Purpose and Scope

This report summarizes the hydrogeology of sand and gravel aquifers (primarily the valley-fill aquifer system) in the 117-mi² Owasco Inlet watershed in Tompkins and Cayuga Counties, New York (fig. 1). An interactive map plate was constructed with layers of information:

Figures depict the study area extent, upland areas of thick glacial deposits, moraines and other glacial landforms, hydrogeologic sections, sand and gravel aquifer type and distribution within the Owasco Inlet watershed, and a conceptual block diagram of groundwater flow within the hydrogeologic framework. This report summarizes the glacial geology, hydrogeologic framework of glacial deposits within the valley areas, sand and gravel aquifer types and distribution, and municipal groundwater resource use as of 2017. Two data releases accompany the report: the first includes geospatial data used in the map plate and map figures (Heisig, 2023a) and the second includes the HVSR seismic data (Heisig, 2023b).

Previous Work

Ice retreat and surficial geology within the study area have been interpreted in detail by Carney (1909) in his Moravia 15-minute quadrangle map. Tarr, in Williams and others (1909), mapped and described glacial deposits and their origin in the adjacent Cayuga Inlet valley. Fairchild (1934) provided a comprehensive interpretation of deglacial events in western New York (with emphasis on the Cayuga Lake valley) including the sequence of regional proglacial lakes. Denny and Lyford (1963, plate 3) depicted Fairchild’s Valley Heads ice extent and deposits across the region. Mullins and others (1996) carried out seismic reflection surveys of Owasco Lake and seven other Finger Lakes and interpreted thickness of sediment fill and the deposits therein. Randall (2001) conceptualized and described the hydrogeologic framework of glacial aquifers within the local hydrophysiographic region as part of a larger report on the glaciated northeastern United States. Bird and Kozlowski (2014) and Bird (2014) used lidar elevation data to refine the extent of proglacial lakes described by Fairchild in the Cayuga County part of the study area. Bird and Kozlowski (2016a, 2016b) produced surficial geology maps of the Moravia and Sempronius 7.5-minute quadrangles. Most recently, Karig and Miller (2020) have documented features in the Cayuga Lake basin that imply northward drainage into the ice after the Brooktondale Readvance (a post Valley Heads readvance), which is counter to previous interpretations.

Data Sources and Methods

Surficial geology, subsurface deposits, and aquifer boundaries were based on well-log and test-boring data, the digital Soil-Survey Geographic database (SSURGO) (U.S. Department of Agriculture [USDA], 2008), lidar imagery, previously compiled geologic maps and reports, and field visits. The aquifer system boundary or aquifer extent is the extent of contiguous stratified-drift deposits within valleys, including mostly unsaturated deposits on adjacent hillside areas. Upland confined sand and gravel aquifers are limited in extent and potential productivity and are not included in this definition.

Well and test boring data used in this report are stored in the USGS National Water Information System (NWIS; USGS, 2022). A “CY” in the site name of the well or test boring indicates the well is located in Cayuga County, and a “TM” is for Tompkins County. Outside sources of well and test-boring records include the NYSDEC (which administers the New York State Water Well Drillers Registration Program), the New York State Department of Transportation, the Cayuga and Tompkins County health departments, and village or town records. All well and test boring data used in this report are compiled in Heisig (2023a).

The HVSR seismic soundings were completed at 31 sites to estimate the thickness of sediments overlying bedrock in the Owasco Inlet valley (Heisig, 2023b). The passive seismic method uses a single broadband three-component (x, y, and z directions) 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 (f), which can be interpreted using a regression equation to estimate sediment thickness overlying bedrock (Lane and others, 2008).

At present (2021), two regression equations (Heisig and Fleisher, 2022) are used to estimate sediment thickness over bedrock. The first equation is for high shear-wave velocities in all thicknesses of sand and gravel or till, and mixed predominantly fine-grained stratified sediments greater than or equal to 275 ft thick:

and the second equation is for lower shear-wave velocities in mixed, predominantly fine-grained stratified sediments less than 275 ft thick: where The equation used for each measurement is indicated in Heisig (2023b).

Regional Overview

Most glacial deposits within the Finger Lakes region are derived from two primary Laurentide ice sheet advances and retreats during the late Wisconsin substage at the end of the Pleistocene Epoch. These include (1) the late Wisconsin last glacial maximum ice advance that extended as far south as Pennsylvania (Nissouri Stade, about 25,000 years before present [B.P.]) (Corbett and others, 2017; Stokes, 2017); (2) the Erie Interstade (retreat), peaking about 18,000 years B.P. (Dyke, 2004); and (3) the Port Bruce Stade, marked by the Valley Heads Readvance (Fairchild, 1932; detailed in Denny and Lyford, 1963, plate 3), which peaked about 17,000 calendar years ago, as extrapolated by Karig and Isacks (2019) from western New York (Muller and Calkin, 1993) and the Mohawk valley of east-central New York (Ridge, 2003). This readvance deposited moraines mostly in the form of outwash heads in valleys across much of western New York. The Owasco watershed was completely covered during both the last glacial maximum and the Valley Heads Readvance, and most surficial deposits were deposited during glacial advance to, and retreat from, the Valley Heads ice margin. Minor subsequent readvances, such as the Brooktondale (Karig, 2015), likely altered surficial geology in the southernmost part of the study area (Karig and Miller, 2020).

The Finger Lakes region is classified as part of the “northern rim of the Appalachian Plateau” hydrophysiographic region defined by Randall (2001), a region of high relief that sloped toward the ice sheet. This region has smooth uplands punctuated by deep Finger Lake valleys. Steep slopes upvalley (south) from deeply scoured Finger Lake troughs were an impediment to ice movement and to the free southward drainage of meltwater. Valley fill in this region is characterized by a predominance of thick fine-grained lacustrine deposits with or without intervening fine-grained till (or diamicton) units and a lack of surficial outwash deposits (Randall, 2001).

As the ice margin advanced or retreated through the Finger Lake troughs, the natural northward drainage was blocked, such that meltwater ponded in a series of proglacial lakes with initial outlets to the south, controlled by spillways at through-valley drainage divides at the heads of the Seneca and Cayuga valleys (Fairchild, 1934; Muller and Prest, 1985; Bird and Kozlowski, 2014; Bird, 2014). Karig (2015) and Karig and Miller (2020) have questioned that the ice margin acted as a dam after the Brooktondale Readvance (approximately [~] 16,300 calendar years ago) and suspect drainage into the ice, citing a lack of outlet channels at within-valley divides of some valleys and indications of northward drainage within the Six Mile Creek and Cayuga Inlet valleys.

It is generally accepted that lake levels subsequently dropped as northward ice retreat exposed isostatically depressed lower-elevation divides. As lake levels declined, the lower base levels caused stream incision and erosion of recently exposed glacial deposits with subsequent deposition in shallow residual lakes or exposed valley floors. With retreat of the ice margin, subsequent isostatic rebound has been greatest to the northeast (Lewis and others, 2005). Since the Finger Lakes drain northward, this greater isostatic rebound of the northern part of the Finger Lake basins has subsequently acted like a rising dam and increased the Finger Lake water levels, resulting in fine-grained postglacial deposition at the south end of the Finger Lake troughs (including the Owasco Lake trough) as described by Tarr (in Williams and others, 1909, p. 32) at Cayuga Lake.

Valley Heads Readvance and Retreat in the Owasco Watershed—Glacial and Postglacial Deposits

The ice that moved across the study area during the Valley Heads Readvance was part of the eastern edge of the Cayuga lobe (Fairchild, 1932; Denny and Lyford, 1963, plate 3; Karig and Miller, 2020, their fig. 2). The direction of ice movement, indicated by streamlined till (mega flutes; Clark, 1993) in upland areas (fig. 2), was initially parallel to southern Owasco Lake and the Owasco Inlet valley nearly as far south as Moravia, but shifted progressively more eastward until flow was almost due east at the southern edge of the study area (Freeville). Ice from the Cayuga lobe thus raked across much of the Owasco Inlet valley. A smaller number of less pronounced mega flutes continue eastward to the Fall Creek valley, smoothing over what appear to be ice-contact deposits. The Valley Heads (Fairchild, 1932) Readvance completely covered the Owasco Inlet watershed (fig. 2 of Karig and Miller, 2020) and extended several miles south and northwest of Freeville (fig. 2).

As Valley Heads ice thinned and the southeastern upland divide was exposed, a brief period of meltwater ponding occurred because of surrounding, stagnant ice, and eventual flow to the east likely took place across several notched divides (fig. 2): at 1,424-ft elevation (floored with coarse stratified sediments) east of Groton, an adjacent one immediately south at 1,430 ft, and less developed ones at 1,416 ft and 1,357 ft farther south, east and a little north of Peruville. As the ice continued to thin, it appears meltwater was conveyed southward along the east side of the watershed as indicated by scattered ice-contact deposits, ice-marginal drainage channels, and washed till (plate 1, figs. 7 and 8). The southeasternmost and lowest (~1,210 ft) drainage outlet with stratified material is the gap between the bedrock ridge of the southeastern divide and an isolated hill south of it (at the west end of section B–B′, on plate 1) at the junction with the broad Fall Creek valley, just northeast of Freeville (plate 1, fig. 2). It has the form of a small ice position (concave on ice side, complete with spring, transitioning to kame terrace or moraine that abuts extensive hummocky ice-contact deposits [a stagnant ice setting]) in the Fall Creek valley (plate 1). Carney (1909, p. 358) stated that the eastern side of the Owasco Inlet valley “is in general uniformly more gravelly” than on the west side and that the drift was amassed on that side by ice marginal drainage. Soil surveys and well logs support this observation (plate 1, figs. 7 and 8). Drainage likely continued northeast towards Cortland, N.Y. A similar asymmetric distribution of stratified deposits also occurs in the upper northwest-southeast trending Dresserville Creek valley (plate 1).

The lowest potential meltwater outlet at the Owasco Inlet headwater divide just north of Freeville is at ~1,040 ft elevation, but the surficial deposits are mostly till (underlain by lacustrine sediments) and there is no drainage channel. This indicates that ice persisted in the valley at this location (and perhaps as far as about 4 mi northward), as no lacustrine deposits have been observed over the till in soil borings and a test pit in the area (D. Karig, Cornell University, written commun., 2021).

Pauses in ice-margin retreat are best marked by moraines in the uplands and by accumulations of coarse stratified material (ice-contact deposits) in the valleys (plate 1, fig. 2). Both Carney (1909) and Bird and Kozlowski (2014) depict ice margins oriented southwest to northeast across the watershed, with small ice tongues in the main valleys. Ice margins in this and other valleys of the Finger Lakes basin impounded water in a series of proglacial lakes as proposed by Fairchild (1934) and subsequently refined by Muller and Prest (1985) and Bird and Kozlowski (2014). Thick fine-grained lacustrine deposits make up most of the valley fill in the Owasco Inlet valley (fig. 4). Sediment-laden meltwater discharge from beneath and within the ice into proglacial lakes resulted in the deposition of coarse-grained subaqueous fans at the ice margin that were eventually buried by finer lacustrine sand and then silt and clay as the ice margin retreated to the north (Randall, 2001). Deep sand and gravel beneath thick lacustrine fine sand, silt, and clay is common throughout the Owasco Inlet valley and to the south across the divide with Fall Creek at least into the village of Freeville (see all hydrogeologic sections [figs. 4–10], especially section A–A′ on plate 1, and fig. 4).

Very thick lacustrine deposits fill the inlet valley along two valley reaches between Locke and Groton (plate 1, figs. 8 and 9). From Locke southward about 2.5 mi, lacustrine deposits form dissected terraces as high as 65 ft above the floodplain and around 100 ft beneath it. A second localized area south of the one just described and 1 mi north of Groton has lacustrine deposits rising nearly 90 ft above the floodplain and capped by an alluvial fan.

Soils derived from lacustrine deposits are indicated by the USDA in the Owasco Inlet valley bottom between Freeville nearly to Groton (USDA, 2008). However, recent coring and a test pit in the valley bottom between Peruville and Freeville indicate a thin layer of ochre to brownish-gray till, derived from lacustrine deposits, overlying gray lacustrine deposits (fig. 11; Karig, 2022). Inspection of well logs in the area indicate surficial clay or hardpan between 6 and 50 ft thick and brown in color, when described, which is interpreted as till. This deposit represents a readvance (Brooktondale Readvance, Karig, 2015) during final retreat of ice from the area. The lack of lacustrine deposits above the till indicates that remnant ice likely occupied the valley or that there was no lake at least locally during final retreat.

Pauses in retreat of the ice margin allowed buildup of coarse sediments from subaqueous or ice-marginal meltwater drainage into ice-contact deltas and development of lacustrine deltas from tributary streams (MacNish and Randall, 1982), which mark proglacial lake levels (Bird and Kozlowski, 2014). The most significant pause in ice retreat was at Groton, where the bedrock floor of the valley rises abruptly (fig. 4). Only at Groton are lacustrine deposits largely absent in the Owasco Inlet valley fill. Coarse stratified sediments make up most of the valley fill here, with the remainder either resedimented (flow) till or minor lacustrine deposits (figs. 4 and 7).

At other locations, it is uncertain that ice-contact deposits along valley walls are continuous from land surface downward through and beneath the lacustrine deposits, which may indicate ice readvance over lacustrine deposits. About 1 to 1.8 mi south of Moravia, just south of the transition to a narrower, shallower, and more southerly oriented valley reach (fig. 4), two features that rise as much as 40 and 75 ft, respectively, above the floodplain are identified as glaciofluvial by soil survey data (USDA, 2008) and Carney (1909). If these deposits predate the lacustrine deposits, a drape of lacustrine deposits (not evident) might be expected to overlie these deposits. The log from well CY 614 (plate 1, fig. 4) on the eastern feature indicates 20 ft of sandy clay underlain by 60 ft of lacustrine clay. This appears to indicate poorly sorted ice-contact sediments or till rather than glaciofluvial deposits at this location. The ice-contact delta about 3.1 mi south of Locke (Kozlowski and Graham, 2014; p. 31) likely spanned the valley and was later breached as lake levels declined. A test boring log (TM1083, plate 1, fig. 4) on the narrow floodplain between the east and west valley wall exposures of the ice-contact delta shows no subsurface indication of deltaic deposits—at least 105 ft of lacustrine fines (silty clay) beneath 6 ft of alluvium and shallow lake deposits (fig. 4). If the delta simply prograded over lacustrine fines, coarser (fine sand) lacustrine deposits might be expected below the delta at this location, as the delta was not far from the ice margin. Additionally, deformed lacustrine deposits exist near land surface beneath about 3 ft of coarse alluvial deposits at streambank exposures in Owasco Inlet about 1.3 mi south of Locke (fig. 12), which might indicate ice advance over the lacustrine deposits. Carney (1909) also described a similar outcrop and discussed possible evidence for glacial deposits older than the surficial deposits in the valley and surrounding area.

Some delta-like features are outwash or fluvial gravel plains over fine-grained lacustrine deposits rather than complete deltaic sequences. The high terrace feature at the junction of the Owasco Inlet valley and the Hemlock Creek valley at Locke indicates inflow from the Hemlock Creek valley that deposited coarse fluvial or outwash sand and gravel over lacustrine fines (figs. 4 and 9). The feature is highest and coarsest near the Hemlock Creek valley. A homeowner indicated that his well (CY 149) is about 200 ft deep because a thick clay unit was penetrated beneath the surficial gravel. A gravel pit along the northwest edge of the feature penetrated clay beneath very coarse gravel that facilitated creation of a pond. Likewise, another pond and contact springs are evident on the east side of this feature.

Another gravel plain over lacustrine fines is at the junction of the Dresserville Creek and Decker Creek valleys east of Moravia. Most of the coarse-grained sediment was derived from southward drainage from ice margin retreating up, and at the head of, Decker Creek valley. This gravel plain merges into thick deltaic deposits at the confluence of the Decker Creek and Dresserville Creek valleys with the Owasco Inlet valley at Moravia. About 15 ft of gravel overlays fines in a played-out gravel pit at the northernmost edge of where the tributary valleys join (south of Wilson Corners), and 0.2 mi due west, a well log (CY1604, plate 1) indicates 20 ft of sand and gravel over 40 ft of gray clay.

Ice in the Owasco Inlet watershed was part of the eastern flank of the Cayuga ice lobe centered in the Cayuga trough (Fairchild, 1932; see Denny and Lyford, 1963, plate 3; Karig and Miller, 2020). Thus, ice retreat across the watershed was toward the Cayuga trough—nearly due west in the south, and more northwestward to the north as indicated by traces of moraines (fig. 2, plate 1; Carney, 1909; Bird and Kozlowski, 2014). This orientation is best illustrated by the upland moraines that can be traced to the small valley tongue associated with the ice-contact delta about 3.7 mi south of Locke (fig. 2). As ice retreated northward in the valley reach from Freeville to Locke, meltwater drained over an unvegetated and unstable landscape toward the valley until the margin retreated west of the watershed divide. Sediments conveyed by the largest drainage pathway from the ice margin (Pleasant Valley, plate 1, fig. 2) created a large fan feature at Peruville. Carney (1909, p. 341) stated that the large alluvial fan and small upstream gorge at Peruville are out of proportion with the small drainage area of the present-day stream and thus represent abnormal drainage conditions—most likely drainage from when the ice was “near at hand.” This feature has a more irregular surface than large, present-day alluvial fans such as the two at Moravia (plate 1) and is incised by the current stream, which indicates they are older features. Church and Ryder (1972) describe fans deposited during the early part of the paraglacial (recently deglaciated) period as more debris-flow rich than fluvial. Therefore, the term “debris-flow dominated alluvial fan” (alfd map unit) is used for these deposits in this report. These features are noted in smaller upland drainages such as Pine Hollow Creek, Pleasant Valley (upgradient of Peruville), and the drainage south of Sayles Corners (plate 1, fig. 2).

Apart from the large fan at Peruville, small fan features exist at the base of steep drainage channels on the walls of the Owasco Inlet valley. Where drainages are closely spaced, the smaller features form an apron along the base of the valley wall. The alfd units are not well defined and may include adjacent till or slumped till, likely derived from lacustrine deposits. These units and adjoining deposits mapped as till have variable soil survey parent materials (glaciofluvial, till, or lacustrine), and well logs report clay, hardpan with gravel layers, and clayey, dirty, or clean gravel. There are no detailed geologic logs of these materials. Without such geologic control, deposits at the base of the valley wall areas mapped as lacustrine silt and clay, till or alfd are speculative, especially as till is likely derived from mostly lacustrine deposits.

Retreat of the ice and exposure of lower lake outlets (Bird, 2014; Bird and Kozlowski, 2014) led to periodic drops in proglacial lake levels and incision of Owasco Inlet valley, especially between Locke and Groton, where lacustrine deposits and the ice-contact delta are 60–90 ft to nearly 200 ft, respectively, above the present-day floodplain. Lowering of base level led to incision in the other drainages within the watershed and deposition of numerous small alluvial fans in the valleys. Two large alluvial fans from Mill Creek (the Decker and Dresserville Creek watersheds) and from the stream at Fillmore Glen State Park have merged and dominate the valley bottom at Moravia. From Moravia north to Owasco Lake, Owasco Flats is a mix of floodplain and swamp lands, with at least 25 ft of organic, fine-grained alluvial deposition over a basal coarser grained deposit that is early alluvium or of alluvial fan origin, as noted at Moravia in figure 4.

At the southern end of the watershed, near the divide north of Freeville, the valley floor is also largely swamp, much of which is underlain by a thin layer of till, over lacustrine deposits. This valley area has probably changed least in postglacial time because of low relief and high elevation (limited drainage area).

Sand and Gravel Aquifer Type and Distribution

Several sand and gravel aquifer settings are recognized within the Owasco watershed: confined (including confined by till), semiconfined, and unconfined/semiconfined aquifers within the Owasco Inlet valley, and unconfined and confined aquifers in upland areas (fig. 13). The distributions of these aquifers (especially confined and semiconfined) within delineated areas are not continuous—they may be locally absent or poorly sorted (of poor permeability).

Unconfined aquifers are the most susceptible to contamination from nearby activities at land surface because precipitation and subsequent recharge can transport contaminants directly through the unsaturated zone to the water table. Confined aquifers are less susceptible to contamination from local land surface because confining units largely prevent downward movement of water. Recharge of confined aquifers occurs elsewhere—at unconfined upvalley locations and along valley walls where alluvial fans, ice-contact deposits, or stream incision into the valley wall may provide pathways for downward movement of groundwater.

A deep, confined aquifer underlies much of the Owasco Inlet valley within the study area (figs. 4 and 13). A smaller, separate, and largely untested confined aquifer is also present in the lower reach of Dresserville Creek. The Owasco Inlet valley confined aquifer is known to extend south and likely north of the study area. Aquifer material extends south across the low divide with Fall Creek at least to Freeville, where flowing artesian conditions exist in many domestic wells. The aquifer also likely continues north of Moravia and beneath Owasco Lake, as northward ice retreat continued to generate sediment-laden meltwater that was deposited as subaqueous fans immediately in front of the ice or as eskers within the ice, which may have fed the subaqueous fans. Water-borne seismic surveys of the Finger Lakes conducted by Mullins and others (1996) lend support to this interpretation. Owasco Lake and other Finger Lakes commonly showed chaotic reflectors that form hummock and ridge features above the bedrock and below lacustrine deposits at the north ends of lakes that were interpreted as eskers or ice-contact deposits. These units transitioned into more regular reflections to the south that were interpreted as more sorted, finer grained “subaqueous outwash” units.

The thickness of this aquifer is rarely documented because drillers obtain sufficient water for domestic purposes shortly after reaching it and thus do not drill to bedrock. Aquifer thicknesses are generally estimated to be 20 ft or less from HVSR depth-to-bedrock estimates and well logs in the Owasco Inlet valley (fig. 4).

Hydraulic heads in the aquifer have historically been flowing artesian—above floodplain elevation in the valley bottom—especially north of Groton (the Locke area in particular) and at the south-end drainage divide and beyond in the Freeville area. Some estimates of head from New York State Department of Transportation borings in the Owasco Inlet valley since the 1950s were 100 ft above land surface with flow rates as high as 150 gallons per minute. These conditions indicate that drillers should expect and be prepared to control high water pressures in this aquifer.

The Village of Locke’s water supply taps this aquifer and serves a population of 600 (U.S. Environmental Protection Agency [EPA], 2017). The two production wells are 179 and 194 ft deep. Total water withdrawn through these wells in 2015 was about 10.7 million gallons (29,300 gallons per day [gal/d]; NYSDEC, 2016). The yields of these wells were not reported.

Groundwater Recharge, Flow, and Vulnerability to Contamination

Groundwater recharge, flow in the subsurface, and discharge to low points on land surface is the subsurface part of the hydrologic cycle. Groundwater recharge is rainfall and snowmelt that infiltrates land surface, moves through the unsaturated zone, and joins groundwater at the water table—the top of the zone of saturated sediments or fractured bedrock. During a given year, the addition of water to the groundwater system induces groundwater to flow in directions of lower hydraulic head, and to ultimately discharge to low points on land surface that act as drains (streams, lakes, wetlands, and springs; fig. 14). If groundwater did not flow and discharge, water levels would rise and eventually flood land surface. Groundwater flows through aquifers, which are water-saturated sediments or rock that are thick and permeable enough to supply water to wells, and through materials that are too thin or low in permeability to be considered aquifers. Groundwater residence times can be relatively short (months, years), such as recharge from precipitation on valley bottoms that flows within floodplain sediments a short distance and discharges to the valley stream. Other groundwater residence times may be very long (hundreds to thousands of years) in some regional aquifers (Walter and others, 2020) and possibly in local, highly confined aquifers.

Recharge can only occur where the sediments or bedrock at and below land surface are permeable enough to allow infiltration. In the Owasco Inlet watershed, for example, thick glacial till and fine-grained lacustrine deposits have low permeability and thus allow little, if any, recharge, whereas stratified coarse-grained deposits (ice-contact, outwash, and some coarser alluvium and alluvial fans) can be areas of high recharge (fig. 14). Fractured bedrock can vary widely in permeability, but in general it is less permeable than coarse stratified sediments; most fracture permeability in clastic (noncarbonate) sedimentary rock in hillside areas is approximately within the upper 100 ft of bedrock (fig. 14; see Heisig, 1999, p. 46 for an example). Recharge is most likely to occur in upland areas with thin, discontinuous till over bedrock, but recharge in upland areas of thick till is likely close to nil. Thus, upland sand and gravel aquifers confined by thick till are probably a limited resource.

Unconfined and semiconfined sand and gravel aquifers in valleys may receive recharge from direct precipitation and two additional sources: infiltration of streamwater through alluvial fan deposits where tributary streams enter valleys (Randall, 1978), and unchanneled surface runoff or subsurface groundwater flow from adjacent, unchanneled valley walls into valley-fill deposits (Morrissey and others, 1988). A fourth, and potentially major, source of recharge to these aquifers is induced infiltration of stream or river water into aquifers where groundwater withdrawals from wellfields intercept, and then reverse, natural groundwater flow to the stream. This may account for the high well yields at the Moravia wellfield.

Confined aquifers receive recharge indirectly through hydraulic connection with permeable deposits (sand and gravel, or bedrock), either in valley reaches where stratified sand and gravel are the primary valley fill (unconfined aquifers) such as at Groton, or where ice-contact, deltaic, or alluvial fan deposits along valley walls extend down to the confined aquifer (figs. 7 and 14).

The likelihood of groundwater contamination from human activities at land surface is highest in unconfined aquifers since infiltrating water (including induced infiltration) can move through the unsaturated zone directly to the aquifer. Upland areas of thin till may also provide short pathways for contaminants into the underlying fractured bedrock. Common contaminants include road-deicing salts, spills along roadways, and septic leachate. Landfills may also be a source of leachate to groundwater if protective caps or liners fail. Supply wells in unconfined and semiconfined aquifers that induce water from rivers are susceptible to negative changes in river water quality such as turbidity, algal blooms, and uncontrolled wastewater discharges.

Confined aquifers are more protected from contamination by overlying land-use activities because the confining unit prevents downward groundwater flow. For example, contaminants infiltrating from land surface in a valley floodplain underlain by fine lacustrine deposits (fig. 14) will be entrained by shallow groundwater flow in the alluvium and eventually discharge to the valley stream. Confined aquifers are susceptible to contamination from activities in their recharge areas—preventing high-contamination-risk land uses in the most likely recharge areas can help safeguard these aquifers.