'Operational Approaches to Weather Risk: from hours to decades' Workshop

LSE, 22 June 2004


The Centre for the Analysis of Time Series (CATS) at LSE is leading one of the Smith Institute Faraday research projects: DIME (Direct and Inverse Methods in End-to-End Environmental Estimation). The principal investigator on this project and Director of CATS, Dr Leonard Smith, is also involved with the NOAA THORPEX Research Program, which is directed towards serving society's needs for weather and water information, the Framework 5 DEMETER project on seasonal forecasting, and the Framework 6 ENSEMBLES project on forecast reliability on all time scales.

There is technology emerging from these projects, namely probabilistic forecasting, which can place environmentally induced risk in the user domain to aid decision-making. Whilst there exist a number of important impact studies, the existing plethora of environmental data and emerging forecasting tools remain under-utilised by government and commercial organisations. There is a need to 'translate' this technology into society and commerce in general.

The Smith Institute and CATS are scoping applications of this technology and plan to increase awareness of its potential benefits in the wider industrial and commercial communities. The workshop held at LSE on 22 June 2004 was the first in a series of awareness raising workshops for industry; future meetings will be held both in the UK and in the US. Follow-up activity will be offered to help those who wish to apply these methods in their businesses. To express interest in a future workshop please follow the registration link below.

Aims and Purpose

Mitigation of weather risk through direct action is an operational reality, even if hedging and insurance options have been put in place. This meeting will survey the application of current and future sources of weather information towards the active management of weather risk in practice. The aim is to provide a realistic evaluation of what is possible with current forecast information, and what is likely to be available in the near future. This includes combining observations, multiple forecasts, and historical information in the context of specific user applications. The value of weather forecast information depends on the particular user application. We will use case studies as spring boards for discussing problems of interest to the participants.

Attendance was strictly limited to allow the consideration of specific challenges of interest to the participants.


End-to-end forecasting, weather risk, decision making under uncertainty, applications.

Expected Outcomes

  • Increased industry awareness of the methods emerging from DIME and of the potential impact of applications.
  • Higher profile for sources of state of the art meteorological information.
  • Methods for comparative analysis of commercially available forecasts.
  • Increased number of application domains.

Format and Programme

The event was organized to provide, in the morning, an overview of new options for dealing with weather risk; the afternoon provided an opportunity both for detailed illustrations (through case studies) and the presentation by industry of particular challenges.

10:00-10:30 Coffee and Registration

10:30-11:30 Introduction and Overview

Beyond hedging: examples of active management of weather risk

11:30-12:00 Information Providers

Weather forecast information in the 21st Century

12:00-12:30 New Challenges, New Needs

12:30-13:30 Lunch

13:30-15:00 Illustrations through Applications (with Discussion)

Using Ensemble Forecasts

(Forecasting wind energy generation)

(Wave height forecasts and off-shore safety)

Using Multi-Model Forecasts

(Demand forecasting: moving beyond the consensus forecast)

(Likely precipitation in the medium range)

Early Warning of High-Impact Events

(Extended hot or cold spells)

(Potential storms)

Options on More Complex Decision Making

(Ship routing)

(Contingency chains)

Correlations, Climate, and the Use of Simulation Models

(Its always windy somewhere: repackaging risk)

(Seasonal forecast information & DEMETER)

(How reliable are current climate forecasts?)

15:00-15:10 Tea/Coffee

15:10-15:45 Forecast Evaluation & DIME

15:45-16:00 Summary and Follow-up

Event Brochure

Participants received a bound copy of the event presentation material.

Background Reading

These articles and reports provide background reading on various aspects of weather risk.

End-to-end ensemble forecasting
Reference: Smith, L.A., Roulston, M. and von Hardenberg, J. 'End-to-end ensemble forecasting: towards evaluating the economic value of an ensemble prediction system', ECMWF Technical Memorandum 336 (2001), 29pp.

Disentangling uncertainty and error
Reference: Smith, L.A. 'Disentangling uncertainty and error: on the predictability of nonlinear systems', in Nonlinear Dynamics and Statistics, ed. Alistair I. Mees, Birkhauser, 31--64 (2000).

Selecting supplementary observations
Reference: Hansen, J.A. and Smith, L.A. 'The role of operational constraints in selecting supplementary observations', J. Atm. Sci. 57, 2859-2871 (2000).

Uncertainty dynamics and predictability in chaotic systems
Reference: Smith, L.A., Ziehmann, C. and Fraedrich, K. 'Uncertainty dynamics and predictability in chaotic systems', Quart. J. Royal Meteorological Soc. 125, 2855-2886 (1999).