Parameterizing “uncertainty” in the numerical weather prediction of storms

Supervisors – Prof. Ian Renfrew (ENV), Dr Glenn Shutts (Met Office), Dr Stefano Migliorini (Univ. of Reading)

                Numerical weather prediction (NWP) models contain prognostic equations which can be solved to predict future weather at scales resolvable by the model grid. However many processes in the atmosphere occur at scales too small to be resolved: these sub-grid scale processes have to be parameterized (examples include convection, turbulence and cloud microphysics). Parameterization schemes use a simplification of the physics of such processes and, as such, are imperfect. One way of representing this imperfection is to deliberately introduce an element of “uncertainty” into the model. This can be done through a stochastic perturbation of existing parameterization schemes (see Palmer et al. 2005); or through a separate parameterization scheme, such as the Stochastic Kinetic Energy Backscatter scheme (Shutts 2005), which has been developed to include the upscale energy transfer relating to the geostrophic turbulence view of the large-scale flow (as well as the upscale influence of deep convection in mesoscale convective systems and the statistical uncertainty of orographic drag representations), i.e. accounting for the fact that small-scale Reynolds' stresses can act directly on the energy-containing synoptic scale aspects of the flow.

                In this PhD project, the impact of these relatively new parameterization approaches will be calibrated against new idealised high resolution simulations of the extra-tropics and tested through case studies of high-impact storms over the UK. NWP simulations at a relatively coarse scale (e.g. 12-24 km resolution) will be carried out, with a variety of parameterization settings tested, and with the results compared against properties of high-resolution (100 m to 1 km) simulations and field campaign observations. It is anticipated that these stochastic-dynamic parameterizations will have a significant effect on the mesoscale structure of the extratropical storms. Results will be used to fine-tune the parameterizations, with particular attention paid to the atmospheric boundary layer.  

                The project will contribute to a major NERC Programme on Storm Risk Mitigation and in particular to the DIAMET consortium (Diabatic influences on mesoscale structures in extratropical storms). There may be the opportunity to take part in the DIAMET aircraft-based field campaigns. This PhD project has dedicated funding from the National Centre for Atmospheric Sciences (NCAS) in support of DIAMET. It is hoped the project will be a CASE Award with the Met Office providing CASE funding and co-supervision. 

Further details

                Stochastic-dynamics parameterization schemes are a relatively new approach for NWP. So far they have primarily been used for medium-range weather forecasting within Ensemble Prediction Systems (e.g. Buizza et al. 1999; Palmer et al. 2005; Jung et al. 2005). In the main they have been calibrated against idealised high-resolution simulations of the tropics; for example, in “coarse-grain” experiments where statistical properties of the high-resolution simulations are used to tune the coarse resolution parameterization (e.g. Shutts and Palmer 2007).

                In this project further calibration and validation will be carried out with a specific focus on the extratropics and the atmospheric boundary layer. There is a great deal of mesoscale structure in the extratropics, often associated with particular weather systems, e.g. secondary frontal cyclones, polar mesoscale cyclones, orographic jets, mesoscale convective systems, etc; some of which are important for the ocean and climate too (e.g. Renfrew et al. 2008).  Such structures have not been a focus of previous coarse-graining studies, so a set of idealised high resolution experiments will be carried out to examine their statistical properties and validate the stochastic-dynamic parameterizations. These experiments will be run with the Met Office’s LEM (Large Eddy Model) run on UEA’s high performance computing cluster, and with the Met Office Unified Model (MetUM) on joint NERC/MetOffice Monsoon HPC facilities. The coarse-graining work will focus on properties of the required tendency perturbations and their relation to mesoscale weather features, in particular on the spatial and temporal decorrelation scales and the correlation between different parameters (e.g. u,v,T and q). This work as aimed at clarifying the nature of the backscatter process at a local, phenomenological level rather than the global spectral approach that has been used to analyse and formulate backscatter so far.

Secondly a statistical comparison against properties of high resolution simulations of several real case studies will be carried out. The cases will be chosen from the DIAMET field campaigns of Autumn 2011 and Summer 2012. Comparisons against the aircraft observations will also be undertaken, in particular against boundary-layer turbulence properties. One focus will be on the interaction of the stochastic-physics schemes and the ABL parameterization scheme, where it is known there are some long-standing issues (e.g. Brown et al. 2008).  

Further Reading

Brown A.R. et al., 2008: Upgrades to the boundary-layer scheme in the Met Office numerical weather prediction model, Bound.-Layer Meteorol. 128, 117-132.

Buizza, R.,  M. Miller and T. N. Palmer 1999: Stochastic representation of model uncertainty in the ECMWF Ensemble Prediction System, Q. J. R. Meteorol. Soc., 125, 2887-2908.

Jung, T., T.N. Palmer, G.J. Shutts, 2005: Influence of a stochastic parameterization on the frequency of occurrence of North Pacific weather regimes in the ECMWF model, Geophys. Res. Lett., 32, L23811, doi:10.1029/2005GL024248.

Palmer, T.N., GJ Shutts, R Hagedorn, FJ Doblas-Reyes, T Jung and M. Leutbecher, 2005: Representing model uncertainty in weather and climate prediction, Annual Rev. Earth Planet. Sci, 33, 163-93.

Renfrew, I.A., G.W.K. Moore, J.E. Kristjánsson, H. Ólafsson, S.L. Gray, G.N. Petersen, K. Bovis, P.R.A. Brown, I. Føre, T. Haine, C. Hay, E.A. Irvine, A. Lawrence, T. Ohigashi, S. Outten, R.S. Pickart, M. Shapiro, D. Sproson, R. Swinbank, A. Woolley, S. Zhang, 2008: The Greenland Flow Distortion experiment, Bull. Amer. Meteorol. Soc., 89, 1307-1324. 

Shutts, G. J. 2005: A kinetic energy backscatter algorithm for use in ensemble prediction systems, Quart. J. R. Meteorol. Soc. 131, 3079-3102.

Shutts, G. J. and T.N. Palmer, 2007: Convective forcing fluctuations in a cloud-resolving model: Relevance to the stochastic parameterization problem,  J. Climate, 20, 187-202.