Structural inconsistency in cloud microphysics limits emergent constraints on aerosol-cloud radiative forcing

The magnitude of aerosol–cloud radiative forcing remains one of the dominant uncertainties in climate projections. Emergent constraints are increasingly used to reduce this uncertainty by linking observable cloud properties to modelled aerosol–cloud interactions. Their physical validity, however, depends critically on whether models reproduce the same cloud–aerosol coupling mechanisms as the real atmosphere. Here we show that current Earth system models exhibit a systematic structural inconsistency in cloud microphysics that undermines the physical interpretability of emergent constraints on aerosol–cloud radiative forcing (ΔF_aci).

Using perturbed-parameter ensembles (PPEs) of UKESM1, we analyse observed and modelled relationships between cloud droplet number concentration (N_d), liquid water path (LWP), and aerosol perturbations across key marine stratocumulus regimes. We show that the N_d–LWP sensitivity, which controls how strongly clouds brighten in response to aerosol, varies by a factor of ∼4 across model parameterisations but is tightly constrained by observations. As a result, models that reproduce present-day mean cloud properties can require physically implausible N_d–LWP responses to generate their aerosol forcing, leading to equally plausible yet physically incompatible ΔF_aci estimates.

This structural degeneracy implies that conventional emergent constraints targeting mean cloud states cannot uniquely constrain aerosol forcing. Instead, physically meaningful constraints must explicitly account for the microphysical response pathways linking aerosols, cloud water, and radiation. Our results reveal a previously under-recognised structural source of uncertainty in aerosol–cloud interactions and provide a new physically grounded diagnostic for evaluating and constraining modelled aerosol–cloud radiative forcing.