Not all plant species mixtures are created equal on green roofs: Additive partitioning reveals complementary groups of species and growth forms

Background/Question/Methods

Green roof ecosystem processes such as biomass production, evapotranspiration, and nutrient uptake influence useful services such as urban temperature reductions and stormwater runoff retention.   Here we report on a four-year biodiversity-ecosystem functioning study involving 13 species from five plant growth forms in shallow-substrate green roof modules.   Plant species used were divided into five life-form groups: succulents, grasses, tall forms, creeping forbs and dwarf shrubs.  Monocultures of each species and life-form group were compared with life-form mixtures and growth-medium-only controls.

Results/Conclusions

All major services besides stormwater retention were correlated with aboveground biomass, and some mixtures of plant species species or growth forms outperformed the best monocultures.  Stormwater retention also showed significant positive effects of plant diversity (overyielding) compared with the best monocultures, but was not correlated with biomass, likely due to conservative water use by the high-biomass succulent growth form.  High biomass in mixtures can result from dominance by a single species or growth form (selection effects), or better average performance across all species in mixtures (complementarity), which can result from synergistic effects of species interactions or resource partitioning.  Additive partitioning allows quantification of these different effects, but the relative strength of selection vs. complementarity effects in this experiment depended greatly on the particular mixtures of species and growth forms tested.  Higher biomass yields than expected in the most diverse mixtures (all five growth forms) resulted from complementarity and selection operating in roughly equal magnitudes.  These phenomena are important in improving the performance of constructed ecosystems as they suggest different mechanisms by which ecosystem services can be optimized.