Analysis and Data Problems in Relation to Numerical Prediction

Abstract

Instantaneous height tendencies computed from different analyses of the same 500-mb chart illustrate that discrepancies in analysis or observations can have a serious effect on the accuracy of numerical forecasts. The present status of the Northern-Hemisphere aerological network is reviewed, stressing the inhomogeneity of the radiosonde network, the need for observations in fringe areas surrounding the present dense networks, and the desirability of more radio-wind stations. The suggestion is made that machine prognoses incorporating variations of fictitious observations could be used to evaluate the improvements needed and the desirability of setting up new aerological stations in particular locations.