Evaluating the Accuracy of Different Monitoring Designs by Subsampling a 14-Year R. boylii Eggmass Census

Sarah J. Kupferberg

Quest Engineering Corp.

Point Richmond, CA

This video and the slide show below it work together.
Stop and advance the slide show by clicking its controls at the top of its window.

ABSTRACT

KUPFERBERG, SARAH J.1, and WENDY J. PALEN2
1Quest Engineering Corp., Point Richmond, CA, skupferberg@pacbell.net; 2Department of Integrative Biology, University of California, Berkeley, CA

Evaluating the Accuracy of Different Monitoring Designs by Subsampling a 14-Year R. boylii Eggmass Census.

Rana boylii eggmass surveys have been conducted annually on a 5 km reach of the South Fork Eel River (Mendocino Co., CA) since 1992. These detailed spatial and temporal data provide the opportunity to evaluate the consequences of choosing among a variety of sampling strategies for estimating 1) site occupancy, and 2) population trends through time. Such choices are often made by managers of regulated rivers. For the purposes of hydroelectric dam relicensing utilities and regulators are faced with the task of determining species distributions and site occupancy over very long reaches, (up to 50 river miles). Further, they must monitor the status of populations to detect long term trends and decide whether to invoke adaptive management of flow regime. We evaluated the effects of specific spatial and temporal sampling biases which reflect potential real world constraints. For example, over the life of a 30 year dam license, censusing of R. boylii may only occur periodically, e.g. once every five years, or selectively at populous breeding sites located in proximity to large tributary confluences. We find that the underlying spatial and temporal pattern in R. boylii breeding strongly influences the accuracy of monitoring, because nearly 50% of the population breeds within one riffle-pool sequence of perennial tributary confluences consistently across years. One of the consequences of this pattern is that if surveys and monitoring efforts solely focused on these sites, we would overestimate both the population size and population stability through time. In the absence of a priori knowledge about the longitudinal distribution of individuals (e.g. a full reach survey) it would be unlikely to accurately capture the spatial and temporal dynamics of R. boylii populations.