Mechanisms determining the coexistence of open- and closed-canopy biomes

Open- (e.g. grassland, savanna, shrubland) and closed-canopy (e.g. forest) biomes frequently coexist in the same landscape, where open environments tend to be fire-prone with higher light, but lower nutrient and water availability than closed environments. Environmental heterogeneity could select for divergent floristic assemblages and adaptive traits, from which emergent differences in resource availability and fire incidence contribute to excluding species from the alternate habitat. In this thesis, I investigated whether the coexistence of open–closed canopy biomes, such as forest and fynbos in the Cape Floristic Region, is contingent on environmental heterogeneity coupled with contrasting species traits. Given the heterogeneity in multiple environmental properties between open- and closed-canopy biomes, I hypothesized that boundaries between open- and closed-canopy biomes will display greater floristic turnover compared to boundaries between structurally similar biomes (e.g. open- and opencanopy biomes). To explore this, genus- and family-level turnover were correlated with climate, fire, leaf area index (LAI: proxy for understorey light) and soil properties across biome boundaries in South Africa. Both genus- and family-level turnovers were highest across open–closed boundaries and most strongly predicted by increased differences in LAI, suggesting that contrasting light regimes provide significant adaptive challenges for plants. The potential effect of contrasting light regimes is highlighted by the absence of open-canopy species from forest understoreys, where low, dynamic light could limit the ability of plants to acquire sufficient carbon. This apparent shade intolerance led to the hypothesis that open-canopy species lack the traits to maintain a positive carbon balance under low and dynamic light. To test this, leaf traits and photosynthetic response to continuous or dynamic light were compared between forest and fynbos species grown under three light treatments. Fynbos species experienced high mortality under shade treatments, produced leaves that were thicker, up to 1000 times smaller, had lower photosynthetic rates (0.8 versus 3.4mol m-2 s -1 ) under continuous low light (400 mol m-2 s -1 ) and lower light-use efficiency during dynamic light sequences than forest species. These differences imply that shade intolerance in fynbos species is associated with traits that are inefficient at harvesting light and require relatively continuous high intensity light for carbon assimilation. Moreover, these inefficiencies would make it difficult to support the carbon intensive traits (e.g. cluster roots, lignotubers, sclerophyllous leaves) that facilitate fire survival and nutrient acquisition/conservation in open habitats. In contrast, forest species are able to colonize open habitats during the long-term absence of fire, implying that they are able to tolerate high light and low nutrient conditions. Given that plants frequently cope with contrasting conditions through the expression of phenotypic plasticity, it was hypothesized that closed-canopy species possess greater plasticity than open-canopy species. To assess this, the response of leaf traits and foliar nutrition to changes in LAI and soil nutrition were compared between forest and fynbos species in the field. Leaf size and specific leaf area in forest species correlated positively with LAI and soil nutrition, whereas fynbos species response was weak, suggesting that forest species are more plastic. This plasticity may be realised by the variable light conditions forest species experience through their canopy and the occupation of higher nutrient soils, which alleviate belowground constraints. By comparison, the occupation of low nutrient soils by fynbos may inhibit plasticity given the selection of inflexible, conservative leaves. Consequently, I propose that the coexistence of open- and closed-canopy biomes arises from the steep turnover in selective regimes, which together with the contrasting adaptive traits and degrees of phenotypic plasticity they require, act together to competitively exclude species from the alternate habitat