Forecast 4/6/09

• 1/1/2009-3/22/2009--FVCOM 2009 flow fields (high res), assimilating PCCS data
• 3/22/2009-4/6/2009--climatological FVCOM fields (lower res), no assimilation
• 4/6/2009-4/15/2009--climatological FVCOM fields, climatological satellite data

I mapped the adult abundances for Calanus, Pseudocalanus, and Centropages for the 10d assimilation windows and uploaded the images to Picassa.  You should be able to click through the figures.  Each figure contains three panels.  The two on the left are the initial conditions for the 10d period.  The leftmost is the initial guess (usually, the output from the previous 10d window).  The second, labeled "post" (for posterior) is the initial condition estimated by the EnKS algorithm.  The panel on the right is at the end of the 10d period.  OK, here are the images:

Some comments on the figures:

• If the two initial conditions look the same, there was likely no PCCS cruise in that period
• If the two initial conditions are similar, then the PCCS data and model agree well in that period
• If the two initial conditions are wildly different, then the model required significant adjustment to reproduce the data.

We plan to try a few things to try to minimize the "case 3" situations.  In particular, using better parameters from Nick's genetic algorithm work, using BCs from our Gulf of Maine model (esp. for Calanus), and trying different analysis intervals.