MeteoCogent is a configurable weather forecast generation tool, which transforms weather data into English textual forecasts. Input to MeteoCogent can be a time series of forecast data from Model Output Statistics (MOS), or from some more interpreted form using "explicit weather elements". The output can be a plain language or coded forecast, according to a standard specified by the user. The content and organization of reports can also be extensively customized, allowing a wide variety of texts to be produced.


  • Automatic generation frees forecasters from the mechanical aspects of producing weather reports, allowing them to concentrate on the aspects of forecasting which most require human knowledge and intuition.
  • Implemented in C++/Java for portability (C++ portions currently being ported to Java).
  • Can be deployed in various configurations (client/server, web server, DLL).


  • Kittredge, Richard; and Lavoie, Benoit (1998). MeteoCogent: A Knowledge-Based Tool For Generating Weather Forecast Texts. In Proceedings of the American Meteorological Society AI Conference (AMS-98), Phoenix, Arizona. [Acrobat, 86 Kb]