Category: Aerial Photographs

Posted on

Remote Sensing and Historical GIS

(Originally posted on The Otter, Jan 10, 2013)

Forty years ago, on Christmas Eve 1972, NASA released a gift to the public, the “Blue Marble” image of the whole Earth from space. This photograph was unplanned and originally unwanted by NASA, but it quickly became one of the most reproduced images on Earth.[1]

The Blue Marble: View of the earth as seen by the Apollo 17 crew traveling toward the moon. Source, Wikipedia
The Blue Marble: View of the earth as seen by the Apollo 17 crew traveling toward the moon. Source, Wikipedia

The astronauts aboard Apollo 17 were the first humans to find themselves in that sweet spot between the earth and the sun that offers a full view of the planet. It was an unprecedented photo op at 28,000 miles, and it has never occurred since. Other photos and compilations were made of Earth from space, but there is something about the clarity, colour, and serendipity of this photograph that sets it apart. It became the iconic image of the environmentalist movement and the epitome of remote sensing in the twentieth century.

In part three of our series on practical Geographic Information Systems (GIS) guides for historians we look at air photos and remote sensing. Remote sensing is the observation of any phenomenon from a distance, or a study made without handling the subject. In a way, all historians practice it. In geospatial terms, remote sensing is the analysis of geographic data captured from the air or space.

We have always been capable of imagining the landscape from above. One of the earliest maps is an oblique aerial view of the Paleolithic town Catalhoyuk, and bird’s eye views of cities became popular in early modern Europe. Of course these were stylized maps and not actually based on remote sensing. But there is a universal desire to see the world from above, or at least to understand the things we can’t reach. Some of the first Canadian maps created from remote sensing data were navigational charts of coasts and river bottoms. In the late nineteenth century, photogrammetry was used for mapping, especially in mountain terrain. Almost as soon as humans took to the air and to space, aerial and satellite photography blanketed the globe and transformed twentieth century cartography.

In this post we deal mostly with aerial photographs, although recent work has also been done on using satellite imagery in historical research.

What can historians do with historical air photos?

In Nature’s Economy, Donald Worster wrote about the historical impact of seeing the planet’s “thin film of life” in the Blue Marble photograph.[2] The earth we touch, taste, and consume seems universal and unlimited, but sensing the planet from above reveals limits and interdependent systems. From 28,000 miles we are fragile, unique, and alone.

Most historians will be interested in historical images taken a little closer to Earth; there are many and they have many uses. Good historical land use data matters, and remote sensing is one of the historian’s and historical geographer’s primary documents for understanding environmental change.

Air photos are also commonly used as supplements to other historical research. For instance, a 1958 air photo of a small island in Prince Edward Island National Park, shows that the island had quite recently been connected to the mainland by a causeway. A later image of the same landscape showed that the island eroded more quickly as a result. By comparing the three images (before, during, and after 1958) in a GIS, historians can measure the precise rate of erosion before and after the construction of the causeway. Thus, air photos contain information about place, single events that occurred in that place, and more gradual trends such as coastline change. All of this is possible with three photos.

some_text
Aerial photo of Robinsons Island, PEI 1935

some_text
Aerial photo of Robinsons Island, PEI 1958

some_text
Aerial photo of Robinsons Island (west end), 1974

What do you need to get started?

First you’ll need some air photos of your area of interest. There are two main forms of aerial photographs, vertical and oblique. Vertical photos are taken at a 90-degree angle and oblique at something closer to 45 degrees. Both forms were used for surveying and mapping. Vertical photos offered a less distorted image of the terrain, and these became standard practice in the inter-war period.

The good news is that Canada was a leader in aerial photography, so its collections date back to the 1920s and it has a lot of them – a lot, as in over 6 million. That’s also the bad news. Locating the air photos for the right time and the right place for your research can be daunting. Most of these documents can be found at the National Air Photo Library (NAPL), but localized collections also exist in government offices, university libraries, and some museums.

Managing the collection will also take some planning; the earliest air photos were taken at a scale of about 1:15,000 (one 9 inch photo represents over 2 miles of terrain) and overlapped, so a single Prairie township occupied up to 50 photos. A GIS is an excellent tool for displaying, managing, and eventually analyzing the images. If you plan to incorporate air photos into a GIS, be aware that large numbers of high resolution images will take up a lot of processing power.

Fortunately, most air photo collections are labeled and well organized, and in some cases they are searchable using a map index. The Canadian National Air Photo Library’s catalogue is available at NAPL Online. This search engine is a bit tricky to get used to (see search tips), but once you’ve identified your research area on the map, it allows you to display the approximate location and outlines for each air photo. Many of NAPL’s images have already been digitized, and the engine will show low-resolution previews of any digitized photo. From there you will have to record the call numbers and request copies from the library, at a cost.

Other providers offer historical imagery for free, such as the Government of Prince Edward Island, although getting the highest resolution may still require ordering originals from NAPL. See the working list of aerial photo collections in Canadian university, provincial, and municipal archives, at the end of this post.

What programs do you need to view air photos?

This depends on what file format the digitized photos came in and what you want to do with them. There are three main types of digitized air photos, and GIS users group them under “raster” based data.

The first and most common is a 2-dimensional raster image (for example .jpg) without geographic coordinates. These images are no different than a photo of your friends on Facebook. They can be opened in any browser or image preview software, and they are easy to explore. However, since they lack geographic coordinates, you could not incorporate them into a GIS and have them line up with other features. In order to do this you must either “georeference” or “orthorectify” the image which converts it into the second and third raster type.

Georeferenced air photos are the same as the 2-dimensional images, except they have been given geographic coordinates and made to fit the 3-dimensional curve of the earth. They are then saved as a new file format (for example TIFF and GeoTIFF are non-proprietary formats that will open in a variety of GIS programs). We will explain how to assign control points and georeference images in another guide, because it applies to any historical map or image that you might want to incorporate into your GIS. However, unlike maps, air photos are images captured with a lens, and like all photographs they contain distortions. Distances and shapes become distorted as features move away from the centre of the image toward the edges. If geographic precision and accurate measurements are important to your research then you need to consider the third format, orthoimages.

Orthorectified images, often just called “orthos,” are 2-dimensional images that have been given both geographic coordinates and geometric correction to account for the curvature of the earth and distortion in the images. Most GIS software will orthorectify air photos, but it requires more than just control points. Information on the altitude, lens type, and angle of the camera are required for each photo. This was usually recorded in flight logs, but in very early air photos it might not be possible to ascertain. Again, if your main interest is in observing features in the landscape or general land use over time, orthorectification is not required.

Georeferenced and orthorectified images are most commonly saved in GeoTiff (.tif) format, and these are easily opened by GIS programs such as Quantum GIS, ArcGIS, and MapInfo. Other formats include MrSID, an early raster format popular among GIS users, and more recently KMZ (a zipped Keyhole Markup Language format). Images stored as KMZ open in Google Earth and allow quick and easy overlays and measurement.

Go ahead, try it!

The University of Waterloo Map Library has an excellent collection of aerial photos from the Waterloo Region and beyond. To explore their collection of digitized air photos you can navigate through the Air Photos Digitization Project. Once you’ve found an area you would like to open in a GIS or Google Earth you can choose a file format and download it. For example, try downloadingthe KMZ in this section of rural Woolwich township (as they instruct, remove .zip from the filename after downloading) and open it in Google Earth. You will be able to zoom in and out, tilt for a 3D effect, and add layers from Google’s collection.

The Waterloo example is one of the best in Canada at the moment, but as you will see from the list below, almost every Canadian centre has an air photo collection. Go ahead and explore, and if you find some images of the world you know from above, you might want to contact the library and the other guides in this series for help.

Josh MacFadyen is a Postdoctoral Fellow at the University of Western Ontario and the Project Coordinator for NiCHE.

 

[1] Donald J. Wuebbles, “Celebrating the ‘Blue Marble,’” Eos 93(49) (4 December 2012); 509-510; Robert Poole, Earthrise: How Man First Saw the Earth (New Haven: Yale University Press, 2008); Al Reinert, “The Blue Marble shot: Our first complete photograph of Earth,” Atlantic, 12 April, 2011. [Available at http://www.theatlantic.com/technology/archive/2011/04/the-blue-marble-shot-our-first-complete-photograph-of-earth/237167/]

[2] Donald Worster, Nature’s Economy: A History of Ecological Ideas (Cambridge: Cambridge University Press, 1977), 358-9.

Federal air photo and remote sensing holdings

University air photo holdings

NB: Since writing this post, it also came to my attention that York has an extensive collection of low-level oblique aerial photographs of conservation areas in Ontario taken by Lou Wise.

Other library and archival collections

Provincial holdings

Municipal holdings

  • City of Toronto Archives
  • The University of Waterloo Library has compiled a list of all the Canadian municipal governments that provide open GIS data online. The municipalities offering air photos include Fredericton, Nanaimo, North Okanagan, North Vancouver, Surrey, and Vancouver.

Many other municipalities provide limited historical air photos for viewing only, such as London,Red Deer, the Region of Waterloo, and many more in this list identified by the University of Waterloo Map Library.

Global datasets of interest

Scotland!

Posted on

Aerial Photographs, Forest History, and P.E.I.’s First Energy Crisis

(First published on The Otter, May 12, 2012)

Recent studies of aerial photographs suggest that Prince Edward Island’s first energy crisis was not the 1973 oil embargo but a firewood shortage that predated OPEC by a century. Two posts on The Otter discuss some of the ways air photos can be used to understand landscape change in the PEI National Park, and a new exhibit stresses the importance of seeing these images as snapshots of time as well as place. Air photos can also be used in a Geographic Information System (GIS) to study land cover and land use change, and the resulting data offer a new tool for understanding the difficult balance between clearing forests for agriculture and leaving enough for fencing, forage, and fuel.

Tourists who approach Prince Edward Island on the ferry in 2012 disembark at “Wood Islands” and take in a coastal landscape of relatively unbroken forest, but the ferry’s first passage in the late 1930s would have revealed a very different and almost completely denuded coastline. (Early tourist invitations to “come play on our island” might have included a suggestion to “bring your own firewood.”) The 20th century forest regeneration at Wood Islands echoes a well known story of outmigration and agricultural downsizing in the Atlantic region, evident in sources like the Census of Agriculture. Historians assume that farm abandonment occurred primarily on marginal lands, as families adapted to poor soils and a poorer economy, but air photos offer a detailed portrait and a spatially explicit model of different stresses such as declining access to woodlots and other wildland resources.

At the beginning of the 20th century Prince Edward Island’s forests had reached a nadir from extensive agricultural and commercial exploitation of the forest. But only a generation or two earlier, there was an abundance of forest. In 1861, the average Island farmer lived on properties which were only about 38 percent “arable,” or cleared to the extent they could be cropped. Taking into consideration the other thirty percent of the colony not in farms, we see that only about a quarter of the Island had actually been deforested. Forest historians argue that by the end of the century over 70 percent of the Province was cleared for agriculture and the forest that remained had been harvested several times for the shipbuilding industry and domestic use.

In a recent Otter post I described how Prince Edward Island farmers responded to a critical food shortage, when the Province’s ruminant population outpaced its marsh and upland hay production. Farmers took to the ice and spent a large part of the winter dredging “mussel mud” to enrich their fields and expand the production of hay and cereals. However, this practice was prevalent in areas where farmers had already cleared most of their land, both because growth required either new land or more productive land and because farms in these areas were running out of firewood. With little forest left, many of these farmers turned their winter activities to hauling mussel mud and began to buy coal or commercial firewood for domestic energy.

Using air photos it is possible to identify places where this likely occurred. Figure 1 shows a composite image of four map overlays (labelled A-D on the right) and the location of every home in Lot 30 (labelled on the left), a rural township to the west of Charlottetown. These layers allow me to identify the precise location of homes, the location of the forest, and the places where the density of homes and scarcity of forest would have caused a biomass energy crisis. These data are all derived from PEI’s aerial photographs which were first flown for many Canadian jurisdictions in the 1920s and 1930s. The original air photos were subsequently used to create maps for the National Topographic Series, and in 1990 the Province of PEI also used them to create historical forest and land use inventories. In the final layer, I used a GIS to measure the amount of forest located within a 2.5 Km radius of each home and made an “energy crisis map” based on the density of points in selected PEI study sites.

pei-macfadyenmap1935
Figure 1: Forest history data derived from air photos, Bonshaw, Lot 30, P.E.I. 1935. Photos and inventory courtesy of PEI Dept of Forestry.

This map shows areas where rural and small town resettlement was putting pressure on natural ecosystems. For example, Lot 30 had approximately 14 hectares of forest for every home, but the homes marked with blue dots had direct access to less than 5. Clearly these farmers lived in an interconnected world and had access to fuel from local outfits and importers, but what is important here is that their immediate fuel supplies would not have lasted more than a few years at even a conservative rate of consumption.

Additional research and a presentation at the Canadian Historical Association / Canadian Association of Geographers at Congress 2012, in Waterloo, will show what kinds of woodlots remained in the most densely inhabited areas. We might expect that harvesting was most intense in those areas and regenerating plots were not. The forest inventories recorded not only the outline but the cover type of each forest parcel, and GIS lets us query the inventories for these sorts of questions.

The 1935 PEI inventory may be the earliest in Canada, and due to the Province’s small size and complete air photo coverage it is certainly the most comprehensive. Behind the inventories, however, are ordinary air photos, and researchers interested in recreating topographic features can find great coverage of local areas at university libraries such as Brock, Dalhousie, Toronto, Waterloo, and Western. For larger areas consult the National Air Photo Library’s enormous collection and online search engine in Ottawa. All air photos contain distortions, and if your research relies on accurate locations, distances, and measurements it is important to correctly “georectify” and digitize these features with the proper projection in a GIS. Many <href=”#/Fundamentals_for_georeferencing_a_raster_dataset/009t000000mn000000/”>online tutorials are available to help with these processes. If you only require the general description and location of topographic features then you might want to use the closest National Topographic Series (NTS) map for your period of interest.

Aerial photographs, NTS maps, and historical forest and land use inventories offer a new way to study the relationship between farms and forests. If the NTS maps were a sort of census of the Canadian environment in certain years, then aerial photos were the original manuscripts behind the printed census. Just as social and economic historians often need to consult the nominal censuses for a new level of detail, historians of the environment can find information in the original air photos (e.g. forest types, land use, coastal erosion, building location/orientation, sub-urban spaces, pollution and brownfields, etc.) that may not have interested the cartographers.

Further reading:

  • D. G. Sobey and W. M. Glen, “A Mapping of the Present and Past Forest-types of Prince Edward Island,” Canadian Field-Naturalist 118 (4) (2004): 504-520.
  • There are several chapters of interest to NiCHE readers in Historical Geographic Information Systems in Canada a forthcoming book edited by Jennifer Bonnell and Marcel Fortin. The chapter by Joanna Dean and Jon Pasher uses air photos and other sources to measure urban forests, and Joshua MacFadyen and William Glen have presented a longer discussion of the PEI forest inventories in “Top-down history: Delimiting forests, farms, and the Agricultural Census on Prince Edward Island using Aerial Photography, c.1900-2000.”
  • Don Valley Historical Mapping Project
  • GeoWATCH: Geospatial Workshops in Atlantic Canadian History
  • NiCHE Resources for learning and developing Historical Geographic Information System (HGIS) projects.
  • National Air Photo Library