Effective reasoning, analysis and communication regarding natural phenomena require the use of models to render tractable the complexities of nature. This paper attempts to put into perspective the proper roles of different types of models to maximize the effectiveness of their utilization. The advances in short term forecasting envisioned for the 1970's from full implementation of new knowledge, models and technology will materialize only if the managers and researchers join in an interagency effort to provide the operational meteorologists with the education, techniques, tools and, particularly, the challenging working environment needed to fully develop man's role in forecasting. A program to meet these requirements is outlined.
The types of models discussed include: descriptive or synoptic, dynamic or analytic, numerical or physical, statistical or optimized. The uses of models discussed include: education (basic concepts), research (experimental), operations (customized). Since the operational meteorologist is responsible for the intelligent use of these types of models, he must continually update his training and properly understand the potential contributions of the models.
It is anticipated that during the 1970's routine computer models will become more refined and specialized data such as trajectories and probabilities will become more common. Highly specialized products will be available from special purpose models on a special request basis as field forecasters gain access to remote terminals. Also, forecasters will have access to specialized consultants when unusual events or unusual forecast requirements arise. Background materials will be provided to the applied meteorologists so that he may gain physical understanding from educational and research models including systematic numerical experiments. Communication advances will provide for dynamic (motion picture) displays of radar, synchronous satellite, weather map and weather forecast data.
Only if the operational forecasters do receive the necessary management and scientific support, will their jobs be challenging and attractive to highly motivated and qualified students; only then will the customers of specialized short term forecasts receive the benefits made feasible by science and technology.
A Novel LSTM Model with Interaction Dual Attention for Radar Echo Extrapolation