Forums

Back

Urban calm bias

RSS (Opens New Window)(Opens New Window)
MD
Matthew Day, modified 7 Years ago.

Urban calm bias

Youngling Posts: 9 Join Date: 4/26/15 Recent Posts

One of the forecasters here pointed out a calm bias in the URMA wind fields over Oklahoma City. I read in another thread that the HRRRv2 had some friction issues which may be contributing to this, but that doesn't look like it can account for all of it. When you compare the URMA final result to either the URMA first guess or the HRRR field, the calm bias is much more pronounced in the final version. Any idea what could be going on?

 

The images are from April 3, 18z. The GFE screencapture is our interpolated Obs grid from that same time.

Attachments:

Attachment

Attachment

Attachment

JC
Jacob Carley, modified 7 Years ago.

RE: Urban calm bias

Youngling Posts: 69 Join Date: 12/17/14 Recent Posts
Hi Matthew,

What you're most likely seeing is the impact from nearby mesonet stations overwhelming the influence of the typically more windy METAR sites.  Mesonet winds are typically more dense near urban areas (note I'm *not* referring to the OK mesonet here, more along the lines of urbanet and CWOP which may not be at 10m AGL, could be somewhat obstructed, etc.).  This topic has been brought up a few times on the forum before [see some older threads earlier on this year].  We are working on addressing this via a variety of pathways: improved obs QC (some coming in the v2.7 package we are finalizing now), an ongoing effort toward improving mesonet metadata so we can assimilate the winds more effectively/appropriately, analyzing stats stratified by urban and rural areas, etc.

For more information, see the the following slide deck that was presented at last week's Model Evaluation Group meeting regarding the v2.7 upgrade package:


Slides 18-21 address the topic in question.

Hope this helps,
Jacob



On Fri, Apr 20, 2018 at 3:08 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:

One of the forecasters here pointed out a calm bias in the URMA wind fields over Oklahoma City. I read in another thread that the HRRRv2 had some friction issues which may be contributing to this, but that doesn't look like it can account for all of it. When you compare the URMA final result to either the URMA first guess or the HRRR field, the calm bias is much more pronounced in the final version. Any idea what could be going on?

 

The images are from April 3, 18z. The GFE screencapture is our interpolated Obs grid from that same time.


--
Matthew Day RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/3929302 VLab.Notifications@noaa.gov

Jeffrey Craven, modified 7 Years ago.

RE: Urban calm bias

Youngling Posts: 90 Join Date: 9/24/12 Recent Posts
I zoomed in and if you select the URMA observation option from the "+" menu that I highlighted, you'll see several sub 10 knot observations in the north metro.   In these areas the first guess field was higher and was pulled down by the observations.  

For example, the 6.1 knot observation to the north had an URMA FG of 22 knots, and the final URMA analysis was 13 knots.




Jeff Craven
Chief, Statistical Modeling Branch
National Weather Service, W/STI-12
Meteorological Development Laboratory (MDL)
Room 10410, SSMC2
Silver Spring, MD 20910
(301) 427-9475 office
(816) 506-9783 cell/text
@jpcstorm

On Fri, Apr 20, 2018 at 3:24 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Hi Matthew,

What you're most likely seeing is the impact from nearby mesonet stations overwhelming the influence of the typically more windy METAR sites.  Mesonet winds are typically more dense near urban areas (note I'm *not* referring to the OK mesonet here, more along the lines of urbanet and CWOP which may not be at 10m AGL, could be somewhat obstructed, etc.).  This topic has been brought up a few times on the forum before [see some older threads earlier on this year].  We are working on addressing this via a variety of pathways: improved obs QC (some coming in the v2.7 package we are finalizing now), an ongoing effort toward improving mesonet metadata so we can assimilate the winds more effectively/appropriately, analyzing stats stratified by urban and rural areas, etc.

For more information, see the the following slide deck that was presented at last week's Model Evaluation Group meeting regarding the v2.7 upgrade package:


Slides 18-21 address the topic in question.

Hope this helps,
Jacob



On Fri, Apr 20, 2018 at 3:08 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:

One of the forecasters here pointed out a calm bias in the URMA wind fields over Oklahoma City. I read in another thread that the HRRRv2 had some friction issues which may be contributing to this, but that doesn't look like it can account for all of it. When you compare the URMA final result to either the URMA first guess or the HRRR field, the calm bias is much more pronounced in the final version. Any idea what could be going on?

 

The images are from April 3, 18z. The GFE screencapture is our interpolated Obs grid from that same time.


--
Matthew Day RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/3929302VLab.Notifications@noaa.gov


--
Jacob Carley RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov:8080/web/715073/home/-/message_boards/view_message/3929498VLab.Notifications@noaa.gov

Attachments:

Attachment

MD
Matthew Day, modified 7 Years ago.

RE: Urban calm bias

Youngling Posts: 9 Join Date: 4/26/15 Recent Posts
That makes sense. Thanks for the response.

--Matthew

On Fri, Apr 20, 2018 at 2:47 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
I zoomed in and if you select the URMA observation option from the "+" menu that I highlighted, you'll see several sub 10 knot observations in the north metro.   In these areas the first guess field was higher and was pulled down by the observations.  

For example, the 6.1 knot observation to the north had an URMA FG of 22 knots, and the final URMA analysis was 13 knots.




Jeff Craven
Chief, Statistical Modeling Branch
National Weather Service, W/STI-12
Meteorological Development Laboratory (MDL)
Room 10410, SSMC2
Silver Spring, MD 20910
(301) 427-9475 office
(816) 506-9783 cell/text
@jpcstorm

On Fri, Apr 20, 2018 at 3:24 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Hi Matthew,

What you're most likely seeing is the impact from nearby mesonet stations overwhelming the influence of the typically more windy METAR sites.  Mesonet winds are typically more dense near urban areas (note I'm *not* referring to the OK mesonet here, more along the lines of urbanet and CWOP which may not be at 10m AGL, could be somewhat obstructed, etc.).  This topic has been brought up a few times on the forum before [see some older threads earlier on this year].  We are working on addressing this via a variety of pathways: improved obs QC (some coming in the v2.7 package we are finalizing now), an ongoing effort toward improving mesonet metadata so we can assimilate the winds more effectively/appropriately, analyzing stats stratified by urban and rural areas, etc.

For more information, see the the following slide deck that was presented at last week's Model Evaluation Group meeting regarding the v2.7 upgrade package:


Slides 18-21 address the topic in question.

Hope this helps,
Jacob



On Fri, Apr 20, 2018 at 3:08 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:

One of the forecasters here pointed out a calm bias in the URMA wind fields over Oklahoma City. I read in another thread that the HRRRv2 had some friction issues which may be contributing to this, but that doesn't look like it can account for all of it. When you compare the URMA final result to either the URMA first guess or the HRRR field, the calm bias is much more pronounced in the final version. Any idea what could be going on?

 

The images are from April 3, 18z. The GFE screencapture is our interpolated Obs grid from that same time.


--
Matthew Day RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/3929302VLab.Notifications@noaa.gov


--
Jacob Carley RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov:8080/web/715073/home/-/message_boards/view_message/3929498VLab.Notifications@noaa.gov


--
Jeffrey Craven RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov:8080/web/715073/home/-/message_boards/view_message/3929691VLab.Notifications@noaa.gov

Attachments:

Attachment

MD
Matthew Day, modified 7 Years ago.

RE: Urban calm bias

Youngling Posts: 9 Join Date: 4/26/15 Recent Posts
I've got a couple of follow-up questions on the use of mesonet winds in URMA. After looking through several of the threads on the VLAB, I read couple times that these mesonet stations are supposedly rejected by default. If that's the case, what caused several of them to be added back into the analysis for Oklahoma City? Also, in digging into the data, I noticed there were no Oklahoma Mesonet stations listed (though another thread mentioned that these should be on the allowed list). I was wondering why this might be?

--Matthew

On Fri, Apr 20, 2018 at 4:38 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
That makes sense. Thanks for the response.

--Matthew

On Fri, Apr 20, 2018 at 2:47 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
I zoomed in and if you select the URMA observation option from the "+" menu that I highlighted, you'll see several sub 10 knot observations in the north metro.   In these areas the first guess field was higher and was pulled down by the observations.  

For example, the 6.1 knot observation to the north had an URMA FG of 22 knots, and the final URMA analysis was 13 knots.




Jeff Craven
Chief, Statistical Modeling Branch
National Weather Service, W/STI-12
Meteorological Development Laboratory (MDL)
Room 10410, SSMC2
Silver Spring, MD 20910
(301) 427-9475 office
(816) 506-9783 cell/text
@jpcstorm

On Fri, Apr 20, 2018 at 3:24 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Hi Matthew,

What you're most likely seeing is the impact from nearby mesonet stations overwhelming the influence of the typically more windy METAR sites.  Mesonet winds are typically more dense near urban areas (note I'm *not* referring to the OK mesonet here, more along the lines of urbanet and CWOP which may not be at 10m AGL, could be somewhat obstructed, etc.).  This topic has been brought up a few times on the forum before [see some older threads earlier on this year].  We are working on addressing this via a variety of pathways: improved obs QC (some coming in the v2.7 package we are finalizing now), an ongoing effort toward improving mesonet metadata so we can assimilate the winds more effectively/appropriately, analyzing stats stratified by urban and rural areas, etc.

For more information, see the the following slide deck that was presented at last week's Model Evaluation Group meeting regarding the v2.7 upgrade package:


Slides 18-21 address the topic in question.

Hope this helps,
Jacob



On Fri, Apr 20, 2018 at 3:08 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:

One of the forecasters here pointed out a calm bias in the URMA wind fields over Oklahoma City. I read in another thread that the HRRRv2 had some friction issues which may be contributing to this, but that doesn't look like it can account for all of it. When you compare the URMA final result to either the URMA first guess or the HRRR field, the calm bias is much more pronounced in the final version. Any idea what could be going on?

 

The images are from April 3, 18z. The GFE screencapture is our interpolated Obs grid from that same time.


--
Matthew Day RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/3929302VLab.Notifications@noaa.gov


--
Jacob Carley RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov:8080/web/715073/home/-/message_boards/view_message/3929498VLab.Notifications@noaa.gov


--
Jeffrey Craven RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov:8080/web/715073/home/-/message_boards/view_message/3929691VLab.Notifications@noaa.gov


--
Matthew Day RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov:8080/web/715073/home/-/message_boards/view_message/3930665VLab.Notifications@noaa.gov

Attachments:

Attachment

JC
Jacob Carley, modified 7 Years ago.

RE: Urban calm bias

Youngling Posts: 69 Join Date: 12/17/14 Recent Posts

Hi Matthew,

 

We do use OK Mesonet winds, but they're restricted data and not available to the public.  So we can't share those obs online or on public pages (and hence why you can't see them).

 

You are correct, mesonet winds are not used by default in the RTMA/URMA.  Instead we've developed provider and station-based use-lists.  The lists were created leveraging available observation network metadata as well as stats related to a particular station.  We are updating this for the upcoming v2.7 implementation, but in the meantime you can read more about the mesonet wind QC and how the winds are used on the write up on QC from the Wiki page on our VLab site.  

 

And sorry about the formatting on that Wiki page.  It looks like the VLab collaboration services upgrade was a bit incompatible with our original formatting.  We need to get that fixed up.

 

If you can't find what you're looking for on the Wiki, please let us know.

 

Thanks,

Jacob