Regeneration in a Pole-Stage Forest of Eucalyptus regnans Subjected to Different Fire Intensities in 1982

In 1982, fire burnt stands of Eucalyptus regnans F.Muell. in a relatively dry site in Victoria. In one area, the fire killed both canopy and understorey; in an adjacent area, only understorey was destroyed. Regeneration in the two areas was similar over the following year, but diverged thereafter to produce understoreys with different species dominance. In both stands, a poor supply of mature E. regnans seed in the crowns at the time of the fire resulted in relatively low initial density of seedlings: in the firekilled stand, this meant that closure of the canopy of the stratum was delayed for 5 or 6 years; however, in the understorey-killed stand, none of the E. regnans seedlings survived for 2 years. The soil seed bank was reduced more severely in the fire-killed than in the understorey-killed stand, although not all seed germinated in the first year. Vegetative regeneration of herbs and shrubs occurred from shallower layers of soil in the understorey-killed than in the fire-killed stand. An increase in soil fertility after the fire, as measured by seedling bioassay, was apparent only in the first season after the fire and was correlated with higher levels of available P. In the understorey-killed stand, fertility in the topsoil was greater than that in the fire-killed stand, and growth in diameter at breast height of dominant trees was significantly greater than in adjacent unburnt stands in the first few years after the fire. By comparison, when fire burnt through a site of higher rainfall after the maturation of the current crop of canopy-stored seed, regeneration was initially denser and growth considerably greater than that in the drier site. The study demonstrated that the course of secondary succession depends on site quality, timing of the fire in relation to seed production, soil seed germination, vegetative growth from protected organs, the severity of the fire, the presence or absence of browsing, and, in the long term, the frequency of recurrent fire.