ESTIMATING DORMANCY AND SURVIVAL OF A RARE HERBACEOUS PERENNIAL USING MARK–RECAPTURE MODELS

Mark–recapture statistics have rarely been applied to plants, yet they can be useful for estimating plant demographic traits where individuals may be missed or unobservable. We applied mark–recapture statistics based on an information theory approach to estimate annual probabilities of dormancy in a five-year study of a threatened, perennial plant, the small yellow lady's slipper orchid (Cypripedium calceolus ssp. parviflorum (Salisb.) Fernald). The aboveground states of 548 genets in eight patches were monitored over five years. Apparent survival probabilities (φ), corrected for dormancy, were calculated. The best-fit model suggested that apparent survival was constant throughout the study for all patches, while dormancy varied additively with time among patches. The mean probability of dormancy was 0.320 ± 0.024, with a mean maximum overestimation of 0.067 as calculated using an estimate of the probability of detection. Dormancy typically lasted for no longer than two consecutive years, although dormancy as long as four years was observed. Dormancy displayed a strong covariate relationship with spring frost days, although effects of precipitation and mean spring temperature were almost equally strong. Mean apparent survival probabilities were high in each patch (φ = 0.878), but dormancy probabilities varied considerably among patches (d = 0.188–0.672). Conventional resprouting probabilities underestimated apparent survival by a mean difference of 0.288 (range: 0.150–0.589). This novel application of mark–recapture statistics to plant demography allowed robust survival estimates that accounted for uncertainty due to an unobservable, dormant life stage.