The Water Quality Portal consolidates hundreds of millions of data points — but navigating the largest water data repository in the U.S. comes with challenges. Water entities rely on data to make critical source water protection decisions, provide Clean Water Act regulatory guidance, and understand water quality changes. Despite the potential of these water quality observations, the data are often difficult to access and analyze. Collaboration between data providers and data users can unlock data repositories and take them from storage to actionable insights.
The Radical Open Science Syndicate (ROSS) — a watershed data science lab at Colorado State University (CSU) — partnered with the U.S. Environmental Protection Agency (EPA) to improve water data access through Tools for Automated Data Analysis (TADA), a toolkit empowering water resource managers to use the data they collect.
When Duke University and the Internet of Water Coalition funded ROSS to build a tool to make water quality data more accessible, they found that the EPA’s Office of Water already had built TADA. Rather than reinvent the wheel, ROSS directly partnered with the EPA. Staff scientists Matt Brousil and Katie Willi applied their open-source and geospatial science expertise to improve TADA, adding functions like simplified spatial data queries to streamline data access in specific locations.
We sat down with Dr. Matt Ross, ROSS Principal Investigator and Director of the Geospatial Centroid at CSU, and Dr. Martin Doyle, Internet of Water Coalition (IoW) expert and Director of the Water Policy Program and professor at Duke University’s Nicholas School of the Environment, to talk about the success of this partnership and water quality data accessibility.
Martin: I beat the crap out of Matt Ross as a PhD student at Duke, so I figured I could trust him. I knew that he had a deep command of water quality data, and he knew how it was used and what you could do with it. But in his new role at CSU, I knew that he also had access to a team that could translate from database to useful data.
With funding from the BHP Foundation for the Internet of Water Coalition, we had the chance to give Matt and his team some runway to work on this problem and see what was possible. So, the alignment really worked for content, capacity, and timing.
Matt: Martin reached out to us to build a water quality data extraction and visualization tool. When we found that the EPA was working on something very similar, we asked Martin to adjust our scope a little to amplify the work of the EPA, instead of building a parallel tool.
Dr. Martin Doyle, Nicholas Institute for Energy, Enviornment, and
Sustainability (left) and Dr. Matt Ross, Radical Open Science Syndicate (right).
So this became a three-part team, with IoW providing funding and vision, EPA providing architecture and needs assessment, and ROSS delivering code development for specific aspects of the TADA tool. Through this partnership and strong connections at EPA, we focused on geospatial workflows for the EPA and helped them expand the capabilities of TADA.
Martin: The greatest barrier to water quality data has been that it has been stored in a way that makes usability and interoperability super difficult. I often find that it takes at least two class sessions in my hydrology class for students to be able to use the EPA water quality data.
Making the database more usable has been something that has been on the ‘to do’ list for many freshwater scientists for years because we, as a nation, have invested so much time and resources into this database yet it is largely impenetrable outside of the research community. The more usable it is, the more it might get used by people who are in water-relevant realms and need broader insight on water quality patterns and trends.
Matt: Absolutely. My other PhD adviser Dr. Emily Bernhardt likes to talk about how data storage can become a data vault, where holding the data almost becomes the priority over extracting and using it. In the case of the WQP, the system has a long history of design around regulatory compliance concerns, not necessarily easy extraction and use. In addition to some of the structural design problems, water quality data is just highly variable by its very nature and harmonizing across providers is an incredibly difficult task.
We're making TADA more usable in a variety of ways: using geospatial tools to help users select data by location, making it easier to retrieve data you’ve submitted to the system, harmonizing data inputted from different sources so you can compare data at scale, and easing the burden of pulling a lot of data at once.
Matt: The open-source component is critical for many reasons. First, this work is ultimately elevating the value of taxpayer funded water quality collection programs. As a result, it’s vital that we keep the tools meant to amplify the value of that data free and open to the public. Second, we only were able to find out so much about TADA because it was an open-source code repository, publicly visible to all. Third, the EPA TADA team, led by Christina Mullin, had an open-source mentality in the construction of TADA, meaning they built their repository in a way that invited external collaborators. This well-organized template and the open-source collaborative spirit made it possible for us to jump into a brand new, large project and contribute significantly.
Martin: These tools help non-researchers actually use the data that already exists. By adding options to query with spatial objects or tribal area names and improving retrieval for queries with mass amounts of data, users will find the data they need and identify actual data gaps more readily. This is key. The motto of the Internet of Water is ‘better water data for better water management,’ and water quality management depends on having some sense for patterns and trends. The more breadth of data you have, the more informed you are as a manager.
Matt: Exactly as Martin says. We can’t use data for decision making if we don’t know how to access it, analyze it, and contextualize it. There are a lot of examples of what this looks like in a workflow for water quality monitoring and standard specialists for states or tribal nations, researchers, or anyone using regulatory data to make decisions.
For instance, the new TADA mapping tools now allow users to quickly identify upstream and downstream monitoring sites during a query, eliminating the need to export data into a separate GIS platform. TADA can also enable faster regulatory assessment. State and Tribal agencies can use it to streamline data retrieval, basic QA/QC, and visualization for 303(d) and 305(b) reporting, reducing the time and expertise needed to clean and compile data. Built in functions can also simplify data harmonization, reconciling inconsistent units and formats across datasets, so users can focus on analysis instead of data wrangling.
Martin: The next big challenge is to get the agencies within states to standardize their data platforms and programs. We often find that the same state can have water-related agencies with completely different data systems and practices. Because states are the real locus for water management in the U.S., we need states to adopt better water data policies. A few states are leading the way on this, such as New Mexico, Oregon, California, and Texas. Idaho and Indiana are starting to look into water data policies. We hope to keep building the number of states, and over time we’ll have built a big library of how to best complete data modernization transitions so that latter states can take advantage of these first movers. We can’t use data for decision making if we don’t know how to access it, analyze it, and contextualize it. The TADA tools, and all tools that make it easier to work with water data, encourage using the data and adopting it for decision support.
New Mexico has been an early adopter of open water data management.
Matt: I think there are two kinds of data collected for water quality: 1) Regulatory collected data, which these tools are meant to help with, and 2) Data collected directly for immediate decision support by utilities and other water providers. This data is often collected by sensors, is proprietary or not public, and even more difficult to harmonize than regulatory grab sample data. Despite these difficulties, real-time decision making requires real-time and forecasted data, and if we really want to harness the potential of big data in water, AI, digital twins, etc., we will need to have equally sophisticated tools to collect, harmonize, and amplify the value of these real time sensor systems.
Like Martin said — a growing number of states and organizations see how new data systems and practices pave the way for better water data management. Rather than reinventing the wheel, this partnership exemplifies what can happen when institutions participate in transparent dialogue, build on each other’s strengths, and prioritize efficiency over propriety. Using open science, paired with flexible funding and execution models, can elevate tools like TADA far beyond their original scope. As we collect more data than ever before, the real opportunity lies not in building new platforms but investing in the care and long-term maintenance of the tools we already have.