S. Schwinning et Aj. Parsons, A SPATIALLY EXPLICIT POPULATION-MODEL OF STOLONIFEROUS N-FIXING LEGUMES IN MIXED PASTURE WITH GRASS, Journal of Ecology, 84(6), 1996, pp. 815-826
1 In a previous paper, we outlined the physiological prerequisites for
population oscillations between a grass and a nitrogen-fixing legume,
such as clover. Here, we examine the field-scale consequences of patc
h-scale oscillations in legume content, using a cellular automaton wit
h variable hierarchy between the two species. 2 We define cell states
in the automaton by species content and soil N status. Grass-legume os
cillations at the patch scale are represented as an alternation betwee
n states of grass dominance (high N) and legume dominance (low N). To
this physiologically based population oscillation, we add local extinc
tions of legume and state-dependent success in legume invasion. 3 Legu
me populations oscillate at the field scale, given arbitrary initial c
onditions. However, spatially random perturbations to the soil N statu
s (e.g. urine) establishes a pasture structure that dampens the field
scale oscillation. The stabilizing pasture structure comprises moving
patches of legume dominance. This pattern was not predicted by our pre
vious, purely physiological model.4 The model highlights that a patchy
species distribution does not in itself mean the species is dispersal
limited. In this model, changing the dispersal ability of legumes pla
ys only a limited role in determining their proportion in the mixture.
Legume abundance depends as much on the rate at which favourable (low
N) sites become available for invasion. 5 Seasonal disturbances that
act uniformly across the field, such as winter (legume) mortality and/
or springtime fertilizer application, can lead to sustained field scal
e variations in legume content that are only partly explained by the l
evel of seasonal disturbance itself. Another large part is explained b
y previous years' legume contents. Pastures may therefore exhibit a 'm
emory' for legume performance which helps to explain the perceived 'un
predictability' of some grass-legume associations. 6 We argue that leg
ume dynamics in mixed pastures cannot be fully understood without comb
ining ecological and physiological concepts of species interactions at
three different scales: competitive interactions at the patch-scale,
dispersal at the between-patch scale, and seasonality at the field-sca
le.