Alternative Metrics of Green Roof Hydrologic Performance: Evapotranspiration and Peak Flow Reduction

Stormwater runoff presents an issue for many urban areas, triggering sewer overflows and water body pollution. Green roofs, engineered vegetative systems that replicate the stormwater absorption properties of natural landscapes, have become an attractive strategy for attenuating stormwater runoff. Historically, green roof hydrologic research has been focused on stormwater volume retention with less emphasis on evapotranspiration (ET) and stormwater detention. ET is associated with green roof environmental benefits, including stormwater runoff attenuation and urban heat island mitigation, and is an important parameter in hydrologic and energy models. Stormwater detention limits flow rate of stormwater into sewer systems, reducing the chance of sewer overflow. The aim of this research is to investigate green roof ET and stormwater detention behavior and develop methods to predict performance based on readily available environmental data. In order to study these hydrologic performance metrics, a series of four New York City green roofs were instrumented with sensors to measure rainfall, runoff, ET, and other environmental data. The green roofs span several extensive green roof installation types, specifically the vegetated mat, built-in-place, and modular tray systems. Environmental monitoring for this analysis began in January 2009 and concluded in October 2013. In the first study, a dynamic chamber method was developed to conduct high-resolution measurements of ET. Results show that monthly ET depths range from 2.2 to 153.6 mm. Chamber results were compared to two ET estimation methods, specifically the Penman-Monteith equation and an energy balance model. Dynamic chamber results were similar to Penman-Monteith estimates; however, the Penman-Monteith equation over-predicted bottommost ET fluxes during the winter, and under-predicted peak summer fluxes. In the second study, dynamic chamber measurements were used to investigate green roof behavior and the effectiveness of various predictive models, particularly in water-limited conditions. Comparison of Hargreaves, Priestley-Taylor, Penman, and Penman-Monteith equation results to chamber measurements reveals that the Priestley-Taylor equation best estimates ET. However, the Priestley-Taylor equation can still overestimate lower fluxes and underestimate high fluxes. Application of a storage model, antecedent precipitation index, and advection-aridity model indicates that the antecedent precipitation index best estimates ET in water-stressed conditions. In the third study, 501 rainfall events were used to characterize green roof stormwater detention behavior, through analysis of event peak rainfall rate reductions. Empirical models relating event peak runoff rate to rainfall depth and peak rainfall rate were developed. Roof-specific models allow for the comparison of peak reduction behavior among roofs, while a combined model allows for designers to estimate green roof event peak rainfall reduction performance. Model application shows that the modular tray system is most effective at reducing peak rainfall rate. Overall, this research provides valuable insight into green roof hydrologic performance. Analysis of environmental data reveals not only the ET and peak rainfall rate reduction performance of green roofs, but also the environmental factors that affect performance. Additionally, predictive models for ET and peak runoff rate investigated in this dissertation can be valuable tools for researchers, practitioners, and policymakers to estimate green roof hydrologic performance