<br><font size=2 face="sans-serif">Don,</font>
<br>
<br><font size=2 face="sans-serif">Even thought the observation might be
read in it may not be used. If the observation is 5 standard deviations
away from the background (for surface temperature I believe the error standard
deviation is 2 degrees C). Also if the model terrain is significantly
different that the actual terrain the observation will not be used. One
quick way to tell if your ob was used is to look at the fort.12 file. This
file has the fit of the background field to the observations (by observation
type) and the fit of the resulting analysis to the observations. If
3DVAR worked then you should see a better fit with the analysis than the
background. Also the fort.60 file will tell you which obs were rejected
by 3DVAR . </font>
<br><font size=2 face="sans-serif">Hope this helps.</font>
<br>
<br><font size=2 face="sans-serif">Mike</font>
<br><font size=2 face="sans-serif">___________________________<br>
Michael D. McAtee<br>
The Aerospace Corporation<br>
106 Peacekeeper Dr. STE 2N3<br>
Offutt, AFB, NE 68113-4039<br>
email: michael.mcatee@aero.org<br>
Phone: (402) 292-1017<br>
Fax: (402) 291-3137</font>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td width=40%><font size=1 face="sans-serif"><b>"Don Morton"
<Don.Morton@umontana.edu></b> </font>
<br><font size=1 face="sans-serif">Sent by: wrf-users-bounces@ucar.edu</font>
<p><font size=1 face="sans-serif">01/01/2007 06:49 PM</font>
<td width=59%>
<table width=100%>
<tr>
<td>
<div align=right><font size=1 face="sans-serif">To</font></div>
<td valign=top><font size=1 face="sans-serif">wrf-users@ucar.edu</font>
<tr>
<td>
<div align=right><font size=1 face="sans-serif">cc</font></div>
<td valign=top>
<tr>
<td>
<div align=right><font size=1 face="sans-serif">Subject</font></div>
<td valign=top><font size=1 face="sans-serif">[Wrf-users] Seeking WRF-Var
Guidance</font></table>
<br>
<table>
<tr valign=top>
<td>
<td></table>
<br></table>
<br>
<br>
<br><font size=2><tt>Howdy, now that it's "vacation" time again
I'm back to<br>
trying to muddle my way through WRF-Var. I feel like I<br>
may have my procedures down correctly, but I don't have<br>
the practical experience to know if my results are reasonable<br>
or not.<br>
<br>
To date, I have gone through the WRF-Var tutorial successfully<br>
and understand the basic procedures. In order to gain a better<br>
understanding of all this, my efforts are now focused on assimilating<br>
a single surface observation on a 50km x 50km domain centered on<br>
Deadhorse, Alaska. The domain is 1km resolution with 31 levels.<br>
<br>
My intent is to ingest a single surface temperature obs at lat/lon<br>
70.20/-148.47 (Deadhorse). I wrote up a Fortran program that<br>
will take a netCDF wrfout (or wrfinput) filename, lat, and lon as<br>
arguments, find the four surrounding grid points, then do a bilinear<br>
interpolation to estimate a "point" value. For this particular
location<br>
I find that the initial temperature in my wrfinput_d01 file is -23.8C.<br>
<br>
So, I've created a little_r obs file (thank you Mike McAtee and Xuanli
for<br>
responding to my post a couple of months ago about how to get my<br>
surface obs into little_r format) that presents a perturbed temperature<br>
at this point (I've tried several values), then gone through the assimilation<br>
process. In all cases, when I compare the initial temperature in
the<br>
original wrfinput_d01 with the one in the wrfvar output file, they're<br>
identical out to 7 significant digits. An excerpt of the wrf-var
output<br>
from using an obs temperature of 8K higher than the initial temperature<br>
from wrfinput_d01 is:<br>
<br>
----------------------------------------------------------<br>
Inner iteration stopped after 1 iterations<br>
Final cost function : 7.76275575D+00 and Gradient: 7.96228229D-14<br>
----------------------------------------------------------<br>
<br>
Final cost function J = 7.762755752421293<br>
<br>
Total number of obs. =
1<br>
Final 3DVAR value of J =
7.76276<br>
Final 3DVAR value of Jo = 7.53804<br>
Final 3DVAR value of Jb = 0.22471<br>
Final J / total num_obs = 7.76276<br>
3DVAR Jb factor used (1)= 1.00000<br>
3DVAR Jb factor used (2)= 1.00000<br>
3DVAR Jb factor used (3)= 1.00000<br>
3DVAR Jb factor used (4)= 1.00000<br>
3DVAR Jb factor used (5)= 1.00000<br>
<br>
I've also tried perturbations of 4K, 23K, etc. It seems that<br>
if my perturbed obs is too large (e.g. greater than 12K), the<br>
obs isn't even used.<br>
<br>
Now I'm stuck. I have a number of theories, but I'd like to avoid<br>
too much blind trial and error. What I'd like to do is introduce
an<br>
observation (or set of observations) that will actually result in my<br>
wrf-var output showing an initial temperature that's different from<br>
the one in my original wrfinput_d01.<br>
<br>
I'm using the default background error files that came with<br>
WRF-Var. I do realize that I need to modify these at some<br>
point, but right now I'm just on an incremental learning curve<br>
and I'm looking for qualitative differences.<br>
<br>
I'm guessing that maybe introducing a single obs into this domain<br>
just isn't "enough" for WRF-Var to assimilate it, and the gradient<br>
I see in the previous example sort of strengthens that theory. So,
my<br>
question is, given my 50x50 1km-resolution domain, is it reasonable<br>
to expect WRF-Var (with the default background error files) to produce<br>
some perturbed initial temperatures near Deadhorse, based on a single<br>
surface temperature obs, or am I asking for too much?<br>
<br>
If some kind and patient soul would be willing to help guide me through<br>
this process, I'd be very appreciative, and I'd be very happy to write
up<br>
a tutorial - sort of a Dummy's Guide on how - after you've gone through<br>
the WRF-Var Tutorial - you'd go about setting up your own simple domain<br>
with some basic obs for assimilation. In my opinion, the current
WRF-Var<br>
tutorial is great for understanding some of the basic procedures, but it's<br>
a huge jump from there to ingesting one's own obs and getting them<br>
to assimilate effectively into one's own domain. I'd like to think
that<br>
this Dummy's Guide would help to bridge that.<br>
<br>
I've gone through an AMS paper and Tech Report by Barker, et. al on<br>
"A Three-Dimensional Variational Data Assimilation System for MM5,"<br>
but estimate I need to go through these another 758 times before I<br>
understand most of the details. I have some background in Finite
Element<br>
Methods, so the idea of variational methods isn't totally foreign. At
this<br>
point, I "think" what I need is access to some practical expertise.
At this<br>
point, blindly trying lots of different things doesn't seem very productive,
and<br>
I'm not sure where else to turn.<br>
<br>
Also, I'm getting the impression that WRF-Var is pretty much an extension<br>
of the MM5 data assimilation code. Does anybody know of "Dummy
Guide"<br>
types of references for MM5 that might help me?<br>
<br>
Thanks for your patience and time on reading what I believe may be<br>
the first wrf-users post of 2007!<br>
<br>
Don Morton<br>
<br>
-- <br>
Don Morton<br>
Department of Computer Science<br>
The University of Montana - Missoula<br>
http://www.cs.umt.edu/~morton/<br>
_______________________________________________<br>
Wrf-users mailing list<br>
Wrf-users@ucar.edu<br>
http://mailman.ucar.edu/mailman/listinfo/wrf-users<br>
</tt></font>
<br>