Main content area

Estimating reference evapotranspiration with the FAO Penman-Monteith equation using daily weather forecast messages

Cai, J., Liu, Y., Lei, T., Pereira, L.S.
Agricultural and forest meteorology 2007 v.145 no.1-2 pp. 22-35
evapotranspiration, weather forecasting, meteorological parameters, weather, irrigation management, water management, mathematical models, equations, diurnal variation, solar radiation, air temperature, wind speed, accuracy, decision making, China
Real-time irrigation management and water resources allocation need real-time prediction of daily reference evapotranspiration (ET(o)). Thus, adopting the FAO Penman-Monteith (FAO-PM) equation as the standard for ET(o) estimation, an attempt was made to predict daily ET(o) using the public weather forecast messages available in China. These involve the cloudiness conditions, daily maximum and minimum temperature, and wind speed scales. An analytical method (AM) was developed to translate daily weather forecast messages into the variables needed to estimate ET(o). Daily weather data for the period 1984-1998 at eight meteorological stations representing a wide range of climatic conditions of China were used to compute the FAO-PM ET(o) and to serve as reference data sets for comparison with the variables obtained from daily weather forecast messages at the same locations and period. Several statistical indicators were used for the respective comparisons. The sunshine duration (n) estimated from the forecasted cloudiness agree well with those observed as indicated by the Willmott indices of agreement, d, and the determination coefficient, R2, higher than 0.99 and 0.96, respectively, at all locations. The translation of wind speed scales into wind speed values shows adequate, with a relative error, RE, near 0.10 and high values for d and R2. The comparison of the actual vapour pressure estimated from the minimum temperature and computed from air humidity observations yielded d and R2 values higher than 0.85 for all the locations except one, Ejina, located in an arid region. The estimated weather parameters were then used to compute the daily ET(o) with the FAO-PM equation for the eight locations. The weather forecasted ET(o) estimates agree well with the ET(o) values computed with full data sets, with d and R2 for all locations larger than 0.95 and 0.91, respectively. However, the accuracy of these ET(o) estimations depends upon the availability of accurate weather forecast messages. Results indicate that daily ET(o) predictions using the public weather forecast messages are appropriate to be used for real-time water allocation and irrigation management.