Impact of different solar penetration depths on climate simulations

Three different estimates of shortwave attenuation depth (SWAD) of photosynthetically active radiation (PAR) derived from remotely sensed ocean colour data have been tested in an ocean general circulation model (OGCM) forced with interannual atmospheric forcings. Two estimates (referred to as [K_d(PAR)]⁻¹₁ and [K_d(PAR)]⁻¹₂) are calculated from different algorithms based on the diffusive attenuation coefficient at 490 nm and the third one (K_d(AVE)]⁻¹₁) is just an average of [K_d(PAR)]⁻¹₁ and [K_d(PAR)]⁻¹₂. [K_d(PAR)]⁻¹₂ is larger than [K_d(PAR)]⁻¹₁ almost everywhere in the tropical oceans. Our results show that the OGCM with [K_d(PAR)]⁻¹₂ produces warmer sea surface temperature (SST) in the eastern equatorial Pacific and Atlantic and leads to reduce a cold bias in the equatorial cold tongue regions. It has warmer subsurface temperatures in the low latitude, a slower meridional velocity and Pacific equatorial undercurrent (EUC) than the model with [K_d(PAR)]⁻¹₁. These results are similar to previous studies, although we use a different model and different methods. This study has further analysis and firstly reveals that slower EUC and meridional velocity in the model with [K_d(PAR)]⁻¹₂ are mainly related to the changes of the acceleration due to zonal density gradient. This acceleration driving the EUC eastward in the subsurface becomes smaller in the subsurface along the equatorial Pacific. However, near the sea surface, the zonally averaged accelerations over the different ocean basins are larger in the model with [K_d(PAR)]⁻¹₂ than that with [K_d(PAR)]⁻¹₁, which pushes back the poleward meridional transport. The interannual variability in the model with [K_d(PAR)]⁻¹₂ is generally weaker than that in the experiment with [K_d(PAR)]⁻¹₁ due to a deeper mixed layer depth. The vertical temperature errors averaged horizontally within the domain of 30°S to 30°N in the experiment with [K_d(AVE)]⁻¹₁ are almost in the middle of errors of the other two experiments. This indicates that the effect of the SWAD on the simulation of the vertical temperature profile is largely linear.