Chronic anthropogenic disturbance (CAD) and climate change represent two of the major threats to biodiversity globally, but their combined effects are not well understood. Here we investigate the individual and interactive effects of increasing CAD and decreasing rainfall on the composition and taxonomic (TD), functional (FD) and phylogenetic diversity (PD) of plants possessing extrafloral nectaries (EFNs) in semi-arid Brazilian Caatinga. EFNs attract ants that protect plants against insect herbivore attack and are extremely prevalent in the Caatinga flora. EFN-bearing plants were censused along gradients of disturbance and rainfall in Catimbau National Park in north-eastern Brazil. We recorded a total of 2,243 individuals belonging to 21 species. Taxonomic and functional composition varied along the rainfall gradient, but not along the disturbance gradient. There was a significant interaction between increasing disturbance and decreasing rainfall, with CAD leading to decreased TD, FD and PD in the most arid areas, and to increased TD, FD and PD in the wettest areas. We found a strong phylogenetic signal in the EFN traits we analysed, which explains the strong matching between patterns of FD and PD along the environmental gradients. The interactive effects of disturbance and rainfall revealed by our study indicate that the decreased rainfall forecast for Caatinga under climate change will increase the sensitivity of EFN-bearing plants to anthropogenic disturbance. This has important implications for the availability of a key food resource, which would likely have cascading effects on higher trophic levels.

We characterised CAD using a global multi-metric disturbance index previously established for the study plots (Arnan et al. 2018b). The index combines three types of source information: (i) indirect landscape measures using satellite imagery in ArcGIS 10.1 software: proximity to the nearest house, and proximity to the nearest road; (ii) socio-ecological information obtained by interviewing local households: number of people, number of goats, number of cattle and firewood use; and (iii) direct field measures of the length of goat trails, the density of goat and cattle dung, extents of live-wood extraction and fire-wood collection. 

We obtained data on mean annual rainfall for each plot from the WorldClim global climate data repository for 1970-2000 (Hijmans et al. 2005) with 1 km resolution using the “maptools” package for R 3.1.2 (R Core Team 2014).

We used woody plant inventory data that have previously been published for the 20 0.1-ha plots (Rito et al. 2017a). Species were assessed as bearing extra floral nectaries (EFNs) on the basis of published literature (e.g. Elias 1983; Rico-Gray and Oliveira 2007; Melo et al. 2010) and information available on the World List of Plants with Extrafloral Nectaries (Weber et al. 2015). When the information was unclear or contradictory (such as for some species of the euphorb genera Cnidosculus, Croton and Jatropha), we performed our own field collections and samples were sent to Estadual University of Bahia herbarium (HUNEB Herbarium) and Vale do São Francisco herbarium (HVASF Herbarium), where they were examined under a stereomicroscope for extra floral nectaries.

For each species we used information from Weber et al. (2015), Queiroz (2009) and Melo et al. (2010) to document three traits related to nectar secretion and attraction/access to ant attendants: EFN morphological type, location and arrangement. Eight of the EFN morphological types recognized by Weber et al. (2015) occurred in our study species: (i) elevated; (ii) flat-elevated; (iii) digitiform-elevated; (iv) concave-elevated; (v) stipitate; (vi) globose; (vii) impregnated in slot and (viii) glandular trichomes. We recognized eight EFN locations in our study species: (i) base of petiole; (ii) petiole (other than the base); (iii) interpetiolar; (iv) rachis; (v) base of leaflet; (vi) leaf tip; (vii) leaf margin and (viii) floral pedicel. EFNs can occur either in isolation or arranged in groups (Díaz-Castelazo et al. 2005), categorised as: (i) isolated; (ii) variably isolated/paired; (iii) paired and (iv) grouped (> 2 EFNs), with multiple EFNs providing greater total nectar resources (Apple and Fenner 2001; Lange et al. 2017).