The inland waterway system of the U.S. is a vital network for transporting key goods and commodities from the point of production to manufacturers and consumers. Shipping materials via the inland waterways is arguably the most economical and environmentally friendly option (compared to hauling freight by trains or railways). Despite the advantages the inland waterways enjoys over competing modes, key infrastructure – such as locks and dams, which help to control water levels on a number of rivers and make navigation possible – is declining. Limited funds have been allocated to make the necessary repairs to lock and dam facilities. Over the past 10 years Inland Waterways Trust Fund resources (which historically funded maintenance and improvement projects) has steadily declined.
Locks and dams are of particular importance, because they assist in the maintenance of navigable depths on many of the major inland waterways (Ohio River, Upper Mississippi River, Tennessee River). To better understand the operation of the inland waterway system, this report examines a portion of the Ohio River (extending from Markland Locks and Dam to Lock 53). The specific focus is to determine what delays barge tows as they attempt to lock through these critical facilities. The Ohio River is a particularly important study area. In many ways it is representative of the conditions present throughout the inland waterways system. The average age of the lock and dam facilities exceed 50 years along our study segment. Most of these facilities are operating beyond their intended design life. As locks age, they increasingly demand more scheduled and unscheduled maintenance activities. Maintenance activities often require temporarily shuttering a lock chamber and diverting traffic through another onsite chamber (often of smaller capacity). All of the facilities included in the research area have two lock chambers ‐ thus, if one goes down for maintenance all vessels are diverted through the second chamber. In many cases this situation can produce extensive delays, which precludes cargo from reaching the destination in a timely manner.
Recently, the aggregate number of hours that shippers and carriers lose due to delays has escalated. Although the U.S. Army Corps of Engineers – the agency responsible for the management and oversight of locks and dams – has worked to keep traffic flowing on the river, tightening budgets hamper efforts. For shippers and carriers to make informed decisions about when and where to deploy freight on the river, they require knowledge that illuminates factors that are most significant in affecting transit times. In particular this applies to certain conditions that are likely to create delays at lock and dam facilities.
The purpose of this report is to 1) develop a comprehensive profile of the Ohio River that provides an overview of how it is integral to U.S. economic security 2) identify salient river characteristics or externally‐driven variables that influence the amount of water flowing through the main channel which consequently impacts vessels’ capacity to navigate 3) use this information (along with a 10‐year data set encompassing over 600,000 observations) to develop an Inland Waterways Operational Model (IWOM). The IWOM objective is to provide the U.S. Army Corps of Engineers, shippers, carriers, and other interested parties with access to8 a robust method that aids in the prediction of where and when conditions will arise on the river that have the potential to significantly impact lockage times and queue times (i.e. how long a vessel has to wait after it arrives at a facility to lock through).
After qualitatively reviewing different features of the river system that affect vessel traffic, this report outlines two approaches to modeling inland waterway system behavior – a discrete event simulation (DES) model which uses proprietary software, and the IWOM. Although the DES produced robust findings that aligned with the historical data (because it relies upon proprietary software), it does not offer an ideal platform to distribute knowledge to stakeholders. Indeed, this is the major drawback of the DES given a critical objective of this project is to generate usable information for key stakeholders who are involved with inland waterway operations. Conversely, the IWOM is a preferable option given it relies on statistical analysis – in this sense, it is more of an open‐source solution. The IWOM uses linear regression to determine key variables affecting variation in lockage time. The final model accounts for over two‐thirds of the observed variation in lockage times from 2002‐2012, which is our study period. Practically, this means that the difference between predicted values and observed delay times is significantly less than how the delays vary around the composite average seen in the river system (R2 = 0.69).
The IWOM confirms that variations in river conditions significantly affect vessel travel times. For example, river discharge ‐ the direction a vessel moves up or down a river ‐ meaningfully influences lockage times. The freight amount a vessel carries, which is represented by the amount of draft and newness of a vessel, influences lockage times. Larger vessels with more draft tend to wait longer and take longer to complete their lockage. The IWOM is less successful at predicting delay times. Because there is greater instability in this data only a modest amount of variation is explained by the model (R2 = 0.23). This, in turn, partly reflects in spillover from one vessel to the next that is difficult for the simulation to impose and account for therefore requiring additional logic.
Once completed, the IWOM was used to parameterize a simulation model. This provided a graphical representation of vessels moving along the river. Users have the capability of adjusting the effects of different variables to anticipate how the system may react, and what changes in vessel traffic patterns emerge. This information will be of great use for stakeholders wanting to gain a better understanding of what conditions lockage times will increase or decrease, why delays emerge, and consequently how these impact traffic flows on the river. In programming a simulation model, users are able to visualize and intuit what causes vessel travel times to vary. Although the regression model accomplishes this, for many users this would prove unwieldy and difficult to grasp beyond a conceptual, abstract level. Matching up regression results with a visual counterpart lets users gain immediate and intimate knowledge of river and vessel behavior – this in turn can positively affect shipper and carrier modal choices. The report concludes with some recommendations for IWOM implementation and thoughts on future research needs. Also discussed are the implications results from the present study have for improving our ability to safely, securely, and swiftly move freight on the inland waterways network.
Digital Object Identifier
Kreis, Doug; Sturgill, Roy E.; Howell, Brian K.; Van Dyke, Christopher W.; and Voss, D. Stephen, "Inland Waterway Operational Model & Simulation Along the Ohio River" (2014). Kentucky Transportation Center Research Report. 1458.