This study investigated how shifting rainfall regimes will influence species and trait composition of chaparral communities in San Diego County. Precipitation in the region is projected to decline in the coming decades, but with increasing interannual variability. Chaparral communities regenerating after fire or other disturbance may be particularly sensitive to this variation. Drought, acting as a habitat filter, could restrict community trait composition both through shifting species composition and plastic trait responses within species. To test this hypothesis we employed shelters and irrigation to achieve five levels of rainfall: 0%, 50%, 100%, 150% or 200% of ambient rainfall. Each treatment had 5 replicate plots, each 3.5 x 3.5 m. Three seedlings of 8 focal species (4 each evergreen and deciduous) were planted in each plot, and monitored for survivorship. Functional traits were measured on these planted seedlings, as well as mature native shrubs and exotic annual species that were abundant in the plots. There is a growing interest in the potential for trait-based approaches to predict community and ecosystem responses to global change. Here, we evaluated whether community-level trait composition in response to shifting precipitation was similarly influenced by variable survivorship among species, and trait responses within species.
Results/Conclusions
Specific leaf area (SLA) decreased under drought conditions (p=0.015), but with variation among species (species x soil moisture interaction p <0.001) that was not consistent within functional groups. Water-use efficiency (WUE) varied among species but did respond to altered rainfall, potentially reflecting the fundamentally conservative strategies of plant species in semi-arid ecosystems. Survivorship of planted seedlings varied both among species and rainfall treatments (species x soil moisture interaction p=0.005), but did not differ consistently between deciduous and evergreen species. Survivorship was low under drought conditions and peaked for some species high rainfall, but peaked for others at ambient rainfall and declined under high rainfall conditions. We originally hypothesized that thick-leaved evergreen species with high WUE would be the most drought tolerant, but we found no correlation between species mean SLA or WUE and survivorship under drought conditions. In this case we were unable to predict the response of community-level trait composition in response to shifting rainfall regimes, because there was not alignment between the trait responses within species and the patterns of survivorship among species with different traits. These results highlight how individualistic species responses challenge trait-based approaches seeking to predict how communities will respond to global change.