Atmospheric Single Column Models (SCMs) provide an economical modeling framework for diagnostic studies. In these models, vertical profiles of temperature and humidity evolve in response to diabatic interactions within the column and adiabatic tendencies produced by the large scale circulation. Often, physical parameterizations in the SCM are obtained directly from a general circulation model (GCM). This makes the SCM a valued resource for the development of comprehensive numerical climate forecast models. Typically in SCMs, the adiabatic tendencies are prescribed and, thus, decoupled from the diabatic tendencies. This decoupling can lead to the rapid development of atmospheric states that are not found in reality or in the corresponding GCM. This makes both the diagnosis of SCMs and its implications for GCM development difficult, it not impossible, to interpret.

We have modified the SCM framework to include coupling between the adiabatic and diabatic tendencies. In this 'coupled' SCM, vertical temperature advection is parameterized in terms of the time-history of diabatic heating rates. The remaining adiabatic tendencies are then calculated under the additional assumption that the column is embedded in an uniform environment. This coupled framework stabilizes the SCM and allows the SCM to maintain a realistic climate, but damps out high frequency variability in long runs.

To construct the 'stochastic climate simulator' (SCS), high frequency variability is introduced to the coupled SCM by adding multivariate red noise to the SCM tendency equation. The linear operator and white noise covariance for the stochastic forcing are calculated from error statistics gathered from short runs of the coupled SCM. This process exploits the stability of the coupled SCM and its near linearity for short integrations. When tested for tropical conditions, the resulting SCS produces realistic fluctuations of temperature and humidity compared to observations from TOGA COARE and maintains a stable climate in year-long runs. The SCS provides a economical diagnostic framework for extensive sensitivity testing. Furthermore, since we can construct SCSs from both observational data and GCM data, the SCS also provides an economical testing framework for GCM development.