Improving initial conditions for convective weather forecasts

Thunderstorms, squall lines and other severe convective weather develop on time scales of less than an hour and space scales of less than 10 km, scales which are not resolved by most models used in numerical weather prediction. In addition to higher-resolution models, extending numerical weather prediction to severe convective weather requires accurate estimates of the state of the atmosphere on those scales, to serve as initial conditions for forecasts a few hours ahead.

MMM researchers are investigating the use of several novel types of atmospheric observations to enhance the quality of initial conditions for convective-weather forecasts. These observations come from sensors on the ground and aboard geostationary satellites and, owing to recent advances in supporting technologies, now provide sufficient spatial and temporal resolution for convective weather. Observations from space of outgoing infrared radiation offer information on atmospheric temperature and are also very sensitive to cloud. Observations of lightning, which come from both ground- and space-based sensors, help identify where upward motion is strongest in convective storms. Important research issues include quantifying the bias in these observations and developing estimation algorithms that can handle the strongly nonlinear relations between these observables and desired atmospheric quantities, such as temperature and winds. MMM is performing this research under the support of NASA, the Institute of Urban Meteorology of Chinese Meteorological Administration, and the Short-Term Explicit Prediction Program at NCAR.

Impact of assimilating lightning event density from geostationary satellite (upper right) for a squall Line case (upper left). The analysis fields of updraft/downdraft and horizontal  winds are shown for an experiment assimilating radar alone (lower left) and radar  plus lightning data (lower right).
Figure: Impact of assimilating lightning event density from geostationary satellite (upper right) for a squall Line case (upper left). The analysis fields of updraft/downdraft and horizontal winds are shown for an experiment assimilating radar alone (lower left) and radar plus lightning data (lower right).