Sorption Isotherms - Transport and Fate of Chemicals in Soils: Principles and Applications - page 30

1.7 Predictions
The models proposed by Sauvé, Hendershot, and Allen (2000) to predict K d
values given in Equations 1.3 and 1.4 were tested for the ability to predict
measured K d values for the five heavy metals and 10 soils discussed above.
As discussed earlier, for model I, K d predictions are based on soil pH only.
In model II, K d predictions were based on two variables, pH and percent soil
organic carbon (SOC). Comparison of measured K d values from this study
for all 10 soils using models I and II are shown in Figures 1.18 to 1.22. For cop-
per, predictions were overestimated at the low K d range. For both cadmium
and zinc, the predictions did not illustrate any pattern and were highly inad-
equate. Somewhat improved predictions were obtained for nickel and lead
when model I was used. Overall, both models yielded inadequate predictions
for all heavy metals used in this study. In contrast, when the Buchter et al.
(1989) model was used to predict the Freundlich parameter b , good overall
predictions were obtained (see Figure 1.23). For all heavy metals extremely
good trends were observed. The best prediction was obtained for Zn. Based
on this investigation, the following conclusions can be drawn.

1 Adsorption of all five heavy metals was nonlinear.

2 For all 10 soils used in this study, adsorption of heavy metals follows
the order Pb > Cu > Cd > Zn > Ni. In the presence of carbonates,
adsorption of heavy metals follows the order Pb > Cu > Zn > Cd > Ni.

10

Cu Predictions

8

6

4

Model 1
Model 2
1:1 Line

2

0

0 1 2 3 4 5 6 7 8 9 10

Measures log Kd

FIGURE 1.18
Measured and calculated K d values for Cu for all soils. Calculated values for K d were obtained
using models 1 and 2 of Sauvé, Hendershot, and Allen (2000).