Diurnal greenhouse gas emissions and substrate temperatures from blue-green roofs in north-eastern Italy during a dry-hot summer season

Abstract

Covering building rooftops with vegetation [Green roofs (GRs)] holds promise for lowering building temperatures, reducing stormwater runoff, and providing other ecosystem services, but it is unclear how this will impact greenhouse gas (GHG) emissions. The latter may also be influenced by vegetation type, substrate depth, and irrigation regime and we sought to test this by comparing daytime GHG emissions (i.e., CH4, CO2, and N2O) and daily substrate temperatures in 48 GR microcosms in North-eastern Italy during a dry-hot summer season (June to September). Four vegetation types (Sedum mixture, cold season grasses, warm season grasses, or wildflowers), two substrate depths (8 cm or 14 cm), and two irrigation levels (1 or 2 mm d−1) were evaluated, for a total of 16 treatments with 3 replicates. We found that vegetation type had a significant effect on temperature [average temp. of 24.8 °C (Sedum) vs 25.5 °C (warm season grasses)] and CH4, CO2, and N2O emissions. While all vegetation types had net CO2 emissions (median values from 147 to 671 mg m−2 h−1) and net N2O uptake (median values from -0.06 to -0.28 mg m−2 h−1), CH4 flux had net negative values (capture) only in microcosms with wildflowers (-0.07 mg m−2 h−1), whereas other treatments had a median CH4 emissions of 0.09 mg m−2 h−1. Substrate depth significantly affected CO2 and N2O fluxes with deeper substrate leading to higher CO2 emission (+ 60.7%) and greater N2O uptake (+ 30.8%). Irrigation level only significantly influenced N2O fluxes with 2 mm irrigation resulting in higher fluxes (-0.20 mg m−2 h−1) than 1 mm irrigation (-0.09 mg m−2 h−1). Our study suggests that under heat induced plant-stress conditions, GRs can improve N2O and CH4 capture but might increase the emissions of CO2 fixed by plants in the previous years in the substrate and that vegetation type and substrate depth can significantly alter emissions and are thus important design parameters.