NASA - National Aeronautics and Space Administration

The accurate prediction of convective cloud development in advance of thunderstorm formation (so‐called “convective initiation,” CI) is a challenging forecast problem, one in which the processing of 5–15 min interval imagery from geostationary satellites (e.g., Meteosat Second Generation) offers considerable promise. A present drawback to using sequences of visible or infrared (IR) satellite images to monitor growing cumulus clouds is that higher altitude cirrus clouds often obscure the view of the low‐level cumulus in the pre‐convective environment. In particular, cirrus anvils from pre‐existing convection, and cirrus caused by deep layer quasi‐geostrophic ascent, are very common in pre‐CI environments. Cloud derived parameters from GOES are used here to demonstrate how quantities like visible optical depth (τ), emittance, liquid water path, and effective particle size can be used to quantify cumulus cloud growth in advance of CI. Time rates of change of these derived quantities, as well as IR interest fields that describe cumulus cloud development rates beneath cirrus, are analyzed as τ of the cirrus are binned from 1 to >50. Results confirm that if cirrus possess τb 20, up to >90% of the “signal” in the IR interest field remains, compared to clear‐sky values, and it is proposed that CI can still be adequately nowcasted using IR channel data similar to what is done in the absence of cirrus. Similarly, cloud derived parameters become valuable as their time rates of change measure cumulus cloud growth beneath the higher clouds. In contrast, once τ values increase beyond ∼20, cumulus cloud growth signals decrease significantly through cirrus, and as τ becomes >40, little information from the cumulus remains.