Comparison of size vs. life-state classification in demographic models for the terrestrial orchid Cleistes bifaria

Abstract

Recent studies using capture–recapture modelling in terrestrial orchids have shown relationships among survival, life-state, climate and dormancy, but to date, there has been no rigorous study of the effects of size on demographic rates. Here we use multistate capture–recapture models to investigate the relationships between survival, ramet size, life-state, and propensity for extended dormancy of Cleistes bifaria in two populations in Florida and West Virginia. Using estimated leaf area as a measure of ramet size, we investigated relative fit to the data of models in which ramets were classified by life-state alone (dormant, vegetative, and flowering), size alone (dormant, small, and large), or by a combination of size and state (dormant, small vegetative, large vegetative, small flowering, and large flowering ramets). This scheme allowed us to compare the effects of a size versus state classification on survival and transition probabilities, including the propensity for dormancy. For survival rates, the size-classified model was preferable in terms of parsimony (AIC) to models classified either by state or by a combination of size and state. In C. bifaria, annual survival rate for large ramets was close to 1.0 and for small and dormant ramets, 0.91 and 0.76, respectively. This highlights the importance of protecting large plants. In contrast, a model combining life-state with ramet size was most parsimonious for explaining variation in transition rates. As the ‘size + state’ classification is convenient, holds more information, and seems more closely tied to fitness than classification by either size or life-state alone, it may be particularly useful in planning and assessing conservation management of orchids.