The impacts of human-mediated disturbances on birds and reptiles in tropical savannas

Disturbances influence the structure of many ecosystems, determining environmental\ud and biological heterogeneity. Human-mediated disturbances, including introduced plant\ud species and fire, have the ability to alter ecosystem-level processes and properties,\ud modify habitat structure and, as a consequence, influence faunal assemblages. This\ud thesis examines the impacts of introduced plant species and fire management practices\ud on vertebrate assemblages in grazed tropical savannas in northern Queensland.\ud \ud Invasive introduced plant species pose a major threat to native environments. Rubber\ud vine (Cryptostegia grandiflora) is an environmental weed that invades native riparian\ud habitats in northern Australia. Small ground-dwelling lizards may be negatively\ud affected as rubber vine replaces and fragments native habitat. Field observations of\ud reptiles in habitat invaded by rubber vine recorded only a single lizard in rubber vine\ud vegetation, compared to 131 lizards in nearby native vegetation. As rubber vine\ud vegetation contains features that superficially resemble native habitat, such as leaf litter,\ud the avoidance of rubber vine suggests that rubber vine has underlying characteristics\ud that create a suboptimal environmental for lizards.\ud \ud Two species of native skinks (Carlia munda and C. pectoralis) and the invasive plant\ud rubber vine were used as a model system to determine possible underlying mechanisms\ud driving avoidance of non-native plants by fauna. In semi-natural enclosures, lizards\ud discriminated between leaf litter types: 85% of C. pectoralis and 80% of C. munda\ud chose native leaf litter over rubber vine, indicating a clear preference for native habitat.\ud In comparison to native habitat, rubber vine provided a suboptimal environment for\ud litter-dwelling lizards with lower ambient temperatures, reduced availability of prey and\ud a reduction in camouflage from predators (dissimilar leaf and lizard shapes). Thus,\ud three possible mechanisms were identified by which an introduced plant species can\ud alter the availability of resources in an environment, making it less attractive to native\ud fauna. As rubber vine is a Weed of National Significance, management of this species\ud is a priority.\ud \ud Fire plays a pivotal role in structuring ecosystems and often occurs as a humanmediated\ud disturbance for land management purposes, including management of\ud introduced plants. Rubber vine is susceptible to fire, and burning for weed control may\ud be implemented in riparian zones of tropical savannas where rubber vine is prevalent.\ud Although tropical savannas are considered fire-adapted ecosystems, riparian vegetation\ud and associated fauna may be less resilient to the effects of fire. Variations in fire\ud regimes alter the environment in different ways, and the type of fire may govern the\ud response of faunal assemblages. Using replicated experimental fire treatments, imposed\ud on two habitats (riparian and adjacent woodland), I examined the responses of reptiles\ud and birds in the short- and longer-term to a range of fire management practices used to\ud control rubber vine.\ud \ud An important component of fire regime is the season of burn. In tropical savannas,\ud most fire management occurs during the dry season; however, wet season burning is\ud often used for pastoral management and may be useful for controlling introduced plant\ud species. Initially, only one species of reptile responded strongly to burning, with few\ud differences detected between burning seasons. Abundances of the skink C. munda were\ud higher in burnt sites and may reflect temporary changes in food availability, or a\ud reduction in rubber vine. However, the overall structure of the reptile community was\ud driven by habitat type (riparian versus woodland) rather than burning, suggesting most\ud reptiles were responding to broader environmental factors. Within three years of\ud burning, reptiles were least abundant in dry season burnt sites, a result mostly driven by\ud the abundance of the small terrestrial gecko, Heteronotia binoei, which was commonly\ud observed in unburnt and wet season burnt sites. In addition, litter-associated species,\ud including the skink C. pectoralis, were rarely observed in burnt habitat and fewer\ud species were detected in the wet season burnt sites.\ud \ud Both season of burn and time since fire also significantly influenced bird assemblage\ud responses. Within 12 months of fire, burning tended to benefit several bird species and\ud feeding groups, with higher overall abundances of birds observed in the sites, although\ud species that favoured dense vegetation (e.g. red-backed fairy-wren, Malurus\ud melanocephalus) were rarely observed in burnt habitat. Responses of feeding groups,\ud including insectivores, nectarivores and carnivores, suggest that burning may have\ud temporarily increased food resources. In the short-term, assemblage of birds tended to\ud reflect whether or not a site was burnt, rather than burning season. However, four years\ud following burning, dry season burnt sites were composed of a different bird assemblage\ud than unburnt and wet season burnt sites. In addition, dry season burnt treatments were\ud characterised by lower bird abundances, especially nectarivore and granivore feeding\ud groups and the insectivorous white-throated honeyeater (Melithreptus albogularis). As\ud dry season burning removed more vegetation than wet season burning, birds may be\ud responding to a reduction in habitat complexity, and subsequent changes in food\ud resources.\ud \ud The frequency with which a habitat is burnt is another critical component of fire regime,\ud and may ultimately determine faunal assemblage responses. Using a fully replicated\ud Before-After-Control-Impact (BACI) design I examined the impacts of repeated\ud burning on bird assemblages. In contrast to unburnt or singularly burnt sites, the\ud repeated burning significantly reduced bird abundance and species richness. Repeat\ud burning also altered the feeding group structure of sites. In particular, frugivorous and\ud insectivorous birds were adversely affected by the second fire. Vegetation complexity\ud was lower in both burning treatments, but the repeatedly burnt sites contained less\ud native fruiting shrubs, especially currant bush (Carissa ovata), which was an important\ud food and shelter source for several species. Repeatedly burning an area in a short time\ud frame may reduce key resources, other than vegetation complexity, such as food\ud availability or foraging opportunities.\ud \ud The use of fire is considered necessary for the maintenance of tropical savannas.\ud However, high impact individual fires may detrimentally affect habitat structure and\ud faunal assemblages at a local scale. In particular, my results suggest that overall bird\ud and reptile assemblages are strongly influenced by management burning, including\ud variations in burning season and fire frequency. In an attempt to overcome potential\ud negative impacts of burning, ecologists have suggested implementing mosaic burning,\ud where a variety of burning regimes are employed. Although mosaic burning\ud theoretically provides a diversity of habitat types that consequently maintain high faunal\ud diversity, my results suggest that some measures of diversity (e.g. species evenness)\ud may be compromised by mosaic burning.\ud \ud In summary, this study provides evidence that introduced invasive plants and\ud management burning play a key role in shaping landscapes and associated faunal\ud communities. Fauna respond to disturbance-induced changes in microhabitat and\ud vegetation structure, food availability or foraging opportunities, and habitat\ud requirements (e.g. temperature). The role of multiple human-mediated disturbances in\ud influencing faunal responses in my study is of particular importance. My research was\ud conducted in landscapes already disturbed by grazing and invasive species, and the\ud responses of fauna may be caused by cumulative impacts. In areas where multiple\ud disturbances already influence landscapes, the resilience of faunal assemblages to fire\ud management practices may be lower than previously predicted