tf3 6.6S-. C^^cA oxxftjo t^l STATE OF ILLINOIS DEPARTMENT OF REGISTRATION AND EDUCATION Temperature Prospecting for Shallow Glacial and Alluvial Aquifers in Illinois Keros Cartwright ILLINOIS STATE GEOLOGICAL SURVEY John C. Frye, Chief URBANA CIRCULAR 433 1968 URBANA ILLINOIS ST GEOLOGICAL SURVLY LIBRARY TEMPERATURE PROSPECTING FOR SHALLOW GLACIAL AND ALLUVIAL AQUIFERS IN ILLINOIS Keros Cartwright ABSTRACT Theoretical considerations of the thermal properties of glacial and alluvial deposits in Illinois suggest that a shallow aquifer might form a heat sink (or source) that would influence the temperature effects on the soil of heat origi- nating at the land surface and within the crust. If shallow aquifers are nonuniformly distributed laterally, and if their effects on the temperatures of surface soil can be measured and distinguished from other factors that affect soil temper- ature, a possible exploration method for shallow aquifers is suggested. A positive (warm) anomaly would be expected over such aquifers in the winter and a negative anomaly in the summer. The size of the anomalies is dependent upon the thermal properties of the overburden, temperature differ- ence between the surface and aquifer, and the depth of bur- ial of the aquifer. An electronic thermometer, utilizing a thermister at the end of an aluminum-tipped stake and a transistor-ampli- fied bridge circuit, was used to measure soil temperatures at a depth of about 18 inches in seven areas in Illinois where shallow aquifers were known in some detail. The data from these surveys, presented here, show maximum anomalies of about 2° C over the aquifer. Surveys were made both in summer and winter; summer anomalies generally are of a greater magnitude than winter anomalies. Soil differences, vegetation, and ice in the soil also affect the soil temperatures. The effect of vegetation in the summer can be as great as the anomaly produced by an aq- uifer. Frozen soil tends to eliminate anomalies, but this effect can be partly overcome by taking readings at a great- er depth. In general, however, field data show close agree- ment between the location of shallow aquifers and thermal anomalies. 1 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 INTRODUCTION Most geophysical techniques used in the exploration for ground water meas- ure some property of the rocks, rather than properties of water. However, the pres- ence of water in the rocks affects the results somewhat in electrical earth resistivity and seismic methods, the most commonly used exploration techniques. A property that may be exploited in ground-water exploration is the high specific heat of water or its resistance to changes in temperature. Theoretical considerations of glacial and alluvial deposits suggest that a saturated aquifer may disturb the geothermal gradient by acting as a heat sink (ab- sorbing heat) (Lovering and Goode, 1963) or heat source. This disturbance may in- fluence the surface soil temperature. If surface soil temperature variations result- ing from disturbance of the geothermal gradient by a shallow aquifer can be meas- ured, the presence of the aquifer might be detected, provided the aquifer has later- al boundaries across which will be temperature contrasts and provided the tempera- ture effects of other heat sources (or losses) can be eliminated or evaluated. The glacial drift of Illinois contains many shallow linear deposits; temperature prospec- ting over several of them has suggested that the method bears further investigation. Nomenclature The following symbols are used throughout the paper in equations, figures, and tables: A Area — cm^" a Thermal diffusivity — cm/sec (3 Variable of integration, in this case (3 = . | c Specific heat — cal/(gm)(°C) k Thermal conductivity — cal/(sec) (cm) (°C) JL Lower boundary of the slab (overburden) above the heat source m Upper boundary of the slab above the heat source n Degrees of freedom — n = 1 for one dimension, n = 2 for two dimensions, and n = 3 for three dimensions of heat flow Q Quantity of heat — cal 0' Rate of heat production in a permanent heat source — cal/sec H^ for one dimension p Density — gm/cm^ S' Strength of heat source t Time — seconds T Temperature— °C TEMPERATURE PROSPECTING FOR AQUIFERS T r Surface temperature range IX x Depth variable — cm B Constant of integration C Constant of integration. AQUIFER TEMPERATURES Geothermal Gradients Bedrock The exact geothermal gradients in Illinois are not known, but they can be expected to vary from one area to another. A commonly accepted value is 1 ° F in- crease in temperature per 100 feet of depth (18.23° C per kilometer). Suter et al. (1959) reported a gradient of 1 ° F per 100 feet in the deep aq- uifers in northern Illinois and only a slight gradient in the shallow aquifers. The shallow aquifers — glacial drift and shallow dolomite — contained water only slight- ly warmer than the mean annual air temperatures. These temperatures were ob- tained mainly from wells in aquifers from which large amounts of water are often pumped. Estimates of geothermal gradient in the Illinois Basin vary slightly. Pryor (1956) used 1 ° F per 100 feet as the gradient in the Illinois Basin. McGinnis (1968) using a large number of data from the oil fields, found the gradient to be about 1.1 ° F per 100 feet (20.05° C per kilometer) with the gradient increasing slightly in the shallower rocks. Loofbourow (1966) suggested an average gradient of 2° F per 100 feet (36.46° C per kilometer) in the oil-producing area of the basin. The geothermal gradient varies from place to place, depending upon rock type, age, and moisture content. In general, the older and more compact the rock, the lower the geothermal gradient; thus, gradients around 0.5° F per 100 feet (9.12° C per kilometer) are recorded for the Canadian Shield, whereas gradients of 2 ° F per 100 feet (36.46° C per kilometer) are common in the Mississippi Embayment of Louisiana (Loofbourow, 1966; Spicer, 1942). Variations in temperature gradients are best explained by considering the nature of heat flow. The quantity of heat (Q) that flows to the surface per unit area depends upon the thermal conductivity (k) of the rocks and the thermal gra- dient ( 8 T/8x): Q=-k(8T/8x) (1) or for a constant gradient through the vertical interval x - x and surface area A: (T 2 ' V Q = kA (2) (X 1 " X 2 } 4 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 where T 2 > T 1 . Under steady-state conditions, the quantity of heat transmitted vertically across any unit thickness of rock is constant. The temperature gradi. ent within each rock unit will vary inversely with its thermal conductivity: k l < S V Sx 2> k 2~< S V Sx i> Equation (3) can be applied to the problem of gradients in the Illinois Basin. The section of rocks from the top of the Cambrian Mt. Simon Sandstone megagroup to the top of the Mississippian Valmeyeran Series is in large part limestone and do- lomite with lesser amounts of shale and sandstone. Above this section, the Mis- sissippian Chesterian and the Pennsylvanian rocks are predominantly clastic, with shale as the major rock type. The thermal conductivities of individual rocks vary. The thermal conductivities reported (presumably water saturated) for limestone and dolomite (Lovering and Goode, 1963; Loofbourow, 1966; Spicer, 1942; Handbook of Chemistry and Physics, 1967) range from 0.0048 to 0.0115 cgs units, with most in excess of 0.0065; sandstones have thermal conductivities of about 0.0055 and shale 0.0042 to 0.0052. Assuming a geothermal gradient of 18.23° C per kilo- meter (1° F per 100 feet), a thermal conductivity of 0.007 in the predominantly car- bonate section of the Illinois Basin, and a thermal conductivity of 0.005 in the pre- dominantly shale section, a gradient of 25.52° C per kilometer (1.4° F per 100 feet) is obtained for the shale. This probably is a reasonable figure because when the two gradients are combined, a geothermal gradient of about 2 0.96° to 21.88° C per kilometer (1.15° to 1.2° F per 100 feet), depending on the ratio of the thick- nesses assumed, is obtained for the rocks of the Illinois Basin. Glacial Drift Measurements of geothermal gradients have not been made in glacial drift and, therefore, the geothermal gradients must be estimated from theoretical consid- erations. Observations of water temperature, generally measured in pumping wells, have been made in a number of areas in the state. In the Chicago region, Suter et al. (1959) reported that the temperature of water from 213 drift and shallow do- lomite wells ranged from 46° to 54° F (7.8° to 12.3° C), averaging 51.6° F (11.0° C), with 71 percent of the temperatures between 50.5° and 52.5° F (10.4° and 11.5° C). The mean annual air temperature of the Chicago region ranges from 48° to 51° F (9.0° to 10.6° C). Walker, Bergstrom, and Walton (1965) reported water temperatures ranging between 53° and 57° F (11.8° and 14.0° C) and averaging 55° F (12.9° C) in the Havana region of Illinois, in areas unaffected by river infiltration. The mean annual air temperature of this region is about 51° F (10.6° C). A tabulation of the data presented by Hanson (1950, 1958, 1961) for munici- pal water supplies in the east-central region of Illinois shows the water tempera- tures from wells 50 to 400 feet deep to range from 53.5° to 55.5° F (12.0° to 13.2° C), averaging 54.5° F (12.6° C). The approximate mean annual air temperature at Champaign-Urbana is 52 ° F (1 1 . 2 ° C) . A linear regression line drawn through the data for east-central Illinois (fig. 1) suggests a gradient of 0.15° F per 100 feet (2.73° C per kilometer), although an argument could be made for a higher gradient by ignoring the points less than 100 feet and greater than 3 00 feet deep. However, TEMPERATURE PROSPECTING FOR AQUIFERS o-- 100- a 200- 300 eters Temperature 12 13 14 i °C °F 52 ' - t Mean annuo/ temperature 54 • 56 • - 25 > > mm • • -50 • • • - 75 • m. ^Regression line -100 • • -125 all temperatures are for water pumped from water wells and suggest either that there is a low geothermal gradient (less than 18.23° C per kilometer, or 1 ° F per 100 feet) or that the aquifers form anoma- lous temperature bodies in the earth. The latter case is probably correct, as will be shown later. Also of significance is the fact that the water temperatures are about 2° F (1.1° C) above mean annual air tem- peratures. If the geothermal gradient in the glacial drift is approached theoretically in the same manner that the discrepan- cies of reported geothermal gradients in the bedrock were resolved, the results are quite different from the results given by water temperatures. Bredehoeft and Papadopulos (1965) and Birch (1942, p. 259) used 0.002 cgs units as a typical value of the thermal conductivity of wa- ter-saturated clay. Loofbourow (1966), Misener, Thompson, and Uffen (1951), and Spicer (1942) give values of thermal conductivity ranging between 0.0021 and 0.0037 for moist to wet soils. Penrod, Elliott, and Brown (1960) give a value of 0.0017 for dry clay soil. Lovering and Goode (1963) found a value of 0.0024 in dry Quaternary gravel in Utah. A value of 0.0025 cgs units is assumed an aver- age value for moist to wet glacial till in Illinois in this report. Assuming values of 0.0025 for the thermal conductivity of glacial till, a thermal conductivity of 0.005 and a geo- thermal gradient of 25.52° C per kilometer (1.4° F per 100 feet) for the underlying Pennsylvanian bedrock, and substituting these values in equation (3), a geothermal gradient of 51.04° C per kilometer (2.8° F per 100 feet) is obtained for glacial till. If the value of thermal conductivity of 0.002 is used for glacial till, a geothermal gradient of 63.81° C per kilometer (3.5° F per 100 feet) is obtained, and if 0.003 is used, a gradient of 43.75° C per kilometer (2.4° F per 100 feet) is obtained. Or, if an average gradient for the whole basin is assumed to be 21.88° C per kilometer (1.2° F per 100 feet), with an average thermal conductivity of 0.006 for bedrock and a drift thermal conductivity of 0.0025, the geothermal gradient will be 52.50° C per kilometer (2.88° F per 100 feet). All of these calculations suggest a geothermal gradient of slightly less than 54.69° C per kilometer (3° F per 100 feet); a value of 51.04° C per kilometer (2.8° F per 100 feet) probably is a good average value in areas of clayey glacial till directly over Pennsylvanian bedrock. 400 Figure 1 - Water temperatures from pump- ing wells in east-central Illinois (from Hanson, 1950, 1958, 1961). 6 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 The discrepancy between the theoretical value of geothermal gradient just obtained and the observed temperatures of water from shallow wells may be used to provide a clue to the presence of shallow glacial and alluvial aquifers. SURFACE EFFECTS Two obvious factors that affect shallow soil temperatures are the diur- nal and seasonal temperature changes. These are both periodic variations, and their fluctuation can be approximated by a sinusoidal curve. A third factor affect- ing near surface temperatures is nonperiodic weather variations, such as warm or cold periods of long duration. The soil stores heat during warm periods and re- leases it during cold periods. The depth to which changes of air temperature affect ground temperatures was investigated in detail by Lovering and Goode (1963) and Penrod, Elliott, and Brown (196 0). Jaeger (1965) states that in most cases, surface effects are negli- gible below a depth of about 65 feet. Penrod, Elliott, and Brown (1960) gave this depth as 100 feet. Lovering and Goode (1963) estimated this depth to be between 3 and 130 feet, depending upon the thermal diffusivity constant and the duration and magnitude of the surface -temperature fluctuations. The thermal diffusivity constant (a) equals the rise in temperature per unit volume produced by a given quantity of heat and is proportional to the ther- mal conductivity (k) and inversely proportional to the specific heat (c) and den- sity (p): a = k/c p . (4) Thermal diffusivities of some common rocks and soils are given in table 1. TABLE 1 - THERMAL DIFFUSIVITIES OF SOME COMMON ROCK AND SOIL IN cgs UNITS* Soils and unconsolidated material: Calcareous earth, 437. water 0.0019 Quartz sand, medium, dry 0.0020 Quartz sand, 8.37. moisture 0.0033 Sandy clay, 157. moisture 0.0037 Soil, very dry 0.0020-0.0030 Some wet soils 0.0040-0.0100 Wet mud 0.0022 Soil, Lexington, Ky. 0.0021 Soil, Lexington, Ky. (avg. 0-10* in place) 0.0072 Gravel 0.0057-0.0062 Rocks; Shale 0.0040 Dolomite 0.0080 Limestone 0.0050-0.0110 (0.0080 avg.) Sandstone 0.0113-0.0140 Granite 0.0060-0.0130 Water: At 0° C 0.00131 At 8° C 0.00169 Commonly used average 0.00143 Air; At 0° C and 1 atm. 178, 000.0000 *From Ingersoll, Zobel, and Ingersoll, 1954; Lovering and Goode, 1963; Penrod, Elliott, and Brown, 1960; National Research Council, 1927. TEMPERATURE PROSPECTING FOR AQUIFERS The mathematical treatment of temperature fluctuations at various depths in the earth is based on the assumption that surface temperature changes generate a heat wave that can be approximated by a sine curve. The distribution of tempera- ture in the earth will approximate a sinusoidal curve and the temperature range at any given depth can be calculated from the following equation, assuming heat moves into the earth material, a semi-infinite body, only by conduction (Ingersoll, Zobel, and Ingersoll, 1954; Lovering and Goode, 1963): OT e r -xNtt/ccP (5) where T x is the temperature range at depth x, T r is^tne total temperature range at the surface, P is period, and a is diffusivity. The above equation is in cgs units: x in centimeters and P in seconds. Table 2 (after Lovering and Goode, 1963) is calculated from equation (5). It shows the effective depth of temperature variation due to diurnal and annual at- TABLE 2 - DEPTH AT WHICH ANNUAL AND DIURNAL TEMPERATURE FLUCTUATIONS ARE 0.1 PERCENT OF SURFACE FLUCTUATIONS, FOR VARIOUS VALUES OF DIFFUSIVITY* Depth to nearest 0.1 foot Depth to nearest foot at at wh LCh diurnal range is which annual range is Diffusivity 0. 17. Jf surface range 0.1% of surface range 0.0016 29 0.0025 1.9 36 0.0036 2.0 43 0.0049 2.6 50 0.0064 3.0 58 0.0081 3.4 65 0.0100 3.8 72 0.0121 4.2 79 0.0144 86 0.0169 94 0.0196 101 0.0225 5.7 108 ♦After Lovering and Goode, 1963. mospheric temperature variation for various values of diffusivity. Temperature var- iations are given as a percentage of temperature fluctuations of the surface. In central Illinois, the mean monthly temperature fluctuation is about 50° F (28.0° C) 0. 1 percent of this is 0.05 ° F (0.03° C). For most considerations, this may be the effective limit of the surface temperature fluctuation. The effect of nonperiodic surface temperature fluctuations also can be approximated using equation (5). This is done by assuming that the long-duration warm or cold period is one -half cycle of a periodic wave. Thus, a one -week heat 8 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 wave, during which the average temperature is 10° C (18° F) above the monthly average, has values of T r = 10° C (18° F) X=£P=(2 x 7 x 86,400). Table 3 shows the approximate temperature or fluctuation, at various depths, in rocks of different diffusivities for a one -week heat wave. TABLE 3 - APPROXIMATE TEMPERATURE RANGE IN DEGREES FAHRENHEIT (° F) AND CENTIGRADE (° C) CAUSED BY A ONE-WEEK HEAT WAVE OF 10° C (18° F) ABOVE MONTHLY AVERAGE* Dif fusivity Depth (in feet) 3.3 6.6 9.9 13.2 16.5 0.0049 2.16° C (3.9° F) 0.23° C (0.4° F) 0.02° C (0.04° F) 0.0100 0.89° C (1.6° F) 0.19° C (0.3° F) 0.04° C (0.07° F) 0.0144 1.65° C (3.0° F) 0.45° C (0.8° F) 0.13° C (0.2° F) 0.04° C (0.07° F) *After Lover ing and Goode, 1963. The temperature resulting from the annual wave front can be calculated for any given point beneath the land surface. To derive such an equation, it is necessary to assume that the soil is homogeneous, its surface flat, and the flow of heat in a direction perpendicular to the surface (Ingersoll, Zobel, and Inger- soll, 1954; Lovering and Goode, 1963; Penrod, Walton, and Terrell, 1958; Penrod, Elliott, and Brown, 1960). It is necessary to solve the Fourier heat equation: Sjr St 8 2 t Sx 2' (6) subject to the boundary condition: m 2tT T_ sin — t, at x = (7) where T is temperature, T is surface temperature range, t is time, P is period, a is thermal diffusivity, and x is the distance from the surface. A particular solution to equation (6) is given by the equation: m + T e r -x WaP sin t - x Wc (8) where T x is the temperature at depth x and T m is the mean annual temperature. Observed soil temperatures closely follow the theoretical curve result- ing from equation (8). Penrod, Elliott, and Brown (196 0) and Flucker (195 8) made a series of soil temperature measurements over a 5 -year period and evaluated equa tion (8) empirically. The main problem involved is the measurement of T m , which TEMPERATURE PROSPECTING FOR AQUIFERS 9 is the surface soil temperature and is slightly different from mean annual air tem- perature. Penrod, Elliott, and Brown (1960) found at Lexington, Kentucky, an av- erage soil temperature of 0.67° C (1.2° F) above mean annual air temperature; Lovering and Goode (1963) in the East Tintic District, Utah, found differences as great as 7.78° C (13.9° F) between surface soil temperatures and air temperatures. After smoothing the air temperature curves, they obtained an average soil temper- ature of 1.90° C (3.4° F) above mean annual air temperature. Flucker (1958) meas- ured soil temperature to a depth of 10 feet over a period of 5 years at College Sta- tion, Texas, and found the average soil temperature to be 2.91° C (5.2° F) above the average air temperature; of significance was the fact that soil temperatures de- creased continuously with depth, and the soil temperatures were higher than the average soil temperature at a depth of 2 feet or less, while at 3 feet and more they were below average soil temperature. Many of the differences and discrepancies between soil temperature measurements from area to area can be attributed to the same factors that cause the differences between air and soil temperatures. These factors are soil cover and color, prevalence of sunshine, wind, snow cover (an insulation from further temperature effects), and soil moisture. Frost in the soil can have a significant effect, as it will hold soil temperature near 0° C (32° F) during cold periods and cause a significant lag in the spring temperature rise because of the latent heat of fusion of the ice. Because they were measured at too shallow a depth, soil temperatures at Champaign-Urbana (table 4) cannot be fitted with confidence to theoretical data. However, the Champaign-Urbana data suggest that equations (5) and (8) are of the correct form (fig. 2) for these data. TABLE 4 - MEAN MONTHLY SOIL AND AIR TEMPERATURES AT CHAMPAIGN-URBANA TO THE NEAREST DEGREE FAHRENHEIT* Month Air Depth below surface 4" 12" 24" 36" January 27 27 33 38 41 February 29 27 32 37 38 March 40 33 38 39 41 April 51 44 47 47 48 May 62 57 52 56 54 June 71 68 68 66 61 July 76 72 71 70 67 August 73 72 71 71 69 September 67 65 64 68 67 October 55 44 53 61 60 November 42 38 47 51 52 December 30 30 37 43 44 Average 52.0 48.1 51.1 53.9 53.5 *After Changnon, 1959. Comparing these data with that of Penrod, Elliott, and Brown (1960) and Flucker (1958), the probable mean annual surface soil temperature is about 10 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 (a) 100 (b) ^* - 10 ^V. 60 / \^ 50 N. - 40 1 1 1 1 1 1 i i i i i r o 5 25 20 15 - 10 Figure 2 - (a) Some temperature depth profiles at Champaign- Urbana (after Changnon, 1959). (b) Mean monthly temperature 24 inches below surface at Champaign - Urbana (after Changnon, 1959). TEMPERATURE PROSPECTING FOR AQUIFERS 11 -50 -40 -30 -20 Percent of annuol temperature range % 15- 30 53.7° F (12.15° C), 1.7° F (0.95° C) above the mean air temperature. The lower temperatures of the 4- and 12 -inch measurements may be due to cooling by evaporation of moisture and the cooling effect of rain, which mostly affects the top of the soil (Flucker, 1958). From equation (8) a set of curves can be calculated, showing soil temper- atures at any depth and any time knowing the surface temperature fluctuation (wave). An envelope that contains all the curves can be calculated, using equation (5). Figure 3 shows three envelopes for dif- ferent diffusivity values. Soil temperature measurements may be used to calculate the value of diffusivity (a) of soils in place. Lovering and Goode (1963) present a graph from which values of (a) may be obtained. Using this graph and the soil tempera- tures at Champaign -Urbana reported by Changnon (1959), the values of diffusi- vity range from 0.004 to 0.0080, with 0.005 the average value. The temper- ature data are not at great enough depth for accurate diffusivity determinations. Figure 4 shows the envelope and some oi the temperature curves for a diffusivity of 0.0049. In calculating the soil tempera- ture curves it is assumed that the land surface (annual wave) is the only heat source or sink (that is, all heat is either gained or lost at soil-air interface), which actually is not the case. The rock below is a constant heat source, as is shown by the geothermal gradient. Therefore, the envelopes should be modified to take account of this factor; this can be done graphically. Figure 4 can be modified to meet the geothermal gradient expected in a glacial till in east-central Illinois. Lovering and Goode (1963) made a temperature measurement through the zone affected by the sur- face wave. Those data show a slight slope of the curves, as might be expected. ^j- «— Below M - 100 v^s - 200 \ V --300 \ \\ x V -400 \ \\ eon Above— ► sy cv'/ il 1 ' /' \ \\ -500 \ \\ \ l i i 1 1 . .-600 \ \ ! / \\ - 700 \\ !j / !• / | M -800~ | 1 i 1 .-900 | 1 !i i -1000 I 1 I i -1100 1 l -1200 1 I -1300 I I -1400 I I I -1500 i Figure 3 - Maximum temperature fluctu- ation at depth in three soils of differ- ent thermal diffusivity (a). Tempera- ture fluctuation is given as percentage of surface annual temperature range (after Lovering and Goode, 1963). TEMPERATURE PROSPECTING The theoretical equations cited above adequately describe the ther- mal properties and expected temperatures in a glacial till sequence. The presence of a saturated sand aquifer in a normal sequence of till will upset the existing ther- mal balance, or, rather, create a new thermal system, which is different from areas where no shallow aquifers exist. The low values of (a) for water (0.0014) and quartz 12 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 Percent of annual temperature range (%T r ) 50 10 15 20 25- 30 35 40 45 50 months 100 200 300 - 400 - 500 600 - 700 800 -900 1000 1100 1200 -1300 -1400 -1500 (Enlarged T r scale) Figure 4 - Theoretical depth-temperature curves and maximum temperature fluc- tuation for soil with thermal diffusivity (a) of 0.0049. The curves show the soil temperature at various times (in months) after the spring crossover when soil temperature at the surface is equal to the annual mean temperature. Temperatures are given as percentage of surface annual temperature range (after Lovering and Goode, 1963). TEMPERATURE PROSPECTING FOR AQUIFERS 13 sand (0.0020 to 0.0033) and the high specific heat of water make the aquifer a heat sink (or source at certain times). Levering and Goode (1963), in attempting to detect high thermal gradients from ore bodies in the East Tintic District, Utah, found that a perched water layer (water table was at 1000 feet) would absorb the heat generating in oxidizing ore bodies and therefore reduce the gradient detected at the surface. The fact that near-surface soil temperatures will be affected by the change in material is obvious; however, the size and nature of the anomaly is of some question. The anomaly either must be large enough to be distinguishable from extraneous temperature fluctuations, or the extraneous fluctuations must be evaluated and/or eliminated in some way from the field data. Much of this can be accomplished by taking readings at a single instant in time or by using a base station to record changes in soil temperatures with time. In all cases, the aquifer will act as a flexure point in the temperature gradient curve, serving as the maxi- mum depth of penetration of surface temperature fluctuation (wave) and the point at which the geothermal gradient is near, or at, mean annual temperature. Thermal Anomaly Caused by Shallow Aquifers \ L ii h ! !! — y- — \ \ ■ To evaluate the effect of a shallow aquifer on the soil temperature near the land surface, the following assumptions are made: (1) the buried aquifer is of sufficient extent that the flow of heat between the top of the aquifer and the land surface can be considered as a one- dimensional transfer process at the point T 2 where the temperature is measured; (2) the aquifer is overlain by nonwater-yield- ing material, which can be considered as a slab of uniform thickness and ther- mal properties and through which heat is transferred between the land surface and the aquifer; (3) the temperature at the upper boundary of the slab (land surface) varies with the sinusoidal yearly temperature wave, and the lower boundary (the aquifer) may be considered a constant temperature boundary. The effect of a shallow aquifer on the temperature of the overlying material can be illustrated graphically by a series of theoretical isotherm and flow lines. The illustration in figure 5, from Ingersoll, Zobel, and Ingersoll (1954, p. 203), graph- ically shows the heat flow through a wall as affected by the presence of an internal projecting rib. It is assumed that the rib has a high conductivity as compared with the wall so that it is an isothermal sur- face taking the temperature T]^ of the sur- face of the wall that it adjoins. This would be similar to an aquifer in the zone Equal potential line (isotherm) Heat flow line > T. T„ < T, i n summer in winter Figure 5 - Isotherms and flow lines for steady heat conduction through a soil (after Ingersoll et al. , 1954, p. 203). 14 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 affected by surface temperature fluctuations where during the cold and hot months the aquifer maintains a temperature that is the same as the temperature at some greater depth. Quantitative analysis of the expected anomaly can be approached by using the general heat equation for a permanent heat source (Ingersoll, Zobel, and Inger- soil, 1954): s- x (2 - n) r™ 9 T = R (n-3) -(3 Z 9 n/2 / P e H d(3 l-n a J (9) 2 J ta where T is the temperature at depth x at time t, S" is the strength of the heat source ; a is diffusivity, and (3 is a variable of integration. For a one -dimensional flow — that is, in the vertical direction only between the heat source and land surface — n = 1 and equation (9) becomes p- 2 e- p2 d P 2au J (10) 2\| ta ' The strength of the heat source, S' , is proportional to the quantity of heat, Q* , lost through the surface and inversely proportional to the density and specific heat (c) of the material transmitting the heat: s- -.2 (ii) cp If the material is uniform, the quantity of heat, Q 1 , is obtained by Q'=Sa < 12) where x is the interval of measurement of the temperature difference (AT), and k is the thermal conductivity. Equation (10) describes the unsteady flow of heat through an infinite slab that is uniformly and suddenly heated at one surface while the other surface is held at a uniform temperature. The equation shows that there is no steady state because the value of the integral becomes very large as time approaches infinity. TEMPERATURE PROSPECTING FOR AQUIFERS 15 The steady- state equation can be derived from the basic Fourier equation, which for one dimension is dx 2 (i3: When integrated, the equation becomes T = Bx + C (14 If the boundary conditions T = Tj at x -J, and T = T2 at x = m are inserted in equa- tion (14) where m and d are, respectively, the distances from the reference datum (yz plane) of the upper and lower surfaces of the slab surfaces (soil-air interface and aquifer-overburden interface), equation (14) becomes T = mT l" iT 2 m (T 1 " T 2 )r (15) where T is the temperature at any point r within the slab (fig. 6). The following assumed reason- able values for the parameters of equa- tions (5) and (15) are based on data ob- tained by McGinnis, Kempton, and Hei- gold (1963), Loofbourow (1966), Inger- soll, Zobel, and Ingersoll (1954), and Clark (1966): plane +VL diffusivity (a), 0.005 cgs units depth to top of the aquifer (x), 500 cm density of overburden {p ), 2.35 specific heat of overburden (c), 0.2 temperature difference (AT), 15° C thermal conductivity of over- burden (k), 0.002. Figure 6 - Geometry of the slab used in equation 15. Plane m is the air-soil interface, plane l/ is the aquifer-over- burden interface, and plane r is a ran- dom plane in the slab parallel to plane m where the temperature is calculated. When these values are substituted in equation (15), a temperature increase of 1 5° C is obtained. This is the max- imum temperature anomaly that can be expected during the coldest and warmest 16 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 months of the year. Figure 7 shows the monthly average temperature variations at 50 cm below the land surface from an aquifer buried 500 cm below the land surface, assuming steady-state condi- tions for each month for the average monthly deviation from the mean annual temperature, based on temperatures ta- ken at Champaign, Illinois. If equation (10) is set equal to equation (15), the approximate time required before a steady- state condition is reached can be solved. For the conditions previously described — that is, for x = 500 cm — a time of 1.4 7 x 10 6 seconds (about 17 days) is required to reach equilibrium. For a larger x (greater depth) the time is greater, and, conversely, for a smal- ler x the time is less. In addition, the aquifer should act as a heat sink (Lovering and Goode, 1963), interrupting the flow of geother- mal heat to the earth 1 s surface. The re- sulting decrease of the geothermal gra- dient should cause a slight cooling of the surface soil, and this is added to the previously described effect of the heat transfer. If a temperature gradient in the glacial till of 54.69° C per kilo- meter (3° F per 100 feet) is used, a de- crease in soil temperature of about 0.05° C (0.09° F) is estimated. This increases the summer anomaly and decreases the winter anomaly, and possibly could be used to detect deeper aquifers. Jan Feb Mar Apr May June July Aug Sept Oct Nov. Dec Jan Figure 7 - Calculated departure from nor- mal temperature of soil at 50 cm depth, resulting from an aquifer atdepthof 500 cm and 15° C maximum difference from surface temperature. Effect of Depth From inspection of equations (9), (10), and (15), it can be seen that increasing the depth to the aquifer will have the effect of decreasing the surface temperature anomaly. The curve in figure 8 was made by solving equations (10) and (15) for numer- ous slab thicknesses (with the IBM 7094 computer at the University of Illinois) for the temperature at 50 cm (19.7 inches) below the surface, which is held at uniform temperature in a material with the properties assumed in the previous section. The curve shows the relation of depth to the size of the anomaly that can be expected at 5 cm below the land surface. To these values must be added the loss of heat due to interruption of geothermal heat flow by the aquifer. The maximum effective limit of the depth of an aquifer that can be detected is probably between 1000 and 2000 cm (32.8 to 65.6 feet), depending on the ability of the operator to distinguish the desired anom- aly from surface temperature changes caused by such surface factors as changes in soil color and character, vegetation, and direction and amount of slope. TEMPERATURE PROSPECTING FOR AQUIFERS 17 Depth Estimates From the discussion of the effect of depth of the buried body on the surface temperature and figure 8, it is immediately obvious that the size of the anomaly is di- rectly related to depth of burial. Actual field observation confirms this general re- lationship. However, the data also sug- gest that thickness and permeability of the aquifer may affect the size of the anomaly. A more accurate method of deter- mining depth can be obtained using equa- tions (8) and (15), although this is sub- ject to similar problems as is the use of the size of the anomaly to determine depth; that is, it ignores the aquifer properties. This method involves the difference in temperature obtained by two simultaneous readings at different depths at the same site. The temperature difference will be equal to the temperature difference due to the normal surface effect (equation 8) plus the temperature difference due to heat flow from the buried aquifer (equation 15): 300 60 80 OO Figure 8 - Relation of depth of an aquifer with a temperature 15° C different from that at surface to the maximum temperature anomaly theoretically observed at a depth of 50 cm. AT = lir/aP . /2-rrt I 7 _V sin (— - x 2 NlTT/aP) - AT 18 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 Equation (16) can be solved for various thickness and temperature dif- ferences between the surface and the aquifer assuming reasonable values for the factors involved. Figure 9 shows a plot of the increased temperature result- ing from a buried body with a tempera- ture 15° C different from the surface, and also the expected difference between readings at 50 and 100 cm, taking into account heat from the anomalous body and normal temperature increases (equa- tion 8). Several factors are difficult to evaluate because they cannot be direct- ly measured. Temperature difference between the surface and aquifer (T1-T2) is the most critical factor and one that is not directly measurable. Values of diffusivity (a) and time of the year after the temperature crossover (t) can be rea- sonably estimated, and small differences are not critical in the final results. Observed temperature difference between depths of 50 and 100 cm 120 1.10 100 90 080 7- 10- - 300 ■ - 500 ^20- Q. S 3 °- 'e - 1000 60- - 2000 PREVIOUS WORK ON TEMPERATURE PROSPECTING Increase in temperature per 50 cm due to buried aquifer 15° C different from surface Figure 9 - Temperature increase per 5 cm (equation 8) and observed temperature difference between depths of 50 and 100 cm (equation 16) resulting from an aquifer with a temperature 15° C different from that at the surface and occurring at vari- ous depths. The curve can be used to estimate the depth of the aquifer. A number of studies have been made of geothermal gradients. Work on the use of temperature prospecting has been limited. Van Orstrand (1940) measured temperatures in wells in the Salt Creek Field in Wyoming and found an increase in geothermal gradient over the structure; he at- tributed the increased gradient to the upwarping of hotter strata toward the land sur- face. Guyod (1946) also noted similar anomalies associated with salt domes and suggested that thermal measurement might be used to detect salt domes; Dobrin (1952) suggested that the reason for this is the very high thermal conductivity of the salt. Stallman (1965) and Bredehoeft and Papadopulos (1965) used soil temperature changes and thermal profiles to determine vertical velocity of ground- water flow in the soil and the vertical permeability of the soil. Kintzinger (1956) noted an anomaly of 12° C (21.6° F) over a hot water area near Lordsburg, New Mexico, where no surface expression of the hot water was present, but super-heated water was encountered in wells at a depth of 78 feet. Lovering and Goode (1963) studied the possibility of detecting oxidizing ore bodies from abnormally high geothermal gradients in the East Tintic District, Utah, and concluded that it was possible, but not practical. Strangway and Holmer (1966), using infrared photography and soil temperature surveys, described thermal anom- alies over several geologic structures and thermal water areas; a 10° F (5.6° C) anomaly was found over an area where 150° F ( 04. U J C) water was reported at a depth of 450 feet (137 meters). &S.Q> C TEMPERATURE PROSPECTING FOR AQUIFERS -/ y ) -y J r O Chicago ,* Spring VH • Mazon Valley — - L T" V "4-r- ,. Mt Pulaski N • \ J Champaign-Urbana Hurricane .Creek Birman (1965) describes a potential method of geothermal prospecting in which temperature probes are buried at a depth of about three meters, well below the zone affected by diurnal temperature variations. Readings are made about a month apart. This method measures the effect of the aquifer on the annual temperature wave. Carr and Blakely (1966) employ a technique similar to Birman' s, but use variation in the diurnal temperature wave. In this method of prospecting, the tempera- ture probe is only 3 cm below the surface and measures the diffusivity (a) of the sur- face material. This paper complements previous preliminary studies in Illinois (Cartwright, 1966, 1968). FIELD STUDIES A number of field studies have been conducted to verify the theoretical calcu- lations given above. To date, nine field studies have been made, of which seven are discussed in this report (fig. 10); the other two studies were exploratory for new well fields, and the data collected thus far are not adequate enough to confirm the re- sults. The size of the anomalies encoun- tered in the field closely approximate the theoretical estimates made in the previous section of this paper. However, the change from winter to summer type anomalies and back again may be more rapid and earlier Figure 10 Location of the field tempera- ture studies. in the seasons than predicted. Measurement and Instrumentation For useful temperature prospecting, the anomaly must be detectable. As has been shown, the main source of rapid temperature fluctuation is the diurnal fluctua- tion, which is effective to a depth of about 3 feet. Strangway and Holmer (1966) made their temperature readings at a depth of 75 to 80 cm (30 to 32 inches) in the soil. In general, diurnal temperature fluctuations are very small at depths of 45 cm (18 inches) or greater, and they probably occur rather slowly. Temperature data for Champaign-Urbana show a summer fluctuation of 5° to 10° F (2.8° to 5.6° C) in the summer and 1 ° to 4 ° F (0.6° to 2.2° C) in the winter at a 4 -inch depth (Changnon, 1959). At a 12 -inch depth, however, the summer fluctuations are about 2° to 3° F 20 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 (1.2° to 1.8° C), and winter fluctuations are generally less than 1° F (0.6° C), commonly less than 0.5° F (0.3° C). Thus, it is possible to eliminate most of the diurnal fluctuations by taking readings at 45 cm (18 inches) or greater in depth. The small temperature changes can be eliminated by taking a large number of measurements at an instant in time. This is not difficult, as the slow fluctuation of temperature at these depths will allow a considerable time lapse, without a significant temperature drift. If, however, the survey is to be large, made over a considerable span of time, it is advisable to set up a base station at which to measure temperature drift. A second method of controlling temperature drift is to make each series of readings overlap. Longer period fluctuations such as hot or cold spells also can be adjusted in the same manner. Other factors affecting soil temperatures, such as color of the soil, vege- tation, etc., cannot be evaluated systematically, nor quantitative relationships given at the present state of knowledge. These factors can be minimized in the field by the choice of station sites. The instrument for measuring soil temperatures is quite simple. It consists of a thermister at the end of an aluminum -tipped stake or probe, which can be driven into the ground (fig. 11), and a transistor-amplified bridge circuit (modified from Radio and Electronics , 1963). This is calibrated in the laboratory to convert micro- volts to temperature. In order to obtain a whole profile at an instant in time, a series of stakes can be connected to a multistrand wire and read individually by turning a selector switch. In general, however, the soil temperature remains sufficiently constant for a period of several weeks so that readings can be made over this span of time with a single probe. Hurricane Creek Hurricane Creek is an alluviated valley 8 miles southeast of Charleston, Illinois (fig. 10). The valley is cut in- to the Illinoian till plain and trends south from the outermost Wisconsinan Moraine, the Shelbyville. The stream, a tributary to the Embarras River, carried outwash gravel, sand, and silt away from the Shelbyville Moraine. The area investigated lies about l| miles south of the moraine. The alluvial fill of the valley is about 45 feet thick (13.7 meters). The maximum thickness of the aquifer is about 38 feet (11.6 meters). The depos- it is underlain by impermeable Pennsyl- vanian age shale and sandstone and overlain by silty, sandy alluvium. The aquifer was located and outlined by an electrical earth resistivity survey (Buhle, 1953). Several large -capacity water wells have been developed in the aquifer by the Forest Oil Company. \ Figure 11 - Instrument used in the field temperature studies (plans modified from Radio and Electronics, 1963). TEMPERATURE PROSPECTING FOR AQUIFERS 21 [a) Resistivity (b) Temperature • Resistivity or temperature station ^/^ Valley wall Line of equal apparent resistivity or temperature, contour interval 2000 ohm -cm and 0.5° C x Wells or test holes A— A* Section or profile, figures 13 and 14 Scale 1000 feet 300 meters Figure 12 - (a) Resistivity (apparent) at 50-foot electrode spacing (from Buhle 1953), and (b) Temperature 18 inches below land surface in June 1966 on Hurricane Creek flat, sec. 28 and 33, T. UN., R. 10 E., Cumberland County, Illinois. 22 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 The resistivity map (fig. 12a) is an isoresistivity map using the apparent resistivity at an electrode spacing of 50 feet in the Wenner configuration. The main part of the aquifer lies within the 8000 ohm -cm contour. The isotemperature map (fig. 12b) of the same deposit is based on 169 readings made during early June 1966. The data are corrected for soil temperature drift to June 8. The contour interval is 0.5° C (0.9° F) . The maximum anomaly is about 1.25° C, which is close to values predicted theoretically (the maximum winter anomaly is 0.75° C — 1.4° F) . The cool anomaly closely fits the resistivity anomaly (fig. 12a, b) and test boring data relative to the location of the aquifer. The two large production wells presently in use lie within the two strong temperature anomalies. The small temperature low on the western side of the valley coincides with a small sand body, The geologic cross section at A-A' (fig. 13) is made from test boring and resistivity data. Directly above the geologic cross section are two temperature West 18.5 r 18.0 17.5 A' East 1.5 1.0 0.5 ! 600 \ ~ 580 - V January 1967 Temperature profile 500 feet — I 150 meters Bedrock B North B' South East West Longitudinal temperature profile Figure 13 - Temperature profiles (winter and summer) and geologic cross section along A-A' (fig. 12) and north-south longitudinal temperature profiles along B-B' (fig. 12), Hurricane Creek. TEMPERATURE PROSPECTING FOR AQUIFERS 23 profiles, one made in the summer and one in the winter. The temperature data are nearly an exact fit. A series of longitudinal temperature profiles were made down the valley (B-B' , fig. 13). The temperatures in the center of the anomaly are com- pared with the temperatures on the eastern and western sides. Not enough is known of the hydrogeology of the deposit to relate the variations in temperature with var- iations in permeability; however, the limited data available suggest this is a pos- sible explanation. The temperature profiles at C-C (fig. 14) are approximately 180 meters (approximately 6 00 feet) south of A-A' . The three profiles, made about a week apart between May 26 and June 8, show the general rise in temperature of the soil. These profiles show that the results are fairly reproducible. A fourth profile made in the winter is also shown. The May 26, June 8, and winter profiles match quite well considering the differences in the number of stations of the profiles. The June 3 profile is a bit erratic, and the thermal high near the center is almost indis- C West 20 c' East 16 - May 26 Scale 250 feet 75meters Figure 14 - Temperature profiles (summer and winter) across the valley along C-C (fig. 12), Hurricane Creek. 24 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 tinguishable. This is attributed to the fact that during the night there was a mod- erate rain, which strongly affected the soil temperature, although it did not com- pletely eliminate the anomaly. Niantic Niantic is in central Illinois about 25 miles east of Springfield (fig. 10). The site is on an outwash plain in front of the Shelbyville Moraine, about half a mile west of the moraine. In the western part of the area, the gravel deposits are confined to the valley of the small stream; the stream valley is essentially indiscernible in the eastern part of the mapped area. The aquifer lies below 15 to 20 feet (4.5 to 6 meters) of alluvium, and overlies Illinoian glacial till. The area has been one of extensive resistivity surveys (Buhle, 195 3; Emery, 1942); the resistivity and temperature maps of the deposit are shown in figure 15. The two maps are strikingly similar but not identical. Test drilling suggests that neither map precisely delineates the aquifer, but they are close. The cross sec- tion at A-A' (fig. 16) shows a close correlation between the geology and the tem- perature anomaly. One of the most prominent features of both maps is the increasing size of the anomaly toward the east, the source of the outwash. This is also illustrated in the longitudinal cross section and temperature profile (fig. 17). This suggests a close relation between the size of the anomaly and the coarseness (and probably permeability) of the material. The anomaly of 1 ° to 1 . 5° C in the summer is approximately what is theo- retically expected. The winter anomaly of 1° C is also reasonable. The winter curves at A-A 1 (fig. 16) illustrate a problem of working in winter months when there is considerable frost in the ground. The profile at 45 cm (18 inches) was strongly affected by the frost, which extended almost to that depth. The profile made at 60 cm (24 inches) the same day shows only a small anomaly. The pro- file made at 100 cm (39 inches) five weeks later, after the construction of a longer probe, shows an anomaly as expected. This points out the need to have the probe well below the frost line. Mazon Mazon is a small town 50 miles southwest of Chicago, situated in a Pleistocene lake flat (fig. 10). The area is extremely flat with less than 10 feet (3 meters) of relief, except where creeks have cut into the surface. The aquifer lies at a depth of about 15 feet (4.5 meters), and there is no surface expression of its presence. The aquifer is fine- to coarse-grained sand, which coarsens to a gravelly sand at the base. The maximum thickness of the aquifer is 13 feet (4 meters). It is underlain by Wisconsinan glacial till and overlain by late Wisconsinan lake silts and clay. The exact origin of the aquifer is not known, but it is presumed to be a pre -lake stream deposit. The resistivity map (fig. 18a) was made at the time of the discovery of the aquifer in 1938 (Buhle, 1938), with a small amount of additional work in the southern part of the area in 1966, after the mapping of the deposit by temperature methods. The temperature anomaly map (fig. 18b) is based on data from 154 stations taken a number of times, making it difficult to correct the temperatures to any one date. TEMPERATURE PROSPECTING FOR AQUIFERS Ufi TTi 25 (b) Temperature -z* — TB 1 Wells or test holes Temperature stations Resistivity profile Line of equal apparent resistivity or temperature, contour interval 2000 ohm cm and l°C Scale 2000 feet 1 i'i i i i r 600 meters Figure 15 - (a) Resistivity (apparent) at 40-foot electrode spacing (from Buhle, 1953, and Emory, 1942), and (b) Temperature 18 inches below land surface in the Niantic area, T. 16 N., R. IE., Macon County, Illinois. 26 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 A West A' East -J&' -February J< * •, s~ X / JanuaryC__J^~-^^/ \ Scole 29 O feet . 50 meters Figure 16 - Temperature profiles (winter and summer) and geologic cross section along A-A' (fig. 15), Niantic area. A geologic cross section of the deposit at B-B' and three temperature pro- files made during different times of the year are shown in figure 19. The profiles show a winter anomaly of about 1.5° C and a summer anomaly of about 1.75° C. The winter profile (3/2 8/66) shows a normal warm anomaly over the deposit. How- ever, the two summer profiles (5/25/66 and 7/6/66) show a warm spot in the cold anomaly in the immediate vicinity of the pumping wells. This is attributed to the draining of the moderately permeable silt, which overlies the aquifer. Profiles at A-A' and C-C (fig. 19), 1000 feet (300 meters) north and south of the well field, show normal cool summer anomalies. TEMPERATURE PROSPECTING FOR AQUIFERS 27 2000 Feet 600 Meters Figure 17 - Longitudinal temperature profile and geologic cross section along B-B' (fig. 15) along the middle of the deposit, Niantic area. The longitudinal profile, D-D' (fig. 19), shows the difference in tempera- tures between the center and the edges of the deposit. For the purpose of this illustration, an attempt was made to correct the temperature to the 7/6/66 tempera ture values. Morrisonville Morrisonville is a small town on the Illinoian till plain about 25 miles southeast of Springfield in central Illinois (fig. 10). The aquifer is thought to be an ice-crevasse deposit, which has been traced almost continuously over a dis- tance of 75 miles by electrical earth resistivity methods. The crevasse deposit seems to bear some relationship to present-day topography, apparently acting as a focus for the location of streams. In the Morrisonville area, the aquifer has a thickness of about 20 feet (6 meters). It is underlain by Illinoian glacial till and overlain by 15 to 30 feet (4.5 to 9 meters) of Illinoian till or, where streams have cut through till, by 10 to 13 feet (3 to 4 meters) of silty alluvium. The map showing the location of the aquifer (fig. 2 0a) was made using well data (Cartwright, 1962) and a reinterpretation of the original resistivity data (Buhle, 1942); the apparent resistivity map does not give an accurate picture of the deposit. The temperature map (fig. 20b) gives a close approximation of the deposit. By test drilling in the vicinity of the aquifer, it was found that the material in the center of the deposit where the wells were built is much more permeable than the material nearer the edges. The maximum temperature anomaly, about 1 ° C, is in the immedi- ate vicinity of the present city well field. The cross section (fig. 21) shows the relationship of the aquifer, as inter- preted from geologic and resistivity data, to the temperature and apparent resis- tivity profiles. The most permeable part of the deposit again seems to coincide with the greatest temperature anomaly. ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 (a) Resistivity anomaly (b) Temperature anomaly 10 II X 12 15 X 14 13 22 ilk 23 24 27 X 26 25 34 36 Anomalous area Wells or test holes • Resistivity or temperature station A A' Section or profile, figure 19 Scole 3000 fee. 1000 meters Figure 18 - (a) Resistivity (apparent) anomaly at 30-foot electrode spacing (from Buhle, 1938), and (b) Temperature anomaly 18 inches below land surface near Mazon, T. 32 N., R. 7 E., Grundy County, Illinois. TEMPERATURE PROSPECTING FOR AQUIFERS 29 — 580 / City wells 560 c o I 540 Till — Bedrock — ^^Aquifer Scole 2000 feel T . '. i ■ '. .' 600 meters A West A' Easf 16 ^M 13 ^* \ / r 14 L v j H n D North Figure 19 - Temperature profiles (summer and winter) along A-A' ,B-B' , and C-C (fig. 18), a geologic cross section along B-B' , and longitudinal temperature profile along D-D' through the western and eastern edges and center of the deposit near Mazon. Mulberry Grove Mulberry Grove is a small town on the Illinoian till plain 55 miles north- east of East St. Louis in south-central Illinois (fig. 10). The deposit is in Hurricane Creek (not the same Hurricane Creek as the first example) east of the town (Pryor, 1955). The creek has cut through the till and into the impermeable Pennsylvanian age shale and sandstone, over which is a maximum of about 3 feet (9 meters) of alluvial material in the valley. The aquifer, with a maximum thickness of 1 1 feet (3.3 meters), varies rapidly in character from fine-grained sand to coarse-grained sand and gravel 30 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 < -\*yr /x./jQ\ o Olifi / / v\ ff> cn|*o o \ o N^o>X^ / "< ^ (T> Iz. 2lK! % 00 v^Vv/ ■f~~\&> F N to < 00 5 o V> ■o c 1 Morrisonville % //////////////A u i* N 1° CM r- n \ "s te" ciin ,_ O "•- "5 >< C -M (0 c p - o m O to c °> (0 » OT x; i- & ^ £ o $ . fl* u -— CO (0 • ti &. (0 T3 - a 2 Cn O ^ o Rh T3 • (0 Z3 "-- > 2 c ^ o en £ u, 1/1 0» F P tH o o CQ ro w CD £ a) > o ■rj 0) U xi c £ o j2 -a :> 2 a> -Q ^3 « £ 4-i 540 Till ^^ Alluvium / c,f y wells Till y f^?222i— \ : :°:{o Aquifer >.'.'••; °. • •T r ~~?~ Bedrock 600 feet I — "-r 1 — S— 1 — \ — '-i 200meters Figure 21 - Temperature and resistivity (apparent) profiles and geologic cross section along A-A' (fig. 20) near Morrisonville. (Cartwright, 1963). It is generally underlain by the bedrock and overlain by silty clay alluvium. The deposit appears to be very narrow, about 100 feet (30 meters) wide or less, and sinuous in pattern, probably being point bar on stream channel deposits. It is also possible that the deposit may be a series of dis- continuous sand bars. The Mulberry Grove deposit (fig. 22a) is the only one on which the tempera- ture survey was not entirely successful, as was also the case with the resistivity surveys of the site. Both summer and winter surveys were made (fig. 2 2b, c). An anomaly of as much as 1.5° C was observed in both summer and winter. Some of the same warming effect around pumping wells was observed as at Mazon. The profiles at the line of cross section (fig. 23) show an anomaly of 1.0° C in the summer and 0.8° C in the winter over the deposit. The western side of the deposit is not as sharply defined as the eastern side. Mt. Pulaski Mt. Pulaski is a small town on the Illinoian till plain 25 miles northeast of Springfield (fig. 10). The well field is located l\ miles north of town in the valley of Salt Creek (Buhle, 1959), which carried outwash from the Wisconsinan ice at its maximum extent located about 12 miles to the east. 32 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 / ^ .^ S~ contour interval 5 feet x Wells or test holes A-A' Section or profile, figure 23 „ Winter temperature, **" contour interval 0.5° C • Temperature station _ Summer temperature contour interval 1° C • Temperature station Scale 100 meters Figure 22 - (a) Thickness of sand, (b) Summer temperature, and (c) Winter temperature 18 inches below land surface near Mulberry Grove, sec. 6, T. 5 N., R. 1 W. , Fayette County, Illinois. TEMPERATURE PROSPECTING FOR AOUIFERS 33 60 meters A 490 ^480 a 470 > February Road Hurricane Creek 460 L Clayey and s illy Alluvium *>.' - < > N J V /Aquifer-^;, t >'-'"" ^ Impermeabl e >'' \ ? /.-■ / •■„ : ' i v k V/ bedrock Figure 23 - Temperature profiles (summer and winter) and geologic cross section along A- A' (fig. 2 2) near Mulberry Grove. The aquifer lies at a depth of 6 to 10 feet (2 to 3 meters) below the surface, and ranges up to about 30 feet (9 meters) thick. The deposit consists mostly of clean, medium- to coarse-grained sand commonly with fine-grained gravel in the upper half of the aquifer. Sand and gravel is present to some extent over the en- tire valley. The Mt. Pulaski survey is the only temperature survey made entirely in the winter months (February-March 1967). The resistivity map (fig. 24a), thermal map (fig. 24b), and cross section (fig. 25) show close agreement. The maximum anomaly is about 3° C, although it is generally slightly less than 2° C. There is an addi- tional 1° C temperature difference between the upland and valley areas. The boun- dary of the principal water-yielding area is drawn on the 10, 000 ohm-cm apparent resistivity at the 40 feet spacing of the electrodes; this matches very well with the area of maximum temperature anomaly. Spring Valley Spring Valley is in Bureau County on the bluff north of the Illinois River, about 90 miles west -southwest of Chicago (fig. 10). The city draws water from three ground -water sources, one of which is a shallow sand and gravel deposit on the western edge of the city. 34 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 Center — *. Sec. 35 ^\00 . — Contour interval 50 ohm-meters with supplemental 25 ohm -meter contour. A~A Section or profile, figure 25 — 2- Contour interval 0.5° C, temperature March 10, 1967 X Test holes and wells • Resistivity or temperature station ^ Gravel pit 600 feet | L_l L_I l_ Scale 200 meters Figure 24 - (a) Resistivity (apparent) at a 40-foot electrode spacing, and (b) Temperature 18 inches below land surface near Mt. Pulaski, sec. 35, T. 19 N. , R. 2 W. f Logan County, Illinois. TEMPERATURE PROSPECTING FOR AQUIFERS 35 Temperature (Feb.) w 150 i ioo H E 50H A North 620- % 600 Of 0) uj 580 Apparent resistivity 560 A' South © * -6 . • -.?• : Till ? -*- ^- • o V 77 777 Bedrock Scole 600 feet i ■ i ■ 'i ' ' — H „ 200 meters Figure 25 - Temperature and resistivity (apparent) profiles and geologic cross section along A-A' (fig. 24) near Mt. Pulaski. 36 ILLINOIS STATE GEOLOGICAL SURVEY CIRCULAR 433 The area is one of Wisconsinan ground moraine with Illinoian till exposed in low ground (Cady, 1919). The aquifer lies under the sloping ground where most of the Wisconsinan till has been removed. The deposit itself lies at a depth of 2 to 40 feet (6 to 12 meters), depending upon surface elevations. It has a maximum thickness of about 15 feet (4.5 meters) and is overlain by glacial till and underlain by Pennsylvanian shales. The deposit is fairly well outlined by resistivity surveys and drilling (Buhle, 1945) (fig. 26a). The temperature survey, made late in March, was not very ex- tensive. It shows a warm summer type anomaly (fig. 26b) instead of a winter type anomaly, although the soil is still rather cool. The temperature profile (fig. 2 7) shows a rather erratic pattern, suggesting the system had not yet entirely stabi- lized. The early change to a summer type anomaly is attributed to a warm March, during which the soil had generally begun to warm up; the area not underlain by the aquifer generally warmed faster than the areas over the aquifer. Some frost was still in the ground over the aquifer, and large variation was due to different soil colors and soil cover. CONCLUSIONS The theoretical consideration of some of the thermal properties of shallow alluvial and glacial aquifers and the properties of the overburden suggests that aquifers can be detected at the land surface by anomalies in the soil temperature. The theory is reasonably well substantiated by field studies of known aquifers. Calculations of depth to the top of the aquifer are estimates at best by either meth- od proposed (size of the anomaly or two-point method), as many of the data needed are based on an estimate of the parameters involved. The field studies show that large areas can be covered in a relatively short time. However, long time lapses (in excess of three or four weeks) cannot be com- pensated for by simple measurements of soil temperature drift; changes in the whole system must then be taken into account. Temperature variations caused by changes in vegetation and shade were the two most difficult problems faced in the field. Variations in soil temperature caused by these factors were as large as, and sometimes larger than, the anomaly due to the aquifer. Buried pipelines or other conduits may also cause problems, and are not as readily identified in the field. By careful field work, these problems generally can be reduced. Rain can also reduce or completely erase the soil temperature anomaly re- sulting from a shallow buried aquifer. The downward movement of water of uniform temperature will erase the anomaly temporarily, or reduce it to a size that is almost indistinguishable. This is partly dependent on the permeability of the soil and the amount of rain; generally the more permeable the soil, the more easily the anomaly is lost by infiltration of rain water, and the less precipitation necessary to erase the anomaly. Conversely, areas of ground-water discharge can produce an anomaly similar to that expected from an aquifer, when no aquifer is present. Frost also is a problem in the winter because the formation of ice crystals acts to hold the soil temperature close to the freezing point of water. This prob- lem is easily overcome by measuring the soil temperatures well below the frost zone (at least 25 cm and preferably 5 cm below). TEMPERATURE PROSPECTING FOR AQUIFERS 37 • 1 L \ s*\ \ */ i 9c • J "< *** \ \ \ \ \ \ \ \ \ ** / * < "3 m "' s to CM en 3 7 <-> a> c O a o 2 o > *o o 7n ? a> Q. a> ^ c O 3 o n 0) D O a) Q. o Q. t- r O h F o CO r 1 ^\ < < £ E o.o 04 CD T-j -H s *° .u > 9> c c -3 -rH U u a (0 oo a ,.. w (0 U (0 0) ^ +j c O (0 o -< O O ""> t! CD ■ M J3 (0 a) W ■r-( ^ CD C -C CD O 1 1 (0 -1h 1 — 1 a a 00 (0 ^ CD C >^ h JJ £ o > (0 o -rt u w a It) w g