Recognition and detection issues will be treated elsewhere in this volume, by Burgess et al. (1993).
 It is noteworthy that the composite chart, as employed in severe weather forecasting, specifically attempts to establish the interaction between features aloft and at the surface. Thus, it is a product with a long history of addressing what Mass (1991) considers a common deficiency in synoptic analysis; namely, the failure to depict 3-dimensional relationships among features at different levels.
 Some of the research results have had an impact in some operational detection and warning programs, although even there, progress has been slow.
 For official purposes, a severe thunderstorm is defined as one which produces one or more of the following: hail > 3/4 in (2 cm) in diameter, measured winds > 50 kt (25 m s-1), "damaging" winds (involving some subjective judgment of effects required to meet the threshold), a tornado. Heavy rain, large quantities of sub-threshold hail, funnel clouds, frequent lightning, etc. are not considered to meet the official criteria (see discussion in Doswell 1985).
 As noted in Doswell and Burgess (1993), some atmospheric vortices are not associated with deep, moist convection. These are not considered to be tornadoes.
 Note that a situation we label as synoptically "evident" should not be automatically equated with an "easy" forecast. No real-world forecast situation is ever easy, except in retrospect!
 Since the grid is defined on a polar stereographic map projection, the grid boxes vary in size across the map by as much as about 10%. The nominal size applies only at 60 deg N latitude, where the map scale factor is unity.
 As is the case in accounting for currency inflation, this does not imply that there is anything special about 1955. It simply represents a base, or reference state. We could just as easily have adjusted toward a 1989 standard, with no material difference in our conclusions.