Search

Machine learning (ML) algorithms are extensively used with outstanding prediction accuracy. However, in some cases, their overfitting capabilities, along with inadvertent biases, might produce overly optimistic results. Spatial data are a special kind of data that could introduce biases to ML due to their intrinsic spatial autocorrelation. To address this issue, a special resampling method has emerged called spatial cross-validation (SCV). The purpose of this study was to evaluate the performance of SCV compared with conventional random cross-validation (CCV) used in most ML studies. Multiple ML models were created with CCV and SCV to predict groundwater electrical conductivity (EC) with data (A) from Rhodope, Greece, in the summer of 2020; (B) from the same area but at a different time (summer 2019); and (C) from a new area (the Salento peninsula, Italy). The results showed that the SCV provides ML models with superior generalization capabilities and, hence, better prediction results in new unknown data. The SCV seems to be able to capture the spatial patterns in the data while also reducing the over-optimism bias that is often associated with CCV methods. Based on the results, SCV could be applied with ML in studies that use spatial data. [ABSTRACT FROM AUTHOR]

Studies dating almost a century relate clay properties with the structure of the diffuse double layer (DDL), where the charged surfaces of clay crystal behave like an electric capacitor, whose dielectric is the interstitial fluid. The intensity of the inner electric field relates to the concentration and type of ions in the DDL. Other important implications of the model are less stressed: this part of the clay soil system, energetically speaking, is conservative. External contribution of energy, work of overburden or sun driven capillarity and long exposure to border low salinity waters can modify the concentration of pore-waters, thus affecting the DDL geometry, with electric field and energy storage variations. The study of clay soils coming from various natural geomorphological and hydrogeological contexts, determining a different salinity of interacting groundwater, shows how the clay interaction with freely circulating waters at the boundaries produces alterations in the native pore water salinity, and, at the nano-scale, variations of electric field and stored energy from external work. The swelling and the shrinkage of clay soil with their volumetric and geotechnical implications should be regarded as variations of the electrostatic and mechanical energy of the system. The study is based on tests on natural clay soil samples coming from a formation of stiff blue clays, widespread in southern Italy. Geotechnical identification and oedometer tests have been performed, and pore waters squeezed out from the specimens have been analyzed. Tested samples have similar grain size, clay fraction and plasticity; sorted according to the classified geomorphological/hydrogeological contexts, they highlight good correlations among dry density, mechanical work performed in selected stages of the oedometric test, swelling and non-swelling behaviour, and electrical conductivity of the squeezed pore waters. The work performed for swelling and non-swelling samples shows well-defined differences; this endorse the relevance of pore-water salinity in determining the volumetric state of clay soils under overburden and specific hydrogeological border conditions, which together define a specific energetic state. [ABSTRACT FROM AUTHOR]