Georeferencing a Palaeogeographic Map

There a a number of resources on the web and in the literature when it comes to palaeogeographic reconstructions. Two of the most commonly used sites for images are Chris Scotese’s PALEOMAP project and Ron Blakey’s home page. While the images that you can find on them are great, the format the images are in are a bit of a problem when it comes to loading them into a GIS program, as they aren’t in a typical GIS format. For example, the common GeoTIFF format that many GIS images are distributed as contains information about the image’s location, allowing it to be easily opened in GIS program. This makes it easier when trying to combine multiple georeferenced files, as the program will automatically place them in the proper location (and projection) relative to one another.  For my research, this is a very useful tool as it allows me to quickly place and update fossil localities on a map, without having to do everything by hand. Luckily, there are several easy ways to georeference an image.

The first method involves using a cross-platform GIS program called Quantum GIS. Once you have the program installed, open it up and the first thing you will do is enable the Georeferencer plugin. To do this, go to  Plugins > Manage Plugins on the top menu bar. Scroll down and select the Georeferencer plugin, and hit OK. Now, go to Plugins > Georeferencer > Georeferencer to open up the plugin. You will bepresented with two windows. Go to the window titled Georeferencer and navigate to the location of your image that you want to georeference and select it. Once you’ve done this, your image will pop up in the other window. In this window, you will select points on the map for which you know the location, and these points will allow the GIS program to position your image accordingly. To select a point, choose the button with the three red dots and then click on the image. It will ask for the X (typically longitude) and Y (typically latitude) of the point, which you need to enter. You can use the zoom options (top left) to zoom in and out to be more precise when choosing points as well. Once you’re done, you can hit Create to save or Create and Load Layer to save and load it in the program. Your original image should be unaltered, but there will now be a world file with a similar name (but slightly different ending) that contains the coordinates for the image, which the GIS can then use.

While the first method may seem easiest, as it just involves clicking pictures, I prefer a second method. I have installed a command line program (well really, a library and associated utilities) called the Geospatial Data Abstraction Library or GDAL for short. Here again, there are packages for all platforms available for download. Within GDAL, there is a utility called gdal_translate which allows you to georeference an image. To try it out once you have GDAL installed, fire up a terminal (or console, or whatever you like to call your command-line interface). From here, a single line of text can georeference your image. For example, if I have a full world map that is 375 pixels high and 750 pixels wide (as are all of Blakey’s palaeogeographic maps), and I know that the corners of the map are at -180 and 180 degrees longitude and -90 and 90 degrees latitude (as are Blakey’s), I can use the line:

gdal_translate -of GTiff -gcp 0 0 -180 90  -gcp 0 375 -180 -90  -gcp 750 0 180 90  -gcp 750 375 180 -90 90mya.jpg 90.tif

To break down the line a bit, gdal_translate calls the function, -of GTiff tells it I want a GeoTIFF as the output format, and the -gcp 0 0 -180 90 etc. each refer to a corner of the image and it’s longitude (or X) and latitude (or Y) (ie the top left corner or 0 0 is at -180 degrees longitude and 90 degrees latitude). Finally, the last two parts refer to the original image name (90mya.jpg) and output image name (90.tif). While this may seem a little more complex at first, it is far more precise than trying to click exact points on maps, and is much faster to use, especially with multiple images of the same size and position.

Either way, in the end you should have a much more useful image that you can do lots of other things which, like changing to a new projection, which I’ll talk about a bit more in a little while.