Minimum number alive (MNA) is a popular index of population size, but it is biased. It is always negatively biased, in that it usually underestimates population size. MNA also has a “tapering bias”, in that the negative bias is greater near the first and last capture sessions. We developed a simulation model to study the properties of MNA to determine if it was possible to obtain unbiased estimates of population size. We manipulated population size, population trends such as a declining or increasing population, and the number of capture sessions used in the MNA analysis.
For a given capture rate and number of capture sessions, the MNA curve had the same shape, which reflected a known percentage of the true population size at each sample, regardless of the population size. Thus, one can estimate population size using the expected proportions from the minimum number alive model. This modification provided an unbiased estimate of population sizes, overcoming the negative bias that is commonly attributed to MNA. While the estimate is unbiased, error increases with decreasing numbers of capture sessions and at the beginning and end of the capture sessions.
We used our model to assess population trends from a twenty one year record of captures of ornate box turtles from two different sites and found that that the population of turtles appears to be stable at one site, but declining at the other.
Nikita Martinson, ’15
Major: Psychology – Psychological Scientist Concentration
Jeremy Novak, ’17
S. Andrew McCollum
Sponsor: S. Andrew McCollum