Abstract

Environmental simulation models usually require weather data as boundary conditions. Observed time series of weather data sometimes are absent or too short for long term model runs. In cases of predictions, future weather data are needed. For both purposes an automated procedure for generation of daily weather data, with realistic statistical properties, is useful. In this report a description is given of a method to generate time series of daily weather data: temperature, evaporative demand and precipitation. Statistical characteristics for these weather variables are derived from measured data series in the Netherlands. Using these statistical properties, artificial data series can be generated by the computer program METEOGEN. In order to generate weather data for a specific location in the Netherlands, annual means of temperature, evaporation demand and precipitation are required. The generated time series are tested against observed data, and prove to be compatible in most respects. Where discrepancies are detected, recommendations are made to improve METEOGEN. User guidelines and a program description are given of METEOGEN version 1.0. METEOGEN is written in portable code: ANSI standard FORTRAN77, and can be used on a variety of systems.