Incremental (distance-dependent) costs and benefits of dispersal have received less attention than those that are qualitative. We present a dynamic programming model of settling behavior using parameters estimated from a field study of dispersal in desert isopods, Hemilepistus reaumuri, which walk up to thousands of meters before settling in new or already-established burrows each spring. The model shows that incremental costs of dispersal lead to right-skewed population dispersal patterns, in contrast to cost-free systems or those with unitary costs (i.e., a one-time cost of leaving a group or patch). In the model, continuous variation in habitat quality, as opposed to discrete suitable vs. unsuitable sites, allows for trade-offs with dispersal costs that lead to shifts in the likelihood of settling in a patch of a given quality. Thus, measurement of quantitative costs and benefits of movement are needed to understand population dispersal distributions. These costs or benefits may be observable during or after movement, and we examined both pre- and postsettling incremental consequences of dispersal. The daily mortality rate of traveling isopods was 4.2% during the dispersal season, higher than that of settled individuals. Successful settlers traveled more slowly, but burrows started in midseason were most likely to succeed. More distant burrows failed more often, suggesting either an additional cost of movement or a difference in the quality of individuals traveling different distances. The predicted mean dispersal duration from the simulations matched observed values closely, but was based on an unrealistic assumption of perfect knowledge of habitat quality, suggesting some other factor favors longer times before settling. Reproductive success was much higher in re-used burrows than in new burrows, making them a possible incentive for long-distance movements.
