On the Practical Use of Graphical Prediction METHODS1

Useful forecasts may be obtained by graphical integrations of the dynamical prediction equations for a barotropic and a two-level baroclinic atmosphere. Such forecasts may be prepared without the aid of special equipment and are therefore particularly valuable as a means of training forecasters in physical prognosis. The present paper reviews the physical principles, modeling assumptions, and methods of solution used in graphical prediction and introduces a method of obtaining surface forecasts which is considerably faster and simpler than previous methods. The predicted surface pressure is shown to be the sum of two components: (1) the pressure advected to the spot by one-half the 500-mb. wind and (2) a pressure change reflected down from aloft (actually one-half the 500-mb. height change expressed in equivalent pressure units at 1000 mb.). The movement of surface pressure systems is thus seen to be largely dependent on upper-level steering, while the deepening is found to be related to the vorticity advection at high levels, since t.his mainly determines the 500-mb. height changes. Twenty sample surface forecasts prepared by the graphical method during July 1959 are presented and compared with the forecasts for the same dates issued by the National Weather Analysis Center. Little difference in accuracy is apparent. Typical shortcomings and failures of the graphical prognoses are discussed. It is believed that the most serious errors are due to the use of only the init’ial500-mbxharts in advect.ing the pressure systems. If the 500-mb. forecasts had been available earlier, it appears that a significant increase in accuracy could have been achieved by using both initial and forecast 500-mb. contours in performing the advections.