Climate change sentinel or false prophet? The case of A loe dichotoma

Aim Previous research suggests that equatorward populations of the iconic arborescent succulent Aloe dichotoma Masson are contracting in response to recent anthropogenic climate change (ACC) in southern Africa. However, previous studies did not account for small‐scale spatial heterogeneity, latitudinal climatic disjunctions or when mortality occurred. We reassess A. dichotoma's suitability as an indicator species and comment on pitfalls in the selection of species with similar life history characteristics. Location South‐western Africa. Methods A 15,000 km distribution‐wide roadside visual survey was conducted to capture demographic patterns and provide the means for extracting tightly coupled climate profiles for the species. Rainfall and temperature gradients were assessed for latitudinal discontinuities indicating distinct equatorward‐poleward climate zones. Repeat photographs and reference individuals were used to develop an index for dating the approximate time‐since‐death of dead trees at 14 latitudinally spaced populations. Results The proportion of dead individuals was greatest within the southern third of the latitudinal distribution rather than at the equatorward range limit. The equatorward summer rainfall zone (SRZ) was significantly drier, rainfall more variable and temperatures hotter compared to the poleward winter rainfall zone (WRZ). This difference was associated with a generally greater proportion of dead individuals in the SRZ and greater proportion of juveniles in the WRZ. Furthermore, juveniles appeared more sensitive to drought stress while adults were more resilient. Most mortality occurred several decades ago, was not more recent at the equatorward limit and was drawn from the more drought resilient adult stage class. Main conclusions Aloe dichotoma mortality (and recruitment) patterns reflect prevailing differences in the SRZ and WRZ climate, while differing juvenile and adult drought stress tolerances and windthrow‐mediated mortality reinforces this pattern. This and strong indications of non‐recent death suggest that ACC is not responsible for observed mortality. An alternative hypothesis for current demographic patterns, which incorporates palaeoclimatic evidence and A. dichotoma life history characteristics is proposed. Long‐term climate trends, recruitment‐mortality dynamics and potential climatic discontinuities should be considered before advancing species as indicators of ACC.