Species diversity is expected to promote stability in ecosystem functions such as productivity. One mechanism that may contribute to stability is compensatory dynamics. Examining species covariances over time or space can reveal whether population dynamics are compensatory. Negative species covariances, indicating compensatory dynamics, occur when an increase in density (biomass) of one species is associated with decreased density of another. These may be driven by competitive interactions or negatively correlated species responses to environmental drivers. Positive species covariances can be caused by facilitative interactions or positively correlated species responses to environmental drivers. Despite the widespread assumption that species abundances covary negatively and compensate to maintain stability in ecosystem function, several recent papers documented generally positive species covariances. Moreover, few studies have addressed how environmental factors influence compensatory dynamics. Using 18 years of data from different experimental treatments in old-field plant communities at the Kellogg Biological Station’s Long Term Ecological Research site, we investigated compensatory dynamics in productivity. We also investigated how two forces believed to be important to community structure (disturbance and nutrient availability) influence species covariances. We compared species covariances within and among groups of functionally similar species.

Results/Conclusions

We found that both spatial and temporal species covariances tended to be negative, unlike previous studies. Furthermore, we found more negative species covariances in annually tilled than untilled treatments and more negative covariances in fertilized than unfertilized treatments. These patterns suggest that tillage and fertilization increased compensatory dynamics, possibly by increasing the strength of competitive interactions or causing species to trade off more strongly in their tolerance of environmental conditions. In tilled fertilized plots, species covariances became less negative over time, suggesting that as communities assembled, the strength of competition decreased. However, in untilled fertilized plots, species covariances became more negative over time, suggesting that the strength of competition increased over the course of community assembly. Species covariances tended to be more negative among plant functional groups than within, suggesting that functional groups responded differently to environmental drivers. Our results indicate that in some communities, species covary negatively, allowing compensatory dynamics to stabilize ecosystem function. In old-fields, these compensatory dynamics operated among functional groups. In successional communities, time for community assembly, disturbance, and nutrient availability can affect patterns of species covariances and explain variation in compensatory dynamics observed in different communities.