A few more fixes today:

The 00z and 12z RUC2 model runs crashed because we only had 1.e-7 precision for water vapor mixing ratio (g/g) in the GRIB table and the actual values at the top level on the north boundary, due to the very cold temps there, finally went below 0.5e-7.

This was rounded off in the GRIB to zero, so the fcst model choked on that. This has now been fixed and even bullet-proofed in another location.

Now for the boring changes:
surface runoff added to isobaric main output file
surface relative humidity added to isobaric main output file
sfc temp, dew point, rh, specific humidity now described as 2 m AGL
sfc winds now described as 10 m AGL

Stan