Habitat loss and acidification cause a pattern of selective extinctions in a model system

Background/Question/Methods

Environmental and anthropogenic disturbances can significantly alter the diversity and structure of ecological communities, causing local and sometimes regional extinctions.  Whether local extirpations result in regional extinctions depends partially on how local extirpations are structured among taxa with respect to their commonness or rarity across a landscape.  Thus, for a disturbance with a given magnitude of local extirpation, no, few, or many regional extinctions may occur, depending on how these extirpations are patterned among species and locations.  Specifically, if a disturbance causes a random pattern of local extirpations it will cause relatively few regional extinctions, whereas if a disturbance causes a selective pattern of local extirpations (i.e., the same species are lost across sites) it will cause a relatively large number of regional extinctions.  Here, we tested these predictions in a model system, laboratory protist communities.  We applied two distinct disturbances to experimental landscapes of local communities: one random (habitat reduction) and one selective (acidification).  We predicted that, when compared to a null model of random extirpation, the effects of habitat reduction on biodiversity would be indistinguishable from a random extinction process, whereas acidification would result in significantly greater rates of regional extinction than predicted by a random extinction process.

Results/Conclusions

When compared to pre-disturbance species richness, average local species richness in a single flask and average regional richness across 10 flasks decreased for both disturbance types, habitat reduction and acidification.  For habitat reduction, we predicted an 11.7% regional extinction rate based on random extinction (from 12 to 10.6 species across 10 flasks), whereas we actually observed a significantly higher extinction rate of 26.7% of species (from 12 to 8.8 species on average).  Similarly, for the acidification disturbance, the observed regional extinction rate of 34.0% (from 10.6 to 7 species) was significantly greater than the rate predicted under a random extinction scenario (17.9%, from 10.6 to 8.7 species).  Both of these results suggest that local extirpations were occurring selectively, leading to more extinctions than would be expected based on non-selective, or random, extinctions.  While it is not surprising that acidification was a selective disturbance that disproportionately affected some species over others, it would not necessarily be expected that habitat loss per se would also be selective, leading to a disproportionately high extinction rate.  Although this research was conducted in a model system, we suggest that understanding whether extinctions occur randomly or selectively is critical to understanding biodiversity loss in general.