Open Collections

Risk assessment for conservation under ecological uncertainty : a case study with a stream-dwelling amphibian in managed forests

University of British Columbia

Quantifying risks of forest and habitat management options are often limited by uncertainties in habitat associations, life history and population trends for resident species. Using the tailed frog (Ascaphus truei), a headwater stream-dependent amphibian, I: 1) developed hierarchical models of habitat relationships across this species' range in British Columbia; 2) examined plausible life history responses to habitat change, and 3) investigated population persistence outcomes within and among streams to uncertain effects of forest harvesting and disturbance scenarios. To develop habitat association relationships I used classification and regression trees (CART) together with simple and partial Mantel tests. Variables describing biophysical setting at meso- and micro-scales had a greater influence on occurrence and abundance of larval frogs than did adjacent forest practices, possibly because most sampling sites were in disturbed watersheds. Underlying geology was a consistently important determinant of occurrence patterns, with fine-scale stream structure more important in determining abundance. Using life stage-based population matrix models, I explored consequences of different life history strategies and ranges of habitat productivity and environmental variation on population persistence. Persistence is decreasingly sensitive to changes in growth rates, tadpole and adult survival, and fecundity. Populations also appear more sensitive to changes in survivorship of in-stream stages (eggs, hatchlings and tadpoles) than riparian stages (juveniles/adults). Clinal, elevational, and local reductions in habitat productivity (e.g., shorter growing seasons, reduced light penetration in mid-seral forests) appear as dominant factors mediating how local fluctuations in demographic rates determine risks of loss of small populations, even if environmental variation is relatively low. Using a spatially explicit metapopulation-landscape dynamics model, I explored six plausible hypotheses linking habitat alterations to population dynamics in response to forest harvesting and disturbances. Risks to populations from disturbance depend strongly on assumptions about age at first reproduction. Incremental effects of forest harvesting on risks were small compared with those already incurred through stochastic events (floods, debris flows, climatic variation) or state of the landscape. Once extirpated, population recovery through dispersal appears unlikely. However, key uncertainties about the impacts of harvesting and forest succession on demographic rates render evaluation of alternative riparian protection systems difficult with present knowledge.

Thesis/Dissertation

application/pdf

2009-09-25

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/13164

Forestry

University of British Columbia

UBCV

DSpace