Climate change, population size, and time since observation create considerable uncertainty regarding the current status of wild plant populations recorded in various rare species databases. Several agencies and organizations (e.g., US Bureau of Land Management, Fish and Wildlife Service, Forest Service, NatureServe, state Natural Heritage Programs) maintain databases of rare plant occurrences and many of these occurrences may not have been visited recently. Therefore, the number of populations in the wild could be smaller than the number listed in databases due to extinctions that have not yet been detected. We used information on repeated surveys for a rare but widespread orchid, Cypripedium fasciculatum (clustered ladies slipper), to test if extinction probability is affected by elevation, population size, and time since observation. We applied the resulting model to populations in southwestern Oregon in the Geographic Biotic Observations data base maintained by the US Bureau of Land Management and the US Forest Service Natural Resource Information System to estimate the number of populations that are still extant.
Results/Conclusions
Of the 236 populations in our data sets of site resurveys, we found that 61% declined in size and 34% fell to zero. Elevation, time between surveys, and population size were important factors in predicting extinction probability. Probability of extinction was best explained by all of these factors, including a significant (p=<0.001) interaction between elevation and years between surveys. A general linear model suggested that populations at lower elevations were more likely to go extinct than high elevation populations, but only as the length of time between surveys increased. Small populations had a greater probability of extinction than those with large populations, and extinction probability was near zero for populations with >100 individuals, regardless of the length of time between samples. Further, extinction probability increased as the time between surveys increased, most notably for smaller populations. We estimated that of the 963 populations reported in GeoBOB and NRIS-Terra databases for southwestern Oregon as of 2015, a median of only 250 (95% quantiles: 172-350) were likely extant. This study demonstrates the need for additional and more frequent surveys of rare plant populations to improve the reliability of information in databases used by land management agencies. Land managers who make decisions on how best to conserve rare species often base their decisions in part on the abundance and distribution of those organisms as reported in databases. However, many reported populations may no longer be present.