Welcome and thank you for standing by. At this time all participants will be on a listen-only mode. During the Question and Answer Session you may press star and 1 on your touch tone telephone. You will be prompted to record your name so that you may be introduced. I would like to turn the call over to Jocelyn Bauer. Ma'am, you may begin. Hello and welcome to the national transportation operation coalition talking operations webinar on Work Zone Performance measurement. My name is Jocelyn Bauer, and I will be giving a brief introduction to the web conferencing environment before turning the session over to Darren Buck from the federal highway administration who will be our moderator for today's seminar. Please be advised today's seminar is being recorded. It will last approximately an hour and 30 minutes with the first 60 minutes allocated for the presenters and the final 30 minutes for audience question and answer. During the presentations if you think of a question you can type it into the smaller text box in the chat area on the lower left side of your screen. Please also make sure you're typing into the thin text box and not the large white area. Make sure you send your question to everyone rather than just the presenters. The presenters will be unable to answer your questions during their presentation, but Darren Buck will use some of the questions typed into the chat box for the Question and Answer Session in the last 30 minutes of the seminar. A file containing the audio and visual portion of this seminar will be posted to the national transportation operation co-lix or NTOX website within the next week. I will type that address into the chat box shortly. We encourage to you direct others in your office who may not have been able to attend this conference to access the recording. The presentation used today or the presentations used today rather are available for download in the file download box on the left side of your screen. To download a file click with your mouse on the name of the file you would like to download and click the button at the bottom of the download box that says save to my computer. The presentations will also be available within the next week on the national transportation operations coalition website. Attendees will be notified of the availability of the presentations, the recording, and the closed captioning of this seminar through email. At this time I would like to introduce Darren Buck, the moderator of today's webcast. Darren is the marketing specialist for the if I recall highway office of operation. His duties including overseeing the outreach activities of the national transportation operations coalition. Prior to joining federal highway in 2008 Oludare lean worked in similar roles with a local small business and in other federal agency. Now I will turn things over to Darren who will start us off. Thank you, Jocelyn. We're very happy to be here. My colleague Traci Scriba has lined up a great slate of speakers the topic of Work Zone Performance measurement. Traci Scriba is a program manager here in the office of operations within federal highway administration headquarters work zone mobility and safety team. She leads FHWA's efforts for implementation of the work zone safety and mobility rule and is responsible for FHWA's program areas related to work zone data and performance measures, work zone best practices, and ITS in work zones among other things. Prior to her work at FHWA she worked as a consult ability on both transportation and environmental issues for over ten years. She holds a systems engineering degree from the University of Virginia, go Cavaliers, so without further delay I will durn it over to Tracy. Thaws,Thank you, Darren. Are we ready to go with the slides, Jocelyn? There we go. I am going to kick it off here and kind of get us introduction to the topic of performance monitoring and the safety and mobility rule and then the following two speakers, Jerry and Dave will give a lot more kind of some of the practical details in information behind that and give you an example of how performance monitoring can be used. I am going to kind of cover two areas in my presentation. The first is just why we need work zone data and performance measures, and then how performance monitoring relates to the work zone safety and mobility rule. So the first area I would say why we need data and performance measures is basically to understand and improve performance. Work zone as best as we might try to minimize impact they do cause some safety mobility impact often times, and those impacts do have an effect on how our system operates. So if we have data and measures we can better assess how those impacts are affecting our system and make decisions that would help mitigate those impacts. The second area I am going to talk about is that they provide a basis for decisions and spending, both making policy kinds of decisions as well as making decisions about how to spend resources so we can spend the resources effectively, so I will talk briefly about that. Then the last area I want to talk a little bit about is using data and performance measures as a tool for communication. Work zone performance measures can really help us communicate with the public, with elected officials, with the media, et cetera, about how we're doing because as we see with the internet, with face book, various different kinds of online groups that information gets out one way or the other, and it would be good if we have some good information to help share. So the first area just to briefly talk about how work zones do affected performance, you can see on the right there that is a pie chart showing an estimate of various sources of congestion, and work zones are estimated to cause about 10% of overall congestion on our transportation system, surface transportation system. On the left there is showing the larger portion that work zones tend to cause as far as nonrecurring congestion, so work zones are thought to cause about a quarter that far kind of congestion which because it is more unpredictable, it can cause more issues both for safety and mobility than say a typical bottleneck where commuters might be used to having congestion. This is just one example of how an agency has tried to track this kind of information to better understand their performance. This is Missouri DOT, and to meet their goal of improving mobility and safety for both travelers and workers in work zones, they just identified this need to better understand and track their performance, so they looked at both traffic flow and some safety measures. They actually initially started asking for this input on a form from their nontechnical employees in the DOT, but recently have actually posted the survey form on the website and allow the public to also give feedback in this area, so these results are compiled quarterly, and reported to both the public and the senior management, and they use those results to look at how they might need to update things in their practices and policies and specs. Another way of looking at it is kind of trying to ask a question, can we answer the question because with you don't have data, it is tough sometimes, so for example in the number of work zone fatalities increase from one year to the next in an agency we doing better or are we doing worse at work zone safety? Can we really say? It is the typical lawyer's answer. It depends. If the number of fatalities went up by a certain amount but maybe the amount of work increased by a whole lot more, maybe in reality we're doing in some ways a better job at safety because even though there is more exposure we made some improvement. So we need some data to help answer those kinds of questions. Then the last one in this area is to help us better understand and quantify impacts of work zones and the effectiveness of the type of strategies and measures we're taking to help manage those impacts. That can help us be more effective on future work zones if we have that understanding, we need some information and data to be able to do that. We also need to have some understanding of what are the measures measures we're looking at in order to know what data to collect and how to measure our progress. Kind of one-way to look at this is which work zone will I get? If we have some data and some information from past work zones and we can use that to predict impacts our upcoming work zone, then we have a means of trying to maybe not always going to get the right work zone on the right there but we have hopefully a means of helping minimize the times we get the work zone on the lower left there. The second area we can use data to provide a basis for our decision making and our spending so we can guide investment decisions, what kind of policies we set. I know some agencies have a maximum allowable delay or a maximum queue. Are those types of things useful policies? How do we know if they're actually getting met? We need data to do that. That would be another way of looking at how we can use data and performance monitoring. In this case how do we decide how much to spend, for example, on upcoming work zone? We know it is going to cause problems, going to spend money to help mitigate the impacts as best we can but where do we draw the line? Sometimes the way of doing that has been looking at someone who is very experienced who just has kind of been there and done that and knows what to expect based on the past which only goes so far. It is useful, but it only goes so far. What happens when Joe retires or the John doesn't go as smoothly as planned and the Mayor's office wants to see the planning for the project and manage the traffic? We need to have something to show for that. Lastly, as I mentioned a tool for communications, aid in communication and outreach, agencies have a story to tell about we're doing a lot of hard work to try to make the transportation experience better for motorists, for commercial vehicles out there, for all of those using the system, and it is good if we can get our message out there and we have data to help tell our story. Both for the public as well as the elected officials that makes the decisions that do affect agencies. The example I showed earlier from Missouri DOT is would be an example of doing this. They release the information quarterly, and it does get available, made available to elected officials and the public. This is maybe a humor us way of looking at it, but it is an example of reality. The cartoon with the general there on the phone taking action as a result of a work zone backup isn't only in the cartoon cartoon. As the quote shows on the slide, but also even more than that one state has ended up with a quickly conceived work zone delay or Q length performance measure as a result of an official's unpleasant work zone travel experience. This is not the recommended approach to determining work zone performance measures or to telling our story, but if we don't tell it, others will. Another aspect is the accountability aspect of telling our story. In general performance measures in all public agencies including state DOTs has been increasing for a little while now, and so in the scheme of all of these various ways of measuring performance work zone performance measurement needs to be coordinated and included in those efforts. Even though the use of performance measures has been increasing, work zone performance monitoring really is still a developing area. I kind of like this picture from that sense of how does the driver see our work zone? This is trying to have some of that sense when we're looking at performance monitoring. There is that kind of angle of performance monitoring as well as the things I mentioned earlier about the internal use of performance monitoring and data. As far as it being a developing area. We do a work zone assessment in our division offices in all the states, and there are six main areas to the self assessment. As you can see from there the program evaluation area is flagged a good bit behind the ratings in the other areas. It is overall a 15 point scale, and some of the lowest rated items fall in that program evaluation area such as does the agency collect data to track work zone and congestion and delay? Does the agency conduct customer surveys, for example? Has the agency established measures to track work zone congestion and delay and to track work zone performance? So these areas definitely still have a lot of room to develop. However, one very hopeful note in that area is that the question about tracking work zone congestion and delay while it has the lowest score of all the 46 questions of the self assessment, it does have the biggest increase between the last two years, so there is a lot more attention certainly being paid to that area. And then the last little piece here talks a little bit about the work zone rule and how data and performance measures fit into the work zone rule. This is the work zone safety and mobility rule. The overall goal of the rule is really in a nutshell to better understand and manage work zone impacts, and there are three provisions in the rule that I would say most closely kind of relate to that. One is the impact assessment area, one is the use of data, and one is the area of process reviews. Now, overall part of the idea of including them in the rule is to help move practice forward in that area to nudge it along, so I am want going to talk in detail on these but just to briefly touch on them, in the impact assessment area the rule recommends that agencies have procedures to assess impacts during project development as the project is going through preliminary engineering and design and then developing the transportation management plan that there would be some element of assessing impacts as the project goes forward. In practice that requires some data and performance metrics. If this project we think this project is all set to go, do we have a policy that, for instance, has a we don't want to have more than 20 minutes delay? Well, how do we know that project is probably going to fall within our policy? There may be some use of analytical tools in some work zones to help us do that. In the data area the rule addresses data on two different levels, the project level and the process level, so the project level is meant more while the project is under way, and the process level is meant more over time what do we learn from looking at multiple projects? And then at the project level the rule requires agencies to use basically whatever available data, field observations, crash data, operational information, to help manage impacts during projects during implementation. Certainly the idea being that if you notice a problem and you can do something about it, why wait until you've done a study later or something like that, but trying to address issues as they come along. There is no requirement to collect additional data, but in reality if there is little or no data available currently, then some new collection of data or tracking might be appropriate. At the process level as I mentioned it is kind of look at multiple projects over time to identify trends, see if there are issues that come up only seem to jump out when looking at multiple projects, and I know Ohio has done some of this. Dave is speaking later, but they've done a good job at identifying trends overlooking at multiple work zones. The last item, the process reviews, is essentially kind of a higher level view, stepback kind of view every two years of looking at whatever information the agency has available, some of that might be from your traffic control inspections, field reviews, post construction, et cetera, but looking at that and taking a step back and saying, okay, what does it tell us, you know, and we have performance kind of tracking how are we meeting that over time and do we need to make any changes, so it is a little bit more that process level of the work zone data could fit very well in with the process review every two years. I will close with just mentioning a few challenges. Certainly there are several to this, and first and foremost is a lack of data, and then a correspondingly a or limit of personnel and infrastructure to gather that data, and I know that is something we encounter in the effort Jerry will talk about next with looking at what DOTs do have available and what are some various ways that maybe can try and get data to help be able to do some performance monitoring. Certainly having agreed upon metrics can be a challenge as well to decide if you're going to track a couple of things what would you track? So just close here with saying a couple of our current efforts here at federal highway are trying to get a number of these issues, both with the pilot test Jerry will talk about, with sharing examples and best practices such as what Dave is going to talk about is certainly another effort, and then also trying to kind of track the practice and what is going on there and what might be some ways that we can help with supporting and moving that practice forward. At that point I will turn it back to Darren here. Great. Thank you, Traci. We did have one question. Do you want to address it now or wait for the question and answer period regarding what tools Missouri DOT uses to measure traffic flow? Why don't we wait. Sure thing. All right. Our next speaker is Gerald L Al mode, the senior research engineer and manager of the work zone and DMS program for the Texas transportation institute. Doctor Ulman has over 24 years of work zone safety and mobility research and technology transfer and has published extensively on this topic. He is currently the chairman of the transportation research board work zone traffic control technical committee, and is a member of the temporary traffic control technical committee of the national committee on uniform traffic control devices. He is also a member of the institute of transportation engineers and a registered professional engineer in the state of Texas. With that, Dr. ulman, take it away. Thanks, Darren. Appreciate the opportunity Traci allowed me to come here this afternoon and give you a brief update on our efforts on a task that's been about a couple of years in the process where we're currently wrapping up on work zone performance measure pilot testing. Our goal of the pilot test as Tracy alluded to was to see if we can identify a few useful performance metrics that states could consider looking at as established program monitoring programs in their jurisdictions, and then look at if we identify these as useful metrics, how we can go about capturing the data we would need to compute them and then on the way document what we learn in terms of what to do, whatnot to do, things that states might want to think about as far as things to watch out for. We'll note that what I will talk about today is focuses primarily on measurement of mobility type measures, certainly the safety measures, also play into this. Dave will talk more about what Ohio does in the safety arena, and so I will targeting the measurement of mobility types of issues. We did start this process with reaching out and trying to talk to various individuals at state and local level about what metrics need to be and what would characterize a good performance metrics. We also relayed on a recently completed study we did for Texas DOT looking at this issue to gather some insights on that. Have a list of the high points of what we heard across the country, so some seemed fairly transparent, straight forward. I think one of the biggest things to take away from this slide is that the measures as you do define them need to truly reflect the realities of what we're trying to do here which is get a particular work or a series of work activities completed. A lot of times trade offs have to be made, get in, get out, get the work done itself as quickly and cost effectively as possible. We're certainly worried about mobility, safety for the traveling public, but it is just one of the many things that go into this process, and we don't want to make sure or we do want to make sure the metrics we use characterize the fact that sometimes things have to be done in a way that maybe less than optimum if it would just only considering mobility, for example. In a sense we need a whole suite of measures. We didn't identify any one or even two measures that would do the job in that regard. Go through a listing of what we have identified as sort of the big Is as far as metric measures that need to be looked at. Certainly exposure is one of the biggest ones that needs to be captured. Tracy alluded to some of the read it while you would, without exposure we have a very difficult time characterizing how well something we get a number, crashes, for example, but knowing whether that is going up or down relative to level of work activity without exposure, you just can't do that. Traffic related exposures important. Work activity itself exposure is important, so that we can characterize it in that regard as well. Another category if you will is the amount of capacity losses that we're experiencing as a result of this work activities both in terms of how much, when, and the length of such capacity losses can be very critical, very important from a monitoring standpoint when we go in and want to compute some additional terms. The other big category of measure that is popped up is that of delay. Certainly in terms of both global measures if you will, vehicle hours, ultimately relates back to road user costs, incentive, disincentive of measures and that kind of thing that agencies utilize. It is a big category of metrics, and at the same time we are concerned about how our work zones impact the individual traveler, and so we also need metrics that relate it back to that level of detail as well. The other category which is related to delay but is also gives us a slightly different and interesting perspective as well is that of queueing. We do have some agencies that themselves have established policy statements that relate to the avoidance of queues entirely or if we're allowed define what lengths and durations of queueing are acceptable. Like I said earlier, they do relate direct queues do relate to delays generated, so that's important from there. From a safety standpoint we know that presence of queues does tend to increase cash rates, particularly rear end, so we are concerned from that perspective, so the documentation of how often we have these, how long they last, how far back they go, and which get beyond our advanced warning, for example, all of these things come into play. They're important things to measure. Finally the last category is travel time reliability. Reliability is getting quite a bit of attention across the country as a useful tool performing operational performance, drivers desire reliable travel time as much as they desire quick travel, rural areas, reliable, it is not a big issue. You generally have free flow unless an event such as a work zone activity disrupts the flow, but in urban areas we can have problems existing even prior to the initiation of a project and monitoring how things such as reliability are influenced during the project is an important aspect or perspective that we need to look at or should be looking at from an operation standpoint. As far as our pilot test, we identified these categories of metrics we were interested in. We took at look at well how are you going to go about getting data that you could use to estimate these things? Three basic categories of data jump out at when when we talk about this, that of project information itself, if you're in an area where you have traffic data traffic sensor data, that is obviously a potential useful source, and unfortunately in many case as cross the country our projects are located in areas where we don't have such instrumentation and in which case what do we do? There are things we can do that we're showing through this pilot test can do manual methods if you will using things that field personnel can document for you. Project information itself is what we're finding out you have to have this regardless of which approach you're taking to measure mobility or the things like delays and queues. Have you to have the project information itself. You need traffic control plan for capacity, long-term capacity losses, if you're shutting down lanes on a continuous basis over the duration of the project. You need obviously limits themselves, one of the other things that comes into play is that even a particular work work zone can change how it looks and how it influences traffic on several times through its life depending on how it has been dissected into various phases and that kind of thing, so having phased changes and those types of things captured allow you to do the type of analysis and measurements that truly reflect what is happening and what did you in those particular phases. Daily project diary information is the final category that comes into play, very important, the exposure type of data we use to make our other come upages in performance measurement wise. From a traffic sensor standpoint, you basically got a couple of categories or ways that exist. One-way is the use of spot sensors, they may be in the payment more recently now, typical non-in-true sigh devices located periodically along the roadway segment and measure speeds and volumes and occupancies from that. For our purposes in related to metric that is we identified earlier, we then need to do some come pu taigsal processes to get at what we would also like to know, in particularly the travel times and the queue lengths that we would be interested in from a work zone monitoring standpoint. Travel time estimations fairly straight forward obviously, the segment or the approximate distances over which you think the sensor data has given you realistic data or represents that data is the speed and you can perfect the travel time on a segment by segment basis and compare it to what your threshold or your desired flow or rate is, your travel time is to get your delays from a queueing standpoint you really have to go back to the looking at the speed profile at each of the sensors, looking for periods which speeds drop below what you define as acute flow conditions and then using the locations of those sensors, how far away they are from your work zone, to indicate lengths of queues, how long the duration of the queues, those kinds of things as we illustrate on the bottom of this slide. Another approach, some locations have their systems developed around more of a traffic vehicle probe system, so the Houston district of text relies on vehicle for supporting the transportation surveillance network. There are other sources as well for our purposes we actually were able to make use of the freight operations and management data set, FHWA's office of operations, freight management operations, group, has truck transponder data, it is gathering across the country, for another effort that was made available for our purposes to look at one of our locations, how well we can use the probe data for monitoring traffic conditions. This way you actually got travel times measured. You have to estimate queue presentation and detection or length based on the extent to which you segment your roadway sections and looking at speeds in those sections over time as well. The third approach, the manual approach that I had alluded to, it is something that a few states have started looking at. We were very successful in getting individuals hat our pilot test locations to agree to do this and basically rely on someone in the field that's there on a daily basis or if you work at night, nightly basis obviously, they're documenting if queues occur on a particular work shift, when they start, when they dissipate, on the duration. They estimate the best abilities, whether it is a matter of looking upstream and guessing, if it is a matter of driving through there with a work vehicle, whatever it is, you periodically do that, and you document the length that it was at every hour or however frequently you can do this while you're taking care of your other duties. From there you then can also estimate delay, and combined with either estimated historical traffic volume section by section you can get at more of our global measures of delay and vehicle hours, that kind of thing. The process for converting maybe queued length estimates to travel time drey is highlighted here. Won't spend a whole lot of time talking about. Basically the approach is to assume you can approximate or compute a what you believe to be the expected speed through the queue, this is a very simple traffic flow theory, comes from some linear relationship between assumed between speed and density, and then assuming that once you're into the work zone because you're at capacity flow, that the speed through the work zone is also at capacity flow and therefore is contributing some reduced travel time, increased delay and we compute that periodically for the measurements and then calculate across the duration to get at our other measures of delay. I have a little data I want to share with you from pilot tests and locations we were able to identify and gather some useful data from. You can see the locations here. The Las Vegas, Seattle, and Philadelphia areas are in urban areas, Seattle and Philadelphia were night time projects, Las Vegas one is a little different. We'll get to that in a minute. It involves some long-term lane closure activity for that project. The lumber ton prog in North Carolina was more rural but still on interstate facility that during daytime hour when is they were doing the work periodically did create some congestion. Given our indicator of the importance of exposure data, we did go through the process of collecting the types of information I alluded to earlier, project diary information, limit, that kind of stuff and have shown here a some of the exposure measures we were able to calculate, and as you can see there we characterizing how much of the time of the project is being attributed by work activity for example can be a very useful measure, how much capacity by lane and a number of lanes closed also very important, able to access sources of traffic data to look at what effects we're having from a standpoint of number of vehicles or vehicle miles traveled exposure to these work zones. One of the things we did, reasons we went to the four sights that I listed here is the fact that those these are the locations gave us a chance to collect the or try both approaches to collecting the mobility operation data that we identified earlier as being ways of collecting this data so that we could do a direct comparison and see, well, how well do these two approaches agree with each other? Here is some data from the Philadelphia project on I-95 which is a nighttime resurfacing and some other type of work activity project, field crews were documenting each night when queues occurred and how far they went, same time we were using traffic sensor data to see whether by knowing that there was work that night could we identify queues and how long and all of those kind of things. Something agreed pretty well as you can tell. Some things do not. This particular location things we learned was that sensor spacing of your surveillance system really has a big impact on how well you can be able to track and estimate work zone performance measures in the locations where we were doing this, they were doing this work, our sensor imagings were anywhere from a half-mile to 4 miles, and since you're going to be assuming that whatever you're measuring at sensors is representative of a certain length of sensor which is tied to how long between sensors, your accuracy is going to go down when you have the long distances between sensors. Same would occur if you had a portable ITS system deployed and you were trying depending how far apart you have the sensors you have the same trade offs between accuracy here. The 405 site had a few more sensors deployed. Unfortunately many of these were being turned off late at night as doing various work activities and in the vicinity of where we needed the data would be when they would turn off a set of detector stations if you will or power is disrupted and the data was not available, not real sure quite why the reasons were, but it does also highlight another thing that you have to consider relative to using sensor data is that you rely on that and your work activities require the loss of power, you're going to have holes in your data in that regard as well. We have data from the I-95 project in North Carolina. We're just getting the traffic sensor data. This is location where we -- the office of freight operations is providing us truck transponder date A we did have data prior to day for one particular day where we knew we had congestion, so I wanted to share that with you, just to illustrate that the probe vehicle approach is also going to be or at least from what we can see here a viable approach. You see a little bit of disparity, but overall the numbers do tend to agree with in terms of how long we were able to detect congestion, how long of the queues were there, what delays we were getting, those kinds of things all fairly reasons well, and the key on this one is from the sensor standpoint, the roadways, what you have available in terms of market penetration of probe vehicles, highly depends or highly affects how accurate and detailed you can do this kind of analysis. On I-95 we have a large truck population, so a number of those were instrumented with transponder that is provided data if we had been on a facility with much lower volume or less truck traffic. I do not believe we would have been able to get to this level of resolution. Something to think about as well. In addition, I have been somewhat harping on the maybe seems like about the concerns about traffic sensors and what you have to be watching out for and using that data. That's really not the only consideration that at least we found that we also have to be careful or concerned about when we're doing our analysis. Our assumptions relative to how we convert our data into these other tools is also important. This slide here I've got was a -- is a comparison of speeds estimating in queue via our simple traffic flow model, what we have measured in via some sensors in these locations, and then also we did a number of some travel time studies at these lox and actually measured speeds in queue and lengths and that kind of thing. The range that I show for the traffic flow theory is what we got or would get depending on whether we assume the queue was occurring upstream, the lane closure if you will and backing up across available lanes and really a true stop and go congested environment versus if we as we encountered or looked or found when we were looking at our data from our pilot test sites, sometimes the queues develop within the limits of the work zone itself, so it is really not a bottleneck being created but could be work vehicles, drivers slowing down just in the vicinity of the work activity to do a little rubbernecking or whatever, and anyway regardless of the reasons if they're within the work area we don't necessarily expect speeds to be dropped as low because we're basically assuming there will be flowing through about capacity flow or a little bit less, so you get a very big range depending on where you assume your queueing is occurring. We did not have our folks in the field documenting where the queues were occurring. We think that's something that should be added to future agencies that are going to try this approach in the future, document whether it is within or occurring upstream will help you be a much more accurate in your estimates of speeds and ultimately then delays that would be resulting from that. The last slide I wanted to highlight is the related to travel time reliability, and measuring that in Las Vegas I indicated they have a project where they have closed one lane in each direction to do some basically median or express line widening in a section near the casino area downtown. They're doing it in multiple phases, and one phase they were closing right, the outer lanes and merging areas there, and then Phase II they're actually closing the median lanes in both directions. I have a little bit of data here that shows you the average and 95 percentile travel times for various periods of time from this site. This is typically how you assess buffer indices which is the ratio or the difference in between average and 95th percentile travel times. The things I wanted to point out is you do see between the two phases differences in terms of how that construction decision or that traffic control plan if you will is affecting traffic flow. In Phase II in the median they're getting much less disruption, traffic flow is flowing much more smoothly, so consequently the increases in travel times are not nearly what you saw in Phase I where they were affecting the right lane, lane, merging vehicles on and off and much more turbulent, so again using these tools or using this indicator, a a useful -- we believe a wear of characterizing performance. It is also interesting, though, you do see some odd kinds of things occurring here such as in some cases during construction the buffer indices is going down which would indicate that you're getting more reliable travel times during construction because basically you're elevating that average and the real excessively high travel times are not getting much higher, so you're getting a decrease there and very interesting characteristic there. Whether that is because of truly that we are getting that or in this case we have an issue in terms of not having instrumentation the full length of the project limits that we think are being influenced which could also influence what we're measuring, so ensuring that your data sources are extending beyond the limits of the impacts you believe to be occurring is another important consideration computing this kind of data or metrics. So that is what I want to update you with on the process here. A couple of key findings, lessons I listed here, we do have I think we have shown methods do exist that we can monitor performance. We do need several measures to fully capture that as you see here. They need to be defined relative to what agencies goals and policies are, objectives are. I think it is important to recognize that even if your traffic analysis impact analysis prior to construction are indicating you're not going to have congestion that, that can occur. Your estimates of demand may be a little off, capacities are a little bit off, or can be off, and just normal daily demands and fluctuations are important or can influence this. A couple other things as far as data collection tips to take away, documentation of where the queue is occurring is important, consistency of documenting the data on a daily basis itself is important obviously and from a traffic sensor standpoint spacings and nonfunctioning sensors both play into the extent to which you can accurately measure these things. That's all I had. Thank you very much, Jerry. We'll punch it directly to David our next speaker from the Ohio Department of Transportation. He is an engineer there working on work zone system performance issues. Take it away. Hi, everyone. Thank you, Tracy, forgiving me the chance to be involved in your presentation. I appreciate it. Work zone date A we're big believers in work zone data in Ohio not only because it is required by the 630j regulations but as we look at crash data we're taking a lot of lessons learned from past mistakes, changing standards, changing policies based on the analysis, and I will go into some of these later. Basically the three kinds of data I will talk about today are historic data which is crash, near realtime data also crash, and something we're kind of starting to move toward is the realtime mobility speed data, so those are the three general categories of work zone performance measures data, whatever you want to call it we look at here in Ohio. Starting with historic data, okay, here is one form format that we looked at. This graph just shows the overall number of work zone crashes in Ohio. Let me turn on my pointer here. Back in 2001 we had 8,039 work zone crashes, in 2006 that number had dropped to 5,772. These other bars represent the size of our construction program in terms of dollars, so as you can see from back in 2001, the size of our program actually grew considerably while the work zone crashes went down, and we attribute that a lot to some of the policies and procedures that we started phasing in back in 2001, 2002 timeframe. Another way we looked at crash data was a simplistic before/after crash rate of large work zones we had, usually on the interstate, what we did is went back and looked at the crash rates in this same sections where the work zones were, looked at the crash rate before the work zone was present, and then the crash rate that resulted when the work zone was actually in place, and you can see here back in 2002 we went from a 1.04 without the work zone to a 1. 68 with it of the that was very disturbing to us. 62% increase. By 2006, and I would like to believe this is possible, but we actually improved the crash rate by installing a work zone N 2006 we had broken even, so I am not sure the data is 100% accurate. I think you can see there is a very positive trend here. This is just the to show you the calculations we went through. I know you can't read T you can download the presentation and look at it if you feel like it, but this is just a sample of the analysis we went through. There is another way we look at historic data how we're doing compared to other states. What this is is sort of a work zone crash rate if you will based on size of construction. This is the number of work zone fatals per $1 billion of construction program. We had to use bars and some other sources of information, but basically being out on the right as far as you can is obviously better than being on the left, and if you're one of the the states on the left, I apologize. This is a way of looking at comparing work zone crash rates in terms of construction program size because I think Jerry or Tracy was talking about exposure being a major element in all of this analysis, and the bigger the program you have, probably the more exposure you have. Here is another way we look at it, and we go through and do this part of the regulation. We go back and look every two years at historical work zone crashes. What this one represents is from the periods of 2000, 2005 we're focusing on work zone crashes, and some of these I guess were a bit surprising to us. A lot of our fatals involve construction equipment, and workers, not so surprising was the fact that a lot of them involve speeding, and we were a little surprised by the amount of nighttime crashes we're having in our work zones, and one of the things that resulted from this was a vast improvement of our policies, procedures, and standards for delineating our work zones at night, so if you ever drive through an Ohio work zone these days, it is extraordinarily well delineated. We changed all of our specs and material prequalification for the striping. We really put an effort into it. This is another way we're looking at the data. For instance, this first one, percent of fatal crash that is are rear end, the number not in parentheses out here on the average was the percentage that are in the work zone, the number that is in the percentage on the general system, so in this case we have a much higher percentage of work zone crash fatal crash that is are rear end than we do on the general system. Again what, we do with this is we start looking for ways to improve our standards, our specs, our policies, our planning processes, things of those nature as we go through the historical data and find out where the problems are. Moving to the next kind of format or form of data we look at is near realtime. What this is is something we set up before the construction season every year. We pick work zones, high profile work zones, busy work zones, work zones with potential to have issues, and what we do sigh a three-year prework zone crash summary, and we break the work zone in half-mile segments. As you can see here from this half-mile segment we would expect 16 crashes during the construction season if there was no work zone present, and that's just the historical number, and this half-mile section 22, and this half-mile section 31. Now, the yellow bar then represents how many work zone crashes there are. These are the months here at the bottom, April, May, June, July, and so on. What you can see is here in May we would have expected to have about eight crashes, just historical, no work zone present, expect to have eight crashes but what in reality and that May we had 23, and that's an alarm for us. To get the data in realtime, we have to work out in advance all the enforcement agencies basically set aside the crash reports in a basket and we pick them every two weeks and input them into this application we write that creates all of these graphs and charts. This turned out to be a very powerful thing for us because sometimes things go wrong in the work zones, and they lead to crashes. Before when we looked at historical data, it was usually two or three years after the fact and too late to change anything and the crashes continue W this near realtime, it is near realtime because there is a lag for two weeks to pick up the crashes what, we can do is identify when something went wrong within two weeks of it, just based on the data lag and go out in the field and figure out what it is and change it, and we have done this any number of times where the data said, look, something has gone orably wrong in the half-mile section of the work zone. We need to go out and fix it, and we have, and I really believe this is contributed as much as anything to that decline in work zone crashes because when something develops in the field on these large projects we're generally become aware of it, we go out in the field, and we make changes, physical changes, phasing changes, whatever it takes to alleviate the problem, and I think we have avoided hundreds if not thousands of crashes in our work zones in Ohio by monitoring things here in mere realtime. The application we put in data in, it has automated statistics for us, just in the office where we could take a glance before we go out and see what's going on. These are just samples of some of the data outputs and graphs and charts that we can get from our application. This is nothing more than just for 2008 these were the work zones that we were monitoring with this near time process. This is just to give you an idea how many work zone that is we look at each year. Okay. The new thing, mobility, speed, realtime. We have a very large construction project in Dayton, Ohio, on I-75. It is immediately adjacent to the urban area of downtown Dayton, with with the potential to affect basically the whole region, so what we do is decided we wanted realtime data on mobility, went through the normal planning process and because of physical constraints we knew there would be issues, so we wanted to step up efforts to inform motorists. It happens that in two years from now we're going to build an ITS system in the region. What we did is pull certain elements, changeable message, radios and cameras and pulled them forward and deployed them in time for this project. To go that we're leasing data at $132,000 a year for the entire area, so by taking elements of future ITS system, putting them in the field, having realtime traffic data, we're really making a large effort to keep motorists informed of travel conditions flew this work zone at this very large work zone in the heart of Dayton. Just in the interest of time, on this map this is from the vendor's map. All the blue dots you see are individual sensors, and Jerry was talking about sensor spacing being critical. Generally we're a mile apart in the sensor spacing. You can see this the work zones here what we have done is strategically placed changeable message changes at alternate routes, higher advisory radio as you enter the region, and so now people can not only get realtime travel conditions through the work zone but they can also get it on all the alternate routes, and we think this is a cost effective way. We're getting the entire Dayton region for about $132,000 a year, something we're looking at expanding to other areas. The work zone and the information, the cameras and everything is monitored by a full time TMC we have down in Cincinnati traffic management center. That's it. I think I got us put pretty close to being back on time. Any questions? Tracy, take it away. Great. This is Darren. We're going to rundown the queue of questions. If you have any more, please feel free to type them into the chat box. Our first question was for Tracy. It was regarding the Missouri DOT example she provided on measuring traffic flow, and the question was what tools they actually use to do that. And the answer is that the measurement is more of a qualitative measurement based on what was expected and what occurred that's not so much a monitoring with sensors and the like, it is done more via a survey that Missouri DOT does use to gather that data, so it is one form of performance monitoring that can be done to give maybe a less perhaps more easily implementable way of getting some of this kind of information, so it is not an actual measured quantitative measurement from sensors and the like. All right. The next question, David talked quite a bit about some of the metric that is Ohio is using to measure safety performance, and just the question was what the accepted metrics were performing safety performance, so either one of you if you could just outline some of those. Do you want to take that, Dave? I can tell what you we do. As we went through on the slides, in some cases we're comparing prework zone to during work zone data. Other times we're comparing ourselves to other states. Other times we're comparing ourselves to general state wide safety data, so we look at it a number of ways, and each two years we just sort of dig into the data and start comparing it and look for problem areas, and one of the things we did in the very original one was we went back and looked at all work zone crashes all the interstate projects we had for a couple of years, and then we found problem areas and concentrations, and then we tunneled down further and logged individual crashes, and what we found was in place in work zones where we critically under shot how much capacity was needed, it obviously created large traffic jams and invariably the resulted was a massive number of crashes. It was something you just couldn't miss. Apparently we did. When we went back to the project engineers two years later and said were there a lot of crashes here, he said we had two, three crashes every single day, nobody did anything about it, so what we did is went in, figured out what went wrong, and we changed a number of policies. We created geometric work standards for work zones which was a tremendous problem, the ramps in our work zones were being slammed in as T enter sections, and now we have geometric standards requiring they be designed and take into consideration decision sight distance, minimum lane widths and shoulder widths, there was a number of lessons we learned by going back and looking at date A I don't know that there is one-way that we look at T it is actually the opposite. We Troy look at it from as many different points of view as we can. This is Jerry. I think that's a good point that most times I will go back to the term suite of measures. I didn't talk about it here, but we recently completed an NCHRP project looking at the intent offs look at nighttime work zones and whether from a traveler standpoint which ones are safer. We looked at rates as Dave was talking about on a per vehicle mile travel basis. Something else we went through of crash costs to get at the fact that we expect between daytime and nighttime conditions we have different operating speeds and if you have maybe fewer crashes but higher impact speeds you would expect costs to be higher for fewer crashes and so that was a way to get at normalizing again the differences in severity and then we also took a look at the vehicle mile traveled sort of represents an individual motorist's risk if you will. The other perspective though is that have you to get the work done, and by switching between day and nighttime, have you a much maybe a higher individual risk at night due to the construction project because you're working at night, but when you convert it back to take crash risks, crash costs, per unit time to get the work done, you can get much different answers, so I hardly or agree with Dave in needing multiple measures to look at safety. I would tag onto that one of the I guess most commonly used measures over the time in work zones has been work zone fatalities and sometimes that's been the primary or only measure used, and one difficulty with that is obviously there is a lot of crashes have come up already in the presentations, a lot of crashes that occur that don't result in fatalities, and if we just look at fatalities we may have special small sample that we aren't able to learn a lot from it, so I think as Dave showed looking at the crashes themselves and when they occur and then digging into it a little deep to her find out the reasons is maybe a better in some ways a better measure than looking at the the more commonly reported fatality numbers. All right. Our next question was regarding the definition of what constitutes a queue and as an example provided is that when say vehicle travel speeds below 10 miles per hour or is there any one accepted definition? We define it in our policy, in our work zone policy as being ten miles an hour or lessor characterized by stop and go conditions, and it is a little open, but that's the way we chose to do it back in 2001. This is Jerry. I think you will find that different agencies, it goes back to what the preferences are. I know of some locations that 30 or 35 miles an hour, anything below that because it represents a significant drop from 55, 65 miles an hour operating speeds, is used as an indication of queue as well, so it really does depend. Okay. Jerry, somebody asked what exactly is the buffer index? I was going to get in a little bit more detail about that and just decided not to. The buffer index is the difference between the average time it would take you to make a trip at a given period of time of day to the travel time you would have to plan for in order to avoid being late if you -- if something happened on the facility, and initially it started as, well I don't want to be late more than once a month because my boss will get mad at me if I am late more than once a month, that corresponds to basically a 95th percentile travel time that might say being during the morning peak rush or it could be in the p.m. as well, but that ratio of what it takes to make sure you're not late, have you to plan for it so that you're not chronically late, the average, the buffer index, the greater it is the more unreliable you could say your trip is ongoing to and from on a particular trip. Next question is a pretty important one. How can we better communicate the actual travel delay expected as motorists to basically increase the satisfaction, reduce anger about the actual travel delay? Dave, I want to know if you want to take the first stab at that being in the DOT there as to how you look at that issue. A couple of ways we have done it. We have used portable IT S systems. There is several of them on the market. One said basically X number of miles to the end of the work zone, travel time to the end of work zone, and we actually had a re research project to see if that caused any effect on the way people behaved T turned out it actually did. When we were able to communicate have you to go X miles, but it is going to take a long time to get there, there were quite a fairly large percentage of people diverted and the reason we knew this we found and did license plate studies to get permission from the bureau of motor vehicles to send letters, it was kind of an expansive effort just to get the data but we were able to show at least in satisfaction that these kinds of ITS systems do a fair job of letting people know because we're not only telling them their travel time but also their travel time to a specific distance, and most people in their heads can say that's unreasonable, something has gone wrong N the case of the Dayton, it is not really a portable ITS system. We're giving travel times through the region. We have a large over head changeable message signs, highway advisory radios, and we're using the speed data. We're providing travel time to this road, X miles, X minutes, so that's how we're doing T we're giving them trying to give as specific information on what lies ahead for them. Are you doing that on certain work zones, certain types of facilities doing that on all work zones? No. I didn't get into any of the work zone planning processes, but we plan our work zones fairly extensively way up front in the project development process, and we typically reserve these efforts for ones where we're just anticipating a problem. We try to design them so we don't have massive tie ups, but sometimes there are just physical or cost constraints where we know it is going to get ugly, and those are the situations we typically do this kind of over and beyond the normal notification. One thing I would mention is in my experience what I have seen more is except in certain work zones but it seems like more often in delay is provided it tends tore more qualitatively, just expect major clays, minor delays, or just expect delays, and it might be on a message board a few weeks before or it might object a website for the project. The DOT might have a general website for all of their work going on that summer and include some of these qualitative estimates of delay. It still seems like there is not extensive use of quantitative estimates, and part of that is probably gets us the subject of this webinar that use of the data and performance monitoring which helps feed into having more of that information to provide to motorists, so I would agree it could help people choose an alternate route like Dave mentioned or might help them be less frustrated or plan ahead better, but allow themselves more time if they know what to expect. I think it is a combination of data and a combination of how best to communicate it to the motorists so they might know before it is too late. That's not giving data -- I guess it is a form of data, but a lot of our major work zones will install web cams. They're not particularly difficult to install. They're fairly portable, and they're quite inexpensive actually and will create a project specific web page they can look at the web cam. It is simple to do, cost effective, and people seem to love them. Before they start their trip we get an enormous number of hits on the work zone web cam, so it is another way you're not telling them but they can make their own inferences. That's a good point. Great. This question was for Dave N regards to your Dayton project you mentioned that leasing data, and somebody asks where exactly you're leasing that data from? We went out on a competitive bid, so the winning company was a company called Speed Info. They have got a number of sites around the country they do, but their bid came in basically drags 110 per site, so however many sites. We had preselected based on our travel time we had planned all along the travel time so we had an idea of how many sensors and sex or densities -- sensor densities and a number of points and put it out at generic to technology. The next low lowest bid was over a million dollars, so it was pretty clear cut. We've had an unsolicited proposal from a vendor of cell phone data cell phone tracking to give us all their information statewide for and I don't know how this is possible, but for 35,000 dollars a year, so we're looking into that. We might be able to get data on a much larger regional if not statewide basis for 35,000 a year. Okay. The next question was I think this is a clarification question with regards to our earlier discussion about safety metrics. The questioner was interested in safety metrics during the operation, and not just pre-and post. Anything to say about that? Go ahead, Dave. I guess we don't really -- during the work zone basically all we're comparing to is historical what would be expected out there. We go back three years of data, and I know that's still kind of a pre-and comparison, but that's pretty much the only way we look at it. We don't provide like a target crash rate for our work zones. What we're looking at is just how many crashes would we expect to be in that same section of road if the work zone wasn't there? I think that there are also like looking at the crashes is getting at an outcome type of measure where this is the final outcome is a crash has occurred, what can we do about it. There are other surrogate measures sometimes used which may be in some cases more perhaps more readily available rather than trying to get crash reports in some states, and things like looking at speeding or speed differential information, if you have some sensors out there you might be able to look at that information and it is again a surrogate. It doesn't necessarily result in crashes, but it can give you some indication that there might be maybe there is an issue that might benefit particularly from enforcement in that area. I know with some things we looked at, for instance, we have done some evaluations of dynamic lane merge systems which is a form of ITS, and some of the measures we looked at for that were by looking at camera data and analyzing it, things like near misses and aggressive maneuvers, things along those lines, so some other types of measures that could be used in the during the work zone to try and address some initial issues as well as just sometimes somewhat random things where the TMC might be looking at a camera in an area and see people having -- I know in New Mexico it happened on one project where they could see that vehicles were having trouble maneuvering a certain area and had to make adjustments, so that's not so much a formal measure, but there can be ways of observing those things and in kind of just individual situations like that, so I think there are other ones, just maybe a little more indicators rather than kind of the final outcome measure. Okay. Next question. Questioner asks does portability, i.e. repositioning sensors help improve the work zone data? Yes. All right. I don't think this is rocket science, but what we found is you need a higher density of sensors in the area of merges and transition areas, things that really after reflect the flow of traffic. Once you get into the main body of the work zone you don't need anywhere near as dense as sensor place at some time get accurate data. You have to measure very closely is what we found in the merge and the queue area if you will where we're affecting the speeds most. We've had to change sensor locations because we were getting faulty travel times because with you don't have sensors basically where we were impacting the cars the most. Jerry, do you want to comment on that? I know you found different queue estimates from the different data sources. I think Dave is right on the mark there that if you don't have -- the sensors adequate density where the impacts are occurring, you're just happen to assume conditions, you know, larger and larger regions for which they may or may not be accurate, I would also comment that this could play -- what it may mean is if your activity itself is what's causing the impacts and you have a large length project of which you're moving up and down periodically doing the work, yeah, you may need to have your where your sensors are relative to the work shift itself would be an important consideration moving them around or at least having the sensors and dense enough level that regardless of where you move your congestion point if you will that you're getting data is an important -- will defend the work that would define how well how accurate your data will be. Sure. Next question was with regards to crash rates and work zones, does anyone look at the crashes that would normally be present in this segment of road during the time construction and perhaps by time of day perhaps, during the time construction is occurring if the construction was not there? Dave, I know you were doing some analysis. That's exactly one of the ways we looked at it, calculating the crash rate, prior to the work zone, and then the crash rate during the life of the work zone, and just made a direct comparison. I think it is slightly different in this question, but I think Michigan is doing a little bit of this now as they're developing their transportation management plans, TMPs, there is a small -- a section in that that looks at crash -- does this crash analysis and I think one of the means is there is some type of comparisons in there just to try and kind of get some sense of whether this would be a particular issue in a given upcoming work zone, so I think there is a little bit more of that occurring starting to occur as the plans are developing. All right. That's it for our chat box questions. Do we have anybody queued on the voice line for questions? At this time I am showing no questions, but at any time they may press star and 1 on their touch tone phone. They will need to record their names so they will be introduced. Once again, please press star 1. All right F we don't have any further questions. We do have a question. One moment. Great. Our question comes from Bill Robins. Your line is open. It is Bill well man. I do have a question about the new Intel drive initiative, and the vehicle to vehicle applications as they go forward to work zones. I have seen a lot of discussion about vehicle to vehicle and vehicle to infrastructure, and it seems like the infrastructure back to the vehicle is being left out and that's an area. Do we have any comments on that from the panel members? Sounds like a federal question to me, Tracy. I am involved a little bit with two of the -- well, one effort that's going on in our ITS office called safe trip 21 which is looking at some different uses of ITS and there is two test beds there an effort to look at using relation -- some of that communication effort with in work zone settings as part of the test bed, but I am not familiar with specifically what's being done from infrastructure to vehicle efforts overall in the larger Intel drive program. That would be something in our ITS office I could get you a point of contact if you wanted to ask them that question. Thank you, Tracy. Okay. How many states presently are purchasing sensor probe data from private sources not including contractor or consultant efforts? Not studies, things other than studies I guess is what the question is getting at? I think so. You know, I don't have a calendar or tracking of how many states are doing that, and then how many are doing that specifically in any type of work zones. I guess I can't give you an estimate on that. I imagine more of what I heard about is purchasing the data perhaps in permanent situations and not necessarily work zone situations, but I don't know if that is prevelant at this point in time. Michigan, Jason from Michigan DOT commented that Michigan is doing that on the chat pod there. Okay. Do we have any more voice questions queued up? At this time I am showing no audio questions. All right. We have another comment that came in that vendor called enrich provides data to eleven states. I am sure all of the vendors would be happy to tell you guys how many states they're providing data for, Speed Info is the vendor Dave mentioned. We're showing just a few minutes to go, so, Jocelyn(y) don't we wrap it up? Sounds good. Thanks, Darren. to wrap up webcast I will give you information on the national transportation operations coalition or NTOC. You will see the member organizations of NTOC. We encourage to you go to the website, www.NTOC. com to find out more about these organizations. On this slide you will see that the NTOC website contains information will upcoming webcasts. The site also contains a webcast archive page and slides and recordings of the previous operations webcast. We'll have the slides and recording of today's presentation up within about a week. N TelePresenceOC also -- NTOC also has two discussions forums, one on high level or strategic issues and the other on ITS deployment and lessons learned. You can also sign up on the website for the NTOC newsletter by email. It is sent out twice monthly and contains ( indiscernible ) so that concludes our webcast today. I will hand it over to Darren to sign off. Great. Thank you all for coming. We have another webinar Jocelyn mentioned this coming week regarding value pricing, and we hope you tune in for the whole series. Thank you very much. Thank you. Thank you. That does conclude today's conference call. Thank you for joining, and you may disconnect at this time. (end)