The Politics of Map Making; Katrina and Google Earth

How are new tools for representing physical geography altering the way we think about ourselves and our place in the world? What differences are sites like Google Earth making in the ways we cognitively map the environments around us? What new issues do they raise as ways of representing the world and how do they relate to older traditions of cartography? These are the questions which CMS Masters Student Amanda Finkelberg took up in her thesis project. Finkelberg came to CMS having worked extensively in the special effects industry. She had spent the last few years of her work erasing wires, rigs, and other elements from shots, not exactly the most compelling work in the industry, and wanted to find ways to make the work of media more visible to the people who consumed it. She worked as part of the Project nml teams, producing a segment for our exemplar library dealing with special effects and advertising; she's been doing some early planning work on an educational games project which is still under wraps and she has been working as a research and teaching assistant for a new course which explores the relationship between theater and science in the early modern period, being taught by Diana Henderson and Janet Sonenberg in collaboration with the Royal Shakespeare Company. Like many of our graduate students, she has taken advantage of an exchange program between Harvard and MIT to take courses down the river, much of which has informed her current interests in maps and systems of spatial representation.

From my own point of view, her work on cartography comes at just the right time since one of our projects for the New Media Literacies team next year centers on thinking through how these new digital mapping technologies can be integrated into the social science classroom. As we pursue that project, I know that I will be re-reading her thesis for new insights. This selection from her thesis centers on a recent controversy surrounding how Google Earth represents the area around New Orleans, how this figures into post-Katrina politics, and what this tells us about the "neutrality" of maps. Enjoy!

Google Earth and Katrina Images

By Amanda Finkelberg

A tradition of accuracy is strongly reflected in issues surrounding contemporary spatial representations. As cartographic technology improves, generating more convincing depictions of space from above, the debate over accuracy becomes marginalized in favor of an acquiescence in a sort of "truthiness" or sense that the data is true enough. This phenomenon is illuminated by the continuing problem of how Google Earth has chosen to represent the Katrina devastated Gulf Coast in its satellite image database.

Immediately following the 2005 hurricane, Google Earth became a valuable tool for evacuees hoping to estimate damage to their property. By using amateur fly over photographs as "overlays" in Google Earth, networks of people banded together to determine which areas had been hit by flooding and posted the information to bulletin boards. Photographs taken from a Cessna Citation jet were available within 24 hours after the storm. Although not entirely clear, the images, when matched to GPS coordinates in Google Earth, gave a good sense of the condition of a home, street, or neighborhood. The novelty of the software encouraged many new users to lend a hand to the effort, according to a September 5th, 2005 New York Times article . This example provokes interesting questions about temporality and global imaging. The fly-over images were clearly being taken as valid real-time information about the state of the disaster area, lent additional credibility from their alignment with GPS coordinates in Google Earth's global framework. While this is an extreme circumstance, it does clearly indicate a new type of amateur cartographic behavior. A plane flew over, photographed and posted aerial images to the internet. The images were obtained by neo-geographers at home who responded to queries from displaced Gulf residents and generated mostly-accurate maps of the real-time flooding in their communities. These maps were either reposted to the internet or described in email .

Google Earth's representation of Katrina's damage did not stop there and has, in fact, recently become the center of a heated debate within GIS communities, raising serious questions of accuracy, politics and digital cartography. Last week (March 25, 2007) Google came under heavy criticism from user, Geographic, and journalistic communities for its unexplained swapping the post-Katrina images with the pre-disaster ones seen here. In a letter to Google, D-N.C., chairman of the House Committee on Science and Technology's Subcommittee on Investigations and Oversight, Brad Miller demanded an explanation for the imagery switch. "To use older, pre-Katrina imagery when more recent images are available without some explanation as to why appears to be fundamentally dishonest ," the letter states. It goes on to directly inquire if the Federal Emergency Management Assistance (FEMA) agency had contacted Google requesting the images be changed. This suspicion reflects not only a deep distrust of the agency's attempt to cover up mismanagement of the crisis but also impugns Google as a potential collaborator in revisionist history.

Google's quick and thorough response earlier this week came in the form of replacing the previous images accompanied by a blog posting by Maps/Local/Earth director, John Hanke. Hanke writes:

In 2005, shortly after Hurricane Katrina hit New Orleans and the Gulf Coast, a very motivated group of volunteers at Google worked with NOAA, NASA, and others to post updated imagery of the affected areas in Google Maps and Google Earth as quickly as possible. This data served as a useful reference for many people... Several months later, in September 2006, the storm imagery was replaced with pre-Katrina aerial photography of much higher resolution as part of a regular series of global data enhancements. We continued to make available the Katrina imagery, and associated overlays such as damage assessments and Red Cross shelters, on a dedicated site (earth.google.com/katrina.html). Our goal throughout has been to produce a global earth database of the best quality -- accounting for timeliness, resolution, cloud cover, light conditions, and color balancing .

The suggestions that Google operated out of malfeasance by replacing the images of devastation were connected to a range of conspiracy theories, from the FEMA implication made by Rep. Miller to the suggestion that the maps had been changed to indicate successful recovery in hopes of revitalizing tourism to the area. These accusations were quickly dismissed by Google's blog statement and reinstitution of the lower quality images of flooding. The reasonable explanation may explain the overlay switches but the theoretical problem indicated by this event is not so easily dismissed.

The problem is perhaps best illustrated by the final paragraph of Representative Miller's letter to Google:

Digital technology has any number of benefits, as Google's healthy balance sheet demonstrates. However, experience has also shown that such technologies pose a particular threat to photography as a representation of reality. While we can understand that Google would prefer the most recent imagery of the New Orleans region for its Web site, to use older, pre-Katrina imagery when more recent images are available without some explanation as to why appears to be fundamentally dishonest. The entire country knows that New Orleans is a great American city struggling to recover from an unprecedented disaster. Google's use of old imagery appears to be doing the victims of Hurricane Katrina a great injustice by airbrushing history .

Several key points stand out from this succinct accusatory paragraph. First, Miller indicates Google's "healthy balance sheet," a barb not missed by the blogosphere which immediately picked up on the implications of a misuse of power in postings like "Is Google Finally Evil? ". The relationship between power and wealth and cartography is not a new one by any stretch of the imagination, and this accusation squares with a historic reading of maps and spatial control. There is, however, a much bigger accusation being directed by this paragraph; one that impugns the digital nature of the digital map, making the same very equation that I make within this essay: the homogenization of digital data that enables a cross-media discussion. Miller asserts that digital maps are not only akin to photography, but can also be "airbrushed " to correct imperfections in the same way a model is corrected on a magazine cover. This is an equivalence that illuminates the ease and completeness with which digital maps can be altered as well as the lens of scrutiny and skepticism through which they should be evaluated.

There is still another issue raised by this little story, indicated by John Hanke's list of priorities for the Google satellite photos. Hanke mentions "timeliness" as a key factor in determining an image's quality for inclusion in the database. What on earth could he have meant by this? Most likely it is a response to the problem at hand and refers to images that are the best representation of "what is" but this explanation is deeply problematic when explored even slightly superficially. The use of time as an evaluative characteristic, particularly when coupled by issues of cloud cover or resolution, imply that there is a time that is more neutral than another. Presumably this is a time that is sunlit, cloudless, and well photographed. These may be the best conditions for viewing the surface of Earth from above but they should by no accounts be mistaken for neutral conditions.

This event highlights cartographic non-neutrality with regards to the representation of time in the hurricane torn Gulf region. Pre-hurricane images of the Gulf area may better meet the qualities of ideal viewing conditions while disregarding cartography's mission of accuracy with regards to representing land as it is. This new problem is created by the type of semi-instantaneousness of satellite images which on one hand provide a sense of being real-time (as the overlays were used immediately following the storm to locate damaged property) but on the other provide an uncanny representation of ambiguous yesterdays: an indeterminable time that was, of course, cloudless and sunny. The crucial decision about which time to represent is unique to satellite imagery, a departure from the abstracted, time-independent, spatial representations of the paper map.

A fruitful example of embedded subjectivity, this story about Google's difficulty in representing the current state of Gulf region clearly indicates an emergent difficulty facing the challenge of cartographic objectivity. Because satellite images are essentially photographs they are now faced with the inherent complexities of photography, a discourse ranging from problematic apparatus to manipulated artifacts. Moreover, the satellite image can never be entirely neutral or objective, regardless of Google's hopes for ideal viewing conditions. There will never be a way to clear off the planet for a few minutes while Google takes a neutral picture for its database. We are all, therefore, permanently embedded in some way, within the pixels of Google Maps/Earth. This is abundantly clear to these Dutch sunbathers (52° 4′43.38″N, 4°19′58.02″E ), imprinted forever or for the time being, within the satellite overlay currently used in the default Google Earth.

While these people could, of course, be removed digitally, "airbrushed" as it were, there is no end to the ouroboros initiated by even thinking about that process: however would the airbrush artist determine how a neutral planet should look?

We see that even at the level of objective data the human is back on the map, so to speak. On a paper map, areas of pastel shading demark territory. This type of abstraction provides a clear symbolic separation between the actual land and what is represented by the map and, I offer, a buffer between the image and the illusion of reality. The satellite image, however, exchanges that abstraction in favor of a far less noticeable one, digitization. Rather than Texas symbolized by a large pink area in the shape of Texas, it is now represented by an overhead photograph laid upon a three dimensional terrain wireframe. Reliance on satellite images as realistic or accurate representations of space is inherently problematic because of this digital nature which gives the appearance of real but is in actuality, a mere slice of space and time. This problem may be temporary as technology improves to allow access to real time streaming satellite images of the planet. This likely possibility opens the door to a host of mind boggling privacy issues only suggested by the current capabilities of Google's spatial representation tools.

Amanda Finkelberg

Pomona College, BA Media Arts/Political Science, 1996

Prior to CMS, Finkelberg worked as a Visual Effects artist for such films as Spiderman 2 and Star Wars Episode 2. In 2004, she started a 2D effects boutique, Rig-Out, which specialized in erasing wires, rigs, and other background details. She has a strong interest in social media and has facilitated media education projects in a parole recovery center in downtown Los Angeles.

Finkelberg;s thesis "Space, Place, Database: Digital Cartography in the Network Age" addresses issues of reality and representation with regards to data. Other interests include sustainability, perception, and information/interface design. She intends to graduate this spring and return to California.