Map-Based Ensemble Forecasts for Maritime Operations: An Interactive Usability Assessment with Decision Scenarios

Jelmer Jeuring aNorwegian Meteorological Institute, Bergen, Norway

Search for other papers by Jelmer Jeuring in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0001-7530-7160
,
Eirik Mikal Samuelsen bNorwegian Meteorological Institute, Tromsø, Norway

Search for other papers by Eirik Mikal Samuelsen in
Current site
Google Scholar
PubMed
Close
,
Machiel Lamers cWageningen University and Research, Wageningen, Netherlands

Search for other papers by Machiel Lamers in
Current site
Google Scholar
PubMed
Close
,
Malte Müller dNorwegian Meteorological Institute, Oslo, Norway

Search for other papers by Malte Müller in
Current site
Google Scholar
PubMed
Close
,
Bjørn Åge Hjøllo eNAVTOR, Bergen, Norway

Search for other papers by Bjørn Åge Hjøllo in
Current site
Google Scholar
PubMed
Close
,
Laurent Bertino fNansen Environmental and Remote Sensing Centre, Bergen, Norway

Search for other papers by Laurent Bertino in
Current site
Google Scholar
PubMed
Close
, and
Berit Hagen aNorwegian Meteorological Institute, Bergen, Norway

Search for other papers by Berit Hagen in
Current site
Google Scholar
PubMed
Close
Open access

Abstract

Previous research indicates that forecast uncertainty can, in certain formats and decision contexts, provide actionable insights that help users in their decision-making. However, how to best disseminate forecast uncertainty, which factors affect successful uptake, and how forecast uncertainty transforms into better decision-making remains an ongoing topic for discussion in both academic and operational contexts. Interpreting and using visualizations of forecast uncertainty are not straightforward, and choosing how to represent uncertainty in forecast products should be dependent on the specific audience in mind. We present findings from an interactive stakeholder workshop that aimed to advance context-based insights on the usability of graphical representations of forecast uncertainty in the field of maritime operations. The workshop involved participants from various maritime sectors, including cruise tourism, fisheries, government, private forecast service providers, and research/academia. Geographically situated in Norway, the workshop employed sea spray icing as a use case for various decision scenario exercises, using both fixed probability and fixed threshold formats, supplemented with temporal ensemble diagrams. Accumulated operational expertise and characteristics of the forecast information itself, such as color coding and different forms of forecast uncertainty visualizations, were found to affect perceptions of decision-making quality. Findings can inform codesign processes of translating ensemble forecasts into usable and useful public and commercial forecast information services. The collaborative nature of the workshop facilitated knowledge sharing and coproduction between forecast providers and users. Overall, the study highlights the importance of incorporating methodological approaches that consider the complex and dynamic operational contexts of ensemble-based forecast provision, communication, and use.

Significance Statement

We wanted to understand how maps showing uncertainty in weather forecasts can help maritime users in their operational decisions. We organized a workshop with Norwegian maritime stakeholders and forecasters, who interpreted maps that combined layers of maritime operational activities and the likelihood of sea spray icing (an important hazard for ships operating on higher latitudes). The results show that contextual knowledge, and the use visual formats such as traffic light colors may help users to understand the maps. The results will help to better communicate weather forecasts to maritime users and gives suggestions about how to involve users in codesigning forecast products. Follow-up research could use our approach to investigate other hazardous conditions, such as wind, waves and sea ice.

© 2024 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Jelmer Jeuring, jelmer.jeuring@met.no

Abstract

Previous research indicates that forecast uncertainty can, in certain formats and decision contexts, provide actionable insights that help users in their decision-making. However, how to best disseminate forecast uncertainty, which factors affect successful uptake, and how forecast uncertainty transforms into better decision-making remains an ongoing topic for discussion in both academic and operational contexts. Interpreting and using visualizations of forecast uncertainty are not straightforward, and choosing how to represent uncertainty in forecast products should be dependent on the specific audience in mind. We present findings from an interactive stakeholder workshop that aimed to advance context-based insights on the usability of graphical representations of forecast uncertainty in the field of maritime operations. The workshop involved participants from various maritime sectors, including cruise tourism, fisheries, government, private forecast service providers, and research/academia. Geographically situated in Norway, the workshop employed sea spray icing as a use case for various decision scenario exercises, using both fixed probability and fixed threshold formats, supplemented with temporal ensemble diagrams. Accumulated operational expertise and characteristics of the forecast information itself, such as color coding and different forms of forecast uncertainty visualizations, were found to affect perceptions of decision-making quality. Findings can inform codesign processes of translating ensemble forecasts into usable and useful public and commercial forecast information services. The collaborative nature of the workshop facilitated knowledge sharing and coproduction between forecast providers and users. Overall, the study highlights the importance of incorporating methodological approaches that consider the complex and dynamic operational contexts of ensemble-based forecast provision, communication, and use.

Significance Statement

We wanted to understand how maps showing uncertainty in weather forecasts can help maritime users in their operational decisions. We organized a workshop with Norwegian maritime stakeholders and forecasters, who interpreted maps that combined layers of maritime operational activities and the likelihood of sea spray icing (an important hazard for ships operating on higher latitudes). The results show that contextual knowledge, and the use visual formats such as traffic light colors may help users to understand the maps. The results will help to better communicate weather forecasts to maritime users and gives suggestions about how to involve users in codesigning forecast products. Follow-up research could use our approach to investigate other hazardous conditions, such as wind, waves and sea ice.

© 2024 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Jelmer Jeuring, jelmer.jeuring@met.no

1. Introduction

Ensemble prediction systems (EPSs) have been available for over three decades in operational forecasting and have gradually become more commonplace in the forecaster portfolio (Novak et al. 2008). Partly due to the increased use of ensemble-based forecast models, forecast uncertainty has become a more integrative part of weather forecast information products that are communicated to various audiences (Taylor et al. 2018). Yet, how to best disseminate forecast uncertainty, which factors affect successful uptake, and how forecast uncertainty transforms into better decision-making remain ongoing topics for discussion in both academic and operational contexts. Context-sensitive insights, derived from naturalistic research settings, about understanding and use of forecast uncertainty can help to corroborate the validity of quantitative evidence from controlled experimental settings (Stephens et al. 2019). In this paper we present findings from an interactive stakeholder workshop that employed a mixed methods approach (Mulder et al. 2017) consisting of individual paper-based exercises and group discussions around graphical representations of weather forecast uncertainty in the context of maritime operations.

Interpretation and understanding of visualizations of forecast uncertainty are far from straightforward. More generally, reasoning about uncertainty expressions such as probabilities suffers from various biases to the extent that some have argued that systematic flaws in human logic hinders people from correctly interpreting uncertainties (Padilla et al. 2021). Stephens et al. (2019) conclude, similar to Savelli and Joslyn (2012), that most people infer uncertainty also in a deterministic forecast. Therefore, they argue, it would be inappropriate to only communicate a forecast that implies an absence of uncertainty. On the other hand, Joslyn and Savelli (2021) found that deterministic construal error (“a tendency to interpret the image as representing some deterministic quantity,” p. 1) acts as a key psychological bias that hampers correct understanding of uncertainty representations. A study on interpretations of various forecast uncertainty formats that are used in the Norwegian weather service “yr.no” (based on a collaboration between the Norwegian Meteorological Institute and the Norwegian Broadcasting Corporation NRK) showed that uncertainty information was interpreted in a rather pragmatic way, since people struggled to meaningfully fit it to their context and perceptions (Sivle et al. 2014). Other factors that have been found to influence decision-making quality using uncertainty information include people’s numerical literacy, as well as characteristics of the forecast information itself, such as the use of cones, colors, and error bars (Grounds and Joslyn 2018; Ruginski et al. 2016). Stephens et al. (2019) found that adding certain graphical elements (e.g., icons) in addition to numerical representations may support understanding.

Despite the cognitive challenges around assessing uncertainties, various studies show that forecast uncertainty as an information characteristic can in certain formats and in certain decision contexts provide actionable insights that help users in their decision-making, arguably complementing more conventional deterministic forecast information (Grounds and Joslyn 2018; Morss et al. 2010; Sivle et al. 2014). However, research that has explored understanding of numerical and visual representations of forecast uncertainty show mixed results in terms of optimal formats (Ripberger et al. 2022). The extensive scholarship on visualization of geospatial information uncertainty (MacEachren et al. 2005) was found to be relevant within the weather communication context, with various studies exploring cartographic uncertainty (Kinkeldey et al. 2014; Klockow-McClain et al. 2020; Ruginski et al. 2016). An experimental study on cartographic framing of tornado risk found that risk perceptions relate to how the likelihood of a hazard is graphically represented in terms of proximity, warning-level boundaries, and colors (Klockow-McClain et al. 2020). In a study on people’s interpretation of flood forecasts and warning products in the United States, Carr et al. (2016) found that while graphical clarity of map-based information is essential, a combination of graphic and textual elements was preferred, as each provides a different layer of actionable information. However, the use of colors may challenge some people, for example, with visual impairments or color blindness (Engeset et al. 2022). From the wide range of research on forecast uncertainty visualizations few generic conclusions can be drawn beyond that there is no one-size-fits-all solution (Ripberger et al. 2022). Instead, choosing how to represent uncertainty in forecast products should be dependent on the specific audience (Mulder et al. 2020; Ripberger et al. 2022). Similarly, Joslyn and Savelli (2021) state that the relative value of different forms of visual representations depends on the decision context at hand.

The challenge to account for forecast uncertainty as a usable quality of information that benefits end users is embedded in the ongoing need to get more in-depth insights in the use of specific operational products that are provided by national meteorological and hydrological services (NMHSs) and that are targeted at specific audiences. Such situated insights can be gained from naturalistic settings that facilitate collaborative learning and knowledge sharing across providers and users of forecast information services, while accounting for the complex and dynamic operational contexts of forecast provision, communication, and use (Blair et al. 2022a). Similar needs arise from the increasing interest among private actors that provide maritime forecast services to use ensemble predictions as input to weather routing forecasting, to support vessel performance optimization and vessel crossing guidance (e.g., the North Atlantic) and to navigate safely and efficiently. Here, ensemble predictions are increasingly used in new and more complex calculations and situated decision contexts (Akkermann et al. 2020). Bringing together providers and users in an interactive, naturalistic setting has been done in several formats, for example, in testbeds. A testbed can be defined as a “physical space and a research framework that fosters collaboration” (Calhoun et al. 2021, p. 2230) to test and evaluate tools, procedures, and information services. Built around principles of coproduction to “support decision making in the user’s particular application” (Parker et al. 2022, p. 365), testbeds have been applied as real-time forecasting exercises, for example, in the United States (Obermeier et al. 2022) and in Africa (Parker et al. 2022).

The scholarship in usability studies has advanced the conceptual basis for research on how producers and users perceive and understand weather and climate information, as well as providing theoretical and practical guidelines for codesign and codevelopment of usable information services. Under definitions like “the degree to which something is able or fit to be used” (Stephenson 2010, cited in Tan et al. 2020) and “the extent to which a system, product, or service can be used by specified users to achieve specified goals with effectiveness, efficiency, and satisfaction in a specific context of use” (Argyle et al. 2017, p. 373), usability as a concept is originally rooted in the field of website and software design. Conceptualizations of usability have been tailored to and operationalized in different fields, with applications in a wide range of user-centered (technological) service contexts. For example, Lemos et al. (2012) discern between “fit,” “interplay,” and “interaction” as three quality dimensions of iterative provider–user dynamics through which climate information can transform from being useful to usable. Along that line, usability research methods applied in the field of weather and climate services aim to facilitate preoperational testing, involving end-user perspectives and closing gaps between research and operations (Vincent et al. 2018; Fujisaki-Manome et al. 2019; Matthews et al. 2023).

The stakeholder workshop from which findings are presented in this paper was set up based on this rationale of bringing together providers and users and build on the variety of expertise and perspectives about provision, communication, and use of ensemble-based forecast information. The workshop was organized as part of the Accounting for Forecast Uncertainties in Communicating Sea Ice and Weather Information in the Arctic (FOCUS) project, funded by the Norwegian Research Council and running from 2020 until 2024. In the transdisciplinary project, public and private actors, as well as academia, collaborate to codevelop science-based solutions that transgress the public and private forecast services domains. The aim of the FOCUS project is to improve the quality of coupled (ocean–atmosphere–ice) ensemble forecast models and advance their usability in public and commercial forecast information services for maritime operators. Within this more general scope of coupling of ocean–atmosphere–ice ensemble forecast models, the workshop used sea spray icing as a use case to explore the usability of map-based representations of forecast uncertainty. A more detailed description of sea spray icing, the challenges related to its forecasting, communication and its potential impacts, and the specific case-study context is outlined in the following section. Our main interest for the workshop was to get a better understanding about 1) how providers and users of maritime forecast information interpret map-based forecast uncertainty, and 2) which aspects of map-based forecast graphics are found useful or difficult to implement in certain decision contexts. In addition, an important process-based objective was also to use the exercises as an instrument to facilitate a coproduction process between providers and users that are involved in the information value chain of (Norwegian) maritime forecast information (Jeuring et al. 2020; Bremer et al. 2022; Blair et al. 2022a). As such, the sharing of knowledge and experience during the workshop directly fed into the broader research and development objectives of the FOCUS project.

Study context

Our study is geographically situated in Norway and employs as research context the higher-latitude waters along the Norwegian coast and the northern Atlantic Ocean. In these waters, a potential hazard for vessels and other maritime infrastructure is accumulation of ice on infrastructure through sea spray. Conditions favorable for sea spray icing are typically a combination of strong winds, high waves, and temperatures below the freezing point of seawater (i.e., approximately −2°C), in combination with various parameters related to the vessel (including relative heading and building structure) (Samuelsen and Graversen 2019). Spray-icing conditions occur regularly, especially in the winter months. Ship observations from 1976 to 2006 show icing occurrences in these waters typically from the end of September to the middle of May (Samuelsen and Graversen 2019). Samuelsen and Graversen (2019) show that particularly a synoptic situation with a low pressure anomaly located in the Barents Sea, and high pressure anomaly in the Greenland Sea together with low temperature anomalies at around 1200–1400 m above sea level (850 hPa) are favorable conditions for icing to occur in this area. In some regions (e.g., west of Svalbard; see Fig. 1) icing conditions occur above 50% of the time from November until April (Naseri and Samuelsen 2019). For moderate or severe icing conditions the frequency of occurrence is above 20% from December until March in the same area.

Fig. 1.
Fig. 1.

MET Norway area of responsibility for navigational warnings (including METAREA XIX: areas coded A1–F3 and the named high-latitude areas up to the North Pole). Coastal water along the Norwegian mainland include areas from Indre Skagerak in the south to Kildinbanken in the north. Source: Norwegian Meteorological Institute (2020).

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Consequences of ice accumulation on vessels include decreased floating capabilities, as well as increased instability due to the rise of center of gravity. This can result in ships tipping over and sinking. Other operational impacts include slippery decks, ladders, and handrails as well as the disabling of lifeboats and constrained radio communication due to frozen antennas. While some vessels may have deicing capabilities, using these systems strongly raises energy consumption and thereby the environmental impact of its operations.

Other potentially hazardous conditions in the area include high wind speed, high waves, and polar lows, as well as the presence of sea ice (Noer et al. 2011). The Norwegian Meteorological Institute (MET Norway) issues regular warnings for spray icing in the waters around Norway and the ocean area named METAREA XIX (Fig. 1). For waters along the coast of mainland Norway (named areas from Indre Skagerak in the south to Kildinbanken in the north), a spray-icing warning is issued whenever moderate or severe icing is expected. For METAREA XIX (areas coded A1–F3 and the named high-latitude areas up to the North Pole) the threshold for issuing warnings is set to severe icing. The logic behind the differentiation of thresholds for sending out warnings in these two areas is to avoid sending out too many warnings in areas with high frequency of moderate-icing events. Also, the open ocean area is in wintertime more trafficked by larger vessels in which the consequences of moderate icing are lower compared to the consequences of moderate icing on smaller vessels. However, closer to mainland Norway there is more year-round traffic by smaller vessels, including fishing, on which moderate icing has a higher impact. In the forecast for METAREA XIX, warnings are also issued for polar lows and winds meeting or exceeding gale force (17.2 m s−1; Beaufort scale number 8). While our study uses coastal northern Norway and the northern Atlantic as a context, similar conditions may occur in other (high latitude) waters across the world, in both the Northern and Southern Hemispheres (Jeuring and Knol-Kauffman 2019; Stewart et al. 2020). Moreover, even though this study is scoped around probabilistic information of sea spray icing in the North Atlantic, many of the principles are valid for other parameters as well, such as the probability for wave or wind to exceed given thresholds.

2. Methods

The workshop gathered a total of 17 participants and was conducted in May 2022. Participants were invited with the aim to gather a broad range of expertise around the use and provision of weather, ocean, and sea ice information in the Norwegian maritime operations context. Participants represented various maritime sectors as well as weather information service providers, with professional roles spread among both the public and private sectors. Representatives from the FOCUS project, including academic researchers with topic expertise, were included among the workshop participants (Table 1). Some participants (e.g., captains) can be categorized as end users of forecast information, while other participants (e.g., public weather forecasters) are typical providers. In addition, several participants (e.g., nautical advisors, itinerary planners, and navigational service providers) had professional roles that deal both with the provision and the communication of weather information. Their relation to weather forecast data or information transgresses thereby the provider–end user duality, as they are intermediary translators of weather information for “downstream” end users in a more understandable or tailored format (Jeuring and Knol-Kauffman 2019).

Table 1.

Sectors represented by workshop participants.

Table 1.

Maritime stakeholders brought in an array of operational expertise from a range of contexts including cruise tourism and fisheries, as well as governmental/state maritime actors such as emergency response and coast guard representatives. Experience with maritime operational/tactical decision-making was on average well over 10 years. Two participants had about 30 years of experience. Public forecast providers included operational forecasters from MET Norway as well as researchers involved in service development. Except for one female participant, all workshop participants were male. Average age was 45 years, with the youngest participant being 32 years old and the oldest being 57 years old. The participants signed a consent form at the start of the workshop.

The exercise format was broadly inspired by methods developed by Mulder et al. (2017) in the context of volcanic ash probability visualizations and was modified to suit the maritime decision context. The workshop exercises were developed by the first author in close collaboration with two operational forecasters from the Norwegian Meteorological Institute (one of whom is part of the author team). Feedback was collected from the rest of the author team (academic and operational experts in fields ranging from knowledge coproduction to ship routing decision support) during pilot testing of preliminary versions of the exercises. Testing iterations consisted of sharing intermediate versions of the workshop exercises together with a request for filling out the exercises. Feedback was received in written form and during online follow-up meetings. These iterations significantly improved the final version of the exercises.

During the workshop, participants were asked about their current familiarity with, and their trust in, a selection of currently operational ensemble-based forecast products that in different forms communicate forecast uncertainty. This gave us a general impression about how aware participants are of uncertainty characteristics in forecast information, but also helped in getting a contextual insight into how participants perceive vessel-icing warning information, relative to other representations of forecast uncertainty. Familiarity was measured with the question “How familiar are you with the following products?,” which was answered on a bipolar scale (not at all familiar–extremely familiar). Trustworthiness was measured with the question “How trustworthy do you find the following products?” (not at all trustworthy–extremely trustworthy). Answers on both the familiarity and trustworthiness were graded from 1 to 10, with higher gradings indicating higher levels of familiarity/trustworthiness.

The included forecast products assessed in this part of the workshop were: the MET Norway vessel-icing warning, the ECMWF Marinogram, the MET Norway polar low warning, and the Windy multimodel location forecast (Figs. 25, respectively). The MET Norway vessel-icing warning is based on the icing rate of freezing sea spray generated by ship–wave interaction. The warnings are based on the ship-icing model of Samuelsen (2017), with some slight modifications. The traffic light visualization shows light, moderate or severe icing, that is, various thresholds of icing rates developed from the frequency and severity of observed icing events [see Samuelsen (2018) for more details on threshold levels]. Vessel icing warnings are available through various online channels, including MET Norway’s website (https://www.yr.no) and Barentswatch (https://www.barentswatch.no), an environmental information portal for maritime operations. The ECMWF Marinogram (also called Wave ENSgram, available at https://charts.ecmwf.int/) shows the temporal evolution at a specific location for meteorological parameters and its uncertainties in a set of diagrams. Uncertainties are visualized by a box-and-whisker plot, including information on the 10th, 25th, 75th, and 90th percentiles; the median; and the mean of the ECMWF ensemble members. The MET Norway polar low warning is based on a cyclone detection and tracking method described in Landgren et al. (2019). Individual polar low tracks are calculated for each ensemble member forecast and summarized as a “track probability” on a map, which is available through channels similar to those used for the MET Norway vessel-icing warnings. In addition, the forecaster’s knowledge and experience about the uncertainty in the forecast product is a subjective assessment that informs the track visualization. The Windy forecast website (https://www.windy.com) includes, as one of the visualization options, a multimodel location-based forecast. This visualization shows different global prediction systems and visualizes each model output for wind and wave conditions by an individual symbol at a specified location, as well as showing a comparative timeline for each model. While the four products are not representative of the portfolio of forecast information that is available to maritime operators, they were selected based on their relevance for maritime operations in the project’s research context. As such the selection suited the purpose of our study.

Fig. 2.
Fig. 2.

MET Norway vessel icing. Forecast uncertainty is shown as light (yellow), moderate (orange), or severe (red) icing for various thresholds of icing rates.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 3.
Fig. 3.

ECMWF Wave ENSgram. Forecast uncertainty is visualized by a box-and-whisker plot including information on the 10th, 25th, 75th, and 90th percentiles; median; and mean of the ECMWF ensemble members.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 4.
Fig. 4.

MET Norway polar low warning. Forecast uncertainty is visualized as a “track probability” on a map, based on a summary of ensemble forecast members.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 5.
Fig. 5.

Windy multimodel location forecast. Forecast uncertainty is visualized by location-based forecasts from different global prediction systems, with model output for wind and wave conditions visualized by an individual symbol at a specified location. Multimodel forecasts are also shown along a comparative timeline.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

The workshop participants were also presented a set of decision scenarios with map-based representations of an ensemble forecast for sea-spray-icing conditions. Participants had to make a safety assessment for three different tactical navigation contexts: a routing scenario and two location-based scenarios. Two forms of map-based representations for forecast uncertainty were used in each scenario type (thus resulting in six scenarios in total, Table 2). The first format [the fixed probability representation (FP)] used preset probabilities for a spray-icing index of three different threshold levels (light, moderate, severe). The second format [the fixed threshold representation (FT)] used probabilities for spray-icing conditions for one preset threshold (moderate icing). In the routing scenarios, a one-directional route was plotted on both the FP maps and the FT maps (Figs. 6 and 7, respectively). The route was divided in four parts, with levels of risk for being impacted by vessel icing varying between segments. Part A represents a low risk, parts B and C represent medium risk, and part D a high risk. In both the FP and FT setup, participants made for each route segment a safety assessment, as well as an assessment of confidence about their own respective decisions. In the location-based scenarios (FP in Fig. 8 and FT in Fig. 9), participants were invited to imagine a maritime operation that is typical for their professional context. In both the FP and FT setup, participants made for four locations (again, each representing a varying level of risk) a safety assessment, as well as an assessment of confidence about their own respective decisions. In the third set of scenarios, an ensemble “spaghetti” diagram complemented the FP and FT maps, thereby adding temporal uncertainty as a factor for safety assessments. Contrary to the previous decision scenarios, instead of asking for spatially defined safety assessments, participants were here asked to assign temporal windows for safe operations. The FP scenario used a 100% probability level, with approval for operations under light spray-icing conditions (Fig. 10). Similarly, the FT scenario used a location with over 90% probability for moderate vessel-icing conditions (Fig. 11). The diagrams also included threshold lines that indicate severity of spray icing. Note that the same ensemble spaghetti plot was used in both scenarios to focus participants’ attention toward the different map formats, and to avoid information overload. In addition to the forecast graphs, contextual information available to the participants included wind direction, wind speed, and a range of vessel characteristics. It was assumed that the forecast was valid for the entire duration of the decision task at hand.

Table 2.

Overview of decision scenarios and representations of forecast uncertainty used.

Table 2.
Fig. 6.
Fig. 6.

Routing scenario 1 with varying threshold levels for vessel icing (FP).

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 7.
Fig. 7.

Routing scenario 2 with varying probabilities for moderate vessel icing (FT).

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 8.
Fig. 8.

Location-based operation scenario 1 (FP) with varying threshold levels for vessel icing.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 9.
Fig. 9.

Location-based operation scenario 2 (FT) with varying probabilities for vessel icing.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 10.
Fig. 10.

Temporal window task scenario 1 (FP) with varying threshold levels for vessel icing. Identical spaghetti diagrams were used for both tasks (see Fig. 11, below).

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 11.
Fig. 11.

Temporal window task scenario 2 (FT) with varying probabilities for vessel icing. Identical spaghetti diagrams were used for both tasks (see Fig. 10).

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

In each scenario, participants were asked a set of evaluative questions. First, they were asked to make safety calls for the different operational contexts (e.g., “All other meteorological and marine conditions equal, would you consider each part A–D of the route safe enough for execution?”; answer options yes or no). Then, participants had to report confidence for each of their go/no go decisions (“How confident are you in your decision?”) on a 1–10 scale with higher ratings indicating higher levels of confidence. Last, in open-ended form, participants could share their perceived decision uncertainties (“Is there any information in the map that you were particularly uncertain about?”) and decision rationales (“What information in the map influenced your judgements of the different parts?”) for each operational context. The safety calls and decision confidence responses were quantitatively summarized. Answers to open-ended questions provided across all six scenarios were thematically categorized based on the stepwise procedure outlined by Boeije (2009). The thematic analysis involved several rounds of summarizing, rereading and coding through which a total of seven themes emerged. Background information of participants (e.g., current affiliation, years of experience with navigational decision-making, age, gender) were also collected. The results presented in this paper aggregate responses from all participants that provided sufficient data (13 of 17 participants).

Three groups of 5–6 persons were formed, with each group sitting around their own table. The exercises were provided in paper format and were filled out individually by each participant in three rounds. Reflective discussions of about 5–10 min took place in between each round, as well as in plenary at the end of the exercises. During the discussions, participants shared their experiences and answers with others at their table. Notes were taken by FOCUS project participants sitting at each table. The quantitative findings from the exercises are complemented by illustrative notes from these reflective discussions.

3. Results

a. Familiarity with and trust in ensemble-based forecast information

For each ensemble-based forecast product, average scores for perceived familiarity and perceived trustworthiness were calculated based on the ratings of 10 of 17 respondents (Fig. 12). The findings show that respondents vary especially in their familiarity with the respective forecast products (scores ranging from 5.7 for the Windy product to 7.3 for the polar low warning product), while their levels of trust were relatively similar (ranging from 5.9 for the Windy product to 6.7 for the polar low warning product). The Windy multimodel forecast received the lowest scores on both trustworthiness and familiarity. The MET Norway polar low warning product was best known and also got the highest ratings of trustworthiness. From the findings, a positive relation between familiarity with, and trust in, the four different forecast products can be discerned, though this was not tested for statistical significance.

Fig. 12.
Fig. 12.

Average scores for trustworthiness and familiarity for four ensemble-based forecast products (1–10 scale, with 1 being the lowest score and 10 being the highest score).

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Group discussions revealed that most participants who work in a maritime context have some freedom in choosing which forecast information to use, signified by a pragmatic approach toward a preference for forecast information that “does the job.” In this context, the Windy forecast app and website was often mentioned as intuitively useful (and easily available) (“Used by navigators in the fleet”) but was also criticized for not giving the best forecasts. As noted by one participant, Windy “Gives an idea of how predictable the general weather situation is, [but I w]ould not use it alone in a go/no go decision.” However, also the polar low warning product that was rated highest on trustworthiness and familiarity was criticized, for example, by a representative from the fishery sector who noted that “often too late warning[s]” are given. Overall, participants’ answers reflect that their own experience with using a product in their profession affects how they value different forecast products.

b. Safety assessments and decision confidence

We first discuss how participants assess safety and confidence across the first two forecast scenarios (routing and operation zones). Then, we discuss for all three scenarios which actionable information is distilled from the maps and diagrams and which decision uncertainties participants experience when digesting the forecast information.

1) Part 1: Routing scenarios

Assessment scores for safe execution of each part of the depicted routes are shown in Fig. 13. An assessment pattern is visible in both FP and FT scenarios in which the low-risk segments (A) are perceived as safe by all participants and, on average, participants have a high confidence in their assessments (Fig. 14). The medium-risk parts (B and C) are also seen as safe areas for navigation by most participants, but decision confidence ratings are lower, which is especially the case for part C in the FP scenario.

Fig. 13.
Fig. 13.

Operational safety assessment for routing scenarios 1 (FP) and 2 (FT), with percentage of participants assessing each part of the indicated route as safe enough for execution.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 14.
Fig. 14.

Decision confidence for safety assessments for routing scenarios 1 (FP) and 2 (FT). Each route part indicated on the map is rated on a 1–10 scale (1: no at all confident; 10: extremely confident).

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

In turn, for the high-risk segments (D) in both scenario 1 and scenario 2, most respondents agree on this stretch not being safe enough for navigation under the given forecast. The level of confidence in the assessment of the high-risk segment not being safe for navigation was rated lowest of all parts in the FP scenario, which reflects that participants increasingly infer uncertainty in their assessments of this decision scenario. Some participants referred to uncertainty about the possible presence of sea ice (“Spray condition near the ice edge. Is that accurate?”), while another participant added that “sea-ice moves fast in these wind conditions.” In turn, the decision confidence of the high-risk segment (D) in the FT scenario was rated at a similar level as the medium-risk assessments. The coastal context and the relatively small area with high icing probability in the FT scenario was mentioned to influence the level of confidence: “No problem to go through the icing if its close to the shore. Then you select a route that is closer to shore to be safe.”

2) Part 2: Location-based operations

The safety assessment scores in Fig. 15 show that the low-risk locations (A) were rated as safe for operations by most participants. The safety ratings for the medium-risk (B and C) locations were lower, especially in the FT scenario in which for location B only 25% of the participants assessed operations to be safe given the sea-spray-icing forecast. For both the medium-risk location (C) and the high-risk location (D), almost all participants assessed operations under the given forecast to be unsafe in both scenarios. The decision confidence (Fig. 16) across the operational locations ranges for both the FP and FT scenarios between average ratings of 6.1 (position B; FP) and 7.6 (position D; FT) on a 1–10 scale.

Fig. 15.
Fig. 15.

Safety assessment for location-based operation scenarios 1 (FP) and 2 (FT), with percentage of participants assessing each indicated position as safe enough for execution.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 16.
Fig. 16.

Average decision confidence ratings for location-based operation scenarios 1 (FP) and 2 (FT). Each position indicated on the map is rated on a 1–10 scale (1: no at all confident; 10: extremely confident).

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Qualitative answers and the discussion after this part of the exercises revealed that key factors influencing the operational safety assessments included the proximity to areas with varying forecast conditions (“how close to yellow or red”), as well as information about the wind conditions (“[Location] B: close to border light/moderate, but wind direction ok”). A forecast being able to accurately predict local conditions at the operational locations was mentioned most often as cause for uncertainty. Both proximity and accuracy factors reflected that operators tend to accept to a certain extent the possibility of being exposed to spray icing. However, it is the duration of its exposure and being able to keep control of this duration that is central for operational go/no go decisions. This is exemplified by notions about the time needed to end an operation, or to escape to a safer area: “How fast can you stop operation? Would wait for new forecast before starting operation.”

3) Part 3: Temporal operation windows

The accumulated drawings of safe weather-windows are plotted in Fig. 17 (FP scenario) and Fig. 18 (FT scenario). A similar pattern of assessments for safe operations over time can be observed across the two scenarios. Thus, even though the decision tasks have different threshold levels for safe operations, and the two scenarios have both different representations and levels of forecast uncertainty, the information derived from the graphics is translated into similar decision patterns. Indeed, the position of the forecast members plotted against the threshold line is reflected in the drawings. In addition, the yellow threshold line (moderate icing) was mentioned during the discussions as an intuitive graphical indicator for participants’ assessments: “Some model runs are over and the rest is close to operation limit.”

Fig. 17.
Fig. 17.

Assessment of safe temporal weather windows for scenario 1 (FP). Lighter or darker shading respectively indicates fewer or more participants assessing conditions to be safe for operations.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Fig. 18.
Fig. 18.

Assessment of safe temporal weather windows for scenario 2 (FT). Lighter or darker shading respectively indicates fewer or more participants assessing conditions to be safe for operations.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

To get more insight into why participants’ ended up with their particular drawings of safe weather-windows, they were asked to reflect on their decision rationales. Three factors emerged from the answers to open-ended questions (Fig. 19) that were found across the FT and FP scenarios. A “risk oriented” rationale refers to participants inferring (un)safety with a focus on the operational implications of the conditions that were forecast. The risk rationale was, for example, reflecting a concern about the likelihood to (not) being exposed to icing conditions. In turn, some participants used a “member-oriented” rationale. This reflects participants paying attention to trends of one or more ensemble members in the forecast diagram and the level of (un)certainty they derive from that trend over time. For instance, one participant noted for the period from 1500 to 2130 central European time (CET) Friday the relevance of the control member: “Short lead time: control member highly trustworthy.” A “threshold oriented” rationale relates to the pattern of ensemble members but is primarily focused on operational limits. For example, several participants noted that their safety assessments depended on whether ensemble members were exceeding the thresholds as set in the scenarios (light icing in FT, moderate icing in FP). This was exemplified by one participant who noted that “some model runs are over and the rest is close to operation limit” (time stamp from 0300 to 2300 CET Saturday). All three rationales were used in both the FP and FT scenarios. However, in the FP scenario participants’ concerns were mostly with operational thresholds being exceeded. The FT scenario was mostly inferred by risk-oriented choices, for example, the likelihood of being exposed to severe icing.

Fig. 19.
Fig. 19.

Frequencies of decision rationales used in assessing safe temporal weather windows for FP and FT scenarios.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

4) Assessments of actionable information and uncertainty factors

Thematic analysis of participants’ qualitative expressions about which information provided in each of the scenarios influenced their safety judgments (i.e., actionable information), as well as which information they were uncertain about (i.e., uncertainty factors) resulted in seven themes (Table 3).

Table 3.

Themes used to indicate actionable information (“What information in the map influenced your judgements of the different parts?”) and uncertainty factors (“Is there any information in the map that you were particularly uncertain about?”) with illustrative examples of responses.

Table 3.

Responses were summarized by type of forecast uncertainty representation (FP or FT). All themes of actionable information were mentioned across the FP and the FT scenarios (except “other”), while the overall number of actionable information aspects mentioned was highest in the FP scenarios (Fig. 20). This higher number was mainly derived from probability expressions and references to colors. Other often mentioned aspects refer to the exposure and severity of the conditions that were forecast and to circumstantial conditions. The latter theme includes, for example, geographical details derived from the maps, as well as people being able to infer how hyperlocal (e.g., inside a fjord) conditions will be, given the conditions that were forecast.

Fig. 20.
Fig. 20.

Number of times each factor of actionable information is mentioned across FP and FT decision scenarios.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

Similar to the analysis of actionable information, the analysis of uncertainty factors shows few differences between the FP and FT scenarios (Fig. 21). The two main aspects causing uncertainty are knowledge (or lack thereof) about local circumstantial conditions (“Wind speed inshore”), as well as dynamics in environmental conditions (“Sea ice move fast in these wind conditions”; “What way is the [weather] system moving”?) or the extent these dynamics are accurately reflected in the forecast (“average wind and the accuracy”). Colors were only mentioned once as a factor causing uncertainty. One participant mentioned that colorblindness affected noting color coding in warning systems.

Fig. 21.
Fig. 21.

Number of times each type of uncertainty factors is mentioned across FP and FT decision scenarios.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0076.1

4. Discussion

Previous research has shown that disseminating uncertainty explicitly in forecast information can help decision-making (Morss et al. 2010; Sivle et al. 2014; Grounds and Joslyn 2018). How uncertainty is inferred, and to what extent it adds actionable information is highly dependent on both forecast format, decision context and individual characteristics of the decision-maker. To that end, the results presented above give rise to various points for discussion.

a. Familiarity and trust

Ensemble-based forecasts have been around for several decades. However, the spread in uptake and use is arguably challenged by the complexity of communicating their probabilistic features. This may have implications for the pace by which familiarity with and trust in EPS forecasts is built up among its (potential) scope of users. The somewhat positive relation between familiarity with, and trust in, forecast products has also been found elsewhere (e.g., Car et al. 2012), while other research has found mixed results regarding the relation between trust in and familiarity with forecast products, or the direction of such a causal relation (e.g., Mulder et al. 2017; Losee and Joslyn 2018). The participants’ perceptions of familiarity with and trust in several examples of ensemble-based forecasts show a variation on the product level. From the perspective of maritime operators, a possible underlying factor that explains this variation pertains to the seemingly large relative freedom they have in choosing which forecast products to consult for their daily planning and operations. In other words, personal preferences and routines seem to matter.

Another relevant factor that affects familiarity, trust and uptake of forecast products pertains to the need to prove their usability in the operational context (Blair et al. 2022b), while accounting for the significant cognitive effort required from users to engage with novel forms of information (Padilla et al. 2021). Several workshop participants reported that they struggled with understanding the probabilistic information and translating it into deterministic decisions. From a group discussion at one of the tables it was noted that one “easily misread[s] the maps and side info if you are not used to it: takes practice.” A relevant challenge for complex forecast information such as EPS-based products therefore is to advance familiarity while investing time in trialing these products in practice (Fundel et al. 2019; Blair et al. 2022b). The evolution of familiarity and trust over time may therefore be relevant performance indicators to monitor for forecast service providers (Losee and Joslyn 2018). Such insights are particularly pertinent in the light of the emerging obligations for maritime operators to comply with regulations (e.g., the polar code) for safe and environmentally friendly navigation on higher latitudes (Müller et al. 2023). Possible tensions and paradoxes between pragmatic routines of forecast information use and formalized requirements to consult certain products need to be minimized as much as possible, which may take time in the case of complex or novel types of forecast products.

b. Safety assessments and decision confidence

Decisions in maritime operations are often taken under a certain degree of time constraint. The trade-off between the amount or complexity of forecast information that can be considered and how much time it takes to access, digest, and apply this information implies the importance of an interface that facilitates intuitive and efficient uptake of information. As was mentioned during one of the group discussions, “the probability-based map makes you want to see other conditions as well.” An integrated visualization interface with “all the information in one place” that allows operators to overlay maps with various data would therefore be preferable (Thompson et al. 2015; Mulder et al. 2017; Knol et al. 2018). Likewise, providing insight in how forecast uncertainty evolves over time across model runs may facilitate a more nuanced and probability-based identification of operational safety over time. A similar need for sensitivity to communicating the temporal evolution of weather conditions has recently been discussed in the context of weather warnings (Krocak et al. 2023). Further research is needed to give more conclusive insights about how integrated visualization formats suit different operational decision contexts.

Participants showed higher confidence in their safety assessments for low-risk areas compared to medium and high-risk areas. This may be an implication of the research design: the differences between the risk exposure, route segments and locations are relative, and not based on a quantified assessment of exposure. However, such an approach is supported by research concluding that visualization of uncertainty in a forecast product should adequately communicate the levels of risk associated with specific areas and relative to the decision scenarios at hand (e.g., Niclasen et al. 2010). Contextual tailoring of uncertainty as an information quality of high impact conditions will likely help users make informed decisions, and may improve confidence in operational safety assessments. For example, the use of EPS-based forecasts can be beneficial for maritime weather routing. When having 50 ensemble members for a given set of required weather parameters (e.g., wind, wave, swell, current), the conventional approach would be to run 50 routings and then cluster the outcomes. Optimal routing suggestions should then also consider certain safety settings such as avoiding probability of waves higher than 10 m, or sea-spray-icing indexes based on vessel characteristics. As with the exercises in our study, thresholds can typically be visualized with traffic lights presentation giving a probability of exceedance. Such interfaces may be further tested in simulation settings (Akkermann et al. 2020).

c. Actionable information and uncertainty perceptions

Even though the use of color coding was overall seen as helpful to assess likelihood or severity of spray-icing conditions, participants also inferred uncertainty when assessing colors. While green and red levels provide a sense of confidence about the severity of the situation, yellow levels are difficult to interpret since in maritime operations “it is never red or green, it is always yellow.” In other words, there is always some risk, and all parameters have to be taken into account all the time. However, and in line with various other research, the use of colors was overall found to be particularly helpful to infer risk levels in an intuitive way (MacEachren et al. 2005; Weyrich 2020; Krocak et al. 2023). While our research found different preferences between FT and FP formats, future research should investigate in more detail in which formats colors provide additional values in the case of EPS-based maps.

An important factor that contributed to higher levels of perceived uncertainty among participants pertains to how a given forecast product could be interpreted in a specific decision scenario at hand. This highlights that the scope of actionable information (as well as the perceived uncertainty) in maritime operational decisions is larger than forecast information alone and is rather emerging from a combination of factors. Moreover, the themes along which actionability and uncertainty are inferred reflect both sides of the “use” coin: these sides are indicative of the possible outcome of using forecast information for creating situation awareness in a given decision context. For example, whereas information about context-specific conditions was seen as actionable information, the fact that operations occur in a given locality also was a factor that causes uncertainty, especially when forecast information did not match the spatial or temporal resolution of that operational decision context. Participants’ own accumulated expertise of performing sociotechnical tasks in certain weather conditions and in a specific area was therefore noted as key in how an operational situation is assessed (de Vries 2017). This highlights the bridging function of local expertise between generic information and its contextual application.

The specific condition of sea spray icing itself was also a source of uncertainty. Contextual factors such as ship size and duration of exposure are characteristic for icing hazards and operational considerations reflected such factors. For example, smaller ships are more easily affected by spray icing and may therefore choose a route closer to the coast. At the same time, spray-icing conditions are tolerable for most vessels, if conditions do not last for too long. These contextual considerations underline the dynamic nature of decision-making processes, and how forecast information is embedded in and should be tailored toward these dynamics.

d. Methodological strengths and limitations

The scenario exercises were an engaging way for forecast providers and users to explore the implications of providing forecast uncertainty information for maritime decision contexts. The exercises that were used aimed at finding a balance between representing a realistic scenario and simplifying the decision tasks, as to allow for forecast uncertainty visualizations to take a central, albeit contextualized, place in the discussion. This has been at least partially successful, despite (or exemplified by) several participants who noted that the exercises took a lot of cognitive effort to understand. One reason was the large amount of information included in the scenarios, which had to be processed simultaneously. Paradoxically, both the complexity of the information at hand and the simplification of the task meant that participants had to carefully consider what tools they had available to answer the exercise questions. A second reason for high cognitive demand was related to the content of the exercises, and specifically to grasping the different visualizations of forecast uncertainty. To that end, we argue that initial perceptions of complexity of novel forecast products should not immediately lead to discarding these. Similar to other studies (Argyle et al. 2017), a process of familiarization was visible within the workshop setting. Whereas participants initially required significant time and cognitive effort to make sense of the information and tasks provided in the decision scenarios, later on, a certain familiarity emerged with the “logic” of the exercises and participants became more comfortable with responding to the questions. In the process of codesigning novel weather information services together with users, it may therefore be helpful to explicitly allow for familiarity to be built up over time, which eventually may enhance operational uptake (Blair et al. 2022b).

While the study provides valuable insights into the topic of interpreting and communicating forecast uncertainty, there are some methodological weaknesses that can be identified. First of all, the study primarily relies on a mixed-methods approach, using a workshop format with participants from various sectors. While qualitative research can provide valuable in-depth and contextual insights, the relatively small sample limits the ability to generalize the findings and draw statistically significant conclusions. Similarly, a small sample makes it challenging to capture a wider range of perspectives and experiences that may exist outside of those represented by the participants. The qualitative setting with self-reporting methods may also have facilitated the possibility of response biases, such as social desirability bias, pursuing agendas, or participants’ responses not reflecting their true thoughts and behaviors. In addition, the study focuses on a specific geographical context (Norway) and a specific hazard (sea spray icing). While this allows for a more detailed examination of the topic, it may limit the generalizability of the findings to other regions and hazards. Follow-up research could benefit from considering a broader range of contexts to enhance the external validity of the results. The methodological challenge to balance complexity and simplification in the research design impacts how operational locations and route segments differ according to exposure to spray icing. As mentioned earlier in the discussion, the exercises were designed in such a way that this risk is not objective or quantifiable. Comparability of findings between scenarios is therefore limited and should be done with this limitation in mind. Finally, there is also a potential bias in the participant selection, which was at least partly based on availability and existing contacts. While this may limit representativity, we think that the participants together formed a group of topic experts that was able to provide in-depth insights with perspectives from across the wide range of maritime stakeholders and forecast service providers within the Norwegian context.

5. Conclusions

Using a contextual mixed-method approach, this study explored interpretations of map-based representations of uncertainty in spray-icing forecasts for decisions in maritime operations in Norwegian coastal waters and the northern Atlantic Ocean. Findings were presented from an interactive stakeholder workshop that gathered maritime users, forecast information providers and researchers. We conclude that preferences, routines, and freedom that maritime actors enjoy in choosing which forecast information channels and products to use seem to matter in terms of the extent to which certain products are intuitive, familiar, or trusted. Being more familiar with a (provider of a) forecast product may be positively related to trust in that forecast, at least in cases where use is evaluated positively. In the context of assessing sea-spray-icing as an operational risk, participants appeared risk averse, while showing higher confidence in safety assessments for low-risk areas compared to medium and high-risk areas. While generally the usability of either fixed probability or fixed threshold representations depends on the operational context at hand, further tailoring of possible formats to specific decision contexts is needed to capitalize on the potential of uncertainty as a usable characteristic in forecast information. An integrated solution where operators can overlay risk or warning maps with other layers, such as in weather routing services, may facilitate the usability of probabilistic forecast information. Integrating color-based visualizations that are tailored to the specific operational context may enable users to extract and apply the necessary information for decision-making. Uptake of such visualizations depends at least partly on participants’ own accumulated expertise of performing sociotechnical tasks in certain weather conditions or in a specific area. To avoid suboptimal uptake or increased decision uncertainty, the pace of the familiarization process must be accounted for in testing and codesign activities.

The collaborative nature of the workshop facilitated knowledge sharing and coproduction between forecast providers and users. The study highlighted the importance of incorporating qualitative approaches and considering the complex and dynamic operational contexts of ensemble-based forecast provision, communication, and use. Findings can inform codesign processes of translating ensemble forecast models into usable and useful public and commercial forecast information services.

Acknowledgments.

The authors are grateful for all workshop participants sharing their expertise. All authors acknowledge funding from the Norwegian Research Council under project 301450.

Data availability statement.

Because of privacy and ethical concerns outlined in human-subjects research guidelines and adherence to confidentiality agreements with the participants of this study, a limited and anonymized survey response dataset is available from the first author at request.

REFERENCES

  • Akkermann, A., B. Å. Hjøllo, and M. Siegel, 2020: Maritime co-simulation framework: Challenges and results. Validation and Verification of Automated Systems, A. Leitner, D. Watzenig and J. Ibanez-Guzman, Eds., Springer, 251–269, https://doi.org/10.1007/978-3-030-14628-3_19.

  • Argyle, E. M., J. J. Gourley, Z. L. Flamig, T. Hansen, and K. Manross, 2017: Toward a user-centered design of a weather forecasting decision-support tool. Bull. Amer. Meteor. Soc., 98, 373382, https://doi.org/10.1175/BAMS-D-16-0031.1.

    • Search Google Scholar
    • Export Citation
  • Blair, B., M. Müller, C. Palerme, R. Blair, D. Crookall, M. Knol-Kauffman, and M. Lamers, 2022a: Coproducing sea ice predictions with stakeholders using simulation. Wea. Climate Soc., 14, 399413, https://doi.org/10.1175/WCAS-D-21-0048.1.

    • Search Google Scholar
    • Export Citation
  • Blair, B., A. M. Gierisch, J. Jeuring, S. M. Olsen, and M. Lamers, 2022b: Mind the gap! A consensus analysis of users and producers on trust in new sea ice information products. Climate Serv., 28, 100323, https://doi.org/10.1016/j.cliser.2022.100323.

    • Search Google Scholar
    • Export Citation
  • Boeije, H., 2009: Analysis in Qualitative Research. Sage Publications, 240 pp.

  • Bremer, S., A. Wardekker, M. Baldissera Pacchetti, M. Bruno Soares, and J. van der Sluijs, 2022: Editorial: High-quality knowledge for climate adaptation: Revisiting criteria of credibility, legitimacy, salience, and usability. Front. Climate, 4, 905786, https://doi.org/10.3389/fclim.2022.905786.

    • Search Google Scholar
    • Export Citation
  • Calhoun, K., K. L. Berry, D. M. Kingfield, T. Meyer, M. J. Krocak, T. M. Smith, G. Stumpf, and A. Gerard, 2021: The experimental warning program of NOAA’s hazardous weather testbed. Bull. Amer. Meteor. Soc., 102, E2229E2246, https://doi.org/10.1175/BAMS-D-21-0017.1.

    • Search Google Scholar
    • Export Citation
  • Car, N. J., E. W. Christen, J. W. Hornbuckle, and G. A. Moore, 2012: Using a mobile phone short messaging service (SMS) for irrigation scheduling in Australia—Farmers’ participation and utility evaluation. Comput. Electron. Agric., 84, 132143, https://doi.org/10.1016/j.compag.2012.03.003.

    • Search Google Scholar
    • Export Citation
  • Carr, R. H., B. Montz, K. Maxfield, S. Hoekstra, K. Semmens, and E. Goldman, 2016: Effectively communicating risk and uncertainty to the public: Assessing the National Weather Service’s flood forecast and warning tools. Bull. Amer. Meteor. Soc., 97, 16491665, https://doi.org/10.1175/BAMS-D-14-00248.1.

    • Search Google Scholar
    • Export Citation
  • de Vries, L., 2017: Work as done? Understanding the practice of sociotechnical work in the maritime domain. J. Cognit. Eng. Decis. Making, 11, 270295, https://doi.org/10.1177/1555343417707664.

    • Search Google Scholar
    • Export Citation
  • Engeset, R. V., G. Pfuhl, C. Orten, J. Hendrikx, and A. Hetland, 2022: Colours and maps for communicating natural hazards to users with and without colour vision deficiency. Int. J. Disaster Risk Reduct., 76, 103034, https://doi.org/10.1016/j.ijdrr.2022.103034.

    • Search Google Scholar
    • Export Citation
  • Fujisaki-Manome, A., D. Gill, T. Guo, E. J. Anderson, and M. C. Lemos, 2019: Knowledge co-production in a research-to-operation (R2O) process for development of a Great Lakes ice forecast: Reflection from a stakeholder engagement workshop. ESS Open Archive preprint, 9 pp., https://doi.org/10.1002/essoar.10501135.1.

    • Search Google Scholar
    • Export Citation
  • Fundel, V. J., N. Fleischhut, S. M. Herzog, M. Göber, and R. Hagedorn, 2019: Promoting the use of probabilistic weather forecasts through a dialogue between scientists, developers and end‐users. Quart. J. Roy. Meteor. Soc., 145, 210231, https://doi.org/10.1002/qj.3482.

    • Search Google Scholar
    • Export Citation
  • Grounds, M. A., and S. L. Joslyn, 2018: Communicating weather forecast uncertainty: Do individual differences matter? J. Exp. Psychol. Appl., 24, 1833, https://doi.org/10.1037/xap0000165.

    • Search Google Scholar
    • Export Citation
  • Jeuring, J., and M. Knol-Kauffman, 2019: Mapping weather, water, ice and climate knowledge & information needs for maritime activities in the Arctic. Survey Rep., 44 pp., https://munin.uit.no/bitstream/handle/10037/16508/article.pdf.

  • Jeuring, J., M. Knol-Kauffman, and A. Sivle, 2020: Toward valuable weather and sea-ice services for the marine Arctic: Exploring user–producer interfaces of the Norwegian Meteorological Institute. Polar Geogr., 43, 139159, https://doi.org/10.1080/1088937X.2019.1679270.

    • Search Google Scholar
    • Export Citation
  • Joslyn, S., and S. Savelli, 2021: Visualizing uncertainty for non-expert end users: The challenge of the deterministic construal error. Front. Comput. Sci., 2, 590232, https://doi.org/10.3389/fcomp.2020.590232.

    • Search Google Scholar
    • Export Citation
  • Kinkeldey, C., A. M. MacEachren, and J. Schiewe, 2014: How to assess visual communication of uncertainty? A systematic review of geospatial uncertainty visualization user studies. Cartographic J., 51, 372386, https://doi.org/10.1179/1743277414Y.0000000099.

    • Search Google Scholar
    • Export Citation
  • Klockow-McClain, K. E., R. A. McPherson, and R. P. Thomas, 2020: Cartographic design for improved decision-making: Trade-offs in uncertainty visualization for tornado threats. Ann. Amer. Assoc. Geogr., 110, 314333, https://doi.org/10.1080/24694452.2019.1602467.

    • Search Google Scholar
    • Export Citation
  • Knol, M., P. Arbo, P. Duske, S. Gerland, M. Lamers, O. Pavlova, A. Sivle, and S. Tronstad, 2018: Making the Arctic predictable: The changing information infrastructure of Arctic weather and sea ice services. Polar Geogr., 41, 279293, https://doi.org/10.1080/1088937X.2018.1522382.

    • Search Google Scholar
    • Export Citation
  • Krocak, M. J., J. T. Ripberger, S. Ernst, C. Silva, H. Jenkins-Smith, and A. Bitterman, 2023: Public information priorities across weather hazards and time scales. Bull. Amer. Meteor. Soc., 104, E768E780, https://doi.org/10.1175/BAMS-D-22-0190.1.

    • Search Google Scholar
    • Export Citation
  • Landgren, O. A., Y. Batrak, J. E. Haugen, E. Støylen, and T. Iversen, 2019: Polar low variability and future projections for the Nordic and Barents Seas. Quart. J. Roy. Meteor. Soc., 145, 31163128, https://doi.org/10.1002/qj.3608.

    • Search Google Scholar
    • Export Citation
  • Lemos, M. C., C. J. Kirchhoff, and V. Ramprasad, 2012: Narrowing the climate information usability gap. Nat. Climate Change, 2, 789794, https://doi.org/10.1038/nclimate1614.

    • Search Google Scholar
    • Export Citation
  • Losee, J. E., and S. L. Joslyn, 2018: The need to trust: How features of the forecasted weather influence forecast trust. Int. J. Disaster Risk Reduct., 30, 95104, https://doi.org/10.1016/j.ijdrr.2018.02.032.

    • Search Google Scholar
    • Export Citation
  • MacEachren, A. M., A. Robinson, S. Hopper, S. Gardner, R. Murray, M. Gahegan, and E. Hetzler, 2005: Visualizing geospatial information uncertainty: What we know and what we need to know. Cartography Geogr. Inf. Sci., 32, 139160, https://doi.org/10.1559/1523040054738936.

    • Search Google Scholar
    • Export Citation
  • Matthews, G., H. L. Cloke, S. L. Dance, E. Hansford, C. Mazzetti, and C. Prudhomme, 2023: Co-design and co-production of flood forecast products: Summary of a hybrid workshop. Bull. Amer. Meteor. Soc., 104, E1058E1066, https://doi.org/10.1175/BAMS-D-23-0061.1.

    • Search Google Scholar
    • Export Citation
  • Morss, R. E., J. K. Lazo, and J. L. Demuth, 2010: Examining the use of weather forecasts in decision scenarios: Results from a US survey with implications for uncertainty communication. Meteor. Appl., 17, 149162, https://doi.org/10.1002/met.196.

    • Search Google Scholar
    • Export Citation
  • Mulder, K. J., M. Lickiss, N. Harvey, A. Black, A. Charlton-Perez, H. Dacre, and R. McCloy, 2017: Visualizing volcanic ash forecasts: Scientist and stakeholder decisions using different graphical representations and conflicting forecasts. Wea. Climate Soc., 9, 333348, https://doi.org/10.1175/WCAS-D-16-0062.1.

    • Search Google Scholar
    • Export Citation
  • Mulder, K. J., M. Lickiss, A. Black, A. J. Charlton‐Perez, R. McCloy, and J. S. Young, 2020: Designing environmental uncertainty information for experts and non‐experts: Does data presentation affect users’ decisions and interpretations? Meteor. Appl., 27, e1821, https://doi.org/10.1002/met.1821.

    • Search Google Scholar
    • Export Citation
  • Müller, M., Knol-Kauffman, M., Jeuring, J., and C. Palerme, 2023: Arctic shipping trends during hazardous weather and sea-ice conditions and the polar code’s effectiveness. npj Ocean Sustainability, 2, 12, https://doi.org/10.1038/s44183-023-00021-x.

    • Search Google Scholar
    • Export Citation
  • Naseri, M., and E. M. Samuelsen, 2019: Unprecedented vessel-icing climatology based on spray-icing modelling and reanalysis data: A risk-based decision-making input for Arctic offshore industries. Atmosphere, 10, 197, https://doi.org/10.3390/atmos10040197.

    • Search Google Scholar
    • Export Citation
  • Niclasen, B. A., K. Simonsen, and A. K. Magnusson, 2010: Wave forecasts and small-vessel safety: A review of operational warning parameters. Mar. Struct., 23 (1), 121, https://doi.org/10.1016/j.marstruc.2010.02.001.

    • Search Google Scholar
    • Export Citation
  • Noer, G., Ø. Saetra, T. Lien, and Y. Gusdal, 2011: A climatological study of polar lows in the Nordic Seas. Quart. J. Roy. Meteor. Soc., 137, 17621772, https://doi.org/10.1002/qj.846.

    • Search Google Scholar
    • Export Citation
  • Norwegian Meteorological Institute, 2020: Varslingsområder kyst og hav (Coastal and marine warning areas). Accessed 12 November 2023, https://www.met.no/vaer-og-klima/ekstremvaervarsler-og-andre-farevarsler/varslingsomrader-kyst-og-hav.

  • Novak, D. R., D. R. Bright, and M. J. Brennan, 2008: Operational forecaster uncertainty needs and future roles. Wea. Forecasting, 23, 10691084, https://doi.org/10.1175/2008WAF2222142.1.

    • Search Google Scholar
    • Export Citation
  • Obermeier, H., K. L. Berry, K. E. Klockow-McClain, A. Campbell, C. Carithers, A. Gerard, and J. E. Trujillo-Falcón, 2022: The creation of a research television studio to test probabilistic hazard information with broadcast meteorologists in NOAA’s hazardous weather testbed. Wea. Climate Soc., 14, 949963, https://doi.org/10.1175/WCAS-D-21-0171.1.

    • Search Google Scholar
    • Export Citation
  • Padilla, L. M., M. Powell, M. Kay, and J. Hullman, 2021: Uncertain about uncertainty: How qualitative expressions of forecaster confidence impact decision-making with uncertainty visualizations. Front. Psychol., 11, 579267, https://doi.org/10.3389/fpsyg.2020.579267.

    • Search Google Scholar
    • Export Citation
  • Parker, D. J., and Coauthors, 2022: The African SWIFT project: Growing science capability to bring about a revolution in weather prediction. Bull. Amer. Meteor. Soc., 103, E349E369, https://doi.org/10.1175/BAMS-D-20-0047.1.

    • Search Google Scholar
    • Export Citation
  • Ripberger, J., A. Bell, A. Fox, A. Forney, W. Livingston, C. Gaddie, C. Silva, and H. Jenkins-Smith, 2022: Communicating probability information in weather forecasts: Findings and recommendations from a living systematic review of the research literature. Wea. Climate Soc., 14, 481498, https://doi.org/10.1175/WCAS-D-21-0034.1.

    • Search Google Scholar
    • Export Citation
  • Ruginski, I. T., and Coauthors, 2016: Non-expert interpretations of hurricane forecast uncertainty visualizations. Spat. Cognit. Comput., 16, 154172, https://doi.org/10.1080/13875868.2015.1137577.

    • Search Google Scholar
    • Export Citation
  • Samuelsen, E. M., 2017: Prediction of ship icing in Arctic waters—Observations and modelling for application in operational weather forecasting. Ph.D. thesis, UiT The Arctic University of Norway, 93 pp., https://munin.uit.no/handle/10037/11801.

  • Samuelsen, E. M., 2018: Ship‐icing prediction methods applied in operational weather forecasting. Quart. J. Roy. Meteor. Soc., 144, 1333, https://doi.org/10.1002/qj.3174.

    • Search Google Scholar
    • Export Citation
  • Samuelsen, E. M., and R. G. Graversen, 2019: Weather situation during observed ship-icing events off the coast of northern Norway and the Svalbard archipelago. Wea. Climate Extremes, 24, 100200, https://doi.org/10.1016/j.wace.2019.100200.

    • Search Google Scholar
    • Export Citation
  • Savelli, S., and S. Joslyn, 2012: Boater safety: Communicating weather forecast information to high-stakes end users. Wea. Climate Soc., 4, 719, https://doi.org/10.1175/WCAS-D-11-00025.1.

    • Search Google Scholar
    • Export Citation
  • Sivle, A. D., S. D. Kolstø, P. J. K. Hansen, and J. Kristiansen, 2014: How do laypeople evaluate the degree of certainty in a weather report? A case study of the use of the web service yr. no. Wea. Climate Soc., 6, 399412, https://doi.org/10.1175/WCAS-D-12-00054.1.

    • Search Google Scholar
    • Export Citation
  • Stephens, E. M., D. J. Spiegelhalter, K. Mylne, and M. Harrison, 2019: The Met Office Weather Game: Investigating how different methods for presenting probabilistic weather forecasts influence decision-making. Geosci. Commun., 2, 101116, https://doi.org/10.5194/gc-2-101-2019.

    • Search Google Scholar
    • Export Citation
  • Stewart, E. J., D. Liggett, M. Lamers, G. Ljubicic, J. Dawson, R. Thoman, R. Haavisto, and J. Carrasco, 2020: Characterizing polar mobilities to understand the role of weather, water, ice and climate (WWIC) information. Polar Geogr., 43, 95119, https://doi.org/10.1080/1088937X.2019.1707319.

    • Search Google Scholar
    • Export Citation
  • Tan, M. L., R. Prasanna, K. Stock, E. E. Doyle, G. Leonard, and D. Johnston, 2020: Usability factors influencing the continuance intention of disaster apps: A mixed-methods study. Int. J. Disaster Risk Reduct., 50, 101874, https://doi.org/10.1016/j.ijdrr.2020.101874.

    • Search Google Scholar
    • Export Citation
  • Taylor, A. L., T. Kox, and D. Johnston, 2018: Communicating high impact weather: Improving warnings and decision making processes. Int. J. Disaster Risk Reduct., 30, 14, https://doi.org/10.1016/j.ijdrr.2018.04.002.

    • Search Google Scholar
    • Export Citation
  • Thompson, M. A., J. M. Lindsay, and J. C. Gaillard, 2015: The influence of probabilistic volcanic hazard map properties on hazard communication. J. Appl. Volcanol., 4, 6, https://doi.org/10.1186/s13617-015-0023-0.

    • Search Google Scholar
    • Export Citation
  • Vincent, K., M. Daly, C. Scannell, and B. Leathes, 2018: What can climate services learn from theory and practice of co-production? Climate Serv., 12, 4858, https://doi.org/10.1016/j.cliser.2018.11.001.

    • Search Google Scholar
    • Export Citation
  • Weyrich, P., 2020: To act or not to act: Warning communication and decision-making in response to weather-related hazards. Ph.D. thesis, University of Bayreuth, 156 pp., https://www.research-collection.ethz.ch/handle/20.500.11850/404058.

Save
  • Akkermann, A., B. Å. Hjøllo, and M. Siegel, 2020: Maritime co-simulation framework: Challenges and results. Validation and Verification of Automated Systems, A. Leitner, D. Watzenig and J. Ibanez-Guzman, Eds., Springer, 251–269, https://doi.org/10.1007/978-3-030-14628-3_19.

  • Argyle, E. M., J. J. Gourley, Z. L. Flamig, T. Hansen, and K. Manross, 2017: Toward a user-centered design of a weather forecasting decision-support tool. Bull. Amer. Meteor. Soc., 98, 373382, https://doi.org/10.1175/BAMS-D-16-0031.1.

    • Search Google Scholar
    • Export Citation
  • Blair, B., M. Müller, C. Palerme, R. Blair, D. Crookall, M. Knol-Kauffman, and M. Lamers, 2022a: Coproducing sea ice predictions with stakeholders using simulation. Wea. Climate Soc., 14, 399413, https://doi.org/10.1175/WCAS-D-21-0048.1.

    • Search Google Scholar
    • Export Citation
  • Blair, B., A. M. Gierisch, J. Jeuring, S. M. Olsen, and M. Lamers, 2022b: Mind the gap! A consensus analysis of users and producers on trust in new sea ice information products. Climate Serv., 28, 100323, https://doi.org/10.1016/j.cliser.2022.100323.

    • Search Google Scholar
    • Export Citation
  • Boeije, H., 2009: Analysis in Qualitative Research. Sage Publications, 240 pp.

  • Bremer, S., A. Wardekker, M. Baldissera Pacchetti, M. Bruno Soares, and J. van der Sluijs, 2022: Editorial: High-quality knowledge for climate adaptation: Revisiting criteria of credibility, legitimacy, salience, and usability. Front. Climate, 4, 905786, https://doi.org/10.3389/fclim.2022.905786.

    • Search Google Scholar
    • Export Citation
  • Calhoun, K., K. L. Berry, D. M. Kingfield, T. Meyer, M. J. Krocak, T. M. Smith, G. Stumpf, and A. Gerard, 2021: The experimental warning program of NOAA’s hazardous weather testbed. Bull. Amer. Meteor. Soc., 102, E2229E2246, https://doi.org/10.1175/BAMS-D-21-0017.1.

    • Search Google Scholar
    • Export Citation
  • Car, N. J., E. W. Christen, J. W. Hornbuckle, and G. A. Moore, 2012: Using a mobile phone short messaging service (SMS) for irrigation scheduling in Australia—Farmers’ participation and utility evaluation. Comput. Electron. Agric., 84, 132143, https://doi.org/10.1016/j.compag.2012.03.003.

    • Search Google Scholar
    • Export Citation
  • Carr, R. H., B. Montz, K. Maxfield, S. Hoekstra, K. Semmens, and E. Goldman, 2016: Effectively communicating risk and uncertainty to the public: Assessing the National Weather Service’s flood forecast and warning tools. Bull. Amer. Meteor. Soc., 97, 16491665, https://doi.org/10.1175/BAMS-D-14-00248.1.

    • Search Google Scholar
    • Export Citation
  • de Vries, L., 2017: Work as done? Understanding the practice of sociotechnical work in the maritime domain. J. Cognit. Eng. Decis. Making, 11, 270295, https://doi.org/10.1177/1555343417707664.

    • Search Google Scholar
    • Export Citation
  • Engeset, R. V., G. Pfuhl, C. Orten, J. Hendrikx, and A. Hetland, 2022: Colours and maps for communicating natural hazards to users with and without colour vision deficiency. Int. J. Disaster Risk Reduct., 76, 103034, https://doi.org/10.1016/j.ijdrr.2022.103034.

    • Search Google Scholar
    • Export Citation
  • Fujisaki-Manome, A., D. Gill, T. Guo, E. J. Anderson, and M. C. Lemos, 2019: Knowledge co-production in a research-to-operation (R2O) process for development of a Great Lakes ice forecast: Reflection from a stakeholder engagement workshop. ESS Open Archive preprint, 9 pp., https://doi.org/10.1002/essoar.10501135.1.

    • Search Google Scholar
    • Export Citation
  • Fundel, V. J., N. Fleischhut, S. M. Herzog, M. Göber, and R. Hagedorn, 2019: Promoting the use of probabilistic weather forecasts through a dialogue between scientists, developers and end‐users. Quart. J. Roy. Meteor. Soc., 145, 210231, https://doi.org/10.1002/qj.3482.

    • Search Google Scholar
    • Export Citation
  • Grounds, M. A., and S. L. Joslyn, 2018: Communicating weather forecast uncertainty: Do individual differences matter? J. Exp. Psychol. Appl., 24, 1833, https://doi.org/10.1037/xap0000165.

    • Search Google Scholar
    • Export Citation
  • Jeuring, J., and M. Knol-Kauffman, 2019: Mapping weather, water, ice and climate knowledge & information needs for maritime activities in the Arctic. Survey Rep., 44 pp., https://munin.uit.no/bitstream/handle/10037/16508/article.pdf.

  • Jeuring, J., M. Knol-Kauffman, and A. Sivle, 2020: Toward valuable weather and sea-ice services for the marine Arctic: Exploring user–producer interfaces of the Norwegian Meteorological Institute. Polar Geogr., 43, 139159, https://doi.org/10.1080/1088937X.2019.1679270.

    • Search Google Scholar
    • Export Citation
  • Joslyn, S., and S. Savelli, 2021: Visualizing uncertainty for non-expert end users: The challenge of the deterministic construal error. Front. Comput. Sci., 2, 590232, https://doi.org/10.3389/fcomp.2020.590232.

    • Search Google Scholar
    • Export Citation
  • Kinkeldey, C., A. M. MacEachren, and J. Schiewe, 2014: How to assess visual communication of uncertainty? A systematic review of geospatial uncertainty visualization user studies. Cartographic J., 51, 372386, https://doi.org/10.1179/1743277414Y.0000000099.

    • Search Google Scholar
    • Export Citation
  • Klockow-McClain, K. E., R. A. McPherson, and R. P. Thomas, 2020: Cartographic design for improved decision-making: Trade-offs in uncertainty visualization for tornado threats. Ann. Amer. Assoc. Geogr., 110, 314333, https://doi.org/10.1080/24694452.2019.1602467.

    • Search Google Scholar
    • Export Citation
  • Knol, M., P. Arbo, P. Duske, S. Gerland, M. Lamers, O. Pavlova, A. Sivle, and S. Tronstad, 2018: Making the Arctic predictable: The changing information infrastructure of Arctic weather and sea ice services. Polar Geogr., 41, 279293, https://doi.org/10.1080/1088937X.2018.1522382.

    • Search Google Scholar
    • Export Citation
  • Krocak, M. J., J. T. Ripberger, S. Ernst, C. Silva, H. Jenkins-Smith, and A. Bitterman, 2023: Public information priorities across weather hazards and time scales. Bull. Amer. Meteor. Soc., 104, E768E780, https://doi.org/10.1175/BAMS-D-22-0190.1.

    • Search Google Scholar
    • Export Citation
  • Landgren, O. A., Y. Batrak, J. E. Haugen, E. Støylen, and T. Iversen, 2019: Polar low variability and future projections for the Nordic and Barents Seas. Quart. J. Roy. Meteor. Soc., 145, 31163128, https://doi.org/10.1002/qj.3608.

    • Search Google Scholar
    • Export Citation
  • Lemos, M. C., C. J. Kirchhoff, and V. Ramprasad, 2012: Narrowing the climate information usability gap. Nat. Climate Change, 2, 789794, https://doi.org/10.1038/nclimate1614.

    • Search Google Scholar
    • Export Citation
  • Losee, J. E., and S. L. Joslyn, 2018: The need to trust: How features of the forecasted weather influence forecast trust. Int. J. Disaster Risk Reduct., 30, 95104, https://doi.org/10.1016/j.ijdrr.2018.02.032.

    • Search Google Scholar
    • Export Citation
  • MacEachren, A. M., A. Robinson, S. Hopper, S. Gardner, R. Murray, M. Gahegan, and E. Hetzler, 2005: Visualizing geospatial information uncertainty: What we know and what we need to know. Cartography Geogr. Inf. Sci., 32, 139160, https://doi.org/10.1559/1523040054738936.

    • Search Google Scholar
    • Export Citation
  • Matthews, G., H. L. Cloke, S. L. Dance, E. Hansford, C. Mazzetti, and C. Prudhomme, 2023: Co-design and co-production of flood forecast products: Summary of a hybrid workshop. Bull. Amer. Meteor. Soc., 104, E1058E1066, https://doi.org/10.1175/BAMS-D-23-0061.1.

    • Search Google Scholar
    • Export Citation
  • Morss, R. E., J. K. Lazo, and J. L. Demuth, 2010: Examining the use of weather forecasts in decision scenarios: Results from a US survey with implications for uncertainty communication. Meteor. Appl., 17, 149162, https://doi.org/10.1002/met.196.

    • Search Google Scholar
    • Export Citation
  • Mulder, K. J., M. Lickiss, N. Harvey, A. Black, A. Charlton-Perez, H. Dacre, and R. McCloy, 2017: Visualizing volcanic ash forecasts: Scientist and stakeholder decisions using different graphical representations and conflicting forecasts. Wea. Climate Soc., 9, 333348, https://doi.org/10.1175/WCAS-D-16-0062.1.

    • Search Google Scholar
    • Export Citation
  • Mulder, K. J., M. Lickiss, A. Black, A. J. Charlton‐Perez, R. McCloy, and J. S. Young, 2020: Designing environmental uncertainty information for experts and non‐experts: Does data presentation affect users’ decisions and interpretations? Meteor. Appl., 27, e1821, https://doi.org/10.1002/met.1821.

    • Search Google Scholar
    • Export Citation
  • Müller, M., Knol-Kauffman, M., Jeuring, J., and C. Palerme, 2023: Arctic shipping trends during hazardous weather and sea-ice conditions and the polar code’s effectiveness. npj Ocean Sustainability, 2, 12, https://doi.org/10.1038/s44183-023-00021-x.

    • Search Google Scholar
    • Export Citation
  • Naseri, M., and E. M. Samuelsen, 2019: Unprecedented vessel-icing climatology based on spray-icing modelling and reanalysis data: A risk-based decision-making input for Arctic offshore industries. Atmosphere, 10, 197, https://doi.org/10.3390/atmos10040197.

    • Search Google Scholar
    • Export Citation
  • Niclasen, B. A., K. Simonsen, and A. K. Magnusson, 2010: Wave forecasts and small-vessel safety: A review of operational warning parameters. Mar. Struct., 23 (1), 121, https://doi.org/10.1016/j.marstruc.2010.02.001.

    • Search Google Scholar
    • Export Citation
  • Noer, G., Ø. Saetra, T. Lien, and Y. Gusdal, 2011: A climatological study of polar lows in the Nordic Seas. Quart. J. Roy. Meteor. Soc., 137, 17621772, https://doi.org/10.1002/qj.846.

    • Search Google Scholar
    • Export Citation
  • Norwegian Meteorological Institute, 2020: Varslingsområder kyst og hav (Coastal and marine warning areas). Accessed 12 November 2023, https://www.met.no/vaer-og-klima/ekstremvaervarsler-og-andre-farevarsler/varslingsomrader-kyst-og-hav.

  • Novak, D. R., D. R. Bright, and M. J. Brennan, 2008: Operational forecaster uncertainty needs and future roles. Wea. Forecasting, 23, 10691084, https://doi.org/10.1175/2008WAF2222142.1.

    • Search Google Scholar
    • Export Citation
  • Obermeier, H., K. L. Berry, K. E. Klockow-McClain, A. Campbell, C. Carithers, A. Gerard, and J. E. Trujillo-Falcón, 2022: The creation of a research television studio to test probabilistic hazard information with broadcast meteorologists in NOAA’s hazardous weather testbed. Wea. Climate Soc., 14, 949963, https://doi.org/10.1175/WCAS-D-21-0171.1.

    • Search Google Scholar
    • Export Citation
  • Padilla, L. M., M. Powell, M. Kay, and J. Hullman, 2021: Uncertain about uncertainty: How qualitative expressions of forecaster confidence impact decision-making with uncertainty visualizations. Front. Psychol., 11, 579267, https://doi.org/10.3389/fpsyg.2020.579267.

    • Search Google Scholar
    • Export Citation