Trait-based sensitivity of large mammals to a catastrophic tropical cyclone.

Extreme weather events perturb ecosystems and increasingly threaten biodiversity1. Ecologists emphasize the need to forecast and mitigate the impacts of these events, which requires knowledge of how risk is distributed among species and environments. However, the scale and unpredictability of extreme events complicate risk assessment1-4-especially for large animals (megafauna), which are ecologically important and disproportionately threatened but are wide-ranging and difficult to monitor5. Traits such as body size, dispersal ability and habitat affiliation are hypothesized to determine the vulnerability of animals to natural hazards1,6,7. Yet it has rarely been possible to test these hypotheses or, more generally, to link the short-term and long-term ecological effects of weather-related disturbance8,9. Here we show how large herbivores and carnivores in Mozambique responded to Intense Tropical Cyclone Idai, the deadliest storm on record in Africa, across scales ranging from individual decisions in the hours after landfall to changes in community composition nearly 2 years later. Animals responded behaviourally to rising floodwaters by moving upslope and shifting their diets. Body size and habitat association independently predicted population-level impacts: five of the smallest and most lowland-affiliated herbivore species declined by an average of 28% in the 20 months after landfall, while four of the largest and most upland-affiliated species increased by an average of 26%. We attribute the sensitivity of small-bodied species to their limited mobility and physiological constraints, which restricted their ability to avoid the flood and endure subsequent reductions in the quantity and quality of food. Our results identify general traits that govern animal responses to severe weather, which may help to inform wildlife conservation in a volatile climate.

Anthropogenically driven spatial niche partitioning in a large herbivore assemblage.

Understanding how human activity can influence species distributions and spatial niche partitioning between sympatric species is a key area of contemporary ecology. Extirpations of large mammalian populations, the result of a 15-year civil war, within the Gorongosa National Park in Mozambique, have been followed by an extended period of restoration. The species-specific recovery of these populations has provided an ideal system to identify how niche partitioning between coexisting species is altered as a consequence of extreme disturbance events. Here, we aimed to understand how distribution patterns of grazing herbivores, as well as spatial niche overlap between them, changed between the pre- and post-war scenarios. We focused on the following four grazer species: buffalo (Syncerus caffer); sable (Hippotragus niger); waterbuck (Kobus ellipsiprymnus); and zebra (Equus quagga). Using long-term aerial survey data, we quantified range size for each species, as well as spatial niche overlap between each species pair, for pre- and post-war periods. Range size of buffalo and zebra decreased drastically from the pre-war period; with both species inhabiting subsets of their historical distribution in the park. Sable and waterbuck have both colonised historically avoided habitat, with waterbuck doubling their pre-war range size. Spatial overlap between all four grazers pre-war was significantly high, indicating niche similarity; however, this decreased in the post-war period, with some species pairs displaying spatial niche dissimilarity. Our findings highlight how population responses to anthropogenic disturbance can result in significant alterations to species' distributions, with consequences for patterns of niche similarity.

Gene flow and population structure of a solitary top carnivore in a human-dominated landscape.

While African leopard populations are considered to be continuous as demonstrated by their high genetic variation, the southernmost leopard population exists in the Eastern and Western Cape, South Africa, where anthropogenic activities may be affecting this population's structure. Little is known about the elusive, last free-roaming top predator in the region and this study is the first to report on leopard population structuring using nuclear DNA. By analyzing 14 microsatellite markers from 40 leopard tissue samples, we aimed to understand the populations' structure, genetic distance, and gene flow (Nm). Our results, based on spatially explicit analysis with Bayesian methods, indicate that leopards in the region exist in a fragmented population structure with lower than expected genetic diversity. Three population groups were identified, between which low to moderate levels of gene flow were observed (Nm 0.5 to 3.6). One subpopulation exhibited low genetic differentiation, suggesting a continuous population structure, while the remaining two appear to be less connected, with low emigration and immigration between these populations. Therefore, genetic barriers are present between the subpopulations, and while leopards in the study region may function as a metapopulation, anthropogenic activities threaten to decrease habitat and movement further. Our results indicate that the leopard population may become isolated within a few generations and suggest that management actions should aim to increase habitat connectivity and reduce human-carnivore conflict. Understanding genetic diversity and connectivity of populations has important conservation implications that can highlight management of priority populations to reverse the effects of human-caused extinctions.