Florida DEP Approves Predictive Modeling for LSLI

Florida Department of Environmental Protection (FDEP) recently released updated Lead Service Line Inventory (LSLI) Guidance that outlines the approved methods for service line material identification in the LSLI, including predictive modeling and statistical methods. BlueConduit applauds FDEP for providing water systems with clear guidance well ahead of the October 16, 2024 compliance deadline.

The FDEP guidance specifically advises water systems on Florida’s three approved methods for classifying service lines as non-lead: 1) physical verification; 2) historical record review; and 3) statistical methods / predictive modeling.

Predictive modeling is especially attractive in Florida because FDEP’s guidance explicitly limits the historical records that can be used to classify service lines as non-lead to 1) built after 1989 and 2) pipe diameter >2 inches.

To utilize statistical methods or predictive modeling as a method for classifying service lines of unknown material as “non-lead,” FDEP requires that the water system provide:

  • A summary report describing the representative source data used to train the predictive model;
  • Information describing field verification methods;
  • Information describing methods used to validate historical records;
  • A report showing that predictive model results were evaluated and tested to show that model has:
    • A minimum of 95% confidence level and 
    • An accuracy rate of 90% or higher for verified model predictions
  • Information describing how model avoids potential biases, as well as potential false negatives and positives;
  • A summary of model predictions including the translation of predictive model results to parcel level

FDEP explicitly does not require water systems to submit a plan for approval in advance of using statistical methods as a classification method for the LSLI. BlueConduit is in ongoing contact with FDEP regarding acceptable criteria for material classification and threshold decision making for classifying non-lead; we will continue to provide Florida water systems with updates as FDEP provides continued updates.

Want to learn more about using statistical analysis to re-classify “unknown” service lines as “non-lead?”* Get in touch!


About BlueConduit

BlueConduit pioneered the predictive modeling approach to lead service line identification and replacement. Through BlueConduit data science, utilities, municipalities, government agencies, and consultants standardize, predict, report, and communicate key information about lead.