Rainfall predictive models for building simulation II –rainfall estimation

Rainfall simulation is key problem for the evaluation of the energy saving potential of passive cooling by natural rainfall evaporation in humid and hot areas. This paper used a set of monthly, daily and hourly rainfall data in Guangzhou Area to identify the statistic feathers and distribution of the monthly, daily and hourly rainfall. A one-order AR model was found to be suitable for monthly rainfall estimation by AIC criterion judgment. Using a Gamma distribution model and monthly rainfall, daily rainfall was estimated by stochastically assignment. Using a Beta distribution model and daily rainfall, hourly rainfall was estimated by stochastically assignment. A comparison of statistics features of simulated results to that of measured results validates the estimation model.