Floristic diversity and patch dynamics in hyper-disturbed tropical lowland rainforest fragments in north Queensland: no evidence for cyclone-disturbance and fragmentation driving species assemblages towards early-successional states

Abstract

Tropical cyclones reshape the structure, composition and successional trajectories of forest ecosystems. Climate warming is expected to increase the intensity and impact of these storms while habitat fragmentation further modifies trajectories of response. Many studies over the past decade suggest that tropical forest fragments are locked into a future dominated by edge-favoured pioneer species, dramatic loss of late-successional, large-fruited species, and invasion by exotic weeds. However, this study shows that severely storm-damaged and fragmented tropical forests are remarkably resilient, with a capacity for rebuilding and maintaining plant species composition and diversity. My study followed two severe tropical cyclone events, Cyclones Larry (2006) and Yasi (2011) in the Australian Wet Tropics. To date, no comparable investigation has measured the effects of successive severe tropical cyclone events on different forest habitats in lowland rainforest. My study investigated: 1) immediate effects of a second severe cyclone on the structural characteristics of different forest habitat types; 2) family and species-level responses to damage and short-term survival rates; 3) plant community and species-level assemblages across different forest habitats; and 4) life-history successional characteristics of species for different habitat types. My results showed all trees sustained some level of damage (i.e. minor to severe) due to the effects of these two severe cyclone events. About 75% of trees had their main stem snapped compared to 11% of trees with major breakage of branches and less than 4% were uprooted. 10% of all trees sustained only minor damage including partial defoliation, twig-snapping and minor branches snapped. Snapping of tree trunks was higher in fragmented forest compared to continuous forest whereas snapping of major branches was significantly higher in continuous forest and edges. Trees resprouted within weeks of the cyclone and 96% of standing stems continued to show vigourous growth after eighteen months. Although there was a dramatic loss of large canopy and emergent trees during Cyclone Larry five years earlier, 83% of all stems in my plots were identified as belonging to mid-late and late successional species, while only 13% of stems were early-mid successional species. These stems are mainly saplings (≤ 10-20cm DBH) which have survived successive severe cyclone events within the understorey 'vine tangles.' I found no evidence for proliferation of early successional or pioneer species in any of the habitat types, including fragmented forest sites, within the time-frame of this study. No evidence was found for elevated levels of exotic weed invasion following these events with these species comprising less than 1% of total assemblages. All forest habitat types showed an unexpected capacity for resistance and resilience following the combined effects of fragmentation and two severe tropical cyclones. My data suggest that forest habitat types influence successional life-history characteristics but are not driving species assemblages in fragmented forests towards proliferation of short-lived, edge-favoured generalists (i.e. pioneer species).