code table 4.2: surface adjusted wind (UK)

[marqh]

Summary and purpose

Request for a new code table 4.2 entry set for surface adjusted wind

Action proposed

The Team is requested to review the proposed new parameter and approve it for implementation.

Discussions

The UK Met Office generates data from its atmosphere models with a model diagnostics of surface adjusted wind, which do not appear in the current GRIB parameter code tables. The respesentation this quantity within WMO GRIB2 tables would assist in the sharing of this data.

The atmosphere model generates values which are intended to be directly comparable to observed 10m winds, equivalent to BUFR table b entries

• 11 81 Model wind direction at 10 m
• 11 82 Model wind speed at 10 m

The naming could use the term surface adjusted, or could include the explicit text '10m'

Detailed proposal

• U
  • Discipline: 0
  • Category: 2
  • Name: u-component of wind (surface adjusted)

Full details are given in the attached document
Code.table.4.2.-.Parameter.number.by.product.discipline.and.parameter.category.-.Operational-1.docx

* Units: m/s

• V
  • Discipline: 0
  • Category: 2
  • Name: v-component of wind (surface adjusted)
  • Units: m/s
• speed
  • Discipline: 0
  • Category: 2
  • Name: Wind speed (surface adjusted)
  • Units: m/s
• direction
  • Discipline: 0
  • Category: 2
  • Name: Wind direction (from which blowing) (surface adjusted)
  • Units: degree true

[marqh]

@richardweedon
@mgange659

[sebvi]

I think these can be achieved using a type of first fixed surface set to "height above ground in metres" (103) and set the level value at 10. This is why I am not in favour of creating new parameters

[sebvi]

To be more explicit, the usual way to encode these parameters is:

10 m u component of wind:

GRIB key	Code table	Value	Meaning
discipline	0.0	0	meteorological products
category	4.1	2	momentum
parameter	4.2.0.2	2	U component of wind
first fixed surface	4.5	103	specified height above ground
scale factor	n/a	0	multiply by 10^-0 --> 1
scaled value	n/a	10	10 metres (10 * 10^-0 )

10 m v component of wind:

GRIB key	Code table	Value	Meaning
discipline	0.0	0	meteorological products
category	4.1	2	momentum
parameter	4.2.0.2	3	V component of wind
first fixed surface	4.5	103	specified height above ground
scale factor	n/a	0	multiply by 10^-0 --> 1
scaled value	n/a	10	10 metres (10 * 10^-0 )

10 m Wind speed

GRIB key	Code table	Value	Meaning
discipline	0.0	0	meteorological products
category	4.1	2	momentum
parameter	4.2.0.2	1	wind speed
first fixed surface	4.5	103	specified height above ground
scale factor	n/a	0	multiply by 10^-0 --> 1
scaled value	n/a	10	10 metres (10 * 10^-0 )

10m wind direction

GRIB key	Code table	Value	Meaning
discipline	0.0	0	meteorological products
category	4.1	2	momentum
parameter	4.2.0.2	0	wind direction
first fixed surface	4.5	103	specified height above ground
scale factor	n/a	0	multiply by 10^-0 --> 1
scaled value	n/a	10	10 metres (10 * 10^-0 )

[marqh]

Hello @sebvi

thank you for the feedback.
We are a bit concerned about the approach you are proposing here for encoding these parameters.

Our model provides wind outputs at all levels within the model.
We use the approach you have detailed to encode data for the standard model wind parameters.

Our model also includes specific code which implements surface adjustments to provide a specific surface adjusted wind parameter set.

So, there are two alternative formulations for these wind parameters in our model. We are already using the settings described for the ‘standard’ wind diagnostics from the model. We want to have alternative settings for the ‘surface adjusted’ versions of these 4 parameters so that downstream customers can access both sets and differentiate between the differently derived parameters.

These are not deemed to wind parameters which happen to be at a vertical level of 10m. These are specifically adjusted parameters to represent wind at the surface. They are different parameter codes in our model.

There are specific BUFR codes for this concept:

• 11 81 Model wind direction at 10 m
• 11 82 Model wind speed at 10 m

but no specific GRIB2 concepts.

thank you for your consideration
mark

[sebvi]

Hello @marqh,

if I understand you correctly, the model is producing 10m wind parameters in 2 different ways but at the end, it is still the same parameters computed/derived. If it is 2 different processes you should be able to distinguish the two sets using the process identifiers octets that are made for this I believe.
We have similar cases at ECMWF, for instance we produce "ozone mass mixing ratio" using different atmospheric composition schemes but we are not going to propose several ozone entries in the chemical constituent table, ozone is ozone!
This is why I am not really in favor of this proposal at the moment. But I am not the one deciding, we need the point of view from others in the team and from @efucile .

[richardweedon]

Request the following entries -
Product discipline 0 - Meteorological product parameter category 2: momentum 47 U-component of wind (surface adjusted) M/S
Product discipline 0 - Meteorological product parameter category 2: momentum 48 V-component of wind (surface adjusted) M/S
Product discipline 0 - Meteorological product parameter category 2: momentum 49 Wind Speed Surface Adjusted M/S
Product discipline 0 - Meteorological product parameter category 2: momentum 59 - 191 Reserved

[efucile]

Hello, @marqh @richardweedon and @sebvi
Here is my opinion. A surface adjusted wind speed is still a wind speed. I mean, the parameter or the observed quantity is the same, and therefore there should not be the need to add another entry. However, I understand that there is the need to convey the message to the users that there are two wind products and they are different. Unfortunately, we don't have any element in GRIB2 to express this difference, and therefore I think that we have to accept the compromise and add these extra entries in table 4.2.
I also have another comment. As a user, if I read surface adjusted, I am not sure of the meaning and the aim of the adjustment. I believe that we need a clear description of these elements, without going into the specific process of adjustment which may change in the future, but expressing clearly what is the purpose of the adjustment.

[sebvi]

@efucile I thought the way to convey the particularity of these parameters would be the background generating process identifier keys which are freely defined by the data producer.

[richardweedon]

Apologies all this is out of my comfort zone so please bear with me. I have attempted to summarise a rather long email from my colleagues -

A short explanation is:
10m winds should be virtually identical in the UK models
Single level 10m wind in the global models will be stronger than the 10m multi-level wind, as the former attempts to undo some spurious slower down in wind from overzealous turbulent orographic form drag parametrization.

The aim being -
10m wind is re-derived from the diagnosed surface turbulence to provide a more accurate prediction. This addresses the spurious slowing of the resolved wind profile at low levels that the parametrization of turbulent orographic form drag can introduce in some configurations of the model

[amilan17]

@marqh @richardweedon It looks like the conversation is still open and will not be ready for FT22-1. If you need this sooner, then please finalize the proposal within a couple days and I will create a branch.

[amilan17]

@marqh @richardweedon @sebvi -- Will this proposal be ready for FT2022-2 before June 8?

[amilan17]

to be determined for during another fast-track cycle

[amilan17]

https://github.com/wmo-im/CCT/wiki/Teleconference-4.10.2022  meeting notes: not discussed

[amilan17]

https://github.com/wmo-im/CCT/wiki/Teleconference-21.11.2022 note: not discussed