DEALING with UNCERTAINTIES
This session will also include examples of how science can and should support decision-making. In this context a special section this year will be dedicated to the highly important issue of Dealing with Uncertainties:
Weather forecasts have matured substantially in providing reliable probabilistic predictions, with a useful quantification of forecast uncertainties. Including this information in the communication of forecasts and warnings, and integrating it into downstream models and decision-making processes has become increasingly common practice.
Including uncertainties not only implies the interpretation of ‘raw’ uncertainty information in ensemble forecasts, their post-processing, and visualization, but also the integration of a wide range of non-meteorological aspects such as vulnerability and exposure data to estimate risk and the social, psychological and economic aspects which affect human decision-making.
In this session, we aim to support a holistic perspective on issues that arise when making use of uncertainty information of weather forecasts in decision processes and applications.
Weather forecasting science has long been developing ensemble forecasts as a way to improve forecast capability and provide better information to support users’ decisions. The science is well proven and the Met Office will soon move to an ensemble-only NWP (Numerical Weather Prediction) system. Ensemble forecasts can be used in a number of ways, but fundamentally they provide a probabilistic picture of the weather forecast which includes a most likely outcome but also information on the confidence, uncertainty or risks associated with forecast outcomes. In order to pull through the full benefits of this information it is important to communicate this information effectively to as many users as possible so that they can make appropriate risk-based decisions. There is a widely-held belief that people will find probabilistic information hard to understand or make use of – “People don’t understand probabilities” – which provides a significant obstacle to communicating it.
This challenge for ensemble forecasts has long been recognised and there has been extensive research conducted into effective communication and people’s understanding of such forecasts. This paper offers a review of that research to help guide future communications of forecasts. The overwhelming and consistent conclusion found in the literature is that people do understand the probabilistic information and make better decisions when presented with it, provided that the information is presented appropriately.
The literature review provides strong support for communicating probabilistic information to forecast users, including the general public. It does not support the idea that people’s understanding should be a barrier to communicating such information. While not every single person will understand or take full advantage of the additional information, most people will benefit and make better decisions as a result.
How to cite: Mylne, K. and Roberts, H.: “People don’t understand probabilities” – or do they? , EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-141, https://doi.org/10.5194/ems2024-141, 2024.
Ensemble forecasting has become the standard approach to represent uncertainties and produce multiple scenarios. Ensembles therefore provide rich information, but it often remains challenging to identify the useful signal for decision-making.
Ensemble information can be synthesized using clustering algorithms, which gather similar members in a limited number of meaningful scenarios. The current work builds on the strategy proposed by Mounier et al. (from EMS Annual Meeting 2023) for clustering convective-scale precipitation forecasts of the French Arome-EPS. We demonstrate that the approach can be applied to other meteorological fields, and we propose a way to adapt the design of the scenarios to users' needs.
In our methodology, scenarios are obtained by assigning ensemble members to climatological classes defined after applying a classification method to a meteorological database. Members assigned to the same class are grouped together to form a scenario. To reduce the dimensionality of the problem, the first step of the methodology performs a dimension reduction of input meteorological fields, using a convolutional autoencoder. The clustering algorithm is then applied to the meteorological training database projected in the autoencoder's latent space in order to issue climatological clusters.
To obtain scenarios that best fit the users' needs, the baseline autoencoder has been modified to account for users' variables.
This is achieved by adding a dense neural network in the decoder part of the architecture, that aims at predicting the user variable. We show how we can control the latent representation learned by the autoencoder by adding this supervised learning term to the reconstruction.
This user-oriented clustering approach will be illustrated for the case study of wind energy production. Thanks to a collaboration with the Compagnie Nationale du Rhône, the method has been trained considering a 5-year dataset of 100m wind analysis from the Arome model as meteorological inputs, and wind energy production values in north-western France as the user variable.
How to cite: Roubelat, F., Joly, B., Mounier, A., and Raynaud, L.: Ensemble weather forecasting scenarios tailored to users' needs, EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-310, https://doi.org/10.5194/ems2024-310, 2024.
Most end-users are keen to exploit ensemble forecast products for decision-making, such as workflow planning, human resources management, preventive actions. An improper use of a probabilistic product can lead to taking bad decisions, which may result in consistent economic loss. Stake-holders using ensemble forecast product may be classified into two categories: the ones who integrate probabilistic products into decision-making models or leverage human expertise to provide valuable information, the ones who are not able to exploit probabilistic information and need a translation to binary forecasts. One of the current objectives of Meteo-France is to enhance the current range of products and services offered by the public institution by integrating uncertainty, reliability and probabilistic forecast information in order to meet the specific needs of each end- user. Some new approaches are presented that aims at exploring ensemble based binaries forecasts synthetising vulnerability and uncertainty to build simple and adapted decision making tools. One has been experimented to categorize wave heats risk at one day ahead. Another one has been held for vineyards protection against spring frost and developed the use of costs and losts economic balance related to the event. Both have in common that the end user impact associated to the event is the issue that is followed. The ensemble forecasting part of information that could help to determine when mitigation decision has to be taken can then be disciminated. Standardizing these approaches is also investigated in order to adress a large number of commercial and institutional sectors that are left behind the use of probabilistic forecasting today.
How to cite: Joly, B., Ponzano, M., and Beau, I.: Approaches for making ensemble forecasting based services more suitable to end-users needs, EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-851, https://doi.org/10.5194/ems2024-851, 2024.
Modern weather forecasts and warnings are essential for billions of people around the globe, providing information about daily conditions as well as possible hazards. Despite the significance, they struggle with the challenge of in-complete accuracy in communication due to several factors, such as atmospheric unpredictability, data interpretation, uncertainty of composing forecasts, and forecast interpretation (Gill et al., 2008, p. 6). This study critically examines the prevalent deterministic approach in public meteorological communication, often presenting predictions as definite. The author explores the necessity for a shift towards uncertainty-informed impact-based forecasts, advocated by the World Meteorological Organization (WMO, 2021). Through multidisciplinary inquiry, this research reveals deterministic forecasts' limitations in conveying uncertainty sources, eroding public trust when forecasts do not align with reality, subsequently hindering decision-making (Burgeno and Joslyn, 2023). In contrast, probabilistic forecasts with transparent uncertainty margins might offer a more nuanced depiction of atmospheric variability, fostering public confidence and reducing bias adjustments. A multidisciplinary approach from cognitive studies, meteorology, and geography is employed to understand how communication induces behavioural responses to ambient and severe weather.
Embracing uncertainty not only enhances forecast accuracy but also facilitates informed prioritisation of precautionary measures, bolstering societal resilience to weather-related hazards. Moreover, integrating uncertainty communication into impact-based forecasts informs broader activities such as transportation planning and emergency preparedness. This is especially important in the case of unlikely yet high-disruption events, as well as highly likely yet limited-disruption events.
The current overview of forecast communication practices from meteorological institutes and agencies across Europe is presented and scrutinised against the existing theoretical basis. Accessible strategies for communication advancement are discussed, and possible future solutions in personalised forecasting are theorised. From a wider perspective, this research aims to guide informed activity in space and in relation to transport choices, ad-hoc plan adjustment, and protective action towards self and personal property.
How to cite: Bukowski, F.: From deterministic forecasts to probabilistic impact-based forecasts in meteorology: Theory and practice, EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-341, https://doi.org/10.5194/ems2024-341, 2024.
Ensemble Numerical Weather Prediction (NWP) systems have now been used operationally for over 30 years, mainly as a supplement to single realisation “deterministic” forecasts based on a higher resolution NWP model. Ensembles sample a range of possible outcomes and allow the estimation of probabilities and uncertainties in the forecast. Recognising the greater forecast skill of ensemble forecasts, the Met Office is moving towards a completely ensemble-based NWP system with no separate higher resolution deterministic forecasts, so users and developers will need to consider how these ensemble-based forecasts can best be used to cover the wide variety of applications and use cases. Ensembles offer a wealth of information which can be used in many different ways, from picking preferred or representative members to basing forecasts on a full probability distribution; forecasts may also be presented in many ways, either deterministically or taking account of the forecast uncertainty, depending on the use case and the decisions to be supported. In order to help users get the most benefit from ensembles, we present a framework of “classes of use cases” that classifies different user requirements of an NWP system and describe how these can be met. Which class is most appropriate for a particular user and application will depend on the nature of the decision(s) to be taken, the vulnerability of the user to weather impacts and the risk appetite of the user. The use cases are not fundamentally new, but by classifying them we aim to help users and service providers to identify the best approaches to address the needs of different users, and to identify synergies between the needs of different users. By classifying the use cases in this way we also highlight the implications for how we should be developing and processing ensemble data to deliver to these use cases, to ensure that the maximum benefits from ensemble NWP forecasts are pulled through for the whole of society.
How to cite: Mylne, K., Walters, D., Roberts, N., Barciela, R., Gray, M., Petch, J., Wells, O., Sachon, P., Skea, A., Davies, P., Willington, S., and Titley, H.: Ensemble Classes of Use Cases, EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-143, https://doi.org/10.5194/ems2024-143, 2024.
Climate services are on the cusp of becoming mainstream decision support tools. Many present accurate climate information within the bounds of scientific knowledge and technological development, yet some present climate information of limited “quality”, that is often “too good to be true”: i.e., scientific and technological constraints render it impossible to be as precise or as confident as suggested. This fidelity is rarely apparent when climate services are used to support decision making.
Alongside pursuing academic and technological advances, traditional efforts to counter this disconnect (between what climate scientists know to be the boundaries of what their work shows, and the way in which climate information is used in some decision making situations) has focussed on two groups of actors at two different moments in the production of climate services. Most established is training users how to interpret the climate information, occurring after it has been produced. More recently, climate scientists have begun to articulate guidelines of how to produce “high-quality” information, for other climate scientists to follow during the production of climate information.
We fear that demand for climate services will outpace the dissemination and use of good-practice standards. More positively, we believe the decisions taken to produce the data that underlies climate services could be made understandable for decision makers, making them active interrogators and providing a complementary route to counter the spread of meaningless climate information.
For the production of climate information, we use the metaphor of a jigsaw puzzle consisting of distinct, interlocking pieces. We illustrate the importance of user context in framing the puzzle, and for each of the constituent parts (e.g. timescales, spatial resolution, indicators) explain the production process and suggest guiding questions those commissioning climate services should ask to probe the fidelity of information presented in climate services.
How to cite: Pickard, S., Bojovic, D., Llabres, A., Perez Zanon, N., Garikoitz Munoz Solorzano, A., Gonzalez Romero, C., Duzenli, E., Lacima Nadolnik, A., Ruprich Robert, Y., Trascasa Castro, P., Urquiza, D., Cos, J., Lera St Clair, A., and Doblas Reyes, F.: Avoiding too good to be true: Guiding decision makers toward more meaningful climate information, EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-1063, https://doi.org/10.5194/ems2024-1063, 2024.
Digital Twins have become a buzzword often associated with innovation and transformation. The notion of digital twin includes a tight integration between models, data and decisions, which has resulted in a diverse and proliferating set of use cases, with applications across multiple areas of science, technology, and society. In the field of earth sciences, the digital twin for climate adaptation promises the potential to predict the impact of climate change with unprecedented reliability at regional and national scales. However, despite the opportunities and challenges it may bring to the climate adaptation community, researchers and practitioners working in the climate adaptation field may not be aware of such a digital twin and the added value of the information it can provide, and may even hold unrealistic expectations for purposes it can be applied to. To assess the awareness and expectations from the climate adaptation digital twin, a survey was launched targeted to adaptation researchers and practitioners. The analysis of results shows that the community sample reached with the survey (25 answers) is generally unaware of digital twins, let alone the one for adaptation to climate change, except for some academics involved in research projects of technical nature. The increase in the spatial and temporal resolution of climate and impact models, together with the simulation of different climate adaptation options to develop better regional and local adaptation strategies are the main highlighted opportunities, although for many it remains unclear who will be able to access the data and how to handle it. We suggest that a more tailored engagement and communication strategy is needed to raise awareness on the climate adaptation digital twin, the opportunities it may provide to the climate adaptation community and to the general public, and the interaction modes available for the uptake of the information generated. This calls for a stronger involvement of other disciplines, including experts in communication and dissemination, social sciences and humanities, user experience and data visualisation, in this undertaking.
How to cite: Terrado, M., Baulenas, E., Versteeg, G., and Bojovic, D.: Awareness and expectations from the climate change adaptation digital twin, EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-601, https://doi.org/10.5194/ems2024-601, 2024.
Rapid urbanization and climate change have significantly altered the microclimate of northern cities, leading to unfavorable anomalies that threaten cities' sustainability and residents' comfort. Although it is crucial to address these anomalies, the impact of large public buildings on the microclimate needs to be addressed, with assessments of the surrounding morphology taking precedence. This study aims to fill this gap by conducting land surface temperature (LST) and geographic information systems (GIS) analyses in four northern cities: Fairbanks, Tromsø, Bergen, and Nadym. We use remote sensing techniques and GIS tools to assess the fragmentation of urban morphology and distinguish between large and small buildings to evaluate their impact on LST variations. Understanding the microclimate impacts of large public buildings is necessary to optimize spatial configurations and find a balance between vegetation and buildings to mitigate the effects of local heat islands.
The study highlights the importance of effective communication in translating scientific findings into practical ideas. To achieve this, we examine the potential of two innovative communication strategies, web-based geographic information system (GIS) applications and storytelling, to raise awareness of urban climate issues. WebGIS is a powerful tool that combines stories, visualizations, engagement tools, briefings, and surveys to present content in an easy-to-understand format for all users. This framework integrates impact analysis, risk assessment, and options exploration, allowing the assimilation of scenarios and the attribution of trends to changes in society and the land surface. WebGIS effectively communicates complex scientific results to a broad audience by combining model output with thematic city layers. Storytelling techniques complement WebGIS by providing narrative context that makes scientific data more accessible and relevant to non-experts. Framing scientific evidence into compelling narratives promotes understanding, empathy, and action by highlighting the human dimensions of urban problems. In conclusion, effectively disseminating science is crucial for solving urbanization's problems. By utilizing the power of WebGIS and storytelling, we can make complex scientific data accessible and meaningful, facilitating informed decision-making and driving positive change in urban environments.
How to cite: Miles, V., Ezau, I., and Pettersson, L.: Cold City Microclimates: Enhancing Understanding and Communication, EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-961, https://doi.org/10.5194/ems2024-961, 2024.
