6 Findings and Lessons Learned
6.1 Introduction
Compared with typical winters in Maine, the winter of 2006-2007 was milder and there were fewer severe storm events. There were 21 recorded winter storm events between December 2006 and April 2007, and most of them produced less snow than would have been expected based on historical weather patterns. A total of 12 of these winter storm events were evaluated in detail, a process referred to in this report as event reconstruction. This approach to understanding how the MDSS was accessed and used to support maintenance decision making by the Scarborough crew was described previously in Section 5.2. This chapter discusses how MaineDOT used the MDSS in their storm management, presents findings based on these reconstructed events, and presents lessons that can be drawn from MaineDOT’s experiences with their initial use of an MDSS.
6.2 The MDSS Role in Road Maintenance
6.2.1 Background
For each of the events examined during the winter of 2006-2007 in this evaluation, the Scarborough crew, with the support of MaineDOT management, sought to include the MDSS in their decision making in whatever ways they thought would be helpful. Initially they specified the alerts they wanted to receive for several forecast points along the test route (see Section 3.4 for details). In addition, the Scarborough supervisor and members of the crew learned how to access the MDSS on their computers, which they could do in the office, and in some cases at home as well. Also, the operations room in Augusta was set up with access to the MDSS on their office computer. Augusta primarily used the MDSS to provide the NWS radar image display and warnings (as opposed to monitoring weather and pavement conditions at specific forecast points across the state) so they could easily track storms as they approached and progressed across the State of Maine. They played a key role in notifying their regional maintenance crews, particularly during night time hours.
MaineDOT management and the Scarborough crew participated throughout this assessment with a very positive, supportive attitude and a commitment to work with the MDSS. At the initial meeting between MaineDOT representatives, the evaluation team, FHWA, and representatives of the MDSS vendor, the MDSS was discussed in detail. From that point on, it was largely a trial period in which the Scarborough crew attempted to apply the tool on a storm-by-storm basis and “test” out how it might be able to help them in their maintenance decision making. The vendor team made themselves readily available to MaineDOT by phone and e-mail for inquiries and consultation as needed to support this deployment. The vendor encouraged the Scarborough supervisor to contact their meteorologists at the beginning of every event for help interpreting the MDSS forecast output and as a way to better fine tune the decision guidance being offered.
6.2.2 Summary of Findings
This winter provided an opportunity for the Scarborough crew to experiment with the use of the MDSS. MaineDOT believes that deploying this system provided an important learning experience and valuable new information they formally lacked. Overall, they are quite positive about the future potential for an MDSS to support their winter maintenance operations. They acknowledge that they need to explore ways to modify their approach to and use of the MDSS to better take advantage of the benefits it offers. They also would like to see improvements in the MDSS tool to increase the accuracy of its forecasts, and they want to concurrently fine tune the set of treatment recommendations that they provided for use in the MDSS. The net benefit of the MDSS is difficult to assess in this initial application period. Over each of the observed storm events, the crew considered a combination of their real time observations of conditions on the road and input from a variety of weather information sources in addition to the treatment recommendations and related forecast information provided by the MDSS. Initially they were willing to “experiment” with the MDSS and implement its treatment recommendations as provided. Later they were inclined to review those recommendations and adjust their decisions based on all the other information that was available to them at the time.
This winter season was the first opportunity for the Scarborough crew of MaineDOT to try out the MDSS which had only been made available to them a few months earlier. The adoption of new technologies and new ways of making operational decisions involves changes and acceptance by crews and management that can take a relatively long period of time. It is reasonable to expect that the pace of internal organizational change, coupled with improvements to the MDSS tool, will lead to greater usage and success in future applications of such tools.
This section discusses a number of the findings derived from the reconstruction of the winter’s storm events, along with additional explanation and interpretation. Each of the findings below includes discussion of the finding with regard to one or more of the objectives established for this case study evaluation.
- Assess the operational use of an MDSS by reconstructing each of the winter storm events and Maine DOT’s management of those events.
- Identify institutional issues and implementation costs faced by MaineDOT in integrating an MDSS into their winter maintenance decision processes and how these issues were addressed.
- Identify effective strategies for deploying an MDSS.
- Assess how an MDSS supplements the suite of tools available to MaineDOT for making winter maintenance decisions.
- Assess the potential benefits that an MDSS could offer MaineDOT if it were utilized more comprehensively.
- Identify lessons learned from MaineDOT’s experience using an MDSS.
Receiving too many alerts provided by the MDSS is of limited benefit and can be distracting. By the third recorded event MaineDOT had set up and subscribed to receive alerts at all four forecast points in the study area, resulting in numerous alerts being generated and received by the crew. In most of the events such a large volume of alerts was generated that the Scarborough crew found them to be overwhelming and at times distracting. In one storm event, for example, 87 alerts were generated. These alerts usually were triggered by the same weather front at all four forecast points in short succession (due to the relatively close proximity among the forecast points in the study area), that was found to be duplicative and not particularly useful. These alerts usually consisted of a mixture of forecasts (e.g., snow expected at 2:30 PM at Falmouth), observed conditions (e.g., snow observed at Falmouth), and NWS advisories and warnings. The crew indicated they would not go back to check on old messages when they showed up in a large batch or rapid succession. They noted a lot of variability in what these alerts were telling them about the timing and type of precipitation to expect over the course of a storm. That is, they felt that the alerts did not always provide consistent, actionable information.
The reconstruction of these events suggests that the use and value of the alerts varied across the events. The initial alerts were the most useful as they offered an early warning of an impending storm, and provided an initial forecast of the precipitation type and timing. A value of these alerts is to prompt the supervisor and crew to go to the MDSS on line and assess the more detailed information about the coming storm. It is this early information that provides guidance for deciding when to call in the crew and the storm conditions to anticipate as a basis for making pre-treatment decisions. Many of the alerts were not used, either because they arrived during the crew’s “quiet period” at night, because the number of alerts seemed overwhelming, or because the information didn’t appear to agree with the crew’s observations.
Some of the Scarborough crew’s storm management practices benefited more by using the MDSS than others. The routine of the crew remained very much as in prior winters, in which they accessed a variety of tools, including various weather forecasting systems, consulted as needed with their Augusta operations center and exchanged information, when appropriate, with NHDOT and neighboring crews. The prime MDSS component that offered pavement condition and temperature forecasts was of some value to the crew in practice and would likely have been of greater benefit with additional training in how to make best use of this information. The Scarborough crew noted that the pavement temperature forecasts provided by the MDSS were consistently close to the readings they obtained directly from the IR sensors on their trucks. This allowed them to develop some confidence in the pavement temperature forecast trends that were offered by the MDSS and consider that information in making their road treatment decisions. However, the crew entered into this winter season with only a limited understanding about how they could take advantage of the availability of a pavement temperature trend forecast to improve their treatment decisions. Their level of understanding, coupled with an inability of the MDSS to adjust treatment recommendations based on prior actions taken by the crew on the road (and hence an understanding of actual road surface conditions during the storm), limited the usefulness of this element of the MDSS.
The crew regularly consulted the MDSS ahead of each storm event, whether prompted by an alert or their knowledge from other sources that a storm was coming. When the supervisor arrived at the Scarborough shed, he would check the MDSS vendor’s web site on his computer for the latest information and forecasts. The Augusta operations center also tracked the NWS radar images and forecasts via computer.
MaineDOT maintenance crews are very interested in having access to accurate forecasts of the type and amount of precipitation to expect before the start of a storm and the precipitation start time. This allows them to time pre-treatment, efficiently schedule their crews, and have consistent guidance regarding the type and amount of chemicals that will be most appropriate to apply to their pavement. They initially followed the MDSS treatment recommendation at the start of a storm closely. In some instances, they used less materials based on the MDSS recommendation than they would have otherwise, and said that worked well. However, they found in those instances that they usually needed to increase the subsequent treatment amounts to compensate for the lesser initial treatment. After some experimentation in the early storm events, they subsequently assessed a variety of weather information sources along with the MDSS, including private sector forecast services, television broadcasts of local news weather reports, and a network of weather information sharing with nearby crews and with sources in neighboring New Hampshire. Their decisions were then based on all these sources rather than simply following the MDSS recommended treatment.
The Scarborough crew also relied on night patrols and a weather observation station at the Portland Jetport that provided weather observation data. The closest ESS to the Scarborough crew was outside of the study area and therefore offered less useful observational data for decision making. Finally, their own crews on the road provided a constant source of input regarding prevailing conditions. It is within this context that the MDSS sought to offer enhanced road weather forecasting capabilities and decision guidance regarding treatment timing and choice of material types and amounts.
It is difficult to provide consistently accurate weather forecasts for the complex maritime climate in southern Maine. This case study offered an opportunity to test the MDSS under challenging weather and climate conditions. If it proved accurate, offered helpful information, and was easy to use, then the crew could be expected to rely on it more. In some event situations it provided good forecasts of either the start time or end time of the storm, or the precipitation type, or pavement temperature trends. In others it missed the mark by too wide a margin for it to be useful. The general consensus over the course of the winter was that the MDSS was better at forecasting the timing of the start of a storm than it was at forecasting the precipitation type, but this was quite variable across storms as well. Local geography coupled with a maritime climate and the confounding effects of ocean temperatures apparently made it very difficult for the MDSS to generate accurate forecasts of precipitation type. Two of the storm paths were over open water before coming on shore in southern Maine, resulting in a wetter precipitation pattern than the MDSS was forecasting. Often, when air temperatures hovered close to the freezing point, the MDSS would forecast snow but rain or freezing rain would actually occur.
It was evident that other weather services (other Internet weather information services, local radio and television stations) as well as the MDSS had limited success in forecasting the often tricky ocean effect weather events of the type that occurred in the study area. While most weather information service providers take input from NWS forecasts and other shared resources (e.g., radar, satellite images), mission critical (e.g., maritime, aviation, or MDSS) local forecasts usually require routine, and manual refinement by in-house meteorologists employed by the respective weather information providers. MDSS meteorologists may lack the advantage of on site, real-time feedback and knowledge of the local area to make timely and effective adjustments, especially when complicated geographic and micro-climatic scenarios are involved. This is a common challenge for all weather service providers and is not unique to an MDSS. Accurate local forecasts are absolutely essential for the provision of MDSS services. One limitation in this study was the lack of any ESS in the immediate study area, which, as demonstrated in the federal MDSS prototype, could be used to provide real-time atmospheric and pavement observations in support of “forward correction” of the forecasts.
The Scarborough crew primarily relied on multiple sources of weather information, including the MDSS, their local experience, and crew observations on the road to make treatment decisions during a storm. During the event reconstruction interviews after a storm event, the crew emphasized how they really didn’t have the time in the middle of an event to consult an MDSS or any of their other standard weather forecast tools because they were so busy “fighting the storm.” In addition, while the crew is out on the road dealing with the storm effects, they base their treatment and related operational decisions primarily on their observations of conditions and how traffic is responding. For example, they can observe how the materials on the pavement surface are being deflected by vehicle tires to judge the effectiveness of their treatments and the next most appropriate treatment strategies. They are much more inclined to rely on their years of experience and real-time observations than on a forecasting tool that is not closely connected with the actual conditions.
The Scarborough crew tended to mostly apply greater material amounts to the pavement than the MDSS recommended. A central purpose of an MDSS is to offer the road maintenance crew treatment recommendations that are based on forecasted weather and pavement conditions and updated as actual conditions change. An MDSS ideally offers an alternative to ad hoc decision making that may be based substantially on anecdotal information. A scientifically grounded treatment guidance system should offer benefits in terms of a more efficient use of costly materials, labor and equipment resources, assuming the forecast data are accurate. It has been a long-standing practice of the Scarborough crews to apply their chemical treatments conservatively; that is, they typically apply a higher amount than might be recommended by an MDSS, especially in the first treatment, because they want to be sure that adverse conditions don’t get ahead of them. They would rather not find themselves having to play “catch up” in subsequent treatments. They believe that if the initial treatment(s) prove to be inadequate, the later treatments will need to be substantially greater to recover LOS, likely resulting in higher resource usage over the course of the storm. This conservative approach suggests that there is an opportunity for a decision support system that can accurately and consistently guide the best possible treatment decisions, thereby offering the potential for resource and cost savings.
In selected storm events the Scarborough crew was able to use the MDSS forecast to make a decision not to pretreat the pavement. In some others they were able to use less materials based on the MDSS forecast. It is difficult however to assess the amount of savings over an entire event attributable to the MDSS because often the crew reported that their subsequent treatments were increased over what they might otherwise have done in the absence of an MDSS. It is also difficult to separate out the influence of the MDSS from all the other forecast and observational tools that they used.
The MDSS typically recommended less material than the crew was comfortable applying under a given set of conditions. These recommendations were arrived at as a function of the standard treatment protocols provided by MaineDOT and placed in the MDSS, and the MDSS forecasts. As noted earlier, the crew followed several lower-bound treatment recommendations, which actually performed well under the conditions as they turned out. The crews later tended to apply significantly more than the MDSS recommended amounts in subsequent treatment decisions. In their experience with almost all of these events, their decision was supported both by the experiences of other crews in the area and by the demonstrated success on the road of the chemicals applied at those higher levels. However, since the crew didn’t consistently follow the MDSS recommended treatment amounts, it is not possible to fully assess either the efficacy of the recommended treatments nor the benefits that would have accrued had they been closely followed.
The resolution of the weather forecasts was not high enough to allow the MDSS to differentiate conditions across the set of forecast points along the relatively short distances covered in this study area. One of the expectations from this assessment was that MaineDOT would be able to use the multiple forecast points along the study route to identify sufficient differences in forecast weather effects (timing, precipitation amount, and air and pavement temperature trends) to be able to apply differential treatment regimes along the route. Such an approach also would offer the potential to save in material usage and costs by varying the level of treatment by locale. This did not turn out to be the case. First, the sets of alerts provided for each of the forecast points tended to provide identical or very similar information. Second, the forecasting was not precise enough to allow for point forecasts that could accurately differentiate conditions at these different locations. An illustrative example was provided by one of the reconstructed storm events in which observed snow fall was much more pronounced in the southern end of the route; however, MDSS forecasts for the two extreme south and north forecast points were unable to capture these differences sufficiently to support different maintenance treatment decision making by the Scarborough crew.
Viewing their experience with the MDSS as a whole, MaineDOT found benefits in the approach and potential for the future. While the Scarborough crew did not find the MDSS to be sufficiently accurate and consistent enough to rely on it as a tactical tool for their treatment decision making during the storm, they saw the introduction of this tool as a valuable first step toward integrating route-specific weather and pavement condition information on a GIS platform. They felt that this preliminary deployment provided an experience base upon which they can learn to enhance their use of this tool to derive increased benefits in the future. They plan to make modifications to the alert structure and precipitation timer, and are considering how they might modify the treatment recommendations imbedded in the MDSS to better address their state’s needs by maintenance shed. They want to work with the MDSS vendor to enhance their use of the pavement temperature and bridge frost information components. This past season was viewed as a valuable learning experience, and MaineDOT is looking to the future to enhance their management of storm events through the better use of an MDSS.
6.3 Summary of MDSS Benefits
MaineDOT has had a proactive approach to maintenance operations for quite a few years. They prepare for a storm event by having their trucks already loaded and ready to go, they have their night patrol, their operations center providing a statewide perspective, and their crews on call or ready to go when they know a storm is coming. The MDSS supplemented this readiness in the following ways.
- It offered useful features not previously accessible to MaineDOT. These included a GIS radar and NWS weather forecasting platform that their operations center could use to track storm events across the state. This provided enhanced notification capability for the various maintenance sheds.
- It provided a consolidated set of treatment recommendations and made them easily accessible to the crews to support their pavement treatment decisions. The MDSS framework allows a DOT to customize these treatment recommendations to the particular needs of their maintenance shed locations. MaineDOT relies heavily on their crews’ treatment judgments because there are so many variables involved, such as time of day, day of week, traffic levels, air and pavement temperatures, precipitation type and amount, and the effects of prior treatments.
- The GIS platform allows inclusion of RWIS/ESS, ASOS and other observational data to be represented and assessed. They liked the ability to be able to overlay pavement temperatures on a road map. MaineDOT is interested in including data for their state and adjacent states in this way.
- New data were available on bridge frost potential, pavement temperature trend forecasts, and wind and blowing snow potential.
- MaineDOT’s maintenance strategy is to first look to their crews and learn from them what is happening out on the road, then from there go to forecasts of what is expected to happen in the future.
- MaineDOT’s treatment decisions rely significantly on their crews’ observations of pavement temperature. The MDSS provided the ability to extend those observations with trend forecasts.
- MaineDOT saw potential for improving the alert notification provided by the MDSS that could help them in storm preparedness.
- MaineDOT believes the MDSS offers a very useful training tool to help all DOT maintenance crews throughout the state operate on a par with their best crews.
MaineDOT viewed the MDSS in this first season’s experience as a planning tool to help them identify precipitation type and timing as a storm is approaching. They did not tend to view it or use is as a tactical tool to manage their activities during the storm event.
The Scarborough crew found that their traditional sources of weather forecast information were at least as accurate as those provided by the MDSS. Although the crew reported that the MDSS did not change most of their decisions from what they would have been in the absence of the MDSS, they did make a concerted effort to examine and explore ways to use MDSS information for every event.
The MaineDOT crews who used the MDSS came to better appreciate the potential of this kind of decision support tool, and they appreciate more than at the start how an MDSS could offer them benefits. MaineDOT believes they have derived benefits from having a weather forecasting tool that provides a set of capabilities they have not had available previously. As the MDSS is refined and further deployed, trust and confidence in its use can be expected to increase. The MDSS tool as deployed this first season in Maine has provided benefits, and with further refinement, user training, and consistent application it can be expected to offer greater accuracy, consistency and ability to differentiate variable conditions across Maine’s complex geography and local climates. The benefits it has provided constitute a platform upon which MaineDOT can build over upcoming winter seasons.
6.4 Lessons Learned for State DOTs
MaineDOT worked with the initial MDSS deployment over the 2006-2007 winter season in a single interstate corridor around the Portland, ME metropolitan area. The MaineDOT Scarborough crew enthusiastically took on the challenge of incorporating the MDSS into their existing suite of weather support tools. While they had decades of experience dealing with winter storms, they had essentially no training or advance preparation for how to use an MDSS. The initiation of this assessment just as the winter season was getting under way meant that all the stakeholders in this process were learning together as the storms were upon them how an MDSS might affect the crews’ maintenance and road treatment decisions.
This section summarizes a number of lessons that were learned in the course of three months of experience with the MDSS.
An MDSS that offers accurate and consistent forecasts and treatment recommendations is more likely to engender trust and confidence in the users.
A complete MDSS is more than a weather forecasting tool. It is a decision support tool that seeks to guide the timing, type, and amount of materials a DOT will apply to their pavement throughout a storm based on accurate forecasting of weather and pavement conditions. In order for the MDSS to offer full benefits, maintenance crews need to have the confidence to follow the MDSS treatment recommendations. In practice, this means that the crews first need to be willing to apply the recommended treatment type, amount and timing, then, after giving it a fair trial over successive storm events, they can evaluate the performance of the MDSS. If it proves to be effective and offers measurable benefits, then the crews are much more likely to build trust in the tool and use it consistently. If the MDSS forecasts and recommendations appear off the mark, compared to the other available tools or to the crew’s experience, then the MDSS recommendations are less likely to be consistently followed.
MaineDOT experienced an unusually mild winter, which presented its own set of challenges for weather forecasting tools. While there was considerable variability in type of precipitation, temperature, and wind conditions, the weather events typically hovered around the border between rain and snow. This made it quite difficult for the MDSS, and the other available weather forecasting tools, to forecast precipitation type, amount and timing with accuracy. Any such tool, including an MDSS, must seek to demonstrate accuracy and consistency across all the precipitation types that a DOT can expect to experience. Also, the DOT agencies need to ensure that the treatment recommendations offered by the MDSS are appropriately customized for their environment and operations. Then they can evaluate the benefits of modifying their standard maintenance practices to reduce costs associated with labor, materials and equipment usage. However, user trust in the basic functions of the tool must evolve before the more advanced capabilities and potential can be fully realized.
Provide training to maintenance crews before MDSS introduction, and offer additional training and support thereafter in the use of an MDSS to derive full benefits to improve maintenance decision making.
The pathways to benefits presented earlier in this report described many different ways that an MDSS could lead to decision making that could either reduce operational costs, increase benefits, or accomplish both. This assessment made it clear that in order to operate along these pathways, the users of the MDSS need to fully understand how it works, how best to apply it in support of decision making, and how to interpret the information it offers. To accomplish this takes more than enthusiasm and support for the use of the MDSS; it also takes time to learn, time to accept news ways of operating with the MDSS, and time to develop the level of trust in the MDSS to fully accept it into regular operational practice.
The MDSS vendor can offer active support to the maintenance crew that is using the MDSS to explain its capabilities, answer questions that arise, and suggest effective ways to take best advantage of its capabilities. This is the role that was played by the MDSS vendor in working with MaineDOT. The more active this relationship between the MDSS vendor and the DOT, the more effective the MDSS will be in supporting the DOT’s maintenance operations. The MDSS vendor offered MaineDOT the services of their meteorological staff and encouraged the Scarborough crew to call before every storm event to obtain further guidance and interpretation of the forecasts that were being provided by the MDSS. In this sense, the crew can play a role in improving the quality and effectiveness of these MDSS forecasts by helping the vendor fine tune the forecasts based on the observations, experience and interpretations of the local crews. In addition, the crew needs to be willing to act on the recommendations of the MDSS in order to realize the potential benefits it offers.
DOTs should expect that it will take time for their management and maintenance crews to adopt and accept an MDSS into their standard operations.
Organizational change of any significance is usually slow to occur because it is human nature to hold on to what has worked well in the past and to resist change that presents uncertainty and the risk of failure. A basic ingredient for organizational change is strong leadership backing the change and a willingness on the part of the organization’s personnel to try new ways of performing their jobs. MaineDOT had each of these prerequisites for change in their favor. At every level of the DOT, there was strong support for the adoption of an MDSS. Nevertheless, a single winter season, with no advance preparation or training, is not conducive to adoption and acceptance of such a new and different way of operational decision making. As noted in several of the accompanying lessons, adoption and acceptance depend on forecast accuracy in practice, crew training in how to best make use of what the MDSS has to offer, and an MDSS that is closely customized to the conditions where it is being applied.
Configure MDSS treatment recommendations in close consultation with the DOT, and customize them to fit the conditions, needs and practices of the crews in the district where the MDSS is to be used.
In providing a treatment protocol for incorporation into an MDSS, MaineDOT decided to select an average set of treatment guidelines under a set of weather and pavement temperature ranges that could apply across the state. In Maine there are variations in winter weather patterns and treatment strategies across the state. The Scarborough region tends to experience less harsh winters than other parts of Maine and they also have a heavily urbanized portion of the state. The Scarborough crew tends to apply more chemicals under a given set of weather and pavement conditions than the state average treatments that were configured in the MDSS. This predisposition to heavily treat the roads (for example, to apply higher amounts initially to assure effectiveness and reduce the need to “catch up” with subsequent heavy treatments) caused the Scarborough crew to exceed the initial recommended treatment levels on a fairly consistent basis. The MDSS derived its treatment recommendations primarily from forecasts of air and pavement temperature that were linked with MaineDOT’s “average” guidelines that had been incorporated into the MDSS. This resulted in fairly consistently under-recommending the amount of chemicals needed in the judgment and experience of the Scarborough crew.
While a possible strategy would be to tailor the treatment protocols in the MDSS to the prevailing conditions where the MDSS is going to be applied, the DOT needs to be cautious not to negate the potential benefits of more finely tuned treatment recommendations from the MDSS compared with the DOT’s standard operating procedure. A better approach would be to work closely with the MDSS vendor to implement a protocol of treatment recommendations that offer the potential for effectiveness and cost savings, consistent with the DOT’s standards for road maintenance. Arriving at an optimal configuration may be an iterative process, with the DOT and the vendor reflecting on the performance of the MDSS after trial periods. The DOT may discover in this process that their existing treatment protocols can be modified to provide greater efficiencies and savings.
DOTs should select the alert topics and alert timing that will be most helpful in making their road treatment decisions while avoiding generating too many alerts that become more distracting than helpful.
The MDSS offered MaineDOT a wide variety of possible alerts covering all the possible event types for each forecast point created for the study region. The more alert topics, forecast points for which alerts are requested, and frequency with which alerts are generated, the more total alerts that will be sent and received. Across the events reviewed in this evaluation for which alerts were provided by the MDSS, an average of 50 alerts per event and about 3.3 alerts per hour were generated based on the criteria specified by MaineDOT. In the judgment of the Scarborough crew at the end of the winter season, this was such a large number of alerts that they found themselves disregarding many of them and at some points finding them distracting. Later in the winter, the supervisor elected to cut back on the number of alerts and stopped providing them to every crew member.
The number of alerts has to strike a balance between too few to provide timely and useful warnings, and too many such that they become a distraction and end up being ignored. In part due to the small geographic area covered by this study, alerts for each of the designated forecast points proved to not be necessary. Alerts issued for different forecast points are only helpful when the information at each forecast point is significantly different from the others. Receiving multiple alerts that all say the same thing is less helpful. The DOT will need to assess their experiences with the MDSS alerts and fine tune them to best meet their needs.
MaineDOT’s experience suggests that it would be helpful to configure forecast points that are outside the maintenance area along storm paths to offer improved advance warning of storms. In this way alerts can be generated based on weather conditions that are known to be likely to precede the conditions that will be experienced by the maintenance crews. MaineDOT’s traditional strategy for alerting crews has been to communicate with NHDOT and with other crews adjacent to their crew shed area to get advanced warnings of precipitation type, intensity and timing. The MDSS can effectively supplement these warning strategies by configuring alerts efficiently.
Maine DOT learned that alerts do not need to be continually issued over time when the information content has not changed significantly. Issuing alerts only when conditions have changed is one way to reduce their numbers and maintain their salience to maintenance operations. Alert wording should be short and clear, with the main objective of alerting the operator to consult the MDSS for more detailed information. Alerts serve as a warning of an impending storm event and should prompt the DOT to actively monitor the MDSS as the storm progresses. Alerts need to be configured for users according to the times of day when the crews are prepared to receive them and make best use of them. When a crew member is off duty or asleep, the alert can be directed to a supervisor or radio operator who is on duty and can act on them immediately. This is how MaineDOT provided 24/7 coverage.
MaineDOT’s experience with the MDSS alerts illustrates the value of evaluating alert information in conjunction with other sources of data on impending or current weather events in order to derive the best information from both forecast and observational data in support of decision making.
Provide a mechanism in MDSS treatment recommendations to incorporate the effectiveness of prior treatments, along with observational data, in order to better adjust forward-looking recommendations.
The MDSS used by MaineDOT was of most value at the front end of a storm event and less so during the event, partly because of its lack of ability to factor current conditions into its recommendations, partly because data on crew actions and resulting road conditions were not being fed back into the MDSS, and partly because the crew felt they were too busy “fighting the storm” to pay closer attention to the MDSS during the storm. The Scarborough crew evaluated the MDSS before every storm event, along with their other sources of road weather information, and made their decisions whether to pre-treat the road and what type and amount of chemicals to apply. While they didn’t necessarily follow verbatim the recommendation of the MDSS, they incorporated it into their decision making. Other than for pre-treatment before a storm began, much of the treatment decision making that occurred during the storm was made by truck operators on the road under the stress and pressures of “fighting the storm.”
Maintenance crews need concise, timely information mainly about variations in conditions across their region of responsibility and forecasts of changes in the timing and type of precipitation, air and pavement temperature trends, and guidance for optimal treatment types and amounts based in part on actions they already have taken. The supervisors as well were often not at their desks in front of their computers during storm events to monitor the MDSS. As a result, the Scarborough crew consistently consulted the MDSS before the event but not regularly during the event. DOTs can enhance the potential accuracy and benefit of an MDSS by making RWIS/ESS data available for the geographic areas to be covered by the MDSS forecasts and treatment recommendations.
State DOTs can leverage the potential value of an MDSS deployment by offering it initially to one or more of their more progressive crews.
MaineDOT viewed this experience with an MDSS as a way to show their crews statewide that there are potential benefits and advantages to using an MDSS in support of winter maintenance operations. The more advanced crews can then serve as an example as well as offering training to other crews.
An MDSS is a very different and more complex technology compared with many of the systems used throughout Maine and in other states. Maintenance personnel who are uncomfortable with computers and related technologies may be reluctant to accept and work with an MDSS initially. MaineDOT offered the MDSS to a number of their maintenance crews and selected the Scarborough crew for this assessment project based in part on their enthusiasm and willingness to work with the MDSS throughout the winter season.
An MDSS should be viewed as a valuable tool in a DOT’s winter maintenance operations tool box, offering the benefits of an integrated GIS platform and significant educational value.
MaineDOT routinely accesses a variety of weather forecasting and informational sources that include local weather station broadcasts, private weather services, local ASOS, and regular communications with NHDOT. They access data from five RWIS/ESS located throughout the state. This past winter season they also had access to the MDSS vendor’s staff of meteorologists to provide interpretative guidance that formerly was obtained through a contract meteorologist.
MaineDOT found that the MDSS supplemented these resources in some important ways. First, the MDSS added capabilities that they previously didn’t have from their other tools. These included pavement temperature forecast trends, bridge and road surface frost forecasts, and a tool that could provide pavement treatment recommendations based on an analysis of multiple weather parameters. Second, the MDSS offered an integrated platform for the display and analysis of NWS forecasts in a user friendly GIS format. This capability was particularly valuable for use in their Augusta operations center where they were able to track storms across the entire state and notify their crews selectively and strategically in advance of severe weather. Third, the MDSS as used in this first winter season offered significant educational value for the crews that used it. Crews became much more aware of the variety of valuable weather information that is now becoming available and how to interpret that information in the context of their existing tools and extensive on-the-road experience fighting storms. This experience with the MDSS has positioned these crews to be able to more effectively utilize MDSS capabilities in the future.
The pavement temperature forecast module is an example of an important element in the MDSS that must be well understood by a DOT to provide benefits. The MaineDOT Scarborough crew regularly tracked pavement temperature with their IR pavement sensors on their trucks and used that information in making their treatment decisions. What was new in this study was the availability of pavement temperature trend forecasts that could help the crews anticipate whether the pavement was expected to warm up, cool down, or stay about the same for several hours into the future. The evaluation team prepared graphic representations of these trends, coupled with ASOS readings of air temperature from the Portland Jetport to illustrate what the MDSS data were showing in each storm event (see Figure 6 for an example). Also, Table 1 on MDSS Pathways to Benefits outlines some of the key decision pathways associated with pavement temperature trend forecasts provided by an MDSS. A DOT can use this information from an MDSS to help them decide where to treat, when to start treatment, what types and amounts of materials to use, whether pretreatment is necessary, and the likelihood of road or bridge frost occurring.