Many of the most damaging and life threatening types of weather, such as torrential rains and severe thunderstorms, begin quickly, strike suddenly, and disappear rapidly, destroying small regions while leaving neighboring areas untouched. Conventional computer models of the atmosphere have limited value in predicting short-lived local storms because the available weather data are generally not detailed enough to allow computers to see clearly the small atmospheric changes that come before these storms.In most nations, for example, weather-balloon observations are taken just once ever3r twelve hours at location typically separated by hundreds of miles. With such limited data, conventional forecasting models do a much better job predicting general weather conditions over large regions than they do forecasting specific local events.
        Until recently, the observation-intensive approach needed for accurate, very short range forecasts, or "nowcasts", was impracticable. The cost of equipping and operating many thousands of conventional weather stations was extremely high, and the difficulties involved in rapidly collecting and processing the raw weather data from such a network were beyond overcoming.Fortunately, scientific and technological advance have overcome most of these problems. Radar systems, automated weather instruments, and satellites are all capable of making detailed, nearly continuous observations over large regions at a relatively low cost. Communication satellites can transmit data around the world cheaply and immediately, and modern computers can quickly compile and analyze this large volume of weather information. Meteorologists and computer scientists now work together to design computer programs and video equipment capable of transforming raw weather data into words, symbols, and vivid graphic displays that forecasters can interpret easily and quickly. As meteorologists have begun using these new technologies in weather forecasting offices, nowcasting is becoming a reality.

很多种最具破坏性和威胁生命性的天气，诸如倾盆大雨和猛烈的雷暴，它们开始得迅速，袭击得突然，消失得也很快，破坏小范围地区而相邻的地区却丝毫不动。在短期的地区性暴风雨预报中，大气的普通计算机模型价值很有限，因为可以得到的天气数据通常没有详细到能够让计算机看清楚暴风雨来临前大气的微小变化。例如，在大多数国家里，通常相隔几百米远的地方，每12小时进行一次天气气球观测。由于数据有限，相对于预测特定地区的天气状况，普通预测模型在预测大范围地区的总体天气状况上做得更好。

直到最近，由于密集观测方法需要较高的精确性，所以短期预测，或者“短期天气预报”是没有可行性的。设备和上千个普通天气站运作的成本是非常高的，而且从这样一个网络中快速收集和处理原始数据的困难是难以克服的。幸运的是，科学和技术进步已经克服了大部分问题。雷达系统、自动天气工具和卫星都能够在大多数地区以相对较低的成本进行详细的、近乎连续的观测。通讯卫星能够在世界范围内便宜、及时地传送数据，现代计算机能够很快地编译和分析这个大容量的天气信息。气象学家和计算机科学家现在一起工作来设计一些计算机程序和视频设备，它们能够把原始天气数据转换成文字、符号和生动的图像展示，预测人员能够很容易、很快地解释这些转换后的信息。随着气象工作者开始使用这些天气预报的新技术，短期天气预报也正在变成现实。