The variability of the midlatitude ocean is
predominantly driven by
atmospheric forcing.
Estimates of the large-scale midlatitude oceanic state
are limited by few observations, except for surface
fields of SST (from AVHRR) and sea level height (from TOPEX).
Under the ECPC, we are attempting to nowcast the
present ocean state and to forecast some fraction
of it future state.
Nowcasts are made using the OPYC ocean model
driven by NCEP forcing (wind stress, total surface heat flux,
TKE input to the mixed layer, and fresh water flux).
The nowcast is a continuing run, driven over decadal
timescales up to the present.
No oceanic observations
are included in the present nowcast, only the information
supplied to the ocean model by the forcing.
Forecasts are made using the OPYC ocean model
driven by predicted forcing fields from the ECPC GSM.
There are two main sources of error in the
nowcasts and the forecasts.
The first is that no oceanic observations are included
to improve the nowcast (which also serves as the
initial condition for the forecast).
We are now developing a strategy
(under NASA funding) for mapping NASA-GFSC ocean analyses to the OPYC grid
to improve the initial state.
The forecasts are also deterioted by drifts
in the atmospheric predicted forcing, a problem we are
also trying to remedy.
Ocean forecasts are issued
every month, 3 months in advance. The procedure
involves the following
4 steps:
| Step |
Tool |
Product |
Comments |
| 1 |
ECPC Global Spectral
Model |
Atmospheric Forecast |
(NCEP initialization) |
| 2 |
OPYC ocean model |
Ocean State Forecast |
(forced by atmospheric
forecast) |
| 3 |
GMT graphics package |
Maps of several oceanic
variables |
(done for total and
anomalous fields) |
| 4 |
Locally developed
software |
HTML documents |
(displayed on the
WWW every month) |
Forcing functions:
The ocean model is forced by
the overlying atmosphere by four different variables: (a) Wind Stress,
(b)
Surface heat fluxes, (c) Turbulent Kinetic Energy, and (d) Freshwater
fluxes, i.e., Precipitation minus evaporation.
Initialization:
OPYC is initialized using NCEP reanalysis and analysis data. The first
initial
conditions (January 1999) were obtained by running OPYC in hindcast
mode
since 1958 using anomalous
reanalysis data from NCEP. From that point in time and
before every monthly forecast, OPYC is advanced using NCEP analysis
forcing
functions.
For instance, our first forecast was initialized with ocean conditions obtained
from a 41-year long run initiated in 1958 under NCEP reanalysis forcing
(forecast for January, February and March of 1999).
The next forecast was issued in early February of 1999. To obtain the initial
conditions for February first 1999 we ran OPYC for one month starting
in January 1st 1999 forcing it with NCEP analysis fluxes (i.e., wind
stresses, heat fluxes, etc.). From there on the process repeats itself
every month.
In order to further improve our forecasting system we will use initial
conditions obtained from (a) NASA models and, (b) sensitivity runs
aimed to obtain optimal initial conditions and parameters by minimizing
the difference between model solutions and different sets of observations.
This experimental forecast stage is in preparation for the development of
a coupled ocean-atmospheric model at ECPC.
Available Forecast Variables:
Ten different variables are forecast every month.
1) Sea
surface temperature
2) Surface
heat flux
3) Wind
stress
4) Freshwater
fluxes
5) Surface
currents
6) Surface
Geostrophic Currents
7) Mixed
Layer Depth
8) Thermocline
Depth
9) Vertical
Profile of Equatorial Temperature
10) Heat
Storage (upper 400 m)
They are displayed along with the most recent available observations
for comparison.
