Transformation of intact vegetation into new kinds and configurations of human-modified habitats is a well established driver of biodiversity loss. Following initial conversion, many human-dominated landscapes are then subject to further large-scale changes in land use. The impacts on biodiversity of repeated changes in land use remain poorly known, particularly how changes in the matrix interact with initial patterns of vegetation clearing. We used an 18-year study of birds in remnant patches of endangered temperate woodland in south-eastern Australia to quantify the spatial and temporal effects of successive land use transformation in the surrounding landscape. We examined bird response to: (1) initial patterns of landscape modification (creating semi-cleared grazing land dominated by pastures that surrounded remnant woodland patches), (2) subsequent establishment and maturation of exotic tree plantations on the pastures surrounding woodland patches, and (3) additive and interactive effects of both types of landscape transformation. The majority of the 57 bird species modelled responded to conversion of grazing land to exotic plantations, either independently from initial patterns of landscape transformation (20 species), or interactively (18 species) or additively (15 species) with initial landscape transformation. The occurrence of only one species (the Common Bronzewing) was related to patterns of initial transformation but not subsequent transformation due to plantation establishment. Thus, despite many characteristics of the woodland patches within the plantation remaining largely unaltered throughout our 18 year investigation, the matrix had a profound effect on the kinds of species inhabiting them, with such impacts often magnified over time as the matrix continued to change. Plantation establishment triggered new regional-level spatial processes with effects on birds detected in woodland patches up to 2 km away from the plantation. Matrix conversion selected for species with different traits (size, diet and movement patterns) compared to the initial transformation, suggesting it is acting as a different filter on the bird community. New kinds of landscape transformation (such as plantation establishment on previously cleared land) can radically affect the species that have persisted for many decades in previously modified landscapes. This highlights the challenges, but also opportunities, for conserving taxa in ever changing human-dominated environments.

The formal testing of mating system theories with empirical data is important for evaluating the relative importance of different processes in shaping mating systems in wild populations. Here we present a generally applicable probability modelling framework to test the role of local mate availability in determining a population’s level of genetic monogamy. We provide a significance test for detecting departures in observed mating patterns from model expectations based on mate availability alone, allowing the presence and direction of behavioural effects to be inferred. The assessment of mate availability can be flexible and in this study it was based on population density, sex ratio and spatial arrangement. This approach provides a useful tool for (1) isolating the effect of mate availability in variable mating systems and (2) in combination with genetic parentage analyses, gaining insights into the nature of mating behaviours in elusive species. To illustrate this modelling approach, we have applied it to investigate the variable mating system of the mountain brushtail possum (Trichosurus cunninghami) and compared the model expectations with the outcomes of genetic parentage analysis over an 18 year study. The observed level of monogamy was higher than predicted under the model. Thus, behavioural traits, such as mate guarding or selective mate choice, may increase the population level of monogamy. We show that combining genetic parentage data with probability modelling can facilitate an improved understanding of the complex interactions between behavioural adaptations and demographic dynamics in driving mating system variation.

Environmental disturbance is predicted to play a key role in the evolution of animal social behaviour. This is because disturbance affects key factors underlying social systems, such as demography, resource availability and genetic structure. However, because natural disturbances are unpredictable there is little information on their effects on social behaviour in wild populations. Here, we investigated how a major wildfire affected cooperation (sharing of hollow trees) by a hollow-dependent marsupial. We based two alternative social predictions on the impacts of fire on population density, genetic structure and resources. We predicted an adaptive social response from previous work showing that kin selection in den-sharing develops as competition for den resources increases. Thus, kin selection should occur in burnt areas because the fire caused the loss of 80% of hollow-bearing trees, but no detectable mortality. Alternatively, fire may have a disruptive social effect, whereby post-fire home range-shifts ‘neutralise’ fine-scale genetic structure, thereby removing opportunities for kin selection between neighbours. Both predictions occurred: the disruptive social effect in burnt habitat and the adaptive social response in adjacent unburnt habitat. The latter followed a massive demographic influx to unburnt ‘refuge’ habitat that increased competition for dens, leading to a density-related kin selection response. Our results show remarkable short-term plasticity of animal social behaviour and demonstrate how the social effects of disturbance extend into undisturbed habitat due to landscape-scale demographic shifts. We predicted long-term changes in kinship-based cooperative behaviour resulting from the genetic and resource impacts of forecast changes to fire regimes in these forests.