Pedotransfer functions of soil physical properties: methodological approaches to development, model catalogue, and local validation schemes

Abstract

Pedotransfer function development is widely used as an alternative approach for estimating soil properties whose direct measurement is technically demanding, time-consuming, and costly. Soil physical properties such as bulk density, particle density, and total porosity are frequently estimated using pedotransfer functions (PTFs) and are subsequently applied in soil quality assessment, detection of soil degradation processes, and parameterization of process-based models. The development of reliable PTFs requires adherence to well-defined methodological principles, whereas the transfer of models developed in other regions to local datasets requires prior harmonization of input variables and independent validation, since differences in predictor determination methods and particle-size class boundaries may significantly affect prediction accuracy. The aim of this study was to synthesize methodological approaches to the development of PTFs for soil physical properties and to compile a catalogue of candidate models together with validation schemes for their application to local soil datasets. A theoretical and analytical approach was used to review and generalize methodological practices and to identify relevant pedotransfer functions. The analysis shows that locally specific PTFs are typically developed using data stratification procedures combined with statistical regression and machine-learning techniques. Model performance is assessed through different validation strategies, including statistical validation and comparison with PTFs developed for other regions. A catalogue containing 14 PTFs for bulk density and 5 PTFs for particle density was compiled. These models employ soil organic carbon, soil organic matter, and/or particle-size distribution as predictors. Validation schemes are proposed depending on the availability of input data and the need to convert variables due to inconsistencies between national and international particle-size classification systems. Priority directions for
the advancement of PTF development in Ukraine include the systematic expansion of national soil databases,
the development of open thematic soil datasets, the creation of PTFs for specific ecosystems, the adoption of innovative approaches for acquiring predictor variables (e.g. spectroscopic methods and ISP+), and the development of accessible software tools for practical model implementation. To improve the compatibility of the national particle-size classification system with widely used international standards (FAO/USDA), it is proposed to complement the standard analytical protocol with additional measurements of particle fractions with diameters of 2 mm and 0.002 mm. Such an approach would ensure that analytical results remain compatible with existing national soil quality assessment frameworks while also enabling their integration into European and global soil information systems.