Survey coverage, recording duration, and community composition affect observed species richness in passive acoustic surveys