Abstract
A 12 h nested-grid numerical simulation of a warm-season mesoscale convective weather system (Zhang and Fritsch, 1986) is utilized as a control run in order to 1) test the sensitivity of the numerical simulation to different types of initial conditions; 2) examine the need for an observing system that would resolve mesoscale features; and 3) determine which meteorological variables need to be most carefully considered in observing system design and preprocessing analysis.
It is found that improved observational capabilities are likely to have an important impact on the successful prediction of the timing and location of summertime mesoscale convective weather systems if mesoscale features can be resolved. In particular, the resolution of the moisture field significantly affects the prediction of the evolution of the convective weather systems. Correspondingly, the mesoscale distribution of precipitation is substantially affected, especially the location of the areas of heavy rain. It is also found that procedures to account for the effects of convective systems that are in progress at the time of initialization can make significant contributions to the prediction of the evolution of the meteorological events and to the improvement of the quantitative precipitation forecasts. In particular, in weak-gradient summertime situations, mesoscale convective systems can severely alter their near environment within a short time period by producing strong mesoscale circulations, thermal boundaries, moist adiabatic stratification etc.
For summertime situations where the large-scale gradients are weak, detailed temperature and moisture fields appear to be more important than the detailed wind fields in determining the development and evolution of deep convection. However, poor resolution of the wind field such that wind speed magnitudes and gradients are underestimated tends to reduce the degree of mesoscale organization. It also alters the magnitude and distribution of low-level convergence, and this affects the evolution of the thermodynamic fields and the deep convection.
Incorporation of dense surface observations into the initial conditions can be very important in improving forecasts of meso-β-scale structures such as moist (dry) tongues, thermal boundaries, and, in particular, pressure distribution. Most significantly, the large (meso-α)-scale environment appears to contain some type of signal such that the general evolution of events is similar, even when the initial mesoscale structure and the simulated meso-β-scale evolution of events are significantly different. On the other hand, poor resolution of meso-α-scale gradients can substantially alter the predicted evolution of meso-β-scale features and the location of heavy rain.
