Grouping Soil Fertility Indicators using Within-Field Spatial Variability Parameters and Multivariate Analysis.

Getting knowledge of the within-field spatial variability of soil fertility indicators is crucial before implementing precision conservation. A study was conducted on two contiguous fields (8.7 ha) of the Bras d’Henri watershed, near Quebec City (Canada). The objective of this study was: 1) to describe the within-field spatial variability of selected soil fertility indicators, and 2) to group these soil properties according to their spatial variability.

A very detailed soil survey was conducted in these fields. The soil surface layer (0-20 cm) of 91 soil profiles was sampled using an unaligned grid (30-m x 30-m). Soil analyses include OM, soil humus fractions, i.e. humic (HA) and fulvic acid (FA), texture, Mehlich-3 extractable P (M3P), free forms of Fe and Al, P sorption capacity (PSC) and degree of P saturation (DPS). Parameters used for characterizing the spatial variability of these soil properties are the coefficient of variation, the range (A₀), the ratio of structural variance over sill (C/C₀+C), the cross-validation regression coefficient (R²_cv), and the standard error of prediction expressed as a percentage of the soil property mean. The Ward’s clustering method and the Euclidean distance were used for grouping soil properties showing similar spatial variability.

The spatial distribution of HA, FA, Fe, Al and PSC was highly related to soil map and taxonomy. Three groups of soil properties have been identified in these fields according to their spatial variability: 1) M3P and DPS with A₀=100-200m (small spatial structure) and R²_cv=0.15-0.30 (weak prediction reliability), 2) humus fractions and soil texture with A₀=100-200m and R²_cv=0.30-0.60 (moderate reliability), and 3) PSC and free forms of Fe and Al with A₀>300m (moderate structure) and R²_cv=0.30-0.60. Soil properties related to soil nature and genesis are more reliably interpolated than those highly influenced by soil management and therefore should be included in precision conservation prescription.