Multiple Dimensions Define Thresholds for Population Resilience of the Eastern Oyster, Crassostrea virginica

Abstract

A species' distribution depends on its tolerance to environmental conditions. These conditions are defined by a minimum, maximum, and optimal ranges of single and combined factors. Forays into environmental conditions outside the minimum or maximum tolerance of a species (i.e., thresholds) are predicted to have large effects on a species' population and may help predict population resilience in the face of changing conditions. Here, we explore ecological thresholds for an important fisheries species and ecosystem engineer, Crassostrea virginica (eastern oyster). In coastal Louisiana, extreme freshwater inputs from rivers and precipitation events impact estuarine salinity, which is a key driver of oyster population dynamics. Using daily salinity and monthly oyster abundance monitoring data across Louisiana estuaries, we explore low salinity exposure threshold levels for oysters. Two statistical approaches were applied, with each model highlighting a different operational definition of a threshold: random forest models identified a threshold as an abrupt change in the oyster abundance- salinity relationship, while Bayesian models identified an increased probability of oyster abundance dropping below a critical threshold, defined here as less than 50% of the 5-year mean. All model results indicate oysters in coastal Louisiana experience low salinity exposure thresholds, defined as the number of consecutive summer days of salinity levels less than 5. However, actual number of days and salinity threshold differed by statistical approach, oyster life stage, and estuary highlighting the multiple dimensions defining ecological thresholds. While thresholds are considered important benchmarks to inform management and assess population or ecosystem vulnerability, our results reveal the need to carefully relate threshold definition to management goals and to acknowledge that thresholds may be highly context dependent.