OSA1.2 | Value Chains for Early Warning Systems
Value Chains for Early Warning Systems
Conveners: Brian Golding, Robert Neal, Jeff Da Costa, David Hoffmann | Co-convener: Chiara Marsigli
| Wed, 06 Sep, 11:00–13:00 (CEST)|Lecture room B1.05
Wed, 11:00
Successful hazardous weather warnings require information and expertise to be integrated across a multitude of domains including environmental observation, weather and hazard modelling, impact prediction, warning communication and decision making. This comes with many challenges including building effective partnerships between the different players involved in the warning process who may have different expectations about the spatio-temporal detail of the warning, different needs for uncertainty information, different abilities to handle missing information, and so on.

The value chain (or the value cycle or network) provides a useful framework for describing and understanding the many different groups, skills, tools, relationships, and data/information flows that combine to produce and deliver warnings. It can characterise who does what and how groups interact and exchange data and information to provide critical services during a warning situation (information flow mainly "down the chain"). It can also support the co-design, co-creation and co-provision of services during the service development phase (user needs propagated "up the chain"). The effectiveness of the value chain may be measured using different, yet complementary, methods and metrics that emphasise different characteristics of the value chain such as accuracy, timeliness, relevance, and socioeconomic outcomes.

Case studies of existing warning chains/cycles and high impact events can apply value chain approaches to characterise and measure the effectiveness of the tools, processes, partnerships, and infrastructure. This provides the evidence to identify shortfalls and propose investments in new capability and partnerships.

This session welcomes contributions on:
• Assessments of high-impact weather case study events using value chain/cycle approaches
• Challenges, gaps and opportunities arising from using value chains/cycles
• Value chain/cycle approaches, metrics and measures

Orals: Wed, 6 Sep | Lecture room B1.05

Chairpersons: David Hoffmann, Robert Neal, Jeff Da Costa
Block I - Chairs: Rob Neal, Jeff Da Costa
Onsite presentation
David Hoffmann, Beth Ebert, Carla Mooney, Sharan Majumdar, Martin Goeber, and Brian Golding

The weather information value chain provides a framework for characterising the production, communication, and use of information by all stakeholders in an end-to-end warning system. It covers weather and hazard monitoring, modelling and forecasting, risk assessment, communication, and preparedness activities.  

A 4-year international project under the WMO World Weather Research Programme is using value chain approaches to describe and evaluate warning systems for high-impact weather by integrating physical and social science. One of the project’s key outputs is a database questionnaire for high-impact weather event case study collection and analysis. The questionnaire is primarily aimed at scientists and practitioners to review, analyse and learn from previous experience using value chain approaches. Project scientists are using it to analyse high impact weather events that have occurred in recent years.  

Beyond the professional use for severe event assessment, the questionnaire has proven to be an effective educational tool for university students to learn about high-impact events. Undergraduate students at the University of Miami used the questionnaire to study the warning value chain for Hurricanes Ida (2021) and Ian (2022) as an assignment in an undergraduate tropical meteorology course. Similarly, undergraduate interns at the Bureau of Meteorology completed the questionnaire for the Black Summer Bushfires in south-east Australia (2019/2020) and did a comparative study of the warning value chains for Hurricane Isaias (2020) for the Caribbean and the US. Using available online resources, the students prepared their responses and collaborated in teams to present syntheses of their evaluations. Engaging students in such a cross-disciplinary study enhanced their critical thinking about high-impact weather event forecasting, impacts, warning communication and response. In this presentation we introduce the database questionnaire and how it can be used for educational purposes. 

The questionnaire and accompanying guide are freely available for anyone to use and can be downloaded at http://hiweather.net/Lists/130.html. We encourage not only the research and operational communities but also academic institutions to participate in this project by contributing case studies of high impact events and collaborating in their analysis.  


 Corresponding/presenting author: David Hoffmann, Bureau of Meteorology, Melbourne, Australia; david.hoffmann@bom.gov.au  

How to cite: Hoffmann, D., Ebert, B., Mooney, C., Majumdar, S., Goeber, M., and Golding, B.: Evaluation of the Warning Value Chain as an Educational Approach for University Students, EMS Annual Meeting 2023, Bratislava, Slovakia, 4–8 Sep 2023, EMS2023-359, https://doi.org/10.5194/ems2023-359, 2023.

Online presentation
Elizabeth Ebert, David Hoffmann, Jeff Da Costa, Xudong Liang, Brian Mills, Carla Mooney, Hellen Msemo, Jacob Pastor-Paz, Adriaan Perrels, and Andrew Tupper

Early warning systems can be conceptualised as information value chains or cycles consisting of a complex and dynamic web of nodes—where information is produced, interpreted, and used—and flows representing the communication of information, movement of resources, and nature of relationships among actors operating at each node. Value is created when information flowing through the chain supports decisions and actions that result in improved social, environmental and/or economic outcomes.

Value chain studies can provide useful insights for groups involved in early warnings. National weather services and their partners have a strong stake in understanding and improving the warning value chain because it directly affects their activities and their stakeholders. Authorities and funding bodies need to ensure that the warning services are operated according to agreed regulations and that they represent value for money. User communities in all parts of the chain receive and transmit warning information that assists them to take appropriate action at the right time; they also provide important feedback on warning effectiveness, thereby contributing to their improvement.

The WMO WWRP Value Chain project is developing a framework and guide for using value chain approaches to understand, improve, measure, and design early warning systems. Building on the seminal work of WMO (2015), Golding et al. (2019) and Lazo & Mills (2021), it draws on expertise from practitioners in the broader warning community and researchers in the natural and social sciences. It brings together process-oriented “top-down” perspectives and people-oriented “bottom-up” perspectives, offering a variety of approaches that are suitable for different types of value chain studies. Unlike most cost/benefit studies, value chain studies emphasize the means of getting to the benefits.

The framework begins with describing an existing service chain, then progresses to describe approaches for guiding service improvements, assessing the social and economic value of service improvements using quantitative and qualitative methods, and designing a new service. It includes tools and workshop ideas as well as examples of how value chain approaches are being successfully applied in the field of hydrometeorology.

The framework is currently undergoing review and is expected to be released by the WWRP around the end of 2023.


Golding, B., M. Mittermaier, C. Ross, B. Ebert, S. Panchuk, A. Scolobig, D. Johnston (2019). A value chain approach to optimizing early warning systems. Global Assessment Report on Disaster Risk Reduction, 30 pp. 

Lazo, J. K., & Mills, B. (2021). Weather-Water-Climate Value Chain(s): Giving VOICE to the Characterization of the Economic Benefits of Hydro-Met Services and Products. American Meteorological Society.

WMO (2015). Valuing weather and climate: Economic assessment of meteorological and hydrological services. WMO-No. 1153, 286 pp.

How to cite: Ebert, E., Hoffmann, D., Da Costa, J., Liang, X., Mills, B., Mooney, C., Msemo, H., Pastor-Paz, J., Perrels, A., and Tupper, A.: A framework and guide for using value chain approaches to understand, improve, measure, and design early warning systems, EMS Annual Meeting 2023, Bratislava, Slovakia, 4–8 Sep 2023, EMS2023-498, https://doi.org/10.5194/ems2023-498, 2023.

Onsite presentation
Brian Golding

Intense convective storms cause many deaths around the world each year from flash floods, landslides, lightning, tornadoes and hail. These are also some of the most difficult to forecast weather hazards and are generally not predictable in a deterministic sense by Numerical Weather Prediction models. In the UK the incidence of such storms has increased significantly in recent years, as a result of climate change, and is projected to increase further with increasing summer temperatures and corresponding increases in absolute humidity. The current UK weather warning service is focused on providing early warnings of six or more hours lead time to enable people to plan ahead for safety. However, the small spatial and temporal scales of convective storms mean that only a very general probabilistic indication can be given of timing and location in these warnings. As a result, most recipients do not prepare for the possibility of severe impacts, even when the accompanying message indicates that they are possible. One option being considered for enhancing the current warning service is to developed a complementary very short range warning focused on storms that present a risk to life. This implies a particular set of responses from warning recipients which make specific demands on the information communicated and the means of communication. On the other hand, current forecasting capabilities for these storms are limited and there are known constraints to forecast improvement. This presentation will demonstrate how a value chain approach has helped to identify the critical components of such a warning system, mapping them on to existing and projected capabilities. Results of a preliminary feasibility study will be shown.

How to cite: Golding, B.: Application of the warning value chain in designing a new warning capability for the UK, EMS Annual Meeting 2023, Bratislava, Slovakia, 4–8 Sep 2023, EMS2023-116, https://doi.org/10.5194/ems2023-116, 2023.

Onsite presentation
Solfrid Agersten and Anders Sivle

Several changes have been made in Norway over the last five years. The warnings are now issued in both Norwegian and English, and include color/warning level, advises, expected impacts, illustrations and probability information. The CAP format is used and warning dissemination is also made available through a broader range of media and specifically in social media channels. The warnings also include wording regarding impacts and wanted behavior.

MET Norway is strongly involved in risk assessments regarding probabilities of weather exceeding given thresholds i.e. traditional forecasting, and now routinely participates in video conferences with important contingency actors on incident handling. Verification reports are also written regularly after orange and red warnings, and a weekly report of the issued warnings and their impacts are made.

Gradually more phenomena have been included in the warning service by expert teams (meteorologists specialized on a certain phenomenon), based on feedback on user needs as well as the effects of a changing Norwegian climate. Flash-floods are an example of a phenomena that has occurred more often in recent years with a lot of damage and impact on transportation and other sectors.

Even though the thresholds for the warnings are still meteorologically based, they are based on return-values and/or discussed with large user-segments as the road-authorities, seen in the perspective of expected impacts. Norway is a country with a lot of mountain roads so to warn the different drivers about possible problems or a closed road is of large socio-economic importance.

A lot of the impacts and advice used are based on dialogue with relevant user groups. In the presentation the weather warning process and service in Norway will be shown and a glimpse into our plans to even improve it more.

How to cite: Agersten, S. and Sivle, A.: The way towards impact based warnings in Norway, EMS Annual Meeting 2023, Bratislava, Slovakia, 4–8 Sep 2023, EMS2023-521, https://doi.org/10.5194/ems2023-521, 2023.

Block II - Chairs: David Hoffmann, Brian Golding
Onsite presentation
Robert Neal, Helen Titley, Brian Golding, Caroline Jones, Seshagirirao Kolusu, John Mooney, Joanne Robbins, and Faye Wyatt

The warning value chain is a concept whereby the process of producing a warning is represented by a chain of sources of expertise (components), connected by bridges that convey bidirectional information exchanges. Uncertainties exist at all stages of the warning value chain. For example, uncertainties exist in the current (observational) state of the atmosphere used to initialise the numerical weather prediction models. This in turn contributes towards weather forecast uncertainties (e.g., ensemble-generated forecast probabilities and run-to-run variability). Weather forecast uncertainties then feed into the hazard and impact forecasts where they can be amplified – such as through uncertainties in defining hazard footprints or impact assessments. Often it comes down to operational meteorologists to examine the varying levels of forecast uncertainty across several value chain components and assess the real likelihood of high impact weather and its potential impacts. This presentation will focus on the challenges of forecast uncertainty within the value chain, using the forecasts and warnings associated with Storm Eunice which affected southern parts of the UK in February 2022 as an example. The Met Office Weather Impacts team recently used the warning value chain questionnaire produced by the WMO’s High Impact Weather (HIWeather) Warning Value Chain Flagship Project, to carry out a detailed warning value chain assessment for this event, where evidence was considered from across the value chain. Results showed that all components of the value chain performed well overall, and it was clearly a highly successful set of warnings as shown by the large reach and public/emergency response. However, several recommendations were still made and challenges resulting from forecast uncertainty were evident across many value chain components. A focus of this presentation will be on how operational meteorologists interpreted the changing forecast signal and how this affected warning issuance and communication.

How to cite: Neal, R., Titley, H., Golding, B., Jones, C., Kolusu, S., Mooney, J., Robbins, J., and Wyatt, F.: Challenges of forecast uncertainty within the warning value chain – a UK example from Storm Eunice, EMS Annual Meeting 2023, Bratislava, Slovakia, 4–8 Sep 2023, EMS2023-42, https://doi.org/10.5194/ems2023-42, 2023.

Onsite presentation
Jeff Da Costa, Hannah Cloke, Jessica Neumann, and Steve Robinson

The European Flood disaster caused widespread devastation across national borders between the 14-15th  July 2021 with an estimated total damage of EUR 32 billion. The disaster brought human suffering to the communities hardest hit with hundreds of reported casualties in western Germany and eastern Belgium. Serious doubts about the efficacy of Early Warning Systems (EWSs) were raised by scientists, politicians and flood victims. The response to the hazard was largely delayed and deficient in terms of warning dissemination to vulnerable groups. Institutionally, the disaster was often deemed unpredictable and extraordinary by decision-makers. However, knowledge risk on the predictability of the hazard has since demonstrated that, precursor signs of a major flood event were detected by the EFAS almost a week ahead of the event. Signs of a potential extreme hydrometeorological event were identified by July 10th by ECMWF’s Extreme forecast Index (EFI). The DWD accurately forecasted the potential for an extraordinary precipitation event at least 48 hours in advance in Germany and in neighbouring affected countries. Despite initial uncertainties related to precise prediction of total rainfall amounts, this was a well forecasted hazard and not unexpected. In the Aftermath of the disaster,  Germany and in Belgium were faced with investigations against neglect of officials during the floods and manslaughter accusations over flood deaths. Meanwhile, Luxembourg, a small country nestled between the most affected regions of Germany and Belgium respectively went largely unnoticed and questions related to the flood response mostly unanswered. The paper will focus on Luxembourg, addressing knowledge gaps relating to the existing disaster management mechanisms and policies in place during the event. Preliminary research on the existing disaster management mechanisms and policies in place will examine Luxembourg’s conceptualisation of what EWSs are and how they performed during the flood event in 2021. Aspects of people-centred EWSs (Risk Knowledge, Monitoring and Warning, Warning Dissemination and Communication, Response Capability) will be considered. A systems thinking approach will enable for a detailed timeline of events related to the functioning of EWSs. The main objective is the exploration of the complex links and dynamics involved in the flow of information during critical phases of disaster management. The case study questionnaire made available by the WMO WWRP Value Chain Framework Project will act as a guide to extract and conceptualise information flows. Expected outcomes are the identification of potential shortcomings and strengths of the system. The absence of a universal definition for EWSs and their place within disaster management frameworks highlights the need for future research of disaster risk in the context of extreme events. Root causes leading to gridlocks in currently operational EWSs are non-linear and call for integrated disaster risk research approach.

How to cite: Da Costa, J., Cloke, H., Neumann, J., and Robinson, S.: Unprecedented but not unexpected: A case study of the European Flood disaster 2021 in Luxembourg using a value chain approach., EMS Annual Meeting 2023, Bratislava, Slovakia, 4–8 Sep 2023, EMS2023-435, https://doi.org/10.5194/ems2023-435, 2023.

Online presentation
Case study evaluation of the prediction and communication of flash flooding risk using a value chain approach.
Carla Mooney, David Wilke, Paul Fox-Hughes, Beth Ebert, and David Hoffmann
Onsite presentation
Which prognostic minimum temperatures are most appropriate for cold wave warnings
Vjeran Magjarevic and Lidija Srnec