Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

1 9.13 Geomorphic Controls on Hyporheic Exchange Across Scales - Watersheds to Particles 2 3 Steven M. Wondzell 4 U.S. Forest Service, 5 Pacific Northwest Research Station, 6 Olympia Forest Sciences Laboratory, 7 Olympia, WA 98512 USA. 8 Phone: 360-753-7691 9 E-mail: [email protected] 10 11 Michael N. Gooseff 12 Civil & Environmental Engineering Department, 13 Pennsylvania State University, 14 University Park, PA 16802 USA 15 Phone: 814- 867-0044 16 E-mail: [email protected]

1 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

17 Abstract 18 19 We examined the relationship between fluvial geomorphology and hyporheic exchange flows. 20 We use geomorphology as a framework to understand hyporheic process and how these 21 processes change with location within a stream network, and over time in response to changes in 22 stream discharge and catchment wetness. We focus primarily on hydostatic and hydrodynamic 23 processes – the processes where linkages to fluvial geomorphology are most direct. Hydrostatic 24 processes result from morphologic features that create elevational head gradients whereas 25 hydrodynamic processes result from the interaction between stream flow and channel 26 morphologic features. We provide examples of the specific morphologic features that drive or 27 enable hyporheic exchange and we examine how these processes interact in real stream networks 28 to create complex subsurface flow nets through the hyporheic zone. 29 30 31 Key words 32 33 Hyporheic, step-pool sequence, pool-riffle sequence, meander bends, back channels, floodplain 34 spring brooks, mid-channel islands, stream bedforms, pumping exchange, saturated hydraulic 35 conductivity.

2 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

36 9.13.1. Introduction 37 38 Hyporheic exchange flow (HEF) is the movement of stream water from the surface channel into 39 the subsurface and back to the stream (Figure 1). Stream water in hyporheic flow paths may mix 40 with groundwater so that the relative proportion of stream-source water in the hyporheic zone is 41 highly variable, ranging from 100% stream water to nearly 100% groundwater. Also the 42 residence time distribution of stream water in the hyporheic zone tends to be highly skewed, with 43 most of the stream water moving along short flow paths and thus having short residence times 44 (hours), but some water either moving on long flow paths or encountering relatively immobile 45 regions having very extended residence times (weeks to months, or longer). The boundaries of 46 the hyporheic zone are arbitrary, usually defined by the amount of stream-source water present in 47 the subsurface. Triska et al. (1989) set a threshold of 10% stream-source water to define the 48 limits of the hyporheic zone so that regions with <10% stream-source water were defined as 49 groundwater. Alternatively, the extent of the hyporheic zone can be delimited by water residence 50 time, for example, the subsurface zone delineated by hyporheic exchange flows with residence 51 times less than 24 hours (the 24-h hyporheic zone; Gooseff, in press). 52 53 The objective of this chapter is to examine the relation between geomorphology and hyporheic 54 processes. The two primary controls on hyporheic exchange are the gradients in total head 55 established along and across streambeds and the hydraulic conductivity of the streambed and 56 adjacent aquifer, both of which are significantly influenced by geomorphology. Total head (also 57 known as potential) is the sum of pressure head, elevation head, and velocity head. Pressure head 58 represents height of a column of fluid to produce pressure. Velocity head represents the vertical 59 distance needed for the fluid to fall freely (neglecting friction) to reach a particular velocity from 60 rest. Elevation head represents the potential energy of a fluid particle in terms of its height from 61 reference datum. Hydrostatic head is referred to as the sum of elevation and pressure head. 62 Groundwater tables in unconfined aquifers represent the spatial gradients in hydrostatic head. A 63 number of processes either drive or enable HEF, several of which are based on changes in head 64 gradients. We follow the organizational structure presented by Käser et al. (2009), who divided 65 these processes into five distinct classes: 66

3 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

67 1. Transient exchange – the temporary movement of stream water into stream banks due to 68 short-term increases in stream stage (i.e., bank storage processes due to changes in 69 hydrostatic head gradients between stream and lateral riparian aquifer; Lewandowski et al. 70 2009; Sawyer et al. 2009a). 71 72 2. Turn-over exchange – the trapping of stream water in the streambed during times of 73 significant bed mobility (Elliot and Brooks, 1997b; Packman and Brooks 2001). 74 75 3. Turbulent diffusion – exchange driven by slip velocity that is created at the surface of the 76 porous medium of the bed where streamwise velocity vectors continue to propagate into the 77 surface layers of the bed (Packman and Bencala, 2000). 78 79 4. Hydrostatic-driven exchange – exchange driven by static hydraulic gradients which are 80 determined by changes in water surface elevation (Harvey and Bencala, 1993), spatial 81 heterogeneity in saturated hydraulic conductivity, or changes in the saturated cross-sectional 82 area of floodplain alluvium through which hyporheic flow occurs. 83 84 5. Hydrodynamic-driven exchange – exchange driven by the velocity head component of the 85 total head gradient on the bed surface (i.e., pumping exchange; Elliott and Brooks, 1997a,b) 86 and exchange induced by momentum gradients across beds and banks. 87 88 These classes of HEF processes are coupled to geomorphic processes in many ways. This is most 89 obvious for hydrostatic effects, which are directly dependent on channel and valley-floor 90 morphology and the depositional environment that controls spatial heterogeneity in saturated 91 hydraulic conductivity (K). However, turnover of streambed sediment is also related to fluvial 92 geomorphic processes. Similarly, hydrodynamic effects result from the interaction of flow over 93 stream bedforms. Geomorphic processes build stream bedforms and determine channel 94 morphology, especially longitudinal gradient, bed roughness, and water depth all of which 95 influence flow velocity. The relationship between geomorphology and the other classes of 96 processes is less direct, but still plays a role in controlling these processes through channel form 97 and the size distribution of sediment that makes up the streambed. This chapter focuses primarily

4 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

98 on the hydostatic and hydrodynamic processes where linkages to geomorphic processes are most 99 direct. 100 101 We organize our discussion of the interactions between geomorphology and HEF using a 102 hierarchical scaling framework developed for river networks (Frissell et al. 1986; Bisson and 103 Montgomery, 1996), starting at the whole network, through the stream segment, to the stream 104 reach, to the channel unit, and down to the sub-channel unit scale. We recognize that describing 105 any given process or related flow path at a single “scale” is somewhat arbitrary because of the 106 nested structure of the hyporheic flow net and dispersion among HEF flow paths. Despite that, 107 the concept of scale is an important heuristic tool to organize our understanding of hyporheic 108 processes. In many senses, the reach scale is the most informative scale at which to consider 109 HEF. A single reach, by definition, has characteristic channel morphology so that the factors 110 driving HEF within the reach are relatively consistent. However, only a few of the geomorphic 111 factors driving HEF actually operate at this scale. Most of the drivers work at the channel unit or 112 smaller scales. And to understand the importance of HEF in stream ecosystem processes, the 113 cumulative effects of HEF must be evaluated at scales much larger than a single reach. 114 115 9.13.2. The effect of geomorphology on hyporheic exchange flows 116 117 9.13.2.1. The whole network to segment scale 118 119 The geologic setting of the stream network is an important factor determining the likely 120 occurrence of HEF, but there have been few attempts to study HEF at this broad scale. Rather, 121 our expectations are pieced together by drawing comparisons among HZ studies that have been 122 conducted in widely varying geologic settings, at different locations in the stream network, or 123 under widely varying flow conditions. We expect that geomorphic-hyporheic relationships will 124 differ substantially among different geologic settings. 125 126 Fluvial geomorphic studies have examined the factors that determine the types of channel 127 morphologies present within stream networks (Montgomery and Buffington, 1997; Wohl and 128 Merritt, 2005; Brardinoni and Hassan, 2007). Montgomery and Buffington (1997) presented one

5 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

129 such description of the distribution of channel morphologies typical of many mountainous 130 landscapes. They showed that catchment area and channel longitudinal gradient controlled the 131 development of distinct channel types such that the channel types tended to follow a 132 characteristic sequence within a catchment (Figure 2A). In their example, this sequence starts 133 with bedrock and colluvial channels in the steepest, upper-most headwaters. As longitudinal 134 gradients decrease, channels change to cascades, to step-pool, to plane-bed, to pool-riffle, and the 135 largest, lowest gradient rivers were typified as dune-ripple channels. Along with these changes in 136 channel morphology, the following would be expected: decreased longitudinal gradient and 137 mean grain size of streambed sediment, and increased depth, width, hydraulic radius, and flow 138 velocity (Leopold and Maddock, 1953; Wohl and Merritt, 2008). 139 140 In this paper, we use Montgomery and Buffington’s (1997) description of the sequence of 141 channel types within a catchment as a simple heuristic model to organize our examination of the 142 relative importance of the different processes that drive HEF within stream networks. We 143 recognize that local controlling factors often interrupt simple sequencing of channel types. For 144 example, landslides may block large mainstem channels, creating locally steep gradients over the 145 landslide debris and uncharacteristically low gradients in the depositional reach immediately 146 upstream (Benda et al. 2003). We also recognize that regional differences in geology and 147 geomorphology will lead to dramatically different spatial organization of channel types (see for 148 example characteristic channel type in glaciated mountainous regions as described by Brardinoni 149 and Hassan, 1997). Our descriptions of the spatial organization of stream types and the resulting 150 HEF processes will have to be modified for any specific landscape. 151 152 Most hyporheic exchange results from head gradients pushing water through the streambed. The 153 amount of stream water entering the hyporheic zone is thus a function of the steepness of the 154 head gradient and the saturated hydraulic conductivity of the streambed and underlying aquifer. 155 The head gradients can be induced in many ways, but the two of primary influence are the 156 hydrostatic and hydrodynamic processes. The relative importance of each of these processes is 157 expected to vary among channel types and with longitudinal gradient. In high gradient streams, 158 channel forms such as step-pool sequences or pool-riffle sequences can create very steep 159 hydrostatic head gradients. Further, because of high bed roughness and relatively shallow water

6 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

160 depth, flow velocities tend to be lower in small steep streams than in larger, low gradient streams 161 (Leopold and Maddock, 1953; Wondzell et al. 2007). In contrast, it is difficult for natural 162 processes to create steep changes in the longitudinal gradient in low gradient streams. Instead, 163 stream flow interacts with stream bedforms, such as dunes or ripples, such that hydrodynamic 164 forces dominate the development of head gradients through the streambed. Thus, we expect that 165 hydrostatic effects will dominate in high gradient channels and that hydrodynamic processes will 166 dominate in low gradient channels (Figure 2B). Further, because channel types and longitudinal 167 gradients generally vary systematically within stream networks, we further expect that 168 hydrostatic effects will tend to dominate in the upper portions of stream networks and that the 169 relative importance of hydrodynamic processes will increase down the stream network. 170 171 9.13.2.2. The reach scale – setting the potential for hyporheic exchange 172 173 The potential for HEF to occur varies within any given stream reach. Roughly speaking, this 174 potential is determined by the factors that generate head differences that drive HEF, the 175 properties of the subsurface alluvium through which HEF occurs, and the potential effect of 176 lateral groundwater inputs from adjacent hillslopes that might limit hyporheic expression. 177 178 9.13.2.2.1. Losing and gaining reaches 179 180 Hyporheic exchange is likely to be more limited in strongly gaining reaches than in neutral 181 reaches because of steep streamward hydrologic gradients surrounding the channel (Wroblicky et 182 al. 1998; Storey et al. 2003; Malcolm et al. 2003 and 2005; Cardenas, 2009). Similarly, where 183 water is lost to regional aquifers in strongly losing reaches, return flows of stream water back to 184 the stream are likely to be severely restricted and thus also limit the expression of the hyporheic 185 zone (Cardenas, 2009). These patterns of gains and/or losses are controlled, at some level, by 186 regional groundwater and catchment characteristics interacting with smaller scale effects. In 187 large gaining rivers, Larkin and Sharp (1992) demonstrated that the relative dominance of cross- 188 valley vs. down valley flow paths through valley-floor aquifers varied depending on the 189 longitudinal gradient of the valley floor and the hydraulic conductivity of the valley floor 190 alluvium. In higher gradient reaches (>0.004 m/m) and in areas with coarser substrate, flow was

7 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

191 predominantly down valley. Conversely, where valley floor gradients were shallower or 192 sediment more finely textured, flow tended to be toward the stream. Thus, the way in which 193 lateral inputs influence hyporheic exchange is not solely a function of their magnitude, but also a 194 function of the ability of subsurface water to move down-valley (Storey et al. 2003). The ratio 195 between these two factors – the magnitude of the inputs relative to down valley flow – 196 determines how hyporheic exchange is affected. 197 198 As a first approximation, the potential for down valley flow can be estimated using the 199 relationships summarized in Darcy’s Law – that is, the product of the longitudinal valley 200 gradient, the saturated cross-sectional area of the floodplain perpendicular to the direction of 201 subsurface flow, and the hydraulic conductivity of the alluvium. As lateral inputs increase, 202 several factors may change: (1) water tables may rise, thus increasing the saturated thickness and 203 the cross-sectional area through which water flows allowing the transmission of more water, or 204 (2) flow paths may begin to turn obliquely toward the stream, which also increases the saturated 205 cross-sectional area and may also increase head gradients. Consequently, under dry conditions 206 when lateral inputs are relatively small, the potential extent and magnitude of hyporheic 207 exchange can be fully expressed (Figure 3A). As subsurface flows turn toward the channel they 208 begin to limit the extent of the hyporheic zone with only minor effect on the HEF (Wondzell and 209 Swanson, 1996). If sufficiently large, lateral inputs can severely limit both the spatial extent and 210 magnitude of hyporheic exchange (Figure 3B; Harvey and Bencala, 1993; Wroblicky et al. 1998; 211 Storey et al. 2003; Cardenas and Wilson, 2007; Malcolm et al. 2003, Soulsby et al. 2009). 212 213 Simple generalizations of where and when lateral inputs will limit HEF are difficult because of 214 the wide range of geomorphic settings in which HEF occurs and because the magnitude of lateral 215 inputs changes with catchment wetness. Lateral inputs are expected to be high when catchments 216 are wet and decrease as catchments dry out. However, lateral inputs are not spatially uniform. In 217 steep mountainous settings, the size of the upslope area draining directly to the valley floor is 218 important, concentrating lateral inputs in zones at the base of hillslope hollows (Jencso et al. 219 2009). Lateral inputs may persist the entire year at the bases of the largest hillslope hollows. 220 Most hillslope hollows are small, however, so that most of the stream network would be 221 disconnected from lateral inputs except for short periods of time when catchments are very wet,

8 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

222 for example after large storms or during peak snowmelt. We are unaware of similar studies 223 relating topography to spatial patterns of hillslope inputs in areas of low relief with humid 224 climates. However, Storey et al. (2003) reported that an extensive shallow surfical aquifer was 225 present along their lowland, low-gradient study reach and that lateral inputs of groundwater 226 substantially reduced both the extent and the amount of hyporheic exchange flows except during 227 summer baseflow. Clearly, the influence of lateral inputs may be much different in lowland 228 catchments than in steep mountainous catchments. 229 230 Changes in lateral inputs to streams do not occur in isolation. Rather, they are likely to be 231 accompanied by corresponding changes in stream stage (and discharge). The change in water 232 table elevations resulting from changed lateral inputs must be considered relative to the 233 accompanying changes in stream stage. Although the number of studies examining changes in 234 hyporheic flow paths with changing catchment wetness is limited, studies in small mountain 235 streams suggest that water table elevations in the floodplain increase more than stream stage so 236 that HEF is typically more restricted when catchments are wet (Figure 4A and 4B; Harvey and 237 Bencala, 1993; Wondzell and Swanson, 1996; Stednick and Fernald, 1999). Storey et al. (2003) 238 reported similar results for a lowland, low-gradient river. 239 240 In some cases, however, stream stage may change markedly without corresponding changes in 241 precipitation recharge or changes in lateral inputs. Most examples of these processes come from 242 large, lowland rivers because river stage is controlled by processes far upstream. These “bank 243 storage” processes (Pinder and Sauer, 1971) have been recognized as a form of transient 244 hyporheic exchange (Figure 4C and 4D) that can result from both in-bank or over-bank floods 245 (Bates et al. 2000; Burt et al. 2002). In some situations, increased stream stage may even lead to 246 groundwater ridging in the floodplain, reversing head gradients and limiting lateral groundwater 247 inputs. Similarly, hyporheic exchange through stream banks can result from diel variations in 248 stream stage (and discharge) during snow melt periods (Loheide and Lundquist, 2009) or from 249 tidally induced changes in water elevations in coastal streams and rivers (Bianchin et al. in 250 press). 251

9 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

252 Transient hyporheic exchange may be especially evident in regulated rivers where releases from 253 dams (or other control structures) can result in large and rapid changes in river stage without 254 corresponding local precipitation to recharge floodplain aquifers (e.g., Fritz and Arntzen, 2007; 255 Lewandowski et al. 2009; Sawyer et al. 2009a; Francis et al., in press). However, transient 256 hyporheic exchange may not always result from fluctuations in river stage. For example, 257 Hanrahan (2008) studied vertical HEF through the streambed of a large, regulated gravel bed 258 river where stage sometimes changed by nearly 2 m in an hour. For the most part, they did not 259 observe transient hyporheic exchange related to changes in stage. They concluded that 260 hydrostatic and hydrodynamic processes remained the dominant control on HEF. Notably, 261 Hanrahan (2008) did not examine lateral exchanges through the stream banks, which can be 262 more responsive to changes in stage than are locations in the stream channel itself (Storey et al. 263 2003). Water table fluctuations in the floodplain at long distances from the stream are not 264 necessarily indicative of extensive HEF because pressure fluctuations can propagate through 265 surficial (unconfined) aquifers much faster than does the actual flow of stream water. This was 266 clearly demonstrated by Lewandowski et al. (2009) who showed that river water penetrated, at 267 most, only 4 m into the stream bank even though water table fluctuations were observed more 268 than 300 m from the river. 269 270 HEF can occur in strongly gaining and losing reaches because of the nested structure of 271 hyporheic flow paths, and because HEF can occur at a variety of spatial scales. Thus an envelope 272 of the HZ can be set within larger non-hyporheic flow paths (Figure 3B; Cardenas and Wilson, 273 2007). Similarly, smaller-scale HEF can occur as a result of smaller scale geomorphic drivers, 274 even within a reach that is, overall, strongly losing (Payn et al. 2009). Further, because HEF is 275 dominated by relatively near-stream flow paths that are short in length and residence time 276 (Kasahara and Wondzell, 2003), the magnitude of HEF can be substantial, even in strongly 277 gaining reaches where the spatial extent of the hyporheic zone is greatly restricted (Wondzell and 278 Swanson, 1996; Cardenas and Wilson, 2007; Payn et al. 2009). 279 280 9.13.2.2.2. Changes in saturated cross-sectional area 281

10 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

282 The saturated cross sectional area of the floodplain (orthogonal to groundwater flow path 283 direction) is one of the factors determining the amount of groundwater transmitted down valley 284 through the valley floor alluvium. Thus, any change in the cross sectional area along the length 285 of a stream reach will lead to parallel changes in the down valley flow of water through the 286 floodplain, thereby driving downwelling from, or upwelling to the stream (Stanford and Ward, 287 1993). Downwelling occurs where valley floors increase in width, for example, downstream of 288 bedrock-constrained reaches (Figure 5A; Poole et al. 2004 and 2006; Acuna and Tockner, 2009). 289 Conversely, upwelling occurs where valley floors narrow at the lower end of wide unconstrained 290 reaches (Figure 5A; Baxter and Hauer, 2000; Acuna and Tockner, 2009). Similarly, variations in 291 the thickness of the surficial aquifer, caused by variations in depth to bedrock or other confining 292 layers drive similar patterns of upwelling and downwelling. For example, upwelling commonly 293 occurs just upstream of bedrock sills with a subsequent transition to downwelling just 294 downstream of such bedrock sills as the surficial aquifer again thickens (Figure 5B; Valett, 295 1993). This is easily observed in streams in arid regions during the dry season, where perennial 296 flow may only occur above bedrock sills, which force the subsurface flow to the surface. 297 298 9.13.2.3. The sub-reach to channel-unit scale – hydrostatic processes 299 300 Geomorphic features of the stream channel and valley floor within stream reaches control the 301 elevation of surface water and can thereby create significant head gradients through the valley 302 floor alluvium, driving HEF. Because these geomorphic features are static on the time scales 303 typical of hyporheic exchange (hours to weeks) they are broadly recognized as “hydrostatic 304 processes”. 305 306 9.13.2.3.1. Step-pool and pool-riffle sequences 307 308 One of the best-studied examples of hydrostatic processes involves the changes in water surface 309 elevation along a pool-step sequence and the resulting head gradients that drive HEF (Figure 1; 310 Harvey and Bencala, 1993). Harvey and Bencala (1993) showed that the change in the 311 longitudinal gradient of the stream channel (which approximates the stream energy profile) drove 312 HEF. They also observed that HEF flow paths tended to be curved – first curving away from the

11 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

313 stream above the step or riffle and then curving back to the stream below the step or riffle. 314 Building from their observations, model analyses show that along an idealized straight channel 315 with homogeneous isotropic porous sediment, hyporheic flow paths around a change in the 316 longitudinal gradient will exploit the full 3-dimensional saturated volume along the channel, thus 317 extending both vertically beneath the streambed and horizontally through the streambanks and 318 near stream aquifer (Figure 1A and 1B). Real streams are substantially more complicated, 319 however, such that changes in hydraulic conductivity of the alluvium, bends in the channel, and 320 the spatial location of lateral groundwater inputs lead to the development of a complicated flow 321 net through the valley floor (e.g., Cardenas and Zlotnik, 2003). Despite these complexities, the 322 steepness of the hydraulic head gradient imposed by the change in the longitudinal gradient and 323 the saturated hydraulic conductivity control the amount of stream water exchanged with the 324 subsurface. 325 326 Many factors can modify the effect of steps or riffles on HEF. For example, the height of the step 327 (or steepness of the riffle) determines the head gradient available to drive HEF so that a single 328 very large step has the potential to drive more HEF than if the same amount of elevational 329 change is spread over several smaller steps (Kasahara, 2000). Because of this, large wood can be 330 important in determining the amount of HEF in forest streams. Single logs tend to create 331 frequent, small obstructions that collect and store small amounts of sediment, forming pool-step 332 sequences in which the extent of the hyporheic zone tends to be small (Wondzell, 2006). 333 Although log jams are less common, they can create large obstructions storing sediment in 334 wedges several meters deep and 10 or more meters in length, and significantly widen constrained 335 stream channels. Consequently, log jams can form extensive hyporheic zones in steep, confined 336 mountain streams (Wondzell, 2006). 337 338 Large, channel-spanning logs can wedge into steep narrow channels, forcing the accumulation of 339 sediment in channels, converting bedrock reaches to alluvial reaches with a step-pool 340 morphology (Montgomery et al. 1996), thereby greatly enhancing HEF. Similarly, large wood 341 can force plane-bed channels into a pool-riffle morphology (Montgomery et al. 1996) which 342 should lead to more HEF than would be present in a comparable wood-free channel. Large wood 343 can have the opposite effect in channels that would have a free-formed pool-riffle morphology.

12 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

344 In one documented case, accumulations of large wood tended to force a pool-riffle channel 345 toward a step-pool morphology (Wondzell et al. 2009). The channel adjusted to removal of all 346 large, in-stream wood by developing a better defined pool-riffle structure around meander bends, 347 leading to increased sediment storage. Continued channel adjustment over time following the 348 removal of large wood eventually led to substantial increases in HEF. 349 350 The size, spacing, and sequence of channel units (e.g., pools and riffles) along the stream 351 longitudinal profile can also affect HEF (Anderson et al. 2005; Gooseff et al. 2006). Anderson et 352 al. (2005) made detailed measurements of channel profiles and patterns of HEF, and showed that 353 channel unit size and spacing increased as did the length of channel characterized by 354 downwelling with increasing drainage area in a mountainous stream catchment. Gooseff et al. 355 (2006) built on these results, examining HEF using 2-D groundwater models of idealized 356 longitudinal profiles of mountain streams. Gooseff et al.’s (2006) modeling results confirmed 357 that both channel unit spacing and size were important in determining hyporheic exchange 358 patterns of upwelling and downwelling. Perhaps more surprising, however, was the observation 359 that the sequence of channel units also affected simulated HEF. Gooseff et al. (2006) compared 360 pairs of idealized stream reaches that varied only by the way the longitudinal gradient changed 361 over the pool-riffle sequence – i.e., the slope of the riffle was gradual on its upstream end and 362 steepest at its downstream end (described as a pool–riffle–step sequence) versus riffles that were 363 initially steep with the slope decreasing toward the downstream end (described as a pool–step– 364 riffle sequences). Simulated downwelling lengths were substantially longer for pool–riffle–step 365 sequences than for pool–step–riffle sequences. 366 367 9.13.2.3.2. Meander bends and point bars 368 369 A variety of channel and valley floor morphologic features, in addition to changes in the 370 longitudinal gradient, create head gradients with the potential to drive HEF. These include 371 channel meander bends and associated point bars, back channels or floodplain spring brooks, and 372 islands set between main and secondary channels. In all these cases, differences in the 373 elevational head of surface water between two channels, between different points in a single 374 channel around a meander bend, or between points on opposite sides of an island create head

13 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

375 gradients that drive HEF. For example, head gradients through the point bar in a meander bend 376 are steeper than the longitudinal gradient of the stream channel around the point bar (Peterson 377 and Sickbert, 2006) so that stream water infiltrates the upper end of the point bar and is returned 378 to the channel at the lower end of the point bar (Figure 6A; Vervier and Naiman, 1992). More 379 generally, these exchange flows occur across the full length of meander bends and are influenced 380 by both the change in stream water elevation around the meander bend and the plan-view shape 381 of the meander bend. Highly evolved meander bends support steep head gradients across the 382 mender neck because of the close proximity of the stream channels (Figure 6B; Boano et al. 383 2006; Revelli et al. 2008) so that HEF is dominantly located in the meander neck, with much 384 reduced HEF across the remainder of the meander where head gradients are much lower. In other 385 cases, meanders develop a characteristic pattern of alternating pools and riffles, with riffles 386 located at the thalweg cross-overs in the inflections between adjacent meanders and pools or low 387 gradient runs wrapping around the point bar (Figure 6C). This combination of channel 388 morphologic features can create complex HEF flow paths within meander bends. The residence 389 times of HEF traversing meander bends can be quite short where meanders are small and 390 saturated hydraulic conductivities are high (Pinay et al. 2009). Conversely, residence times of 391 HEF may be extremely long in meander bends of low gradient rivers with fine textured sediment 392 (Boano et al. 2006; Peterson and Sickbert, 2006). 393 394 9.13.2.3.3. Back channels and floodplain spring brooks 395 396 Channel planforms are often complex in wide floodplains, including a network of old or 397 abandoned channels. If the upstream ends of these channels are plugged with sediment and if the 398 downstream ends are sufficiently incised to intercept the water table and are connected back to 399 the river at their downstream ends, they will act as drains, imposing head gradients from the 400 stream to the old channel (Figure 7A; Wondzell and Swanson, 1996; Poole et al. 2006). These 401 channels are also known as floodplain spring brooks because water upwells into the channel, 402 forming a spring at its head. In addition to creating HEF, these channels will capture whatever 403 water is in the surficial aquifer of the floodplain, including down valley flows from upstream 404 locations, and lateral inputs of groundwater or hillslope water from the valley margin. However, 405 because lateral inputs tend to be small and spatially isolated (Jencso et al. 2009; and as discussed

14 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

406 above), floodplain spring brooks will most often be fed by HEF (Wondzell and Swanson, 1996; 407 Jones et al. 2007). 408 409 Abandoned channels can also be plugged at their downstream ends and open to the river at their 410 upstream ends. In this case, stream water can flow into the abandoned channel, infiltrate the 411 channel bed and raise the water table in the middle of the floodplain, thereby creating head 412 gradients and driving HEF from the abandoned channel back to the main stream channel (Figure 413 7B). More complex situations arise when the longitudinal gradients in either the back channel or 414 mainstem channel are interrupted by steeper riffles or steps. Figure 7C shows the interactions 415 between a back channel and riffle. Above the riffle, water in the main channel is higher than the 416 back channel so water flows towards the spring brook. Downstream of the riffle, the main 417 channel is lower than the back channel so that the back channel loses water over its downstream 418 extent, eventually going dry before reaching the main channel. 419 420 The channel planform features that drive HEF can occur over a range of spatial scales, and their 421 influence may change through time as the stage height of water in the main channel changes. For 422 example, a small gravel bar may have low points along the stream bank. At high stage, the entire 423 gravel bar may be submerged. As stage decreases the center of the bar may become exposed, 424 creating a secondary channel along the bank. As stage decreases further, flow may become 425 discontinuous through the secondary channel such that it functions as a drain if it is plugged at 426 the upstream end, or functions as a conduit allowing stream water to infiltrate the surface of the 427 gravel bar if it is plugged at its downstream end. Old channels in large floodplains may act 428 similarly, with continuous flow along their full length during floods, but becoming disconnected 429 at intermediate to low stage, or even dry completely during periods of minimum discharge. In 430 large floodplain reaches, these channels can be 100’s of meters to kilometers in length, extending 431 nearly the full length of the stream reach (Poole et al. 2006; Arrigoni et al. 2008). 432 433 9.13.2.3.4. Secondary channels and islands 434 435 Islands present a special case of back channels in which the channel is continuously connected to 436 the main channel over its full length. Hyporheic hydrology of islands has not been extensively

15 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

437 studied. However, we expect that the surface water elevations in channels bounding the island 438 create boundary conditions for total head and control HEF through islands as is generally 439 indicated by the available literature (Dent et al. 2007; Francis et al., in press). If channels along 440 both sides of the island are parallel and symmetric with constant longitudinal gradient, then flow 441 through the island will parallel the channels and the head gradient driving flow will equal the 442 overall longitudinal gradient of the stream reach (Figure 8A). If riffles are present in the 443 channels, the head gradient through the island adjacent to the riffles can be much steeper than the 444 reach averaged longitudinal gradient (Figure 8B). Also, if riffles are displaced along the primary 445 and secondary channels surrounding an elongated island such that a riffle is located near the head 446 of the island in one channel and near the tail of the island in the second channel, the resulting 447 head gradients would tend to drive flows laterally through the island, leading to very large cross- 448 sectional areas experiencing HEF, and therefore large amounts of HEF, albeit, with shorter 449 length flow paths (Figure 8C). While islands may be uncommon in most channel types, they may 450 dominate HEF in braided and anastomosing stream reaches (Ward et al. 1999; Arscott et al. 451 2001). Given the complexities of potential sizes and shapes of islands and patterns in 452 longitudinal gradients in the bounding channels, the resulting flow nets, residence times, and 453 amounts of HEF are likely to vary widely. 454 455 9.13.2.3.5. Spatial heterogeneity in saturated hydraulic conductivity 456 457 Fluvial processes control the depositional environment on the streambed and across the 458 floodplain creating spatial heterogeneity in the texture of deposited and reworked sediment 459 across a range of scales, from the surface of the streambed to the entire floodplain. Because 460 sediment texture is closely related to saturated hydraulic conductivity (K), these processes can 461 substantially influence HEF. However, because of the difficulties in quantifying these patterns at 462 the scales at which they influence HEF, they have been relatively little studied. At fine scales, 463 streambed roughness can control the depositional environment across the streambed (Buffington 464 and Montgomery, 1999), which lead to spatial patterns in the distribution of K within the 465 streambed (Genereaux et al. 2008), which in turn can influence both the location and amount of 466 HEF. HEF will be restricted where the streambed is clogged with fine sediment and 467 preferentially located in zones with higher K. Experiments in flumes have also shown that HEF

16 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

468 can also influence patterns of fine-sediment deposition, with fine sediment preferentially 469 deposited in downwelling zones (Packman and MacKay, 2003; Rehg et al. 2005) which may 470 explain differences in K between upwelling and downwelling zones observed in a steep 471 headwater stream (Scordo and Moore, 2009). 472 473 Spatially heterogeneous patterns in K influence HEF. For example, groundwater flow modeling 474 studies using homogeneous vs. heterogeneous K showed that spatial heterogeneity may add 475 substantial complexity to the spatial patterns of the hyporheic flow net (Woessner 2000). When 476 relatively high K regions are aligned parallel with head gradients they create preferential flow 477 pathways (Wagner and Bretschko, 2002) that can increase the total amount of HEF (Cardenas 478 and Zlotnik, 2003; Cardenas et al. 2004). Results from Cardenas et al. (2004) showed that 479 influence of heterogeneity in K was relatively greater in lower gradient streams and where head 480 gradients driving HEF were reduced. To our knowledge, the influence of fine-grained 481 heterogeneity has not been studied in steeper channels where hydrostatic processes dominate. 482 483 Fluvial processes also influence spatial patterns in K at the scale of the entire floodplain. 484 Especially important is the layering of stream and floodplain alluvium. Layering can create 485 strong vertical anisotropy (Chen, 2004), limiting vertical exchange and promoting lateral flows 486 through the streambed and floodplain (Packman et al. 2006; Marion et al. 2008). Overbank 487 deposition can also bury back channels creating “paleochannels” where coarse streambed 488 alluvium is buried under finer floodplain soils (Stanford and Ward, 1993; Stanford et al. 1994; 489 Poole et al. 2004). If these paleochannels intercept the water table, they will function as large 490 preferential-flow pathways that can route water the full length of a floodplain. In this regard they 491 function much like a subsurface version of back channels or floodplain spring brooks – either 492 acting as drains lowering the water table in the floodplain and imposing head gradients from the 493 stream to the paleochannel, or acting as distributaries, routing water into the floodplain and 494 imposing head gradients from the paleochannel to the stream. Locations of paleochannels are 495 sometimes evident from shallow depressions along the floodplain. In other cases, over-bank 496 deposition will have completely filled old channels so that there is no surficial indication on the 497 flat floodplain surface. The influence of paleochannels is difficult to discern because networks of

17 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

498 widely spaced wells are unlikely to find and trace the location of these features along the length 499 of the floodplain. As a consequence, their influence on HEF has not been widely studied. 500 501 9.13.2.4. The bedform scale – hydrodynamic processes 502 503 Channel hydraulics, and the spatial and temporal distribution of velocity (kinetic energy) across 504 streambeds are significantly influenced by the form of the channel and the bedforms that occur in 505 channels. The continuous feedback between pressure distribution and shear stress across the bed 506 surface and the potential to erode the bed will cause turn-over exchange to occur during times of 507 high flows. During lower flows, when bed sediment is relatively stable, bedforms cause some 508 level of form drag on the flows, inducing pressure distributions across the bedforms, thereby 509 driving HEF at a scale smaller than the bedform (Figure 9). The size of the bedform is set by 510 both the energy regime of the reach and the material that makes up the reach, and the form drag 511 induced on the water column by the bedform is of course partly controlled by its size. Thus, the 512 scale of HEF flowpaths induced by hydrodynamic exchange across the bedforms will scale in 513 part with the size of bedforms present (Cardenas et al. 2004). Finally, the heterogeneity of the 514 bed material that makes up the bedforms will have a distinct control on the flux rate and actual 515 flowpaths through and around the bedforms (Sawyer et al. 2009b). 516 517 In sand bed streams, hydrodynamic HEF has been extensively studied both theoretically and 518 empirically. Typical bed forms in sand bed streams are dunes and ripples, which have a fairly 519 predictable geometry and spacing, based on bed sediment composition and flow rate. 520 Thibodeaux and Boyle (1987) pioneered investigations of the hydrodynamic pressure 521 distribution across dunes, noting the penetration of channel water into the porous bed forms. 522 Further development of a ‘pumping exchange’ model by Elliot and Brooks (1997a,b) expanded 523 the ability to predict HEF and associated solute dynamics in channel-bed systems. Whereas most 524 studies of hydrodynamic exchange processes were generally carried out in or applied to flume 525 studies, there has been at least one application of incorporating the pumping exchange model to 526 tracer transport in field studies. Salehin et al. (2003) studied the transport of tracer along several 527 km of Sava Brook in Sweden and successfully applied a solute transport model to the observed 528 data to explain long time residence time distributions using the pumping exchange model theory.

18 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

529 The predictability of dune and ripple sizing and spacing makes the pumping exchange model a 530 useful tool to explore HEF in sand bed streams and rivers. 531 532 In gravel bed streams, bed form types may be generally predictable (i.e., Montgomery and 533 Buffington, 1997; Wohl and Merritt, 2008; Chin, 2002), but the exact geometry and spacing of 534 bed forms is less predictable, particularly at a scale that will directly influence head distributions 535 across and along the channel. Hence, the velocity distribution in the channel and around the bed 536 form, which contributes to hydrodynamic exchange, is also unpredictable. Tonina and 537 Buffington (2007) conducted careful studies of total pressure distribution across streambeds in 538 flumes that had ‘realistic’ geometry of a pool-riffle sequence in a gravel bed channel. Their 539 results indicated that total head distribution (i.e., incorporating velocity head in addition to 540 hydrostatic head) was important to exchange at focused points in the channel where high velocity 541 occurred. Further, they confirmed that in general, there was little or no contribution of velocity 542 head to parts of the bed that were overlain by deeper, slower flow, and therefore a hydrostatic 543 representation of exchange will likely be more applicable in these locations. 544 545 Regardless of the predictability of bed form geometry and spacing, the associated hydrodynamic 546 HEF may induce only limited lengths of exchange in the subsurface because much of the 547 exchange dynamics are expected to be vertical rather than lateral. Exchange lateral to the channel 548 is more likely to be driven by hydrostatic gradients set up across meander bends or bars (as 549 described above). Hydrodynamic HEF will contribute to, but be only one component of, total 550 HEF in natural channels, and its importance will be dictated by both channel hydraulics and, if 551 present, competing hydrostatic factors that can create steeper head gradients. 552 553 9.13.2.4. The particle scale – turbulent diffusion 554 555 At the particle scale on streambeds, turbulent diffusion is significantly influenced by the size and 556 arrangement of surface sediment. Because turbulent diffusion is induced by the momentum 557 transfer between the water column and the porous media, HEF due to turbulent diffusion is a 558 function of the decreasing velocity profile within the surface layers of the porous media (Shimizu 559 et al. 1990). Thus, the distribution of sediment at the surface will greatly influence the potential

19 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

560 for energy and mass transfer within this zone. Turbulent diffusion HEF is prominent in gravel 561 bed streams where surface pores are more likely to accommodate such open exchanges of 562 momentum across the bed (Tonina and Buffington, 2009). Beds composed of sand particle sizes 563 and smaller provide too much resistance to the momentum exchange between the water column 564 and the bed. Hence, turbulent diffusion is more likely to be an important component of HEF in 565 low order, high-gradient streams (Figure 2B). Careful theoretical and empirical research on 566 turbulent diffusion has been conducted largely on planar beds (Shimizu et al. 1990; Habel et al. 567 2002). Therefore, in the complex bed topography of typical gravel channels, turbulent diffusion 568 will be a component of HEF, likely not the singular driver of HEF. 569 570 9.13.3. Discussion 571 572 9.13.3.1 Multiple features acting in concert 573 574 In the examples presented above (Figures 1, 3–9), we have mostly focused on single types of 575 channel morphologic features that drive or enable hydrostatic and hydrodynamic HEF. However, 576 these features never occur in isolation. Rather, a single stream reach will typically contain many 577 of the morphologic features described above. Interactions among these features are likely to be 578 important in determining the actual HEF in any given stream reach. In some cases, the effects of 579 multiple features could be additive and result in higher HEF than if they did not co-occur. For 580 example, cross-valley flow paths between main channels and floodplain spring brooks can be 581 accentuated by riffles (Figure 7C). However, interactive effects could also cancel, for example 582 where riffles at the inflection points of meander bends reduce head gradients through point bars 583 (Figure 6C). The interactions between different processes driving or enabling HEF is complex, 584 and to some degree, site specific, making it difficult to quantify the effects of these interactions. 585 Because of these difficulties, there are relatively few comparative studies that have examined 586 multiple processes concurrently, within natural stream channels and attempted to evaluate the net 587 effect of each process on the total HEF within stream reaches. 588 589 Sensitivity analyses with groundwater flow models calibrated to simulate HEF in a studied 590 stream reach provide one opportunity to examine the relative importance of channel morphologic

20 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

591 features on HEF where multiple features are present in a single reach. For example, Kasahara 592 and Wondzell (2003) examined a number of channel morphologic features among stream reaches 593 of different sizes in a mountainous stream network under conditions of summer baseflow 594 discharge. In all cases, the single strongest driver determining the amount of HEF occurring 595 within the simulated stream reaches was the change in longitudinal gradient over step-pool 596 sequences in the 2nd-order channel (Figure 10A) and pool-riffle sequences in the 5th-order 597 channel. The shape of the hyporheic flow net in the 5th-order stream, however, was strongly 598 controlled by the presence, location, and relative elevation difference between water in the main 599 channel and the back channels (Figure 10B). Similarly, Cardenas et al. (2004) examined 600 sediment heterogeneity, size of bedforms, and both longitudinal and lateral head gradients in a 601 low gradient, sand bed stream. They found that HEF was greater where beforms had higher 602 amplitude and were more closely spaced. Spatial heterogeniety in K increased HEF relative to 603 homogeneous simulations, as did inclusion of lateral head gradients, but the effect was small 604 relative to the effect of the size and spacing of bedforms. 605 606 Channel morphologic features can interact with changes in steam stage and lateral groundwater 607 inputs in ways that can substantially influence the amount of HEF over time, across seasons or 608 within a single storm event. Storey et al. (2003) examined HEF in a pool-riffle sequence at both 609 high- and low-baseflow discharge. At high stage, the stream tended to “drown” the riffle, 610 substantially reducing the change in the longitudinal gradient over the pool-riffle sequence and 611 thus reducing HEF. In contrast, at low stage, the water surface more closely followed the 612 streambed topography, thus creating steeper head gradients that supported more HEF. Storey et 613 al. (2003) also showed that lateral inputs during the wet season were sufficient to eliminate most 614 of the HEF through the riffle. Cardenas and Wilson (2006) showed that low rates of groundwater 615 discharge limited the extent of the HZ formed by the hydrodynamics of stream bedforms, and 616 that high rates of groundwater discharge could completely eliminate HEF. 617 618 We know of only one study comparing the relative influence of hydrostatic and hydrodynamic 619 effects. In a flume, Tonina and Buffington (2007) investigated the control of total head (i.e., 620 including dynamic head) in driving hyporheic exchange. Their results suggested that there are

21 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

621 specific locations within channels where the velocity head can provide additional potential and 622 thereby influence the pattern of hyporheic exchange. 623 624 9.13.3.2 Change in processes driving HEF through the stream network 625 626 Hyporheic exchange will vary widely across the sequence of channel types found in stream 627 networks (Figure 2; Buffington and Tonina, 2009). Channel networks generally follow a pattern 628 of steep headwaters to low-gradient reaches downstream. In mountain stream networks in 629 particular, gradient changes are expected to be accompanied by channel morphology changes 630 resulting in a sequence of distinct channel morphologies (Figure 2A). Obviously, bedrock 631 reaches have negligible hyporheic zones (Gooseff et al. 2005; Wondzell, 2006). We are unaware 632 of any studies of HEF in colluvial and cascade channel morphologies, however the extremely 633 high longitudinal gradients of these channels likely result in high velocity underflow which has 634 been shown to restrict the extent of the hyporheic zone (Storey et al. 2003). Also, the relatively 635 disorganized structure of the bed sediment prevents development of stepped water surface 636 profiles so that hydrostatically driven exchange due to longitudinal changes in gradient will 637 likely be low. Turbulent diffusion is likely to be a primary driver of HEF (Figure 2B). 638 639 Free-formed step-pool channels occur at slightly lower gradients (Figure 2A). These channels 640 have well-organized structure with periodic spacing of both steps and pools (Chin, 2002; Wohl 641 and Merrit, 2008) that have been shown to be primary drivers of HEF (Figure 2B; Kasahara and 642 Wondzell, 2003). The addition of large wood can substantially increase sediment storage 643 (Nakamura and Swanson, 1993; Montgomery et al. 1996), the development of step-pool 644 structure, and the extent, amount and residence times of HEF in these stream reaches (Wondzell, 645 2006). Other hydrostatic factors tend to have less dominance on HEF; these reaches have low 646 sinuosity so meander bends are uncommon and steep longitudinal gradients limit the potential 647 for back channels to create lateral HEF flow paths. 648 649 We are unaware of any published studies examining HEF in plane-bed channels. However, we 650 expect HEF to be lower than in either step-pool or pool-riffle channels (Figure 2B). The 651 streambed tends to be smoothly graded in these channels as suggested by their name, and there is

22 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

652 low spatial heterogeneity in surface texture (Buffington and Montgomery, 1999). Pools are 653 widely spaced, and both steps and riffles are rare. While these channels occur as free-formed 654 morphologies, pool-riffle channels can be converted to plane-bed channels by land use practices 655 that increase sediment supply and through the direct removal of large wood, with concurrent 656 decreases in HEF. 657 658 Lower in the stream network, channels tend have lower longitudinal gradients (Figure 2A), and 659 even in mountainous areas, unconstrained stream reaches become increasingly common. Channel 660 planforms can be quite complex in these rivers and as a consequence, a wide array of channel 661 geomorphic features influences HEF. Braided and anastomosing channels may form where 662 sediment loads are high and stream banks are erodible; the complex of channels likely leads to 663 substantial HEF through islands. Meandering channels form under lower sediment loads and 664 where banks are more stable. Meandering channels typically have pool-riffle morphologies, 665 although complexes of secondary channels, back channels, and paleochannels are common, a 666 legacy of past floods, channel avulsions and overbank deposition. Because most HEF occurs 667 along short, near stream flow paths, riffles are the dominant feature determining the amount of 668 HEF (Kasahara and Wondzell, 2003). However, the shape of the hyporheic flow net and the 669 residence time distribution of HEF will be strongly influenced by the complex of channel 670 planforms. Finally, hydrodynamic processes are expected to dominate in streams with relatively 671 mobile streambeds characterized by dune-ripple bedforms. These streams have low longitudinal 672 gradients and therefore channel morphologic features tend not to create steep hydrostatic head 673 gradients (Figure 2B). 674 675 Other exchange processes are likely to be related to specific conditions. Turn-over exchange will 676 only occur when bed material is mobile – a characteristic feature of both anastomosing and dune- 677 ripple channels. Transient exchange will only be appreciable during wet catchment conditions, 678 when channel stage is high and surrounding groundwater tables are comparatively low. 679 However, transient exchange may be a dominant form of HEF in regulated rivers where stage 680 fluctuates over daily cycles due to hydroelectric generation. Turbulent diffusion, on the other 681 hand, is likely to occur in gravel bed sections of the network, likely with the greatest potential

23 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

682 influence in either cascade or plane-bed sections of stream networks where stream velocities are 683 expected to be high. 684 685 9.13.4. Conclusion 686 687 Hyporheic exchange results from distinct processes, and the relations between those processes 688 and geomorphology are well understood from a mechanistic perspective. Thus, geomorphology 689 provides a critical framework to understand hyporheic processes and how they change with 690 location within a stream network, and over time in response to changes in stream discharge and 691 catchment wetness. To the degree that these geomorphic patterns are predictable, they provide 692 the foundation for hydrologists to make general predictions of the relative importance of the 693 hyporheic zone at the scale of entire catchments. Reach to reach variability is high in stream 694 networks, however, so understanding HEF at the reach scale continues to require detailed study 695 of specific stream reaches. These studies are difficult and current methodological approaches are 696 insufficient to fully examine the full suite of processes that account for patterns of HEF in any 697 specific stream reach. Consequently, hyporheic studies tend to focus on a single factor, or at 698 most a small subset of the factors driving HEF. Hyporheic researchers recognize that such 699 studies are incomplete. Detailed, holistic understanding of the importance of different processes 700 in driving HEF, how the relative importance of these processes changes with location in the 701 stream network, with the specific structure of any given stream reach, and with changes in 702 discharge and lateral groundwater inputs remains elusive.

24 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

703 References 704 705 Acuna, V. and Tockner, K. 2009. Surface-subsurface water exchange rates along alluvial river 706 reaches control the thermal patterns in an Alpine river network. Freshwater Biology 54:306-320. 707 708 Anderson, J. K., Wondzell, S. M., Gooseff, M. N., and Haggerty, R. 2005. Patterns in stream 709 longitudinal profiles and implications for hyporheic exchange flow at the H.J. Andrews 710 Experimental Forest, Oregon, USA. Hydrological Processes 19:2931-2949. 711 712 Arrigoni, A. S., Poole, G. C., Mertes, L. A. K., O’Daniel, S. J., Woessner, W. W., and Thomas, 713 S. A. 2008. Buffered, lagged, or cooled? Disentangling hyporheic influences on temperature 714 cycles in stream channels. Water Resources Research 44, W09418, 715 doi:10.1029/2007WR006480 716 717 Arscott, D. B., Tockner, K., and Ward, J. V. 2001. Thermal heterogeneity along a braided 718 floodplain river (Tagliamento River, Northeastern Italy). Canadian Journal of Fisheries and 719 Aquatic Sciences 58:2359-2373. 720 721 Bates, P. D., Stewart, M. D., Desitter, A., Anderson, M. G., Renaud, J. P., and Smith, J. A. 2000. 722 Numerical simulation of floodplain hydrology. Water Resources Research 36:2517-2529. 723 724 Baxter, C. V. and Hauer, F. R. 2000. Geomorphology, hyporheic exchange, and selection of 725 spawning habitat by bull trout (Salvelinus confluentus). Canadian Journal of Fisheries and 726 Aquatic Sciences 57:1470-1481 727 728 Bencala, K. E. and Walters, R. A. 1983. Simulation of solute transport in a mountain pool-and- 729 riffle stream: A transient storage model. Water Resources Research 19:718-724 730 731 Benda, L., Miller, D., Bigelow, P., and Andras, K. 2003. Effects of post-wildfire erosion on 732 channel environments, Boise River, Idaho. Forest Ecology and Management 178:105-119. 733

25 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

734 Bianchin, M., L. Smith, and R. Beckie. in press. Quantifying hyporheic exchange in a tidal river 735 using temperature time series. Water Resour. Res., doi:10.1029/2009WR008365. 736 737 Bisson, P. A. and Montgomery, D. R. 1996. Chapter 2: Valley segments, stream reaches, and 738 channel units. Pages 23-52. In: R. R. Hauer and G. A. Lamberti (eds). Methods in Stream 739 Ecology. Academic Press, New York, NY. 740 741 Boano, F., Camporeale, C., Revelli, R., and Ridolfi, L. 2006. Sinuosity-driven hyporheic 742 exchange in meandering rivers. Geophysical Research Letters 33, L18406, 743 doi:10.1029/2006GL027630 744 745 Brardinoni, F. and M.A. Hassan, 2007. Glacially-induced organization of channel-reach 746 morphology in mountain streams. Journal of Geophysical Research, 112, F03013, 747 doi:10.1029/2006JF000741. 748 749 Buffington, J. M. and Montgomery, D. R. 1999. Effects of hydraulic roughness on surface 750 textures of gravel-bed rivers. Water Resources Research 35:3507-3521. 751 752 Buffington, J. M. and Tonina, D. 2009. Hyporheic exchange in mountain rivers II: effects of 753 channel morphology on mechanics, scales, and rates of exchange. Geography Compass 3:1038- 754 1062. 755 756 Burt, T. P., Bates, P. D., Stewart, M. D., Claxton, A. J., Anderson, M. G., and Price, D. A. 2002. 757 Water table fluctuaions within the floodplain of the River Severn, England. Journal of Hydrology 758 262:1-20. 759 760 Cardenas, M. B. 2009. Stream-aquifer interactions and hyporheic exchange in gaining and losing 761 sinuous streams. Water Resources Research 45, W06429, doi:10.1029/2008WR007651 762 763 Cardenas, M. B. and Wilson, J. L. 2006. The influence of ambient groundwater discharge on 764 exchange zones induced by current-bedform interactions. Journal of Hydrology 331:103-109.

26 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

765 766 Cardenas, M. B., and Wilson, J. L. 2007. Exchange across a sediment-water interface with 767 ambient groundwater discharge. Journal of Hydrology 346:69-80. 768 769 Cardenas, M. B., Wilson, J. L., and Zlotnik, V. A. 2004. Impact of heterogeneity, bed forms, and 770 stream curvature on subchannel hyporheic exchange. Water Resources Research 40:doi: 771 10.1029/2004WR003008. 772 773 Cardenas, M. B. and Zlotnik, V. A. 2003. Three-dimensional model of modern channel bend 774 deposits. Water Resources Research, 39: 1141, doi:10.1029/2002WR001383. 775 776 Chen, X. H., 2004. Streambed hydraulic conductivity for rivers in south-central Nebraska. 777 Journal of the American Water Resource Association 40(3): 561-574. 778 779 Chin, A. 2002. The periodic nature of step-pool mountain streams. American Journal of Science, 780 302:144-167. 781 782 Dent, C. L., Grimm, N. B., Martı, E., Edmonds, J. W., Henry, J. C., and Welter, J. R. 2007. 783 Variability in surface-subsurface hydrologic interactions and implications for nutrient retention 784 in an arid-land stream. Journal of Geophysical Research 112, G04004, 785 doi:10.1029/2007JG000467. 786 787 Elliott, A. H., and Brooks, N. H. 1997a. Transfer of non-sorbing solutes to a streambed with bed 788 forms: Theory. Water Resources Research 33:123-136. 789 790 Elliott, A. H., and Brooks, N. H. 1997b. Transfer of non-sorbing solutes to a streambed with bed 791 forms: Laboratory experiments. Water Resources Research 33:137-151 792 793 Francis, B. A., Francis, L. K., and Cardenas, M. B. in press. Water table dynamics and 794 groundwater-surface water interactions during filling and draining of a large fluvial island due to 795 dam-induced river stage fluctuations. Water Resources Research.

27 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

796 797 Frissell, C. A., Liss, W. J., Warren, C. E., and Hurley, M. D. 1986. A hierarchical framework for 798 stream classification: Viewing streams in a watershed context. Environmental Management 799 10:199-214 800 801 Fritz, B. G. and Arntzen, E. V. 2007. Effect of rapidly changing river stage on Uranium flux 802 through the hyporheic zone. Groundwater 45:753-760. 803 804 Genereux, D. P., Leahy, S., Mitasova, H., Kennedy, C. D., and Corbett, D. R. 2008. Spatial and 805 temporal variability of streambed hydraulic conductivity in West Bear Creek, North Carolina, 806 USA. Journal of Hydrology 358:332-353 807 808 Gooseff, M. N. in press. Defining hyporheic zones - Advancing our conceptual and operational 809 definitions of where stream water and groundwater meet. Geography Compass 810 811 Gooseff, M. N., Anderson, J. K., Wondzell, S. M., LaNier, J., and Haggerty, R. 2006. A 812 modeling study of hyporheic exchange pattern and sequence, size, and spacing of stream 813 bedforms in mountain stream networks. Hydrological Processes 20:2443-2457. 814 815 Gooseff, M. N., LaNier, J., Haggerty, R., and Kokkeler, K. 2005. Determining in-channel (dead 816 zone) transient storage by comparing solute transport in a bedrock channel-alluvial channel 817 sequence, Oregon. Water Resources Research 41:W06014, doi:10.1029/2004WR003513 818 819 Habel, F., Mendoza, C., and Bagtzoglou, A. C. 2002. Solute transport in open channel flows and 820 porous streambeds. Advances in Water Resources 25:455-469. 821 822 Hanrahan, T. P. 2008. Effects of river discharge on hyporheic exchange flows in salmon 823 spawning areas of a large gravel-bed river. Hydrological Processes 22:127-141. 824 825 Harvey, J. W. and Bencala, K. E. 1993. The effect of streambed topography on surface- 826 subsurface water exchange in mountain catchments. Water Resources Research 29:89-98

28 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

827 828 Jencso, K. G., McGlynn, B. L., Gooseff, M. N., Wondzell, S. M., Bencala, K. E., and Marshall, 829 L. A. 2009. Hydrologic connectivity between landscapes and streams: Transferring reach- and 830 plot-scale understanding to the catchment scale. Water Resources Research 45, W04428, 831 doi:10.1029/2008WR007225. 832 833 Jones, K. J., Poole, G. C., Woessner, W. W., Vitale, M. V., Boer, B. R., O’Daniel, S. J., Thomas, 834 S. A., and Geffen, B. A. 2007. Geomorphology, hydrology, and aquatic vegetation drive seasonal 835 hyporheic flow patterns across a gravel-dominated floodplain. Hydrological Processes 22:2105- 836 2113. 837 838 Kasahara, T. 2000. Geomorphic controls on hyporheic exchange flow in mountain streams. MS. 839 Thesis, Oregon State University. 103 p. 840 841 Kasahara, T. and Wondzell, S. M. 2003. Geomorphic controls on hyporheic exchange flow in 842 mountain streams. Water Resources Research 39, 1005, doi:10.1029/2002WR001386. 843 844 Käser, D. H., Binley, A., Heathwaite, A. L., and Krause, S. 2009. Spatio-temporal variations of 845 hyporheic flow in a riffle-step-pool sequence. Hydrological Processes 23:2138-2149 846 847 Larkin, R. G. and Sharp, J. M.Jr. 1992. On the relationship between river-basin geomorphology, 848 aquifer hydraulics, and ground-water flow direction in alluvial aquifers. Geological Society of 849 America Bulletin 104:1608-1620. 850 851 Leopold, L. B. and Maddock, T. Jr. 1953. The hydraulic geometry of stream channels and some 852 physiographic implications. U.S. Geological Survey, Professional Paper No. 252. 57 p. 853 854 Leopold, L. B., Wolman, M. G., and Miller, J. P. 1964. Fluvial Processes in Geomorphology, 855 W.H. Freeman and Co., San Francisco. 522p. 856

29 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

857 Lewandowski, J., Lischeid, G., and Nutzmann, G. 2009. Drivers of water level fluctuations and 858 hydrological exchange between groundwater and surface water at the lowland River Spree 859 (Germany): field study and statistical analyses. Hydrological Processes 23:2117-2128 860 861 Loheide, S.P. and Lundquist, J.D. 2009. Snowmelt-induced diel fluxes through the hyporheic 862 zone. Water Resources Research 45, W07404. 863 864 Malcolm, I. A., Soulsby, C., Youngson, A. F., and Petry, J. 2003. Heterogeneity in ground 865 water - surface water interactions in the hyporheic zone of a salmonid spawning stream. 866 Hydrological Processes 17:601-617. 867 868 Malcolm, I. A., Soulsby, C., Youngson, A. F., and Hannah, D. M. 2005. Catchment-scale 869 controls on groundwater-surface water interactions in the hyporheic zone: Implications for 870 salmon embryo survival. River Research and Applications 21:977-989 871 872 Marion, A., Packman, A. I., Zaramella, M., and Bottacin-Busolin, A. 2008. Hyporheic flows in 873 stratified beds. Water Resources Research 44, W09433, doi:10.1029/2007WR006079. 874 875 Montgomery, D. R., Abbe, T. B., Buffington, J. M., Peterson, N. P., Schmidt, K. M., and Stock, 876 J. D. 1996. Distribution of bedrock and alluvial channels in forested mountain drainage basins. 877 Nature 381:587-589. 878 879 Montgomery, D. R. and Buffington, J. M. 1997. Channel-reach morphology in mountain 880 drainage basins. Geological Society of America Bulletin 109:596-611. 881 882 Nakamura, F. and Swanson, F. J. 1993. Effects of coarse woody debris on morphology and 883 sediment storage of a mountain stream in western Oregon. Earth Surface Processes and 884 Landforms 18:43-61. 885 886 Packman, A. I., and Bencala, K. E. 2000. Modeling surface-subsurface hydrological interactions. 887 In Streams and Ground Waters, Jones, J. B.; Mulholland, P. J., Eds. Academic Press: San Diego.

30 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

888 889 Packman, A. I. and Brooks, N. H. 2001. Hyporheic exchange of solutes and colloids with 890 moving bed forms. Water Resources Research 37:2591–2606. 891 892 Packman, A. I. and MacKay, J. S. 2003. Interplay of stream-subsurface exchange, clay particle 893 deposition, and streambed evolution. Water Resources Research 39, 1097, 894 doi:10.1029/2002WR001432. 895 896 Packman, A. I., Marion, A., Zaramella, M., Chen, C., Gaillard, J-F., and Keane, D. T. 2006. 897 Development of layered sediment structure and its effects on pore water transport and hyporheic 898 exchange. Water, Air, and Soil Pollution: Focus 6:433-442. 899 900 Payn, R. A., Gooseff, M. N., McGlynn, B. L., Bencala, K. E., and Wondzell, S. M. 2009. 901 Channel water balance and exchange with subsurface flow along a mountain headwater stream in 902 Montana, United States. Water Resources Research 45, W11427, doi:10.1029/2008WR007644. 903 904 Peterson, E. W. and Sickbert, T. B. 2006. Stream water bypass through a meander neck, laterally 905 extending the hyporheic zone. Hydrogeology Journal 14:1443-1451 906 907 Pinay, G., O’keefe, T. C., Edwards, R. T., and Naiman, R. J. 2009. Nitrate removal in the 908 hyporheic zone of a salmon river in Alaska. River Research and Applications 25:367-375 909 910 Pinder, G. F. and Sauer, S. P. 1971. Numerical simulation of flood wave modification due to 911 bank storage effects. Water Resources Research 7:63-70. 912 913 Poole, G. C., Stanford, J. A., Running, S. W., and Frissell, C. A. 2006. Multiscale geomorphic 914 drivers of groundwater flow paths: subsurface hydrologic dynamics and hyporheic habitat 915 diversity. Journal of the North American Benthological Society 25:288-303 916

31 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

917 Poole, G. C., Stanford, J. A., Running, S. W., Frissell, C. A., Woessner, W. W., and Ellis, B. K. 918 2004. A Patch hierarchy approach to modeling surface and subsurface hydrology in complex 919 flood-plain environments. Earth Surface Processes and Landforms 29:1259-1274. 920 921 Rehg, K. J., Packman, A. I., and Ren, J. 2005. Effects of suspended sediment characteristics and 922 bed sediment transport on streambed clogging. Hydrological Processes 19:413-427. 923 924 Revelli, R., Boano, F., Camporeale, C., and Ridolfi, L. 2008. Intra-meander hyporheic flow in 925 alluvial rivers. Water Resources Research 44, W12428, doi:10.1029/2008WR007081. 926 927 Salehin, M., A. I. Packman, and A. Wörman. 2003. Comparison of transient storage in vegetated 928 and unvegetated reaches of a small agricultural stream in Sweden: seasonal variation and 929 anthropogenic manipulation. Advances in Water Resources 26:951-964. 930 931 Sawyer, A. H., Cardenas, M. B., Bomar, A., and Mackey, M. 2009a. Impact of dam operations 932 on hyporheic exchange in the riparian zone of a regulated river. Hydrological Processes 23:2129- 933 2137. 934 935 Sawyer, A. H., and M. B. Cardenas. 2009b. Hyporheic flow and residence time distributions in 936 heterogeneous cross-bedded sediment, Water Resources Research, 45, W08406, 937 doi:10.1029/2008WR007632. 938 939 Scordo, E.B. and Moore, R.D. 2009. Transient storage processes in a steep headwater stream. 940 Hydrological Processes 23:2671-2685. 941 942 Shimizu, Y., Tsujimoto, T., and Nakagawa, H. 1990. Experiment and macroscopic modelling of 943 flow in highly permeable porous medium under free-surface flow. Journal of Hydrosciences and 944 Hydraulic Engineering 8:69-78. 945

32 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

946 Soulsby, C., Malcolm, I. A., Tetzlaff, D., and Youngson, A. F. 2009. Seasonal and inter-annual 947 variability in hyporheic water quality revealed by continuous monitoring in a salmon spawning 948 stream. River Research And Applications 25:1304-1319. 949 950 Stanford, J. A. and Ward, J. V. 1993. An ecosystem perspective of alluvial rivers: Connectivity 951 and the hyporheic corridor. Journal of the North American Benthological Society 12:48-60. 952 953 Stanford, J. A., Ward, J. V., and Ellis, B. K. 1994. Ecology of the alluvial aquifers of the 954 Flathead River, Montana. Pgs 367-390. In: J. Gibert, D. L. Danielopol and J. A. Stanford (eds.) 955 Groundwater ecology. Academic Press, San Diego CA. 956 957 Stednick, J. D. and Fernald, A. G. 1999. Nitrogen dynamics in stream and soil waters. Journal of 958 Range Management 52:615-620. 959 960 Storey, R. G., Howard, K. W. F., and Williams, D. D. 2003. Factors controlling riffle-scale 961 hyporheic exchange flows and their seasonal changes in a gaining stream: A three-dimensional 962 groundwater flow model. Water Resources Research 39:1034, doi:10.1029/2002WR001367. 963 964 Thibodeaux, L. J. and Boyle, J. O. 1987. Bed-form generated convective transport in bottom 965 sediment. Nature 325:341-343. 966 967 Tonina, D. and Buffington, J. M. 2007. Hyporheic exchange in gravel bed rivers with pool-riffle 968 morphology: Laboratory experiments and three-dimensional modeling. Water Resources 969 Research 43, DOI: W01421, 10.1029/2005WR004328. 970 971 Tonina, D. and Buffington, J. M. 2009. Hyporheic exchange in mountain rivers I: mechanics 972 and environmental effects. Geography Compass 3:1063-1086. 973 974 Triska, F. J., Kennedy, V. C., Avanzio, R. J., Zellweger, G. W., and Bencala, K. E. 1989. 975 Retention and transport of nutrients in a third-order stream in northwestern California: Hyporheic 976 processes. Ecology 70:1893-1905

33 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

977 978 Valett, H. M. 1993. Surface-hyporheic interactions in a Sonoran Desert stream: Hydrologic 979 exchange and diel periodicity. Hydrobiologia 259:133-144 980 981 Vervier, P. and Naiman, R. J. 1992. Spatial and temporal fluctuations of dissolved organic 982 carbon in subsurface flow of the Stillaguamish River, (Washington, USA). Archiv fur 983 Hydrobiologie 123:401-412. 984 985 Wagner, F. H. and Bretschko, G. 2002. Interstitial flow through preferential flow paths in the 986 hyporheic zone of the Oberer Seebach, Austria. Aquatic Sciences 64:307-316. 987 988 Ward, J. V., Malard, F., Tockner, K., and Uehlinger, U. 1999. Influence of groundwater on 989 surface water conditions in a glacial flood plain of the Swiss Alps. Hydrological Processes 990 13:277-293. 991 992 Woessner, W. W. 2000. Stream and fluvial plain ground water interactions: Rescaling 993 hydrogeological thought. Groundwater 38:423-429. 994 995 Wohl, E. and Merritt, D. M. 2008. Reach-scale channel geometry of mountain streams. 996 Geomorphology 93:168-185. 997 998 Wohl, E. and Merritt, D. 2005. Prediction of mountain stream geomorphology. Water Resources 999 Research 41:W08419, doi:10.1029/2004WR003779. 1000 1001 Wondzell, S. M. 2006. Effect of morphology and discharge on hyporheic exchange flows in two 1002 small streams in the Cascade Mountains of Oregon, USA. Hydrological Processes 20:267-287. 1003 1004 Wondzell, S. M., Gooseff, M. N., and McGlynn, B. L. 2007. Flow velocity and the hydrologic 1005 behavior of streams during baseflow. Geophysical Research Letters 34, L24404, 1006 doi:10.1029/2007GL031256. 1007

34 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

1008 Wondzell, S. M., LaNier, J., Haggerty, R., Woodsmith, R. D., and Edwards, R. T. 2009. Changes 1009 in hyporheic exchange flow following experimental wood removal in a small, low-gradient 1010 stream. Water Resources Research 45, W05406, 31 doi:10.1029/2008WR007214. 1011 1012 Wondzell, S. M. and Swanson, F. J. 1996. Seasonal and storm dynamics of the hyporheic zone 1013 of a 4th-order mountain stream. I: Hydrologic processes. Journal of the North American 1014 Benthological Society 15:1-19. 1015 1016 Wroblicky G. J., Campana M. E., Valett H. M., and Dahm C. N. 1998. Seasonal variation in 1017 surface-subsurface water exchange and lateral hyporheic area of two stream-aquifer systems. 1018 Water Resources Research 34:317-328. 1019

35 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

1020 Figure Legends: 1021 1022 Figure 1. Idealized conceptual model of nested hyporheic flow paths as influenced by step-pool 1023 or pool-riffle sequences. A) Plan view showing arcuate HEF flow paths through the adjacent 1024 floodplain created by the change in the longitudinal gradient over the pool-riffle sequence where 1025 the amount of HEF is proportional to the head gradient. B) Longitudinal-section along the 1026 thalweg of the stream showing the vertical component of HEF flows through the streambed. 1027 1028 Figure 2. A) Hypothetical distribution of channel types along a stream profile in a mountainous 1029 stream catchment (redrawn from Montgomery and Buffington, 1997), and B) the corresponding 1030 relative contribution of turbulent diffusion and both hydrostatic or hydrodynamic processes to 1031 the total amount of HEF occurring within a stream reach. (Note that boundaries between channel 1032 types are often less distinct than shown here and that a range of conditions occurs within each 1033 category, thus the contribution of each process varies both among and within each channel type). 1034 1035 Figure 3. Idealized conceptual model of the influence of lateral inflows on hyporheic exchange 1036 flows. A) A high gradient stream where floodplain alluvium has relatively high saturated 1037 hydraulic conductivity under relatively dry conditions when lateral inputs are low and easily 1038 transported down valley via subsurface flow. Lateral inputs still reach the stream, but are 1039 diverted towards zones with hyporheic upwelling. B) A low gradient stream where floodplain 1040 alluvium has relatively low saturated hydraulic conductivity under relatively wet conditions 1041 when lateral inputs are sufficiently large to overwhelm down valley transport, causing lateral 1042 inputs to cross the valley toward the stream. Lateral inputs severely restrict hyporheic exchange 1043 flows. Legend follows Figure 1. 1044 1045 Figure 4. Idealized conceptual model of the influence changing stream stage on transient 1046 hyporheic exchange. A and B) A losing reach at low baseflow is converted to a gaining reach 1047 during a storm because precipitation recharge and lateral inputs of hillslope water increase water 1048 table elevations more than the corresponding increase in the stream stage. The original stream 1049 and water table position from 4A is shown in 4B for reference (light grey line). C & D) An 1050 example of a river where changes in stream stage result from snow melt, tidal influences, or dam

36 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

1051 releases far upstream. Increased stream stage causes stream water to flow into the adjacent 1052 aquifer creating a losing stream reach. Conversely, decreased stage leads to drainage of the 1053 aquifer creating a gaining reach. Alternating increases and decreases in stream stage leads to 1054 transient hyporheic exchange. The neutral condition (where stream stage is equal to the water 1055 table elevation) is shown for reference (black and white dashed line). Legend follows Figure 1. 1056 1057 Figure 5. Idealized conceptual model of the influence of the change in saturated cross-sectional 1058 area of the floodplain on hyporheic exchange flows. A) The influence of change in valley 1059 constraint with downwelling at the head of an unconstrained reach and upwelling at the 1060 downstream end of the reach caused by the transition from narrow bedrock gorges to wide 1061 alluvial valley floors. B) The influence of variations in depth to bedrock forcing upwelling 1062 upstream of a bedrock sill and downwelling downstream, where the depth of alluvium again 1063 increases. Legend follows Figure 1. 1064 1065 Figure 6. Idealized conceptual model of the influence of meander bends on hyporheic exchange 1066 flow. A) Simple, low radius meander with HEF traversing the point bar and floodplain. B) High 1067 radius meander with incipient meander-cutoff, where the short distance across the neck leads to 1068 much higher head gradients and thus greater HEF through the neck than the remainder of the 1069 meander bar. C) Meander bend with riffles located at the inflections between adjacent meanders 1070 so that head gradients through the point bar are low and much of the HEF occurs around the 1071 riffles, driven by longitudinal changes in gradient. Legend follows Figure 1. 1072 1073 Figure 7. Idealized conceptual model of the influence of back channels on hyporheic exchange 1074 flows. A) A back channel is incised below the water table, acts as a drain, and creats head 1075 gradients from the main channel to the back channel. B) A back channel is plugged near its 1076 downstream end, conducts water onto the floodplain, raises the water table and creats head 1077 gradienets from the back channel to the main channel. C) Complex pattern of HEF caused by 1078 interactions between a riffle in the main channel and a back channel. Paleochannels (dashed 1079 lines) support preferential flow. Legend follows Figure 1. 1080

37 Wondzell and Gooseff: Treatise in Fluvial Geomorpholgy – Geomorphic Controls on Hyporheic Exchange

1081 Figure 8. Idealized conceptual model of the influence of mid-stream islands on hyporheic 1082 exchange flows. A) Parallel and smooth longitudinal gradients in the channels on both sides of 1083 the island create HEF flow paths that parallel stream flow. B) Riffles at the head of the island 1084 enhance head gradients leading to greater HEF. C) Offset riffles create strong cross-island head 1085 gradients and flow paths, resulting in more HEF but with shorter flow path lengths and residence 1086 times. Legend follows Figure 1. 1087 1088 Figure 9. Idealized longitudinal-section in the center of a straight stream channel with bedforms 1089 (triangular dunes) showing the interaction with stream flow that creates regions of low- and high- 1090 pressure on the streambed which drive HEF. Non-hyporheic subsurface flows, known as 1091 underflow (dashed arrows), are present beneath the hyporheic zone. Legend follows Figure 1. 1092 1093 Figure 10. Examples of complex hyporheic flow paths resulting from interactions between 1094 channel morphologic features: A) a steep, 2nd-order step-pool channel with abundant large wood, 1095 and B) a moderate gradient, 5th-order pool-riffle channel with two major spring brooks. Note the 1096 difference in spatial scale between the two stream reaches. Letters indicate morphologic features 1097 driving HEF: S – steps; R – riffles; M – meander bends; B – back channels / spring brooks; I – 1098 islands; and T – a steep riffle at the mouth of a tributary. Equipotential intervals (dashed lines) 1099 are 0.2 m. Hyporheic flow paths (arrows) are hand drawn to indicate general direction of 1100 hyporheic flow through the valley floor.

38 A

Pool - riffle - pool sequence

B Riffle Pool Pool

Floodplain Riffle

Active Subsurface Channel flow path Wetted Channel A

Hill- slope Hollow

Colluvial- bedrock

Cascade Step- pool Plane- bed Pool-riffle Dune-ripple

Hydrostatic B contribution Hydrodynamic Turbulent contribution diffusion

Relative HEF

Longitudinal stream profile through stream network A

Higher gradient, low lateral inputs

B

Lower gradient, high lateral inputs ABPrecipitation Stage Increase

s

Lateral input

CD Stage Increase

Stage Decrease A

Bedrock Unconstrained stream reach gorge

B

Bedrock sill A

B

C Pool / Run A

B

C A

B

C Stream flow A

S S S B I S S S S S S

Scale = 20 m

Log Bedrock B M Wetted stream channel Valley floor alluvium Back channel Hillslope or terrace Equipotential (0.2 m) Hyporheic flow path M R B T R R

R I R R

Scale = 50 m B