People seem to be misunderstanding this paper. It doesn't claim that any previous papers have overestimated contamination. That would only happen if scientists didn't routinely use blanks as a comparison, which they do.
E.g. "A procedural filter blank was created during each sample batch and analysed alongside the samples, to enumerate potential contamination that could have been introduced during the extraction process."
https://eprints.soton.ac.uk/476076/1/1_s2.0_S014765132300286...