Biology Reference
In-Depth Information
personal communications), we will not discuss that work here. We will instead
focus on management of habitats invaded by
M. minutiflora
and
S. condensatum
.
M. minutiflora
was introduced to the Hawaiian Islands as livestock forage
(Parsons 1972). It invades mesic shrubland and open woodland ecosystems where
it promotes an increase in fire occurrence and intensity (Tunison et al. 2001).
Individual
M. minutiflora
are typically killed by fire, but the abundant seedbank is
a source of resilience and high, rapid postfire seed production of first-year plants
results in rapid recovery of
M. minutiflora
populations (D'Antonio et al. 2001a).
M. minutiflora
co-occurs with
S. condensatum
in many natural areas. This latter
species is from Central and South America where it is not typically associated with
fire-prone ecosystems. It readily invades submontane forests, competes with native
vegetation before fire (D'Antonio et al. 1998), and resprouts rapidly after fire
(D'Antonio et al. 2001b). Its relative,
Andropogon virginicus
(broomsedge), from
the southeastern USA, also occurs across these environments and is also associated
with increased fire frequency (Tunison et al. 2001).
Within Hawaii Volcanoes National Park, fires are associated primarily with vol-
canic activity and humans. Fire regimes began to change after the spread of
A. vir-
ginicus
,
S. condensatum
, and
M. minutiflora
within park boundaries in the 1960s.
Tunison et al. (2001) documented that fire frequency has increased by threefold and
fire size by sixtyfold since the establishment of these grasses. D'Antonio et al.
(2000) analyzed impacts of grass-fueled fires on native species diversity and cover
in 19 sites from the coastal lowlands to the upper submontane seasonal zone of
the Park and found the strongest impacts in the submontane zone where the prefire
dominants are intolerant of fire. These dominants include the native tree,
Metrosideros polymorpha
, and the native shrub,
Leptecophylla tameiameia
(for-
merly
Styphelia tameiameia
, Wagner et al. 1999). In the coastal lowlands
A. virgin-
icus
and
S. condensatum
-fueled fires regenerate toward at least some native species
although native diversity is reduced (Tunison et al. 1994).
M. minutiflora
-fueled
fires in both the coastal lowlands and submontane seasonal zone greatly reduce
native species.
By sampling burned vs. unburned forests across the same elevation and rainfall
within the submontane zone, Friefelder et al. (1998) documented that the homoge-
nous structure of the grass canopy in burned sites resulted in an approximately
threefold increase in wind speeds than were found above the canopy of unburned
forests (with grasses). This resulted in modeled fire spread rates that were 20 times
higher than those in unburned forest with exotic grasses in the understory. Mack
et al. (2001) and Mack and D'Antonio (2003a, b) documented extensive changes in
productivity, microclimate, and nitrogen cycling in burned compared with unburned
woodland. They found that by greatly altering species composition including the
elimination of native woody species, fire created an ecosystem that has much lower
primary production and is much leakier for nitrogen. For example, Mack et al.
(2001) report that net primary production is reduced by 55% in burned sites com-
pared with nearby unburned counterparts, but that annual net nitrogen mineraliza-
tion rates are an order of magnitude higher in burned sites. The lack of primary
production during periods of high soil N mineralization resulted in periods with
