Oceanography
The Oceanography Section (OS) has again spent much of its resources in support of the Climate System Model (CSM) project in the past year. New versions of the NCAR CSM Ocean Model (NCOM) and the CSM flux coupler code, were released on the World Wide Web for public use.
The OS members have extensively participated in both the integration and the analyses of the 300-year coupled CSM simulation. The coupled system uses no surface flux corrections in momentum, heat, or freshwater. The ocean component has virtually no trend in the surface temperatures over the 300-year period. However, there are significant trends in other model fields, especially in the deep ocean. The analyses of the oceanic results are documented in several manuscripts that will appear in a special issue of the Journal of Climate in the spring of 1998. In particular, the global ocean-only equilibrium solutions are presented in Gent et al. (1998); the spin-up procedure for the coupled integration is given in Boville and Gent (1998); the abyssal ocean drift is detailed in Bryan (1998); surface fluxes and water-mass transformation rates are investigated in Doney et al. (1998); a comparison of the wind driven circulation of the coupled and uncoupled solutions is given in Danabasoglu (1998); and the details of the third-order upwind scheme used in NCOM are presented in Holland et al. (1998). The abstracts of these papers are available on the CSM web page.
Deficiencies of both the coupled and uncoupled ocean solutions are the details of the equatorial current structure. Primarily due to the meridional grid spacing and the associated large horizontal viscosity, the equatorial currents are not well represented. To alleviate this problem, the OS members have modified the meridional resolution of the model: the grid spacing is now uniform at 0.6º near the Equator. The horizontal viscosity coefficient has also been reduced accordingly. Preliminary results with these two changes show excellent agreement with the observed current structures, as well as speed.
Another long-standing deficiency has also been addressed. Annual mean river runoff now forces the ocean as a surface boundary condition of the salinity conservation equation. The runoff is distributed according to climatological data from gauged rivers and from estimates of continental balances of evaporation and precipitation.
Gokhan Danabasoglu and James McWilliams have continued their development of a sigma-coordinate version of NCOM in which the model vertical coordinate follows the ocean bottom topography. This coordinate system eliminates the step-like representation of the bottom topography associated with the conventional height following coordinate system. In addition, the bottom boundary conditions can be implemented more naturally. The present design allows the model to be configured as an upper-ocean model that can be used in intra-annual to decadal climate variability studies. The well-known pressure gradient error of this type of terrain following coordinate systems has been minimized using high order numerical schemes in its computation. Several idealized test cases have been studied to assess this pressure gradient error (comparable or less than in similar existing models) and to demonstrate the model's ability to simulate flow over bottom topography. This sigma-coordinate version of NCOM will be a part of the CSM project and will be available for general use in the spring of 1998.
The equilibrium x3 ocean-alone solutions forced by mean monthly International Satellite Cloud Climatology Program (ISCCP) solar flux, National Centers for Environmental Prediction (NCEP) reanalyses fields, and Microwave Sounding Unit (MSU) precipitation were documented in Large et al. (1997). Danabasoglu, William Large, Scott Doney, and McWilliams extended this work to study both the effects of high-frequency forcing and the annual cycle. In particular, two new equilibrium x3 ocean-alone solutions were obtained using daily and 6-hourly NCEP reanalyses fields, respectively. The ISCCP and MSU datasets were still mean monthly. These integrations were accomplished using a dummy atmospheric model that reads and manipulates the datasets. The flux coupler is then used to pass the surface forcing data to the ocean model. The results show that the high-frequency forcing does not effect the equilibrium solution appreciably, i.e., the monthly, daily, and 6-hourly forced solutions are very similar. Naturally, the high-frequency forcing increases the variance of the model variables. However, this increase in variance does not necessarily occur at the annual or semi-annual harmonics.
Further progress has been made by Frank Bryan, Dailin Wang, and William Holland in studies of forced sea level variability using a non-eddy-resolving version of the Los Alamos National Laboratory (LANL) global ocean model (POP). Emphasis in the last year has been placed on the oceanic response to atmospheric pressure loading and in applying the results of these simulations to problems in geodesy, such as excitation of polar motion and variations in the external gravity field of the Earth. The former analysis shows that for periods up to annual, the oceanic excitation of polar motion is comparable to that of the atmosphere, and that for periods below about two weeks, the dynamic response of the ocean to pressure loading contributes significantly to the total oceanic excitation. The ocean model results were also used in observing system simulation studies (conducted in collaboration with John Wahr and graduate student Mery Molenaar, University of Colorado) that contributed to the design of the Gravity Recovery and Climate Experiment (GRACE) satellite gravity mission recently approved for launch in the year 2001.
Further progress on the simulation of the Gulf Stream system was made in a new very high resolution (1/10 degree Mercator grid) model of the North Atlantic. The simulations carried out with this model (in collaboration with Richard Smith and Mat Maltrud, LANL) show a more realistic mean path and variability of the Stream than has heretofore been achieved in basin- to global-scale models. In contrast to recent simulations at horizontal resolutions up to 1/6 degree, the model has a good representation of the northern recirculation gyre of the Gulf Stream (an important contribution to the net transport of the jet) and a realistic representation of the path of the North Atlantic Current as it rounds the Grand Banks and turns sharply northward east of Flemish Cap. In addition, the distribution of eddy energy away from the boundary current system looks far more realistic than has heretofore been achieved. Dynamical analysis of the results of this experiment and a suite of companion experiments at resolutions of 1/5 and 2/5 degree are currently underway.
Holland and Antonietta Capotondi have continued their investigation
of issues concerning the role of the ocean in North Atlantic climate
variability at decadal timescales. Previous experiments with a
stochastically-forced North Atlantic model showed pronounced
variability with a period of approximately
Ocean model solutions often depend critically on the surface boundary conditions with which they are forced. A comprehensive dataset of such boundary conditions has been prepared by Large, Doney, and Jan Morzel. It is based on winds, air temperature, and humidity from the NCEP reanalysis, cloud fractions and surface insulation from the ISCCP, and a composite of precipitation from satellite, observations, and models. The data cover for the time period 1973 through 1997 are being used to study ocean intra-annual variability and to compare model results with observations from the same time frame (in a collaboration with Michael Spall, Woods Hole Oceanographic Institution).
Ocean surface winds from the NASA Scatterometer (NSCAT) first became available in 1997. Ralph Milliff, Large, and Morzel have collaborated with Toshio (Mike) Chin (Rosenstiel School of Marine and Atmospherice Sciences (RSMAS), University of Miami) to merge these data with European Remote Sensing Satellite-1 (ERS-1) and European Remote Sensing Satellite-2 (ERS-2) scatterometer wind measurements to produce surface wind datasets required for various aspects of OS research. A wavelet-based interpolation method for combining scatterometer surface wind fields with NCEP reanalysis fields (Chin et al., 1996) has been used to produce a consistent global wind dataset for the year 1 August 1996 to 31 July 1997. The new methodology emphasizes the high-wavenumber content of the scatterometer observations that might otherwise be sacrificed (e.g., through temporal averaging or degradation of spatial resolution) to offset sampling biases inherent in the swath-based schemes of present and planned scatterometer missions. These winds are now available for forcing ocean models and studying the effect of the high spatial resolution. The ERS-1 surface zonal winds from the Tropical Ocean and Global Atmosphere Program Coupled Ocean-Atmosphere Response Experiment (TOGA COARE) intensive observation period (IOP) have been combined by Chris Wikle (Geophysical Statistic Project, GSP/OS) with numerical weather prediction (NWP) products and aircraft observations to produce near surface wind spectra covering spatial scales from 1 km to 1000 km. These results suggest the existence of inertial sub-ranges supporting the forward cascade of enstrophy and the inverse cascade of energy. Wikle and Milliff use this multi-scale result in the prescription of a prior probability distribution for a correlated noise model that comprises one level in a Bayesian hierarchical model to blend NSCAT and NCEP data for surface wind fields in the tropical Pacific. The new surface winds are expected to be free of unrealistic features present in analysis winds alone. These winds will contribute to a study of tropical cyclogenesis associated with the Madden-Julian Oscillation that is a collaboration between Milliff, Roland Madden (Climate Analysis Section, CAS), Akira Kasahara (Global Dyanmics Section, GDS), and Tetsuo Nakazawa (visitor, Japan Meteorological Agency, JMA).
Milliff and Large are collaborating with Xiaoqing Wu, Wojciech
Grabowski, and Mitch Moncrieff (all of the Mesoscale and Microscale
Meteorology Division, MMM) to couple the Climate Research Model (CRM)
with an ocean model. The 39-day dynamically consistent surface forcing
from the CRM was used to force a
Milliff has collaborated with Andy Royle, Wikle, and Mark Berliner (all of GSP) to develop a methodology for generating high-resolution 10 m winds and surface pressure fields for the Labrador Sea region of the North Atlantic (Royle et al., 1997). A Bayesian hierarchical model uses the geostrophic relation to constrain the deduction of a surface pressure field from measured scatterometer winds. The pressure field is then extended to regions of the Labrador Sea not sampled by the scatterometer at a given time, and a consistent wind field for these regions is also inferred. The product of this research includes a time series of high-resolution (50 km, 2x daily) wind and surface pressure fields for the Labrador Sea for the period October through November, 1996. This coincides with extensive field programs and a NSCAT calibration and validation exercise in the Labrador Sea.
Milliff, Harry van Loon (CAS), and Marilyn Raphael (UCLA) have
evaluated the 500 millibars NCEP reanalyses (1968-1997) for zonally
asymmetric structures that document the seasonal and interannual
variability of the quasi-stationary waves (QSW) of the Southern
Hemisphere. Spectral peaks occur on
Marine biogeochemical processes as they relate to the global carbon cycle and climate system are under study by Doney. Two paths are currently being pursued. A simple, nitrogen-based ecosystem model has been developed and embedded in a 1D version of the KPP upper ocean boundary layer scheme. The resulting coupled biological-physical model has been compared extensively with available time-series data for the subtropical North Atlantic, showing good skill at replicating the major features of the seasonal cycle and vertical depth structure. The 1D model is currently being used to explore the biological response to high-frequency physical forcing and to develop improved biological parameterizations for the global modeling effort.
On a second front, a series of tracer and carbon cycle calculations are underway in the global CSM ocean model as a prelude to a full 3D ocean biology model. The tracer work provides valuable information on the ventilation rates of the ocean model on timescales from several years out to centuries. To date, two tracer runs are available and are being analyzed: a natural equilibrium radiocarbon simulation and an anthropogenic perturbation radiocarbon simulation, which includes both the Suess effect (the century-scale dilution of radiocarbon isotopic ratio by fossil-fuel emission) and the bomb-radiocarbon transient. The physically-driven, ocean carbon solubility pump is also being studied in the global model, a version of which now incorporates seawater carbonate thermodynamics and air-sea gas exchange. Abiotic ocean carbon transport is closely linked to the large-scale meridional heat and freshwater fluxes, and we expect corresponding improvements in the carbon transport associated with the new mesoscale eddy flux parameterization and surface boundary conditions. A perturbation simulation has also been carried out to estimate the oceanic uptake of anthropogenic carbon dioxide (CO2) over the industrial period.
Geophysical Fluid Dynamics and Turbulence
Geophysical fluid dynamics (GFD) investigations of decadal variability of the thermohaline circulation and its sensitivity to the oceanic convection parameterization are continuing in a collaboration between Danabasoglu, Ramalingam Saravanan (GDS), and McWilliams. As part of a hierarchy of coupled ocean-atmosphere models to study low-frequency variability in the climate system, they have developed a coupled atmosphere-ocean sector model with an idealized land surface and ocean basin geometry. Extended integrations of this idealized coupled model show preferred interdecadal timescales in many of the oceanic variables. This interdecadal signal in the ocean is linked to a North Atlantic Oscillation-like mode of variability in the atmosphere. The regions of most active variability coincide with regions of deep convective activity in the ocean, suggesting that the variability may be sensitive to the parameterization of convection in the ocean model. Subsequent sensitivity studies have shown that the characteristics of the oceanic variability are significantly influenced by the particular choice of the convection parameterization used in the ocean model, where the less plausible choice induces a spurious centennial modulation of the interdecadal variability.
Robert Kerr (MMM, Climate and Global Dynamics Division (CGD), and High Altitude Observatory, HAO), Large, and Eileen Saiki (Advanced Study Program, ASP) are collaborating on direct numerical simulations (DNS) of oceanic double diffusion. Their goal is to provide a tested parameterization, consistent with the few available observations, for direct use in the CSM ocean component. Low-resolution numerical salt fingers and staircases have been achieved. The techniques of obtaining similar results at high resolution are being perfected.