Succession and the stability of communities in response to rainfall variation

Background/Question/Methods

For decades ecologists have thought that more mature communities, those which haven’t been disturbed for a long period of time, should be more stable than those which have been disturbed more recently. Conceptually this is very interesting, but we have remarkably little data that directly addresses this hypothesis. To address this issue, I used long-term vegetation data to test whether the stability of the community changes as succession proceeds.  The data came from the Buell Small Succession Study (BSS) a 53-year study of dynamics in abandoned agricultural land. In the BSS, vegetation changes in each of ten fields have been monitored in 48, 1 m² plots. I related year to year changes in total plant cover and plot species richness for each field to the amount of rainfall in a given year, the change in rainfall from one year to the next, and an index of drought severity. The data were split into four successional age classes, 1-10, 11-20, 21-30 and 31+ to determine the responsiveness of community structure to rainfall over time.  If stability is linked with succession, the relationship between rainfall variation and change in community structure should decrease in older fields.

Results/Conclusions

Neither measure of within-year weather, total rainfall or the Palmer Drought Severity Index, had any relationship with changes in richness or total plant cover. However, year to year variation in rainfall was associated with changes in community structure. Following theoretical predictions, there was a strong relationship between rainfall variation and community stability for early successional communities. As the change in rainfall became more positive (i.e. a greater increase) from year to year, cover and richness both increased. Similarly, the larger the decrease in rainfall from one year to the next, the larger the depression of richness and cover.  In mid successional communities (years 11-20 and 21-30) there was no linkage between richness or cover and variation in rainfall. This shows that the older communities, as expected, were more resistant to environmental variation. However, in the oldest successional stages (31+ years after abandonment) the linkage between environment and the plant community returned.  Change in rainfall was again positively correlated with the change in the number of species per plot. These data show that succession-stability relationships are not necessarily as predicted. Stability in response to environmental variation may decrease in late successional systems as the forest understory develops.