Abstract:
In this study, the performance of the output data from the latest high-resolution Coordinated Regional Climate Downscaling Experiment-Coordinated Output for Regional Evaluations (CORDEX-CORE) model simulations is assessed with respect to the corresponding gridded Indian Meteorological Department (IMD) and ERA5 observations in representing the monsoonal precipitation over northeast India (NEI) for the period 1979-2005. Here, in this study, three different RCM model simulations (COSMO, RegCM4.7 and REMO) downscaled over the South Asian CORDEX domain are used. To understand how these RCMs perform over a data scarce region like NEI, various statistical techniques were used to evaluate their performance. Our results show that the COSMO model experiments perform better (moderate bias) with respect to both the observations as compared to other two RCMs. In contrast, the RegCM model experiments show very wet bias, whereas the REMO model experiments show more dry bias across a significant part of the study area. Quantitatively, the COSMO suite of models exhibits a little overestimation (~7 to 13%) in comparison to IMD, while an underestimation (~14 to 18%) is seen with ERA5. However, with respect to both the observations, the monsoon precipitation is overestimated (~15 to 80%) by the RegCM model and is underestimated (~15 to 50%) by the REMO model experiments. Overall, the output data from the CORDEX-CORE model experiments reproduces the monsoon precipitation over the study region but with biases that vary spatially over the study region. Finally, the results from this study offer a thorough assessment of the output data from the CORDEX-CORE model experiments over the largely unexplored NEI and hence can be considered a good tool to study climatological processes over the data-scare regions.
