From structure to function: understanding shrub encroachment in drylands using hydrological and sediment connectivity

Abstract

Hydrological and sediment connectivity can help us to understand better how physical and process-based linkages govern ecogeomorphic feedbacks and identify locations where degradation is likely to be pronounced. In this study we investigate how hydrological and sediment connectivity affect ecogeomorphic feedbacks in transitional landscapes. We propose a novel approach, the Relative Connectivity Index (RCI), to quantify landscape connectivity which explicitly integrates structural measures of landscape connectivity (SC) with functional measures of hydrological and sediment connectivity (FC) that are derived from runoff and sediment-transport modelling. We use the RCI calculated for runoff (RCIH) and sediment (RCIS) to identify locations and times when functional connectivity exceeds structural connectivity thresholds – where land degradation is likely to be pronounced – and explore how these thresholds are affected by rainfall-event size and antecedent soil-moisture content. We find that there are non-linear increases in RCIH values with an increase in shrub cover, which suggest that ecogeomorphic feedbacks become more important in modifying system structure and function during late stages of shrub encroachment. Thresholds of sediment connectivity appear to be directly related to thresholds of hydrological connectivity, although rainsplash appears to be an important mechanism in creating connected sediment transport where there is no connected runoff. High RCIH values are most widely distributed for the largest (45 mm) rainfall event, whilst high RCIS values are observed to some extent across all stages of the grass to shrub transition for rainfall events as small as 10 mm. Whilst particularly large events have a low return period, they appear to be particularly instrumental in shaping ecogeomorphic feedbacks that are likely to drive catastrophic shifts in ecosystem state. The strength of the indicator approach used here is that it enables identification of regions with pronounced ecogeomorphic feedbacks, which act as potential trigger points for catastrophic shifts in ecosystem state, and thus, we demonstrate how the static limitations of existing approaches to developing connectivity indices may be overcome. The dynamic RCI allows the evaluation of the vulnerability or resilience of a particular system to variable driving mechanisms. The RCI can therefore be used to guide management interventions aimed at reducing or mitigating undesirable ecosystem state change, by focussing on specific locations/regions with high RCI values, to prevent further chances in system structure and function and to maximise the provision of ecosystem services.