Forums

Back

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

RSS (Opens New Window)(Opens New Window)
DB
David Bonnette, modified 1 Year ago.

High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 4 Join Date: 3/31/18 Recent Posts

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)

JS
Jack Settelmaier, modified 1 Year ago.

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 23 Join Date: 11/2/12 Recent Posts
David Bonnette:

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)

Not sure my Reply via email made it through, so I'd thought I'd drop it in here via the VLab login, and hopefully, share a visual of what you describe.

 

Awesome analysis David!

 

It seems like another option might be to adjust the weighting, and less weighting for distance, for CWOP station Td contributions.  Or, as you say, they might be best to just not use, and take a more whitelisting approach, if their input is desired. 

 

Thanks so much for reporting this.

 

 

 

Attachments:

Attachment

MM
Matthew Morris, modified 1 Year ago.

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 169 Join Date: 12/6/17 Recent Posts
Hi David and Jack,

We have received some feedback from other offices regarding CWOP dewpoint observations, but it is not clear whether these issues are localized or more widespread.  Keep in mind that the RTMA/URMA suite is configured to reject all mesonet wind observations by default.  Wind speed and gust observations are only used if the provider is trusted or if the individual station is on a trusted station list.  On the other hand, mesonet observations of other fields (e.g., temperature and moisture) are used unless the individual station appears on a reject list.  Thus, updating RTMA/URMA to perform a bulk blacklisting of CWOP dewpoint observations is not a viable option at this time, as it would constitute a science change.  Thus, we will need to handle this on a case-by-case basis.

In principle, we could add the new CWOP sites to the SDM reject list, but, as you've already noted, any new CWOP stations will likely reintroduce the problem, so it will require frequent manual intervention to remove the observations from the analysis.  If you would like to add the new stations to the reject list, we can work with you on that, but it will also take some time.  There is a limitation on the number of mesonet stations that can be read off the SDM reject list, and we are very close to reaching that limit.  A fix is in the works for this, but it will take several months to be implemented.  In the meantime, if you have a list of the new CWOP stations in the area, please share it with us so that we can add the stations to the SDM reject list once it's feasible to do so.

The 3D-RTMA/URMA system will include an automated quality control package that will hopefully better capture problematic observations.  In the coming months, we will begin sharing experimental 3D-RTMA/URMA products for stakeholder feedback.

Thanks,
Matt

On Thu, Jul 27, 2023 at 11:33 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
David Bonnette:

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)

Not sure my Reply via email made it through, so I'd thought I'd drop it in here via the VLab login, and hopefully, share a visual of what you describe.

 

Awesome analysis David!

 

It seems like another option might be to adjust the weighting, and less weighting for distance, for CWOP station Td contributions.  Or, as you say, they might be best to just not use, and take a more whitelisting approach, if their input is desired. 

 

Thanks so much for reporting this.

 

 

 


--
Jack Settelmaier RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/30999304VLab.Notifications@noaa.gov


--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758
DB
David Bonnette, modified 1 Year ago.

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 4 Join Date: 3/31/18 Recent Posts

Thanks for the replies. The issue doesn't seem to be localized to North Texas. I ran some METAR vs URMA stats from last summer and May 2023 to see if there was a similar trend across the county. It appears that the highest Td bias coincides with CWAs that are more highly populated (i.e. more CWOP/Meso/backyard) stations drowning out the analysis. This is still just speculation on our part, but FWD did see improvement last year after we bulk-blacklisted the unreliable Td locations. We could continue to blacklist each site as it crops up, but I fear attempting to do so would be futile and expend resources that WFOs simply don't have.

These slides have the data mentioned above:

https://docs.google.com/presentation/d/1vS0zkKlcpJaDcCVn4r22uTkxzOfYG8S6Yf1Zbwpc4d8/

 

Since mesonet winds are now rejected by default, would it be worth exploring something similar for Td? If not that, then a science change in future iterations of the RTMA/URMA that provides "more weight" to METAR values than mesonet values or a comparison with nearby METAR values? Or something else? I really don't know what the eventual solution could/should be, but I do know that the current URMA analysis is not representative of what we are observing at our reliable observation locations and at the end of the day, this is causing downstream IDSS implications where no forecaster intervention takes place.

 

Thank you!

David Bonnette

WFO FWD 

Matthew Morris:
Hi David and Jack,
 
We have received some feedback from other offices regarding CWOP dewpoint observations, but it is not clear whether these issues are localized or more widespread.  Keep in mind that the RTMA/URMA suite is configured to reject all mesonet wind observations by default.  Wind speed and gust observations are only used if the provider is trusted or if the individual station is on a trusted station list.  On the other hand, mesonet observations of other fields (e.g., temperature and moisture) are used unless the individual station appears on a reject list.  Thus, updating RTMA/URMA to perform a bulk blacklisting of CWOP dewpoint observations is not a viable option at this time, as it would constitute a science change.  Thus, we will need to handle this on a case-by-case basis.
 
In principle, we could add the new CWOP sites to the SDM reject list, but, as you've already noted, any new CWOP stations will likely reintroduce the problem, so it will require frequent manual intervention to remove the observations from the analysis.  If you would like to add the new stations to the reject list, we can work with you on that, but it will also take some time.  There is a limitation on the number of mesonet stations that can be read off the SDM reject list, and we are very close to reaching that limit.  A fix is in the works for this, but it will take several months to be implemented.  In the meantime, if you have a list of the new CWOP stations in the area, please share it with us so that we can add the stations to the SDM reject list once it's feasible to do so.
 
The 3D-RTMA/URMA system will include an automated quality control package that will hopefully better capture problematic observations.  In the coming months, we will begin sharing experimental 3D-RTMA/URMA products for stakeholder feedback.
 
Thanks,
Matt
 
On Thu, Jul 27, 2023 at 11:33 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
David Bonnette:

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)

Not sure my Reply via email made it through, so I'd thought I'd drop it in here via the VLab login, and hopefully, share a visual of what you describe.

 

Awesome analysis David!

 

It seems like another option might be to adjust the weighting, and less weighting for distance, for CWOP station Td contributions.  Or, as you say, they might be best to just not use, and take a more whitelisting approach, if their input is desired. 

 

Thanks so much for reporting this.

 

 

 


--
Jack Settelmaier RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/30999304VLab.Notifications@noaa.gov
 
--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758

 

 

 

MM
Matthew Morris, modified 1 Year ago.

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 169 Join Date: 12/6/17 Recent Posts
Hi David,

We will have some internal discussions regarding this issue in order to determine the best path forward.  As noted before, updating the RTMA/URMA code to reject CWOP dewpoint observations by default would constitute a science change and require a system implementation, which will likely not be possible as the operational system is largely frozen while work continues on 3D-RTMA/URMA.  Once we are ready to share experimental 3D-RTMA/URMA products with stakeholders, we will also resume regular telecons with users to gather feedback such as this.

Thanks,
Matt

On Tue, Aug 1, 2023 at 3:25 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Thanks for the replies. The issue doesn't seem to be localized to North Texas. I ran some METAR vs URMA stats from last summer and May 2023 to see if there was a similar trend across the county. It appears that the highest Td bias coincides with CWAs that are more highly populated (i.e. more CWOP/Meso/backyard) stations drowning out the analysis. This is still just speculation on our part, but FWD did see improvement last year after we bulk-blacklisted the unreliable Td locations. We could continue to blacklist each site as it crops up, but I fear attempting to do so would be futile and expend resources that WFOs simply don't have.

These slides have the data mentioned above:

https://docs.google.com/presentation/d/1vS0zkKlcpJaDcCVn4r22uTkxzOfYG8S6Yf1Zbwpc4d8/

 

Since mesonet winds are now rejected by default, would it be worth exploring something similar for Td? If not that, then a science change in future iterations of the RTMA/URMA that provides "more weight" to METAR values than mesonet values or a comparison with nearby METAR values? Or something else? I really don't know what the eventual solution could/should be, but I do know that the current URMA analysis is not representative of what we are observing at our reliable observation locations and at the end of the day, this is causing downstream IDSS implications where no forecaster intervention takes place.

 

Thank you!

David Bonnette

WFO FWD 

Matthew Morris:
Hi David and Jack,
 
We have received some feedback from other offices regarding CWOP dewpoint observations, but it is not clear whether these issues are localized or more widespread.  Keep in mind that the RTMA/URMA suite is configured to reject all mesonet wind observations by default.  Wind speed and gust observations are only used if the provider is trusted or if the individual station is on a trusted station list.  On the other hand, mesonet observations of other fields (e.g., temperature and moisture) are used unless the individual station appears on a reject list.  Thus, updating RTMA/URMA to perform a bulk blacklisting of CWOP dewpoint observations is not a viable option at this time, as it would constitute a science change.  Thus, we will need to handle this on a case-by-case basis.
 
In principle, we could add the new CWOP sites to the SDM reject list, but, as you've already noted, any new CWOP stations will likely reintroduce the problem, so it will require frequent manual intervention to remove the observations from the analysis.  If you would like to add the new stations to the reject list, we can work with you on that, but it will also take some time.  There is a limitation on the number of mesonet stations that can be read off the SDM reject list, and we are very close to reaching that limit.  A fix is in the works for this, but it will take several months to be implemented.  In the meantime, if you have a list of the new CWOP stations in the area, please share it with us so that we can add the stations to the SDM reject list once it's feasible to do so.
 
The 3D-RTMA/URMA system will include an automated quality control package that will hopefully better capture problematic observations.  In the coming months, we will begin sharing experimental 3D-RTMA/URMA products for stakeholder feedback.
 
Thanks,
Matt
 
On Thu, Jul 27, 2023 at 11:33 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
David Bonnette:

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)

Not sure my Reply via email made it through, so I'd thought I'd drop it in here via the VLab login, and hopefully, share a visual of what you describe.

 

Awesome analysis David!

 

It seems like another option might be to adjust the weighting, and less weighting for distance, for CWOP station Td contributions.  Or, as you say, they might be best to just not use, and take a more whitelisting approach, if their input is desired. 

 

Thanks so much for reporting this.

 

 

 


--
Jack Settelmaier RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/30999304VLab.Notifications@noaa.gov
 
--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758

 

 

 


--
David Bonnette RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31049184VLab.Notifications@noaa.gov


--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758
TR
Ted Ryan, modified 1 Year ago.

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 11 Join Date: 1/4/16 Recent Posts
Matt, 

Copy that. Is there a general timeline on when some kind of meaningful action could be taken with regard to this issue?  Are we talking about a year? two years?

The good news is that this seems to be more of a summertime problem, the bad news is that the NBM has trained itself to be too humid which feeds into heat index and fire weather forecasts, both of which are unfortunately a big issue for us this year.

On Wed, Aug 2, 2023 at 2:04 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Hi David,

We will have some internal discussions regarding this issue in order to determine the best path forward.  As noted before, updating the RTMA/URMA code to reject CWOP dewpoint observations by default would constitute a science change and require a system implementation, which will likely not be possible as the operational system is largely frozen while work continues on 3D-RTMA/URMA.  Once we are ready to share experimental 3D-RTMA/URMA products with stakeholders, we will also resume regular telecons with users to gather feedback such as this.

Thanks,
Matt

On Tue, Aug 1, 2023 at 3:25 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Thanks for the replies. The issue doesn't seem to be localized to North Texas. I ran some METAR vs URMA stats from last summer and May 2023 to see if there was a similar trend across the county. It appears that the highest Td bias coincides with CWAs that are more highly populated (i.e. more CWOP/Meso/backyard) stations drowning out the analysis. This is still just speculation on our part, but FWD did see improvement last year after we bulk-blacklisted the unreliable Td locations. We could continue to blacklist each site as it crops up, but I fear attempting to do so would be futile and expend resources that WFOs simply don't have.

These slides have the data mentioned above:

https://docs.google.com/presentation/d/1vS0zkKlcpJaDcCVn4r22uTkxzOfYG8S6Yf1Zbwpc4d8/

 

Since mesonet winds are now rejected by default, would it be worth exploring something similar for Td? If not that, then a science change in future iterations of the RTMA/URMA that provides "more weight" to METAR values than mesonet values or a comparison with nearby METAR values? Or something else? I really don't know what the eventual solution could/should be, but I do know that the current URMA analysis is not representative of what we are observing at our reliable observation locations and at the end of the day, this is causing downstream IDSS implications where no forecaster intervention takes place.

 

Thank you!

David Bonnette

WFO FWD 

Matthew Morris:
Hi David and Jack,
 
We have received some feedback from other offices regarding CWOP dewpoint observations, but it is not clear whether these issues are localized or more widespread.  Keep in mind that the RTMA/URMA suite is configured to reject all mesonet wind observations by default.  Wind speed and gust observations are only used if the provider is trusted or if the individual station is on a trusted station list.  On the other hand, mesonet observations of other fields (e.g., temperature and moisture) are used unless the individual station appears on a reject list.  Thus, updating RTMA/URMA to perform a bulk blacklisting of CWOP dewpoint observations is not a viable option at this time, as it would constitute a science change.  Thus, we will need to handle this on a case-by-case basis.
 
In principle, we could add the new CWOP sites to the SDM reject list, but, as you've already noted, any new CWOP stations will likely reintroduce the problem, so it will require frequent manual intervention to remove the observations from the analysis.  If you would like to add the new stations to the reject list, we can work with you on that, but it will also take some time.  There is a limitation on the number of mesonet stations that can be read off the SDM reject list, and we are very close to reaching that limit.  A fix is in the works for this, but it will take several months to be implemented.  In the meantime, if you have a list of the new CWOP stations in the area, please share it with us so that we can add the stations to the SDM reject list once it's feasible to do so.
 
The 3D-RTMA/URMA system will include an automated quality control package that will hopefully better capture problematic observations.  In the coming months, we will begin sharing experimental 3D-RTMA/URMA products for stakeholder feedback.
 
Thanks,
Matt
 
On Thu, Jul 27, 2023 at 11:33 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
David Bonnette:

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)

Not sure my Reply via email made it through, so I'd thought I'd drop it in here via the VLab login, and hopefully, share a visual of what you describe.

 

Awesome analysis David!

 

It seems like another option might be to adjust the weighting, and less weighting for distance, for CWOP station Td contributions.  Or, as you say, they might be best to just not use, and take a more whitelisting approach, if their input is desired. 

 

Thanks so much for reporting this.

 

 

 


--
Jack Settelmaier RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/30999304VLab.Notifications@noaa.gov
 
--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758

 

 

 


--
David Bonnette RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31049184VLab.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 RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31069456VLab.Notifications@noaa.gov


--


Ted Ryan
Science and Operations Officer
National Weather Service Forecast Office
Fort Worth, Texas
817-429-2631 x224
MM
Matthew Morris, modified 1 Year ago.

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 169 Join Date: 12/6/17 Recent Posts
Hi Ted,

At this point, we will likely be looking ahead to 3D-RTMA/URMA for any meaningful action to be taken given that the operational RTMA/URMA system is largely frozen at this point.  The first implementation of 3D-RTMA is currently scheduled for early 2025.  As we continue to have internal discussions on how to address this issue, we will keep you updated.

Thanks,
Matt

On Fri, Aug 4, 2023 at 11:01 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Matt, 

Copy that. Is there a general timeline on when some kind of meaningful action could be taken with regard to this issue?  Are we talking about a year? two years?

The good news is that this seems to be more of a summertime problem, the bad news is that the NBM has trained itself to be too humid which feeds into heat index and fire weather forecasts, both of which are unfortunately a big issue for us this year.

On Wed, Aug 2, 2023 at 2:04 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Hi David,

We will have some internal discussions regarding this issue in order to determine the best path forward.  As noted before, updating the RTMA/URMA code to reject CWOP dewpoint observations by default would constitute a science change and require a system implementation, which will likely not be possible as the operational system is largely frozen while work continues on 3D-RTMA/URMA.  Once we are ready to share experimental 3D-RTMA/URMA products with stakeholders, we will also resume regular telecons with users to gather feedback such as this.

Thanks,
Matt

On Tue, Aug 1, 2023 at 3:25 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Thanks for the replies. The issue doesn't seem to be localized to North Texas. I ran some METAR vs URMA stats from last summer and May 2023 to see if there was a similar trend across the county. It appears that the highest Td bias coincides with CWAs that are more highly populated (i.e. more CWOP/Meso/backyard) stations drowning out the analysis. This is still just speculation on our part, but FWD did see improvement last year after we bulk-blacklisted the unreliable Td locations. We could continue to blacklist each site as it crops up, but I fear attempting to do so would be futile and expend resources that WFOs simply don't have.

These slides have the data mentioned above:

https://docs.google.com/presentation/d/1vS0zkKlcpJaDcCVn4r22uTkxzOfYG8S6Yf1Zbwpc4d8/

 

Since mesonet winds are now rejected by default, would it be worth exploring something similar for Td? If not that, then a science change in future iterations of the RTMA/URMA that provides "more weight" to METAR values than mesonet values or a comparison with nearby METAR values? Or something else? I really don't know what the eventual solution could/should be, but I do know that the current URMA analysis is not representative of what we are observing at our reliable observation locations and at the end of the day, this is causing downstream IDSS implications where no forecaster intervention takes place.

 

Thank you!

David Bonnette

WFO FWD 

Matthew Morris:
Hi David and Jack,
 
We have received some feedback from other offices regarding CWOP dewpoint observations, but it is not clear whether these issues are localized or more widespread.  Keep in mind that the RTMA/URMA suite is configured to reject all mesonet wind observations by default.  Wind speed and gust observations are only used if the provider is trusted or if the individual station is on a trusted station list.  On the other hand, mesonet observations of other fields (e.g., temperature and moisture) are used unless the individual station appears on a reject list.  Thus, updating RTMA/URMA to perform a bulk blacklisting of CWOP dewpoint observations is not a viable option at this time, as it would constitute a science change.  Thus, we will need to handle this on a case-by-case basis.
 
In principle, we could add the new CWOP sites to the SDM reject list, but, as you've already noted, any new CWOP stations will likely reintroduce the problem, so it will require frequent manual intervention to remove the observations from the analysis.  If you would like to add the new stations to the reject list, we can work with you on that, but it will also take some time.  There is a limitation on the number of mesonet stations that can be read off the SDM reject list, and we are very close to reaching that limit.  A fix is in the works for this, but it will take several months to be implemented.  In the meantime, if you have a list of the new CWOP stations in the area, please share it with us so that we can add the stations to the SDM reject list once it's feasible to do so.
 
The 3D-RTMA/URMA system will include an automated quality control package that will hopefully better capture problematic observations.  In the coming months, we will begin sharing experimental 3D-RTMA/URMA products for stakeholder feedback.
 
Thanks,
Matt
 
On Thu, Jul 27, 2023 at 11:33 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
David Bonnette:

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)

Not sure my Reply via email made it through, so I'd thought I'd drop it in here via the VLab login, and hopefully, share a visual of what you describe.

 

Awesome analysis David!

 

It seems like another option might be to adjust the weighting, and less weighting for distance, for CWOP station Td contributions.  Or, as you say, they might be best to just not use, and take a more whitelisting approach, if their input is desired. 

 

Thanks so much for reporting this.

 

 

 


--
Jack Settelmaier RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/30999304VLab.Notifications@noaa.gov
 
--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758

 

 

 


--
David Bonnette RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31049184VLab.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 RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31069456VLab.Notifications@noaa.gov


--


Ted Ryan
Science and Operations Officer
National Weather Service Forecast Office
Fort Worth, Texas
817-429-2631 x224

--
Ted Ryan RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31102418VLab.Notifications@noaa.gov


--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758
MM
Matthew Morris, modified 1 Year ago.

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 169 Join Date: 12/6/17 Recent Posts
Hi Ted and David,

The 2D-RTMA/URMA code is frozen at this point as we are preparing for the transition to 3D-RTMA/URMA.  The first implementation is expected in early 2025, but this is subject to change.  If there are delays with 3D-RTMA/URMA, then we may have a narrow window for a small "science change" upgrade to 2D-RTMA/URMA.  With that said, the 2D-RTMA/URMA code is currently not configured to reject moisture observations from an entire mesonet provider so this would constitute a major science change.  Since the error tables for the data assimilation system --the GSI-- assign the same error value to all mesonet observations, it wouldn't be feasible to assign a higher value to CWOP observations at this point.  We could explore the possibility of flagging the moisture observations from this network at the upstream data processing step, but we must caution that we don't know at this point if all of the forecast offices would be supportive of this change; for instance, isolated observations in remote areas could offer some value to the analysis.  Some coordination with the forecast offices through the Regional Centers would, therefore, be necessary.

We are aiming to share experimental 3D-RTMA output with users in the coming weeks/months.  An automated QC package is under development for this system, which we hope will lead to improvements over the operational system.

Thanks,
Matt

On Wed, Aug 9, 2023 at 8:58 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Hi Ted,

At this point, we will likely be looking ahead to 3D-RTMA/URMA for any meaningful action to be taken given that the operational RTMA/URMA system is largely frozen at this point.  The first implementation of 3D-RTMA is currently scheduled for early 2025.  As we continue to have internal discussions on how to address this issue, we will keep you updated.

Thanks,
Matt

On Fri, Aug 4, 2023 at 11:01 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Matt, 

Copy that. Is there a general timeline on when some kind of meaningful action could be taken with regard to this issue?  Are we talking about a year? two years?

The good news is that this seems to be more of a summertime problem, the bad news is that the NBM has trained itself to be too humid which feeds into heat index and fire weather forecasts, both of which are unfortunately a big issue for us this year.

On Wed, Aug 2, 2023 at 2:04 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Hi David,

We will have some internal discussions regarding this issue in order to determine the best path forward.  As noted before, updating the RTMA/URMA code to reject CWOP dewpoint observations by default would constitute a science change and require a system implementation, which will likely not be possible as the operational system is largely frozen while work continues on 3D-RTMA/URMA.  Once we are ready to share experimental 3D-RTMA/URMA products with stakeholders, we will also resume regular telecons with users to gather feedback such as this.

Thanks,
Matt

On Tue, Aug 1, 2023 at 3:25 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Thanks for the replies. The issue doesn't seem to be localized to North Texas. I ran some METAR vs URMA stats from last summer and May 2023 to see if there was a similar trend across the county. It appears that the highest Td bias coincides with CWAs that are more highly populated (i.e. more CWOP/Meso/backyard) stations drowning out the analysis. This is still just speculation on our part, but FWD did see improvement last year after we bulk-blacklisted the unreliable Td locations. We could continue to blacklist each site as it crops up, but I fear attempting to do so would be futile and expend resources that WFOs simply don't have.

These slides have the data mentioned above:

https://docs.google.com/presentation/d/1vS0zkKlcpJaDcCVn4r22uTkxzOfYG8S6Yf1Zbwpc4d8/

 

Since mesonet winds are now rejected by default, would it be worth exploring something similar for Td? If not that, then a science change in future iterations of the RTMA/URMA that provides "more weight" to METAR values than mesonet values or a comparison with nearby METAR values? Or something else? I really don't know what the eventual solution could/should be, but I do know that the current URMA analysis is not representative of what we are observing at our reliable observation locations and at the end of the day, this is causing downstream IDSS implications where no forecaster intervention takes place.

 

Thank you!

David Bonnette

WFO FWD 

Matthew Morris:
Hi David and Jack,
 
We have received some feedback from other offices regarding CWOP dewpoint observations, but it is not clear whether these issues are localized or more widespread.  Keep in mind that the RTMA/URMA suite is configured to reject all mesonet wind observations by default.  Wind speed and gust observations are only used if the provider is trusted or if the individual station is on a trusted station list.  On the other hand, mesonet observations of other fields (e.g., temperature and moisture) are used unless the individual station appears on a reject list.  Thus, updating RTMA/URMA to perform a bulk blacklisting of CWOP dewpoint observations is not a viable option at this time, as it would constitute a science change.  Thus, we will need to handle this on a case-by-case basis.
 
In principle, we could add the new CWOP sites to the SDM reject list, but, as you've already noted, any new CWOP stations will likely reintroduce the problem, so it will require frequent manual intervention to remove the observations from the analysis.  If you would like to add the new stations to the reject list, we can work with you on that, but it will also take some time.  There is a limitation on the number of mesonet stations that can be read off the SDM reject list, and we are very close to reaching that limit.  A fix is in the works for this, but it will take several months to be implemented.  In the meantime, if you have a list of the new CWOP stations in the area, please share it with us so that we can add the stations to the SDM reject list once it's feasible to do so.
 
The 3D-RTMA/URMA system will include an automated quality control package that will hopefully better capture problematic observations.  In the coming months, we will begin sharing experimental 3D-RTMA/URMA products for stakeholder feedback.
 
Thanks,
Matt
 
On Thu, Jul 27, 2023 at 11:33 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:
David Bonnette:

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)

Not sure my Reply via email made it through, so I'd thought I'd drop it in here via the VLab login, and hopefully, share a visual of what you describe.

 

Awesome analysis David!

 

It seems like another option might be to adjust the weighting, and less weighting for distance, for CWOP station Td contributions.  Or, as you say, they might be best to just not use, and take a more whitelisting approach, if their input is desired. 

 

Thanks so much for reporting this.

 

 

 


--
Jack Settelmaier RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/30999304VLab.Notifications@noaa.gov
 
--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758

 

 

 


--
David Bonnette RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31049184VLab.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 RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31069456VLab.Notifications@noaa.gov


--


Ted Ryan
Science and Operations Officer
National Weather Service Forecast Office
Fort Worth, Texas
817-429-2631 x224

--
Ted Ryan RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31102418VLab.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 RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/31192538VLab.Notifications@noaa.gov


--
Matthew Morris
SAIC at NOAA/NWS/NCEP/EMC
5830 University Research Ct., Rm. 2038
College Park, MD 20740
301-683-3758
JS
Jack Settelmaier, modified 1 Year ago.

RE: High Dew Point Bias - CWOP Sites Drowning out ASOS/AWOS Values

Youngling Posts: 23 Join Date: 11/2/12 Recent Posts
Awesome analysis David!

It seems like another option might be to adjust the weighting, and less weighting for distance, for CWOP station Td contributions.  Or, as you say, they might be best to just not use, and take a more whitelisting approach, if their input is desired. 

Thanks so much for reporting this.

On Thu, Jul 27, 2023 at 2:00 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

We had a problem last summer where the URMA analysis was several degrees higher than the METAR values. The same issue from last year is occurring again. This is causing a high-bias with the NBM at non-ASOS/AWOS points, therefore it seems the underlying cause is the URMA. Scrubbing through the KML files, it seems that the CWOP Td points are swamping out the more reliable ASOS/AWOS Td values (example images/examples in the attached presentation).

https://docs.google.com/presentation/d/1085DzsGcRXn1jAzH_Az2QKZs4_FSwtkufFxMQ-7HQIA

Last year, we did a bulk blacklist request of CWOP Td values. It appears that we've had a massive influx of CWOP Td observations in the past 12 months and the issue has returned. Is it worth doing a bulk blacklisting of all CWOP sites for dewpoint similar to what is currently done for wind? Is this simply a local/regional issue that can be handled by SOOs or is it more widespread across the country?

 

Thanks!

David Bonnette

WFO FWD (Fort Worth, TX)


--
David Bonnette RTMA/URMA Discussion Group Virtual Lab Forum https://vlab.noaa.gov/web/715073/discussions-forums-/-/message_boards/view_message/30996721VLab.Notifications@noaa.gov


--
Jack Settelmaier
Digital Techniques Meteorologist
NOAA/NWS, Southern Region HQ
Fort Worth, TX