Rain forest invasion of eucalypt-dominated woodland savanna, Iron Range, north-eastern Australia: I. Successional processes

Aim: To explore successional processes associated with rain forest expansion in Eucalyptus-dominated woodland savanna vegetation in north-eastern Australia. Location: Iron Range National Park and environs, northeast Queensland, Australia. This remote region supports probably the largest extent of lowland (< 300 m) rainforest remnant in Australia. Rainfall (c. 1700 mm p.a.) occurs mostly between November and June, with some rain typically occurring even in the driest months July-October. Methods: (1) Sampling of rain forest seedling distributions, and other vegetation structural attributes, in fifteen 10 x 10 m quadrats distributed equi-distantly between mature rain forest margins (range: 70-840 m), at each of 10 sites which were open-canopied vegetation in 1943. (2) Assessment of relationships between rain forest seedling densities and structural characteristics, including distance-to-rain forest-margin, canopy height, stem density. (3) Assessment of lifeform and dispersal spectra for defined vegetation structural types. Results: Rates of rain forest invasion were found to be substrate-mediated. Transects established on hematite schist, diorite, riverine alluvium, and granite developed closed canopies (termed phase III sites) by 1991. The remainder (four transects on poorly drained colluvial/alluvial sediments; one on dune sands) continued to occur either as grassy woodland (phase I), or with developing rain forest understoreys (phase II). Rain forest seedlings were observed at maximum sampled distances from mature rain forest margins at all sites. Lifeform and dispersal spectra data illustrated that: (1) the proportions of woodland trees, shrubs and graminoids declined with successional phase, with concomitant increases in rain forest primary trees and all other lifeform categories save rain forest trees; (2) the proportions of major dispersal syndromes did not vary between successional phases, neither for rain forest nor woodland taxa. Main conclusions: Rain forest seedling distribution data for phases I and II sites illustrate three successional processes: margin extension - seedling density significantly negatively correlated with distance from mature rain forest margins at two sites; nucleation - seedling densities significantly positively correlated with tall trees at two sites; and irruption - seedling densities at two sites neither correlated with distance from mature rain forest margins, nor with measured vegetation structural features. The observation of irruptive rain forest regeneration at these sites, combined with decadal-scale rain forest canopy development at the five remaining sites, illustrates that under conditions conducive to growth (moisture, substrate), low fire disturbance, and maintenance of diverse dispersal processes (high frugivore richness), rain forest can rapidly invade regional landscapes.