Objective The research aims to fully depict the dynamic processes of rainfall interception by tree crown and the effects of crown and rainfall traits on rainfall interception process, and to develop the dynamic interception process simulation model on such basis.

Method A process-based experiment was conducted under five simulated rainfall intensities (10, 20, 50, 100, 150 mm/hour) to directly quantify tree crown interception and analyze the effects of rainfall traits and crown structure characteristics on interception.

Result (1) The interception process was composed of three phases, a rapid increase phase, a relatively-stable phase, and a post-rainfall drainage phase, in which 40% (±16%) of maximum interception storage (C_max) drained off to reach the minimum interception storage (C_min). (2)Interception is a threshold process, cumulative interception remained relatively stable when cumulative precipitation (P_c) surpassed 12-13 mm on a rainfall-event basis. (3) Both leaf traits (i.e., leaf area, leaf area index, leaf biomass) and branch traits (i.e., branch surface area, branch count, branch length, and woody biomass) significantly affected C_max and C_min. The LAI, as an easily-measurable parameter, affects the C_max and C_min following a power function.(4) A cumulative interception during rainfall (CIDR) model had been developed based on P_c and LAI: the model worked well and can simulate and predict the cumulative interception amount during rainfall.

Conclusion Crown interception is a dynamic process with three stages, which can be simulated by a model incorporated cumulative precipitation and LAI as major variables. The model is of great importance to quantify hydrologic processes and water balance in forest ecosystems.