Skip to main content
NCBI home page
EPA Author Manuscripts logoLink to EPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Aug 27.
Published in final edited form as: J Am Water Resour Assoc. 2018;54(4):899–913. doi: 10.1111/1752-1688.12649

Characterization of River Networks: A GIS Approach and Its Applications

Martin Thoms 1, Murray Scown 2, Joseph Flotemersch 3
PMCID: PMC6711203  NIHMSID: NIHMS1041238  PMID: 31456632

Abstract

Fluvial geomorphology provides the basis for characterizing complex river networks and evaluating biophysical processes within watersheds. Understanding the spatial organization of morphological features, their influencing processes, and resultant geomorphic diversity in stream networks are important for efficient restoration, river health assessment, and improving our knowledge of the resilience of riverine landscapes. River characterization is a means to determine the biophysical character of river networks but many methods are fraught with pitfalls, such as the use of incorrect variables and limited acknowledgment of the hierarchical organization of rivers. In this paper, a top‐down geographic information system‐based approach for determining the physical typology of river networks is outlined. A suite of multivariate analyses are used to develop a nomenclature for functional process zones (FPZs) — large tracts of the river network with similar hydro‐geomorphological character. Applied to the Little Miami River, Ohio, six distinct FPZs emerged, which had a nonuniform distribution along the river network. Some FPZs repeated downstream; others were rare in terms of total length and number of FPZ segments. The physical structure of the Little Miami River network was analyzed using a series of community metrics. Application of this approach for river monitoring, establishing reference conditions, as well as management of threatened and endangered species and asset trading is highlighted.

Introduction

A notable shift in the study and management of river systems has occurred from a reach/site‐based focus to more holistic larger landscape‐ or watershed‐scale focus, in the last two decades (Thoms et al. 2016). This shift was, inter alia, a result of increasing recognition that reach/site‐based approaches to the study of rivers often failed to address problems that contribute to longer term declines in the structure and function of rivers at the watershed scale (Thorp et al. 2008; Likens et al. 2009). Understanding and effective management of river systems at the watershed‐scale — the river network — requires data, information, tools, and actions to occur at appropriate scales; however, the traditional focus of river research, monitoring, and assessment has been limited to the scale of the site or reach (Delong and Thoms 2016; Thoms et al. 2016). The use of data and/or information collected at the site‐ or reach‐scale to assess watershed‐scale attributes is scientifically inappropriate because of the hierarchical nature of river networks and the principles of hierarchy theory (Dollar et al. 2007). Moreover, the focus on individual habitats or resources within river networks usually fails to recognize them as complete, intra‐connected systems. River networks are complex, hierarchically nested systems that function across multiple levels of organization and scale (Thoms et al. 2007; Milner et al. 2016). As such, river networks exhibit emergent properties, whereby the structure and function at higher levels of organization (e.g., entire river networks) cannot be simply deduced from the collective knowledge of components at lower levels of organization (e.g., individual reaches or sites) (Allen and Starr 1982). In response, efforts to develop and implement scale‐aware, interdisciplinary approaches to river management have emerged (Mitchell 1990; Cairns and Crawford 1991; Noss 1995; Karr 1996; Kenney 1997; Hull et al. 2003). Such challenges require frameworks and approaches to river network characterization and evaluation at appropriate scales for the questions being set and link the physical structure of a river with its ecosystem functioning and provision of services (Dollar et al. 2007; Steel et al. 2010).

The structure of river networks at the watershed scale is typically viewed as a simple, continuous gradient of physical conditions from headwaters to great rivers (cf. the River Continuum Concept [RCC]; Vannote et al. 1980). River network characterization approaches have been influenced and developed from clinal conceptual models (e.g., stream ordering; Strahler 1957). For some questions, knowledge of the stream order/size and position downstream (as in the RCC) or the number of upstream tributary connections (as in the network dynamics hypothesis; Benda et al. 2004) provides information required for river management. However, recent conceptual models and research of river system structure and function suggest they are more accurately portrayed as downstream arrays of large hydrogeomorphic zones formed by factors such as regional geology, valley conditions, geomorphic structure of the channel and valley, climate, and hydrologic patterns (cf. Montgomery 1999; Poole 2002; Thoms and Parsons 2002, 2003; Thorp et al. 2006, 2008; Fitzpatrick et al. 2016; Thoms et al. 2016). These zones at the valley‐scale are described in the River Ecosystem Synthesis (RES; Thorp et al. 2006, 2008) as functional process zones (FPZs). FPZs are delineated and named based on statistically derived features from the river network of the river channel and surrounding valley along with geological and precipitation information. According to the RES, FPZs are repeatable longitudinally and only partially predictable in location, especially at scales above the ecoregional level (Omernik 1987), which contrasts the clinal downstream changes predicted by the RCC. While derived from physical datasets, the character and location of FPZs with a stream network have been shown to be significantly associated with different flow regimes (Thoms and Parsons 2003), sediment regimes (Collins et al. 2014), and geological history (Thoms et al. 2007). These are key independent variables determining the structure and function of river channel morphology (Schumm 1977). The physical character of individual FPZs has also been shown to be associated with significantly different fish and invertebrate communities (Arthington et al. 2010; Milner et al. 2015), biogeochemical conditions (Kreiling et al. 2018), and aquatic food web character (Thoms et al. 2017). Hence these physically derived river zones have functional relevance.

River networks are complex social–ecological systems that are hierarchically organized (sensu Berkes 2007). As such, river networks require research and management that integrate their hydrogeomorphic character, ecosystem services, and societal priorities and preferences (Parsons et al. 2016). Combining knowledge of the character of river networks — for example, through identification of FPZs — with the ecosystem services expected to be provided by different FPZs (Thorp et al. 2010) is one such approach toward integrating the biophysical, ecological, and social aspects of river networks for management. Studies that have characterized FPZs within river networks with limited application or utility to such integrated river management exist (Thoms et al. 2007), and there is a need to progress these avenues by linking concepts of river network characterization, ecology, and ecosystem services.

Tools exist for analyzing and characterizing river networks (Kondolf et al. 2003) and FPZs (Thoms et al. 2007; Thorp et al. 2008; Milner et al. 2016). FPZs can be identified with either bottom‐up field methods (e.g., traditional on‐site methods in fluvial geomorphology) or with increasingly accessible top‐down techniques (e.g., geospatial‐based analyses) with readily available high resolution digital elevation model (DEM) data. Bottom‐up methods can be labor intensive, time consuming, and costly. Conversely, the top‐down methods, if data are available and of sufficient quality, are emerging as general techniques for delineating FPZs. Pilot efforts using top‐down approaches have been applied to rivers in the United States (U.S.) (e.g., Kanawha River; Williams et al. 2013), and those of the Murray‐Darling Basin, Australia (Thoms et al. 2007). Although these approaches have not as yet been widely applied, Milner et al. (2016) advocate for a paradigm shift in river science and its applications, especially in terms of river characterization — one that explicitly tackles the issues of multiple spatiotemporal scales and heterogeneity. The use and availability of remote sensing technologies and the increasing use of multivariate statistics for river typing should underpin this paradigm shift. Such a transition will improve conceptual understanding of river ecosystems from a hydrogeomorphic and eco‐geomorphological perspective, and improve river management, restoration, and conservation activity.

In this paper, we examine the spatial arrangement of hydrogeomorphic zones — FPZs — throughout a river network delineated using a series of automated geographic information system (GIS) tools. This top‐down approach is applied to the Little Miami River, in southwestern Ohio, which traverses a landscape dominated by the influence of glacial process during the Quaternary. We examine the structure and occurrence of FPZs in the Little Miami River (i.e., frequency, distribution, and diversity), provide a quantitative approach for the nomenclature of FPZs as well as a discussion of what FPZ characterization can tell us about floodplain occurrence within and across all FPZs within a watershed. The various components of the diversity of FPZs throughout the watershed are examined to glean information about rare physical habitats or potential threats, in a similar manner to how components of species diversity are used in ecological studies. We also discuss the applications of river network characterization to management and conservation of habitats and ecosystem services.

Study Area

The Little Miami River, located in southwestern Ohio, is a tributary of the Ohio River and has a watershed area of approximately 4,550 km2 (Figure 1). Its watershed is a complex social–ecological system containing 5,642 km of stream network, at the 1:24,000 scale, draining natural grasslands, urban, suburban, forested, and agricultural lands. It has four main sub‐catchments — Caesar Creek, Todd Fork, East Fork, and the Little Miami. The headwaters of the main channel of the Little Miami River are located southeast of Springfield, Ohio. From here the Little Miami River flows southwest to its confluence with the Ohio River — approximately 6 km east of downtown Cincinnati, Ohio. Elevation across the watershed ranges from a maximum of 490 m a.s.l. in the headwater regions near Springfield, Ohio, to 121 m a.s.l. at the confluence of the Little Miami and Ohio Rivers (Debrewer et al. 2000). Thus a distinct north–south topographic gradient exists across the watershed. The Little Miami River is a designated State of Ohio and National Scenic River because of its high‐quality aquatic communities (mainly fish), a largely intact riparian corridor, scenic views, and historic sites.

Figure 1.

Figure 1

Open in a new tab

The Little Miami River catchment.

Glaciation during the Quaternary influenced the regional landscape and presence of different landforms within the watershed (Debrewer et al. 2000). Regional variations in the thickness of glacial deposits are evident across the watershed because of differential advances and retreat of glacial ice sheets during the Quaternary. In the southern regions of the watershed, glacial deposits are relatively thin (<1 m) or absent compared to those in the northern regions (up to 100 m). The Quaternary glacial deposits overlie thick sequences of Ordovician, Silurian, and Devonian sedimentary rocks (Casey 1996). Overall, variations in the presence of unconsolidated glacial deposits and the character of the underlying bedrock influence the physical characteristics of Little Miami watershed landscape (Debrewer et al. 2000). The Little Miami River watershed can be separated, topographically into two distinct regions (Figure 1). The Southern Ohio Loamy Till Plain in the northern regions of the watershed is characterized by moderately low relief associated with end and recessional moraines, flat to rolling topography, well‐drained soils, and generally low‐gradient streams (Schiefer 2002). Stream valleys alternate between broad floodplains and constrained zones and are commonly filled with glacial outwash, although the main river channel has incised into local bedrock in locations, forming deep gorges. This region of the watershed is dominated by agricultural activities. By comparison, the Pre‐Wisconsin Drift Plain in the southern regions of the watershed was unglaciated during the last glacial maximum and is characterized by increased topographic relief, although in the East Fork subwatershed, there are areas of swampland. The river valleys in this region of the Little Miami River watershed are typically relatively narrow and bordered by rock buffs. However, where the rivers traverse preglacial drainage lines, valley bottoms are broader (Schiefer 2002). The Drift Plain region, in the southern portion of the watershed, is a mix of developed, forested, and pasture land. In many parts of the Little Miami watershed, river channels have eroded through the glacial deposits exposing underlying bedrock sequences. Overall, these spatially distinct landforms and land uses pose different threats to the health and integrity of river systems that drain the watershed. Numerous streams in the lower watershed have been impaired by the effects of urbanization, urban runoff, and/or municipal wastewater discharge. The agricultural areas of the Little Miami River watershed have seen problems with siltation and nitrogen and phosphorus loading (Ohio EPA 2009).

The Little Miami River has a strong seasonal flow regime, the peak of which is associated with spring snowmelt during March and April. On average, flows range from 11 m3/s in September to 70 m3/s in March for the 100‐year mean monthly discharge. Soil drainage also affects hydrology differently between the Till and Drift Plains in the watershed, with loamy soils in the Till Plain typically having good natural drainage resulting in seasonally wet areas, whereas loess and fragipan soils in the Drift Plain typically result in more sustained stream runoff than in the Till Plain. The flow regime of some regions of the Little Miami River watershed is influenced by impoundments. The largest of these impoundments is Caesar Creek Lake, followed by Harsha Lake, Cowan Lake, Stonelick Lake, and Shawnee Lake, and have a combined reservoir surface area of approximately 23 km2. Approximately 75 km of stream length in the Little Miami River watershed has been affected by these five reservoirs (Figure 1).

Methods

The drainage network of the Little Miami River was obtained from the National Hydrography Dataset (NHD), which provides a 1:24,000‐scale digital representation of streamlines (drainage network) within watersheds across the conterminous U.S. (Simley and Carswell 2009). The NHD streamlines were reconditioned to provide a single continuous line for each unique stream; such reconditioned streams are now available for the conterminous U.S. in the National Stream Internet (Nagel et al. 2015). Initially a series of sites along the Little Miami River drainage network were created at 5‐km intervals. Thus, 427 sites were allocated along 1,800 km of the Little Miami River drainage network. These sites became the focus for the extraction of 15 geomorphic variables (Table 1), which describe the physical character of the riverine landscape. At each site, variables from three scales of organization (watershed, valley, and channel scales as recommended by Thoms et al. 2007) were extracted using a revised version of the ArcGIS functions and tools contained within the RESoNate program (Williams et al. 2013). The watershed‐scale variables included elevation, dominant geology, and annual rainfall. Elevation was determined from the digital National Elevation Dataset (NED — 10 m) (Gesch et al. 2009) and mean long‐term annual rainfall (30‐year) was derived from the PRISM Spatial Climate Datasets (PRISM Climate Group 2004). Geology was measured from the 1:25,000‐scale vector geologic map of the U.S. (USGS 2015) and this was aggregated into five basic geological categories within the watershed, which were ranked based on their rock type, erodibility, and potential sediment yield (cf. Renwick 1996). The valley scale variables were: valley width, valley trough width, the ratio of valley width to the valley trough width, the left and right valley slopes, and down valley slope, and these were determined from the 10 m NED DEM of the watershed. The valley troughs were delineated from the DEM using the FLDPLN Model Valley Floor Mapper 1.0 (KARS 2015). The six channel‐scale variables were: channel belt width, channel belt sinuosity, channel belt wavelength, channel sinuosity, wavelength, planform class, and the number of river channels, which were determined from either the original or reconditioned NHD streamlines. Data sources used to derive the 15 geomorphic variables used for the river characterization of the Little Miami River are given in Table 1.

Table 1.

Variables used for the physical characterization of the Little Miami River network

Variable Scale Data source
Elevation (m) Watershed National Elevation Dataset (NED). The NED is the primary elevation data product produced and distributed by the United States (U.S.) Geological Survey (Gesch et al. 2009); data available at https://pubs.usgs.gov/fs/2009/3053/
Dominant geology Watershed National geological maps of the U.S.; available at https://mrdata.usgs.gov/geology/state/
Mean annual rainfall (mm) Watershed PRISM Spatial Climate Dataset for the U.S., which contains historical annual rainfall; available at http://prism.oregonstate.edu
Valley width (m) Valley Digital elevation model (DEM) derived from the NED; data available at https://pubs.usgs.gov/fs/2009/3053/
Valley floor (trough) width (m) Valley DEM derived from the NED; data available at https://pubs.usgs.gov/fs/2009/3053/
Ratio of valley to valley trough width Valley DEM derived from the NED; data available at https://pubs.usgs.gov/fs/2009/3053/
Down valley slope Valley DEM derived from the NED; data available at https://pubs.usgs.gov/fs/2009/3053/
Right valley sideslope Valley DEM derived from the NED; data available at https://pubs.usgs.gov/fs/2009/3053/
Left valley sideslope Valley DEM derived from the NED; data available at https://pubs.usgs.gov/fs/2009/3053/
Width of the river channel belt (m) Channel The National Hydrography Dataset (NHD). The NHDFlowline dataset provides the spatial geometry of streamlines with watersheds of the U.S. The NHDWaterbody dataset provides the spatial geometry of basic water bodies such as lakes, reservoirs, and pond features. These datasets are available at https://nhd.usgs.gov/data.html
Wavelength of the channel belt Channel The NHD. The NHDFlowline dataset provides the spatial geometry of streamlines with watersheds of the U.S. The NHDWaterbody dataset provides the spatial geometry of basic water bodies such as lakes, reservoirs, and pond features. These datasets are available at https://nhd.usgs.gov/data.html
Sinuosity of the channel belt Channel The NHD. The NHDFlowline dataset provides the spatial geometry of streamlines with watersheds of the U.S. The NHDWaterbody dataset provides the spatial geometry of basic water bodies such as lakes, reservoirs, and pond features. These datasets are available at https://nhd.usgs.gov/data.html
Sinuosity of the main river channel Channel The NHD. The NHDFlowline dataset provides the spatial geometry of streamlines with watersheds of the U.S. The NHDWaterbody dataset provides the spatial geometry of basic water bodies such as lakes, reservoirs, and pond features. These datasets are available at https://nhd.usgs.gov/data.html
River channel planform class Channel The NHD. The NHDFlowline dataset provides the spatial geometry of streamlines with watersheds of the U.S. The NHDWaterbody dataset provides the spatial geometry of basic water bodies such as lakes, reservoirs, and pond features. These datasets are available at https://nhd.usgs.gov/data.html
Number of river channels Channel The NHD. The NHDFlowline dataset provides the spatial geometry of streamlines with watersheds of the U.S. The NHDWaterbody dataset provides the spatial geometry of basic water bodies such as lakes, reservoirs, and pond features. These datasets are available at https://nhd.usgs.gov/data.html
Open in a new tab

The large dataset generated for this study (427 sites by 15 variables) was analyzed using a variety of multivariate statistical techniques that identified groups of sites with similar physical characteristics. Initial analyses show the Little Miami River to have a single thread channel; therefore, the number of river channels variable had no discriminatory power among the sites. Thus, this variable was removed from further analysis. Next, the data were classified using the flexible unweighted pair‐group method with arithmetic averages (UPGMA) fusion strategy, as recommended by Belbin and McDonald (1993), based on the final 14 variables. The Gower association 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 from the dendrogram representation of the cluster analysis, whereby the least number of groups with the maximum similarity was chosen. This step required the identification of an inflexion point in the relationship between the number of groups in the classification and their corresponding similarity value. Dendrograms rarely display a linear increase in the number of groups; more commonly, the number of groups increase exponentially with increasing similarity (Thoms et al. 2007). This approach has been successfully applied in different environmental applications, including sediment textural analysis (Forrest and Clark 1989), catchment hydrology (Thoms and Parsons 2003), and stream network morphology (Thoms et al. 2017). Once identified, these similar groups were then arrayed onto the streamlines of the Little Miami River, to delineate the position of sites with similar physical character (using standard GIS mapping techniques). These groups of sites identified in the classification analysis represent FPZs — lengths of river with similar valley‐floodplain settings and river morphologies — and are inferred to be influenced by similar geomorphic processes (Thoms et al. 2016). To further elucidate FPZs, a semi‐strong hybrid multidimensional scaling ordination was performed on the data. Sites were arrayed in ordination space and then an ANalysis Of SIMilarity (ANOSIM) was used to assess differences between FPZs. Finally, a SIMilarity PERcentage analysis (SIMPER) was undertaken to determine which geomorphic variables contributed to the within‐group similarity. This analysis was used to construct a FPZ nomenclature for the Little Miami River drainage network. There are five reservoirs within the Little Miami River watershed. These were identified from the water bodies shapefile of the NHD and added as an additional FPZ in the characterization.

The typology data for the Little Miami River drainage network (number of segments, total length of segments for each FPZ) were analyzed with a series of parameters commonly used for determining the diversity of ecological communities (Magurran 2004). This requires the recognition of the drainage network as a community of FPZs, with a “functional process zone” being analogous to a “species” in ecology. Thus, not only could the overall diversity of a community of FPZs within a drainage network be determined, but also the individual components of abundance, evenness, and richness that make up diversity (Harris et al. 2009). Diversity was measured at the whole‐network scale, where richness was calculated as the number of FPZs present, and abundance as the total length of the channel of each FPZ. Evenness was measured using Simpson’s evenness index that provides a value between 0 and 1 representing the overall distribution of channel lengths between different FPZs. When an evenness value approaches 1, channel lengths are more evenly distributed between FPZs. A lumped diversity measurement for the Little Miami River network was measured using the Shannon–Weiner diversity index (H′); calculated as:

H1=∑pilnpi (1)

where pi is the proportion of channel lengths found in the ith FPZ (adapted from Magurran 2004).

The potential floodplain area within each continuous section of individual FPZs was also determined. The valley floor areas delineated using the FLDPLN Valley Floor Mapper were used as a surrogate for potential floodplain area. The valley floor polygons were split laterally across the valley at each node where two different FPZs met. Valley floor polygons remained continuous and intact within each continuous section of FPZ. The total floodplain area within each FPZ type was then calculated in ArcGIS 10.2 as well as the distribution of floodplain areas among individual segments of each FPZ.

Verification of the location of FPZs that emerged in the Little Miami River network was undertaken. The field‐based studies of Schiefer (2002) and Ritter (2012) on the regional variability of landforms and river systems provide information on topographic relief, regional geology, down valley slopes, valley dimensions as well as general descriptions of the physical character of river networks within watersheds of Ohio. The data from Schiefer (2002) and Ritter (2012) were combined to form a field‐based river network characterization of the Little Miami River. Albeit limited in terms of scale, it does allow a comparison of a field‐based characterization of the river network for the Little Miami River (cf. Schiefer 2002; Ritter 2012) with that generated from the top‐down GIS method employed in this study.

Results

Five distinct FPZs emerged from the statistical classification of the 427 sites within the Little Miami River drainage network (Figure 2). These five FPZs explain 72.1% of the similarity between sites within this drainage network. Each FPZ was significantly different to one another in terms of its physical character (ANOSIM: Global R = 0.673). FPZs 1 and 2 are characterized by lower downstream slopes, bedrock geology, and are generally located in the lowland regions of the watershed (Figure 2a), thus they are referred to as “Lowland open valley” and “Lowland constrained” types, respectively (Figure 2b). The difference between these two lowland FPZs is the degree of valley confinement. By comparison, FPZs 4 and 5 are the “Upland constrained” and “Upland open valley lower energy” zones located predominantly in the upper Till Plain regions of the Little Miami River watershed, a region of lower relief, lower slopes, and alternating broad and narrow valley floors (Figure 2a). The part of the drainage network characterized by FPZ 3 (the “Open valley floodplain” zone) has distinctly wider valley floors that are associated with the presence of more extensive floodplain. This FPZ is located in the mid‐regions of the watershed, where the Till Plain transitions to the Drift Plain, and also at the mouth of the watershed (Figure 2a). Thus, similar FPZs (as defined by group membership derived from the cluster analysis) tend to group together spatially within the Little Miami River watershed (Figure 2a).

Figure 2.

Figure 2

Open in a new tab

The functional process zones (FPZs) of the Little Miami River. (a) Spatial organization of FPZs within the Little Miami River drainage network and (b) the classification dendrogram of the Little Miami River. Dashed line shows the boundary between the Till and Drift Plains.

The cluster analysis for a three to six group solution revealed clear separation of the physical variables contributing highly to the within‐group similarity (Figure 2b). The first split in the dendrogram separated FPZs 1, 2, and 3 from FPZs 4 and 5. The former zones are mainly located in the mid to lowland, relatively higher energy regions of the watershed, while the latter two zones are located in the upper, lower energy regions of the watershed.

The spatial location of the five FPZs that emerged in the Little Miami River network (Figure 2a) corresponds to the field‐based river network characterizations reported by Schiefer (2002) and Ritter (2012). At a regional scale, there is clear separation of upland (Upland constrained and Upland open valley lower energy) and lowland FPZs (Lowland open valley and Lowland constrained) and this conforms to those river and stream types of the Till Plain and Drift Plain regions (cf. Figure 2a). Two of the FPZs (i.e., Upland constrained and Open valley floodplain) conform directly to the location of specific river and stream types within the Little Miami River. The Upland constrained FPZ, for example, corresponds with the Clifton Gorge (39°47′41.5″N, 83°49′42.5″W) and the river gorge in the John Bryan State Park (39°47′23.5″N, 83°51′34.5″W); the geomorphology of both have been described by Ritter (2012). In addition, the transition area between the Till Plain and Drift Plain regions is associated with a marked increase in the width of the valley floor (Schiefer 2002; Ritter 2012) and it is part of the Little Miami River network dominated Open valley floodplain FPZ (Figure 2a).

The spatial organization of FPZs in the Little Miami River network displayed broad patterns, with FPZs spatially distributed in generally discrete areas of the stream network (Figure 2a). However, transitions between FPZs were not always orderly downstream, along the river network. All FPZ types transitioned into three to five different types downstream (Figure 3), indicating that downstream changes between FPZs are not predictable in this drainage network. The Open valley floodplain and Upland constrained zones most frequently transitioned into Lowland constrained zones downstream. Lowland constrained zones were most frequently followed by Lowland open valley zones, although transitions to Upland constrained zones were also common, indicating that zones with morphology more typical of upland areas can also occur further downstream influenced by regional physiography. Upland open valley low energy zones mainly transitioned into either Upland or Lowland constrained zones, reflecting the typical alternation between constrained and unconstrained valleys associated with Till and Drift Plain physiography. Lowland open valley zones were frequently followed by reservoirs downstream, while reservoirs were frequently followed by Upland constrained zones. Constrained zones dominated the most upstream segments of the network, representing about 90% of source segments.

Figure 3.

Figure 3

Open in a new tab

FPZ transitions within the Little Miami River drainage network. For each FPZ the adjacent downstream FPZ is given.

The composition of the different FPZs, in terms of the abundance, richness, and evenness of individual segments comprising each FPZ, varied (Figure 4). A summary of the character of the five FPZs is provided in Table 2. The most abundant FPZ was the Upland constrained zone, with a total channel length of 808.8 km, which represents 43.9% of the drainage network. This was followed by the Lowland constrained zone (511.5 km — 27.7%), the Lowland open valley zone (259.5 km — 14.1%), the Upland open valley lower energy zone (162.8 km — 8.8%), and the Open valley floodplain zone (100 km — 5.4%). In terms of the number of individual segments comprising each FPZ (richness), the Lowland constrained and the Upland constrained had 69 individual segments with an average segment length of 7.4 and 11.7 km, respectively. This was followed by the Lowland open valley, the Upland open valley lower energy, and the Open valley floodplain zones with 19, 14, and 4 segments that averaged 13.7 km, 11.6 km, and 25 km, respectively. Evenness values between the FPZs of the Little Miami River ranged between 0.808 and 0.935 (Table 1), hence most FPZs had individual segments that are similar in length. The Open valley floodplain zone had the highest evenness value (Figure 4), corresponding to a small number of very long segments associated with this FPZ.

Figure 4.

Figure 4

Open in a new tab

The character of diversity components of FPZs in the Little Miami River.

Table 2.

Composition of the five FPZs in the Little Miami River watershed

River type Description Abundance (%) Total channel length (km) Richness (no. individual segments) Evenness (Simpson’s evenness index) Diversity (Shannon–Weiner value)
1 Lowland open valley 14.1 259.5 19 0.808 2.38
2 Lowland constrained 27.7 511.5 69 0.925 3.91
3 Open valley floodplain 5.4 100.0 4 0.935 1.29
4 Upland constrained 43.9 808.8 69 0.868 3.67
5 Upland open valley lower energy 8.8 162.9 14 0.829 2.19
Open in a new tab

In terms of the overall diversity of FPZs in the Little Miami River, Shannon–Weiner values varied markedly. The Lowland constrained and Upland constrained had diversity values >3.0 (Table 2) while the Lowland open valley and Upland open valley lower energy zones had diversity values of 2.38 and 2.19, respectively. The Open valley floodplain zone had the lowest diversity value, with 1.29. In terms of the influence of the components of segment abundance, richness, and evenness on FPZ diversity, there are two compositional groups of FPZs (Figure 4). Based on the relative composition of the abundance, richness, and evenness of segments comprising each FPZ, the Lowland open valley, Lowland constrained, and the Upland open valley lower energy zones form one group. This group is characterized by the relatively similar influences of abundance, richness, and evenness on diversity. The second group comprises the Open valley floodplain and Upland constrained zones, which are dominated by higher evenness values compared to those of abundance and richness.

Delineation of the valley floor areas throughout a watershed can be used as a surrogate for floodplain area (Thoms et al. 2007). In the Little Miami River there is 363.8 km2 of floodplain, representing 8.2% of the watershed surface area. This area is distributed among 173 contiguous FPZ segments in the watershed (Table 3); however, the spatial distribution of floodplain areas is not evenly distributed among the five FPZs (Figure 5). The Upland open valley FPZ is associated with the largest combined floodplain area (132.4 km2 or 36.4% of the total floodplain area within the watershed). Floodplains associated with this FPZ are relatively small in area (mean = 1.92 km2) but numerous (n = 69). By comparison, the Open valley floodplain FPZ contains only 49.5 km2 or 13.6% of the total watershed floodplain area but this is comprised of four larger continuous floodplain segments, with a mean area of 12.39 km2 (Table 2). Overall, floodplain area in the Till Plain region (total area = 197.1 km2) of the watershed is slightly more than that of the Drift Plain region (total area = 166.7 km2).

Table 3.

The area of valley bottom/floodplain in the five FPZs of the Little Miami River watershed

FPZ Total floodplain area (km2) Number of individual floodplain segments Mean area per FPZ segment (km2) Percentage of watershed area that is floodplain
Little Miami River watershed 363.8 173 2.1 8.21
Lowland open valley 89.8 19 4.72 2.03
Lowland constrained 54.8 67 0.82 1.24
Open valley floodplain 49.5 4 12.39 0.84
Upland constrained 37.3 14 2.66 1.12
Upland open valley lower energy 132.4 69 1.92 2.98
Open in a new tab

Figure 5.

Figure 5

Open in a new tab

Distribution of valley floors, used as a surrogate for floodplains, in the Little Miami River watershed. Dashed line indicates boundary between Till and Drift Plains.

Discussion and Conclusions

There is an increasing number of approaches and data being used to characterize riverine landscapes (Kondolf et al. 2003; Milner et al. 2016). Technological advances and the increasing availability of high resolution data allow rivers to be mapped, and processes inferred at multiple scales. This has facilitated identifying and quantifying relationships among landscape patterns, anthropogenic disturbances and freshwater ecosystems as a rapidly emerging area of river science (Gilvear et al. 2016). The choice of what river characterization or classification scheme to employ will depend on the nature of the task or question (s) being asked. It is essential to recognize that river networks are hierarchical systems and according to hierarchy theory, each level of organization within the hierarchy is a discrete unit of the levels above and an agglomeration of lower levels. In river networks, each level of organization, whether it is a reach or site, can be distinguished by different rates of processing and morphological character (Dollar et al. 2007). Thus, higher levels in the river hierarchy provide a constraint on lower levels in the organization, especially that level immediately above the level under investigation, while lower levels in the organization have an upward influence through emergent properties. The recognition of top‐down constraints and bottom‐up influences has implications for undertaking river characterization (Milner et al. 2016). Characterization of entire river networks therefore dictates the resolution and the selection of variables. Thus, the suite of variables used will depend on the scale of characterization, for example the network, reach, or the microhabitat scale (Dollar et al. 2007). The scale of observation inflicts differing limitations on system structure, form, and function. Moreover, relationships between spatiotemporal scales and their influence on geomorphic and ecological functioning also depend on the purpose of the inquiry (Brierley and Fryirs 2005). For instance, dispersal mechanisms, predator–prey, and species interactions occur at a different spatial scale to geomorphic processes that govern channel morphology (Brierley and Fryirs 2005). A future challenge for river characterizations is to use the spatial scale of the observations and experiments relevant to the phenomena under investigation.

This study demonstrates an application of GIS hydrogeomorphic characterization for understanding and managing the structure and function of entire river networks, using the Little Miami River watershed as a case study. In contrast to traditional views of the structure of rivers as a continuum (e.g., Leopold et al. 1964; Vannote et al. 1980), this study shows the Little Miami River network to be structured as a series of large‐scale patches (FPZs), the spatial arrangement of which is not clinal and therefore does not support conventional river models. Other studies of river morphology in heavily glaciated landscapes have also shown the lack of a longitudinal continuum (e.g., Fitzpatrick et al. 2016). The results indicate that downstream change in the hydrogeomorphic character of river networks can be diverse and complex, rather than linear and predictable. Thus, conventional approaches are limited in their applicability for managing the distribution of riverine habitats and the ecosystem services provided by FPZs. Diversity measures have been effectively applied to understanding and managing the structure of ecological communities (Magurran 2004) and can similarly be applied to the study of FPZ assemblages in river networks as shown in this study. The hydrogeomorphic character of FPZs directly influences the habitats that occur within them (Thoms et al. 2007); thus, the abundance, richness, and evenness of FPZ types or “species” can inform about physical habitat diversity, which is important for ecological structure and function in river networks (Harris et al. 2009). These diversity measures provide information on potential or impending threats to the various FPZ types and the resources they provide. For example, FPZs with low abundance, such as the Open valley floodplain zone in the Little Miami River network, may indicate rare and endangered habitats that are at risk of being lost from the network. Conversely, FPZs with high abundance, such as the Upland constrained zone, that are widespread and common throughout the Little Miami River network, may be susceptible to fragmentation and a subsequent loss of connectivity throughout large parts of the network. The empirical approach provided in this study of the Little Miami River, via the use of diversity measures, adds to the traditional quantitative geomorphological approach first outlined by Strahler (1957).

The physical character of the Little Miami River watershed and channel network has been previously described from field‐based studies (Casey 1996; Debrewer et al. 2000; Schiefer 2002; Ritter 2012). Key morphological features of the Little Miami River network, such as the Clifton and John Bryan State Park gorges and the large floodplain systems, identified in field‐based studies are associated with the Upland constrained FPZ and Open valley floodplain FPZ, respectively. While a more detailed comparative study is required, these preliminary associations between field‐based observations and the GIS method outlined in this study, suggest a degree of similarity in the spatial organization of distinct river forms, and the applicability of the GIS approach to characterizing the stream network of the Little Miami River. Detailed comparisons of field‐based and GIS approaches for the physical characterization of river networks have been previously undertaken elsewhere (Thoms et al. 2007; Thorp et al. 2008). Most notable is the statistical comparison of the GIS approach used in this study and field‐based river data outlined in Thorp et al. (2008). Analyses support the fact that FPZs can be identified from a variety of data and methods. In the example of FPZs in Brungle Creek, Australia, Thorp et al. (2008) using two different approaches and datasets showed the same FPZ distribution with a top‐down approach which uses larger scale catchment, valley, and channel planform variables and a bottom‐up approach using smaller scale bankfull cross‐sectional variables. Based on the results of this study in the Little Miami River and studies undertaken in other river networks (cf. Thoms et al. 2007), it appears that the scale of FPZs integrates both the constraint and processes operating at higher level influences — catchment and valley scales — as well the influence of smaller scale process responsible for shaping the character of the bankfull river channel. Thus the approach outlined and applied in the Little Miami River adds to the increasing number of tools for river characterization.

The hydrogeomorphic character of river systems directly influences ecosystem structure and function by altering the spatial and temporal components of the habitat template (e.g., riffles, pools, wetted channels, slackwaters, and floodplains) at a range of scales within the riverine landscape (Thorp et al. 2008). Thus the spatial organization of FPZs dictates the availability, partitioning, and distribution of resources, as well as the provision of ecosystem services throughout a river network. Ecosystem structure, function, and services can vary significantly (and somewhat predictably) among FPZs (Thorp et al. 2008). River networks and watersheds are social–ecological systems, and their ability to provide a range of services is dependent on each component functioning and interacting properly (Bunch et al. 2011; Walker and Salt 2012). In other words, society depends on ecological systems to provide ecosystem services (Malinga et al. 2015 and references therein), and in turn, exploited ecological systems depend on the society to maintain them in a way that ensures their long‐term functioning (Berkes 2007). Examples of ecosystem services provided by different FPZs in river networks include flood mitigation, nutrient cycling, carbon sequestration, and recreation (Thorp et al. 2010). Characterization of FPZs using variables at appropriate scales and with an emergent nomenclature framework, using multivariate statistics, enables FPZs to be delineated and classified with relevance to their ecosystem structure, function, and services. Knowledge of the distribution of particular ecosystem services throughout river networks is important for effective management, as is knowledge of the past, present, and future river conditions. Delineation of FPZs can aid in mapping and predicting where particular ecosystem services are provided, as well as in the management of these services and discussions of asset trading. Applications of river network characterization, such as that provided in this paper, to management and conservation at the watershed scale are gaining momentum (cf. Milner et al. 2016).

Conservation and management of riverine landscapes entails a combination of monitoring, assessment, restoration, and rehabilitation activities. The effectiveness of these activities can be increased by better accounting for hydrogeomorphic variability within and across river networks. For example, river monitoring and assessment designs are typically: (1) unit‐based (e.g., samples per a set distance); (2) stratified by some natural or anthropogenic feature of the river (e.g., above and below reservoirs) or landscape (e.g., ecoregions or political boundaries); and (3) randomly distributed (Flotemersch et al. 2006). These approaches do not effectively account for observed hydrogeomorphic variability among assessment units. Delineation of FPZs within a river network accounts for this variability, thus supporting the development of more efficient and effective sampling designs and informative assessments. Furthermore, a “reference” condition is often required, or at a minimum, desired for use as a benchmark for conservation and management activities (Stoddard et al. 2006). A problem can arise if the site used as reference is not hydrogeomorphically analogous to the site of interest. The delineation of FPZs offers a scientifically defensible method for the characterization of river sections that facilitates comparison to other sections of river that are equivalent in both structure and function. These “comparable” sections may be within the same river network or in another. Regardless, the ability to make a defensible case for a given reference is improved; and thus the likelihood of positive research or management outcomes is increased.

Knowledge of the spatial distribution of ecosystem services is important for the management of natural resources, particularly in the context of asset trading at the broader landscape scale (Naidoo et al. 2008). FPZs provide an array of ecosystem services and resources throughout river networks (Thorp et al. 2010) and some FPZs may have a higher “value” for particular services they provide (e.g., food and fiber production), whereas others may have limited value for that service, but higher value for other services (e.g., transportation). For example, floodplains are physical features of river networks that provide a suite of services to humans, including water supply, disturbance regulation, waste treatment, and cultural and recreational amenities. Floodplains are among the most diverse and dynamic features of riverine landscapes (Tockner and Stanford 2002; Thoms 2003) and the monetary value of the ecosystem services they provide to humans was estimated at over $25,000 per hectare per year in 2011 — twice the estimated unit value of rivers and lakes (Costanza et al. 2014). However, the spatial distribution of floodplains, and thus their ecosystem services and resources, has traditionally been viewed under the river continuum perspective, which has emphasized the accumulation of floodplain resources in downstream reaches of a river network. This study quantified the spatial distribution of floodplain resources and the results show that floodplains are distributed patchily through the Little Miami River network and unevenly among the FPZs identified, which contrasts traditional perspectives. Current river models (e.g., the RCC of Vannote et al. 1980) hypothesize an increase in floodplain area along river networks, with the lowermost regions being the dominant floodplain zone.

Heterogeneity in the spatial distribution of floodplain resources throughout river networks has implications for research and management. The disturbance buffering capacity provided by floodplains is an important but often overlooked service provided by these riverine landscape features (Costanza et al. 1997). When floodplains are distributed diversely throughout a river network, the buffering capacity and hydrological stability of the network at the watershed scale is increased compared to when floodplains are accumulated in one part of the network (Moore et al. 2015); for example, in the lower reaches. Bulk removal of floodplains in one area may have disproportionately large impacts downstream because each part of the network plays an important role in dynamics at the watershed scale (Moore et al. 2015). In addition, different FPZs will require different management in relation to floodplain ecosystem services. For example, the Lowland constrained and Open valley floodplain FPZs contain approximately the same total area of floodplain throughout the Little Miami River network; however, floodplain segments in the former FPZ are frequent in number but an order of magnitude smaller in area than floodplain segments in the latter (Table 2). Thus, ecosystem services provided by floodplains are limited in their spatial distribution within Lowland constrained FPZ segments and removal of floodplain from this FPZ may have disproportionately large effects on this already limited resource.

The FPZ characterization approach applied to the Little Miami River can be used to support trading of such assets in both space and time. For example, levee construction to facilitate transportation and protect against flooding can disconnect a river channel from its adjacent floodplain features and inhibit its ability to process nutrients (Barth and Veizer 1999). The consequences of resultant excess nutrient export downstream can be offset by reconnecting floodplain areas in other parts of the network, maintaining the network’s overall ability to process nutrient inputs from the watershed. Such is an example of asset trading. Two of the components necessary for a fair basis of trading are: (1) a regional or national framework for classifying rivers at appropriate scales (including at least the valley‐to‐reach scale of FPZs); and (2) an understanding of the link between river character and resulting differences in ecosystem structure and function for the river sections being considered. Frameworks for characterizing rivers and understanding their likely responses to disturbances are emerging (cf. Dollar et al. 2007; Thorp et al. 2008). However, our ability to predict ecosystem responses within and between FPZs to a range of single and multiple drivers remains limited (Ormerod et al. 2010; Sutherland et al. 2013). Increasing our knowledge of, and developing tools for, characterizing rivers at multiple scales, especially those for FPZs, will assist in providing better empirical models for linking the complexity of FPZs within river networks and their ecosystem structure, function, and ability to absorb disturbances. By doing so, it will effectively support river management applications, as discussed in this study.

Acknowledgments

MCT was supported by a Senior Fulbright Fellowship while undertaking part of this research. The U.S. Environmental Protection Agency through its Office of Research and Development funded and managed part of the research described here. This research was also supported in part by an appointment for MWS to the Oak Ridge Institute for Science and Education Research Participation Program supported by an interagency agreement between the U.S. Environmental Protection Agency (USEPA) and the U.S. Department of Energy. The findings and conclusions of this research may not necessarily reflect the views of the USEPA and no official endorsement should be inferred. The comments by three anonymous referees and the Associate Editor on an earlier version of this manuscript are gratefully acknowledged.

References

  1. Allen TFH, and Starr TB. 1982. Hierarchy: Perspectives for Ecological Diversity. Chicago, IL: University of Chicago Press. [Google Scholar]
  2. Arthington A, Olden J, Balcombe S, and Thoms MC. 2010. “Multi‐Scale Environmental Factors Explain Fish Losses and Refuge Quality in Drying Waterholes of Cooper Creek, an Australian Arid‐Zone River” Marine and Freshwater Research 61: 842–56. . [Google Scholar]
  3. Barth JA, and J Veizer. 1999. “Carbon Cycle in St. Lawrence Aquatic Ecosystems at Cornwall (Ontario), Canada: Seasonal and Spatial Variations.” Chemical Geology 159 (1): 107–28. [Google Scholar]
  4. Belbin L 1993. “Environmental Representativeness: Regional Partitioning and Reserve Selection.” Biological Conservation 66: 223–30. [Google Scholar]
  5. Belbin L, and McDonald C. 1993. “Comparing Three Classification Strategies for Use in Ecology.” Journal of Vegetation Science 4: 341–48. [Google Scholar]
  6. Benda L, Poff NL, Miller D, Dunne T, Reeves G, Pess G, and Pollock M. 2004. “The Network Dynamics Hypothesis: How Channel Networks Structure Riverine Habitats.” BioScience 54 (5): 413–27. [Google Scholar]
  7. Berkes F 2007. “Understanding Uncertainty and Reducing Vulnerability: Lessons from Resilience Thinking.” Natural Hazards 41: 283–95. 10.1007/s11069-006-9036-7. [DOI] [Google Scholar]
  8. Brierley GJ, and Fryirs KA. 2005. Geomorphology and River Management: Applications of the River Styles Framework. Oxford: Blackwell Publications. [Google Scholar]
  9. Bunch MJ, Morrison KE, Parkes MW, and Venema HD. 2011. “Promoting Health and Well‐Being by Managing for Social–Ecological Resilience: The Potential of Integrating Ecohealth and Water Resources Management Approaches.” Ecology and Society 16:6 http://www.ecologyandsociety.org/vol16/iss1/art6/. [Google Scholar]
  10. Cairns Jr J. , and Crawford TV (Editors). 1991. Integrated Environmental Management. Chelsea, MI: Lewis Publishers, Inc. [Google Scholar]
  11. Casey GD 1996. “Hydrogeologic Framework of the Midwestern Basins and Arches Region in Parts of Indiana, Ohio, Michigan, and Illinois.” U.S. Geological Survey Professional Paper 1423‐B, p. 46. [Google Scholar]
  12. Collins SE, Thoms MC, and Flotemersch JA. 2014. “Hydrogeomorphic Zones Characterize Riverbed Sediment Patterns Within a River Network?” River Systems 21: 203–13. 10.1127/1868-5749/2014/0084. [DOI] [Google Scholar]
  13. Costanza R, d’Arge R, de Groot R, Farber S, Grasso M, Hannon B, Limburg K, Naeem S, O’Neill RV, Paruelo J, Raskin RG, Sutton P, and van den Belt M. 1997. “The Value of the World’s Ecosystem Services and Natural Capital.” Nature 387: 253–60. [Google Scholar]
  14. Costanza R, de Groot R, Sutton P, van der Ploeg S, Anderson SJ, Kubiszewski I, Farber S, and Turner RK. 2014. “Changes in the Global Value of Ecosystem Services.” Global Environmental Change 26: 152–58. [Google Scholar]
  15. Debrewer LM, Rowe GL, Reutter DC, Moore RC, Hambrook JA, and Baker NT. 2000. “Environmental Setting and Effects on Water Quality in the Great and Little Miami River Basin, Ohio and Indiana.” U.S. Geological Survey National Water‐Quality Assessment Program, Water Resources Investigations Report 99‐4201, p. 110. [Google Scholar]
  16. Delong MD, and Thoms MC. 2016. “An Ecosystems Framework for River Science, Management and Policy” In River Science, Management and Policy for 21st Century, edited by Gilvear DA, Greenwood MW, Thoms MC, and Wood PA, 15–36. Chichester, United Kingdom: Wiley. [Google Scholar]
  17. Dollar ESJ, James CS, Rogers KH, and Thoms MC. 2007. “A Framework for Interdisciplinary Understanding of Rivers as Ecosystems.” Geomorphology 89:147–62. 10.1016/j.geomorph2006.07.022. [DOI] [Google Scholar]
  18. Fitzpatrick FA, Ellison CA, Czuba CR, Young BM, and McCool MN. 2016. “Geomorphic Responses of Duluth‐Area Streams to the June 2012 Flood, Minnesota.” U.S. Geological Survey Scientific Investigations Report 2016‐5105, p. 53. [Google Scholar]
  19. Flotemersch JE, Stribling JB, and Paul MJ. 2006. “Concepts and Approaches for the Bioassessment of Non‐Wadeable Streams and Rivers” EPA 600‐R‐06‐127. Cincinnati, OH: U.S. Environmental Protection Agency, p. 154. [Google Scholar]
  20. Forrest J, and Clark NR. 1989. “Characterising Grain Size Distributions: Evaluation of a New Approach Using a Multivariate Extension of Entropy Analysis.” Sedimentology 36: 711–22. [Google Scholar]
  21. Gesch D, Evans G, Mauck J, Hutchinson J, and Carswell WJ. 2009. “The National Map — Elevation.” U.S. Geological Survey Fact Sheet 2009‐3053, p. 4. [Google Scholar]
  22. Gilvear DJ, Greenwood MW, Thoms MC, and Wood PA. 2016. River Science, Management and Policy for 21st Century. Chichester, United Kingdom: Wiley. [Google Scholar]
  23. Harris CD, Thoms MC, and Scown MW. 2009. “The Ecohydrology of Stream Networks.” International Association of Hydrological Sciences 328: 127–36. [Google Scholar]
  24. Hull RB, Richert D, Seekamp E, Robertson D, and Buhyoff GJ. 2003. “Understandings of Environmental Quality: Ambiguities and Values Held by Environmental Professionals.” Environmental Management 31: 1–13. [DOI] [PubMed] [Google Scholar]
  25. Karr JR 1996. “Ecological Integrity and Ecological Health Are Not the Same Thing” In Engineering Within Ecological Constraints, edited by Schulze P, 97–110. Washington, D.C: National Academy of Science. [Google Scholar]
  26. KARS (Kansas Applied Remote Sensing). 2015. “FLDPLN Model Software Development” Lawrence, KS: Kansas Biological Survey, University of Kansas; http://kars.ku.edu/research/fldpln-model-software-development/. [Google Scholar]
  27. Kenney DS 1997. Resource Management at the Watershed Level: An Assessment of the Changing Federal Role in the Emerging Era of Community‐Based Watershed Management. Denver, CO: Western Water Policy Review Advisory Commission. [Google Scholar]
  28. Kondolf GM, Montgomery DR, Piégay H, and Schmitt L. 2003. “Geomorphic Classification of Rivers and Streams” In Tools in Fluvial Geomorphology, edited by Kondolf GM and Piégay H, 171–204. Chichester, United Kingdom: Wiley. [Google Scholar]
  29. Kreiling RM, Thoms MC, and Richardson WR. 2018. “Beyond the Edge: Linking Agricultural Landscapes, Stream Networks, and Best Management Practices.” Journal of Environmental Quality 47 (1): 42–53. 10.2134/jeq2017.08.0319. [DOI] [PubMed] [Google Scholar]
  30. Leopold LD, Wolman MG, and Miller JP. 1964. Fluvial Processes in Geomorphology. New York: W.H. Freeman and Company. [Google Scholar]
  31. Likens GE, Walker KF, Davies PE, Brookes J, Olley J, Young W, Thoms MC, Lake PS, Davis J, Arthington A, Thompson R, and Oliver RL. 2009. “Ecosystem Science: Toward a New Paradigm for Managing Australia’s Inland Aquatic Ecosystems.” Marine and Freshwater Research 60: 271–79. [Google Scholar]
  32. Magurran A 2004. Measuring Biological Diversity. Oxford: Blackwell Science Ltd. [Google Scholar]
  33. Malinga R, Gordon LJ, Jewitt G, and Lindborg R. 2015. “Mapping Ecosystem Services Across Scales and Continents — A Review.” Ecosystem Services 13: 57–63. 10.1016/j.ecoser.2015.01.006. [DOI] [Google Scholar]
  34. Milner VS, Gilvear DJ, and Thoms MC. 2016. “Characterisation Riverine Landscapes; History, Applications and Future Challenges” In River Science, Management and Policy for 21st Century, edited by Gilvear DA, Greenwood MW, Thoms MC, and Wood PA, 239–58. Chichester, United Kingdom: Wiley. [Google Scholar]
  35. Milner VS, Wilby NJ, Gilvear DJ, and Perfect C. 2015. “Linkages between Reach‐Scale Physical Habitat and Invertebrate Assemblages in Upland Streams.” Marine and Freshwater 66: 438–48. [Google Scholar]
  36. Mitchell B (Editor). 1990. Integrated Water Management: International Experiences and Perspectives. London: Belhaven Press. [Google Scholar]
  37. Montgomery DR 1999. “Process Domains and the River Continuum.” Journal of the American Water Resources Association 35 (2): 397–410. [Google Scholar]
  38. Moore JW, Beakes MP, Nesbitt HK, Yeakel JD, Patterson DA, Thompson LA, Phillis CC, Braun DC, Favaro C, Scott D, and Carr‐Harris C. 2015. “Emergent Stability in a Large, Free‐Flowing Watershed.” Ecology 96 (2): 340–47. [DOI] [PubMed] [Google Scholar]
  39. Nagel DE, Peterson E, Isaak DJ, Ver Hoef JM, and Horan DL. 2015. “National Stream Internet Protocol and User Guide” U.S. Forest Service, Rocky Mountain Research Station Air, Water, and Aquatic Environments Program, Boise, ID. [Google Scholar]
  40. Naidoo R, Balmford A, Costanza R, Fisher B, Green RE, Lehner B, Malcolm TR, and Ricketts TH. 2008. “Global Mapping of Ecosystem Services and Conservation Priorities.” Proceedings of the National Academy of Sciences of the United States of America 105 (28): 9495–500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Noss RF 1995. “Ecological Integrity and Sustainability: Buzzwords in Conflict?” In Perspectives on Ecological Integrity, edited by Westra L and Lemons J, 60–76. Dordrecht: Kluwer Academic Publishers. [Google Scholar]
  42. Ohio EPA. 2009. Biological and Water Quality Study of the Lower Little Miami River and Selected Tributaries: Including the Todd Fork Sub‐Watershed. Columbus, OH: EPA Division of Water Quality Monitoring and Assessment, Surface Water Section; http://www.epa.state.oh.us/portals/35/documents/LowerLittleMiamiRiverTSD2007.pdf [Google Scholar]
  43. Omernik JM 1987. “Ecoregions of the Conterminous United States.” Annals of the Association of American Geographers 77:118–25. [Google Scholar]
  44. Ormerod SJ, Dobson M, Hildrew AG, and Townsend CR. 2010. “Multiple Stressors in Freshwater Ecosystems.” Freshwater Biology 55: 1–4. [Google Scholar]
  45. Parsons M, Thoms MC, Flotemersch J, and Reid MA. 2016. “Monitoring the Resilience of Rivers as Social–Ecological Systems: A Paradigm Shift for River Assessment in the Twenty‐First Century” In River Science, Management and Policy for 21st Century, edited by Gilvear DA, Greenwood MW, Thoms MC, and Wood PA, 197–220. Chichester, United Kingdom: Wiley. [Google Scholar]
  46. Poole GC 2002. “Fluvial Landscape Ecology: Addressing Uniqueness Within the River Discontinuum.” Freshwater Biology 47 (4): 641–60. [Google Scholar]
  47. PRISM Climate Group. 2004. PRISM Gridded Climate Data. Corvallis, OR: Oregon State University; http://prism.oregonstate.edu. [Google Scholar]
  48. Renwick WH 1996. “Continental‐Scale Reservoir Sedimentation Patterns in the United States.” The International Association of Hydrological Sciences 236: 513–22. [Google Scholar]
  49. Ritter JB 2012. “Temporal and Spatial Scales of Stream Study, West‐Central Ohio.” The Geological Society of America, Field Guide 27: 33–53. [Google Scholar]
  50. Schiefer MC 2002. “Basin Descriptions and Flow Characteristics of Ohio Streams.” Bulletin 47,Ohio Department of Natural Resources, Division of Water, 161 pp. [Google Scholar]
  51. Schumm SA 1977. The Fluvial System. New York: Wiley. [Google Scholar]
  52. Simley JD, and Carswell WJ Jr. 2009. “The National Map — Hydrography.” U.S. Geological Survey Fact Sheet 2009‐3054, p. 4. [Google Scholar]
  53. Steel EA, Hughes RM, Fullerton AH, Schmutz S, Young JA, Fukushima M, Muhar S, Poppe M, Feist BE, and Trautwein C. 2010. “Are We Meeting the Challenges of Landscape‐Scale Riverine Research? A Review.” Living Reviews Landscape Research 4: 5–36. [Google Scholar]
  54. Stoddard JL, Larsen DP, Hawkins CP, Johnson RK, and Norris RH. 2006. “Setting Expectations for the Ecological Condition of Streams: The Concept of Reference Condition.” Ecological Applications 16: 1267–76. 10.1890/1051-0761(2006)016[1267:SEFTEC]2.0.CO;2. [DOI] [PubMed] [Google Scholar]
  55. Strahler AN 1957. “Quantitative Analysis of Watershed Geomorphology.” Transactions of the American Geophysical Union 38: 913–25. [Google Scholar]
  56. Sutherland WJ, Freckleton RP, Godfray HCJ, Beissinger SR, Benton T, Cameron DD, Carmel Y, Coomes DA, Coulson T, Emmerson MC, Hails RS, Hays GC, Hodgson DJ, Hutchings MJ, Johnson D, Jones JPG, Keeling MJ, Kokko H, Kunin WE, Lambin X, Lewis OT, Malhi Y, Mieszkowska N, Milner‐Gulland EJ, Norris K, Phillimore AB, Purves DW, Reid JM, Reuman DC, Thompson K, Travis JMJ, Turnbull LA, Wardle DA, and Wiegand T. 2013. “Identification of 100 Fundamental Ecological Questions.” Journal of Ecology 101:58–67. [Google Scholar]
  57. Thoms MC 2003. “Floodplain–River Ecosystems: Lateral Connections and the Implications of Human Interference.” Geomorphology 56: 335–49. [Google Scholar]
  58. Thoms MC, Delong MA, Collins SE, and Flotemersch JH. 2017. “Physical Heterogeneity and Aquatic Community Function in River Networks: A Case Study of the Kanawha River, USA.” Geomorphology 290: 277–87. [Google Scholar]
  59. Thoms MC, and Parsons ME. 2002. “Ecogeomorphology: An Interdisciplinary Approach to River Science.” The International Association of Hydrological Sciences 276: 113–19. [Google Scholar]
  60. Thoms MC, and Parsons ME. 2003. “Identifying Spatial and Temporal Patterns in the Hydrological Character of the Condamine‐Balonne River, Australia, Using Multivariate Statistics.” River Research and Applications 19 (5–6): 443–57. [Google Scholar]
  61. Thoms MC, Parsons ME, and Southwell MA. 2016. “The Physical Template of Australia’s Rivers” In Vegetation of Australian Riverine Landscapes: Biology, Ecology and Management, edited by Capon S, James C, and Reid MA, 5–26. Melbourne: CSIRO. [Google Scholar]
  62. Thoms MC, Rayburg S, and Neave M. 2007. “The Physical Diversity and Assessment of a Large River System: The Murray‐Darling Basin, Australia” In Large Rivers, edited by Gupta A, 587–608. Chichester, United Kingdom: Wiley. [Google Scholar]
  63. Thorp JH, Flotemersch JE, Delong MD, Casper AF, Thoms MC, Ballantyne F, Williams BS, O’Neill BJ, and Haase CS. 2010. “Linking Ecosystem Services, Rehabilitation, and River Hydrogeomorphology.” BioScience 59 (1): 67–74. [Google Scholar]
  64. Thorp JH, Thoms MC, and Delong MD. 2006. “The Riverine Ecosystem Synthesis: Biocomplexity in River Networks Across Space and Time.” River Research and Applications 22 (2): 123–47. [Google Scholar]
  65. Thorp JH, Thoms MC, and Delong MD. 2008. The Riverine Ecosystem Synthesis: Towards Conceptual Cohesiveness in River Science. San Diego, CA: Elsevier Academic Press. [Google Scholar]
  66. Tockner K, and Stanford JA. 2002. “Riverine Flood Plains: Present State and Future Trends.” Environmental Conservation 29: 308–30. [Google Scholar]
  67. USGS (United States Geological Survey). 2015. “Geologic Maps of US States.” http://mrdata.usgs.gov/geology/state/..
  68. Vannote RL, Minshall GW, Cummins KW, Sedell JR, and Cushing CE. 1980. “The River Continuum Concept.” Canadian Journal of Fisheries and Aquatic Sciences 37 (1): 130–37. [Google Scholar]
  69. Walker B, and Salt D. 2012. Resilience Practice: Building Capacity to Absorb Disturbance and Maintain Function. Washington, D.C.: Island Press. [Google Scholar]
  70. Williams BS, D’Amico E, Kastens JH, Thorp JH, Flotemersch JE, and Thoms MC. 2013. “Automated Riverine Landscape Characterization: GIS‐Based Tools for Watershed‐Scale Research, Assessment, and Management.” Environmental Monitoring and Assessment 185: 7485–99. 10.1007/s10661-013-3114-6. [DOI] [PubMed] [Google Scholar]

RESOURCES