Storylines are physically based model recreations of impactful weather events, often chosen through iterative discussions between scientists and stakeholders, that are then simulated under plausible past and future climate scenarios. The New Year's flood event of 1997 represents the costliest in California's history and a well-monitored, decision relevant case study that can be used to stress test the fidelity of our cutting-edge modeling methods. The 1997 flood event featured an extreme atmospheric river that produced significant rainfall that fell on a ripened snowpack and saturated soils, leading to widespread inundation throughout northern California. This event has become a "design storm" for water resource managers and flood hazard engineers. We use the 1997 flood to investigate how well the Regionally Refined Mesh capabilities in the Energy Exascale Earth System Model (RRM-E3SM) recreates this decision-relevant compound extreme event. Specifically, we investigate RRM-E3SM sensitivities in horizontal grid spacing (14km to 3.5km) and forecast lead time (2-to-4 days prior flood onset). We find that planetary-to-synoptic scale circulation and vapor transport weakly depend on grid spacing over California, yet horizontal resolution does influence mesoscale circulations (e.g., Sierra barrier jet) that influence the hydrometeorological characteristics (e.g., storm total precipitation, storm duration changes in snowpack, and reservoir inflows). We employ both time series and causal analysis frameworks to decompose the contribution of the main hydrometeorological variables to runoff and reservoir inflows. We then utilize the best performing model setup to isolate the effects of IPCC-relevant past and future warming level impacts on the 1997 flood event drivers, such as changes in the storm characteristics, land-surface response, and flood potentials.