Up to now, Global Carbon Cycle Models (GCCM) have only represented the soil and biosphere in a very simplified way. For example, the High Resolution Biosphere Model (HRBM) (Esser 1987; Esser and Lautenschlager 1994) distinguishes five subreservoirs and determines the fluxes between them for selected area of a global grid. The models have not yet been sufficiently tested against global observations. Such testing is difficult because the observed variables are in turn dependent on the behavior of other geological carbon pools, e.g., the atmosphere and ocean.

Measurements of cosmogenic nuclides made in situ in the Earth's surface are being used to help resolve a wide range of geologic and chronologic questions. Cosmogenic nuclides (3He, 10Be, 14C, 21Ne, 26Al 36C1 are presently used) can reveal rock exposure history information leading to estimates of timing of surface forming events, rates and styles of erosion, and timing and durations of episodes of burial. Depending on the problems being tackled, a significant source of error (±10–25%) for any cosmogenic nuclide method is the present uncertainty in the spatial and temporal variability of the rates of production of these in-situ nuclides.

Although soil contains about three times the amount of carbon present in the preindustrial atmosphere, determining how perturbations (e.g., changing land use, CO2 fertilization, changing climate and anthropogenic nitrogen deposition) alter soil carbon storage and influence atmospheric CO2 levels has proved elusive. Not knowing the soil carbon turnover times causes part of this uncertainty. I outline a strategy for using radiocarbon measurements to estimate soil organic matter turnover times and inventories in native soil. The resulting estimates of carbon exchange produce reasonable agreement with measurements of CO2 fluxes from soil. Furthermore, derivatives of the model are used to explore soil carbon dynamics of cultivated and recovering soil. Because the models can reproduce observed soil 14C measurements in native, cultivated, and recovering ecosystems (i.e., the underlying assumptions appear reasonable), the native model was modified to estimate the potential rate of additional carbon storage because of CO2 fertilization. This process may account for 45–65% of the “missing CO2 sink.”

Bomb 14C from nuclear tests in the atmosphere has proved to be a particularly useful tool in the study of the carbon cycle. We provide here a ca. 30-yr time series of 14C concentrations in the atmosphere between 28°N and 71°N and in the ocean surface between 45°S and 45°N. More recently (since 1990), a north-south profile also has been obtained for 14C in the surface waters of the Atlantic Ocean. The measurements were performed using the conventional technique of beta counting or large samples (4 to 5 liter CO2) in CO2 proportional counters. These data show that the 14C concentration in the atmosphere is leveling off with a time constant of 0.055 yr-1, and is now approaching that of the ocean surface at lower latitudes.

Additional tracer studies have been concerned especially with the penetration of bomb 14C into the deep ocean. The Norwegian and Greenland seas are of interest as a sink for atmospheric CO2 and also a source of water for the deep Atlantic Ocean. During the last five years, several 14C depth profiles have been measured from the Fram Strait (79°N) to south of Iceland (62°N), using the AMS technique available at the University of Arizona AMS Facility. We considered it important to repeat and compare a few of the profiles with those produced by the GEOSECS expedition in 1972 and the TTO expedition in 1981. The profiles show that water descending to the deep Atlantic Ocean is originating mainly from intermediate and surface depths in the Nordic Seas. However, the ventilation rate of the Norwegian Sea deepwater is too slow to be an important component in the transfer of water over the Greenland-Scotland Ridge.

The distribution of bomb-produced 14C in the ocean provides a powerful constraint for circulation models of upper ocean mixing. We report 14C measurements from an east-west section of the main thermocline at 24°N latitude in the subtropical North Atlantic Ocean in summer 1992, and one profile from the Gulf of Mexico in 1993. Observed gradients reflect the transient invasion of bomb 14C into the thermocline via mixing along isopycnals from the poleward outcrop, with progressively more sluggish mixing at greater depths. A slight deepening of the profile is observed over the 20-yr period since the GEOSECS survey at one location where the comparison is possible.

This paper presents carbon isotope data measured in three soil profiles from the Salitre area, Central Brazil. The study forms part of a research project on tropical and subtropical soils in Brazil, in which the main objective is to use carbon isotopes to provide information about vegetation changes that have occurred in relation to climate changes during the Holocene. 14C data from charcoal samples and soil organic matter (SOM) indicate that the organic matter in the soils studied is of Holocene age at least. Furthermore, the presence of a significant amount of charcoal in the soils suggests that forest fire was a significant ocurrence during the Holocene and probably had an important role in determining the dynamics of forest vegetation in the study area. Correspondingly, 13C data indicate that C3 plants provided the dominant vegetation of the study area, even during the dry periods when savanna vegetation is supposed to have replaced the forest communities. This study contributes to our better understanding of the relation between climatic changes and vegetation in the subtropical region of Brazil.

Fieldwork for the World Ocean Circulation Experiment (WOCE) radiocarbon program was recently completed. Ca. 9000 samples were collected for analysis using both conventional β-counting techniques and the newer AMS technique. The mean uncertainty for the β analyses is 3‰; for AMS analyses, ca. 4.5‰.