Changes in the geometry of ocean meridional overturning circulation (MOC) are crucial in controlling past changes of climate and the carbon inventory of the atmosphere. However, the accurate timing and global correlation of short-term glacial-to-deglacial changes of MOC in different ocean basins still present a major challenge. The fine structure of jumps and plateaus in atmospheric and planktic radiocarbon (14C) concentration reflects changes in atmospheric 14C production, ocean–atmosphere 14C exchange, and ocean mixing. Plateau boundaries in the atmospheric 14C record of Lake Suigetsu, now tied to Hulu Cave U∕Th model ages instead of optical varve counts, provide a stratigraphic “rung ladder” of up to 30 age tie points from 29 to 10 cal ka for accurate dating of planktic oceanic 14C records. The age differences between contemporary planktic and atmospheric 14C plateaus record the global distribution of 14C reservoir ages for surface waters of the Last Glacial Maximum (LGM) and deglacial Heinrich Stadial 1 (HS-1), as documented in 19 and 20 planktic 14C records, respectively. Elevated and variable reservoir ages mark both upwelling regions and high-latitude sites covered by sea ice and/or meltwater. 14C ventilation ages of LGM deep waters reveal opposed geometries of Atlantic and Pacific MOC. Like today, Atlantic deep-water formation went along with an estuarine inflow of old abyssal waters from the Southern Ocean up to the northern North Pacific and an outflow of upper deep waters. During early HS-1, 14C ventilation ages suggest a reversed MOC and ∼1500-year flushing of the deep North Pacific up to the South China Sea, when estuarine circulation geometry marked the North Atlantic, gradually starting near 19 ka. High 14C ventilation ages of LGM deep waters reflect a major drawdown of carbon from the atmosphere. The subsequent major deglacial age drop reflects changes in MOC accompanied by massive carbon releases to the atmosphere as recorded in Antarctic ice cores. These new features of MOC and the carbon cycle provide detailed evidence in space and time to test and refine ocean models that, in part because of insufficient spatial model resolution and reference data, still poorly reproduce our data sets.