Empirical estimation of saturated soil-paste electrical conductivity in the EU using pedotransfer functions and Quantile Regression Forests: A mapping approach based on LUCAS topsoil data

Abstract

Soil Electrical conductivity (EC) is a measure of the ability of soil to conduct an electric current, which is primarily influenced by the concentration of soluble salts in the soil solution that takes place principally through water-filled pores. Ions (Ca²⁺, Mg ²⁺, K ⁺, Na ⁺, and NH ⁴⁺, SO₄^2-, Cl^-, NO₃^–, and HCO₃^–) from soluble salts dissolved in soil water carry electrical charges and conduct the electrical current. EC is considered a proxy of soil salinity and other soil characteristics, whose monitoring is much needed in the context of climate change, increasing irrigation in agricultural areas and sea level rise. The pan-European LUCAS soil monitoring scheme, established in 2009, provided EC_1:5 in the topsoil (0–20 cm) in the surveys of the years 2015 and 2018 for almost 20,000 samples. In this work, using the LUCAS 2018 dataset, we provide an empirically-derivedpedotransfer function to convert diluted EC_1:5 to saturated EC_e using the LUCAS soil texture and soil organic carbon, and a framework for EC_e mapping with a machine-learning algorithm named Quantile Regression Forest. The final model resulted in an R² of 0.302 with an RMSE of 0.265 dS m⁻¹ for soil samples not used for model calibration. The results are presented as predicted EC_e in the topsoil, and they reveal that in Atlantic and Northern Europe, salts may accumulate in soils through several natural processes, i.e., primary salinization, but in Mediterranean and Southern Europe, they accumulate because of human interventions on the soil water and solute regimes. Among these interventions, seawater intrusion into coastal aquifers, irrigation with waters containing soluble salts, poor drainage of irrigated lands and of naturally saline soils, stand out. Hotspot analysis per country or Nomenclature of Territorial Units for Statistics (NUTS 0) revealed high topsoil EC_e levels occurred in Spain 0.11 %. Increasing EC_e can led to constrained crop productivity in irrigated farming. With this assessment, we try to determine the hotspots for future monitoring and understanding the main drivers for sustainable soil management. Future challenges for EC_e mapping that need to be address are sample numerosity and depth and availability of a consistent set of EC_e data measured to provide a regression based PTF for the use of diluted EC_1:5.