A Numerical Weather Prediction Based Road Icing Index for Informed Winter Road Maintenance and Management Decision-Making

A Numerical Weather Prediction Based Road Icing Index for Informed Winter Road Maintenance and Management Decision-Making | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article A Numerical Weather Prediction Based Road Icing Index for Informed Winter Road Maintenance and Management Decision-Making William Hatheway, Matias Ezequiel Suarez This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6443854/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 11 Feb, 2026 Read the published version in Natural Hazards → Version 1 posted 5 You are reading this latest preprint version Abstract Timely and accurate assessment of road surface conditions during winter weather events is critical for ensuring safety, reducing crash rates, and optimizing roadway maintenance activities. To address this need, we propose the Road Icing Index (RII), a new metric that identifies regions with elevated road icing risk based on outputs from numerical weather models such as the Weather Research and Forecasting (WRF) model. The RII integrates multiple atmospheric parameters—including near-surface air temperature (T 2 ), wet-bulb temperature (T w ), precipitation type and rate (QRAIN, QSNOW, RAINNC), relative humidity (RH 2 ), wind speed (U 10 , V 10 ), and top-layer soil temperature (TLST)—into a single spatially explicit metric. These variables are validated by studies such as Tamang et al., (2019) for wet-bulb temperature significance, Stewart et al. (2015) for precipitation type impacts, and Gustavsson et al. (2007) for road weather modeling. Drawing on established meteorological principles, transportation meteorology research, and operational insights, this paper outlines the RII’s methodological framework, implementation process, and applications. In addition, a use case demonstrates its utility in identifying areas of heightened icing potential, offering practical guidance for road maintenance decision-makers and enhancing traveler safety. Future work aims to incorporate vehicle-based observations, advanced precipitation-type algorithms, and refined parameter weighting to improve forecast accuracy and decision-making. Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Winter weather events pose significant challenges to road safety and mobility, contributing to increased accident rates, vehicle damage, travel delays, and economic losses (Pisano et al., 2008; American Meteorological Society, 2004, Andreescu & Frost, 1998). The Federal Highway Administration (FHWA) reports that over 24% of weather-related crashes in the United States occur on snowy, slushy, or icy pavement (FHWA, n.d.). Transportation agencies employ a range of strategies, including Road Weather Information Systems (RWIS), pre-treatment with de-icing agents, sanding, plowing, and issuing public advisories, to help mitigate these risks (Crevier & Delage, 2001; Jin et al., 2024). Despite these efforts, interpreting the multitude of meteorological variables that affect road conditions remains a complex task. Traditional numerical weather prediction (NWP) models do not directly provide a unified measure of road icing potential. The Road Icing Index (RII) addresses this complexity by consolidating key meteorological factors into a single, actionable metric. Building on existing RWIS frameworks (Gustavsson et al., 2007; Boselly et al., 1993), the RII utilizes high-resolution WRF model outputs to evaluate icing risk across spatially continuous domains. By integrating temperature, wet-bulb temperature, precipitation type, humidity, soil temperature, and wind speed, the RII bridges the gap between complex meteorological data and practical road maintenance decisions. This approach is grounded in established research and validated by studies in transportation meteorology (Jin et al., 2024; FHWA, n.d.). This study details the RII’s methodology, including data sources, parameter weighting, and scoring. A use case illustrates its application during a winter storm event, demonstrating its potential to guide resource allocation and proactive measures for road safety. Background and Rationale 2.1. Winter Road Condition Forecasting and Decision Support Winter road conditions arise from a dynamic interplay of atmospheric and surface processes—most notably temperature, precipitation type, and moisture content (Ishii et al,. 2024). Road Weather Information Systems (RWIS) address these challenges by combining real-time sensor data, mobile observations, and NWP model outputs to guide proactive actions such as anti-icing, salting, and plowing (Gustavsson et al., 2007; Jin et al., 2024). In many developed nations, RWIS have considerably lowered crash rates. Studies from Sweden and Finland, for instance, highlight the value of pairing RWIS with localized thermal mapping for more efficient resource use (Gopalakrishna & Gestwick, 2019; Gustavsson et al., 2007). Despite their benefits, RWIS primarily yields point-based information, making it difficult to represent spatial variability across an extensive roadway network (FHWA, n.d.). Moreover, rural or under-resourced areas often lack RWIS infrastructure (Kaiser & Barstow, 2022). By contrast, a gridded Road Icing Index offers a detailed, spatially explicit snapshot of icing risk, thus, enabling prioritization of high-risk corridors by improving the allocation of winter maintenance resources, and facilitating more targeted traveler advisories. To better understand the multi-faceted nature of road icing, the following subsection examines the key atmospheric and surface variables that collectively determine icing risk. 2.2. Key Variables Influencing Road Icing ● Air Temperature (T 2 ): Near-surface air temperature is a fundamental determinant of icing risk. Temperatures below freezing are critical for ice formation, but roads can still become icy at marginally above-freezing temperatures due to latent heat effects (Gustavsson et al., 2007; Jin et al., 2024). ● Wet-Bulb Temperature (T w ): The wet-bulb temperature (T w ) is a more dependable measure of icing potential than air temperature alone because it accounts for both temperature and moisture—two factors essential for ice formation. Specifically, Tw reflects the lowest temperature achievable through evaporation and directly considers humidity. When water evaporates into unsaturated air, it draws heat from the surroundings, cooling the air (and any surfaces) below the ambient temperature. If T w falls below 0 °C , the likelihood of ice formation increases. Moreover, humidity significantly affects icing risk. While air temperature alone ignores how moisture interacts with freezing processes, Tw incorporates those effects, indicating whether water will remain liquid, evaporate, or freeze. This integration of temperature, moisture, and latent heat exchange provides a more precise view of the thermodynamic processes that lead to icing, making the wet-bulb temperature a more accurate gauge for predicting when and where ice is likely to form (Theriault & Stewart, 2005). ● Precipitation Type and Rate: Precipitation type (e.g., freezing rain, snow, sleet) is a critical factor influencing road conditions. Freezing rain, in particular, is associated with hazardous glaze ice formation (Stewart et al. 2015). Rates of precipitation are categorized into low, moderate, and high to capture varying icing risks. Low precipitation can still contribute to icing under freezing conditions, while moderate and heavy rates substantially increase the likelihood of road surface accumulation and hazardous conditions (Stewart et al. 2015; Gustavsson et al., 2007). ● Relative Humidity (RH 2 ): High relative humidity supports ice formation by slowing sublimation and retaining moisture on road surfaces. Conversely, low relative humidity allows roads to dry more quickly, reducing icing risk (Jin et al., 2024). ● Wind Speed (U 10 , V 10 ): Wind speed affects evaporative cooling and surface energy exchange. While moderate winds can enhance cooling and promote ice formation, strong winds can reduce icing potential by drying surfaces (Gustavsson et al., 2007). ● Top-Layer Soil Temperature (TSLT): Soil temperatures below 0 °C exacerbate icing potential by facilitating heat loss from the road surface, further increasing the likelihood of surface icing (Karsisto et al., 2016). Methodology 3.1. Model Configuration and Data Sources The RII is derived from atmospheric conditions that can be simulated by numerical weather models such as the Weather Research and Forecasting Advanced Research WRF (WRF-ARW) model, specifically configured to capture key physical processes influencing road icing potential. The model employs physical parameters designed for boundary-layer dynamics, cloud microphysics, and surface energy exchange (Skamarock et al., 2008). These parameters are essential for accurately simulating near-surface atmospheric and surface conditions that govern icing risks. The model was compiled and installed using self-installation scripts (Hatheway et al., 2023). 1) Model Output Variables: ● 2-Meter Air Temperature (T 2 ): Represents the near-surface air temperature, critical for identifying freezing and near-freezing conditions that lead to road icing. Extracted directly from WRF output, T 2 is provided in degrees Kelvin (K). ● 2-Meter Relative Humidity (RH 2 ): Quantifies the moisture content in the air and its ability to sustain surface wetness. RH 2 is extracted as a percentage (%) from WRF output. ● 10-Meter Wind Components (U 10 , V 10 ): Provide information on wind speed and direction. These components are used to calculate the resultant wind speed at 10 meters above ground level, in meters per second (m/s). ● Precipitation Fields (QGRAUP, QSNOW, QRAIN): Capture rain, snow, and other precipitation types in kilograms of water per kilogram of dry air. ● Temperature Profile Up to 700 mb: Captures atmospheric conditions critical for precipitation type determination. Note: In meteorological terms, 1 kgm -2 of liquid precipitation is equivalent to 1 mm of measurable precipitation (World Meteorological Organization (WMO), 2008). 2) Model Output Derived Variables: ● Wet-Bulb Temperature (T w ): Integrates air temperature and relative humidity to represent the lowest temperature air can reach through evaporation. Calculated using MetPy's wet_bulb_temperature function (May et al., 2021). ● Hybrid Precipitation Type Algorithm (Bourgouin/Ramer/Baldwin): Each method predicts a precipitation type (rain, snow, freezing rain, sleet/graupel). If two or more methods agree, that type is assigned. Remaining ambiguities are resolved by wet-bulb temperature checks at the surface. The Bourgouin method excels at detecting warm layers aloft and shallow cold layers near the surface, making it effective for diagnosing freezing rain events with a pronounced melting layer. However, it can struggle in marginal temperature profiles where small changes can lead to misclassification (Bourgouin, 2000). The Ramer method provides a more detailed view of melting/refreezing processes but may be overly sensitive to temperature errors (Ramer, 1993; Theriault & Stewart, 2005). The Baldwin method is computationally straightforward and works well with clear thermal gradients, but can misclassify precipitation under complex temperature structures. By employing an ensemble prediction approach, each methods’ strengths counteract one another’s weaknesses (Baldwin et al., 1994, Theriault & Stewart, 2005), reducing misclassification in marginal layers. 3) Computational Tools: ● WRF-Python: Utilized to extract model variables, perform diagnostics, and interpolate data onto desired levels. ● MetPy: Provides unit-aware meteorological calculation tools for derived variables (May et al., 2021). ● CartoPy and Matplotlib: Employed for geospatial data visualization ● GeoPandas: Facilitates geographic data manipulation and plotting. 4) Uncertainty Quantification: The WRF model is a powerful tool for simulating atmospheric conditions but does not directly account for road surface conditions or types, which are crucial for accurately predicting road icing. 3.2. Road Icing Index Scoring Each meteorological variable is categorized and assigned a score based on thresholds from literature and operational guidelines. The following table (Table 1) summarizes the Road Icing Index (RII) parameters, their thresholds, associated scores, and references. The scores can be customized and adapted to the area of interest, allowing for region-specific adjustments based on local climatology, infrastructure, and operational needs. The formation of ice on road surfaces is governed by a multifaceted interaction of atmospheric and environmental parameters. Among the most influential of these is ambient air temperature. Extremely low temperatures markedly increase the probability of ice development, particularly when surface moisture is present. Temperatures near the freezing point also pose a substantial risk, as they facilitate the phase transition of moisture into ice under favorable conditions. Although slightly above-freezing temperatures generally diminish this potential, ice formation may still occur, especially during periods of radiative cooling or in areas with persistent surface moisture. Conversely, temperatures well above freezing are typically insufficient to support ice formation on roadways (Stewart 1992; Gustavsson et al., 2007; Stewart et al. 2015; Jin et al., 2024). Wet-bulb temperature, which integrates the effects of ambient temperature and humidity, serves as a critical indicator of the atmosphere’s cooling potential. Low wet-bulb values reflect both a high moisture content and a capacity for evaporative cooling, thereby enhancing the likelihood of ice accumulation. As wet-bulb temperatures increase, the efficiency of evaporative cooling diminishes, correspondingly reducing the risk of ice formation. Precipitation type and intensity further modulate road surface conditions. Heavy freezing rain, for example, can generate hazardous glaze ice, while moderate to heavy snowfall or sleet may lead to significant surface accumulation, exacerbating slipperiness. Lighter forms of precipitation exert a comparatively minimal influence unless ambient conditions are near freezing (Stewart 1992). Rainfall introduces additional complexity, particularly when accompanied by marginal temperatures. Heavy or moderate rain can saturate road surfaces, and if coinciding with subfreezing or near-freezing conditions, this can significantly elevate the risk of ice formation. Even light rain can contribute to hazardous conditions under the right thermal profile (Pisano et al., 2008; Stewart et al. 2015; Jin et al., 2024). Relative humidity and wind speed also exert notable effects. Elevated humidity supports prolonged moisture retention on road surfaces, thereby facilitating ice development. Wind influence is bidirectional: gentle winds may promote evaporative cooling and moisture persistence, while stronger winds generally enhance surface drying, mitigating ice risk (Thordarson & Olafsson, 2008). Finally, subsurface conditions such as soil temperature play an important thermodynamic role. Colder soil temperatures accelerate the loss of surface heat, increasing the probability of surface freezing. In contrast, warmer soil temperatures act to buffer against rapid heat loss, thereby reducing the propensity for ice formation (Karsisto et al., 2016). Taken together, these factors form a dynamic system in which ice risk is determined by the concurrent state of multiple meteorological and environmental variables. A comprehensive understanding of these interactions is essential for accurate forecasting and effective mitigation of roadway icing events. Table 1. Presents the Road Icing Index (RII) parameters, their thresholds, corresponding scores, and references. Parameter Threshold Score References Air Temperature (T 2 ) T2< -5°C 5 Stewart 1992; Gustavsson et al., 2007; Stewart et al. 2015; Jin et al., 2024 –5°C ≤ T2 ≤ 0°C 3 0°C < T2 ≤ 3°C 1 3°C < T2 ≤ 5°C -1 5°C < T2 ≤ 10°C -3 10°C < T2 -5 Wet-Bulb Temperature (T w ) Tw 2°C 0 Precipitation Type Freezing Rain 3 Pisano et al., 2008; Stewart et al. 2015 Jin et al., 2024 Snow or Sleet 2 Rain 1 No Precipitation 0 Precipitation Rate – Freezing Rain Rate > 7.5 mm/hr 3 Pisano et al., 2008; Stewart et al. 2015 2.5 mm/hr 3.0 mm/hr 3 1.0 mm/hr Rate ≤ 3.0 mm/hr 2 Rate ≤ 1.0 mm/hr 1 Precipitation Rate – Rain Rate > 15.0 mm/hr 3 5.0 mm/hr < Rate ≤ 15.0 mm/hr 2 Rate ≤ 5.0 mm/hr 1 Relative Humidity (RH 2 ) RH2 ≥ 90% 1 Thordarson & Olafsson, 2008 RH2 < 90% 0 Wind Speed (U 10 , V 10 ) 20 m/s -1 Top Layer Soil Temperature (TLST) TLST < 0°C 1 Karsisto et al., 2016 TLST ≥ 0°C 0 To enhance the understanding of how precipitation types are determined within the RII framework, (Table 2) outlines the decision-making process using the hybrid approach that integrates the Bourgouin, Ramer, and Baldwin methods. These structured criteria ensure accurate classification of precipitation types, thereby improving the reliability of the RII. Table 2. Outlines the decision-making process for determining precipitation types using a hybrid approach that leverages the strengths of the Bourgouin, Ramer, and Baldwin methods. This ensemble prediction mechanism enhances the accuracy and reliability of precipitation classification, which is critical for calculating the Road Icing Index (RII). Precipitation Type Bourgouin Method Criteria Ramer Method Criteria Baldwin Method Criteria Final Classification References Rain - No significant warm layer aloft. - Surface temperature > 0°C. - Surface temperature > 0°C. - No warm layer aloft. - Minimal or no melting/refreezing layer. - Clear thermal gradients. If Bourgouin and Ramer agree, classify it as Rain. Bourgouin, 2000; Ramer, 1993; Baldwin et al., 1994 Freezing Rain - Presence of a warm layer aloft (> 0°C). - Shallow cold layer near the surface (≤ 0°C). - Warm layer aloft (> 0°C). - Surface temperature ≤ 0°C. - Thin melting/refreezing layer. - Specific thickness and temperature differences. If Bourgouin and Ramer both indicate freezing conditions, classify as Freezing Rain. Bourgouin, 2000; Ramer, 1993 Snow - No significant warm layer aloft. - Entire temperature profile ≤ 0°C. - Entire temperature profile ≤ 0°C. - Consistent below-freezing temperatures. - No melting layer. If all methods consistently indicate below-freezing temperatures without warm layers, classify as Snow. Bourgouin, 2000; Ramer, 1993; Baldwin et al., 1994 Sleet/Graupel - Presence of a warm layer aloft followed by a cold layer. - Significant graupel mixing. - Mixed temperature profiles with layers fluctuating around 0°C. - Presence of graupel. - Intermediate melting/refreezing processes. - Variable thermal gradients. If Bourgouin, Ramer, and Baldwin collectively indicate mixed precipitation with graupel, classify as Sleet/Graupel. Bourgouin, 2000; Ramer, 1993; Baldwin et al., 1994., None - No significant precipitation indicators. - Surface conditions dry. - No precipitation detected. - No melting/refreezing layers. - Clear skies or minimal moisture. If none of the methods detect precipitation conditions, classify as None. Jin et al., 2024 3.3. Total Road Icing Index Score and Risk Categories The total RII score is calculated by adding individual variable scores together, as visualized in Fig. 1. and is categorized as follows: 0–3: Minimal Risk – No significant icing threat; minimal intervention needed. 3–6: Low Risk – Minor icing potential; routine preventive measures may be sufficient. 6–9: Moderate Risk – Higher icing potential; proactive maintenance and monitoring are advised. 9–12: High Risk – High likelihood of road icing; immediate maintenance is required. ≥ 12: Severe Risk – Severe icing threat; extensive and urgent mitigation is essential. Fig 1 provides a visual representation of the RII calculation workflow, including ensemble voting for precipitation type classification. The flowchart outlines the sequence of steps, starting from input data extraction to the final determination of risk levels. The integration of voting ensures that multiple classification methods contribute to the accuracy of the precipitation type assessment, forming a core component of the RII calculation process. A heuristic approach is a simplified rule-based method—often referred to as a “rule-of-thumb” strategy—that delivers practical, detection and solutions without incurring the complexity of advanced or purely data-driven algorithms. In the context of road icing prediction, heuristics generally establish threshold values and weighting factors for each meteorological parameter (e.g., air temperature, wind speed, and precipitation rate) based on physical principles, empirical observations, and operational practice. This design ensures computational simplicity by avoiding nested calculations or resource-intensive machine learning models, making real-time deployment more feasible and useful to the agencies responsible for road conditions. At the same time, it preserves operational relevance because thresholds can be readily calibrated to match local climatological characteristics. For instance, a region prone to freezing rain can adopt lower temperature thresholds compared to one more frequently affected by heavy snow. While heuristics may not always achieve the same level of precision as highly sophisticated modeling techniques, their transparency, reliability, and adaptability render them well-suited for day-to-day road maintenance decisions. Impact of Treated and Untreated Roads on the RII 4.1. Treated Roads Treatments such as salting, sanding, and chemical de-icing lower the freezing point of water, reducing the likelihood of road surface icing (Crevier & Delage, 2001). By preventing or delaying ice formation, these treatments effectively alter the thermal and moisture balance at the road surface, thereby lowering the RII’s practical impact—even if meteorological parameters (e.g., sub-freezing temperatures, precipitation) remain unchanged. For instance, calcium chloride is effective down to approximately –20°C, making it valuable in severe winter conditions (FHWA, n.d.). However, treatment effectiveness can vary based on several factors: Temperature Extremes – Extremely cold conditions can slow chemical reactions, reducing melting capability (Crevier & Delage, 2001). Precipitation Intensity – Heavy or sustained precipitation may wash away chemicals, necessitating reapplication (FHWA, n.d.). Traffic Volume – High traffic distributes de-icing chemicals but can also deplete them faster. Application Timing – Pre-treatments are most effective if applied prior to the onset of precipitation. Even treated roads may register high RII scores if chemicals are overwhelmed by intense precipitation rates, rapidly dropping temperatures, heavy traffic, or prolonged events. Monitoring these factors helps agencies tailor responses, such as recommending additional salting or sanding. 4.2. Untreated Roads Untreated roads, especially in rural or low-maintenance regions, rely entirely on meteorological and surface conditions to dictate icing risk. When near-freezing temperatures align with precipitation, roads can quickly shift from wet to icy, creating unexpected hazards for motorists. Rapid Response – Deploying maintenance units to proactively treat priority segments.. Traveler Warnings – Using message boards, media outlets, social media, and applications to advise reduced speeds and alternative routes. Preventive Measures – Instituting variable speed limits or temporary closures in high-risk areas. Future work can refine the RII by incorporating and analyzing historical real-time treatment logs and traffic accident records to check the RII accuracy. If a segment was recently salted under temperatures that favor the chemical’s effectiveness, the RII score could be slightly reduced. Conversely, if no maintenance has occurred despite high precipitation and sub-freezing temperatures, the RII may remain elevated. Such treatment-aware modeling ensures more efficient resource allocation and better reflection of on-the-ground conditions.: Even treated roads may register high RII scores if chemicals are overwhelmed by intense precipitation rates, rapidly dropping temperatures, heavy traffic, or prolonged events. Monitoring these factors helps agencies tailor responses, such as recommending additional salting or sanding. Use Case: Application to a Winter Storm Event As an example of the application of the Road Icing Index (RII), a cold weather event that occurred in the central region of Argentina was simulated using the Weather Research and Forecasting (WRF) model. The RII successfully identified critical icing zones within the affected areas, particularly highlighting regions forecasted to experience freezing rain and near-surface freezing conditions. The highest RII scores were concentrated along mountain roads, where the risk of hazardous driving conditions was most significant. To further validate the RII predictions, the outputs were compared against data from the network of meteorological stations in the province of Córdoba (Argentina) and reports from news sources. These comparisons provided an observational reference for assessing the accuracy of RII predictions regarding near-surface temperature drops and road icing conditions. Notably, the RII was able to detect the imminent risk of freezing along the Camino de las Altas Cumbres, a heavily trafficked mountain road that connects key regions and is frequently used by drivers. This detection highlights the RII's operational utility in identifying high-risk areas that are critical for transportation safety. 5.1. Description of the area of application The study area corresponds to the province of Córdoba, located in central Argentina, covering a surface area of 165,321 km² (see Fig. 3.). This extensive territory features a diverse landscape, including plains, mountain ranges, and valleys, which shape its unique geographic identity. The eastern part of the province consists of a vast plain with a gentle slope towards the east. In contrast, the western region is characterized by three mountain ranges collectively known as the Sierras de Córdoba. The easternmost range, Sierras Chicas Mountains, has elevations ranging from around 1,000 m in the southern end to 1,950 m at Cerro Uritorco in the northern end, both measured above sea level. Further west, two significant valleys—Calamuchita to the south and Punilla to the north—play an essential role in the regional economy, primarily due to tourism. Beyond these valleys, the next mountain range includes Champaquí Hill, the highest peak in the province at 2,880 m. A key route that traverses the mountainous region of the Sierras de Córdoba is the Camino de las Altas Cumbres (Altas Cumbres Road), a major roadway widely used for transportation and tourism. However, due to its high-altitude location, this route is particularly susceptible to adverse weather conditions, especially during winter. Low temperatures frequently cause ice formation and freezing conditions, leading to hazardous driving situations and temporary road closures. The climate in Córdoba is temperate to subtropical, with a distinct dry season. Both temperature and precipitation generally decrease from north to south and from east to west, except along the eastern slopes of the mountains, where humid easterly winds contribute to higher rainfall levels. Rainfall is seasonal, occurring mainly between October and April. Temperature variations are notable throughout the year. In summer, Córdoba experiences warm conditions, with average maximum temperatures around 30°C and peaks that can exceed 40°C. Winters, on the other hand, are mild and relatively dry. The average temperature during this season ranges between 10°C and 12°C, with maximum values around 18°C and minimum temperatures between 4°C and 5°C. A significant winter feature is the occurrence of snowfall, particularly in the higher-altitude areas of the province. 5.2. Data Collection Meteorological variables necessary for RII calculation were obtained from WRF model outputs corresponding to the freezing event. Key atmospheric and surface parameters—including near-surface air temperature, wet-bulb temperature, precipitation type and rate, relative humidity, and wind speed—were extracted for further analysis. Given that these events were characterized by significant cloud cover, satellite-derived Land Surface Temperature (LST) data from GOES-16 could not be utilized, as the cloud coverage obstructed LST observations. Instead, the validation relied on temperature, dew point, relative humidity, precipitation and wind speed readings from the meteorological station network in Córdoba and information extracted from news reports that highlighted the severity of the event. The total number of available meteorological stations in Córdoba is approximately 250, from which only those that recorded data during the studied event were used. 5.3. WRF Model Configuration The WRF model version 4.2 was used (Skamarock et al., 2008). Initial conditions were provided by the GFS global model with a horizontal resolution of 0.25 degrees. The WRF domain consists of 270×270 grid points with a horizontal resolution of 4 km and 35 vertical levels. The physical parameterizations used were: Thompson microphysics (option 8), Mellor–Yamada–Janjic planetary boundary layer scheme (option 2), Dudhia shortwave radiation scheme (option 2), RRTM longwave radiation scheme (option 1), Unified Noah land surface model (option 2), and Eta similarity surface layer scheme (option 2). 5.4. RII Calculation A Python script was developed to integrate all relevant atmospheric and surface variables from the WRF outputs to compute the RII (see Appendix B). This script generated spatially explicit RII maps, pinpointing regions with elevated road icing risks during both events. On June 16, 2021, an extreme cold event affected the central region of Argentina, particularly Córdoba, leading to subzero temperatures. This unusual weather phenomenon resulted from a strong cold air mass intrusion, causing widespread frost and significant temperature drops across the region. Fig. 3. shows the meteorological variables measured at the surface by the network of weather stations on June 16, 2021 at 11Z, with black dots representing the locations of the weather stations. The thick solid line indicates the political boundary of the province of Córdoba, while the thin lines represent contour lines at 1000-meter intervals above sea level, providing a better reference for high-altitude areas. To enhance visualization, the measured data were interpolated, and a characteristic color map was applied for each variable to approximate its distribution across the territory. Temperatures below 6°C were observed across most of the region. In particular, the highest areas of the region exhibit the lowest temperatures, reaching values between -4°C and -6°C. The dew point temperature and relative humidity maps indicate that the higher-altitude regions are the areas closest to saturation. The precipitation map shows that low precipitation rates occurred at certain localized points in the region. The wind speed map reveals that low wind speeds were recorded in high-altitude areas, which could increase the chance of the road freezing. Fig. 4. shows the RII calculated 24 hours in advance using the methodology described in this paper. Three panels are presented at 03Z, 11Z, and 19Z to capture the temporal evolution of the index: the night before (marking the onset of temperature decline), the morning (when the risk is at its peak), and the afternoon (when the risk diminishes due to rising temperatures). As shown in this figure, the RII obtained from the WRF model and the procedure explained in this paper reveals that the areas with the highest risk of freezing are the highest-altitude regions of the territory. This finding aligns with the observed meteorological variables and news reports, which indicated that the roads traversing the high-altitude areas of the region were the ones affected by freezing. The following links correspond to official reports and news articles in Spanish about the event: https://prensa.cba.gov.ar/informacion-general/historica-nevada-en-cordoba-por-que-se-dio-este-impactante-fenomeno/ and https://www.minutouno.com/sociedad/nieve/frio-polar-el-pais-las-mejores-fotos-y-videos-la-nevada-cordoba-n5202922. 5.5. Results Interpretation The results showed that areas with high RII values closely corresponded to the actual regions that experienced significant road icing and higher accident rates. Importantly, the RII successfully identified critical icing risks on the Camino de las Altas Cumbres, a vital mountain route heavily used by drivers. This capability demonstrates the RII’s operational relevance, providing actionable insights for optimizing winter road maintenance activities and improving driver safety during cold weather events. Discussion and Conclusion The RII offers a scalable, actionable metric for assessing road icing potential, particularly in regions lacking extensive RWIS infrastructure. Its integration of multiple meteorological parameters ensures a comprehensive evaluation of icing risks. Moreover, the RII can be customized to specific regions by adjusting the scoring criteria for each variable, allowing for better alignment with local climatological and geographical conditions. This use case illustrates the RII’s capability, showing its value in guiding maintenance decisions. In particular, this use case applied to an extreme cold weather event in Córdoba, along the Altas Cumbres Road, a high-mountain route with significant traffic flow, demonstrates that the RII is capable of identifying areas with a high likelihood of road surface freezing. 6.1. Practical Implementation The RII can be easily incorporated into existing transportation management systems, delivering real-time updates and forecasts. Visualization dashboards can help decision-makers apply proactive measures and allocate resources effectively. 6.2. Validation and Future Work Further validation against broader observational datasets is warranted. Future efforts may include integrating pavement temperature models, refining precipitation-type classification, and employing real-time maintenance logs to dynamically adjust the RII (Jin et al., 2024; Theriault & Stewart, 2005; FHWA, n.d.). 6.3. Economic and Policy Implications More accurate road icing indices can yield substantial economic benefits by optimizing maintenance operations helping to reduce crash-related expenses, and enhance public safety (Wu et al., 2021). Policymakers can use the RII to inform infrastructure investments and develop holistic winter road management strategies. The RII’s accuracy depends on numerical weather model outputs, which may be affected by model resolution and the parameterization schemes used. While this study utilizes the WRF model, these outputs can also be obtained from other forecasting models. Additionally, the scoring framework, though effective, may require localized calibration to account for regional climate and infrastructure. Overall, the Road Icing Index represents a significant advance in winter road condition forecasting. Its successful application in the use case underscores its potential to enhance road safety and operational efficiency. Continued refinement and validation will further solidify the RII as a vital tool in winter transportation meteorology. Declarations Acknowledgments: The authors want to thank the anonymous reviewers for helping to improve the quality of the manuscript. Matías Ezequiel Suárez gratefully acknowledges the PhD scholarship awarded by the National Scientific and Technical Research Council of Argentina (CONICET). Author’s contributions: WH designed and drafted the manuscript and initially came up with the first attempt of the calculations. WH and MS conceived and coordinated the study, as well as refined the calculations. MS ran the use study. All authors read and approved the final manuscript. Funding: The research authors did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Availability of data and material : All data used in this study are generated and archived by Matias Ezequiel Suarez of the National University of Córdoba. The meteorological station network data were downloaded from the following website: https://cordoba.redesclimaticas.com/, which requires prior registration, but access is free of charge. The calculations and python code used to generate the graphics are adapted from the WRF-Python public website (https://wrf-python.readthedocs.io/en/latest/plot.html). Calculations used and scoring for the Road Icing Index are found in the appendix below. Conflict of interests: The authors have no conflicts of interest to declare that are relevant to the content of this article. Ethics approval and consent to participate: Not applicable. References American Meteorological Society, 2004: Weather and highways: Report of a policy forum. Atmospheric Policy Program, https://www.ametsoc.org/sites/ams/assets/File/policy/WxHighways_2004.pdf. Andreescu, M. P., and D. B. Frost, 1998: Weather and traffic accidents in Montreal, Canada. Climate Res., 9(3), 225–230, https://doi.org/10.3354/cr009225. Baldwin, M., R. Treadon, and S. Contorno, 1994: Precipitation type prediction using a decision tree approach with NMC’s Mesoscale Eta Model. Preprints, Tenth Conf. on Numerical Weather Prediction, Amer. Meteor. Soc., Portland, OR, 30–31. Boselly, S. E., J. Thornes, C. Ulberg, and D. Ernst, 1993: Road weather information systems, Volume I. Strategic Highway Research Program Publication-SHRP-H-350, Natl. Res. Council, Washington, DC, 90–93, https://onlinepubs.trb.org/onlinepubs/shrp/shrp-h-350.pdf. Bourgouin, P., 2000: A method to determine precipitation types. Wea. Forecasting, 15(5), 583–592, https://doi.org/10.1175/1520-0434(2000)0152.0.CO;2. Crevier, L. P., and Y. Delage, 2001: METRo: A new model for road-condition forecasting in Canada J. Appl. Meteor., 40(11), 2026–2037, https://doi.org/10.1175/1520-0450(2001)0402.0.CO;2. Federal Highway Administration (FHWA), n.d.: Weather impacts on roads. https://ops.fhwa.dot.gov/weather. Gopalakrishna, D., and T. Gestwick, 2019: Road Weather Management Performance Measures Update (FHWA-HOP-19-089). Federal Highway Administration, U.S. Dept. of Transportation, https://ops.fhwa.dot.gov/publications/fhwahop19089/fhwahop19089.pdf. Gustavsson, T., and J. Bogren, 2007: Information not data: Future development of road weather information systems. Geogr. Ann. A, 89(4), 263–271, https://doi.org/10.1111/j.1468-0459.2007.00325.x. Hatheway, W., H. Snoun, H. ur Rehman, and others, 2023: WRF-MOSIT: A modular and cross-platform tool for configuring and installing the WRF model. Earth Sci. Inform., 16, 4327–4336, https://doi.org/10.1007/s12145-023-01136-y. Ishii, K., S. Ono, T. Masago, M. Ishizuki, T. Mori, and Y. Hanatsuka, 2024: Fine-scaled predictive modeling of road surface conditions and temperature in urban areas. IEEE Trans. Intell. Transp. Syst., https://doi.org/10.1109/TITS.2024.3433004. Jin, M., and D. G. McBroom, 2024: Investigating road ice formation mechanisms using Road Weather Information System (RWIS) observations. Climate, 12(5), 63, https://doi.org/10.3390/cli12050063. Kaiser, N., and C. K. Barstow, 2022: Rural transportation infrastructure in low-and middle-income countries: A review of impacts, implications, and interventions. Sustainability, 14(4), 2149, https://doi.org/10.3390/su14042149. Karsisto, V., P. Nurmi, M. Kangas, M. Hippi, C. Fortelius, S. Niemelä, and H. Järvinen, 2016: Improving road weather model forecasts by adjusting the radiation input. Meteor. Appl., 23(3), 503–513, https://doi.org/10.1002/met.1574. May, R., B. Reed, and G. Thomas, 2021: MetPy: A Python package for meteorological calculations. https://unidata.github.io/MetPy/latest/. Pisano, P. A., L. C. Goodwin, and M. A. Rossetti, 2008: US highway crashes in adverse road weather conditions. 24th Conf. on International Interactive Information and Processing Systems for Meteorology, Oceanography and Hydrology, New Orleans, LA, Amer. Meteor. Soc., 20–24. Ramer, J., 1993: An empirical technique for diagnosing precipitation type from model output. 5th Intl. Conf. on Aviation Weather Systems, Vienna, VA, 227–230. Skamarock, W. C., J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, W. Wang, and J. G. Powers, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, https://doi.org/10.5065/D68S4MVH. Stewart, R. E., 1992: Precipitation types in the transition region of winter storms. Bull. Amer. Meteor. Soc., 73(3), 287–296, https://doi.org/10.1175/1520-0477(1992)0732.0.CO;2. Stewart, R. E., J. M. Thériault, and W. Henson, 2015: On the characteristics of and processes producing winter precipitation types near 0°C. Bull. Amer. Meteor. Soc., 96(4), 623–639, https://doi.org/10.1175/BAMS-D-14-00032.1. Tamang, S. K., A. M. Ebtehaj, A. F. Prein, and A. J. Heymsfield, 2019: On changes of global wet-bulb temperature and snowfall regimes. arXiv, https://doi.org/10.48550/arXiv.1905.07776. Theriault, J., and R. Stewart, 2005: Winter precipitation types and icing at the surface. Proc. 11th Intl. Workshop on Atmos. Icing of Structures, 6. Thordarson, S., and B. Olafsson, 2008: Weather induced road accidents, winter maintenance and user information. Transport Research Arena Europe 2008, 72. WMO (World Meteorological Organization), 2008: Guide to Meteorological Instruments and Methods of Observation. Weather‐Climate‐Water. Wu, J., S. Yang, F. Yang, and X. Yin, 2021: Road weather monitoring system shows high cost-effectiveness in mitigating malfunction losses. Sustainability, 13(22), 12437, https://doi.org/10.3390/su132212437. Supplementary Files Appendix.docx ClimaReporte1606202116062021.xlsx Cite Share Download PDF Status: Published Journal Publication published 11 Feb, 2026 Read the published version in Natural Hazards → Version 1 posted Editorial decision: Minor revisions 25 Aug, 2025 Reviewers agreed at journal 14 May, 2025 Reviewers invited by journal 07 May, 2025 Editor assigned by journal 16 Apr, 2025 First submitted to journal 15 Apr, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6443854","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":453230337,"identity":"331da26b-345f-4232-80af-fd8064d73922","order_by":0,"name":"William Hatheway","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA80lEQVRIiWNgGAWjYBACAwYGNiiTsfHAgwogzczcQEALM1xLw4GEMyAtjERrYWA4kNgG0YtXizn7+WMPPrbZJTZIJDccSJxXG83fDtTyo2IbTi2WPcnshjPbkoFaEoFath3PnXGYsYGx58xt3A47kMwmzbuNObFBGqzlWG4DUAszYxseLecfs0n/3VYP1TLnWO58glpuAG1h3HYYqqWhJncDYS2PzSR7/x03bpN/CAzkYwdyNwK1HMTrl/OJzyR+nKmW7ec5/vDBh5q63HnnDx988KMCtxYYcGyD0IfB5AGC6oHAHkrXEaN4FIyCUTAKRhgAABhnZEMoAIHeAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-7814-5336","institution":"Watchdog Weather","correspondingAuthor":true,"prefix":"","firstName":"William","middleName":"","lastName":"Hatheway","suffix":""},{"id":453230338,"identity":"eebbf2c4-fe8f-49e9-a4e5-5a7a64c07966","order_by":1,"name":"Matias Ezequiel Suarez","email":"","orcid":"","institution":"University of Cordoba: Universidad de Cordoba","correspondingAuthor":false,"prefix":"","firstName":"Matias","middleName":"Ezequiel","lastName":"Suarez","suffix":""}],"badges":[],"createdAt":"2025-04-14 08:21:35","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6443854/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6443854/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11069-025-07883-z","type":"published","date":"2026-02-11T15:58:43+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":82606580,"identity":"d88e0cb2-e6f8-41fa-b395-32fdd516b173","added_by":"auto","created_at":"2025-05-13 10:08:26","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":47510,"visible":true,"origin":"","legend":"\u003cp\u003eVisual representation of RII calculation.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6443854/v1/a940812b89d03fa6d76df882.png"},{"id":82605216,"identity":"3b17a1d2-a093-48a8-94a6-3afe9ddb64b1","added_by":"auto","created_at":"2025-05-13 10:00:24","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":212070,"visible":true,"origin":"","legend":"\u003cp\u003eTopographic map of the province of Córdoba, Argentina. The Altas Cumbres Road, a major mountain road, is highlighted in pink, traversing the Sierras de Córdoba.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6443854/v1/a982aac3fd6e21da564d1cd8.png"},{"id":82605222,"identity":"d3a7733b-0f4e-4bee-9de0-93dcfadc0467","added_by":"auto","created_at":"2025-05-13 10:00:24","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1164587,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of meteorological variables recorded by the station network on June 16, 2021 at 11Z and interpolated across the region. The panels (from top to bottom, left to right) show temperature (°C), dew point (°C), relative humidity (%), precipitation (mm), and wind speed (m/s). Black dots indicate the locations of weather stations.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6443854/v1/160945c5f7b155a7cfe4f047.png"},{"id":82605242,"identity":"ff69b174-fbc1-4645-84b6-3260b3921894","added_by":"auto","created_at":"2025-05-13 10:00:25","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":281855,"visible":true,"origin":"","legend":"\u003cp\u003eTemporal evolution of the Road Icing Index (RII) forecasted using the WRF model. The RII is calculated using the output from the WRF forecast. The panels correspond to 03Z (left), 11Z (middle), and 19Z (right) on June 16, 2021. The Camino de las Altas Cumbres, a major mountain road, is highlighted in dark blue, traversing the Sierras de Córdoba.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6443854/v1/95390e591889da940d132b5a.png"},{"id":102785462,"identity":"ffab2eab-1b00-45f7-b16b-5dca88f38c7b","added_by":"auto","created_at":"2026-02-16 16:07:04","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2168578,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6443854/v1/0e14f87b-78af-4ac7-9a15-d636a6acd3fb.pdf"},{"id":82605217,"identity":"e0d5716b-66af-47ac-8b49-73ffb1c9a72c","added_by":"auto","created_at":"2025-05-13 10:00:24","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":21810,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-6443854/v1/a261545b384d7bf3b476c26a.docx"},{"id":82606560,"identity":"ccbe893e-d162-4d3c-be71-7dc83539d121","added_by":"auto","created_at":"2025-05-13 10:08:25","extension":"xlsx","order_by":15,"title":"","display":"","copyAsset":false,"role":"supplement","size":214068,"visible":true,"origin":"","legend":"","description":"","filename":"ClimaReporte1606202116062021.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6443854/v1/69e2df37af8d62b9314806cc.xlsx"}],"financialInterests":"","formattedTitle":"A Numerical Weather Prediction Based Road Icing Index for Informed Winter Road Maintenance and Management Decision-Making","fulltext":[{"header":"Introduction","content":"\u003cp\u003eWinter weather events pose significant challenges to road safety and mobility, contributing to increased accident rates, vehicle damage, travel delays, and economic losses (Pisano et al., 2008; American Meteorological Society, 2004, Andreescu \u0026amp; Frost, 1998). The Federal Highway Administration (FHWA) reports that over 24% of weather-related crashes in the United States occur on snowy, slushy, or icy pavement (FHWA, n.d.). Transportation agencies employ a range of strategies, including Road Weather Information Systems (RWIS), pre-treatment with de-icing agents, sanding, plowing, and issuing public advisories, to help mitigate these risks (Crevier \u0026amp; Delage, 2001; Jin et al., 2024).\u003c/p\u003e\n\u003cp\u003eDespite these efforts, interpreting the multitude of meteorological variables that affect road conditions remains a complex task. Traditional numerical weather prediction (NWP) models do not directly provide a unified measure of road icing potential.\u003c/p\u003e\n\u003cp\u003eThe Road Icing Index (RII) addresses this complexity by consolidating key meteorological factors into a single, actionable metric. Building on existing RWIS frameworks (Gustavsson et al., 2007; Boselly et al., 1993), the RII utilizes high-resolution WRF model outputs to evaluate icing risk across spatially continuous domains. By integrating temperature, wet-bulb temperature, precipitation type, humidity, soil temperature, and wind speed, the RII bridges the gap between complex meteorological data and practical road maintenance decisions. This approach is grounded in established research and validated by studies in transportation meteorology (Jin et al., 2024; FHWA, n.d.).\u003c/p\u003e\n\u003cp\u003eThis study details the RII’s methodology, including data sources, parameter weighting, and scoring. A use case illustrates its application during a winter storm event, demonstrating its potential to guide resource allocation and proactive measures for road safety.\u0026nbsp;\u003c/p\u003e"},{"header":"Background and Rationale","content":"\u003cp\u003e\u003cstrong\u003e2.1. Winter Road Condition Forecasting and Decision Support\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWinter road conditions arise from a dynamic interplay of atmospheric and surface processes—most notably temperature, precipitation type, and moisture content (Ishii et al,. 2024). Road Weather Information Systems (RWIS) address these challenges by combining real-time sensor data, mobile observations, and NWP model outputs to guide proactive actions such as anti-icing, salting, and plowing (Gustavsson et al., 2007; Jin et al., 2024).\u003c/p\u003e\n\u003cp\u003eIn many developed nations, RWIS have considerably lowered crash rates. Studies from Sweden and Finland, for instance, highlight the value of pairing RWIS with localized thermal mapping for more efficient resource use (Gopalakrishna \u0026amp; Gestwick, 2019; Gustavsson et al., 2007).\u003c/p\u003e\n\u003cp\u003eDespite their benefits, RWIS primarily yields point-based information, making it difficult to represent spatial variability across an extensive roadway network (FHWA, n.d.). Moreover, rural or under-resourced areas often lack RWIS infrastructure (Kaiser \u0026amp; Barstow, 2022).\u003c/p\u003e\n\u003cp\u003eBy contrast, a gridded Road Icing Index offers a detailed, spatially explicit snapshot of icing risk, thus, enabling prioritization of high-risk corridors by improving the allocation of winter maintenance resources, and facilitating more targeted traveler advisories. To better understand the multi-faceted nature of road icing, the following subsection examines the key atmospheric and surface variables that collectively determine icing risk.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2. Key Variables Influencing Road Icing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e● Air Temperature (T\u003csub\u003e2\u003c/sub\u003e):\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNear-surface air temperature is a fundamental determinant of icing risk. Temperatures below freezing are critical for ice formation, but roads can still become icy at marginally above-freezing temperatures due to latent heat effects (Gustavsson et al., 2007; Jin et al., 2024).\u003c/p\u003e\n\u003cp\u003e● Wet-Bulb Temperature (T\u003csub\u003ew\u003c/sub\u003e):\u003c/p\u003e\n\u003cp\u003eThe wet-bulb temperature (T\u003csub\u003ew\u003c/sub\u003e) is a more dependable measure of icing potential than air temperature alone because it accounts for both temperature and moisture—two factors essential for ice formation. Specifically, Tw reflects the lowest temperature achievable through evaporation and directly considers humidity. When water evaporates into unsaturated air, it draws heat from the surroundings, cooling the air (and any surfaces) below the ambient temperature. If T\u003csub\u003ew\u003c/sub\u003e falls below 0\u003csup\u003e°C\u003c/sup\u003e, the likelihood of ice formation increases.\u003c/p\u003e\n\u003cp\u003eMoreover, humidity significantly affects icing risk. While air temperature alone ignores how moisture interacts with freezing processes, Tw incorporates those effects, indicating whether water will remain liquid, evaporate, or freeze. This integration of temperature, moisture, and latent heat exchange provides a more precise view of the thermodynamic processes that lead to icing, making the wet-bulb temperature a more accurate gauge for predicting when and where ice is likely to form (Theriault \u0026amp; Stewart, 2005).\u003c/p\u003e\n\u003cp\u003e● Precipitation Type and Rate:\u003c/p\u003e\n\u003cp\u003ePrecipitation type (e.g., freezing rain, snow, sleet) is a critical factor influencing road conditions. Freezing rain, in particular, is associated with hazardous glaze ice formation (Stewart et al. 2015). Rates of precipitation are categorized into low, moderate, and high to capture varying icing risks. Low precipitation can still contribute to icing under freezing conditions, while moderate and heavy rates substantially increase the likelihood of road surface accumulation and hazardous conditions (Stewart et al. 2015; Gustavsson et al., 2007).\u003c/p\u003e\n\u003cp\u003e● Relative Humidity (RH\u003csub\u003e2\u003c/sub\u003e):\u003c/p\u003e\n\u003cp\u003eHigh relative humidity supports ice formation by slowing sublimation and retaining moisture on road surfaces. Conversely, low relative humidity allows roads to dry more quickly, reducing icing risk (Jin et al., 2024).\u003c/p\u003e\n\u003cp\u003e● Wind Speed (U\u003csub\u003e10\u003c/sub\u003e, V\u003csub\u003e10\u003c/sub\u003e):\u003c/p\u003e\n\u003cp\u003eWind speed affects evaporative cooling and surface energy exchange. While moderate winds can enhance cooling and promote ice formation, strong winds can reduce icing potential by drying surfaces (Gustavsson et al., 2007).\u003c/p\u003e\n\u003cp\u003e● Top-Layer Soil Temperature (TSLT):\u003c/p\u003e\n\u003cp\u003eSoil temperatures below 0\u003csup\u003e°C\u003c/sup\u003e exacerbate icing potential by facilitating heat loss from the road surface, further increasing the likelihood of surface icing (Karsisto et al., 2016).\u003c/p\u003e"},{"header":"Methodology","content":"\u003cp\u003e\u003cstrong\u003e3.1. Model Configuration and Data Sources\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe RII is derived from atmospheric conditions that can be simulated by numerical weather models such as the Weather Research and Forecasting Advanced Research WRF (WRF-ARW) model, specifically configured to capture key physical processes influencing road icing potential. The model employs physical parameters designed for boundary-layer dynamics, cloud microphysics, and surface energy exchange (Skamarock et al., 2008). These parameters are essential for accurately simulating near-surface atmospheric and surface conditions that govern icing risks. The model was compiled and installed using self-installation scripts (Hatheway et al., 2023).\u003c/p\u003e\n\u003cp\u003e1) \u0026nbsp; Model Output Variables:\u003c/p\u003e\n\u003cp\u003e● 2-Meter Air Temperature (T\u003csub\u003e2\u003c/sub\u003e):\u003c/p\u003e\n\u003cp\u003eRepresents the near-surface air temperature, critical for identifying freezing and near-freezing conditions that lead to road icing. Extracted directly from WRF output, T\u003csub\u003e2\u003c/sub\u003e is provided in degrees Kelvin (K).\u003c/p\u003e\n\u003cp\u003e● 2-Meter Relative Humidity (RH\u003csub\u003e2\u003c/sub\u003e):\u003c/p\u003e\n\u003cp\u003eQuantifies the moisture content in the air and its ability to sustain surface wetness. RH\u003csub\u003e2\u003c/sub\u003e is extracted as a percentage (%) from WRF output.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e● 10-Meter Wind Components (U\u003csub\u003e10\u003c/sub\u003e, V\u003csub\u003e10\u003c/sub\u003e):\u003c/p\u003e\n\u003cp\u003eProvide information on wind speed and direction. These components are used to calculate the resultant wind speed at 10 meters above ground level, in meters per second (m/s).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e● Precipitation Fields (QGRAUP, QSNOW, QRAIN):\u003c/p\u003e\n\u003cp\u003eCapture rain, snow, and other precipitation types in kilograms of water per kilogram of dry air.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e● Temperature Profile Up to 700 mb:\u003c/p\u003e\n\u003cp\u003eCaptures atmospheric conditions critical for precipitation type determination.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNote: In meteorological terms, 1 kgm\u003csup\u003e-2\u003c/sup\u003e of liquid precipitation is equivalent to 1 mm of measurable precipitation (World Meteorological Organization (WMO), 2008).\u003c/p\u003e\n\u003cp\u003e2) Model Output Derived Variables:\u003c/p\u003e\n\u003cp\u003e● Wet-Bulb Temperature (T\u003csub\u003ew\u003c/sub\u003e):\u003c/p\u003e\n\u003cp\u003eIntegrates air temperature and relative humidity to represent the lowest temperature air can reach through evaporation. Calculated using MetPy\u0026apos;s wet_bulb_temperature function (May et al., 2021).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e● Hybrid Precipitation Type Algorithm (Bourgouin/Ramer/Baldwin):\u003c/p\u003e\n\u003cp\u003eEach method predicts a precipitation type (rain, snow, freezing rain, sleet/graupel). If two or more methods agree, that type is assigned. Remaining ambiguities are resolved by wet-bulb temperature checks at the surface. The Bourgouin method excels at detecting warm layers aloft and shallow cold layers near the surface, making it effective for diagnosing freezing rain events with a pronounced melting layer. However, it can struggle in marginal temperature profiles where small changes can lead to misclassification (Bourgouin, 2000). The Ramer method provides a more detailed view of melting/refreezing processes but may be overly sensitive to temperature errors (Ramer, 1993; Theriault \u0026amp; Stewart, 2005). The Baldwin method is computationally straightforward and works well with clear thermal gradients, but can misclassify precipitation under complex temperature structures. By employing an ensemble prediction approach, each methods\u0026rsquo; strengths counteract one another\u0026rsquo;s weaknesses (Baldwin et al., 1994, Theriault \u0026amp; Stewart, 2005), reducing misclassification in marginal layers.\u003c/p\u003e\n\u003cp\u003e3) Computational Tools:\u003c/p\u003e\n\u003cp\u003e● WRF-Python:\u003c/p\u003e\n\u003cp\u003eUtilized to extract model variables, perform diagnostics, and interpolate data onto desired levels.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e● MetPy:\u003c/p\u003e\n\u003cp\u003eProvides unit-aware meteorological calculation tools for derived variables (May et al., 2021).\u003c/p\u003e\n\u003cp\u003e● CartoPy and Matplotlib:\u003c/p\u003e\n\u003cp\u003eEmployed for geospatial data visualization\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e● GeoPandas:\u003c/p\u003e\n\u003cp\u003eFacilitates geographic data manipulation and plotting.\u003c/p\u003e\n\u003cp\u003e4) Uncertainty Quantification:\u003c/p\u003e\n\u003cp\u003eThe WRF model is a powerful tool for simulating atmospheric conditions but does not directly account for road surface conditions or types, which are crucial for accurately predicting road icing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2. Road Icing Index Scoring\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEach meteorological variable is categorized and assigned a score based on thresholds from literature and operational guidelines. The following table (Table 1) summarizes the Road Icing Index (RII) parameters, their thresholds, associated scores, and references. The scores can be customized and adapted to the area of interest, allowing for region-specific adjustments based on local climatology, infrastructure, and operational needs. The formation of ice on road surfaces is governed by a multifaceted interaction of atmospheric and environmental parameters.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAmong the most influential of these is ambient air temperature. Extremely low temperatures markedly increase the probability of ice development, particularly when surface moisture is present. Temperatures near the freezing point also pose a substantial risk, as they facilitate the phase transition of moisture into ice under favorable conditions. Although slightly above-freezing temperatures generally diminish this potential, ice formation may still occur, especially during periods of radiative cooling or in areas with persistent surface moisture. Conversely, temperatures well above freezing are typically insufficient to support ice formation on roadways (Stewart 1992; Gustavsson et al., 2007; Stewart et al. 2015; Jin et al., 2024).\u003c/p\u003e\n\u003cp\u003eWet-bulb temperature, which integrates the effects of ambient temperature and humidity, serves as a critical indicator of the atmosphere\u0026rsquo;s cooling potential. Low wet-bulb values reflect both a high moisture content and a capacity for evaporative cooling, thereby enhancing the likelihood of ice accumulation. As wet-bulb temperatures increase, the efficiency of evaporative cooling diminishes, correspondingly reducing the risk of ice formation. Precipitation type and intensity further modulate road surface conditions. Heavy freezing rain, for example, can generate hazardous glaze ice, while moderate to heavy snowfall or sleet may lead to significant surface accumulation, exacerbating slipperiness. Lighter forms of precipitation exert a comparatively minimal influence unless ambient conditions are near freezing (Stewart 1992).\u003c/p\u003e\n\u003cp\u003eRainfall introduces additional complexity, particularly when accompanied by marginal temperatures. Heavy or moderate rain can saturate road surfaces, and if coinciding with subfreezing or near-freezing conditions, this can significantly elevate the risk of ice formation. Even light rain can contribute to hazardous conditions under the right thermal profile (Pisano et al., 2008; Stewart et al. 2015; Jin et al., 2024). Relative humidity and wind speed also exert notable effects. Elevated humidity supports prolonged moisture retention on road surfaces, thereby facilitating ice development. Wind influence is bidirectional: gentle winds may promote evaporative cooling and moisture persistence, while stronger winds generally enhance surface drying, mitigating ice risk (Thordarson \u0026amp; Olafsson, 2008).\u003c/p\u003e\n\u003cp\u003eFinally, subsurface conditions such as soil temperature play an important thermodynamic role. Colder soil temperatures accelerate the loss of surface heat, increasing the probability of surface freezing. In contrast, warmer soil temperatures act to buffer against rapid heat loss, thereby reducing the propensity for ice formation (Karsisto et al., 2016). Taken together, these factors form a dynamic system in which ice risk is determined by the concurrent state of multiple meteorological and environmental variables. A comprehensive understanding of these interactions is essential for accurate forecasting and effective mitigation of roadway icing events.\u003c/p\u003e\n\u003cp\u003eTable 1. Presents the Road Icing Index (RII) parameters, their thresholds, corresponding scores, and references.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"603\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eParameter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eThreshold\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eScore\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eReferences\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"6\" valign=\"top\"\u003e\n \u003cp\u003eAir Temperature (T\u003csub\u003e2\u003c/sub\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eT2\u0026lt; -5\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"6\" valign=\"top\"\u003e\n \u003cp\u003eStewart 1992; Gustavsson et al., 2007; Stewart et al. 2015; Jin et al., 2024\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026ndash;5\u0026deg;C \u0026le; T2 \u0026le; 0\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u0026deg;C \u0026lt; T2 \u0026le; 3\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u0026deg;C \u0026lt; T2 \u0026le; 5\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5\u0026deg;C \u0026lt; T2 \u0026le; 10\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e10\u0026deg;C \u0026lt; T2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003eWet-Bulb Temperature (T\u003csub\u003ew\u003c/sub\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTw \u0026lt; 0\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003eStewart 1992\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u0026deg;C \u0026le; Tw \u0026le; 2\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTw \u0026gt; 2\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"4\" valign=\"top\"\u003e\n \u003cp\u003ePrecipitation Type\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFreezing Rain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"4\" valign=\"top\"\u003e\n \u003cp\u003ePisano et al., 2008; Stewart et al. 2015 Jin et al., 2024\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSnow or Sleet\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo Precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003ePrecipitation Rate \u0026ndash; Freezing Rain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRate \u0026gt; 7.5 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"9\" valign=\"top\"\u003e\n \u003cp\u003ePisano et al., 2008; Stewart et al. 2015\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.5 mm/hr \u0026lt; Rate \u0026le; 7.5 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRate \u0026le; 2.5 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003ePrecipitation Rate \u0026ndash; Snow/Sleet/Graupel\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRate \u0026gt; 3.0 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.0 mm/hr Rate \u0026le; 3.0 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRate \u0026le; 1.0 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003ePrecipitation Rate \u0026ndash; Rain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRate \u0026gt; 15.0 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.0 mm/hr \u0026lt; Rate \u0026le; 15.0 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRate \u0026le; 5.0 mm/hr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eRelative Humidity (RH\u003csub\u003e2\u003c/sub\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRH2 \u0026ge; 90%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"5\" valign=\"top\"\u003e\n \u003cp\u003eThordarson \u0026amp; Olafsson, 2008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRH2 \u0026lt; 90%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003eWind Speed (U\u003csub\u003e10\u003c/sub\u003e, V\u003csub\u003e10\u003c/sub\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026lt; 10 m/s\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e10 \u0026ndash; 20 m/s\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026gt; 20 m/s\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eTop Layer Soil Temperature (TLST)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTLST \u0026lt; 0\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eKarsisto et al., 2016\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTLST \u0026ge; 0\u0026deg;C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eTo enhance the understanding of how precipitation types are determined within the RII framework, (Table 2) outlines the decision-making process using the hybrid approach that integrates the Bourgouin, Ramer, and Baldwin methods. These structured criteria ensure accurate classification of precipitation types, thereby improving the reliability of the RII.\u003c/p\u003e\n\u003cp\u003eTable 2. Outlines the decision-making process for determining precipitation types using a hybrid approach that leverages the strengths of the Bourgouin, Ramer, and Baldwin methods. This ensemble prediction mechanism enhances the accuracy and reliability of precipitation classification, which is critical for calculating the Road Icing Index (RII).\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"599\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePrecipitation Type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBourgouin Method Criteria\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRamer Method Criteria\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBaldwin Method Criteria\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFinal Classification\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eReferences\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- No significant warm layer aloft.\u003cbr\u003e\u0026nbsp;- Surface temperature \u0026gt; 0\u0026deg;C.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Surface temperature \u0026gt; 0\u0026deg;C.\u003cbr\u003e\u0026nbsp;- No warm layer aloft.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Minimal or no melting/refreezing layer.\u003cbr\u003e\u0026nbsp;- Clear thermal gradients.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIf Bourgouin and Ramer agree, classify it as Rain.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBourgouin, 2000; Ramer, 1993; Baldwin et al., 1994\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFreezing Rain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Presence of a warm layer aloft (\u0026gt; 0\u0026deg;C).\u003cbr\u003e\u0026nbsp;- Shallow cold layer near the surface (\u0026le; 0\u0026deg;C).\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Warm layer aloft (\u0026gt; 0\u0026deg;C).\u003cbr\u003e\u0026nbsp;- Surface temperature \u0026le; 0\u0026deg;C.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Thin melting/refreezing layer.\u003cbr\u003e\u0026nbsp;- Specific thickness and temperature differences.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIf Bourgouin and Ramer both indicate freezing conditions, classify as Freezing Rain.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBourgouin, 2000; Ramer, 1993\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSnow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- No significant warm layer aloft.\u003cbr\u003e\u0026nbsp;- Entire temperature profile \u0026le; 0\u0026deg;C.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Entire temperature profile \u0026le; 0\u0026deg;C.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Consistent below-freezing temperatures.\u003cbr\u003e\u0026nbsp;- No melting layer.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIf all methods consistently indicate below-freezing temperatures without warm layers, classify as Snow.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBourgouin, 2000; Ramer, 1993; Baldwin et al., 1994\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSleet/Graupel\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Presence of a warm layer aloft followed by a cold layer.\u003cbr\u003e\u0026nbsp;- Significant graupel mixing.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Mixed temperature profiles with layers fluctuating around 0\u0026deg;C.\u003cbr\u003e\u0026nbsp;- Presence of graupel.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- Intermediate melting/refreezing processes.\u003cbr\u003e\u0026nbsp;- Variable thermal gradients.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIf Bourgouin, Ramer, and Baldwin collectively indicate mixed precipitation with graupel, classify as Sleet/Graupel.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBourgouin, 2000; Ramer, 1993; Baldwin et al., 1994.,\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- No significant precipitation indicators.\u003cbr\u003e\u0026nbsp;- Surface conditions dry.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- No precipitation detected.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e- No melting/refreezing layers.\u003cbr\u003e\u0026nbsp;- Clear skies or minimal moisture.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIf none of the methods detect precipitation conditions, classify as None.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJin et al., 2024\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3. Total Road Icing Index Score and Risk Categories\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe total RII score is calculated by adding individual variable scores together, as visualized in Fig. 1. and is categorized as follows:\u003c/p\u003e\n\u003cp\u003e0\u0026ndash;3: Minimal Risk \u0026ndash; No significant icing threat; minimal intervention needed.\u003c/p\u003e\n\u003cp\u003e3\u0026ndash;6: Low Risk \u0026ndash; Minor icing potential; routine preventive measures may be sufficient.\u003c/p\u003e\n\u003cp\u003e6\u0026ndash;9: Moderate Risk \u0026ndash; Higher icing potential; proactive maintenance and monitoring are advised.\u003c/p\u003e\n\u003cp\u003e9\u0026ndash;12: High Risk \u0026ndash; High likelihood of road icing; immediate maintenance is required.\u003c/p\u003e\n\u003cp\u003e\u0026ge; 12: Severe Risk \u0026ndash; Severe icing threat; extensive and urgent mitigation is essential.\u003c/p\u003e\n\u003cp\u003eFig 1 provides a visual representation of the RII calculation workflow, including ensemble voting for precipitation type classification. The flowchart outlines the sequence of steps, starting from input data extraction to the final determination of risk levels. The integration of voting ensures that multiple classification methods contribute to the accuracy of the precipitation type assessment, forming a core component of the RII calculation process.\u003c/p\u003e\n\u003cp\u003eA heuristic approach is a simplified rule-based method\u0026mdash;often referred to as a \u0026ldquo;rule-of-thumb\u0026rdquo; strategy\u0026mdash;that delivers practical, detection and solutions without incurring the complexity of advanced or purely data-driven algorithms. In the context of road icing prediction, heuristics generally establish threshold values and weighting factors for each meteorological parameter (e.g., air temperature, wind speed, and precipitation rate) based on physical principles, empirical observations, and operational practice.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis design ensures computational simplicity by avoiding nested calculations or resource-intensive machine learning models, making real-time deployment more feasible and useful to the agencies responsible for road conditions. At the same time, it preserves operational relevance because thresholds can be readily calibrated to match local climatological characteristics. For instance, a region prone to freezing rain can adopt lower temperature thresholds compared to one more frequently affected by heavy snow. While heuristics may not always achieve the same level of precision as highly sophisticated modeling techniques, their transparency, reliability, and adaptability render them well-suited for day-to-day road maintenance decisions.\u003c/p\u003e"},{"header":"Impact of Treated and Untreated Roads on the RII","content":"\u003cp\u003e\u003cstrong\u003e4.1. Treated Roads\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTreatments such as salting, sanding, and chemical de-icing lower the freezing point of water, reducing the likelihood of road surface icing (Crevier \u0026amp; Delage, 2001). By preventing or delaying ice formation, these treatments effectively alter the thermal and moisture balance at the road surface, thereby lowering the RII’s practical impact—even if meteorological parameters (e.g., sub-freezing temperatures, precipitation) remain unchanged. For instance, calcium chloride is effective down to approximately –20°C, making it valuable in severe winter conditions (FHWA, n.d.). \u0026nbsp; However, treatment effectiveness can vary based on several factors:\u003c/p\u003e\n\u003cp\u003eTemperature Extremes – Extremely cold conditions can slow chemical reactions, reducing melting capability (Crevier \u0026amp; Delage, 2001).\u003c/p\u003e\n\u003cp\u003ePrecipitation Intensity – Heavy or sustained precipitation may wash away chemicals, necessitating reapplication (FHWA, n.d.).\u003c/p\u003e\n\u003cp\u003eTraffic Volume – High traffic distributes de-icing chemicals but can also deplete them faster.\u003c/p\u003e\n\u003cp\u003eApplication Timing – Pre-treatments are most effective if applied prior to the onset of precipitation.\u003c/p\u003e\n\u003cp\u003eEven treated roads may register high RII scores if chemicals are overwhelmed by intense precipitation rates, rapidly dropping temperatures, heavy traffic, or prolonged events. Monitoring these factors helps agencies tailor responses, such as recommending additional salting or sanding.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.2. Untreated Roads\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUntreated roads, especially in rural or low-maintenance regions, rely entirely on meteorological and surface conditions to dictate icing risk. When near-freezing temperatures align with precipitation, roads can quickly shift from wet to icy, creating unexpected hazards for motorists.\u003c/p\u003e\n\u003cp\u003eRapid Response – Deploying maintenance units to proactively treat priority segments..\u003c/p\u003e\n\u003cp\u003eTraveler Warnings – Using message boards, media outlets, social media, and applications to advise reduced speeds and alternative routes.\u003c/p\u003e\n\u003cp\u003ePreventive Measures – Instituting variable speed limits or temporary closures in high-risk areas.\u003c/p\u003e\n\u003cp\u003eFuture work can refine the RII by incorporating and analyzing historical real-time treatment logs and traffic accident records to check the RII accuracy. If a segment was recently salted under temperatures that favor the chemical’s effectiveness, the RII score could be slightly reduced. Conversely, if no maintenance has occurred despite high precipitation and sub-freezing temperatures, the RII may remain elevated. Such treatment-aware modeling ensures more efficient resource allocation and better reflection of on-the-ground conditions.:\u003c/p\u003e\n\u003cp\u003eEven treated roads may register high RII scores if chemicals are overwhelmed by intense precipitation rates, rapidly dropping temperatures, heavy traffic, or prolonged events. Monitoring these factors helps agencies tailor responses, such as recommending additional salting or sanding.\u003c/p\u003e"},{"header":"Use Case: Application to a Winter Storm Event","content":"\u003cp\u003eAs an example of the application of the Road Icing Index (RII), a cold weather event that occurred in the central region of Argentina was simulated using the Weather Research and Forecasting (WRF) model. The RII successfully identified critical icing zones within the affected areas, particularly highlighting regions forecasted to experience freezing rain and near-surface freezing conditions. The highest RII scores were concentrated along mountain roads, where the risk of hazardous driving conditions was most significant.\u003c/p\u003e\n\u003cp\u003eTo further validate the RII predictions, the outputs were compared against data from the network of meteorological stations in the province of C\u0026oacute;rdoba (Argentina) and reports from news sources. These comparisons provided an observational reference for assessing the accuracy of RII predictions regarding near-surface temperature drops and road icing conditions. Notably, the RII was able to detect the imminent risk of freezing along the Camino de las Altas Cumbres, a heavily trafficked mountain road that connects key regions and is frequently used by drivers. This detection highlights the RII\u0026apos;s operational utility in identifying high-risk areas that are critical for transportation safety.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5.1. Description of the area of application\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study area corresponds to the province of C\u0026oacute;rdoba, located in central Argentina, covering a surface area of 165,321 km\u0026sup2; (see Fig. 3.). This extensive territory features a diverse landscape, including plains, mountain ranges, and valleys, which shape its unique geographic identity. The eastern part of the province consists of a vast plain with a gentle slope towards the east. In contrast, the western region is characterized by three mountain ranges collectively known as the Sierras de C\u0026oacute;rdoba. The easternmost range, Sierras Chicas Mountains, has elevations ranging from around 1,000 m in the southern end to 1,950 m at Cerro Uritorco in the northern end, both measured above sea level. Further west, two significant valleys\u0026mdash;Calamuchita to the south and Punilla to the north\u0026mdash;play an essential role in the regional economy, primarily due to tourism. Beyond these valleys, the next mountain range includes Champaqu\u0026iacute; Hill, the highest peak in the province at 2,880 m. A key route that traverses the mountainous region of the Sierras de C\u0026oacute;rdoba is the Camino de las Altas Cumbres (Altas Cumbres Road), a major roadway widely used for transportation and tourism. However, due to its high-altitude location, this route is particularly susceptible to adverse weather conditions, especially during winter. Low temperatures frequently cause ice formation and freezing conditions, leading to hazardous driving situations and temporary road closures.\u003c/p\u003e\n\u003cp\u003eThe climate in C\u0026oacute;rdoba is temperate to subtropical, with a distinct dry season. Both temperature and precipitation generally decrease from north to south and from east to west, except along the eastern slopes of the mountains, where humid easterly winds contribute to higher rainfall levels. Rainfall is seasonal, occurring mainly between October and April. Temperature variations are notable throughout the year. In summer, C\u0026oacute;rdoba experiences warm conditions, with average maximum temperatures around 30\u0026deg;C and peaks that can exceed 40\u0026deg;C. Winters, on the other hand, are mild and relatively dry. The average temperature during this season ranges between 10\u0026deg;C and 12\u0026deg;C, with maximum values around 18\u0026deg;C and minimum temperatures between 4\u0026deg;C and 5\u0026deg;C. A significant winter feature is the occurrence of snowfall, particularly in the higher-altitude areas of the province.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5.2. Data Collection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMeteorological variables necessary for RII calculation were obtained from WRF model outputs corresponding to the freezing event. Key atmospheric and surface parameters\u0026mdash;including near-surface air temperature, wet-bulb temperature, precipitation type and rate, relative humidity, and wind speed\u0026mdash;were extracted for further analysis. Given that these events were characterized by significant cloud cover, satellite-derived Land Surface Temperature (LST) data from GOES-16 could not be utilized, as the cloud coverage obstructed LST observations. Instead, the validation relied on temperature, dew point, relative humidity, precipitation and wind speed readings from the meteorological station network in C\u0026oacute;rdoba and information extracted from news reports that highlighted the severity of the event. The total number of available meteorological stations in C\u0026oacute;rdoba is approximately 250, from which only those that recorded data during the studied event were used.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5.3. WRF Model Configuration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe WRF model version 4.2 was used (Skamarock et al., 2008). Initial conditions were provided by the GFS global model with a horizontal resolution of 0.25 degrees. The WRF domain consists of 270\u0026times;270 grid points with a horizontal resolution of 4 km and 35 vertical levels. The physical parameterizations used were: Thompson microphysics (option 8), Mellor\u0026ndash;Yamada\u0026ndash;Janjic planetary boundary layer scheme (option 2), Dudhia shortwave radiation scheme (option 2), RRTM longwave radiation scheme (option 1), Unified Noah land surface model (option 2), and Eta similarity surface layer scheme (option 2).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5.4. RII Calculation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA Python script was developed to integrate all relevant atmospheric and surface variables from the WRF outputs to compute the RII (see Appendix B). This script generated spatially explicit RII maps, pinpointing regions with elevated road icing risks during both events.\u003c/p\u003e\n\u003cp\u003eOn June 16, 2021, an extreme cold event affected the central region of Argentina, particularly C\u0026oacute;rdoba, leading to subzero temperatures. This unusual weather phenomenon resulted from a strong cold air mass intrusion, causing widespread frost and significant temperature drops across the region.\u003c/p\u003e\n\u003cp\u003eFig. 3. shows the meteorological variables measured at the surface by the network of weather stations on June 16, 2021 at 11Z, with black dots representing the locations of the weather stations. The thick solid line indicates the political boundary of the province of C\u0026oacute;rdoba, while the thin lines represent contour lines at 1000-meter intervals above sea level, providing a better reference for high-altitude areas. To enhance visualization, the measured data were interpolated, and a characteristic color map was applied for each variable to approximate its distribution across the territory.\u003c/p\u003e\n\u003cp\u003eTemperatures below 6\u0026deg;C were observed across most of the region. In particular, the highest areas of the region exhibit the lowest temperatures, reaching values between -4\u0026deg;C and -6\u0026deg;C. The dew point temperature and relative humidity maps indicate that the higher-altitude regions are the areas closest to saturation. The precipitation map shows that low precipitation rates occurred at certain localized points in the region. The wind speed map reveals that low wind speeds were recorded in high-altitude areas, which could increase the chance of the road freezing.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003eFig. 4. shows the RII calculated 24 hours in advance using the methodology described in this paper. Three panels are presented at 03Z, 11Z, and 19Z to capture the temporal evolution of the index: the night before (marking the onset of temperature decline), the morning (when the risk is at its peak), and the afternoon (when the risk diminishes due to rising temperatures). As shown in this figure, the RII obtained from the WRF model and the procedure explained in this paper reveals that the areas with the highest risk of freezing are the highest-altitude regions of the territory. This finding aligns with the observed meteorological variables and news reports, which indicated that the roads traversing the high-altitude areas of the region were the ones affected by freezing. The following links correspond to official reports and news articles in Spanish about the event: https://prensa.cba.gov.ar/informacion-general/historica-nevada-en-cordoba-por-que-se-dio-este-impactante-fenomeno/ and https://www.minutouno.com/sociedad/nieve/frio-polar-el-pais-las-mejores-fotos-y-videos-la-nevada-cordoba-n5202922.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5.5. Results Interpretation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe results showed that areas with high RII values closely corresponded to the actual regions that experienced significant road icing and higher accident rates. Importantly, the RII successfully identified critical icing risks on the Camino de las Altas Cumbres, a vital mountain route heavily used by drivers. This capability demonstrates the RII\u0026rsquo;s operational relevance, providing actionable insights for optimizing winter road maintenance activities and improving driver safety during cold weather events.\u003c/p\u003e"},{"header":"Discussion and Conclusion","content":"\u003cp\u003eThe RII offers a scalable, actionable metric for assessing road icing potential, particularly in regions lacking extensive RWIS infrastructure. Its integration of multiple meteorological parameters ensures a comprehensive evaluation of icing risks. Moreover, the RII can be customized to specific regions by adjusting the scoring criteria for each variable, allowing for better alignment with local climatological and geographical conditions. This use case illustrates the RII’s capability, showing its value in guiding maintenance decisions. In particular, this use case applied to an extreme cold weather event in Córdoba, along the Altas Cumbres Road, a high-mountain route with significant traffic flow, demonstrates that the RII is capable of identifying areas with a high likelihood of road surface freezing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e6.1. Practical Implementation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe RII can be easily incorporated into existing transportation management systems, delivering real-time updates and forecasts. Visualization dashboards can help decision-makers apply proactive measures and allocate resources effectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e6.2. Validation and Future Work\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFurther validation against broader observational datasets is warranted. Future efforts may include integrating pavement temperature models, refining precipitation-type classification, and employing real-time maintenance logs to dynamically adjust the RII (Jin et al., 2024; Theriault \u0026amp; Stewart, 2005; FHWA, n.d.).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e6.3. Economic and Policy Implications\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMore accurate road icing indices can yield substantial economic benefits by optimizing maintenance operations helping to reduce crash-related expenses, and enhance public safety (Wu et al., 2021). Policymakers can use the RII to inform infrastructure investments and develop holistic winter road management strategies. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe RII’s accuracy depends on numerical weather model outputs, which may be affected by model resolution and the parameterization schemes used. While this study utilizes the WRF model, these outputs can also be obtained from other forecasting models. Additionally, the scoring framework, though effective, may require localized calibration to account for regional climate and infrastructure.\u003c/p\u003e\n\u003cp\u003eOverall, the Road Icing Index represents a significant advance in winter road condition forecasting. Its successful application in the use case underscores its potential to enhance road safety and operational efficiency. Continued refinement and validation will further solidify the RII as a vital tool in winter transportation meteorology.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments:\u0026nbsp;\u003c/strong\u003eThe authors want to thank the anonymous reviewers for helping to improve the quality of the manuscript. Mat\u0026iacute;as Ezequiel Su\u0026aacute;rez gratefully acknowledges the PhD scholarship awarded by the National Scientific and Technical Research Council of Argentina (CONICET).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor\u0026rsquo;s contributions:\u003c/strong\u003e\u0026nbsp; WH designed and drafted the manuscript and initially came up with the first attempt of the calculations. WH and MS conceived and coordinated the study, as well as refined the calculations. \u0026nbsp;MS ran the use study. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e\u0026nbsp; The research authors did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e:\u003c/p\u003e\n\u003cp\u003eAll data used in this study are generated and archived by Matias Ezequiel Suarez of the National University of C\u0026oacute;rdoba. The meteorological station network data were downloaded from the following website: https://cordoba.redesclimaticas.com/, which requires prior registration, but access is free of charge. The calculations and python code used to generate the graphics are adapted from the WRF-Python public website (https://wrf-python.readthedocs.io/en/latest/plot.html). \u0026nbsp;Calculations used and scoring for the Road Icing Index are found in the appendix below.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interests:\u003c/strong\u003e The authors have no conflicts of interest to declare that are relevant to the content of this article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u003c/strong\u003e\u0026nbsp; Not applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAmerican Meteorological Society, 2004: Weather and highways: Report of a policy forum. Atmospheric Policy Program, https://www.ametsoc.org/sites/ams/assets/File/policy/WxHighways_2004.pdf.\u003c/li\u003e\n\u003cli\u003eAndreescu, M. P., and D. B. Frost, 1998: Weather and traffic accidents in Montreal, Canada. Climate Res., 9(3), 225\u0026ndash;230, https://doi.org/10.3354/cr009225.\u003c/li\u003e\n\u003cli\u003eBaldwin, M., R. Treadon, and S. Contorno, 1994: Precipitation type prediction using a decision tree\u0026nbsp;\u0026nbsp;approach with NMC\u0026rsquo;s Mesoscale Eta Model. Preprints, Tenth Conf. on Numerical Weather\u0026nbsp;Prediction, Amer. Meteor. Soc., Portland, OR, 30\u0026ndash;31.\u003c/li\u003e\n\u003cli\u003eBoselly, S. E., J. Thornes, C. Ulberg, and D. Ernst, 1993: Road weather information systems, Volume I. Strategic Highway Research Program Publication-SHRP-H-350, Natl. Res. Council, Washington, DC, 90\u0026ndash;93, https://onlinepubs.trb.org/onlinepubs/shrp/shrp-h-350.pdf.\u003c/li\u003e\n\u003cli\u003eBourgouin, P., 2000: A method to determine precipitation types. Wea. Forecasting, 15(5), 583\u0026ndash;592,\u0026nbsp; https://doi.org/10.1175/1520-0434(2000)015\u0026lt;0583:AMTDPT\u0026gt;2.0.CO;2.\u003c/li\u003e\n\u003cli\u003eCrevier, L. P., and Y. Delage, 2001: METRo: A new model for road-condition forecasting in Canada J. Appl. Meteor., 40(11), 2026\u0026ndash;2037,\u0026nbsp;\u0026nbsp;https://doi.org/10.1175/1520-0450(2001)040\u0026lt;2026:MANMFR\u0026gt;2.0.CO;2.\u003c/li\u003e\n\u003cli\u003eFederal Highway Administration (FHWA), n.d.: Weather impacts on roads.\u0026nbsp;\u0026nbsp;https://ops.fhwa.dot.gov/weather.\u003c/li\u003e\n\u003cli\u003eGopalakrishna, D., and T. Gestwick, 2019: Road Weather Management Performance Measures\u0026nbsp;\u0026nbsp;Update (FHWA-HOP-19-089). Federal Highway Administration, U.S. Dept. of\u0026nbsp;\u0026nbsp;Transportation, https://ops.fhwa.dot.gov/publications/fhwahop19089/fhwahop19089.pdf.\u003c/li\u003e\n\u003cli\u003eGustavsson, T., and J. Bogren, 2007: Information not data: Future development of road weather information systems. Geogr. Ann. A, 89(4), 263\u0026ndash;271, https://doi.org/10.1111/j.1468-0459.2007.00325.x.\u003c/li\u003e\n\u003cli\u003eHatheway, W., H. Snoun, H. ur Rehman, and others, 2023: WRF-MOSIT: A modular and cross-platform tool for configuring and installing the WRF model. Earth Sci. Inform., 16,\u0026nbsp;\u0026nbsp;4327\u0026ndash;4336, https://doi.org/10.1007/s12145-023-01136-y.\u003c/li\u003e\n\u003cli\u003eIshii, K., S. Ono, T. Masago, M. Ishizuki, T. Mori, and Y. Hanatsuka, 2024: Fine-scaled predictive modeling of road surface conditions and temperature in urban areas. IEEE Trans. Intell. Transp. Syst., https://doi.org/10.1109/TITS.2024.3433004.\u003c/li\u003e\n\u003cli\u003eJin, M., and D. G. McBroom, 2024: Investigating road ice formation mechanisms using Road Weather Information System (RWIS) observations. Climate, 12(5), 63, https://doi.org/10.3390/cli12050063.\u003c/li\u003e\n\u003cli\u003eKaiser, N., and C. K. Barstow, 2022: Rural transportation infrastructure in low-and middle-income countries: A review of impacts, implications, and interventions. Sustainability, 14(4), 2149, https://doi.org/10.3390/su14042149.\u003c/li\u003e\n\u003cli\u003eKarsisto, V., P. Nurmi, M. Kangas, M. Hippi, C. Fortelius, S. Niemel\u0026auml;, and H. J\u0026auml;rvinen, 2016: Improving road weather model forecasts by adjusting the radiation input. Meteor. Appl., 23(3), 503\u0026ndash;513, https://doi.org/10.1002/met.1574.\u003c/li\u003e\n\u003cli\u003eMay, R., B. Reed, and G. Thomas, 2021: MetPy: A Python package for meteorological calculations. https://unidata.github.io/MetPy/latest/.\u003c/li\u003e\n\u003cli\u003ePisano, P. A., L. C. Goodwin, and M. A. Rossetti, 2008: US highway crashes in adverse road weather conditions. 24th Conf. on International Interactive Information and Processing Systems for Meteorology, Oceanography and Hydrology, New Orleans, LA, Amer. Meteor. Soc., 20\u0026ndash;24.\u003c/li\u003e\n\u003cli\u003eRamer, J., 1993: An empirical technique for diagnosing precipitation type from model output. 5th Intl. Conf. on Aviation Weather Systems, Vienna, VA, 227\u0026ndash;230.\u003c/li\u003e\n\u003cli\u003eSkamarock, W. C., J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, W. Wang, and J. G. Powers, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, https://doi.org/10.5065/D68S4MVH.\u003c/li\u003e\n\u003cli\u003eStewart, R. E., 1992: Precipitation types in the transition region of winter storms. Bull. Amer. Meteor. Soc., 73(3), 287\u0026ndash;296, https://doi.org/10.1175/1520-0477(1992)073\u0026lt;0287:PTITTR\u0026gt;2.0.CO;2.\u003c/li\u003e\n\u003cli\u003eStewart, R. E., J. M. Th\u0026eacute;riault, and W. Henson, 2015: On the characteristics of and processes producing winter precipitation types near 0\u0026deg;C. Bull. Amer. Meteor. Soc., 96(4), 623\u0026ndash;639, https://doi.org/10.1175/BAMS-D-14-00032.1.\u003c/li\u003e\n\u003cli\u003eTamang, S. K., A. M. Ebtehaj, A. F. Prein, and A. J. Heymsfield, 2019: On changes of global wet-bulb temperature and snowfall regimes. arXiv, https://doi.org/10.48550/arXiv.1905.07776.\u003c/li\u003e\n\u003cli\u003eTheriault, J., and R. Stewart, 2005: Winter precipitation types and icing at the surface. Proc. 11th Intl. Workshop on Atmos. Icing of Structures, 6.\u003c/li\u003e\n\u003cli\u003eThordarson, S., and B. Olafsson, 2008: Weather induced road accidents, winter maintenance and user information. Transport Research Arena Europe 2008, 72.\u003c/li\u003e\n\u003cli\u003eWMO (World Meteorological Organization), 2008: Guide to Meteorological Instruments and Methods of Observation. Weather‐Climate‐Water.\u003c/li\u003e\n\u003cli\u003eWu, J., S. Yang, F. Yang, and X. Yin, 2021: Road weather monitoring system shows high cost-effectiveness in mitigating malfunction losses. Sustainability, 13(22), 12437, https://doi.org/10.3390/su132212437.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nhaz","sideBox":"Learn more about [Natural Hazards](https://www.springer.com/journal/11069)","snPcode":"11069","submissionUrl":"https://submission.nature.com/new-submission/11069/3","title":"Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6443854/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6443854/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eTimely and accurate assessment of road surface conditions during winter weather events is critical for ensuring safety, reducing crash rates, and optimizing roadway maintenance activities. To address this need, we propose the Road Icing Index (RII), a new metric that identifies regions with elevated road icing risk based on outputs from numerical weather models such as the Weather Research and Forecasting (WRF) model. The RII integrates multiple atmospheric parameters—including near-surface air temperature (T\u003csub\u003e2\u003c/sub\u003e), wet-bulb temperature (T\u003csub\u003ew\u003c/sub\u003e), precipitation type and rate (QRAIN, QSNOW, RAINNC), relative humidity (RH\u003csub\u003e2\u003c/sub\u003e), wind speed (U\u003csub\u003e10\u003c/sub\u003e, V\u003csub\u003e10\u003c/sub\u003e), and top-layer soil temperature (TLST)—into a single spatially explicit metric. These variables are validated by studies such as Tamang et al., (2019) for wet-bulb temperature significance, Stewart et al. (2015) for precipitation type impacts, and Gustavsson et al. (2007) for road weather modeling.\u003c/p\u003e\n\u003cp\u003eDrawing on established meteorological principles, transportation meteorology research, and operational insights, this paper outlines the RII’s methodological framework, implementation process, and applications. In addition, a use case demonstrates its utility in identifying areas of heightened icing potential, offering practical guidance for road maintenance decision-makers and enhancing traveler safety. Future work aims to incorporate vehicle-based observations, advanced precipitation-type algorithms, and refined parameter weighting to improve forecast accuracy and decision-making.\u003c/p\u003e","manuscriptTitle":"A Numerical Weather Prediction Based Road Icing Index for Informed Winter Road Maintenance and Management Decision-Making","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-13 10:00:19","doi":"10.21203/rs.3.rs-6443854/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Minor revisions","date":"2025-08-25T07:15:40+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"","date":"2025-05-14T22:05:59+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-05-07T11:43:05+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-04-16T11:51:51+00:00","index":"","fulltext":""},{"type":"submitted","content":"Natural Hazards","date":"2025-04-15T07:53:34+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"natural-hazards","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nhaz","sideBox":"Learn more about [Natural Hazards](https://www.springer.com/journal/11069)","snPcode":"11069","submissionUrl":"https://submission.nature.com/new-submission/11069/3","title":"Natural Hazards","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"ae0d2c42-e0e7-4f7a-8fcf-6850ec4cbc12","owner":[],"postedDate":"May 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-02-16T16:04:15+00:00","versionOfRecord":{"articleIdentity":"rs-6443854","link":"https://doi.org/10.1007/s11069-025-07883-z","journal":{"identity":"natural-hazards","isVorOnly":false,"title":"Natural Hazards"},"publishedOn":"2026-02-11 15:58:43","publishedOnDateReadable":"February 11th, 2026"},"versionCreatedAt":"2025-05-13 10:00:19","video":"","vorDoi":"10.1007/s11069-025-07883-z","vorDoiUrl":"https://doi.org/10.1007/s11069-025-07883-z","workflowStages":[]},"version":"v1","identity":"rs-6443854","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6443854","identity":"rs-6443854","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}