ECCO 1/6 Global Model

Grid

east-west	0-360	2,160 GP
north-south	78S-81N	960 GP
vertical	0-6525 m	30 GP
	total	62,208,000 GP

dz / m	z / m	dz /m	z /m	dz / m	z /m
10.0	10.	40.0	225.	245.0	1455.
10.0	20.	50.0	275.	290.0	1745.
10.0	30.	60.0	335.	340.0	2085.
12.5	42.	70.0	405.	395.0	2480.
15.0	58.	80.0	485.	460.0	2940.
17.5	75.	95.0	580.	535.0	3475.
20.0	95.	115.0	695.	625.0	4100.
25.0	120.	140.0	835.	725.0	4825.
30.0	150.	170.0	1005.	825.0	5650.
35.0	185.	205.0	1210.	875.0	6525.

Density Calculation

New polynominal coefficients (POLY3.COEFFS) that reflect the new vertical discretization have been calculated using util/scripts/knudsen2/knudsen2.f

Topography

The interpolated topography is already fine enough that only minimal handtuning has to be done. An example of the fine resolution is the Indonesian Throughflow region.

Interpolation from ETOPO5 onto the model grid, no smoothing
Masking of regions:

some marginal seas (Baltic Sea, Hudson Bay, Melville Sound, Persian Gulf, Black Sea)
lakes and (due to the interpolation) closed bights

Widening of some passages up to 3 grid points

Strait of Gibraltar
entrance to Red Sea

Rebuild of some islands:

Canary Islands
Madeira
Azores
Jan Mayen
Bermuda
Bahamas
Caribic
Capeverdes
Comores

Because of the limited availibility of the surface forcing data the domain was limited to 79N. North of that latitude all points were masked out (the domain north of 79N was kept for technical issues, to keep an appropriate number for the parallelization)

Initialization

Potential temperatures and salinities of the Levitus et al. (1994) climatology were gap-filled and interpolated onto the model grid
The data were checked against the pmask

Surface Forcing

Daily zonal and meridional wind stress data and net heat and salt surface fluxes of the period 1992-1997 were taken from the global 2 degree run of the global state estimation (iteration 60) and interpolated onto the model grid. Because of the lack of a mixed layer model the heat and salt surface fluxes are equally distributed over the first two vertical level.
An additional restoring (time scale 60 days) to surface values of the monthly Levitus data is applied to the first two vertical level. This is done because of the remaining drift in the forcing fields at this number of interations.
Two restoring zones, a 12 grid point wide at the northern limit and a 18 grid point wide at the southern limit of the domain (time scales vary from 5 to 60 days), are used to restore the full water column to monthly Levitus values. [Complete 3D restoring mask]

Data

The model should be spinned up over two years of the period 1992 and 1993. Afterwards the full range 1992-1997 is simulated/

	online (GB)	permanent (GB)
input data (64 bit)
initialization (3D temp and salt)	1	1
monthly climatology (4D temp and salt)	12	12
daily forcing (3D wind stress and fluxes)	145	145
output data
thirdyearly restart (64 bit, 15GB)	30	360
daily snapshot (32 bit, 1GB)	180	2,923
total	368	3,441

It is assumed to run the model in pieces of 1/3 year and produce daily snapshots. The total run length is 8 years.

This page is maintained by Arne Biastoch, March 13, 2000

ECCO 1/6 Global Model

Scientific Justification

Computational Requirements

Test Runs