Mapping Observations in a Dynamical Framework: A New Ocean Atlas.

story by Helen Hill

 

This month we focus on work by Gael Forget and the ECCO team who have been using MITgcm in an effort to construct a new ocean atlas.

Ocean atlases typically contain climatological distributions of quantities such as, say, temperature, salinity or sea surface height (from which scientists can deduce the characteristics of the flow, its speed and direction). Data-sets are generally provided as time averaged fields, for example as a set of monthly means. Such estimates of ocean state have a variety of uses  including providing the framework within which to initialise and test model solutions, as well as offering a resource to help identify and understand oceanic flows and transports.

Because of the extreme spatial and temporal inhomogeneity of the underlying data, most climatologies necessarily involve some kind of objective mapping, typically through weighted averaging in space and time. However, because this process knows nothing about the physics of the ocean (the way in which, say, the distribution of temperature relates to the circulation) it cannot be expected to result in a physically consistent set.

ECCO‘s newest atlas gets around this by using MITgcm as a means of optimally synthesising data within the framework of a physically accurate general circulation model. By also limiting the time-period over which data was included (the atlas strictly only applies to the period between December 2003 to November 2006 – the beginning of the era of near global Argo float coverage), OCCA (short for OCean Comprehensible Atlas, the term ”comprehensible” denoting that the underlying equations of motion are known) (short for OCean Comprehensible Atlas) provides a singularly accurate 3-year “snap-shot” of the world ocean state for that period.

About the data:

OCCA draws on the following datasets

Data Set	Source
Argo Profiles	Coriolis Argo Data Center
(*)XBT profiles	D. Behringer, NCEP
(*) CTD profiles	WOCE
(*) Southern Ocean profiles	SEaOS
(*) Tropical mooring time series	TAO-PIRATA
infrared SST	Reynolds and Smith (1994), NOAA
microwave SST	Remote Sensing System
Topex-Poseidon SLA	PODAAC
Geosat-Follow-On SLA	US Navy, NOAA
Envisat SLA	AVISO
surface mean dynamic topography	M. Rio, Rio et al. (2005)
WOA 05 – Argo blend	Locarnini et al. (2006)
near-surface atmospheric state	Kalnay et al. (1996), NCEP
scatterometer windstress	PODAAC

Asterisks denote data sets with fairly regional or sparse data. SST stands for satellite measured sea surface temperature; SLA for satellite altimeter measured sea level anomaly.

The MITgcm-Interpolation Problem:

The constrained least-squares problem is similar to that solved by Wunsch and Heimbach (2007).  Unlike in that work, however, the OCCA configuration additionally includes MITgcm’s sea-ice thermodynamic model (Hibler, 1980) and bulk formulae atmospheric surface layer scheme (Large and Yeager 2004). The other important difference from Wunsch and Heimbach (2007, 2008) is a reduction of the estimation interval(s), from 14-16 years to sixteen months.

The near-global model domain extends from 80^◦ S to 80^◦ N, with a horizontal resolution of 1 degree, 50 vertical levels, and a time step of one hour. The model uses a third order upwind advection scheme, Adams-Bashforth explicit time-stepping, and an implicit scheme for vertical diffusion. The model uses hydrostatic and Boussinesq approximations, and a linear implicit free surface. Sub-gridscale processes are parameterized using KPP vertical mixing (Large et al. 1994), GM eddy thickness diffusivity (1000m²/s) (Gent and Mcwilliams 1990), vertical (104 m²/s) and isopycnic (10−5 m²/s) tracer diffusivity, vertical (10-3 m²/s) and horizontal (10-4 m²/s) friction, with a quadratic bottom drag of 0.002. Free-slip side and no-slip bottom boundary conditions round out the implementation. The ocean bulk formulae scheme handles freshwater as a virtual salt flux. while the 6-hourly atmospheric air-temperature, specific humidity, wind velocity, short wave downward flux and precipitation from the NCEP re-analysis are used as the starting components for atmospheric bulk formulae calculation.

Each calculation estimates the oceanic state over one calendar year plus the previous four months, to cover slightly more than one seasonal cycle. The results from three such (overlapping) calculations are then time-averaged, first to form a three year daily time series and then to form a mean monthly atlas. This atlas, representing an average seasonal cycle for the period from 2004 to 2006, is OCCA, which stands for OCean Comprehensible Atlas, the term ”comprehensible” denoting that the underlying equations of motion are known.

Physically consistent interpolation through the use of MITgcm then results in an atlas of both optimally mapped observed quantities as well as corresponding derived quantities including, for example, mid-depth dynamic topography and ocean fluxes of heat and salt/fresh water. The OCCA atlas consists of a set of fields for each month of the year which can be readily accessed through the MIT ECCO project LAS server. Annual mean fields as well as a daily time series can be found at the same location.

Look out for Forget’s paper “Mapping ocean observations in a dynamical framework: a 2004-2006 ocean atlas” accepted for publication in JPO later this Spring. Want to know more? Contact Gael…

Publications:

Forget, G., 2010: Mapping ocean observations in a dynamical framework: a 2004-2006 ocean atlas. Journal of
Physical Oceanography – to appear (doi: 10.1175/2009JPO4043.1)

References:

Gent, P. and J. Mcwilliams, 1990: Isopycnal Mixing in Ocean Circulation Models. Journal of
Physical Oceanography, 20 (1), 150–155.

Hibler, I. W., 1980: Modeling a Variable Thickness Sea Ice Cover. Monthly Weather Review,
108 (12), 1943–1973.

Kalnay, E., et al., 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bulletin of the Amer-
ican Meteorological Society, 77 (3), 437–471.

Large, W. and S. Yeager, 2004: Diurnal to decadal global forcing for ocean and sea-ice models:
The data sets and flux climatologies. Technical Report TN-460+STR, NCAR.

Large, W., J. McWilliams, and S. Doney, 1994: Oceanic vertical mixing: A review and a model
with a nonlocal boundary layer parameterization. Rev. Geophys, 32 (4), 363–403.

Locarnini, R., A. Mishonov, J. Antonov, T. Boyer, H. Garcia, and S. Levitus, 2006: World
Ocean Atlas 2005 Volume 1: Temperature. NOAA Atlas NESDIS.

Rio, M., P. Schaeffer, F. Hernandez, and J. Lemoine, 2005: The estimation of the ocean Mean
Dynamic Topography through the combination of altimetric data, in-situ measurements and
GRACE geoid: From global to regional studies.”. Proceedings of the GOCINA international
workshop, Luxembourg.

Wunsch, C. and P. Heimbach, 2007: Practical global oceanic state estimation. Physica D:
Nonlinear Phenomena, 230 (1-2), 197–208.

Wunsch, C. and P. Heimbach, 2008: The Global Zonally Integrated Ocean Circulation (MOC),
1992-2006: Seasonal and Decadal Variability. Journal of Physical Oceanography, (submit-
ted).