DC Archives Scramble to Replace Thousands of Corrupted Digital Images

The District of Columbia Public Library system confirmed this week that a routine audit of its digital collections uncovered thousands of duplicate and corrupted image files spread across its servers, prompting an emergency remediation effort that began Monday at the Martin Luther King Jr. Memorial Library on G Street NW. The problem, which archivists say stems from years of inconsistent file migration protocols, has affected image records dating back to scans completed between 2018 and 2022.

The timing matters. DC's cultural institutions are in the middle of a broader push to digitize and make publicly accessible historical photographs, maps, and documents, efforts that gained urgency after federal funding uncertainty under ongoing restructuring left some grant-dependent programs on shakier ground. When duplicate or corrupted images sit in a collection, they don't just waste storage space; they can surface in public-facing portals, displacing the correct files and undermining the credibility of the entire archive.

What Went Wrong, and Where

The issue is not confined to the main library branch. Staff at the Washingtoniana Division, the specialized local history unit housed within the MLK library system, identified at least three separate instances this week where replacement images uploaded to the DC Digital Collections portal had themselves been duplicated, overwriting original high-resolution scans with lower-quality copies. The Washingtoniana Division holds more than 100,000 photographs documenting DC neighborhoods including Anacostia, Shaw, and LeDroit Park, communities whose visual histories are already underrepresented in national archives.

The Historical Society of Washington DC, based at the Carnegie Library on Mount Vernon Square, flagged a related problem on Tuesday. Its online catalog, which shares a back-end content management system with several partner institutions, began displaying mismatched thumbnail images, a 1940s photograph of U Street NW appearing under a caption for a Capitol Hill streetcar line, for instance. Staff pulled approximately 340 catalog entries from public view while the mismatch is investigated and correct source files are located and re-uploaded.

Duplicate image problems in digital archives are more common than institutions typically acknowledge publicly. A 2023 report from the Digital Preservation Coalition found that roughly 23 percent of mid-sized institutional repositories surveyed had experienced at least one significant duplicate-file incident requiring manual remediation in the preceding two years. For smaller municipal archives operating with limited IT staff, the rate was higher. DC's library system employs a digital services team that, according to its published 2025 budget documentation, received $1.4 million in operational funding, a figure that covers software licensing, storage infrastructure, and personnel across all branch locations.

The Fix, and What Comes Next

Remediation at the MLK library is being handled in two phases. The first, which library staff say will run through July 11, involves running automated deduplication scripts across the primary server cluster to flag identical or near-identical files. The second phase requires human review: archivists must compare flagged files against physical originals or verified master copies to confirm which version is authoritative before the corrupted or duplicate file is permanently deleted.

The Historical Society is taking a parallel approach, working with its content management vendor to restore catalog entries from a backup snapshot taken in early June. That snapshot predates the most recent batch of uploads that introduced the mismatched files, meaning roughly six weeks of catalog updates will need to be manually re-entered once the clean version is restored.

For residents or researchers who rely on DC Digital Collections, a portal that logged more than 80,000 unique visits in 2025, according to the library system's annual report, the practical advice is straightforward: if you download an image from the portal this week and the metadata seems inconsistent with the content, flag it using the feedback form on the site rather than assuming the record is correct. Both institutions say they are prioritizing corrections for the Anacostia and NoMa neighborhood collections, where redevelopment pressures have made historical documentation especially contested and valuable. The full audit is expected to conclude before the end of July.