Aurora forecasting apps as communicators of space weather: an exploratory perspective

1 Introduction: aurora apps as interfaces to space weather

Aurora viewing has become a widely shared public interest, supported by large online communities and a growing tourism economy (Grand View ResearchInc., 2024). Social media groups allow observers to share space weather data, auroral images, and local sky conditions in real time, and these discussions can quickly spread both accurate and inaccurate interpretations of space weather products (Case et al., 2015). Photography also shapes the ways in which many people engage with the aurora and what they count as a successful viewing experience (Diehl et al., 2016; Kislinger and Kotrschal, 2021). Community reporting platforms such as Aurorasaurus provide location-based, timestamped observations that can serve as practical confirmation of activity (MacDonald et al., 2015).

Mobile applications are now a common means by which the public encounters space weather information. Many apps place geomagnetic indices, solar wind conditions, auroral probability maps, local weather, webcams, and user reports into a single view of current conditions. For users without a technical background, these dashboards often function as interpretive environments rather than simple data displays. Design choices such as labels, color scales, thresholds, and alert phrasing can shape how users understand the data and how confidently they act on it.

This paper examines aurora apps as public-facing communication platforms and decision-support tools. Drawing on a community poll and an exploratory review of commonly used apps, we analyze how space weather data streams are used in apps to communicate auroral visibility. This study focuses on communication patterns rather than on forecast verification or app ranking, and considers how those patterns may affect comprehension, trust, and expectation-setting during a period of growing public engagement with heliophysics (Space Weather Advisory Group, 2024; Frei, 2021; Ledvina et al., 2023; Grandin et al., 2025).

2 App survey and review

Our app review team, a group of experienced aurora chasers, conducted a poll within the Upper Midwest Aurora Chasers (UMAC), a large community for aurora chasing in the five-state area of the US Upper Midwest. The members of the app review team are administrators of this group. We received approval from the University of Minnesota Institutional Review Board (IRB) to conduct the study with human participants. The poll asked respondents which apps they use to inform their aurora chasing, and included write-in responses. Its full content is available in the Supplementary Material. The apps identified by this survey are listed in Table 1 along with their developers, number of downloads, and poll votes. In general, the more popular apps in the poll also had more downloads. Because the poll is a convenience sample, we treat it as an exploratory guide to commonly used apps rather than as a comprehensive survey of the global aurora chasing community.

TABLE 1. Aurora applications reviewed in this study, UMAC poll votes, and availability/readability of in-app explanatory text. Download counts and google play store metadata are approximate and represent a snapshot at the time of review; “n/a” indicates that the app did not provide sufficient explanatory text for quantitative readability assessment.

The app review team investigated the free mobile versions of the following apps identified by the poll: Aurora/My Aurora, Aurora Alerts, SpaceWeatherLive, Aurora Forecast, Aurora Notifier, Aurora Pro, Hello Aurora, Aurora Forecast 3D, Glendale, Amazing Aurora, and Aurora Compass (Table 1). The review focused on the interface components most likely to shape how users translate space weather data streams into estimates of auroral visibility, including geomagnetic indices, empirical models, contextual layers such as webcams and reports, and alert systems.

The app review team also assessed the readability of explanatory text using the Flesch–Kincaid Grade Level formula (Kincaid et al., 1975) for the apps with sufficient text volume. We analyzed passages of at least 100 words from in-app help pages, educational panels, and glossary sections. This step was included as a separate assessment because explanatory text is one of the main ways apps support or constrain interpretation by non-specialist users.

3 How aurora apps communicate space weather conditions and estimates of visible aurora

Aurora apps aggregate heterogeneous space weather observations and model outputs to help users determine expected auroral visibility. In the reviewed sample, this translation generally followed a recognizable sequence: apps summarized activity with indices and related data streams, linked those activity measures to visibility expectations through models or heuristics, added real-time context through reports or cameras, and then converted changing conditions into alerts.

Across the reviewed apps, most that displayed space weather information fell into five groups: (i) geomagnetic indices such as Kp, (ii) solar wind data such as total interplanetary magnetic field strength (Bt), (iii) probability maps based on empirical auroral oval models such as OVATION (Newell et al., 2014), (iv) webcams and user reports, and (v) alerts based on preset or custom thresholds. Geomagnetic indices and auroral oval models often prominently displayed data streams in the reviewed apps.

Several geomagnetic indices were displayed by the apps. Geomagnetic indices serve as a summary of global disturbances of Earth’s magnetic field based on measurements from ground magnetometers. Stronger geomagnetic disturbances are typically associated with equatorward expansion of the auroral oval, and geomagnetic indices are often used as indicators of auroral activity. Kp index, one of the most commonly reported in the apps included in this study, is derived from the readings of a subauroral magnetometer network. Kp index correlates with the equatorward boundary of the auroral oval and has been widely used as an indicator of auroral visibility although it is not its primary purpose (Troyer et al., 2025; Gussenhoven et al., 1983). The definitive Kp index is not available in real time; however, several similar indices developed for operational monitoring are reported by several different organizations. These indices are derived from different magnetometer networks, have a different reporting structure, or are reported for different synoptic windows. As a result, identical values for these indices may represent different conditions or different time periods. The update cadence of geomagnetic indices is also important for aurora chasing considerations because it influences how quickly any given index can reflect a change in geomagnetic conditions and, to some degree, auroral visibility. The traditional Kp index summarizes activity over a completed 3-h synoptic window and provides historical context rather than current conditions. Operational products, such as Nowcast Kp and Hp30, incorporate newer magnetometer data calculations in near-real time, but still summarize activity over the lapsed time within the synoptic interval rather than providing up-to-date information. In eight out of ten apps that mentioned Kp, different versions were used, including 1-min Kp and Estimated Planetary K from NOAA SWPC and Nowcast Kp from the GFZ Helmholtz Centre for Geosciences (Matzka et al., 2021). Some apps displayed the Hp-30 index, which is similar to Kp but uncapped and reported at 30-min cadence (Yamazaki et al., 2022). Different Kp values could therefore be seen at the same time in different apps (Figure 1). Two apps, Glendale and Aurora Compass, displayed custom proprietary geomagnetic indices. These indices were derived from ground magnetometer readings and reported in near real time, likely resulting in a more rapid response to changing conditions than the conventional 3-h Kp index. However, the methodologies for these indices were not publicly disclosed. Users may find it difficult to compare values of these indices to conventional geomagnetic indices or understand what these values represent. Several apps, including SpaceWeatherLive, Aurora Compass, and Hello Aurora, also provided real-time ground magnetometer data.

Many apps used these indices as compact summaries of recent activity, often alongside view lines or other visual cues implying a relationship between a single geomagnetic value and likely auroral visibility. The original Kp index reflecting recent geomagnetic activity is reported on a 3-h cadence and was not designed to track short bursts of auroral activity such as substorms (Kosar et al., 2018). Several frequently displayed indices, including Nowcast Kp and Hp-30, update within a synoptic period as new magnetometer data arrive, but those updates do not necessarily coincide with physical changes in auroral visibility. Some custom geomagnetic indices derived from near-real-time ground magnetometer data may track changes in activity more closely, although their methodologies were not always publicly described.

Empirical auroral oval models were another common way apps displayed auroral visibility estimates. OVATION Prime (Newell et al., 2014) was the most common model in the reviewed apps. It is an empirical model that estimates the probability of visible aurora by calculating the intensity of auroral precipitation across the auroral oval using solar wind data. Aurora Forecast 3D instead implemented three other auroral oval models (Starkov, 1994; Breedveld, 2020; Zhang and Paxton, 2008), and Glendale displayed an auroral oval product based on high- and mid-latitude magnetometer data.

Maps based on OVATION and similar models were visually prominent and could strongly shape user expectations. These maps were sometimes presented without a clear statement of the forecast time, creation time, or update cadence. OVATION-based products are tied to a defined time window and a set of inputs based on solar wind travel time (Newell et al., 2007), so the absence of this timing context changes how a user is likely to read the map.

The user may interpret the probability indicated by the OVATION output as a direct prediction of whether the aurora will be visible from the ground. In reality, the model estimated the likelihood of auroral precipitation overhead that may be sufficient for visibility. The model summarizes solar wind coupling and the expected size and intensity of the auroral precipitation region, not the full set of factors that control ground-level visibility (MacDonald et al., 2015). One common output is a probability of overhead aurora expressed as a percentage. Many apps displayed this as a general chance of seeing aurora at a location, without clarifying that the percentage refers specifically to overhead aurora. Aurora can be visible hundreds of kilometers away from where it is produced, depending on emission altitude, brightness, sky clarity, camera use, and viewing geometry (Case et al., 2016). As a result, an overhead probability map can underestimate how far equatorward aurora may be visible on a clear horizon. Related outputs, such as the hemispheric power index, can also be interpreted as direct visibility indicators, even though changes in these quantities do not translate one-to-one into what a person will see at a specific place and time (Case et al., 2016).

Several apps added contextual layers intended to anchor proxy indicators and model outputs to current observing conditions. User-submitted aurora reports, time-stamped photographs, webcam feeds, and ground magnetometer traces provided a way to compare modeled expectations with conditions observed in practice. In some apps, these features were displayed alongside indices and model maps rather than in separate menus. These contextual layers shift interpretation from reliance on a single-number summary toward broader situational awareness. Reports and camera views can function as a practical ground-truth layer, especially when they include time-stamped images and locations (MacDonald et al., 2015). Their inclusion also helps users learn how different activity levels and model outputs relate to actual sky conditions.

Alerts often reach users when they are not actively monitoring relevant data. Across the reviewed apps, alerts were based on threshold crossings such as Kp exceeding a user-set value, probability estimates of auroral visibility, or selected solar wind conditions such as prolonged negative Bz. Some alert systems also incorporated custom geomagnetic indices or logic intended to capture rapid changes in auroral intensity, including transitions between substorm phases. In these cases, alerts functioned as a direct trigger for user action by converting changing space weather conditions into discrete prompts. Alert systems, however, varied in whether they explained the threshold, the time window, or the event type behind the notification. Differences in alert logic can affect the effectiveness of alerts in decision-making. Alerts based on slowly updating indices, such as conventional Kp, may arrive after the peak activity has ended. Alerts based on rapidly updating data streams can better capture evolving conditions but may generate frequent notifications. Clear explanation of trigger criteria for alerts can help users understand not only the strengths and limitations of individual alerts, but their role in decision support.

4 Communication frictions in app-based space weather

The preceding section describes how apps package data into visibility guidance. The next question is how apps facilitate the interpretation of these data streams for aurora viewing. Across the reviewed apps, recurring frictions were associated with ambiguous labels and provenance, the temporal behavior of commonly displayed indices, and uneven accessibility of explanatory text.

A recurring communication challenge was inconsistent or incomplete labeling across apps. Phrases such as “current Kp” and “probability of visible aurora” were sometimes presented without basic context about whether the value was reported, estimated, or model-derived, which product was being shown, what time window it represented, and when it was last updated.

Provenance and update metadata were also not consistently displayed. Most of the apps disclosed the data source in the user guide or Supplementary Information. However, some errors in data provenance were still present in the form of incorrect attribution of geomagnetic indices to the sources. Most apps did not address possible discrepancies between the values of geomagnetic indices for the same geomagnetic condition. Update cadence for solar wind data was not addressed or disclosed by most apps. In some apps, information about data sources was difficult to locate because it was not clearly marked (i.e., located in the Settings section rather than in an easily accessible dedicated user guide). When this information is missing from primary displays, users may compare values that originate from different sources or reflect different processes. When the source or model name, update cadence, or timestamp is unclear or missing, inconsistencies between apps can appear to be disagreements even when they instead reflect different products or different update timing. Differences in update cadence between the app and a data product can also increase the risk that users interpret unchanged values as stable conditions when the app has simply not been refreshed and no longer displays current data.

Temporal framing emerged as another major source of interpretation difficulty. Kp and related geomagnetic indices are reported on fixed schedules and can obscure rapid changes in activity (Matzka et al., 2021; Troyer et al., 2025). During a synoptic interval, values may also change based on data availability. Many apps indicate “0” both when data actually indicate 0, and as the default value when no data are present. During high geomagnetic activity, users may check apps frequently and expect real-time latency from geomagnetic indices that have fixed reporting intervals. The artifacts of how indices are calculated are often not communicated in the apps and may result in users misinterpreting low values for low geomagnetic activity.

Readability was another source of variability across the reviewed apps. The volume of text was sufficient for a Flesch–Kincaid assessment in five apps, while other apps provided fewer than 100 words of explanatory text and could not be tested quantitatively (Table 1). The Flesch-Kincaid Grade Level provides an estimate of the educational level required to understand written material. It is widely used as an indicator of the accessibility of public-facing communication resources (Ferguson et al., 2021; Zainurrahman et al., 2024). Because aurora apps often serve as an entry point to space weather for non-specialist users, readability is an important measure of the accessibility of the information provided. When enough text was available, the reading level often fell in the range of secondary school through college, and sometimes higher. Even when apps included explanations, those explanations may therefore have been difficult for some users to process (National Literacy Institute, 2024). In apps with little or no explanatory text, and for readers for whom the text was difficult to read, users were left to infer meaning from interface cues such as color coding, single-number visibility scores, or alert frequency. Those simplification strategies can support quick reading, but because they are not standardized across apps, they do not necessarily provide stable interpretation on their own. In addition, graphic representation of data may not be sufficient for understanding data and may, in fact, increase the risk of misinterpretation, as seen with terrestrial weather (Abraham et al., 2015; Broad et al., 2007). Descriptions of specific indices and parameters also varied. Some apps described Kp as a broad indicator of activity and visibility, others explicitly noted its latency for real-time use, and a few correlated probability-style outputs to specific viewing thresholds (e.g., suggesting auroral visibility on the horizon at around 5%–6% chance of overhead activity). These differences change how much interpretive support a user receives before making a decision. The readability assessment was not intended to evaluate the overall effectiveness of individual apps or their success in helping users see aurora. It was used as an indicator of how accessible the written content of aurora apps may be for a non-specialist. Multimodal communication of terrestrial weather that incorporates graphical and text formats has been shown to be beneficial for the general public’s understanding of terrestrial forecasts and may be true for space weather communication (Sivle et al., 2014). Accessible explanatory guidance provides an important foundation for understanding complex space weather products and their intended interpretation, complementing other interface elements, including graphical components and alerts.

5 Design ideas for clearer aurora app communication

Once the main communication frictions are identified, design responses become easier to define. The reviewed apps suggest that relatively modest interface and messaging changes could improve interpretation without requiring substantially more complex dashboards.

To support interpretation of common space-weather data streams in aurora chasing and other public-facing use cases, we highlight a short set of communication elements that may reduce ambiguity:

• Update time and refresh rate: indicate when the value or map was last updated and how often it updates.
• Time window: specify whether the product describes recent conditions, the next hour or two, or a broader outlook.
• Observed versus model-based: label whether the displayed quantity is a direct measurement, a derived index, or a model output.
• Accessible explanations: provide brief plain-language descriptions of what a parameter represents and how it is commonly interpreted for auroral visibility.
• Source information: name the data source and, for maps, identify the model and what it represents.
• Alerts and thresholds: explain what triggers an alert and what the threshold is intended to indicate for auroral visibility.

When present, these elements help users compare apps, interpret differences between app outputs, and avoid common misunderstandings while keeping interfaces streamlined. Our review suggests several practical evaluation strategies that could be applied when assessing aurora apps for usability. These include the following steps: reviewing app content for readability; assessing whether introductory explanations are understandable to non-specialists; checking whether key metadata such as source, time window, and update cadence are easy to find; and evaluating whether multiple complementary indicators are shown instead of a single-number summary.

User guidance also has a role. App users can be encouraged to compare several data streams, learn what commonly displayed products represent, and consult established educational or citizen-science resources when different apps appear to disagree. Links to resources such as Aurorasaurus can provide additional pathways to context, definitions, and subject matter expertise (MacDonald et al., 2015).

6 Implications for the space weather communication ecosystem

The recommendations above apply at the level of interface design, but the broader significance of aurora apps extends beyond any single product. Aurora apps now represent an increasingly visible layer in the public space weather information environment. By combining operational data products, empirical models, and user-generated observations into simple decision interfaces, they influence how non-specialist audiences interpret geomagnetic activity and translate it into expectations of local visibility.

This landscape also creates opportunities for collaboration among app developers, data providers, space weather scientists, and experienced aurora chasers. These groups often work with different goals, terminology, and design constraints, yet they share an interest in clearer expectation-setting. Input from a broader range of experienced, knowledgeable app users can help address many of the identified communication gaps. Lightweight review practices, such as periodic checks of labels, definitions, and alert logic, could improve clarity as products evolve. The app ecosystem may also serve as a practical lens into how end users interpret and operationalize space weather data. Structured feedback from aurora chasing communities, together with examples from citizen science and related public-private efforts (Frei, 2021; Ledvina et al., 2023; Grandin et al., 2025), could support clearer communication, better uncertainty framing, and broader public literacy in heliophysics.

7 Conclusion: aurora apps as emerging communicators of space weather

Aurora applications are an increasingly important part of the public space weather landscape. Most apps in the review (Table 1) showed hundreds of thousands of downloads. In the reviewed sample, they commonly combine geomagnetic indices, empirical models, contextual observations, and alerts into simplified visibility guidance for time-sensitive user decisions. That translation process was not only technical but communicative.

Future studies should focus on interface-based evaluations and user-centered assessments, including comprehension of written guidance and usability studies. Surveys of how space weather products are interpreted for decision-making during aurora chasing may also be informative for app design.

The exploratory review indicates that the reviewed aurora apps differ significantly in labeling, provenance, time framing, and explanatory accessibility. These design elements may shape how users interpret similar underlying data products. Clearer metadata, more explicit interpretive cues, and better plain-language explanations would improve consistency without removing the speed and convenience that make these apps useful. As interest in aurora and smartphone usage continues to grow, these apps will remain an important interface between operational space weather products and public understanding.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fspas.2026.1749567/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Gareth Perry, New Jersey Institute of Technology, United States

Reviewed by: Megan Gillies, Mount Royal University, Canada