The effects of use of average instead of daily weather data in crop growth simulation models

Abstract

Development and use of crop growth simulation models has increased in the last decades. Most crop growth models require daily weather data as input values. These data are not easy to obtain and therefore in many studies daily data are generated, or average values are used as input data for these models. In crop growth models non-linear relations often occur. Thus the simulation result with average data can be different from the average result with daily data. In this study the effects of using average weather data on simulated potential and water-limited yields were investigated with a spring wheat crop growth model. It was expected that deviation in simulation results was related to the variability of the weather. Therefore effects were studied for sites in three different climates: temperate maritime, mediterranean and humid tropical. Variability of the weather during the growing season on these sites was quantified. Intuitively the weather in the mediterranean and humid tropical climates is far more constant than the weather in the temperate maritime climates. However, for all locations the variability of the weather during the growing season was nearly the same. The explanation for this unexpected result was found in the fact that on all sites crops were grown in that part of the year in which it rains. The existence of dry and wet days during the growing season causes a large day-to-day variation in weather. For all sites an overestimation of the simulated potential yield of 5–15% was found as a result of using average weather data. For water-limited production the use of average data resulted in overestimation of yield in the wet conditions and underestimation of yield in dry conditions (up to 50%).