precipitation, as accessed by OpenDAP, now is accumulative over
the entire forecast. Previously it was accumulation per timestep.
The GFS variable (
, surface total
precipitation [kg/m^2]) exists with the same identifier, but now
is in this different structure.
The fix is to
add <externUnit parameterId="PC.nwp"
unit="mm" cumulativeSum="true"/> to the
precipitation data import.
This change was
intentional but not very advertised. A notice can be found
here effective June 12:
II) New total
precipitation and convective precipitation, with continuous
accumulation, will be added to the output. Examples of the new
variables: APCP:surface:0-10 day acc fcst ACPCP:surface:0-10 day
It is confusing that
the same variable query from the NAM forecast returns the
accumulation per timestep.