Dr Tigga Kingston

Assistant Professor

Local and landscape level: ecological structuring processes and response to disturbance

I work with the insectivorous bats of South East Asia; a species-rich group that occurs as highly diverse assemblages in the threatened rainforests of the region.  In particular, the insectivorous bat assemblages in Malaysia are amongst the most species-rich in the world (Kingston et al. 2003, Kingston et al. 2006); even the assemblages confined to the forest interior can include more than 30 species. With a few exceptions, Old World insectivorous bats depend upon sophisticated echolocation systems to both orientate themselves and locate and capture prey.  Signal design is often viewed as the bat equivalent of a bird’s bill, with adaptations that enable a species to exploit certain resources and habitats but limit access to others, consequently much of my early work focused on the role that echolocation plays in coexistence mechanisms. Closely related syntopic species do indeed differ in their signal design (Kingston et al. 1999, J. Zool., 249:359-374; Kingston et al.2003, J. Mamm, 84: 205-215)), and when flight morphology is considered in tandem, there is clear evidence of competitive structuring in some guilds (Kingston et al. 2000, Oecologia, 124:332-342).

However, although a community is thought of as an assemblage of interacting organisms utilizing the same resource space, in practice studies of terrestrial vertebrate assemblages, particularly those in the tropics, are often spatio-temporal snapshots, missing many of the spatial and longitudinal interactions. In an attempt to overcome this, I have initiated a long-term, spatially-explicit study of insectivorous bat assemblages in a large area (620 km²) of contiguous primary rainforest in Peninsular Malaysia (Krau Wildlife Reserve). The objective is to determine the patterns and processes behind spatio-temporal variability in assemblage composition, and, to the best of my knowledge, is the largest study of any vertebrate community undertaken in an intact system. It is an ambitious project both logistically and in terms of theoretical advances needed for analyses and has taken four years to develop and generate the first round of publications. I have established a system of five 1 km² study grids, each at least 7 km apart and providing a network of over 20 km of trails. The bat communities at each grid are surveyed on a rotational basis (each survey lasting 6 weeks to 3 months), and we are currently implementing the fourth round of surveying. All study grids and trap locations have been mapped and brought into a GIS where they are linked to the capture records, generating a globally unique database of over 16,000 spatially-explicit, individualized records of 36 species. Each grid provides for an estimate of local and temporal population and assemblage variability, and across grid comparisons measure landscape variability. Now that the system has been established my current/ongoing research goals are as follows:

1. To determine the spatial and temporal variability in assemblage composition at the local and landscape scale.

Preliminary findings indicated that community composition varies enormously at both the local and landscape scale and over time, but quantification of these patterns required: the development of the spatially explicit capture database; a full review of assemblage analyses in bats (as bats violate many of the assumptions underlying popular diversity measures) (Kingston,  in press); and the development of new statistical approaches to identify and locate departures from randomness in the local distribution of species (Kingston & Gopal in revision for Ecology) (See BOX 1).  Having achieved these prerequisite objectives, I am now in the process of fully elucidating spatial variability in assemblage structure at the local and landscape scales for this system, and am currently devising new diversity metrics that can incorporate the spatial distribution of species using social network analysis. Repeated sampling in the coming years across the landscape will provide a measure of population fluctuations and the extent to which they are spatially autocorrelated and synchronous across the landscape

2. To investigate the ecological processes behind this spatio-temporal variability.

Spatial variation in the distribution may reflect spatially dependent or autocorrelated responses to a structured environmental variable, endogenous clustering processes, and spatially-mediated competitive interactions. The local autocorrelation surfaces (BOX 1) provide for tests of local relationships across data sets that can distinguish among these processes or indicate interactions between them. Relatively simple correlations within GIS can be implemented because the surfaces explicitly represent the significant spatial pattern, negating the need to partition out spatial autocorrelation effects before analysis. I am currently using these species surfaces to test whether spatially-mediated competitive interactions are structuring the assemblages (by identifying positive or negative spatial associations among species layers) and/or whether it is the distribution and availability of key niche variables that is driving spatial patterns. Two of the most important resources in bat assemblages are food and roosts; geo-referenced field data on roosting and foraging ecology have been collected for the past three years by my Malaysian and UK graduate students through an extensive radio-tracking program. Over 80 individuals of 7 target species have been studied and each year additional species are studied. Ultimately, the goal is to fully describe (and map where appropriate) the roosting ecology, foraging ecology and social biology of all species in the system, and to elucidate the extent to which individual behavior influences species- and community-level interactions.  

Whereas roosting ecology is likely a major determinant of landscape-level spatial variations in populations, temporal variability may be more closely associated with the availability of food resources. Insect populations are heavily influenced by climatic conditions, and our reproductive phenology datasets suggest that many species in this system time their breeding to ensure that lactation is coincident with a short period of rain in April/May, and that unusual weather patterns have a pronounced effect on reproductive activity and may lead to fluctuations in populations. This will be investigated further in the coming years, and forms the basis for research into the effects of climate change on bat populations and diversity.

3. To determine the impact of social organization on genetic structure at the landscape scale.

The radio-tracking and spatial analyses suggest that co-distributed species exhibit a range of social structures and roosting needs and are thus predicted to show contrasting patterns of gene flow. Because these populations are contiguous and the study site has been buffered from radical shifts in habitat distribution for thousands of (bat) generations, we can test predictions regarding the role of group size, roosting habits and local distribution in establishing genetic differentiation at the landscape-level, without encountering many of the confounding variables that have plagued other studies (e.g. historical vicariance). This work is being conducted in collaboration with Dr Stephen Rossiter (Queen Mary University of London).

4. To predict the relative vulnerability of species of the rainforest interior

The insectivorous bat species of the rainforest interior that are the focus of my research are extremely susceptible to environmental disturbance and are dependent upon intact stands of undisturbed forest. Post-hoc evaluations of species loss from disturbed and fragmented habitats can provide valuable insights into the dynamics of local extinction, but there is an urgent need to move from explanatory models and to develop predictive frameworks that can identify vulnerable species in intact communities before populations begin to decline. Both theoretical and empirical studies suggest that certain ecological traits or variables confer a greater risk of extinction. However, single traits are rarely sufficient as predictors; traits interact or may reinforce or replace each other in determining extinction proneness, but more accurate assessments of vulnerability can be made if the predictors combined into composite profiles. In this study I am developing composite profiles of vulnerability using those predictors for which there is greatest support and which can be derived from the trapping and radio-tracking protocols (e.g. abundance, spatial distribution, reproductive phenology, home range, longevity and population turnover rates, and landscape and temporal population variability). These risk profiles will be used to predict species’ survival in disturbed and fragmented habitats, and the predictions tested against actual species representation in degraded and fragmented areas that surround the reserve (PhD dissertation – Matthew Struebig, Queen Mary University of London). 

 

Community Ecology

Coelops robinsoni

Megaderma spasma

Acerodon celebensis

Nyctimene cephalotis

Rhinolophus shameli

Hipposideros ridleyi

Rhinolophus euryotis

Harpionycteris celebensis

Hipposideros cervinus