007. Evaluat ion o f streambed hydraulic conductivity heterogeneity in an urban watershed. Stochastic Environmental Research and R i s k Assessment 21: 309-316. Savant, S . A . , D . D . R e i b l e and L . J . Thibodeaux. 1987. Convec t ive transport w i t h i n stable r iver sediments. Water Resources Research 23: 1763-1768. Scha l ch l i , U . 1992. The c l o g g i n g o f course gravel r iver beds b y fine sediment. H y d r o b i o l o g i a 235/236: 189-197. Song, J . , X . C h e n , C . Cheng , S. Summerside and F . W e n . 2007. Effects o f hyporhe ic processes on streambed vert ical hydrau l ic conduct iv i ty i n three r ivers o f Nebraska . Geophys i ca l Research Letters 34: L 0 7 4 0 9 , do i : 10 .1029/2007/GL029254 . Stanford, J . A . , and J . V . W a r d . 1988. The hyporhe ic habitat o f r iver ecosystems. Nature 3 3 5 : 6 4 - 6 6 . Stanford, J . A . , and J . V . W a r d . 1993. A n ecosystem perspective o f a l luv ia l r ivers: Connec t iv i ty and the hyporheic corridor. Journal o f the N o r t h A m e r i c a n Ben tho log ica l Socie ty 12: 48-60. v Storey, R . G . , K . W . F . H o w a r d and D . D . W i l l i a m s . 2003. Factors con t ro l l ing riff le-scale hyporheic exchange f lows and their seasonal changes i n a ga in ing stream: A three-dimensional groundwater f low mode l . Water Resources Research 39: 1034. do i : 10 .1029/2002/WR001367 . Story, A . C . 2002. Creek temperatures in shaded reaches downstream o f forestry activit ies, Centra l B r i t i s h C o l u m b i a . M . S c . thesis, Un ive r s i ty o f B r i t i s h C o l u m b i a , Vancouve r , B r i t i s h C o l u m b i a . Story, A . C , R . D . M o o r e and J .S. M a c d o n a l d . 2003. Stream temperature i n two shaded reaches be low cutblocks and logg ing roads: downstream c o o l i n g l i nked to subsurface hydro logy . Canadian Journal o f Forest Research 33: 1383-1396. Stream Solute W o r k s h o p . 1990. Concepts and methods for assessing solute dynamics i n stream ecosystems. Journal o f N o r t h A m e r i c a n Ben tho log ica l Socie ty 9: 95-119. Thibodeaux , L . J . , and J .O . B o y l e . 1987. B e d - f o r m generated convect ive transport i n bot tom sediment. Nature 325 :341-3 43 . Ton ina , D . , and J . M . Buff ington . 2005. B iogeomorpho logy : Effects o f sa lmon redds on r iver hydraul ics and hyporheic f low i n gravel-bed rivers. A G U F a l l M e e t i n g . San Franc isco . 12/2005. T r i ska , F . J . , V . C Kennedy , R . J . A v a n z i n o , G . W . Ze l lwege r and K . E . Benca la . 1989. Retent ion and transport o f nutrients i n a third order stream i n northwestern Ca l i fo rn i a : hyporheic processes. E c o l o g y 70: 1893-1905. T r i ska , F . J . , J . H . D u f f and R . J . A v a n z i n o . 1993. Patterns o f hydro log ica l exchange arid nutrient transformation i n the hyporheic zone o f a gravel bot tom stream: E x a m i n i n g terrestrial-aquatic l inkages. Freshwater B i o l o g y 29: 259-274. Valet t , H . M . , S . G . Fisher^ N . B . G r i m m and P . C a m i l l . 1994. V e r t i c a l hyd ro log ic exchange and ecologic stability o f a desert stream ecosystem. E c o l o g y : 75: 548-560. 91 Valet t , H . M . , J . A . M o r r i c e , C . N . D a m n and M . E . Campana . 1996. Parent l i tho logy, surface-groundwater exchange and nitrate retention,in headwater streams. L i m n o l o g y and Oceanography 41 : 333-345. Verv ie r , P . , J . Gibert , P . M a r m o n i e r and M . J . D o l e - O l i v i e r . 1993. A perspective on the permeabi l i ty o f the surface freshwater-groundwater ecotone. Journal o f the N o r t h A m e r i c a n Ben tho log ica l Socie ty 11: 93-102. Wagner , B . J . , and J . W . Harvey . 1997. Exper imenta l design for estimating parameters o f rate-l imited mass transfer: A n a l y s i s o f stream tracer studies. Water Resources Research 33: 1731-1741. Wagner , F . H . , and G . Bre tschko. 2002. Interstitial f l ow through preferential f low paths i n the hyporheic zone o f the Oberer Seebach, Aus t r i a . A q u a t i c Sciences 64: 307-316. Webster, J .R. , and B . C . Patten. 1979. Effects o f watershed perturbation on stream potassium and c a l c i u m dynamics . E c o l o g i c a l Monographs 19: 51-72. Webster, J .R . , and T . P . Eh rman . 1996. Solute dynamics . In: Haver , F . R . , and G . Lamber t i . (Eds.) . Me thods i n stream ecology. A c a d e m i c Press, Toronto , O N . W h i t e , D . S . 1993. Perspectives on def ining and del ineat ing hyporheic zones. Journal o f N o r t h A m e r i c a n Ben tho log ica l Socie ty 12: 472-475. W o n d z e l l , S . M . , and F . J . Swanson. 1996a. Seasonal and storm dynamics o f the hyporhe ic zone o f a 4th-order mounta in stream I: H y d r o l o g i c a l processes. Journal o f the N o r t h A m e r i c a n Ben tho log ica l Socie ty 15: 3-19. W o n d z e l l , S . M . , and F . J . Swanson. 1996b. Seasonal and storm dynamics o f the hyporhe ic zone o f a 4th-order mounta in stream II: N i t rogen c y c l i n g . Journal o f the N o r t h A m e r i c a n Ben tho log ica l Socie ty 15: 20-34. W o n d z e l l , S . M . 2005. Effect o f morpho logy and discharge on hyporheic exchange f lows i n two smal l streams i n the Cascade Moun ta ins o f Oregon, U S A . H y d r o l o g i c a l Processes 20: 267-287. W o r m a n , A . , A . I . Packman , H . Johansson and K . Jonsson. 2002. Effect o f f low- induced exchange i n hyporheic zones on longi tudinal transport o f solutes i n streams and rivers. Water Resources Research 3 8 : 1 - 1 5 . W r o b l i c k y , G . J . , M . E . Campana , H . M . Vale t t and C . N . D a h m . 1998. Seasonal var ia t ion i n surface-subsurface water exchange and lateral hyporheic area o f two stream-aquifer systems. Water Resources Research 34: 317-328. Zarnetske, J . P . , M . N . Gooseff, T . R . Bros ten , J . H . Bradford , J .P. M c N a m a r a and W . B . B o w d e n . 2007. Transient storage as a function o f geomorphology, discharge, and permafrost active layer condi t ions i n A r c t i c tundra streams. Water Resources Research 43: W 0 7 4 1 0 , do i : 10 .1029 /2005WR004816 . Z immerrhann , A . , and M . C h u r c h . 2001 . Channe l morphology , gradient profiles and bed stresses dur ing f lood i n a step-pool channel . G e o m o r p h o l o g y 40: 311-327. 92 A P P E N D I X A : M O D E L S I M U L A T I O N S 0.0 0.5 1.0 1.5 2.0 Time (hour) Figure A . l . M o d e l s imulat ions us ing O T I S - P for M a y 31 for the lower reach F igure A . 2 . M o d e l s imulat ions us ing O T I S - P for June 27 for the lower reach Figure A . 3 . M o d e l s imulat ions us ing O T I S - P for Sept 21 for the upper reach (a) and the lower reach (b) Time (hour) Figure A . 4 . M o d e l s imulat ions us ing O T I S - P for Sept 29 for the upper reach mAA / £ A * A ^ z & A Upper boundary & D Lower boundary Time (hour) Figure A . 5 . M o d e l s imulat ions us ing O T I S - P for Sept 30 for the lower reach o ro o c o O ro cu a: o o o o o o CO o o o CM o o o A A ^ A ^ 0.0 A Upper boundary D Lower boundary — Simulated ( A ) -a-—r~ 2.5 Time (hour) c 0) o c o O ro CD rr o o o r^ o o o co o o o CM o o . o ZAA A A A ^ A A A A A A A Upper boundary D Lower boundary — Simulated " I — 0.0 Time(hour) Figure A . 6 . M o d e l s imulat ions us ing O T I S - P for October 20 for the upper reach (a) and the lower reach (b) 96 c o 1 <3 '16 o 8 A A ^ A A A A A Upper boundary ° Lower boundary Simulated (A) n Time (hour) Time (hour) Figure A . 7 . M o d e l s imulat ions us ing O T I S - P for September 29. Results are f rom two pools located i n the upper reach 97 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0101232/manifest

Comment

Related Items