Question about GOFS 3.1 and ESPC-D-V02

[Tongtong Xu - NOAA Affiliate]

Hi,

My name is Tongtong Xu. I am a research scientist working at CIRES, University of Colorado Boulder and NOAA Physical Sciences Laboratory. We are a team looking to develop statistical models for predicting coastal flooding in U.S. And we are thinking of training the model based on HYCOM specifically the GOFS 3.1 and then moving to real-time forecasts later. 

I see that GOFS 3.1 has been discontinued around 2024 and a new model called ESPC-D-V02 has global analysis from 2024 to present. May I ask how different you think the GOFS 3.1 and ESPC-D-V02 are? Is it scientifically and operationally reasonable to train on GOFS 3.1 and predict based on ESPC-D-V02?

Also do you have someone specific that I could contact?

Best regards,

Tongtong

————————

Tongtong Xu

Research Scientist

CIRES, University of Colorado Boulder

NOAA Physical Sciences Laboratory

[Alan Wallcraft]

In deep water, GOFS 3.1 and ESPC-D are broadly similar because they use virtually identical ocean data assimilation.  The US Navy does not measure the skill of its global prediction products in coastal regions, because it has higher resolution models for such regions that only use the global products for boundary conditions. 

For coastal regions, I would expect ESPC-D to the different, and likely better, that GOFS 3.1, because:

a) Native horizontal resolution is 2x

b) ESPC-D includes tides

c) ESPC-D includes surface pressure forcing

d) GOFS is forced by stand-alone NAVGEM (i.e. with a data atmosphere), ESPC-D is coupled to NAVGEM

NCEP's RTOFS is similar if GOFS 3.1 with different atmospheric forcing and slightly different data assimilation and is still running today.

Alan.