Hydrologic analysis of an earthen embankment dam in southern Westchester County, New York

In 2001, the New York City Department of Environmental Protection installed 25 wells on the southern embankment of the Hillview Reservoir in Westchester County in an unsuccessful attempt to locate the source of a large seep (seep A) that began flowing continuously in 1999. In 2005, the U.S. Geological Survey began a cooperative study with the NYCDEP to characterize the hydrology of the local groundwater system and identify potential sources of seep A and other seeps on the embankment.

At least two groundwater-flow zones—one shallow and the other deep—overlie the bedrock at the Hillview Reservoir in southern Westchester County, New York. Analyses of slug tests of wells drilled into the southern embankment of the reservoir were used to determine the three-dimensional distribution of hydraulic conductivity of the embankment materials. The wells with the minimum and maximum hydraulic conductivity values are in the deep saturated zone on the southern embankment, where hydraulic conductivity ranges from 0.0012 to 2 feet per day. Hydraulic conductivity ranges from 0.0026 to 1 foot per day in the shallow saturated zone and from 0.021 to 0.27 foot per day in the toe of the embankment. A hydraulic conductivity of 0.016 foot per day was determined for one toe well partially screened in the crystalline-bedrock aquifer. In 2005, the U.S. Geological Survey began a cooperative study with New York City Department of Environmental Protection to characterize the local groundwater-flow system and identify potential sources of seeps on the southern embankment of the Hillview Reservoir in southern Westchester County, New York.

Long-term hydrologic data indicated that water levels trended downward in 29 of 41 sites, including the reservoir basin that was monitored during the 12-year study period; data from a National Weather Service precipitation gage at Central Park indicated annual precipitation also trended downward during the same 12-year period. Of the seven wells in which water levels trended upward during the study, two of the wells are on the west side of the southern embankment, proximal to a major water supply conduit, whereas the five remaining wells are screened in the toe. These data indicate an increasing hydrostatic pressure within the deep system and the toe of the dam, which could result in future seeps on the southern embankment near these wells.

Results of 11 suspended-sediment samples collected from seeps along the southern embankment at 234.1- and 221.6-feet elevation, and another drainage outflow point between 2007 and 2015 indicate a poor correlation between suspended-sediment concentration and discharge. From the flowing seep at 234.1 feet, suspended-sediment concentrations ranged from 1 milligram per liter at a flow of 2.6 gallons per minute (that is, 1 milligram per 0.26 gallons) during March 2008 to 16 milligrams per liter at 12 gallons per minute during July 2014. At about 12 gallons per minute discharge, suspended-sediment concentration from samples collected at that seep during different sampling events, ranged from 3 to 16 milligrams per liter. From the seep at 221.6 feet elevation, the suspended sediment concentration was 2 milligrams per liter at a discharge of 3.4 gallons per minute and 2 milligrams per liter at a discharge of 1.1 gallons per minute. Only one sample was collected at the drainage outflow point, for which the suspended sediment concentration was 2 milligrams per liter at a discharge of 2.4 gallons per minute.

Anomalously high-water levels were recorded in deep-system wells between June 5, 2013, and January 14, 2014. The period for the increase and the decrease back to more typical water-level elevations occurred rapidly during a 13-hour period in each instance. The sudden and rapid changes, in addition to the spatial distribution of magnitude of water-level response indicate that leaky water infrastructure was the source of recharge to the affected wells.

A major water supply conduit was drained for repairs between July 7 and 10, 2010. The seeps indicated an immediate response and a substantial hydraulic connection to the water supply conduit. Approximately 10.5 hours after the water supply conduit was drained, flow from a seep on the southern embankment decreased from about 20 gallons per minute to less than 1 gallon per minute. This seep is located at about the same elevation and within the vicinity of the water supply conduit. A travel-time of about 10.5 hours from the source to the seep at 234.1 feet elevation was estimated from the dewatering timeline. During the 3-month shutdown of the water supply conduit, the previously flowing seeps remained dry until precipitation resulted in discharge of about 0.7 gallon per minute at the higher elevation seep, indicating a minor contribution from precipitation to the total seepage discharge. Discharge from the seeps resumed almost immediately coincident with the refilling of the water supply conduit, supporting the hydraulic connection observations during the drainage stage. In addition, during the refilling of the water supply conduit on September 21, 2010, a new seep (I) was observed on the southern embankment. Discharge from this new seep remained relatively constant until it became inaccessible under construction stone from subsequent embankment repairs by the New York City Department of Environmental Protection. Precipitation after the refilling stage of the shutdown seemed to have induced a rise in water levels in the toe wells and an increase in discharge from the seep at 234.1 feet elevation. The post shutdown discharge was less than 12 gallons per minute, compared to a discharge of about 20 gallons per minute before the repairs. The lower discharge rate measured during the period of historically higher discharge rates for the fall season indicates that the repair of the major water supply conduit may have contributed to a reduced discharge from the seeps. There were no definitive responses to the shutdown in any of the wells near the major water supply conduit.

The more transmissive deep system of the southern embankment near the major water supply conduit and its associated infrastructure seems to be the preferential flow path for leaking infrastructure. The wells screened in this system showed a response during the deep system anomaly and have some of the highest hydraulic conductivities of the tested wells. All the seeps are in the elevation range of the deep system from approximately the crystalline bedrock surface around 200 feet elevation to the contact between the deep and shallow saturated zones of the reservoir at about 250 feet elevation.