Summary: FracFocus disclosures that contain unintentional duplication of records distort the quantity of the chemicals used. Open-FF1 found 149 such disclosures by Camino Natural Resources, LLC. On average, these duplicates inflated chemical mass in disclosures by 60,800 pounds. We describe the details of those errors here.
Background on duplication errors
The national disclosure instrument, FracFocus, documents the fracking chemicals that have been used in over 220,000 fracking jobs in the US since 2011. This extensive data set is especially important because it is one of the few resources available to the public about fracking chemicals. As such, FracFocus should be central to analyses and debates about fracking patterns and impacts. However, poor data integrity can undermine the usefulness of this resource.
The Open-FF project is trying to make errors in FracFocus more visible to stakeholders. Our aim is to alert users of FracFocus data to weaknesses and to encourage operators to correct problems and prevent future errors.
We recently summarized an odd problem in a large number of FracFocus disclosures. Many disclosures have duplicate records, that is, two single lines share identical values in many identifying columns. Because such records can be confused with legitimate records, users of the data are confronted with significant ambiguity. If they are errors, these records distort the quantity of chemical usage.
We’ve found evidence that these duplicates are unintentional. They have also gone largely uncorrected, some for over 8 years. As we’ve noted, because our project is based only on publicly available data, we cannot be certain that the duplicates are not legitimate; only feedback from the companies can clarify this issue.
What a duplicate from Camino looks like
The duplicates we have detected are present in all forms of the FracFocus disclosures: PDFs, online “Find-a-Well” summaries and the bulk data downloads. The following is an example FracFocus disclosure summary2 from Camino Natural Resources, LLC. Duplicates are visually easiest to find in this format because the records are sorted by one of the duplicated columns, making those records adjacent. We have highlighted detected duplicates. (The FracFocus online disclosure can be found at this link, where you can also find the PDF version).
While this FracFocus format shows only three columns, the duplicates are identical to their matching record in five columns. Those columns are: “CASNumber,” “IngredientName.” “PercentHighAdditive,” “PercentHFJob” and “MassIngredient.” “MassIngredient” is only available in the bulk download version. Open-FF uses an algorithm to flag these records across FracFocus in the bulk data download. The algorithm only flags records when the first 4 are not blank or zero and all five are identical.
We use the “Purpose,” “TradeName” and “Supplier” columns to determine which of the two records to flag. Often the duplicates will be together in one section of PDF. However in most cases, not all in that section are duplicates.
Affected disclosures of Camino
We have detected 149 FracFocus disclosures published by Camino Natural Resources, LLC with duplication errors. Of a total of the company’s 164 disclosures, 90.9% contain duplicates. These disclosures are found exclusively in Oklahoma.
This report is based on a snapshot of the FracFocus data (Sept. 12, 2024). The number of disclosures with duplicates will likely change as operators add more disclosures and correct mistaken ones. To see the most updated status of Camino Natural Resources, LLC’s disclosures with duplicate records and a comparison to this baseline snapshot, visit this link.
Duplicate records in Open-FF data sets
Users of FracFocus data via of Open-FF data sets can remove duplicates by using a standard data set. However, because we only have access to publicly available FracFocus data, we can’t be sure that the records we flag are truly mistakes. In some cases, our detection process may be flagging legitimate records. We leave it to the users of our data sets to weigh the costs and benefits of keeping vs. removing these records. In our experience, it is safest to remove the duplicates.
Title image credit: Becky Mansfield (modified by author)
- This work is part of an effort at The Open-FF Project to improve the quality and accessibility of the FracFocus data. ↩︎
- screenshot from Oct. 2024 ↩︎