June 9, 2020

COVID-19 pandemic

Disputed Wuhan parking-lot preprint draws worldwide attention

Researchers tried to look backward through a pandemic using cars visible from space. The study never found the virus — its signal was fuller parking lots and more searches for 'diarrhea.'

9 June 2020: satellites, searches, and a headline

On 9 June 2020, news organizations worldwide reported on a preprint hosted in Harvard's DASH repository that used satellite counts of vehicles outside six Wuhan hospitals, plus Baidu search trends for symptom terms, to look for signs of unusual disease activity before December 2019.

The authors — led by Boston University's Elaine Nsoesie, with Harvard Medical School's John S. Brownstein — reported rising hospital traffic and 'diarrhea' searches beginning in late summer 2019, with hospital traffic peaking in December. The manuscript was posted on 8 June; broad news coverage and expert reaction followed on 9 June.

What the data actually was

The study analyzed 111 satellite images spanning January 2018 to April 2020. In one published comparison, a hospital lot held 171 cars in October 2018 and 285 about a year later — a 67% increase. Crucially, the authors themselves framed the signals as possible evidence of an unidentified respiratory outbreak, not detection of SARS-CoV-2, and acknowledged the data could not establish a cause.

The same day, independent experts pressed the weaknesses: not peer reviewed, purely correlational, no control analyses from comparable cities outside Hubei — and critics found crowded images from other dates that complicated the comparisons.

How a hedge became a headline

The coverage introduced two notable slippages. A manuscript stored in a Harvard repository became a 'Harvard University study,' though hosting is neither peer review nor institutional endorsement. And the authors' hedged conclusion about possible fall-2019 disease activity sharpened, in headlines, into a claim that COVID-19 was spreading in August.

The preprint drew immediate scientific criticism and rejection by Chinese officials, and it did not revise the accepted outbreak timeline — no August patient, clinical specimen, or viral genome was ever part of it.

Sources

Researched 23 Aug 2026 5 sources 3 audit passes Date disputed

How this was checked

Researched from the web into a fact sheet, rewritten from that sheet, then audited against it by a different model. How the pipeline works →

Date disputed. Sources disagree about when this happened.

What the sources leave uncertain

The manuscript was posted June 8; June 9 is the date on which the claim received broad news coverage and expert reaction.
The study was a non-peer-reviewed preprint and did not test any August patient, clinical specimen or viral genome.
Parking occupancy and symptom searches have many possible causes. The observed correlation cannot identify COVID-19 specifically.
Experts also questioned image selection, vehicle-count comparability and the lack of control cities. Consequently, the claim of August circulation remains unsupported by this study alone.