Hydrologic performance quantification of green roofs using an analytical stochastic approach based on kernel distribution estimation: Extensive case studies in Shandong Province, northern China

Abstract

Study region: Nine representative cities in Shandong Province, northern China Study focus: The analytical stochastic approach is computationally-efficient in green roofs (GRs)’ hydrologic performances. However, use of rainfall event separation methods and the resulting rainfall statistic affect the accuracy of this approach. This study integrates a Kernel distribution estimation (KDE)-based rainfall event separation method with the analytical stochastic model (ASM) developed for GRs. The proposed approach was tested for 198 design cases of GRs with considering different soil types and depths at 9 cities in the Shandong Province, northern China. New hydrologic insights for the region: Poisson and Kolmogorov–Smirnov tests can be performed to obtain the available pairs of minimum interevent time and threshold rainfall event depth. The optimal pairs of MIET-vt can be further determined based on the standardized procedure by the KDE-based rainfall event separation and characterization approach. Exponential distributions fit well the observed frequency distributions of rainfall event characteristics of the study area. ASM results using rainfall statistics obtained from a KDE-based method agree very well with those obtained from continuous simulations. The proposed integration of KDE-based rainfall statistics and ASM is accurate and useful for the planning, design and assessment of green roofs in regions of northern China.

Keywords

• Rainfall event separation
• Green roof
• Kernel density estimation
• Stormwater management
• Analytical stochastic model