Data Dilemma: Waterway Infrastructure Is Founded On Data
(Editor’s note: This is the first of a planned periodic series on the data infrastructure underlying the physical infrastructure of the U.S. inland waterways system.)
The symbol and logo for the U.S. Army Corps of Engineers is a turreted castle. Shawn Komlos compares the work his team does to the castle’s foundation. Komlos has served for four years as director of the Corps of Engineers Navigation and Civil Works Decision Support Center (NDC). As NDC director, he leads a multi-disciplinary staff of roughly 29 scientists, engineers, statisticians and computer scientists assigned to offices in Alexandria, Va., and New Orleans, La., in an organization whose workforce chart allows for 50 positions.
The staff is dedicated, but under-funding and under-staffing has left it struggling to provide the data that users need to make informed decisions about waterways operations. A former commander of a Corps of Engineers port district said the Corps’ information services is “one of the most under-resourced units in the organization.”
The data the support centercollects and provides supports and underpins key decisions on the entire inland water transport infrastructure the Corps maintains and operates. A variety of federal customers, including the Department of Homeland Security, U.S. Coast Guard and the Maritime Administration, as well as members of Congress and their staffs, rely on that data.
Foundations support the entire structure, yet they can be invisible to observers. Those who work in the inland waterways sector often note the entire sector faces a visibility problem with the public compared to rail and trucking. The work of the data collectors and producers, whose facts and figures provide support for vital decisions for waterway infrastructure, can go unremarked, even by those who use it.
Lock performance, dredging data and notices to navigation are among the real-time data systems the NDC manages. In addition to those real-time systems, the team also manages waterborne commerce data out of the Waterborne Commerce Statistics Center, which are more “after-the-fact.”
NDC is responsible for management of waterborne commerce, navigation infrastructure performance and civil works mission data and associated data systems. The center works with its stakeholders at Corps headquarters and in the field to deliver data and data products that support investment decisions, strategic planning and operational needs.
The data the Institute for Water Resources (IWR) collects from vessel operators is legally mandated. All vessel operators and carriers must fill out a document known as an ENG Form 3926, “Vessel Operation Report Statement of Freight and Passengers Carried.”
oday, the forms are mostly submitted electronically, with only a few operators holding out for paper. Some vessel operating software packages have built-in capabilities that allow operators to fill out and send in the forms more quickly. The forms solicit information on type of vessel, destination, departure and arrival times, whether it loads or unloads cargo, type of commodity and tonnage.
Komlos said the response rate from vessel operators is in the high 90th percentile range, which is similar to response rates provided by farm producers to the survey-based efforts of the U.S. Department of Agriculture’s National Agricultural Statistics Service (USDA-NASS), whose periodically released crop reports are used in the World Agricultural Supply and Demand Estimates (WASDE) report, a market-moving benchmark public data product.
Komlos does note, though, that some new operators might not be reporting and that others can vary in how quickly they file their reports. Occasionally, a charterer or smaller operator may need to be reminded of the obligation to fill out the information. Once received, the data are then reviewed for quality assurance and publicly published once each year.
At one time, the NDC performed a broader array of functions, which included sending a team of engineers into the field to gather data about ports and terminal characteristics. Due to budget constraints, though, the center has come to rely on other sources of information. NDC staff support activities of the Inland Waterways Users Board and deliver technical support to Corps headquarters and users of data managed by the center.
The NDC is funded by Congress under a budget line known as “Remaining Items.” Each of its four centers has a separate funding line, and funding can vary from year to year. According to Komlos, funding for center operations has been fairly consistent for the past 18 years. Though there was a roughly 2.5-year hiatus in funding for modernization of its data systems, there has been increased support for completing system improvements.
“We have seen some upward movement in budget numbers,” Komlos said.
That upward movement is necessary, he said, as the NDC faces increased challenges with cybersecurity and keeping its systems up to date.
The NDC is in the middle of a multi-year upgrade to its data systems, some of which were initially built 20 to 30 years ago. While the systems have been updated periodically, a decision to modernize the entire data system was made around 2018. The upgrade into a “cloud-smart” system will incorporate new geospatial features that should make it easier to produce new data products, Komlos said. The upgrade should be completed in late 2025 and will support the processing of 2026 data, he said. His group is also looking for opportunities to leverage artificial intelligence to speed up and improve the collection, processing and interpretation of data.
“There are lots of opportunities for improvements based on AI,” Komlos said. “We’re frequently approached by third parties interested in partnering. We’d like to do what we can to automate data entry where possible to reduce time involved in processing the data.”
Komlos said he and his staff are motivated by a “passion” for the data that supports the entire waterborne commerce system.
“I’ve never seen anyone telling stakeholders, ‘We can’t do that,’ ” said Amy Tujague, a project manager for the Waterborne Commerce Statistics Center (WCSC) who is looking at ways to streamline and improve data delivery on the Corps’ websites.
“We’re always looking for ways to do more with the funds that are provided to deliver the data required by our stakeholders,” Komlos said. “Unused data are useless data, so effectively delivering the data mission is important to us.”
To communicate the center’s data mission, Komlos pointed to an old slogan for chemical firm BASF: “We don’t make the products you buy, but we make the products you buy better.”
Komlos’ version is, “We deliver the data that makes the infrastructure better.”
Gaps, Delays In Data
Some customers who rely on Corps data, though, have expressed concerns about the completeness and timeliness of the information.
Ken Eriksen, a longtime senior leader and consultant on commodities, supply chain and logistics, and a contributor to The Waterways Journal, notes one example, saying, “The Monthly Commodity Indicators Report has been a valuable resource to track commodity flows on the inland waterways. Many industry participants use this data for planning and monitoring to make strategic decisions in their businesses and for policy and regulation.”
But he notes concerns about the timeliness of the data. Reports that are supposed to be monthly are often delayed. Users are guided through several clicks to various websites instead of being taken straight through to the data.
According to Eriksen, the Corps is now including “final” history by month from the WCSCthrough December 2022. There is a “helpful and commendable” explanation of what the data represents and how it is compiled. However, Eriksen noticed some important data discrepancies going back to 2016.
“The most egregious difference is coal and coke,” he said. “Using unofficial data from the lock performance monitoring system (LPMS), the lock data, counting barges with identified cargo commodity and estimated tons per barge overstated volumes by an average of 26.2 percent from a range of no difference to being 56.1 percent too high as compared to final. Coal is moved in open barges and easily identifiable. There is an extraordinary level of over-counting the volumes.”
Eriksen also pointed out an apparent under-counting of other commodities.
“In contrast to coal, food and farm products were understated an average of 10.5 percent,” he said. “Petroleum volumes were overstated an average of 4.4 percent, and chemicals were understated an average of 1 percent.”
Eriksen tracks an “other” category to reflect total volumes reported, less the sum of the volume of commodities provided. “Other” is overstated an average of 14.5 percent, he said.
According to the Corps, the monthly “indicators are estimated using linear regression models that mathematically delineate the relationship between the tonnages reported for a given month at key LPMS locks to the monthly WBC (Waterborne Commerce) value.” In spite of the deficiencies, that WBC data is what the industry must accept and use.
“That the Corps is one-and-one-half years late reporting final data, and big errors are showing through the data, is concerning,” Eriksen said.
Carriers and operators need to recognize the importance of submitting their vessel operation reports in a timely and complete manner, he said, adding, “Reliance on this data could make or break investment decisions.”
