Hi Kris,
Upon further examination, this behavior can be traced back to two
observations being assimilated in URMA. Please see the attached
slide deck for our findings:
First, recall that the maxT background for URMA is derived by
taking the maximum of the hourly background and analysis fields
valid 7A-7P LST. For June 13th, this corresponds to 14Z June 13th
through 02Z June 14th. The maxT background and analysis for this
particular case are shown on slide 3. The pixel in question has a
background value between 116 and 120F, which is cooled only slightly
in the analysis to 112 to 116F.
A review of the hourly temperature background and analysis fields
traced this hot pixel to the 22Z analysis (see slide 7).
A bias correction technique is applied to the first guess fields
for the URMA 2-m temperature analyses. The bias correction figure
valid 22Z June 13th is shown on slide 10. There is a positive
adjustment of ~7-8 degrees Celsius at the pixel in question. It
appears that two observations are responsible for this behavior.
Slides 11-12 show the KML file output for the same cycle. Note: The
"BG VALUE" listed in the screenshots is valid
after the bias correction has been applied, so these
observations were having a considerable impact on the analyses.
You can view the recent evolution of the bias correction valid at
22Z on slides 13-17. Since flagging KMPC1 on June 14th, it appears
that the situation has improved somewhat, as the station's influence
on the bias correction will decrease over time. However, there is
likely still some influence from the nearby station (EKIC1). With
your approval, we would like to add this station to the reject list
for temperature observations.
Thank you for bringing this matter to our attention, and please
let us know if you have any questions/comments.
Thanks,
Matt
Hi Kris,
The requested stations have been added to the SDM reject
list, effective with today's 12Z URMA.
The NBM bias corrects directly to select observations (e.g.,
METAR, RAWS) at the corresponding grid point. This process is
likely responsible for the hot pixels in the attached figure. I
would suggest reaching out to the NBM team for assistance with
identifying the problematic station(s). We can then have them
removed on our end, which should also remove them from the bias correction.
TSHC1 was added to the reject list in July 2019. Based on
this, the station should not be used in the NBM bias correction.
Thanks,
Matt
Hello,
I just submitted several stations to be removed from the
WFO HNX office.
However, I'm not sure if that will cure the problem.
Look at the MaxT image below for today. These spots on
our grid should be upper 90s at best, they are about 15F
too warm. This is also polluting the HeatRisk. I
believe that station TSHC1 (38.49, -118.83) is leading
to this horrible issue but I'm not sure. We've had
issues with this station TSHC1 (Ash Mountain) before, and it
is in the list of blacklisted stations. Could you
provide any further insight?
Thank you.
--
Kris Mattarochia (he/him/his)
Science and Operations Officer
National Weather Service San
Joaquin Valley
- HNX
Hanford, CA
(559) 584-0583 x224
--
Kristian Mattarochia RTMA/URMA Discussion Group
Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/37879598VLab.Notifications@noaa.gov
--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758
--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758
