2.1 Data collection

Using bibliographies of MDV publications (including grey literature) (Antarctic Division, D.S.I.R., 1985; Mead, 1978; New Zealand Antarctic Programme, 1995), recommendations from other researchers, and outreach to “Old Antarctic Explorer” organizations, we compiled a list of individuals who have worked in the MDV. We then contacted these individuals with requests for historical photographs and documents related to the MDV (Fig. 1). When we received interested replies, we collected the data using one of the following approaches. If data (especially photographs) were already available in digital format, we arranged for files to be sent to us electronically. When data were not available digitally, we either arranged for the contributors to digitize their documents locally through commercial scanning services, or we visited the researchers in person with a scanner (either a slide scanner or a flatbed scanner depending on the nature of the data). The in-person visits also facilitated oral history interviews with researchers, which we recorded, transcribed, and included in the archive (Fig. 2).

Figure 1Examples of photographs from the archive showing human activity in the MDV. Unless otherwise stated, copyright for each item remains with the respective contributor. Detailed metadata for each item are available on the archive website (http://mcmurdohistory.lternet.edu) and the Environmental Data Initiative (Howkins et al., 2019).

In addition to individuals, we also worked with universities, libraries, archives, and national Antarctic programs. Sometimes data were already easily accessible online, such as the Antarctica New Zealand's digital photograph collection (Antarctica New Zealand, 2017). More often, however, data were only available in non-digital formats, in which case we followed a digitization process similar to our work with the individual researchers. The resolution and provenance of the photographs and other data are variable, but an inclusive approach was taken to maximize the scope of the database. The full database is accessible via a browser-based interface hosted by the McMurdo Long Term Ecological Research site: http://mcmurdohistory.lternet.edu/. The complete metadata for all resources in the database are also available at the Environmental Data Initiative: https://doi.org/10.6073/pasta/6744cb28a544fda827805db123d36557 (Howkins et al., 2019).

2.2 Metadata structure

All data in the archive are organized according to the Dublin Core (DC), an international metadata element set intended to facilitate the discovery of electronic assets (Weibel, 1997; Weibel et al., 1998). The DC is designed to be simple and flexible, and we have tailored its core fields for the specific purposes of the archive. This provides a consistent metadata structure among the various types of resources while maintaining interoperability with other DC databases. The full list of fields and the associated definitions for each type of resource is available in the Supplement (Table S1).

2.3 Georeferencing

Although the DC metadata schema includes a “coverage” field for storing geospatial information, the MDV archival data cover a wide range of spatial scales with highly variable locational certainty. We therefore chose to assign each resource a geographic “place”; these range from the scale of individual huts to entire valleys. The majority of these places are associated with a set of existing geospatial vector layers acquired from the MDV Long Term Ecological Research site (Gardner, 2016), which we augmented to include additional features. The layer list includes the following geospatial vectors:

• camps, stations, and huts;

• glaciers;

• lakes and ponds;

• streams (monitored);

• streams (not monitored);

• stream gauges;

• meteorological stations;

• Dry Valleys Drilling Project sites;

• miscellaneous human features;

• miscellaneous natural features.

We decided that places such as valleys or mountain ranges were too general to assign a specific geospatial layer. However, these were still assigned a searchable term in the relational database and may be assigned a geospatial location as additional layers become available. The general bounding coordinates (decimal degrees) for the MDV and the data in the archive are as follows: north (lat −77.504), south (lat −77.642), east (long 164.319), west (long 161.111).

Figure 2Examples of non-photographic data available in the archive. Unless otherwise stated, copyright for each item remains with the respective contributor. Detailed metadata for each item are available on the archive website (http://mcmurdohistory.lternet.edu) and from the Environmental Data Initiative (Howkins et al., 2019).