Context. Feral pigs (Sus scrofa) have a wide global distribution that includes large parts of Australia and New Zealand. There is concern about the impacts of feral pigs on above- and below-ground flora and fauna, but little is known about their habitat use and feeding activity in temperate rainforests.

Aims. We evaluated the importance of abiotic and biotic variables hypothesised to influence seasonal and annual feral pig habitat use and feeding activity in a montane conifer–angiosperm rainforest in Te Urewera, North Island, New Zealand.

Methods. We used a grid of 25 remote-camera locations to collect feral pig images in a 100-ha area during the winters and summers of 2010 and 2011. Plant composition, solar radiation and soil fertility variables were determined for each camera-image area. Multiseason, multistate occupancy models and information-theoretic methods were used to evaluate how these variables related to feral pig occupancy and feeding.

Key results. Feral pigs occupied more camera locations in summer than in winter, and detection probabilities increased if piglets were present and with increasing soil phosphorus (P). Piglets were detected only in summer, and their detection probability increased with increasing soil P. The probability of detecting feral pigs feeding also increased with soil P and was higher in 2010 than 2011.

Conclusions. Feral pigs selected locations with high soil P, probably because those sites had more food than did locations with low soil P. Mast fruiting of tawa (Beilschmiedia tawa) has been hypothesised to increase feral pig recruitment, and the higher detection probability of piglets in summer 2010 followed a heavier tawa fruit fall.

Implications. Our study highlighted the usefulness of camera traps and occupancy models for understanding seasonal and annual dynamics of cryptic ungulate species in remote, rugged forests, and suggests that the impacts of feral pigs will be greatest in areas of high soil P following widespread tawa masting.