The 237th R-CCS Cafe - part 3

The diurnal cycle of convection is an essential trigger that develops and organizes precipitation systems and contributes to the spatial and temporal distribution of monsoonal or tropical convection. The developmental features of these diurnal convection and precipitation systems are not well represented in global and regional climate models across different monsoon regions. Such uncertainty in the diurnal convection representation undermines the fundamental reliability of the simulated physical processes in the climate models. This has encouraged an examination of the unpredictability of the monsoon and tropical diurnal convection in climate models.My research has focused on observations, modeling, and data assimilation techniques to diagnose the diurnal convection cycle in a tropical monsoon and understand the importance of its representation in climate models. In this research, I found the ubiquitous nature of diurnal convection across spatial and temporal scales. None of the current generation climate models could show such a ubiquitous (universal) nature of diurnal convection, which could be due to model-based errors.Most of the model-based errors in diurnal convection precipitation representation are mainly associated with the horizontal grid resolution and cumulus parameterization of the model. High-resolution climate models also cannot capture the diurnal cycle of precipitation accurately since these also tend to utilize cumulus parameterization. Cumulus parameterization has simulated varying diurnal cycle spatial patterns in different climate models, challenging the impact of various cumulus schemes. In the present generation, cumulus parameterizations in climate models poorly represent the diurnal convection and organization of precipitation systems, unlike observations. These fundamental problems in the cumulus parameterizations arise due to closure assumptions and trigger functions that promote the linear growth of diurnal convection. Diurnal growth is explicitly non-linear, which fundamentally means that the climate models must simulate the conditions explicitly without cumulus parameterization to avoid these assumptions. A clear difference in the simulated precipitation systems was evident when convection was explicitly represented. For example, high-intensity localized and organized precipitation systems were simulated by high-resolution models without cumulus parameterization (convection-permitting) experiments.
My primary research will determine how the ubiquitous nature of diurnal convection can be represented in climate models and how much high-performance computation and data assimilation can support achieving the target.