PREDICTING BREEDING SUCCESS OF ATLANTIC POPULATION CANADA GEESE FROM METEOROLOGICAL VARIABLES

Management strategies for sustained harvest and long-term viability of Atlantic Population Canada geese (Branta canadensis) require evaluati ons of annual breeding success before establishing fall harvest regula tions. The only quantitative measure of the annual breeding success of this population is the proportion of young geese in the fall harvest that is not available when harvest regulations are set in late July. B ecause the majority of Atlantic Population Canada geese breed in the s ub-arctic regions of the Ungava Peninsula in northern Quebec, spring c limatic conditions are potential predictors of annual production for t his population. We used tail-fan data from the Maryland harvest to cal culate an index of the proportion of young geese (Y-i) in the fall pop ulation, 1963-94. We used 1963-87 weather data to develop multiple lin ear regression models to predict Y-i and validated these models by pre dicting Y-i for 1988-94. Models with the greatest predictive ability i ncluded the average daily mean temperature and the number of days of s nowfall in May and June. The final model included 6 parameters and acc ounted for 78.7% of the total variability in Y-i (P = 0.001). This ana lysis demonstrates the potential use of climatic data to predict an in dex of annual production derived from harvest age ratios. The usefulne ss of this technique will depend on periodic assessment of predictive models as more data is gathered, and evaluation of harvest tail-fan su rveys as indices to breeding success.