Tools for mitigating extra zeros

The primary tool for reducing the bias associated with false negatives is to shift the design of a survey so that sample points are surveyed more than once. Even without more complex statistics, this approach reduces bias by providing more opportunities to detect the species at a given sample point. This does induce a severe tradeoff: visiting the same points more than once reduces the number of points that can be sampled if the amount of effort is held constant. It is possible to identify an optimal level for this tradeoff that depends on the goals of the study (Field, Tyre, and Possingham 2005, MacKenzie and Royle (2005)).

In the context of mapping a disease, if the sample point is represented by a population of hosts, then sampling > 1 host per point has the same effect as repeat visits. The same tradeoff also exists – the more individuals sampled at one site the fewer different sites can be sampled with a fixed amount of effort.

how to estimate false negatives? repeat observations?

Dynamic detection error models –

Joint occurrence / co-occurence SDM