Lithologic and hydrologic controls of mixed alluvial–bedrock channels in flood-prone fluvial systems: Bankfull and macrochannels in the Llano River watershed, central Texas, USA

Geomorphology 232 (2015) 1–19

Contents lists available at ScienceDirect

Geomorphology journal homepage: www.elsevier.com/locate/geomorph

Lithologic and hydrologic controls of mixed alluvial–bedrock channels in flood-prone fluvial systems: Bankfull and macrochannels in the Llano River watershed, central Texas, USA Franklin T. Heitmuller a,⁎, Paul F. Hudson b, William H. Asquith c a b c

The University of Southern Mississippi, Department of Geography and Geology, 118 College Drive #5051, Hattiesburg, MS 39406-5051, USA Leiden University College The Hague, P.O. Box 13228, 2501 EE The Hague, The Netherlands U.S. Geological Survey, Texas Water Science Center, Science Building MS-1053, Texas Tech University, Lubbock, TX 79409, USA

a r t i c l e

i n f o

Article history: Received 24 June 2014 Received in revised form 19 December 2014 Accepted 20 December 2014 Available online 24 December 2014 Keywords: Bankfull discharge Channel geometry Llano River Macrochannels Mixed alluvial–bedrock rivers Texas

a b s t r a c t The rural and unregulated Llano River watershed located in central Texas, USA, has a highly variable flow regime and a wide range of instantaneous peak flows. Abrupt transitions in surface lithology exist along the main-stem channel course. Both of these characteristics afford an opportunity to examine hydrologic, lithologic, and sedimentary controls on downstream changes in channel morphology. Field surveys of channel topography and boundary composition are coupled with sediment analyses, hydraulic computations, flood-frequency analyses, and geographic information system mapping to discern controls on channel geometry (profile, pattern, and shape) and dimensions along the mixed alluvial–bedrock Llano River and key tributaries. Four categories of channel classification in a downstream direction include: (i) uppermost ephemeral reaches, (ii) straight or sinuous gravel-bed channels in Cretaceous carbonate sedimentary zones, (iii) straight or sinuous gravel-bed or bedrock channels in Paleozoic sedimentary zones, and (iv) straight, braided, or multithread mixed alluvial–bedrock channels with sandy beds in Precambrian igneous and metamorphic zones. Principal findings include: (i) a nearly linear channel profile attributed to resistant bedrock incision checkpoints; (ii) statistically significant correlations of both alluvial sinuosity and valley confinement to relatively high f (mean depth) hydraulic geometry values; (iii) relatively high b (width) hydraulic geometry values in partly confined settings with sinuous channels upstream from a prominent incision checkpoint; (iv) different functional flow categories including frequently occurring events (b1.5-year return periods) that mobilize channel-bed material and less frequent events that determine bankfull channel (1.5- to 3-year return periods) and macrochannel (10- to 40-year return periods) dimensions; (v) macrochannels with high f values (mostly ≥ 0.45) that develop at sites with unit stream power values in excess of 200 watts per square meter (W/m2); and (vi) downstream convergence of hydraulic geometry exponents for bankfull and macrochannels, explained by co-increases of flood magnitude and noncohesive sandy sediments that collectively minimize development of alluvial bankfull indicators. Collectively, these findings indicate that mixed alluvial–bedrock channels exhibit first-order lithologic controls (lithologic resistance and valley confinement) of channel geometry, second-order hydrologic (flow regime) control of channel dimensions, and third-order sedimentary controls that exert subsidiary influence on channel shape and bed configuration. © 2014 Elsevier B.V. All rights reserved.

1. Introduction River channel adjustment to hydrologic, hydraulic, and sedimentary controls is a topic that has long received considerable attention by fluvial geomorphologists (e.g., Leopold and Wolman, 1957; Schumm, 1960, 1985; Ferguson, 1987; Pitlick and Cress, 2002; Kleinhans, 2010) and is of importance to river restoration efforts, aquatic and riparian ecosystem functions, surface and groundwater exchange dynamics, ⁎ Corresponding author. Tel.: +1 601 266 5423; fax: +1 601 266 6219. E-mail addresses: [email protected] (F.T. Heitmuller), [email protected] (P.F. Hudson), [email protected] (W.H. Asquith).

http://dx.doi.org/10.1016/j.geomorph.2014.12.033 0169-555X/© 2014 Elsevier B.V. All rights reserved.

flood prediction and management, and infrastructure design and maintenance. A typical hydrologic index is bankfull discharge (m3/s), which implicitly is associated with channel dimensions. Sedimentary indices of channel morphology include bed-material composition (e.g., Osterkamp and Hedman, 1982; Howard, 1987; van den Berg, 1995), bedload transport (e.g., Parker, 1979; Bettess and White, 1983; Ferguson, 1987), and bank composition (e.g., Schumm, 1960, 1963; Simpson and Smith, 2001). The general knowledge relating these various indices to channel morphology is primarily derived from rivers in humid settings or from river systems lacking highly variable flow regimes (Baker, 1977; Doyle et al., 2007). Less is known about modes of adjustment for rivers with highly variable flow and sediment transport

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

regimes, although evidence indicates that complex arrangements of channel pattern, shape, and profile occur in arid (e.g., Huckleberry, 1994; Bourke and Pickup, 1999), semiarid (e.g., García, 1995; Heritage et al., 1999), or seasonal (e.g., Gupta, 1995; Kale and Hire, 2007) fluvial systems. Additional differences in fluvial forms and processes are evident for bedrock-dominated (e.g., Baker and Kale, 1998; Tinkler and Wohl, 1998; Erskine and Livingstone, 1999) or geologically complex (Schumm, 2005) fluvial systems. Lastly, unlike this study, relatively few investigations have focused on adjustments of channel pattern and shape imposed by abrupt downstream changes in flood hydrology and surface lithology. This article investigates the roles of hydrology, hydraulics, lithology, and sedimentary characteristics on downstream changes in channel morphology in the Llano River watershed (11,568 km2), central Texas, USA (Fig. 1). This rural and unregulated watershed occurs at a transition between semiarid and subhumid climates, and is subject to extreme instantaneous peak floods (Beard, 1975; Burnett, 2008) ranging nearly three orders of magnitude. Further, a lithologic transition from Cretaceous carbonates in the upper watershed to Precambrian igneous and metamorphic rocks in the lower watershed complicates an assessment of channel adjustment based solely on hydrologic considerations.

geometry focused on independent hydraulic or sedimentary controls of pattern (Leopold and Wolman, 1957; Schumm, 1963; Schumm and Khan, 1972) and shape (Leopold and Maddock, 1953). Later investigations generally concluded that hydraulic and sedimentary controls jointly determine channel geometric configurations, most notably channel patterns (Ferguson, 1987; van den Berg, 1995). Further advances in scientific description and understanding of channel patterns are likely to come from rivers where abrupt transitions in either hydraulic or sedimentary controls occur (Kleinhans, 2010). Additionally, process-based classifications for multichannel and bedrock-incised channels are not fully developed at present (Carling et al., 2014), although Meshkova et al. (2014) introduces a simple classification of bedrock channels and mixed alluvial–bedrock channels that are either laterally confined or vertically constrained. For example, confined channels are restricted from lateral migration by resistant bedrock banks but have erodible bed material, whereas constrained channels are restricted from incision by an indurated bed but have erodible bank material. Mixtures of the two mixed alluvial–bedrock classifications could arguably exist, but are not considered further for purposes of this investigation. Empirical analyses of cross-sectional morphology commonly utilize hydraulic geometry equations (Leopold and Maddock, 1953), which are uncomplicated power relations associating water-surface width (m), mean depth (m), and mean velocity (m/s) to discharge (m3/s). Hydraulic geometry variables have been successfully used to derive predictive models of runoff discharge and velocity at ungauged locations in flash-flood-prone regions (e.g., Asquith et al., 2013), which lends credence to their utility for associating cross-sectional morphology with formative flows in this study. Exponents (b, f, and m) represent

2. Theoretical framework The theoretical framework to associate channel geometry in the Llano River watershed to hydrologic and boundary conditions is derived from previously published research on channel pattern, hydraulic geometry, and dominant discharge concepts. Early research on channel

9 9 °3 0 ' W

9 9 °W

Ck .

L EL

ry

ST

ck o

Ck .

.

R.

R. Llano

Lake LBJ

er

30°N

vil s

e De

R

KINGSLAND

av

Ri v er

a Sab

Hi

CA

Sa n

28

Be

Ri

29°N

R Fo rk n so n

Li t t l

rd

na

ra w

26

s

Lampasas Cut Plain

P ed

0

25

o

G ra nd

50

75

e

DEL RIO 100 KILOMETERS

105°W

TEXAS

City or community

e rn a les R.

Guadalu

AUSTIN

pe R.

na

i

D

J am es

Ck .

R.

L l an o

r

Med

ry

ive

es R. Nu e c

Paleozoic sedimentary Precambrian igneous and metamorphic

98°W

ho R. Conc

Sabinal R.

Dr y

oR

25

R.

D

27

rad

R. vil s De

Joh

3 0 °3 0 ' N

Co lo

23 24 ney C k . Ho

99°W

B ra zo

os Pec

3 0 °N

Colorado R.

O

k. oC

Lithology Quaternary deposits Cretaceous carbonate

40 KILOMETERS

100°W

l an

LLAN

21

le L

30

Lit t

20

20

101°W

(B)

Llano River watershed 11,568 km2

Spring 10

Llan o Riv er

Lake Buchanan

EXPLANATION

ROCKSPRINGS

0

k. nC . Ck

Ck .

5

o

6

so

Big Paint Springs

13

r i ve

TELEGRAPH

Seven Hundred Springs

nd

12 11 10

ay

Ri ver no

4

8

t Pain

u th L l a

IVY CHAPEL

9

hn

7

. Ck

M

3

a

22

17 18

14

JUNCTION Maynard Spring

ch e

. Ck

2

So

Ck .

s

o

er h L n o Ri v la

eC k.

19

. w Ck

ra

N

rt

li n

MASON 16 15

B ea r

pe

ROOSEVELT

Sa

INCISION CHECKPOINT

p Co

1

W il lo

SONORA

Co m a n

Bi g

San Fe rn

Jo

Study sites in the Llano River watershed, central Texas

102°W

9 8 °3 0 ' W

32°N

10 0 °W

(A)

R.

SAN ANTONIO

Fri oR .

100°W

95°W

(C) 35°N

10 0 °3 0 ' W

31°N

2

Sites with sediment samples Sites without sediment samples

TEXAS Color

30°N

ive r

ado

Edwards Plateau

ec

MEXICO Plateau delineation from Griffith et al. (2004) 0

100

200

300

R

r

sR

ive

Pe co

.

Jam es River near M ason (2) Llano River at Jam es River Crossing near M ason (3) LCRA Com anche Creek near M ason (1) 08150700 Llano River near M ason (3) 08150800 Beaver Creek near M ason (2) LCRA Willow Creek near M ason (1) Llano River at Cast ell (3) LCRA Hickory Creek near Castell (1) LCRA San Fernando Creek near Llano (1) LCRA Johnson Creek near Llano (1) 08151500 Llano River at Llano (3) LCRA Litt le Llano River near Llano (1) Llano River near Kingsland (2) Honey Creek at KDK Ranch near Kingsland (3) LCRA Honey Creek near Kingsland (1)

sR

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

o az

Nort h Llano Draw near Sonora (2) Nort h Llano River near Roosevelt (3) 08148500 Nort h Llano River near Junct ion (4) Sout h Llano River at Baker Ranch near Rocksprings (4) Sout h Llano River at U.S. Highw ay 377 near Rocksprings (2) Sout h Llano River at 700 Springs Ranch near Telegraph (3) Sout h Llano River at Sout h Llano River St at e Park (3) Sout h Llano River at Texas Tech Universit y - Junct ion (4) 08150000 Llano River near Junct ion (4) LCRA Johnson Fork near Junct ion (1) Johnson Fork at Low lands Crossing near Junct ion (3) Llano River near Ivy Chapel (3) LCRA Jam es River near M ason (1)

Nu

1 2 3 4 5 6 7 8 9 10 11 12 13

Br

St udy sites (num ber of cross sect ions)

e s R.

GULF OF MEXICO

400 KILOMETERS

Fig. 1. (A) Study sites, number of cross sections per site, hydrography, and lithology (Barnes, 1981) of the Llano River watershed in central Texas, USA. Site names preceded by an 8-digit identification number are located at active U.S. Geological Survey (USGS) streamflow-gauging stations. Site names preceded by (LCRA) are located at active Lower Colorado River Authority streamflow-gauging stations. All other sites are currently (as of 2014) ungauged stream locations. (B) Hydrography of the Edwards Plateau region in central Texas, USA. (C) Hydrography of Texas and location of the Edwards Plateau ecoregion. Edwards Plateau boundary is derived from ecoregion boundaries of Griffith et al. (2004).

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

the rate of change of width, depth, and velocity, respectively, to discharge. Algebraic relations commonly used in channel morphology are: b

w ¼ aQ ; f d ¼ cQ ; and m

U ¼ kQ ; where w is water-surface width (m); d is mean depth (m); U is mean velocity (m/s); Q is discharge (m3/s); a, c, and k are empirically derived regression coefficients (intercepts); and b, f, and m are empirically derived coefficients (exponents). Hydraulic geometry analyses are useful when adequate discharge and cross-sectional data are available. Analytical results are readily compatible with common hydrologic analyses such as flood frequency (Gregory and Madew, 1982). Hydraulic geometry can be applied to the range of flows at one cross section, termed at-a-station, or for multiple stations or sites downstream for a user-specified index of discharge (e.g., bankfull conditions). The latter technique is subject to interpretive error if sites are representative of anomalous boundary conditions (e.g., isolated bedrock zone, artificial bank reinforcement structures). Use of downstream hydraulic geometry has further been criticized in several areas including: (i) adherence to log-linear relations and neglect of flow resistance (Richards, 1973), (ii) interpretive limitations for rivers with relatively low ratios of stream power to particle size (Wohl, 2004), (iii) interpretive limitations for rivers with diverse sedimentary (e.g., Parker, 1979; Knighton, 1987; Huang and Warner, 1995) and vegetative (e.g., Hey and Thorne, 1986) controls, and (iv) instability of cross-sectional shape (Phillips, 1990; Fonstad and Marcus, 2003). The technique, however, has been successfully applied to derive regional regression equations to predict channel morphology (e.g., Betson, 1979; Castro and Jackson, 2001), infer morphologic changes resulting from a single flood (e.g., Merritt and Wohl, 2003), discriminate between channel patterns (e.g., Xu, 2004), and distinguish among various channel boundary and hydraulic controls on channel morphology (e.g., Pitlick and Cress, 2002; Torizzo and Pitlick, 2004; Wohl and Wilcox, 2005; Arp et al., 2007; Eaton and Church, 2007; Latrubesse, 2008). Downstream hydraulic geometry techniques assume a dominant, channel-forming discharge. Alluvial channels exhibit this discharge at bankfull stage (Wolman and Leopold, 1957), which is commonly associated with return periods of about 1 to 2 years (Wolman and Miller, 1960; Dury, 1973; Andrews, 1980; Biedenharn et al., 1999). Others have shown, however, that channel-forming discharge of rivers with highly variable flow regimes is likely to occur less frequently (e.g., Schick, 1974; Pickup and Warner, 1976; Baker, 1977; López-Bermúdez et al., 2002; Doyle et al., 2007). Additionally, the complex arrangements of alluvial features that occur at various flow depths in systems characterized by high-magnitude floods, including seasonal or monsoonal rivers (e.g., Gupta, 1995; van Niekerk et al., 1999; Heritage et al., 2001) can challenge prevailing assumptions about dominant discharge. For example, macrochannels with defined breaks in cross section slope are formed by extreme floods and commonly include a lower, smaller bankfull channel that is maintained by frequent flows and partly or wholly destroyed by flows that fill the macrochannel (Gupta, 1995). 3. The Llano River watershed The Llano River watershed (Fig. 1) is located in the Edwards Plateau of central Texas; the Llano River is a tributary to the larger Colorado River. A complex surface lithology, given the relatively small size of the watershed (11,568 km2), has been exposed by exhumation of the Llano Uplift (Murray, 1961). At the surface, the structural uplift is composed of Precambrian intrusive igneous and metamorphic rocks centered in the eastern part of the watershed, and is rimmed by lower

3

Cretaceous limestone and dolostone rock units (Barnes, 1981). A transitional zone of Paleozoic sedimentary rocks is present between the lower Cretaceous rim and the Precambrian basement rocks. The western tributaries of the Llano River initiate at elevations exceeding 700 m. The channels dissect the Edwards Plateau and are composed of subhorizontally bedded limestone and dolostone sequences with varying amounts of nodular chert (Barnes, 1981). Based on a georeferenced version of the Barnes (1981) geologic map of the region, lower Cretaceous rock formations compose 66% (7629 km2) of the watershed area. Further downstream, Paleozoic rock formations of various lithologies are present, mostly Ordovician carbonate rocks and Cambrian sandstone, and compose almost 12% (1369 km2) of the watershed area. In the lower and eastern side of the watershed, exposed Precambrian granite, gneiss, and schist dominate and compose 19% (2180 km2) of the watershed area. The remaining 3% (390 km2) includes Quaternary alluvium. Differential rates and patterns of bedrock weathering throughout the watershed appreciably affect valley confinement, alluvial development, and sediment composition. According to Shepherd (1975), a general classification of channels in central Texas ranges from those with relatively low width–depth ratios, relatively sinuous pool–riffle sequences, and poorly sorted gravel beds in limestone-dominated zones to wider, straighter, well sorted sand-bed channels in granitedominated zones. In the study area, limestone and dolostone in the upper watershed intrinsically weather more readily and result in a relatively wide alluvial valley near Junction, Texas (Figs. 1, 2, and 3A). The eastward transition to more resistant Paleozoic sedimentary rock is associated with a decrease in valley width and more frequent bedrock exposures that laterally confine and vertically constrain the main-stem channel. Alluvial development is limited to localized, hydraulically favorable zones of deposition, including discontinuous floodplains and mid-channel bars within macrochannels. Precambrian crystalline rocks in the lower watershed exhibit varying resistance (Shepherd, 1979), commonly weather to grus, and provide mostly sand-sized sediment to channels (Heitmuller and Hudson, 2009). The valley is laterally confined in its lower reaches and supports limited depositional features that thinly drape over the shallow bedrock, which constitutes the alluvial overprint described in Carling (2009). The plateau-based setting, geologic structure, lithology, and incision history have resulted in a nearly linear streambed profile of the Llano River (Fig. 4). An abrupt increase in bedrock resistance in the Paleozoic zone downstream from Junction is an incision checkpoint that forces alluvial deposition upstream from the checkpoint to maintain grade and, therefore, results in a partly confined setting favoring the development of floodplains and sinuous channels. The combined South Llano and Llano River channel descends from ~ 700 to 250 m over a distance of about 300 km, which is an overall dimensionless channel slope of 0.0015. Aside from minor curvature at the uppermost reach and subtle deviations thereafter, the linear profile negates explanations for major downstream channel adjustments to slope-based changes in stream power. The annual precipitation regime of the watershed ranges from semiarid in the west (~580 mm) to subhumid in the east (~760 mm) (Texas Parks and Wildlife Department, 2013). Rainfall, however, is highly variable, and droughts can rapidly transition to floods (Bomar, 1983). As a result of locally steep slopes and very thin soils (Cooke et al., 2003), rapid runoff rates coupled with large depths of rainfall (Asquith, 1998; Asquith and Roussel, 2004) lead to extreme flash floods (Beard, 1975; Asquith et al., 1996; Tinkler, 2001; Burnett, 2008) (Table 1) capable of transporting substantial quantities of sediment (Heitmuller and Asquith, 2008) and greatly modifying channel morphology (Baker, 1977). Heitmuller (2011) determined that bankfull flows are less frequent (return periods commonly 1.5 to 3 years) than those responsible for entraining bed material (return periods commonly b 1.5 years) along low-flow channels and channel bars in the Llano River watershed, a finding compatible with that of Pitlick and Cress (2002) along the

4

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

610

500 South Llano River at Texas Tech University—Junction Study site 8 (Fig. 1) Drainage area: 2271 km2 Downstream distance: 110.1 km

Elevation (m)

600

475

Llano River at James River Crossing near Mason Study site 15 (Fig. 1) Drainage area: 8032 km2 Downstream distance: 192.7 km

575 450 550 425 Bankfull

Alluvium

510 0

500

1000

1500

Cretaceous bedrock 2000

2500

475

Elevation (m)

450

Paleozoic bedrock

Alluvium

525

Bankfull

400 500

1000

1500

2000

2500

3000

375 Beaver Creek near Mason Study site 18 (Fig. 1) Drainage area: 558 km2 Downstream distance: 57.2 km

350

425

Llano River at Llano Study site 24 (Fig. 1) Drainage area: 10,885 km2 Downstream distance: 261.2 km

325

Alluvium Bankfull

Alluvium

400

300

Precambrian bedrock

Bankfull

Precambrian bedrock 375 500

275 1000

1500

2000

2500

3000

Distance from left (m)

500

1000

1500

2000

2500

Distance from left (m)

Fig. 2. Valley cross sections of selected study sites in the Llano River watershed in central Texas, USA, derived in GIS from a 10-m digital elevation model. Downstream distance is relative to the incipient head of the uppermost first-order stream. Bankfull stages are determined from cross-sectional surveys. The alluvial valley in the upper and middle parts of the watershed transitions to a bedrock-controlled valley in the lower watershed. Modified from Heitmuller and Hudson (2009).

gravel-bed Colorado River. Reservoirs regulating the flow regime in the watershed are not substantial — although low-water control structures in Junction and Llano, Texas, allow ponded water to be used for municipal supply. 4. Data and methods This investigation uses data derived from field surveys of channel topography and boundary composition, laboratory sediment analyses, flow-resistance analyses, flood-frequency analyses, and geographic information system (GIS) mapping. 4.1. Field surveys Multiple cross sections were surveyed between December 2004 and June 2007 at each of 19 sites along the Llano River and selected tributaries (Fig. 1) using a total-station surveying instrument. Additionally, a single cross section at each of nine tributary sites was provided by the Lower Colorado River Authority (LCRA). All of the selected field sites for this study represent variations in drainage area, tributary inputs, and lithology in the watershed (Table 2). Five sites also are coincident with U.S. Geological Survey (USGS) streamflow-gauging stations, and surveyed elevations are assigned the corresponding USGS-supported vertical datum. The remaining 14 sites were assigned a local and arbitrary vertical datum. Common hydraulic properties including hydraulic radius, crosssectional area, and Froude number were computed from surveyed cross sections using WinXSPRO (USDA Forest Service, 2013). Channel slopes for 11 sites were derived from total-station surveys of watersurface or thalweg elevations between cross sections. The GIS-based

longitudinal profile data (see Section 4.3) were used for the remaining eight sites where the measured distance between water-surface elevation points was not sufficient to obtain reliable estimates. For these instances, the GIS-based slope is assumed more representative of the so-called true energy grade line at bankfull stage relative to an otherwise flat channel slope computed with survey data. Channel bed and bank sediments were sampled along the same cross sections that were used for topographic surveys. Multiple bed and bank samples were obtained along each cross section and were spatially distributed to account for various geomorphic surfaces, including the low-flow channel, channel bars, banks, and inset floodplains. Gravel-bed material was sampled using a modified Wolman (1954) pebble count procedure (e.g., Heitmuller and Asquith, 2008). Sandsized or finer sediments were sampled with a hand scoop and bagged for further analyses. 4.2. Laboratory sediment analyses Bagged sediment samples were analyzed for particle size by hydrometer and wet-sieve methods (Gee and Bauder, 1986) in the Applied Geomorphology and Geo-Archaeology Laboratory at The University of Texas at Austin (UT). Data were entered into a preformatted spreadsheet, and cumulative particle-size distribution curves were developed. 4.3. GIS and statistical analyses A variety of morphologic evaluations and products were made using GIS, including longitudinal profiles, valley cross sections, a channel classification scheme, and other planform values. Profiles and valley

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

Cretaceous carbonate bedrock

Low -flow

Chan nel b a

Fl

oo

dp

la

in

r

Terrace deposit

chan ne

l

(A)

BANKFULL Alluvium

(B) Inset floodplains

Paleozoic sedimentary bedrock

MACROCHANNEL BANKFULL Alluvium

(C) Precambrian igneous or metamorphic bedrock

Channel bars

MACROCHANNEL

5

data, including area and hydraulic radius, were computed at 0.01-m stage increments using WinXSPRO. Mean flow velocity was computed at 0.3048-m (1-foot original units) increments by dividing the discharge by the cross-sectional area, and flow-resistance coefficients were estimated. The hydraulic values and flow-resistance coefficients listed in Table 3 correspond with the stage at observed breaks in slope along alluvial banks or bridge-apron tops of the cross section at the gauge location. Values either correspond with bankfull or macrochannel conditions, depending on the channel morphology at the study site. The use of cross sections at bridges is appropriate because the structures fully span the bankfull or macrochannels at those sites. Lastly, flow-resistance computations were not solved at USGS station 08150000 Llano River near Junction, Tex., because a surveyed cross section was not obtained at the exact gauge location. Manning's n values (Barnes, 1967) at or near bankfull stage range from 0.029 to 0.055 for channels with cobble- and pebble-sized bed material, 0.027 to 0.047 for channels with sand-sized bed material, and 0.041 to 0.073 for channels with numerous bedrock exposures. Darcy– Weisbach f friction factors (Robert, 2003) range from 0.047 to 0.165 for channels with cobble- and pebble-sized bed material, 0.042 to 0.152 for channels with sand-sized bed material, and 0.076 to 0.279 for channels with numerous bedrock exposures. Variability is evident given the range of values, in large part because of the presence of channel bars and bedrock exposures. Unless otherwise noted below, flood-frequency analyses at all study sites and hydraulic geometry computations of bankfull stage at ungauged sites used an approximate mean Darcy–Weisbach f value of 0.115. The corresponding approximate mean n value of 0.045 is slightly higher than the n value of 0.035 used for bedmaterial entrainment computations in Heitmuller and Asquith (2008) that were largely based on authors' qualitative judgment, but it is within the quantitatively determined range (0.04 to 0.06) of Conyers and Fonstad (2005). 4.5. Flood-frequency analyses

BANKFULL Alluvium

Fig. 3. Cross-sectional representations of bankfull and macrochannels in the Llano River watershed in central Texas, USA, including: (A) a partly confined sinuous channel in the upper watershed, (B) a bedrock-constrained macrochannel with inset floodplains defining the bankfull channel, and (C) a multithread, bedrock-constrained and confined macrochannel with minor alluvial deposits defining the vertical extent of bankfull and macrochannels.

cross sections were generated from 10-m digital elevation models (DEMs) and the high-resolution National Hydrography Dataset (NHD) (USGS, 2008a,b). The North Llano, South Llano, and Llano rivers were classified through a combination of field observations and GIS analyses of digital orthoimagery and surface geology (Barnes, 1981). These data sets also were used to quantify alluvial sinuosity (channel length/valley axis length) and valley width. Simple statistical analyses, including particle-size descriptors (e.g., median diameter) and various hydraulic values (e.g., unit stream power), were computed in spreadsheets. Statistical analyses, including hydraulic geometry, trend lines, and flood frequency, were done using R (R Development Core Team, 2013). 4.4. Flow resistance analyses Hydraulic geometry analyses at ungauged sites require an appropriate flow-resistance coefficient to compute mean flow velocity and discharge. Manning's n and Darcy–Weisbach f flow-resistance coefficients were estimated using surveyed cross sections coupled with expanded stage-discharge rating curve tables at USGS and LCRA streamflowgauging stations (Table 3). Expanded stage-discharge rating tables list incremental stage and discharge relations for every 0.003 m (0.01 ft, native unit increments of the rating tables) of stage. Cross-sectional

Flood-frequency analyses include various statistical methods to estimate discharges associated with given T-year return periods and are necessary to determine the frequency at which bankfull or macrochannel flow events occur. Two different flood-frequency methods were used in this investigation: (i) partial-duration analyses at sites with USGS stations and (ii) regionally applicable regression equations at ungauged sites. Partial-duration analysis uses period-of-record peak flows above a designated base discharge (Stedinger et al., 1993). The partialduration series is preferred over the annual-maximum series to estimate the frequency of relatively small or moderate events (return periods of 1 to 10 years) or for stations with a short period of record (Soong et al., 2004) when it is reasonable to assume that partialduration analysis is appropriate for investigations of frequently occurring bankfull flows. Further, the episodic flood regime in the study area commonly has one year with numerous peaks above the base discharge and other years with no peaks above base discharge. This variability is an implicit issue for the study area, but is a statistical problem largely circumvented through use of the partial-duration series. Base discharges and peaks above base for the five USGS stations in the Llano River watershed were published in USGS annual water-data reports (e.g., USGS, 2007). Peaks above base discharge were manually digitized, and empirical return periods were computed by Weibull plotting positions using the following equation: T ¼ ðN þ 1Þ=M where T is the return period in years (y), N is the period of record (y), and M is the rank of the discharge in order from largest to smallest. Peaks above the USGS-established base discharge were read into the lmomco package (Asquith, 2013) of R (R Development Core Team, 2013), and the four-parameter Κ distribution was fit using L-moments

6

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

750

South Llano and Llano rivers

750

700

700

650

650

Incision checkpoint

Elevation (m)

600

600

550

450

400

400

350

350 Llano

300 0

50

100

150

200

250

300

350

250 750

700

700

650

650

no

an

dL

lan

op

rof

ile

0

50

100

150

200

250

300

350

300

350

Little Devils and James rivers

600

600

Elevation (m)

Lla

300

Johnson Fork

So

550

uth

an

Lla

no

op

lan

op

400

rof

ile

350

an

dL

450

lan

400

uth

500

dL

450

So

550

Lla

no

500

rof

ile

350 300

300 250 0

uth

500

450

750

So

550 Junction

500

250

North Llano River

50

100

150

200

250

300

350

250

0

50

100

Downstream distance (km)

150

200

250

Downstream distance (km)

Fig. 4. Longitudinal profiles of the combined South Llano and Llano rivers, the North Llano River, Johnson Fork, and the combined Little Devils and James rivers in central Texas, USA, were rendered from GIS analysis of 10-m digital elevation models (DEMs). The main-stem South Llano and Llano rivers, as well as the North Llano River, have a remarkably linear streambed profile, although the overall slope is greater for the North Llano River. Major tributaries of the Llano River show a subtle slope curvature with increasing distance and are steeper than mainstem channels. Note the incision checkpoint of resistant Paleozoic bedrock along the Llano River that results in alluvial deposition upstream.

of observed data (Asquith, 2011). The sample sizes were deemed large enough by the authors to reliably use a four-parameter distribution. Because partial-duration series include a greater number of peaks than years in the period of record, the general assumption of a stationary

Poisson process that is described in Stedinger et al. (1993, sec. 18.6.1) was used to transform average arrival rates into annual exceedance probabilities. Flood magnitudes for various return periods were calculated from the Κ distribution (Table 4).

Table 1 Hydrologic data for USGS streamflow-gauging stations in the Llano River watershed.a Modified from Heitmuller and Hudson (2009). USGS station Station name number 08148500 08150000 08150700 08150800 08151500

DA Period of record used (km2)b

North Llano River near 2335 Junction, Tex. Llano River near 4815 Junction, Tex. Llano River near 8418 Mason, Tex. Beaver Creek near 558 Mason, Tex. Llano River at 10,885 Llano, Tex.

October 1, 1915 to October 26, 1977; June 13, 2001 to April 5, 2008 October 1, 1915 to May 10, 1993; October 1, 1997 to April 5, 2008 March 7, 1968 to May 9, 1993; October 1, 1997 to April 5, 2008 August 1, 1963 to September 30, 2007 September 17, 1939 to April 5, 2008

Qmnmx and Qmdmx Qmax (m3 s−1) and date Qmn and Qmd (m3 s−1; m3 s−1) (m3 s−1; m3 s−1) 1.94; 0.57

595d; 173d e

e

5.65; 2.89

982 ; 374

9.24; 4.81

f

Qmnmn and Qmdmn (m3 s−1; m3 s−1)c

2888; September 16, 1936 0.149; 0.050 9033; June 14, 1935

1.69; 1.44

1374 ; 714

10,760; June 14, 1935

2.59; 2.24

0.55; 0.10

278g; 213g

1894; August 3, 1978

0.0092; 0.0011

10.9; 4.45

1458h; 796h

10,760; June 14, 1935i

1.29; 0.934

f

a DA, drainage area; Q mn, mean daily mean discharge; Q md, median daily mean discharge; Q mnmx, mean annual maximum discharge; Q mdmx, median annual maximum discharge; Q max, maximum instantaneous discharge; Q mnmn, mean annual minimum discharge; Q mdmn, median annual minimum discharge. b Drainage area derived from GIS analysis of 10-m digital elevation model. c From Asquith et al. (2007) using daily mean discharge values from the beginning of the period of record for each gauging station to December 31, 2003. d Period of record used for annual maximum series from October 1, 1915 to September 30, 1978; June 13, 2001 to September 30, 2006. e Period of record used for annual maximum series from October 1, 1915 to September 30, 2007. f Period of record used for annual maximum series from March 7, 1968 to September 30, 2007. g Period of record used for annual maximum series from October 1, 1963 to September 30, 2007. h Period of record used for annual maximum series from October 1, 1939 to September 30, 2007. i Maximum instantaneous discharge value determined from nearby gauging station (08151000 Llano River at Castell, Tex.) and indirect estimation methods.

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

7

Table 2 Selected morphometric and sedimentary characteristics of study sites in the Llano River watershed in central Texas; various sites near LCRA stations are not included if cross-sectional morphology measurements contain bridge structures; particle size values represent averages of multiple samples collected at each site.a Site

Category

DA (km2)

DD (km)

VW (m)

Alluvial sinuosityb

S

Channel bar d50 (mm)

Bank d50 (mm)c

North Llano Draw near Sonora North Llano River near Roosevelt North Llano River near Junction South Llano River at Baker Ranch near Rocksprings South Llano River at U.S. Highway 377 near Rocksprings South Llano River at 700 Springs Ranch near Telegraph South Llano River at South Llano River State Park South Llano River at Texas Tech University—Junction Llano River near Junction LCRA Johnson Fork near Junction Johnson Fork at Lowlands Crossing near Junction Llano River near Ivy Chapel LCRA James River near Mason James River near Mason Llano River at James River Crossing near Mason Llano River near Mason Beaver Creek near Mason Llano River at Castell Llano River at Llano Llano River near Kingsland Honey Creek at KDK Ranch near Kingsland

ED KG KG ED ED KG KG KG KG KG KG KG P P P pC pC pC pC pC P

7.79 1145 2335 417 1134 1352 2256 2271 4815 758 778 5939 845 877 8032 8418 558 9429 10,885 11,406 28.6

0.90 56.4 89.2 29.5 64.5 76.3 105 110 121 52.7 57.5 140 53.7 65.6 193 210 57.2 230 261 292 11.7

–d 250–375 750–1000 400–415 100–250 225–350 1000–1150 1250–1600 700–1000i 650–850 850–950 300–650 –d –d 400–500 –d –d –d –d –d –d

1.00 1.13 1.13 1.19 1.06 1.01 1.31 1.31 1.38 1.33 1.33 1.24 1.00 1.00 1.08 1.00 1.00 1.00 1.00 1.00 1.00

0.0054e 0.0021g 0.0029h 0.0015h 0.0036e 0.0030e 0.0012e 0.0014e 0.0071g 0.0026e 0.0053g 0.0015e 0.0034e 0.0059g 0.0025g 0.0025g 0.0031e 0.0013e 0.0014g 0.0027g 0.0220g

– 36.6 18.0 0.039 42.2 44.9 25.5 21.8 16.3 – 23.8 23.7 – 46.0 28.0 14.2 0.884j 2.09 0.457 0.629 39.6

0.0042f 0.044 0.110 0.033 – 0.076 0.039 0.127 0.066 – 0.353 0.062 – 0.228 0.104 0.160 0.264 0.090 0.146 0.198 0.464

a ED, ephemeral draw; KG, Cretaceous gravel-bed; P, Paleozoic bedrock or gravel-bed; pC, Precambrian straight, braided, or multithread, mixed alluvial–bedrock; DA, drainage area; DD, downstream distance; VW, alluvial valley width; S, dimensionless channel slope; d50, median particle size; –, not available. b Alluvial sinuosity was measured by dividing the channel length by the valley-axis length. In some cases, alluvial sinousity is ~1.0 because of confined valley settings, even though the valley meanders across the land surface. c Composite of channel bank and floodplain surface deposits. d Valley confined by bedrock. Alluvial deposits thinly overlie bedrock, but alluvial floodplains are absent. e Channel slope derived from GIS analysis of 10-m DEMs. f Particle size for material collected from the base of the draw, which adequately represents all nearby deposits. g Channel slope derived from total-station survey of water-surface elevations at low-flow conditions. h Channel slope derived from total-station survey of thalweg elevations, which were at similar depths below the water surface at the upper and lower ends of the reach. i Valley width is associated with abandoned valley segment, not with present-day channel avulsion through bedrock exposure. j Median particle size only for sand-sized fraction, although material is bimodal with some proportion of gravel-sized material.

For ungauged sites and two LCRA sites with short record lengths, regionally applicable regression equations specific to undeveloped Texas watersheds were used. Six equations by Asquith and Thompson (2008) were applied to all study sites to estimate the 2-, 5-, 10-, 25-, 50-, and 100-year annual flood discharges; and the equations are based on drainage area, channel slope, and mean annual precipitation (Table 5). Channel slope was based on longitudinal profile data derived from DEMs, and specifically not local total-station survey data. This is because the regression equations by Asquith and Thompson (2008) use a non-local definition of slope based on a whole water channel.

Regression-based annual peak discharge values were compared with those computed by partial-duration analyses at gauged locations, and the regression-based values were consistently less than those derived from partial-duration analyses. The average residual standard errors (log10) between regression-based and partial-duration analysis values were computed for each return period (0.3706, 0.3444, 0.3221, 0.2974, 0.2782, and 0.2595 for the 2-, 5-, 10-, 25-, 50-, and 100-year return periods, respectively), and these scale factors were applied to regression-based results (Asquith and Thompson, 2008) at ungauged sites (Table 4).

Table 3 Manning's n and Darcy–Weisbach f flow-resistance coefficients. The coefficients were solved for using surveyed cross sections coupled with expanded stage-discharge rating curve tables of instantaneous values at selected USGS and LCRA gauging stations in the study area; values at or near bankfull or macrochannel stage are shown below.a Station 08148500 North Llano River near Junction, Tex. LCRA Johnson Fork near Junction LCRA James River near Mason LCRA Comanche Creek near Mason 08150700 Llano River near Mason, Tex. LCRA Willow Creek near Masonf LCRA Hickory Creek near Castellf LCRA San Fernanado Creek near Llanof LCRA Johnson Creek near Llanof 08151500 Llano River at Llano, Tex. LCRA Honey Creek near Kingslandf

Bed materialb Cobbles, pebbles Cobbles, pebbles Cobbles, pebbles Sand Cobbles, pebbles Sand Sand, bedrock Sand, bedrock Sand Sand, bedrock Cobbles, pebbles

Stage (m) c

5.49 6.10c 5.49e 3.96c 6.71e 4.57c 5.79c 5.49c 4.88c 7.62e 5.49c

Q (m3/s)

S

A (m2)

w (m)

R (m)

d (m)

U (m/s)

n

f

Fr

462 501 1250 207 2470 456 708 651 510 3060 756

0.0029 0.0026d 0.0034d 0.0048d 0.0025 0.0030d 0.0051d 0.0025d 0.0029d 0.0014 0.0067d

232 267 318 105 736 190 318 283 142 1080 169

88.6 92.1 117 66.8 124 80.7 90.4 94.7 57.6 198 58.6

2.59 2.85 2.71 1.56 5.74 2.32 3.45 2.94 2.41 5.48 2.83

2.62 2.90 2.73 1.58 5.92 2.35 3.52 2.99 2.46 5.55 2.89

1.99 1.87 3.94 1.97 3.36 2.40 2.23 2.30 3.60 2.82 4.47

0.051 0.055 0.029 0.047 0.048 0.040 0.073 0.045 0.027 0.041 0.037

0.149 0.165 0.047 0.152 0.099 0.095 0.279 0.109 0.042 0.076 0.074

0.39 0.35 0.76 0.50 0.44 0.50 0.38 0.42 0.73 0.38 0.84

a Q, instantaneous discharge; S, dimensionless channel slope; A, cross-sectional area; w, water-surface width; R, hydraulic radius; d, mean depth; U, mean flow velocity; n, Manning's n flow-resistance coefficient; f, Darcy–Weisbach friction factor; Fr, Froude number. b Bed material size categories are based on the Wentworth scale (Wentworth, 1922; Guy, 1969), where clay b0.004 mm, silt is between 0.004 and 0.062 mm, sand is between 0.062 and 2.0 mm, a pebble is between 2.0 and 64 mm, and a cobble is between 64 and 256 mm. c Stage and hydraulic values for bankfull conditions. d Dimensionless channel slope values computed by using 10-m DEMs and high-resolution NHD in GIS. e Stage and hydraulic values for macrochannel conditions. f Values given for top of concrete bridge apron and not for morphologic bankfull stage. The top of the apron, however, is used to approximate bankfull stage.

8

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

Table 4 Flood magnitudes at various return periods for gauged and ungauged sites in the Llano River watershed, central Texas, based on partial-duration and regionally tuned regression analyses; for gauged locations, flood magnitudes do not exactly match discharges of bankfull and macrochannel conditions reported in Table 3 because the mean f value of 0.115 was applied to compute discharge and the average of multiple cross sections at each site was used.a Site

North Llano Draw near Sonorab North Llano River near Rooseveltb North Llano River near Junctionc South Llano River at Baker Ranch near Rockspringsb South Llano River at U.S. Highway 377 near Rockspringsb South Llano River at 700 Springs Ranch near Telegraphb South Llano River at South Llano River State Parkb South Llano River at Texas Tech University—Junctionb Llano River near Junctionc LCRA Johnson Fork near Junctionb Johnson Fork at Lowlands Crossing near Junctionb Llano River near Ivy Chapelb LCRA James River near Masonb James River near Masonb Llano River at James River Crossing near Masonb Llano River near Masonc Beaver Creek near Masonc Llano River at Castellb Llano River at Llanoc Llano River near Kingslandb Honey Creek at KDK Ranch near Kingslandb a b c

Return period 1-year (m3/s)

1.5-year (m3/s)

2-year (m3/s)

3-year (m3/s)

4-year (m3/s)

5-year (m3/s)

10-year (m3/s)

25-year (m3/s)

50-year (m3/s)

100-year (m3/s)

– – 171 – – – – – 242 – – – – – – 428 107 – 503 – –

– – 355 – – – – – 488 – – – – – – 793 173 – 928 – –

14 283 519 134 272 305 424 426 710 251 255 734 312 320 1047 1111 227 1221 1302 1450 75

– – 780 – – – – – 1076 – – – – – – 1635 314 – 1895 – –

– – 979 – – – – – 1369 – – – – – – 2057 383 – 2345 – –

30 638 1138 290 589 660 915 919 1613 563 572 1481 717 736 2160 2412 440 2492 2707 2925 194

45 970 1648 433 879 984 1363 1369 2459 854 867 2137 1099 1127 3145 3675 645 3609 3871 4211 315

66 1484 2329 651 1321 1478 2042 2051 3772 1302 1322 3105 1694 1738 4613 5738 984 5264 5456 6106 514

84 1929 2838 838 1698 1900 2623 2634 4914 1689 1715 3914 2212 2269 5850 7638 1302 6654 6664 7693 694

103 2426 3336 1045 2115 2367 3265 3278 6193 2120 2153 4796 2793 2865 7204 9879 1686 8171 7871 9418 902

–, not available. Flood magnitudes computed by regionally tuned regression analysis. Flood magnitudes computed by partial-duration analysis.

5. Channel classification and descriptions Channels are classified into four broad categories based on hydrology, planform morphology, boundary lithology, and alluvial development: (i) uppermost ephemeral reaches in the study area, also referred to as ephemeral draws; (ii) straight or sinuous gravel-bed channels in Cretaceous carbonate zones; (iii) straight or sinuous gravel-bed or bedrock channels in Paleozoic sedimentary zones; and (iv) straight, braided, or multithread mixed alluvial–bedrock channels with sandy beds in Precambrian igneous and metamorphic zones (Table 6, Figs. 5 and 6). Descriptions of these categories are presented subsequently in downstream order. Various morphometric and sedimentary characteristics for individual study sites are provided in Table 2. Site numbers correspond with those shown in Fig. 1. 5.1. Ephemeral draws Uppermost ephemeral draws of the North and South Llano rivers occur at higher elevations of the Edwards Plateau. Three study sites (1, 4, and 5) are representative of the ephemeral draw category. All three sites are characterized by flood flows contemporaneous with storm-derived runoff, and morphologic differences between the three sites can be attributed to increasing hydraulic energy associated with drainage area. Therefore, two subcategories of ephemeral draws are defined for this study: (i) ephemeral aggraded and (ii) ephemeral bedrock incised (Table 6, Fig. 5).

The channels of ephemeral aggraded draws are characterized by finegrained (silt and clay) boundaries that have subtle topographic transitions to surrounding valley fill deposits (Fig. 7A). During low flows, disconnected pools with varying proportions of subangular cobbles in the bed interrupt dry reaches and are readily distinguished from the surrounding valley fill. At ~ 20 km of downstream distance, ephemeral draws of the North and South Llano rivers develop alluvial valleys ranging from ~ 250 to 750 m wide, and a sinuous channel ranges from ~ 75 to 100 m wide. Fine-grained deposits along aggraded reaches of ephemeral draws are attributed to extensive upland soil erosion during the early and middle Holocene (Cooke et al., 2003) that filled valleys incised during the late Pleistocene (Blum et al., 1994; Mear, 1995). As drainage area and downstream distance increase, the ephemeral draws have enough unit stream power during runoff events to remove fine-grained sediment and incise the carbonate plateau, and bedrock exposures are numerous. Ephemeral draws incised in bedrock have relatively narrow valleys of ~50 to 250 m and narrow straight channels from ~25 to 75 m wide. They are characterized by step-like bedrock banks that descend to a bedrock bed with limited cobble- and pebble-sized deposits and form a transition from channels along the top of the Edwards Plateau to those with incised valleys. 5.2. Gravel-bed channels in the Cretaceous zone Ephemeral draws incised in bedrock transition to perennial, partially confined straight or sinuous channels near the confluence of large

Table 5 Values used in regression-based flood–frequency equations (Asquith and Thompson, 2008) based on a power transformation of drainage area and three predictors (drainage area, channel slope, and mean annual precipitation); channel slope values are compared to values listed in Asquith and Slade (1997) for cross validation. USGS station

Drainage area (km2)

Mean annual precipitation (mm)

Channel slope

Asquith and Slade (1997) channel slope

08148500 North Llano River near Junction, Tex. 08150000 Llano River near Junction, Tex. 08150700 Llano River near Mason, Tex. 08150800 Beaver Creek near Mason, Tex. 08151500 Llano River at Llano, Tex.

2335 4815 8418 558 10,885

635 635 711 711 737

0.0021 0.0011 0.0015 0.0034 0.0015

0.0022 0.0019 0.0017 0.0049 0.0016

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19 Table 6 Percentage of channel length classified into various geomorphologic categories, based on hydrology, planform morphology, lithology, and alluvial development. River

North Llano River

Channel Geomorphologic classification and length percentage of length (km) 95.6

South Llano 115 River

Llano River

187

Sum of channel segment lengths (km)

29.6% ephemeral aggraded 14.4% ephemeral bedrock incised 44.0% partially confined Cretaceous sinuous 12.0% partially confined Cretaceous straight 45.6% ephemeral aggraded 14.6% ephemeral bedrock incised 25.7% partially confined Cretaceous sinuous 14.1% partially confined Cretaceous straight 2.1% confined Cretaceous 18.3% partially confined Cretaceous sinuous 12.7% confined Paleozoic bedrock 4.4% partially confined Paleozoic sinuous 10.1% confined Paleozoic straight 10.9% Precambrian braided 13.3% Precambrian multithread mixed alluvial–bedrock 24.6% Precambrian straight 3.6% Lake LBJ

28.3 13.8 42.0 11.5 52.4 16.8 29.6 16.2 3.9 34.2 23.8 8.2 18.9 20.4 24.9 46.0 6.7

tributaries (e.g., Dry Llano River, Paint Creek) and springs along the North Llano and South Llano rivers (Fig. 1). Bankfull channel widths range from ~ 75 to 175 m and generally increase downstream. Eight study sites surveyed for this investigation (2, 3, 6, 7, 8, 9, 11, and 12) are representative of the Cretaceous zone category. The eight sites are characterized by mostly perennial flow and cobble- and pebble-sized bed material, and morphologic differences between them can be attributed to alluvial sinuosity and confinement by bedrock valley walls.

Three subcategories therefore are defined for this study: (i) partially confined straight, (ii) partially confined sinuous, and (iii) confined (Table 6, Fig. 5). Partially confined straight channels are characterized by an alluvial sinuosity b1.1 and have gravel or bedrock beds. Banks are either gradually sloping (≤5°), where coarse channel-lag deposits are dominant, or steep (≥ 15°) where overbank fine-grained material is dominant. At some sites, steep bedrock valley walls occur along one side of the channel. Alluvial valleys are relatively narrow (200 to 400 m) and are composed of channel-lag deposits capped with fine-grained overbank sediment. As the North and South Llano rivers approach Junction, drainage area exceeds 2000 km2 and the alluvial valley widens at some locations to N1.5 km. Sinuous channels occur within a Holocene floodplain surrounded by earlier terrace deposits, and frequently occurring longitudinal channel bars occur alongside the low-flow (thalweg) channel. Partially confined sinuous channels are characterized by alluvial sinuosity N 1.1, and banks are composed of various combinations of coarse channel-lag deposits, fine-grained overbank deposits, or sand (Fig. 7B–C). Channel banks gently slope on point bars of gradual meander bends or where the sand percentage is high or steeply slope at cutbanks of gradual meander bends or where the sand percentage is low. Downstream from Junction, the Llano River exhibits enlarged meander bends, similar to those documented along the nearby Pedernales River by Blum and Valastro (1989) who attribute such bends to relatively more humid conditions between 4500 and 1000 YBP. Confined channels in the Cretaceous zone occur along two reaches of the Llano River and correspond with the study sites near Junction and Ivy Chapel (Figs. 5 and 7D). The alluvial valley narrows at these two locations to b 500 m, and bedrock valley walls are composed of the Hensell Sand Member of the Glen Rose Formation of the Trinity Group. At the site near Junction, an avulsion through the bedrock during a previous high-magnitude flood results in a confined channel b100 m

100°W 31°N

9

99°W

Llano River watershed (11,568 km2)

Zone III Llano R

iver LLANO

iv e r North Lla no R

Zone II

JUNCTION

Zone I

30°N

no South Lla Ri v e r

30-m DEM background 0

10

20

30

40 KILOMETERS

GEOMORPHOLOGIC CLASSIFICATION

Zone I Ephemeral draws

Cretaceous gravel bed

Zone II Paleozoic gravel bed or bedrock

Zone III Precambrian sand bed Precambrian braided Precambrian multithread mixed alluvial-bedrock

Ephemeral aggraded

Partly confined Cretaceous straight

Partly confined Paleozoic sinuous

Ephemeral bedrock incised

Partly confined Cretaceous sinuous

Confined Paleozoic bedrock

Precambrian straight

Confined Cretaceous

Confined Paleozoic straight

Lake LBJ

Fig. 5. The North Llano, South Llano, and Llano rivers in central Texas, USA, traverse three geologic zones and are classified into four general geomorphologic categories: (i) uppermost ephemeral reaches, commonly referred to as ephemeral draws in the study area, (ii) Cretaceous straight or sinuous gravel-bed channels, (iii) Paleozoic straight or sinuous gravel-bed or bedrock channels, and (iv) Precambrian straight, braided, or multithread, mixed alluvial–bedrock channels with sandy beds. Straight or sinuous channels are determined by an alluvial sinuosity threshold of 1.1 irrespective of bends in the valley.

10

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

Fig. 6. Selected oblique aerial photographs from February 2008. (A) South Llano River near Junction, representative of a partly confined sinuous channel in the Cretaceous zone. (B) Llano River downstream from Junction, representative of a partly confined sinuous channel in the Cretaceous zone. (C) Llano River between Big Saline Creek and James River, representative of a laterally confined and vertically constrained bedrock channel in the Paleozoic zone. (D) Llano River upstream from Llano, representative of a multithread, mixed alluvial–bedrock channel in the Precambrian zone. (E) Llano River downstream from Llano, representative of a braided channel in the Precambrian zone.

wide. Bedrock also is exposed along the channel bed and banks, and alluvium is associated with narrow inset floodplains (Fig. 6C). 5.3. Bedrock and gravel-bed channels in the Paleozoic zone As the Llano River enters the Paleozoic sedimentary zone, the valley becomes more confined; and following a brief sinuous reach, the channel becomes straight. The alluvial sinuosity effectively is reduced to 1.0 because the channel boundary consists entirely of bedrock. Cobble- and pebble-sized bar deposits are less frequent and occur in hydraulically favorable areas protected from high-velocity flows (e.g., inside of very gradual bends) or where abrupt increases in channel width occur (e.g., confluence of Llano and James rivers). Bankfull width is ~ 150 m, and macrochannel width is ~250 m along the Llano River. Three study sites surveyed for this investigation (14, 15, and 27) are representative of the Paleozoic zone category. Most channel reaches are characterized by laterally confined and vertically constrained bedrock valleys and varying amounts of cobble- and pebble-sized bed material, and morphologic differences can be attributed to alluvial sinuosity and abundance of observed alluvial deposits. Three subcategories therefore are defined for the study: (i) partially confined sinuous, (ii) confined bedrock, and (iii) confined straight (Table 6, Fig. 5). Partially confined sinuous reaches are characterized by an alluvial sinuosity N1.1, and are associated with ~ 8 km of the Llano River

downstream from the Cretaceous–Paleozoic contact and upstream from an incision checkpoint. The valley width gradually decreases along this reach, and Pennsylvanian limestone and shale units are exposed at the surface. The sinuous reaches in this subcategory mark the downstream extent of enlarged meander bends (Blum and Valastro, 1989) observed along the Llano River in the Cretaceous zone. Confined bedrock reaches in the Paleozoic zone downstream from the incision checkpoint are remarkably straight and channel boundaries are composed of Ordovician limestone and dolostone bedrock. A few small inset floodplains and channel-bar deposits are observed along bedrock reaches, but overall the channel is hydraulically efficient and readily transports sediment to downstream reaches. Confined straight reaches begin where the Llano River exits the Ordovician bedrock and enters a zone of Cambrian siltstone, carbonate, and sandstone (Fig. 7E). After entry into the Cambrian lithologic ensemble, the alluvial valley widens to ~300 to 500 m. Cobble- and pebble-sized channel-bar deposits and fine-grained inset floodplains become more numerous, and alluvial sinuosity slightly increases but remains b 1.1. 5.4. Mixed alluvial–bedrock and sand-bed channels in the Precambrian zone Planform morphology of the Llano River becomes more complex upon entering the Precambrian igneous and metamorphic zone of the

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

11

Cretaceous carbonate zone 10

10

(A)

8 7 6 5

3

Bankfull stage (37.1 m /s) < 2-year return period

4

Cretaceous 3 Segovia Member of the 2 Edwards 1 Limestone 0

0

Fluvial terrace deposit

9 Height above gauge datum (m)

Arbitrary height (m)

9

50

Fine-grained alluvium

Buried gravel-sized channel-lag deposit? 100

150

200

250

300

350

400

8 7

Bankfull stage (536 m3/s)

6 5

~ 2-year return period

4

Cretaceous Hensell Sand

3 2

Gravel-bed material

1 0 0

50

100

150

200

250

300

350

400

Distance from left (m)

Distance from left (m) 12

10

(C)

8

10

Fluvial terrace deposit

6 5

Bankfull stage (351 m3/s) < 2-year return period

Arbitrary height (m)

7

Fluvial terrace deposit

4

Fine-grained alluvium

3 2 0 0

50

100

150

200

250

8 6 4 2

Gravel-sized channel lag deposit

1

Fine-grained alluvium

9

Arbitrary height (m)

(B)

Fine-grained alluvium

Macrochannel stage (2360 m3/s) ~ 12-year return period

(D) Cretaceous Hensell Sand (with conglomerate)

Gravel-bed material 0 300

350

0

400

50

100

Distance from left (m)

150

200

250

300

350

400

Distance from left (m)

Paleozoic sedimentary zone Arbitrary height (m)

25

(E)

20 15

Macrochannel stage (4130 m3/s) ~ 20-year return period Cambrian Fine-grained Riley alluvium Formation Bankfull stage (1180 m3/s) ~ 2.5-year return period 5 March 2007 (325 m3/s) Gravel and sand mid < 2-year return period channel bar

10

0 0

50

100

150

200

250

300

350

400

Distance from left (m)

20

10 9

(F)

15

Arbitrary height (m)

Height above gauge datum (m)

Precambrian igneous and metamorphic zone Macrochannel stage (6040 m3/s) ~ 40-year return period 10

Bankfull stage (970 m3/s)

5

Precambrian Packsaddle Schist

~ 1.5-year return period Sand-bar

7 6

Fine-grained alluvium Bankfull stage (778 m3/s) < 2-year return period

5 4 3

0

50

100

150

200

250

300

350

400

Distance from left (m)

0

June 2007 (3350 m3/s) ~ 6-year return period Precambrian Town Mountain Granite

March 2007 (502 m3/s) < 2-year return period

2 1

-5

(G)

NOTE: Survey did not reach top of macrochannel Macrochannel stage (3690 m3/s) ~ 8-year return period

8

Sand mid-channel bar 0

50

100

150

200

250

300

350

400

Distance from left (m)

Fig. 7. Boundary lithology, sedimentary composition, and flow stages of various return periods at selected cross sections including: (A) an ephemeral aggraded channel of the South Llano River at Baker Ranch near Rocksprings, Texas (cross section 2); (B) a partially confined sinuous channel of the North Llano River near Junction, Texas (cross section 2); (C) a partially confined sinuous channel of the South Llano River at Texas Tech University, Junction, Texas (cross section 2); (D) a confined macrochannel of the Llano River near Ivy Chapel, Texas (cross section 2); (E) a straight bankfull and macrochannel of the Llano River with a large mid-channel bar at James River Crossing near Mason, Texas (cross section 2); (F) a braided bankfull and macrochannel of the Llano River at Llano, Texas (cross section 2); and (G) a transitional multithread, mixed alluvial–bedrock and braided bankfull and macrochannel reach of the Llano River near Kingsland, Texas (cross section 2). Additional information about these sites is provided in Supplementary data 1, 2, 3, 4, 5, 6, and 7.

watershed. Similar to reaches in the Paleozoic zone, the valley is largely confined, and alluvial sinuosity is 1.0 at most locations. Bends in the river are inherited from preexisting variability of the exhumed crystalline rock, possibly related to relatively weak lithologic seams or lineaments (Shepherd, 1979). Additions of sand to the channel result in braided reaches, and numerous irregular outcrops along the bed result in multithread mixed alluvial–bedrock reaches. Bankfull width ranges from 125 to 250 m, and macrochannel width ranges from 250 to 450 m. Five study sites (17, 18, 20, 24, and 26) are representative of the category. Although most channel reaches are characterized by confined bedrock valleys and low alluvial sinuosity (b1.1) in the Precambrian zone, morphologic differences can be attributed to channel-bed

composition and the presence or absence of multithread, low-flow channels. Three subcategories therefore are defined for this study: (i) braided; (ii) multithread, mixed alluvial–bedrock; and (iii) straight (Table 6, Fig. 5). Braided reaches are characterized by a sand-bed, multithread, lowflow channel absent of Precambrian bedrock exposures along the channel bed (Fig. 7F). Braided reaches constitute about 20 km of the Llano River, and mostly these occur downstream from Llano because of more sand in the channel. Multithread, mixed alluvial–bedrock reaches are characterized by a multithread, low-flow channel with exposed, irregular exposures of Precambrian bedrock along the channel bed (Fig. 7G). These reaches

12

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

constitute about 25 km of the Llano River and are uniformly distributed along the channel in the Precambrian zone. Straight reaches are characterized by a sand-bed, single channel with fewer exposures of Precambrian bedrock along the channel bed. Straight reaches constitute about 46 km of the Llano River in the Precambrian zone, and mostly occur upstream from Llano because the quantity of sand is insufficient to create mid-channel bars and induce braiding. All three channel categories in the Precambrian zone have limited inset floodplains composed of sand, and channel banks commonly are exposed bedrock.

6. Results and discussion Channel geometry along main-stem river channels in the Llano River watershed (including the North Llano, South Llano, and Llano rivers) is influenced by downstream changes in boundary lithology and hydrology. Tributary sites serve to reinforce results. The following themes resulting from this study, which are discussed in detail, are (i) channel pattern adjustments related to abrupt transitions in lithology and associated composition of alluvial deposits, (ii) highly variable hydraulic geometry that necessitates site-averaged values of multiple cross sections

(B)

Width

50

100

150

PRECAMBRIAN

200

250

0.8 0.6

Depth

Width

0.0

0.0 0

PALEOZOIC

b (Rate of change of width) f (Rate of change of depth) m (Rate of change of velocity)

0.4

Macrochannel b, f, and m

0.8 0.6 0.4

Depth

CRETACEOUS

Macrochannel at−a−station b, f, and m exponents

1.0

b, f, and m

1.0

Major downstream adjustments of channel pattern occur at abrupt lithologic transitions, whereas sedimentary variability within those lithologic zones exerts a subsidiary influence on channel morphology. The initial headwater ephemeral draws of the North Llano and South Llano rivers take a sinuous path across the higher elevations of the Edwards Plateau before incising into the Cretaceous carbonate bedrock. The incision results in confined channels with low sinuosity. As drainage area and alluvial valley width gradually increase, the cobble- and pebblebed channels become laterally active (see Fig. 7C) within fine-grained banks composed mostly of silt and sand; and maximum alluvial sinuosity in the watershed occurs. Along most reaches, a low-flow channel (thalweg) can be distinguished from longitudinal channel-bar deposits. Downstream from the confluence of the North Llano and South Llano rivers, the Llano River is characterized by enlarged meander bends. Relatively wide alluvial valleys become confined by bedrock along two short reaches (3.9 km total) of the Llano River in the Cretaceous zone.

b (Rate of change of width) f (Rate of change of depth) m (Rate of change of velocity)

0.2

Bankfull b, f, and m

6.1. Channel pattern

0.2

(A)

to infer downstream trends in cross-sectional shape, and (iii) hydrologic controls of bankfull and macrochannel dimensions.

300

CRETACEOUS

0

50

Downstream distance (km)

(D)

0.0

CRETACEOUS

100

1.0 PALEOZOIC

150

250

300

Macrochannel at−a−station b, f, and m exponents

PRECAMBRIAN

200

Downstream distance (km)

250

0.6

0.8

b (Rate of change of width) f (Rate of change of depth) m (Rate of change of velocity)

0.4

Depth (r2=0.00)

Width (r2=0.00) Velocity (r 2=0.00)

0.0

Velocity (r2=0.04)

Macrochannel b, f, and m

Width (r 2=0.06)

50

200

0.2

0.6 0.4

2 Depth (r =0.06)

0.2

Bankfull b, f, and m

0.8

b (Rate of change of width) f (Rate of change of depth) m (Rate of change of velocity)

0

150

PRECAMBRIAN

Downstream distance (km)

Bankfull at−a−station b, f, and m exponents

1.0

(C)

100

PALEOZOIC

300

CRETACEOUS

0

50

100

PALEOZOIC

150

PRECAMBRIAN

200

250

300

Downstream distance (km)

Fig. 8. At-a-station hydraulic geometry b (top width), f (mean depth), and m (mean velocity) values along main-stem channels in the Llano River watershed in central Texas, USA, are displayed in downstream order for (A) individually surveyed cross sections at bankfull stage, (B) individually surveyed cross sections at macrochannel stage, (C) site-averaged bankfull stage, and (D) site-averaged macrochannel stage. Regression lines are displayed for reference and are not statistically significant.

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

Median bank material size (mm)

Channel slope (m/m)

Alluvial sinuosity

Bankfull channels (including tributaries) b 45 0.304 0.988 f 45 0.287 0.992 m 45 0.298 0.923

0.667 0.669 0.692

0.279 0.267 0.323

0.020b 0.018b 0.022b

Bankfull channels (only main-stem channels) b 37 0.194 0.970 f 37 0.174 0.954 m 37 0.212 0.985

0.625 0.614 0.613

0.583 0.546 0.673

0.071 0.061 0.082

Macrochannels (including tributaries) b 25 0.123 0.073 f 25 0.118 0.081 m 25 0.140 0.086

0.119 0.125 0.107

0.124 0.117 0.144

0.617 0.642 0.614

Macrochannels (only main-stem channels) b 19 0.926 0.494 f 19 0.953 0.514 m 19 0.895 0.543

0.086 0.091 0.078

0.703 0.724 0.704

0.790 0.768 0.806

n

Downstream distance (km)

Median bed material size (mm)

a n, number of cross sections; b, at-a-station hydraulic geometry exponent for rate of change of bankfull width; f, at-a-station hydraulic geometry exponent for rate of change of bankfull mean depth; m, at-a-station hydraulic geometry exponent for rate of change of bankfull mean velocity. b Statistically significant correlation (p-value b 0.05).

whereas increases in width become more important along partly confined reaches. As a result of considerable variability in hydraulic geometry exponent values, site-averaged b, f, and m values (Fig. 8C–D) were computed from multiple cross sections. Hydraulic variability along channel reaches is commonly minimized by averaging parameters from multiple cross sections (e.g., slope-area discharge computations) (Dalrymple and Benson, 1968). Furthermore, substantial variability in channel shape parameters is commonly reported for areas of relatively homogeneous controls in contrast to the Llano River watershed (e.g., Phillips, 1990; Fonstad and Marcus, 2003). Thus, discussion below focuses on site-averaged values (Tables 8 and 9) to infer possible controls on channel geometry.

1.0

Bankfull at−a−station b and f exponents

0.6

0.8

b (Rate of change of width) f (Rate of change of depth)

)

Width

0.4

Bankfull b and f

At-a-station hydraulic geometry exponent values (b, f, and m) for all surveyed cross sections are highly variable (Fig. 8A–B) and no statistically significant trends are associated with downstream distance (km) (Table 7). Furthermore, bankfull and macrochannel hydraulic geometry exponents display no statistically significant relations with bed material particle size (mm), bank material particle size (mm), and channel slope. Exclusion of tributary sites does not affect the lack of statistical significance among the variables. The only statistically significant relations (p-value b 0.05) are for bankfull hydraulic geometry exponents (b, f, and m) and alluvial sinuosity (Table 7, Fig. 9); this relation is not statistically significant for macrochannels. These findings collectively reinforce an interpretation that lithology is a primary control of channel shape. Alluvial sinuosity essentially quantifies the degree of bedrock control in the Llano River watershed. Bankfull at-a-station hydraulic geometry data for individual cross sections (i.e., not site-averaged) were categorized into (i) confined or (ii) partly confined boundary conditions, irrespective of their broader channel classification (Table 6, Figs. 5 and 6). The mean b (width) values for confined and partly confined cross sections are 0.33 and 0.42, respectively. The mean f (mean depth) values for confined and partly confined cross sections are 0.45 and 0.38, respectively. Two-sample t-tests indicate statistically significant differences in mean b (p-value = 0.043) and f (p-value = 0.039) exponents between the two categories. These statistically significant differences coupled with the statistically significant correlation of bankfull hydraulic geometry with alluvial sinuosity indicate that valley confinement by bedrock exerts a primary influence on cross-sectional shape in the Llano River watershed. Specifically, mean depth of bankfull channels increases at a greater rate than width along confined reaches,

Hydraulic geometry exponent

.12 (r2 = 0

Depth (r 2 = 0.12)

0.2

6.2. Hydraulic geometry

Table 7 Pearson correlation p-values of at-a-station hydraulic geometry exponents (b, f, and m) and possible explanatory variables at individual cross sections in the Llano River watershed; sediment particle size values were separately computed for individual cross sections; other variables (downstream distance, channel slope, and alluvial sinuosity) were computed for each site (i.e., not for individual cross sections).a

0.0

Although a short reach of the Llano River (8.2 km) retains a sinuous course in the uppermost Paleozoic zone, the channel immediately straightens at an erosion-resistant incision checkpoint within a confined valley composed of Ordovician limestone and dolostone. A narrow alluvial valley forms downstream from the Ordovician–Cambrian contact, but confinement between the bedrock valley walls maintains an alluvial sinuosity b1.1. Cobble- and pebble-sized channel bars are more frequent and contribute to slight variations in the straight channel, such as the large mid-channel bar at the Llano River at James River Crossing near Mason (see Fig. 7E). Channel pattern is complex in the Precambrian zone of the watershed, but all reaches are confined within the exhumed igneous and metamorphic bedrock. Similar to the Paleozoic sedimentary zone, observed bends in the channel are likely associated with lithologic contacts or preferential zones of bedrock weathering such as lineaments (Shepherd, 1979). Straight channel reaches devoid of exposed midchannel bars and irregular bedrock exposures characterize most of the Llano River in the Precambrian zone upstream from Llano. As introduction of sand-sized sediment increases (Shepherd, 1975), the single straight channel transitions to a braided sand-bed channel downstream from Llano. Along some reaches upstream and downstream from Llano, irregular exposures of granite, schist, and gneiss force a multithread, mixed alluvial–bedrock channel pattern characterized by a network of low-flow channels with high bifurcation angles (many ~ 90° or more in fractured bedrock) (e.g., Meshkova and Carling, 2013). These lowflow channels are usually separated by b25 m of cross section distance and rejoin after 10 to 100 m of downstream length. The network does not conform to the anabranching classification of Nanson and Knighton (1996). Furthermore, these multithread reaches are less extensive than the bedrock-anastomosing reaches of the Sabie River in South Africa (Heritage et al., 1999; van Niekerk et al., 1999), mixed alluvial–bedrock anastomosed channels of the Mekong River (Meshkova and Carling, 2012), or the multichannel Narmada River in India (Kale et al., 1996).

13

1.0

1.1

1.2

1.3

1.4

1.5

Alluvial sinuosity Fig. 9. At-a-station hydraulic geometry b (top width) and f (mean depth) values for individually surveyed cross sections at bankfull stage in the Llano River watershed in central Texas, USA. Although r2 values are low, correlations are statistically significant at the 5% confidence interval.

14

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

Table 8 Selected hydraulic characteristics for bankfull stage at study sites in the Llano River watershed in central Texas; various sites near LCRA stations are not included if cross-sectional morphology measurements contain bridge-construction features, including concrete aprons and artificial embankments; values represent averages of multiple cross sections at each site.a Site

BF Q (m3/s)

BF Ω (W/m)

BF ω (W/m2)

BF w (m)

BF d (m)

BF U (m/s)

b

f

m

North Llano Draw near Sonorab North Llano River near Rooseveltc North Llano River near Junctionc South Llano River at Baker Ranch near Rockspringsc South Llano River at U.S. Highway 377 near Rockspringsc South Llano River at 700 Springs Ranch near Telegraphc South Llano River at South Llano River State Parkc South Llano River at Texas Tech University—Junctionc LCRA Johnson Fork near Junctionc Johnson Fork at Lowlands Crossing near Junctionc LCRA James River near Masonc Llano River at James River Crossing near Masonc Llano River near Masonc Beaver Creek near Masonc Llano River at Castellc Llano River at Llanoc Llano River near Kingslandc Honey Creek at KDK Ranch near Kingslandd

– 494 536 37.1 157 556 382 351 606 350 371 1180 594 193 906 970 778 25.3

– 10,200 15,200 545 5520 16,300 4490 4820 15,400 18,200 12,300 28,900 14,600 5850 11,500 13,300 20,600 5530

– 142 142 6.47 175 139 25.7 31.5 167 180 119 191 127 89.6 82.1 80.6 89.6 205

– 71.8 107 84.2 31.6 117 175 153 92.1 101 103 151 115 65.3 140 165 230 27.0

– 3.05 2.44 0.62 2.01 2.21 1.99 1.79 2.90 1.49 1.78 3.22 2.50 1.57 3.57 3.31 1.83 0.73

– 2.06 2.19 0.78 2.20 2.12 1.25 1.31 2.27 2.32 2.03 2.33 2.05 1.82 1.78 1.78 1.84 1.30

– 0.42 0.47 0.48 0.22 0.34 0.47 0.38 0.34 0.46 0.30 0.32 0.27 0.38 0.20 0.36 0.46 0.27

– 0.39 0.36 0.34 0.52 0.44 0.35 0.41 0.44 0.36 0.47 0.45 0.48 0.41 0.53 0.43 0.36 0.49

– 0.19 0.18 0.18 0.26 0.22 0.21 0.21 0.22 0.18 0.23 0.23 0.24 0.21 0.27 0.21 0.18 0.24

a BF Q, bankfull discharge; BF Ω, bankfull stream power; BF ω, bankfull unit stream power; BF w, bankfull width; BF d, bankfull mean depth; BF U, bankfull mean velocity; b, at-a-station hydraulic geometry exponent for rate of change of bankfull width; f, at-a-station hydraulic geometry exponent for rate of change of bankfull mean depth; m, at-a-station hydraulic geometry exponent for rate of change of bankfull mean velocity; –, not available. b No data provided for North Llano draw near Sonora because a definable channel is not present. c Bankfull discharge, stream power, mean velocity, and hydraulic geometry exponents computed from hydraulic analyses using a Darcy–Weisbach f factor of 0.115. d Bankfull discharge, stream power, mean velocity, and hydraulic geometry exponents computed from hydraulic analyses using a Darcy–Weisbach f factor of 0.750.

6.3. Bankfull and macrochannel shape Reaches along the North Llano and South Llano rivers generally have a distinct scarp that facilitates identification of bankfull stage. For the South Llano River at Baker Ranch, an ephemeral draw, the greater rate of increase in width (b = 0.48) relative to mean depth ( f = 0.34) (Table 10) compensates for increasing discharge. Proceeding downstream and contrasting with aggraded ephemeral draws, the South Llano River at U.S. Highway 377 near Rocksprings is incised into bedrock and has a greater rate of increase in mean depth (f = 0.52) relative to width (b = 0.22) as discharge increases (Table 10). Partially confined, gravel-bed channels in the Cretaceous zone of the watershed have variable relations of bankfull width to depth, but mean at-a-station hydraulic geometry exponents indicate that a slightly greater rate of increase in width (b = 0.42) relative to mean depth ( f = 0.39) compensates for increasing discharge (Table 10). Bankfull flows along ephemeral draws have return periods b 2 years, but occur less frequently (~ 4-year return period) along partially

confined channels just downstream from bedrock-incised reaches (e.g., North Llano River near Roosevelt, South Llano River at 700 Springs Ranch near Telegraph) (Tables 4 and 8). The increase in bankfull capacity is associated with a combination of relatively steep channel slopes resulting in powerful flows (total stream power N10,000 watts per meter (W/m)) and increased valley dimensions for channel enlargement. Proceeding downstream, bankfull conditions of sinuous gravelbed channels near Junction have return periods of ~ 2 years, and the Johnson Fork tributary exhibits bankfull conditions every 3 to 5 years, on average (Tables 4 and 8). Downstream from the confluence of the North Llano and South Llano rivers, the Llano River exhibits macrochannel morphology along confined reaches, which demonstrates the sensitivity of channel dimensions to valley confinement (Magilligan, 1992; Fuller, 2007). The macrochannel extent is identified by distinct breaks in slope along the channel banks, regardless of whether they are composed of bedrock or alluvium. Evidence of a lower bankfull stage along these reaches is minor or absent. The development of macrochannel morphology is

Table 9 Selected hydraulic characteristics for macrochannels at study sites in the Llano River watershed in central Texas; values represent averages of multiple cross sections at each site, if more than one cross section was available for the macrochannel; values are not comparable with one another because they represent individual heights above bankfull stage, and do not necessarily extend to the top of the macrochannel.a Site

Height above BF (m) b

Llano River near Junction Llano River near Ivy Chapelb LCRA James River near Masonb James River near Masonb Llano River at James River Crossing near Masonb Llano River near Masonb Beaver Creek near Masonb Llano River at Castellb Llano River at Llanob Llano River near Kingslandb Honey Creek at KDK Ranch near Kingslande

c

– –c 2.7c –c 3.9c 7.1c 3.4d 5.0d 4.9c 3.6d 2.5c

Q (m3/s)

Ω (W/m)

ω (W/m2)

w (m)

d (m)

U (m/s)

b

f

m

2400 2360 1520 5410 4130 4570 947 3720 6040 3690 286

167,000 34,700 50,600 313,000 101,000 112,000 28,700 47,300 82,900 97,600 62,500

1800 208 386 1750 409 463 293 190 287 330 1280

93.0 167 131 179 247 242 97.9 249 289 296 48.9

5.33 5.75 3.88 6.04 5.48 5.91 3.54 6.31 7.70 4.39 2.47

5.07 2.43 3.00 4.92 3.06 3.17 2.74 2.37 2.71 2.84 2.40

0.23 0.35 0.16 0.18 0.24 0.40 0.26 0.43 0.32 0.20 0.21

0.51 0.43 0.56 0.55 0.50 0.40 0.49 0.38 0.45 0.53 0.53

0.26 0.22 0.28 0.27 0.25 0.20 0.25 0.19 0.23 0.27 0.26

a BF, bankfull stage; Q, discharge; Ω, stream power; ω, macrochannel unit stream power; w, width; d, mean depth; U, mean velocity; b, at-a-station hydraulic geometry exponent for rate of change of width above bankfull stage; f, at-a-station hydraulic geometry exponent for rate of change of mean depth above bankfull stage; m, at-a-station hydraulic geometry exponent for rate of change of mean velocity above bankfull stage; –, no morphologic indicators of bankfull stage. b Discharge, stream power, mean velocity, and hydraulic geometry exponents computed from hydraulic analyses using a Darcy–Weisbach f factor of 0.115. c Top of macrochannel. d Below top of macrochannel. e Discharge, stream power, mean velocity, and hydraulic geometry exponents computed from hydraulic analyses using a Darcy–Weisbach f factor of 0.750.

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19 Table 10 Mean values of at-a-station hydraulic geometry for sites representative of the four categories of channel classification along the North Llano, South Llano, and Llano rivers in central Texas; mean values of tributaries also are provided for comparison.a Channel classification category Main-stem channels Ephemeral aggraded (bankfull)b Ephemeral bedrock incised (bankfull)c Cretaceous gravel-bed (bankfull) Cretaceous gravel-bed (macrochannel) Paleozoic bedrock or gravel-bed (bankfull)d Paleozoic bedrock or gravel-bed (macrochannel)d Precambrian straight, braided, or multithread mixed alluvial–bedrock (bankfull) Precambrian straight, braided, or multithread mixed alluvial–bedrock (macrochannel) Tributaries Cretaceous gravel-bed tributary (bankfull) Paleozoic bedrock or gravel-bed tributary (bankfull) Paleozoic bedrock or gravel-bed tributary (macrochannel) Precambrian straight, braided, or multithread mixed alluvial–bedrock tributary (bankfull) Precambrian straight, braided, or multithread mixed alluvial–bedrock tributary (macrochannel)

b

f

m

0.48 0.22 0.42 0.29 0.32 0.24 0.32

0.34 0.52 0.39 0.47 0.45 0.50 0.45

0.18 0.26 0.20 0.24 0.23 0.25 0.23

0.34

0.44

0.22

0.40 0.28 0.18 0.38

0.40 0.48 0.55 0.41

0.20 0.24 0.27 0.21

0.26

0.49

0.25

a b, at-a-station hydraulic geometry exponent for rate of change of width; f, at-a-station hydraulic geometry exponent for rate of change of mean depth; m, at-a-station hydraulic geometry exponent for rate of change of mean velocity. b Values only for South Llano River at Baker Ranch near Rocksprings. A bankfull channel condition could not be established at North Llano Draw near Sonora. c The South Llano River at U.S. Highway 377 near Rocksprings is considered separately from ephemeral aggraded channels because of the disparity of bankfull hydraulic geometry exponents. d Values only for Llano River at James River Crossing near Mason. This is the only site representative of main-stem channels in the Paleozoic bedrock or gravel-bed category.

attributed to infrequent high-magnitude flows N 2000 m3/s (Fig. 10B), total stream power exceeding 30,000 W/m, and unit stream power exceeding 200 W/m2 (Table 9, Fig. 10D). Macrochannel unit stream power for the Llano River in the Cretaceous zone is greater than values reported for the gently sloped Mekong River in Cambodia (Meshkova et al., 2014), comparable to the Auranga River (a smaller seasonal fluvial system in eastern India), and is considerably less than reported for the larger Narmada River in western India (Gupta, 1995). Confined macrochannels in the Cretaceous zone exhibit a greater rate of increase in mean depth ( f = 0.47) relative to width (b = 0.29) (Table 10) to compensate for increasing discharge, a notable shift from bankfull reaches located upstream. Macrochannel flows along confined reaches in the Cretaceous zone have return periods of ~ 10 to 12 years (Tables 4 and 9), alluding to the importance of high-magnitude flows in shaping channel morphology along the Llano River. In summary, macrochannel development is a mode of channel adjustment whereby unit stream power in excess of bankfull flow conditions is sufficient to erode a large channel along confined reaches rather than accrete what would otherwise become a depositional floodplain. Unit stream power necessary for macrochannel development is not attributed to downstream inflections in channel slope (e.g., Lecce, 1997; Knighton, 1999; Reinfelds et al., 2004) along the Llano River, but rather an overall increase in discharge with drainage area. Downstream from the Ordovician bedrock-lined reaches of the Llano River, the channel emerges into a narrow valley with distinctive inset floodplains abutting against Cambrian sedimentary strata (Barnes, 1981). Although distinct breaks in slope along banks identify macrochannel conditions, subtle forms of morphologic evidence can be used to infer lower bankfull stages. For the Llano River at James River Crossing, bankfull conditions are associated with the top of a large mid-channel bar, and macrochannel conditions are associated with the height of inset floodplains at the top of the bank. For bankfull conditions along the Llano River at James River Crossing, the greater rate of increase in mean depth ( f = 0.45) relative to width (b = 0.32) (Table 10) compensates for increasing discharge. For macrochannel

15

conditions above bankfull stage, an even greater rate of increase in mean depth ( f = 0.50) relative to width (b = 0.24) (Table 10) compensates for increasing discharge. Three sites along the James River and Honey Creek, tributaries in the Paleozoic zone, also are characterized by a greater rate of increase in bankfull mean depth ( f = 0.48) relative to width (b = 0.28) (Table 10). Similar to the main-stem Llano River, tributary macrochannels have a greater rate of increase in mean depth ( f = 0.55) relative to width (b = 0.18) (Table 10) above bankfull stages. In comparison to upstream reaches along the North Llano and South Llano rivers, the more prominent increase of mean depth for bankfull and macrochannel conditions along Paleozoic reaches is explained by increasingly confined valleys and limitations for lateral channel adjustment. Bankfull flows at the Llano River at James River Crossing have a return period of ~2.5 years (Tables 4 and 8), and macrochannel conditions occur with a return period of ~20 years (Tables 4 and 9), introducing a dichotomy of relatively frequent flows that maintain alluvial indicators of bankfull stage and less frequent flows that maintain macrochannel dimensions. Farther downstream, channel reaches in the Precambrian zone of the watershed are confined by variably resistant granite, gneiss, and schist. Similar to reaches in the Paleozoic zone, bankfull conditions are identified by subtle breaks in the slope along sandy alluvial banks, occurring no higher than ~5 m above the thalweg. At bankfull conditions, sandbed and mixed alluvial–bedrock channels of the Llano River in the Precambrian zone have a greater rate of increase in mean depth ( f = 0.45) relative to width (b = 0.32) (Table 10). For macrochannel conditions above bankfull stage, mean depth ( f = 0.44) also increases at a greater rate than width (b = 0.34) (Table 10). The minimal difference between at-a-station exponents of bankfull channels and macrochannels in the Precambrian zone of the watershed indicates that macrochannel adjustment is different from that observed for Cretaceous and Paleozoic reaches. The resistant lithology of the igneous and metamorphic rock in this zone limits lateral development of inset floodplains. The bankattached alluvial deposits observed at study sites along Precambrian reaches do not have any sharp demarcation identifying an inset floodplain surface, possibly resulting from minimal cohesion of the sandsized material. Therefore, the slope of sub-bankfull, bank-attached deposits closely follows that of higher alluvial deposits and the surrounding bedrock. Bankfull flows along Precambrian reaches have a return period of ~1 to 1.5 years (Tables 4 and 8) and macrochannel conditions have return periods of ~20 to 40 years (Tables 4 and 9), reinforcing the dichotomy of relatively frequent flows that maintain alluvial indicators of bankfull stage and less frequent flows that maintain macrochannel morphology. For at-a-station hydraulic geometry at bankfull conditions along the main-stem channels (Table 10), a gradual downstream increase in f and decrease in b indicate that mean depth increasingly compensates for discharge relative to width. For macrochannel conditions, mean depth compensates for discharge more than width; and width–depth relations are similar from the upper to lower watershed. The similarity of bankfull width–depth and macrochannel width–depth relations in the Precambrian zone of the watershed reiterates that breaks in slope at the bankfull stage are subtle. Downstream hydraulic geometry plots (Fig. 10E–F) of bankfull and macrochannel conditions do not provide a complete explanation for downstream patterns of channel adjustment because bankfull discharge did not necessarily increase in a downstream direction. Aside from this caveat, bankfull channel morphology of the North Llano, South Llano, and Llano rivers has a greater rate of increase in mean depth ( f = 0.44) relative to width (b = 0.31) (Fig. 10E). This finding stands in contrast to the importance of width relative to depth for single-thread, alluvial, gravel-bed rivers analyzed by Parker et al. (2007), which serves to reinforce the conclusion that bedrock-controlled rivers are less sensitive to hydraulic (e.g., shear stress) and sedimentary controls (e.g., bedsediment resistance) compared to adjustable alluvial rivers. The higher f value, however, is consistent with the findings of Pitlick and Cress

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

(A)

(C)

(E)

Bankfull discharge

Bankfull unit stream power (ω)

1200

Bankfull unit stream power (W/m2)

1000

800

600

40

20

150

100

50

PALEOZOIC

PALEOZOIC

0 0

50

100

0

PRECAMBRIAN

CRETACEOUS

150

200

250

CRETACEOUS

0

300

50

Downstream distance (km)

PRECAMBRIAN

150

200

250

300

2000

Macrochannel unit stream power (W/m2)

PRECAMBRIAN PALEOZOIC

8000

6000

4000

2000

CRETACEOUS

PALEOZOIC

PRECAMBRIAN

1500

1000

500

200 0

50

100

150

200

Downstream distance (km)

100

Width

250

300

1)

(r2 =0.3

10

(r2 =0.50) 0) 2 =0.8 Velocity th (r Dep

1

0.1 10

50

100

500 1000

5000

(F) Macrochannel unit stream power (ω)

CRETACEOUS

Width (b=0.31) Depth (f=0.44) Velocity (m=0.24)

Bankfull discharge (m3/s)

(D) Macrochannel discharge

10,000

Bankfull downstream hydraulic geometry 1000

Downstream distance (km)

(B) Macrochannel discharge (Qmc) (m3/s)

100

0

50

100

150

200

Downstream distance (km)

250

300

Width (m), mean depth (m), and mean velocity (m/s)

Bankfull discharge (Qbf) (m3/s)

200

Width (m), mean depth (m), and mean velocity (m/s)

16

Macrochannel downstream hydraulic geometry 1000

Width (b=0.58) Depth (f=0.26) Velocity (m=0.17) Outlier not used in analysis

500

2 Width (r =0.98)

100 50

10 2 .43) Depth (r =0

5

2 Velocity (r =0.31)

1 1000

2000

5000

10,000

Macrochannel discharge (m3/s)

Fig. 10. Hydraulic variables computed for main-stem channels in the Llano River watershed in central Texas, USA, including the North Llano, South Llano, and Llano rivers. (A) Bankfull discharge with downstream distance. (B) Macrochannel discharge with downstream distance. For macrochannel discharge, sites were excluded if the topographic survey did not reach the top of the macrochannel. (C) Bankfull unit stream power with downstream distance. Bankfull stream power generally increases downstream and is b200 W/m2. Considerable variability is attributed to local differences in channel slope, which is an influential variable in computation of stream power. (D) Macrochannel stream power with downstream distance. Macrochannel stream power ranges from ~200 to 1800 W/m2 with most values b500 W/m2. Sites were excluded if the topographic survey did not reach the top of the macrochannel. (E) Downstream hydraulic geometry for bankfull conditions. (F) Downstream hydraulic geometry for macrochannel conditions. One outlier, Llano River near Junction, was not included in macrochannel analyses because the locally steep slope associated with a cascade at the site resulted in an anomalously high mean velocity. For downstream hydraulic geometry plots (E and F), it should not be assumed that the left-to-right progression of points represents sites in order from upstream to downstream because macrochannel discharge did not always increase in a downstream direction.

(2002) along alluvial reaches of the Colorado River, which is interpreted to result from excessive shear stress capable of transporting bed sediment. Again, however, channel shape along the bedrock-controlled Llano River is relatively insensitive to adjustment and bed material mobilization is a subsidiary control of channel geometry. Unlike bankfull channels, macrochannels have a greater rate of increase in width (b = 0.58) relative to mean depth ( f = 0.26) (Fig. 10F). To summarize, channel dimensions that convey flows with return periods typically b2 years in the Llano River watershed become relatively narrow and deep in a downstream direction. The width of macrochannels above the bankfull stage, however, increases downstream at a greater rate than mean depth, although not enough to result in a detectable shift of at-a-station width compensating for discharge more than mean depth. 7. Conclusions Three orders of controls on channel geometry of the mixed alluvial– bedrock Llano River are postulated: (i) first-order lithologic resistance and valley confinement, (ii) second-order hydrologic (flow regime), and (iii) third-order composition of alluvial deposits. Lithologic resistance is directly associated with valley confinement and determines the longitudinal profile, planform geometry, and

cross-sectional channel shapes; a conclusion consistent with that of Shepherd (1979) in the central Texas region. The consistent longitudinal profile results from abrupt downstream increases in bedrock resistance to channel incision, notably at an incision checkpoint separating confined, straight channels downstream from partly confined, sinuous channels with floodplains upstream. Lithology further explains multithread, mixed alluvial–bedrock reaches in the Precambrian zone. Additionally, channel shape is controlled by lithology because mean depth is dominant relative to width at sites with substantial bedrock confinement, which is reinforced by a general downstream increase in f exponents of bankfull channels relative to b exponents. The second-order control of channel configuration is the flow regime, which can be separated into three functional categories: (i) bedmaterial entrainment, (ii) bankfull, and (iii) macrochannel. First, flows that mobilize channel bed material occur relatively frequently (b1.5 years) in the watershed, indicating that sub-bankfull events can effectively transport sediment supplied to the fluvial system and rework in-channel deposits that serve as temporary storage. Second, alluvial features and morphologic indicators of bankfull stage, including large channel bars and banks, are associated with return periods of ~ 1.5 to 3 years. Third, macrochannel dimensions are maintained by less frequent, high-magnitude flows with return periods that generally increase downstream from ~10 to 40 years.

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

Macrochannel conditions emerge downstream from the confluence of the North Llano and South Llano rivers and are detected by relatively high inset floodplains or definitive breaks in slope to the upland landscape. Macrochannels occur when unit stream power exceeds 200 W/m2. Ata-station hydraulic geometry of macrochannels indicates that mean depth compensates for increasing discharge at a greater rate than width. A downstream convergence of width–depth hydraulic geometry relations for bankfull and macrochannel flows indicates that the transverse slope of sub-bankfull, bank-attached deposits closely follows that of higher alluvial deposits and the surrounding bedrock, especially in the Precambrian zone. Again, valley confinement is important because inset floodplain development is limited in the Precambrian zone; and combined with minimal cohesion of sand-sized material (a third-order control), breaks in slope at the bankfull stage are subtle. The dichotomy of bankfull channels and macrochannels in the Llano River watershed underscores the difficulties in associating channel morphology with one dominant discharge, especially in fluvial systems with highly variable flow regimes. Although the oft-cited average return period range of 1 to 2 years is supported for bankfull channels, many representative reaches are best described as macrochannels with less frequent formative flows. For much of the Llano River, the work accomplished by severe floods in association with valley confinement obscures the cumulative imprint of more frequent, moderate floods on observed channel morphology. Moderate floods, however, are important to redistribute sediment from low-lying alluvial benches, channel bars, and the channel bed. Various practitioners, including aquatic biologists and managers interested in sediment-transport dynamics, are likely to have considerable interest in the more frequent bedentrainment and bankfull flow events. Finally, the third-order control on channel geometry is sedimentary composition, which is associated with (i) channel shape along partly confined reaches, (ii) braided channel reaches, and (iii) minor bankfull indicators in the Precambrian zone. Although silt-clay content of banks is highest along channels in the Cretaceous zone, bankfull shape is characterized by relatively high b (width) exponents. Relatively broad alluvial valleys in the Cretaceous zone of the watershed facilitate lateral channel migration that, when coupled with the abundance of cobblesized bed material, results in wide channels. Therefore, sedimentary composition is a subsidiary control of channel shape relative to valley confinement. Farther downstream, inputs of sand in the Precambrian zone lead to the development of braided channel reaches. Additionally, the infusion of noncohesive sand along channel margins, coupled with an increasingly powerful flood regime, results in a downstream obscurity of bankfull indicators and a convergence of width–depth relations for bankfull and macrochannels. Acknowledgments This research would not have been possible without support from the following agencies and contributors: National Science Foundation (DDRI Award 0623230), USGS, Texas Department of Transportation (Project 0-4695), LCRA, Braun & Associates and UT Environmental Science Institute Grant for Research on Private Lands, Robert E. Veselka Endowed Fellowship for Graduate Research Travel — and in memory of Stephen T. Moore, and the David Bruton, Jr., Graduate Fellowship. The authors also express gratitude to the following individuals for their assistance and encouragement: Karl W. Butzer, Francisco L. Pérez, John M. Sharp, Kenneth R. Young, Clay Spivey, George R. Herrmann, David B. Thompson, Theodore G. Cleveland, Xing Fang, K.H. Wang, Meghan C. Roussel, Lance Christian, Ben Connor, Delbert Humberson, Joe Beauchamp, Erin Atkinson, and Terry Hollister. Reviews of various stages of this manuscript were made by Victor R. Baker, Paul M. Bradley, Johnathan R. Bumgarner, Paul A. Carling, John D. Gordon, Keith J. Lucey, John Pitlick, Karl E. Winters (who doubly served as pilot during a lowaltitude aerial flight), and other anonymous peers who substantially improved the manuscript. Gratitude is expressed to these reviewers and

17

Associate Editor Richard Marston for their helpful comments and suggestions. Additional gratitude is expressed to the landowners who hospitably permitted access to their properties. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.geomorph.2014.12.033.

References Andrews, E.D., 1980. Effective and bankfull discharges of streams in the Yampa River basin, Colorado and Wyoming. J. Hydrol. 46 (3–4), 311–330. http://dx.doi.org/10. 1016/0022-1694(80)90084-0. Arp, C.D., Schmidt, J.C., Baker, M.A., Myers, A.K., 2007. Stream geomorphology in a mountain lake district: hydraulic geometry, sediment sources and sinks, and downstream lake effects. Earth Surf. Process. Landf. 32 (4), 525–543. http://dx.doi.org/10.1002/ esp.1421. Asquith, W.H., 1998. Depth–duration frequency for precipitation in Texas. U.S. Geological Survey Water-Resources Investigations Report 98-4044, Austin, TX. Asquith, W.H., 2011. Distributional Analysis with L-moment Statistics Using the R Environment for Statistical Computing. CreateSpace Independent Publishing Platform. Asquith, W.H., 2013. lmomco: L-moments, Trimmed L-moments, L-comoments, Censored L-moments, and Many Distributions. R Package Version 1.7.7. http://www.cran.rproject.org/package=lmomco. Asquith, W.H., Roussel, M.C., 2004. Atlas of depth–duration frequency of precipitation annual maxima for Texas. U.S. Geological Survey Scientific Investigations Report 2004-5041 (Austin, TX). Asquith, W.H., Slade, R.M., 1997. Regional equations for estimation of peak-streamflow frequency for natural basins in Texas. U.S. Geological Survey Water-Resources Investigations Report 96-4307, Austin, TX. Asquith, W.H., Thompson, D.B., 2008. Alternative regression equations for estimation of annual-peak streamflow frequency for undeveloped watersheds in Texas using PRESS minimization. U.S. Geological Survey Scientific Investigations Report 20085084 (Austin, TX). Asquith, W.H., Slade, R.M., Lanning-Rush, J., 1996. Peak-flow frequency and extreme flood potential for streams in the vicinity of the Highland Lakes, central Texas. U.S. Geological Survey Water-Resources Investigations Report 96-4072. Austin, TX. Asquith, W.H., Vrabel, J., Roussel, M.C., 2007. Summary of annual mean, maximum, minimum, and L-scale statistics of daily mean streamflow for 712 U.S. Geological Survey streamflow-gaging stations in Texas through 2003. U.S. Geological Survey Data Series 248, Austin, TX. Asquith, W.H., Herrmann, G.R., Cleveland, T.G., 2013. Generalized additive regression models of discharge and mean velocity associated with direct-runoff conditions in Texas: utility of the U.S. Geological Survey discharge measurement database. J. Hydrol. Eng. 18 (10), 1331–1348. http://dx.doi.org/10.1061/(ASCE)HE.1943-5584. 0000635. Baker, V.R., 1977. Stream-channel response to floods, with examples from central Texas. Geol. Soc. Am. Bull. 88 (8), 1057–1071. http://dx.doi.org/10.1130/0016-7606(1977) 88b1057:SRTFWEN2.0.CO;2. Baker, V.R., Kale, V.S., 1998. The role of extreme floods in shaping bedrock channels. In: Tinkler, K.J., Wohl, E.E. (Eds.), Rivers Over Rock: Fluvial Processes in Bedrock Channels. American Geophysical Union, Washington, DC, pp. 153–165. Barnes, H.H., 1967. Roughness characteristics of natural channels. U.S. Geological Survey Water-Supply Paper 1849, Washington, DC. Barnes, V.E., 1981. Geologic Atlas of Texas: Llano Sheet (1:250,000). The University of Texas at Austin Bureau of Economic Geology, Austin, TX. Beard, L.R., 1975. Generalized evaluation of flash-flood potential. The University of Texas at Austin Center for Research in Water Resources Report CRWR-124 (Austin, TX). Betson, R.P., 1979. A geomorphic model for use in streamflow routing. Water Resour. Res. 15 (1), 95–101. http://dx.doi.org/10.1029/WR015i001p00095. Bettess, R., White, W.R., 1983. Meandering and braiding of alluvial channels. Inst. Civ. Eng. Proc. 75 (3), 525–538. http://dx.doi.org/10.1680/iicep.1983.1443. Biedenharn, D.S., Little, C.D., Thorne, C.R., 1999. Magnitude–frequency analysis of sediment transport in the Lower Mississippi River. U.S. Army Corps of Engineers Miscellaneous Paper CHL-99-2, Vicksburg, MS. Blum, M.D., Valastro, S., 1989. Response of the Pedernales River of central Texas to Late Holocene climatic change. Ann. Assoc. Am. Geogr. 79 (3), 435–456. http://dx.doi. org/10.1111/j.1467-8306.1989.tb00271.x. Blum, M.D., Toomey, R.S., Valastro, S., 1994. Fluvial response to Late Quaternary climatic and environmental change, Edwards Plateau, Texas. Palaeogeogr. Palaeoclimatol. Palaeoecol. 108 (1–2), 1–21. http://dx.doi.org/10.1016/0031-0182(94)90019-1. Bomar, G.W., 1983. Texas Weather. University of Texas Press, Austin, TX. Bourke, M.C., Pickup, G., 1999. Fluvial form variability in arid central Australia. In: Miller, A.J., Gupta, A. (Eds.), Varieties of Fluvial Form. Wiley, Chichester, UK, pp. 249–271. Burnett, J., 2008. Flash Floods in Texas. Texas A&M University Press, College Station, TX. Carling, P., 2009. Geomorphology and sedimentology of the lower Mekong River. In: Campbell, I.C. (Ed.), The Mekong: Biophysical Environment of an International River. Academic Press, Oxford, UK, pp. 77–111.

18

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19

Carling, P., Jansen, J., Meshkova, L., 2014. Multichannel rivers: their definition and classification. Earth Surf. Process. Landf. 39 (1), 26–37. http://dx.doi.org/10.1002/esp.3419. Castro, J.M., Jackson, P.L., 2001. Bankfull discharge recurrence intervals and regional hydraulic geometry relationships: patterns in the Pacific Northwest, USA. J. Am. Water Resour. Assoc. 37 (5), 1249–1262. http://dx.doi.org/10.1111/j.1752-1688.2001. tb03636.x. Conyers, M.M., Fonstad, M.A., 2005. The unusual channel resistance of the Texas Hill Country and its effect on flood flow predictions. Phys. Geogr. 26 (5), 379–395. http://dx.doi.org/10.2747/0272-3646.26.5.379. Cooke, M.J., Stern, L.A., Banner, J.L., Mack, L.E., Stafford, T.W., Toomey, R.S., 2003. Precise timing and rate of massive late Quaternary soil denudation. Geology 31 (10), 853–856. http://dx.doi.org/10.1130/G19749.1. Dalrymple, T., Benson, M.A., 1968. Measurement of peak discharge by the slope-area method. U.S. Geological Survey Techniques of Water-Resource Investigation 03-A2 (Washington, DC). Doyle, M.W., Shields, D., Boyd, K.F., Skidmore, P.B., Dominick, D., 2007. Channel-forming discharge selection in river restoration design. J. Hydraul. Eng. 133 (7), 831–837. http://dx.doi.org/10.1061/(ASCE)0733-9429(2007)133:7(831). Dury, G.H., 1973. Magnitude–frequency analysis and channel morphometry. In: Morisawa, M. (Ed.), Fluvial Geomorphology. State University of New York, Binghamton, NY, pp. 91–121. Eaton, B.C., Church, M., 2007. Predicting downstream hydraulic geometry: a test of rational regime theory. J. Geophys. Res. 112 (F3), F03025. http://dx.doi.org/10.1029/ 2006JF000734. Erskine, W.D., Livingstone, E.A., 1999. In-channel benches: the role of floods in their formation and destruction on bedrock-confined rivers. In: Miller, A.J., Gupta, A. (Eds.), Varieties of Fluvial Form. Wiley, Chichester, UK, pp. 445–475. Ferguson, R.I., 1987. Hydraulic and sedimentary controls of channel pattern. In: Richards, K.S. (Ed.), River Channels: Environment and Process. Blackwell, Oxford, UK, pp. 129–158. Fonstad, M., Marcus, W.A., 2003. Self-organized criticality in riverbank systems. Ann. Assoc. Am. Geogr. 93 (2), 281–296. http://dx.doi.org/10.1111/1467-8306.9302002. Fuller, I.C., 2007. Geomorphic work during a “150-year” storm: contrasting behaviors of river channels in a New Zealand catchment. Ann. Assoc. Am. Geogr. 97 (4), 665–676. http://dx.doi.org/10.1111/j.1467-8306.2007.00576.x. García, C.C., 1995. Torrential flow frequency and morphological adjustments of ephemeral channels in south-east Spain. In: Hickin, E.J. (Ed.), River Geomorphology. Wiley, Chichester, UK, pp. 169–192. Gee, G.W., Bauder, J.W., 1986. Particle-size analysis. In: Klute, A. (Ed.), Methods of Soil Analysis: Part 1: Physical and Mineralogical Methods, 3rd edition Soil Science Society of America, Madison, WI, pp. 383–411. Gregory, K.J., Madew, J.R., 1982. Land use change, flood frequency and channel adjustments. In: Hey, R.D., Bathurst, J.C., Thorne, C.R. (Eds.), Gravel-Bed Rivers. Wiley, Chichester, UK, pp. 757–780. Griffith, G.E., et al., 2004. Ecoregions of Texas (1:2,500,000). U.S. Geological Survey, Reston, VA. Gupta, A., 1995. Magnitude, frequency, and special factors affecting channel form and processes in the seasonal tropics. In: Costa, J.E., Miller, A.J., Potter, K.W., Wilcock, P.R. (Eds.), Natural and Anthropogenic Influences in Fluvial Geomorphology. American Geophysical Union, Washington, DC, pp. 125–136. Guy, H.P., 1969. Laboratory theory and methods for sediment analysis. U.S. Geological Survey Techniques of Water-Resource Investigation 05-C1, Washington, DC. Heitmuller, F.T., 2011. Magnitude and frequency relations of bankfull flows and bedmaterial entrainment in the Llano River watershed, Central Texas, USA. Phys. Geogr. 32 (6), 533–560. http://dx.doi.org/10.2747/0272-3646.32.6.533. Heitmuller, F.T., Asquith, W.H., 2008. Potential for bed-material entrainment in selected streams of the Edwards Plateau: Edwards, Kimble, and Real counties, Texas, and vicinity. U.S. Geological Survey Scientific Investigations Report 20085017 (Austin, TX). Heitmuller, F.T., Hudson, P.F., 2009. Downstream trends in sediment size and composition of channel bed, bar, and bank deposits related to hydrologic and lithologic controls in the Llano River watershed, central Texas, USA. Geomorphology 112 (3–4), 246–260. http://dx.doi.org/10.1016/j.geomorph.2009.06.010. Heritage, G.L., van Niekerk, A.W., Moon, B.P., 1999. Geomorphology of the Sabie River, South Africa: an incised bedrock-influenced channel. In: Miller, A.J., Gupta, A. (Eds.), Varieties of Fluvial Form. Wiley, Chichester, UK, pp. 53–79. Heritage, G.L., Broadhurst, L.J., Birkhead, A.L., 2001. The influence of contemporary flow regime on the geomorphology of the Sabie River, South Africa. Geomorphology 38 (3–4), 197–211. http://dx.doi.org/10.1016/S0169-555X(00)00090-8. Hey, R.D., Thorne, C.R., 1986. Stable channels with mobile gravel beds. J. Hydraul. Eng. 112 (8), 671–689. http://dx.doi.org/10.1061/(ASCE)0733-9429(1986)112:8(671). Howard, A.D., 1987. Modelling fluvial systems: rock-, gravel-, and sand-bed channels. In: Richards, K.S. (Ed.), River Channels: Environment and Process. Blackwell, Oxford, UK, pp. 69–94. Huang, H.Q., Warner, R.F., 1995. The multivariate controls of hydraulic geometry: a causal investigation in terms of boundary shear distribution. Earth Surf. Process. Landf. 20 (2), 115–130. http://dx.doi.org/10.1002/esp.3290200203. Huckleberry, G., 1994. Contrasting channel response to floods on the middle Gila River, Arizona. Geology 22 (12), 1083–1086. http://dx.doi.org/10.1130/0091-7613(1994) 022b1083:CCRTFON2.3.CO;2. Kale, V.S., Hire, P.S., 2007. Temporal variations in the specific stream power and total energy expenditure of a monsoonal river: the Tapi River, India. Geomorphology 92 (3–4), 134–146. http://dx.doi.org/10.1016/j.geomorph.2006.06.047. Kale, V.S., Baker, V.R., Mishra, S., 1996. Multi-channel patterns of bedrock rivers: an example from the central Narmada basin, India. Catena 26 (1–2), 85–98. http://dx.doi.org/ 10.1016/0341-8162(95)00035-6.

Kleinhans, M.G., 2010. Sorting out river channel patterns. Prog. Phys. Geogr. 34 (3), 287–326. http://dx.doi.org/10.1177/0309133310365300. Knighton, A.D., 1987. River channel adjustment: the downstream dimension. In: Richards, K.S. (Ed.), River Channels: Environment and Process. Blackwell, Oxford, UK, pp. 95–128. Knighton, A.D., 1999. Downstream variation in stream power. Geomorphology 29 (3–4), 293–306. http://dx.doi.org/10.1016/S0169-555X(99)00015-X. Latrubesse, E.M., 2008. Patterns of anabranching channels: the ultimate end-member adjustment of mega rivers. Geomorphology 101 (1–2), 130–145. http://dx.doi.org/ 10.1016/j.geomorph.2008.05.035. Lecce, S.A., 1997. Nonlinear downstream changes in stream power on Wisconsin's Blue River. Ann. Assoc. Am. Geogr. 87 (3), 471–486. http://dx.doi.org/10.1111/14678306.00064. Leopold, L.B., Maddock, T., 1953. The hydraulic geometry of stream channels and some physiographic implications. U.S. Geological Survey Professional Paper 252, Washington, DC. Leopold, L.B., Wolman, M.G., 1957. River channel patterns: braided, meandering, and straight. U.S. Geological Survey Professional Paper 282-B, Washington, DC, pp. 39–85. López-Bermúdez, F., Conesa-García, C., Alonso-Sarría, F., 2002. Floods: magnitude and frequency in ephemeral streams of the Spanish Mediterranean region. In: Bull, L.J., Kirkby, M.J. (Eds.), Dryland Rivers: Hydrology and Geomorphology of Semi-arid Channels. Wiley, Chichester, UK, pp. 329–350. Magilligan, F.J., 1992. Thresholds and spatial variability of flood power during extreme floods. Geomorphology 5 (3–5), 373–390. http://dx.doi.org/10.1016/0169-555X(92)90014-F. Mear, C.E., 1995. Quaternary geology of the upper Sabinal River valley, Uvalde and Bandera counties, Texas. Geoarchaeology 10 (6), 457–480. http://dx.doi.org/10. 1002/gea.3340100604. Merritt, D.M., Wohl, E.E., 2003. Downstream hydraulic geometry and channel adjustment during a flood along an ephemeral, arid-region drainage. Geomorphology 52 (3–4), 165–180. http://dx.doi.org/10.1016/S0169-555X(02)00241-6. Meshkova, L.V., Carling, P.A., 2012. The geomorphological characteristics of the Mekong River in northern Cambodia: a mixed bedrock–alluvial multi-channel network. Geomorphology 147–148, 2–17. http://dx.doi.org/10.1016/j.geomorph.2011.06.041. Meshkova, L.V., Carling, P.A., 2013. Discrimination of alluvial and mixed bedrock–alluvial multichannel river networks. Earth Surf. Process. Landf. 38 (11), 1299–1316. http:// dx.doi.org/10.1002/esp.3417. Meshkova, L.V., Carling, P.A., Buffin-Bélanger, T., 2014. Nomenclature, complexity, semialluvial channels and sediment-flux-driven bedrock erosion. In: Church, M., Biron, P.M., Roy, A.G. (Eds.), Gravel-bed Rivers: Processes, Tools, Environments. Wiley, Chichester, UK, pp. 424–431 http://dx.doi.org/10.1002/9781119952497.ch31. Murray, G.E., 1961. Geology of the Atlantic and Gulf Coastal Province of North America. Harper & Brothers, New York, NY. Nanson, G.C., Knighton, A.D., 1996. Anabranching rivers: their cause, character and classification. Earth Surf. Process. Landf. 21 (3), 217–239. http://dx.doi.org/10.1002/(SICI) 1096-9837(199603)21:3b217::AID-ESP611N3.0.CO;2-U. Osterkamp, W.R., Hedman, E.R., 1982. Perennial-streamflow characteristics related to channel geometry and sediment in Missouri River Basin. U.S. Geological Survey Professional Paper 1242, Washington, DC. Parker, G., 1979. Hydraulic geometry of active gravel rivers. J. Hydraul. Div. 105 (9), 1185–1201. Parker, G., Wilcock, P.R., Paola, C., Dietrich, W.E., Pitlick, J., 2007. Physical basis for quasiuniversal relations describing bankfull hydraulic geometry of single-thread gravel bed rivers. J. Geophys. Res. 112 (F4), F04005. http://dx.doi.org/10.1029/2006JF000549. Phillips, J.D., 1990. The instability of hydraulic geometry. Water Resour. Res. 26 (4), 739–744. http://dx.doi.org/10.1029/WR026i004p00739. Pickup, G., Warner, R.F., 1976. Effects of hydrologic regime on magnitude and frequency of dominant discharge. J. Hydrol. 29 (1–2), 51–75. http://dx.doi.org/10.1016/00221694(76)90005-6. Pitlick, J., Cress, R., 2002. Downstream changes in the channel geometry of a large gravel bed river. Water Resour. Res. 38 (10), 34-1–34-11. http://dx.doi.org/10.1029/ 2001WR000898. R Development Core Team, 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria (http://www.r-project. org/ (last accessed on 3 June 2013)). Reinfelds, I., Cohen, T., Batten, P., Brierley, G., 2004. Assessment of downstream trends in channel gradient, total and specific stream power: a GIS approach. Geomorphology 60 (3–4), 403–416. http://dx.doi.org/10.1016/j.geomorph.2003.10.003. Richards, K.S., 1973. Hydraulic geometry and channel roughness: a non-linear system. Am. J. Sci. 273 (10), 877–896. http://dx.doi.org/10.2475/ajs.273.10.877. Robert, A., 2003. River Processes: An Introduction to Fluvial Dynamics. Arnold, London, UK. Schick, A.P., 1974. Formation and obliteration of desert stream terraces: a conceptual analysis. Z. Geomorphol. Suppl. 21, 88–105. Schumm, S.A., 1960. The shape of alluvial channels in relation to sediment type. U.S. Geological Survey Professional Paper 352-B, Washington, DC, pp. 17–30. Schumm, S.A., 1963. Sinuosity of alluvial rivers on the Great Plains. Geol. Soc. Am. Bull. 74 (9), 1089–1100. http://dx.doi.org/10.1130/0016-7606(1963)74 [1089:SOAROT]2.0.CO;2. Schumm, S.A., 1985. Patterns of alluvial rivers. Annu. Rev. Earth Planet. Sci. 13, 5–27. http://dx.doi.org/10.1146/annurev.ea.13.050185.000253. Schumm, S.A., 2005. River Variability and Complexity. Cambridge University Press, Cambridge, UK. Schumm, S.A., Khan, H.R., 1972. Experimental study of channel patterns. Geol. Soc. Am. Bull. 83 (6), 1755–1770. http://dx.doi.org/10.1130/0016-7606(1972)83[1755: ESOCP]2.0.CO;2. Shepherd, R.G., 1975. Geomorphic Operation, Evolution, and Equilibria, Sandy Creek Watershed, Llano Region, Central Texas. Ph.D. Dissertation. The University of Texas at Austin, USA.

F.T. Heitmuller et al. / Geomorphology 232 (2015) 1–19 Shepherd, R.G., 1979. River channel and sediment responses to bedrock lithology and stream capture, Sandy Creek drainage, central Texas. In: Rhodes, D.P., Williams, G.P. (Eds.), Adjustments of the Fluvial System. Kendall/Hunt, Dubuque, IA, pp. 255–275. Simpson, C.J., Smith, D.G., 2001. The braided Milk River, northern Montana, fails the Leopold–Wolman, gradient-discharge test. Geomorphology 41 (4), 337–353. http:// dx.doi.org/10.1016/S0169-555X(01)00066-6. Soong, D.T., Ishii, A.L., Sharpe, J.B., Avery, C.F., 2004. Estimating flood-peak discharge magnitudes and frequencies for rural streams in Illinois. U.S. Geological Survey Scientific Investigations Report 2004-5103, Urbana, IL. Stedinger, J.R., Vogel, R.M., Foufoula-Georgiou, E., 1993. Frequency analysis of extreme events. In: Maidment, D.A. (Ed.), Handbook of Hydrology. McGraw-Hill, New York, NY, pp. 18.1–18.66. Texas Parks and Wildlife Department, 2013. Precipitation in Texas. https://www.tpwd. state.tx.us/publications/pwdpubs/media/pwd_mp_e0100_1070e_08.pdf (last accessed 17 September 2013). Tinkler, K.J., 2001. The case of the missing flood: the unrecorded flood of 1935 on the James River, Mason County, Texas. Geomorphology 39 (3–4), 239–250. http://dx. doi.org/10.1016/S0169-555X(01)00030-7. Tinkler, K., Wohl, E., 1998. A primer on bedrock channels. In: Tinkler, K.J., Wohl, E.E. (Eds.), Rivers Over Rock: Fluvial Processes in Bedrock Channels. American Geophysical Union, Washington, DC, pp. 1–18. Torizzo, M., Pitlick, J., 2004. Magnitude–frequency of bed load transport in mountain streams in Colorado. J. Hydrol. 290 (1–2), 137–151. http://dx.doi.org/10.1016/j. jhydrol.2003.12.001. U.S. Department of Agriculture (USDA) Forest Service, 2013. WinXSPRO Version 3.0. http://www.stream.fs.fed.us/publications/winxspro.html (last accessed 3 June 2013). U.S. Geological Survey (USGS), 2007. Water resources data for the United States, water year 2006. U.S. Geological Survey Water-Data Report WDR-US-2006 (http://wdr. water.usgs.gov/wy2006/search.jsp (last accessed 15 March 2007)).

19

U.S. Geological Survey (USGS), 2008a. Earth Resources Observation and Science (EROS): Seamless Data Distribution System. (http://seamless.usgs.gov (last accessed 19 December 2008)). U.S. Geological Survey (USGS), 2008b. National Hydrography Dataset. http://nhd.usgs.gov (last accessed 19 December 2008). van den Berg, J.H., 1995. Prediction of alluvial channel pattern of perennial rivers. Geomorphology 12 (4), 259–279. http://dx.doi.org/10.1016/0169-555X(95)00014-V. van Niekerk, A.W., Heritage, G.L., Broadhurst, L.J., Moon, B.P., 1999. Bedrock anastomosing channel systems: morphology and dynamics in the Sabie River, Mpumalanga Province, South Africa. In: Miller, A.J., Gupta, A. (Eds.), Varieties of Fluvial Form. Wiley, Chichester, UK, pp. 33–51. Wentworth, C.K., 1922. A scale of grade and class terms for clastic sediments. J. Geol. 30 (5), 377–392. Wohl, E.E., 2004. Limits of downstream hydraulic geometry. Geology 32 (10), 897–900. http://dx.doi.org/10.1130/G20738.1. Wohl, E.E., Wilcox, A., 2005. Channel geometry of mountain streams in New Zealand. J. Hydrol. 300 (1–4), 252–266. http://dx.doi.org/10.1016/j.jhydrol.2004.06.006. Wolman, M.G., 1954. A method for sampling coarse bed material. Trans. Am. Geophys. Union 35 (6), 951–956. Wolman, M.G., Leopold, L.B., 1957. River flood plains: some observations on their formation. U.S. Geological Survey Professional Paper 282-C, Washington, DC, pp. 87–109. Wolman, M.G., Miller, J.P., 1960. Magnitude and frequency of forces in geomorphic processes. J. Geol. 68 (1), 54–74. http://dx.doi.org/10.1086/626637. Xu, J., 2004. Comparison of hydraulic geometry between sand- and gravel-bed rivers in relation to channel pattern discrimination. Earth Surf. Process. Landf. 29 (5), 645–657. http://dx.doi.org/10.1002/esp.1059.