Principal component analysis of ferrasols from the Central Plateau of southern Cameroon

Soil is a non renewable resource playing critical roles in food production and  many other life-support functions. This worldwide recognition calls for a better understanding of soil functioning, soil qualities (SQ) and their variation at various spatial scales. We applied inferential statistics and  principal component analysis (PCA) on 96 pedons from a 40.000 km² area in southern Cameroon in order  to i) assess soil property spatial change, ii) get more insight into the relative importance of pedogenetic processes and iii) empirically model soil qualities and their variation in humid forest soilscape. Analysis of variance highlighted  significant regional differences across soilscapes (P<0.0001) and soil depth (P<0.0001). Properties for which the interaction term was not significant included soil textural fractions, available phosphorus, soil melanization index, pH water, exchangeable Ca, K and Al. Pedon statistical modelling through PCA resulted in a 4-component model which accounted for 75.6 % of total soil variance (TSV). Their interpretation in terms of soil genesis and functional relationships among variables provided quantitative information on driving regional soil processes. Ferralitic pedogenesis emerged as the main zonal factor of soil differentiation accounting for  45 % of TSV. This process operative on all parent material has strong controls on nutrient availability and retention capacity.  The second most important process (15.4 % of TSV) was assimilated to soil potential acidification as modulated by parent material geochemistry. The 3^rd and 4^th source of variation (10.7% and 6.9 % of TSV) reflected soil age or time elapsed and soil color differentiation respectively.  Empirical  modeling of surface horizon generated 4 key functions which identified nutrient availability, water retention capacity, K availability and P fertility qualities.  These SQ accounted for 38.1, 17.4, 8.9 and 8.2 of TSV respectively with significant (P<0.05) regional differences. Grouping of soils on the basis of the identified SQ  stratified the studied area into functional groups with strong pedogenetic significance. Subsoil horizons produced similar results indicating that either the surface or subsoil horizons can efficiently be used for SQ evaluation. Keywords. South Cameroon, soil genesis, land qualities, PCA.