Use of point counts to estimate butterfly densities

Background/Question/Methods

Rigorous monitoring of rare species is an essential aspect of conservation efforts. Established monitoring methods for butterflies have been designed for open, accessible habitats such as grasslands. However, not all rare species occupy habitats that are easy to see through and navigate, therefore, a new approach to monitoring is necessary for some species. We present a novel use of distance sampling from point transects to monitor the Miami blue butterfly, a highly endangered butterfly that inhabits dense shrub habitat in the Key West National Wildlife Refuge.  Little is known about this butterfly population because of its remote location and logistical constraints involved in accessing the butterflies. To monitor this butterfly’s population size, we developed surveys consisting of two-minute point counts in semi-circular points along the shrub edge. We then used program DISTANCE to fit detection functions to our data and estimate Miami blue butterfly densities. Because survey duration influences density estimates, we examined the relationship between survey time and density to determine the most accurate survey duration.

Results/Conclusions

In 2012, we observed peak Miami blue butterfly densities in early April and again in late July/early August. Peak abundances appear to be closely tied to the phenology of the butterfly’s Pithecellobium host. During these periods, we have analyzed data from a two minute survey and a 30 second survey, the resulting density estimates are 1100 Miami blue butterflies/hectare and 600 butterflies/hectare respectively. We recommend using estimates from 30 second (nearly instantaneous) surveys to monitor Miami blue butterfly density as these are the most conservative density estimates at this time. One limitation of the method is the need for 50-60 detections to fit models to the data, a number that may be difficult to obtain for some rare species.  For butterfly species that exist in high enough densities, we recommend using point counts in habitats where traditional methods are impossible to implement.