Atmospheric convection is one of the most important part of climate systems. Ranging from a few kilometers to tropospheric top, atmospheric convection is manifested by cumulus clouds with significant vertical extension compared with its horizontal scale. By transporting heat, moisture, and momentum vertically, it serves as an important driver for atmospheric circulations. Its coupling with surface processes and hydrological cycles has strong impacts on human society and ecological systems.

However, the climate models used for future climate projection still have difficulty predicting convective rainfall variability. For example, figure 1 shows the geographical distribution of diurnal rainfall amplitude from satellite observations (left). Convection rainfall varies strongly within one day over tropical lands and storm tracks over the oceans (figure 1 right). But large model biases of diurnal amplitude still exist, which can be up to 50% of mean rainfall among the state-of-the-art climate models. The CMIP6 archive show large amount of underestimation of diurnal amplitude over maritime continent and central America (figure 1 left). As productive ecological systems existing over the tropical lands, such rainfall uncertainty in future projection can lead to large uncertainty in future governance in preparation for future climate changes in these regions.

Our lab aims to understand the physical understanding of rainfall and atmospheric convection by utilizing growing satellite observations (figure 2 left) and increasing computational power with numerical models (figure 2 right). Moreover, such understanding can help Such effort will contribute to improvements of weather forecast, climate projection, and other societal and economical applications.