PROCEEDINGS OF THE 18TH SIRWEC CONFERENCE, FT. COLLINS, COLORADO, USA
27th – 29th APRIL 2016
TOPIC 1: ROAD WEATHER MANAGEMENT SYSTEMS
The ASSIST Feasibility Study intends to investigate a suite of potential high-end innovative services aiming to support the execution of winter road maintenance (in different operational scenarios, e.g. snow plowing and gritting) and also to provide an effective assistance to the drivers involved in these as well as the management overseeing the operations. These services are enabled by a robust and accurate real-time positioning of the vehicles based on satellite navigation, e.g. GPS and Galileo – and by road geometry information and road weather forecast – through integrating Earth Observation with in-situ technologies.
The architectural design goes in direction to realize an end-to-end solution taking advantages on the space assets domain for advanced winter service provisioning. ASSIST strategy proposed to build up a high fidelity and updated model of the road to clear collecting acquired by external service providers in a cloud-computing platform. Such model allows evaluating analytically the geo localized work parameters (both in terms of salt spreading and snow plow control) in order to maximize the effectiveness of the winter maintenance treatment. The application of such parameters is in charge to a NAV\COM intelligent device mounted on the truck, called On Board Unit (OBU)) that is able to automatically control the mechanical equipment for the effective actuation.
Thanks to the conceived intelligent system architecture, ASSIST solution integrates additional function blocks (e.g. tracking\telemetry of the trucks, Intelligent Device Management) that permits to provide a comprehensive support to the overall winter maintenance activity.
The ASSIST Feasibility Study has been concluded in May 2015. A suite of innovative space-based "assistance services" in supporting the major operational scenarios for the winter maintenance has been deeply investigated with positive results. The technical feasibility and the users' acceptance of such services have been assessed through an actual Proof of Concept that took place in the 2014-2015 winter season as an extensive field test in Sweden and a showcase in Norway, with a direct and active involvement of two Nordic public administration and their contractors.
The viability analysis performed during the Feasibility Study confirmed the ASSIST potential, showing benefits coming from its adoption for all key actors of the winter maintenance value chain. The ten years business plan also showed clear opportunities for all partners, especially for Giletta S.p.A. as equipment manufacturer acting as "ASSIST service provider". Following the positive feedback on both the technical feasibility and the market attractiveness, the same partnership is now fully committed in a 24-month Demonstration Project by ESA, with a start 2016, in sight of the roll out of full operational "assistance services" in 4 different European countries; Sweden, Norway, Austria and Italy.
These numbers continue to tell a story that road weather affects the safety and mobility of the traveling public. In order to mitigate these road weather impacts, near real-time atmospheric and road weather observations must be available; decision support tools must be utilized; and impact messages must be disseminated. With the advent of connected vehicle technology, a wealth of data will be made available. Most, however, will need to be converted into usable information for the road weather community, decision support tools, connected and automated vehicles, and the traveling public.
In this presentation, we will discuss the following:
Pathfinder Project: A collaborated effort between departments of transportation, National Weather Service, and private sector weather provides to create a collaborated impact message for the traveling public.
Integrated Modeling for Road Condition Prediction: A tool that incorporates real-time and/or archived data plus ensembles of forecast and probabilistic models then applies algorithms to predict real-time and forecasted road conditions.
Road Weather Performance Management Tool: A tool that utilizes transportation data, weather and road weather data, and connected vehicle data to determine the road condition and notify drivers inside their vehicle about slick conditions, backups (queue) and speed recommendations.
On one side a comprehensive winter maintenance management system needs high quality inputs from weather services and Road Weather Information Systems, which shows the road manager the different situations of weather and road surface conditions in the road network.
That gives on the other side the decision maker the possibility to plan the next winter service operations, to decide when and with how many vehicles to start the necessary treatments. Modern AVL systems show the road manager all vehicles in operation on a map.
Both together, RWIS and vehicle operation data are the basis of a comprehensive winter maintenance management system and to manage winter service in an efficient way. But nowadays there is more and more the need to be able to have access to the data everywhere and every time. Therefore it is necessary to visualize the most important data of RWIS and vehicles live on smartphones and tablets to gives the responsible persons the possibility to combine and control the actual situation and the measurement activities.
An efficient solution is the RWIS-App, which combines live data of RWIS station with an integrated graph of the last 2 hours and the service vehicles in operation, showing different treatments like spreading, plowing, mowing or sweeping on an interactive Google-Map. That tool gives responsible staff in road service the possibility to have an overview about the actual situation in their road network wherever they are, on road, in office or travelling at all times and to manage necessary treatments. The RWIS App is used for operations on roadside as well as on airports and is very welcome by personnel as an additional tool next to the standard winter maintenance management system.
The operational data collected by MDOT's AVL system is processed and automatically fed into MDSS. The goal of MDSS is to provide a decision-support tool for MDOT staff involved in winter maintenance operations. MDSS recommends maintenance treatments, application rates, and suggested times to apply material to maximize its effectiveness for the snowplow operators. These route-specific treatment recommendations are provided to maintenance garage supervisors on a website as well as transmitted directly to a screen within the truck cab for plow operators to view in near-real time.
The use of AVL to update the actions performed by maintenance operations enhances the validity and accuracy of the MDSS treatment recommendations. In addition to live maintenance information, the MDSS provides decision maker's with treatment strategies, taking into account the incoming weather situations on a route-specific basis.
Now that the AVL/GPS/MDSS project has been implemented and available across the state, MDOT direct garages are finding ways to incorporate the tools from the project as standard practice in winter operations. Using the resources available through MDSS, maintenance garage supervisors can view road condition forecasts for specific snow routes, as opposed to the regional atmospheric forecasts provided by other sources of weather information, to optimize staff and resource deployment planning. The focused nature of the localized forecasts enable MDOT to become proactive in identifying and responding to troublesome areas instead of being reactive during the winter storm event. MDSS forecasts are also being used to assist in the staffing decisions by allowing supervisors to call in more or less operators several hours sooner to formulate an operational plan ahead of the storm rather than an hour or two after the storm started.
In addition to improving winter operation efficiencies, MDOT is looking at ways of disseminating winter operation information to the public through a pilot to display snowplow information on the state's traveler information site, MiDrive. MDOT is continually engaging stakeholders and looking for ways to incorporate relevant information (such as CCTV cameras and ESS sites) all in one location as a tool for winter operations. Looking to the future, MDOT plans to expand access to weather information and maintenance recommendations to include areas of the state maintained by contract county agencies.
TOPIC 2: NOVEL TECHNOLOGIES IN ROAD WEATHER
Winter maintenance engineers base their nightly decision making by consulting a Road Weather Information System (RWIS) which combines weather forecast data with road temperature and condition data. The latest generation of RWIS is based on route based forecasts which take into account how the local geography interacts with the regional climate to produce a detailed model of road surface temperatures for every 50m section of road. By knowing which sections of road are likely to fall below the 0°C threshold on a night-by-night basis, highway engineers can selectively treat just the affected routes and thus make significant savings in salt usage. However, this is presently not happening. In an environment of increasing litigation, practitioners are nervous about making decisions based on model output as opposed to ground truth. This now means that the verification of route based road weather forecasts is urgently needed at an unprecedented scale. Research at the University of Birmingham seeks to solve this problem by producing and deploying a new generation of low cost, internet enabled, road surface temperature sensors embedded within an Internet of Things ecosystem. This approach will not only provide a monitoring and verification solution, but also has the potential to form the basis of a new generation of ‘nowcasting’ (i.e. forecast for the next 3 hours) models for winter road maintenance. This paper investigates the feasibility of harnessing the emerging Internet of Things (IoT) to develop a high resolution, but low cost, road surface temperature monitoring network. It focuses on the description and evaluation of a low-cost (<$200), self-contained sensor, which has been tested with positive results in both a lab and field setting. Hide Abstract
In this study five different sensors, two stationary and three mobile, were tested in a laboratory environment. This was done in order to better understand how reliable such sensors are, and to get an idea of how consistent results different sensors give compared to each other. The parameters that we compared were the classified contaminant as well as the derived friction and water-film thickness. Measurements were made on two different asphalt substrates, one with old gray and one with new black asphalt. These substrates were prepared with different amounts of water and ice, as well as different types and densities of snow.
The results for the road condition classification showed that the two stationary sensors performed very well, and correctly managed to classify all “simulated” contaminants. The mobile sensors were in general good, but they had some issues distinguishing between liquid water and ice. One challenge for optical classification of road conditions seems to be snow. Snow, ranging from loose fluffy snow to very hard dense snow reminiscent of ice, was produced in the study. While these two extremes clearly present very different road conditions, all tested sensors classified them both as snow with similar friction values.
Taking misclassifications into account, the derived friction from all sensors agreed well with each other. To actually validate the friction values, breaking tests in the field would be required, but in most cases the values obtained from the sensors seemed reasonable. The exception is the above-mentioned case with very dense snow, where all sensors gave a much too high friction value.
Water-film thickness was highly variable between the sensors; the results varied by a factor of 2 – 3. Naturally, this meant that the accuracy was limited, correspondingly. Absolute numbers aside, all sensors could register increased amounts of water on the substrate, as long as the substrate did not change. A change from old to new black asphalt, however, produced a 50-200% change of the measured water-film thickness. Overall, water-film thickness seems to be the most immature algorithm that was tested, and the results should be interpreted with caution. However, considering there are few options to reliably measure water-film thickness, and the large span of film-thicknesses on roads (30 µm - 3 mm), an accuracy of a factor two may be adequate.
In mobile phones and also modern vehicles there are cameras that takes pictures or even video of the road ahead of the vehicle. By combining such images or videos with an optical road condition sensor a much better image of the current road condition could be presented.
In this approach the vehicle mounted sensor Road eye is combined with a dash camera application for an IPhone to show the potential of combining the two sensors. By utilizing the three intensities in the Red Green and Blue (RGB) image in combination with the road condition classification from the Road eye sensor it is possible to get a classification of the entire road lane. As the Road eye only classify a area of around 1 cm but with a high frequency, 20 Hz, a line in the image from the camera can be used as a reference to separate different road conditions across the road lane.
This investigation focuses on post processing with the intent of enabling a technology that gives the road maintenance entrepreneurs a possibility to click on a symbol on a map to get an image of the latest road condition classification for that specific position. The image will contain not only an ordinary RGB image but also a color-coded image of the current road condition.
The result of the investigation is a video from real measurements done in Norway covering the road conditions dry, wet, icy and snowy asphalt both across the lane as along the lane.
Road winter maintenance heavily relies on an accurate weather forecast. MeteoGroup's road forecast model translates the weather forecast into a location specific road surface temperature and condition forecast. This information helps the road manager to make decisions on the required actions in his domain. However the road weather forecast is just one element of the complete information, which is needed to make the right decision. The road manager will combine weather information with actual measurements of road surface temperature and residual salt. Based on these data, the treatment policy and his own rules of practice, the road manager makes a final decision. The goal of the new DSS is to bundle all relevant information into a clear treatment advice for the next 24 hours in order to support the road manager in his decision.
The DSS provides a treatment advice; basically "where", "when" and "what":
– Where: For each route DSS presents a treatment advice. The latest measurements and site-specific forecast from a representative RWIS station are used as input. Also freezing point and residual salt information based on observation data from this particular location is automatically taken into account in the advice. In a next version of DSS we will extend the system by using information from our route based forecast model.
– When: DSS shows a time window for each route in which the salting (and/or ploughing) action needs to take place. It shows an optional time window and a final hour to start salting in order to be safe. It takes into account forecasted factors like amount of water on the road surface, amount of residual salt and dilution due to precipitation.
– What: The system shows – based on the policy and rules of practice of the individual road authority – the type and amount of chemical to be used. Rules of practice are provided by the road manager. These contain for example the time required to treat a certain route, the type and dosage of chemical to be used in case of certain road conditions and the snow amount at which ploughing needs to start.
The new DSS provides an overview in one glance of the advised treatment actions for each route, in order to facilitate an objective decision by the road manager. The system is currently in pilot phase and tested by some of our main customers. First feedback is very positive: the system gives valuable information in a clear overview.
The forecast model for road surface temperature developed uses a surface energy balance equation. The meteorological input parameters are forecasts taken from the Met Office Datapoint website, and other parameters in the model are derived from the literature, GIS data, and traffic flow information. All parameters in the model have an associated uncertainty. A distribution is placed over the input parameters of the model based on the error of the forecast when compared with observation data or ranges determined from the literature. Sampling from this distribution we generate an ensemble of each of these parameters and run the model for each ensemble member. The output of the model is an ensemble of road surface temperature forecasts at each point.
The spatial aspect of the model is dependent upon the different meteorological forecasts taken at different areas along the route of the Hagley Road and also the geographical parameters. These are included as a layer in a GIS and so the appropriate value can be included in the model at each point along the route, including associated uncertainty. A global sensitivity analysis is used to show which of the input parameters are the most important in this model across the range of conditions experienced therefore indicating which sources of uncertainty we should be aiming to reduce on average. We also investigate the forecast sensitivity and show that under different conditions different factors become important in determining the model response. We demonstrate the probabilistic forecast system, and show that constructing a statistical emulator of this probabilistic model will allow the forecasts to be made more quickly, allowing dynamic updating of route based forecasts using a near real time data assimilation approach.
Design and Testing of a Decision Support System for Deploying Weather Responsive Traffic Signal Operations in Texas
K. Balke, S. Sunkari, H. Charara, D. Florence, N. Chaudhary, P. Songchitruksa, G. Pesti
Abstract | PDF | Presentation
The Texas A&M Transportation Institute, with assistance from the Texas Department of Transportation, is developing and testing a system for developing and deploying weather responsive traffic signal timings. The system takes information about roadway surface and weather conditions in real-time, and determines the changes to traffic signal timing and intersection detector settings to improve traffic flow and safety during inclement weather conditions. Weather and roadway surface condition information is integrated with traffic data and fed into the decision support system designed to assist the traffic signal system operator in: (1) assessing the potential impacts and effects of current (and projected) weather and roadway surface conditions on operations, and (2) determining the most appropriate type of traffic signal strategy to deploy in response to deteriorating conditions. Potential traffic signal timing strategies include altering traffic detector configurations and settings to account for obscured lane markings, or reduced travel speeds; lengthening all-red clearance intervals between conflicting movements to account for increased stopping distances and reduced visibility; altering phase patterns or phase splits to reduce the potential of stopping vehicles at locations where roadway surface conditions have diminished (i.e., at a crest of up-slopes where reduced pavement friction might make it difficult to accelerate from a stop or at the bottom of down-slopes where reduced pavement friction may cause drivers to slide through intersection during clearance intervals, etc.); and implementing new coordination plans designed to provide progressions for increased traffic demands at slower speeds prior to or during a severe weather event. These new settings are sensitive to the changes in vehicle operating characteristics caused by adverse weather. The vision is to use these strategies to better match signal timing plans and parameters to the prevailing travel conditions to promote more efficient traffic operations and reduce the potential of some weather-related vehicle crashes.
Decisions during inclement weather are typically made by the individual driver. There are many aspects of bias that go into a naturalistic decision including length of time driving trucks and past experiences. There is not currently a method to assist truck drivers with that decision. Decision support systems (DSS) have been used in many areas such as winter maintenance and for IFF (Identify Friend or Foe) by the Department of Defense. There are many types of DSS in use today. A Bayesian Belief Network (BBN) can be used as a DSS and can assign a probability of delay or crash based on empirical data. The BBN uses a graphical representation of a probability dependency model. By using existing data of delay and crash the BBN assigns a probability of delay or crash for a given trip that can then be compared to the naturalistic decision for that trip. Because the DSS is evidence based, it provides more robust decision making and may thus increase safety and decrease delays, potentially saving billions.
This research explores the difficulties associated with inclement weather for vehicles, in particular tractor trailer trucks. By using a Bayesian Belief Network (BBN), weather data (snow, ice, wind, rain/thunderstorms and visibility) and probability for specific aspects of the weather event are input into the network based on previous studies and empirical data. Truck scenario delivery routes are used to assign dependent probabilities or so called risk of delay or crashes. These risks are assigned a decision using an index that incorporates the geometric mean of the probabilities. A survey of trucking firms was done to determine if assumptions within the model were accurate, by examining the decision of the truck driver in storm scenarios. The survey also examined how drivers would react if they were given the level of risk associated with inclement weather, and in particular sought to determine how much risk it would take for them to stop.
The results show that the BBN is successful at predicting risk of delay and crash. From that, an index was created to assign a decision based on those outputs from the model. The survey confirms the assumptions made in the model and that if truck drivers are provided a level of risk, their decisions are more conservative than using their own experiences. Quantifying the collective risk associated with inclement weather and providing an optimized decision has not been done in the past. The research does this and provides a more robust decision using a data driven process.
With the advancement of connected vehicle technology and automated vehicles, this freight and weather decision support system (FWDSS) becomes a critical tool in trip decision making of freight vehicles. In the example of a convoy of trucks, the decision of one truck then affects the rest in the convoy. Using FWDSS can apply a value added decision to all the trucks and reduce delay and crash during inclement weather events increasing safety and mobility as we move to automated vehicles.
TOPIC 4: INTEGRATING ROAD WEATHER INFORMATION AND OPERATIONS
RWIS Automated Advisory System: Centralized advisory system for the control of Dynamic Message Signs
Abstract | PDF | Presentation
To combat this issue, Alberta Transportation wanted to extend the Road Weather Information System (RWIS) delivered and maintained by Schneider Electric, with a means of automatically controlling Dynamic Message Signs (DMS) installed in this region. Collaboration between Alberta Transportation, district staff, the Alberta 511 group and Schneider Electric resulted in the development of the RWIS Automated Advisory System (RAAS). An RWIS station installed at the high wind location on Hwy 22 is monitored by the RAAS, which in turn controls 6 DMS signs installed at various diversion points North and South on the highway, covering a total distance of more than 90 kilometers. Three DMS are NTCIP compatible, and three are solar powered, Modbus based beacons.
Covered in this presentation will be the design of the RAAS and the benefits of a centralized, cloud-based field-hardware agnostic, data management system including the ability to control any number of DMS, with any kind of alerting hardware. Also covered in the presentation will be the benefits of employing a prioritized rule-based system so that any measured or derived RWIS parameter can control the DMS signs either individually or as a group. Benefits also covered will include a system that is easy to configure and flexible enough to handle mobile assets deployed on demand to assist with emerging conditions or events. Finally, a system status report will be included regarding the operation of the system over the winter of 2015-2016.
The paper contrasts 2 different approaches, one being a daytime only treatment regime with the other looking at a fully proactive pre-treatment regime with a back to black policy 24 hours a day. This is referenced against a do nothing option. By analysing the performance at each observation location assumptions are made as to the number of hours of mobility saved by each treatment methodology and assigns a value in term of GDP that is saved through these efforts.
Data was gathered from existing stations in Scotland and France where very different treatment practices are operating. However the stations chosen were similar in their winters regarding the occurrence or possible occurrence of icy conditions to allow a closer comparison than their geography suggests.
The methodology of this way of assessing data from fixed RWIS allows very specific hourly analysis to be scaled up to seasonal statistics. Figure 1 shows the frequency of hours where ice could have formed at the location versus where it actually was detected (effective treatment versus ineffective treatment) for a site where daytime only treatments are applied. What is most notable here is the likelihood a driver will encounter ice when conditions are correct is nearly twice that which they will find a treated road during the early hours of the morning.
Scaling this kind of data up to the country level and applying a cost of delay gives an outcome that it is possible to obtain a return on investment value that is approximately 13:1 for a daytime only treatment regime with up to 22:1 return for a fully proactive 24hr treatment policy. The paper also shows that these values are of course weather dependent with a very mild winter with only single routes being affected at any one time seeing an overall cost to benefit ratio of 3:1.
In the State of Queensland, Australia, flooding, rather than snow and ice, is the most prominent weather related hazard to impact road users, and road weather information across most of the State largely consists of water level data provided through strategically located Roadway Flood Monitoring Systems (RFMS). Whilst many of these in-situ stations were initially installed as standalone systems communicating via third party software applications, there is now a concerted effort within the Department of Transport and Main Roads to integrate these existing stations and any newly deployed systems into the STREAMS Intelligent Transportation System (ITS) Traffic Management System used by the Queensland Government Road Agency.
STREAMS is a fully integrated ITS Traffic Management System. Most ITS software platforms run on an inter-operability model, where multiple ITS systems work in parallel, each performing a discrete function. STREAMS however employs a distributed computing software architecture, where field hardware such as intersection controllers, video cameras, speed detectors and environmental monitoring systems are connected via field processors back to a central application server, providing a truly integrated Advanced Traffic Management System.
This paper explores the system architecture behind STREAMS integrated Roadway Flood Monitoring Systems, and examines the operational benefits of having road weather information available to Traffic Management Centres through an integrated Traffic Management System.
This study aims to develop the methodology to provide the variable speed limits in real time by rainfall intensity. For this study, algorism about variable speed limits was developed, which reflects the methodologies of Stopping Sight Distance (SSD) and Visibility Distance (VD) by using rainfall data. This algorism provides variable speeds according to rainfall intensity and warn it to road users when SSD value is greater than VD value.
Weather data from 190 weather stations in Seoul were used for analysis. The 33 of 190 weather stations were used for validation purpose, which sites are based on 20 meter away from roads in Seoul. The data of remaining 157 weather stations were used for model fitting. Seoul road data from Urban Transit Information System operated by Korean National Police Agency were used. Seoul Road data consist of 22,183 nod links (lines of roads) which are connected to 177,599 points. Finally spatial pattern analysis for rainfall information of Seoul city were expressed based on results of interpolation by using representative 22,184 points of each nod links which are reproduced by 177,599 points. This rain information on nod links is generated every 10 minutes time by using 190 weather observations based 250m space resolution in Seoul.
For the statistical verification, the Root Mean Squared Error (RMSE) and correlation coefficient were considered by using IDW (Inverse Distance Weight) method which estimates values of cells by weighting of values (point) of geometric data in the neighborhood of each processed cell. This results showed significant statistically and prediction performance of rainfall in Seoul indicated good performance.
Further study will be needed to consider SSD value which can be determined by types of road alignment like difference in grades and horizontal factor, however tangent factor is only considered in this study.
This study will be expected to make driving safer and more convenient by providing the rain information required for road drivers, when heavy precipitation events occur on the roads.
A case study is done for a runway at Oslo Airport in Norway during the winter of 2010-2011. A surface temperature and condition prediction model is built which predicts the surface condition a couple of hours ahead of time. The model takes into account the effect of the weather, air traffic, pavement properties and winter maintenance measured such as snow ploughing and salting. The following heat fluxes are taking into account: the conductive heat flux through the pavement, the longwave and shortwave radiative heat fluxes, the convective heat flux, the heat fluxes of rainfall and snowfall, the heat flux for evaporation and condensation and the heat flux due to sublimation and deposition. The effect of air traffic is included in the convective and latent heat fluxes. The effect of chemicals on the melting temperature is included in the latent heat fluxes.
The model is run in now-casting, where the surface temperature and condition were predicted three hours ahead of time and in long term mode, both with predicted and observed surface conditions. When running the model with SNOWTAM data in "now-casting mode" the average error of the surface temperature during the entire winter season, which is one of the main input parameters for the surface condition prediction, is 0.25 °C and the RMSE is 1.65 °C. The surface prediction model is in almost half on the times predicted accurately. The prediction is most accurate for dry and icy conditions. The results show that the use of the observed surface conditions (SNOWTAM data) increases the accuracy of the surface condition with 30% and the accuracy of the RMSE of the temperature prediction with 13% compared to using predicted surface conditions.
TOPIC 5: MOVING RWIS FORWARD
Maturity of Finnish Road Weather and Maintenance Network – challenges for innovation capability
Zulkarnain, A. Aapaoja, R. Hautala, T. Kinnunen
Abstract | PDF | Presentation
Over the years there have been significant developments, including the introduction of remote pavement sensors that have allowed additional parameters to be measured and also given rise to fully mobile sensing when mounted on vehicles. The earliest of these measured just pavement surface temperature, but nowadays can also measure a multitude of parameters, for example surface state and grip. One of the main advantages of these mobile observation platforms is that the entire road network can be measured rather than just single point data from the static systems. Since these mobile systems generally cost less there is a common thought that mobile measurements will spell the decline of static systems, especially since a number of weather parameters can be derived from the increasing intelligence of modern vehicles, such as taking canbus information including the likes of traction control, windshield wipers and temperature sensors etc.
This paper intends to explain how the need for static systems will remain in to the foreseeable future, as well as suggestions for their improvement.
MODE 1, whole road-network viewer (http://meteoruta.aemet.es). Forecasts up to 36 hours of hourly severe traffic conditions are displayed using vector-polylines in different colors classified by the weather severity. The user can click on the road or any other place to get more information on weather conditions and evolution. Weather forecasting data are obtained from the operational AEMET-HARMONIE numerical model, running at 2.5 km resolution for the Iberian Peninsula and surroundings. Model runs 4 times a day (00, 06, 12 and 18 UTC) and is going to be updated up to 8 times a day this summer.
MODE 2, pathfinding mode, (http://meteoruta2.aemet.es). The user, from the client browser, may choose his own way from the starting point to the final point, and one optional intermediate point. Defining the departure time, the server returns a vector-polyline fitting the shortest path, with route directions, timetables and forecast weather conditions. The polyline is colored according to the severity of the forecast. The algorithm used for the shortest path calculus is the heuristic A*, and it is almost entirely based in OSRM (Open Source Routing Machine). The forecast is obtained from AEMET-HARMONIE numerical model and, after reaching its longest forecast, from ECMWF numerical model.
This new tool is being used by the Spanish Road Authorities to include weather forecasting in their contingency daily operations.
The mass flux of blowing snow is the mass of snow particles passing through a unit cross-section per unit time. The value obtained by integrating the mass flux of blowing snow in the height direction equals the snow transport rate. A snow particle counter (SPC) is a device that can measure mass flux continuously. Use of SPCs has been increasing in blowing snow research. It would be possible to determine the cumulative annual snow transport rate based on continuous measurements, if the mass flux of blowing snow could be measured by installing a number of SPCs in a vertical array. However, such a technique is impractical, because SPCs are expensive. Therefore, the authors investigated the relationship between the mass flux at a given height and the snow transport rate to clarify the feasibility of measuring the cumulative snow transport with a single SPC.
The authors manually measured the mass flux of blowing snow by installing several net-type blowing-snow traps in a vertical array at the Ishikari Blowing Snow Test Field, which is about 20km north of downtown Sapporo. The snow transport rate Q for heights of up to 5m was determined by using the measured data. Then, the relational equation Q=kq (k is a proportional constant), which expresses the relationship between the snow transport rate Q and the mass flux q, was obtained for each of the heights of 0.5m and 1.0m. The resulting relational equations were as follows: Q=3.0q for the mass flux at the height of 0.5m, and Q=5.4q for the mass flux at the height of 1.0m. It is thought that the empirical equation obtained in this study enables the estimation of the snow transport rate Q by using the mass flux q at a given height.
The different mass and heat fluxes during the melting process of snow on a heated pavement system are calculated based on measured temperatures, weather data and estimated pavement properties. An experiment was conducted in the cold lab at the NTNU. On the bottom side of the 5cm thick asphalt plate heating films were connected to the slabs. The setup was equipped with sensors measuring the pavement, snow and air temperatures, the humidity, incoming longwave radiation and wind speed. Additionally the height of the snow layer was measured and the surface condition was registered.
The height of the snow layer changed from 40 to 4 mm during the melting process. As soon as the snow started to melt the height of the snow layer decreased with a constant rate till it was changed into slush and reached a height of 4 mm. The density changes from 66 till 999.8 kg m3 during the melting process. The specific heat capacity is calculated based on the specific heat capacity, density and volume fractions of ice, water and air. It changes from 2.1 to 4.2 kJ kg-1K-1. The thermal conductivity depends mostly on the density and snow microstructure. It can be described as a function of density or as a function of the thermal conductivity and volume fractions of ice, water and air. During the melting process the thermal conductivity changes from 2.6*10-2 W m-1K-1 for dry snow till 0.58 W m-1K-1 for water.
The surface layer changes from a dry well insulated layer of snow into a much thinner layer and a better conductor when it is melted. Measured temperature differences within a 4 cm snow layer were up to 6°C. As soon as the melting process started the snow surface temperature started to rise and the temperature difference started to decrease until it was transformed into slush after which the temperatures of the surface layer and the top of the asphalt layer were almost the same. During the melting process the surface flux increases, mostly due to the changes in the radiative and convective heat fluxes. The results clearly show that even a thin snow layer on a pavement can significantly affect the surface temperature and thermal balance.
The CLEAN-ROADS project: a combined environmental-cost-user gain from the application of a MDSS to winter road maintenance operations
I. Pretto, C. Di Napoli, E. Malloci, R. Cavaliere, R. Apolloni, G. Benedetti, A. Piazza
Abstract | PDF | Poster
Over three consecutive winter seasons road weather and surface condition data have been collected through a mobile probe vehicle and a state-of-the-art road weather information system (RWIS) of fixed stations. From this information base, a novel MDSS tool has been developed to report past and present roadway status, assess current weather conditions, predict icy conditions on the short-term and nowcast range, and automatically deliver alarms in case of road icing risk.
The novel MDSS has been applied to assist a road maintenance unit in de-icing operations operated in the Province along a test route. Its benefits to the environment have been empirically evaluated. By integrating the MDSS with an environmental monitoring system, the impact of de-icing salt (sodium chloride) on aquatic systems and air quality has been quantified, and the environmental gain due to a MDSS estimated for the first time.
The impact of a MDSS to the management of anti-ice treatments has also been analyzed. The implemented MDSS recommends the necessity and timing of gritting the road with salt. A comparison between MDSS recommendations and the treatments actually carried out has permitted to (i) identify possible weaknesses in the current winter maintenance management system and (ii) improve the system itself in terms of efficiency and costs. In this paper the economic gain due to the novel MDSS is presented and quantified using a cost/benefit analysis.
Finally, implemented strategies to increase the level of responsibility in road users are described. Benefits achieved from conducting awareness-raising actions about road ice and disseminating road weather information to local travelers through a MDSS are assessed and discussed.
The huge old one system was divided to 9 new separate applications. There are now four different metadata applications: one for weather stations, road weather cameras, traffic measurement stations and infrastructure related to observation stations. Correspondingly data collection was divided to separate independent applications: for road weather stations, cameras and traffic stations. Road weather data is managed in its own new data storage system. Road camera images are stored in special image storage. Traffic information is forwarded to the old storage. With the help of this new software architecture it is now much easier to develop each feature separately.
Very big changes were done to data dissemination. The old system had lots and lots of point to point data connections between different systems and storages. The new system is based on SOA (Service Oriented Architecture) and all data is transferred with the help of, or via, the service bus. This was a huge improvement, data is easily accessible to everyone.
The renewal project of Finnish RWIS started in year 2012 and ended 2015. The design phase took one year from March 2012 to March 2013. The development phase including procurement process took two and half years ending September 2015. The consultant costs were: design 150 000 euros and development 550 000 euros.
There was no significant difference of the temperature between the entrance and exit in every month. In winter, the net radiation is lower than summer. The peak of the net radiation of the exit is earlier than the entrance in winter. This result is related with the topographical characteristics of the Jeongneung tunnel. The road surface temperature of the exit is higher than the entrance in winter, similar in spring and summer, and lower in fall. The change of the road surface temperature is related with heat of inside the tunnel. The diurnal pattern of the difference between road surface temperature and air temperature (Ts-Ta) was also analyzed. The results of the Ts-Ta mean that the road surface temperature might be related with the traffic volume, vehicle speed and net radiation.
Snowfall refers to the depth of snow that accumulates in a given period of time. A snow depth gauge is generally used for its observation. The Japan Meteorological Agency defines an hour’s snowfall as the difference in snow depth measured every hour. According to this system, ‘no snowfall’ is recorded when fresh snow is blown away by strong winds or when the snow depth does not increase due to densification. Moreover, observation using a precipitation gauge cannot give an accurate snowfall intensity, because the ratio of snow particles caught by the precipitation gauge to total fallen particles decreases with increase in wind speed. As mentioned above, observed values of snowfall and snowfall intensity include errors from various causes. Therefore, correction is necessary in order for the output value to be closer to the true value.
The purpose of this study is to accurately understand the snowfall intensity at times of strong winds. We have been conducting comparison observations of snowfall intensity using a precipitation gauges (tipping bucket type) and the Double Fence Intercomparison Reference (DFIR) since January 2015. DFIR is a secondary standard designated by the WMO. A measurement taken by DFIR can be converted to a value which can be regarded as a true value by using a conversion equation (Goodison et al., 1998). A DFIR gauge consists of a pair of octagonal wind shield fences of different sizes (the smaller fence is inside the larger one) with a precipitation gauge placed at the center of the inner (the smaller) fence. The diagonal lengths of the outer and the inner fences are 12m and 4m, respectively. The inlet of the precipitation gauge is set at the same hight as the top edge of the inner fence. The cumulative precipitation as measured by precipitation gauges (tipping bucket type) was compared to that measured by DFIR for the previous winter (2014/2015). The measurement values of the precipitation gauges were found to be about 50% those of the DFIR gauge.
The winter seasons in Russia are characterized by the alternation of the air temperatures over zero degrees, thus the snow volume near roads depends on the air temperature change. For the account of snow volume change it is proposed to calculate the "Coefficient of snow losses from melting and evaporation". The definition of this parameter is done according to the data observation of the snow cover characteristics, i.e. the snow height and density.
The proposed models allow to calculate the coefficient of snow losses from melting and evaporation which is important parameter for snow protection maintenance and design on the roads of Russia. The computer program "Blizzards" was specially developed for that kind of calculations.
For solving the practical problems the most convenient way of submitting the calculations results are in the form of maps form with isolines representing the distribution of the coefficient of snow losses from melting and evaporation during thaws. The calculation of isolines and drawing them on the map was automated by using the methods of creation of the digital terrain models. To calculate the program complex "CREDO Road" from joint venture "Credo-Dialogue" was used. The calculations results in the program "CREDO Roads" can be exported into a GIS system.
The core of the presentation is derivation and description of a technique for generating ensembles and evaluation of the ensemble forecasts. The obtained results can be briefly summarized in the following way:
• Temperature, humidity and wind speed in the near-surface air layer are not sufficient to describe reasonably the uncertainty of RST forecasts. Radiation fluxes, which are derived from total cloud cover, are crucial for RST development and this quantity should be included in the ensemble generation. Even after the inclusion, the ensemble is too confident (the spread is not sufficient) and cannot describe natural uncertainty of the forecasts.
• There are several main reasons for the insufficient ensemble spread and following underestimation of the forecast uncertainty. One of the reasons is not considering the errors of forecasted rain/snowfall by a NWP model in generating ensembles. Other important reasons have technical character, for example wrong information about a sky view factor, unavailable or inaccurate information on the current state of surface condition on roads, inaccurate information about technological parameters of roads, etc.
• Although the ensemble technique underestimates the uncertainty of RT forecasts, it can be used in decision making of road authorities providing winter road maintenance. The outputs can be subjectively "corrected" or a technique like model output statistics could be applied to objectively correct the forecasts.
We suppose that the RST ensemble forecast will be a very useful supplement of special road meteorological information presented in RWIS of the Czech Republic and an interesting tool for meteorological forecasts.
The 500 hPa is often referred to as a steering level to the weather systems of synoptic scale as the mid-latitude low pressures and ridges. On average, Iceland is situated on the eastern fringe of cold higher level trough over eastern Canada. This trough maintains an average WSW wind at this level over Iceland. Fluctuations are of all dimensions in both temporal and spatial fields. The formation, movements and dissipating of mid-latitude systems are strongly connected to the horizontal east propagating planetary Rossby waves near the level of 500 hPa. Depending also to the higher level coherent North Atlantic jet stream.
A multitude of circulation classification methods has been introduced in the literature. In the 1950's the Swedish meteorologist Ernst Hovmøller classified how to use weather types based on the circulation pattern for employing parameters such as temperature or precipitation. Recently, as reanalysis of standardized upper air data is available, there is again focus on such simple statistical methods for classification. It is shown that such a method is valuable for measuring effort and cost of winter service operations and it could benefit the cost probably better than use of widely supported winter-index.
Road weather forecasts were produced for 23 selected points along the bus route using the FMI road weather model. For initialization the model used air temperature, humidity and wind speed obtained from the observational grid of 1 km x 1 km, which values were calculated by the Kriging method using observations from nearby weather stations. The forcing atmospheric forecasts were obtained from the HARMONIE model with a grid size of 2.5 km x 2.5 km. The first results show that using car observations for the initialization have indeed some potential to improve the road surface temperature forecasts. Different ways to include observations in the model and also the effects of the sky view factor will be studied further.
Over the past year, Foreca has been developing computer generated maintenance recommendations. During the winter 2015-2016, selected pilot customers have evaluated the recommendations in two very different settings: in a meteorologist-supervised area, and in an area with completely automated forecast production with no meteorologist intervention. We wish to highlight two key features of the system, one technical and one philosophical. On the technical side, to support the short lead time operations in the area with no meteorologist supervision we leveraged a previously developed automatic nowcasting system, which uses satellite and radar data. A key enabling technology for the nowcasting system is the Meteosat-PRIME satellite. Its multi-spectral data makes it possible to detect clouds sufficiently well even in the presence of strong inversions with very low surface temperatures. On the philosophical side, when creating the recommendations we never aimed for completely robotized operations. We aimed for a system, which would also act as a teaching tool for personnel. Thus, the recommendations are not only about the sequence of maintenance tasks but also include practical advice regarding the uncertainties in the situation at hand. The recommendations help ensure consistent quality via practical weather situation dependent advice, reducing the chance for human error.
We will present verification results for the automatic nowcasting system along with some success and failure case examples. We will also present case examples of the recommendations and feedback from the pilot customers.
Furthermore, thermal mappings on the whole Slovenian motorways were performed in 2016 with equipment, developed for this project. Results were used to provide better route-based forecasts and for road weather stations location and its sensors optimisation.
Because of this project we expect safer motorways and higher efficiency of winter service on Slovenian roads in the future.