Nature retracts its global drought paper after the authors found the error in their own code
Nature published a retraction note on 2 September for the paper that appeared online on 4 June 2025 and concluded global drought is accelerating. The authors write that after being notified of errors they found further problems themselves: averages that were not area-weighted, results from an older evaporation model in one figure, a timestepping error in the code computing regional drought areas, and an unexplained masking method. Lead author Solomon Gebrechorkos says the new number will be slightly lower.
Science··Night
The retraction note lists four separate problems
Nature published a retraction note on 2 September for the paper that appeared online on 4 June 2025 and concluded that global drought is accelerating. The authors write that they identified further issues themselves after being notified of errors. The note lists four of them: study-wide averages that did not use area-weighted grid cells, results from GLEAM v3 used instead of GLEAM4 in one extended-data figure, a timestepping error in the code computing regional drought-affected areas, and an unclear account of how masking was applied. Lead author Solomon H. Gebrechorkos works at the universities of Oxford and Southampton; three of the paper's 11 authors, including senior author Simon Dadson, are at the University of Oxford.[1], [2]
Dates were misread across the 1981-2017 and 2018-2022 slices
With the timestepping error, dates were not parsed correctly when the data were subset for the 1981-2017 and 2018-2022 periods. An arid-regions mask — areas with mean annual precipitation below 180 millimetres — was applied in one figure but not in another. In a statement sent through the University of Oxford, Gebrechorkos told Retraction Watch that the problem was a coding error rather than one of data or methodology. The authors say they were willing to make the corrections but, given their scope, chose to retract the paper and resubmit a fully revised version for review; all authors agree with the decision.[1], [2]
The paper had reported drought severity worsening by 40 per cent
The retracted paper reported that a rise in the atmosphere's evaporative demand had worsened drought severity by about 40 per cent; it has been cited 337 times according to Clarivate's Web of Science, and The New York Times covered the work. Nature had issued an expression of concern for the paper on May 14. Karl Ziemelis, the journal's chief applied and physical sciences editor, said the authors were invited to submit a new manuscript with corrected calculations, that it has not yet arrived and that it would be treated independently of the retracted paper. Gebrechorkos says the errors changed the reported percentages, that the new number will be slightly lower and that the paper's conclusion stands. It is the first retraction for any of the authors and the seventh for the journal this year.[2]
