Groundwater flow conditions below beaches of Lake Huron have been examined at eight sites. Groundwater below beaches flows towards the lake throughout the year. However, water table conditions vary between beaches with a shallow water table (< 1 m) and beaches with a deep water table (> 2 m). Where the water table is deep, the water table across a beach will rise and fall proportionally to fluctuations in lake levels. The hydraulic gradient is essentially linear across a beach. Beaches with a shallow water table exhibit very little change in the elevation of the water table over time, do not fluctuate proportionally with lake level changes, and have a break in the hydraulic gradient at ∼ 20 m from shore. Long term and season fluctuations in the surface of Lake Huron will change the elevation and location of the groundwater discharge area, resulting in the water table beneath the entire beach where the water table is deep to rise and fall proportionally and maintaining essentially the same hydraulic gradient. However, fluctuations in the lake levels have little impact on water table conditions at beaches where the water table is shallow.
Increased residential development and recreational activities at beaches along the shore of Lake Huron has caused heightened awareness and concerns about water quality issues along the shoreline. Water quality issues at beaches have focused on near-shore lake water quality; elevated nutrient levels that promote algae and high levels of E. coli that lead to beach closures (Byappanahalli et al., 2003; Haack et al., 2003; Science Committee, 2005; Alm et al., 2006, Higgins et al., 2008; Edge et al., 2007). However, groundwater quality is also a concern at beaches. Most beach front residents obtain their domestic water from shallow water table wells and sand points, and these wells may be susceptible to contamination from septic systems (Robertson and Cherry, 1992; Ptacek, 1998). This has lead to septic system inspection programs in Huron County, along the eastern shore of Lake Huron (Huron County, 1993), and Tiny Township, along the southern shore of Georgian Bay (Vella, 2002) to protect drinking water sources. Groundwater below the beach may act as a mechanism for transporting contaminants, such a nutrients and bacteria, to the lake (Chen, 1988; Boehm et al., 2004; Haack et al., 2003; de Sieyes et al., 2008).
Several studies have assessed or quantified the seepage of groundwater into the Great Lakes on both a local scale (Lee and Harvey, 1996; Harvey et al., 1997a, 1997b), and regional scale (Cherkauer and Hensel, 1986; Cherkauer and Taylor, 1990; Cherkauer and McKreeghan, 1991; Sellinger, 1995). There have also been a few studies that investigated the role of groundwater in the hydrology of coastal wetlands of the Great Lakes (Doss, 1993; Crowe et al., 2004; Crowe and Shikaze, 2004; Bailey and Bedford, 2003). There have been numerous studies that investigated groundwater conditions below beaches in marine environments (see reviews by Baird and Horn, 1996; Horn, 2002; 2006), but these studies typically investigated the short-term effects of tides and waves on groundwater conditions and sediment transport at the swash zone. Currently, there is no information that specifically describes groundwater conditions below beaches of the Great Lakes, including Lake Huron, or how groundwater below beaches is affected by water level fluctuations of the Great Lakes.
The objectives of this paper are to present an overview of water table fluctuations below beaches along the shoreline of Lake Huron, and how they are affected by the seasonally changing levels of Lake Huron. Because this paper is focused on the beaches of Lake Huron, the groundwater conditions are assessed to a distance of generally less than 80 m from shore. The groundwater regime below beaches adjacent to the shoreline can be affected by wave run up and infiltration at the swash zone during storms, resulting in a water table rise and a hydraulic gradient reversal (Horn, 2002; Meek, 2007), but this will not be discussed in this paper. The information presented in this paper is based on field observations of water table characteristics and hydraulic testing at eight sites at five different beaches along the shoreline of Lake Huron, including southern Georgian Bay (Figure 1).
Along the Canadian shoreline of Lake Huron, beach and dune systems are located primarily from Sarnia to Sauble Beach and along southern Georgian Bay from south of Collingwood to west of Midland (Figure 1). However, these beaches are not continuous, but separated by bedrock outcrops, rocky shoals, and bluffs (e.g. Kettle Point, Point Clark, Douglas Point, Spratt Point). Smaller and isolated beaches are located along the coast of Manitoulin Island (e.g. Providence Bay).
Beaches along the shore of Lake Huron and Georgian Bay are built upon the wave cut terraces of lacustrine clay or till formed by glacial Lakes Algonquin and Nipissing (Chapman and Putnam, 1985). The beach environment varies along the coast from large beach and dune systems that extend a few kilometers inland, to narrow fringe beaches at the base of bluffs. The texture of the beach deposits of Lake Huron is generally gravel; while beaches of southern Georgian Bay are composed of well sorted coarse sand. The dune system on the leeward side of the beaches is composed of a fine to medium grained aeolian sand. The width of the beaches vary both seasonally and long-term because of fluctuations in the elevation of Lake Huron.
Long-term records of lake level fluctuations in the surface of Lake Huron/Georgian Bay exhibit both long-term and seasonal cycles (Department of Fisheries and Oceans (DFO), 2007). The maximum and minimum lake levels were 177.50 m asl in October 1986, and 175.88 m asl in March 1964. Seasonal fluctuations of 0.2 to 0.6 m show a maximum lake level during the late spring and a steady decline during the summer, fall and winter. During the years of the current groundwater study, 2005 to 2007, the annual lake levels fluctuations were 0.32 m during 2005, 0.26 m during 2006, and 0.40 m during 2007 with minimum lake levels in December/January and maximum levels in June (Table 1).
Eight beach sites were studied at five beaches along the shores of Lake Huron (Figure 1). The sites are Ashfield Township Park Beach, Jackson Park Beach and two sites each at Amberley Beach, Balm Beach, and Woodland Beach. At all sites, the beach sand overlies lacustrine clay or till.
The Ashfield Township Park Beach (ASHF) on Lake Huron is a public beach located adjacent to the mouth of 18 Mile River. The actually study site here is located 100 m north of the river. The beach is approximately 55 m wide and a series of low dunes (1.5 m high) are present at the leeward side of the beach. Beach grass (Ammophila breviligulata) is present within the dunes.
The two beach sites at Amberley Beach (AMBN, AMBS), along the shore of Lake Huron, are very similar. These sites are approximately 410 m apart. In both cases the beach is quite narrow with low sand dune development with beach grass starting approximately 8 m from shore. The sand dunes rise to 3.5 to 4 m above the lake about 30 m from shore. Beach front residences are located 50 m from the shoreline. Amberley Beach is non-public beach, and accessible only to the local beach community.
Two sites, approximately 1,300 m apart, were studies at Balm Beach on Georgian Bay. The north site at Balm Beach (BALMN) is a public beach approximately 50 to 60 m wide, with a road, parking and stores along the leeward edge of the beach. Sand dunes are absent from this beach. Two small creeks cross the beach about 60 to70 m apart on either side of the study area and often meander across the beach during the fall and spring. Given the low ground surface elevation and frequently wet-sand conditions here, invasive vegetation (turf grass, sedge grass) is present. However, beach grooming in the summer removes this vegetation and straightens the creeks. The south site (BALMS) is a natural beach with well developed sand dunes, and beach grass within the dunes. The beach is over 120 m wide, and the dunes rise 9 m above the beach. Residences are set back approximately 90 to 100 m from the shoreline. This beach is a non-public beach and accessible to only the local shoreline community.
Two sites, approximately 520 m apart, were studied at Woodland Beach on Georgian Bay. The south site (WOODS1) has a beach approximately 40 m wide. Sand dunes and natural beach grass have been removed by the beach front residents and retaining walls have been erected at the edge of the beach. As a result, the maximum elevation change across the beach is only 1.5 m above the lake. Invasive phreatophytes and turf grass has become established. The north site (WOODN) has well developed sand dunes and natural beach grass. The beach is approximately 55 m wide and has dunes that rise 5 to 6 m above the lake. Both beaches are non-public beaches and accessible to only the shoreline community.
The beach at Jackson Park (JACK) is a public beach located in a sheltered bay, with parking facilities north of the beach and houses along the leeward side of the beach. The beach is approximately 55 m wide, but sand dunes and beach grass are absent. The low elevation of the ground surface (maximum elevation is 0.85 m above the lake) of the beach typically produces wet and damp sand conditions, and invasive phreatophytes and turf grass are common.
Groundwater conditions below the beaches were investigated by measuring the elevation of the water table in a grid of boreholes across the beach, or in a series of boreholes dug along a single transect perpendicular to the shoreline. Permanent water table wells and piezometers could not be installed across the beach because of issues related to public safety and lake-ice frequently is pushed across the beach during the spring. Given that beach sand has a high hydraulic conductivity, the thickness of saturated sand below a beach (between the underlying till and the water table) is small relative to the width of a beach, the slope of the water table is relatively small, and hence, groundwater flow below beaches will be essentially horizontal (Freeze and Cherry, 1979; Horn, 2002; Meek, 2007). Vertical gradients at a specific location will be very small and the hydraulic head a depth at a given location will be essentially the same as the elevation of the water table. Hence groundwater flow below a beach is essentially one dimensional with flow lines essentially parallel to the water table. However, groundwater flow at the shoreline is two dimensional and the elevation of the water table here does not represent hydraulic head with depth (Freeze and Cherry, 1979; Baird and Horn, 2002; Cartwright et al., 2006). Two-dimensional groundwater flow occurs near the lake because the slope of the water table increases and groundwater flows upward to discharge into the lake. Thus, water table elevations can be used to accurately characterize groundwater flow conditions below beaches beyond approximately 5 m from the shoreline.
Figure 2 illustrates a typical bore-hole grid for measuring water table elevations across a beach, both parallel to and perpendicular to lake. Between 5 and 10 lines of boreholes perpendicular to the lake were used, and the spacing between lines were either 5 m or 10 m. The centre line of boreholes perpendicular to the lake was selected for the following hydraulic gradient calculations, and borehole spacing and water tables along this transect are shown in Figure 3. When the water table elevations were measured along a single transect perpendicular to the shoreline, the number of boreholes and the spacing between boreholes varied from site to site (Figure 3). The locations of all boreholes were recorded using GPS and measured to a reference point at each site to ensure that locations of water table measurements could be exactly compared at different times. The depth to the water table was determined by measuring the depth to the water table from the top of a plank securely placed across the borehole after the water level in the borehole stabilized. The elevation at the point where the water table depth was measured (representing ground surface) was measured using an electronic level (model NA200 by Wild), and these elevations were referenced to the elevation of local bench mark near each of the five beaches. Also, at each sampling time, the elevation of the lake was estimated by leveling the surface of the lake. For this study, the ground surface and water tables are referenced to the elevation of the surface of the lake (International Great Lakes Datum, 1985) so that the elevations among the five beaches can be compared. Surface elevations (IGLD, 1985) of Lake Huron at Goderich and Georgian Bay at Collingwood were obtained from the Canadian Hydrographic water level gauging stations (DFO, 2007).
Hydraulic conductivity of the beach sands at four of the beaches were measured using a Guelph Permeameter (model 2800KI by Soilmoisture Equipment Corp.). Infiltration measurements were taken at depths of 10 to 72 cm below ground surface and converted to field saturated hydraulic conductivities using the method described by Reynolds and Elrick (1987).
Characteristics of water tables below beaches of Lake Huron
The elevation of the water table below Jackson Park Beach shown on Figure 2 is typical of other beaches that were studied with a grid of boreholes; the water table slopes throughout a beach towards the lake at all beaches and at all times during the year. Hence, water table elevations along a single transect of boreholes perpendicular to the shoreline is representative of groundwater conditions at a beach. Figure 3 shows the elevations of the beach surfaces and the water tables measured at different times of the year at all eight beach sites, and referenced to a common vertical datum and horizontal scale.
The water tables below the eight beaches, shown in Figure 3, exhibit two trends: (1) beaches with a deep water table (Amberley Beach north and south, Ashfield Twp Park Beach, Woodland Beach north, Balm Beach south), and (2) beaches with a shallow water table (Jackson Park, Woodland Beach south, Balm Beach north). The water tables at the beaches with a deep water table do not follow the slope of the ground surface but its depth varies with the height of the ground surface (Figure 4a). These beaches have sand dunes at the leeward side of the beach and the water table is between 2.5 and 4 m below the dunes. At the beaches with a shallow water table, the water table follows the ground surface and the depth to the water table is less than 1 m across the beach (Figure 4b). The change in the elevation of the ground surface across these beaches is less relative to the other beaches, rising only 1 to 2 m above the lake.
Hydraulic conductivity of the beach sand was measured at five beaches; four with deep water tables (AMBN, ASHF, WOODN, BALMS) and one beach with a shallow water table (WOODS). Hydraulic conductivities were all within an order of magnitude of each other, both among all measurements at a beach, and among the geometric means of all the beaches (Table 2). The geometric mean of the hydraulic conductivity measurements at AMBN, ASHF, WOODN, BALMS and WOODS are 3.68 × 10−2 cm s−1, 2.91 × 10−2 cm s−1, 3.96 × 10−2 cm s−1, 8.69 × 10−2 cm s−1 and 4.55 × 10−2 cm s−1, respectively. These values are similar to hydraulic conductivities measured at beaches at Point Pelee on Lake Erie (2.5 × 10−2 cm s−1) by Crowe et al. (2004) and Indiana Dunes on Lake Michigan (1.8× 10−2 to 3.5× 10−2 cm s−1) by Doss (1993), and similar to hydraulic conductivities at marine beaches (Horn, 2002; 2006). However, hydraulic conductivity at Balm Beach south is about twice as high as the other four beaches.
The hydraulic gradient of the water tables are different between beaches with a deep or shallow water table (Table 3). At beaches with deep water tables, the water table exhibits a fairly linear slope across the beaches (Figure 4a), with hydraulic gradients at different times of the year between 0.00291 (ASHF) and 0.0183 (AMBS) and a geometric mean of 0.0100. The hydraulic gradient of the water tables at beaches with a shallow water table exhibit two slopes (Figure 4b). This trend is most pronounced at WOODS and BALMN (Figure 3) where between 20 to 25 m from the lake and the leeward edge of the beach, hydraulic gradients are between 0.0107 and 0.0158, with a geometric mean of 0.0139. These gradients are within the range of the water table gradients at beaches with deep water tables. However, between about 5 m to about 20 to 25 m from the lake at beaches with shallow water tables, the hydraulic gradient is greater, ranging from 0.0444 to 0.0618, with a geometric mean of 0.0482. Jackson Park Beach also exhibits a change in the hydraulic gradient at about 20 m from the lake (Table 3), but only during July 2006. The hydraulic gradients of the beach at 5 to 20 m from the lake (0.0115) and beyond 20 m from the lake (0.00714) are much lower than those at WOODS and BALMN. Hydraulic conductivity is not the main cause of the change in the hydraulic gradients at beaches with a shallow water table. The hydraulic conductivity of the beach sand at WOODS is essentially the same both beyond ∼ 20 m from the shore is 5.93 × 10−2 cm s−1 and within 5 to 20 m of the shore is 4.36 × 10−2 cm s−1, and is essentially the same as the hydraulic conductivities of the beaches with deep water tables (Table 2).
The elevation of the water table fluctuates throughout the year and from one year to the next. The water table is at its maximum elevation during the spring due to infiltration of snow melt and spring rains, and a high lake level. The water table is at a minimum elevation in the winter, following an extended period of little groundwater recharge, continual discharge to Lake Huron, and low lake levels. The largest water level changes were observed at beaches with a deep water table; for example approximately 0.41 m at AMBS, 0.24 m at AMBN, 0.32 at WOODN and 0.30 m at BALMS. There was very little change in the elevation of the water tables at beaches with a shallow water table; for example 0.14 m at WOODS, 0.043 m at BALMN and 0.064 m at JACK. Fluctuations in the elevation of the water tables at a beach are caused in part due to infiltration (e.g. rainfall and spring snow melt), and in part due to changes to the surface levels of Lake Huron, as discussed below. The water table rise and fall at each beach is essentially the same across the beach. Hence, although the hydraulic gradients of the water tables differ slightly from one beach to the next, the hydraulic gradients at each beach are consistent at each beach from one season to the next (Table 3).
Groundwater velocity and groundwater flux can be calculated at the five beaches where hydraulic conductivity was measured (AMBN, ASNF, BALMS, WOODN, WOODS). Because the hydraulic conductivity and hydraulic gradients are consistent at each of the beaches, calculations use a geometric mean hydraulic gradient for each beach using all sampling dates, and a geometric mean hydraulic conductivity using all measurements at a beach. Because the hydraulic gradient at WOODS exhibits two slopes, calculations use the hydraulic gradient from the portion of the beach beyond approximately 20 m from the shore. An assumed porosity of 0.35 was used from Freeze and Cherry (1979). The calculated groundwater fluxes range from 7.31 × 10−2 m d−1 at ASHF to 8.3 × 10−1 m d−1 at BALMS (Table 4). Groundwater velocities range from 0.21 m d−1 at ASHF to 2.4 m d−1 at BALMS. These high fluxes and velocities would indicate that changes to the water table would occur quickly and throughout the beach in response to changes in infiltration rates or lake level fluctuations.
All beaches exhibit an increased downward slope of the water table within 5 m of the lake. The increase in the slope of the water table here is due to the increased slope of the ground surface causing the water table to intersect and follow the surface of the beach forming a seepage face. Ground surface elevations across the beaches and dunes, shown in Figure 3, were relatively stable between the dates at which measurements were taken. But the ground surface elevations at the shoreline changed considerably among sampling times due to changes in the lake levels that moved the shoreline lakeward or inland across the shore face of the beach and due to erosion and deposition of sand by wave action along the shoreline. For example, at Woodland Beach south site, the shoreline moved lakeward 8.4 m between May 2006 and November 2006. Changes in the ground surface elevation near the shoreline are indicated in Figure 3 with a dashed line, changes in the elevation of the lake are indicated with a horizontal dotted line, and the location of the groundwater-lake-beach interface is shown by a solid black symbol.
Impact of Lake Huron on water table conditions below beaches
Infiltration of precipitation, both at the beach and the areas adjacent to the beach, will maintain the water table at an elevation above the level of Lake Huron throughout the year. Lake Huron acts as a drainage point for groundwater below the beaches, resulting in a water table that slopes towards the lake throughout the year (Figure 3). Groundwater discharge is concentrated within a few metres of the lake-beach interface (Figure 4), with discharge occurring at the seepage face and into the lake adjacent to the shore (McBride and Pfannkuch, 1975; Lee, 1977; Cherkauer and McBride, 1988; Winter, 1999; Simpkins, 2006; Meek, 2007).
The primary effect of Lake Huron on water table conditions below beaches relates to how fluctuating lake levels will change the location of the groundwater discharge area at the shoreline. The seasonal changes in the elevation of Lake Huron are approximately 0.2 to 0.4 m (DFO, 2007), with highest lake levels during the time of this study (2005–2007) in June, and lowest lake levels in late winter (Table 1). Because the surface of a beach slopes towards the lake, a rising lake level will cause the lake to move laterally inland across the shoreline, and a falling lake level will move the shoreline laterally lakeward. Thus, because groundwater discharges at the shoreline, as the lake level rises and falls, the groundwater discharge area will move both laterally inland and lakeward across the beach and to a higher elevation corresponding to the elevation of the lake. This is observed at the beaches of Lake Huron, where the locations of the groundwater-lake-beach interface at different dates, as shown by the solid black symbols on Figure 3, move laterally and vertically from season to season due to changes in the lake level as shown by the dotted horizontal line. The lateral extent of the movement of the shoreline during the lake-level fluctuations depends on the slope and elevation of the beach at the shoreline. For example, at WOODS, WOODN, BALMS and BALMN the lateral movement was 8.4 m, 4.3 m, 8.1 m and 5.3 m, respectively.
As Lake Huron and the accompanying groundwater discharge areas rises and falls, the water table across a beach with a deep water table will also rise and fall (Figure 3). The change in the elevation of the water table at these beaches is approximately the same as the change in the lake level. The change in the elevation of the water table at beaches with a deep water table was 0.41 m at AMBS, 0.24 m at AMBN, 0.32 at WOODN and 0.30 m at BALMS, this corresponds to lake level changes of 0.23 m, 0.23 m, 0.36 m and 0.29 m at AMBS, AMBN, WOODN and BALMS, respectively. The hydraulic gradient of the water table is controlled by the hydraulic conductivity of the beach sand and the amount of recharge entering the groundwater at and adjacent to the beach because lake water does not move into the beach with seasonal changes in lake levels. But the elevation of the water table is controlled by the elevation of the drainage point (the elevation of the groundwater-lake interface). Therefore, the elevation of the water table across the beach will respond proportionally to a rising and falling groundwater drainage point caused by the seasonal lake level fluctuation. The high hydraulic conductivity of the sand will allow groundwater to readily drain from the beach, and thus rise and fall of the water table occurs at essentially the same time scale as the rise and fall of the lake level (Cartwright et al., 2006; Li et al., 1997). Hence, there is no movement of lake water into the beach as the lake level rises and there is no mounding of the water table below the beach as the lake level falls.
At beaches with a shallow water table (BALMN, JACK, WOODS), changes in the lake level resulted in a water table rise and fall near the shoreline, but further back from the shoreline, the lake level changes had little impact of the elevation of the water table. Thus seasonal lake level fluctuations have little impact on the elevation of the water table below beaches with a shallow water table. It is suspected that at these beaches, groundwater recharge has a greater impact than the lake on water table conditions. Because these beaches have a water table close to ground surface (< 1 m) and the sand is near residual saturation from the water table to ground surface, rainfall will infiltrate to the water table maintaining a high water table (Crowe and Milne, 2007). Beaches with a deep water table (> 2 m) also have a lower moisture content in the sand and this will limit the ability of rainfall to infiltrate to the water table.
Implications for water quality at beaches
Because groundwater below beaches always flows towards the lake and at high groundwater velocities, any contaminant that enters the groundwater at a beach will move towards the lake. Whether the contaminant actually reaches the shoreline and discharges into the lake is dependent on the transport and persistence properties of the contaminant (e.g. degradation rate, attenuation rate, dilution). If the contaminant does reach the shoreline, it will discharge into the lake at the shoreline and not 10s to 100s of metres from shore. The time for a non-reactive contaminant to travel across the width of a beach (approximately 60 m) by groundwater flow varies from beach to beach, for example 25 days and Balm Beach south to 280 days at Ashfield Twp Park Beach. Thus contamination of groundwater below a beach and contaminant discharge to the lake will occur relatively quickly after a contaminant has reached the water table. Beaches potentially have many sources of contaminants from both human activities and wildlife. Numerous studies have already documented groundwater contamination at beaches from a variety of sources, including nutrients and bacterial from septic systems or sewage lines (Robertson and Cherry, 1992; Ptacek, 1998; Boehm et al., 2004; Crowe and Milne, 2007), and bacteria from birds (Whitman and Nevers, 2003; Edge et al., 2007; Crowe and Milne, 2007; Kon et al., 2007). These studies also indicate that groundwater can transport contaminants to the shore.
This study illustrates the importance of understanding conditions affecting the flow of groundwater at beaches, and the potential impact that groundwater flow has on water quality conditions both in groundwater below the beach and in the surface water along the beach (nearshore zones). In particular, given the potential impact of contamination of groundwater below beaches and at the shore, care should be taken to prevent contaminants from entering groundwater that eventually impacts surface water at the beach.
Water table and groundwater flow conditions have been examined at eight sites at five beaches along the shore of Lake Huron. Although the elevation of the water table fluctuates from season to season (highest in the spring, lowest in late winter), groundwater below beaches will flow towards the lake throughout the year. Water table conditions vary between beaches with a shallow water table (< 1 m) and beaches with a deep water table (> 2 m) mainly due to the surface characteristics of a beach (i.e. presence or absence of sand dunes). Where the water table is deep, the hydraulic gradient is essentially linear across a beach, but varies from beach to beach from 0.00291 to 0.0183. Long term and seasonal fluctuations in the surface of Lake Huron will change the elevation and location of the groundwater discharge area. Groundwater levels below beaches with a deep water table will rise and fall (0.2 to 0.4 m) proportionally to lake levels but will maintain essentially the same hydraulic gradient. However, beaches with a shallow water table exhibit very little change in the elevation of the water table over time (< 0.15 m). The water table does not fluctuate proportionally with lake level changes and has a change in the hydraulic gradient at ∼ 20 m from shore. The hydraulic gradient between approximately 5 m and 20 m from shore is 2–4 times greater than the hydraulic gradient beyond approximately 20 m from shore. A change in hydraulic conductivity across these beaches is not the cause of the change in the hydraulic gradient because the hydraulic conductivity is essentially the same at both portions of the beach and the same as beaches with a deep water table. Hence, fluctuations in the lake levels have little impact on water table conditions at beaches where the water table is shallow.
The work could not have been undertaken without the field assistance of Jacqui Milne, Charlie Talbot, Earl Walker, Mike Benner, and Adam Morden of Environment Canada!s National Water Research Institute. Funding was provided in part by Environment Canada!s Great Lakes Restoration Programs.