HDCP2 - EMCL

HD(CP)² - High definition clouds and precipitation for advancing climate prediction

Aims & Objectives

Our lack of understanding of cloud and precipitation processes is arguably the foremost problem of climate simulation, and by inference prediction; we propose an initiative for addressing this cloud- precipitation (CP) problem. If successful we expect a significant reduction in the uncertainty of climate change projections, both on global and regional scales. Eliminating the CP roadblock in its entirety stands to reduce uncertainties in climate change projections by more than 50 %, even more on regional scales and with respect to climate extremes, which are now very uncertain. Hence a significant, i.e., factor of two, reduction in CP associated uncertainty is a worthy ambition.

Our proposed initiative would leverage rapid developments in simulation and measurement science to provide both new insights and alternative strategies, for resolving or circumventing the CP roadblock. Specifically we envision three goals:

Improved climate prediction: to significantly improve the CP representation in climate models.
Quantification of uncertainty: to quantify how much of the imprecision in current climate projections one can expect to eliminate through an improved CP representation.
A great leap: to place national research efforts at the forefront of international efforts to harness new breakthroughs in environmental computing and sensing.

Research Topics

Numerical simulations of climate, which are capable of resolving clouds and precipitation processes at the desired level of accuracy, require a very fine spatial resolution at a horizontal grid scale of 100 m and vertical grid spacings, which are considerably finer. The execution of such simulations over a 1000 km by 1000 km area over Germany and Central Europe therefore leads to tremendous computational costs and requires the efficient utilization of the most powerful supercomputers that are available today. Important aspects in this context are:

Numerical analysis of the deployed algorithms concerning accuracy, scalability and robustness
Physically consistent numerical schemes
Efficient and scalable time-stepping schemes