Abstract:Coupled atmosphere–ocean general circulation models (AOGCMs) commonly fail to simulate the eastern equatorial Atlantic boreal summer cold tongue and produce a westerly equatorial trade wind bias. This tropical Atlantic bias problem is investigated with a high-resolution (27-km atmosphere represented by the Weather Research and Forecasting Model, 9-km ocean represented by the Regional Ocean Modeling System) coupled regional climate model. Uncoupled atmospheric simulations test climate sensitivity to cumulus, land-surface, planetary boundary layer, microphysics, and radiation parameterizations and reveal that the radiation scheme has a pronounced impact in the tropical Atlantic. The CAM radiation simulates a dry precipitation (up to −90%) and cold land-surface temperature (up to −8 K) bias over the Amazon related to an over-representation of low-level clouds and almost basin-wide westerly trade wind bias. The Rapid Radiative Transfer Model and Goddard radiation simulates doubled Amazon and Congo Basin precipitation rates and a weak eastern Atlantic trade wind bias. Season-long high-resolution coupled regional model experiments indicate that the initiation of the warm eastern equatorial Atlantic sea surface temperature (SST) bias is more sensitive to the local rather than basin-wide trade wind bias and to a wet Congo Basin instead of dry Amazon—which differs from AOGCM simulations. Comparisons between coupled and uncoupled simulations suggest a regional Bjerknes feedback confined to the eastern equatorial Atlantic amplifies the initial SST, wind, and deepened thermocline bias, while barrier layer feedbacks are relatively unimportant. The SST bias in some CRCM simulations resembles the typical AOGCM bias indicating that increasing resolution is unlikely a simple solution to this problem.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to investigate the tropical Atlantic bias issue in high-resolution coupled regional climate models (CRCM). Specifically, the paper focuses on the following points: 1. **Atmospheric Model Bias**: Exploring the impact of different physical parameterization schemes on the atmospheric model's simulation bias in the tropical Atlantic region through parameterization scheme testing. 2. **Air-Sea Coupling Bias**: Based on the sensitivity analysis results of the atmospheric model, designing high-resolution coupled regional climate model experiments to test whether the spring westerly equatorial trade wind bias leads to the development of summer eastern equatorial Atlantic sea surface temperature (SST) warm bias through unrealistic thermocline deepening, and exploring the connection between this trade wind bias and precipitation biases in the Amazon (dry) and Congo Basin (wet) regions. 3. **Feedback Mechanisms**: Identifying the role of atmosphere-ocean feedback mechanisms in amplifying biases, particularly the Bjerknes feedback and barrier layer feedback. 4. **Sensitivity to Model Initialization Time**: Understanding the sensitivity of biases to the model initialization time. ### Background Coupled atmosphere-ocean general circulation models (AOGCMs) have long exhibited biases in the tropical Atlantic region. Observational data show that the sea surface temperature (SST) in the eastern equatorial Atlantic drops by about 5°C from May to August, forming the Atlantic cold tongue, but AOGCMs without flux correction fail to capture this phenomenon, instead simulating an increasing zonal SST gradient towards the east. The formation of the Atlantic cold tongue is related to enhanced vertical mixing and upwelling, processes influenced by the West African monsoon. However, coupled models fail to realistically represent these processes, increasing the uncertainty in future climate change simulations. ### Main Findings 1. **Atmospheric Model Bias**: - Using different radiation schemes significantly affects the simulation results in the tropical Atlantic. The CAM radiation scheme simulates dry (up to -90%) and cold surface temperature (up to -8 K) biases in the Amazon region, while the RRTM/Goddard radiation scheme simulates higher precipitation (up to 130% in the Amazon and Congo Basin) and weaker eastern Atlantic trade wind biases. - Biases in low-level cloud cover are the main cause of radiation biases, particularly in the Amazon region, where the CAM radiation scheme simulates excessive low-level cloud cover, leading to reduced surface shortwave radiation (up to -100 W/m²). 2. **Air-Sea Coupling Bias**: - Long-term high-resolution coupled regional climate model experiments show that the initial development of the eastern equatorial Atlantic SST warm bias is more sensitive to local rather than basin-wide trade wind biases and is related to the wet Congo Basin rather than the dry Amazon region. - Comparing coupled and uncoupled simulation results reveals that the Bjerknes feedback in the eastern equatorial Atlantic amplifies the initial SST, wind speed, and thermocline depth biases, while the barrier layer feedback is relatively unimportant. 3. **Feedback Mechanisms**: - The Bjerknes feedback mechanism plays a crucial role in amplifying SST, wind speed, and thermocline depth biases. - The barrier layer feedback may be a secondary error source in some models, but no significant barrier layer bias was found in other models. 4. **Sensitivity to Model Initialization Time**: - Improving wind stress in spring (MAM) can significantly reduce the SST warm bias in the eastern equatorial Atlantic and maintain it through summer (JJA) via the Bjerknes feedback mechanism. - Applying observed wind stress across the entire basin (30°S–30°N) yields the best improvement in SST bias. ### Conclusion This study, through high-resolution coupled regional climate model experiments, thoroughly investigates the causes and feedback mechanisms of tropical Atlantic biases. The results show that the choice of radiation scheme significantly affects simulation results, with the CAM radiation scheme performing poorly in tropical regions, while the RRTM/Goddard radiation scheme, despite simulating higher precipitation, better reproduces the climate characteristics of the tropical Atlantic. Additionally, the Bjerknes feedback mechanism plays a key role in amplifying biases, while the barrier layer feedback is relatively secondary. The choice of model initialization time also has a significant impact on biases.

An investigation of tropical Atlantic bias in a high-resolution coupled regional climate model

Structure of the Atlantic Meridional Overturning Circulation in Three Generations of Climate Models

Oceanic origin of southeast tropical Atlantic biases

A Study of Biases in Simulation of the Indian Ocean Basin Mode and Its Capacitor Effect in CMIP3/CMIP5 Models

Effects of Climatological Model Biases on the Projection of Tropical Climate Change

Origin of Warm SST Bias over the Atlantic Cold Tongue in the Coupled Climate Model FGOALS-g2

Alleviating tropical Atlantic sector biases in the Kiel climate model by enhancing horizontal and vertical atmosphere model resolution: climatology and interannual variability

Warm Bias of Sea Surface Temperature in Eastern Boundary Current Regions—a Study of Effects of Horizontal Resolution in CESM

An overview of the performance of CMIP6 models in the tropical Atlantic: mean state, variability, and remote impacts

Dynamical Downscaling of Climate Simulations in the Tropics

Convection Parameterization, Tropical Pacific Double ITCZ, and Upper-Ocean Biases in the NCAR CCSM3. Part I: Climatology and Atmospheric Feedback

Mechanisms of model bias impacting responses of the Atlantic cold tongue to greenhouse warming

Inter-model Spread of the Climatological Annual Mean Hadley Circulation and Its Relationship with the Double ITCZ Bias in CMIP5

Summertime Rossby waves in climate models: substantial biases in surface imprint associated with small biases in upper-level circulation

Tropical Cyclogenesis Bias over the Central North Pacific in CMIP6 Simulations

Extra-tropical origin of equatorial Pacific cold bias in climate models with links to cloud albedo

Impacts of North Atlantic Model Biases on Natural Decadal Climate Variability

Ocean Dynamics Causes the Equatorial Pacific SST Bias in CAS‐ESM2‐0

The Double ITCZ Syndrome in GCMs: A Coupled Feedback Problem among Convection, Clouds, Atmospheric and Ocean Circulations

Intermodel Biases of the Western North Pacific Monsoon Trough in CMIP6 Models

Evaluation of the Causes of Wet‐Season Dry Biases over Amazonia in CAM5