Developing an integrated model relating substrate water content to indoor temperature reduction for irrigation-decision support of a green roof

This study analyzes the influences of evapotranspiration or substrate moisture variation on the indoor-temperature reduction of green roofs compared to the control group. A multiple linear regression (MLR) model for the operation stage based on observation and an integrated MLR model for the planning stage based on simulation are verified. The MLR model shows 0.64 °C of the Root Mean Square Error (RMSE) in predicting the hourly difference of temperature reduction based on the measured change in evapotranspiration and air temperature. The contributions of the hourly increment of air temperature (ΔTa) and increment of evapotranspiration (ΔET) are similar to the hourly increment of temperature reduction (ΔTdif). Then, the feasibility of the integrated MLR model is demonstrated based on the evapotranspiration and substrate moisture of a green roof simulated by a hydrological model as well as the indoor-temperature reduction simulated by a building energy model, which has fair performances in capturing the heat-transfer and water-balance physical process within a green roof. The integrated MLR model shows that evapotranspiration is relatively essential, followed by substrate moisture, air temperature, and vapor pressure. Despite the modeling bias, the integrated model quantitatively relates the influential factors to temperature reduction and predicts temperature reduction with an RMSE of 1.02 °C. The integrated model can quantify the influence of irrigation on temperature reduction under various climate conditions and green roof structures. This study demonstrates the procedure of establishing the integrated model. It shows the potential of the integrated model to provide decision support on irrigation for multi-purpose optimization of green roof performances.