What feds must do to get a handle on water data

The federal government does not know how much water Americans used this week, this month or even this year. Its most recent account of water consumption and availability covers 2010, before Texas entered a four-year drought (which has since ended) and before California's drought crisis even began.

The problem is data. The most comprehensive analysis of the nation's water consumption, compiled by the U.S. Geological Survey, is released at painfully slow intervals with sizable gaps. The 2010 data was released in a 2014 report.

The agency relies almost solely on state-collected figures, often gathered via mail-in forms, and relates to information that can be hard to pin down. Figuring out groundwater withdrawals from high-capacity wells in Wisconsin, for instance, can come down to guessing how thirsty a cow might be.

Robert Smail, a water supply specialist at the Wisconsin Department of Natural Resources, acknowledged that there is a "range of precision" in the water data his state collects.

"About 6 percent of the high-capacity wells in the state do report dairy usage, but they don't have meters so they measure it based on an estimation of how big their herd is or how many cows they have," he said.

The federal government has never put a priority on water data. In 2003, the Government Accountability Office noted that national water availability and use had "not been comprehensively assessed in more than 30 years."

That's a big policy problem, said Charles Fishman, a water policy expert and author of "The Big Thirst: The Secret Life and Turbulent Future of Water." He calls the federal water data system "ridiculously primitive."

"We are facing a whole set of very important and very expensive water problems: what systems to repair and replace, what technology to use, which water users to tap for help and to pay for what needs to be done, even what problems to tackle most urgently," he said. "We make decisions about water every day using old information or deeply inadequate information or even using educated guesses. That's irresponsible and unnecessary."

Lawmakers have sought to change the situation. The Science and Engineering to Comprehensively Understand and Responsibly Enhance (SECURE) Water Act of 2009 endeavored to update the landscape of water data in the U.S. Guided by the law and the $12 million it promised to the agency, USGS officials launched a program that is among the biggest efforts to improve water data in the country's history.

The agency created the Water Use Data and Research program and last year doled out $26,000 per state to help officials write plans for improving their water data. Drafts were due in June.

Sen. Maria Cantwell (D-Wash.), ranking member of the Energy and Natural Resources Committee and one of the senators who introduced and championed the SECURE Water Act, said, "Communities across the United States are facing extraordinary water challenges, especially across the West. [The law] has helped communities plan for and respond to these challenges, including in the Yakima Basin in my home state of Washington."

But she added that this is only the beginning. "Building on this foundation, we must continue to modernize and improve federal programs and advance water science and data, which are key to transforming water management and empowering communities to build a more water resilient future," Cantwell said.

Technology challenges

Scaling up water data collection means overcoming some serious technology hurdles. David Blodgett, a civil engineer and project coordinator at USGS' Center for Integrated Data Analytics, explained what the government is contending with now.

"Our data challenges are largely around privacy and trust with use of furnished data," he said. "Another significant issue is around compatibility and interpretation of information that can inform a water-use estimate but requires some level of integration or interpretation to be meaningful. Normalizing these kinds of data across the country is a huge task that is, necessarily, a manual effort."

USGS software and IT services are largely provided in-house. The agency worked with a contractor to build database software, and USGS maintains the code.

"Recently, we have started to use the open-source R programming language to build [quality assurance/quality control] tools," Blodgett said. "This has allowed internal IT professionals to help water-use specialists craft and maintain their own tools in a way that follows best practices and should be more sustainable."

USGS has two main water-use databases, said Nancy Barber, a USGS hydrologist who serves on the agency's National Water Use Leadership Team. For the major water reports, information is stored in the Aggregate Water-Use Data System (AWUDS). USGS also maintains the Site-Specific Water-Use Data System (SWUDS) for information on water withdrawal. Both rely on an Oracle database and form part of the National Water Information System (NWIS).

"[For SWUDS], a move to a single central server is planned, as is a complete NWIS modernization, which will replace the outdated code with more modern code, and probably a redesign," Barber said. "As with AWUDS, the database has a long history with several architectural changes, although a larger share of the design and coding has been handled by USGS personnel."

No state is perfect

As noted above, USGS relies almost entirely on state data for its water reports. That's why the agency's effort to improve data collection has focused, in large part, on improving what happens in the states.

Melinda Dalton, a staff scientist at USGS, is one of the staffers who will wade through the work plans states have formulated to improve their sometimes antiquated data systems.

"Some states collect [data] digitally, but they only collect data for certain categories," Dalton said. "No state is the perfect example of everything."

For the draft plans, states were asked to outline the status of their water-use programs, including the information they collect, the scale at which it is collected and the frequency of reporting. They also defined their priorities for improving their activities, whether through better data collection or new methods of estimating water use.

However, resources vary from state to state, according to Dalton.

"There are some states that put a lot of resources toward water-use data collection," she said. "It depends on how much water they have, how many people and what the competition over resources [is]. A lot of times they are limited by legislation: They collect some things, and they can't collect others. You sometimes see a change in collection activities when water is becoming more scarce during times of drought."

Furthermore, some states lack the technological readiness to run a sophisticated water data program.

"Some of them talked about how they are still collecting information on paper," she said. "They'd like to do it electronically. We'd like to be able to report data as often as we can. That's our long-term goal. We get our data from the states. We can't report on use as often as we would like to. If states are still collecting data on paper, it creates a timing issue."

Texas: Collaborative data collection

Texas used its federal water data funds to hire an outside consultant, Freese and Nichols, to interview stakeholders and develop a work plan to submit to USGS, said Kevin Kluge, manager of water use, projections and planning at the Texas Water Development Board (TWDB).

The consultants floated the idea of integrating data to move the state closer to a single database for water data. Kluge said there were hurdles, however. "The alternative would be to integrate existing agencies and organizations," he added. "Even if we can't put it all into one database, we can utilize information collaboratively."

He sees the future of water data as "collaborative data collection" rather than "big data."

"Agencies integrating their existing data collection give us the largest chance of success in the future," he said. A more centralized approach "is not the way it exists for water."

At TWDB, integrating water data can be complex. The board collects three sets of water data: water loss audits, water use surveys and water conservation reports. TWDB undertook a major project to integrate that data two years ago.

"Even though we are on the same floor, it took a significant amount of time for staff to come together to meld these three programs," Kluge said. "We didn't push them all into one database. Instead, we developed ways to reduce duplication and develop standardization. That took a significant amount of coordination."

Kluge described how his agency approached the integration effort: "We hired employees and used regular employees within the agency to develop the screens and the data tables and the coding. We already had an online water use survey, and we had an existing water loss audit. The project required tweaking those and transferring data, rather than melding them all into one online application."

A single application might not have served the diversity of the data, he added.

"Different programs have different targets," he said. "The idea of squishing all these programs into one online application when you have three different targets set for the program didn't make sense. We ended up keeping existing applications and developed coding from the water use survey and pushed it into the water loss audit."

Single database in Nevada

Nevada already has a single database for water-use information, but populating it will take a couple years, said Adam Sullivan, chief of the hydrology section at the Nevada Division of Water Resources.

"One step is entering existing historic data, which just takes staff time," he added. "More interestingly, water users throughout the state can now log on and enter their totalizing meter readings. The challenge on our end is data QA/QC and making sure that the database functions correctly in generating summary reports."

Another hurdle is maintaining all the data in a common format, he said.

The work plan the state developed for USGS proposes supporting an employee "dedicated to water-use data QA/QC and timely reporting, and to improve the functionality of the pumpage database to meet the needs of data contributors and users, particularly with the USGS Nevada Water Science Center," Sullivan said.

The state spent part of its $26,000 federal allocation to hire a contractor for software development.

Wisconsin: 'Data is good when it's used'

A key proposal in Wisconsin's work plan is improving data collection and delivery. According to Smail, that means making it easier for stakeholders to report data and delivering high-quality applications that use the state's data.

"Data is good when it's used," Smail said. "We want to make sure we are being useful to those reporting data, instead of just being a regulatory requirement. There are a lot of opportunities for us to give data back to customers in a way they can use it, and that, in turn, improves the data because you wind up having more contact with the people giving you data. My top priority is adding context and meaning to the data so it's not just a set of numbers."

Gaps in internet coverage in rural areas pose a challenge, though. "A lot of our people reporting live in areas of the state that don't have high-quality internet access," he said. "You might have a retired person who owns a piece of land and rents it out."

Once the information comes in, tending to it can take significant resources. "The IT challenges are less insurmountable than the maintenance challenges: maintaining contacts with customers, entering new data into the system," he said. "Database structure gives us a great deal of flexibility."

To aid in that effort, Smail's staff focuses on maintaining continuity in the data even when variables, such as well owners, often change.

Database limitations in Washington state

Washington state is wrestling with a database structure that was created to store water measurement information.

"Over time, the needs have changed, and we're limited by the existing database structure," said Kristin Johnson-Waggoner, communications manager for the Water Resources Program at the Washington Department of Ecology.

The department wants to increase its access to automated reporting to improve the accuracy of its data. However, Johnson-Waggoner cited "political, social and economic challenges."

"In the work plan, we identified both IT needs as well as procedural and policy changes that we would need to implement to have a more robust water measurement program," she said. "Obtaining the grant allowed us to spend dedicated staff resources to evaluate our needs."

A vision of the future

At USGS, Dalton can imagine a future when the federal government has a clearer picture of water use. The goal is getting "an idea of water use on a monthly timestamp, but that's many years down the road," she said.

It would require significant improvements to data collection and estimation techniques, but states seem keen on improving their water data, she added.

"The response has been overwhelming," Dalton said. "States are very eager. The response to the program has been phenomenal."

As part of the program, USGS organized stakeholder meetings and conference calls to connect teams from around the country. The agency will continue making grants, now on the basis of merit, in the coming years. More than 40 states applied for federal money to write their work plans, and dozens of state aides join the USGS conference calls.

"One of the things we heard was that there's not really an outlet for state water-use people to talk about water use and the work they do," Dalton said.

Fishman said he imagines even bigger possibilities for federal water data. He said the government could develop a water version of the Energy Information Administration, which was created during the 1970s energy crisis.

"The goal is simple: We need to get the most out of the water we have available, at the most reasonable cost," Fishman said. "We can't possibly do that without information about who is using water and how much they're using every day."