Ver 1.1 Bugs & issues

Discuss any other weather software not covered above, including data feeds, weather web pages, and more!
Post Reply
Craig Setzer
Member
Posts: 23
Joined: Thu Aug 05, 2004 5:13 am
Location: Miami, FL

Ver 1.1 Bugs & issues

Post by Craig Setzer »

Some of the things I've found with 1.1:

-NGRD2PCG32 has problems with NOGAPS grib files. For some reason after it builds the grid file and places it in GRIDDATA directory, even if the /NWS directory is empty, it still checks all the fields from the earlier version of the file in the GRIDDATA directory.

-PCGRIDDS32 has several issues with the ANIM command. First, if it builds a loop based on 3 hour interval data and loops that data, a switch to a different GRIB file with different forecast hour intervals results in the ANIM command still looking for 3 hourly interval data. The way to get around this is to exit PCGRIDDS32 and delete all the BMP files in the ANIMATE directory. A good fix would be to have PCGRIDDS32 reset the animation parameters when a new GRIB file is loaded.

Another issue with the ANIM command is the lack of an abort option when ANIM is loading parameters that are not found. For example if you run the command "bknt 105m clr1&anim" because there is no 105m wind, the ANIM command will keep looking throughout the whole forecast interval time series with no way to stop the process. This is not the case when data is found, the "S" key stops the loading.

-NGRB2PCG32, GFDL nested grib files create a unique grid project and file for each forecast interval. Because NGRB2PCG32 processes the grib files in the order they were saved on the computer, it's possible the 36 hour forecast and the 12 hour forecast be labled in or out of sequential order. This makes it quite challanging to load a GFDL file and find the forecast hour for that specific file. Also, this prevents the use of macros because the labeling is not time specific.

-NGRB2PCG32, the change to the naming of the files now overwrites the date digit. For example, if NGRB2PCG32 creates a file called OCT200412.AVN211 and then different data for the same model is processed, it gets the name OCT2A0412.AVN211 but you can't see if it's October 20th or 21st or 22nd, etc.

All this being said, this is a great program and I look forward to using it well into the future. Thanks Jeff for the hard work.

Craig Setzer
jkrob
Senior Member
Posts: 300
Joined: Fri Jun 25, 2004 11:42 pm
Contact:

Post by jkrob »

Craig,

I'll address these in order received...

NOGAPS Problems: this is a new on on me. If the NWS directory is empty, how can it be processing GRIB data? Yes, for every GRIB parameter it processes, it checks those which it has already processed to see if it has already processed the same parameter. I have not processed any NOGAPS in a while so there may be an issue here.

ANIM #1 : the routine to build the animation is supposed to check the number of forecast hours in the current file being processed. However...I just checked the code & I have it check that info the first time the routine is called & not every time. So, yes, this is a bug & will be fixed in the next release. Good catch!! Also, with the next release coming shortly, it will also deleate the files in the ANIMATE directory before it builds a new animation sequence.

ANIM #2 : Valid bug as well... when the routine does not find a requested grid parameter, it jumps to the portion of the code which is past the area which checks for the key-stroke to stop the animation build.

GFDL : This is not a bug...this is a feature...The "problem" is that the GFDL grid is a *mobile-domain* grid which moves with the storm it is forecasting. The NGRB2PCG32 concept is grid oriented and most all other grids are fixed-domain. So, as NGRB2PCG32 processes from one forecast hour to the next, it sees the grid has moved and so it says it needs to create an new output file. Sorry, such is life in PCGRIDDS32. Now, this is something which can be changed to remap and interpolate the GRIB data to a grid projection the user wants but it would require a whole restructuring of the ingest process. This is something I'm actually cosidering for PCGRIDDS32 farther down the road.

NGRB2PCG32 Name : this came as a request for 3 letter months vs. 2 letter months & is a valid bug which will be addressed.


WHEW!! I know I keep promising a new PCGRIDDS32 beta release "shortly", and I was real close to releasing it a week or two ago. However, a gentleman has come along & offered to help get printing going and I know how many have been asking for that feature, so I decided to delay the release to get it worked in. Things are real close and it will be soon...I promise!!! Also, it will allow me to integrate these fixes as well.

Thanks for your patience and contributions,

Jeff
kinetic
Member
Posts: 60
Joined: Sat Feb 07, 2004 3:04 pm

Post by kinetic »

Kind of new to the program, but could the fetch routine be incorporated directly into the program.
Post Reply