Morphology of partially vegetated desert dunes affects their response to climatic perturbations and disturbance

Longitudinal dune activity and vegetation cover in the northern Simpson Desert, Australia, was investigated with remote sensing data to develop new insights into the process of desert dune activation and stabilisation. Using topographic data, high resolution satellite imagery, and multi-temporal Landsat fractional vegetation cover, dune crest activity was studied during drought conditions and following wild-fire (2002–2009), and following a rainfall pulse (2010 onwards). The key finding was that desert dunefield response to climatic perturbations is locally variable because of local variation in wind strength, moderated by dune topography, and antecedent vegetation cover and sediment mobility. Many dunes were active during drought conditions, more commonly so if they were also burnt beforehand. Taller and steeper dunes were more likely to (a) be observed as active following a reduction in cover, and (b) remain bare and active following rainfall. High rainfall in 2010 enabled the colonisation of many active dunes by vegetation, but more regrowth occurred on dunes that were previously stable. This restriction of vegetation establishment on mobile sand was long-lasting (up to 10 years). Logistic regression was used to model the relationship between vegetation cover and dune height and the probability of slipface development. Model parameters were affected by climatic conditions and seemingly by the prior state of the dune surface (i.e. active or stable). The results from the Simpson Desert were integrated into a new conceptual model, in which the response of longitudinal desert dunes to a stabilising climatic perturbation depends on the spatially variable negative feedback (hysteresis) imparted by active dunes on vegetation regrowth. • Variable response in vegetation cover and dune activity seen after wet period • Dune shape and fire impacted likelihood of dune activity during drought. • Post-drought dune activity related to dune height and antecedent vegetation cover • Logistic regression predicts slipface occurrence from ground cover and dune height.