A geostatistical approach to estimate probability of occurrence of Tuber melanosporum in relation to some soil properties

This study concerned an experimental truffle bed of downy oaks infected by the ectomycorrhizal fungus Tuber melanosporum and planted in 1983. The presence of T. melanosporum creates rounded areas with little herbaceous cover, termed "brulis", where carpophores are found. The investigation was aimed at relating the occurrence and carpophore production of T. melanosporum to soil properties, i.e., organic matter, structure, aeration and fertility, expressed in terms of total organic C, aggregate size classes, DTPA-extractable Fe and Mn, and host plant height. Data were processed by multivariate geostatistical techniques. A linear model of coregionalization (LMC) including (i) a nugget effect, (ii) a short-range spherical structure with a range of 7 m and (iii) a long-range spherical structure with a range of 32 m, was fitted to the experimental direct and cross-variograms of the investigated properties. Factorial kriging analysis (FKA) was used to separate the sources of variation of the data according to the spatial scale at which they operate, and to summarize and map them in terms of spatial factors. An indicator approach was adopted to estimate and map the conditional probability of presence and fructification of T. melanosporum. The visual comparison between the spatial pattern of the long-range structure of the first regionalized factor with the probability map of finding brulis plus carpophore production revealed some interesting relationship and suggested that T. melanosporum may prefer a soft and well-aerated soil environment to grow. (C) 2000 Elsevier Science B.V. All rights reserved.