interactions terms in MARK's nest survival feature

[bmaslo21]

I am working on a study involving the effectiveness of predator exclosures in increasing nest success of Piping Plovers. My initial analysis indicated that both high mammalian predator pressure and high human disturbance were led to poor nest success. This additive model reported negative parameter estimates of the same magnitude for each of the 2 variables. The predator pressure variable had a CI straddling 0.

When I added a model containing the above variable and an interaction term (predator X disturbance), this became the top model, and the next highest delta AIC score was 4. However, the effect of predator pressure became positive. Disturbance was still negative, but its CI straddled 0. And the interaction term was negative with a somewhat large CI.

I am unsure how to interpret these results. How does MARK handle interaction terms within the nest survival feature? Is an additive model enough to justify an effect?