(Dis)connectivity in hydro-geomorphic systems – emerging concepts and their applications

In geomorphology, connectivity has emerged as a framework for understanding the transfer of water and sediment through landscapes. Over the past decade, sessions on (dis)connectivity at the General Assembly of the European Geosciences Union (EGU), and more recently, three mini-conferences in 2020 and 2021 called ‘ Connectivity Conversations ’ , organized by the International Association of Geomorphologists (IAG) working group on ‘ Connectivity in Geomorphology ’ , have created a space for the exchange of ideas relating to (dis)connectivity in geomorphology and related disciplines. The result of these initiatives has been a collection of research articles related to a special issue (SI) entitled ‘ (Dis)connectivity in hydro-geomorphic systems – emerging concepts and their applications ’ . In this article, we provide a syn-thesis that embraces the SI contributions related to the application of the connectivity concept in different environments and geomorphic process domains, spatial and temporal scales, types and spatial dimensions of connectivity and the role of human impacts and associated river and catchment management aspects.


| INTRODUCTION
In the past two decades, connectivity has emerged as a useful conceptual framework for understanding the transfer of water and sediment through landscapes. Connectivity thinking in (hydro-)geomorphic research entails a range of benefits for investigating the spatial and temporal variability of material fluxes and (complex) system behaviour (Poeppl et al., In press;Wohl et al., 2019), by focusing on the interactions among system components, geomorphic response to varying inputs, and the role of humans in influencing the behaviour of geomorphic systems. The effects of widespread disruptions in connectivity in hydro-geomorphic systems have been recognized, especially with research on natural, leaky dams (e.g., beaver dams and log jams).
We use the term (dis)connectivity to refer to the different levels of connectivity, from the highly disconnected, to the highly connected, combining the study of fluxes between geomorphic features with that of the role of disruptions in the transfer of mass in hydro-geomorphic systems. Over the past decade, sessions on (dis)connectivity at the General Assembly of the European Geosciences Union (EGU), and more recently, three mini-conferences in 2020 and 2021 called 'Connectivity Conversations' (CC), organized by the International Association of Geomorphologists (IAG) 'Connectivity in Geomorphology' working group, have created a space for the exchange of ideas relating to (dis)connectivity, surrounding methodological approaches, spatial and temporal scales, process domains, spatial dimensions of connectivity, structural connectivity (SC) and functional connectivity (FC), drivers of change in connectivity, and the different processes/ materials for which connectivity is a useful concept.
Applications of the connectivity concept in geomorphology are incredibly diverse in terms of their goals, the approaches used, whether or not (and how) they differentiate between SC and FC or the type of material (e.g., water, sediment, biota) they are concerned with. The special issue (SI) '(Dis)connectivity in hydro-geomorphic systemsemerging concepts and their applications' brings together 10 papers that were presented at the EGU (dis)connectivity sessions and the CC in 2020 and 2021, that explore various aspects of these key themes (see overview in Table 1).

| ENVIRONMENTS AND GEOMORPHIC PROCESS DOMAINS
Hydro-geomorphic connectivity emerges from a complex interplay between different environmental factors such as geology, relief, landscape position, climate, soils, biota, and human activity (e.g., Bracken & Croke, 2007;Keesstra et al., 2018;Poeppl et al., 2017). These factors further govern water and sediment dynamics in different environmental settings and geomorphic process domains (i.e., fluvial, hillslope, coasts/deltas, periglacial, glacial, aeolian; cf. Poeppl et al., In press). In this SI, five contributions have dealt with connectivity in mountain environments, with a specific focus on fluvial processes by Hinshaw et al. (2022;South Fork McKenzie River basin, Cascade Mountains, Oregon, USA), Kemper et al. (2022;Yampa River basin, Rocky Mountains, Colorado, USA), and Khan et al. (2021;Rich-mond River catchment, Great Dividing Range, Australia), on hillslope processes by Martini et al. (2022;Rio Cordon catchment, Italian Alps), or on both environments as shown by Turley et al. (2021;Tahoma Creek catchment, Cascade Mountains, Washington, USA). Two papers investigated connectivity relationships in river and catchment systems (Bizzi et al., 2021: Vjosa River, Albania;González-Romero et al., 2021: Segura River catchment, Spain), while two contributions had a specific focus on delta, estuary and wetland environments Singh et al., 2022;Sonke et al., 2022).

| TYPE OF MATERIAL
A common definition of connectivity in geomorphology is 'the degree to which matter and organisms can move among patches in a landscape or ecosystem' (Wohl, 2017), and the 'matter' studied can be water, sediment, solutes or organic material. Focusing solely on longitudinal connectivity of water through a catchment often lends itself to pure hydrological studies, whereas the role of various types of disconnectivity on water fluxes has been a recent theme in hydrogeomorphic research. Sediment (dis)connectivity controls rates of landscape erosion and types of channel features and thus habitat in rivers. Ecological questions can be answered using a hydrogeomorphic framework of (dis)connectivity for studies on hydrochory (seed/plant propagule dispersal by water) and other passive dispersal of organisms (e.g., macroinvertebrates and pelagic-spawning fish), or nutrient and carbon storage.
In this SI, two studies focus on connectivity of only water Singh et al., 2022). Sonke et al. (2022) examine connectivity of both water and sediment. Most studies focus on sediment connectivity González-Romero et al., 2021;Martini et al., 2022), including Kemper et al. (2022) who concentrate on catchment-scale connectivity using vegetation establishment as a proxy for when sediment was eroded or deposited in particular locations in the catchment. Only one study focuses on other types of matter than sediment and water: Hinshaw et al. (2022) examine changes in connectivity of water, sediment and organic matter through time after a restoration method that aims to increase lateral connectivity.

| SPATIAL DIMENSIONS
Within hydro-geomorphic systems, connectivity and the lack thereof can be studied in several spatial dimensions, but it has traditionally been tied to longitudinal transport and continuity of material. In fact, the very reason that rivers are able to act as conduits for water and sediment and thus erode landscapes is based on downstream T A B L E 1 Overview of the special issue (SI) papers. connectivity (Ferguson, 1981). Fundamental fluvial geomorphic theories, such as downstream hydraulic geometry (Leopold & Maddock, 1953), rely on downstream connectivity of water flows.

SI papers
The effects of disconnectivity on longitudinal fluxes originally received most attention in terms of anthropogenic effects of building dams (e.g., Graf, 1999Graf, , 2006Nilsson et al., 2005); however, natural forms of longitudinal disconnectivity, in terms of, for example, beaver dams, wood jams, and lakes, can play a pivotal role in steering fluxes of water (Puttock et al., 2021), sediment (Wohl & Scott, 2017)

| FUNCTIONAL CONNECTIVITY (FC) AND STRUCTURAL CONNECTIVITY (SC)
Within geomorphology a distinction is often made between SC (network architecture) and FC (dynamical processes; e.g., Bracken et al., 2015;Wainwright et al., 2011;Wohl et al., 2019). This distinction between SC and FC is an artificial one that tries to separate the influence of system structure on dynamical processes (Turnbull et al., 2018), yet is nevertheless useful as it allows simplification of the characterization/quantification of connectivity. As seen from the collection of studies within this SI, there is a general tendency to focus within geomorphology on either SC (e.g., Martini et al., 2022) or FC (e.g., Bizzi et al., 2021;Khan et al., 2021), and rarely do studies focus on both (but see Hiatt et al., 2022), often because the approaches used make it challenging to make such a distinction.
What is still less common in geomorphology is the quantification of FC. It is this combined analysis of SC and FC that has the potential to allow for greater insights into key locations within the landscape where feedbacks between form and function are particularly pronounced, which cannot be obtained without a connectivity-oriented approach (as demonstrated in Turnbull and Wainwright [2019]). In order for useful comparisons between SC and FC to be made, the template over which SC and FC are compared also needs to be the same. Graph theory approaches are very powerful in this regard, as network topology enables SC metrics to be determined, while also allowing for FC to be quantified. An example of this application is given in Hiatt et al. (2022) where they explore how FC metrics vary with different network characteristics.

| ANTHROPOGENIC INFLUENCE ON CONNECTIVITY
Different types of human activity can alter the connectivity in hydrogeomorphic systems and thus their sensitivity to change (Fryirs, 2017;Poeppl et al., 2017). Human impacts on water and sediment connectivity in river systems, for example, can be either direct (e.g., due to dam construction/river engineering) or indirect (e.g., due to land cover/land-use changes; cf. Gregory, 2006;Poeppl et al., 2017).
Human disturbance of connectivity in geomorphic systems can further be differentiated between ramped (i.e., sustained; e.g., due to dam construction) and pulsed (i.e., event; e.g., a flash flood after dam breach) types of inputs (Brunsden & Thornes, 1979;Poeppl et al., 2017). Moreover, the recognition of (past) human disturbances and how these have modified natural connectivity relationships, and in how far they can or should be managed, is of major importance, especially in river and catchment management contexts (Fryirs & Brierley, 2009;Keesstra et al., 2018;Poeppl et al., 2017Poeppl et al., , 2020Wohl et al., 2019).

| ADVANCES AND REMAINING CHALLENGES IN UNDERSTANDING GEOMORPHIC SYSTEMS USING CONNECTIVITY AS A FRAMEWORK
One of the main benefits of studying a system from a connectivity perspective is that it allows the influence of local-scale processes on large-scale system form and function to be disentangled. Within geomorphology, studies of connectivity range across spatial and temporal scales, from plot-based experiments to large river basin assessments and network-based approaches as well as from static representations of connectivity (also referred to as SC derived from network topology) to dynamical representations of connectivity (also referred to as process-based, or FC), ranging from seconds to millennia.
Within the geomorphological community, there has been a multitude of connectivity overview papers that have helped to bring together ways to operationalize this framework and bring quantitative meaning to discussions surrounding connectivity in geomorphology.
Nevertheless, these frameworks continue to evolvefor instance, in this SI, Kemper et al. (2022) present a sediment-ecological framework for large river networks that links geomorphic and ecologic processes across space and time at catchment scales. The importance of the coupling between geomorphological and ecological processes is a recurrent theme (for example Hinshaw et al., 2022;Kemper et al., 2022;Singh et al., 2022) and speaks to the importance of multidisciplinary perspectives (Turnbull et al., 2018).
Both the scope of a study (i.e., in terms of network topology or dynamical processes and their interactions, and the system boundary) along with its scale determine the methodical approach to assess connectivity (e.g., Keesstra et al., 2018;Singh & Sinha, 2020;Turnbull et al., 2018). One of the main challenges in quantifying connectivity from a geomorphological perspective is the definition of fundamental units of study as well as the availability and generation of suitable datasets that allow for detailed characterizations of landscape form and function over relevant space and timescales Turnbull et al., 2018). These datasets are essential in order for the geomorphic community to move beyond conceptual frameworks for the study of connectivity, to operationalize these frameworks, and to allow for an improved understanding of emergent patterns and characterization of connectivity. However, it still remains common within geomorphology to allude to the concept of connectivity, rather than operationalize (and thus quantify) the concept. A common approach to attempt to surmount this challenge in geomorphology is the use of indices that are focused on the structure of the landscape (SC) and use a relatively narrow set of parameters (e.g., topography, vegetation cover). A key example of this is the Index of Connectivity (Borselli et al., 2008) which has been widely applied (see, e.g., Martini et al., 2022). There have been attempts to expand this index further, for example to capture the potential influence of dynamical processes by the addition of further parameters (see González-Romero et al., 2021).
Characterizing FC within geomorphology is still challenging. Given the inherent challenge of collecting data suitable for characterizing FC, one approach is to model the processes for which we are interested in quantifying FC. Modelling is particularly powerful as it allows us to fill in the spatial and temporal gaps in observed records, and capture dynamical processes at relevant spatial and temporal scales for to the quantification of SC and FC. They find that the approaches that utilize high-resolution input data, tend to generate the best characterization of connectivity when compared with field observations, and conclude that characterizations of FC that account for dynamical processes nevertheless remain important, which is clearly an area where progress still needs to be made within geomorphology.
An important development within geomorphology has been the application of network analysis approaches derived from graph theory. The value of the implementation of graph theory, and the suite of connectivity-based measures of graph characteristics derived from network theory, is that it allows a more robust quantification of connectivity that enables the operationalization of the concept of connectivitysomething that is still often lacking within geomorphological studies. A detailed review of the recent developments in the application of graph theory in geomorphology is given by Heckmann et al. (2015). In order to apply graph theory within geomorphology, one of the first tasks is to represent the system as a network.
Approaches to network representation are well developed for relatively simple and directed networks, but less so for multi-channel networks, typical of, for example, estuaries. In this SI, Sonke et al. (2022) explore two approaches (one local, one global) to identifying channel networks from the terrain of a river bed, based on the volume of alluvium between them. The resulting channel networks (and thus quantification of SC) are very sensitive to the approach used, with the local approach yielding optimal results. The insight provided by this local approach is that it allows for a new understanding of topological (or structural) connectivity of the channel networks, whereby alluvial connectivity is the inverse of the volume of sediment that needs to be removed in order for two channels to be connected. Several authors have highlighted that (dis)connectivity as a concept should form a basis for informed decision-making in river and catchment management including river restoration (Keesstra et al., 2018;Poeppl et al., 2020;Wohl et al., 2019). Still, in practice most management plans overlook the role of (dis)connectivity in driving natural and human-induced disturbance and treatment responses in river and catchment systems (Poeppl et al., 2020), often further lacking holistic approaches to manage these systems in a sustainable manner. However, especially in integrated catchment management and recent river restoration efforts, aspects of lateral connectivity are increasingly being incorporated (e.g., Cluer & Thorne, 2014). In this SI, Hinshaw et al. (2022) addressed the problem of how to examine SC and FC after a new restoration strategy of Stage-0 restoration and to accurately determine how the newly restored SC translates to FC.

| CONCLUSIONS
The concept of connectivity has shown to be applied to different environments and geomorphic process domains, on different spatial and temporal scales, spatial dimensions of connectivity as well as in diverse contexts using different (methodical) approaches (e.g., measuring, indices, modelling). The SI contributions have dealt with connectivity in mountain environments, river and catchment systems, as well as delta, estuary and wetland environments focusing on fluvial and/or hillslope processes. Moreover, different types of human impact (e.g., dams, land-use changes) have been addressed in different contexts (e.g., river process and form, river and catchment management). Most of the articles focus on water and/or sediment connectivity studying either FC or SC, but rarely botha general tendency in geomorphological connectivity research often related to lacking data on processes, their dynamics as well as on process-form relationships.
Furthermore, it can be observed that many (if not most) connectivity studies in geomorphology rather tend to infer connectivity from different types of (obtained) data and results instead of operationalizing the concept of connectivity by actually quantifying connectivity. An important development within geomorphology has been the application of network analysis approaches derived from graph theory which allows for a more robust quantification of connectivity (e.g., see Sonke et al. [2022] and Hiatt et al. [2022] in this SI). Network-based approaches are promising as they allow for a quantitative interrogation of how dynamical processes are related to system structure (i.e., SC-FC relations). Nevertheless, applications of network-based approaches require some rethinking of how to most appropriately represent geomorphic systems as networks, and critical evaluation of suitable network-based metrics to improve our understanding of these systems.

AUTHOR CONTRIBUTIONS
Conceptualization and writing: Ronald E. Poeppl, Lina E. Polvi and Laura Turnbull.