1 Erratum to: Clim Dyn (2016) 47:3335–3357 DOI:10.1007/s00382-016-3029-3

The original article has been published online with 2 errors. The errors have been corrected as below:

  1. 1.

    The definition of pre-industrial oceanic potential oxygen (OPOpi) given in Eq. (7) reads:

$$ {\text{OPO}}_{{{\text{pi}}}} = {\text{ O}}_{{\text{2}}}{}^{*} + {\text{1}}.{\text{1}} \times {\text{C}}_{{{\text{pi}}}}{} ^{*} $$

but should appear as:

$$ {\text{OPO}}_{{{\text{pi}}}} = {\text{ O}}_{{\text{2}}} {}^{*} + {\text{1}}.{\text{1}} \times {\text{C}}_{{{\text{pi}}}}{} ^{*} - \left( {{\text{1}}/{\text{X}}_{{{\text{N}}_{{\text{2}}} }} } \right) \times {\text{N}}_{{\text{2}}} $$

where \({\text{O}}_{2}{}^{*}\) and \({\text{C}}_{{{\text{pi}}}}{}^{*}\) are the quasi-conservative ocean tracer concentrations tracking air–sea exchanges of O2 and pre-industrial CO2, N2 is the dissolved N2 ocean concentration and \({\text{X}}_{{{\text{N}}_{{\text{2}}} }} \) is the atmospheric mole fraction of N2 (\({\text{X}}_{{{\text{N}}_{{\text{2}}} }} \) = 0.7808).

This definition of OPOpi and the OPOpi data shown on Fig. 1 therefore include the changes in OPO due to air–sea fluxes of O2 and pre-industrial CO2 but also due to the thermally driven air–sea fluxes of N2, which are important for understanding changes in pre-industrial atmospheric potential oxygen (APOpi). Including the N2 contribution makes OPOpi a mirror tracer of APOpi across the air–sea interface, i.e. with equivalent air–sea fluxes. The data based OPOpi to potential temperature ratio derived from Fig. 1 includes the N2 contribution and is equal to −3.9 nmol/J. Note that without the N2 contribution, this ratio would equal −4.4 nmol/J.

  1. 2.

    In Table 1, the isopycnal defining the base of the thermocline reads \({\sigma _\theta }\) = 24.7 and should read \({\sigma _\theta }\) = 27.4.