On the Wave Spectrum Generated by Latent Heating

Waves generated by convective heating in the troposphere propagate upward through the equatorial stratosphere. These waves are responsible for driving the quasibiennial (QBO) and semiannual (SAO) oscillations. The equatorial wave spectrum is very rich, consisting of Rossby, Kelvin, and gravity waves over a large range of wavelengths and frequencies. It remains an open question as to what range of scales is required to provide the necessary wave driving for the QBO and SAO. In order to examine this question, we use observations of convective activity to drive a realistic simulation of the equatorial wave spectrum. High resolution data of rainfall rate and infrared cloud top brightness temperature derived from TRMM (Tropical Rainfall Measurement Mission) and geostationary satellite observations have been used to construct estimates of latent heating and cloud top height every 3 hours on a .25 x .25 longitude-latitude grid. Heating profiles are constructed using a model of convective heating. (The effect of stratiform heating is considered in another paper submitted to this session). This heating is then used to force a primitive equation model, and the wave spectrum is derived from the response. In this study we examine how the wave spectrum depends on model resolution. We find that the waves so generated in a model with T40 horizontal resolution with vertical grid spacing at .75 km is insufficient to drive a QBO. However, a QBO-like oscillation can be produced by artificially enhancing the eastward wave amplitudes by a factor of 2 and the westward wave amplitudes by a factor of 3. Higher resolution simulations (up to T120) do not contribute sufficient additional momentum flux. One possibility is that the 'missing flux' arises from gravity waves at horizontal scales much smaller than we can simulate or at higher frequencies than the TRMM sampling rate. Another possibility is that the wave spectrum is significantly strengthened by the inclusion of stratiform heating.