Artificial Neural Network (ANN) modelling for the estimation of soil microbial biomass in vineyards soils
We are glad to announce that we have contributed together with Elisa Pellegrini, Nicola Rovere, Irene Franco, Maria de Nobili and Marco Contin to this new study, published in the journal “Biology and Fertility of Soil” (2020)
To evaluate the quality of the soil, one of the most used parameters is the microbial biomass (SMB-C). But since threshold values have never been indicated and there is no reliable numerical estimate of SMB-C, a study was launched in order to evaluate the physicochemical properties of the soil through the “Artificial Neural Network” (ANN).
In this research, the soil of 231 areas with different characteristics of temperature, exposure and humidity were analysed by identifying ten physico-chemical parameters for each one. Thanks to the dataset obtained, the ANN model confirmed the primary importance of soil organic matter (SOM) for the prediction of SMB-C and therefore showed a improved fit than the linear model. This means we have a unique tool to estimate the microbial biomass: the comparison with the laboratory measurement allows us to deal agronomically with the biological fertility of the single soil.
We would like to thank in particular the Association “VinNatur” for contributing and believing in the project from the beginning, making resources available and wineries interested in participating in the experiment. Furthermore the experiment wouldn’t be possible without the people who created the estimated Biomass model.