Irrigation supports agricultural production, but widespread use of Irrigation can perturb the regional and global water cycle. Different assumptions have been used to represent irrigation water demand, water supply, supply sources, and water redistribution by the reservoirs. However, most studies ignored the surface water constraints at sub-daily time scales, which could lead to over-estimation of surface water usage, under-estimation of groundwater pumping, and alter their seasonal variability. In this study, we investigate the seasonal irrigation withdrawal at global scale using a water management model that dynamically constrains the irrigation surface water supply on a 3-hourly time scale. Extending the one-way coupled irrigation module in the Energy Exascale Earth System Model (E3SM) framework, this newly developed coupled model consists of three components: a land surface model (E3SM Land Model, ELM), in which irrigation water demand is estimated and applied and a river routing model coupled with a water management model (MOSART-WM) that simulate streamflow, reservoir operations, and irrigation water supply. Simulations are being performed globally for 30 years (1975-2004) at ~100 km grid spacing to estimate irrigation surface water and groundwater use in major global river basins and the seasonal characteristics. We will also conduct a series of simulations with different assumptions commonly used in large scale irrigation studies to evaluate their spatial and temporal impacts on estimations of global irrigation withdrawals from surface water and groundwater and understand their implications on groundwater sustainability.