RE: Re: [RTMA/URMA Discussion Group - RTMA/URMA Feedback Forum] RE: Max, RT - Forums - RTMA/URMA Discussion Group

Forums

RE: Re: [RTMA/URMA Discussion Group - RTMA/URMA Feedback Forum] RE: Max, RT

RSS (Opens New Window)

Roger Gass, modified 7 Years ago.

Max, RTMA/URMA concerns along central California coast

Youngling Posts: 1 Join Date: 12/29/13 Recent Posts

RTMA/URMA Team,

We wanted to pass along some concerns with the RTMA/URMA, especially along the Big Sur Coast south of Monterey, CA. During the past week, the RTMA/URMA has been consistently off (too cold) by 9° to 14° in Big Sur, CA where a well established RAWS is located. This type of discrepancy is very common year around at this location, especially during offshore flow.

This is likely due to the complex terrain and influences of the Pacific we deal with here across our forecast area. However, this location in particularly is highly visible due to tourism. Thus, was wondering if there is anything we can do to get the RAWS to be considered a reliable data point in the RTMA/URMA or if something else needs to be considered. As you would guess, this to shows up in the NMB forecast as a cold biases in these situations.

Open to feedback.

Sincerely,

Roger

Roger Gass

Lead Forecaster

National Weather Service

San Francisco Bay Area/Monterey, CA

www.weather.gov/sanfrancisco

(831) 656-1710

Steven Levine, modified 7 Years ago.

RE: Max, RTMA/URMA concerns along central California coast

Youngling Posts: 174 Join Date: 11/13/14 Recent Posts

Hi Roger,

It looks like you are talking about site PPSC1. While we get and use data from this station, for some reason we don't have a maxT for it valid on the 3rd (we have it for all other days going back to the 1st). This will happen if we are missing data for a few hours that day - we don't compute a min or maxT for obs where we are missing several hours of data. In that situation, we're stuck with the background value.

For the days we do have it, the background maxT is a good 5-6 F cooler than the observed maximum temperature. This is a common problem in these areas of complex terrain - the HRRR really struggles to get these small valleys/ridges fully cooled/warmed. We are working on a new background error metric that should give individual obs more weight in cases like this, but from what I see over the past few days it hasn't made much of an impact at this site.

I do notice that there is another CWOP site nearby (D1685) about 4 km SE of the RAWS site. That site is reporting 8-10 F cooler than PPSC1 for maxT over the past few days. Also worth noting is the differing terrain height levels we have at the two points. Rpt terrain is metadata-station provided, mdl terrain is the terrain at that point in the RTMA's dataset (in meters):

Station Lat/Lon Rpt Terrain Mdl Terrain

PPSC1 36.25/121.78 108 415.7

D1685 36.22/121.76 246 86.1

Despite the terrain difference, the background maxT values for the two sites are generally within 1-2 F of each other.

It is worth noting that we give equal weighting to all mesonet stations. We are working on ways to change that, but for now they are the same.

In the shorter term, it might be worth trying removing the temperature ob for D1685 from the analysis, and seeing if that has any effect. The terrain difference is large enough that these two obs should not be significantly counter-balancing each other, but they are close enough that I feel like it is worth taking a look at for a few days. If you want, please ask Warren (we want to limit this to SOO's) to fill out the reject list form on the VLab page, and I can flag the temperature and put a trace on temperature at PPSC1 for a few days. If it helps, we can keep the CWOP temperature out. If not, we can put it back in.

Steve

On Mon, Feb 5, 2018 at 2:11 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:

RTMA/URMA Team,

We wanted to pass along some concerns with the RTMA/URMA, especially along the Big Sur Coast south of Monterey, CA. During the past week, the RTMA/URMA has been consistently off (too cold) by 9° to 14° in Big Sur, CA where a well established RAWS is located. This type of discrepancy is very common year around at this location, especially during offshore flow.

This is likely due to the complex terrain and influences of the Pacific we deal with here across our forecast area. However, this location in particularly is highly visible due to tourism. Thus, was wondering if there is anything we can do to get the RAWS to be considered a reliable data point in the RTMA/URMA or if something else needs to be considered. As you would guess, this to shows up in the NMB forecast as a cold biases in these situations.

Open to feedback.

Sincerely,

Roger

--
Roger Gass
Lead Forecaster
National Weather Service
San Francisco Bay Area/Monterey, CA
www.weather.gov/sanfrancisco
(831) 656-1710

--
Roger Gass RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/3569692 VLab.Notifications@noaa.gov

Warren Blier, modified 7 Years ago.

Re: [RTMA/URMA Discussion Group - RTMA/URMA Feedback Forum] RE: Max, RTMA/U

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

Hi Steve,

Wanted to follow-up on this. A few things to note:

(1) Yes, it does appear PPSC1 was missing a few hours of obs on Feb 3:

(2) Both this site and D1685 appear to generally be reporting accurately. Large differences in temperatures between the two are characteristic of the microclimate of the area on warm sunny days. Here, take a look at their respective settings:

PPSC1 is down in the Big Sur Valley floor, commonly protected from marine air influence by the ridge to its west. On the other hand, D1685 is at an altitude of about 900 ft directly above the water. What gets even more fun are those summer days when the marine layer gets very shallow -- say 300-400 ft. The difference in temperature between where D1685 is and the beach directly below can then easily be > 20 deg F (over a horizontal distance of ~ 1/2 km). When the NDFD evolves down to ~ 100-m on a side grid pixels we should be able to resolve this. In the meantime, can see where it will be a challenge :)

Warren

On Mon, Feb 5, 2018 at 12:35 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Hi Roger,

It looks like you are talking about site PPSC1. While we get and use data from this station, for some reason we don't have a maxT for it valid on the 3rd (we have it for all other days going back to the 1st). This will happen if we are missing data for a few hours that day - we don't compute a min or maxT for obs where we are missing several hours of data. In that situation, we're stuck with the background value.

For the days we do have it, the background maxT is a good 5-6 F cooler than the observed maximum temperature. This is a common problem in these areas of complex terrain - the HRRR really struggles to get these small valleys/ridges fully cooled/warmed. We are working on a new background error metric that should give individual obs more weight in cases like this, but from what I see over the past few days it hasn't made much of an impact at this site.

I do notice that there is another CWOP site nearby (D1685) about 4 km SE of the RAWS site. That site is reporting 8-10 F cooler than PPSC1 for maxT over the past few days. Also worth noting is the differing terrain height levels we have at the two points. Rpt terrain is metadata-station provided, mdl terrain is the terrain at that point in the RTMA's dataset (in meters):
Station      Lat/Lon         Rpt Terrain   Mdl Terrain
PPSC1 36.25/121.78        108             415.7
D1685   36.22/121.76        246              86.1
Despite the terrain difference, the background maxT values for the two sites are generally within 1-2 F of each other.

It is worth noting that we give equal weighting to all mesonet stations. We are working on ways to change that, but for now they are the same.

In the shorter term, it might be worth trying removing the temperature ob for D1685 from the analysis, and seeing if that has any effect. The terrain difference is large enough that these two obs should not be significantly counter-balancing each other, but they are close enough that I feel like it is worth taking a look at for a few days. If you want, please ask Warren (we want to limit this to SOO's) to fill out the reject list form on the VLab page, and I can flag the temperature and put a trace on temperature at PPSC1 for a few days. If it helps, we can keep the CWOP temperature out. If not, we can put it back in.

Steve

On Mon, Feb 5, 2018 at 2:11 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
RTMA/URMA Team,

We wanted to pass along some concerns with the RTMA/URMA, especially along the Big Sur Coast south of Monterey, CA. During the past week, the RTMA/URMA has been consistently off (too cold) by 9° to 14° in Big Sur, CA where a well established RAWS is located. This type of discrepancy is very common year around at this location, especially during offshore flow.

This is likely due to the complex terrain and influences of the Pacific we deal with here across our forecast area. However, this location in particularly is highly visible due to tourism. Thus, was wondering if there is anything we can do to get the RAWS to be considered a reliable data point in the RTMA/URMA or if something else needs to be considered. As you would guess, this to shows up in the NMB forecast as a cold biases in these situations.

Open to feedback.

Sincerely,

Roger

--
Roger Gass
Lead Forecaster
National Weather Service
San Francisco Bay Area/Monterey, CA
www.weather.gov/sanfrancisco
(831) 656-1710

--
Roger Gass RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/3569692 VLab.Notifications@noaa.gov

--
Steven Levine RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/3569896 VLab.Notifications@noaa.gov

Steven Levine, modified 7 Years ago.

RE: Re: [RTMA/URMA Discussion Group - RTMA/URMA Feedback Forum] RE: Max, RT

Youngling Posts: 174 Join Date: 11/13/14 Recent Posts

Warren,

Thanks for explaining this issue. Clearly, a 20 F temperature change over less than the length of a grid box is something the RTMA (or it's upstream dependencies) will not be able to handle. To be perfectly honest, I am not aware of anything in the pipeline at NCEP that will be able to resolve something like over the next few years.

The outage at PPSC1 is notable, though as you can imagine there is not much we can do here when this sort of temporary outage happens. This is something we are trying to document on the wiki page (though this point isn't up there yet).

Steve

On Tue, Feb 6, 2018 at 1:36 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:

Hi Steve,
Wanted to follow-up on this. A few things to note:

(1) Yes, it does appear PPSC1 was missing a few hours of obs on Feb 3:

(2) Both this site and D1685 appear to generally be reporting accurately. Large differences in temperatures between the two are characteristic of the microclimate of the area on warm sunny days. Here, take a look at their respective settings:

PPSC1 is down in the Big Sur Valley floor, commonly protected from marine air influence by the ridge to its west. On the other hand, D1685 is at an altitude of about 900 ft directly above the water. What gets even more fun are those summer days when the marine layer gets very shallow -- say 300-400 ft. The difference in temperature between where D1685 is and the beach directly below can then easily be > 20 deg F (over a horizontal distance of ~ 1/2 km). When the NDFD evolves down to ~ 100-m on a side grid pixels we should be able to resolve this. In the meantime, can see where it will be a challenge :)

Warren

On Mon, Feb 5, 2018 at 12:35 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
Hi Roger,

It looks like you are talking about site PPSC1. While we get and use data from this station, for some reason we don't have a maxT for it valid on the 3rd (we have it for all other days going back to the 1st). This will happen if we are missing data for a few hours that day - we don't compute a min or maxT for obs where we are missing several hours of data. In that situation, we're stuck with the background value.

For the days we do have it, the background maxT is a good 5-6 F cooler than the observed maximum temperature. This is a common problem in these areas of complex terrain - the HRRR really struggles to get these small valleys/ridges fully cooled/warmed. We are working on a new background error metric that should give individual obs more weight in cases like this, but from what I see over the past few days it hasn't made much of an impact at this site.

I do notice that there is another CWOP site nearby (D1685) about 4 km SE of the RAWS site. That site is reporting 8-10 F cooler than PPSC1 for maxT over the past few days. Also worth noting is the differing terrain height levels we have at the two points. Rpt terrain is metadata-station provided, mdl terrain is the terrain at that point in the RTMA's dataset (in meters):
Station      Lat/Lon         Rpt Terrain   Mdl Terrain
PPSC1 36.25/121.78        108             415.7
D1685   36.22/121.76        246              86.1
Despite the terrain difference, the background maxT values for the two sites are generally within 1-2 F of each other.

It is worth noting that we give equal weighting to all mesonet stations. We are working on ways to change that, but for now they are the same.

In the shorter term, it might be worth trying removing the temperature ob for D1685 from the analysis, and seeing if that has any effect. The terrain difference is large enough that these two obs should not be significantly counter-balancing each other, but they are close enough that I feel like it is worth taking a look at for a few days. If you want, please ask Warren (we want to limit this to SOO's) to fill out the reject list form on the VLab page, and I can flag the temperature and put a trace on temperature at PPSC1 for a few days. If it helps, we can keep the CWOP temperature out. If not, we can put it back in.

Steve

On Mon, Feb 5, 2018 at 2:11 PM, VLab Notifications <VLab.Notifications@noaa.gov> wrote:
RTMA/URMA Team,

We wanted to pass along some concerns with the RTMA/URMA, especially along the Big Sur Coast south of Monterey, CA. During the past week, the RTMA/URMA has been consistently off (too cold) by 9° to 14° in Big Sur, CA where a well established RAWS is located. This type of discrepancy is very common year around at this location, especially during offshore flow.

This is likely due to the complex terrain and influences of the Pacific we deal with here across our forecast area. However, this location in particularly is highly visible due to tourism. Thus, was wondering if there is anything we can do to get the RAWS to be considered a reliable data point in the RTMA/URMA or if something else needs to be considered. As you would guess, this to shows up in the NMB forecast as a cold biases in these situations.

Open to feedback.

Sincerely,

Roger

--
Roger Gass
Lead Forecaster
National Weather Service
San Francisco Bay Area/Monterey, CA
www.weather.gov/sanfrancisco
(831) 656-1710

--
Roger Gass RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/3569692 VLab.Notifications@noaa.gov

--
Steven Levine RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/3569896 VLab.Notifications@noaa.gov

--
Warren Blier RTMA/URMA Discussion Group Virtual Lab Forum http://vlab.noaa.gov/web/715073/home/-/message_boards/view_message/3571596 VLab.Notifications@noaa.gov

RE: Re: [RTMA/URMA Discussion Group - RTMA/URMA Feedback Forum] RE: Max, RT - Forums - RTMA/URMA Discussion Group

Forums