Planning for the future of FCSAP post-2020 required having reliable information on the number of sites and the expected remaining costs for assessment, remediation/risk management and long-term management of the FCSAP eligible sites that will not have been closed by that time horizon. While custodians have this information on their sites that are active and have passed Step 3, there are still approximately 2,000 suspected sites, and 900 more at the initial testing step that are unclassified or have no liability estimate. The FCSAP Secretariat solicited the development of a defensible method to project the remaining funding demand through to 2020, and post 2020. The developed predictive tool uses the characteristics of similar sites to calculate the cost to remediate sites which have not yet started, and the proportion of those sites which will become Class 1 and 2, and can, according to current eligibility rules, be FCSAP funded. The tool uses the data in the Federal Contaminated Sites Inventory, plus additional information requested from custodians and a predictive algorithm to calculate number of sites and costs post-2020. The premise at the heart of the algorithm is that sites that are similar by function will have a similar contaminant profile, site class distribution, and similar assessment, risk management/remediation (R/RM) and long-term monitoring (LTM) costs. This presentation will describe how the model uses the characteristics of closed sites or nearly closed sites in each category to predict the following information on sites that are unclassified or have no liability estimate: Distribution of unclassified sites into Classes 1, 2, 3 and N; Proportion of sites that will require remediation after assessment and the proportion of sites that will pass onto long-term monitoring after remediation; The expected duration of assessment and remediation of contaminated sites in the Category; and, Costs for assessment, R/RM and LTM for contaminated sites in the Category. Cost projections were initially performed only for the sites which were predicted to become Class 1 and Class 2 (based on projections). In a parallel analysis to the predictive modelling, the sites with a known or expected liability of over $5M were identified as high cost-high risk (HCHR) sites, and were subject to a risk-based analysis of costs. Custodians of those sites were convened in a workshop and assisted in filling out a template that estimated the cost uncertainty by key stage in the project. The known liabilities (estimated by custodians) were added to the outputs of the predictive model’s liabilities and the HCHR risk adjusted liabilities to determine the future funding demand for FCSAP eligible contaminated sites. The predictive model has recently been enhanced to predict the liability of Class 3 sites at the request of the Office of the Auditor General.