Diabatic evolution of clouds in a Lagrangian framework: turbulence, vorticity dynamics and precipitation effects

The importance of clouds in weather and climate has long been recognised, yet accurate cloud modelling is immensely difficult due to the wide range of energetic, spatial and temporal scales involved in cloud formation, growth and decay, and the highly nonlinear nature of these turbulent cloud processes. Indeed, the turbulent behaviour of clouds is responsible for many of the uncertainties in atmospheric models, affecting cloud formation and thereby disrupting the global energy balance in climate simulations, as well as the timing and intensity of precipitation in weather forecasts. Weather and climate models fail to resolve the details of the interactions between clouds and their environment and suffer from a crude representation of important microphysical processes, such as rain and snow formation.

These processes can be studied in detail using `large eddy simulation' (LES), a widely-used computational approach employing a fixed discrete grid - a so called `Eulerian' approach. Typically, LES uses grid spacings below 100 metres to resolve such cloud-environment interactions. However, LES still suffers from substantial sensitivity to numerical mixing. This is due to the highly nonlinear nature of the dynamics and thermodynamics of clouds. For the formation of precipitation, regions of high liquid water content are crucial, and numerical mixing can prevent the occurrence of such regions in Eulerian models.

We have recently developed a promising new computational model, MPIC (Moist Parcel-In-Cell), which largely overcomes these numerical errors. MPIC deals with the dynamics of clouds in an essentially `Lagrangian' framework, i.e. by explicitly following parcels of fluid rather than approximating this motion on a fixed grid as in LES. The MPIC model represents both dynamics and processes explicitly using Lagrangian parcels that carry a volume, circulation and thermodynamic properties (e.g. potential temperature and moisture content). This approach accurately preserves key parcel properties and avoids problems due to numerical mixing at small scales that are inherent to Eulerian models.

Our research aims to adapt the MPIC model for use in realistic atmospheric case studies as well as advance our understanding of the fundamental dynamics of cloud turbulence. In particular, we will investigate how discontinuous diabatic forcing associated with condensation and evaporation modifies turbulence relative to adiabatic conditions. This research presents a novel opportunity to improve our theoretical understanding of generic features of cloud processes. It is especially timely due to the development of a massively parallel version of MPIC, which can resolve detailed processes far beyond the reach of existing numerical models.

The potential impacts of this research are vast. Since entrainment in convective clouds strongly influences heavy rainfall, new discoveries about entrainment should enable improved flood forecasting, potentially saving lives, property and business assets. Beyond this project, aspects of the MPIC model may be adopted by numerical weather models to improve the representation of cloud processes and rainfall events. The pathway to this is first to engage scientists at the Met Office, who have agreed to be a project partner. Aspects of the MPIC model will need to be integrated into the latest Met Office model, and the proposed research will make significant progress towards this goal.

The proposed research will make substantial advancements in our understanding of fundamental turbulent processes, ultimately leading to improved weather and global climate models. Extensions to the MPIC model will make it well-suited to a range of geophysical applications including cloud simulations and atmospheric chemistry experiments, making it attractive to a wide scientific and industrial audience.

The proposed research will exploit a recent computational modelling breakthrough, developed in an EPSRC-funded feasibility study and eCSE project, to advance our understanding of clouds, a critically important component of weather, climate and environmental modelling.

The work will benefit the community of scientists studying a variety of problems related to turbulent clouds and their transport and mixing properties, and using models to make environmental predictions. These include challenges of very high socio-economic importance, notably prediction of extreme rain, global climate modelling, and pollution dispersion. This community will benefit through the provision of the model, and associated research demonstrating its capabilities.

We will engage scientists in universities and forecasting centres by holding a series of workshops, which will train participants to use the modelling approach and develop further applications. We will make all software available through a public online repository, together with documentation and examples of use. We will also promote MPIC using detailed visualisations developed by a trained software developer.

Research conducted as part of the proposal, such as studying entrainment occurring at the edges of clouds, is directly relevant to the representation of clouds in Numerical Weather Prediction models. A particular application is improving forecasts for heavy precipitation, which is strongly influenced by entrainment. Our work will contribute to improved forecasts: earlier warnings save lives, property and business assets. The annual economic benefit to the UK of the Met Office and its forecasts has been calculated to be £3 billion.

While our research focuses on mathematical aspects of cloud evolution, the changes to the MPIC model that we will implement will make it possible in principle to embed MPIC in a larger-scale model for fine-scale forecasting. There is further potential for developing the model for use in weather and climate prediction, as a subgrid model or as part of the dynamical core. These ideas will be developed with our project partner, the Met Office. While development of such a model in the timescale of this project is unrealistic, the approach offers new directions for the years to come, when pressures on computing resources will demand new ideas.

The proposed framework is particularly suitable for problems involving aerosols or chemistry, as it models advection very efficiently. Aerosol-induced radiative forcing and cloud feedbacks are the largest cause of uncertainty in current estimates of climate sensitivity (IPCC report, 2014). A more accurate representation of cloud processes in climate models will allow for better quantitative arguments to be made to influence mitigation strategies to be adopted globally.

Worldwide, air pollution is known to cause a significant number of deaths. Being able to model the transport and dispersion of pollutants out of the near-surface atmospheric layer is an important part of the problem. Using the modelling approach proposed, it will be possible to efficiently model the motion and behaviour of the pollutants in clouds. The model could be further developed for real-time dispersion modelling. The pathway to this impact will be through scientists at the Met Office and the National Centre for Atmospheric Sciences in Leeds.

The applicability of our approach may extend much further and have impacts in areas far from cloud turbulence. One topical area concerns the mixing of bio-geochemical tracers by turbulent fluid motions in the oceanic sub-mesoscales (100m-10km). Our approach may enable a much more accurate representation of this flow, which would have significant impact. The pathway to impact is first through engagement with experts, already under way, and subsequently with interested stakeholders.