Reconstruction of atmospheric and surface ocean 14C content during the last deglaciation is crucial to understanding changes in 14C production and carbon cycle mixing rates. Tropical and cold-water fossil corals provided excellent archives of atmospheric and ocean 14C content, because they can be dated by 14C and 230Th/234U/238U dating methods with high precision and high accuracy. Chapter two of this dissertation presents a radiocarbon reservoir age record of high latitude western North Atlantic Ocean during the last deglaciation via the use of fossil cold-water corals growing in waters that are rapidly exchanged with nearby surface waters. The high latitude North Atlantic Ocean reservoir age data combined with recalculated reservoir ages based on published coexisting terrestrial and marine material and Vedde ash radiocarbon dates from central and eastern North Atlantic show modern values (380+/-140 year, n=14) during the Bolling/Allerod (BA) warm period and a 200 year increase in reservoir age (590+/-130 year, n=10) during the entire Younger Dryas (YD) cold episode. The reservoir age then decreased to 270+/-20 year (n=2) at the Preboreal/YD transition, although the dates are too sparse to place confidence in this estimate. Intrusion of 14C depleted Antarctic Intermediate Water (AAIW) to the high latitude North Atlantic and reduction of NADW formation are possible causes for the coincident shift to high reservoir ages in the North Atlantic surface ocean and increased atmospheric Delta 14C during the beginning of the YD event. Chapter three presents a reconstruction of tropical ocean surface radiocarbon reservoir ages for the Holocene deglaciation. Coral data from Barbados, Kiritimati, and Vanuatu combined with previously published coral data from Tahiti, Papua New Guinea, and Marquesas provide coverage of the tropical Atlantic and Pacific Oceans (Bard et al., 1998; Cutler et al., 2004; Edwards et al., 1993; Fairbanks et al., 2005; Paterne et al., 2004). Reservoir age calculated during last glaciation (Barbados, 335 +/- 110 (n=26); Kiritimati, 335 +/- 50 (n=23); Vanuatu, 385 +/- 140 (n=27); Tahiti, 260 +/- 120 (n=14); Papua New Guinea, 490 +/- 190 (n=16); Marquesas, 510 +/- 160 (n=5)) are similar to modern values. Based on these results, it is reasonable to apply a constant reservoir age to correct the marine radiocarbon age during the last deglaciation. We also re-anchored a floating tree ring chronology using Barbados and Kiritimati coral radiocarbon record (Kromer et al., 2004). The agreement between the re-anchored floating tree ring and reservoir age corrected Vanuatu and Tahiti coral record confirmed the absence of significant reservoir age variation during the BA warm interval (Bard et al., 1998, Cutler et al., 2004). Finally, we explored parameters that might influence the reservoir age of tropical surface water. The variation of trade wind stress rather than variation of deep-ocean circulation is the most important factor of tropical surface reservoir age. These results should be helpful in constraining global ocean circulation models and simulations. Chapter four displays an atmospheric Delta14C record from paired 230Th/234U/238U and 14C dates of tropical surface fossil corals from Kiritimati and Barbados during the last deglaciation. The high resolution Delta 14C record reveals a series centennial to millennial scale fluctuations during the last deglaciation. We illustrate the sensitivity of Delta 14C changes to the uncertainty in possible 14C half-lives, and further evaluate the millennial and centennial time scale variability by also considering the effect of the geomagnetic field change. Our half-life and...