Artificial neural networks to investigate the bioavailability of selenium nanoparticles in soil–crop systems†
Abstract
While selenium nanoparticles (Se NPs) can effectively enrich crop yield and quality, the limited research on the interactions between Se NPs and soil–crop systems hinders their potential use in agriculture. Hence, the soil application of Se NPs (0 [control] and 0.5 mg kg−1) and Na2SeO3 (1.11 mg kg−1) was used to enhance rice quality and yield. The artificial neural network (ANN) approach was used to model and simulate the response of soil properties (SPs) and plant physiological activities (PPAs) under different treatments at different time stages (30, 60, 90, and 120 days). The results indicate that Se NPs can enhance photosynthesis, leading to increased yield (1.33-fold) and quality of rice (Se-enriched rice, 3.46-fold). The effects of Se NPs on rice growth and development were found to be time-dependent. Soil properties, including soil organic matter (TOC), ammonium nitrogen (NH4+), pH, redox potential (Eh), and conductivity (Ec), emerged as crucial factors influencing the observed effects. With the progression of time, plant physiological activities, including chlorophyll (Chl), net photosynthetic rate (Pn), stomatal conductance (Gs), and optimal/maximal photochemical efficiency of PS II in the dark (Fv/Fm), exhibited an increasing level of importance. Moreover, the processes of Se NPs affecting the yield and quality were distinct, with TOC being more important for rice yield and Ec being more significant for quality. Therefore, this study offers a novel approach to assess the bioavailability of Se NPs in soil–crop systems and provides valuable insights into the potential for using Se NPs to enhance rice productivity and quality. The use of model-based interpretation methods combined with experimental data allows for a more comprehensive understanding of the advantages and disadvantages of NPs in soil–plant systems and facilitates the implementation of safe design options for NPs in agriculture.