4x2 timeseries plot of various humidity metrics and datasets. Coloured (red/blue) vertical bars indicate ENSO state. Left column - in situ data (HadISDH), Right column - reanalyses data (ERA5, MERRA-2, JRA-3Q). Top two rows - specificy humidity (land, then ocean), bottom two rows - relative humidity (land, then ocean).
Here's the BAMS SotC timeseries image I referred to. Asking Kate Willett (HadISDH developer & Humidity expert), that rise in RH is mostly driven by the marine RH which "went a bit wierd".
(Not a record land value in RH - I was looking at specific humidity by mistake)