Caution: more on using different landuse data sets

Dataset specific topics involving WPS.

Caution: more on using different landuse data sets

Postby cthomas » Thu Mar 12, 2009 2:08 am

Hi,

I thought I had figured out all the things required to use a different landuse dataset (see earlier threads "Documentation for LANDUSE.TBL?" and "Using custom topo and landuse data" in this forum). However:

If I follow the procedures for producing the datasets and index files and modifying LANDUSE.TBL, as detailed in ARWUsersGuideV3.pdf and the information provided by wrfhelp in the above threads, everything seems to work great as far as WPS goes, and both the geo_em and met_em files reflect the new landuse categories and contain a correct landmask (I can check this, for example, by using ncview).

BUT:

When I use real.exe to produce a wrfinput file, I find that the landuse and landmask in wrfinput is rubbish. This is a rather insidious problem because many people might, as I did, check their landuse and landmask in the geo_em or met_em files and just assume that this is carried over correctly to the wrfinput file. It appears to me that it is not. To fix this I had to set the namelist variable num_land_cat in the physics section of namelist.input to my new value (which happens to be 13). The default value of this namelist variable appears to be 24. If you have never heard of this namelist variable, don't worry: it does not seem to be mentioned anywhere in ARWUsersGuideV3.pdf and does not appear in the sample namelist.input file distributed with WRF. I could only discover it by examining the code.

The number of categories in the landuse data is contained in both the geo_em files (dimension land_cat) and the met_em files (it is the size of the second dimension of the LANDUSEF variable). Real.exe should be able to read it from one of these files. I can't see why it needs to be stipulated again in the namelist.input. However, as far as I can see, it does.

At the very least this needs to be documented, because anyone trying to use a different landuse dataset will potentially have a problem.

Am I missing something basic here? Has anyone else had similar experience? Can wrfhelp confirm that num_land_cat must be set in namelist.input (unless you are using the default value of 24)?

Regards

Chris
cthomas
 
Posts: 14
Joined: Tue Sep 30, 2008 7:14 pm

Re: Caution: more on using different landuse data sets

Postby pcampra » Wed Jan 26, 2011 8:04 pm

Well you are right. I am having lots of problems just trying to enter modis data in wrf. Some problems are:

1. Entering the variable num_land_cat to run real.exe. No matter is I change the namelist.input, an error shows saying NUM_LAND_CAT is set to 24, and I can not find were

2. In geogrid, after changing the namelist.wps, the modis land use categoris are loaded, but their albedo and many other variables dont show at 30s resolution

So you are right, many problems changing land use database in wrf

Pablo Campra
Berkeley LBNL
pcampra
 
Posts: 2
Joined: Wed Jan 26, 2011 7:56 pm

Re: Caution: more on using different landuse data sets

Postby pcampra » Sat Mar 12, 2011 12:21 pm

Additional problems when changing to Modis land use and running real.exe and wrf.exe.in WRFV321

No problems arise when running WPS. Met_em* files seem to be ok, with MODIS categories correctly changed from default USGS. I have entered num_land_use_cat = 20 in the namelist.input file.

But when running wrf.exe a msj error such as this shows:
INPUT LandUse = "MODIFIED_IGBP_MODIS_NOAH"
PGFIO-F-217/list-directed read/unit=29/attempt to read past end of file.
File name = LANDUSE.TBL formatted, sequential access record = 197
In source file module_physics_init.f90, at line number 632

I have checked the wrfinput files generated by real.exe, and there is something wrong with them, as I cannot open the LU_INDEX variable with NCVIEW. I get this error in ncview: min and max both 0 for variable LU_INDEX.

Any help to use MODIS land use data when runninf wrf???
pcampra
 
Posts: 2
Joined: Wed Jan 26, 2011 7:56 pm

Re: Caution: more on using different landuse data sets

Postby mwong » Wed Apr 13, 2011 4:25 pm

Hi!

Have you found a fix for your problem yet? I've just started using MODIS land use data for my runs (I've been testing out the new V3.3 so I decided I might as well look into the MODIS land use dataset).

Regarding your problem, pcampra, I think I encountered a similar one (although my LU_INDEX field looks fine in my wrf_input files, so you may be having some other problems as well). But regarding the error message, one thing I checked was whether the "name" of my land use data set matched those in the files LANDUSE.TBL and VEGPARM.TBL. It turns out mine didn't. For example, my MMINLU is defined to be "MODIS" when I use the MODIS data set, but in my .TBL files, they're called "MODIFIED_IGBP_MODIS_NOAH". My wrf.exe runs fine after I've changed it so the "name" of the dataset in the .TBL files are the same as MMINLU = "MODIS".

I'm not sure if this is the correct fix for this, but it seems to have worked for me.

Cheers,
-May
mwong
 
Posts: 5
Joined: Wed Jan 07, 2009 8:45 pm

Re: Caution: more on using different landuse data sets

Postby c2happle » Sat Jul 07, 2012 4:57 am

Hello, I a new one here.
I face some similar questions as above when I use MODIS as landuse data.
When I run real.exe, eveything is fine, and its output 'wrfinput.d01.nc' contains right information 'MMINLU = "MODIFIED_IGBP_MODIS_NOAH" ;NUM_LAND_CAT = 20 ;'
But when running wrf.exe, the error such as this shows:

*************************************
DYNAMICS OPTION: Eulerian Mass Coordinate
med_initialdata_input: calling input_model_input
INPUT LandUse = "MODIFIED_IGBP_MODIS_NOAH"
PGFIO-F-231/list-directed read/unit=29/error on data conversion.
File name = LANDUSE.TBL formatted, sequential access record = 118
In source file module_physics_init.f90, at line number 565


So how you solved your problem, and thank you very much for your knid suggestion.




pcampra wrote:Additional problems when changing to Modis land use and running real.exe and wrf.exe.in WRFV321

No problems arise when running WPS. Met_em* files seem to be ok, with MODIS categories correctly changed from default USGS. I have entered num_land_use_cat = 20 in the namelist.input file.

But when running wrf.exe a msj error such as this shows:
INPUT LandUse = "MODIFIED_IGBP_MODIS_NOAH"
PGFIO-F-217/list-directed read/unit=29/attempt to read past end of file.
File name = LANDUSE.TBL formatted, sequential access record = 197
In source file module_physics_init.f90, at line number 632

I have checked the wrfinput files generated by real.exe, and there is something wrong with them, as I cannot open the LU_INDEX variable with NCVIEW. I get this error in ncview: min and max both 0 for variable LU_INDEX.

Any help to use MODIS land use data when runninf wrf???
c2happle
 
Posts: 1
Joined: Sat Jul 07, 2012 4:49 am


Return to Working with Various Datasets

Who is online

Users browsing this forum: Google [Bot] and 1 guest