Yitagesu F.A., Van DerMeer F., Van DerWerff H., Seged H.
Faculty of Geo-information Science, Earth Observation (ITC), University of Twente, 99 Hengelosestraat, 7500AA Enschede, Netherlands; Ethiopian Roads Authority (ERA), P.O. Box 7129, Addis Ababa, Ethiopia; Addis Ababa Univ. (AAU), Faculty of Technology, Dep. of Civil Engineering, P.O. Box 385, Addis Ababa, Ethiopia
Yitagesu, F.A., Faculty of Geo-information Science, Earth Observation (ITC), University of Twente, 99 Hengelosestraat, 7500AA Enschede, Netherlands, Ethiopian Roads Authority (ERA), P.O. Box 7129, Addis Ababa, Ethiopia; Van DerMeer, F., Faculty of Geo-information Science, Earth Observation (ITC), University of Twente, 99 Hengelosestraat, 7500AA Enschede, Netherlands; Van DerWerff, H., Faculty of Geo-information Science, Earth Observation (ITC), University of Twente, 99 Hengelosestraat, 7500AA Enschede, Netherlands; Seged, H., Addis Ababa Univ. (AAU), Faculty of Technology, Dep. of Civil Engineering, P.O. Box 385, Addis Ababa, Ethiopia
Correlations are essential to obtain information on soil geotechnical parameters, which are costly and time consuming to measure directly, such as expansion potential. A common procedure for evaluating and rating soil expansion potential is the expansion index (EI) test. The purpose in this study was to establish a multivariate regression model to predict soil EI, thereby classify and rate soil expansiveness. Soil samples were collected from the newly planned expressway connecting the city of Addis Ababa with the town of Nazret in Ethiopia. A regression equation was established from liquid limit (LL), plasticity index (PI), and soil fine fraction (percentage of material passing the ASTM 0.075-mm sieve aperture), using a partial least squares (PLS) multivariate calibration method. A coefficient of determination (R2) of 0.92 accompanied with a root mean square error of prediction (RMSEP) of 9.87, standard error of performance (SEP) of 9.91, offset of 5.31 and bias of 0.04 was obtained. Response surface models showing three-way relationships among the predictors (Atterberg limits and fine fraction) and response variable (EI) may serve as classification systems for evaluating soil expansion potential. Apart from its basic scientific value as a simple method for estimating and rating soil expansiveness, the approach has the advantage of employing easily and routinely determined soil properties, to get information on soil expansion potential at minimal cost and time requirements. © Soil Science Society of America.
A-coefficient; Atterberg limits; Classification system; Ethiopia; Expansion index; Fine fraction; Geotechnical parameters; Liquid limits; Minimal cost; Multivariate calibration methods; Multivariate regression models; Partial least squares; Plasticity indices; Regression equation; Response surface models; Root-mean-square error of predictions; Scientific values; SIMPLE method; Soil property; Soil sample; Standard errors; Time requirements; Expansion; Least squares approximations; Mean square error; Regression analysis; Soil testing; Soils; Geologic models; calibration; correlation; expansion; geotechnical mapping; liquid limit; multivariate analysis; rating curve; regression analysis; Addis Ababa; Ethiopia