JAD69 and the GPS
topographical maps will soon discover that when
it comes to latitude and longitude all is not as it seems. Mankind’s models
of the earth, upon which all maps are dependent, were until recently
determined by long baseline land surveys. Given the difficulties involved,
every geodesist arrived at a different result, in fact at times two different
results, (Clarke1866, Clarke 1880). Satellites have changed all of that.
Today, the world standard, the most accurate, most geocentric ellipsoid, is
known as World Geodetic System 84 and this is the datum that the GPS uses.
Although Jamaica now has a new datum, JAD2001, that is based on the WGS84 ellipsoid, the 1:50,000 Jamaica metric topo maps use an older datum called JAD69.
JAD69, the Jamaican datum, is based upon a non-geocentric Clarke ellipsoid, (example of ellipsoid), established in 1866 that in Jamaica differs from WGS84 by hundreds of metres. The datum shift, (example of datum shift), causes the natural origin of the Jamaica Metre Grid, 18º N, 77º W, to be situated apx. 311 metres to the NNE of what a GPS will call the same latitude and longitude. In short, the latitudes and longitudes found on the Jamaican topos must be converted to WGS84, or the GPS datum must be user-defined to JAD69, to have the two agree. This presents us with a challenge.
At this time, no handheld GPS units include JAD69 in the datum list and the user-defined datum parameters are poorly defined. In fact, the author of this article currently has three different sets of parameters. The best of them, when run through Geotrans, is inaccurate when compared to known points. Because of this, it is not advisable to collect positional data in user-defined JAD69 coordinates. The better solution is to collect data in WGS84 and then, when necessary, do a datum transformation to JAD69.
In the past, the JAD69 datum parameters seem to have been regarded as a "state secret" by certain departments of the Jamaican government, and were not freely available, but the recent publishing of a new set has established a precedent, (the cat is out of the bag). I offer the following walk-through of the datum transformation in hopes that it will be of help to others; it represents the cheapest solution for those using a good GPS who would like to plot positions on the 1:50,000 Jamaican metric topo maps. All three sets of received datum parameters are given. None of the transformed JMG Eastings and Northings should be assumed to have an accuracy greater than +/- 10 m.
I have assigned the three sets of received parameters the names JAD69_1, JAD69_2, JAD69_3, for my own purposes. They are numbered in the order by which I received them. All use the Clarke 1866 ellipsoid, (the 1:12,500 topo series uses an older datum based on Clarke 1880 - conversion method here).
A selection of Jamaican datum parameters (and your best bet is number 3):
JAD69_1: dx 70.0892, dy 197.3584, 393.0896 (received from a Jamaican Govt source)
JAD69_2: dx 69, dy 207, dz 389 (received from someone who has a friend at ESRI)
JAD69_3: dx 65.33 +/- 0.96 m, dy 212.46 +/- 1.49 m, dz 387.63 +/-0.69 m, published by Mugnier (.pdf file) from Newsome and Harvey
To use these numbers, download Geotrans, and then, in Geotrans, create a new datum with a Clarke 1866 ellipsoid, (CC). Enter the dx, dy, dz, and make the valid domain a few degrees either side of 77 W, 18 N. Click OK. You can also accomplish this by replacing the existing file "3_param.dat" with this file. Save the old one with an altered file name in case something goes wrong. The new datum list will show JAD_1, JAD_2, and JAD_3.
In the lower window, set the datum to the one you've just created, then set the projection to Lambert Conformal Conic. Enter 77 W for the meridian, 18 N for the origin latitude, and 18 N for both standard parallels. Make the False Easting 250000 and the False Northing 150000. Enter your GPS lat/long positions into the upper window, with the datum at WGS84, (WGE), then convert upper to lower. The Easting and Northing displayed in the lower window are the Jamaica Metre Grid coordinates used on the 1:50k topo maps. The conversion can of course be run in reverse to turn JMG coords into WGS84 L/L.
You can test the transform by entering 77 W, 18 N, in the upper window, (WGS84), and converting to JMG. The output using JAD_1 should show an Easting of 249887, and a Northing of 149711. If a dialog box appears warning that the output is outside of the valid domain for the datum, just click ok and if the JMG numbers are right, ignore it; it seems to be a bug in Geotrans that will cure itself for the new datum after a restart. If the JMG doesn't match the numbers given, then check all your input parameters carefully, especially with regard to E/W and N/S.
An option to Geotrans is using Chuck Taylor's online coordinate and datum transformation tools. He's recently added the Jamaica 1969 datum based on the figures given by Mugnier. If you find Geotrans confusing, this is the best way to do the datum transform itself, but keep in mind that the output will be in lat/long or UTM, not Ja Metre Grid.
Users of Arcview 3 who have the digitized topos referenced to JMG can use any of these of sets of parameters to transform WGS84 positions, and then enter them into a dbf spreadsheet, and then display positions on the topos to within several pixels, (6m/pixel), but to do this with Arcview in WGS84, (i.e. without a prior datum transformation), requires version 8. This means muchos dineros because ESRI has the entire field of GIS securely in its pocket. They will charge not only an arm and a leg, but your first-born and the family dog to let you use their user-unfriendly programs, (the M$ of GIS), and if you don't even have Arcview 3.x, or work at a University or something, forget about ESRI... you can't afford it.
The upshot of it all is that if you have buckets of money and/or research funding, then you can use ESRI for the GIS, and you can throw thousands more at Trimble for the GPS, and have a JAD69 approximation of WGS84 reality. ESRI and Trimble will work together to give you coordinate transformations and positions that very nicely reflect the accuracy of the poorly defined Jamaican datum.
Fortunately, until the Jamaican government makes the island more GPS friendly, (see the link to Kirk Haughton at the end of the text), and someone brings E$RI down a few notches, we still have an easy and inexpensive solution available; reference the maps yourself.
Referencing the Jamaican topo maps requires two things: you to be on the island and in possession of a good GPS WAAS enabled receiver. Don’t bother throwing money at Trimble, a Garmin will give 1.5m versus Trimble's 1m. The WAAS version of Trimble is said to be around 10,000US$; the Garmin is around 300$. You’ll need an external antenna in most situations in Jamaica so make sure your receiver can accommodate that. Make sure the GPS can output NMEA strings, (a WAAS enabled Garmin will).
The actual referencing involves the use of a GPS mapping program and the determination of accurate calibration points at sites easily located on the map, usually road intersections, (the middle of the intersection is best; you'll find that on the digitized topos that the roads are shown as several times their actual width). Obtain averaged GPS WAAS WGS84 positions for these calibration points, (see wpt 070 on the map shown to the right), while trying to not be run over. Calibrate your digitized topo maps with the new latitudes and longitudes and disregard those found on the maps. Three calibration points will do but more will increase the accuracy by averaging out residual errors, (try for 6 to 9 calibration points per 1:50k topo sheet). Although this approach is time-consuming it does have the advantage of getting you out and about on the island and will result in accurately GPS referenced maps if done carefully. There are various GPS mapping programs available; OziExplorer and Fugawi are the most popular. If traveling with a laptop is out of the question, the hard-copy topo maps can be calibrated by hand if you’re the sort of person who is very, very patient.
If the foregoing seems a little intimidating, then geo-referencing files for the digitized topo sheets, in OziExplorer format, are available from the author of this article for free, but requiring a description of their intended use, by emailing the address found at the top of this page. The map calibration files will need a registered version of OziExplorer to function.
The Jamaican Topographic Maps are available in hardcopy from The National Land Agency of Jamaica. Unfortunately, although the 1:50K topo maps do exist in a digitized raster form, (20 tiff files with accompanying tfw files), they are not publicly available from Jamaican Government sources. You can digitize them yourself by scanning the hardcopy sheets, (unless you luck into a copy of the existing digitized set), and then writing simple 6 line tfw files in JMG or GPS referencing them in WGS84.
The digitized 1:12,500 topo maps have recently become available in both vector and raster formats, but the price of the raster maps for the entire island is brutal, (this is understandable because there are 232 sheets). The vector maps, (much cheaper), are available in DXF and Arc/Info but, of course, you'll need E$RI Arcview 8 to use them with a GPS and the datum used for the maps is even more poorly defined than JAD69.
At the very worst, JAD69 presents us with difficulties only when the topographical maps are used; a WAAS enabled GPS position, in WGS84, for anywhere in Jamaica, will return a GPS user to within metres of the desired location. Find it once, give it a waypoint, and you'll always find it again.
Many thanks to Chuck Taylor for his great assistance with the JAD69/WGS84 datum transformation problem.
By: R.S.Stewart of the Jamaican Caves Organisation
|More info on Datums,
Projections, and Grids in Jamaica|
The needs for and problems associated with the adoption of a geocentric datum for Jamaica: Kirk Haughton, MSc Surveying 2001
|Maps, Georeferencing, and GIS in Jamaica|