1 Correction to: Climate Dynamics (2022) 59:753–766 https://doi.org/10.1007/s00382-022-06153-z

The following Erratum aims at correcting Panel (a) in Fig. 1 and a typo in Fig. 2.

(a) Correcting Fig. 1

In Fig. 1a we showed the mid- to late Holocene evolution in annual mean insolation at the top of the atmosphere in the Northern and Southern Hemisphere (hereafter NH and SH). Following Kepler 2nd law, the insolation at the top of the atmosphere should be symmetric about the equator.

Figure 1a shows very small differences in the annual insolation for the North and South Hemisphere. This was an oversight. The absolute differences between the NH and SH curves in Fig. 1a has the following statistics:

mean = 0.92 W/m2; standard deviation = 0.10 W/m2; max = 1.02 W/m2; min = 0.67 W/m2.

We found two errors in the computation of annual insolation in the NH and SH. We used the CDO package (https://code.mpimet.mpg.de/projects/cdo/) and realized that the annual mean was computed simply by summing up the monthly values and dividing by 12 (function yearmean), while it should account for differences in number of days in each month (function yearmonmean). Moreover, another error comes from the way the package dealt with data at the poles.

In the new Fig. 1 (see below), we show the correct annual isolation, and has been compared using both Python-xarray and the Ferret packages. Results for seasons JJAS and DJFM in panel (b) show no change. We confirm that all other computations are not affected.

The absolute differences between the NH and SH curves in Fig. 1a now has the following corrected statistics:

mean = 0.20 W/m2; standard deviation = 0.10 W/m2; max = 0.30 W/m2; min = 6.1 × 10–5 W/m2.

Such small differences are expected and come from the fact that the date of the vernal equinox is prescribed to be March 21 at noon. The number of days between the perihelia and the vernal equinox varies with time due to precession. The differences between hemisphere results from small rounding of the calculation given that this version of the model has a complex noleap calendar that is also used for the paleoclimate transient simulation. These differences are negligible if compared to the seasonal differences as shown in the second Figure in this Erratum.

figure a

Correct Fig. 1. Evolution of the incoming solar radiation [W/m2] at the top of the atmosphere and trace gases in the last 6000 years (0 kyr BP = 1950). Panel a: annual mean incoming solar radiation averaged in the Northern Hemisphere (NH) [0–90 N] and Southern Hemisphere (SH) [0–90 S]. Panel b: solar radiation averaged for the months of June to September (JJAS) in the NH and December to March (DJFM) in the SH. Panel c: annual mean of CO2 [ppm], CH4 [ppb], and N2O [ppb] averaged globally. Both radiation and trace gases forcing are updated yearly. Details on the forcing can be found in Braconnot et al. (2019b)

figure b

Figure. Evolution of the incoming solar radiation [W/m2] at the top of the atmosphere and trace gases in the last 6000 years (0 kyr BP = 1950). Panel a: annual mean incoming solar radiation averaged in the Northern Hemisphere (NH) [0–90 N] and Southern Hemisphere (SH) [0–90 S] in blue and orange (dashed) lines. Solar radiation averaged over the months of June to September (JJAS) in the NH and December to March (DJFM) in the SH in green and red, respectively. Panel b: annual mean of CO2 [ppm], CH4 [ppb], and N2O [ppb] averaged globally. Both radiation and trace gases forcing are updated yearly. Details on the forcing can be found in Braconnot et al. (2019b)

We acknowledge Juan Lora for pointing out the error in Fig. 1.

(b) Typo in Fig. 2

In Fig. 2, Panel b we show the first and second EOF of the IPSL model in the Indian Ocean. The subplot on the bottom right shows EOF2 in the Indian Ocean in the IPSL model. The label “IPSL model. 1850–1950, EOF1, 12.6%” should be corrected to “IPSL model. 1850–1950, EOF2, 12.6%”.