Performance evaluation of daily Weather Generator ClimGen model in a mountainous watershed of Himalayan foothills

Long-term series of daily weather data are required for the analysis of weatherimpacted systems such as cropping management systems, hydrologic studies, environmental studies etc. Weather generators are used to produce long series of synthetic daily climatic data using existing weather records where observed climate data are limited. But before evaluating any hydrological, crop growth and other models, it is first required to evaluate the weather generator in-built in the model for its better efficiency. Therefore, in this study, mostly used weather data generator, ClimGen, was evaluated for generating daily information taking daily precipitation (1970–2008), maximum and minimum temperatures of two existing weather stations of a hilly watershed, located in Uttarakhand i.e. Tamadhaun (1989–2008) and Kedar (1989–2003). After doing parameterization, ClimGen was used to generate daily series of rainfall from 2004–2008, maximum and minimum temperature data for Tamadhaun from 2006–2008 and for Kedar from 2001–2003. The generated data then used to evaluate the performance of ClimGen model using minimum root mean square error (RMSE), maximum correlation coefficient (r) and coefficient of efficiency (CE). ClimGen performed well in generating the minimum and maximum temperatures. However, for precipitation, it clearly underestimated the daily rainfall in all two locations. Therefore, based on actual and generated rainfall, system response models may be evaluated very carefully for the hilly watersheds.