The proportion of area occupied (PAO) by an amphibian species has been adopted by the Amphibian Research and Monitoring Initiative (ARMI) as the metric by which many amphibian populations will be measured. The PAO statistical approach, developed by Drs. James D. Nichols, Andy Royle, and Darryl MacKenzie, evaluates the fraction of the landscape that is occupied by a species of interest, but not the actual abundance of the population across the landscape. It is therefore a monitoring technique that is less costly than methods that attempt to estimate population sizes, and ideal for large-area monitoring programs that seek to identify areas where amphibian species may be in decline. Once identified, these areas can then be monitored more closely, and appropriate action taken to halt the decline.

One challenge to implementing measurement of PAO is that many amphibian species are cryptic and may be undetected even when present at a site, which will lead to serious underestimates of PAO. Given that this probability of detecting the species will vary with environmental conditions, any long-term estimate of the change in PAO is likely to be biased unless detection probability has been accounted for. A method was developed by MacKenzie et al. (2002) that enables PAO to be estimated for a single season. By surveying the monitoring sites for the species more than once each season, a statistical model can be constructed enabling PAO to be estimated while allowing for detection probability. This technique has been extended for application to monitoring data collected over multiple seasons (MacKenzie et al., 2003). This new statistical model permits the direct estimation of rate of change in PAO, a statistic closely associated with the concept of 'trend." In addition, the model permits estimation of seasonal colonization and local extinction probabilities, and thus facilitates the mechanistic modeling of factors influencing population change. The ability to estimate these important parameters of population dynamics, while allowing for detection probability, permits ARMI investigators to accurately monitor amphibian populations.

This research is at the cutting-edge of ecological statistics and promises to be a major methodological step forward in applied ecology. The general approach has the potential to be a very useful monitoring and management tool, although further research is still required with respect to optimizing practical sampling schemes for use in the field. Dr. MacKenzie has been developing Windows-based software that implements these methods. A beta version (PRESENCE) that incorporates the initial method described in MacKenzie et al. (2002) is currently available for download via the Internet URL: http://www.proteus.co.nz or http://www.mbr-pwrc.usgs.gov/software.

Literature Cited:

MacKenzie, D.I., Nichols, J.D., Hines, J.E., Knutson, M.G., and Franklin, A.D., 2003, Estimating site occupancy, colonization and local extinction probabilities when a species is detected imperfectly: Ecology, v. 84, no. 8, p. 2200-2207.

MacKenzie, D.I., Nichols, J.D., Lachman, G.B., Droege, S., Royle, J.A., and Langtimm, C.A., 2002, Estimating site occupancy rates when detection probabilities are less than one: Ecology, v. 83, no. 8, p. 2248-2255.

For further reading, search ARMI publications for a list of articles associated with the keyword “occupancy.”