Introduction

In 2016, the U.S. Geological Survey (USGS), in cooperation with the New York State Department of Environmental Conservation (NYSDEC), began an assessment of the 35 miles (mi) of valley-fill aquifers within a 112-square-mile (mi²) study area in southern Otsego and northwestern Delaware Counties, New York. This area includes a 16-mi-long reach of the Susquehanna River valley and reaches of the Otego Creek, Charlotte Creek, and Schenevus Creek valleys that range from 2 to 7 mi in length. This study is a continuation of a cooperative USGS–NYSDEC aquifer mapping program that began in 1980. The published reports from this program have provided a foundation for wellhead protection efforts, water-resource management and planning decisions, and groundwater remediation in upstate New York (U.S. Geological Survey, undated).

Purpose and Scope

This report summarizes the hydrogeology of 35 linear miles of valley-fill aquifers in a 112-mi² area centered around the City of Oneonta in Otsego and Delaware Counties, N.Y. (fig. 1), and defines the distribution of thin and thick till in the surrounding upland watershed areas.

An interactive map plate (plate 1) was constructed with several layers of information, including the following:

• • Extent of the valley-fill aquifer system

• • Surficial geology

• • Bedrock surface elevations within the valleys

• • Aquifer types

• • Locations of hydrogeologic sections

• • Locations of wells, test holes, and passive seismic measurements

• • High-resolution light detection and ranging (lidar) hillshade imagery of land surface

• • 1:24,000-scale topographic base map

Data Sources and Methods

The surficial geologic mapping of the study area was completed by P. Jay Fleisher in 2016 and 2017, with some adjustments from new detailed lidar elevation data and imagery, data from the Soil Survey Geographic Database (U.S. Department of Agriculture, 2014), and recent well and test-boring data. The boundaries of the valley-fill aquifer system were delineated from the surficial geologic map and lidar imagery. The aquifer boundaries encompass contiguous stratified deposits of glacial and postglacial origin in the valleys, regardless of the degree of saturation because unsaturated deposits may act as conduits for recharge to confined aquifers. Areas containing a mixture of till and stratified deposits were also included. Well data used in this report are stored in the USGS National Water Information System (U.S. Geological Survey, 2020b). Much of the data is from Randall (1972). Additional data sources include well records from the New York State Department of Environmental Conservation (which administers the New York State Water Well Drillers Registration Program; unpub. data, 2018), well records from the New York State Department of Health's Oneonta office (unpub. data, 2018), SUNY Oneonta master’s theses (Morrow, 1989; Kucewicz, 1992), and test-boring data from the New York State Department of Transportation (unpub. data, 2018). Geospatial data for this study (shapefiles of surficial geology, dead-ice sinks, inferred ice tongue positions, within-valley bedrock surface elevation, stratified-drift aquifer extent, aquifer types, wells, and horizontal-to-vertical spectral ratio [HVSR] sites) are available in two USGS data releases (Heisig and Fleisher, 2022; Heisig and Izdebski, 2022).

An HVSR seismic survey (a type of passive seismic survey) can be used to estimate the thickness of unconsolidated sediment over bedrock (Nakamura, 1989). The HVSR seismic survey method uses a single three-component (x, y, and z directions) broadband seismometer to record ambient seismic noise. The ratio of the averaged horizontal (x, y) to vertical (z) frequency spectrum is used to determine the fundamental site resonance frequency, which can be interpreted by using regression equations to estimate the thickness of the sediment overlying bedrock (Lane and others, 2008).

Two regression equations were used to estimate the thickness of sediment over bedrock. These equations were derived from a total of 43 control-point measurements from various study areas in upstate New York (Heisig and others, 2021; Fisher and Keto, 2022; Heisig, 2022; Heisig and Izdebski, 2022). The control-point HVSR measurements were made at current or former well sites at which the depth to bedrock was known and unconsolidated sediment had been described in well logs. The HVSR records also must have shown single, well-defined resonance frequency peaks. Estimates of shear wave velocity (V_S) from these records were calculated by using the following equation:

V_S

= 4

fZ

(1)

where

f

is the resonance frequency (in hertz), and

Z

is depth to bedrock (in feet).

The calculations of V_S for all control sites yielded a wide range in V_S values that were divided into two distinct groups, low and high V_S, which corresponded to specific sediment types. The low-V_S group (V_S approximately [≈] 700 to 1,135 feet per second [ft/s]; median V_S ≈ 969 ft/s) corresponded to sites with mixed, mostly fine grained, stratified sediments between 50 and 275 ft thick. The high-V_S group (V_S ≈ 1,140 to 1,700 ft/s; median V_S ≈ 1,435 ft/s) corresponded to three sediment types: (1) predominantly sand or sand and gravel of any thickness, (2) till (limited data), and (3) mixed stratified sediments greater than 275 ft thick that were likely overridden by readvancing ice. The regression equations are power functions. The low-V_S equation (coefficient of determination [R²] = 0.962) is

Z

= 276.23

f^−1.213

,

(2)

and the High-V_S equation (R² = 0.993) is

Z

= 366.85

f^−1.038

,

(3)

where

Z

is the thickness of sediment over bedrock (in feet), and

f

is the measured resonance frequency (in hertz).

A plot of the control points labeled with USGS HVSR seismic survey and well identification numbers is shown in figure 2.

HVSR measurements were carried out at 71 sites in the study-area valleys to estimate the thickness of sediments overlying bedrock. The most distinct HVSR peaks and best estimates of sediment thickness were at sites where there was stratified drift overlying bedrock. Indistinct or absent HVSR peaks may indicate till overlying bedrock with little or no stratified drift thickness. Steeply sloping bedrock surfaces may also result in indistinct peaks. Raw and analyzed HVSR data, depth to rock estimates, and comments are provided in Heisig and Izdebski (2022).

Glaciation and the Distribution of Glacial and Postglacial Deposits

This section includes a discussion of glaciation and deglaciation in a regional context followed by a characterization of the framework of deposits within the study area. The distribution of glacial and postglacial deposits at land surface is depicted in plate 1. Subsurface glacial deposits are detailed in 13 hydrogeologic sections, including a longitudinal geologic section that covers the entire length of the Susquehanna River valley in the study area (figs. 3, 4A–M). The depth of glacial scouring and the thickness of valley-fill deposits can be inferred from a map of the bedrock elevation beneath the main valleys (fig. 5). The contoured data indicate that the bedrock underlying the Susquehanna River valley has been more deeply eroded than the bedrock underlying its tributary valleys, which can be considered hanging valleys.

Glaciation of the Upper Eastern Susquehanna Basin, With Emphasis on Regional Glacial Retreat

The following sections provide an overview of topographic influences on glacial retreat, environments of deposition, landforms and stratigraphy, and sediment sources and ice-contact lakes.

Timing of Glacial Retreat and the Influence of Regional Topography

The Late Glacial Maximum (LGM) of 26,000 years ago placed the Laurentide Ice Sheet as far south as present-day Long Island, N.Y. From there the ice front extended westward into northern New Jersey and southern Pennsylvania (Peteet and others, 2006; Oakley and Boothroyd, 2013). At this time, all of Otsego County was under approximately 3,000±1,600 ft of ice (Clark, 1992). Global climate change brought on the rapid retreat of the ice sheet. As the temperate ice sheet (Gustavson and Boothroyd, 1987; Ashley and others, 1991) adjusted to the underlying terrain, extensive valley ice tongues and upland reentrants formed that persisted throughout the glacial retreat (fig. 6). As such, the configuration of the ice front more closely resembled alpine conditions rather than the classic sweeping, broad lobes of a continental glacier. By approximately 18,000 years ago, the ice front had withdrawn to the Cassville-Cooperstown moraine in northern Otsego County, 1.8 mi south of Cooperstown (Krall, 1977; Ridge, 2003, 2004; fig. 7).

The Susquehanna River and its tributaries drain the eastern Appalachian Plateau Province of central New York State. Glacially enlarged north-south through valleys are joined by less-developed tributary through valleys from the northeast, separated by gently undulating uplands. Topographic relief of 650 to 1,000 ft accounts for the general appearance of the terrain. However, because the valleys typically contain stratified drift that commonly exceeds 300 ft in thickness, the bedrock relief within the eastern Susquehanna River Basin can approach 1,300 ft.

Environments of Deposition

The uplands and valleys were distinct areas of deposition during glaciation and deglaciation. As a result, upland deposits are largely unstratified and valley deposits are largely stratified.

Uplands

The uplands are mantled by lodgment till, which was deposited beneath the ice. Thicknesses are up to 30 ft on the stoss side (northeastern ice-facing side) of hills and ridges, with generally thicker till on the lee side (southwestern side) of topographic highs where till shadows (Coates, 1966) may exceed 160 ft (plate 1). Small lakes and ponds are commonly located on the down-glacier side of the uplands where glacial plucking excavated sandstone joint blocks.

Although the uplands are generally devoid of stratified drift, a few water-well logs indicate limited pockets of stratified drift in the headwaters of large tributaries. However, the number of available well logs is far too limited to assess the extent of the stratified deposits in the uplands.

Some second- and third-order tributaries in the uplands are anomalously large for the size of the basins they drain. Here, downstream channels contain underfit streams with limited discharge, even during periods of heavy runoff. These channels are interpreted to be sluiceways where high-discharge meltwater streams conveyed runoff from temporarily stable ice-front positions. Examples may be found on many upland divides.

Valleys

Valley-fill sediments consist mostly of lacustrine silt and fine sand that were largely deposited during glacial retreat. These deposits range in thickness from 100 to 200 ft in valleys tributary to the Susquehanna River valley. Sediment within the Susquehanna River valley proper attains a thickness of 400 ft at Oneonta (OG 410) and a thickness exceeding 440 ft 3.5 mi downvalley from Oneonta (OG2392), all presumably the result of the deep scouring that occurred during the convergence of glacial ice from the three tributary valleys with that in the Susquehanna River valley (fig. 4A–D, plate 1). Valley-fill sediments within the main Susquehanna River valley thin farther south in Chenango and Broome Counties to about 200 ft in thickness (Heisig, 2013).

The morphology and orientation of these valleys strongly influenced the nature of deposition during glacial retreat. The ice-tongue model depicting short lobes accounts for the distribution of erratic-rich gravel in valleys (Moss and Ritter, 1962) but has been replaced here by a model that indicates the presence of longer valley tongues that resemble an alpine setting. The adjustment of temperate ice to a terrain of this relief resulted in a configuration of ice-front retreat that included ice tongues 6 to 12 mi long (fig. 6; Fleisher, 1985, 2003).

As MacNish and Randall (1982) proposed, a variety of depositional environments existed as determined by the rate of ice flow (slow versus fast) and the manner of ice retreat (backwasting versus downwasting). Fleisher (2003) distinguished between (1) ice tongues in through valleys that served as troughs for prolonged active ice flow and (2) those in non-through valleys where thinning over upland divides resulted in the stagnation of valley ice (fig. 8). Valleys oriented parallel to the ice front (perpendicular to the direction of ice movement) would preclude ice-tongue development entirely, thereby further altering depositional conditions. Therefore, it is necessary to distinguish through valleys from non-through valleys, a distinction that is based on the topographic expression of their headward divide (figs. 7, 8; W.M. Davis, in Tarr, 1905).

Through valleys are bedrock valleys that cross regional topographic drainage divides. These bedrock troughs are open to the north and effectively served to channel the southward-flowing glacial ice. Here, confining valley walls do not converge headward, but instead diverge because of the absence of an upland bedrock divide. As such, postglacial drainage flowed to the south from a valley-floor divide. The open nature of through valleys provided a continuous connection to the overriding ice sheet that fed the valley ice tongues. Local examples include the valleys of the Susquehanna River and Cherry Valley Creek (fig. 7).

Conversely, non-through valleys are not open troughs; instead, they occupy drainage basins with headward divides in upland terrain. A prime example is the valley of Otego Creek (fig. 7). Within through valleys, ice tongues remained dynamically and hydrologically connected to the retreating ice sheet, whereas ice flow within non-through valleys gradually diminished as the ice sheet thinned over the divides. Eventually, the ice reached a critical thickness below which active ice flow ceased (fig. 8). The net result was the disconnecting of the entire downvalley ice tongue from the active ice and meltwater, which in turn led to stagnation and downwasting of the tongue. Thus, meltwater flow after stagnation was limited to the water equivalent of the stagnant ice, which was significantly less than in the through valleys that continued to convey meltwater from the retreating ice sheet.

Similar conditions existed within valleys that were oriented transverse to the Laurentide Ice Sheet’s flow. Here, stagnation was widespread and depositional conditions were similar to those in non-through valleys. Local examples include the valleys of Charlotte and Schenevus Creeks (plate 1, figs. 3, 4L–M, 7).

Landforms and Stratigraphy

Different ice-flow conditions in conjunction with their respective volumes of meltwater discharge resulted in distinct landform assemblages in each of the two valley types. Kame moraines and outwash surfaces of various types are common to both types of valleys. Kame moraines are typically expressed by a hummocky topography with surface relief on the order of about 50 to 75 ft. They stand in conspicuous relief above all other valley-floor landforms and mark the former position of an active, stable ice front. Local exposures indicate that kame moraines consist mainly of poor to well-sorted sand and gravel of fluvial origin with minor mudflow units, all of which were deformed by small-scale collapses. Of particular interest to interpretations of glacial history are data from water wells drilled in kame moraines near Portlandville and West Davenport (fig. 1; plate 1). These data indicate that the moraines are underlain by thick lacustrine sediments, which suggests deposition at an ice front that advanced into or across an older lakebed (fig. 3I). Also present at Portlandville are a pair of eskers connected to the outwash immediately south of the moraine that also sit atop lacustrine deposits.

In the study area, broad outwash surfaces typically extend downvalley from the kame moraines. These deposits consist of well-sorted and distinctly stratified outwash gravels that typically display massive foreset beds indicative of deltaic deposition, as well as associated topset and toeset beds. As indicated by MacNish and Randall (1982), the valley-wide outwash surfaces indicate a slow, active retreat of glacial ice. More commonly, landforms suggest that the retreat was rapid, which limited the growth of deltas to lateral valley positions. When the downvalley dam that impounded the lake eventually failed and the lake emptied, these lateral deltas were preserved in the modern landscape as terraces above a lacustrine plain. Although often referred to in a colloquial sense as kame terraces, they are in fact deltaic outwash terraces.

Eskers, outwash terraces graded to a variety of local base levels, and isolated kame fields lacking an associated downvalley outwash deposit are generally more common features of non-through valleys. In these settings, eskers indicate that the subglacial-stream accumulation of sediment remained unaltered by active ice flow, thus reflecting ice stagnation. The eskers, in conjunction with the isolated kame fields not associated with a downvalley outwash deposit, suggest a downwasting ice front. Furthermore, discontinuous patches of low-relief hummocky topography, formed by clusters of kames and kettles, lack any correlation with glacial landforms in adjacent valleys and suggest downwasting of glacial ice specific to that valley. Additionally, the non-through valleys contain discontinuous outwash terraces that grade to former proglacial lakes. As a group, the eskers, isolated kames and kettles (topographic hummocks and depressions), and isolated outwash segments characterize these non-through valleys.

One feature common to both valley types in the study area, although more commonly present in non-through valleys, is the dead-ice sink (plate 1; Fleisher, 1986a). Dead-ice sinks formed when the retreating ice tongue abandoned massive ice blocks that were large enough to occupy the entire valley floor (Mulholland, 1982). With further retreat and ablation of the ice block, the outwash accumulated around and, to some extent, over the ice block. With continued ablation, the area that the dead-ice block occupied developed into a sink for sedimentation, and a massive kettle began to form in which sediment accumulated in a semicontinuously subsiding depositional environment. The net result was a localized sequence of various sediment types that lacked any lateral continuity with adjacent glacial deposits.

These conditions prevailed until the ice front eventually retreated from the Appalachian Plateau Province and meltwater was redirected eastward into the Mohawk River valley (fig. 7). Consequently, the only ongoing fluvial discharge into the Susquehanna River valley was derived from precipitation within the watershed, which prevented further development of outwash sand-and-gravel deposits. If the meltwater had not been diverted to the Mohawk River valley and its discharge down the Susquehanna River valley had continued, these dead-ice sinks eventually would have filled with sediment and would have little or no topographic expression today.

Sediment Sources and Ice-Contact Lakes

Virtually all field evidence and available water-well data indicate that the active ice tongues retreated from ice-contact lakes. The evidence of these former lakes is the thick deposits of clay, silt, and sand that have accumulated in most of the major southward-draining valleys in the study area (see fig. 4 sections, except fig. 4L–M). These deposits are common throughout the upper Susquehanna River valley (fig. 3A, for example), thus suggesting a late glacial system of lakes, each forming behind a kame moraine and large-scale kame-and-kettle topography at the confluences of the southward-draining streams with major tributaries. Driller’s logs of wells on the valley floors reveal lenses and layers of glaciofluvial sand and gravel interstratified with thick deposits of lacustrine clay, silt, and sand, thus confirming the concurrent deposition of coarser sediments from prograding lateral deltas. The maximum thickness of the lacustrine sediments is generally 100 to 200 ft but can exceed 300 ft in the largest valleys.

The vast majority of glaciofluvial sand and gravel was derived from upvalley sources where sandstone and limestone bedrock is predominant. The clastic sediment load of meltwater streams, therefore, consisted mainly of these rock types, with only a small percentage of gravel derived from crystalline rock of the Adirondack Mountains and from a small exposure in the Mohawk River valley floor (Fleisher, 1993, 2003).

During the deglaciation, the ice-free uplands fed tributaries carrying alluvium to valley-side fans and deltas. However, these deposits are anomalously small compared to the thick deposits of lake sediments commonly present within these valleys. One explanation for the disparity may be found in a study of the present-day rate of sedimentation by the highly turbid water in ice-contact lakes at the margins of the Bering Glacier in Alaska. There, rates of accumulation at the ice front reached 23 to 31 inches per year (Fleisher and others, 2003), which is significantly greater than the sedimentation rates for varved lacustrine deposits (for example 0.25 to 8 inches per year; Ridge and others, 1990).

Such high rates of sedimentation may be attributed to exceptionally turbid water in which the excessive particle concentrations force the physical contact of suspended silt and clay-size particles, resulting in freshwater flocculation and rapid settling of the sediment (Hubbard, 2002). This theory applies to the sediment sources within the Susquehanna River valley, where upvalley Devonian shales and siltstones were overridden and eroded by glacial ice to produce clay-sized sediment particles (Rickard and Zenger, 1964). Subglacial abrasion and hydraulic scouring likely produced highly turbid water entering the lacustrine environment by englacial and subglacial conduits. Older lake sediments across which the Laurentide Ice Sheet may have moved constitute an additional potential source of clay-sized particles. The net result of these processes was a thick accumulation of fine lacustrine sediments in a relatively short period of time—so short as to limit the size of tributary deltas, as is observed in the study area.

Valley-Fill Aquifer Types and Their Groundwater-Resource Potential in the Study Area

Sand-and-gravel aquifers in the valleys have the greatest potential for municipal water supply in the study area, particularly where they are in hydraulic connection with the Susquehanna River or the larger creeks (Charlotte, Otego, and Schenevus Creeks). Under such conditions, sustainable yields as high as 750 gal/min have been achieved through induced infiltration. Fractured bedrock aquifers underlie the entire study area and are the most common source of domestic supply, but the highest reported bedrock well yields are only in the range of tens of gallons per minute. Fractured bedrock is used as a water source in the valleys where sand-and-gravel deposits are absent, unsaturated, or too silty.

The sand-and-gravel aquifers in the study area can be classified as unconfined, semiconfined, or confined, as shown in figure 9. Unconfined and semiconfined or locally confined aquifers are open to the atmosphere to some degree. They are combined into one class in figure 9 because they are likely transitional and because there is insufficient subsurface data to differentiate them in many areas.

The unconfined valley-fill aquifers have limited extents (fig. 9) despite widespread outwash sand-and-gravel and alluvial deposits because saturated thicknesses of terrace landforms and alluvium that make up these aquifer segments are typically thin and are underlain by fine-grained, lacustrine deposits. Many areas of ice-contact deposits are classified as potential unconfined aquifers if they are permeable and extend far enough below the water table to have sufficient saturated thickness. However, limited well-log data indicate that ice-contact deposits that span the Otego Creek valley (plate 1) and northernmost Susquehanna River valley (fig. 4I) are underlain by lacustrine deposits that limit the saturated thickness of these potential aquifers. Because these deposits do not appear to extend very far below the water table, they are not mapped as unconfined aquifers. The delineation of unconfined or semiconfined aquifers is mostly from well logs, and their extents are therefore viewed as conservative estimates. Domestic supply may be possible from thin, permeable saturated deposits; however, there are no recently drilled wells, only older dug wells, completed in this type of setting.

Unconfined and semiconfined aquifers are most productive in the following areas, especially where they are in hydraulic connection with the Susquehanna River or creeks in the other main valleys: (1) along valley sides where ice-contact deposits (usually in the form of kame or deltaic terraces) are well sorted (have a low silt content) and extend below the water table, and (2) midvalley in the Oneonta area, where outwash or ice-contact deposits appear to have collapsed as deep as 90 ft below the present floodplain level, perhaps as buried ice beneath it melted (figs. 4A, D; 9). A well (OG 399) completed in this semiconfined aquifer is the most productive in the area (pump tested as high as 1,000 gal/min) and is likely in hydraulic connection with the Susquehanna River. As shown in other hydrogeologic sections in the valley (fig. 4A, C, F, H), gravel and clay or gravel and silt (with or without sand) above or within lacustrine deposits appear to be at least partly associated with ice meltout at dead-ice sink locations (plate 1) and likely have limited to poor water-resource potential.

Ice-contact deposits (a kame moraine) make up much of the glacial fill in the Charlotte Creek valley. These deposits consist of a variable mixture of collapsed and chaotic layers or bodies of till, poorly stratified and crudely sorted gravels, and intervals of sorted sand, gravel, and silt. Materials encountered in adjacent wells can show extreme variability from (locally) confined or semiconfined to unconfined sediments. Unless identified otherwise from well logs, the aquifer in the Charlotte Creek valley is largely identified as semiconfined in figure 9. Some mobile-home park and commercial wells are completed in sand and gravel with open-ended casing, but the majority of these wells tap bedrock, presumably because the stratified sediments are poorly sorted. Yields in all cases are modest—on the order of tens of gallons per minute—but could likely be improved with screened well completions.

Confined aquifers are isolated from the atmosphere primarily by widespread fine-grained lacustrine deposits; till acts as a confining unit in limited areas, particularly within the Charlotte Creek valley (figs. 4A–M; 9; plate 1 [see well symbols that indicate the presence of confined sand and gravel]). Lacustrine confining units are widespread because proglacial lakes were present in most valley reaches in the study area except for a section of Charlotte Creek (fig. 9). There is insufficient well-log data within the short section of Schenevus Creek to delineate the extent of lacustrine deposits, but they are inferred to be present beneath the surficial outwash deposits. Confining units consist of fine sand and silt to clay. Beneath the confining units, subaqueous fan, esker, and deep deltaic terrace deposits (Fleisher, 1993) form discontinuous and variably sorted sand-and-gravel deposits that constitute the confined aquifer. The extent of the confined aquifer shown in figure 9 is considered a maximum extent of where such aquifers can exist, but the extent does not imply that they are present everywhere (see hydrogeologic sections fig. 4A–M, particularly fig. 4A, for examples of the occurrence of confined aquifer material). Saturated aquifer materials range in thickness from a few feet to over 100 ft. Favorable areas for the occurrence of productive confined aquifers include (1) junctions of large valleys; (2) junctions of large valleys with major upland tributary valleys, where ice-contact and alluvial-fan deposits can receive recharge through streambed infiltration of tributary flow; and (3) other valley-wall areas, where permeable ice-contact units are exposed at the land surface and extend beneath the lacustrine deposits, thereby allowing recharge from direct precipitation and runoff from adjacent valley walls. The reported yields of properly screened wells in such favorable locations range from 75 to 750 gal/min.

Groundwater Flow and Aquifer Susceptibility to Contamination

Under natural conditions, precipitation that falls within the Susquehanna River Basin is the ultimate source of groundwater. Within the study area, groundwater inflow from upgradient valley aquifers is an additional source of groundwater (fig. 10). The portion of precipitation that infiltrates the unsaturated zone and becomes part of the groundwater flow system is known as recharge. In areas with permeable soils, recharge water infiltrates directly to the water table. On hillsides underlain by thick till, little if any water infiltrates deeply to join an aquifer, and most remains in the shallow subsurface until it discharges to upland tributary streams or moves downslope toward the valley aquifers. Dry-weather tributary flow, known as base flow, and storm runoff constitute the flow in tributary streams. As all but the largest tributaries enter the main valleys, their underlying channel materials transition from thin alluvium over till or bedrock to alluvial fans that are moderately permeable and typically less saturated near the valley wall. If the stream channel is above the local water table, a portion of streamflow will infiltrate into the alluvial fan and become a source of recharge to the underlying unconfined aquifer (fig. 10; Randall, 1978). Overland storm runoff and subsurface runoff from valley walls infiltrates as recharge where the permeable valley-fill deposits lie adjacent to the valley wall. If the valley-fill deposits are not permeable, this hillslope water will discharge at springs on the valley floor. Within the main valley, the gradients of natural groundwater flow system are toward the lowest point of the landscape, which is the channel of the main trunk stream (fig. 10). Shallow groundwater flows across the valley to this drain with a component of downgradient (downvalley) movement, which is partly dependent on the slope of the valley floor and the permeability of the valley fill. Groundwater that is deeper within the hillside will take a deeper and longer flow path within the valley fill and eventually flow vertically upward to the stream in the absence of a confining unit. The presence of a confining unit will limit recharge to and groundwater flow within the deeper valley-fill aquifers. Under natural conditions, groundwater flow within such confined aquifers may be a slow downvalley flow toward some distant discharge point because any upward flow is nearly or completely stopped by the confining unit. In some sedimentary terranes where deep, confined groundwater flow is restricted and the water is very old, high mineral content in groundwater can be an issue; however, this problem has not been reported in the study area by the New York State Department of Health. Strongly reducing chemical environments in deep groundwater may also result in naturally high concentrations of manganese, iron, hydrogen sulfide, or methane, but their presence here was not evaluated.

Considerations for protection of groundwater resources vary with aquifer type and hydrogeologic setting. Upland fractured bedrock aquifers are best protected from contamination if activities having a potential for contamination are confined to lower slopes where till cover over bedrock is thick and continuous (greater than at least 40 ft of till; see plate 1, which classifies thick and thin till). Wells completed in upland fractured bedrock are most susceptible to contamination from upslope activities where till is thin or absent.

Unconfined sand and gravel aquifers in valleys are susceptible to contamination from any source of recharge that becomes contaminated by human activities: direct recharge from above, runoff and infiltration from adjacent hillsides, infiltration from tributary streams as they cross alluvial fans in main valleys, and induced infiltration of river water into aquifers caused by groundwater withdrawals. If the valley-fill aquifer includes an alluvial fan through which a tributary’s streamflow infiltrates, activities that generate contaminants not only on the alluvial fan but also in the tributary’s upland watershed (several miles away) may potentially affect an aquifer’s water quality. More direct contamination may occur if groundwater withdrawals from an unconfined aquifer induce the infiltration of contaminated water from the main valley’s river or creek into the aquifer. The contamination of streams and rivers may result from overland runoff from extreme storm events, chemical spills, or wastewater discharges.

Confined aquifers have some protection against groundwater contamination from overlying activities on the land surface because of the widespread lacustrine (and locally, till) units that confine them. Windows of hydraulic connection between valley-side unconfined aquifers and adjacent or underlying fractured-bedrock aquifers are the most likely pathways of recharge to deep confined aquifers. Groundwater in confined aquifers may be replenished (and contaminated) by the same sources of recharge as the unconfined aquifers along the valley sides, although induced infiltration from rivers is unlikely unless the river has impinged upon the valley wall.

In general, protection of valley-fill aquifers can be accomplished by (1) limiting activities that could introduce contaminants into the subsurface in areas of permeable sediments, (2) creating natural buffers around stream courses and limiting nearby activities that could generate contaminated runoff in upland watersheds of the tributary streams that cross the aquifers, and (3) having lines of communication to alert water suppliers in the event of contaminant spills on the land surface or in the streams.

Summary

The valley-fill aquifers within a 112-square-mile area surrounding the City of Oneonta were evaluated by using available well logs, surficial geologic mapping, and horizontal-to-vertical spectral ratio (HVSR) seismic surveys. The study area encompasses the confluence of several large valleys—those of the Susquehanna River, Otego Creek, Charlotte Creek, and Schenevus Creek. The geographic information system (GIS) map layers produced for this project included the locations of wells and test holes, elevation of the bedrock surface within the valleys, surficial geologic units, locations of dead-ice sinks, locations of past ice-tongue margins, and aquifer types and extents. To characterize the subsurface, 13 transverse hydrogeologic sections and one longitudinal geologic section covering the length of the Susquehanna River valley in the study area were constructed. The deglaciation and resulting geologic units, aquifer occurrence and extent, and an overview of local groundwater hydrology and aquifer protection are discussed.

Glacial deposits blanket much of the study area, ranging from a thin veneer or exposed bedrock on hilltops and northeast-facing slopes to thicknesses as great as 440 feet in the most downgradient reach of the Susquehanna River valley. The bedrock surface in the Susquehanna River valley is more deeply eroded than in its tributary valleys. Upland glacial deposits consist of mostly till with localized areas of alluvium or wetland deposits within drainage courses. Glacial deposits in the valleys are largely stratified; lacustrine fine sand, silt, and clay deposits are the most widespread and thickest, but they are punctuated by accumulations of ice-contact or outwash sand-and-gravel deposits where the retreating ice paused or where large, ice-free upland drainages provided a source of sediment. Ice-contact material, which contains variable silt content, is also present along some lower valley walls and commonly beneath lacustrine deposits.

The orientation of the valleys relative to the ice flow and the nature of the headward divides of the valleys (through or non-through valleys) affected the mode of deglaciation. Through valleys, especially those aligned with ice flow such as the Susquehanna River valley, conveyed more meltwater for a longer duration than non-through valleys, such as Otego Creek. The Charlotte Creek valley was oriented perpendicular to ice movement and was thus choked with an extensive kame moraine that resulted in a limited input of meltwater which, in turn, resulted in limited or absent lacustrine sediments. The melting of stagnant, detached blocks of ice in non-through valleys and in sinuous reaches of the Susquehanna River valley occurred after sediment-laden meltwater from upvalley ice tongues was diverted elsewhere. Dead-ice sinks formed as the ice blocks melted out of surrounding outwash deposits and the resulting depressions were partly filled in with postglacial alluvium and organic deposits.

Deglaciation produced large amounts of outwash in the Susquehanna River valley, with the upvalley retreat of glacial ice now marked by a number of outwash or ice-contact heads. Ice-contact deposits are the dominant stratified glacial deposit in the other valleys, representing downwasting ice with less meltwater deposition of outwash.

The vertical stratification of glacial deposits in most valleys is typically: (1) early deglacial sand and gravel deposits that are variable in spatial extent, sorting, and thickness overlain by (2) mid-deglacial fine-grained lacustrine deposits, followed by (3) late deglacial outwash or alluvial deposits (Randall, 2001). Any one deposit may be missing within a given area, but deposits out of sequence (for example, ice-contact overlying lacustrine deposits) in the study area imply that a local readvance of the ice occurred.

Both fractured-bedrock and sand-and-gravel aquifers are used for water supply in the study area. Fractured bedrock is essentially the sole source of water in upland areas, whereas valley water sources include both fractured bedrock and sand-and-gravel aquifers. Where glacial deposits are poorly permeable, fractured bedrock is utilized. Sand-and-gravel aquifers exist under both unconfined and confined conditions, and they are most commonly used for public water supply because their water storage capacity and transmissivity are usually far greater than those of fractured bedrock. Yields of production wells completed in unconfined sand-and-gravel aquifers may be greatly enhanced where well-field withdrawals induce infiltration from nearby large creeks or rivers. Beneath lacustrine deposits, confined sand-and-gravel deposits are common and constitute the most widespread sand-and-gravel aquifer, although its permeability varies greatly. Ice-contact deposits present at the land surface along valley walls may act as conduits of recharge to both unconfined and confined hydraulically connected aquifers. Ice-contact deposits, sometimes intermixed with till, that span the valley’s width are typically continuous down to bedrock and thus may function as both aquifers and as conduits of recharge to confined aquifers. However, in this area, limited subsurface data indicate that ice-contact deposits overlie lacustrine deposits at several locations, which indicates that this assumption is not necessarily correct and that subsurface verification is needed.

Groundwater in the study area is derived from the infiltration of precipitation at the land surface that reaches the zone of saturation. This water (recharge) replenishes groundwater-flow systems in upland and valley areas because groundwater constantly flows from areas of higher hydraulic head to areas of lower hydraulic head. On a local scale, areas of lower hydraulic head are low points on the landscape where groundwater can discharge, such as streams and rivers, which act as groundwater system drains. Unconfined sand-and-gravel or fractured bedrock aquifers can directly receive recharge from precipitation, but this connection also makes them susceptible to contamination from activities on the land surface. In areas where aquifers are confined by relatively impermeable surficial or subsurface material, such as thick till in the uplands or lacustrine deposits in the valleys, little direct recharge occurs, and the aquifers are mostly protected from contamination from land-surface activities in the vicinity of a well. However, land-surface activities located on the recharge areas of confined aquifers, such as permeable deposits along valley walls or fractured bedrock with little till cover, may also result in contamination of hydraulically connected confined aquifers.