Attrition and the design of metacollections for ex situ plant conservation

Living plant collections increasingly contribute to ex situ conservation through the establishment of metacollections: coordinated networks of living collections that, by pooling resources, aim to represent and maintain intraspecific variation of species of conservation concern. Metacollections may be particularly important for the conservation of species that cannot be seed banked (i.e., exceptional species) and require substantial space or care in ex situ collections of whole growing plants. Here we asked: what is the minimum number of initial plants needed to achieve a conservation goal while accounting for attrition due to plant death in a metacollection? We defined conservation goals as the number of plants, sourced from different regions, that should survive after a given time period to represent intraspecific variation ex situ. The emphasis on minimizing the number of initial plants needed to achieve these goals follows from resource limitations common in ex situ collections. The approach to this question in the current literature does not account for the reliability with which conservation goals are achieved. To address this issue we developed stochastic optimization models that estimate the minimum number of initial plants required to achieve a conservation goal with a stated level of reliability. We illustrated the application of these models with hypothetical as well as real-life examples and uncovered management implications of intraspecific variation in attrition rates, including a trade-off between metacollection size (number of initial plants) and resilience to unpredictable events that might affect part of a metacollection.