The simultaneous occurrence and persistence of multiple extremes can be associated with large‐scale surface temperature gradients, regional forcings, and intermediate atmospheric variables such as the potential vorticity (i.e., a measure of cyclone generation), regional blocking frequency, and frequency and intensity of persistent meteorological high and low pressure systems. This research will adopt a novel methodology for identifying the causality and predictability of hydrometeorological (i.e., the water cycle and the transfer of water and energy between the land surface and the lower atmosphere) extreme events. The hydrometeorological extreme events that are the primary focus in this context are extreme precipitation, and the resulting floods due to compounding rainfall events. The two research objectives include: (1) investigation of the physical causality of regional to continental scale hydrometeorological extreme events using Information Theory‐based causality tests in a Bayesian learning framework; and (2) development of stochastic simulation and predictive models using these drivers for the occurrence of extreme precipitation and floods of varying duration. For the first objective, statistical learning frameworks using causality tests, model checking, and hypothesis generation will be designed with both reverse causal inference (what causes an outcome) and forward causal inference (effects of the causes) to verify dependence of the above‐identified hydrometeorological extreme events on the earth system and atmospheric controls. The second objective will employ a stochastic modeling strategy that integrates the above factors and accounts for correlations across multiple events and multiple locations under dynamic earth system conditions at different time scales. Bayesian network models will be used with full uncertainty analysis along the causal chain.