Abstract: The New Jersey Meadowlands Commission has determined that remediation of the Kearny Marsh freshwater ecosystem is a high priority and has partnered with Rutgers University to achieve this goal. Human alterations affecting current hydrologic conditions include municipal stormwater inputs from the Town of Kearny, construction of railways and highways surrounding the marsh, creation of mosquito drainage ditches throughout the marsh, and channeling of marsh drainage to a partially clogged pipe in the northeast corner of the marsh. A bulkhead that is conveying Frank’s Creek stormwater between the Town of Kearny and the Passaic River has also been breached, allowing for water exchanges with the marsh. Because of the surrounding urban land use and the adjacent Keegan Landfill, negative impacts to the marsh were suspected as a result of groundwater and surface water interactions and due to stormwater runoff discharging into the marsh. To determine the amount and source of current contamination a hydrology study was undertaken during 2006 and 2007. Stormwater and groundwater samples were collected and analyzed for nutrients, polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), heavy metals, and volatile organic compounds (VOCs). The results showed elevated levels of heavy metals and in some locations PAHs and PCBs. Ammonia and phosphorus concentrations were elevated during all sampling events, indicating the potential for a high degree of future eutrophication of this ecosystem. Principal component analysis (PCA) and other analytical techniques (contaminant ratios, ANOVAs, and ‘fingerprinting’) were used to determine possible source(s) of elevated contaminant loadings. To analyze the various components of this complex system, a Storm Water Management Model (SWMM) surface water model and a Visual MODFLOW groundwater model were developed for Kearny Marsh, and these models were subsequently calibrated and verified. To project the hydrologic condition(s) of the marsh post-restoration, the models were run using three different precipitation scenarios (“dry”, “wet”, and “normal” conditions).