The first use case of the SAWGraph project is to support identifying gaps in testing for PFAs and understanding the sampling that does exist. This use case contains a number of competency questions which answer the following types of questions:
The goal of this use case is to support state and federal agencies and researchers in prioritizing where to test.
We demonstrate this use case by providing visual results for a number of example competency questions that exemplify specific querying tasks. Each example is also available as a linked jupyter notebook that contains the python code to generate the maps and the SPARQL queries.
State Environmental Toxicologist with a focus on drinking water. From a state/region that has not yet done comprehensive water testing
The first stage of testing prioritization for many states involves identifying facilities that are likely to be utilizing PFAs chemicals. This competency question addresses features of enivronmental concern (surface water bodies) that are near potential contamination sources.
The query could easily be modified to identify other features of concern (water supply wells, streams, etc.) or other potential contamination sources. This query would inform testing prioritization as it would help identify surface water bodies that may be at risk of contamination.
The map was generated by querying the SAWGraph for facilities of the specified type in the state of Maine (that are suspected pfas sources), and then finding the nearby surface water bodies based on the S2 Cell spatial indexing.
The full code is available: here and here
This question demonstrates some more detailed hydrologic connections via the aquifers. First we identify private wells that tested above 200 ng/l in PFOA. Next we determine which aquifers they are likely fed by based on proximity in the same S2 cell. Then we discover all other private water supply wells that are likely fed by the same aquifer.
This question facilitates prioritization of private well water testing by identifying high risk wells that have not been tested based on the existing data. It provides a more sophisticated hydrological modeling than simple proximity.
The code for this example is available here .