Improving climate model simulation of tropical Atlantic sea surface temperature: The importance of enhanced vertical atmosphere model resolution

Abstract

A long‐standing problem in climate modeling is the inaccurate simulation of tropical Atlantic (TA) sea surface temperature (SST), known as the TA SST bias. It has far‐reaching consequences for climate prediction in that area as it goes along, among others, with erroneous precipitation patterns. We show that the TA SST bias can be largely reduced by increasing both the atmospheric horizontal and vertical resolutions in a climate model. At high horizontal resolution, enhanced vertical resolution is indispensable to substantially improve the simulation of TA SST by enhancing surface wind stress. This also reduces biases in the upper ocean thermal structure and precipitation patterns. Although, enhanced horizontal resolution alone leads to some improvement in the mean climate, typical bias patterns characterized by a reversed zonal SST gradient at the equator and too warm SST in the Benguela upwelling region are mostly unchanged at a coarser vertical resolution.

On the relationship between Atlantic Niño variability and ocean dynamics

“…Surprisingly, this behaviour is hardly altered qualitatively in the FLX experiment, indicating that the zonal wind bias depends only weakly on eastern basin SST in the model. In agreement with the recent findings of Richter et al (2014b) and Harlaß et al (2015) we suspect that the zonal wind bias is at least partly due to the insufficient vertical resolution of the atmosphere model and related deficiencies in vertical momentum transport.…”

Causes of the large warm bias in the Angola–Benguela Frontal Zone in the Norwegian Earth System Model

“…Richter et al (2010) also suggest the local impact of surface wind on Benguela Niño events (warm SST anomaly events with inter-annual time scale, e.g., Shannon et al 1986). Other numerical experiments with CGCMs find a reduction of the SST bias in the ABFZ with finer vertical and horizontal resolutions of the atmospheric model component (Patricola et al 2012;Small et al 2014;Harlaß et al 2015).…”

“…The COREv2-IAF data have a resolution of about 2°. Although a finer horizontal and vertical atmospheric model resolution might reduce the ABFZ bias (Harlaß et al 2015), even at a quarter degree resolution MICOM forced with COREv2-IAF exhibits a similar warm SST bias (in Fig. S1).…”

Recent Progress in Understanding and Predicting Atlantic Decadal Climate Variability

“…Model bias is a complex problem that can involve coupled processes that are notoriously difficult to unravel. While many coupled model biases in the Atlantic are probably related to poor Gulf Stream representation in coarse resolution ocean models, others, such as poor upper ocean thermal structure in the tropical Atlantic, appear to originate in the atmosphere [174].…”