Details for floats

[rduran]

I am looking for details of ROMS' floats, I would like to put together a brief description. So far I have found in the manual (Figure 11: Flow chart of the model main program) that floats are stepped in time every baroclinic time step. Presumably the latest time-stepped 3D velocity is integrated to add the latest piece of trajectory.

Could anyone confirm this is accurate or not?

Could anyone point me in the direction of paper (s) that would elaborate on the topic?

thanks many,

R.

p.s. thank you indeed for the manual Kate et. al.

[kate]

The place to look is in step_floats.F, which says:

Code: Select all

!  This routine time-steps  simulated  floats  trajectories using a    !
!  fourth-order Milne predictor and fourth-order Hamming corrector.    !

The last time I wrote up the float code, it was using fourth-order Runge-Kutta, which required four interpolations to the float location. The current code only needs two per timestep.

[rduran]

many thanks Kate, that all makes sense.