The Afro-Asian summer monsoon (AfroASM), encompassing West African, South Asian, and East Asian monsoon regions, is vital for billions of people. Accurate projections of future AfroASM precipitation are crucial for climate change adaptation and mitigation. However, current state-of-the-art climate models exhibit considerable discrepancies in their projections, hindering reliable assessments. This uncertainty stems from various sources including socioeconomic scenarios, internal variability, and model structures. Model uncertainty alone can account for over 70% of the total uncertainty in AfroASM precipitation projections. To address this challenge, emergent constraint techniques are employed. These techniques leverage the relationship between observable present-day climate variables and projected future changes to refine model predictions. Previous studies have identified a link between uncertainty in AfroASM precipitation changes and the increase in interhemispheric thermal contrast (ITC) in projections. This study investigates this relationship further, aiming to constrain the model spread and improve the reliability of AfroASM precipitation projections by focusing on the large-scale interhemispheric or land-sea thermal contrast, which drives the large-scale monsoon circulation, particularly in West Africa and Asia where precipitation exhibits in-phase changes modulated by ITC and North Atlantic SST variations.

Numerous studies using Coupled Model Intercomparison Project (CMIP) models have projected increases in AfroASM precipitation under medium and high emission scenarios, but with large inter-model spreads. The thermodynamic processes related to increased atmospheric moisture consistently enhance precipitation across models, while dynamic processes related to circulation changes contribute substantially to uncertainty. Total uncertainty in future projections arises from socioeconomic scenarios, internal variability, and model structures. Model uncertainty dominates the total uncertainties of AfroASM precipitation changes. Emergent constraint techniques, using physical links between present-day and future climate variables, have been successfully used to reduce projection uncertainty in other monsoon regions. The connection between uncertainty in AfroASM precipitation and increasing interhemispheric thermal contrast (ITC) has been noted previously, but methods to constrain this spread have remained elusive. The current study aims to address this gap by using CMIP6 model outputs.

This study utilizes monthly output from 30 CMIP6 models under the high-emission scenario (SSP5-8.5) and the median emission scenario (SSP2-4.5) for validation. The analysis focuses on June-September (JJAS) precipitation changes from 2050-2099 relative to 1965-2014. The Afro-Asian monsoon region is defined based on precipitation differences between summer and winter and the proportion of annual precipitation in summer. An inter-model empirical orthogonal function (EOF) analysis was performed on the projected precipitation changes to identify dominant patterns of uncertainty. The relationship between the leading principal component (PC1) and present-day and future interhemispheric thermal contrast (ITC) is explored. The ITC is calculated as the difference in area-averaged surface temperatures between 20°N-50°N and 20°S-50°S. An ITC pattern index is defined by projecting present-day surface temperature trends onto the warming pattern associated with PC1. Multiple observational datasets for surface temperature and precipitation are used to constrain the model biases. A hierarchical statistical framework, accounting for uncertainties in both models and observations, is employed to constrain the PC1 using a linear regression between present-day ITC and projected PC1. The constrained PC1 is then used to reconstruct the precipitation and circulation patterns. The impact of the emergent constraint on potential water availability (estimated using runoff) and the areal extent of significant precipitation and runoff increases is assessed.

CMIP6 models generally project an increase in AfroASM precipitation under SSP5-8.5, except for parts of West Africa. The projected regional average increase is 14%, with a large inter-model spread (1%-27%). EOF analysis reveals that the leading mode of uncertainty shows a spatially consistent increase in precipitation across the AfroASM domain, suggesting a large-scale controlling factor. A strong correlation exists between the present-day ITC trend and the future AfroASM precipitation increase. Models with a larger ITC trend project a greater precipitation increase. This relationship is linked to the equilibrium climate sensitivity (ECS); models with higher ECS show larger ITC increases in both historical and future periods. The emergent constraint, based on the relationship between present-day ITC and the projected precipitation, reduces the projected precipitation increase to approximately 10% (0.57 ± 0.38 mm day⁻¹), which is about 70% of the raw projection. The largest reduction is observed in West Africa (49%). The land area experiencing significant precipitation increases is 57% of the raw projection, while the land area with significant runoff increases is 66%. The constrained projection suggests smaller increases in precipitation than the raw projection, which may mitigate flood risk but pose challenges to water resource management.

The findings demonstrate that the raw CMIP6 projections of AfroASM precipitation may overestimate the increase. The emergent constraint, using present-day observed ITC trends, provides a more realistic projection. The reduced precipitation increase, particularly in West Africa, has implications for flood risk and water resource management. While the reduction in flood risk is beneficial, the lower-than-expected increase in water resources poses challenges for water management and food security. The robustness of the emergent constraint was verified using the SSP2-4.5 scenario and different model subsets, showing consistency in the results. Further analysis constraining hydrological sensitivity and global surface air temperature (GSAT) separately supports the main findings, with GSAT warming playing a dominant role in the constraint.

This study highlights the importance of considering present-day model biases when projecting future AfroASM precipitation. The emergent constraint, utilizing the relationship between present-day ITC trends and future precipitation projections, substantially reduces the uncertainty and provides a more reliable estimate of future changes. The reduced projected increases in precipitation, particularly in West Africa, indicate a lower risk of flooding but also highlight potential challenges for water resource management and food security. Future research should focus on improving model representation of ITC and exploring other emergent constraints to further refine regional climate projections.

The study primarily focuses on the SSP5-8.5 scenario, although the robustness is checked using SSP2-4.5. While the hierarchical statistical framework accounts for uncertainties in both models and observations, other sources of uncertainty, such as internal climate variability, might still influence the results. The spatial resolution of the data could also influence the findings. Future research incorporating higher-resolution data and exploring additional emergent constraints could further improve the accuracy of AfroASM precipitation projections. The study focuses on a specific set of models which might not be fully representative of all existing models.