Software now allows archaeologists to document excavations in more detail than ever before through rich, born-digital datasets. In comparison, paper documentation of past excavations (a valuable corpus of legacy data) is prohibitively difficult to work with. This pilot study explores creating custom software to digitize paper field notes from the 1970s excavations of the Gulkana site into machine-readable text and maps to be compatible with born-digital data from subsequent excavations in the 1990s. This site, located in Alaska's Copper River Basin, is important to archaeological understanding of metalworking innovation by precontact Northern Dene people, but is underrepresented in the literature because no comprehensive map of the site exists. The process and results of digitizing this corpus are presented in hopes of aiding similar efforts by other researchers.

The Early and Middle Bronze Age in the Carpathian Basin is often viewed as a long period of transition from a dispersed form of land occupation to one of increasing aggregation, ultimately resulting in the formation of tell settlements and large cemeteries. This developmental trajectory remains a legacy of early 20th century archaeology, where the similarity of material culture recovered from cemeteries and settlements was used to develop a multi-linear scheme of progression of regional chronologies tied to specific archaeological cultures. While typologically conclusive, the recent increase in the availability of radiocarbon determinations suggests that these sequences represent a priori interpretation of social development rather than empirically verified observations. In order to do so, it is necessary to re-evaluate the existing dataset in order to determine whether the formation of tells was a chronologically contemporary development and whether the regional chronological sequences are supported by independent dating.

As archaeologists expand the accessibility of legacy data, they have an opportunity to use these datasets to design future research. We argue that legacy data can be a critical resource to help predict characteristics of sites and socioeconomic systems. In this article, we present a combined geographic information system (GIS) and network analysis methodology that turns site location data into testable hypotheses about site characteristics and the organization of regional settlement systems. We demonstrate the utility of this approach with a case study: Bronze Age (2700–1100 BC) settlement patterns in the mining region of Hunedoara in southwest Transylvania, Romania. We leverage unsystematically collected site location information in legacy datasets to develop testable predictions about sites, regional networks, and socioeconomic systems that can be evaluated through future systematic surveys and large-scale excavations. Such testable hypotheses can inform archaeological research design by providing a quantitative basis for determining where to focus research efforts and can also help secure funding and fieldwork permits. The method developed here can be applied in diverse archaeological contexts to reinvigorate legacy data as part of future archaeological research design.

We consider the problem of providing optimal uncertainty quantification (UQ) – and hencerigorous certification – for partially-observed functions. We present a UQ frameworkwithin which the observations may be small or large in number, and need not carryinformation about the probability distribution of the system in operation. The UQobjectives are posed as optimization problems, the solutions of which are optimal boundson the quantities of interest; we consider two typical settings, namely parametersensitivities (McDiarmid diameters) and output deviation (or failure) probabilities. Thesolutions of these optimization problems depend non-trivially (even non-monotonically anddiscontinuously) upon the specified legacy data. Furthermore, the extreme values are oftendetermined by only a few members of the data set; in our principal physically-motivatedexample, the bounds are determined by just 2 out of 32 data points, and the remaindercarry no information and could be neglected without changing the final answer. We proposean analogue of the simplex algorithm from linear programming that uses these observationsto offer efficient and rigorous UQ for high-dimensional systems with high-cardinalitylegacy data. These findings suggest natural methods for selecting optimal (maximallyinformative) next experiments.