As of September 2026, we have over 28,500 archaeological sites recorded in Pennsylvania. This reflects nearly 100 years of Pennsylvanian’s appreciating and recording information on the archaeological evidence of past populations. Of those sites, over 21,900 are pre-contact period archaeological sites – sites that date to the thousands of years that people lived in what is now Pennsylvania before Europeans arrived.
While that is a lot of archaeological sites, we also know that number is only a fraction of the total number of archaeological sites that exist but have not yet been discovered.
Part of the Environmental Review process is assessing the potential for new ground disturbance to impact archaeological sites. Sometimes that new ground disturbance is proposed in a location where we have a previously recorded archaeological site in our files. But far more often we are assessing the potential for archaeological sites to be present in a project area based on what we know of human behavior and culture, patterning of archaeological sites, and environmental factors.
For the historic period and contact period, we can use historic maps and historic records to help identify the potential for archaeological sites to be present. For the pre-contact period, one way to assess potential for archaeological sites systematically is through archaeological modeling.
The Pennsylvania Statewide Pre-Contact Predictive Model (PA SPPM)
Between 2013 and 2015 the Pennsylvania Statewide Pre-contact Predictive Model set (PA SPPM) was developed for PennDOT, FHWA, and PHMC to aid in assessing archaeological potential during project planning. Using decades of information recorded for pre-contact period archaeological sites, a regional model set was developed for the entire commonwealth. A PA SHPO blog post from 2016 highlights the development of this model and you can check that out here: https://pahistoricpreservation.com/mapping-probability-pre-historic-archaeological-sites/.
If you clicked on the link above, you’ll notice there was some discussion in that post looking forward: “The actual final project deliverable is a set of algorithms that can be run again and again to update the model in the future as more data are received… The models are in the early stages of their lifecycle and will be continually evaluated and occasionally updated.”
Unfortunately, the way the model was developed and shared with PennDOT, FHWA, and PHMC kept us from ever being able to update this specific model set. Still, we needed to check in to evaluate how it was performing. So, in 2025, PA SHPO contracted Penn State University to conduct a re-analysis of the PA SPPM with a decade of new data.
PA SPPM Re-analysis
It’s important to point out that the PA SPPM is not one model, but a regional model set made up of 132 small models that fall into 10 regions. For the re-analysis of the model, Penn State University used these 10 regions to assess how the model sets were performing within each region.
PA SPPM Regions (From Harper et al. 2025)
The report of Penn State University’s analysis (Harper et al. 2025) is available in PA-SHARE. The report includes archaeological site location information; however, a redacted version of the report is available here: Report # 2026RP00288.
If you have archaeological privileges in PA-SHARE, you can see the full report and accompanying data and shapefiles here: Report # 2025RP00305.
Results of the Re-analysis
To assess how the PA SPPM has performed over the last ten years we need to look at several different metrics. That information is summarized in the table below, which lays out some of the nuance of the re-analysis. Some definitions to help with the table below:
Sensitivity/True Positive Rate: The fraction of actual site-present cells correctly classified as site-likely (i.e. sites are found in moderate and high probability areas of the PA SPPM);
Specificity/True Negative Rate: The fraction of actual background cells correctly classified as site-unlikely (i.e., sites are not found after surveying in low probability areas of the PA SPPM);
Kvamme Gain (Kg): A metric assessing how much more likely a site discovery is in the predicted site-likely area compared to random surveying.
Results of the re-analysis.
In short, in Regions 1, 4, 7, and 10, the model performs better than average. The model also performs well in Regions 2, 6, and 8, with some caveats as discussed in the table above. In Regions 3, 5, and 9, archaeologists should use discretion in use of the model, weighting professional judgment much more heavily than the PA SPPM. Large scale projects in Regions 3, 5, and 9 would be best suited to developing project specific predictive models.
Looking Forward
As discussed above, the existing PA SPPM cannot be updated and will continue to live on as a static layer in PA-SHARE. The information in the table above has been added as a new layer in PA-SHARE, mapped to the 10 PA SPPM regions. PA SHPO continues to recommend use of the model for project planning in regions where the model is performing well. We will also continue to consider it as part of our reviews in the Environmental Review process.
Ultimately, as has become a common chorus here at the PA SHPO, the PA SPPM should be considered as one of many tools to assess the potential for precontact period archaeological sites. It should not ever be used without an archaeologist applying professional judgement to its application. In some regions, it should just be considered a more useful tool than in other regions.
A Note on Reporting for Professional Archaeologists
And a final note for professional archaeologists working in Pennsylvania: PA SHPO no longer requires a separate section to discuss the PA SPPM performance in Phase I archaeological survey reports, which is reflected in the most recent update to our Guidelines for Archaeological Investigations revised in August 2026. Any application of the PA SPPM should continue to be discussed in the Research and Field Methodology section. The Negative Survey Form will be updated soon to reflect this change in reporting.
Reference
Thomas K. Harper, S. Ilgin Guler, Faeze Fathijam, and Jacob Sholtis
2025 Pennsylvania Statewide Precontact Probability Model Re-Evaluation, Tasks 3 and 4: Final Report. Prepared for Pennsylvania Historical and Museum Commission, PA SHPO, September 2025.
Comment Policy
PHMC welcomes and encourages topic-related comments on this blog. PHMC reserves the right to remove comments that in PHMC’s discretion do not follow participation guidelines.
Commenters and Comments shall be related to the blog post topic and respectful of others who use this site.
Commenters and Comments shall not: use language that is offensive, inflammatory or provocative (this includes, but is not limited to, using profanity, obscene, or vulgar comments); disparage other commenters or people; condone illegal activity; identify the location of known or suspected archeological sites; post personal information in comments such as addresses, phone numbers, e-mail addresses or other contact details, which may relate to you or other individuals; impersonate or falsely claim to represent a person or an organization; make any commercial endorsement or promotion of any product, service or publication.
If you would like to comment on other topics not related to this blog post but related to PHMC, please fill out the PHMC Contact Us Form.
Leave a Reply