RE: Re: National Blend of Models - Feedback on National Blend of Models: Routine NBM mishandling of coastal stratus temporal evolution - Forums - RTMA/URMA Discussion Group

Forums

RE: Re: National Blend of Models - Feedback on National Blend of Models: Routine NBM mishandling of coastal stratus temporal evolution

RSS (Opens New Window)

Jeffrey Craven, modified 3 Years ago.

Re: National Blend of Models - Feedback on National Blend of Models: Routine NBM mishandling of coastal stratus temporal evolution

Youngling Posts: 90 Join Date: 9/24/12 Recent Posts

Roger and Warren: Thanks for sharing this. When I see systematic items like this the first place I check is URMA.

So at 15z and 02z, there is a significant drop in overall sky cover with URMA each day. I looked at examples just in the past few days and I see it.

My best guess is there is a transition period between using visible satellite and fog product/IR imagery in the morning and the evening where this occurs. I am not sure what really can be done from MDL/NBM stand point. I have cc:d the RTMA/URMA list for their awareness.

So here is a loop from 14-15-16z where we have NDFDnow and BCDG on the left and RTMA/URMA on the right.

Same thing from a few days ago except 01-02-03z.

JPC

Jeff Craven

Chief, Statistical Modeling Division (SMD)

Vice Lead, Innovation, Science & Technology (Evolve PMO)

National Weather Service, W/STI-12

Meteorological Development Laboratory (MDL)

Room 10410, SSMC2

Silver Spring, MD 20910

(816) 506-9783 cell/text

(301) 427-9475 office

@jpcstorm

https://www.weather.gov/mdl/nbm_home
https://vlab.ncep.noaa.gov/web/mdl/home

On Thu, Aug 5, 2021 at 10:11 PM Roger Gass - NWS Federal <roger.gass@noaa.gov> wrote:

All,

Thanks for putting this together and passing it up Warren. I will add that the 15Z artifact extends out through our entire 7-day forecast cycle. Easiest way to remove this is by deleting each 15Z grid before clicking "continue to finalize check" in step-2 of Forecast Builder. The finalize check will interpolate between the 12Z and 18Z grid to remove the artifact.

Roger

On Thu, Aug 5, 2021 at 5:35 PM Warren Blier - NOAA Federal <warren.blier@noaa.gov> wrote:

Thanks to Roger for first bringing this to my attention. In now having looked at it for several days in our current persistent and typical summer coastal stratus regime, seems to be showing up at the same times every day in the NBM output -- as shown here:

and here (from our GFE as WSUP viewer only goes out so far with allowing display of hourly grids)

Mets: just fyi, unless do something other than simply populate from FB with NBM, goes straight into our sky grids.

Warren

-------------------------------------------
Warren Blier, Ph.D.
Science and Operations Officer
NOAA/NWS San Francisco Bay Area Forecast Office
831-238-8109 (cell)
831-656-1724 (ops floor)

--

Roger Gass
Lead Forecaster

National Weather Service

San Francisco Bay Area/Monterey, CA

www.weather.gov/BayArea

(831) 656-1710

Ted Ryan, modified 3 Years ago.

RE: Re: National Blend of Models - Feedback on National Blend of Models: Routine NBM mishandling of coastal stratus temporal evolution

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

Here's an idea. This couldn't be done in the RTMA necessarily. But URMA (with hindsight) should be able to see the sudden jump in values over a large area going from cloudy to not so cloudy back to cloudy. These jumps would be predictable, only occurring at certain hours of the day based on the time of year. So if the algorithm detects the cloudy -> clear -> cloudy and at a time of day we expect to see it in that area, then it just backfills the clear hour(s) with an average (interpolated) value from the two cloudy bookend hours.

On Fri, Aug 6, 2021 at 8:47 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Roger and Warren: Thanks for sharing this. When I see systematic items like this the first place I check is URMA.

So at 15z and 02z, there is a significant drop in overall sky cover with URMA each day. I looked at examples just in the past few days and I see it.

My best guess is there is a transition period between using visible satellite and fog product/IR imagery in the morning and the evening where this occurs. I am not sure what really can be done from MDL/NBM stand point. I have cc:d the RTMA/URMA list for their awareness.

So here is a loop from 14-15-16z where we have NDFDnow and BCDG on the left and RTMA/URMA on the right.

Same thing from a few days ago except 01-02-03z.

JPC

Jeff Craven

Chief, Statistical Modeling Division (SMD)

Vice Lead, Innovation, Science & Technology (Evolve PMO)

National Weather Service, W/STI-12

Meteorological Development Laboratory (MDL)

Room 10410, SSMC2

Silver Spring, MD 20910

(816) 506-9783 cell/text

(301) 427-9475 office

@jpcstorm

https://www.weather.gov/mdl/nbm_home
https://vlab.ncep.noaa.gov/web/mdl/home

On Thu, Aug 5, 2021 at 10:11 PM Roger Gass - NWS Federal <roger.gass@noaa.gov> wrote:

All,

Thanks for putting this together and passing it up Warren. I will add that the 15Z artifact extends out through our entire 7-day forecast cycle. Easiest way to remove this is by deleting each 15Z grid before clicking "continue to finalize check" in step-2 of Forecast Builder. The finalize check will interpolate between the 12Z and 18Z grid to remove the artifact.

Roger

On Thu, Aug 5, 2021 at 5:35 PM Warren Blier - NOAA Federal <warren.blier@noaa.gov> wrote:

Thanks to Roger for first bringing this to my attention. In now having looked at it for several days in our current persistent and typical summer coastal stratus regime, seems to be showing up at the same times every day in the NBM output -- as shown here:

and here (from our GFE as WSUP viewer only goes out so far with allowing display of hourly grids)

Mets: just fyi, unless do something other than simply populate from FB with NBM, goes straight into our sky grids.

Warren

-------------------------------------------
Warren Blier, Ph.D.
Science and Operations Officer
NOAA/NWS San Francisco Bay Area Forecast Office
831-238-8109 (cell)
831-656-1724 (ops floor)

--

Roger Gass
Lead Forecaster

National Weather Service

San Francisco Bay Area/Monterey, CA

www.weather.gov/BayArea

(831) 656-1710

--
Jeffrey Craven RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/16933530 VLab.Notifications@noaa.gov

Ted Ryan

Acting Science and Training Branch Chief

Southern Region Headquarters - STSD

June through September

Science and Operations Officer

National Weather Service Forecast Office

Fort Worth, Texas

817-429-2631 x224

Matthew Morris, modified 3 Years ago.

RE: Re: National Blend of Models - Feedback on National Blend of Models: Routine NBM mishandling of coastal stratus temporal evolution

Youngling Posts: 169 Join Date: 12/6/17 Recent Posts

Hi Jeff and Warren,

Thanks for bringing this case to our attention. We are looking into it and will share our findings once we learn more.

-Matt

On Fri, Aug 6, 2021 at 9:47 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Roger and Warren: Thanks for sharing this. When I see systematic items like this the first place I check is URMA.

So at 15z and 02z, there is a significant drop in overall sky cover with URMA each day. I looked at examples just in the past few days and I see it.

My best guess is there is a transition period between using visible satellite and fog product/IR imagery in the morning and the evening where this occurs. I am not sure what really can be done from MDL/NBM stand point. I have cc:d the RTMA/URMA list for their awareness.

So here is a loop from 14-15-16z where we have NDFDnow and BCDG on the left and RTMA/URMA on the right.

Same thing from a few days ago except 01-02-03z.

JPC

Jeff Craven

Chief, Statistical Modeling Division (SMD)

Vice Lead, Innovation, Science & Technology (Evolve PMO)

National Weather Service, W/STI-12

Meteorological Development Laboratory (MDL)

Room 10410, SSMC2

Silver Spring, MD 20910

(816) 506-9783 cell/text

(301) 427-9475 office

@jpcstorm

https://www.weather.gov/mdl/nbm_home
https://vlab.ncep.noaa.gov/web/mdl/home

On Thu, Aug 5, 2021 at 10:11 PM Roger Gass - NWS Federal <roger.gass@noaa.gov> wrote:

All,

Thanks for putting this together and passing it up Warren. I will add that the 15Z artifact extends out through our entire 7-day forecast cycle. Easiest way to remove this is by deleting each 15Z grid before clicking "continue to finalize check" in step-2 of Forecast Builder. The finalize check will interpolate between the 12Z and 18Z grid to remove the artifact.

Roger

On Thu, Aug 5, 2021 at 5:35 PM Warren Blier - NOAA Federal <warren.blier@noaa.gov> wrote:

Thanks to Roger for first bringing this to my attention. In now having looked at it for several days in our current persistent and typical summer coastal stratus regime, seems to be showing up at the same times every day in the NBM output -- as shown here:

and here (from our GFE as WSUP viewer only goes out so far with allowing display of hourly grids)

Mets: just fyi, unless do something other than simply populate from FB with NBM, goes straight into our sky grids.

Warren

-------------------------------------------
Warren Blier, Ph.D.
Science and Operations Officer
NOAA/NWS San Francisco Bay Area Forecast Office
831-238-8109 (cell)
831-656-1724 (ops floor)

--

Roger Gass
Lead Forecaster

National Weather Service

San Francisco Bay Area/Monterey, CA

www.weather.gov/BayArea

(831) 656-1710

--
Jeffrey Craven RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/16933530 VLab.Notifications@noaa.gov

Matthew Morris, modified 3 Years ago.

RE: Re: National Blend of Models - Feedback on National Blend of Models: Routine NBM mishandling of coastal stratus temporal evolution

Youngling Posts: 169 Join Date: 12/6/17 Recent Posts

Hi Jeff and Warren,

We have taken a closer look at the cases you recently shared regarding the RTMA/URMA sky cover analyses. The following slide deck shows the graphics from 14Z to 17Z on August 5th and 6th:

SFO Sky Cover Case Study

With the latest RTMA/URMA implementation (v2.8; July 2020), we implemented a quality control procedure for the GOES sky cover observations; this QC rejects observations <30% when the solar zenith angle exceeds 80 degrees. This QC likely explains why several offshore observations are rejected at 14Z, but not at 15Z, and leads to a temporary decrease in the sky cover values offshore at 15Z. Then, at 16Z, the sky cover observations offshore increase markedly leading to increases in the analyzed sky cover values, which more closely match the background fields on August 6th. We believe that an adjustment to the cut-off solar zenith angle used for the QC would likely address this issue, but as you know we are unable to make any further code changes to the operational 2D-RTMA/URMA suite.

We will share this case with the GOES Imager team at CIMSS and find out if anything can be done on their end. We expect to see improvements in the upcoming 3D-RTMA system with the inclusion of a complex cloud analysis procedure. That system is expected to be implemented in Q1FY2024.

Please let us know if you have any further questions.

-Matt

On Tue, Aug 10, 2021 at 1:23 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Hi Jeff and Warren,

Thanks for bringing this case to our attention. We are looking into it and will share our findings once we learn more.

-Matt

On Fri, Aug 6, 2021 at 9:47 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Roger and Warren: Thanks for sharing this. When I see systematic items like this the first place I check is URMA.

So at 15z and 02z, there is a significant drop in overall sky cover with URMA each day. I looked at examples just in the past few days and I see it.

My best guess is there is a transition period between using visible satellite and fog product/IR imagery in the morning and the evening where this occurs. I am not sure what really can be done from MDL/NBM stand point. I have cc:d the RTMA/URMA list for their awareness.

So here is a loop from 14-15-16z where we have NDFDnow and BCDG on the left and RTMA/URMA on the right.

Same thing from a few days ago except 01-02-03z.

JPC

Jeff Craven

Chief, Statistical Modeling Division (SMD)

Vice Lead, Innovation, Science & Technology (Evolve PMO)

National Weather Service, W/STI-12

Meteorological Development Laboratory (MDL)

Room 10410, SSMC2

Silver Spring, MD 20910

(816) 506-9783 cell/text

(301) 427-9475 office

@jpcstorm

https://www.weather.gov/mdl/nbm_home
https://vlab.ncep.noaa.gov/web/mdl/home

On Thu, Aug 5, 2021 at 10:11 PM Roger Gass - NWS Federal <roger.gass@noaa.gov> wrote:

All,

Thanks for putting this together and passing it up Warren. I will add that the 15Z artifact extends out through our entire 7-day forecast cycle. Easiest way to remove this is by deleting each 15Z grid before clicking "continue to finalize check" in step-2 of Forecast Builder. The finalize check will interpolate between the 12Z and 18Z grid to remove the artifact.

Roger

On Thu, Aug 5, 2021 at 5:35 PM Warren Blier - NOAA Federal <warren.blier@noaa.gov> wrote:

Thanks to Roger for first bringing this to my attention. In now having looked at it for several days in our current persistent and typical summer coastal stratus regime, seems to be showing up at the same times every day in the NBM output -- as shown here:

and here (from our GFE as WSUP viewer only goes out so far with allowing display of hourly grids)

Mets: just fyi, unless do something other than simply populate from FB with NBM, goes straight into our sky grids.

Warren

-------------------------------------------
Warren Blier, Ph.D.
Science and Operations Officer
NOAA/NWS San Francisco Bay Area Forecast Office
831-238-8109 (cell)
831-656-1724 (ops floor)

--

Roger Gass
Lead Forecaster

National Weather Service

San Francisco Bay Area/Monterey, CA

www.weather.gov/BayArea

(831) 656-1710

--
Jeffrey Craven RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/16933530 VLab.Notifications@noaa.gov

--
Matthew Morris RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/16991258 VLab.Notifications@noaa.gov

Warren Blier, modified 3 Years ago.

RE: Re: National Blend of Models - Feedback on National Blend of Models: Routine NBM mishandling of coastal stratus temporal evolution

Youngling Posts: 14 Join Date: 9/24/12 Recent Posts

Hi Matt,

Really appreciate the in-depth look into this, and think your proposed solution sounds good (once the WCOSS Moratorium ends and it becomes possible to make changes to the operational 2D-RTMA/URMA suite). And appreciate your idea of sharing it with the GOES Imager team at CIMSS in case there's anything that can be done by them.

Thanks,

Warren

On Mon, Aug 16, 2021 at 7:28 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Hi Jeff and Warren,

We have taken a closer look at the cases you recently shared regarding the RTMA/URMA sky cover analyses. The following slide deck shows the graphics from 14Z to 17Z on August 5th and 6th:

SFO Sky Cover Case Study

With the latest RTMA/URMA implementation (v2.8; July 2020), we implemented a quality control procedure for the GOES sky cover observations; this QC rejects observations <30% when the solar zenith angle exceeds 80 degrees. This QC likely explains why several offshore observations are rejected at 14Z, but not at 15Z, and leads to a temporary decrease in the sky cover values offshore at 15Z. Then, at 16Z, the sky cover observations offshore increase markedly leading to increases in the analyzed sky cover values, which more closely match the background fields on August 6th. We believe that an adjustment to the cut-off solar zenith angle used for the QC would likely address this issue, but as you know we are unable to make any further code changes to the operational 2D-RTMA/URMA suite.

We will share this case with the GOES Imager team at CIMSS and find out if anything can be done on their end. We expect to see improvements in the upcoming 3D-RTMA system with the inclusion of a complex cloud analysis procedure. That system is expected to be implemented in Q1FY2024.

Please let us know if you have any further questions.

-Matt

On Tue, Aug 10, 2021 at 1:23 PM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Hi Jeff and Warren,

Thanks for bringing this case to our attention. We are looking into it and will share our findings once we learn more.

-Matt

On Fri, Aug 6, 2021 at 9:47 AM VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Roger and Warren: Thanks for sharing this. When I see systematic items like this the first place I check is URMA.

So at 15z and 02z, there is a significant drop in overall sky cover with URMA each day. I looked at examples just in the past few days and I see it.

My best guess is there is a transition period between using visible satellite and fog product/IR imagery in the morning and the evening where this occurs. I am not sure what really can be done from MDL/NBM stand point. I have cc:d the RTMA/URMA list for their awareness.

So here is a loop from 14-15-16z where we have NDFDnow and BCDG on the left and RTMA/URMA on the right.

Same thing from a few days ago except 01-02-03z.

JPC

Jeff Craven

Chief, Statistical Modeling Division (SMD)

Vice Lead, Innovation, Science & Technology (Evolve PMO)

National Weather Service, W/STI-12

Meteorological Development Laboratory (MDL)

Room 10410, SSMC2

Silver Spring, MD 20910

(816) 506-9783 cell/text

(301) 427-9475 office

@jpcstorm

https://www.weather.gov/mdl/nbm_home
https://vlab.ncep.noaa.gov/web/mdl/home

On Thu, Aug 5, 2021 at 10:11 PM Roger Gass - NWS Federal <roger.gass@noaa.gov> wrote:

All,

Thanks for putting this together and passing it up Warren. I will add that the 15Z artifact extends out through our entire 7-day forecast cycle. Easiest way to remove this is by deleting each 15Z grid before clicking "continue to finalize check" in step-2 of Forecast Builder. The finalize check will interpolate between the 12Z and 18Z grid to remove the artifact.

Roger

On Thu, Aug 5, 2021 at 5:35 PM Warren Blier - NOAA Federal <warren.blier@noaa.gov> wrote:

Thanks to Roger for first bringing this to my attention. In now having looked at it for several days in our current persistent and typical summer coastal stratus regime, seems to be showing up at the same times every day in the NBM output -- as shown here:

and here (from our GFE as WSUP viewer only goes out so far with allowing display of hourly grids)

Mets: just fyi, unless do something other than simply populate from FB with NBM, goes straight into our sky grids.

Warren

-------------------------------------------
Warren Blier, Ph.D.
Science and Operations Officer
NOAA/NWS San Francisco Bay Area Forecast Office
831-238-8109 (cell)
831-656-1724 (ops floor)

--

Roger Gass
Lead Forecaster

National Weather Service

San Francisco Bay Area/Monterey, CA

www.weather.gov/BayArea

(831) 656-1710

--
Jeffrey Craven RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/16933530 VLab.Notifications@noaa.gov

--
Matthew Morris RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/16991258 VLab.Notifications@noaa.gov

--
Matthew Morris RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/17085038 VLab.Notifications@noaa.gov

RE: Re: National Blend of Models - Feedback on National Blend of Models: Routine NBM mishandling of coastal stratus temporal evolution - Forums - RTMA/URMA Discussion Group

Forums