Reexamining the storage effect: Why temporal variation in abiotic factors seems unlikely to cause coexistence

The temporal storage effect-that species coexist by partitioning abiotic niches that vary in time-is thought to be an important explanation for how species coexist. However, empirical studies that measure multiple mechanisms often find the storage effect is weak. We believe this mismatch is because of a shortcoming of theoretical models used to study the storage effect: that while the storage effect is described as having just three requirements (partitioning of temporal variation, buffered population growth, and a covariance between environment and density-dependence), models used to study the storage effect make four assumptions, which are mathematically subtle but biologically important. In this paper, we examine those assumptions. First, models assume that environmental variation leads to a rapid impact on density-dependence. We find that delays in density-dependence (including delays caused by competition between cohorts) weaken the storage effect. Second, models assume that intraspecific competition is almost identical to interspecific competition. We find that unless resource or predator partitioning are virtually absent, then variation-independent mechanisms will overshadow the benefits of the storage effect. Third, models assume even though there is vast variation in the environment, species are equally adapted on average (i.e., zero fitness-differences). We show that fitness differences are particularly problematic in the storage effect because specializing on temporally rare niches is far less effective than specializing on other types of rare niches. Finally, models assume that stochastic extinctions can be ignored, and invader growth can determine coexistence. We show that storage effects tend to reduce mean persistence times, even if invader growth rates are positive. These results suggest that the assumptions needed for the storage effect are strict: if the first or second assumption is relaxed, it will greatly weaken the stabilizing mechanism; if the third or fourth assumption is relaxed, it will create a diversity-destroying effect that may undermine coexistence. We examine three real-world communities-annual plants, tropical forests, and iguanid lizards-and find that empirical studies suggest that all three communities violate multiple assumptions. This suggests that the temporal storage effect is probably not an important explanation for species diversity in most systems.