Droughts that achieve extreme spatial extent over the contiguous United States pose unique challenges because of their potential to strain multiple water resources simultaneously. Two such spatially coherent, reoccurring droughts are (i) those that span the majority of the US (herein pan-CONUS droughts) and (ii) those that span the US Pacific coast (herein pan-coastal droughts). These droughts can have drastic impacts on US agriculture, water resources, and wildfire risk, thus posing serious risks to our food security, infrastructure, and economy. Such events are difficult to characterize due to the relatively short instrumental record and the rarity of observed widespread drought. The combined availability of observations, ensembles of climate model simulations, and high-resolution paleoclimate reconstructions, however, have recently increased the sampling and length of the hydroclimate record. This wealth of climate data makes the time ripe to investigate the causes and dynamics of spatially widespread droughts, with implications for their impacts in the future under a changing climate. Previous studies have established the sensitivity of North American drought variability to large-scale atmosphere-ocean modes. In particular, the El Niño Southern Oscillation (ENSO) and Atlantic Multidecadal Oscillation (AMO) have been linked with widespread drying over the United States. While neither mode alone is likely to cause either pan-CONUS or pan-coastal droughts, the canonical understanding of oceanic influences on North American hydroclimate nevertheless suggest that (i) pan-CONUS droughts are forced by a contemporaneous cold tropical Pacific Ocean and a warm tropical Atlantic Ocean and (ii) pan-coastal droughts are forced by cold tropical and north Pacific conditions. By examining how pan-CONUS and pan-coastal droughts are represented in climate model simulations and comparing them against observation and paleoclimate reconstructions, the work in this dissertation tests the above-mentioned canonical understanding. For pan-CONUS droughts, SST forcing is shown to originate almost entirely from La Niña conditions, with little contribution from the tropical Atlantic. Furthermore, internal atmospheric variability influences pan-CONUS drought occurrence by as much or more than ocean forcing and can alone cause pan-CONUS droughts. Internal atmospheric variability is shown to play an even larger, predominant role in driving pan-coastal droughts, accounting for upwards of 80% of the severity of the events; cold Pacific conditions, while playing a clearly detectable role, are only secondary in their influence relative to internal atmospheric variability. These results are then compared to the observational and/or paleoclimate record, which supports the model-inferred conclusions. Collectively, the work outlined in this dissertation holds important implications regarding (i) mechanistic understandings of North American hydroclimate, (ii) predictability, or lack thereof, of pan-CONUS and pan-coastal droughts, and (iii) how pan-CONUS and pan-costal droughts may change in the future due to increases in greenhouse gas emissions. The research contained herein furthermore demonstrates the precise extent to which large-scale ocean-atmosphere and internal atmospheric variability interact. In so doing, this dissertation contributes to a fundamental understanding of the importance of large-scale ocean-atmosphere modes relative to internal atmospheric variability in North American hydroclimate dynamics.