[Wrf-users] NMM binary cloud cover percentage

[Patrick Reuter]

Dear WRF users,

I succesfully run NMM forecasts with the WPP postprocessor for central  
europe, and I am pretty happy about the wind output results compared  
to the real-world.

However, I have some concerns using the cloud cover output.

Actually, it appears to output very often 0% cloud cover or 100% cloud  
cover, and very rarely something in between (actually it does it only  
at the spatial borders of some cloudy regions). As a consequence, the  
cloud forecast is not reliable at all (compared to directly usin GFS,  
for example), and even the 0% or 100% are not reliable, sometimes it  
shows 100% although it's sunny and vice versa.

This happens for all four types of cloud covers, low, middle, high,  
and total percentage.

For the physics, I use the following entries in the physics section  
(pretty standard I guess) :

  &physics
  mp_physics                          = 5,        5, 5,
  ra_lw_physics                       = 99,       99, 99,
  ra_sw_physics                       = 99,       99, 99,
  nrads                               = 105,      315, 315,
  nradl                               = 105,      315, 315,
  co2tf                               = 1,
  sf_sfclay_physics                   = 2,        2, 2,
  sf_surface_physics                  = 99,       99, 99,
  bl_pbl_physics                      = 2,        2, 2,
  nphs                                = 6,        18, 18,
  cu_physics                          = 2,
  ncnvc                               = 6,
  tprec                               = 1,
  theat                               = 6,
  tclod                               = 1,


My input data is 6 hourly GFS 0.5 degree data (54km), and I run a  
nested grid at 15 and then 5 km horizontal resolution, on a  
multiprocessor machine.

I would appreciate any comment on this topic.

Thanks in advance
   Patrick