K.B. Gregg
M. Kery
2003
<p><span>1. Most plant </span>demographic<span> studies follow marked individuals </span>in<span> permanent plots. Plots tend to be small, so </span>detectability<span> is assumed to be one for every individual. However, </span>detectability<span> could be affected by factors such as plant traits, time, space, observer, previous detection, biotic interactions, and especially by </span>life<span>-</span>state<span>. 2. We used </span>a<span> double-observer survey and closed population capture-recapture modelling to estimate </span>state<span>-specific </span>detectability<span> of the </span>orchid<span> </span>Cleistes<span> </span>bifaria<span> </span>in<span> </span>a<span> long-term </span>study<span> plot of 41.2 m</span><sup>2</sup><span>. Based on AIC</span><sub>c</sub><span> model selection, </span>detectability<span> was different for each </span>life<span>-</span>state<span> and for tagged vs. previously untagged plants. There were no differences </span>in<span> </span>detectability<span> between the two observers. 3. </span>Detectability<span> estimates (SE) for one-leaf vegetative, two-leaf vegetative, and flowering/fruiting states correlated with mean size of these states and were 0.76 (0.05), 0.92 (0.06), and 1 (0.00), respectively, for previously tagged plants, and 0.84 (0.08), 0.75 (0.22), and 0 (0.00), respectively, for previously untagged plants. (We had insufficient data to obtain </span>a<span> satisfactory estimate of previously untagged flowering plants). 4. Our estimates are for </span>a<span> medium-sized plant </span>in<span> </span>a<span> small and intensively surveyed plot. It is possible that </span>detectability<span> is even lower for larger plots and smaller plants or smaller </span>life<span>-states (e.g. seedlings) and that detectabilities < I are widespread </span>in<span> plant </span>demographic<span> studies. 5. </span>State<span>-dependent detectabilities are especially worrying since they will lead to </span>a<span> size- or </span>state<span>-biased sample from the </span>study<span> plot. Failure to incorporate </span>detectability<span> into </span>demographic<span> estimation methods introduces </span>a<span> bias into most estimates of population parameters such as fecundity, recruitment, mortality, and transition rates between </span>life<span>-states. We illustrate this by </span>a<span> simple example using </span>a<span> matrix model, where </span>a<span> hypothetical population was stable but, due to imperfect detection, wrongly projected to be declining at </span>a<span> rate of 8% per year. 6. Almost all plant </span>demographic<span> studies are based on models for discrete states. </span>State<span> and size are important predictors both for </span>demographic<span> rates and </span>detectability<span>. We suggest that even </span>in<span> studies based on small plots, </span>state<span>- or size-specific </span>detectability<span> should be estimated at least at some point to avoid biased inference about the dynamics of the population sampled.</span></p>
application/pdf
10.1046/j.1365-2745.2003.00759.x
en
British Ecological Society
Effects of life-state on detectability in a demographic study of the terrestrial orchid Cleistes bifaria
article