I download everything, I have a little bit problems with the folders, but Martynas help me with that. By now I can't make the run, cause the cluster is down for maintenance. As soon as I can, I'll let you now how is this working. Thanks for everything.

I have two little questions:

1) I'd like to plot the climate files that Laura gave me, I tried to open with my version of Grads1.9 (Config: v1.9b4 32-bit little-endian readline lats printim) with the command:

ga-> open cl00010000
Scanning description file: cl00010000
Open Error: Unknown keyword in description file
--> The invalid description file record is:
--> Q

And I have that mistake, is there anything special to load this data in Grads?

2) Does anybody know if this version of Hirlam works with the MATCH version 4.8.3-5 ? Or at least if you can give an email to contact somebody.

Best regards!

I'll post you and others a few simple scripts to plot and compare fields with grads. You need to create the ctl files first, you cannot open the grib file itself. I'll put a tarball at hirlam.fmi.fi/tmp, will let you know the file name soon.
Hope someone (from SMHI) will know about Match.

O.K., please find hirlam.fmi.fi/tmp/plotkit.tar.gz . There is all kinds of stuff, you could study the scripts with *kuva* to get an idea and be able to edit the scripts. I have just called e.g. the script Erokuvamf like this, to get the result shown below (these are our lake studies).
Erokuvamf an20091115_06+000TOTLAK an20091115_06+000ANALAK 194 105 902 mres_0_-30 194 105 902 196 105 0 0.25
I think I used Pikakuva to draw the orography for you domain earlier.

Well by now, I have working the Toolkit that Laura Gave me. I attached a plot of the total cloud cover from one of the experiment that we have with Martynas! Unfortunately, the final run I've been having problems, that we expect to have a solution during this week :). In that plot you can see the stratocumulus deck of the north part of Chile, which is typical. One more question about the Boundaries, is there any way to plot the .mars files? Just to have a look on them, try to identify if there are some fronts... anyway.

By the moment, thank you to all of you!! We'll keep in touch!

Felix,

The thing is that boundary files mainly contain data from Ewmwf model levels, (T,U,V,Q) and several extra fields Land-sea mask, Sea surface temperature and surface geopotential.
Just list their contents with wgrib or grib_ls which are both present in your system.

If you want to plot some of those things:
GrADS (which you used for ploting Hirlam data), it's also suitable for viewing the mars files. There are also other ways, but with GrADS
You would need a separate .ctl file for them, maybe some extra grads scripts. I can help you with that of course.

Thanks! I check with my wgrib on my personal computer (which is the place where I generate the plot) that the mars data contains exactly what you said, and as I mention in my email, it contains 60 vertical levels.


One more question that I have is about the OBS data, where it come from those observations? is there anyway to plot what it contains? (Grads??). I mean, as I imagine that model is making a data assimilation of this data, I'd like to know where it come from.

In other aspects, when I did the wgrib for the fc* data it gives me this:

felix@asml1-desktop:~/TESIS/HIRLAM/PlotKit_Rontu$ /home/felix/TESIS/HIRLAM/grads-1.9b4/bin/wgrib fc20051202_18+000

Undefined parameter table (center 96-0 table 1), using NCEP-opn ?????
1:12000:d=19010101:RI:kpds5=254:kpds6=109:kpds7=65535:TR=0:P1=1:P2=0:TimeU=1:hybrid lev 65535:1hr fcst:NAve=0
2:12600:d=05120218:GP:kpds5=6:kpds6=105:kpds7=0:TR=1:P1=0:P2=0:TimeU=1:0 m above gnd:anl:NAve=0
3:102840:d=05120218:PRES:kpds5=1:kpds6=105:kpds7=0:TR=1:P1=0:P2=0:TimeU=1:0 m above gnd:anl:NAve=0
4:212400:d=05120218:TMP:kpds5=11:kpds6=105:kpds7=998:TR=1:P1=0:P2=0:TimeU=1:998 m above gnd:anl:NAve=0
5:276960:d=05120218:SOILM:kpds5=86:kpds6=105:kpds7=998:TR=1:P1=0:P2=0:TimeU=1:998 m above gnd:anl:NAve=0


and the continues with all the variables? is there any problem? maybe I have to install an older version of GrADS (the version is grads-1.9b4)

Best regards!

Felix,

the observations come packed in Bufr, mainly from the station/soundings/etc... network available in your desired period in your domain (or maybe global, i did not check that to tell the truth)

Well wgrib gave exactly what it was supposed to give, the contents of your grib file (fc20051202_18+000). You could try grib_ls a/w or ad a wgrib -v option to see more detailed list.
No, there are no problems, as long as you're simply checking what is inside the files (parameters, levels etc.).
Wgrib is not for plotting your data, it just lists the general info about the fields.
Please read more at
www.cpc.ncep.noaa.gov/products/wesley/wgrib.html
or just try
wgrib
You can also have a look at grib api's grib_ls
www.ecmwf.int/publications/manuals/grib_api/grib_ls.html

If you want to plot the fields, you can use GrADS. I'll contact you about that,

Felix, if you have your experiment working and processing the boundries for the forecast run, it will start from the mars.* files and create after horizontal and vertical interpolation the next version, ma* or mf* files. These are normal HIRLAM gribs, which you can plot with the Pikakuva and others. But yes, grads allows you to plot the mars files directly if you know the ECMWF grib codes. There are a couple of routines in src/gcod, of which gbctb2.F contains a list of HIRLAM grib codes and someone else the correspondence to ECMWF grib code numbers.
Laura

Thanks Laura! By the way, do you know an easy way to plot time series with Hirlam data? maybe some code to extract a time series? Match has one called "metgram" that extract a time series at certain positions directly from the grib files.

Another thing is, you show me some cross section from your runs here in Chile, is there any way that you can facilitate me that code? I tried to load the whole Hirlam Grib file, but it failed because there are variables with differents grid size.

And as a final thing, I tried to load the two components of the wind at the first level from my outputs, but I couldn't, it says something abou the grid size, but I couldn't figure it out, any ideas about this?

Best regards.....

Hi,

Could you post the errors that you have received during plotting in a attached file?
At the moment it seems to me that you missed running gribmap on your ctl file.

Hi Felix,

only quickly, sorry, I do not have more time now to go deeper.

You could use metgram from SMHI, or then you need to do full gribmapping with grads and build your grads scripts.

Yes, my cross-section thing used the latter approach, I can send it but it needs some background work to be done for creation of gribmaps, and actually the experiment run with postprocessing to z-levels included.

I guess the wind problem is related to staggering, you could find somewhere about the Arakawa C grid which has mass points for variables like temperature and others and separate half-grid shifted u and v, to make advection calculation easier. Thus, u and v are in different grids which also differ from the mass-points. We have tools to take this into account when doing gribmapping for grads.

Thus, the main thing is correct gribmapping. I can address this also to our colleague Carl, who may have some modern tools for this.

Hi

I saw my name mentioned in a recent posting by Laura. If you are interested, I welcome you to try my attatched perl-script for mapping grib-files for use in grads. It does the following:
1) uses the HARMONIE utility gl to ecamine a given grib-file
2) creates a grads ctl-file referencing a set of given grib-files
3) performs the grads gribmap command
The script includes explanatory comments, but I am happy to answer questions.

If ytou still have an issue with the wind components, it hardly has anything to do with the staggering. Grads is not smart enough to know anything about staggering, but will happily use the grid specified in the ctl-file for all fields.

Carl

It's been a while... but for now Thank you Carl, I think that the gribmapper script works good for me. I just had to install the gl application but Martynas helped me with that and now it works ok!

By now I'm just having some "trouble" because I'd like to change the cycling interval from 6 to three hours. I have some clues from Xiao that it is possible , but we haven't discuss enough about cause he is travelling (as soon as he's back, we'll probably will touch this) anyway the idea was to post the things here in the forum, just in case this info could be useful for somebody else or if somebody has experience on this and can post it too. I'm trying to change this cause I'd like to have more than four points per each day just analyze and compare against Meteorological station over central Chile, jut to see how it works.

By the way, we've been discuss with my advisor about the convection schemes that Hirlam use. The thing is when they did some simulations on (1999) for a similar region, they shut down this schemes, cause any difference in temperature "unleash" convection, so they have problems with the convective precipitation. Any thoughts on this? Does the convection scheme have major improvements from that time to now? Maybe Laura Rontu, that did some modeling here in South America has some info about it.... anyway a description of those old runs you can find it here (page 36/115):

http://www.dim.uchile.cl/~lgallard/S_disp/HIRLAM-MATCH/tech_appendix_pdf.PDF

Best regards from the south!!!

Hi Felix,

why should you change the cycling interval? That will definitely be not that easy as you will need new observations or ready-made analyses. To compare with observations, you only need to change the output definitions, e.g. to hourly interval, or three-hourly for the beginning. Output interval is given in Env_domain WRITUPTIMES and WRITUPTIMESPP.
Then you can compare what you want, also count differences between accumulated variables for comparison of e.g. precipitation or radiation flux observations. This is what is done for the normal verification. By the way, you could use the normal verification package for your comparisons. You can download gl and monitor from hirlam.org repository from the harmonie trunk/util/...

Convrection schemes in HIRLAM have definitely changed since 1999. Now the default is something called Kain-Fritsch scheme + Rasch-Kristjansson microphysics. Another choice is the STRACO scheme, which contains both convection and cloud microphysics. It is in a way a direct successor of the Sundqvist scheme of 1999. You could try both options and compare the results in some cases where you expect problems. The choice is made in the script FCinput, in the array OPTIONS array. Also, you probably use better vertical resolution than in 1999, 60 levels or something I guess.This influences convection parametrisations Sorry for not reading your attachment to understand the details of your problem - it is a bit thick for my present time-schedule ...

One possible suggestion, by the way, might be to use Enviro-HIRLAM for your study. It can do both meteorology and air quality within an integrated framework. You could ask our Danish colleagues Alexander Mahura, This email address is being protected from spambots. You need JavaScript enabled to view it. or Alexander Baklanov, This email address is being protected from spambots. You need JavaScript enabled to view it. or Ulrik Korsholm, This email address is being protected from spambots. You need JavaScript enabled to view it. about this. However, I should perhaps have suggested this in the very beginning. Perhaps it is too late now to rethink about the whole philosophy of your study. It is more heavy to run Enviro-HIRLAM experiments than the regular ones, but then you would not need MATCH at all.

Just saw the last two postings on this thread. I agree with Laura on the preference of using fc...+003 than fc...+000 for non-synoptic hours. In my view, if the purpose is to compare model to observations every three, or even every 1 hour, it is not a good idea, given the current situation, to perform 3 hourly or one hourly analysis.

Currently for the historical episode we are unable to find higher frequency lateral boundary data from ECMWF than 6-hourly analyses. Accordingly, you have received 4 observation/day data package centered around synoptic hour (00/06/12/18 UTC). Changing cycling interval will require quite some additional experiment runs with need for numerous modifications of scripts in order to fetch available observation and lateral boundary data.

On top of these, you are not likely to get better model data if that's what you are interested. I mean,e.g.
fc2006091503+000
will unlikely be of higher quality than
fc2006091500+003

In the above example, obtaining first model data requires quite some extra effort, but it may not be of better quality than the second one, which you already have now. This is due to typical shortcomings in assimilation technique and in moisture spin-up. Furthermore, if the quantity you want to compare is precipitation or other things that are closely related to moisture process, I'd actually recommend use of forecasts with longer forecast lead time than 6 h.