Soil Science Society of America Journal 67:487-493 (2003)
© 2003 Soil Science Society of America
DIVISION S-1—NOTES
Temperature dependence of infiltration rate during large scale water recharge into soils
Chunye Lin,
Dan Greenwald and
Amos Banin*
Dep. of Soil and Water Sci., The Hebrew Univ. of Jerusalem, P.O. Box 12, Rehovot 76100, Israel
* Corresponding author (banin{at}agri.huji.ac.il)
 |
ABSTRACT
|
|---|
We report the results of a systematic long-term study of infiltration rate (IR) in a large scale effluent recharge plant, showing a significant dependence of the infiltration rate on temperature (T). Water level and T were continuously monitored and recorded in several infiltration basins of an operating wastewater treatment plant (WWTP) during the course of a 4-yr study of basin geochemistry and performance. Infiltration rates were calculated from the slope of linear plots of water level vs. time during the drainage phase. Systematic interseasonal variations of IR were observed and were strongly correlated to water T. Calculations showed that the variation of IR with T was generally 1.5 to 2.5 times larger than that predicted from effluent viscosity changes per se, suggesting the possible involvement of other T-dependent factors. This may have profound effects on the overall efficacy of wastewater reclamation and other water-recharge operations.
Abbreviations: IR, infiltration rate • SAT, soil aquifer treatment • WWTP, wastewater treatment plant • RIR, relative infiltration rate • ORIR, observation-based relative infiltration rate • RRKV, relative reciprocal kinematic viscosity • T, temperature • VRIR, viscosity-based relative infiltration rate
|
INTRODUCTION
|
|---|
SOIL AQUIFER TREATMENT (SAT), as an extensive, low technology, and low cost wastewater treatment method, is based on rapid infiltration of pretreated wastewater to a local aquifer by means of infiltration basins. Maintaining high IR is of prime importance in ensuring the efficacy of large scale wastewater renovation plants utilizing the SAT and similar protocols such as river-bank filtration and percolation-infiltration (Lazarova et al., 2001; Banin et al., 2002).
The IR of a natural porous body depends on its sorptivity and saturated hydraulic conductivity, which in turn is a function of the intrinsic permeability of the medium and the fluidity of the penetrating liquid (Hillel, 1980b). At the initial stages of infiltration, soil sorptivity is the primary factor affecting IR; but at prolonged infiltration times, the hydraulic conductivity becomes the controlling factor. Previous studies addressed a number of causes leading to decreases in IR of soils including physical clogging (Siegrist, 1987), biological clogging (Vandevivere and Baveye, 1992), entrapped air (Jarrett and Fritton, 1978; Seymour, 2000; Wangemann et al., 2000), surface sealing (Moore, 1981), and dispersion and swelling of soil clay (Quirk and Schofield, 1955; Keren and Singer, 1998; Moutier et al., 1998). Other research focused on the possible effect of T on hydraulic conductivity of porous bodies (Duley and Domingo, 1943; Constantz, 1982; Hopmans and Dane, 1986a; Constantz et al., 2001). Theoretically, everything else being equal, IR into a given soil will decrease as the T of the system decreases since viscosity of the percolating water will increase. Water viscosity changes by 
2% per degree Celsius in the relevant environmental T range of 15 to 35°C, leading to an estimated 40% change of IR between summer and winter in arid zones. However, laboratory studies have given conflicting results. While some studies showed a greater T dependence of unsaturated hydraulic conductivity than predicted from changes in the water viscosity (Flocker et al., 1968; Constantz, 1982), others concluded that the T effects on the hydraulic conductivity were close to the predictions from viscosity changes albeit somewhat larger than expected (Hopmans and Dane, 1986a,b).
The majority of the previous studies on T effects were conducted under controlled conditions in the laboratory using small soil columns. The present study reports the results of a systematic study of IR and its seasonal change across a 4-yr period in a large scale effluent recharge operation, and shows a repeating pattern of cyclical changes of IR depending on T. We have found that the T effect tends to be larger (by a factor of 1.5–2.5 times) than that expected from effluent viscosity changes per se, suggesting the involvement of other factors. This may have profound effects on the performance of the infiltration basins and the efficacy of the reclamation operation.
|
Materials and Methods
|
|---|
Description of Experimental Site
The Dan Region Reclamation Project (Shafdan) is the largest WWTP in Israel, operating since 1977 and currently serving a population of 1 200 000 people in the Tel-Aviv metropolitan area (Fig. 1)
. The plant treats 100 million m3 of wastewater per year with plans underway to expand its capacity to 140 million m3 yr-1. Raw municipal wastewater is treated in a mechanical-biological plant using an activated sludge process, involving nitrification/denitrification steps (Kanarek and Michail, 1996). As a final step in the process, effluents are recharged to the Coastal Plain Aquifer using large-scale infiltration basins. Each of the basin fields (101–104) consist of 4 to 5 leveled subbasins, separated from one another by earthen dams. Subbasin area ranges between 8 000 to 25 000 m2 (Fig. 1). The recharge cycle consists generally of 1 to 2 d of flooding and 4 to 5 d of drainage and drying. During the recharge phase of the cycle, all or the majority of the subbasin surface area is flooded. The average hydraulic load [defined as the volume of recharged effluents per basin (m3 yr-1) divided by the basin's area (m2)] is 50 to 100 m yr-1, representing an input rate of 100 to 200 times the natural input by rainfall. The overlying coastal sand dunes are loose to partly consolidated deposits, rich in quartz, contain small amounts of accessory aluminosilicate minerals, and have a high hydraulic conductivity. The Coastal Plain Aquifer consists of clastic sediments, mainly calcareous sandstones (aeolianites), of Pleistocene age. Recovery wells collect the treated water, mixed with local groundwater, at a distance of 1 to 2 km from the centers of the recharge sites. The water is used for large-scale unlimited irrigation in the southern part of Israel.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 1. The layout of the infiltration basins in the Soreq site of the Dan wastewater treatment plant near Tel-Aviv, Israel. Infiltration rate was measured from December 1997 to December 1999 in Basin-field 103 (Subbasin 103-4/5, leveled at 40 m above sea level) and from March 2000 to March 2002 in Basin-field 102 (Subbasins 102-1 to 102-5, leveled at 45 m above sea level).
|
|
Measurement of Infiltration Rate and Effluent Temperature in the Basins
Recharge dynamics and IR were measured in several subbasins at the Soreq site in the Shafdan WWTP (Fig. 1) across a 4-yr period, during the course of study of geochemical processes in the basins soils (Banin et al., 2002). Between December 1997 and December 1999, water levels and air T were respectively continuously monitored in basin 103-4/5 by a Sonic Ranger (DT020) and Thermistor NTC (DT012; Fourier Systems Ltd., Rosh Ha'ayn, Israel) and recorded by a portable datalogger (DB-526, Version 4.3, Fourier Systems Ltd.). Between March 2000 and March 2002, level measuring devices (Models D1212 and TD1220 Divers, Van Essen Instruments, Delft, The Netherlands) were installed in the four subbasins of Basin 102 (102-1, -2/3, -4, and -5) to continuously monitor and record water level. One of the devices (TD1220) installed in Subbasin 102-5 (or moved to Basin 102-4 for a short period) also measured effluent T. Effluent level and T readings were taken every 5 min.
Infiltration rate in each recharge cycle was calculated from the slopes of the highly significant linear regression functions fitted to plots of water level vs. time during the drainage phase of each recharge cycle. Average T for the wet period of each cycle was calculated and was used in the data analyses to represent the T of the cycle. Effluent T during the period of December 1997 to December 1999 was calculated from measurements of air T using a regression equation between air T and effluent T calculated from data measured during the period of March 2000 to March 2002.
Data Analysis
Recharge regimes and hydraulic loads were adapted to basin characteristics and were varied during the 4 yr of the study according to the operational needs of the plant. The recharge regimes applied to the studied basins during the observation period are given in Table 1.
Relative infiltration rate (RIR) per cycle was calculated using measured smoothed IR curves for each basin. Average measured IR at 25 ± 0.5°C effluent T (IR25oC) was used as the normalization basis. IR25oC values varied in the basins due to differences in their intrinsic permeability and variations of the recharge regime. The values used as the calculation basis for the various basins at different periods are given in Table 2. The IR of 103-4/5 basin at 25.3°C in April to May 1998 was 15.6 cm d-1, which was used as reference to calculate the RIR for the period of December 1997 to April 1998. The possible cause for this low IR was a flood of drainage rainwater from the adjacent municipality of Rishon-Lezion that penetrated the basin in December 1997 and deposited fine eroded soil material on the surface. Partial removal of the top layer and surface cultivation during the summer of 1998 restored the permeability and IR of 103-4/5 basin. In November 1998 the IR was 26.0 cm d-1 at 25.3°C, much higher than the reference value observed in April to May 1998. To account for this large variation, we estimated the reference IR value for the period of May to November 1998 by the following interpolation equation: IR25°C (cm d-1) = 0.0504 x date - 1795. From December 1998 to December 1999, the IR taken as reference for 103-4/5 basin was the average of IRs measured at 25 ± 0.5°C in this period, as given in Table 2. A major change in the recharge regime in Basin 102 in July 2000 dictated the selection of different IR25°C values for the period of March to June 2000 (1–d flooding/5-d cycle) from those used for the period of July 2000 to February 2002 (2-d flooding/7-d cycle). The average values and their statistical characterization for each of the subbasins are given in Table 2.
|
Results and Discussion
|
|---|
Seasonal Change of Infiltration Rate and Effluent Temperature
Measured IRs and average effluent T per recharge cycle during 4 yr of basin operations in the Shafdan WWTP are presented in Fig. 2
, as smoothed (five points moving average) curves of IR vs. time and T vs. time. Seasonal repetitive changes of IR were clearly observed in the five subbasins across the period of experimentation. Infiltration rate in the 103-4/5 basin increased from 10 cm d-1 in the winter of 1997-1998 to 25 cm d-1 in summer of 1998, then again decreased to 15 cm d-1 in the winter of 1998-1999, increased to 30 cm d-1 in summer of 1999, and finally decreased to 20 cm d-1 in December 1999. The seasonal cyclical changes of IR were seen also in the 102 basins, although the pattern was somewhat affected by the drastic change of the recharge regime on 1 July 2000. A trend of increasing IR from 100 cm d-1 in the spring of 2000 to 160 cm d-1 in the summer of 2000 was observed. During the winter of 2000-2001, IR in the 102 basins decreased to 60 to 80 cm d-1. The pattern of seasonal change of IR continued during 2001, except in the subbasin 102-2/3. This subbasin did not operate between 18 Feb. and 30 Mar. 2001 due to various experimental operations in the basin. As a result, recharge loads in the other 102 subbasins were extremely high (up to 5.6 m per cycle; Table 1b), resulting in erratic operational patterns and large IR variations. Therefore, data for these two months were omitted from Fig. 2 and from the following analysis and discussion. The pronounced and striking correspondence between the repeating pattern of variation of IR and that of the seasonal effluent T observed in Fig. 2 will be the basis for the following analysis of the cause(s) of the seasonal change in IR.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 2. Infiltration rate (IR) per cycle and average effluent temperature (T) per cycle measured in large scale recharge basins at the Shafdan wastewater treatment plant across a 4-yr period. Curves were obtained by smoothing (five-point moving average) of the measured data. Dashed lines indicate recharge regime change in Basin 102 on 1 July 2000 (see Table 1). F = fall, Sm = summer, Sp = spring; W = winter.
|
|
Quantitative Dependence of Infiltration Rate on Effluent Temperature
The increase of hydraulic conductivity with water T is commonly attributed to a decrease in the viscosity of water. Soil hydraulic conductivity (K, cm s-1) may be split phenomenologically into two factors (Hillel, 1980a),
 | [1] |
where k is the intrinsic permeability of the soil (cm2) and f is the fluidity of water (cm-1 s-1). Fluidity is inversely proportional to viscosity and is given by
 | [2] |
where 
w is water density (g cm-3), g is gravitational acceleration (cm s-2), 
is the viscosity (g cm-1s-1) and ' = /w is the kinematic viscosity (cm2 s-1). Thus, T elevation is predicted to decrease viscosity, increase fluidity, and overall, increase hydraulic conductivity. Such effect was indeed reported in a number of laboratory studies as reviewed above. However, the magnitude of the change differed considerably among the reports and in some cases K changed by orders of magnitude more than predicted from viscosity change alone. Constantz (1982) listed several possible causes, related to dependence of various water and soil properties on T, for the greater-than-expected dependence: (i) the surface tension change caused by T; (ii) much greater T dependence of viscosity of soil water than of free water (the anomalous surface water approach); (iii) the change of diffuse double-layer thickness with T; (iv) T-induced structural changes; and (v) isothermal vapor flux being more significant than previously thought. Hopmans and Dane (1986b) observed that the entrapped air volume decreased with increasing T, so T might affect IR by changing both water viscosity and the liquid-conduction properties of the soil due to variations in the entrapped air content. Even though Hopmans and Dane (1986a) concluded that the effect of T on the hydraulic conductivity in their results was close to predictions from viscosity changes, hydraulic conductivity was still more than two to six times larger than the predicted values in the low water content and high T regimes. Several recent studies (e.g., Weir and Kissling, 1992; Shan, 1995; Wang and Feyen, 1997, 1998) have analyzed water infiltration coupled with air movement and release from the soil. It was pointed out that the system is complex but that simplified approaches that enable successful numerical prediction of observed variations in IR in various systems involve consideration of air permeability of the soil. Since air permeability is affected by air viscosity, T variations may affect IR more than predicted by its effect on water viscosity. It is beyond the scope of this note to analyze in depth the causes of the variations in IR in the basins. However, given the obvious cyclical nature of the seasonal variations and their correlation with the seasonal T cycles, we will analyze the data with the aim of semiquantitatively assessing the effect of water viscosity changes on the observed variations. It should be further noted in this context that although slight seasonal changes in effluent salinity do occur, their possible effects on the soil's intrinsic permeability and on the water viscosity are small to negligible compared with the predicted effects of the T changes.
Relative infiltration rate and relative reciprocal kinematic viscosity (RRKV), as a measure of water fluidity, were calculated and their correlation with T are given in Table 3. Infiltration rate and viscosity values at 25°C were taken as the normalization bases. Temperature dependence of effluent viscosity was calculated by the Guzman-Andrade equation
 | [3] |
where the coefficients A1 and B1 have been determined to have values of 1.98404 x 10-6 Pa s and 1825.85°C, respectively, by curve fitting to experimental data (Perry and Chilton, 1973). Measurements in our laboratory have shown that effluent viscosity is equal to pure water viscosity to within 0.1% in the T range relevant to this study (15–35°C) (Greenwald, 2000).
View this table:
[in this window]
[in a new window]
|
Table 3. Linear regression equations correlating the relative infiltration rate (RIR) and effluent temperature (T) across a 4-yr operational period in the Shafdan wastewater treatment plant. The functional change with T of the relative fluidity of the effluent (RRKV = reciprocal of relative kinematic viscosity), normalized to its value at 25°C, is also shown.
|
|
Highly significant linear correlations between the RIR and effluent T were observed for the five subbasins, based on the IR and T measurements during 4 yr (Table 3). Slopes of the regression equations of RIR vs. T were in the range of 0.030 to 0.050, compared with a slope of 0.020 for the dependence of RRKV on T. The increase of the IR with water T is thus 1.5 to 2.5 times higher than the decrease of the kinematic viscosity with water T. The overall calculated T dependence of IR in the effluent recharge basins was similar to that reported by Hopmans and Dane (1986a) for hydraulic conductivity variation in controlled laboratory studies. The repeated seasonal pattern of IR change indicates that the processes controlling the changes in IR are quantitatively connected to the system properties and, most likely, are controlled by its T in a replicated and nonarbitrary mode. Our findings also indicate that changes of viscosity and fluidity are the primary factor causing the seasonal variation in the IRs.
Finally, we define the observation-based T-dependent RIR (ORIRT) by the equation
 | [4] |
where a and b are the slope and the intercept values in the linear regression equations given in Table 3. This way, we isolate the T-dependent IR variations from those due to the various other operational variables, applying the regression equations obtained for each subbasin and using the measured T for each recharge cycle as the independent variable. The results are plotted in Fig. 3
. For comparison, we also include in Fig. 3 the calculated viscosity-based relative infiltration rate (VRIRT) obtained by the equation
 | [5] |
where 25°C and T are water viscosity at 25°C and any given T in the range of 15 to 35°C, respectively. The VRIRT solely takes into account the effects of T changes on water viscosity in the calculation of the RIR and shows the expected seasonal variability of IR in each basin if only water viscosity changes would have affected it. It is clearly observed (Fig. 3) that the amplitude of the ORIRT seasonal wave is always larger than that of the VRIRT wave. It is also observed that the predicted RIR tends to decrease in the winter more than it increases in the summer when compared with the one calculated based on water viscosity only. Changes in the entrapped air content of the basins, due to limited or slow outflow of air during flooding in the cold season might result in significant decrease of the recharge rate in addition to the decrease caused by effluent viscosity changes per se. These changes may create an operational bottleneck in the fall and winter seasons and should be taken into account in the planning and management of large-scale recharge operations of water into soils.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 3. Observation-based seasonal changes of measured relative infiltration rate (ORIR) and calculated viscosity-based relative infiltration rate (VRIR) in the Shafdan wastewater treatment plant. RIR = ORIR + VRIR; F = fall; Sm = summer; Sp = spring; W = winter.
|
|
|
Conclusions
|
|---|
Infiltration rates measured during numerous effluent recharge cycles across a 4-yr period in the Shafdan WWTP showed a repetitive seasonal pattern. Relative infiltration rate, which normalized the measured IR values of each basin to the value measured at dates when the T was 25°C, peaked in the summer, and reached a minimum in the winter. The general shape and the time of the maxima and minima of the seasonal RIR wave were significantly correlated with the shape and timing of maxima and minima of the seasonal effluent-T curves. Temperature-dependent changes in the viscosity of the infiltrating effluents accounted for much of the variability. However, the amplitude of the RIR interseasonal wave was generally 1.5 to 2.5 times larger than predicted by water viscosity changes alone. It is possible that changes of the viscosities of the two counter-flowing fluids, water and air, jointly contribute and affect the change of IR with T. Further studies of the affecting mechanism(s) are needed to fully quantify the T effects on large scale infiltration operations into soils.
|
ACKNOWLEDGMENTS
|
|---|
Financial support by Mekorot National Water Co., Israel, and fellowships from the Hebrew University to C. Lin and D. Greenwald are gratefully acknowledged.
Received for publication May 21, 2002.
|
REFERENCES
|
|---|
- Banin, A., C. Lin, G. Eshel, K.E. Roehl, I. Negev, D. Greenwald, Y. Shachar, and Y. Yablekovitch. 2002. Geochemical processes in recharge basin soils used for municpal effluents reclamation by the soil-aquifer treatment (SAT) system. p. 327–332. In P.J. Dillon (ed.) Management of aquifer recharge for sustainability. Proc. 4th Int. Symp. of Artifical Recharge (ISAR4). A.A. Blakema, Rotterdam, The Netherlands.
- Constantz, J. 1982. Temperature dependence of unsaturated hydraulic conductivity of two soils. Soil Sci. Soc. Am. J. 46:466–470.[Abstract/Free Full Text]
- Constantz, J., D. Stonestrom, A.E. Stewart, R. Niswonger, and T.R. Smith. 2001. Analysis of streambed temperatures in ephemeral channels to determine streamflow frequency and duration. Water Resour. Res. 37(2):317–328.
- Duley, F.L., and C.E. Domingo. 1943. Effect of water temperature on rate of infiltration. Soil Sci. Soc. Am. Proc. 8:129–131.
- Flocker, W.J., M. Yamaguchi, and D.R. Nielsen. 1968. Capillary conductivity and soil water diffusivity values from vertical soil columns. Agron. J. 60:605–610.[Abstract/Free Full Text]
- Greenwald, D. 2000. Effect of chemical and physical processes in the basins soils on recharge rates in the Shafdan plant. M.S. thesis, Faculty of Agricultural, Food, and Environmental Quality Sciences, Hebrew Univ. of Jerusalem, Rehovot, Israel.
- Gupta, R.P., and D. Swartzendruber. 1964. Entrapped air content and hydraulic conductivity of quartz sand during prolonged liquid flow. Soil Sci. Soc. Am. Proc. 28:9–12.
- Hillel, D. 1980a. Fundamentals of soil physics. Academic Press, New York.
- Hillel, D. 1980b. Applications of soil physics. Academic Press, New York.
- Hopmans, J.W., and J.H. Dane. 1986a. Temperature dependence of soil hydraulic properties. Soil Sci. Soc. Am. J. 50:4–9.[Abstract/Free Full Text]
- Hopmans, J.W., and J.H. Dane. 1986b. Temperature dependence of soil water retention curves. Soil Sci. Soc. Am. J. 50:562–567.[Abstract/Free Full Text]
- Jarrett, A.R., and D.D. Fritton. 1978. Effect of entrapped soil air on infiltration. Trans. ASAE 21(5):901–906.
- Kanarek, A., and M. Michail. 1996. Groundwater recharge with municipal effluent: Dan Region Reclamation Project, Israel. Water Sci. Technol. 34(11):227–233.
- Keren, R., and M.J. Singer. 1998. Effect of low electrolyte concentration on hydraulic conductivity of sodium/calcium-montmorillonite-sand system. Soil Sci. Soc. Am. J. 52:368–373.
- Lazarova, V., B. Levine, J. Sack, G. Cirelli, P. Jeffrey, H. Muntau, M. Salgot, and F. Brissaud. 2001. Role of water reuse for enhancing integrated water management in Europe and Mediterranean countries. Water Sci. Technol. 43(10):25–33.
- Moore, I.D. 1981. Effect of surface sealing on infiltration. Trans. ASAE 24:1546–1552.
- Moutier, M., I. Shainberg, and G.J. Levy. 1998. Hydraulic gradient, aging, and water quality effects on hydraulic conductivity of a vertisol. Soil Sci. Soc. Am. J. 62:1488–1496.[Abstract/Free Full Text]
- Perry, R.H., and G.H. Chilton. 1973. Chemical engineer's handbook. 5th ed. McGraw-Hill, New York.
- Quirk, J.P., and R.K. Schofield. 1955. The effect of electrolyte concentration on soil permeability. J. Soil Sci. 6(2):163–178.
- Seymour, R.M. 2000. Air entrapment and consolidation occurring with saturated hydraulic conductivity changes with intermittent wetting. Irrig. Sci. 20:9–14.
- Shan, G. 1995. Analytical solutions for determining vertical air permeability in unsaturated soils. Water Resour. Res. 31(9):2139–2200.
- Siegrist, R.L. 1987. Soil clogging during subsurface wastewater infiltration as affected by effluent composition and loading rate. J. Environ. Qual. 16(2):181–187.
- Vandevivere, P., and P. Baveye. 1992. Saturated hydraulic conductivity reduction caused by aerobic bacteria in sand columns. Soil Sci. Soc. Am. J. 56:1–13.
- Wang, Z., and J. Feyen. 1997. Two-phase flow infiltration equations accounting for air entrapment effects. Water Resour. Res. 33(12):2759–2767.
- Wang, Z., and J. Feyen. 1998. Air entrapment effects on infiltration rate and flow instability. Water Resour. Res. 34(2):213–222.
- Wangemann, S.G., R.A. Kohl, and P.A. Molumeli. 2000. Infiltration and percolation influenced by antecedent soil water content and air entrapment. Trans. ASAE 43(6):1517–1523.
- Weir, G.J., and W.M. Kissling. 1992. The influence of airflow on the vertical infiltration of water into soil. Water Resour. Res. 28(10):2765–2772.
TOP
ABSTRACT
INTRODUCTION
