altitude for barometer calibration

[Greg Troxel]

tl;dr: What elevation is barometric reduced to? 0 meters in NAVD88?
Some other height datum? What about other countries, WMO, etc?

I'm going down the rabbit-hole of barometer calibration. I'm trying to
address it in two parts: calibrating station pressure, and then using
the correct altitude. This is part trip report and part question.

I realize that my question has close to zero practical value. But nerds
like to measure things correctly anyway.

For station pressure, I have a VP2, a Brunton "Atmospheric Data Center"
that measures pressure (also T/H), a BME280, and a BME680.
Provisionally treating the Brunton as a correct reference (because it's
portable, so I can move it to my various sensors and at some point take
it to visit other people's stations), I find the following offsets:

VP2 +0.1 hPa
BME680 +2.5 hPa
BME280 +1.5 hPa

It's encouraging that the Brunton and VP2 are so close (resolution 0.1
and 0.01 hPa respectively), but of course they could be off the same
amount. It is clear that that BME280 family isn't that accurate, but
that's no surprise given the cost and intent. However the offsets are
pretty stable.

For reference

- a height change of 1 meter corresponds to 0.12 hPa. (That's 0.09
mils of Hg per foot :-)

- WMO seems to specify 0.1 hPa as the required uncertainty for field
instruments.

For altitude, I borrowed some 25-year old survey-grade GPS equipment,
and am in the process of measuring from that mark to where my console
is. The data is natively in "NAD83(2011) epoch 2010.0 Height above
Ellipsoid", but I can transform to NAVD88 or to "WGS84 Orthometric
Height". I expect to end up with a height value that is accurate to
about 10 cm in HAE and not too much more in orthometric heights.

My questions are:

- When people talk about reducing a station pressure reading to "sea
level", what do they mean?

- Specifically, what does NWS mean? Given the overall US government
approach, I would guess they mean "0m elevation in NAVD88".

- What is the situation with other countries' national weather
organizations?

- What is the situation with the WMO? It seems there is an idea that
each country might specify a height reference more precisely and is
to report it. There is a hint of EGM96 in CWOP-WM08 -- but EGM96
has long since been superceded. Hence mentioning "WGS84 Orthometric
Height" above.

I would appreciate not only an answer but a pointer to where NWS, other
national agencies, and WMO specify this.

(As a side note for people who find my question boggling, "sea level" is
a concept that really only applies to a particular tide gauge station,
and the elevations of different stations' zero points are different.
Referring to "sea level" in a general way introduces fuzz of around a
meter.)

Greg

[vince]

Greg - does https://www.usna.edu/Users/oceano/pguth/md_help/html/vert_datum.htm help any ?

There are some links at the bottom of the article to NOAA info as well...

[Susan Mackay]

Also doesn't it vary with temperature?

The formula I use is at  Conversion to sea-level pressure Calculator - High accuracy calculation (casio.com) (which is also used in a number of other sites).

Susan

[Greg Troxel]

vince <vince...@gmail.com> writes:

> Greg -
> does https://www.usna.edu/Users/oceano/pguth/md_help/html/vert_datum.htm
> help any ?

No, but it's a great short introduction for people not familiar with the
subtleties of height meaurement. (There are a few minor errors, too:
chart datum is not always MLLW - some US regions use MLW. And I've
never heard of orthometric heights from geoid models called geodetic
before.)

While the various datums are described, it doesn't answer:

Which elevation and which vertical datum does NWS mean when they say
say "sea level".

I am guessing it is "0m NAVD88", but I am just guessing. My second
guess is "0m WGS84 Orthometric Height as defined by EGM2008", to align
with what WMO might want.

> There are some links at the bottom of the article to NOAA info as well...

One of those says

It should be noted that due to effects such as atmospheric pressure,
temperature, prevailing winds and currents, and salinity variations,
MSL can depart from an equipotential surface by a meter or more.

which is what I was getting at in saying that mean sea level, without
qualifications, is a fuzzy concept.

[Greg Troxel]

Susan Mackay <vk3...@gmail.com> writes:

> Also doesn't it vary with temperature?
> The formula I use is at Conversion to sea-level pressure Calculator - High
> accuracy calculation (casio.com)

> <https://keisan.casio.com/exec/system/1224575267> (which is also used in a
> number of other sites).

Yes, for barometric pressure reduction temperature matters.

When weewx does the reduction, the formula is:

https://github.com/weewx/weewx/blob/master/bin/weewx/wxformulas.py#L266

[gszla...@gmail.com]

Looks like you've got the elevation for your location nailed down if you've gotten it down to 10cm. For us ordinary folks we have to rely on topo maps or google earth ( or similar map reference tools) to get down to 1 meter accuacy. Not all countries have gone digital for surveys quite yet.

Here is some general concepts to consider when you are calibrating/setting your barometer:

             

1. Absolute pressure (Abs) is the current air pressure/atmospheric pressure measured at your elevation.

2. Relative pressure (Rel) represents what the atmospheric pressure would indicate if your station was theoretically located at sea level. This is done because it provides a common barometric reference for all weather stations regardless of their elevation.

For our weather stations we usually assume station pressure = atmospheric pressure (raw pressure) measured/observed at the elevation of sensor.

The term "relative pressure" is not used so much these days but this is the SLP you see in METAR reports. Basically, The station pressure is theoretically "reduced" down to sea level elevation. Note: it sis easy to get confused here with the terminology. You are not reducing pressure you are reducing elevation to mean sea level. Yes, high tides, low tide is taken in account.

The calibration process is usually two steps.

1. correct for sensor error

2. set your barometer to your elevation.

For a Davis VP2, SLP is already calculated for you in the console so all you need is the elevation of your sensor. Hwever you still need to check if your barometric sensor is accurate or not,  The BMP280, BME680 are accurate to 1 hPa and so is the VP2 barometric sensor. The newer Bosch barometric sensor (BMP390) is accrate to 0.5 hPa.

You could calibrate your VP2 with the higher spec BMP390. Just match the pressures and you are done! or alternatively you can match Altimeter with a close-by METAR.

So no, we will not be achieving 0.05 to .10 hPa accuracy WMO standards with our consumer grade personal weather stations!

For Ambient Weather/Ecowitt weather stations, SLP or Altimeter (setting) is estimated  by using a fixed offset ( a fixed amount correction) to Absolute pressure/station pressure.

Keep in mind that different countries use different pressures. Most countries don't use SLP or Altimeter - they use QNH in whole integer amounts which makes matching difficult as you would have to match 1013, 1014, 1015, etc. If you luck out, some countries publish decimal QNH like the BOM in Australia.

I couldn't answeryour questions regarding digital survey/elevation but none of the sensors we're talking about come even close to 10 cm accuracy. as these sensors have "noise", drift, , etc. You can only do so much with a $10 sensor. but yeah, it can be a rabbit hole for sure.

[Greg Troxel]

thanks for the comments. I will be digesting all of the replies and
CWOP data and coming up with a cal plan.

"gszla...@gmail.com" <gszla...@gmail.com> writes:

> For a Davis VP2, SLP is already calculated for you in the console so all
> you need is the elevation of your sensor. Hwever you still need to check if
> your barometric sensor is accurate or not,

Agreed.

> The BMP280, BME680 are accurate
> to 1 hPa and so is the VP2 barometric sensor. The newer Bosch barometric
> sensor (BMP390) is accrate to 0.5 hPa.

My data so far indicates that a BME280 and a BME680.
BME680 is +2.4 hPa from VP2.
BME280 is +1.4 hPa from VP2
Brunton ADC is -0.1 hPa from VP2.

So I don't believe those claimed accuracies. I am tentatively
believing that VP2 and Brunton ADC are correct until I can take the ADC
to visit other barometers.

> You could calibrate your VP2 with the higher spec BMP390. Just match the
> pressures and you are done! or alternatively you can match Altimeter with a
> close-by METAR.

I am going to try to compare my barometric and altimeter pressures with
NWS stations on a day with a low rate of change.

> So no, we will not be achieving 0.05 to .10 hPa accuracy WMO standards with
> our consumer grade personal weather stations!

VP2 seems pretty stable so I think 0.2 hPa is not crazy.

> Keep in mind that different countries use different pressures. Most
> countries don't use SLP or Altimeter - they use QNH in whole integer
> amounts which makes matching difficult as you would have to match 1013,
> 1014, 1015, etc. If you luck out, some countries publish decimal QNH like
> the BOM in Australia.

I thought QNH was Altimeter. The US uses that in aviation and SLP or
barometric pressure for weather. But yes, roudning to whole hPa makes
it hard.

[gszla...@gmail.com]

If you are comparing your external sensors to your console's barometer keep in mind that the offset does not mean the sensor is out-of-spec. An absolute offset is not a specification. It is just the sensor correction offset so that all your barometers read the same absolute pressure. I am assuming you using the console as your reference and you have calibrated it with a close-by airport using METAR or mesowest readings or equivalent.

To answer your question if QNH = Altimeter? Altimeter includes a landing gear offset of 10 feet = 0.3mb/.01 inHg. I believe that QNH does not have that peculiar offset. Although most sources equate the two, I think the equations are different.

Lastly you were asking about different definitions of SLP?

It is easy to get confused here because it is confusing.

I know of at least 6 different equations for Altimeter and there might be dozens of different country-specific equations for (M)SLP.

To make things more interesting both Altimeter and (M)SLP can be referred to as "SLP".as both are reduced to sea level elevation.

As you are aware, if you hook up your console to WeeWX, WeeWX will automatically calculate Altimeter and (M)SLP using the proper algorithms. Need to adjust the current algorithm for your region?  You can modify it.

[Greg Troxel]

"gszla...@gmail.com" <gszla...@gmail.com> writes:

> If you are comparing your external sensors to your console's barometer keep
> in mind that the offset does not mean the sensor is out-of-spec. An
> absolute offset is not a specification. It is just the sensor correction
> offset so that all your barometers read the same absolute pressure. I am
> assuming you using the console as your reference and you have calibrated it
> with a close-by airport using METAR or mesowest readings or equivalent.

What I have done so far is that I entered an elevation that I knew was
very approximate, and watched the barometric pressure compared to
official readings at airports around me, and on days when the pressure
was steady and similar adjusted by elevation (because that's all you can
do on VP2 console) to get a matching value. I knew that was bogus
because it's doing station pressure calibration by having the wrong
altitude, but it got me decent values.

However what I should be doing is having a calibration offset for
station pressure, and then using the correct altitude (which I now
know). The pressure sensor has some spec, and if the calibration
magnitude is higher than that, then yes the device doesn't meet spec.
The VP2 is specified to +/- 1.0 hPa. If the calibration offset were 1.5
hPa, then it would be out of spec. But my best guess this minute is
that I need to add 0.1 hPa to the station pressure to get the true
value.

> To answer your question if QNH = Altimeter? Altimeter includes a landing
> gear offset of 10 feet = 0.3mb/.01 inHg. I believe that QNH does not have
> that peculiar offset. Although most sources equate the two, I think the
> equations are different.

Interesting. I'll try to understand this better, but I suspect that the
altimeter pressure used by CWOP (US again) and calculated by weewx does
not have this offset.

> Lastly you were asking about different definitions of SLP?

No, I was asking "what does sea level mean, when weather people say
'pressure reduced to sea level'". How to reduce from a known elevation
is less confusing.

> It is easy to get confused here because it is confusing.

Yes, and particularly because height and sea level are both very
difficult subjects if you want to get them right.

> I know of at least 6 different equations for Altimeter and there might be
> dozens of different country-specific equations for (M)SLP.
> To make things more interesting both Altimeter and (M)SLP can be referred
> to as "SLP".as both are reduced to sea level elevation.

They are, but in US usage "Sea Level Pressure" means "barometric
pressure" meaning a reduction equation that uses temperature.

> As you are aware, if you hook up your console to WeeWX, WeeWX will
> automatically calculate Altimeter and (M)SLP using the proper algorithms.
> Need to adjust the current algorithm for your region? You can modify it.

Sure, but with VP2 it defaults to using hardware calculation for
barometric pressure. I thought station pressure was also hardware, but
https://www.weewx.com/docs/hardware.htm#vantage_notes says S. That may
be the next missing link in my understanding.