Agroforestation of Grasslands in Southeast Asia: WaNuLCAS Model Scenarios for Shade-Based Imperata Control During Tree Establishment - Smallholder Tree Growing for Rural Development and Environmental Services - page 149

6.4 Discussion and Conclusion

The WaNuLCAS model could be used to simulate the various phases of an agro-
forestation process, including the possible interactions between initial tree
growth, the cropping phase till the tree canopy started to capture too much light
for the crop and the subsequent Imperata regrowth window. For any other tree
species with adequate tree parameterization a comparison such as made here can
be used to optimize tree spacing.Results for the simulations in the Philippines
and Indonesia agree that the length of time required for Imperata control varies
in the 3 to >15 year range, depending on the growth rate of the tree and the spac-
ing used.
The predicted duration of the Imperata regrowth window of two to five years
may be linked to the probability of fire-induced tree mortality and tree failure. The
probability for trees to survive this period is (1-p) n where p is the annual probability
of fire and n is the number of years. If the fire probability is 0.1, 0.2 or 0.3, the
probabilities of surviving the Imperata regrowth window of two or four years are
0.81 or 0.66, 0.64 or 0.41 and 0.49 or 0.24, respectively. The probability (p) of fire
entering the plot (p) will depend on the landscape context of the plot, the length of
time of exposure, n, depends on plot management.
For the trees with a relatively slow growth rate, the benefits of a longer initial
cropping period have to be weighed against the longer risk of the Imperata regrowth
window, as the duration of Phase A and Phase B is well correlated. Wider spacing
of trees, similarly has benefits for the cropping period but will lead to slower final
closure of the tree canopy. A landscape mosaic with plots in different phases of
agroforestation may enhance the overall probability of success, and allow for plots
of valuable, slow growing trees to be interspersed with faster growing trees that will
reduce fire risk for the surrounding plots.
While there is a wide range of options for the choice of tree species, most farm-
ers opt for a few well-known species. Out of a range of 83 tree species tested,
Otsamo et al. (1997) found that the exotics Acacia mangium , A. auriculiformis ,
Gmelina arborea , Paraserianthes falcataria and Cassia siamea were among the
most hardy survivors when planted in Imperata grasslands at low levels of manage-
ment intensity, while many trees from the local flora failed. Gmelina arborea is not
as popular among smallholders in Indonesia as Paraserianthes falcataria , Swietenia
macrophylla and Tectona grandis (teak; Roshetko et al. 2004). Otsamo (2000,
2002) described that Acacia mangium as a 'framework species' also provides the
best opportunities for regeneration of native tree flora after the early control of
Imperata . The specific properties of the trees and the dominant weed need to be
known, while generic statements about monocultures and biological diversity have
low predictive value (Collins et al. 2007).
As recommendations and conclusions we propose:

●

The choice of tree species and spacing must balance between benefits of a longer
cropping period with wider tree spacing, and the prolonged risks of fire in the
Imperata regrowth phase.