Eco-geomorphological connectivity and coupling interactions at hillslope scale in drylands: Concepts and critical examples

Highlights

• Structure/function approach evidences erosion effects on vegetation configuration.

• Composition of abiotic patches modulates source-sink dynamics by surface armouring.

• Channel linear incision diffuses the upslope erosion, biasing vegetation.

• Erosional sequence across-scales sets boundary conditions to hillslope connectivity.

• Eco-geomorphic frameworks are needed to address explicitly space-time scale dependencies.

Abstract

The diagnosis of land degradation requires a deep understanding of ecosystem functioning and evolution. In dryland systems, in particular, research efforts must address the redistribution of scarce resources for vegetation, in a context of high spatial heterogeneity and non-linear response. This fact explains the prevalence of eco-hydrological perspectives interested in runoff processes and, the more recent, focused on connectivity as an indicator of system resource optimisation. From a geomorphological perspective and reviewing the concepts of eco-hydro-geomorphological interactions operating in ecosystems, this paper explores the effects of erosion on vegetation configuration through two case studies at different spatio-temporal scales. We focus on the structure-function linkage, specifically on how morphological traits relate with different stages in the erosional sequence, both in the abiotic and the biotic domain. Results suggest that vegetation dynamics are affected by structural boundary conditions at both scales, i.e. by surface armouring related with rock fragments at the patch scale, and by the degree of hillslope-channel coupling at the hillslope scale. Our preliminary results can serve as new working hypotheses about the structure-function interplay on hillslopes. All this, taking advantage of the recent technological achievements for acquiring very high-resolution geospatial data that offer new analytical possibilities in a range of scales.

Introduction

Understanding the functioning and evolution of natural systems is a basic research endeavour, and a necessity for meaningful land degradation assessments (Brandt and Thornes, 1996; Geeson et al., 2001), particularly in the context of climate change where trajectories of ecosystem response are highly uncertain (Maestre et al., 2016). Research needs to deal both with the structural system complexity and with its non-linear response (Wainwright et al., 2011) which is governed by thresholds (Cammeraat, 2004). Particularly so in the drylands, where this complexity is exacerbated by the high spatial heterogeneity (Puigdefabregas et al., 1998; Lavee et al., 1998; Aguiar and Sala, 1999; Puigdefábregas, 2005) and the timing of trigger-events (Schwinning and Sala, 2004), respectively. In the context of water-limited ecosystems, resource redistribution is critical for plant growth and survival and the role of biotic-abiotic interdependence is in the origin of the operating principle of the source-sink dynamics.

Over the last decades, research on dryland ecosystem functioning has thus been dominated by the eco-hydrological perspective, which focuses on the water balance between vegetation and the inter-canopy bare ground (Rodríguez-Iturbe and Porporato, 2004). The role of runoff distribution has commonly been the research focus until the issue of connectivity became a primary challenge (Mayor et al., 2008, 2019; Okin et al., 2015; Saco et al., 2020).

Our aim here is to contribute to this eco-hydrological perspective from the geomorphological viewpoint, delving into the role of erosion processes in the vegetation configuration that affect hillslope connectivity. Specifically, by setting the current conceptual foundations in eco-hydro-geomorphological system operation and presenting preliminary results from two case studies from different spatio-temporal scales in process operation (namely, the patch and the hillslope/valley scales), we extract two new working hypotheses about the structure-function interplay at the hillslope for future empirical-based research.

In water-limited ecosystems, the mechanisms that guarantee vegetation survival are strengthened at the patch scale, where spatial heterogeneity in the soil properties and their corresponding erosional response begins. Puigdefábregas (2005) developed the concept of Vegetation Driving Spatial Heterogeneity (VDSH), which synthesises the process of patch differentiation with contrasted properties through feedback mechanisms promoted by the vegetation. These mechanisms are the result of water and solar radiation, organic carbon, nutrient and sediment balances.

Contrary to bare areas, patches of vegetation function as resource traps (i.e. water, sediment and nutrients) thanks to their favourable conditions enhancing infiltration and preventing soil erosion, namely: (i) soil aggregation and porosity, promoted by the organic matter supplied by plants and the presence of macro-porosity related to the root system; (ii) the absence of surface sealing as the soil is protected from the impact of raindrops, and (iii) the slowdown of the flows by the development of plant mounds.

Therefore, soil-vegetation interactions lead to a spatial mosaic of areas with optimal conditions for infiltration (i.e. vegetated areas) and areas where higher runoff rates are recorded (i.e. bare areas between vegetation). This phenomenon is known as source-sink dynamics, which constitutes the operating principle in drylands allowing the spatial redistribution of resources.

Connectivity, a concept originally termed continuity by Brunsden and Thornes (1979), materialises when the source-sink scheme extends in space and relates to the way the organisation of patches determines the redistribution of overland flow (Yair and Lavee, 1976, 1985; Lavee and Yair, 1990; Bergkamp, 1998; Puigdefabregas et al., 1998; Puigdefábregas, 2005). This fundamental interdependence of bare soil and vegetation can acquire multiple spatial configurations depending on water balance variables, resulting in different degrees of overland flow continuity. Connectivity has been linked to the system efficiency in long-term resource retention, and is closely related to system health in terms of its functioning and, therefore, to its potential state of degradation (Tongway and Hindley, 2004; Kéfi et al., 2007; Okin et al., 2015). Over the last decade, quantifying connectivity and the associated loss of system resources, has become a challenge.

Advances in these issues come mainly from two approaches: (i) in the development of implicit connectivity indices in which the flow routing is considered incorporating the topographical variable, as is the case of the Leakiness index (Ludwig et al., 2007), the Flowlength index (Mayor et al., 2008, 2019) or the Sediment Connectivity Index (Cavalli et al., 2013), and (ii) in the indirect assessment of general spatial continuity of bare areas through the morphological attributes of vegetation patterns, ranging from simple spatial metrics (Imeson and Prinsen, 2004; Boer and Puigdefábregas, 2005) to patch-size frequency distribution inferences (Kéfi et al., 2007; Scanlon et al., 2007; Maestre and Escudero, 2009).

A common feature, however, is the widespread usage of binary vegetation/non-vegetation maps in the analysis justified by the sharp contrasting response of its constituents. This fact is explained by the prevalence of the eco-hydrological perspective in the source-sink conceptualisation, simplifying at this stage the inter-canopy non-vegetated area as homogeneous net water supplier to vegetation.

Nevertheless, there is a growing field of research that emphasizes the role of the Biological Soil Crusts in dryland functioning (Belnap, 2006; Maestre et al., 2013), particularly in water, nutrient and sediment redistribution balances (Kidron and Yair, 1997; Chamizo et al., 2016), thus evidencing the relevance of the diversity in composition as a fundamental dimension of the analysis.

However, in addition to these efforts to broaden considerations from the biotic domain, the complexities added by abiotic factors to the system functioning also need to be considered. In particular, the foundations of the hydro-geomorphological operation of the hillslopes conceptualised from field empirical evidence, could help include the erosional processes involved in hillslope connectivity across-scales. Benchmark contributions established the factors involved in overland flow discontinuities along the hillslope, both spatially (Yair and Lavee, 1976, 1985; Lavee and Yair, 1990; Lavee et al., 1998) and temporally related with rainfall events (Puigdefabregas et al., 1998; Calvo-Cases et al., 2003; Cammeraat, 2004).

In the last decade, the conceptualisation of connectivity (Bracken and Croke, 2007; Bracken et al., 2013), the search for analytical frameworks from different perspectives (Lexartza-Artza and Wainwright, 2009; Bracken et al., 2015) and scale levels (García-Ruiz et al., 2010), have advanced significantly, especially in the field of geomorphology. In this context, there is an ongoing debate related with the functional and structural connectivity (Wainwright et al., 2011; Baartman et al., 2013).

In this regard of linking structural and functional dimensions, and particularly in the erosion/vegetation interplay, the effect of vegetation cover on the exponential reduction of soil erosion rates is well known, as it is one of the earliest identified and most extensively studied processes (Thornes, 1990; Abrahams et al., 1995; Gyssels et al., 2005). Later, interest turned to the effect of spatial configuration of vegetation on the redistribution of surface runoff (Cammeraat, 2004; Puigdefábregas, 2005; Ludwig et al., 2005; Bautista et al., 2007; Kéfi et al., 2007). However, despite these advances on how vegetation spatial patterns traits (i.e. the structural dimension) affect the operation of processes (i.e. the functional dimension), as Thornes (1985) had already pointed out, much less attention has been paid to the influence of erosion over vegetation dynamics.

The contribution of John B. Thornes in the two-way eco-geomorphological interactions was remarkable and was mainly achieved through the framework of the EU MEDALUS projects (Brandt and Thornes, 1996; Geeson et al., 2001). These projects are considered milestones for research on dryland land degradation processes, from which conceptual and methodological bases for improved understanding of such eco-geomorphological interplay were developed (i.e. Sanchez and Puigdefabregas, 1994; Kirkby et al., 1998; Puigdefabregas et al., 1999; Puigdefábregas, 2005). Regarding the issue of connectivity, in specific, Puigdefabregas et al. (1999) added the complexities of the vegetation and erosion interplay to the discourse through the water and sediment redistribution, emphasising the scale and the linkages between processes and patterns. The source-sink dynamics was conceived here as a game of tuning between the flow lengths and the development of the vegetation based on specific boundary conditions in each situation, which in turn is conditioning the connectivity of the fluxes along the hillslope.

Most progress, in relation to how structure traits affect functioning, has been made at the patch scale, and more specifically, in how the high spatial variability of the observed responses in runoff and erosion is related to the coverage of the ground surface (Lavee et al., 1998). Therefore, in addition to the well-known effect of plants, the role of rock fragment position, cover and size (Lavee and Poesen, 1991; Poesen and Lavee, 1994; Poesen et al., 1994; Katra et al., 2008), the dynamics of surface sealing (De Ploey and Poesen, 1984; Roth, 2004) and the role of biological soil crusts (Alexander and Calvo, 1990; Kidron and Yair, 1997; Belnap et al., 2005) have also been considered relevant in modifying the hydrological response. Based on these findings, the concept of Soil Surface Components (SSC) emerged (Kutiel et al., 1998; Lavee et al., 2004) as the most elementary, visible, identifiable and discrete expression of such soil surface properties and, therefore, as homogeneous units of hydrological response (Arnau-Rosalén et al., 2008).

Regarding the two-way vegetation/erosion interaction, this was synthesised in the MEDALUS model by means of three feedback loops and their inter-relationships that govern the dynamics of system operation, i.e. organic, erosion and armour loops (Kirkby et al., 1996; 1998, in Fig. 1A). The relevance of these feedbacks is in settling the linkage between the structural variables (vegetation cover, soil depth, surface armouring, etc.) and the response variables (runoff and erosion rates) in a mechanistic way. Organic and erosion loops reproduce in the model direct and indirect effects between vegetation and erosion processes (Fig. 1). The direct effect is the well-known vegetation control of soil erosion at each rainfall event by protecting the soil surface through rainfall and runoff inception and avoiding soil particle detachment and transport. The indirect effects are related with soil water retention capacity and are both qualitative and quantitative: the state, dependent on soil condition, is altered qualitatively by organic matter content and affects infiltration and soil water retention capacity; and quantitatively by soil erosion affecting soil depth.

In addition, the MEDALUS model incorporates a third feedback, the surface armouring loop, which is related to the increase of rock fragments cover as the soil surface erodes over time. In this case, it has a direct effect on the runoff and erosive response (Fig. 1), but with the particularity of having a bimodal behaviour. This duality means that this loop is subjected to a threshold condition in relation to the stage of development of the surface armouring. This phenomenon is grounded from the empirical results reached on the runoff response to different cover and positions of rock fragments (Poesen and Lavee, 1994; Poesen et al., 1994).

In the model, there is a prevalence of positive feedback loops, with only a negative one when the armour loop switches to an inverse relationship with runoff/erosion rate. Positive feedbacks reinforce the system functioning towards a certain path or trajectory, both individually within each loop, and synergistically. Whereas the negative feedbacks, has a potential regulatory effect in the uni-directional trajectories that the system can reach. In other words, in the MEDALUS model, the armour loop can act as counterbalancing the reinforcing effect of the rest of the positive feedbacks.

When the model is interpreted at the patch-scale (i.e. a single vegetated or bare-soil patch, Fig. 1B), the reinforcing effect of vegetation and erosion positive loops has simultaneous opposite trajectories, with divergent results depending on whether it is a vegetated or a bare patch (as unfolded in Fig. 1B). Therefore, the result is a spatial differentiation or divergence in the soil properties with contrasting hydrologic responses, issue that Puigdefáfregas (2005) synthetized with the concept of Vegetation Driven Spatial Heterogeneity (VDSH): the precursor of the source-sink dynamic mechanism. Nevertheless, the scope of the potential regulatory role of the armour loop in the general system functioning remains understudied.

According to Harvey (2001, p. 226), “despite its importance in influencing how geomorphic systems respond to environmental change, there have been relatively few studies of the role and mechanisms of coupling within fluvial systems”. This is especially pronounced in the coupling between hillslopes and river channels. In relation to channels, hillslopes can be not-coupled or absolutely disconnected, decoupled, or temporally disconnected by a barrier, and coupled, when there is a free transmission of mass and energy between both system parts (Harvey, 2001). From the fluvial point of view, the coupling-decoupling of the hillslope sediment cascades to the stream (Harvey, 2012) has been the focal point of interest for a variety of timescales (Caine and Swanson, 1989; Heckmann and Schwanghart, 2013; Del Vecchio et al., 2018; Michaelides et al., 2018).

Hillslopes coupled to the channel, after channel downcutting or bank erosion, are subject to effective erosion processes, such as rilling, gullying (Harvey and Calvo-Cases, 1991) or landslides (Savi et al., 2013). The coupling-decoupling interplay is very relevant in piping formation (Faulkner et al., 2008) and badlands development (Faulkner, 2008; Calvo-Cases et al., 2014), situations in which the hillslope strength resistance (i.e. lithology) is favouring a fast upslope spread of the waves of aggression.Where channel incision is lowering the hillslope's local base level (i.e. during the last millennia), the morphological resistance of the hillslopes changes as a consequence of the increase of the average slope, especially at the lower part (i.e. basal convexity). Often, morphological changes are not evident in the upslope part (i.e. rills or gullies). In these conditions, it can be assumed that the propagation of the changes is occurring “as diffuse waves of aggression away from the river channels or linear axes of change” (Brunsden and Thornes, 1979, p. 476). Thanks to the energy diffusion and storage (i.e. internal runoff discontinuities), the hillslope internal filters resistance (Brunsden, 2001) are not overpassed to reach an appreciable morphological change.

The strong strength resistance of hillslopes on some lithologies that need long timescales to show morphological changes (e.g. limestones) does not necessarily imply the absence of other changes in the medium-short term. The diffuse waves of aggression are operating and affecting the soil-plant system via the changes in the transmission resistance (i.e. the vegetation size and distribution). This can be done by the influence of diffuse soil erosion over the other factors in the interactive loops (Fig. 1) that include affection over the soil water balance, as described by Loheide and Booth (2011) on hillslopes under the coupling effect.