8/18/2009

36, 12 and 4km control runs have been completed. Analysis graphics are online here.

I also have some verification plots in $control4km/plots/verification that I need to look at with Greg/Rahul. I also need to speak with Rahul regarding post_proc_renkf_*.m.

I have been working on getting radar data ready for assimilation. It must go through the following steps:
1) De-aliasing (unfolding), this has been done through soloii but will eventually need to find a completely automated way of doing this. Possibly with 4DD (James and Houze (2001)).
2) Clutter removal, Stacy has just finished a new version of this script and it will be tested in the next day or so.
3) Change the data into Cartesian coordinates and put into a NetCDF file with Reorder. I have the script and input file for this so once the clutter removal is completed I can move on to Reorder. I was originally unaware of this program and wrote my own 2-step matlab process to do this but it was VERY slow...
4) Quality Control, I will most likely be writing this code from scratch. It will go around a scan at a constant radius and remove rays that have failed to unfold.
5) Super-obing, Rahul has a method for this.

Then finally Rahul and I need to incorporate the radar forward operator into the EnKF and update any of Ryan's scripts for use with radar data.

2/02/2009

Process for Preparing Radar Data

Original Sweep File


1) Bargen-Brown Unfolding

• In the Edit Widget:
  • Commands for each ray:
    • duplicate VE in Vunfold
      
    • remove-aircraft-motion in Vunfold
      
    • BB-unfolding in Vunfold
    
    
  • One time only commands:
    • BB-gates-averaged is 4
      
    • BB-max-pos-folds is 1
      
    • BB-max-neg-folds is 1
      
    • BB-use-ac-wind

Apply to all swp files.

After Unfolding


2) Ground Clutter Removal

See unclutter_ppi.txt

After Uncluttering


3) Remove Zeros/-327.7's from Clutter Removal

• In the Edit Widget:
  • Commands for each ray:
    • set-bad-flags when CVE below -327
      
    • assert-bad-flags in CVE

After Thresholding (Final Image)

1/29/2009

Had a meeting with Socorro the other day to discuss radar data ground clutter removal. I will discuss this tomorrow in a meeting with Greg.

I now have a script to change the swp files to netcdf files.

Finished the Bargen-Brown unfolding all of the radar data I have available. Will be attempting tests of ground clutter removal.

4km domain is running finally, it is on cycle 33 of 38 then it will do the extended forecasts. It should be done in a week and a half or so.

1/22/2009

Still trying to get the 4km run to work, have had several computer issues. Currently working with David and Harry to resolve.

UPDATE: Should be resolved... changed sourcing on zion's queue system... now sources: system.bashrc

Analysis fields are plotted for the 36km & 12km domains. Debugging Verification scripts with the help of Rahul.

1/06/2009

Restarted 4km run on cycle 8 after finding out it shut down over winter break. It doesn't appear as though there was a problem but possible a power outage that stopped it. No signs of errors in any log files.

12/17/2008

Successfully completed 12km run today. I have started the 4km run.

Also have started producing graphics for the 36km and 12km runs. The 36km General Plots are completed but an error occured when producing the statistics. I have started the General Plots for the 12km and it should be done in the morning (with the same error as the 36km).

Note: xanim * will loop .png files.

12/12/2008

I believe that the 36km run is now 100% complete. Now I have restarted the 12km run from cycle 30 with run_free_wrf off. It should finish at cycle 39 over the weekend. Assuming everything goes well, the 4km should be started next week.

11/25/2008

After the 36km failed again Rahul pointed out that the wrfbdyts link (in the members/mean directory) was pointed to the wrond location because I had moved the experiment files. I have started the 36km again...

11/21/2008

Submitted Masters Thesis Proposal.

Found out that the EnKF requires one extra cycle on the parent domain to execute the nested domain. so instead of 30 cycle for all, the 36km domain would have 32 cycles, the 12km would have 31 cycles, and the 4km would have 30 cycles. I have alerted Ryan to this.

I am extending the 36km domain to include the second 24 hour period + 2 extra cycles (for a total of 38 cycles). The I will run the 12km with 37 cycles and the 4km with 36 cycles. There should be 2 24-hr extended forecasts starting a 00Z on the 4th and 00Z on the 5th.

Command for 36km restart: WRF_ENKF_SCRIPT_control.bash /home/disk/zion/rwolcott/wrf_enkf/Version_2.2_control/ > & ~rwolcott/log/control_restart30_38.log &

11/13/2008

Stopped again at the end of cycle 30, before extended forecasts were made. Rahul is going to help me find the error and fix the problem.