UPDRAFT Project is a Marie Curie Postdoctoral Fellowship awarded to Dr. Cristian Vraciu at the University of Reading, running from August 2025 to August 2027. Conducted under the supervision of Prof. Robert Plant, the project aims to develop a new parameterization for shallow and deep convection.
Most numerical weather prediction (NWP) and global climate models (GCMs) still rely on convection parameterization schemes developed over 50 years ago. These schemes assume quasi-equilibrium, meaning that cloud properties depend solely on large-scale forcing. As a result, storm intensity (e.g., precipitation rate) at any moment is linked only to the current atmospheric instability. However, recent studies show that convection also depends on its past state—a phenomenon known as convective memory—especially within the diurnal cycle. Despite numerous updates, current schemes still struggle to represent the rapid development of convective storms over land, leading to forecast errors and climate biases. The ER has proposed a new unified framework in which convective clouds form from multiple overlapping updrafts in space and time, rather than from a single bulk updraft in a homogeneous environment. While recent efforts to improve storm initiation exist, most rely on ad-hoc adjustments rather than revisiting the core assumptions of traditional formulations. Moreover, shallow (fair-weather) and deep (storm) convection are usually treated as separate phenomena, requiring additional assumptions for their interaction. The UPDRAFT project will implement this unified perspective, representing shallow and deep convection consistently, with differences emerging naturally from the number of overlapping updrafts. The new model will explicitly account for cloud–cloud interactions and environmental heterogeneity, first by quantifying their roles in storm development and then by using these insights to design a new cloud–convection parameterization scheme. This unified scheme will represent both vertical convective transport and cloud cover and will be tested in idealized simulations as a proof of concept, paving the way for integration into operational weather and climate models.
The main goal is to provide concrete evidence that the new theoretical framework can lead to improved predictions of convection, with the prospect of breaking some major problems associated with the representation of convection in NWP and GCMs. Specific objectives to this end are:
O1: Understand the mechanisms by which local moisture anomalies develop from pre-existing convection, thereby creating favourable conditions for subsequent convective elements to develop within the same or new clouds and understand the role of the environmental relative humidity in the local moisture preconditioning of deep convection (WP1);
O2: Derive a prognostic bulk plume model (a time-dependent set of equations) for updrafts with explicit updraft-cloud interaction (WP2);
O3: Design and test a toy-model for updraft-cloud interaction that is able to model the shallow-to-deep transition and develop and test the new parameterization (WP3).
The first work package (WP) is designed to improve our understanding of the role of local moisture preconditioning in the deepening of storm convection and how it is controlled by precipitation and the environment. For this task, high-resolution numerical simulations (large-eddy simulations; LES), in which the convection is resolved, will be run in idealized settings, allowing for multiple experiments with various conditions. WP2 and the toy-model to be developed in WP3.1 are designed to offer tests of the conceptual model and to critically assess potential routes for its implementation, before the actual implementation of the concepts into a new parameterization in WP3.2. For example, the plume model development in WP2 will form an important building block of the new parameterization scheme, which will be tested in an idealized single-column model (SCM). WP4 is devoted to the dissemination of findings and concepts of the project to research scientists in the field, with an emphasis on engaging professional modellers working on weather forecasting.
Numerical weather prediction (NWP) and general circulation models (GCMs) are used to provide society with important scientific information to guide decision-making and policy. The heat balance of the climate system is very sensitive to storm clouds, as they are of leading-order importance for the radiation and moisture budgets of the Earth’s atmosphere. Moreover, these clouds produce severe weather events with major societal and economic impacts. Improving the representation of storm convection in models would thus have a large societal impact through: i) better prediction of severe weather events by NWP models, which will help inform the population of the development of storms, allowing them to be less exposed to severe weather events, and ii) better climate predictions by GCMs, which will help with the design of policy to mitigate climate change and adapt to severe weather events in a warmer climate.