Repository logo
 

Monitoring amphibian populations and the status of wood frogs and boreal chorus frogs in the Kawuneeche Valley of Rocky Mountain National Park

Date

2010

Authors

Scherer, Richard Dean, author
Noon, Barry R., advisor
Crooks, Kevin R., committee member
Dean, Denis J., committee member
Doherty, Paul F., 1970-, committee member
Muths, Erin, 1961-, committee member

Journal Title

Journal ISSN

Volume Title

Abstract

The selection of a state variable is an important component of any program for monitoring a wildlife population. Annual counts of egg masses have been promoted as an appropriate state variable for monitoring populations of some amphibian species. In Chapters 1 and 2, I evaluate different aspects of the use of counts of egg masses as a state variable for monitoring wood frogs and other pond-breeding amphibians. Specifically, I assessed patterns of detectability in wood frog egg masses and evaluated the statistical power of using counts of egg masses and linear regression to detect trends. In Chapter 1, I used closed capture-recapture models to estimate detection probability and evaluate potential sources of variation in the detectability of wood frog (Rana sylvatica) egg masses. Model selection results and model-averaged estimates provided evidence that detection probability varied between years which emphasize the importance of accounting for detection probability when using egg mass counts to infer population trends in amphibian populations. In Chapter 2, I used computer simulation to evaluate the statistical power of detecting trends in counts of egg masses. A previous study identified important benefits of using counts of egg masses as the state variable in a monitoring programs for wood frogs (e.g., low cost) but did not evaluate the statistical power of this approach. The results of the simulations indicate that a minimum of nine years of monitoring is necessary to achieve high statistical power (≥ 0.80). For populations experiencing low rates of annual decline (≤ 4% annually), two decades or more of monitoring data were required to achieve high levels of power. Unless populations are changing at high rates, counts of wood frog egg masses will be a poor state variable for monitoring programs interested in detecting trends or evaluating the effects of management action or other perturbations. Jeff Tracey designed the computer simulations for this chapter and co-authored the chapter. In Chapter 3, I estimated the proportion of wetlands that were occupied by breeding wood frogs (Rana sylvatica) and boreal chorus frogs (Pseudacris maculate), as well as extinction and colonization probabilities, from 2004 to 2006 in a valley in Rocky Mountain National Park. I also evaluated hypothesized relationships between occupancy and attributes measured at multiple spatial scales. Erin Muths and Barry Noon were co-authors on this chapter. Breeding wood frogs occupied approximately 12% of the wetlands across the three years of the study, while occupancy by boreal chorus frogs increased from approximately 20% in 2004 to 37% in 2006. Wood frog occupancy was positively associated with the amount of streamside habitat adjacent to a wetland and negatively associated with the cost-based distance to the nearest occupied wetland. Boreal chorus frog occupancy was positively associated with the number of neighboring, occupied wetlands. For wood frogs, model-averaged estimates of colonization probability were < 10% and estimates of extinction probabilities were > 30%. Model-averaged estimates of colonization for boreal chorus frogs, on the other hand, were higher (> 10%) than for wood frogs, while estimates of extinction probabilities were much lower (< 10%). I discuss the management implications of these results. Finally, in Chapter 4, I used eight microsatellite markers to assess the genetic population structure, estimate genetic variability and test for evidence of recent bottlenecks in a population of wood frogs (Rana sylvatica) in Rocky Mountain National Park, Colorado. The results from a genetic clustering algorithm indicate the population is partitioned into two genetic clusters, and an estimate of FST provides strong evidence of differentiation between the clusters. Though I found no evidence of recent population bottlenecks, genetic variability in the population was relatively low. Sara Oyler-McCance completed many of the genetic analyses in this chapter and was a co-author, as were Erin Muths and Barry Noon.

Description

Rights Access

Subject

Citation

Associated Publications