Effects of Quaternary climatic fluctuations on the distribution of Neotropical savanna tree species

Abstract

In order to develop niche models for tree species characteristic of the cerrado vegetation (woody savannas) of central South America, and to hindcast their distributions during the Last Glacial Maximum and Last Inter‐Glacial, we compiled a dataset of tree species checklists for typical cerrado vegetation (n = 282) and other geographically co‐occurring vegetation types, e.g. seasonally dry tropical forest (n = 355). We then performed an indicator species analysis to select ten species that best characterize typical cerrado vegetation and developed niche models for them using the Maxent algorithm. We used these models to assess the probability of occurrence of each species across South America at the following time slices: Current (0 ka pre‐industrial), Holocene (6 ka BP), Last Glacial Maximum (LGM – 21 ka BP), and Last Interglacial (LIG – 130 ka BP). The niche models were robust for all species and showed the highest probability of occurrence in the core area of the Cerrado Domain. The palaeomodels suggested changes in the distributions of cerrado tree species throughout the Quaternary, with expansion during the LIG into the adjacent Amazonian and Atlantic moist forests, as well as connections with other South American savannas. The LGM models suggested a retraction of cerrado vegetation to inter‐tableland depressions and slopes of the Central Brazilian Highlands. Contrary to previous hypotheses, such as the Pleistocene refuge theory, we found that the widest expansion of cerrado tree species seems to have occurred during the LIG, most probably due to its warmer climate. On the other hand, the postulated retractions during the LGM were likely related to both decreased precipitation and temperature. These results are congruent with palynological and phylogeographic studies in the Cerrado Domain.

Vegetation cover of Brazil in the last 21 ka: New insights into the Amazonian refugia and Pleistocenic arc hypotheses

“…The semi‐deciduous forest, however, less demanding in terms of soils, expanded its territory during the Mid‐H, following the drought increase at latitudes > 20° S. In short, the increase (not decrease) in temperature and water stress caused the expansion of these seasonal biomes, where the soil was favourable. The savanna, however, had a greater extent in the LGM (in disagreement with Bueno et al, ) and Mid‐H than nowadays. The higher temperature range in the basin and the compatibility with the leached landscape favoured its expansion, essentially into the open ombrophilous forest, an ecotonal biome of the Amazon border.…”

Ant diversity in Neotropical savannas: Hierarchical processes acting at multiple spatial scales

“…A total of eight variables were used as correlates of ant abundance, species density and rarefied richness, each related to one of four general hypotheses that may explain patterns of biodiversity. The hypotheses and corresponding correlates were as follows: (a) species–energy hypothesis: mean annual rainfall (RAIN) (obtained from the WorldClim database at ~1‐km resolution; Hijmans, Cameron, Parra, Jones, & Jarvis, ) and net primary productivity (NPP) (data averaged over the years 2000–2015 and obtained from NASA Earth Observation System at ~1‐km resolution, product MOD17; Zhao, Heinsch, Nemani, & Running, ); (b) historical factor hypothesis: refugial areas (RFA), historical stability in the Cerrado region based on the averaged values of historical climatic (Werneck et al, ) and environmental (Bueno et al, ) stability maps using three past scenarios plus the current climatic conditions and historical differences in rainfall (HDR) between current conditions (CR) and conditions during the last glacial maximum (LGM), calculated as the mean of three available scenarios: HDR = [(CR − LGM1) + (CR − LGM2) + (CR − LGM3)]/3 (WorldClim; Hijmans et al, ); (c) habitat heterogeneity hypothesis: diversity of land cover (DLC), calculated as the Shannon index of diversity of land cover types in each site (IGBP classes 6–11 of MCD12Q1 data product; Friedl et al, ) and heterogeneity in greenness (HET) – the complement of homogeneity – which is based on the similarity in the enhanced vegetation index between any pixel (~1‐km resolution) and its adjacent pixels (Tuanmu & Jetz, ); and (d) local environment hypothesis: volumetric percentage of coarse fragments (diameter > 2 mm) in the soil (FRAG) and the proportion of sand particles (diameter between 0.05 and 2 mm; parts per ton) in the soil (SAND) (obtained from SoilGrids database at 0.25‐km resolution; Hengl et al, ). We used geographical information systems to extract variables within the coordinates of each transect, which were then averaged per site (further details on variable extraction may be found in the Supplementary material, Appendix S1).…”

Dispersal assembly of rain forest tree communities across the Amazon basin

“…The existence of long-distance, transoceanic dispersal at an intercontinental scale suggests that there should be little to hinder dispersal across the flat, continuously forested Amazon basin given its lack of present-day physical barriers. Although there is some debate about the role of potential historical dispersal barriers in the Amazon, such as forest fragmentation during Pleistocene climate changes (17)(18)(19) and a large wetland complex (Pebas) or marine incursions that occupied much of western Amazonia in the Miocene (20,21), these are far less substantial impediments to plant dispersal than major oceans.…”