How sensitive are aerosol‐precipitation interactions to the warm rain representation?

It is widely acknowledged that aerosol‐cloud interactions are a major uncertainty in climate and numerical weather prediction. One of the sources of uncertainty is the sensitivity of the cloud microphysics parameterization to changes in aerosol, in particular the response of precipitation. In this work, we conduct an idealized, dynamically consistent, intercomparison of warm rain microphysics schemes to understand this source of uncertainty. The aims of this investigation are: (i) investigate how sensitive precipitation susceptibility ( S 0 ) is to cloud microphysics representation and (ii) use S 0 to determine the minimum complexity of microphysics required to produce a consistent precipitation response to changes in cloud drop number concentration ( N d ). The main results from this work are: (i) over a large range of liquid water path and N d , all the bulk schemes, but particularly the single moment schemes, artificially produce rain too rapidly. Relative to a reference bin microphysics scheme, this leads to a low in‐cloud S 0 and impacts the evolution of S 0 over time. (ii) Rain evaporation causes surface S 0 from all schemes to be larger than the cloud base S 0 . The magnitude of the change in S 0 with altitude is dependent on the scheme and the representation of the rain drop size distribution. Overall, we show that single‐moment schemes produce the largest range in the sensitivity of precipitation to changes in N d . Modifying rain production parameterization alone does not reduce this spread. Instead, increasing the complexity of the rain representation to double‐moment significantly improves this behavior and the overall consistency between schemes.