Doctor of Philosophy, University Of New South Wales (2015)
Master of Science, University Of Sydney (2011)
Bachelor of Science, University Of Cape Town (2015)
Tadashi Fukami, Postdoctoral Faculty Sponsor
Detecting state changes for ecosystem conservation with long-term monitoring of species composition.
2017; 27 (2): 458-468
Effective conservation requires an understanding not only of contemporary vegetation distributions in the landscape, but also cognizance of vegetation transitions over time with the goal of maintaining persistence of all states within the landscape. Using a state and transition model framework, we investigated temporal transitions over 31 years in species composition among five upland swamp vegetation communities in southeastern Australia. We applied fuzzy clustering to document transitions across communities; evaluated the resilience and resistance of communities to change; and explored the relationship between ecosystem states and major environmental factors posited to structure the system. We also evaluated the predictive ability of an established vegetation dynamics model. We found that community composition remained stable or underwent reversible or directional transitions depending on the vegetation type. Wetter communities (Ti-tree thicket and Cyperoid heath) were more stable (i.e., resistant) while drier communities showed a greater propensity to transition (i.e., had lower resistance) under the observed disturbance regime (low variance fire intervals). The resilience of drier communities differed under this regime, with Banksia thicket showing reversible compositional change, while Restioid heath and Sedgeland showed directional change. In accord with an established conceptual model, we found that communities were distributed along a hydrological gradient. In addition, vegetation structure, along with light penetration to ground level, differentiated communities. However, internal dynamics of drier communities were complex: differences in fire regime (penultimate fire interval in 2014 and number of fires since 1965) were unable to predict differences in community membership among sites. Aspects of the fire regime are expected to be more important predictors if fire intervals vary more strongly among sites in the future. Fuzzy clustering of compositional data allows managers to track community transitions over time and facilitates planned interventions for conservation purposes.
View details for DOI 10.1002/eap.1449
View details for PubMedID 28207176
Using a model based fourth-corner analysis to explain vegetation change following an extraordinary fire disturbance.
2016; 182 (3): 855-863
In ecosystems where large-scale disturbances are infrequent, the mode of succession may be difficult to discern and floristic surveys alone cannot be used determine the underlying processes causing vegetation change. To determine the causes of vegetation change in response to a large-scale fire event, we combined traditional floristic survey data, plant functional traits and environmental variables in a model-based solution to the fourth-corner problem. This approach allowed us to describe the trait-environment relationship and provides an intuitive matrix of environment by trait interaction coefficients. We could then quantify the strength and direction of associations between plant traits, species life-forms and environmental factors in two alpine plant communities over nine years post-fire. Initially, the fire drastically reduced vegetation cover and species density to very low levels. The fourth-corner analysis interaction coefficients indicated that over the course of the nine-year study a high abundance of graminoids, a low abundance of shrubs, tall species and those with high leaf dry matter content had the strongest associations with the two plant communities. We also found evidence for functional homogenisation between these two communities using this novel technique. Analysing plant traits and species responses post-fire in this manner can be used to infer the ecological processes driving shifts in vegetation.
View details for DOI 10.1007/s00442-016-3700-8
View details for PubMedID 27573617