[MITgcm-support] ctrl-pkg: meaning xx_${varname}_period changed slightly

[Martin Losch]

Dear users of the ctrl-package,

after the today's tag (checkpoint62u (2011/3/15)) I changed the behavior of the run time parameter xx_${varname}period, that controls the sampling interval of a control variable ${varname}, but only for the case when the calendar package "cal" is used (ALLOW_CAL is defined). 

Up to checkpoint62u, a value of 0. (e.g. xx_atemp_period = 0. in data.ctrl) meant 12 cyclic monthly updates and there was no way to specify a constant (in-time) control parameter. In contrast, in pkg/exf you specify time-independent fields by, e.g., atemp_period = 0. and 12 climatological monthly records as -12. I changed the convention of the ctrl-package in this way, so that e.g. xx_atemp_period = 0., in data.ctrl means that the control variable is a constant-in-time  surface temperature correction and xx_atemp_period = -12., means that the control variable is a monthly surface temperature correction that repeats itself after 12 months. Everything else remains the same.

Please adjust you experiments accordingly (change xx_${varname}_period = 0. to xx_${varname}_period = -12.), when you update to code later than checkpoint62u (and use "cal").

Martin