PRESENTATIONS
Negotiating Complexity in the Management of Sensitive Digital Data
David Gadsby
National Park Service
Appropriate stewardship of sensitive archeological data necessarily involves overlapping and intertwined authorities, systems, and institutions. The authorities, in turn have different limits and requirements, while various entities have divergent purposes, needs, and protocols. Archeologists, librarians, data managers, planners, and resource managers who work with sensitive data require a thorough understanding of the nature of both the physical and digital resources and the contexts in which they are embedded. Healthy institutional relationships and the effective communication that comes with those relationships are as critical to ensuring appropriate access to data as they are to protecting them from unwanted disclosure. Legacy data present a particular challenge, especially when methods for their appropriate disposition are ambiguous. I describe some of the complexities of managing these data, and examine practices and procedures for the stewardship of older data and examine some consequences of inappropriate disclosure.
In my work as national inventory coordinator for the NPS’s Archeology Program, I’ve been forced to adjust my orientation with regard to openness about archaeological data. Before I was a data manager, much of my career was devoted to publicly funded and publicly interpreted archaeological work. While I haven’t changed my attitude about the importance of serving the public or of interpreting archaeological work as a service to the data and the discipline, I’ve had to adjust the stance of radical openness and engagement that I developed as a graduate student.
In an essay I wrote with Robert Chidester and Joe Flatman in 2009, (shortly after joining the NPS), I expressed the opinion that we could address most heritage problems in the U.S. through engagement and public education (Flatman, et al. 2012: 65 – 77). In the decade since writing that piece, I’ve come to recognize that uncritical openness can ultimately damage sites, and that the conditions under which archaeological information can be released are highly context dependent, and require the careful work of knowledgeable specialists. New technologies that enable mapping and sharing can make this work both easier –security measures can be built directly into applications and metadata – and more difficult – electronic media are much easier to share and spread more rapidly than they did when my career began.
With my research partner (Chidester), I completed my graduate research in Hampden, a gentrifying working class neighborhood in North-Central Baltimore. Bob and I were interested in using archaeology to foster a genuinely open dialogue about class history and public heritage in a neighborhood that was undergoing rapid social change (Gadsby and Chidester 2009). We spent a lot of effort trying to capture public interest, and we had very little interest in keeping site locations secure – virtually every place in the neighborhood was part of an archaeological site, and the greatest security danger was to our equipment (we once had a camera snatched by a tour participant during the tour) rather than to the sites themselves. Hampden sites were safe because they were in plain sight and they were relatively recent historic sites. We never felt that it was important to take extraordinary measures to protect site locations because the locations were in the middle of an urban space and most people in the community knew where we were working and what we were up to. The sensitivity of the site location information (or rather the lack thereof) was determined by the social and physical conditions of the research environment.
As a steward of archaeological site information for the NPS, I find myself in an entirely different situation; a substantial body of law, regulation, and NPS policy1 – including but not limited to the National Historic Preservation Act (NHPA)2 and Archaeological Resources Protection Act (ARPA)3 – condition how my colleagues and I deal with data that we generate and curate in the name of the public and with the public trust. NHPA charges the NPS with taking inventory of the historic properties, including more than 80,000 archeological sites on its more than 85 million acres in every state, the District of Columbia, some territories. That inventory is a dataset that includes sensitive information. Fortunately, NPS sites, which lie on NPS lands, are protected from release under ARPA which even prevents release of sensitive information even under FOIA (Hitchcock 2006:469).
These various guideposts contain multiple requirements and prohibitions that stack and overlap in some places, and leave gaps in others. For instance, the language of NHPA requires Federal agency officials and others to withhold sensitive information under specific circumstances, while ARPA prohibits disclosure of information concerning the nature and location of protected archeological resources on Federal and Tribal lands under nearly all circumstances, but ARPA only applies to resources older than 99 years. ARPA has a definitive FOIA exemption, while section 304 is broader. ARPA applies only to archeological sites on Federal or Tribal lands, while NHPA, of course, applies to historic properties regardless of their location. Statute generally prevents the release of sensitive information, but not always, and the circumstances of release depend a great deal on the specific situation and location of the site or property under discussion. There will always be data that slip through the cracks, and the cracks, as I’ll explain, can be pretty large.
In general, U.S. government policy and statute directs openness; although ARPA and NHPA require some secrecy, and NPS policy follows ARPA and NHPA. In addition to these legal requirements, we may also have agreements with interested parties (such as tribes and SHPOs) that further our ability to share information.
In the NPS, we sometimes hear expressions of frustration from our colleagues from other disciplines, particularly in planning and disaster response. They want to know why we can’t just open the data. It would help with planning and compliance. (It might.) Don’t we want to be included? (We do.) Other disciplines aren’t nearly so cagey about their sensitive data (some are). Isn’t the trend in government toward openness and transparency? The data aren’t, after all, classified.
Despite these protests, the unrestricted release of site locations would be against the law for reasons I’ve outlined above. Also, in the NPS operating context, it would be very foolish. The consequences of a serious release of sensitive site location data can mean terrible damage or even total destruction of an archeological site. Looters are organized and relentless. They can be very sophisticated, and they can come from within the ranks of the agency and even from law enforcement. For example, an archeologist colleague –now the superintendent of a Civil War park – recalls an incident in which a park’s law enforcement officer brought a friend to “relic hunt” at a site after seeing archeologists working there, digging dozens of holes, and damaging buried features. The damage to the looted features impacted the site’s integrity, requiring additional fieldwork to re-establish it (Bies Pers Comm. 3/28/2018). A search of the NPS Archeological Sites Management Inventory System (ASMIS) reveals more than 8,000 recorded incidents of looting, vandalism, and unauthorized collection on the approximately 80,000 sites in inventory. This is almost certainly just a drop in the bucket.
Site looting and vandalism damage archaeological resources, reducing their integrity and destroying information about the past. But the consequences can reach well beyond the site level. Scholars of heritage looting are beginning document the ways illicit trafficking in looted or stolen antiquities, a practice tied to a widespread international black and gray markets (per Mackenzie and Yates 2017) is also related to other crimes – such as drug and human trafficking– and often has both legitimate and illegitimate international components. Objects of cultural and spiritual importance to American Indian tribes, many looted or collected before the advent of heritage protection laws find their way frequently to auction in Paris (General Accounting Office 2018). We’ve also seen in recent decades, several famous examples in which heritage site destruction – at Bamyan, Nimrud, and Aleppo, for example – can be employed as a tool or weapon in international conflicts, while portable antiquities from those sites appear for sale (Levinthal et al. 2014, Turku 2018). At the same time, organizations like the US. Committee of the Blue Shield work to protect cultural heritage from destruction by the military by gathering and sharing information on heritage sites on “no-strike” lists (USCB 2019).
Part of my point here is this: just as is true in the practice of archeological interpretation, for the purposes of understanding how to handle sensitive archaeological information, context is everything, and understanding context means applying skillful critical thinking. The consequences of unauthorized release of information vary in scale and impact from situation to situation. We can’t necessarily build security protocols that will solve the problem. Some openness seems to work better in non-US contexts than we could expect it to in the United States. For instance, the Scottish Coastal Heritage and the Problem of Erosion (SCAPE) project is able to use crowd sourced information and publicly available GIS mapping to understand the condition of archeological sites along the Scottish coast in a way that is virtually unthinkable in the United States. (SCAPE Trust 2019 a, b).
There are allies that we want to share information with, but need to be cognizant that the sharing is warranted, and is ultimately safe for sites. The United States has a tradition of avocational archaeologists that goes back to the roots of the discipline, but as archaeology has professionalized, we have increasingly required a lot of training for non-professionals who wish to participate in archaeological work or to have access to site locations. Examples include the Archeological Society of Maryland’s Certified Archeological Technician (CAT) program and the Site Stewards Foundation’s Site Stewards program in several South Western states, both of which require significant training in archaeological technique, knowledge, and ethics (Archaeological Society of Maryland 2019, Site Stewards Foundation 2019). NPS relies to some degree on non-professionals who work in parks to monitor sites for signs of deterioration or looting, but that reliance is limited to a few programs like the Midwest Archeological Center (MWAC)’s ranger monitoring program.
Agreement amongst archaeologists is high that unauthorized disclosure of sensitive information poses a risk to resources (Frank et al 2015: 8). Indeed, the SAA code of ethics requires archeologists to consider the protection of sites when publishing and distributing information about them (Society for American Archeology 2019; Lynott 1997). My sense is that most of us understand all of this at an almost instinctual level; it is part of our disciplinary DNA. The risk of archaeologists intentionally releasing data that should be kept secret is relatively low. Complications arise when our sensitive data moves outside of the cultural resource domain and into other domains (GIS, Planning, engineering, libraries and archives), where considerations and sensibilities of the practitioners are different. We also see problems arise with older materials that, while still sensitive, were generated with less consideration for their confidentiality. There are analogs for this situation in other disciplines – (Frank et al. point to personally identifiable information (PII) in public health and other social sciences) and those disciplines have begun to develop best practices for dealing with this information (e.g. Zandbergen 2016)
The success, however, of protocols or best practices, depends on the presence of trained professionals to enforce them. In the NPS, we’re seeing a decrease in specialists of all kinds, as the service moves toward a model that uses more generalized resource managers rather than specialists with particular knowledge about archeological resources. Human Resources reports from 2016 and 2018 show a 6% decline in the NPS’s archeological workforce in just two years. While that’s not a precipitous decline, it is one that, if sustained, will certainly deliver a blow to the service’s capacity to protect archeological sites.
Institutions like the NPS need trained professionals to ensure that our data curation practices foster resource protection, while allowing for preservation planning, and fulfilling as much as possible our mandates to be open with data. The nature of our resources also make it necessary that we pay attention to problems of consultation and ownership. Tribes, SHPOs, and other federal agencies expect us to protect data in which they may claim (rightfully or not) to possess a proprietary interest. So we have a need to withhold sensitive data in most cases from most people, but there are also important instances where sharing data with appropriate consumers can, of course also protect sites.
As I’ve said, the consequences of inappropriate sharing or release are high, particularly for NPS and USG, and need to be careful about balancing the need to interpret our work (which is, of course, very important) with the need to protect resources Consequences include, but probably are not limited to:
1) Violating the law, including the statutes above, as well as a series of NPS policies;
2) Impinging on the protection of/enabling damage to protected resources by exposing them to looting and vandalism; and therefore
3) Damaging to the scientific/archaeological value of sites;
4) Enabling the sale of looted antiquities through legitimate or illegitimate sources; and finally
5) Causing or compounding real pain and trauma to communities that lose heritage (e.g. Phippen 2016, Weston 2016). Archeological sites are, in many cases, a source of cultural identity and memory, and attacking them is a real attack on culture. GAO identified 15 Paris auctions since 2012 of Native American archaeological or ethnographic materials (GAO 2018:7). The impacts of these auctions on tribes and their members continues to be documented.4
On the other hand, the consequences of sharing too little can include the inadvertent destruction of sites by development activities, but can also cause archeologists to appear to be hoarding information and acting in bad faith.
As a profession, we need to think carefully about why we collect and preserve sensitive digital data. Is it in the appropriate environment? Is it in the appropriate format? Is it clearly labelled as sensitive, so that a non-specialist will know that it is not for public release?
I think we are all aware of the possibility that bad actors might gain access to sensitive data and share it for the purpose of damaging sites and committing illegal acts. Take for instance, a 2015 thread from on the looter chat board Treasure Hunter, which shares coordinates for more than a dozen submerged archeological sites and offers willing “treasure hunters” the opportunity to pay a known treasure hunter \$1000 to “work” these cultural resources. (Treasure Hunter2015).
Bad actors with access to site locations can certainly present a problem. Also problematic, however, are those with good intentions but imperfect knowledge of the sensitive nature of archeological resources and data, and circumstances in which such people gain control of archeological data may constitute unforeseen weaknesses in the loose system that controls archaeological information. These instances are a bit more complex with regard to controlling sensitive information. Recently, several NPS archeologists discovered that our online technical library had released a new public-facing website that exposed several un-redacted survey and overview and identification reports. The ensuing conversations with the administrators of that website revealed that 1) they didn’t believe they had done anything wrong, and in fact believed that they were fulfilling government mandates for transparency, 2) they had not communicated with ANY archeologists or cultural resources prior to releasing the data, and 3) they believed they had fulfilled their due diligence prior to releasing the information by performing internet searches. When archaeologists reviewed the website, they found several documents, in particular older survey reports, that were not appropriate for public sharing. The archivists took them down, but by then, of course, the cat was out of the bag. The administrators did agree, after intense discussion, to contact the chief archaeologist before releasing any more documents that might contain sensitive data, and to remove the ones identified by archaeologists as inappropriate.
Archeologists need to educate people who deal with our data, such as librarians, information resource specialists, and GIS specialists about the established norms, rules, and laws that protect our discipline’s sensitive data. These people are educable; but we’ve also encountered circumstances where apparently intelligent people who seem to have good intentions repeatedly gain control of and share sensitive data. For instance, a former NPS historian (Butowsky 2019) on an apparent mission to share as many NPS documents as possible has placed numerous sensitive documents on the web, despite numerous requests that he not do so. While he sometimes asks permission from park personnel before posting documents, our shrinking archaeological workforce means that when he reaches out to parks to request their data, he may wind up contacting a non-specialist who won’t understand or respect the need to keep sensitive data private.
I use these examples to illustrate that sensitive archaeological data released carelessly but with good intentions can be used by those with bad intentions. Particularly vulnerable are older (pre-digital) datasets that may be published in gray literature and lack such safeguards. We can certainly continue to build safeguards for our digital data as our colleagues in other disciplines have done for some time (e.g. Zandbergen 2014) , but non-digital legacy datasets in archives or office drawers often don’t have the same safeguards. There aren’t any pat solutions in in such cases. Archeologists need to be vigilant for improper releases of data, but we also have an opportunity to reach out to others who may have access to our data and educate them on the real-world consequences of such releases.
References:
Archaeological Society of Maryland
2019 Certified Archeological Technician Program
http://marylandarcheology.org/CATprogram.html (Accessed 25 March)
Bies, Brandon
2019 Personal Communication. April 28, 2019.
Butowsky, Harry A. and Randall D Payne
2019 Cultural Resources: Anthropology/Archeology.
http://npshistory.com/cultural_resources.htm#anthro-archeo Accessed
3/25/2019.
Chidester, Robert C. and David Gadsby
2009 One Neighborhood, Two Communities: The Public Archaeology of Class
in a Gentrifying Urban Neighborhood
International Labor and Working-Class History No. 76:3 (127–146)
https://www.academia.edu/539716/One_Neighborhood_Two_Communities_The_Public_Archaeology_of_Class_in_a_Gentrifying_Urban_Neighborhood.
Accessed 3/27/2019.
Flatman, Joe, Robert Chidester, and David Gadsby
2012 What Public Engagement in Archaeology Really Means. In Archaeology
in Society: Its Relevance in the Modern World, Marcy Rockman and Joe
Flatman (eds). Springer, New York.
Frank, Rebecca, Adam Kreisberg, Elizabeth Yakel, and Ixchel M. Faniel
2015 Looting Hoards of Gold and Poaching Spotted Owls: Data
Confidentiality Among Archaeologists & Zoologists
https://www.asist.org/files/meetings/am15/proceedings/openpage15.html
accessed
3/22/2019.
General Accounting Office
2018 NATIVE AMERICAN CULTURAL PROPERTY Additional Agency Actions Needed
to Assist Tribes with Repatriating Items from Overseas Auctions
https://www.gao.gov/products/GAO-18-537, Accessed May 25, 2019
Hitchcock, Ann
2006 FOIA and Protecting Cultural Resources People, Places, and Parks.
Proceedings of the 2005 George Wright Society Conference on Parks,
Protected Areas, and Cultural Sites. David Harmon (ed). Pp. 468 – 474.
http://www.georgewright.org/0581hitchcock.pdf
accessed 3/25/2019.
Levinthal, Richard, Brian I. Daniels, Corine Wegener, and Susan Wolfenbarger
2014 Ancient History, Modern Destruction Assessing the Status of Syria’s World Heritage Sites Using High-Resolution Satellite Imagery. American Association for the Advancement of Science (AAAS), Washington, DC. http://www.aaas.org/sites/default/files/content_files/AAAS-SyrianTWHS-122014.pdf Accessed 3/28/2019.
Lynott, M. J.
1997. Ethical principles and archaeological practice: development of an
ethics policy. American Antiquity, 589–599.
Mackenzie, Simon and Donna Yates
2017 What Is Grey about the “Grey Market” in Antiquities?
In The architecture of illegal markets : towards an economic sociology
of illegality in the economy. Jens Beckert and Matias Dewey eds. Oxford
University Press, New York.
Phippen, J. Weston
2016 The Auction of Native American Artifacts, The
Atlantic MAY 27, 2016
https://www.theatlantic.com/international/archive/2016/05/native-american-auction/484316/
accesssed 3/25/2019.
Society for American Archaeology
1996 Principles of Archaeological Ethics.
https://www.saa.org/career-practice/ethics-in-professional-archaeology
Accessed 3/28/2019
Site Stewards Foundation
2019 Site Stewards Foundation
http://www.sitestewardfoundation.org/index.php Accessed 3/28/2019.
SCAPE Trust
2019a Scotland’s Coastal Heritage At Risk http://www.scharp.co.uk/
accessed 3/25/2019
2019b Sites at Risk (map)
http://www.scharp.co.uk/sites-at-risk/#zoom=6&lat=8089114.80835&lon=-350045.02884&layers=B00000FT
accessed 3/25/2019
TreasureHunter
2015 1715 Treasure Fleet GPS
Cordinates
Online discussion board:
http://www.treasurenet.com/forums/beach-shallow-water/399418-1715-treasure-fleet-gps-cordinates-2.html.
Accessed 3/25/2019.
United States Committee of the Blue Shield (USCBS)
2019 Cultural Heritage Inventories and No-Strike Lists.
https://uscbs.org/cultural-heritage-inventories.html Accessed
3/28/2019.
Zandbergen P. A.
2014 Ensuring Confidentiality of Geocoded Health Data: Assessing
Geographic Masking Strategies for Individual-Level Data. Advances in
medicine, 2014, 567049.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4590956/
Zellmer, Linda R.
2017 Protected Places: A Survey of Laws on Archaeological Site and Cave
Location Confidentiality and Their Potential Impact on Library Reference
Policies and Services. Journal of Map & Geography
Libraries, 13:2, 148-174, DOI: 10.1080⁄15420353.2016.1249446
Including NPS Director’s Order 28a and 2006 Management Policies
[return]NHPA Section 304 (54 USC 307103 et seq, Protection of Historic Properties, at 36 CFR 800.11 ©(1)
[return]ARPA Section 9 16 USC 470hh, Protection of Archaeological Resources at 43 CFR 7.18
[return]An auction of Arts premiers - Art Precolombien et Amerindien, **originally scheduled for early April 2019 - the day of our session - appears to have been rescheduled for June 6. The catalog has not yet been published.
[return]