Data Infrastructure Investments Could Increase Effectiveness of Reservoir Management

One of the largest repositories of historic reservoir data in the United States is the U.S. Army Corps of Engineers. Its districts have been amassing data on hundreds of reservoirs for decades, but, like many other water data gathers in the United States, it cannot always use its own information to support broad-scale decision making. In a new report by Duke University’s Nicholas Institute for Environmental Policy Solutions, Lauren Patterson and Martin Doyle of the Nicholas Institute and Samantha Kuzma of the World Resources Institute point out that the federalist structure of the Army Corps and other U.S. agencies has often led to wide variation in data management, requiring development of protocols for standardizing and integrating those data. Pointing out that water management transcends political boundaries—requiring data sharing within and between agencies at the scale of watersheds or river basins—they describe the challenges of and opportunities for using the Army Corps’ historic reservoir data to understand how reservoirs are performing as environmental and societal needs change.

Why did you want to explore the Army Corps’ data management efforts?

The Army Corps is the predominant steward of waterways and reservoir systems across the United States. Until recently, it had no centralized database containing historic records of reservoir levels, flows, or operations—and consequently, no way to develop a systematic, nationwide analysis of the impacts of changing conditions on reservoir operations. The Army Corps is now constructing that database through inquiries to its districts, each of which has its own data management system. Our effort to integrate currently open but disparate data from the Army Corps’ districts highlights the challenges and potential opportunities of similar efforts at other federal agencies.

What kind of opportunities?

With one agency-wide, standardized dataset, the Army Corps can assess reservoir performance in terms of management goals for reservoir levels over time and can identify reservoirs with operations that are particularly sensitive to changing environmental or societal conditions. The results of such a synthesis are dependent on the underlying data, which were not originally collected for that purpose. You need good metadata to understand relevant decisions, limitations, and caveats.

How did you conduct your case study?

In cooperation with the Army Corps, we developed a national reservoir database of historic reservoir data and the Army Corps’ management goals using primarily data that were accessible and available as of 2015 from the Army Corps districts within the coterminous United States. Our goal was to create a prototype “data as a service.” That is, we attempted to standardize reservior data between districts along with each reservoirs operational targets, which represent the management goals for reservoirs in terms of lake elevation. These uniformly formatted data were placed in a centralized data repository. Synthesizing both historic reservoir levels and operational targets within a single database facilitates immediate use of the data to assess how well reservoirs are meeting their operational targets.

How much of the Army Corps’ historic reservoir data is included in your study?

As of September 2016, the data are available and accessible for 51 percent of districts and account for 65 percent of reservoirs identified as owned and operated by the Army Corps. Some districts had no historical data available for download online, and we had to amass it through direct contact with those districts. Collecting and formatting some of the online data into a usable time series took considerable time and effort. We created a tool to visualize the provisional data.

Aside from the large variation in data availability and accessibility among districts, what are the other main takeaways from the analysis?

The analysis found that to fully benefit from the data the Army Corps is collecting and storing, it must make the data available and accessible. Also, the costs of integrating disparate data systems across the Army Corps are relatively small compared to capital expenditures—but the benefits can be immense. Having available and accessible data as a service will drastically reduce the costs of gathering data from different systems for internal studies, will allow individuals and agencies to develop tools to use the Army Corps data to provide new insights and strategies to address water resource challenges, and will decrease litigation costs and Freedom of Information Act requests as water supply issues and interest in them grow. 

Members of the media interested in speaking with Patterson or Doyle should contact Erin McKenzie, erin.mckenzie@duke.edu or 919.613.3652.