We compared the metapopulation dynamics of predator-prey systems with
(1) adaptive global dispersal, (2) adaptive local dispersal, (3) fixed
global dispersal and (4) fixed local dispersal by predators. Adaptive
dispersal was modelled using the marginal value theorem, such that pr
edators departed patches when the instantaneous rate of prey capture w
as less than the long-term rate of prey capture averaged over all patc
hes, scaled to the movement time between patches. Adaptive dispersal t
ended to stabilize metapopulation dynamics in a similar manner to conv
entional fixed dispersal models, but the temporal dynamics of adaptive
dispersal models were more unpredictable than the smooth oscillations
of fixed dispersal models. Moreover, fixed and adaptive dispersal mod
els responded differently to spatial variation in patch productivity a
nd the degree of compartmentalization of the system. For both adaptive
dispersal and fixed dispersal models, localized ('stepping-stone') di
spersal was more strongly stabilizing than global ('island') dispersal
. Variation among predators in the probability of dispersal in relatio
n to local prey density had a strong stabilizing influence on both wit
hin-patch and metapopulation dynamics. These results suggest that adap
tive space use strategies by predators could have important implicatio
ns for the dynamics of spatially heterogeneous trophic systems.