Reducing station pressure to sea level pressure

Su Campu

I've just noticed something strange this morning while looking through the 9am synop reports for Ireland. It seems that the sea level pressures calculated could be off by up to a couple of hectopascals.

Synop reports give both the actual station pressure measure, and the resulting mean sea level pressure (the 3xxx and 4xxxx groups, respectively).

eg.

Valentia (9m amsl)
AAXX 02091
03953 41558 82308 10128 20107 30043 40079 58004 78086 888//
333 84820 88646 92446=

39920 = 992.0 hPa (station)
40022 = 1002.2 hPa (calculated sea level pressure from station pressure)

The problem is that the sea level pressures calculated are in some cases a few hPa off, going by the normal method of calculating. This involves allowing a normal change of around 1 hPa for every 8-9 metres increase in altitude (or around 6 metres in very cold arctic airmasses), and also takes into acount the temperature 12 hours previously. But if you look at the table I've made up below, you'll see that, although in most cases, the pressure fall is around the 8m/hPa mark, for others (Valentia, Belmullet, Finner and Shannon) it is well below that, and in Valentia's case, it's only 2.5m/hPa metres!

So to take Valentia's 9am report, the sea level pressure should be around 1005.5hPa, not the 1007.9 reported. Remember, the only measured pressure is the station pressure; the sea level pressures we see in reports and on charts are calculated, unless the station happens to be precisely at sea level. With an error of 2-3 hPa feeding into a computer model, it is going to cause a real problem in the model's output.

Deep Easterly

Su Campu wrote: »

So to take Valentia's 9am report, the sea level pressure should be around 1005.5hPa, not the 1007.9 reported. Remember, the only measured pressure is the station pressure; the sea level pressures we see in reports and on charts are calculated, unless the station happens to be precisely at sea level. With an error of 2-3 hPa feeding into a computer model, it is going to cause a real problem in the model's output.

I didn't know that Su, that station level pressure was used and the sea level pressure readings are actually caluculated from these readings at each synoptic station.

Do SLP adjustments take into account the actual conditions at the time when a synoptic report is being issued as well as the temperature readings of 12 hours earlier?

Su Campu

Using station pressure is the only way to get sea level pressure, otherwise how would it be possible??!! Even your own station does the same thing - it measure absolute pressure at the site, you have already set its altitude or calibrated it to some nearby station's sea level pressure, therefore it knows what relationship to apply to its measured pressure to give you your sea level pressure.

There are different ways of doing this, and I'm not sure which one Met Eireann use. In the US, the ASOS stations average the current temperature and that 12hrs ago, then extrapolate a column to sea level using an average lapse rate of 6.5 deg per km. Other stations can use empirically-generated methods, so within the same country you have differing methods! But for Ireland, with no high-altitude stations to consider, I'm sure they must use just the one method.

Maybe MT can shed more light...

Deep Easterly

Su Campu wrote: »

Using station pressure is the only way to get sea level pressure, otherwise how would it be possible??!! Even your own station does the same thing - it measure absolute pressure at the site, you have already set its altitude or calibrated it to some nearby station's sea level pressure, therefore it knows what relationship to apply to its measured pressure to give you your sea level pressure.



Maybe MT can shed more light...

I know my own station does this, but my question was in regard to actual official met readings. (should have made this clear). Does the relationship between station and sea level pressure actually change under certain conditions, i.e stable/unstable airmasses that would be taken into account in an official reading? The question is badly asked I admit, but hoping you'll get the gist of it.

Su Campu

In the lowest levels of the atmosphere (surface to ~5000ft) there is a roughly linear fall of pressure with height (~8m/hPa). In the US case above, they approximate the lapse rate to be the standard atmosphere rate of 6.5 deg/km. In a well mixed turbulent atmosphere, such as a warm sunny afternoon, the actual lapse rate in the boundary can be nearer to the DALR of 9.8deg/km, while in stable conditions, such as early morning, it can actually be negative. The reason they average the current and 12hr-ago temperatures is to negate these diurnal effects, so over time everything should even out. But I'd love to know if they do in fact apply say a summer and winter method!

Su Campu

I think I may have an answer to the Valentia problem. Ogimet have the station altitude as 9m (and Met Eireann's website has it as 11m.....so near enough), but when you look at radiosonde data, the surface is taken as 30m, which would correspond to the 3.6hPa difference we see between the surface and sea level pressures (3.6 x 8.5 = ~30m). So I take it the barometer is at a much higher height than the rest of the instrumentation?? I know they release some of the radiosondes from a rooftop launchpad, but it would hardly be 20m higher than the ground would it? Why didn't I go to the IMS fieldtrip there last month!! :mad:

It still doesn't explain the descrepancies of the other stations though...

Mothman

Google street view of Valentia.
A long sloping site. I haven't been there, but it looks like the yellow building down the hill other side of hedge is part of the observatory

Su Campu

Thanks for that. Indeed the site goes from 30m up at the main road down to sea level, so I assume the barometer near the road.

Su Campu

The new met station at Dublin Airport is at around 78m, not the 68m stated on Ogimet, which is the elevation of the old location. This would explain the discrepancy for that observation.

Su Campu

The WMO say that they haven't agreed on a method of sea-level reduction, except in the case of low-level stations (below 50m). As a lot of the Irish stations are above this height, I wonder if we too use different methods for different stations, depending on their altitude.....

From here

3.11 Adjustment of barometer
readings to other levels

In order to compare barometer readings taken at
stations at different altitudes, it is necessary to
reduce them to the same level. Various methods
are in use for carrying out this reduction,but
WMO has not yet recommended a particular
method, except in the case of low-level stations.
The recommended method is described in WMO
(1954; 1964; 1968). WMO (1966) contains a comprehensive
set of formulae that may be used for
calculations involving pressure.

3.11.1 Standard levels

The observed atmospheric pressure should be
reduced to mean sea level (see Part I, Chapter 1)
for all stations where this can be done with
reasonable accuracy. Where this is not possible, a
station should, by regional agreement, report
either the geopotential of an agreed “constant
pressure level” or the pressure reduced to an
agreed datum for the station. The level chosen
for each station should be reported to the WMO
Secretariat for promulgation.

3.11.2 Low-level stations

At low-level stations (namely, those at a height of
less than 50 m above mean sea level), pressure
readings should be reduced to mean sea level by
adding to the station pressure a reduction
constant C given by the following expression:

C = p . Hp/29.27 Tv

where p is the observed station pressure in hectopascals,
Hp is the station elevation in metres, and
Tv is the mean annual normal value of virtual
temperature at the station in kelvins.

Note: The virtual temperature of damp air is the temperature
at which dry air of the same pressure would have the same
density as the damp air. WMO (1966) contains virtual temperature
increments of saturated moist air for various pressures and
temperatures.

This procedure should be employed only at stations
of such low elevation that when the absolute
extreme values of virtual temperature are substituted
for Tv in the equation, the deviation of the
result due to the other approximations of the equation
(used for height rather than standard
geopotential, and with C to be small compared with
P) is negligible in comparison.

EDIT: But they are in the process of approving a recommended method for stations below 750m, in draft form here, which would appear to be the one I mentioned above.

miner eaten

Su Campu wrote: »

I've just noticed something strange this morning while looking through the 9am synop reports for Ireland. It seems that the sea level pressures calculated could be off by up to a couple of hectopascals.

Synop reports give both the actual station pressure measure, and the resulting mean sea level pressure (the 3xxx and 4xxxx groups, respectively).

eg.


39920 = 992.0 hPa (station)
40022 = 1002.2 hPa (calculated sea level pressure from station pressure)

The problem is that the sea level pressures calculated are in some cases a few hPa off, going by the normal method of calculating. This involves allowing a normal change of around 1 hPa for every 8-9 metres increase in altitude (or around 6 metres in very cold arctic airmasses), and also takes into acount the temperature 12 hours previously. But if you look at the table I've made up below, you'll see that, although in most cases, the pressure fall is around the 8m/hPa mark, for others (Valentia, Belmullet, Finner and Shannon) it is well below that, and in Valentia's case, it's only 2.5m/hPa metres!

So to take Valentia's 9am report, the sea level pressure should be around 1005.5hPa, not the 1007.9 reported. Remember, the only measured pressure is the station pressure; the sea level pressures we see in reports and on charts are calculated, unless the station happens to be precisely at sea level. With an error of 2-3 hPa feeding into a computer model, it is going to cause a real problem in the model's output.

I suspect that these errors are the result of a difference between station levels and Barometer levels. For example the station barometer in Shannon Airport is at 20M elevation (on top of the Arrivals building) and not 14M as you give. Hope this is helpful.

Su Campu

miner eaten wrote: »

I suspect that these errors are the result of a difference between station levels and Barometer levels. For example the station barometer in Shannon Airport is at 20M elevation (on top of the Arrivals building) and not 14M as you give. Hope this is helpful.

Thanks for that, I was actually wondering where the Shannon station was. I see that the actual station elevations quoted by WMO are in some cases slightly different that the Ogimet ones, so as you say, that's where the apparent discrepancies arise.

piuswal

Su

Why not ask Met Eireann how they adjust station level pressure to MSL?

miner eaten

Su Campu wrote: »

I've just noticed something strange this morning while looking through the 9am synop reports for Ireland. It seems that the sea level pressures calculated could be off by up to a couple of hectopascals.

Synop reports give both the actual station pressure measure, and the resulting mean sea level pressure (the 3xxx and 4xxxx groups, respectively).

eg.


39920 = 992.0 hPa (station)
40022 = 1002.2 hPa (calculated sea level pressure from station pressure)

The problem is that the sea level pressures calculated are in some cases a few hPa off, going by the normal method of calculating. This involves allowing a normal change of around 1 hPa for every 8-9 metres increase in altitude (or around 6 metres in very cold arctic airmasses), and also takes into acount the temperature 12 hours previously. But if you look at the table I've made up below, you'll see that, although in most cases, the pressure fall is around the 8m/hPa mark, for others (Valentia, Belmullet, Finner and Shannon) it is well below that, and in Valentia's case, it's only 2.5m/hPa metres!

So to take Valentia's 9am report, the sea level pressure should be around 1005.5hPa, not the 1007.9 reported. Remember, the only measured pressure is the station pressure; the sea level pressures we see in reports and on charts are calculated, unless the station happens to be precisely at sea level. With an error of 2-3 hPa feeding into a computer model, it is going to cause a real problem in the model's output.

Correct barometer levels (WMO)

Su Campu

Thanks for that miner, there's a big difference with some of them. Solves the case!! ;-)