In Defense of the Map Library

Introduction

In recent years, the value of print map collections around the world has come into question resulting in the closure and consolidation of many map libraries and their collections (Keller 2001; Andrew and Weimer 2016). The justification for these closures is often that “print maps are no longer relevant in a digital world.” In this editorial, we (Marcy and Nathan) take this opportunity to introduce ourselves to the readership of the Taylor & Francis Journal of Map & Geography Libraries (hereafter “JMGL”), by outlining four major ways we feel that map libraries and the print cartographic collections that they preserve remain relevant and irreplaceable even in a digital world. There are likely other ways that map libraries and their collections remain relevant today and we invite you, the authors and readers of this journal, to contribute your own ideas and writing on the topic in future issues. In our view, there is a dearth of formal writing and scholarship in response to map library closures and if not in this journal, then where would one look for sound reasoning and justification to keep them open?

The Oxford English Dictionary defines a map as a “drawing or other representation of the earth’s surface or a part of it made on a flat surface, showing the distribution of physical or geographic features (and often also including socio-economic, political, agricultural, meteorological, etc. information), with each point in the representation corresponding to an actual geographical position according to a fixed scale or projection; a similar representation of the positions of the stars in the sky, the surface of the planet, or the like.” For the purposes of this editorial, we refer to “maps” generically to also include the other print cartographic resources often housed in map libraries like atlases, globes, historic aerial photographs, and others.

Print maps (and other print cartographic resources) are often the authoritative geographic and scientific documents of their time, accurately encoding a wealth of information about society, culture and the environment in a spatially-accurate way not easily afforded by other historical documents. In the past, the primary role of the map library was to collect, preserve, and maintain access to this information and because there was no alternative, this was justification enough for map libraries to often hold prominent positions in academic libraries as recently as the 1990s. Since then, computer mapping, geographic information systems (GIS), and related geospatial technologies have evolved quickly and clearly changed the mapping world for the better (Goodchild 1988). The ubiquity and technological sophistication of digital maps has led to a common perception that the born-digital map is now the authoritative and/or best source of geographic information and this is true in some, but still not all cases. Print maps remain irreplaceable under (at least) the following four conditions: (1) For places and/or times where digital information is not available or information accuracy must be known and high; (2) For answering research questions about the past, present, and future; (3) For navigation and route-finding where computerized mapping devices and digital geospatial data do not work, are unreliable, or are prohibitively expensive; (4) When technology is a barrier to teaching the foundations of spatial thinking and literacy.

Place, time, and accuracy

Many people assume that the entire earth has already been digitally mapped with high known positional accuracy. It is true that there is more geospatial data available now than ever before and much of it is generally of high-quality and yet print maps still offer highly accurate geospatial information about places and times for which there is no digital substitute.

Barrington-Leigh and Millard-Ball (2017), estimated that the roads layer of the crowd-sourced geographic database OpenStreetMap (www.openstreet.map.org) was globally more than 80% complete with 40% (usually the most populous) of countries having a complete road network. This is a tremendous accomplishment. And yet, their paper highlights that nearly 20% of the global road network remains digitally unmapped (at least by OSM), and more than half of the countries in the world likely have an incomplete digital road network. It is unclear how much of the current global road network is represented on print maps, but numerous large-scale international map sets of thousands of individual sheets provide a tremendous amount of information about road networks that is current to within the last two or three decades. This is true even of less-populated countries because many of these map sets resulted from decades long efforts to systematically “map the world” (usually the terrestrial earth’s surface), rather than methods that are sensitive to how populous an area may be?

Mapping the contemporary global road network is just one example for one thematic layer and we are still learning about previously classified international map sets that offer a wealth of local-scale information (in addition to road networks) like the Soviet Military maps detailed in the book The Red Atlas: How the Soviet Union secretly mapped the world (Davies and Kent 2017). It is highly likely that even today in 2018, in order to assemble global base-mapping layers (i.e. transportation, populated places, streams and rivers, etc.), one would have to use some information from print maps or digital sources that are based on print maps (e.g. the Vector Map or VMAP). We have already collected and preserved the results of many of these global mapping efforts in map libraries.

Perhaps the most obvious justification for the continued collection, management, and preservation of print maps in maps libraries is that there are few if any viable substitutes of geographic information for historical time-periods prior to the relatively recent past. The LANDSAT satellite is longest running satellite imaging program in history and its inception (1972) generally marks the modern era of global geospatial data availability. Prior to about 50 years ago (i.e. the 1970s), authoritative geographic information was encoded on print maps, many of which are preserved and accessible in map libraries.

A common misconception is that digital geospatial data are always more spatially and geometrically accurate than the information contained on print maps and this reflects misunderstandings about both print map accuracy and the accuracy of contemporary geospatial data sources like satellite imagery and crowd-sourced geospatial datasets. It may also be a side-effect of much of the print map scholarship to date being focused on pre-18th century maps, which were often not very accurate by today’s standards. However, there is at least some suggestion that large-scale mapping techniques of the more recent 18th, 19th, and early 20th centuries produced local-scale results that are as accurate as some of the best geospatial data currently available. Evidence of this is varied and comes from both directed quantitative efforts to investigate this question (e.g. Piekielek 2017), and from the anecdotes of map users. Anecdotes of users include that Allied Forces military operations during the U.S. war in Afghanistan (2001) relied on information contained on large-scale topographic maps rather than or perhaps in addition to the geospatial data that were available at that time (G. Forbes, personal communication, 2018); and that foreign governments protested the U.S. National Intelligence Mapping Agency’s (now National Geospatial-Intelligence Agency) release of its Vector Map data product that was based on print joint operations military maps because it would threaten state monopolies on foreign geospatial data. Both anecdotes are now more than 10 years old, but they deal with military operations and national security which should have very high standards for geographic information quality. Surely, even today the geographic information contained on print maps must have value for non-military operations even in a digital world.

Misconceptions about the accuracy of the information contained on print maps may also come from misunderstandings of the accuracy of some of the most widely available contemporary digital geospatial datasets like satellite imagery and crowd-sourced geospatial data. Although the spatial resolution of satellite imagery has recently decreased dramatically to meters and even centimeters, horizontal positional accuracy has not experienced a similar revolution. For example, Digital Globe’s Worldview IV data that is some of the best and most sought after publicly available satellite imagery, are advertised as being 30 cm spatial resolution and <5 m horizontal positional accuracy at 90% confidence (http://worldview4.digitalglobe.com/). Many historic aerial photographs like those housed in map libraries are currently being digitized at approximately 70 cm spatial resolution and Sanborn Fire Insurance Maps of Pennsylvania were found to have around 3.75-m horizontal positional accuracy at 95% confidence (Piekielek 2017) – both examples of historical collections commonly housed in map libraries compare favorably in their characteristics to many contemporary geospatial data sources. When positional accuracy matters, even contemporary researchers would do well to consider what print resources are available for their area of interest by consulting the holdings of map libraries.

There are many advantages of distributing geospatial data production broadly, including to the “the crowd” (National Research Council 2006). Advantages include labor efficiency, immediacy of response following disaster, and opportunities for under-represented populations to have their communities and interests formalized on maps through the democratization of data production. One disadvantage is that the production methods and accuracy of the data being produced is often times unknown (Goodchild and Glennon 2010). This is in contrast to a former print map world where data producers (i.e. map producers) were almost exclusively authoritative government agencies that understood well and adhered to mapping accuracy standards (e.g. the US National Map Accuracy Standards; USGS 1999). Map accuracy is closely related to map scale and the technological limitations of manual cartographic methods. Despite these limitations, accuracy was often carefully considered in map production and was often explicitly reported on the map itself so that future users could ascertain for what uses it was appropriate. The geospatial information of best known accuracy for an area of interest may to this day still be contained on a print map.

Research on the past, present, and future

It is a given that print maps are required of most any historical spatial analyses and only recently have we begun to appreciate that historical information is also often of benefit to analyses of the present and future. During the print map era, geographic relationships were “tested” through the creation of new maps, often around a theme (e.g. demography or environment). The new maps were called “thematic maps” and due to the laborious nature of creating print maps and how static they were, updates were infrequent and learning was slow. Fast forward to the digital age and the now mature GIS and web-mapping tools have dramatically changed our ability to rapidly overlay geographic information from different sources and on different topics. The types of questions that can be asked and answered of a GIS is remarkable and has led to spatial revolutions in numerous academic fields of study. Although it is possible that the value of print maps was reduced for a short period of time during this transition to digital mapping, the following describes why they remain important to exploring the past, present, and future.

Researchers armed with a GIS quickly realized that not only could they overlay geographic information on different themes, but also on the same theme through time. Interest in mapping change through time gained prominence and led to things like the development and long-term maintenance of earth observation satellites like the U.S.-based LANDSAT and MODIS programs (and many others internationally). Our ability to describe long-term change in a spatially-explicit way has evolved quickly, even to the point of developing quantitative empirical models of change (this ignores deterministic models that do not require training data). Armed with the ability to quantitatively describe spatially-explicit change, researchers from many fields soon found themselves running models into the future to generate projections and scenarios of what could come to pass if many model assumptions hold true. A somewhat recent limitation identified by empirical modelers is the inadequately short time-frame of readily available geospatial data on which to train models (see Dobrowski et al. 2013 for just one example from the climate change impacts literature). Many models of climate change impacts are run to 2100, more than 80 years into the future and yet they are often trained on roughly 30 years of climate-impact data (the LANDSAT record most often used starting with LANDSAT 4 in 1984) or less (the MODIS satellite record is less than 20 years old). Understanding and predicting the long-term spatial dynamics of just about any human or natural system would benefit from historical datasets that reach at least as far into the past as one would like to project into the future (i.e. 80 years or more). Fortunately, there is a wealth of research quality spatial information about the world from the last 80 years and more. Much of this information is contained on print maps and housed in map libraries.

Navigation and reliability

By 2020 more than 80% of the world’s adult population may own a location enabled smartphone (McKinlay 2016). This will have obvious implications for daily navigation and route-finding around the world. And yet, ask anyone involved in outdoor recreation, search and rescue, adventure sports, or who works in a remote field study site what they rely on for location and navigation and most will tell you that a print map and compass is indispensable, even if only used as a backup when a global positioning system (GPS) and cellular or satellite phone are their primary, yet sometimes unreliable, tools. Geospatial technologies offer incredible location and navigation utility, and yet when peoples’ lives depend on it, they often express a preference for print maps and compasses. Even in areas of relatively good GPS and geospatial data coverage, contemporary navigation and route-finding can come with a host of problems that were not as prevalent in a print map world (McKinlay 2016). Many map libraries continue to collect and provide access to maps commonly used in outdoor recreation, navigation, and route-finding.

Teaching spatial thinking and literacy with maps

There are numerous benefits to learning to think spatially and increasing spatial literacy, including that the brain can even be enlarged (Smith 2011). Print maps, in their simplicity, stasis and low cost provide powerful aids to teaching spatial thinking and literacy in ways that computer mapping sometimes cannot. With the large format of a print map you can easily see an area in context in addition to seeing its details; you can see where you have come from and where you are going. Even some of the most technologically advanced geospatial experts and producers of geospatial technologies pay homage to the importance of “the fundamentals” of spatial thinking and teaching spatial literacy, which reinforces that print maps have value even in a digital world (Kerski 2008). Many map libraries collect, manage, and provide access to print maps specifically for the purposes of teaching spatial thinking and literacy. What is more, print maps can provide an instructional gateway to web mapping tools such as Carto or Neatline that are designed to be accessible, user-friendly and are often used to highlight the print map and the wealth of information they contain.

Conclusion

Although this editorial focused on the contemporary value of print collections, map libraries are also at the forefront of new activities like offering centralized geospatial services, involvement in spatial humanities projects, creating new and innovative spatial tools, demonstration projects, and turning their own print collections into geospatial data at “big data” scales. But the rate at which map libraries are losing support is of deep concern with, from our view, insufficient response in the scholarly literature. We do not want to define or create a list of the threats that many libraries are facing because each institution is unique. We are calling for our colleagues to define the threats they see, are facing or have faced and to put the work we do into context within the larger movements in geography/mapping/GIS, higher education, and information literacy to build compelling arguments for the value of map libraries and their collections.

Our hope is that this journal (JMGL) can play an important role in fostering the conversations around these emerging topics while continuing to be the go-to source for authoritative publications on the more traditional topics of map and imagery collection, management, archiving, and reference.

Marcy Bidney and Nathan Piekielek
Co-editors, Journal of Map & Geography Libraries