Monday, August 4, 2008

PS 2-43: Impacts of global and regional climate on population dynamics of an endangered migratory bird

Karine C. Gil1, William E. Grant1, R. Douglas Slack1, and Enrique H. Weir2. (1) Texas A&M University, (2) The Platte River Whooping Crane Maintenance Trust, Inc.

Background/Question/Methods We analyzed long-term demographic and environmental data to understand the role of large scale climatic factors such as the Pacific Decadal Oscillations (PDO) and environmental factors in three regions of North America on natality and mortality of the migratory Whooping Crane (Grus americana) population. This is an endangered species that winters at Aransas National Wildlife Refuge (ANWR) in Texas, breeds at Wood Buffalo National Park (WBNP) in Canada and uses Nebraska (NE) as a main stopover. Demographic data were collected over 27 years, and the finite rate of population increase (lambda) was estimated from 1967 to 2004. Historical data  were obtained for the PDO index, temperature and precipitation (at WBNP, NE and ANWR), pond water depth at WBNP, freshwater inflow, and net evaporation at wintering grounds. A statistical regression analysis was developed between the mean value of the PDO and the coefficient of variation of the intrinsic rate of growth as another measure of lambda.

Results/Conclusions We obtained a good correlation between population growth and PDO effects with one year lag, and we found that almost all periods of population decline occurred during PDO cold phase.  We determined the mechanisms through path relationships. We used simple linear regression to explore relationships and identify variables for inclusion in path analysis. A combination of 6 environmental factors from different regions (PDO, pond water depth at WBNP, freshwater inflow and net evaporation at ANWR, and precipitation and temperatures in WBNP, Nebraska, and ANWR) was useful in explaining population dependent variables linked to lambda variability. Path analysis showed two pathways.  PDO has an effect on temperatures at ANWR and Nebraska, and these factors combined with temperature at WBNP and pond water depth were good predictors of variability in lambda, through the impact on total annual mortality.  Variability in lambda also was explained by effects of a) extreme maximum temperature, low temperature in December and precipitation at ANWR, on mortality at wintering grounds, b) temperature in Nebraska during fall migration, and precipitation all summer at WBNP, on mortality from April to November, and c) WBNP low pond water level, on chick mortality. In addition, we found that environmental factors on the wintering grounds, such as freshwater inflow and extreme minimum temperature, have an effect on subsequent reproduction. This link between winter and breeding grounds was measured through clutch size reduction and brood failure.  Direct and indirect effects of these environmental factors were discussed.