Weak temperature gradient simulations; sensitivity to the thermodynamic environment

We study the sensitivity of weak temperature gradient (WTG) simulations to reference profiles derived from the TOGA COARE field campaign. The WTG approximation is an idealized way to parameterize the large scale environment in a limited domain numerical model. Weak temperature gradients are assumed to hold in the tropics because gravity waves quickly dissipate any perturbations that arise in the potential temperature profile. In a model, the WTG approximation is enforced by relaxing perturbations of potential temperature to a reference temperature profile. This produces a WTG vertical velocity which vertically advects moist or dry air in the model. Moisture can be imported, or exported, in the domain in two ways: 1) relaxing the domain moisture profile to the reference environment moisture profile, or 2) lateral entraining moisture from/to the environment by independently satisfying a mass continuity equation for the WTG vertical velocity. We test the sensitivity of the WTG approximation to both moisture transport treatments with respect to TOGA COARE observations.

Further, we test the sensitivity of the WTG approximation to the source of observed reference profiles. A recent study by Wang et al (2013) tested the skill of the WTG approximation using observations from the TOGA COARE field campaign. In the study, the authors used reference temperature and moisture profiles obtained from spatially averaged soundings from multiple stations. In our work the reference profile is obtained from the TOGA COARE field campaign in two ways: 1) average thermodynamic profiles from multiple stations (as in Wang et al, 2013), and 2) soundings from a single station.

We find better agreement of observed and model precipitation when moisture relaxation is used with spatially averaged thermodynamic profiles, while using lateral entrainment of moisture gives better results with single point thermodynamic profiles. This result suggests that depending on the nature of the source of thermodynamic profiles used in the WTG simulations (area averaged versus single point soundings), one choice of moisture treatment in the WTG approximation might be better than the other!