A spatially explicit metapopulation model was used to examine the effects o
f landscape characteristics (patch sizes, interpatch distances, patch produ
ctivity) and dispersal behavior (maximum dispersal distance) on the source-
sink dynamics of white-tailed deer (Odocoileus virginianus) in a simulated
landscape. Specifically, the model was used to test the hypothesis that pat
terns in primary productivity, which affect the growth of deer populations,
will interact with, and potentially alter, the effects of landscape struct
ure on metapopulation dynamics. The model, constructed in the STELLA modeli
ng environment, combines a density-dependent Leslie matrix population model
with a simple habitat/non-habitat representation of an artificial landscap
e to simulate the growth and spread of deer populations in a spatial framew
ork. A set of 60 simulations were run, measuring the average net emigration
per habitat patch for maximum dispersal distances of 5, 10 and 20 kin and
under four net primary productivity (NPP) conditions: (1) constant NPP per
km(2), (2) NPP varying with patch size; (3) NPP varying inversely with patc
h size, and (4) NPP varying over time. Significant differences were observe
d for effects of patch size, dispersal distance, NPP conditions and all two
- and three-way interactions on patterns of net emigration. Specifically, i
ncreasing the maximum dispersal distance led to magnifications in net emigr
ation patterns (i.e. patches exhibiting a positive trend in net emigration
became more positive and vice versa ). Changes in NPP altered the magnitude
of the source-sink trends, sometimes reversing the relationship, particula
rly for populations in smaller patches. The results illustrate the importan
ce of multiple interacting landscape factors in influencing metapopulation
dynamics, indicating that shifting patterns of productivity can significant
ly alter the source-sink patterns of metapopulations when spatial factors a
re explicitly considered. Implications for management and conservation, as
well as future model improvements, are briefly discussed. (C) 2001 Elsevier
Science B.V. All rights reserved.