Estimation of the soil strength parameters in Tertiary volcanic regolith (NE Turkey) using analytical hierarchy process

Creative Commons License

ERSOY H., Karsli M. B., CELLEK S., KUL B., BAYKAN I., Parsons R. L.

JOURNAL OF EARTH SYSTEM SCIENCE, vol.122, no.6, pp.1545-1555, 2013 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 122 Issue: 6
  • Publication Date: 2013
  • Doi Number: 10.1007/s12040-013-0366-z
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.1545-1555
  • Keywords: Empirical relation, plasticity, regression analysis, strength parameters, Trabzon, ATTERBERG PLASTICITY LIMITS, RESIDUAL SHEAR-STRENGTH, P-WAVE VELOCITY, COHESIVE SOILS, NATURAL SOILS, DECISIONS, CLAYS, ROCKS
  • Karadeniz Technical University Affiliated: Yes


Costly and time consuming testing techniques and the difficulties in providing undisturbed samples for these tests have led researchers to estimate strength parameters of soils with simple index tests. However, the paper focuses on estimation of strength parameters of soils as a function of the index properties. Analytical hierarchy process and multiple regression analysis based methodology were performed on datasets obtained from soil tests on 41 samples in Tertiary volcanic regolith. While the hierarchy model focused on determining the most important index properties affecting on strength parameters, regression analysis established meaningful relationships between strength parameters and index properties. The negative polynomial correlations between the friction angle and plasticity properties, and the positive exponential relations between the cohesion and plasticity properties were determined. These relations are characterized by a regression coefficient of 0.80. However, Terzaghi bearing capacity formulas were used to test the model. It is important to see whether there is any statistically significant relation between the calculated and the observed bearing capacity values for model testing. Based on the model, the positive linear correlation characterized by the regression coefficient of 0.86 were determined between bearing capacity values obtained by direct and indirect methods.