Office of Planning, Environment, & Realty (HEP)
Good day, welcome to Freight Analysis Framework 2: Findings and Products. I'll be your audio coordinator, all participants are in a listen own only mode. There will be a question and answer session at the end of the presentation. You may submit questions throughout the presentation by typing them into the chat area on your screen. I would now like to turn the call over to your host for today, Jennifer Seplow.
Good afternoon or good morning to those of you to the West. Welcome to the Talking Freight Seminar Series. My name is Jennifer Seplow and I will moderate today's seminar. Today's topic is Freight Analysis Framework 2: Findings and Products. Please be advised that today's seminar is being recorded.
Today we'll have three speakers - Tianjia Tang of the Federal Highway Administration Office of Freight Management and Operations, Randy Curlee of the Center for Transportation Analysis at the Oak Ridge National Laboratory, and Felix Ammah-Tagoe of MacroSys Research and Technology.
Dr. Tianjia Tang joined the Federal Highway Administration's Freight office from FHWA's Resource Center in April 2005. In the freight office, his main responsibility is to run the Freight Analysis Framework and the Freight Modeling Improvement Program.
At the resource center, Dr. Tang served in the position of transportation air quality specialist position. His main focuses were urban travel demand modeling and motor vehicle emission modeling in conjunction with transportation conformity issues.
Dr. Tang had worked in the transportation industry in a wide range of positions including senior environmental specialist, project engineer, project manager, and program manager. He is a registered professional engineer in Georgia.
Dr. T. Randall (Randy) Curlee is a Distinguished Research Staff Member at Oak Ridge National Laboratory in Knoxville, Tennessee and an Adjunct Professor of Economics at the University of Tennessee. He received his Ph.D. in Economics from Purdue University. His research has focused on freight transportation, public transit, energy, and waste management. Working with the U.S. Army Corps of Engineers, Dr. Curlee led the development of the Ohio River Navigation Investment Model (ORNIM), which the Corps uses to assess optimal investments in new and improved inland navigation infrastructure. Dr. Curlee also led ORNL's support of the Federal Highway Administration's Freight Analysis Framework (FAF) and the Freight Model Improvement Program (FMIP), which promotes new Freight Analysis Frameworks and models for all freight modes. Dr. Curlee's presentation today provides an overview of the methodology used in the 2002 FAF.
Dr. Felix Ammah-Tagoe has several years of research and analytical professional experience, particularly in transportation, freight, logistics, and transportation energy use. He currently is a Senior Research Consultant and Project Manager at MacroSys Research for the Bureau of Transportation Statistics, serving as the primary analyst on the Commodity Flow Survey and a variety of other freight projects. Dr. Ammah-Tagoe researches the role of freight transportation in the nation's economy, freight transportation and inter-state trade, and the relative roles of domestic transportation of international trade.
Dr. Ammah-Tagoe has performed a variety of analytical research and produced data reports on freight movements on the U.S. transportation system, its relationship with the state and regional economies, and its impacts on freight corridors and gateways. He has designed Commodity Flow Survey reports and worked with the Census Bureau to produce all the major Commodity Flow Survey freight data publications. He designed and produced state freight summaries from the Commodity Flow Survey for the Bureau of Transportation Statistics.
Dr. Ammah-Tagoe earned a doctorate degree in geography and a master's degree in energy and resource management, both from Boston University.
I'd like to go over a few logistical details prior to starting the seminar. Today's seminar will last 90 minutes, with 60 minutes allocated for the speakers, and the final 30 minutes for audience Question and Answer. The Operator will give you instructions on how to ask a question over the phone during the Q&A period. However, if during the presentations you think of a question, you can type it into the smaller text box underneath the chat area on the lower right side of your screen. Please make sure you are typing in the thin text box and not the large white area. Presenters will be unable to answer your questions during their presentations, but I will use some of the questions typed into the chat box to start off the question and answer session in the last half hour of the seminar. Those questions that are not answered will be posted to the Freight Planning LISTSERV. The LISTSERV is an email list and is a great forum for the distribution of information and a place where you can post questions to find out what other subscribers have learned in the area of Freight Planning. If you have not already joined the LISTSERV, the web address at which you can register is provided on the slide on your screen.
Finally, I would like to remind you that this session is being recorded. A file containing the audio and the visual portion of this seminar will be posted to the Talking Freight Web site within the next week. Due to the size of the file, recorded files are available for viewing/listening purposes only and cannot be saved to your own computer. We encourage you to direct others in your office that may have not been able to attend this seminar to access the recorded seminar.
The PowerPoint presentations used during the seminar will also be available within the next week. I will notify all attendees of the availability of the PowerPoints, the recording, and a transcript of this seminar.
Our first presentation of the day will be that of Tianjia Tang of the Federal Highway Administration Office of Freight Management and Operations.
If you think of questions during this presentation or during any of the other presentations, please type them into the chat area on the screen. Questions will be answered in the last 30 minutes of the seminar
Thank you. Good afternoon, or good morning, I'm Tianjia Tang, and I'm a Program Manager here in the Freight Office. I am going to start with an overview of the second generation of the Freight Analysis Framework. Following my presentation, Dr. Randy Curlee will go over the methodology for the 2002 base case commodity OD flow database. Dr. Felix Ammah-Tagoe from MacroSys Research and Technology under contract with BTS will cover the out-of-scope component of the 2002 OD database. The out-of scope piece is freight shipment not covered by 2002 commodity flow survey.
I am going to start with the outline of my presentation. Starting with approaches we are taking, deliverables, structures, and components of the FAF2, and last, Q and A, questions and answers.
Now, let's take a look at the objectives. That's to answer freight shipment volume and congestion questions and also to answer questions on the freight shipment motor issues, and all other usage which we define as -- the key here to notice is that FAF is to provide a big picture of the freight moment. Specifically, this national picture is to enable the Federal Highway Administration to do policy and investment analysis.
However, this is not to say that we are not interested in helping state and local MPOs to utilize teh FAF2 data. I just came back form the State of Florida where the FDOT is going to do a pilot project on how to use teh FAF2 data.
I am still looking for another state to be my pilot project, if you are interested in this pilot effort, please nominate yourself or a friend, send me e-mail or call me and we will proceed from there.
Now, let's look at the FAF2 approach. With regard to data sources, we are using public sources, all are based on publicly available data. With regards to methodology, all methods are transparent and will be disclosed to you, at both the Federal level and state local level. With regard to the deliverables, both input and output are publicly available. So what we have is what you will be able to see.
With regard to the products, the FAF 2 products and deliverables can be separated into two groups, the first group is commodity related products, the second group is network related products.
Commodity related products have 3 key components: The first one is the current 2002 base case origin destination, known is as OD database. For future years from 2010 to 2035, with a five year interval. The third components is annual provision estimates.
Start with the bottom left, on left, right here. We have 2002 basic commodity OD. When you have commodity OD, make sure you are aware of the units we are using. Commodity OD are tonnage and dollar, so that's really key. I want you to remember that. ORNL under the leadership of Dr. randy Curllee has performed this task with assistance from MacroSys Research and technology. Felix Ammah-Tagoe will cover the Macrosys's effort later here. From the 2002 database, we are working on the future commodity output data, future data, -- 2010 and 2035 with a five year interval. The format of the future OD database will be similar with the 2002 base case. To get yourself familiar with the 2002 database, you will know the format and content of the future of the database. We are in the process of evaluating proposals to get a contractor on board to help us in this endeavor. By the end of next week, I hope we will be able to - to do that.
Let's move to the truck OD, -- unit of conversion. As I said earlier, when we talk about commodity OD, the commodity OD has the unit of tonnage and dollar amount, we are not talking about number of vehicles, trucks. Here we need to convert the tonnage to truck payload equivalents.
With regard to faf highway network, we are in good shape, 95% there, so I'm so excited. It's very good. We are in good shape. Once we have both the network and the truck numbers, we will move to assignment of process, all analysis will be conducted and generated here.
Now we have covered all the key components, I'm going to shift gear to geography on the commodity details. There is a total of 138 geographic regions, 114 is from CFS. Diagrams shows the regions. On this map, the green color represents the 64 metropolitan areas, other 50 are either a complete state or part of a state: This diagram here shows you the 13 additional international Gateways. There are over 400 Gateways throughout the country. All major ones are covered in the CFS metropolitan areas, or by the 17 additional international Gateways, as we have them listed here, those cover border crossings, seaports, and one airport, in Anchorage, Alaska. Now we divide the world into 7 regions. Canada, Mexico, South America, Asia, Europe, Middle East, the ones not covered by those 6, they are covered with the remainder of the world. We have several modes here, trucks, rail, water, air, -- others.
We made assumption that for a given commodity, the modal split will stay the same unless we have other information to indicate that it is incorrect, so by default, commodity split will be the same. This is my crown jewel, our 2002 base case database, if you did not attend TRB last month, you can still get the opportunity -- by visiting our website to get teh data. You can download the data at www.ops.fhwa.dot.gov/freight, just make sure you read the User's Guide. From there, look for the Freight Analysis Framework, where you'll be able to find all the data. We have five files here. The first one is a Microsoft Excel file. This file has 3 tables, -- DOM stand for domestic. Origin/destination is in the United States. Second table here is underscore BRD, transporter. U.S., Mexico and Canada. Last one is underscore SEA, which is ocean. This is our import/export with the rest of the world, and other countries, other regions. As with the domestic table, all data are presented in 6 columns, origin, destination, commodity, mode, Mdol (million dollar), Kton is thousand tons, units of it. This one is the international trade file format. We have 7 columns here. Origin, Destination, commodity, mode, port -- port is referring to the point of -- for export of point of entry for import. Then we will have the million dollar and thousand tons.
The status of the FAF 2 project, as I said, we just released 2002 base case commodity OD during TRB last month, 2005 annual projections and estimates will be released later this year, and the rest of the products, flow database, projections will be released by the end of September of this year. Before I yield to my next speaker, I would like to remind you to nominate yourself or a friend for my pilot project. Please feel free to contact me at any time. Thank you very much.
Thank you, Tianjia. We are going to move on to our next presentation, which is that of Randy Curlee of the Center for Transportation Analysis at the Oak Ridge National Laboratory. Thank you to those who are posted information in the chat area, if you do think of additional questions for Tianjia, or if you think of questions for Randy feel free to post them. Randy, you can go ahead.
Good afternoon. I appreciate the opportunity to talk to all of you out there, and tell you about the methodology that we developed working with MacroSys and FHWA. In putting together the Freight Analysis Framework OD matrix, let me stipulate up front that this methodology discussion I'll go through today has to do with the origin designation flows. Tianjia mentioned the other products for the FAF, and I won't talk about the methodologies for those but will be talking about the methodology we developed for the OD flows. I would be remiss if I didn't mention the team members that put this together. Dr. Frank Southworth, Bruce Peterson, Drs. Bush, Chin, Wong, and Ms. Perry. We had an excellent team that worked together to come up with a methodology to meet the needs, ambitious needs of the 2002 FAF. We will see this flow diagram throughout my presentation. I'll refer to it again and again; it is a simplified flow diagram of the methodology that was used to develop the FAF 2 flows table. I'll be giving you a high level summary of the data we used to derive the flows. Let me mention that the end point is at the bottom right, final FAF 2 flows table, and these various other boxes are intermediate points, and the ovals represent methodological steps we took to reach our objective.
Let's start with the final FAF 2 flows table. This is somewhat duplicative of what Tianjia said, but it's better to say it twice if you want to make a point. The dimensions are the 114 flow areas, the 17 international Gateways, and 7 international regions. As he said, we use the two digit SCTG , refined the modes used in the CFS down from 11 down to 7; and what we have then is a very, very large four dimensional set of matrices; one for dollars and one for tons. And this is approximately 12 million cells. So when you get ready to download it on your computer, hopefully you have a high speed connection. This is just a repeat of the FAF regions that Tianjia mentioned. And the FAF 2 commodities, the 43 categories within there are some areas where it gets to fairly large. Alcoholic beverages and others are fairly small Others include some several large quantity commodities. This is a repeat of the modes - truck, rail, water, air, air and truck, truck and rail, other intermodal, and pipeline and unknown.
Let's return now to the methodology that we use, and we are trying to remember to arrive to the bottom line. Where do we start? The starting point is with the commodity flow survey table 17, which is at the top middle. Now, what is this? Tianjia said that we are very committed to keeping the FAF 2 dependent upon publicly available data; so when we began this process, we asked the question, what information is available to the public that can be reproduced, and the closest thing that comes to the objectives of FAF 2 is the commodity flow survey, table 17. This is a four dimensional matrix at the state level. It does have the required detail, it's at the state level, not at the level of geography we need, but we can accomplish that and I'll talk a little bit about that later. This is a very large and very sparse table. And the question is why is it so sparse? Well, it's only about 5% filled. The reasons are -- there are 3 main reasons; first of all, the Census does not report cells that have a coefficient of variation of greater than 50%, and there are disclosure rules they must follow, and in some cases, there are no observations for the cells at this very, very disaggregated level.
The challenge to all of us in developing FAF 2 was to take this CFS table 17 and apply statistical approaches and other data, and whatever we can come up with that's defensible statistically, scientifically, that we can build on to develop a publicly available FAF based on publicly available information. The first step we took was to ask Census to identify those cells in CFS table 17 for which they had no observations in their survey, and they were kind enough to provide this information to us. We identify those as true 0 cells from Census. Census identified these missing cells and by assumption, we concluded that those cells for which there were no observations we would label as true 0's and constrain them to be so. So that cuts of size of our problem down significantly. But then we have the remaining cells that are suppressed for one of these other reasons, and thus must be estimated as part of this process.
How do we go about estimating those? The next step that we take is a statistical step, a log linear model. Some of you are familiar with this, and I won't go into the details of that, but it's fairly straight forward in terms of what we did. If you don't have a cell at the four dimensional level in CFS table 17, then what you do have, is you have information at other levels, whether at the 3-D, 2 dimensional, 1 dimensional level, and you can use statistical techniques, like log linear modeling, to define relationships between the various variables that come into play. And what we have in fact is fully saturated, if you will, log linear model that in the case of where we have data available, it can have up to 127 terms.
We have a distance bucket in the equation which we drew upon because of the problems primarily we have with truck. Truck data are very difficult to come by, and we all know of all the modes, that's probably the mode where we have the least data to work with; so we use the distance buckets, and we related all of these parameters through one, two, three dimensional relationships to identify, if we don't have a number for that particular cell, then we know information about generally what happens with a particular region or commodity, and you can use the statistical relationships to estimate values for missing cells. Now, if we have alternative data that we also use, this is extremely important; in the case of rail and water, we relied on the railcar waybill data, and we relied on the waterborne commerce data. For those cases where we had cells in table 17 for which we could not estimate relationships, we, first of all, converted the railcar waybills and waterborne commerce from the commodity classification schemes they are presented in over to SCTG -- this was not a trivial task, and there's some degree of error that goes along with that. The next step is we have an issue that arises which causes the need for proportional fitting. Why do we have that? After you estimate the cells in CFS table 17, there's no reason or constraint to force them to add up to what the higher level totals are in the CFS table. So proportional fitting is a straightforward process to force the margins, if you will, in table 17 and for that matter, tables at higher levels of aggregation to be consistent with those totals. Computationally, with a number of matrixes as large as this, it is not trivial. We are talking about computational times in hours, not minutes. We did constrain the marginals to be consistent with the totals, and constrained those cells within table 17 that had values to remain unchanged. If they are at the level needed in terms of geography, they did not change. They have been maintained. That then brings us to an enhanced CFS flows matrix. Thus, at this point we have taken the CFS very sparse matrix, we've used the 3 approaches identifying true 0's, using the log linear model approach, and iterative proportional fitting to come up with an enhanced fully filled matrix at the 4 dimensional level. But this doesn't include flows that are out of scope to the CFS or flows that are generally thought to be under counted. And Felix will talk about this more in great detail, but there are several omissions that do exist, and this is something that we decided early on in this process, that this FAF would address those CFS out-of-scope flows and do it the best we can given the resources available.
So we came to the work for the CFS out of scope sectors. we did this in collaboration with MacroSys. There were 15 studies. Let me mention some of the main ones we did. Some things are out of scope to the CFS, for example it doesn't include crude petroleum, doesn't include natural gas, doesn't include municipal solid waste. We had to go out and identify the data that are available in these areas, pull them together, put them in the right format for FAF needs. All of these required independent and special studies.
There are some areas such as, farm based movement and fisheries, which are partially missing. Let me give you an example. For farm based movements, CFS does not capture the freight movements from the farm itself to, for example, the grain elevator. Those are short trips in terms of ton miles, but large in number of tons. Same thing applies to fisheries. We don't capture from the boat to the first point of processing so we did special studies for those. So each of these categories of special studies represents one aspect of one of those 43 SCTG groups; so by doing the studies, filling in aspects of those commodities for which we did not have information from the CFS, we then had the task of taking these national totals and aggregating those down to the geography needs of FAF 2. This was in some cases challenging, and we developed a set of methodologies that had to be dependent on the data we have, the resources we have to do this with, and it had to be defensible from all perspectives. In some cases, we disaggregated the national numbers down to the required FAF regions based upon economic activity in those sectors. In other cases, such as municipal solid waste, we collected information at the FAF regional level. So it was a detailed study that has information that we collected on movements from a city to either nearby disposable facilities or possibly long transport by rail or truck or by barge. So there were a variety of methods we developed; but what we did was to arrive at a regional disaggregation for all of these out of scopes.
Well, that just about gets us to where we need to be for the FAF. Coming back to our main diagram here... We then can straightforwardly add the two matrixes, this at the required flow dimensions. And the CFS out of scope sectors, which we have started out with national totals and through the disaggregation processes we've taken them down to the 4 dimensional level required by FAF. And it's a simple addition, if you call 12 million cells adding to 12 million cells, it's simple. So that allows us to arrive at the final FAF 2 flows -- OD flows for the FAF 2 objectives. So with that, I'll await your questions and thank you very much.
Thank you Randy. Our final presentation of the day will be that of Felix Ammah-Tagoe of MacroSys Research and Technology.
This is Felix Ammah-Tagoe, a consultant with MacroSys. I manage this project from the BTS side. I want to thank you for the invitation to participate in this seminar. I want to thank BTS, the sponsor of this project.
What I would like to do this afternoon is to follow up with what Randy just presented and what Tianjia presented, and give you a feeling for some of the things that we did in terms of the coverage, method we used in getting the data for the out of scope sectors. Seeing the names of the participants, a number of you are aware of the CFS, so I won't bore you too much with that, but if you think of it, the primary factors that are covered are manufacturing, wholesale, and mining. In all the CFS, 93 to 97, and 2002, we tried to cover the retail sector but for different reasons, only a very small part of retail was covered, so in order to arrive at the totality of freight that moved from the national freight transportation system, we had to go outside the CFS and try to catch up the sectors that are not covered.
In the next box, it shows some sectors, now, this is a cross section, some are sectors and some are commodities that what we try to do was look at imports, construction, fisheries, municipal solid waste, and others that are listed. There are afew sectors that are partially covered, and I'll explain what I mean by that. Exports - it's collecting information on outbound shipments, it covers exports, but we know from comparing the CFS data with favorable trade data that CFS doesn't cover all the exports, so there's the need for us to bring in export data from multiple places and we will talk about that in that minute. Household and office goods moves and in-transit, that was also out of scope that we tried to cover, so these sectors and groups together forms the basis for the national estimate that were generated and that became the basis for -- segregated into the difference regions.
I'm going to go into the 8 or so sectors that MacroSys handled and give you a sense of it again without going into details, this is part of the transparency that Tianjia talked about, to give you a sense of what are the public data sources that we used, and what are some steps we used in compiling the data. There are a wide variety of papers for these sectors that will be available online that you can download and look at details. I'm going to starts off with retail. As you can imagine, there are countless retail outlets in the United States, and the survey could not sample them and collect data from those sectors in parts because sometimes from the retail outlet, you business to business retail transactions or you have business to customer retail transactions where I could go to Home Depot or Lowe's, pick up some items, and take it home myself, that is not captured in the CFS, you can order things from Staples, or Office Max, and deliver it to them, that's a retail business to business transaction. That's not captured in the CFS, the goal was to look at the retail sector overall, look at the shipments from that, and see how we can then incorporate a portion that is freight related.
Several data sources we used, we used some data from the input/output models basically we took the value of retail sales trade and took out the profit margin, sales margin to bring the value closer to what the produced purchase price would be which is what you would find. I see a question from one of the participants, how come we don't have ton miles, and we can talk about that question later on. For the construction sector, again, as you can imagine, there are shipments that are related to the construction, and that whole sector is not covered in the CFS as, in part, because the businesses that are classified as construction. If you look in the North American industry classification system, some of them are really construction companies that are involved in handling the architecture part of the construction, and not necessarily handling the freight side, or anything that involves the movement of goods per se, but they are part of the construction sector. Transporting materials from one side in handling construction debris and such, and for this sector, we also went to the -- survey to get the ton miles -- for trucks that are engaged in the construction sector, so from the miles information and tonnage information, we can work our way back to tons and value, and apply ratios.
For services, the source of the data were similar and methods we used were fairly similar to what we did, like from the ton miles to ton, and to the value. With both services and construction, one of the limitations was we were unable to find information on goods handled by the service sector. It is limited to only truck and Federal research, further research is needed to branch out of the truck and look at rail-related shipments.
One of the other sectors that is out of scope in 2002 was the logging sector. The issue here is slightly different because logging was covered in the '97, and '93 CFS, unfortunately, in 2002, logging was not covered because logging as a sector was moved from manufacturing into agriculture, so that moved out of the CFS and became out of scope -- it's a classification that was used. So the goal was to try and get that sector and to complete the data set to do that, we went to the USDA agricultural statistics data -- railway building and Army Corps of Engineers. Having identified the sector, we went to the economic -- to get information on the value and sales receipts from that logging sector. We looked at 97 trade data, and '93, grew it based on the average group rate, -- '97, and back down -- the value, using value to weight ratios again from the '97 CFS.
Publishing was one of the other sectors that got lost in the 2002 because of the transition from SIC. Again, publishing was moved from manufacturing into information and information as a sector is not covered. And you can imagine there are publishers that are involved in printing of the product, involved in printing books and such, and transporting the books from the printing presses to wholesalers or to retailers, so that part of it is not covered, and that's what we are trying to cover here. And essentially, working from the economics sense, we were able to get the value of it using value to ton ratios.
Imports is different from the other ones, and that's because imports is not covered at all because the CFS only samples domestic U.S. business, so anything coming into the ports is not covered. So to get this, we have to go to the official trade data from the U.S. Census Bureau. It's not a straightforward thing because we are doing this as you can imagine at a modal level, and detailed commodity level, and in part, because the CFS is in the standard classification of transported goods, we have to convert it. But essentially, that brought us to where we could get a value that -- to get a ton mile, we have to apply average levels of, from the waybill that is more specific. There is still some remaining issues with imports in part because it's not too early to ask how much imports may have been covered already in the CFS. When an export is transported it is relatively straightforward, but on the freight side, on the surface mode, there is no official data for the tonnage of U.S. export to Canada and to Mexico, we simply don't collect that data. So we have to find appropriate values to convert detailed commodity and mode level and roll it back up to the total subtotal for tonnage. This is a relatively small sector. I'm not going to spend a lots of time.
What does it all mean? After we put all of this data together, I want to briefly describe what the national estimate came to, why should we care, and what did we find out? When it was said and done, the total estimates that we got indicate that nearly over $19 billion tons of goods move on the U.S. transportation system and this was generated almost 4.4 trillion ton-miles. Now, this was significant because if you take this and compare it to the CFS, you could tell much more freight move on the transportation system than indicated by the previous estimates.
This slide shows you the modal breakdown by single mode, and multiple mode. One of the things you find out when you look at the estimates is that on the single mode side, the estimate for rail and water vary from where you would find from the Army Corps of Engineers, part of the reason is because those are single mode estimates on the bottom part where we have multiple mode, once you disaggregate and appropriately reassign components back to rail and back to water, the estimate in the far two are very comparable to what you find for rail and water from the Army Corps of Engineers.
One of the major findings, as you can expect, is that truck remains the mode of choice. Another interesting finding is that nearly 1.7 billion tons of the 19 billion tons is related to international trade, meaning that 9% of the total tonnage has to do with international merchandise trade. And of course, this 9% varies by state, for states like California, Texas, Michigan, and New York that it's a larger percent ranging from 15 to about 20%. One of the other things that we learn from this was that when you compare the composite national systems in -- with the CFS, nearly 36% of that freight is not covered in the CFS, that's by value, by tonnage, about 40%, it's not covered, and by ton-miles, about 29% is not covered in the CFS. And here is the table to show you that.
So just to wrap up, CFS was the main source for the FAF 2. There are reasons why when you do your state to state flows and for those interested in the sub state geography, when you look at those, it's important to use the FAF 2 data only because FAF 2 covers almost all the out of scope areas that are not covered in the CFS, and so when you look at the CFS data, and try to flow that on a network, you are not representing the totality of freight on the network, and hopefully, by bringing all these other things together, we are close to what that totality should be. Are we 2% off? Five percent? We don't know that yet, but we feel strongly that numbers are close to what totality of freight actually is. Thank you very much, and we'll take your questions.
Thank you, Felix. I hope everybody enjoyed the presentations. We are now going to start with the question and answer session with the questions that have been posted online. And if we get through those and have time, we can open the phone lines up for questions. I'm going to try to direct the questions to the appropriate presenter, but because all 3 kind of overlapped on the topic, if any one of you wants to jump in and provide an answer, that's fine. I'll start off with the first questions that were typed in and move on down the list. I think Randy might have addressed this, but I'll ask it again, what are the primary sources of public data being used to develop the FAF-2? Randy, I don't know if you are the appropriate person or somebody else?
Well, the primary sources for the CFS side of it are the table 17 and the waybill data and the waterborne commerce data. And then Felix mentioned the export/import data, and I think he covered that -- out of scope is maybe -- and that really depends by the commodity. So there will be documentation available on that, but there are numerous, numerous sources of data for those out of scopes. You have to really look down into the details of that. But the main sources are CFS, the freight, excuse me, waybill data, and the Corps of Engineers data are the 3 main sources with numerous others, and I mean 20 other data sources that come in to, or even more. I'd be happy to go into anymore specifics if we can be more specific with the question.
I want to add one more thing. The bottom line here is the 2002 CFS data is the backbone of the 2002 FAF , it's really just a combination effort.
Let me adjust one other thing to follow up on Tianjia's point. Recall that while we used other data sources for the CFS components of this to identify missing cells, we constrained the totals to be consistent with CFS. Thus, if we had a total from waterborne commercial or waybill for a particular OD movement that was different than what the total that the aggregated level was for CFS, we went with the CFS total.
The next question is, for the 1997 commodity flow survey, summary reports were prepared for each state. Will state or regional summaries be prepared for the 2002 FAF 2? These would be useful as a FAF 2 users try to organize the data for their own uses.
We are pretty much done with this. During last month's TRB meeting, we released most of the state profiles. We are still in the final stages to put all the state summary data on the web so, if your state is not on web yet, it will be on web soon. I got the raw data, everything, we got the raw data attachment now we just need to put it in the right format. But majority of the states are done.
Thank you. The next question is if all data is being collected, why can't we get a breakdown in ton miles in various modes of freight transport. Felix, I think you addressed that.
I think it goes to a product that Tianjia mentioned, before you can generate ton miles, particularly for the out of scope, which originally we don't have ton miles from the CFS, but if Tianjia wants to follow up on that.
We have OD data. Without assigning into the route, you are -- what effect on which route because it takes, I don't know which route it takes, hard to have a precise distance, so the ton, ton mile -- so it is not there. That's not complete either, we do have a route estimate of ton miles, you know, and the way I use it is a rough estimate in our computation. Database matrixes 2, but as relates to the final ton mile numbers, we feel comfortable, I think the bottom line was that we did calculate those, ton miles, but in evidently, we will have another set of ton miles once it goes to the network, I think the decision was made that having two sets of ton miles estimates might essentially be confusing and not add to general knowledge.
The next question is, for benchmarking purposes, is it anticipated that the FAF effort will be undertaken on a 5 year, recurring cycle? Tianjia? Do you want to address that?
CFS is really our backbone of the FAF 2 project. The next CFS is planned for 2007 so we may do a third generation of FAF then.
The next question is, are the special studies that were done to fill in out of scope sectors available to other DOT modes? Randy, would you be the one to address that?
We have special studies that are in various states of completion in terms of availability, so I'll direct that to Tianjia.
As I said, the beauty of the FAF 2 project is most input and output are publicly available, it doesn't matter if you are outside DOT, all what we have down here as related to FAF 2 will be available. The out of scope studies have been started and we will make sure that all those documents will be available to you.
This next question is for all of presenters. Can you please generally describe the changes in methodology and/or data that were made for FAF2? What lessons were learned from FAF1 that are now applied to FAF2? Tianjia, do you want to start?
As you know, I'm sure from the question, you probably know FAF 1 fairly good, the key difference here is FAF 1 is based on proprietary data. That's why folks outside of the DOT are not able to see anything smaller than a state summary. That is the proprietary nature. But the success of FAF 1 has been so overwhelming, that we took a different approach this time, we said you know, the key here is public availability. We also heard the need for smaller geographic data, county levels.
I think Tianjia covered most of the differences, the differences in the data, and the private data source that was used for the 97 FAF was based in part also on the CFS, but it was based on the 93 CFS, so although the data was coming out in '98, it was coming out of '93 CFS, and while it did try to cover some of the sectors, the approach was different in that, in this case, we making sure that all of individual sectors for crude petroleum is appropriately covered. Shipments are covered, and the other sectors that I went through, from that conception, the methods are different, taking different approaches to arrive at the totality of freight shipment.
I was saying, and just to follow up on that, that FAF 2 is not directly comparable to FAF 1, because of these, the addition of the out of scopes, which I think Tianjia referred to. But, yes, it is a different product, it is, the breadth of coverage is broader, I think the fact that all of these are public data, if you so desired, we would provide you all of the data such that you could recreate this and I hope, our hope is that by having that transparency with FAF, it will allow users to go down into the details in a way that will make it much more usable to a variety of FAF customers familiar with FAF 1.
Thank you. The next question is for Tianjia, is there a way for FHWA to alert us to the final production of each phase of FAF2, such as through the freight planning LISTSERV?
If you are on my mailing list, typically I will send an e-mail out, but the freight planning LISTSERV is also good place, we will do that.
Next question is the public use waybill data base shows BEA to BEA movements. How did you break it down to State-to-State movements?
The question is, I believe, how the BEA activities -- the level of disaggregation that's required for the out of scopes, which this refers to, depends on the methodology, and some indications if you take it down, you take your totals for the nation down to state, and to the, to whatever appropriate level of geography, and you add back up, so you may be taking disaggregating flows down to fairly low level of disaggregation, it varies from product mix, commodity to commodity, so I guess I'm not fully following the question, we did use those kinds of data to disaggregate to the regions, and add back up to the appropriate regions for FAF purposes.
In some cases, when we say data is public, in the case for example of identifying the true 0's in the CFS table 17, there were cases where we made special requests to obtain data that heretofore was not publicly available, in other words, you couldn't open a book, or couldn't go online to obtain it, but when we made the special requests, both of the Corps of Engineers and in the case of census, their approval process said yeah, that's okay, we can give you that information. So it is something, in some cases we relied on information in date that that you can't directly just pull off the web, but it is public, it's just not readily available.
The next question is, in FAF-1 demand for freight transportation was forecast to double by 2020. Is this still the case in FAF-2?
I think the forecast for FAF 1 is about 70% growth all the way to 2020. FAF 2 forecast is -- we don't have that number on hand, but we are expecting by end of May, we will have that number.
The next question is, how does the FAF compare to other databases such as the Transearch database developed by Global Insight (previously Reebie Corporation)?
I can say this, both data components of CFS, Reebie has a lot of their proprietary information from their clients, but the difference here is ours, you know, whatever we add to it, you get to know it.
Randy, you said there are two sets of ton-miles data, is this publicly available?
It's not as of now. I guess that's more a question for Tianjia. I think the philosophy behind that was, and it's not on the CD, that that might be more misleading than informative. When we say the two sources of ton miles, there's, the reliable way to do it is to calculate your ton miles after you sign the flows to the particular network. Just measuring a centroid to centroid doesn't suggest that's the route it's going to take. That is something that will be forthcoming as Tianjia indicated after the assignments step when you can then actually calculate how did it go, and thus calculate the ton miles associated with that particular shipment.
The next question is, I believe this is for you, Tianjia, it would be interesting to see an analysis of all inter-metro traffic between the identified economic areas and those cities that you have collected data for along the inland waterways and gulf coast intercoastal areas. We worked hard to try to do this with FAF 1, but it was not easily available on the less than state level. Would this be possible for a pilot project?
I think, you know, we will talk more about this. Clearly, that is what my goal is, something, what we will be able to do, and what we are thinking here right now with pilot project is the key thing here, some kind of methodology, remember, FAF 2, the true -- data, only has 138 geographic regions, so we are hoping to focus on a large area, such as a State, rather than an MPO.
This is some clarification on the waybill data question. The public use version is a BEA to BEA O/D table. How did you transform the BEA to BEA table to the FAF-2 zone-to-zone or centroid to centroid table?
That is a question that's probably better answered by Frank Southworth on the team who did that. This is not the only case of where that came up. It comes up with respect to the waterborne commerce, and we developed crosswalks to translate this information -- whatever level of this aggregation -- so I guess I would aim to address that off-line as a more detailed something.
Okay, that's all the questions we have typed in, but we still have time left, so at this point, we can open the phone lines for questions, so operator, if you can give directions for asking questions over the phone, we can see if anybody has questions.
If you would like to ask a question, dial star-1 on your touch tone phone. If the answer has been answered or you wants to withdraw, press star-2. First question comes from the line of James McCarville.
I was wanting to ask about a pilot project, but I noticed there are were lots of people asking about what's expected of a pilot project, what do you expect from us, what do you expect, what do you want to accomplish. Can you give us more detail as to where you want to go with those?
Most of the inquiries are from MPOs, you are smaller geographic area than say a DOT, and so to tell the truth, I would like to see a state DOT make this request. For those of you who said you are potentially interested to be my pilot project, I'll get back to you by end of next week, I will send you more specific material and my thoughts. Thank you very much.
Thank you, Tianjia.
Since we don't seem to have anymore questions, we will close out the seminar for today. Thank you to all 3 presenters for great presentations and thanks to everybody who attended. Thank you all for attending today's seminar. The recorded version of this event will be available within the next week on the Talking Freight website.
The next seminar will be held on March 15, and is titled "Commodities: From Origin to Destination." If you haven't done so already, I encourage you to visit the Talking Freight Web Site and sign up for this seminar. I also encourage you to join the Freight Planning LISTSERV if you have not already done so.
Enjoy the rest of your day!