Eta vs. EtaX vs. NESDIS SSTs

The only difference between the operational Eta and EtaX was
the SSTs, shown below.
There are three SST products:
- Reynolds SST, a very smooth 1ºx1º Optimum Interpolation analysis
using ship and buoy data together with satellite data. This product is used
in the AVN/MRF and at the time of this case, in the operational Eta.
- 2D-VAR, a detailed 0.5ºx0.5º variational analysis using ship
and buoy data together with satellite data. This is a new product from the
NCEP Ocean Modeling Branch and was being tested in the experimental EtaX at
the time of this case.
- NESDIS satellite-only product. This also has 0.5ºx0.5º resolution
and captures real features missed in the Reynolds coarse SST analysis, but
it was found to have a very large high bias in areas of strong SST gradients.
The operational Eta used this product in its 32- and 22-km versions, but discontinued
its use when it was found to cause excessive coastal development in the 22-km
Eta, culminating in major snow forecast busts over North Carolina on Dec 3,
2000 and over the mid-Atlantic states 2 weeks later.
Note in particular the following:
-
The coarse resolution SST pattern has east-west oriented isotherms from
the shoreline into the open ocean, which creates a surface baroclinic zone
where the cold land airmass meets the warm water-modified air along the
Virginia coast.
-
The 2D-VAR SST field has cold (3-6ºC) continental shelf water off
the middle Atlantic coast as far south as Wilmington NC, with a warm (temperatures
exceeding 24ºC) Gulf stream eddy well east of the Virginia coast. The
isotherms landward of these warm waters are oriented north-south.
-
The Chesapeake Bay waters are assigned 3-6ºC temperatures in the
2D-VAR analysis, while having temperatures of 12-14ºC in the Reynolds
analysis.
The location and orientation of the SST gradient, and the presence of the very
warm waters in the warm eddy off of the Virginia coast played an important role
in the forecast location of coastal cyclogenesis and precipitation. Cyclogenesis
in the EtaX occurred further out to sea over the warm pool. The SST gradient
also played a role in the orientation of wrap-around snows subsequent to cyclone
passage, with the north-south gradient resulting in less wrapping back of precipitation
into the cold air inland of the mid-Atlantic coast than the east-west SST gradient
in the operational Eta. Finally, the surface temperature of the Chesapeake Bay
may have further exacerbated these differences.
SST comparison, EtaX vs. Eta (row 1) and NESDIS (row 2)

2D-VAR SST (in EtaX 12/29/00)
|

Reynolds SST (in operational Eta 12/29/00)
|
|

Difference 2D-VAR minus Reynolds
|
|
|

2D-VAR SST (in EtaX 12/29/00)
|

NESDIS satellite-only product
(discontinued from operational Eta)
|

Difference 2D-VAR minus NESDIS
|