In 2021, the Federal Chief Data Officers Council conducted a pilot project to explore the use of knowledge graph technology and its applicability to data governance with a focus on wildland fire fuel reduction efforts. One of the primary goals of wildland fire management is to reduce the risk of catastrophic fires. To achieve that goal, land managers remove potential fuel for fires before they can burn. These fuels can include living and dead trees, shrubs, and grasses. Fuel reduction projects often take the form of prescribed fires, but fuels can also be reduced by cutting and removing them or other methods.
Determining the effectiveness of these fuel reduction projects—which take place in a variety of locations with vastly different characteristics and are conducted by different agencies, each with its own process—is a challenge. It is also the key to success. When we know which projects will be the most effective, it enables us to plan fuel reductions that will have the greatest impact at protecting communities and important resources.
To better identify, compare, and share information about the effectiveness of fuel reduction projects, this project applied a novel technique - using fuel data to build a knowledge graph, which captures the meaning of data and their relationship to other concepts. “We function in a very complex and diverse environment,” said Roshelle Pederson, Chief Data Officer for Wildland Fire. “This project demonstrates the power of a knowledge graph to create common understanding for both people and machines.”
The knowledge graph, or semantic network, takes existing data about fuel projects and connects the data elements in ways that reveal new insights about the effectiveness of fuel management and risk reduction. The project demonstrates how to manage and extract value from interagency wildland fire data. It also highlights how to provide highly reliable data that can improve land manager insights and the effectiveness of their operations.
The underlying data reflects a collaboration between the Department of the Interior’s wildland fire management bureaus, the Bureau of Indian Affairs, Bureau of Land Management, Fish and Wildlife Service, National Park Service; the Department of Agriculture’s Forest Service; and CalFire, the California Department of Forestry and Fire Protection. The data managers drew from interagency data available in the National Fire Plan Operations and Reporting System (NFPORS) and agency financial accounting systems to populate the knowledge graph.
“The collaborative nature of the program was a key factor in the project’s success,” said Tod Dabolt, Chief Data Officer at the Department of the Interior.
The pilot project is encouraging for a long list of next steps, from reporting system upgrades, to greater collaboration on wildfire risk mitigation, to database consolidation.
“This project is really exceptional,” said Pederson, “because we aren’t ending with a nice report to put on the shelf. We have tangible products that we can actually use as foundation stones for the whole wildland fire community.”