GPS "Accuracy" AGAIN

[+Just Roger]

 

 

Which would you believe? One of them has to be 325yds wrong and I know which one I'd put my money on.

[+HouseOfDragons]

 

Which would you believe? One of them has to be 325yds wrong and I know which one I'd put my money on.

 

Surely if you allow for different datum <,insert dim emoticon here as I'm not sure what I'm talking about>> and the not pin-point accuracy of a GPSr, they can both be right.

[+The Forester]

In Latitude, my money's on you, Roger.

 

In Longitude, my money would be on the GPS getting it wrong by a lot more than a hundred metres at that Latitude.

[+fat bloke]

 

 

Which would you believe? One of them has to be 325yds wrong and I know which one I'd put my money on.

 

Been there Done that (Kenya) refused the Earth cache http://img.geocaching.com/cache/log/displa...g?rnd=0.9393534

[+The Patrician]

Some info here on the datums used on maps in Africa:

 

http://www.mail-archive.com/mapinfo-l@list...m/msg02339.html

 

"The computations of the adjustment of the Arc Datum 1950 resulted

in a series of relations that all can presently be related to the WGS84

Datum:

 

>From Arc50 to WGS84;

 

Zaire: dX = -169m, dY = -19m, dZ = -278m

 

>From Arc60 to WGS84;

 

Kenya: dX = -157m, dY = -2m, dZ = -299m

 

Tanzania: dX = -175m, dY = -23m, dZ = -303m

 

>From Cape to WGS84;

 

S. Africa: dX = -136m, dY = -108m, dZ = -292m"

 

If I had to guess I'd say it is more likely that the GPS is giving the correct WGS84 equatorial position than the monument, which was probably positioned using a sextant and some arithmetic, and then shifted slightly because they couldn't dig a rock out of the ground, the local leopard was giving them a funny look, and they were in a hurry because the bar was open.

 

Bearing in mind that the earth is a lumpy oblate spheroid and WGS84 is sort of massaged to fit the whole world pretty well, rather than fit a particular part of the globe extremely well; that's why the OS use something else*. Unless of course the monument was erected by one of those amazing Victorian surveyors who could get the height of Everest right(ish) from 100 miles away.

 

Regards

 

Neil

 

*I know the OS pre-dates WGS84, but you know what I mean!

[+The Forester]

I think the rock/leopard/bar explanation is the best fit.

 

The Geoidal Height, ie the difference between the geoid and the spheroid at that location, is +17.01m for WGS84 (the one we use) and -95.05m for the Clarke 1880 spheroid which is quiet certainly the one which was used for the calculation of the location of this survey monument.

 

That wouldn't account for the 298m error shown by the GPSr in the picture. The height at that location is 945m above the geoid, but we don't know what orthometric height the traverse from Cape Town (or wherever) had produced. I would doubt very much that it was out by as much as 200 metres of height, even that far into the hinterland.

 

Measuring Latitude astro-geodetically is a piece o' pish with a theodolite. Digging the foundations for a tourist monument, with rocky ground and leopards and dusky bar-flies who urgently crave your indulgence in certain "Ugandan discussions" while you blow the froth off an ice-cold pint of Tusker, is a quite different matter.

[+The Patrician]

I think the rock/leopard/bar explanation is the best fit.

 

 

Me too!

 

Regards,

 

Neil

[+Just Roger]

Some info here on the datums used on maps in Africa:

 

 

If I had to guess I'd say it is more likely that the GPS is giving the correct WGS84 equatorial position than the monument, which was probably positioned using a sextant and some arithmetic, and then shifted slightly because they couldn't dig a rock out of the ground, the local leopard was giving them a funny look, and they were in a hurry because the bar was open.

 

 

*I know the OS pre-dates WGS84, but you know what I mean!

 

That's pretty much what I thought. Consider the following:-

 

I'm assuming that the monument has been there for some 50-100 years (Definitely pre-GPS.)

 

Using a sextant needs a horizon and this is hundreds of miles from the sea. Other alternative is to use a sextant and bring the object down to its reflection in a dish of mercury (Eat your heart out HSE) and divide the result by 2, but that means that the largest elevation that can be measured is only about 50º.

 

A noon sight for Latitude us next to useless at the equator and is impossible using the HSE's favourite method above

 

Polaris is equally useless as it would be on the horizon hiding behind a tree or an ant hill.

 

The nearest bar is probably not an issue as now it is only a couple of miles away - but only since the Queen's visit in the 50's and the nearest town is 20 miles away - a days donkey ride (unless persued by the local leopard when it might drop to a couple of hours)

[+The Patrician]

 

Using a sextant needs a horizon and this is hundreds of miles from the sea. Other alternative is to use a sextant and bring the object down to its reflection in a dish of mercury (Eat your heart out HSE) and divide the result by 2, but that means that the largest elevation that can be measured is only about 50º.

 

 

You can (or could) get sextants with artificial horizons (mainly for use in aircraft), a theodolite would do just as well. Any of the planets, stars or other bodies listed in a nautical almanac would give an accurate fix.

 

Regards,

 

Neil

[+Bear and Ragged]

 

The nearest bar is probably not an issue as now it is only a couple of miles away - but only since the Queen's visit in the 50's and the nearest town is 20 miles away - a days donkey ride (unless persued by the local leopard when it might drop to a couple of hours)

 

So, the issue is:

The beer/whisky is in a pack on the donkey, which is being eyed-up by the leopard in the bush.

(the donkey is being eyed-up by the leopard, not the alcohol!)

 

"They'll never think of making a hand-held gizmo to check we've got this thing in the correct spot..."

[+fat bloke]

 

The nearest bar is probably not an issue as now it is only a couple of miles away - but only since the Queen's visit in the 50's and the nearest town is 20 miles away - a days donkey ride (unless persued by the local leopard when it might drop to a couple of hours)

 

So, the issue is:

The beer/whisky is in a pack on the donkey, which is being eyed-up by the leopard in the bush.

(the donkey is being eyed-up by the leopard, not the alcohol!)

 

"They'll never think of making a hand-held gizmo to check we've got this thing in the correct spot..."

I'll go for that one! all this techno mumbo jumbo

Why make it hard

[+The Patrician]

 

So, the issue is:

The beer/whisky is in a pack on the donkey, which is being eyed-up by the leopard in the bush.

(the donkey is being eyed-up by the leopard, not the alcohol!)

 

 

I treasure the memory of a sign I saw in Africa:

 

"Please do not feed the leopards, it turns them into dangerous beggars"

 

Regards

 

Neil

Edited September 25, 2009 by Neil Pugh

[+The Forester]

Other alternative is to use a sextant and bring the object down to its reflection in a dish of mercury (Eat your heart out HSE) and divide the result by 2, but that means that the largest elevation that can be measured is only about 50º.

 

60° actually. That's why it's called a sextant.

 

If you haven't got a dish of mercury to hand (very heavy stuff to lug around), just paint the inside of a bucket matt black and fill it with water. It's a trick I teach in my Advanced Land Navigation Course. You can also use the same trick with a theodolite, which is one or two orders of magnitude more precise than a sextant, but then you are limited to stars which are quite low above the horizon because of the time lag in the two-part measurement. I use the trick with a theodolite rather than a sextant as it's easier to demo if I'm teaching a group of four students instead of just one.

 

A noon sight for Latitude us next to useless at the equator and is impossible using the HSE's favourite method above

Polaris is equally useless as it would be on the horizon hiding behind a tree or an ant hill.

 

You can use any star you can identify. Just look up its Declination in the tables. Calculating Latitude is a doddle. Just measure the vertical angle from the horizontal and subtract it from the Declination of the star. Voila: Latitude.