Physical heterogeneity and aquatic community function in river networks: A case study from the Kanawha River Basin, USA

Physical heterogeneity and aquatic community function in river networks: A case study from the Kanawha River Basin, USA

Accepted Manuscript Physical heterogeneity and aquatic community function in river networks: A case study from the Kanawha River Basin, USA M.C. Thom...

2MB Sizes 0 Downloads 41 Views

Accepted Manuscript Physical heterogeneity and aquatic community function in river networks: A case study from the Kanawha River Basin, USA

M.C. Thoms, M.D. Delong, J.E. Flotemersch, S.E. Collins PII: DOI: Reference:

S0169-555X(16)30995-3 doi: 10.1016/j.geomorph.2017.02.027 GEOMOR 5945

To appear in:

Geomorphology

Received date: Revised date: Accepted date:

17 October 2016 23 February 2017 23 February 2017

Please cite this article as: M.C. Thoms, M.D. Delong, J.E. Flotemersch, S.E. Collins , Physical heterogeneity and aquatic community function in river networks: A case study from the Kanawha River Basin, USA. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Geomor(2016), doi: 10.1016/j.geomorph.2017.02.027

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Physical heterogeneity and aquatic community function in river networks: A case study from the Kanawha River Basin, USA.

T

M.C. Thoms1*, M.D., Delong2, J.E., Flotemersch3, and S.E., Collins4,

IP

1. Riverine Landscapes Research Laboratory, University of New England, NSW, Australia

CR

2. Large Rivers Studies Center, Winona State University, MN, USA

3. National Exposure Research Laboratory, U.S. Environmental Protection Agency, Cincinnati,

US

Ohio 45268, USA

AN

4. Lees-McRae College, NC, USA

M

* Corresponding author, Tel.: +61 2 67732768; E-mail: [email protected]

ED

Abstract

PT

The geomorphological character of a river network provides the template upon which evolution acts to create unique biological communities. Deciphering commonly observed

CE

patterns and processes within riverine landscapes resulting from the interplay between

AC

physical and biological components is a central tenet for the interdisciplinary field of river science. Relationships between the physical heterogeneity and food web character of functional process zones (FPZs) -- large tracts of river with a similar geomorphic character --in the Kanawha River (West Virginia, USA) are examined in this study. Food web character was measured as food chain length (FCL), which reflects ecological community structure and ecosystem function. Our results show that the same basal resources were present throughout 1

ACCEPTED MANUSCRIPT the Kanawha River but that their assimilation into the aquatic food web by primary consumers differed between FPZs. Differences in the trophic position of higher consumers (fish) were also recorded between FPZs. Overall, the morphological heterogeneity and heterogeneity of the river bed sediment of FPZs were significantly correlated with FCL. Specifically, FCL

T

increases with greater FPZ physical heterogeneity. The result of this study does not support

IP

the current paradigm that ecosystem size is the primary determinant of food web character in

US

CR

river ecosystems.

AC

CE

PT

ED

M

riverine ecosystem synthesis.

AN

Keywords: riverine landscapes; in-channel complexity; food webs; functional process zones;

2

ACCEPTED MANUSCRIPT Introduction The physical environment provides the template upon which evolution acts to forge characteristic life history strategies (cf. Southwood, 1977, 1988). Accordingly, the physical properties of rivers determine the type, abundance, and arrangement of biological

IP

T

assemblages found within these landscapes. Interactions between physical and biological components generate distinct patterns within riverine landscapes that have been observed at

CR

multiple scales. Conceptual models attempt to depict the structural and functional character

US

of rivers and the interactions between physical and biological components that occur within these landscapes. These include, but are not limited to, the river continuum concept (Vannote

AN

et al., 1980) and its various modifications (e.g., the serial discontinuity concept; Ward and

M

Stanford, 1983); the flood pulse Concept (Junk et cal., 1989); and the network dynamic hypothesis (Benda et al., 2004). Many of these models provided the foundation for

ED

investigating hydrogeomorphic species interactions within riverine landscapes (Allan, 2004;

PT

Bey and Sullivan, 2014). More recently, the riverine ecosystem synthesis (RES) of Thorp et al, (2008) provides a heuristic multiscaled model of riverine landscapes, built from the key

CE

concepts of fluvial geomorphology, landscape ecology, and lotic ecology. The RES specifically

AC

addresses the dynamic interaction of geomorphology, hydrology, and ecology in establishing the basis of physical and biological interaction within river networks. Physical character, disturbance, and energy flow are three primary drivers of the structure and function of river ecosystems (Power and Dietrich, 2002; Power, 2006; Thorp et al., 2008). The physical character of the riverine landscape determines the presence, location, and heterogeneity of morphological features present within river networks that can be used by 3

ACCEPTED MANUSCRIPT organisms as ‘habitat’. River channel and floodplain morphological features are subject to an array of hydrological and hydraulic conditions. Modifications to these conditions through flow and sediment regime changes represent significant ‘disturbances’ that can alter the structure and function of river ecosystems (Delong and Thoms, 2016a). The availability of food

T

resources is important to the trophic structure of river ecosystems, influencing the length of

IP

river food chains and biotic interactions at scales from days to decades. Food chain length

CR

(FCL), defined as the number of transfers between trophic groups (i.e., primary consumers eaten by secondary consumers) within a food web or as the number of trophic groups within a

US

food web (Post, 2002a) is a vertical measure of food web structure that reflects ecological

AN

community structure and ecosystem function (Hoeinghaus et al., 2008). For this reason, FCL is considered a central characteristic of ecological communities and one that is a key indicator of

M

the structure and function of ecosystems. Identifying the factors that determine FCL is a

ED

fundamental issue of ecogeomorphology (Dollar et al., 2007), river science (Thoms et al.,

PT

2016), and ecosystem ecology (Pimm, 2002). Food webs are a functional feature of ecosystems, the study of which provides insights into

CE

the natural drivers and anthropogenic influences on river ecosystems that cannot otherwise

AC

be attained from community-based approaches (Vander Zanden et al., 2003; Hoeinghaus et al., 2008). A paradigm in aquatic ecosystem ecology is that FCL increases with ecosystem size (cf. Post et al., 2000a; McHugh et al., 2004), which is based on the relationship between ecosystem size and habitat availability and species diversity. Coined the ecosystem-size hypothesis (Cohen and Newman, 1992), it supports the traditional view that as rivers increase in size downstream FCL also increases and food webs become larger (Vannote et al., 1980). Although 4

ACCEPTED MANUSCRIPT the influence environmental conditions has on FCL have been studied extensively (e.g., Polis et al., 2004; Belgrano et al., 2005) support for the ecosystem-size hypothesis is not overwhelming. A meta-analysis of the effect of environmental variability on FCL by Takimoto and Post (2012) and the extensive review by Sabo et al. (2010) highlighted that the

T

relationship is more complex. Empirical studies of FCL and ecosystem size reveal negative,

IP

positive, and null effects (e.g., Hoeinghaus et al., 2008). Ecosystem theory suggests that food

CR

webs, as measured by FCL, also respond to changes in environmental heterogeneity (Pimm, 1982), especially physical heterogeneity (Gilvear et al., 2016). Physical heterogeneity can be

US

represented as a function of the frequency, diversity, spatial arrangement, and turnover of

AN

morphological units within a riverine landscape. In terrestrial ecosystems, physical heterogeneity (a key component of landscape biocomplexity) has been shown to have direct

M

and positive influence on the structure, function, and overall productivity (e.g., Cadenasso et

ED

al., 2006). Research on the influence of physical heterogeneity on FCLs is limited despite its significance to understanding the interplay between physical and biological components

PT

within riverine landscapes (Levin, 1992; Fisher, 1993).

CE

The RES of Thorp et al. (2008) proposed a series of tenets intended to act as frameworks

AC

with which to explore the biocomplexity of riverine ecosystems, and these tenets are divided into three groups: individual ecological responses, community regulation, and ecosystem/landscape properties. Some tenets have been formulated into testable hypotheses that consider physical–biological interactions, but these focus in the area of individual ecological responses only. Many have not as yet been tested,, despite being widely claimed as to their usefulness for providing a framework with which to investigate physical and biological 5

ACCEPTED MANUSCRIPT interactions within riverine landscapes (Winemiller et al., 2010). The RES recognizes that hydrogeomorphic-ecological linkages operate at multiple scales. In river networks the presence of functional process zones (FPZs) -– large tracts of river with similar hydrogeomorphic character -– may dictate ecosystem structure and function. Functional

T

Process Zones have been shown to repeat themselves along a river network (Thoms et al., in

IP

press). Concurrent to the repetition of FPZs, ecological form and function shaped by physical

CR

character would be similar within repeating FPZs regardless of their location within a river network. Thus challenging traditional views of the more or less smooth transition in the

US

structure and function of river ecosystems from headwaters to lowland regions.

AN

A goal of river science is to decipher causal mechanisms or processes underlying commonly

M

observed patterns within riverine landscapes (Thoms et al., 2016). Thus, the objective of this manuscript is to test the hypothesis that food chain length (FCL) increases directly with the

ED

hydrogeomorphic complexity of a functional process zone (FPZ) in response to multiple

AC

Study area

CE

PT

factors related to habitat heterogeneity, biodiversity and abundance of food resources.

The Kanawha River Basin, located in the Mid-Atlantic region of the USA (Fig. 1A), has a catchment area of 31,691 km2. The catchment is relatively undisturbed in terms of human activities and contains significant areas (≈80 % of the catchment area) of a mixture of northern hardwoods (e.g., Acer saccharum, A. negundo, Betula nigra, Liriodendron tulipifera, Prunus serotina, Ulmus americana), oak (e.g., Quercus alba, Q. rubra), pine (e.g., Pinus strobus, P. 6

ACCEPTED MANUSCRIPT virginiana), and mixed mesophytic forests that extend from an altitude of 1355 to 164 m (Messinger and Hughes, 2000). The basin has three main physiographic provinces (Figure 1b), with each having a unique mixture of sandstone, shale, limestone, dolostone, chert, and alluvium geologies. The Appalachian Plateau, located in the northern sections of the basin,

T

accounts for ≈60 % of the catchment area. Paleozoic sedimentary rocks dominate this uplifted

IP

province of the Appalachian Plateau. Sedimentary rocks of various ages dominate the Valley

CR

and Ridge Province located in the mid-regions of the basin. This province consists of a series of parallel northeast-trending ridges and intervening valleys that are strongly folded and

US

faulted. The Blue Ridge Province occupies the southern sections of the basin and contains a

AN

series of metamorphic and igneous rocks that are exposed in a series of thrust sheets and, as a

M

result, is a region characterized by relatively deep, narrow valleys. The Kanawha River Basin has a typical continental climate with marked seasonal

ED

differences in precipitation and runoff. Mean annual precipitation (1961-2000) ranges from

PT

910-1500 mm, but there is 50 % more precipitation in the winter months compared to the summer months. This seasonal variation is consistent between the three provinces.

CE

Precipitation varies across the basin, being highly correlated with elevation (Messinger and

AC

Hughes, 2000). As a result, runoff across the basin is highly correlated with rainfall; thus, most runoff is generated in the Blue Ridge Province. The long-term (1961-2000) mean daily discharge at Charleston, WV, in the lowermost reaches of the Kanawha River is 432 m3 s-1, ranging between 31 and 6116 m3 s-1. The Kanawha River is a sixth order stream with a typical trellised network pattern (Fig. 1B) suggesting a relatively strong geological influence on the overall spatial organization of stream 7

ACCEPTED MANUSCRIPT network draining the Kanawha Basin. Williams et al., (2013) characterized six distinct FPZs within 9,326 km of the Kanawha River network -– large river zones of similar hydrogeomorphic character (Thoms et al., 2007). The upland constrained FPZ dominates the stream network of the Kanawha, accounting for 37.3 % of the total network length.

T

Contributions of the other FPZs are lowland constrained (18.8 %), upland high-energy (16.8

IP

%), reservoir (11.7 %), lowland alluvial (11.6 %), and open-valley upland (3.8 %). The spatial

CR

organization of FPZs resembles a mosaic of patches throughout the stream network with the repetitive arrangement of FPZs within the basin (Fig. 1C). Most FPZs are found in each

US

physiographic province; however, the lowland constrained FPZ is only located in the

AN

Appalachian Plateau Province. In general, the sequence of FPZs along the stream network reflects the varying influence of catchment and valley conditions. Upland high-energy FPZs

M

are generally located in the upper regions of the network, while upland constrained and open-

ED

valley upland FPZs tend to be more prominent in the mid sections of the stream network. In

alluvial FPZs dominate.

PT

the lowermost sections of the Kanawha River Basin the lowland constrained and lowland

CE

A total of 126 fish species from 25 families, have been recorded in the Kanawha River

AC

(White et al., 2005). Overall, the fish assemblage of the Kanawha River is representative of the greater Ohio River Basin -- the diversity of which is considered to be high globally (Benke and Cushing, 2011). A feature of the Kanawha River is the Kanawha Falls located ~ 143 km upstream of the Kanawha’s confluence with the Ohio River and immediately downstream of the confluence of the Gauley and New rivers. The Falls have a height of 4.5 m, forming a natural barrier to upstream fish movement at all flows below bankfull discharge. The river 8

ACCEPTED MANUSCRIPT network of the Kanawha Basin is a significant centre for fish dispersal regionally (cf. Hocutt et al., 1978). Despite the presence of potential fish bottlenecks in the Kanawha River, such as the Kanawha Falls, there is no clear evidence that they represent significant barriers to fish migration throughout the basin (Lachner and Jenkins, 1971). Recent surveys by the USEPA on

T

fish diversity in the West Virginia section of the Kanawha River documented 120 species

IP

downstream of the Falls and 87 upstream of the Falls. Additional species are likely present in

CR

the upper reaches of the Kanawha River (personal communication: Stuart Welsh, 27 January

AN

US

2016).

M

Methods

ED

3.1. Data collection

Three data sets were used in this study. The first data set provides information on the

PT

hydrogeomorphic character of different FPZs in the Kanawha River (Williams et al., 2013).

CE

The second data set contains information on the textural character of the river bed substratum at 35 sites located within four of the six FPZs within the Kanawha River Basin. The third is a

AC

biological data set containing information on the character of basal resources, benthic macroinvertebrates, and fish for the same 35 sites obtained from the USEPA. This data set also contains δ13C and δ15N ratios for basal resources, benthic macroinvertebrates and fish species. Details of each data set are provided here. The physical data set comprised 13 different variables for 554 sites located along the main drainage network of the basin and were used to determine the physical or geomorphic 9

ACCEPTED MANUSCRIPT heterogeneity of each FPZ (Table 1). A variety of tools and approaches are available for determining and analyzing FPZs within a river network (cf. Thoms et al., 2007; Milner et al., 2016). Initially, the spatial arrangement of FPZs in the Kanawha River was delineated using a series of automated geographic information system (GIS) tools, the details of which are given

T

in Williams et al. (2013) with a summary provided here. A series of sites along the Kanawha

IP

River drainage network, obtained from the U.S. National Hydrography Dataset (NHD), were

CR

created at 5-km intervals. These sites became the focus for the extraction of 13 geomorphic variables, which describe the physical character of the riverine landscape. At each site,

US

variables from three levels of organization (watershed, valley, and channel scales) were

AN

extracted using ArcGIS functions and tools contained within the RESoNate program (cf.

M

Williams et al., 2013).

This large data set on the physical character of the Kanawha River (554 sites by 13

ED

variables) was analyzed using a suite of multivariate statistical analyses that identified groups

PT

of sites with a similar physical character. This data set was classified using the flexible unweighted pair-group method with arithmetic averages (UPGMA) fusion strategy, as

CE

recommended by Belbin (1993), based on the final 13 variables. The Gower association

AC

measure, which is a range-standardized measure and recommended for nonbiological data (Belbin, 1993) was used. Groups of sites with similar physical character were selected by viewing the dendogram representation of the classification, and the least number of groups with the maximum similarity were chosen. Dendogram groups were then arrayed onto the river network of the Kanawha River to delineate the position of sites with similar physical character (using standard GIS mapping techniques). Groups of sites with similar physical 10

ACCEPTED MANUSCRIPT character equate to functional process zones (FPZs; cf. Thoms et al., 2007). A semi-strong hybrid multidimensional scaling (SSH) ordination was also computed from the Gower association measure of the FPZ data, and centroid distances were calculated between the FPZs to determine how different the four FPZs were. A SIMilarity PERcentage analysis (SIMPER)

T

was then undertaken to determine which hydrogeomorphic variables contributed most to the

IP

within-group similarity of each FPZ. The variability in physical character within the four FPZs

CR

was determined according to the rank dissimilarity used to compute the comparative index of multivariate dispersion (IMD), as described by Warwick and Clarke (1993). These measures

US

were developed to determine if the variability in community structure increased with

AN

disturbance (Warwick and Clarke, 1993). Thus, this multivariate measure of rank dissimilarity (i.e., IMD) was adopted to determine if the variability in physical character (i.e.,

M

the assemblage of hydrogeomorphic components) differed between FPZs. For this study,

ED

component variability within an FPZ was used as a surrogate measure for FPZ heterogeneity.

PT

The second data set contained information on the textural character of the river bed substratum at 35 sites located within four of the six FPZs within the Kanawha River Basin.

CE

These data are reported in a study of river bed sediment texture in different FPZs of the basin

AC

(Collins et al., 2014). Textural data, presented as the percent contribution to a Wentworth sediment class (e.g., fine sand, medium sand, etc.) were available for nine sites in the lowland alluvial, lowland constrained and upland constrained FPZs and eight sites within the upland high-energy FPZ. All sites were randomly located within each FPZ. Two of the FPZs present in the Kanawha River Basin, the open-valley upland FPZ and the Reservoir FPZ, did not have river bed sediment data available (Collins et al., 2014). For our study, we reanalyzed the raw 11

ACCEPTED MANUSCRIPT river bed sediment texture data to determine the complexity of the river bed substratum within each FPZ. For this analysis a similarity matrix, using the Gower association measure from the textural data was derived and an SSH ordination was computed in order to visually determine differences in river bed sediment texture between the FPZs. An analysis of

T

similarity (ANOSIM) was then undertaken to determine statistical differences between FPZs,

CR

heterogeneity in sediment texture within each FPZ.

IP

followed by computation of the IMD. For this data set, the IMD was taken to represent the

US

The third was a biological data set, obtained from the USEPA, containing information on the character of basal resources, benthic macroinvertebrates, and fish for the same 35 sites.

AN

Species present in this data set are those encountered during the collection of species targeted

M

as part of a food web study conducted by Collins (2014). This data set contains δ13C and δ15N ratios for basal resources, benthic macroinvertebrates, and fish species that allowed for food

ED

web calculations to be determined for four FPZs within the basin. The use of the stable

PT

isotopes, specifically δ13C and δ15N, is a commonly applied tool in food web studies to determine trophic position (the location of a consumer on the food chain) and the length of

CE

food chains (Vander Zanden and Rasmussen, 1999; Post, 2002b). The methods of sample

AC

collection, taxonomic resolution, and basic analyses to calculate FCL are described below. 3.2. Data analysis

For the biological data set, initially a series of one-way analysis of variance was performed on the δ13C and δ15N of the basal resources, as well as food chain length and mean trophic position, to establish if significant differences existed between the different FPZs. Community similarity measures were also determined for the macroinvertebrate and fish data sets for 12

ACCEPTED MANUSCRIPT each FPZ using the Sorensen similarity index (Sorensen, 1948). This metric has a range between 0 and 1, and higher values indicate greater similarity between two groups. The δ13C and δ15N for all samples (basal, macroinvertebrate, and fish), and with only the fish collected at sites within each FPZ, were then further examined via a range of multivariate statistical

T

analyses. Initially, a one-way ANOSIM tested for differences between FPZs based on a Bray

IP

Curtis similarity measure using 1000 permutations. Note that the Bonferroni correction for

CR

multiple pairwise tests is inappropriate for ANOSIM (Clarke, 1993), thus significance for

US

multiple pairwise tests for ANOSIM was determined with α = 0.05.

The contribution of basal food resources incorporated into biomass of benthic

AN

invertebrates was determined using the SIAR mixing model (stable isotope analysis in R;

M

Parnell et al., 2010). Mean δ13C and δ15N of aquatic macrophytes, benthic algae, whole leaves from the floodplain, and terrestrial detritus were compared separately to the δ13C and δ15N of

ED

each invertebrate sample within each FPZ to establish basal and primary consumer linkages.

PT

Correction for fractionation of δ13C and δ15N of basal resources was 0.8 and 3.4o/oo, respectively. Stable isotope ratios for each species of fish were used to calculate the trophic

CE

position using a single source mixing model (Jepsen and Winemiller, 2002). In contrast to the

AC

discrete measures of trophic level, trophic position is a continuous measure of trophic status and accounts for omnivory (Vander Zanden and Rasmussen, 1999; Post, 2002b). A mean of δ13C and δ15N of all invertebrate primary consumers within each FPZ served as a surrogate for basal resources to account for natural spatial and temporal variability of δ13C and δ15N of basal resources. Trophic position was calculated for each fish sampled. Individual measures of all fish within an FPZ were used to calculate mean trophic position for each FPZ. Mean trophic 13

ACCEPTED MANUSCRIPT position captures the average status of all consumers within a spatial entity as well as relative predominance of higher vs. lower consumers within the community. Food chain length (FCL) is defined here as the number of transfers of energy from one level of the food chain to the next. Based on this, fish with the highest trophic position (TP) within each FPZ were used to

IP

T

calculate food chain length as FCL = TPmax – 1. Individual relationships between FPZ heterogeneity and FCL as well as between river

Results

M

4.1. Physical character of Kanawha FPZs

AN

US

CR

sediment textural heterogeneity and FCL were determined via least squares regression.

ED

The four FPZs identified in the Kanawha River Basin were statistically different to one another in terms of their hydrogeomorphic character (ANOSIM Global R = 0.47: Table 2). The

PT

degree of difference between those FPZs located in the upland and lowland regions was

CE

marked. Centroid distance between the two upland FPZs (upland high-energy and upland constrained) was lower than that between the two lowland FPZs (lowland alluvial and

AC

lowland constrained), 1.02 and 1.89 respectively, suggesting a higher degree of physical similarity between upland FPZs compared to the lowland FPZs. In contrast, differences in FPZ physical character of the upland and lowland regions were greater (mean = 1.98; range = 0.932.03). The exception was the centroid distance between the upland and lowland constrained FPZs (centroid distance = 0.93). Results of the SIMPER analysis indicate the varying degree of influence of the 13 hydrogeomorphic variables on the four FPZs. Catchment geology, 14

ACCEPTED MANUSCRIPT downvalley slopes, and valley widths distinguish both upland region FPZs; whereas elevation and valley sideslopes were more prominent, in terms of their influence, on within FPZ similarity for the lowland region FPZs. The degree of valley confinement (the ratio of valley width to valley floor width) also had a similar degree of influence between the upland and

T

lowland constrained FPZs. Thus, those FPZs in the upland regions (upland high-energy and

IP

upland constrained) of the basin had higher downvalley slopes and narrower valley widths

CR

compared to those in the lowland regions (lowland alluvial and lowland constrained). A

US

summary of the physical characteristics of each FPZ is provided in Table 1. The total river channel surface area of the four FPZs in the Kanawha River Basin is 130.8

AN

km2. This surrogate measure of ecosystem size was calculated as a basic function of the

M

average bankfull width and the length of each FPZ segment following McHugh et al. (2010). The dominant FPZ in the Kanawha River network, in terms of ecosystem size, is the lowland

ED

alluvial (59.9 km2; 45.8 % of the total river channel surface area), followed by the lowland

PT

constrained (45.3 km2; 34.7 %), then the upland constrained (18.1 km2; 13.8 %) and finally the

CE

upland high-energy FPZ (7.5 km2; 5.7 %). The median grain size (D50) of the river bed sediment varied markedly between the four

AC

FPZs (Table 1) ranging from an average of 1.32 mm (sand) in the lowland alluvial FPZ to 72.1 mm (cobbles) in the upland high energy FPZ. The ANOSIM confirmed significant differences in the river bed sediment textural composition between FPZs of the Kanawha River Basin (Global R = 0.58: Table 2). Multiple pairwise comparisons show significant differences between most of the FPZs (i.e., four of the six pairwise relationships). The lowland alluvial FPZ was significantly different to all FPZs, and the lowland constrained FPZ was different to upland 15

ACCEPTED MANUSCRIPT high-energy. Only two pairwise comparisons had values < 0.3 (i.e., upland constrained vs. lowland constrained and upland constrained vs. upland high-energy) indicating a relatively high degree of overlap between zones in terms of sediment composition. 4.2. Basal resources

IP

T

Four dominant basal resources were identified within the river network of the Kanawha

CR

River: terrestrial detritus, tree leaves, macrophytes, and benthic algae. No significant differences were recorded in δ13C and δ15N of the basal resources between the four FPZs

US

(ANOVA: F1, 4 = 2.37; p = 0.47). However, the overall contribution or assimilation of these

AN

basal resources to primary consumers did differ between FPZs (Fig. 2).

M

4.3 Primary consumers

Twenty-three macroinvertebrate taxa (primary consumers) were collected at the 35 sites

ED

within the Kanawha River network (Table 3). Of these, two were found at all FPZs, and seven

PT

were unique to a particular FPZ. Pairwise similarity community ratios and the number of common taxa in each FPZ are provided in Table 4. The upland and lowland constrained FPZs

CE

have the greatest overall similarity between the four FPZs with a Sorenson similarity index of

AC

0.64, followed by the upland high-energy and upland constrained (0.57) and then upland constrained and lowland constrained (0.56). The largest difference in macroinvertebrates was between the upland high-energy and lowland alluvial FPZs with a Sorenson similarity index of 0.21. Mixing model output demonstrated the contribution of the four basal sources to benthic macroinvertebrate biomass differed across FPZs (Fig. 2). Aquatic macrophytes contributed 16

ACCEPTED MANUSCRIPT the most to the assimilated diet of invertebrates (~35%) in the lowland alluvial FPZ. Macrophytes were also the predominant source in lowland constrained FPZs with benthic algae contributing to a greater proportion of the diet here than in the lowland alluvial FPZ. The relative contributions of the four basal resources in the lowland alluvial FPZ were

T

comparable to the upland constricted FPZ, with benthic algae contributing only slightly more

IP

than the other sources. Benthic algae and terrestrial detritus were the major contributor to

CR

the macroinvertebrates biomass in the upland high-energy FPZ, whereas the contribution of

US

macrophytes was lowest here (18%). 4.4. Higher consumers

AN

Seventy-five of the expected 126 fish species were collected at the 35 sites within the

M

Kanawha River network (Table 3). Ten were common to all four FPZs, and 33 were only

ED

recorded within one FPZ. Pairwise similarity community ratios and the number of common fish species in the four FPZs are provided in Table 4. The upland and lowland Constrained

PT

FPZs have the greatest overall similarity between the four FPZs with Sorenson similarity index

CE

of 0.63. This is also reflected in the relatively high number of common fish species shared between these two FPZs (Table 3).

AC

The ANOSIM had a high Global R (R = 0.719) for the δ13C and δ15N of fish within the FPZs of the Kanawha River Basin. The pairwise ANOSIMs recorded differences in both δ13C and δ15N between all FPZs, except for the upland constrained and lowland constrained FPZs (Table 5).

17

ACCEPTED MANUSCRIPT 4.5. Food chain length and geomorphology Significant differences in food chain length and mean tropic position existed between most FPZs. The exception was for the upland constrained and lowland constrained FPZs (ANOVA: F1, 4 = 1.19; p = 0.51). Food chain lengths were, on average higher, in the lowland constrained

T

FPZ, compared to those in the upland constrained, lowland alluvial and upland high-energy

IP

zones, with median food chain lengths of 2.90, 2.08, 2.59, and 2.09 respectively (Fig. 3A).

CR

Mean trophic position also varied in a similar manner between the four FPZs (median FPZs

US

were 3.56 – upland constrained, 3.35 – lowland constrained, and 2.76 – upland high-energy; Fig. 3B).

AN

A positive relationship was found between food chain length and the physical heterogeneity

M

of the four FPZs in the Kanawha River Basin (Fig. 4). This was evident for the

ED

hydrogeomorphic character of the FPZs (p < 0.01; r2 = 0.74) and river bed sediment heterogeneity (p < 0.01; r2 = 0.71). These preliminary relationships show that there is no

PT

effect of FPZ size on FCL in the Kanawha River. The upland constrained FPZ had the longest

CE

FCL (mean FCL = 3) in the Kanawha River network but represented 13.8% of the total river channel area; whereas, the lowland alluvial FPZ, the dominant FPZ in the river network, had a

AC

mean FCL of only 2.8.

Discussion A positive relationship between physical heterogeneity and FCL, for the Kanawha River, lends support to the previously untested Tenet 8 of the riverine ecosystem synthesis -- food 18

ACCEPTED MANUSCRIPT chain length (FCL) increases directly with the hydrogeomorphic complexity of a functional process zone (FPZ) in response to multiple factors related to habitat heterogeneity, biodiversity, and abundance of food resources. Our results show significant positive relationships between FPZ physical heterogeneity, the heterogeneity of FPZ river bed

T

sediments, and FCL (cf. Fig. 4). Thus, the trophic structure of aquatic communities and their

IP

energy flow are conditioned by the abundance and diversity of components of the physical

CR

template as well as their organization (i.e., geomorphology) in the Kanawha River network. Food chain lengths in the Kanawha River are typical of river systems globally (Vander Zanden

US

and Fetzer, 2007). The range of mean trophic position in the Kanawha River FPZs (range: 2.8-

AN

3.5) are comparable to a global mean of ~3.5. High spatial heterogeneity is a defining feature of food webs in the Kanawha River despite the overall high-energy nature of the systems and

M

the relatively low geomorphic complexity between FPZs.

ED

Attempts to empirically test the influence of the physical character of riverine landscapes

PT

on FCL focus on the effect of ecosystem size (Power and Dietrich, 2002; Power, 2006; Vander Zanden and Fetzer, 2007). The results from the Kanawha River show no relationship between

CE

FPZ size (as a surrogate for ecosystem size) and FCL. Rather, FPZ heterogeneity (complexity)

AC

and heterogeneity of the river bed sediment is significantly associated with FCL. Habitat heterogeneity has been acknowledged to be an important determinant for FCL, but it has not been quantitatively tested. Walters and Post (2008) suggested that, although disturbance may alter individuals within a food chain, overall food chain length does not change with disturbance. Instead, Walters and Post (2008) argued that habitat heterogeneity and food web complexity buffer the effect of disturbance on food chain length. 19

ACCEPTED MANUSCRIPT The ‘habitat heterogeneity hypothesis’ proposed by MacArthur and Wilson (1967) assumes that structural complexity in environments provides more niches and diverse ways of exploiting resources in habitats thus creating more diverse communities (Simpson, 1949). In a review of the habitat heterogeneity- animal species relationship, Tews et al. (2004) found

T

that the majority of studies reported a positive correlation between habitat heterogeneity and

IP

species diversity. However, studies were biased toward terrestrial vertebrates and habitats

CR

under anthropogenic disturbance. Moreover, the review highlighted that the significant effect of habitat heterogeneity for a particular species differs with respect to the scale of study. The

US

influence of physical heterogeneity in riverine ecosystems has been previously documented at

AN

relatively small scales only. For example, spatial variations in hydraulic conditions and associated biological communities are well documented at the scale of individual riffles (e.g.,

M

Thoms, 1998; Lamouroux et al., 2004). Spatial variations in bed roughness along riffles and

ED

their corresponding distinct surface flow types –- areas of similar turbulence and near bed flow velocities (Dyer and Thoms, 2006) –- are the primary driver of macroinvertebrate

PT

assemblages in terms of its taxon richness and assemblage composition (Reid and Thoms,

CE

2008). Whereas, within small river reaches (<1000 m) of the Scioto River, Ohio, altering hydrogeomorphic conditions and changing food web architecture shortened FCL and

AC

destabilized food webs (Kautza and Sullivan 2014). Our study of the physical heterogeneity and FCL relationship in the Kanawha River is the first at the scale of a river network. Physical heterogeneity of riverine landscapes directly and indirectly influences the flow of energy and community structure of aquatic ecosystems. Although the same basal resources occurred throughout the Kanawha River, they are utilized differently by invertebrate primary 20

ACCEPTED MANUSCRIPT consumers based on their availability in each FPZ. Of the seven macroinvertebrate taxa unique to a single FPZ, the assimilated diet of each was consistent with the dominant basal resource for other taxa within the same FPZ. This pattern of shifts in the predominant basal resource was evident in taxa found in two or more FPZs. In contrast, only three common taxa

T

relied on the same basal resource in all FPZs. An analysis of geomorphological influences on

IP

the diet of primary consumers using fatty acids also determined that invertebrate primary

CR

consumers either have a change in or maintain a constant reliance on basal resources (Smits et al., 2015). Differences in physical habitat heterogeneity among FPZs may account for the

US

observed differences in basal resource availability and assimilation. Physical habitat

AN

heterogeneity is influenced by and, in turn, influences variation in flow, albeit at different scales. At the reach scale, areas with high flow variability have been found to have low bed

M

stability and greater habitat diversity (Jowett and Duncan, 1990). Davis and Barmuta (1989)

ED

suggested that as near-bed flow is influenced by substratum heterogeneity at the microhabitat scale, patches are created that support different autotrophic and heterotrophic communities.

PT

The contribution of aquatic macrophytes in the food web was relatively low in the two upland

CE

zones, and these zones are noted by higher instream energies that frequently disturb the river bed substratum. The two upland zones also have limited floodplain development thereby

AC

increasing the connectivity to, and possible dependence on, terrestrial basal resources (Power and Dietrich, 2002).

The influence of the physical character of river channels on fish, the composition of fish communities, and their location within river networks has been well studied. A meta-analysis of 393 fish biogeographical studies by Field et al. (2009) concluded environmental 21

ACCEPTED MANUSCRIPT heterogeneity was the primary factor influencing fish species richness. While Thompson and Townsend (2005), in a study of 18 streams in New Zealand, inferred that the spatial heterogeneity of physical habitat also plays a role in FCL in streams via indirect influences on algal productivity and the standing crops of organic matter. A combination of energy

T

availability, spatial heterogeneity and ecosystem size accounted for the representation of

IP

particular invertebrate feeding groups in the streams and overall food web structure. Fish

CR

communities in the Kanawha River differed among FPZs suggesting that the changes in hydrogeomorphic variables influence the communities in this system. Differences also exist in

US

the stable isotope ratios of higher-level consumers. This is likely a result of the changes in

AN

community composition of prey whose biomass would reflect shifts in the contribution of basal resources. An increase in number of macroinvertebrate genera has been linked to FPZs

M

in Brazilian streams (Godoy et al., 2016), and macroinvertebrate community richness may

ED

play a role in overall food web structure.

PT

The majority of conceptual river models suggest systematic longitudinal changes in river properties including food webs (cf. Vannote et al., 1980). In contrast to this traditional view,

CE

this study shows the Kanawha River network to be structured as a series of large FPZs, the

AC

downstream spatial arrangement of which is not regular and does not support conventional river models (Thorp et al., 2008; Delong and Thoms, 2016b). The results indicate that downstream change in the hydrogeomorphic character of river networks can be diverse and complex rather than linear and predictable. This study has shown each FPZ to have a unique FCL or food web character, and this is driven by the physical heterogeneity of the FPZ. Thus, conventional approaches to understanding the structure and function of river ecosystem 22

ACCEPTED MANUSCRIPT processes at the scale of river networks may be limited. Indeed, future models and studies of the interaction between physical character and food webs should include measures of heterogeneity at the network scale. The hydrogeomorphic character of river networks directly influences ecosystem structure and function by altering the spatial and temporal

T

components of the habitat template (e.g., riffles, pools, wetted channels, slackwaters, and

IP

floodplains) at a range of scales within the riverine landscape (Thoms et al., in press). Thus

CR

the spatial organization of FPZs dictates the availability, partitioning, and distribution of

US

resources (as well as the provision of physical habitat) throughout a river network. The influence environmental conditions have on FCL have been studied extensively (cf.

AN

Belgrano et al., 2005; Polis and Winemiller et al., 2013; Polis et al., 2004). Recent food web

M

studies, including our study of the Kanawha River, challenge the conventional wisdom of the dominant effect of ecosystem size (Takimoto and Post, 2012). Food chain length can be

ED

influenced by an array of variables operating and interacting at multiple spatial and temporal

PT

scales (Fig. 5). An analysis of food webs in 36 North American rivers by Sabo et al. (2010) for example, suggests that hydrological variability to be the primary mechanism underlying the

CE

correlation between ecosystem size and FCL in rivers. Current evidence points to a complex

AC

set of interacting environmental determinants on FCL (Delong and Thoms, 2016a). Identifying the relevant drivers, processes or phenomena, and their scales of influence on FCL is important for advancing our knowledge of the structure and function of river ecosystems. This requires an interdisciplinary approach and a joining of many areas of understanding into a single conceptual-empirical structure. A framework for the interdisciplinary study of FCL is provided in Fig. 5. Frameworks show how facts, hypotheses, 23

ACCEPTED MANUSCRIPT and models may be linked (Pickett et al., 1999), thereby providing a way of ordering phenomena, and thereby revealing patterns of structure and function (Delong and Thoms, 2016b). This framework acknowledges FCL as an integrator of community structure in river ecosystems and recognizes the influence of physical character, disturbance and food resources

T

and their interactions, as primary variables influencing FCL. These primary variables are an

IP

expression of natural and anthropogenic secondary variables operating over a wider range of

CR

scales. The action of the physical character of riverine landscapes on FCL, for example, is expressed not just via ecosystem size but also its physical heterogeneity and the spatial

US

arrangement or location of the various physical components within a river system (Figure 5).

AN

The framework places the collective use of top-down and bottom-up methods in a multidimensional context, such as required for FCL studies (Polis et al., 2004). It also identifies the

M

role of the physical character of the river network (i.e., geomorphology) on shaping FCL,

ED

specifically, and river ecosystem food webs in general. Knowledge of the distribution of

particular ecosystem processes throughout river networks is important for river science and

CE

PT

management.

AC

Acknowledgements

MCT thanks the Large River Studies Center at Winona State for hosting him whilst on study leave. The views and interpretations expressed herein are those of the authors and do not reflect those of the USEPA.

24

ACCEPTED MANUSCRIPT References

CE

PT

ED

M

AN

US

CR

IP

T

Allan, J.D. 2004. Landscapes and riverscapes: the influence of land use on stream ecosystems. Annu. Rev. Ecol. Evol. Syst. 35: 257–284. doi:10.1146/annurev.ecolsys.35.120202.110122. Belgrano, A., Scharler, U.M., Dunne, J., Ulanowicz, R.E., (2005). Aquatic food webs: An ecosystem approach. Oxford University Press. Bey, C. R., Sullivan, S. M. P. (2014). Associations between stream hydrogeomorphology and co‐dependent mussel–fish assemblages: evidence from an Ohio, USA river system. Aquatic Conservation: Marine and Freshwater Ecosystems 25: 555-568. Belbin, L. (1993): Environmental representativeness: regional partitioning and reserve selection. Biological Conservation 66: 223-230. Benda, L., Poff, L.R., Miller, D., Dunne, T., Reeves, G., Pollock, M., Pess, G. (2004). Network dynamics hypothesis: spatial and temporal organization of physical heterogeneity in rivers. BioScience 54: 413-427. Benke, A. C., Cushing, C. E. (Eds.). (2011). Rivers of North America. Academic Press. Cadenasso, M. L.,, Pickett, S. T. A., Grove, J. M. (2006). Dimensions of ecosystem complexity: Heterogeneity, connectivity, and history. Ecological Complexity 3, 1-12. Clarke, K.R. (1993): Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18: 117-143. Collins, S.E., (2014). Comparing hypotheses proposed by two conceptual models for stream ecology. Unpublished PhD Thesis, University of Cincinnati. Collins, S.E, Thoms, M.C., Flotemersch, J.E. (2014). Hydrogeomorphic zones characterize riverbed sediment patterns within a river network. River Systems DOI: http://dx.doi.org/10.1127/1868-5749/2014/0084. Davis, J.A., Barmuta, L.A. (1989). An ecologically useful classification of mean and near-bed flows in streams and rivers. Freshwater Biology 21: 271-282. Delong, M.D., Thoms, M.C., (2016a). Changes in the trophic status of fish feeding guilds in response to flow modification. Journal of Geophysical Research: Biogeosciences 121: 949-964.

AC

Delong, M.D., Thoms, M.C., (2016b). An ecosystems framework for river science, management and policy. In Gilvear, D.A., Greenwood, M.W., Thoms, M.C., Wood, P.A., (eds). River Science, Management and Policy for 21st Century. Wiley, 15-36. Dollar, E. S. J., James, C. S., Rogers, K. H., Thoms, M. C. (2007). A framework for interdisciplinary understanding of rivers as ecosystems. Geomorphology 89:147-162.Dyer, F., Thoms, M.C., (2006). Managing river flows for hydraulic diversity. River Research and Applications 22, 257-267. ield R. aw ins .A. Cornell . . Currie . . iniz- ilho .A. . uegan, J-F., Kaufman, D.M., Kerr .T. Mittelbach . . Oberdorff T. O’ rien E.M. Turner .R. . (2009). Spatial speciesrichness gradients across scales: a meta-analysis. Journal of Biogeography 36: 132-147. 25

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

Fisher, S.G. (1993). Pattern, process, and scale in freshwater systems: some unifying thoughts. in: Giller, P.S., Hildrew, A.G., Raffaelli, D.G. (eds.), Aquatic ecology: scale, pattern, and process. Blackwell Scientific, Oxford, 575-597. Gilvear, D. G., Greenwood, M. A., Thoms, M. C., Wood, P. A. (2016). River science: Research and management for the 21st century. Wiley, Chichester. Godoy, B.S., Simiao-Ferreira, J., Lodi, S., Oliveira, L.G. (2016). Functional process zones characterizing insect communities in streams of the Brazilian Cerrado. Neotropical Entomology 45: 159-169. Hawkes, H.A. (1975). River zonation and classification. in: Whitton, B.A. (ed.), River Ecology. Blackwell Science Publishers, Oxford, UK, 312-374. Hocutt, C. H., Denoncourt, R. F., Stauffer, J. R., (1978). Fishes of the Greenrier river, West Virginia, with drainage history of the Cental Appalachians. Journal of Biogeography, 5, 5980. Hoeinghaus DJ, Winemiller KO, Agostinho AA, (2008). Hydrogeomorphology and river impoundment affect food chain length of diverse Neotropical food webs. Oikos 117: 984995. Jenkins, R. E., Burkhead, N. M. (1994). Freshwater fishes of Virginia. Bethesda, Maryland: American Fisheries Society. 1079pp. Jepsen, D.B., Winemiller, K.O. (2002). Structure of tropical river food webs revealed by stable isotope ratios. Oikos 96: 46-55. Jowett, I.G., Duncan, M.J. (1990). Flow variability in New Zealand rivers and its relationship to in-stream habitat and biota. New Zealand Journal of Marine and Freshwater Research 24: 305-317. Junk, W.J., Bayley, P.B., Sparks, R.E. (1989). The flood-pulse concept in river-floodplain systems. in: D.P. Dodge (ed.), Proceedings of the international large river symposium (LARS). Canadian Special Publication in Fisheries and Aquatic Sciences 106 110-127. Kautza, A., Sullivan, S.M.P., (2016). Anthropogenic and natural determinants of fish food-chain length in a midsize river system. Freshwater Science: 35, 1-14. Lamouroux N. ole’ ec S. ayraud S. (2004). iological traits of stream macroinvertebrates communities: effects of microhabitat, reach and basin filters. Journal of the North American Benthological Society: 23, 449-466. Lachner, E. A., Jenkins, R. E. (1971). Systematics, distribution and evolution of the chub genus Nocomis Girard (Pisces, Cyprinidae) of eastern United States, with descriptions of new species. Smithsonian Contrib. Zool. 85, 1-97. Levin, S.A., (1992). The problem of pattern and scale in ecology. Ecology 73: 1943-1967. MacArthur, R.H., Wilson, E.O., (1967). The theory of island biogeography. Princeton University Press, Princeton. Marks, J. C., Power, M. E. and Parker, M. S. 2000. Flood disturbance, algal productivity, and interannual variation in food chain length. Oikos 90: 20-27. 26

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

McHugh, P.A., McIntosh, A.R., Jellyman, P.G., (2010). Dual influences of ecosystem size and disturbance on food chain length in streams. Ecology Letters 13: 881–890 Messinger, T., Hughes, C.A. (2000): Environmental setting and its relations to water quality in the Kanwaha River Basin. U.S. Geological Survey, Water Resources Investigations, Report 00-402, 66pp. Milner, V. S., Gilvear, D. J., Thoms, M.C. (2016). Characterisation Riverine Landscapes; Hisotry, applications and future challenges. In Thoms, M.C., Gilvear, D.A., Wood, P.A., Greenwood, M.W., (eds). River Science: Management and Policy for 21st Century. Wiley, 239-258. Parnell, A.C., Inger, R., Bearhop, S., Jackson, A.L. (2010). Source partitioning using stable isotopes: coping with too much variation. PlosOne 5:1-5. Pickett, S. T. A., Burch, W. R., Grove, J. M. (1999). Interdisciplinary research: maintaining the constructive impulse in a culture of criticism. Ecosystems 2, 302-307. Pimm SL, 1982. Food webs. University of Chicago Press. Polis, G.A., Winemiller, K.O., (1996). Food webs: Integration of patterns and dynamics. Springer, NY. Polis, G.A., Power, M.E., Huxley, G.R., (2004). Food webs at the landscape level. The University of Chicago Press, Chicago. Post, D. M. (2002a). The long and short of food-chain length. Trends in Ecology and Evolution 17(6): 269-277. Post, D. M. (2002b). Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83: 703-718. Power, M.E., (2006). Environmental controls on food web regimes: A fluvial perspective. Progress in Oceanography: 68, 125-133. Power, M. E., Dietrich, W.E., (2002). Food webs in river networks. Ecological Research 17:451– 471. Reid, M.A., Thoms, M.C., (2008). Near bed hydraulics and associated invertebrate communities. Biogeosciences, 5, 1033-1041. Sabo, J.L., Finlay, J.C., Kennedy, T., Post, D.M., 2010. The role of discharge variation in scaling of drainge area and food chain length in rivers. Science 330: 965-967. Simpson, E.H., (1949). Measurement of diversity. Nature 163; 688. Smits, A.P., Schindler, D.E., Brett, M.T., (2015). Geomorphology controls the trophic base of stream food webs in boreal watershed. Ecology 96; 1775-1782. Sorensen, T., (1948). A method of establishing groups of equal amplitude in plant sociology based on similarity of species and its application to the analyses of vegetation on Danish commons. Kongelige Danske Videnskabernes Selskab 5: 1-34. Southwood, T.R.E., (1977). Habitat, the template for ecological strategies? Journal of Animal Ecology 46: 337–365. Southwood, T.R.E. (1988). Tactics, strategies and templets. Oikos 52: 3-18. 27

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

Takimoto, G., Post, D.M., (2012). Environmental determinants of food chain length: a meta analysis. Ecological Research 28: 675-681. Tews, J., Brose, U., Grimm, V., Tielborger, K., Wichmann, M.C., Schwanger, M., Jeltsch, F., (2004). Animal species diversity driven by habitat heterogeneity/diversity: the importance of keystone structures. Journal of Biogeography 31: 79-92. Thompson, R.M., Townsend, C.R., (2005). Energy availability, spatial heterogeneity and ecosystem size predict food-web structure in streams. Oikos 108: 137-148. Thoms, M.C., (1998). Sand accumulation in a gravel bed river. in Bobrowsky, P., (ed.) Aggregate Resources: A Global Perspective, Balkemma, 157-169. Thoms, M.C., Rayburg, S., Neave, M., (2007). The physical diversity and assessment of a large river system: The Murray-Darling Basin, Australia. in Gupta, A. (Ed.) Large Rivers, Wiley, Chichester. Thoms, M. C., Scown, M. A., Flotemersch, J. (in press). Characterization of river networks: An approach and applications. Journal of the American Water Resources Association. Thoms, M.C., Gilvear, D.G., Wood, P.A., Greenwood, M.A., (2016). River science and management for the 21st Century. In Gilvear, D.A., Greenwood, M.W., Thoms, M.C., Wood, P.A., (eds). River Science, Management and Policy for 21st Century. Wiley, 1-14. Thorp, J.H, Thoms, M.C., Delong, M.D., (2008). The Riverine Ecosystem Synthesis: Toward a Conceptual Cohesiveness in River Science. Academic Press, London. Vander Zanden, M. J., Fetzer, W.W., (2007). Global patterns of aquatic food chain length. Oikos 116: 1378–1388. 13C a 15N and the trophic position of aquatic consumers. Ecology 80: 1395-1404. Vander Zanden MJ, Chandra S, Allen BC, Reuter JE, Goldman C R, (2003). Historical food web structure and restoration of native aquatic communities in the Lake Tahoe (CaliforniaNevada) Basin. Ecosystems 6: 274-288. Vannote, R.L., Minshall, G.W., Cummins, K.W., Sedell, J.R., Cushing, C.E., (1980). The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences 37:130-137. Walters, A.W., Post, D.M. (2008). An experimental disturbance alters fish size structure but not food chain length in streams. Ecology 89: 3261-3267. Walters, A. W., Post, D.M., (2008). An experimental disturbance alters fish size structure but not food chain length in streams. Ecology 89:3261–3267. Ward, J.V., Stanford, J.A. (1983). The serial discontinuity concept of lotic ecosystems. in: Fontaine, T.D., Bartell, S.M. (eds.), Dynamics of lotic ecosystems. Ann Arbor Science, 43-68. Ann Arbor, Michigan, USA. Warwick, R.M., Clarke, K.R., (1993). Comparing the severity of disturbance: a meta-analysis of marine macrobenthic community data. Marine Ecology Progress Series 92: 221-231. White D, Johnston K, Miller, M., (2005). Ohio River basin. In Benke AC, Cushing CE, editors. Rivers of North America. Elsevier Academic Press, San Diego, CA: p375-424. 28

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

Williams, B.S., D'Amico, E., Kastens, J.H., Thorp, J.H., Flotemersch, J.E., Thoms, M.C., (2013). Automated riverine landscape characterization: GIS-based tools for watershed-scale research, assessment, and management. Environmental Monitoring and Assessment 185: 7485-7499. Winemiller, K.O., Flecker, A.S., Hoeinghaus, D.J., (2010). Patch dynamics and environmental heterogeneity in lotic ecosystems. Journal of the North American Benthological Society 29: 84-99.

29

ACCEPTED MANUSCRIPT List Tables. Table 1.

T

Physical measures of Functional Process Zones (FPZs) in the Kanawha River Basin; variables 1 to 13 were used to determine the FPZs; mean and range (in italics) are provided

IP

Table 2.

US

CR

Pairwise analysis of similarity (ANOSIM) and centroid distances between the four functional process zones (FPZs) in the Kanawha River Basin based on hydrogeomorphic character (bold) and river bed sediment texture; centroid distances are in italics

AN

Table 3.

M

Benthic macroinvertebrates and fish collected in each functional process zone of the Kanawha River that were used for stable isotope analysis; fish listed as “unidentified” were

PT

ED

recognizable as a different species but were unable to be identified to species

CE

Table 4.

AC

Pairwise Sørensen coefficients (community similarity index) and the number of common macroinvertebrate taxa (bold) and fish species for the four functional process zones (FPZs) in the Kanawha River; the number of common taxa are in italics

Table 5. Pairwise analysis of similarity (ANOSIM) results for δ13C (bold) and δ15N of the collective fish species between the four FPZs in the Kanawha River Basin 30

ACCEPTED MANUSCRIPT

List of Figures. Figure 1.

CR

IP

T

The Kanawha River Basin. A) The river basin and its approximate location within the continental USA; B) the stream network of the basin where the Kanawha River is in boldface and physiographic provinces are shown; and, Cc) the location of Functional Process Zones (FPZs) and the locations of sampling sites

US

Figure 2.

AN

The mean (+ 1 SE) relative contribution of basal resources to macroinvertebrate primary consumers within each Functional Process Zone (FPZ) of the Kanawha River

M

Figure 3.

CE

Figure 4.

PT

ED

Food web character for each Functional Process Zone (FPZ) in the Kanawha River network: A) food chain length; and, B) mean trophic position; the box and whisker plots provide the mean, 25th and 75th percentiles, and the range of each food chain length and trophic level

AC

The Functional Process Zone (FPZ) heterogeneity – Food Chain Length relationship for the Kanawha River

Figure 5.

31

ACCEPTED MANUSCRIPT A conceptual model of influences on the structure and function of river ecosystems using

AC

CE

PT

ED

M

AN

US

CR

IP

T

food chain as an integrator of community structure and energy flow

32

CE AC

Figure 1

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

33

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

Figure 2

34

Figure 3

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

35

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

Figure 4

36

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Figure 5

37

ACCEPTED MANUSCRIPT Table 1:

Functional process zone

Elevation (NAVD m)

Basi

2. Dominant geology

constrained

n

164-540

167-587

545-922

562-1022

Basi

Alluvium

Sandstone

Granite-

Gneiss-

7.

Ratio of valley to valley trough width

Downvalley slope

Right valley sideslope

AN

914-1502

1044-1640

2286

1420

1253

811

904-3798

280-2633

425-716

471-809

1191

217

254

220

394-2518

168-971

76-339

61-236

4.28

3.16

2.59

2.19

y

1.32-4.91

1.1-13.6

1.98-2.69

2.01-2.32

Valle

0.00032

0.00136

0.00589

0.00756

y

0.000085-

0.00789-

M

1053-1404

PT

Valle

ED

1019-1258

Valle y

Valle

0.00112

8.

dolostone

1353

n

AC

6.

sandstone

1063

CE

5. Valley floor width (m)

shale-

791

1189

y

1116.2

673

high energy

272.9

Basi

4. Valley width (m)

constrained

Upland

269.9

n

3. Mean annual rainfall (mm)

Upland

CR

1.

alluvial

Lowland

T

e

Lowland

IP

Scal

US

Variable

Valle

0.04151-

0.00896

0.023

0.292 38

0.009567

0.263

0.134

ACCEPTED MANUSCRIPT y

0.016-

0.243-

0.530

0.292

0.286

0.184

y

0.026-

0.277-

0.234-

0.165-

256-652

169-192

1.08

1.0

1.0-1.1

US

1.0-1.0

1.0-1.0

1.04

1.03

1.01

1.0-1.05

1.0-1.04

1.0-1.05

1.25

1.14

1.3

1.0-5.0

1.0-2.0

1.0-2.1

1.0-2.3

1.32

5.58

59.3

72.1

0.91-4.23

3.29-8.23

40.9-82.9

68.2-112.3

1.30 1.14-1.89

M

3.7

PT

Rive r bed

66-112

1.0-1.0

AC

CE

14. D50 of river bed sediment (mm)

92-140

98

1.0

Chan nel

135

0.450

1.0

ED

13. River channel planform class

T

180

Chan nel

0.587

451

Chan nel

0.325

AN

nel

12. Sinuosity of the river

0.457

0.035

Chan

11. Sinuosity of the channel belt

0.115-

Valle

0.530 10. Width of the river channel belt (m)

0.564

IP

Left valley sideslope

0.315

CR

9.

0.194-

39

ACCEPTED MANUSCRIPT

Table 2.

FPZ

Upland high energy

Upland constrained

Lowland constrained

Lowland constrained

0.70 ** - 1.89 0.307 * - 1.25

0.38 * - 0.93 0.0ll ns – 1.02

Lowland alluvial

0.81 ** - 2.03 0.349 * - 1.99

0.74 ** - 1.98 0.428 * - 2.11

0.63 ** - 1.89 0.318 * - 1.79

AN M

ns – no significant difference.

- overlap between groups but a significant difference exists.

ED

*

CR

0.53 * - 1.02 0.032 ns – 0.92

US

Upland constrained

IP

T

Upland high energy

AC

CE

PT

** - no overlap between groups and a significant difference.

40

Lowland alluvial

ACCEPTED MANUSCRIPT Table 3: Benthic macroinvertebrates and fish collected in each functional process zone of the Kanawha River that were used for stable isotope analysis. Fish listed as “unidentified” were recognizable as a different species but were unable to be

IP

T

identified to species.

Lowland

d alluvial

constrained

Upland

constrained

Upland high energy

M

AN

Lowlan

ED

Benthic macroinvertebrates

PT

Baetidae

CE

Birgella

AC

Branchycentridae Caenidae

US

CR

Functional process zone

X X X

X

X

X

Campeloma

X

X

X

Corbicula

X

X

X

41

X

ACCEPTED MANUSCRIPT Elimia

X

Ephemeridae

X

Helicopsychidae

X X X

Isonychiidae

X

Limnephilidae

X

Physella

CE

AC

Philopotamidae

X

Planorbella Pleurocera

X

X

X

X

X X

PT

Neoephemeridae Odontoceridae

X X

M

X

X

ED

Leptoxis

X

AN

Hydropsychidae

X

X

CR

X

US

Heptageniidae

X

T

X

IP

Glossosomatidae

X

X

X

X

X

X

X

X

X

X

42

ACCEPTED MANUSCRIPT Polycentropidae

X

X

Uenoidae

X

X

Aplodinotes grunniens

X

X

X

X

X

X

IP

X

CR

Ambloplites ruperstis

T

Fish

anomalum

X

M

Catostomus

ED

commersonii

X

X

X

PT

Chrosomus

X

AN

Carpiodes carpio

US

Campostoma

X

CE

erythrogaster

AC

Chub unidentified 1 Chub unidentified 2

X X

X

X

Chub unidentified 3

X

X

Chub unidentified 4

X

43

ACCEPTED MANUSCRIPT X

X

X

X

Cyprinus carpio

X

X

X

Darter unidentified 1

X

X

Darter unidentified 2

X

X

Darter unidentified 3

X

Darter unidentified 4

X

Darter unidentified 5

M

ED

X X

PT

Erimystax dissimilis

X

US

CR X

X

X X

X

Etheostoma caeruleum

X

X

Etheostoma flabellare

X

X

X

X

Etheostoma nigrum

X

X

X

X

CE

X

AC

blennioides

X

X

Dorosoma cepedianum

Etheostoma

X

IP

Cyprinella spiloptera

X

T

X

AN

Cottus bairdii

Etheostoma olmstedi

X X

X

44

ACCEPTED MANUSCRIPT Etheostoma zonale

X

X

X

X

Hypentelium nigricans

X

X

X

X

Ichthyomyzon X

X

X

Labidesthes sicculus

X

X

X

CE

Lepomis macrochirus

AC

Lepomis megalotis

X

X

PT

Lepomis cyanellus

ED

Lepomis auritus

X

M

Lepisosteus osseus

X

AN

Lampetra aepyptera

CR

Ictiobus bubalus

X

IP

X

US

Ictalurus punctatus

X X

X

X

X

X

X

X

X

X

Lepomis unidentified 1

Lepomis unidentified

X

45

T

unicuspis

ACCEPTED MANUSCRIPT 2 Luxilus cornutus

X

X

X

Madtom unidentified X X

X

X

X

X

X

M

Minnow unidentified

ED

1

PT

Morone chrysops

X

CE

Moxostoma carinatum

AC

Moxostoma duquesni

X X X

Moxostoma erythrurum

X

X

Moxostoma macrolepidotum

X

X

AN

Micropterus salmoides

X

US

Micropterus punctulatus

IP

X

CR

Micropterus dolomieu

T

1

X

46

ACCEPTED MANUSCRIPT Nocomis micropogon

X

X

Nocomis platyrhynchus

X X X

X

IP

X

CR

Notropis atherinoides

US

Notropis heterolepis X

X

Percina phoxocephala

X

CE

Pimephales promelas

Pomoxis

AC

Pomoxis annularis

nigromaculatus Pylodictis olivaris

X

PT

Pimephales notatus

ED

Percina caprodes

X

X

X

M

Noturus insignis

AN

Notropis telescopus

T

Notorus miurus

X X

X

X

X X

47

X

ACCEPTED MANUSCRIPT Rhinichthys cataractae

X

X

Rhinichthys obtusus

X

X X

Salvelinus frontinalis

X

X

CR

Sander vitreus

IP

T

Salmo trutta

X

Shiner unidentified 1

X

X

Shiner unidentified 2

X

X

X

X

X

X

X

X

X

X

X

X

AC

CE

Sucker unidentified 1

PT

Shiner unidentified 4

ED

Shiner unidentified 3

X

AN

X

M

atromaculatus

US

Semotilus

48

ACCEPTED MANUSCRIPT Table 4.

FPZ

Upland high energy

Upland constrained

Lowland constrained

Upland high energy 0.1 - 8 0.63 - 24

Lowland constrained

0.56 - 7 0.49 - 23

0.41 - 1 0.63 - 30

Lowland alluvial

0.21 - 2 0.40 - 15

0.48 - 6 0.47 - 18

AC

CE

PT

ED

M

AN

US

CR

IP

T

Upland constrained

49

0.63 - 7 0.63 - 20

Lowland alluvial

ACCEPTED MANUSCRIPT Table 5.

Upland

Upland

high energy

constrained

Lowland constrained

IP

Upland high

0.528 *

constrained

US

Upland

CR

energy

0.637 *

constrained

ED

0.598 *

0.712 *

M

Lowland

AN

O.421 *

0.349 *

PT

Lowland alluvial

0.428 *

0.618 *

0.561 *

0.701 *

AC

CE

0.439 *

0.698 *

ns – no significant difference. *

Lowland alluvial

T

FPZ

- overlap between groups but a significant difference exists.

** - no overlap between groups and a significant difference. 50