Finding the right Mathematical Model for Atmospheric Convection

Some of the most extreme weather on the planet is associated with deep convective storms driven by the heat released when water vapour turns into liquid cloud water or ice. These storms start out as small turbulent eddies which grow into violent thunderstorms. Storms can organise into much larger structures such as squall lines and tropical cyclones. This ever-larger scale organisation nevertheless remains, in part, controlled by processes occurring at the smallest turbulent scales. The smallest scales are not represented accurately in models used to predict the weather, which compromises the accuracy of weather forecasts. This project is aimed at evaluating new approaches to the representation of the smallest scales.

The student will assess the suitability of two types of averaging for deriving equations that represent the sub-grid scale variability associated with convection. Conditional averaging assumes two values of momentum, temperature and density per grid box in order to represent the conditions in both the convectively active (cloudy) and stable (clear sky) regions of the atmosphere. Volume averaging involves predicting means, standard deviations and correlations between variables. The project will start by assessing which averaging technique can best represent high resolution data at low resolution. However the representation of the data is just the start. The big challenge lies in creating prediction models using volume or conditionally averaged equations and in predicting the transition between small scale, turbulent motion and deep convection. The student will develop existing models using both averaging techniques and will propose and test new ideas for predicting the transition between convectively stable and convectively active regions of the atmosphere. The results will be compared with high resolution models of convection.

This project is part of ``Project Circle-A. Parametrizing Convection in the Hard Grey Zone'' at Reading which is part of the NERC/Met Office programme ``Understanding and Representing Atmospheric Convection across Scales''. The student will therefore be part of a national network. This is an opportunity to contribute to model development that will directly feed in to weather and climate forecasting.

Society and the General Public
The primary impact of this project will be directly on the accuracy and utility of Met Office weather forecasts and climate predictions via improved prediction and simulation of convection in the Met Office's Unified Model. Impacts will occur through other centres; directly through Met Office Partners in MetUM development and indirectly to other centres through promulgation of the methodology and reference data it is based on. This will have impact on end-users, from government, industry and the public, through significant improvements to their products and advice. The Met Office, in common with operational prediction centres worldwide, considers this to be a major priority, made more essential by ongoing efforts to improve the dynamical cores of models to allow best use of increased parallel computing power and, hence, higher spatial resolution, such as the Gung-Ho project in the UK (now LFRic) and the ICON project in Germany.
Weather and flood forecasting: This project will improve forecasts, certainly in regional weather forecast models and hopefully also in coarser-resolution global models. Convection is responsible for the intense rainfall that generates flash floods. The Pitt Review, written after the 2007 UK floods, stressed the need for better prediction heavy precipitation. Short-range MetUM-forecasts are now being produced in the Met Office with grid-spacing of a few km. The Environment Agency/Met Office Flood Forecasting Centre now provides a national severe rainfall and flood warning service that relies on NWP forecasts. The reliability of these forecasts is compromised by deficiencies in the representation of intense rainfall from convective systems, and a more reliable parametrization will have an immediate impact on forecast quality. The importance of parametrization of turbulence and convection cannot be understated; at present, the only remedy for deficiencies in performance is higher-resolution, which is extremely expensive. Improved parametrization will facilitate larger-domains, larger ensembles to quantify forecast uncertainty, and improved data assimilation.
Climate predictions, climate-change impact assessments: The hard grey-zone parametrizion scheme will contribute to correcting major biases in the MetUM, Recent work using the MetUM at convection-permitting resolution has shown that the climate change signal for extreme rainfall cannot be relied upon from models using current convection parametrizations. A different signal emerges from 'convection-permitting' models, but these clearly, at present, have their own significant biases.
The importance of convection goes far beyond precipitation, however. "Realistic parametrizations of cloud processes are a prerequisite for reliable current and future climate simulation" and "differences in cloud response are the primary source of inter-model differences in climate sensitivity" (IPCC, Ch. 8). The representation of convection has a number of impacts on earth-system models. Wash-out by convective rain is a primary uncertainty in the global aerosol cycle, the representation of convection is more important for the summertime Sahara dust than the land surface and rainfall changes control the future of tropical forests, such as the Amazon, where some models predict a die-back with decreased rainfall, coupling convection with the carbon cycle (IPCC, 2007, Ch. 11.6.3.2).
Academic Impact
This project will deliver new understanding of atmospheric convection that will have wide impact in the academic community, as well as improved climate and earth-system models which form the basis of further academic study. It will also deliver valuable database of high-resolution reference simulations that will be used for years to come by the UK and WGNE communities: these will likely have many uses beyond those described in this proposal, as has been found, e.g., in the NERC Cascade consortium