Forecasting and future work - Severe Convective Storms and Tornadoes: Observations and Dynamics - page 418

Figure 7.3. Box-and-whisker overlay plots of effective storm-relative helicity (ESRH) when
CAPE 100 j kg 1 and CIN 250 j kg 1 (solid gray box and thick gray whiskers) and 0-
1 km SRH (dotted black box and think black whiskers) for discrete convective storms that are
significantly tornadic (sigtor), weaker tornadic (weaktor), non-tornadic (nontor), elevated non-
tornadic (elevnt), marginal supercells (mrgl), and non-supercells (nonsup). Sounding sample
sizes are shown in parentheses (from Thompson et al., 2007).
casters. Pattern recognition makes use of often-observed parameters associated
with severe convection, even when their physical relevance is not completely
understood. Climatology is used to alert forecasters for spatial and temporal
biases toward certain ingredients.
7.1.2 Model-based forecasting
The model-based forecasting methodology is in sharp contrast to the ingredients-
based methodology. Through this method, cumulus convection is explicitly
represented in non-hydrostatic cloud models run over regional domains and poss-
ibly over even larger domains. It is beyond the scope of this text to detail the
numerical schemes, initialization procedures, etc. of all, or even just a few, of the
models. For the latter, important issues include deciding how to assimilate
Doppler radar wind data as well as wind, moisture, pressure, and temperature
data from observational networks. The model needs time to ''adjust'' to the per-
turbations inflicted by incorporating new data.