Diagnosing whether the increasing horizontal resolution of regional climate model inevitably capable of adding value: investigation for Indian summer monsoon.

This study demonstrated the influence of downscaling using the regional climate model (RCM) driven by Era-Interim reanalysis (EIN) in simulating different aspects of the Indian summer monsoon (ISM). It is also examined, whether increasing the horizontal resolution of RCM will inevitably be capable of adding more information to ISM characteristics and its spatio-temporal variability. In this regard, two RCM (at 50 km: Reg₅₀ and 25 km: Reg₂₅) simulations were conducted for six years from 2000 to 2005 for the South Asia Coordinated Regional Downscaling Experiment (CORDEX) domain. The added value (AV) is found to be strongly dependent on region and considered metrics. A slight improvement towards increasing spatial resolution is observed in the simulation of the mean ISM characteristics, while considerable improvements are noticed for the frequency distribution of extremes. The notable improvement in the daily climatology of precipitation is observed over the region of northeast India (~ 35%) and the Hilly region (~ 32%) and the lowest improvement over north-central India (~ 8%). The reduction of anomalously strong northeasterly flow over the southeastern Arabian Sea and strengthening of the moisture leaden southeasterly wind flow from the Bay of Bengal in Reg₂₅ compared to Reg₅₀ is consistent with the reduction of dry bias over India in Reg₂₅. The robust improvements are noticed for the heavy precipitation events (probability density function: PDF tails) and mean precipitation due to extreme precipitation events, particularly over the areas characterized by complex topographical features (e.g., the Western Ghats, Indo-Gangetic plains, and northeast India and Hilly regions) as well as over the areas having substantial bias (e.g., central India), indicating its strong sensitivity towards model resolution. The increasing latent heat flux in Reg₂₅ contributes to increasing the moisture and hence rainfall over India. Both simulations apparently simulate many of the ISM characteristics better than the EIN, thereby emphasizing the usefulness of finer resolutions in the better simulation of the Indian monsoon, especially for heavy rainfall. However, the RegCM bias is comparable to or even greater in some places than the EIN bias. This suggests that high-resolution models are important for improving performance; however, it does not necessarily mean that they can have AV for every aspect and all places. Apart from this, the substantial difference in the AV over different regions or aspects highlights the importance of carefully selecting AV matrices for the different areas and characteristics being investigated. RegCM exhibits some systematic biases in precipitation despite substantial improvement due to misrepresentation of dynamical and thermodynamical processes, including northward and eastward propagating convective bands. [ABSTRACT FROM AUTHOR]