Modelling suppressed and active convection. Comparing a numerical weather prediction, cloud-resolving and single-column model

This paper describes the design of and basic results from a case study to compare simulations of convection over the Tropical West Pacific. Simulations are carried out using a cloud-resolving model (CRM), a global numerical weather prediction (NWP) model and a single-column version of the NWP model (SCM). The experimental design for each model type is discussed and then results are compared. The periods simulated each include a regime with strong convective activity, a much more suppressed regime with far less convection, as well as the transition between these regimes. The description of the design and basic results from this study are given in some detail, as a study including all these model types is relatively new. Comparing the local forcing due to the dynamics in the NWP model with the observed forcing used to drive the CRM and SCM it is found that there is good agreement for one period chosen but significant differences for another. This is also seen in fields such as rain rate and top-of-atmosphere radiation. Using the period with good agreement we are able to identify examples of biases in the NWP model that are also reproduced in the SCM. Also discussed are examples of biases in the NWP simulation that are not reproduced in the SCM. It is suggested that understanding which biases in the SCM are consistent with the full NWP model can help focus the use of an SCM in this framework.