In China, excessive nitrogen fertilizer application in sweet maize fields contributes to greenhouse gas emissions. This study used maize straw (MS), cow dung (CD), biogas residue (BR), and straw-based biochar (CB) to substitute the mineral nitrogen fertilizer at 20% and 50% ratios in the Pearl River Delta in China. In comparison with a conventional amount of mineral nitrogen fertilizer (CK), the soil organic carbon (SOC) storages of the different treatments increased by 6.5–183.0%. The CB treatment significantly improved the inert organic carbon pool in the soil, while other types of organic materials promoted the formation of activated carbon pools. The treatments increased the soil carbon pool management index by 21.1–111.0% compared to the CK. Moreover, the CB treatments increased the soil carbon sequestration index by 78.3% and 155.8% compared to the CK. In general, substituting the mineral N fertilizer with BR, CB, and CD could improve the SOC accumulation in sweet maize farmland in South China. The CB at the high substitution level was the best measure for stabilizing carbon sequestration in the sweet maize cropping system. This experiment provides valuable information for ensuring the clean production of sweet maize in a typical subtropical area in East Asia.

No abstract

The economic value and consumer acceptance of Pu-erh tea heavily depend on the production year. The present study aims to evaluate the potential of utilizing laser-induced breakdown spectroscopy (LIBS) in conjunction with chemometric models to identify Pu-erh raw tea from various production years. The research utilizes tea leaves from a common source in 2008, 2013, and 2018 as the analytical samples. One hundred spectral datasets were collected for each type of tea, and these datasets are randomly partitioned into cross-validation and test sets in a 3:2 ratio. Subsequently, by utilizing threshold peak finding to extract features from the baseline-corrected LIBS spectrum, 21 spectral datasets are identified and input into LDA, SVM, EML, and KNN classification models for analysis. Results demonstrate that the LDA model achieves superior performance in identifying tea leaf years, attaining a recognition rate of 98.75%. Additionally, the average recognition rate of the other three algorithms in three-classification tasks exceeds 90%. Overall, this study confirms the feasibility and effectiveness of utilizing LIBS in conjunction with machine learning algorithms for discriminating Pu-erh raw tea originating from different production years.

No abstract

No abstract