Growing our science
Met Office Academic Partnership
The Met Office Academic Partnership is the Met Office, the University of Exeter, the University of Leeds,
the University of Reading and the University of Oxford
Met Office Academic Partnership
What is it?
The Met Office Academic Partnership is a cluster of research excellence. It brings together the Met Office with the
leading UK universities in weather and climate science (University of Exeter, University of Leeds, University of
Reading and University of Oxford) through a formal collaboration to advance the science and skill of weather and
climate prediction.
What are its
aims?
• To draw together world-class expertise around a focused programme of joint research to tackle key challenges
in weather and climate science and prediction.
• To maximise return on the UK’s investment in research and development in its leading institutions, to provide
society with the best possible advice.
• To combine our strengths to secure the UK’s position in leading the world in weather forecasting and climate
prediction.
• To build a cluster of research excellence that is instrumental in determining priorities for future funding and
influencing the European Union Framework agenda.
• To provide an outstanding environment to develop the science leaders of tomorrow in this very challenging
area of research and delivery.
Dr Jon Petch
Jon is the head of Met Office Science Partnerships. He is responsible for
the Met Office's national and international relationships with other science
organisations, including the Unified Model partnerships, relationship with
NERC through the JWCRP and with the academic partnerships. Jon also
continues to carry out research in areas related to atmospheric processes
and parametrizations.
For further information please contact: www.metoffice.gov.uk/research/partnership
or email chiara.piccolo@metoffice.gov.uk or pip.gilbert@metoffice.gov.uk
Challenges in Weather and Climate Prediction
The atmosphere and ocean are complex physical, chemical and biological systems which need to be modelled to
make predictions. The challenge we set our academic partners is to help us to understand these systems in order
to determine how best to model them in the future, especially on computers with very large numbers of processors.
For short range (local) weather forecasting (up to ~36 hours) the main challenges are to understand predictability at
convective scales and to determine how complex the representation of chemical and microphysical processes
needs to be in order to make forecasts of weather parameters and air quality. Initialisation (data assimilation) at
these convection permitting scales (~1km) is also a considerable challenge, being a relatively new and undeveloped
field of endeavour. Data assimilation and ensemble prediction at such scales also require that the model
simulations be unbiased and provide realistic distributions of weather parameters.
The global weather and climate predictions of the next decade will be made at horizontal resolutions approaching
10km, at which the most difficult challenge in atmospheric prediction remains the representation of organised
convection. We will also need to determine how best to couple convective, microphysical and radiative processes
together so that the atmospheric component of the model is as realistic as possible and does not cause drifts away
from reality when coupled to the ocean and land surface, nor when used as a component in data assimilation.
For us to be able to run global forecast models (both weather and climate) at these increasingly high resolutions, the
models will need to be able to exploit the power of the next generation of massively parallel supercomputers. As a
result the GungHo project was initiated two years ago to deliver a step change in the scalability of the UM by moving
away from the current latitude-longitude grid to a more globally uniform grid. This is an exciting five-year
collaborative project between the Met Office, NERC (via various of our academic partners) and STFC.
Improving Projections of Climate. Our ability to make
projections of how future climate will evolve is limited by the
accuracy of climate models and our understanding of feedbacks in
the climate system. We constantly improve climate models but it is
possible to advance our understanding by attempting to quantify
uncertainties in projections. This quantification relies on looking at
multiple climate models and evaluating their ability to reproduce
present-day and past climate, variability and change, with a focus
is on processes which relate to the uncertainty in projections.
More Accurate Weather and Climate Predictions. Weather is a
dynamical phenomena and dramatic changes in weather can
happen over a matter of hours. At the heart of any weather
forecast model is a component which solves approximations to
the dynamical equations of motion of the atmosphere. The
challenge is to both improve the accuracy of the numerical
techniques that are used to solve those equations and to make
most efficient use of current and future supercomputing power.
Scale interactions in weather and climate. Our models
encapsulate our very mature understanding of physical and
dynamical processes in the earth system. However, their
performance is at their weakest when the interaction
between processes, often at very different scales, produces
organised behaviour, relatively long-lived, behaviour.
Examples include very short range forecasting of intense,
organised convection to the global-scale organisation of the
MJO, which have a profound impact on regional weather
and climate. It is imperative that we improve our
understanding of these processes and develop model
representations of processes which interact more
accurately.
Data assimilation in complex models. We make use of
very sophisticated techniques to incorporate observations
into our atmospheric weather forecast systems. However,
the success of these depends a great deal on the close
relationship between variables such as wind and
temperature at large scales. We have entered an era where
we need to develop data assimilation techniques for much
more challenging systems, from very small-scale weather
(thunderstorms, fog), to coupled
ocean/atmosphere/chemistry/land-surface and even spaceweather models.
Clouds and the water cycle. Clouds and water vapour lead to
climate feedbacks which are very large relative to anthropogenic
forcing. We need to quantify these feedbacks, particularly for the
tropics and subtropics where rainfall is critical to the livelihoods of
a large fraction of the world’s population. Atmospheric aerosols
have an influence on regional climates and high-impact weather
patterns by modifying different kinds of cloud and precipitation,
with cloud ice being particularly uncertain. Progress demands new
laboratory, field, theoretical and modelling studies.
Cold weather, low-level clouds and fog. Travel disruption due to
icy roads or poor visibility at airports is very costly to the UK
economy. Better field measurements and theoretical
understanding of the boundary layer are required in order to deliver
reliable local forecasts of low cloud, fog and temperatures on
scales of a few kilometres.
Probabilistic prediction of weather and climate. A
quantitative appreciation of the uncertainties involved in
weather and climate prediction is essential if rational
decision-making is to be based on such predictions.
Communicating and interpreting probabilistic information to
users and decision-makers must also be done with care and
clarity. Ensemble modelling methods provide a possible
way of estimating uncertainties, but only if model variability
represents the true variability of the climate system.
Stochastic parameterization techniques may help to improve
the ability of models to represent climate variability more
faithfully.
Oceans and Climate. The oceans play a major role in the
climate system, storing and transporting heat across the
planet but also sequestering carbon in complex,
biogeochemical cycles. Modelling these processes and
predicting their response to changing climate forcings pose
major challenges to both modellers and observers because
of the immense range of scales active within the oceans and
the complex nonlinear interactions between the oceans,
atmosphere and cryosphere.
University of Leeds
Professor Doug Parker
Doug is a meteorologist who conducts research into the behaviour of
severe storms, including those affecting the UK and Northwest Europe
and those which occur in Africa. His research combines mathematical
theory, computer models and novel field measurements to understand
these weather systems and improve their prediction.
Achievements of the partnership
The University of Leeds has a continuing history of leading major observational activities jointly with
the Met Office, over more than 2 decades. This programme includes a long-term commitment to UK
field measurements, as well as leadership of international observing campaigns. Together, we have
successfully used observations to develop theoretical understanding of the atmosphere, and to
develop the next generation of weather and climate prediction models. Examples of such activities
include a series of mountain-wave experiments conducted in Scotland and the North of England since
the early 1990s, the Arctic Summer Cloud Ocean Study (ASCOS) in 2008, and the recent COLPEX
experiment measuring night-time cold air in the valleys of Shropshire.
An important theme in our partnership is the development of cutting-edge models for atmospheric
composition. Our small-scale research models are capable of representing the interactions of
aerosols and chemical tracers with individual clouds, while our global models are able to scale such
effects up to the Earth’s climate system. We are conducting new research into the assessment of
uncertainty in such models of composition – analysis of uncertainty is necessary for the use of model
results in decision-making.
Weather and climate models have real problems to make accurate forecasts for many parts of Africa.
The University of Leeds is a centre for the study of the dynamics of weather and climate over Africa.
With the Met Office, we are using field observations from major international field experiments such as
the African Monsoon Multidisciplinary Analysis (AMMA), to challenge and improve the models. Our
results are having benefits for various areas of the human environment in Africa, including the
distribution and intensity of rainfall, its impact on crops, and the lifting of desert dust.
Future directions for Leeds – Met Office



We will continue to lead and contribute to
national and international observational
programmes, and use these to challenge
our mathematical theory and our
predictive models.
We are making a major strategic effort to
advance our understanding and modelling
of cloud-aerosol interactions, from the
cloud scale to the global scale.
We are embarking on a major joint effort
in regard to the improved understanding
and modelling of tropical continental water
cycles.
Met Office Lead for
Leeds: Simon Vosper
University of Exeter
Professor Mat Collins
Mat is a climate scientist with interests in climate modeling, climate
variability and projections of future climate change. His research uses
climate models to understand the climate system and to project its future
under scenarios of greenhouse gas changes. Prior to joining the
University of Exeter, Mat was a scientist and manager in the Climate
Prediction group of the Met Office Hadley Centre.
Professor John Thuburn
John’s main research interest is in numerical methods; how can we solve
the equations governing the behaviour of the atmosphere as efficiently and
accurately as possible? For about ten years he has been collaborating
with the Met Office on a project to develop an improved ‘dynamical core’
for the Met Office’s weather and climate prediction model.
Achievements of the partnership
A major focus of the Exeter-Met Office partnership so far has been in the development of numerical techniques
to improve weather and climate prediction. The ENDGame dynamical core of the Unified Model improves on the
previous one in several ways: the optional use of a conservative transport scheme for mass and tracers; the use
of a more stable time scheme based on an iterative approach; and the consistent inclusion of several switchable
options for the geometric and dynamical approximations. ENDGame improves the model's numerical accuracy,
robustness and scalability and it is anticipated to become operational in 2013/14.
Weather and Climate Predictions are imperfect and there is a constant effort to verify, calibrate and utilize
information from prediction systems. Exeter has considerable expertise in environmental statistics that has been
applied to quantify uncertainty in predictions with emphasis on rare and extreme events. A current focus is on
extratropical windstorms. A team from the University and the Met Office are currently putting together a
catalogue of extratropical extreme wind storms which will be of great use to researchers and to industrial
collaborators.
Under scenarios of future increases in greenhouse gases, both physical and biogeochemical processes drive
feedbacks in the system that may result in changes that have significant impact on society. The possibility of
‘tipping points’, where there response of the system is very much faster than the applied forcing, has been a
feature of the Exeter-Met Office collaboration. A recent success has provided a constraint on Amazon dieback
based on the ability of models to reproduce interannual fluxes of CO 2 suggesting that previous scenarios of
Amazon collapse and associated carbon cycle feedback are less likely than previously thought.
Future directions for Exeter – Met Office



The partnership will continue to aid in
climate model development and to enhance
our understanding of climate change using
mathematical and statistical techniques.
The concept of ‘process-based emergent
constraints’; finding metrics of climate
model processes that are correlated with
future projections of different variables, will
be further developed.
Exeter University will provide training for
Met Office staff in numerical modeling
techniques, climate science, environmental
statistics and scientific writing.
Met Office Lead for
Exeter: Nigel Wood
University of Reading
Professor Peter Clark
Peter is a physicist and is particularly interested in the science behind
the short-range forecasting of high impact weather. His research is
primarily focussed on how we represent complex turbulent flows, such
as in thunderstorms and around urban areas, in mathematical models.
Achievements of the partnership
The Meteorology Department of the University of Reading has had a close relationship with the Met Office
since its inception in 1965. It has a long history developing fundamental understanding of the dynamics of
weather and climate through observation and fundamental theory, and now also includes a large group
devoted to the assimilation of observations into models. It is closely joined to the Department of Mathematics,
which undertakes research in the mathematics of numerical methods in fluid dynamical modelling, data
assimilation and dynamical systems.
Together, we have a long history of developing and using high-resolution versions of the Met Office’s climate
models, in particular the HiGEM model. The objective is not just to produce more detailed predictions but also
to find out if deficiencies in models arise from our inability to properly resolve large-scale processes (such as
Rossby or Kelvin waves) or the way we represent small-scale processes such as convection. This has been
combined, in the CASCADE project, with studies of large cloud systems using extremely high-resolution
models. This built on the experience gained by the development and assessment of high-resolution weather
forecast models which now form the basis of short-range weather forecasting over the UK.
Our understanding of the physics of urban areas has improved considerably in recent years. The combination
of laboratory and field measurements at Reading led to the development of a new parametrization of urban
surface exchange (MORUSES) which has been incorporated into the Joint UK Land Surface model (JULES)
and thence into the Unified model and is being used both to improve weather forecasts in the urban
environment and to investigate future urban climate.
Future directions for Reading – Met Office



We shall continue to improve the
representation of basic processes in weather
and climate models and to establish the role
of small-scale processes in modifying largescale weather and climate.
We shall develop advanced methods to
assimilate data and hence improve our
ability to forecast from hours to seasons and
beyond, and will extend our models to
forecast 'Space Weather'.
We are determining the role of
anthropogenic emissions and natural
processes in climate change and variability
and continue to develop strategies for
adaptation and mitigation.
Met Office Lead for
Reading: Mark Ringer
University of Oxford
Professor Peter Read
Peter is a climate physicist with research interests in the fundamental fluid
dynamical processes in atmospheres and oceans that govern long range
transport and predictability. This entails the study of complex turbulent motions in
observations, numerical models and laboratory analogues of atmospheric
circulation which he applies not only to the Earth's atmosphere but also to the
atmospheres of other planets. He began his career as a research scientist in
geophysical fluid dynamics at the Met Office before moving to Oxford.
The three examples below highlight the interdisciplinary approach used to address key research areas such
as ocean modelling and the interactions between weather, climate and climate change interactions.
Representing uncertainty in weather and climate models
We are researching stochastic parametrisation schemes to
represent uncertainty in weather and climate models. Work
ranges from proof-of-concept experiments in simple dynamical
systems, to investigating new stochastic schemes for use in
ocean models, to using global numerical weather prediction
models to test improved stochastic parametrisations of
convection. An exciting new research theme, involving
collaboration between the Physics and Mathematics departments,
explores the use of energy efficient stochastic computer hardware
for use in atmospheric modelling.
RMS Error-Spread for different representations
of convective uncertainty.
a) 3 year average SSH variance (2005-2007) from AVISO altimeter data.
b) 3 year average SSH variance (2005-2007) from ORCA025.
c) 3 year average SSH variance (2005-2007) from ORCA0083.
Bias correction in ocean modelling
Researchers in the University’s Oceanography Group
have worked with the Met Office to provide them with
access to a wealth of detailed understanding of
Southern Ocean circulation and dynamics. This
allowed for the Met Office to better account for the
large warm SST model bias in the Southern Ocean
which was having an adverse impact on climate
simulations, and in particular the representation of
the globally significant heat/carbon uptakes that are
known to occur in the Southern Ocean.
Climate-weather interactions
At the heart of existing collaborations on weather-climate interactions is
understanding the weather as a chaotic system whose statistics, or the
"shape of the chaotic attractor", can be affected by external drivers like
our changing sun, changing patterns of aerosol load in the atmosphere
and rising greenhouse gas levels. Understanding and responding to
changes in weather-related risks is at the heart of the new science of
climate services and is greatly enhanced using our distributed
Extreme weather risks for African infrastructure
computing network climateprediction.net.
Future directions for Oxford – Met Office
 Continue to lead and contribute to advanced remote sensing methods to
obtain key observations of atmospheric composition, clouds and aerosols.
 Develop a quantitative, multidisciplinary approach to model complex
climate interactions between the atmosphere, ocean, cryosphere and land
surface.
 Develop innovative ways of applying probabilistic and stochastic
approaches to the prediction of weather and climate: used for attribution
studies and assessment of the risks to water resources, ecosystems and
infrastructure.
Met Office Lead for
Oxford: John Eyre
Met Office
FitzRoy Road, Exeter
Devon, EX1 3PB
United Kingdom
Tel: 0870 900 0100
Produced by the Met Office
Fax: 0870 900 5050
enquiries@metoffice.gov.uk
www.metoffice.gov.uk
© Crown copyright 2013 10/0340
Met Office and the Met Office logo