Calibrating the Priestley-Taylor model for evapotranspiration across different substrate depths in green roofs.

This study aims to address a common issue in current research: the neglect of the calibrating model parameters when estimating evapotranspiration (ET) from green roofs (GRs) using the Priestley-Taylor model, with most studies limited to a single substrate depth (SD). To overcome this limitation, this research improves the accuracy of ET estimation for different SDs on GRs by calibrating the Priestley-Taylor coefficient α. The study period was 692 days in total, from 25 April 2021 to 26 April 2023. Daily ET data from the outdoor GR experimental group were used to calibrate and validate the model. Uncalibrated models perform well for medium SDs (150 mm) but decline for deeper (300 mm) or shallower (50 mm) ones. NSGA-II optimization significantly improved model performance across all SDs, notably at 300 mm. The research underscores the importance of parameter calibration for water management in GRs and sets a foundation for future research on optimizing water retention and regulation functions in GRs.