Fig. 6. Small, proportional changes in input parameters expected to decrease probability of extinction in a stage-based, discrete, stochastic population model for large-bodied woodpeckers. Each graph represents the set of mean changes from an analysis that focused on variation in particular drivers of population dynamics: (A) Variation in demographic rates: variances of annual adult survival and fecundity were held constant at intermediate values while either initial number of adults increased by 1, annual adult survival increased by 0.008, or fecundity increased by 0.039; (B) Environmental stochasticity: means for annual adult survival and fecundity were held constant at intermediate values while either lognormal variance in fecundity decreased by 0.0089, initial number of adults increased by 1, or the scaling parameter for the logitnormal variance in annual adult survival decreased by 0.0442, and (C) Allee effect: variance in annual adult survival and mean and variance in fecundity were held constant at intermediate values while either the slope of the relationship between annual population size and probability of breeding increased by 0.14, initial number of adults increased by 1, or annual adult survival increased by 0.008. Results displayed in each graph are based on 200 iterations across all possible combinations of input parameters. Whiskers represent 95% confidence intervals. The largest decreases in predicted extinction rate occurred when an Allee effect was present and an initial adult was added or annual adult survival rate was increased by a small, proportional amount.