Statistical relationships among selected properties of northern Cameroon vertisols and associated alfisols

• Published in: Soil Science Society of America journal Vol. 53; no. 6
• Main Authors: Yerimia, B.P.K. (Univ. Centre Dschang, R. Cameroon, Africa), Wilding, L.P, Hallmark, C.T, Calhoun, F.G
• Format: Journal Article
• Language: English
• Published: 01.11.1989
• Subjects: ANALISIS ESTADISTICO; ANALYSE STATISTIQUE; CAMEROUN; CAMERUN; MODELE; MODELOS; PROPIEDADES FISICO-QUIMICAS (SUELO); PROPRIETE PHYSICO-CHIMIQUE DU SOL; REGION DE BASSE ALTITUDE; SOL DE HAUTES TERRES; SUELO DE ALTURA; TIERRAS BAJAS; TIPOS GENETICOS DE SUELOS; TYPE DE SOL GENETIQUE; VERTISOL; VERTISOLES
• DOI: 10.2136/sssaj1989.03615995005300060025x
• ISSN: 0361-5995

Six Usterts and one Ustalf developed from lacustrine sediments and Precambrian schists in northern Cameroon were characterized for physical, chemical, and mineralogical properties. Correlative statistical relationships were established among physical and chemical properties including surface area (SAT), shrink-swell (COLE), cation exchange capacity (CEC), percent total clay (PCL), percent coarse clay (PCCL), percent fine clay (PFCL), water retention at 33 kPa (WAT), organic carbon (OC), calcium carbonate equivalent (CCE), pH, and exchangeable sodium percent (ESP). In general, upland soils, developed from Precambrian schists and sandstones, yielded lower correlation coefficients among these properties than lowland soils developed from Quaternary fluvial deposits. For the lowland soils, COLE was highly correlated with SAT, PCL, and WAT; simple coefficients of determination (r2) were 0.91, 0.80, and 0.74, respectively. Similar correlations were obtained for COLE for upland soils, but neither CCE for these soils, nor ESP for either upland or lowland soils were significantly correlated with COLE. The CEC of soils studied was highly correlated with SAT and PCL; r2 values ranged from 0.59 to 0.91. The SAT of Usterts in lowlands was very highly correlated with PCL, PFCL, PCCL, CEC, and COLE; these independent variables generally accounted for greater than 90% of the variance in SAT. For Usterts and Ustalfs of uplands, the SAT was not so closely correlated with these parameters, especially PFCL and PCCL, which varied in amount and clay mineralogy with depth. Multiple linear regressions did not generally improve predictive abilities for COLE, CEC, and SAT over simple linear relationships. Regression relationships developed constitute useful predictive indices for estimating engineering and agronomic properties from existing physical and chemical data and soil survey reports of northern Cameroon