This change log starts from late August 2002.
HYBPOST - fixes to BUFR output, precip type.
Updated SST/LST file - was using 24 July data for last 3 weeks
HYBCST - (soil123.F, soilsnow.F, sfctmp.F, inithybv.F, POINT2D, SOILVEG) - fix for thin snow problem with near-surface temperatures
HYBPOST - more fixes to BUFR output.
HYBCST - Increase ice particle fall speed from formula which typically gave 10-25 cm/s up to 1 m/s. Done to decrease excessive cirrus cloud.
HYBCST - roughness length for deciduous and evergreen forest vegetation types decrease from 0.8-2.0 m down to maximum of 0.5 m. Done to improve light wind bias in RUC over northeastern US where these land-use types are dominant.
HYBPOST - Yesterday's change also needed in post-processing -- roughness length for deciduous and evergreen forest vegetation types decrease from 0.8-2.0 m down to maximum of 0.5 m. Done to improve light wind bias in RUC over northeastern US where these land-use types are dominant.
HYBCST - Switch to centered differencing for vertical
advection of moisture instead of upstream. Includes
stability check. Improves moisture conservation and
increases overall model precip 5-7% in winter and summer
case studies
Allow sub-grid-scale precipitation processes to co-exist
when resolved scale precipitation is forecast in a given
time step. Previously, this had not been allowed in an
attempt to force more resolved-scale precipitation, but with
the introduction of centered differencing in the vertical
moisture advection (above), the precip behavior now appears
more realistic to remove this previous constraint.
Implement land-surface model changes for thin snow cover,
which avoids cold spots. This is the same package implemented
into the operational RUC20 in early October 2002.