-----------------------------------------------------
Update: April, 2016 V3.8.0
A new namelist option (rinblw)is added to &record4.  
The default radius of influence used to calculate the observation density weighting factor in the WRF model (rinblw) is rather large (250km). 
This parameter is used by the OBSGRID program to calculate the distance from each grid cell to the nearest observation, information which is then used by WRF to calculate the observation weighting factor. This variable can be adjusted by the user in two ways. The first is to use the latest version of the OBSGRID program released along with FASDAS that includes rinblw as a namelist variable (as shown in the example below). 
It is recommended to use a rinblw value of roughly 2Δx (the example below is for a 12-km grid). The second method would be to start with an older version of OBSGRID and manually change the rinblw value from the default 250.0 to the desired value in the ob_density subroutine contained within the qc0_module.F90 code prior to compiling OBSGRID. This requires replacing the 250.0 values in lines 60-63 and line 67 in qc0_module.F90 with the desired rinblw value and recompiling the OBSGRID code.

-----------------------------------------------------
Update: January, 2016 V3.8.0
 - Optimize objective analysis and time sorting of obs
 - Add option to scale decreases in RH introduced via the analysis [see option
   scale_cressman_rh_decreases] (Reen et al. 2016; 
   http://dx.doi.org/10.1175/JAMC-D-14-0301.1)
 - Surface pressure improvements (improve surface pressure estimation of obs,
   add ability to QC surface pressure obs, add ability to create surface
   pressure objective analysis) [see options qc_psfc and oa_psfc] (Reen 2015; 
   http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA621305)
 - Improve "duplicate" obs merging (prefer ob closest to target time and time of
   merged ob is now time of ob closer to target date)   
 - New littler output files (qc_obs_used_earth_relative*) analogous to
   qc_obs_used* but with earth-relative winds rather than grid-relative winds.
   These can be used as input to the Model Evaluation Tools (MET)
 - Allow smaller radii of influence for surface obs [see setting
   radius_influence_sfc_mult]
 - Single-level above-surface observations (e.g. aircraft obs) can be QC'd 
   against nearby pressure levels in the first guess [see settings 
   use_p_tolerence_one_lev, max_p_tolerance_one_lev_qc, and 
   max_p_tolerance_one_lev_oa]
 - Allow quality control of dewpoint [see settings max_error_dewpoint and 
   max_buddy_dewpoint]
 - Improve consistency of littler output files (make dewpoint consistent with
   temperature and humidity, recalculate wind direction after rotation and
   vertical interpolation of u/v)
 - Allow user to choose to do no analysis (specify analysis type "None")
 - Minor bug fixes 
-----------------------------------------------------
Update: July, 2014 V3.7.0
 - Fix error that caused SLP to be erroneously marked as missing in some cases
 - Add NetCDF version of qc_obs* files
 - Modify compilation to deal with NetCDF changes (add -lnetcdff)
-----------------------------------------------------
Update: April, 2013  V3.4.0
 - Add support for NCAR/MMM observation files which contain unusable
 data that must be ignored (e.g. GPSPW, SATEM, etc.).
 - Allow obs plotting on nested domains.
 - Clean up typos and comments.
-----------------------------------------------------
Update: December, 2012 V3.3.0
 - Add some significant changes to the way the OBS_DOMAINdxx files are written.
 - Add a new qc flag or data that could not be qc-ed due to the fact that the 
   data is off analysis times. The non-qc-ed data can be output if needed.
 - Add some fixes from Brian Reed.
-----------------------------------------------------
Update: April, 2010 V3.2.0
 - Increase search area around each observation, ensuring the banana scheme
   has enough space.
 - minor bug fixes
