Hydraulic conductivity prediction for sandy soils

Area of Science:

• Environmental Science
• Soil Science
• Hydrogeology

Background:

• Petroleum-contaminated soils pose a risk to groundwater quality.
• Accurate estimation of saturated hydraulic conductivity (K(S)) is vital for soil leachability models.
• Existing models like Rawls and Brakensiek (1989) have limitations for soils with high sand content (>70%).

Purpose of the Study:

• To develop a new, more accurate equation for estimating K(S) in sandy soils of the southeastern United States.
• To improve the reliability of soil leachability models for assessing groundwater contamination potential.

Main Methods:

• Analyzed 70 datasets of K(S) and particle-size data from literature for southeastern US sandy soils.
• Developed a multiple linear regression model using percent clay and sand particle data.
• Validated the new model with eight additional datasets, comparing it against Rawls and Brakensiek and Rosetta models.

Main Results:

• The new multiple linear regression model achieved an adjusted R2 of 0.65 (p < 0.0001).
• The developed model demonstrated smaller root mean square deviation and maximum squared difference values compared to existing models.
• The new model predicted K(S) within one order of magnitude for all but one test dataset, outperforming Rawls and Brakensiek and Rosetta for these sandy soils.

Conclusions:

• The new K(S) estimation equation is recommended for sandy soils in the southeastern US with limited input data (percent sand and clay).
• This improved K(S) estimation enhances the accuracy of soil leachability models for groundwater contamination assessments.
• The findings address a critical limitation in existing models for a significant soil type in the region.