SOME CONSIDERATIONS INVOLVED IN PREPARING LONG-RANGE FORECASTS BY NUMERICAL METHODS

Abstract

Basic problems encountered in the preparation of medium- and long-range forecasts by numerical methods are revealed through the use of time-averaged charts. It is found that forecasts up to five days in advance may be prepared by use of the vorticity-tendency equation, provided certain broad-scale sources and sinks of vorticity are taken into account. Since these appear in the final prognosis as locally time-averaged values, it is probable that they may be sufficiently well represented by crude approximations. It is not possible to forecast for a month or more in advance by integrating the tendency equation, because of approximate balance between the terms not containing time derivatives. A modified conservation equation is suggested as the solution to this problem. Finally, it is suggested that the eddy transfer of meteorological properties will not be a major problem in long-range forecasting, since this transfer can probably be expressed in terms of the large-scale circulation.