Estimating rates of population change for a neotropical parrot with ratio, mark-recapture and matrix methods

Robust methods for estimating rates of population change ( k ) are necessary for applied and theoretical goals in conservation and evolutionary biology. Traditionally, k has been calculated from either ratios of population counts (obser ved k or k obs), or population models based on projection matrices (asymptotic k or k asy ,). New markrecapture methods permit calculation of k from markresighting information alone (realized k or k rea), but empirical comparisons with other methods are rare. In this paper, rates of population change were calculated for a population of green-rumped parrotlets (Forpus passer inus) that have been studied for more than a decade in central Venezuela. First, a ratio method based on counts of detected birds was used to calculate k à obs . Next, a temporal symmetry method based on markrecapture data (i.e. the k -parameterization introduced by Pradel, 1996) was used to calculate k à rea . Finally, a stage-structured matrix model based on state-speci® c estimates of fecundity, immigration, local survival, and transition rates was used to calculate k à asy . Analyses were conducted separately for females and males. Overall values of k à from the three methods were consistent and all indicated that the ® nite rate of population change was not signi® cantly di þ erent from 1. Annual values of k à from the three methods were also in general agreement for a majority of years. However, k à rea from the temporal symmetry method had the g reatest precision, and apparently better accuracy than k à asy . Unrealistic annual values of k à asy could have been due to poor estimates of the transitional probability of becoming a breeder (cà ) or to a mismatch between the actual and the asymptotic stable stage distribution. In this study, the trade-o þ between biological realism and accuracy was better met by the temporal symmetry than the matrix Correspondence: B. K. Sandercock, Division of Biology, Kansas State University, Manhattan, Kansas 66506-4901, USA. E-mail: [email protected] ISSN 0266-476 3 print; 1360-053 2 online/02/010589-19 © 2002 Taylor & Francis Ltd DOI: 10.1080 /0266476