WGS84 and the Web

2005-10-18 08:32 | Permanent Link | Standards, Locative

Just when everything is moving along nicely and you think you're making progress, someone is bound to stick up his hand and point out the obvious. In this case, Mike Liebhold's posting on a popular mailing list should give us a little pause in our headlong rush towards geo tagging everything.

Mike went and had a chat with the David Doyle, Chief Geodesist of the US National Geodetic Survey (and not related to me as far as I know) about WGS84. Basically, WGS84 is the name of the reference system provided by GPS satellites. It provides an accuracy of a meter or so. Apparently, in about 2015, there's going to be a new Global Navigational Satellite System capable of giving us greater accuracy.

Mike wonders whether we need to worry about the fact that there's a new system coming that might obsolete all the data we've collected so far. In response, Josh Lieberman and I pointed out that tectonic drift is possibly a bigger destabilizing factor than simply changing to a more accurate reference frame.

In a fairly decent web page provided by the Ordnance Survey, they point out that Britain is drifting north-east at a rate of about 2.5cm per year and that New Zealand is moving about 4 times as fast in (I'm guessing in a different direction).

Meanwhile, back on the Web, we're busily coming up with schemes that will let us put location information into our web posts, blogs, etc. and the obvious answer is to use WGS84 since it's not tied to any specific country's local datum and it seems that the Galileo satellite navigation system will also provide WGS84 coordinates.

The geodesists are way ahead of us on this. If you read far enough down the abovementioned web page provided by the Ordnance Survey there is this little gem:

"This adoption of a particular WGS84 epoch to remove the effect of tectonic motion has been done in various places in the world - in fact, everywhere WGS84 has been adopted for mapping. Examples of WGS84-like datums, which are gradually diverging from WGS84, are North American Datum 1983, New Zealand Geodetic Datum 2000 and the Geocentric Datum of Australia."

Mike later asked me why there did not seem to be much of a reaction to his concern. Do people think it's a non-issue? My initial response was to agree that people don't think it's a big issue. But maybe that's too easy. Perhaps we should think about the implications a little more. If we use "raw", i.e. "current" WGS84, we're really ignoring the drift and other issues. If we use any of the derived, named datums, pegged to an epoch, we lose universality.

How do people in New Zealand feel about having their geotagged web data staying behind as they sail around the globe on their tectonic plate?

Comments

2005-10-18 16:28 | Posted by Allan Doyle

Sean Gillies has posted a response in his blog: http://zcologia.com/news/96 and James Fee picked them both up at http://www.spatiallyadjusted.com/2005/10/life_after_wgs84.html

2005-10-20 00:17 | Posted by Marc

I would think that this issue would primarily be of concern to those applications where accurate sub-centimeter precision over a wide area and over time would be critical. For the vast majority of geospatial applications this wouldn't seem to be a particularly big deal. I would suspect that corrections from one geodetic reference system to another would be possible in many cases given a sufficient number of control points. Ensuring that sufficient metadata is included with any coordinate information (e.g. date of data collection, source of coordinates, reference system used, etc.) would help to mitigate any problems in the future use of such location data.

A related issue is that vertical datums are even far less precise that horizontal datums. Geoid models are based on gravity models which don't map well to the actual surface of the earth (due to varying local density of the Earth's crust). Heck, with the rising ocean level, a few centimeters drift due to plate tectonics should be the least of New Zealand's worries ;-)

2005-10-20 05:02 | Posted by Gerhard | www.dotGIS.de

The city limits of Fort Collins have approximately a diameter of 20km (12.6mi). Giving a point location by geographic identifiers as proposed by Sean does not increase any accuracy. It may be different if you refer to the south-east corner of the city hall building. But even then, how precise is that point location defined? This kind of reference is actually given by a coordinate reference system. The CRS is realized by a number of well-defined "fixed" points that are tracked permanently. The providers of GPS are aware of the fact, that their reference system is not as "fixed" as we assume when dealing with global coordinates. That's why they give a date with the name of a geodetic datum: WGS 1984.

The fact that the uncertainty of absolute GPS measurements is still in the range of several meters prevents us from observing the effect, that objects have actually moved between different observation epochs. For high precision survey we are using differential GPS, i.e. determining relative coordinates to points in the proximity of the observed point. The reference points in that case have moved together with the observed points on the same tectonic plate.

Once GPS or Galileo will be good enough to measure cm-accuracy, we should consider recording a time stamp with each global coordinate we measure. From the system monitoring data (like WAAS or EGNOS) it will be possible to calculate a more precise space time location at any epoch.

You don't have to go as far as New Zealand to see that effect. The observation that the 50th state Hawaii is drifting away form the US homeland by 5cm a year should not be interpreted politically.

2005-10-23 10:25 | Posted by XenonofArcticus | http://xenon.arcticus.com

I think it's a tempest in a teapot. We've dealt with multiple datums for a long time. There will be new datums in the future. Any future navigation system will be able to reproject its output into WGS84. If the web collectively decides to adopt a new datum, then new syntax can be added to the XML entities in quesiton to specify which datum is used -- lack of the new tokens will imply old-style implicit WGS84. Remember, the X in XML is eXtensible. What's the big deal? If we are really concerned about it, define the DATUM token now and start filling it in with WGS84.

2005-11-04 07:04 | Posted by Zoran Kovacevic | http://www.kovacevic.nl/blog

How about backtracking to a location noted in WGS84 using the date the location was published? Surely the (continental) drift in time is available :)

2005-12-01 04:09 | Posted by Andrew Russell

I was wondering why my neighbours garage seemed to be a little closer to my boundary this year, thanks for clearing that one up. I suppose what you lose on the left you gain on the right.
But seriously, NZ is tectonically active! about 100 years ago wellington's harbourside waterfront was raised 20m, which allowed us to put our Highway 1 somewhere. I hope I am not driving on it when it goes splosh back into the briny. So there are larger issues here. Similar horisontal displacements are found occationally like in the Napier earthquake (Sometime in the mid 1900's).
So, theoretically we need a spaciotemporal geodetic datum (???) that can relate any historic point on the earth to Day zero locations (maybe the year 2000, NZ really rocked the planet that new years eve). That would probably require scalable accuracy levels, i.e. only accurate to within 1cm per 100years or something (remember this is probably 10yrs out), because the entire earth mapping would be considerably more complex and large than the simple spheriods that drove me batty a couple of months ago.
And point to note, as far as I know NZ typically uses NZGD49 / New Zealand Map Grid still and not NZGD2000, although my experience is very limited.
And the only GIS work I have done relied on everything being mapped to WGS84 for Google Earth among other things.
Andrew

2007-12-20 10:51 | Posted by Bruce | http://www.gis.co.nz

Re: How do people in New Zealand feel about drifting away on our tectonic plate:

We're Ok about it, because we're heading off towards Tonga where the fishing is great, while Australia's fate, poor sods, is to be subducted under Indionesia. Tish tosh...

Comments are closed for this entry.