4.4 Lab Measurements of Column Studies

The samples of the first and second homogeneous column tests used samples (a) and (b), respectively, as described in section 4.2. These column studies were done on 2-25-1999 and 2-26-1999.

The soil column studies for heterogeneous layers tested two different samples. The samples were the same samples (b) and (c) as used in section 4.2. The heterogeneous column studies were conducted on 3-13-1999 and 3-14-1999.

4.4.1 Homogeneous Soil Column Test

The sample of the first trial of homogeneous column test of the Ottawa quartz sand used sample (a) in grain size of 0.725 ± 0.125 mm. This grain size was too large for controlling the steady state recharge rate (0.03 [cm/sec]). The recharge rate was too fast to measure proper data (Table 4.4). The sample setting for this column test was listed in Table 4.3 in the previous section. The experimental mean data of vadose zone contaminant concentrations were measured during 21 minutes at about one minute intervals at every 7.62 cm (0.25 ft) from top to bottom of column after discharge saline (480 ppm) water as listed in Table 4.4. These eight sets of partial data for vadose zone contaminant concentrations in column are shown in Figure 4.13. The mean data of groundwater loading concentrations were measured every 10 minutes and are listed with bottom data from Table 4.4 in Table 4.5. The profile of groundwater loading concentrations is shown in Figure 4.14. 

Figure 4.13 Profile of Lab Measurements of Vadose Zone Contaminant Concentrations in Soil (Sand Sample a: Grain Size of 0.725 ± 0.125 mm) Column (every 0.25 ft from top of the column) during 21 minutes after Discharge of Saline (480 ppm) Water.

Figure 4.14 Profile of Groundwater Loading Measurements after Discharge of Saline (480 ppm) Water for Sand Sample a (Grain Size of 0.725 ± 0.125 mm).

At a second trial of the homogeneous column test, the sample used sample (b) in grain size of 0.337 ± 0.087 mm of Ottawa quartz sand. The recharge rate was 0.004 [cm/sec] which was easier to measurement than the recharge rate of 0.03 [cm/sec] of the sample (a) case. Sample setting for this column test was listed in Table 4.3. The data of vadose zone contaminant concentrations were measured at every 26 minutes and every 7.62 cm (0.25 ft) from top to bottom of column after discharge saline (480 ppm) water (Table 4.5). These nine sets of complete data for vadose zone contaminant concentrations in the columns are shown in Figure 4.15. The data of groundwater loading concentrations were measured every 10 minutes after recharge of saline (Sodium Chloride, NaCl) water (Table 4.7). The profile of groundwater loading concentrations was shown in Figure 4.16.

4.4.2 Heterogeneous Soil Column Test

The samples consisted of three layers which were 30.48 cm (1 ft) of sample (b) (grain size of 0.337 ± 0.087 mm), 30.48 cm (1 ft) of sample (c) (grain size of 0.627 ± 0.552 mm), and 60.96 cm (2 ft) of sample (b) in order from the top surface of the column. The data of vadose zone contaminant concentrations were measured at every 53 minutes after discharge of saline (Sodium Chloride, NaCl, 480 ppm) water (Table 4.8). Figure 4.17 showed nine periodical data sets of vadose zone concentrations. The data of groundwater loading concentrations were measured every 30 minutes after recharge of saline water (Table 4.9). The profile of these data is shown in Figure 4.18. 

Figure 4.15 Profile of Lab Measurements in Soil (Sand Sample b: Grain Size of 0.337 ± 0.087 mm) Column (every 0.25 ft from top of the column) at every 26 min after Discharge of Saline (480 ppm) Water.

 

Table 4.7 Bottom of Groundwater Loading Measurements and Data Analyses every 10 min after Discharge of Saline (480 ppm) Water for Vadose Zone Leaching to Groundwater (Sand Sample b: Grain Size of 0.337 ± 0.087 mm)

Time		Concentration (ppm) of Column Bottom			Standard Deviation
(hour)	(min)	2-25-1999	2-26-1999	Average
0.00	0.0	0.0	0.0	0.0	0.00
0.17	10.0	0.0	0.0	0.0	0.00
0.33	20.0	0.0	0.0	0.0	0.00
0.50	30.0	0.0	0.0	0.0	0.00
0.67	40.0	0.0	0.0	0.0	0.00
0.83	50.0	0.0	0.0	0.0	0.00
1.00	60.0	0.0	0.0	0.0	0.00
1.17	70.0	0.0	0.0	0.0	0.00
1.33	80.0	0.0	0.0	0.0	0.00
1.50	90.0	0.0	0.0	0.0	0.00
1.67	100.0	0.0	0.0	0.0	0.00
1.83	110.0	7.0	9.0	8.0	1.41
2.00	120.0	20.0	20.0	20.0	0.00
2.17	130.0	40.0	40.0	40.0	0.00
2.33	140.0	60.0	70.0	65.0	7.07
2.50	150.0	100.0	120.0	110.0	14.14
2.67	160.0	180.0	160.0	170.0	14.14
2.83	170.0	240.0	240.0	240.0	0.00
3.00	180.0	320.0	320.0	320.0	0.00
3.17	190.0	380.0	400.0	390.0	14.14
3.33	200.0	440.0	460.0	450.0	14.14
3.50	210.0	490.0	500.0	495.0	7.07
3.67	220.0	510.0	510.0	510.0	0.00
3.83	230.0	510.0	510.0	510.0	0.00
4.00	240.0	500.0	500.0	500.0	0.00
4.17	250.0	490.0	490.0	490.0	0.00
4.33	260.0	470.0	480.0	475.0	7.07
4.50	270.0	470.0	470.0	470.0	0.00
4.67	280.0	470.0	470.0	470.0	0.00
4.83	290.0	470.0	480.0	475.0	7.07
5.00	300.0	480.0	480.0	480.0	0.00
5.17	310.0	480.0	480.0	480.0	0.00
5.33	320.0	480.0	480.0	480.0	0.00
5.50	330.0	480.0	480.0	480.0	0.00
5.67	340.0	480.0	480.0	480.0	0.00

 

Figure 4.16 Profile of Bottom Measurements every 10 min after Discharge of Saline (480 ppm) Water for Vadose Zone Leaching to Groundwater (Sand Sample b: Grain Size of 0.337 ± 0.087 mm).

Figure 4.17 Profile of Lab Measurements in Heterogeneous Layers Column (every 0.25 ft from top of the column) at every 53 min after Discharge of Saline (480 ppm) Water.

Figure 4.18 Profile of Bottom Measurements every 30 min after Discharge of Saline (480 ppm) Water for Vadose Zone Leaching to Groundwater in Heterogeneous Layers.

The recharge rate for this heterogeneous study (sample c) was 0.002 [cm/sec] which was half of the recharge rate on homogeneous sample (b) case (0.004 cm/sec). The reduced recharge rate used based on the lower porosity (0.290) of the middle layer of sample (c), as compared with the porosity (0.394) of the homogeneous sample (b). As shown in Figure 4.17, the second layer indicated that the salt concentration (88 ppm at 53 min) was much less than the homogeneous sample (b) case (320 ppm at 53 min of Figure 4.15). The concentration rate of the second layer in Figure 4.17 was about five times less than the homogeneous sample (b) case between 30.48 and 60.96 cm (1 and 2 ft) from the top surface in Figure 4.15.

4.5 Sensitivity Analysis in VG Model

A sensitivity analysis of the VG model was performed to evaluate the impact of longitudinal dispersivity, a_L, parameter in sample (b) case.

The purpose of sensitivity analysis is to quantify the effects of uncertainty in the estimates of model parameters on model results. During a sensitivity analysis, calibrated values for recharge rate (q_w), longitudinal dispersivity (a_L), and the like, are systematically changed within a pre-established range of applicable values. The magnitude of change in concentrations from the calibrated solution is a measure of the sensitivity of the solution to that particular parameter. The results of the sensitivity analysis are expressed as the effects of the parameter change on the average measure of error (mean error or root mean square error) and on the spatial distribution of concentrations. 

The organic carbon distribution coefficient (K_oc), recharge rate (q_w), soil organic carbon content (f_oc), bulk density ({r_b), volumetric water content of soil (q), and porosity (n) in the leaching studies such as LEACHM (Hutson and Wagenet, 1992), HYDRUS (Vogel and et al., 1996), and VLEACH (Ravi and Johnson, 1993) demonstrated greatest impact on either vadose zone contaminant level or groundwater loading. The other parameters, such as maximum water solubility of contaminant (C_max), free air diffusion coefficient (D_air), and Henry's partition coefficient (K_H), were demonstrated to have less significant impact on accurate input parameters. 

The sensitivity analysis of the VG model was performed to evaluate the impact of input parameter of longitudinal dispersivity, a_L, on vadose zone contaminant level and loading to groundwater. The results of the study are depicted in Figures 4.19 and 4.20. It was seen that the input parameter of longitudinal dispersivity (a_L) had a significant impact on both vadose zone contaminant concentration (Figure 4.19) and groundwater loading (Figure 4.20). A qualitative description of the sensitivity of each parameter to the calculated groundwater impact and vadose zone concentration profile are compiled in the Table 4.10. 

Figure 4.19 The Effect of Longitudinal Dispersivity on Vadose Zone Contaminant Profile Sensitivity Analysis for Verification of the First Part (Vadose Zone Leaching) of VG Model at every 26.5 min.

Figure 4.20 The Effect of Longitudinal Dispersivity on Groundwater Loading Sensitivity Analysis for Verification of the Second Part (Saturated Zone Mixing) of VG Model at every 10 min.

4.6 VG Model Simulations

Simulations of the VG model for two sets of homogeneous studies were performed by input data of samples (a) and (b) on 200 MHz 80586 personal computer. Simulations of the VG model for the study of heterogeneous layers were performed by input data of samples (b) and (c) on 200 MHz 80586 personal computer.

4.6.1 Simulation for Homogeneous Column Studies

The purpose of these simulations is to compare the experimental data for verification with the VG model. Data collection of these simulations was described in previous Section 4.2 and Table 4.3. Input parameters of Samples (a) and (b) for the VG model simulation were shown in Tables 4.11 and 4.12, respectively. Basically, the VG model was developed for long term (yearly) impact of vadose zone and groundwater loading. However, input parameters of the time intervals in the vadose zone concentration profile (PRTIME), groundwater impact and mass balance results (PTIME), and plot output (PLTIME) were typed 4.0E-6 (year) for sample (a) and 5.0E-4 (year) for sample (b) which are the same as 2.1 (min) and 26.0 (min), respectively. The horizontal width (WIDTH) and length (LENGTH) of the soil column were set at 0.074 [ft] which allow calculation of the same area as the cross-sectional lab column.

Note:  Cross-sectional area of 1 inch diameter circle column is

 A= p(d/2)² = p(1/24)²  (ft²) = 0.005  (ft²).

          Cross-sectional area of 0.074 ft by 0.074 ft square soil cel1 is

A = 0.074·0.074 (ft²) = 0.005  (ft²).

 

Table 4.11 VG Model Input File for Sample (a: Grain Size of 0.725 · 0.125 mm)

Ottawa sand Sample (a) and Saline (480 mg/l) Water NSCOL 1 DELT STINE TIMPL PTIME PRTIME ANGL 0.0001 0.005 0.005 0.000004 0.000004 P KOC KH CMAX DAIR 0.0001 0.00001 480 0 Ottawa sand Sample (a): Grain Size of 0.725 · 0.125 mm NCELL DELZ WIDTH LENGTH TYPTBC VALTBC ALDISP 17 0.25 0.074 0.074 1 480 0.02 PLT PLTIME DECY1 DECY2 Y 0.000004 0 0 NLAYER 1 J1 J2 RECHRG RHOB POR THETA FOC  XCON 1 17 34032 1.753 0.338 0.319 0.0001 0 FLXAQF CINAQF PDEPTH PCHOIC ADEPTH AVDISP 20 1 1 E 3 5

 

Table 4.12 VG Model Input File for Sample (b: Grain Size of 0.337 · 0.087 mm)

Ottawa sand Sample (b) and Saline (480 mg/l) Water NSCOL 1 DELT STINE TIMPL PTIME PRTIME ANGL 0.0001 0.001 0.001 0.00005 0.00005 P KOC KH CMAX DAIR 0.0001 0.00001 480 0 Ottawa sand Sample (b): Grain Size of 0.337 · 0.087 mm) NCELL DELZ WIDTH LENGTH TYPTBC VALTBC ALDISP 17 0.25 0.074 0.074 1 480 0.02 PLT PLTIME DECY1 DECY2 Y 0.00005 0 0 NLAYER 1 J1 J2 RECHRG RHOB POR THETA FOC  XCON 1 17 4084 1.607 0.394 0.333 0.0001 0 FLXAQF CINAQF PDEPTH PCHOIC ADEPTH AVDISP 20 1 1 E 3 5

The values of the recharge rate (RECHRG), the bulk density (RHOB), the effective porosity (POR), and the volumetric water content (THETA) have been evaluated from lab column tests in the Section 4.2. The value of the longitudinal dispersivity (ALDISP) has been found from the sensitivity analysis in Section 4.5. The value of the top boundary condition (VALTBC) was the same value (480 ppm) of our colored sodium chloride (NaCl) water (480 ppm). The values of the others were zero or the minimum values of the references. 

The results of the VG model simulation of vadose zone contaminant concentration for samples (a) and (b) have been plotted depth [ft] versus concentration [mg/l] of eight complete sets at during 21 minutes and nine complete sets at every 26 minutes as shown in Figures 4.21 and 4.23, respectively. The profiles of the VG model simulation results of the groundwater loading concentration (mg/l) for samples (a) and (b) were shown in Figures 4.22 and 4.24, respectively.

Figure 4.21 Vadose Zone Contaminant Profile of Simulation Results of the First Part (Vadose Zone Leaching) of VG Model at during 21 min for Sample a (Grain Size of 0.725 · 0.125 mm and Input Parameters in Table 4.11).

Figure 4.22 Groundwater Loading Profile of Simulation Results of the Second Part (Saturated Zone Mixing) of VG Model for Sample a (Grain Size of 0.725 · 0.125 mm and Input Parameters in Table 4.11).

Figure 4.23 Vadose Zone Contaminant Profile of Simulation Results of the First Part (Vadose Zone Leaching) of VG Model at every 26 min for Sample b (Grain Size of 0.337 · 0.087 mm and Input Parameters in Table 4.12).

Figure 4.24 Groundwater Loading Profile of Simulation Results of the Second Part (Saturated Zone Mixing) of VG Model for Sample b (Grain Size of 0.337 · 0.087 mm and Input Parameters in Table 4.12).

4.6.2 Simulation for Heterogeneous Column Study

Data collection of the VG model simulation for the heterogeneous study was described in Table 4.3. Table 4.13 presents input parameters of samples (b) (grain size of 0.337 · 0.087 mm) and (c) (grain size of 0.627 · 0.552 mm) for this study. Cell numbering for this study is 1 to 4 for (first layer), 5 to 8 for (second layer), and 9 to 17 for (third layer) as a total of 17 cells including dummy cell. The values of the recharge rate (RECHRG), the bulk density (RHOB), the effective porosity (POR), and the water filled porosity (THETA) of first, second and third layers were 2042, 2042, and 2042 [ft/yr], 1.607, 1.900, and 1.607 [g/cm³], 0.394, 0.290, and 0.394, and 0.333, 0.210, and 0.333, respectively. The value of longitudinal dispersivity (ALDISP) was 0.02 [ft] which was found from sensitivity analysis in the section 4.5. The other values of input parameters were the same as the homogeneous sample (b) case (Table 4.12)

The results of the VG model simulation of vadose zone concentration for this heterogeneous study are nine complete data sets at every 53 minutes time interval and every 7.62 cm (0.25 ft) from top surface after discharge of saline (480 ppm) water as shown in Figure 4.25. The second layer between 30.48 and 60.96 cm (1 and 2 ft) from top surface showed almost a linear curve for the 105 min data set. Figure 4.26 presented the profile of the VG model simulation results of the groundwater loading concentration after discharge of saline (Sodium Chloride, NaCl) water.

 

Table 4.13 VG Model Input File for Heterogeneous Layers using Two different Samples (b), (c), and (b), respectively from Top of the Column

HETEROGENEOUS COLUMN STUDY NSCOL 1 DELT STINE TIMPL PTIME PRTIME ANGL 0.00001 0.001 0.001 0.00005 0.00005 P KOC KH CMAX DAIR 0.0001 0.00001 480 0 Ottawa sand Sample (b), (c), and (b) NCELL DELZ WIDTH LENGTH TYPTBC VALTBC ALDISP 17 0.25 0.074 0.074 1 480 0.02 PLT PLTIME DECY1 DECY2 Y 0.00005 0 0 NLAYER 3 J1 J2 RECHRG RHOB POR THETA FOC  XCON First Layer: Sand sample (b) (grain size of 0.337 · 0.087 mm) 1 4 2042 1.607 0.394 0.333 0.0001 0 Second Layer: Sand sample (c) (grain size of 0.627 · 0.552 mm) 5 8 2042 1.900 0.290 0.210 0.0001 0 Third Layer: Sand sample (b) (grain size of 0.337 · 0.087 mm) 9 17 2042 1.607 0.394 0.333 0.0001 0 FLXAQF CINAQF PDEPTH PCHOIC ADEPTH AVDISP 20 1 1 E 3 5

Figure 4.25 Vadose Zone Contaminant Profile of Simulation Results of the First Part (Vadose Zone Leaching) of VG Model at every 53 min for the Heterogeneous Layers (Parameters listed in Table 4.13).

Figure 4.26 Groundwater Loading Profile of Simulation Results of the Second Part (Saturated Zone Mixing) of VG Model for Heterogeneous Layers (Parameters listed in Table 4.13).

4.7 Comparison of Lab Measurements and Model Simulations

4.7.1 Homogeneous Study

Lab measurements of sample (a) column study included partial eight sets of data due to the fast recharge rate (0.03 cm/sec). Results of comparison for sample (a) are presented in Figures 4.27 and 4.28. As shown in Figure 4.27, all measuring data sets, except the last set of data, were located near places used in simulation data sets. The mismatched last set may involve the effect of capillary fringe. Figure 4.28 presented comparison of the concentrations of groundwater loading. The mismatched peak point may involve the absence of lab data between 21 minutes and 30 minutes or dilution of measuring water samples. However, these plots indicated identical results with sample (a) in lab measurements and simulations.

Figure 4.27 Comparison with Lab Measurements (Sample a) and VG Model Simulations in Vadose Zone Leaching Concentration.

Figure 4.28 Comparison with Lab Measurements (Sample a) and VG Model Simulations for Vadose Zone Leaching Concentration to Groundwater Loading.

Figure 4.29 Comparison with Lab Measurements (Sample b) and VG Model Simulations in Vadose Zone Leaching Concentration.

Figure 4.30 Comparison with Lab Measurements (Sample b) and VG Model Simulations for Vadose Zone Leaching Concentration to Groundwater Loading.

Results of comparison of lab measurements and the VG model simulation for sample (b) column study were presented in Figures 4.30 and 4.31. Figure 4.30 showed a perfect match of lab measurements with the VG model simulation for vadose zone contaminant concentration until 106.68 cm (3.5 ft) from the top surface. About 15.24 cm (0.5 ft) of mismatched bottom part data may involve the effect of capillary fringe. Figure 4.31 shows a perfect match of both for groundwater loading concentration without any exceptions. The peak point of concentration which is higher than 480 (ppm) of saline happened to be matched with measuring data and simulation output. This peak point in Figure 4.30 may be because of density-salinity differentiation of a portion of the saline water during the transport process as it descends down the column. However, results of these comparisons for sample (b) were identified with lab measurements and the VG model simulation.

4.7.2 Heterogeneous Study

Results of comparison of lab measurements and the VG model simulations for the heterogeneous study are presented in Figures 4.31 and 4.32. Figure 4.31 indicated that the simulated vadose zone concentration curves compared very closely with the measured data. The measured concentration data between 106.68 and 121.92 cm (3.5 and 4.0 ft) in Figure 4.31 exhibited considerable scatter which may indicate the effect of capillary fringe in the experimental data.

Figure 4.31 Comparison with Lab Measurements and VG Model Simulations in Vadose Zone Leaching Concentration for Heterogeneous Layers.

Figure 4.32 Comparison with Lab Measurements and VG Model Simulations in Vadose Zone Leaching Concentration to Groundwater Loading for Heterogeneous Layers.

A comparison of the lab measured and simulated groundwater loading concentration curves is shown in Figure 4.32. The lab-measured data are represented by the liquid-phase concentrations after discharge of saline (Sodium Chloride, NaCl, 480 ppm) water. The peak point of the high concentration was measured at 7 hours after discharge of saline water similar to the data in the VG model simulation. Figure 4.32 showed that the simulated groundwater loading curve compares favorably with the lab measured data.