11 Computer models are used by scientists to assess the potential success of a proposed remedy. Well-designed models use current and historical data, coupled with assumptions about what a proposed project will do, to predict future conditions. The ultimate reliability of a model is judged by how well its predictions compare to historical, known conditions. NOAA did not compare its model predictions to actual data, contrary to accepted practice. Had it done so, a comparison would have shown that the NOAA model over-predicts PCB concentrations in both fish and water. Figure 2 shows the NOAA model predictions of PCB levels in water are two to four times higher than actual data. EPA concluded NOAA’s model “is based on analyses that did not reflect the breadth of project sediment data or the variety of fish species data across sampling stations in the Upper and Lower Hudson River, and therefore is not supported by the available evidence.”12 For those reasons, EPA does not believe NOAA’s conclusions are valid.13 Actual data collected since the completion of dredging show PCB levels in fish are near or below pre-dredging levels; indeed, EPA says, the response in fish is in line with what its own model predicted.14 WATER TREATED: NUMBERS OF EMPLOYEE HOURS WORKED: CLEAN BACKFILL LAID DOWN: as of 10/1/2015 2.57 MILLION TONS RAILCARS LOADED: 29,012 BARGE LOADS: 4,800 FILTER PRESS DROPS: 31,831 WATER AND AIR SAMPLES 29,000 2.12 BILLION GALLONS 5,800,000 PLANTS INSTALLED 1,400,000 18.4 ACRES SEEDED TOTAL SEDIMENT REMOVED 2.76 MILLION CUBIC YARDS THE HUDSON DREDGING PROJECT 12 EPA’s White Paper Response to NOAA report, March 2016, Pg. 3 13 EPA’s White Paper Response to NOAA report, March 2016, Pg. 7 14 Proposed Second Five-Year Review Report for Hudson River PCBs Superfund Site, EPA, May 31, 2017, Pg. 58