An introduction to hierachical frameworks for assessing the hydromorphology of river systems

This paper introduces this special issue of Aquatic Sciences by outlining the multi-scale, hierarchical framework that has been developed for improving hydromorphological understanding and informing management of rivers, particularly in a European context. Here the term hydromorphology, which is used widely within Europe, refers to the suite of hydrological and geomorphological processes and forms that occur within catchments and their river systems. This paper provides the rationale behind the development of the framework and briefly overviews its structure and key features including the way that it supports understanding of the hydromorphological behaviour of river reaches in response to temporal changes at catchment to reach scales. It also refers to other papers within this special issue that provide more details on particular aspects of the framework or that illustrate the framework’s application.

River management often focuses on individual reaches of river networks, aiming to improve their ability to support human needs and those of the river ecosystem. However, the form, sedimentary and vegetation structure, dynamism and behaviour of river reaches depend not only upon natural processes and human interventions within the reach but also within the wider catchment. Furthermore, the response of river reaches to changes in processes and human pressures across the catchment is often delayed. This is because it takes time for the effects of changes (e.g., land cover change, dam construction) to propagate from their initial location across catchments and through river networks to individual river reaches. Thus, understanding of reach scale hydromorphology requires knowledge of processes and human pressures at not only the reach scale but at larger spatial scales including the catchment scale. Since human interventions or pressures at one location and time may induce responses at one or more other locations and times, such knowledge needs to relate to both current and past pressures and processes. Without such a spatial and temporal understanding, management interventions cannot be fully informed and so may not be sustainable and may potentially require significant ongoing maintenance.

In response to this complexity, researchers have developed many spatially-hierarchical frameworks to support better understanding of the functioning of river catchments, networks and corridors. These have been developed with a variety of scientific and management purposes in mind. Several authors have reviewed this topic (e.g., Naiman et al. 1992; Kondolf et al. 2003) and a selection of 16 examples of hierarchical frameworks, some specifically focussed on hydromorphology, some with a broader ecological focus, are briefly described in Table 1. These examples illustrate a range of different frameworks for developing understanding or assessing river systems by organizing and interpreting information across a hierarchy of spatial scales. Many frameworks incorporate formal classifications of spatial units such as river reaches or segments (i.e., the units are assigned to distinct categories or classes based on specific attributes). Where frameworks incorporate such classifications, they are briefly described in Table 1. The following generalisations can be drawn from the example frameworks listed in Table 1:

Table 1 Examples of spatially-hierarchical frameworks to support better understanding of the functioning of river catchments, corridors and networks
Full size table
  1. 1.

    Despite its early publication date, the work of Frissell et al. (1986) continues to present the most comprehensive conceptual multi-scale framework for investigating streams and habitats. The spatial units are delineated so that units at smaller spatial scales nest within those at larger spatial scales. The framework incorporates hydromorphological processes and forms and vegetation at all spatial scales in relation to their influence on habitat. Time scales of persistence or adjustment are associated with spatial units at each scale. Indicators of form and process are suggested for spatial units at each scale. The role of the indicators is explained in terms of developing understanding of the functioning of spatial units and the process linkages among units and scales. Although no formal classifications of spatial units are proposed, the way in which indicators could contribute to classification is discussed. All of the methods described in Table 1 consider a hierarchy of spatial units, but the degree to which they develop the other aspects of the conceptual approach proposed by Frissell et al. (1986) varies widely.

  2. 2.

    Many of the frameworks focus entirely on hydromorphological processes and forms that are either directly measured or inferred. This is because interactions between processes and forms control the dynamic morphology or behaviour of rivers and their mosaics of habitats. Hydromorphological processes drive longitudinal and lateral connectivity within river networks and corridors, the assemblage and turnover of physical habitats, and the sedimentary and vegetation structures associated with those habitats.

  3. 3.

    Some frameworks are conceptual, providing a way of thinking about or structuring analyses of river systems, and interpreting their processes, morphology and function (e.g., Frissell et al. 1986; Habersack 2000; Fausch et al. 2002; Thorp et al. 2006; Beechie et al. 2010; McCluney et al. 2014). Some frameworks are more quantitative, generating one or more indices or classifications of spatial units that support assessment of river systems (e.g., Rosgen 1994; González del Tánago and García de Jalón 2004; Merovich et al. 2013; Rinaldi et al. 2013, 2015a). However, some frameworks follow an intermediate course, generating relatively open-ended indices or classes that can be interpreted flexibly (e.g., Brierley and Fryirs 2005).

  4. 4.

    Time scales and temporal changes are not included in all frameworks, particularly where the framework is proposed as an input to further assessment or analysis (e.g., Snelder and Biggs 2002, González del Tánago and García de Jalón 2004). A time scale is included as a dimension of each spatial scale in some approaches (e.g., Habersack 2000; Dollar et al. 2007), whereas others incorporate historical analyses that track human interventions or changes in units through time at some spatial scales (e.g., Rosgen 1994; Montgomery and MacDonald 2002; Brierley and Fryirs 2005; Beechie et al. 2010; Rinaldi et al. 2013, 2015a). In some cases, theoretical or historical analyses or consideration of specific future scenarios are used to develop space–time understanding that can support management decision-making (e.g., Montgomery and Buffington 1997, 1998; Montgomery and MacDonald 2002; Benda et al. 2004; Brierley and Fryirs 2005; McCluney et al. 2014).

  5. 5.

    Although all frameworks incorporate characteristics that are used to delineate spatial units and may indicate how those units function, many provide specific, well-defined indicators of processes, forms or of the condition of spatial units (e.g., Rosgen 1994; Montgomery and Buffington 1997, 1998; Montgomery and MacDonald 2002; Benda et al. 2004; Brierley and Fryirs 2005; Merovich et al. 2013; Rinaldi et al. 2013, 2015a). Furthermore, some of the frameworks include indicators of human pressures and their impacts (e.g., Merovich et al. 2013; McCluney et al. 2014; Rinaldi et al. 2013, 2015a).

  6. 6.

    Finally, although most frameworks could be described as incorporating processes to some degree, some methods are particularly process-based, even when processes are inferred from forms and associations rather than being quantified by direct measurements. Frameworks that consider temporal dynamics and trajectories of historical change (see point 4, above) are particularly effective in developing understanding of processes and the impacts of changed processes cascading through time and across spatial scales.

Although the list of frameworks presented in Table 1 is far from comprehensive, it illustrates that different types of hierarchical framework have been proposed. These previous frameworks have provided a foundation for developing the multi-scale, hierarchical framework for the hydromorphological assessment of European rivers that is described in this paper. This REFORM framework was developed within the REFORM project, which is funded by the European Commission with the aim of supporting sustainable river management and restoration. It has been developed to fit into the context of the European Union’s Water Framework Directive (WFD; European Commission 2000), which constitutes the principal legal instrument for managing and restoring aquatic ecosystems within member states of the European Union, and so it is intended for application by river managers. The following sections of this paper introduce the REFORM framework and describe its key properties; briefly describe the application of the framework; and then introduce this special issue by referring to other papers that provide further details on particular aspects of the framework and its application.

The REFORM framework

The REFORM framework is informed by many previous frameworks (Table 1). Those of Frissell et al. (1986), Montgomery and Buffington (1997, 1998), Habersack (2000), Brierley and Fryirs (2005) and Rinaldi et al. (2013) have been particularly influential. Nevertheless, the REFORM framework has several properties that in combination differentiate it from its predecessors and suit it to application by river managers working in the environmental contexts for which it has been developed.

  1. 1.

    Because the aim of the research was to develop a tool for use by river managers, the framework has been kept as simple to apply as was felt possible. It is a hydromorphological framework which includes relevant information on vegetation.

  2. 2.

    Reflecting the long history of human interventions on European rivers, the framework incorporates human pressures as well as natural processes and forms at all included spatial scales and gives them equal weighting.

  3. 3.

    The framework is open-ended to the extent that European member states can incorporate their own data sets, methods and modelling tools, although specific methods have been proposed and fully-described for consideration by member states. This open-ended nature ensures the framework’s relevance to all member states, and thus maximizes the potential for its process-based ‘way of thinking’ to be widely adopted. It also ensures that elements of the framework methodology can be adapted to local circumstances, reflecting the enormous variety of river environments and data sets found within Europe.

  4. 4.

    The framework includes spatial units at region, catchment, landscape unit, segment, reach, geomorphic unit, hydraulic unit and river element (i.e., patch of sediment, plant stand etc.) scales. However, the core scales are those ranging from catchment to geomorphic unit. Each spatial unit has an indicative temporal scale of persistence/adjustment, but the main temporal element of the framework is a historical analysis of available data sets. A definition of each spatial scale and associated indicative space and time scales are provided in Table 2.

    Table 2 Spatial units included within the REFORM framework: descriptions, indicative time and space scales, delineation criteria
    Full size table
  5. 5.

    The key scale of the framework is the river reach, since this is the scale at which rivers are most often assessed, managed and rehabilitated. A central and unique feature of the REFORM framework is that all reaches are classified into ‘river types’ using clearly-defined, simple criteria. All other elements of the framework are directed at understanding the naturalness or artificiality of these reaches and their types, the processes to which they are subjected, and their morphodynamic behaviour. This involves assessment of (1) the cascade of processes affecting reaches from catchment to reach scales, (2) the degree to which reaches display characteristics at reach and finer scales that are indicative of ‘natural’ function according to their type or of ‘artificiality’, and (3) the ways in which reach morphology has changed or behaved through time in response to changes in processes and direct human interventions at catchment to reach scales. To fit with the long history of human pressures on European rivers, and thus the fact that there is no time in the past for which detailed information is available that can be considered to represent ‘pristine’ conditions, the character of the river in the past is not considered as a ‘reference condition’ that refers to a ‘pristine state’. Instead, the entire space–time analysis assesses the degree to which the morphodynamic behaviour of some river reaches suggests that they are functioning or have functioned in a relatively natural way. This analysis provides process-form information that can inform management of more impacted reaches of otherwise similar type.

  6. 6.

    Recommendations are made on how to delineate spatial units and how processes, forms and human pressures can be represented by indicators. Tables 2 and 3, respectively, provide brief summaries of the properties used for delineation, and the purpose and types of indicators that are estimated. The reach type is the key indicator.

    Table 3 Examples of indicators and the processes they indicate at catchment to reach scales of the REFORM framework (for further details see González del Tánago et al. (2015a))
    Full size table
  7. 7.

    Indicators support the assessment of human pressures, processes, and morphological responses at each spatial scale. They also support the assessment of the past and present behaviour of river reaches and their riparian zones in terms of changes in their form and function in response to changes in processes and human pressures from catchment to reach scales.

  8. 8.

    Space–time understanding of catchments and their river systems is developed from the indicators and provides a basis for estimating potential reach-scale adjustments to future changes across the spatial units (e.g., climate change, land cover change, introduction or removal of channel reinforcement or structures). Such analyses also help to identify whether or not the river type initially defined by simple rules corresponds to the river type that might function most effectively at a given location or whether a different type is more appropriate, so informing the design of any proposed restoration.

Application of the REFORM framework requires a significant data resource. Measurements at the hydraulic unit and river element scales are not widely available. However, collection of such data by purpose-designed field survey contributes to monitoring specific reaches where detail is needed to track changes, particularly following management interventions. Information at all of the other spatial scales can be obtained from national surveys and analyses such as physical habitat surveys, riparian habitat surveys, morphological surveys and hydrological regime assessments (Belletti et al. 2015a); climate, river flow and groundwater data sets; and national scale mapping of, for example, geology, soils and vegetation. Furthermore, many relevant data sets are available at a European scale (Table 4, see also the paper by Bizzi et al. 2015 in this special issue). While contemporary and recent data sets are usually easy to obtain, historical information may be more restricted (for a recent review see Grabowski et al. 2014).

Table 4 Pan European data sources that are mainly freely available and can support delineation and characterisation of spatial units
Full size table

Reflecting the purpose of the application, data availability, and the combination of cost, time and effort that is available, the REFORM framework can be applied in different ways. For catchment assessment and management purposes, the aim should be to sub-divide the entire catchment into a complete set of catchment to reach scale units, and, at a minimum, to include geomorphic units as attributes of each reach. In this way, the assemblage of reach types and their properties can be placed within a catchment and river network context. However, in large catchments, it may not be possible to compile information on a complete set of units for the entire catchment. Under these circumstances, it is necessary to sub-divide the catchment to the scale of its major landscape units, and then isolate representative sub-catchments within each landscape unit where segments and reaches along the main channel and major tributaries can be analysed. In this way, an analysis of the properties of different reach types can be investigated within sub-catchments that are representative of the catchment’s landscape units. If the purpose is to focus on a particular reach or segment, perhaps in the context of designing an intervention or rehabilitation, the assessment still needs to focus on spatial units that contain and are immediately upstream of the reach or segment under consideration so that the processes affecting the reach can be investigated.

Overview of the application of the REFORM framework

Application of the framework involves three main stages:

  1. 1.

    Delineation of the spatial units;

  2. 2.

    Assembly of available data sets to Characterise the spatial units so that Indicators of processes, forms and human pressures can be extracted for units across the spatial scales to represent their present and past state;

  3. 3.

    Assessment of the present and past character of river reaches (a) to understand how they are affected by processes and human pressures from catchment to segment scales; (b) to understand how these affect river behaviour by driving trajectories of change at the reach scale; and (c) to use the knowledge gained to assess the likely impact of future scenarios on catchment to segment processes and reach scale responses

Stage (1): Delineation

The boundaries of each spatial unit are delineated using the criteria listed in Table 2, so that each unit at any particular spatial scale is located entirely within a single unit at the next scale. If delineation of geomorphic units, hydraulic units and river elements is required, it must be obtained from field survey. However, sufficient information on geomorphic units is usually available to include them as reach scale indicators during stage (2). Delineation of other spatial units can be achieved using existing information.

Stage (2): Characterisation and indicators

Once the spatial units are delineated, their properties are characterised using existing data sets. Characterisation involves identifying existing data sets that contain relevant information from which the recommended set of indicators can be extracted. The characterisation process allows incorporation of many local data sets of different types that can help to define a required set of indicators of processes, forms, and human pressures. Some example indicators are listed in Table 3. Further details of the recommended indicators are provided elsewhere in this special issue (González del Tánago et al. 2015a).

Indicators have been devised to represent processes of water and sediment production and delivery at catchment to reach scales, and also human pressures and interventions that may affect water and sediment production and longitudinal continuity through the river system. Indicators also represent the extent and structure of riparian and aquatic vegetation at segment to reach scales and the degree to which these appear to have been impacted by human pressures. At the reach and geomorphic unit scales, indicators refer to flow energy, channel and floodplain dimensions and types, the assemblage of geomorphic units that is present, and the degree to which there are constraints on the lateral continuity of inundation, erosion and deposition of sediment and large wood.

The key scale is the reach scale and the key indicator at this scale is the river type. 23 river types are defined using three criteria: (1) valley confinement: confined, partly confined, unconfined; (2) planform: straight, sinuous, meandering, braiding, anabranching (defined using specific ranges of values of sinuosity, braiding and anabranching indices); (3) bed material: bedrock, colluvial, boulder, cobble, gravel, sand, silt, clay. River types range from ‘confined bedrock’ to ‘unconfined, sand-silt, anabranching’, with reaches with an artificial bed allocated to an ‘artificial’ type. Information is provided on the typical gradient, stability, size and variability in bed material and geomorphic units that may be expected if these types are functioning in a natural way. In addition, the river types are associated with floodplain types and the typical floodplain geomorphic units that may be observed if the floodplain is a product of the long term dynamics of the river type. The 23 river types were developed from previous geomorphological research (e.g., Schumm 1985; Knighton and Nanson 1993; Rosgen 1994; Nanson and Knighton 1996; Montgomery and Buffington 1997; Church 2006; Fuller et al. 2013; Nanson 2013) with additional information on geomorphic units in confined and bedrock river reaches obtained from Grant et al. (1990) and Halwas and Church (2002). The ten floodplain types, with which the river types are associated, are based on those suggested by Nanson and Croke (1992). This brief summary of the river and floodplain types is fully elaborated elsewhere in this special issue (Rinaldi et al. 2015b).

Most of the indicators (e.g., Table 3) have the potential to change through time, so both their contemporary and past values are estimated wherever possible. Historical analysis of indicators extends back as far as reliable sources of information are available, typically up to 100 years. Ideally, indicators should be evaluated for several time periods in the past to allow a trajectory of change to be tracked. Of course, this may not be feasible, and a longer historical time scale and higher temporal resolution may be achieved for some indicators (e.g., planform) but not for others (e.g., bed elevation).

Stage (3): Assessments

The indicators that are extracted from the set of past and present characteristics of each spatial unit are integrated to develop an understanding of how and why river reaches have their current properties and also whether these have changed over time and what may have caused such changes. This is tackled in a sequence of four steps that are fully described in the paper by González del Tánago et al. (2015a) which also appears in this special issue. These are briefly outlined below.

First, the current state of individual reaches is assessed. Four main assessments are made:

Hydromorphological function :

Starting from the reach type indicator, assessment is based on whether the assemblage of geomorphic units within the channel and floodplain indicate that the reach is functioning as would be expected, and whether the stream power appears to be sufficient to maintain functioning.

Hydromorphological alteration :

Given the indicators of human pressures, the degree of disruption of longitudinal and lateral continuity and restriction of bed or bank dynamics within the reach is assessed.

Riparian corridor function/artificiality :

This is assessed using indicators of the size, vegetation age structure, and sources and presence of large wood within the riparian corridor of the reach.

Hydromorphological adjustment :

The degree and way in which the reach appears to be adjusting or behaving at present is assessed using indicators of the presence, extent and spatial pattern of relevant geomorphic units, and the sedimentary structure of bed and banks.

Second, past and present indicators at catchment, landscape unit and segment scales are used to estimate past and present water production and delivery, and river flow regime; and also sediment production and delivery from the catchment and through the river network. Comparison of present and past values of these indicators, preferably including several time periods in the past, helps to quantify the degree to which flow and sediment processes have changed through time and the likely causes of the changes (e.g., land cover changes, dam construction, channel reinforcement etc.). Based upon this information, a space–time inventory of changes is constructed, focusing particularly on human alterations that have impacted on flow and sediment processes delivered to river reaches.

Third, reach scale historical indicators are coupled with the contemporary reach scale indicators to reconstruct, as far as is possible, the nature of morphological changes within a reach and the timing of those changes to indicate the changing behaviour of the reach. For example, based on an analysis of historical maps and air photographs, an individual reach may show a trajectory of channel narrowing, widening, lateral migration, or a change in river type through time, or the reach may switch from one adjustment type to another. Vertical changes (e.g. bed incision or aggradation) can also be reconstructed from cross section or longitudinal profile information as well as from the evolution of the stage-discharge relationship at gauging stations (specific-gauge analysis). The causes of any identified changes can then be interpreted from knowledge obtained about changes in flow and sediment processes across the catchment and river network during the second step. Along a river, different river reaches may show different degrees and types of morphological adjustment or different behavioural responses to specific changes in the processes delivered to them. Such differences in adjustment may relate to the reach river type and to human interventions within the reach.

Fourth, potential responses at the reach scale to future scenarios of change can be considered, usually focussing on reaches of different river type within particular segments or landscape units, and using information on the way reaches of this type have adjusted in the past. By basing the assessment of causes and responses to changes in the past on a defined set of indicators, those same indicators and their likely responses to specific future scenarios can be interpreted at all spatial scales, providing a basis for forecasting how reaches of different type may respond to particular types of intervention or process change. Furthermore, where reaches are heavily modified by human interventions, historical analyses of all reaches and consideration of future scenarios may contribute to identifying a more appropriate reach type that could guide rehabilitation or restoration designs. Future trajectories are usually based on a small number of scenarios relevant to the river in question, with the aim of informing management recommendations. Two core scenarios are the likely trajectories of adjustment behaviour under (1) the present climate and (2) likely climate changes (e.g., over the next 50 years) but with no significant change in catchment management. Other scenarios can reflect proposed or likely future changes in river management, land cover, the implementation of particular projects etc.

Elaborating and applying the REFORM framework

This paper has presented a brief overview of the REFORM framework that has been developed for application by river managers within Europe. It is both flexible and it incorporates many aspects of previous hierarchical frameworks. Therefore, the framework should be applicable to landscapes beyond Europe that have a similar, long history of human pressures, and where a framework for application by river managers is required. Further details of important aspects of the REFORM framework are presented in two other papers in this special issue. The indicators are justified and described and their application is illustrated by González del Tánago et al. (2015a). The paper by Rinaldi et al. (2015b) fully explores three particularly important indicators: the river, floodplain and flow regime types. It also presents a typology of groundwater-surface water interactions that can be linked to the river and floodplain types.

Remotely sensed data provide an increasingly important source of information on river catchments, and so the paper by Bizzi et al. (2015) reviews this topic to aid users of the REFORM framework to gain information on whatever level of complexity they feel is appropriate. Furthermore, modelling can help to characterise river segments and reaches and can also be used to investigate future scenarios. The paper by Camenen et al. (2015) considers different approaches to modelling the sediment budget of a long segment of a large river, the River Loire, France. Ziliani and Surian (2015) also employ modelling at the segment scale to illustrate how this aids understanding of a trajectory of changes and possible future channel evolution within reaches of the lower course of the Tagliamento River, Italy.

Finally, because of the open-ended nature of the REFORM framework, and the way it can be used to incorporate different local data sets and models to address different management issues, three papers illustrate management-specific applications. These papers illustrate how the framework has helped to diagnose management problems resulting from fine sediment delivery and transfer in a low gradient, temperate, agricultural catchment in southern England (Grabowski and Gurnell 2015); problems induced by past gravel mining and other disturbances in an Italian river (Belletti et al. 2015b); and problems induced by flow regulation in two rivers in Spain (González del Tánago et al. 2015b).