Agronomy

Timely and accurate crop yield estimation is vital for food security and management decision-making. Integrating remote sensing with machine learning provides an effective solution. In this study, based on canopy hyperspectral data collected by an ASD FieldSpec 3 handheld spectrometer during the critical growth stages of winter wheat, 18 vegetation indices (VIs) were systematically calculated, and their correlation with yield was analyzed. At the same time, a continuous projection algorithm, Successive Projections Algorithm (SPA), was used to screen the characteristic bands. Recursive Feature Elimination (RFE) was employed to select optimal features from VIs and characteristic spectral bands, facilitating the construction of a multi-temporal fusion feature set. To identify the superior yield estimation approach, a comparative analysis was conducted among four machine learning models: Deep Forest (DF), Support Vector Regression (SVR), Random Forest (RF), and Gaussian Process Regression (GPR). Performance was evaluated using the coefficient of determination (R²), root mean square error (RMSE), and relative root mean square error (rRMSE). Results indicate that the highest correlations between VIs and grain yield were observed during the flowering and grain-filling stages. Independent analysis showed that VIs reached absolute correlations of 0.713 and 0.730 with winter wheat yield during the flowering and grain-filling stages, respectively, while the SPA further identified key bands primarily in the near-infrared and short-wave infrared regions. On this basis, integrating multi-temporal features through RFE significantly improved the accuracy of yield estimation. Among them, the DF model with the fusion of flowering and filling stage features performed best (R² = 0.786, RMSE = 641.470 kg·hm⁻², rRMSE = 15.67%). This study demonstrates that combining hyperspectral data and VIs from different growth stages provides complementary information. These findings provide an effective method for crop yield estimation in precision agriculture.

In Guangxi, China, the area used to plant sugarcane is growing in order to meet the Fourteenth Five-Year Plan’s objective of sugar self-sufficiency (2021–2025). Comprehensive soil heavy metal data are necessary for growing area expansion in order to inform farmers and policymakers. Here, we analyzed soil samples from ten sugarcane-growing counties/districts of Guangxi by employing four different risk assessment indices. Our results indicate that the studied soils are moderately to strongly acidic and are deficient in soil organic matter (<6 g/kg). Single-factor pollution index evaluation revealed detectable heavy metal pollution, with Cd present above reference levels in all ten areas, Cr in six, Pb in four, As in two, and Hg in two areas. The Nemerow comprehensive pollution index indicated that the overall soil pollution level was mild, except for Jiangzhou district (moderate). The geo-accumulation index revealed significant anthropogenic enrichment, with severe Cr pollution (Igeo > 3) across all regions and Pb and As contamination ranging from moderate to severe, particularly in Jiangzhou district. Contrastingly, Cd and Hg showed no significant enrichment (Igeo < 0) relative to the local background, though their sources require further investigation. The potential ecological risk assessment showed a high risk, specifically from As in Jiangzhou district, which was the only area showing a moderate comprehensive potential ecological risk. A significant positive correlation was found between the total and bioavailable contents of all five heavy metals, whereas soil pH and organic matter were significantly negatively correlated with the bioavailability of Cr and Pb, but positively correlated with As and Hg. The availability of Cd, however, was independent of pH and OM, suggesting the influence of other, unmeasured geochemical factors. These results highlight specific and localized environmental risks that may require targeted management to ensure agricultural safety, ecosystem health, and sustainable sugarcane production.

To evaluate the optimal substitution ratio of organic fertilizer for chemical nitrogen fertilizer and its underlying mechanisms, a pot experiment was conducted in the rhizosphere soil of oat (Avena sativa) on the Qinghai–Tibet Plateau. Five treatments were established: CK (control), T1 (chemical fertilizer alone), T2 (100% organic fertilizer substitution for chemical nitrogen fertilizer), T3 (30% organic fertilizer substitution for chemical nitrogen fertilizer), and T4 (60% organic fertilizer substitution for chemical nitrogen fertilizer). We analyzed soil carbon fractions, microbial community structure, carbon-cycling enzyme activities, and yield responses and applied partial least squares–structural equation modeling (PLS-SEM) to identify key regulatory pathways. The results showed that 30% organic substitution (T3) was associated with optimized soil carbon pools, improved microbial community composition, and enhanced carbon-cycling enzyme activities, while reducing the abundance of potentially harmful fungi. Structural equation modeling indicated that β-glucosidase activity and the relative abundance of Proteobacteria were the primary drivers of yield, together explaining 76% of its variation. The ecosystem multifunctionality index (EMF) was significantly and positively correlated with yield. In summary, under the conditions of this experiment, 30% organic fertilizer substitution achieved a favorable balance between soil ecological functions and crop yield, providing a valuable reference for sustainable nutrient management in oat production in high-altitude cold regions.