Evaluation Ofcligenandwxgenweather Data Generators Under Canadian Conditions

In this study, we evaluate for five selected Canadian climatological stations, the weather data generator, CLIGEN, which was developed for water erosion prediction (WEPP) as well as the similar generator, WXGEN, developed for the erosion/productivity impact calculator (EPIC). An important feature of WXGEN and CLIGEN is the small amount of input data required to generate long periods of record. The input parameters include: monthly relative frequency of wet- and dry-day sequences, monthly mean, standard deviation and skewness of precipitation, monthly mean and standard deviation of daily maximum and minimum air temperature, and solar radiation. CLIGEN-generated weather data were found to correspond well with observed mean values, but had consistently lower variation than observed values. The correlation between wet-days and temperature and radiation was not modelled accurately. In contrast, WXGEN was found to more accurately simulate the observed variability and preserve the relationship between temperature, precipitation and solar radiation. This was expected because of the way the model is designed. Agrometeorological parameters such as frost-free period were estimated from the observed data and the generated data. Both CLIGEN and WXGEN weather data were found to underestimate the normal frost-free period by as much as 20 days which could cause problems if the data were used in WEPP or EPIC crop-yield models.