A preliminary model for determining a soil quality index including biological data implemented through a QR code application

Soil plays a central role in delivering several ecosystem services. However, its complex nature, the spatial variability and the timescale of soil processes make it challenging to quantify shifts in soil quality as a result of agronomical practices. A comprehensive indicator that includes parameters from different categories of soil properties, allowing an easy interpretation of soil quality by farmers and land managers, is thus needed. In this context, a class-modelling approach based on the Data-Driven Soft Independent Model of Class Analogy (DD-SIMCA) was tested to develop a soil quality index based on physical, chemical and biological parameters. Three models were built on a dataset composed by physical, chemical and biological soil parameters, which was created basing on ranges of values common to agricultural soils. The algorithm was thus applied to a real dataset obtained from about 9800 soil samples. The models showed very high performance (sensitivity = 1), allowing to classify the samples into quality groups. The model output was incorporated into a coloured QR-code, which allowed to express the quality of a soil sample with a colorimetric scale based on a soil quality index. A preliminary version of the tool is available for further testing and validation through a web platform (https://agritechlab.crea.gov.it/model/ddsimcasoil/ddsimcasoil.html).