Using remote sensing to detect historical disturbances in a tropical forest
In a world increasingly dominated by human disturbances, most forests, tropical and otherwise are being or have been disturbed. In most tropical countries around the world record keeping of forest logging and clearing is poor, so it is difficult to assess the actual extent of damage to the forest caused by activities such as selective logging, clearing for agriculture, poaching etc.. I am using an active remote-sensing technique called LiDAR (Light Detection and Ranging) to identify the effects of historical selective logging, which ceased in 1989, on a moist tropical forest in the Gola Rainforest National Park in Sierra Leone, West Africa. By characterizing the structure of the canopy of old-growth forest and comparing it to canopy characteristics of forest recovering from historical selective logging I manage to accurately classify disturbed patches of forest and assess the level of recovery from disturbance. Such accurate identification of disturbed forests has implications for biodiversity and ecosystem services conservation and implementation of programs aiming to reduce green-house gas emissions, such as the UN-REDD program.
mechanisms controlling Ecological community assembly
Until recently most ecologists viewed the ecological niche theory as the most plausible explanation of community assembly, based on species' habitat requirements. Around the turn of the millennium a new theory emerged - Hubbell's neutral theory of biodiversity, which offered a simplistic and parsimonious model producing many of the predictions of the niche model, while assuming species are ecologically equivalent, and thus community assembly is independent of environmental conditions. This brought forward hotted debate, which still carries on. Recently however, a new concept emerged, suggesting the niche and neutral theory are non-mutually exclusive. In fact, they were suggested to form a single continuum along which all natural occurring ecological assemblages can be found. I am studying the niche-neutral continuum using an array of methodologies, including theoretical modelling, empirical observation analyses and experimental designs using bacteria communities in bio-reactors (also know as chemostats).
Multi-scale analyses of species composition - environment relationships in bird and mammal assemblages
Factors affecting species composition include deterministic niche processes, stochastic dispersal events and inter-specific composition. The effects of these factors depends greatly on the spatial and temporal scale of observation, and are generally operating at multiple scales. We conducted rigorous multivariate analyses to characterize the relative importance of different environmental variable groups (climate, land-use land-cover, topography and productivity) on the composition of breeding birds and mammals in the continental USA and Australia at 12 spatial scales, ranging from local (3km grains) to continental (>100km grains). We found, as expected, that climate and land-use land-cover variables explain most of the variance in species composition related to environmental conditions, at all spatial scales. Further, we broke down the relationship into two major components - ecological affinity of species and range of conditions, showing that generally, climate has a much stronger effect on species composition than any other variable. Our results also re-establish the fact that human related environmental changes, either direct (land-use changes) or indirect (climate change) have a major effect on species composition at maltiple spatial scales.