Jim Clifford, Josh MacFadyen, Daniel Macfarlane

In this lesson, you will learn how to georeference historical maps so that they may be added to a GIS as a raster layer.

edited by

  • Adam Crymble

reviewed by

  • Finn Arne Jørgensen
  • Peter Webster
  • Abby Schreiber

published

2013-12-13

modified

2013-12-13

difficulty

Medium
This lesson is part of a series. You might want to check out the previous lesson.

Contents

Lesson Goals

In this lesson, you will learn how to georeference historical maps so that they may be added to a GIS as a raster layer. Georeferencing is required for anyone who wants to accurately digitize data found on a paper map, and since historians work mostly in the realm of paper, georeferencing is one of our most commonly used tools. The technique uses a series of control points to give a two-dimensional object like a paper map the real world coordinates it needs to align with the three-dimensional features of the earth in GIS software (in Intro to Google Maps and Google Earth we saw an ‘overlay’ which is a Google Earth shortcut version of georeferencing).

Georeferencing a historical map requires a knowledge of both the geography and the history of the place you are studying to ensure accuracy. The built and natural landscapes change over time, and it is important to confirm that the location of your control points — whether they be houses, intersections, or even towns — have remained constant. Entering control points in a GIS is easy, but behind the scenes, georeferencing uses complex transformation and compression processes. These are used to correct the distortions and inaccuracies found in many historical maps and stretch the maps so that they fit geographic coordinates. In cartography this is known as rubber-sheeting because it treats the map as if it were made of rubber and the control points as if they were tacks ‘pinning’ the historical document to a three dimensional surface like the globe.

To offer some examples of georeferenced historical maps, we prepared some National Topographic Series maps hosted on the University of Toronto Map Library website courtesy of Marcel Fortin, and we overlaid them on a Google web map. Viewers can adjust the transparency with the slider bar on the top right, view the historical map as an overlay on terrain or satellite images, or click ‘Earth’ to switch into Google Earth mode and see 3D elevation and modern buildings (in Halifax and Dartmouth). Note: these historical images are large and will appear on the screen slowly, especially as you zoom into the Google map.

Getting Started

Before proceeding with georeferencing in Quantum GIS, we need to activate the appropriate Plugins. On the toolbar go to Plugins -> Manage and Install Plugins

Figure 1

Figure 1

A window titled Plugin Manager will open. Scroll down to Georeference GDAL and check the box beside it, and click OK.

Figure 2

Figure 2

Open the Necessary GIS Layers

For the Prince Edward Island case study, we are going to use the township boundaries as control points because they were established in 1764 by Samuel Holland, they are identified on most maps of PEI, and they have changed very minimally since then.

Download lot_township_polygon:

This is the shapefile containing the modern vector layer we are going to use to georeference the historical map. Note that townships were not given names but rather a lot number in 1764, so they are usually referred to as ‘Lots’ in PEI. Hence the file name ‘lot_township_polygon’.

http://www.gov.pe.ca/gis/license_agreement.php3?name=lot_town&file_format=SHP

Figure 3

Figure 3

Add lot_township_polygon to QGIS:

Figure 4

Figure 4

For more information on adding and visualizing layers see Installing QGIS 2.0 and adding Layers.

Figure 5

Figure 5

Open the Georeferencer Tool

Georeferencer is now available under the Raster menu on the toolbar – select it.

Figure 6

Figure 6

Add your historical map:

Figure 7

Figure 7

The result will look like this:

Figure 8

Figure 8

Adding control points:

Plan the locations you are going to use as control points in advance of the steps that follow. It is much easier to navigate around the historical map first, so get a good idea of the best points to use and keep them in mind.

Some tips for choosing control points:

Add your first control point:

First, navigate to the location of your first control point on the historical map.

Figure 9

Figure 9

Figure 10

Figure 10

Figure 11

Figure 11

Figure 12

Figure 12

Add at least one more control point:

Figure 13

Figure 13

Determine the transformation settings:

Before you click Play and start the automated georeferencing process you need to tell QGIS where to save the new file (this will be a raster file), how it should interpret your control points, and how it should compress the image.

Figure 14

Figure 14

Most of these settings can be left as default: linear transformation type, nearest neighbour resampling method, and LZW compression. (The world file is not necessary, unless you want to georeference the same image again in another GIS or if someone else needs to georeference the image and does not have access to your GIS data, coordinate reference system, etc.) The target SRS is not important, but you could use this feature to give the new raster a different reference system.

Figure 15

Figure 15

Georeference!

Figure 16

Figure 16

Figure 17

Figure 17

Figure 18

Figure 18

Explore your map:

Figure 19

Figure 19

Figure 20

Figure 20

Figure 21

Figure 21

Now that you have a newly georeferenced map in your GIS you can explore the layer, adjust the transparency, contrast and brightness, and go back through Creating New Vector Layers in QGIS to digitize some of the historical information that you have created. For instance, this georeferenced map of PEI shows the locations of all homes in 1863, including the name of the head of household. By assigning points on the map you can enter home locations and owner names and then analyze or share that new geospatial layer as a shapefile.

By digitizing line vectors such as roads or coastlines you can compare the location of these features with other historical data, or simply compare them visually with the lot_township_polygon layer in this GIS.

In more advanced processes you can even drape this georeferenced image over a DEM (digital elevation model) to give it a hillshade terrain or 3D effect and perform a ‘fly-over’ of PEI homes in the nineteenth century.

This lesson is part of the Geospatial Historian.

About the authors

Jim Clifford is an assistant professor in the Department of History at the University of Saskatchewan.   Josh MacFadyen is a Project Coordinator at the Network in Canadian History & Environment.   Daniel Macfarlane is a Visiting Scholar in the School of Canadian Studies at Carleton University.