From the late 1940's until 1977 two General Electric Plants discharged 200,000 - 1.3 million pounds of polychlorinated biphenyls (PCBs) into the upper Hudson River. Field studies and detailed modeling efforts indicate that PCB release from sediments under realistic mixing conditions determines the efficiency of both 'natural recovery' and proposed dredging operations. In this study, Hudson River sediment was resuspended into clean water in large mesocosms. The desorption rates of individual PCB congeners were determined by measuring dissolved PCB concentrations using solid-phase microextraction.

Immediately following the initiation of resuspension, large particles with an average median diameter of 140 ± 14 mm were lifted into the water column.  Dissolved PCBs rose rapidly and after two hours of resuspension 6 to 38% of the PCBs in the water column were in the dissolved phase.  Rate constants for this rapid release ranged from 0.04 to 0.34 hour-1 and decreased significantly as log Kow of the PCBs increased.  Both the total suspended solids concentration and dissolved PCBs reached steady state in 24 hours.  At steady state the flocs volume median diameter averaged 112 ± 3 mm, porosity averaged 0.90 ± 0.02, and 15-50% of the resuspended PCBs were dissolved.  The PCB concentration on resuspended particles was an average of two times greater than the bulk sediment PCB concentration and 8% of the resuspended mass did not settle after twenty hours without mixing.  At steady state the particle-water PCB partition coefficients were similar to values measured in the Hudson River and constant across the range of congeners examined.  With only one-day quiescence between resuspension events the percent of dissolved PCBs at steady state decreased significantly from the first to the third resuspension event (p = 0.02).  When quiescent time was increased to four days, there was no change in the percent dissolved PCBs at steady state for the low molecular weight congeners (Log Kow ≤ 5.85, p = 0.45).  This analysis suggests there was a large release of PCBs from particles when they were initially resuspended; however, chronic resuspension resulted in less PCB release per event due to the slow recharge of a labile pool.