Coastal flooding from landfalling tropical cyclones (TCs) is a major hazard with increasing severity in a warming climate and rising seas. It is difficult to predict because of highly complex compound effects of TC induced heavy rainfall, storm surge, and river discharge. This can be further exacerbated by sequential TCs such as Hurricane Irene and Tropical Storm Lee in late August to mid-September 2011, which caused major coastal flooding in the Mid-Atlantic region. This study focuses on better understanding and improving prediction of the compound effects of rain, storm surge, and river discharge using a high-resolution coupled atmosphere-wave-ocean model, namely the Unified Wave INterface–Coupled Model (UWIN-CM) and observations from NDBC buoys, NOAA tide gauges, and USGS estuary sites. UWIN-CM effectively captured the storm track and intensity, surface winds, surface waves, and ocean surface evolution associated with the two storms, compared with the observations. Compound effects of wind, rain, storm surge, and river-stream discharge on coastal flooding are investigated. The storm surge from Hurricane Irene was observed along the coasts Maryland, New Jersey, and New York, Delaware Bay, the lower reaches of the Delaware River, and in lower Chesapeake Bay. Strong onshore wind pushes water upstream, which has the highest compound effects on coastal flooding. Heavy rain and river-stream discharge into the coastal zone contributes mainly to locations upstream away from the open bay water. A new, indirect machine learning method of estimating the spatial extent of coastal flooding using simulated coastal sea surface height is shown.