CliMathNet Conference 2014 Plenary speakers

Brian Hoskins (University of Reading/Imperial College)

Equatorial waves in the atmosphere (Presentation)

The standard theory of equatorial waves assumes a resting basic state, adiabatic motion and small amplitude perturbations. All of these assumptions are clearly invalid in the lower atmosphere. In this talk it will be shown that projections on to the theoretical horizontal mode structures separately at each level in the atmosphere enables identification of the structures in the atmosphere that correspond to the theoretical waves, and yields interesting results about their behaviour and importance.


Douglas Maraun (GEOMAR, Kiel)

Challenges in Downscaling Research

Many potential impacts of weather events are experienced at the local scale. To simulate changes of these impacts in response to climate change, impact modellers often demand for high resolution climate change simulations as input. These are typically derived by downscaling approaches, that aim to bridge the gap between relatively coarse resolution global climate model simulations and the resolution desired by impact modellers. The resulting high resolution scenarios, however, are afflicted with substantial uncertainties: the input of global climate models is in general biased towards real climate; the downscaling approaches themselves might be deficient, and finally, the results are subject to random fluctuations of internal climate variability. 

Here, I critically review different downscaling approaches with a focus on statistical downscaling and correction methods and highlight knowledge gaps. I discuss major sources of uncertainties and errors for regional climate change simulations. Finally, the European network VALUE will be presented, that aims to bring climate scientists and statisticians together to validate and improve downscaling methods.


Tim Palmer (University of Oxford)

Greater accuracy with less precision: a new paradigm for weather and climate prediction

Weather and climate models have become increasingly important tools for making society more resilient to extremes of weather and for helping society plan for possible changes in future climate.  These models are based on known laws of physics, but, because of computing constraints, are solved over a considerably reduced range of scales than are described in the mathematical expression of these laws. This generically leads to systematic errors when models are compared with observations.  A new paradigm is proposed for solving these equations which sacrifices precision and determinism for small‐scale motions. It is suggested that this sacrifice may allow the truncation scale of weather and climate models to extend down to cloud scales in the coming years, leading to more accurate predictions of future weather and climate.


Ted Shepherd (University of Reading)

Atmospheric circulation: the wild card of climate change (Presentation)

Our confidence in the atmospheric circulation response to climate change is much lower than for energetic/thermodynamic aspects of climate, yet circulation exerts a strong control on regional aspects of climate change, including weather‐related extremes such as drought and flooding. Two major sources of uncertainty are the apparently strong nonlinearity (i.e. state dependence) of the circulation response to external forcings, reflected in the divergence of model projections of circulation‐related fields (including precipitation) in many regions, and low‐frequency chaotic variability. There are good reasons to believe that these two phenomena are somehow related. This talk will address some of the mathematical challenges raised by these issues. 

Jonathan Tawn (Lancaster University)

Multivariate extremes value methods for univariate and spatial flood risk assessment (Presentation)

The talk will cover two distinct problems in flood risk assessment: the estimation of the distribution of flood peaks at a site and the estimation of the distribution of “financial loss” over a region from flooding. Approaches based on univariate extreme value theory exist for each of these, with the one for flood peaks being very widely used. Both of these problems are essentially multivariate problems. In this talk I will present a multivariate extreme value approach to each of the two problems that offers substantial improvements over the existing methods.

Claudia Tebaldi (Climate Central and NCAR), joint with Reto Knutti, Ben Sanderson

Characterizing uncertainty through climate model ensembles: open issues with model dependence, performance, and robustness (Presentation)

An overview of recent coupled model intercomparisons is given along with a set of major challenges in interpreting them. It is shown that uncertainty in climate projections is difficult to quantify, and has not decreased significantly in the past few years, partly as a result of irreducible climate variability. Progress in model evaluation, as well as statistical methods to interpret and combine model projections, is urgently needed, in particular as more models of different quality and complexity, including perturbed physics ensembles and ensembles with structurally different models become available. We show however, with some examples, that many aspects related to the idiosyncratic nature of multi‐model ensembles get in the way of these developments and, ultimately, of a robust quantification – let alone reduction – of uncertainties in future projections. In particular we will try to elaborate on the issue of model weighting on the basis of observables, that of model (in)dependence and that related to the identification of the forced signal amidst the noise generated by internal and natural variability.


John Thuburn (University of Exeter)

Physical fidelity of numerical methods for weather and climate models (Presentation)

The atmosphere exhibits behaviour that is strongly multiscale in space and time, with shallow spectra in wavenumber and frequency space. In numerical models, errors in solving the governing equations are dominated by the marginally resolved (and unresolved) scales. In this situation, increasing the order of accuracy of our numerical methods will not necessarily reduce errors.  Instead, we should seek numerical methods that correctly represent the important physics, as far as possible, on all scales. Some examples that are important for atmospheric modelling include conservation properties, the ability to represent hydrostatic and geostrophic balance, wave propagation and the avoidance of computational modes, and coupling to unresolved scales.

Some of these desirable properties can be especially difficult to achieve on the quasi‐uniform grids that will be needed for parallel scalability on future supercomputers, but progress is possible by ensuring that the numerical method mimics key mathematical properties of the continuous governing equations. This idea will be illustrated using a `mimetic' finite element approach developed under the Met Office ‐ NERC ‐ STFC GungHo project.

Coupling to unresolved processes is particularly challenging because the coupling is typically most strong on marginally resolved scales; we must take account of how numerical errors on the marginally resolved scales interact with the subgrid model representing the unresolved processes. Some ideas will be illustrated with two examples: (i) energy and enstrophy cascades in twodimensional turbulence; (ii) a simple scheme to mimic some of the challenges of convectiondynamics coupling in shallow water models.

Laure Zanna (University of Oxford)

An ocean of uncertainties in climate predictions (Presentation)

The ocean plays an important role in the response of the climate system to external forcing. For example, the rate of ocean heat and carbon uptake will determine the amplitude and pattern of warming over land resulting from CO2 emissions. Furthermore, the ocean with its long memory can influence the predictability horizon of various quantities of climatic interest. Despite the ocean’s crucial role in climate and the improvement of ocean climate models, many of the relevant dynamical processes, involving multi‐scale interactions, remain poorly understood. 

The aim of this talk will be to discuss some of the key ocean processes regulating the natural and forced response of the climate system. I will consider various timescales using observations and a hierarchy of models (from theoretical to state‐of‐the‐art global climate models) in concert with various mathematical tools. I will focus on ways to quantify the predictability of the ocean circulation and associated uncertainties arising from sparse observations (initial conditions uncertainty) and imperfect representation of sub‐grid processes (model uncertainty).