|dc.description.abstract||ENGLISH ABSTRACT: As evidence of climate change and its impact on biodiversity continues to grow, anticipating and understanding ecological responses to climate change is ever more critical. In fire-prone ecosystems, such as those found in the Cape Floristic Region (CFR), a major concern is that changes in climate will likely lead to dramatic shifts in fire activity (e.g. increased fire frequency) that will significantly affect the distribution, composition, and functioning of vegetation. Effectively mitigating and/or adapting to the potential loss of biodiversity and altered ecosystem function in this region hinges on an in-depth understanding of how vegetation in the CFR interacts with the environment and, more importantly, how vegetation will respond to changes in both climate and fire regime.
Our understanding of how climate change may impact vegetation is largely derived from distribution models. Vegetation distribution models have been used for decades to investigate species-environment relationships, predict future distribution patterns, and test ecological theories. These models are founded on the premise that vegetation distributions are determined by the spatial distribution of environmental variables that are significantly correlated with, or limit, plant distributions. However, while much emphasis is placed on the role of climate and topography as key determinants of vegetation distributions, other critical ecosystem components (e.g. fire regime) that have significant effects on the composition and distribution of vegetation are rarely incorporated in vegetation distribution models. Given the importance of fire as a driver of vegetation formations and assemblages in the CFR, the exclusion of fire variables from vegetation distribution models potentially constrains the generation of accurate and appropriate information, critical for the management and conservation of biodiversity in the region.
The exclusion of fire covariates from distribution studies is partly a result of a lack of fire data, coupled with the widely accepted, but limited, view that climate is the chief determinant of species distributions and also a key determinant of fire regime. To this end, a proxy for fire return interval data, derived from vegetation recovery rates estimated from satellite data, combined with climate and edaphic data, was used to model and analyse the distribution of fynbos vegetation in CFR. Firstly, the importance of fire as a determinant of fynbos species distributions, relative to climate, was evaluated. This was complimented by an assessment of the effect of life history traits on plant species sensitivity to changes in ecological regime. To achieve this first objective, the distributions of 52 closely related fynbos plant species pairs (104 species), classified across two growth forms
Stellenbosch University https://scholar.sun.ac.za
(graminoids and shrubs) and their respective fire response strategies (seeders and resprouters) were modelled using Maxent, and subsequently analyzed. Secondly, the potential impacts of changes in climate and fire regime on future fynbos distributions were assessed by modelling the future distributions of 22 fynbos vegetation types under 44 Phase 5 Coupled Model Intercomparison Project (CMIP5) general circulation models (GCMs) using multinomial linear regression. Lastly, an overlay analysis of projected distributions of fynbos species and vegetation types was used to assess whether fynbos species respond in unison or as individuals to changes in climate and fire regime. The selected species in this final analysis consisted of 74 endemic species and 358 important species (species that are either high in abundance or frequency of occurrence, or predominant in a given vegetation unit), which facilitated a comparison of the potential impacts of changes in climate and fire regime between the two sets of species.
Findings from this research identified fire return interval as a major determinant of fynbos species distributions. Although, the predictive power of the fire variable was greatly reduced when considered in conjunction with the other climate and edaphic variables, it was still among the most important predictors, and including fire data has the potential to add to our understanding of plant species distributions in fire-prone ecosystems. This was particularly apparent in the case of seeder graminoids and shrubs, where both graminoids and shrubs were found to be negatively associated with longer fire return intervals, while seeder species were significantly more sensitive to fire than resprouters.
Projected changes in fire return interval and temperature will potentially have a significant impact on future vegetation distributions, with vegetation types with longer fire return intervals and warmer summer and winter temperatures being at most risk. It was also noted that projected changes in fire regime will likely have a greater impact on vegetation distributions than changes in rainfall regime. Comparing the distribution models of endemic and important species with models for the major vegetation types highlighted that species responses to changes in climate and fire regime largely conformed to the Clementsian concept of communities as organisms, with less than 30% of the species showing individualistic responses. As a result, the species composition of all vegetation types was altered when projected under future scenarios, with species from the present-day vegetation types either being lost or retained, while others were gained, in the projected vegetation types. The change in species composition largely stemmed from the replacement of some species in the present-day vegetation type by the same number of different species in the corresponding future vegetation type. The implications of this is that the underlying
species composition of fynbos vegetation types will likely be altered under future climate and fire regime, thus disrupting the functioning of those vegetation types.
The role of natural disturbance regimes, such as fire, in determining species distributions is generally overshadowed by the long-standing view of climate being the chief driver of species distributions. This research provides evidence for the contribution of fire in shaping species' distributions in fire-prone ecosystems, and highlights the need for the development and inclusion of estimates of fire regime components in vegetation distribution studies and vulnerability assessments. This is especially important since fire regimes are sensitive to a range of global change drivers beyond just changing climate (e.g. land use and invasive species).||en_ZA