Predictability + School on Data Assimilation

Apr 26, 2011, 9:00 AM → May 27, 2011, 7:00 PM Europe/Stockholm

132:028 (Nordita)

Allan Sacha BRUN (DSM/IRFU/SAp, UMR AIM Paris Saclay), Axel Brandenburg (Nordita), Erik Lindborg (KTH), Geert Brethouwer (KTH), Jenny Brandefelt (KTH Mechanics), Jonas Nycander (Stockholm University)

Important dates

School: 26-30 April (use this link for registration to the school)

Conference: 25-27 May

Predicting the unpredictable is a challenge that is common to various physical systems whose dynamics is governed by the equations of fluid dynamics. The oldest example is weather prediction and was developed since the 1950s by various people including the Norwegian meteorologist Ragnar Fjörtoft. Other examples include climate prediction, space weather forecast, and solar cycle forecast. The mathematics developed for these applications is extremely interesting and deserves more detailed understanding, so that these techniques can be used also in other areas where the application of this technique is less well developed.

A major difference between weather forecasting and climate or solar cycle forecasting is the long time scale and an additional shortage of diagnostic data. Climate models are also more complex and involve coupling between atmosphere and oceans. Solar cycle modeling, on the other hand, is still only at a rather exploratory level. It was only recently that a proper data assimilation method using the so-called Ensemble Kalman Filter to take into account uncertainties of dynamo model and measurements has been used for solar cycle prediction However, significant progress is expected within the next few years.

→ link to graphical representation of participants (and click on "Predictability + School on Data Assimilation")

Registration deadline: 22 March 2011 (there is no registration fee)

Papers Pires_etal96.pdf Swanson_etal98.pdf

papers (arXiv) http://xxx.lanl.gov/abs/1105.0626

Monday, May 2

10:30 AM → 11:15 AM Adaptive Sampling: theory and applications

(assumes knowledge of EnKFs and the basic concept of adjoints)

Speaker: Prof. Sharan Majumdar

3:30 PM → 4:30 PM Data assimilation of the solar cycle

Speakers: A. Sacha Brun, Laurene Jouve, Mausumi Dikpati

Tuesday, May 3

10:30 AM → 10:50 AM Things they didn't tell you last week: DA in Maths, Physics, and Decision Support

Speaker: Prof. Leonard Smith

2:00 PM → 3:00 PM Timescales and (in)predictability.

TALAGRAND: About timescales for assimilation. KANEDA: On the limits of predicting turbulence. MAGNUSSON: On estimates of predictability.

Speakers: Dr Linus MAGNUSSON, Prof. Olivier Talagrand, Prof. Yukio KANEDA

Wednesday, May 4

1:30 PM → 1:50 PM Hurricane Predictability

MAJUMDAR: Hurricane Predictability

Speaker: Prof. Sharan MAJUMDAR

Thursday, May 5

10:30 AM → 10:50 AM Data assimilation with uncertain static parameters; or Why it is hard to calibrate dynamical models.

ROUGIER: Data assimilation with uncertain static parameters Abstract: Using the link between variational methods and maximum likelihood, I explore in a non-technical way the conditions under which data assimilation produces consistent estimators, and show that with a perfect model these conditions distinguish clearly between learning about the state vector and learning about the static parameters ('calibration'). Interestingly -- perhaps paradoxically -- the situation for calibration improves when the model is imperfect, and correctly modelled as such. These results are only a sketch in response to the discussions we have been having about imperfect models, and the pros and cons of stochastic models.

Speaker: Dr Jonathan ROUGIER

Friday, May 6

11:15 AM → 12:05 PM Why are there zonal flows on Jupiter

Speaker: Prof. Jonas Nycander (Dept. of Meteorology, Stockholm University)

Monday, May 9

10:30 AM → 11:30 AM Sequential variational method for data assimilation without tangent linear and adjoint model integrations

DOBRICIC: Sequential variational method for data assimilation without tangent linear and adjoint model integrations

Speaker: Dr Srdjan DOBRICIC

Slides SrdjanDobricicSVAR_1.ppt

1:30 PM → 2:20 PM Does model resolution matter for climate predictions?

YANG: Does model resolution matter for climate predictions?

Speaker: Dr Shuting YANG

Tuesday, May 10

1:30 PM → 2:20 PM Systematic Strategies for Stochastic Climate Modeling

FRANZKE: Systematic Strategies for Stochastic Climate Modeling The climate system has a wide range of temporal and spatial scales for important physical processes. Examples include convective activity with an hourly time scale, organized synoptic scale weather systems on a daily time scale, extra-tropical low-frequency variability on a time scale of 10 days to months, to decadal time scales of the coupled atmosphere-ocean system. An understanding of the processes acting on different spatial and temporal scales is important since all these processes interact with each other due to the nonlinearities in the governing equations. Most of the current problems in understanding and predicting the climate system stem from the multi-scale nature of the climate system in that all of the above processes interact with each other and the neglect and/or misrepresentation of some of the processes lead to systematic biases of the resolved processes and uncertainties in the climate response. A better understanding of the multi-scale nature of the climate system will be crucial in making more accurate and reliable weather and climate predictions. In my presentation I will discuss systematic strategies to derive stochastic models for climate prediction. The stochastic mode reduction strategy accounts systematically for the effect of the unresolved degrees of freedom and predicts the functional form of the effective reduced equations. These procedures extend beyond simple Langevin equations with additive noise by predicting nonlinear effective equations with both additive and multiplicative (state-dependent) noises. The stochastic mode reduction strategy predicts rigorously closed form stochastic models for the slow variables in the limit of infinite separation of time-scales.

Speaker: Dr Christian Franzke

Wednesday, May 11

10:30 AM → 11:20 AM Ensembles in Data Assimilation for Numerical Weather Prediction

Speaker: Jelena Bojarova

1:30 PM → 2:20 PM The beginnings of data assimilation in geomagnetism

Speaker: Prof. Andrew Jackson

3:30 PM → 4:20 PM Special Astrophysical Fluids Seminar: Some "Non"s in Turbulence -- Nonlinear, Nonlocal, Nonequilibrium

Speaker: Prof. Yukio KANEDA

Friday, May 13

2:00 PM → 2:50 PM spontaneous discussion on practical 3DVAR implementation with some introductory remarks by Srdjan

Slides 3dvar_1.pdf Adjoint_methods.pdf

Monday, May 16

10:30 AM → 11:30 AM A Bayesian Hierarchical Modelling approach to quantify wind forcing uncertainties in ocean ensemble forecasting

DOBRICIC: A Bayesian Hierarchical Modelling approach to quantify wind forcing uncertainties in ocean ensemble forecasting

Speaker: Dr Srdjan DOBRICIC

Slides BHM-wind-ensemble-overview.pdf

Tuesday, May 17

10:30 AM → 11:20 AM Uncertainty in projections of climate change

Speaker: Jouni Räisänen

3:00 PM → 3:50 PM Parameter esimation indistinguable states

Cuellar: Parameter esimation indistinguable states

Speaker: Dr Milena Cuellar

Wednesday, May 18

2:30 PM → 2:50 PM Float trajectory assimilation in the Mediterranean forecasting system

Speaker: Jenny Nilsson

Thursday, May 19

10:30 AM → 11:20 AM An operational 3-4D-Var - practical aspects including the hybrid var/ensemble approach and performance

Speaker: Dr Nils GUSTAFSSON

Friday, May 20

10:30 AM → 11:30 AM Sequential estimation as a dynamical systems problem, including parameter estimation

Speaker: Michael Ghil

1:00 PM → 1:50 PM Rising carbon dioxide, a never ending story?

Poster

Speaker: Ralph Keeling

Monday, May 23

10:30 AM → 11:20 AM Impact of stochastic perturbations on deterministic dynamical systems

LUCARINI: Impact of stochastic perturbations on deterministic dynamical systems

Speaker: Dr Valerio LUCARINI

1:30 PM → 2:20 PM Informal discussion about solar irradiance variability etc

Speaker: Michael Ghil

Tuesday, May 24

1:00 PM → 1:50 PM Nowcasting via Gradient Descent: Moving from Theory towards Reality

Speaker: Hailiang Du

3:00 PM → 3:50 PM Thermodynamics of the climate system

http://www.misu.su.se/

Speaker: Valerio Lucarini

Wednesday, May 25

1:30 PM → 3:00 PM Session one

13:30 Michael Ghil (Ecole Normale Supérieure),
14:15 Peter Ditlevsen, (Niels Bohr Institute),
15:00 Coffee.

1:30 PM Empirical model reduction and applications

Michael Ghil (Ecole Normale Supérieure) Empirical model reduction and applications

Speaker: Prof. Michael GHIL

Slides EMR-Nordita-May'11-Main.pdf

2:15 PM Tipping points: Early warning and wishful thinking

Peter Ditlevsen, (Niels Bohr Institute) Tipping points: Early warning and wishful thinking

Speaker: Prof. Peter DITLEVSEN

3:30 PM → 6:00 PM Session 2

15:30 Milena Cuellar, 16:10 Jonathan Rougier, 16:50 Jouni
Räisänen, 17:40 bus 44 to Skansen, 18:30 Dinner (Hasselbacken).

3:30 PM Parameter and state estimation nonlinear models

Milena Cuellar (Bronx Community College: Parameter and state estimation nonlinear models

Speaker: Dr Milena CUELLAR

4:15 PM Emulating the behaviour of a large climate simulator

Jonathan Rougier, (University of Bristol), Emulating the behavior of a large climate simulator

Speaker: Dr Jonathan ROUGIER

5:00 PM Can model weighting improve projections of climate change?

Jouni Räisänen, (Department of Physics): Can model weighting improve projections of climate change?

Speaker: Dr Jouni RäISäNEN

Thursday, May 26

9:00 AM → 7:00 PM Session 3

All talks will take place in Uppsala:
http://www2.math.uu.se/cim/events/climate2011.html

Friday, May 27

9:00 AM → 10:30 AM Session 4

9:00 Hailiang Du, 9:45 Frank Kwasniok, 10:30 Coffee.

9:00 AM Nowcasting via Gradient Descent: Moving from Theory towards Reality

Hailiang Du, (London School of Economics), Nowcasting via Gradient Descent: Moving from Theory towards Reality

Speaker: Dr Hailiang DU

9:45 AM Data-based stochastic subgrid-scale parametrisation using cluster-weighted modelling.

Frank Kwasniok, (University of Exeter), Data-based stochastic subgrid-scale parametrization using cluster-weighted modeling.

Speaker: Dr Frank KWASNIOK

11:00 AM → 2:00 PM Session 5

11:00 Andreas Svedin, 11:45 Mihaela Caian, 12:30 Lunch,
14:00 End of conference.

11:00 AM Nowcasting the solar cycle, using low dimensional models.

Andreas Svedin, (Columbia University), Nowcasting the solar cycle, using low dimensional models.

Speaker: Andreas SVEDIN

11:45 AM Decadal prediction initialisation using anomaly method

Mihaela Caian, (SMHI), Decadal prediction initialization using anomaly method

Speaker: Dr Mihaela CAIAN