Evaluation of RegCM4 climate model for assessment of climate change impact on crop production

For evaluating the impacts of climate change on crop yields regional climate models (RCMs) are now considered better than general circulation models (GCMs). In order to assess what extent the climate output of RCM-RegCM4.0 is biased, this is analysed by comparing the base line simulated daily weather with the observed weather for the corresponding year (1971-2000) over Varanasi. The result shows that the RCM output is biased towards lower annual maximum and minimum temperature by 5.4 degrees C and 1.7 degrees C respectively. Seasonal analysis shows that the RCM output is underestimating the kharif (Rice) season maximum and minimum temperature by 3.0 degrees C and 1.5 degrees C respectively and the rabi (wheat) season maximum and minimum temperature by 6.7 degrees C and 1.4 degrees C respectively. The RCM output overestimates the annual and rabi rainfall while it underestimates kharif rainfall. It is also overestimating the annual, kharif and rabi season rainy days. Most importantly, model underestimates the extreme events, i.e., extreme temperature and heavy rainfall. The study also includes assessment of biasness in yields of wheat and rice simulated using CERES-wheat and CERES-rice crop models employing observed and RCM simulated weather data. Due to biasness in the extreme events in RCM baseline data the simulated wheat and rice grain yield during several years were overestimated compared to observed yield. The present RCM output is overestimating the different climatic variables in comparison to present observed climate for annual as well as seasonal. Therefore, framing of better management practices, mitigation programme and planning and policy making based on climate model output must ensure to get the reliable and validated RCM climate output. For that we need more precise and improved regional climate models through more research in climate modelling.