Geostatistical modeling of the spatial variability of coffee fine roots under Erythrina shade trees and contrasting soil management

Abstract

Spatial relationships between root length density of Coffea arabica (coffee RLD) and soil nutrient- related factors at plot scale in a coffee- Erythrina poeppigiana system was studied by geostatistics. In a 24 × 29 m area, (organic and conventional management), coffee and Erythrina fine roots and soil chemical properties were sampled on an irregular grid in the topsoil. A factor analysis explained 83 % of the total variation of the soil attributes. Soil factors were identified: Chemical fertility (CF), Micronutrients, Organic matter, and Acidity (Ac). Based on the spherical model, all the attributes presented a strong spatial structure. The scale of spatial correlation for CF was lesser than for Ac, but similar to coffee RLD. Erythrina RLD had a short-range variation. Patchy areas of high spots of coffee RLD were greater in organic plot. Cross-semivariogram analysis estimated a correlation between soil factors and coffee RLD over a scale of 5.50 m; but 4.23 m with Erythrina RLD. Nutrients linked to P, Zn, exchangeable bases and acidity soil affected the scale of spatial aggregation pattern of coffee RLD. The spatial response of coffee RLD suggests a differential nutrient uptake strategy for acquiring soil nutrients induced by the quality of organic and inorganic fertilizer inputs. The fact that coffee RLD had higher scale of spatial variation than Erythrina RLD and a negative spatial correlation indicate that pruned Erythrina trees are not so competitive for acquiring shared nutrients in an agroforestry system.