The ocean component of the Community Climate System Model version 4 (CCSM4) is described, and its solutions from the twentieth-century (20C) simulations are documented in comparison with observations and those of CCSM3. The improvements to the ocean model physical processes include new parameterizations to represent previously missing physics and modifications of existing parameterizations to incorporate recent new developments. In comparison with CCSM3, the new solutions show some significant improvements that can be attributed to these model changes. These include a better equatorial current structure, a sharper thermocline, and elimination of the cold bias of the equatorial cold tongue all in the Pacific Ocean; reduced sea surface temperature (SST) and salinity biases along the North Atlantic Current path; and much smaller potential temperature and salinity biases in the near-surface Pacific Ocean. Other improvements include a global-mean SST that is more consistent with the present-day observations due to a different spinup procedure from that used in CCSM3. Despite these improvements, many of the biases present in CCSM3 still exist in CCSM4. A major concern continues to be the substantial heat content loss in the ocean during the preindustrial control simulation from which the 20C cases start. This heat loss largely reflects the top of the atmospheric model heat loss rate in the coupled system, and it essentially determines the abyssal ocean potential temperature biases in the 20C simulations. There is also a deep salty bias in all basins. As a result of this latter bias in the deep North Atlantic, the parameterized overflow waters cannot penetrate much deeper than in CCSM3.

A parameterization for the restratification by finite-amplitude, submesoscale, mixed layer eddies, formulated as an overturning streamfunction, has been recently proposed to approximate eddy fluxes of density and other tracers. Here, the technicalities of implementing the parameterization in the coarseresolution ocean component of global climate models are made explicit, and the primary impacts on model solutions of implementing the parameterization are discussed. Three global ocean general circulation models including this parameterization are contrasted with control simulations lacking the parameterization. The MLE parameterization behaves as expected and fairly consistently in models differing in discretization, boundary layer mixing, resolution, and other parameterizations. The primary impact of the parameterization is a shoaling of the mixed layer, with the largest effect in polar winter regions. Secondary impacts include strengthening the Atlantic meridional overturning while reducing its variability, reducing CFC and tracer ventilation, modest changes to sea surface temperature and air-sea fluxes, and an apparent reduction of sea ice basal melting.

There has over recent years emerged a controversy as to how much of the radiocarbon released into the atmosphere by nuclear weapons testing has been taken up by the ocean. Hesshaimer et al. [1994] made a case based on stratospheric and tropospheric measurements coupled with estimates of total bomb radiocarbon yield that it was not possible to explain both atmospheric observations and existing ocean‐based bomb radiocarbon uptake estimates [Broecker et al., 1985, 1995]. They therefore proposed that estimates of the oceanic sink should be revised downward by about 25%. One reason for concern over this discrepancy is that the widely used wind speed dependent air‐sea gas exchange parameterization of Wanninkhof [1992] is scaled to give an average exchange rate matching that given by the ocean bomb‐radiocarbon budget. An example of an application of the Wanninkhof [1992] parameterization is in estimating ocean CO2 uptake based on direct measurements of the air‐sea pCO2 difference [Takahashi et al., 1997]. Such estimates scale linearly with the air‐sea gas exchange coefficient. Further, as has been highlighted by the Hesshaimer et al. [1994] study, an understanding of the global budget of radiocarbon is an important issue in and of itself. In this paper, a number of new approaches to assessing the size of the ocean bomb radiocarbon sink are explored, and estimates are given for the total ocean bomb radiocarbon inventory during both the mid‐1970s and the mid‐1990s. The revised estimates of bomb‐radiocarbon ocean uptake yield a mid‐1970s inventory in closer agreement with that proposed by Hesshaimer et al. [1994] than the inventory obtained by the Broecker et al. [1995] study.

Ocean carbon uptake and storage simulated by the Community Earth System Model, version 1–Biogeochemistry [CESM1(BGC)], is described and compared to observations. Fully coupled and ocean-ice configurations are examined; both capture many aspects of the spatial structure and seasonality of surface carbon fields. Nearly ubiquitous negative biases in surface alkalinity result from the prescribed carbonate dissolution profile. The modeled sea–air CO2 fluxes match observationally based estimates over much of the ocean; significant deviations appear in the Southern Ocean. Surface ocean pCO2 is biased high in the subantarctic and low in the sea ice zone. Formation of the water masses dominating anthropogenic CO2 (Cant) uptake in the Southern Hemisphere is weak in the model, leading to significant negative biases in Cant and chlorofluorocarbon (CFC) storage at intermediate depths. Column inventories of Cant appear too high, by contrast, in the North Atlantic. In spite of the positive bias, this marks an improvement over prior versions of the model, which underestimated North Atlantic uptake. The change in behavior is attributable to a new parameterization of density-driven overflows. CESM1(BGC) provides a relatively robust representation of the ocean–carbon cycle response to climate variability. Statistical metrics of modeled interannual variability in sea–air CO2 fluxes compare reasonably well to observationally based estimates. The carbon cycle response to key modes of climate variability is basically similar in the coupled and forced ocean-ice models; however, the two differ in regional detail and in the strength of teleconnections.

No abstract

Results are presented from a century-long 1/10 • global ocean simulation that included a suite of age-related passive tracers. In particular, an ensemble of five global Boundary Impulse Response functions (BIRs, which are statistically related to the more fundamental Transit Time Distributions, TTDs) was included to quantify the character of the TTD when mesoscale eddies are explicitly simulated rather than parameterized. We also seek to characterize the level of variability in water mass ventilation timescales arising from eddy motions. The statistics of the BIR timeseries are described, and it is shown that the greatest variability occurs at early times, followed by a remarkable conformity between ensemble members at longer timescales. The statistics of the first moment of the BIRs are presented, and the upper-ocean spatial distribution of the standard deviation of the first moment of the BIRs discussed. It is shown that variations in the BIR first moment with respect to the ensemble average are typically only a few percent, and that the variability slightly decreases with increasing ensemble size, implying that only a few ensemble members may be necessary for a reasonable estimate of the TTD. The completeness of the estimated TTD, i.e., the degree to which the century long BIRs capture the range of global ocean ventilation timescales is discussed, and the potential for extrapolation of the BIR to longer times is briefly explored. Several regional BIRs were also simulated in order to quantify the relative abundance of fluid parcels that originate in specific geographical locations.

[1] There is not yet widespread agreement as to the underlying cause of the 80-100 ppmv roughly 100-kyr-duration glacial-interglacial cycles in atmospheric pCO 2 . Most of the mechanisms which have been proposed to account for the observed pCO 2 variations appear to in some way violate interpretations of paleo proxy data. The inability of a single mechanism to explain the observed cycles in atmospheric CO 2 (which show amazing similarity over the past 430,000 years) is perplexing, and leads us to consider whether a combination of mechanisms might be consistent with available evidence. Consistent with previous work, we find that physical changes (ocean circulation, temperature, mixing) can only explain part of the observed atmospheric pCO 2 variability; changes in ocean chemistry are invoked to explain the remainder. In order to account for the initial pCO 2 drawdown (from ''interglacial'' to ''intermediate'' levels), we invoke physical changes in the ocean (mixing, temperature). The transition from intermediate atmospheric pCO 2 levels to full glacial conditions involves a small increase in mean ocean nutrient levels and mean ocean alkalinity, accomplished by falling sea level and subsequent erosion of organic-rich shelf sediments. The first part of the transition out of full glacial conditions is achieved through increased temperature and increased mixing in the Southern Ocean. The final part of the atmospheric pCO 2 rise up to full interglacial conditions is accomplished through rising sea level and the subsequent change in mean ocean alkalinity and phosphate, and a rise in the Northern Hemisphere temperature and ocean mixing. The proposed sequence of events is consistent with most existing proxy evidence for paleo-nutrient levels and changes in export production over the last glacial-interglacial cycle. Furthermore, it is consistent with evidence for a whole-ocean shift in d 13C toward significantly more negative values in the late glacial. The proposed scenario is also consistent with ice core-based timing constraints, as summarized by Broecker and Henderson (1998). We show that we are able to explain the full magnitude of the glacial-interglacial cycle in atmospheric pCO 2 without the need to invoke iron-fertilization in the Southern Ocean.

Results from a simulation of the ocean "transit-time distribution" ("TTD") for global and regional ocean surface boundary conditions are presented based on a 5000-yr integration using the Parallel Ocean Program ocean general circulation model. The TTD describes the probability that water at a given interior point in the ocean was at some point on the ocean surface a given amount of time ago. It is shown that the spatial distribution of ocean TTDs can be understood in terms of conventional wisdom regarding time scales and pathways of the ventilated thermocline and the thermohaline circulation-driven deep-ocean circulation. The true mean age from the model (the first moment of the TTD) is demonstrated to be very large everywhere, because of very long-tailed distributions. Regional TTD distributions are presented for distinct surface boundary subregions, and it is shown how these can help in the interpretation of the global TTD. The spatial structure of each regional TTD is shown to become essentially the same at relatively long times. The form of the TTD at a given point in the ocean can be very simple, but some regions do exhibit more complicated multimodal distributions. The degree to which a simple functional approximation to the TTD is able to predict the spatial and temporal evolution of selected idealized tracers (for which interior sources and sinks are known or zero) with knowledge of only the tracer surface boundary condition is explored.