Fertilizer nitrogen (N) is a main pollutant in the agricultural ecosystem, while the fate of fertilizer N influenced by different irrigation modes is not well comparatively investigated. In this study, the distribution of fertilizer N in soil layers and tomato organs as well as its loss under drip, spray and flood irrigation with different quotas of 140, 180 and 220 m³ ha⁻¹ were evaluated quantitatively by using nitrogen-15 (¹⁵N) labeled urea (abundance of 19.6%) as fertilizer source. The results showed that the plant ¹⁵N, soil ¹⁵N and ¹⁵N loss accounted for 27.9–47.8%, 38.8–54.0% and 10.3–21.9% of the total applied ¹⁵N, respectively. The amount of ¹⁵N absorbed by plants was significantly (p < 0.05) higher under drip and spray irrigation in comparison to flood irrigation with the same irrigation quota. The maximum ¹⁵N use efficiency and the minimum ¹⁵N residual were detected under drip irrigation with quota of 180 m³ ha⁻¹, indicating that the supply and demand of urea-¹⁵N was more synchronized under such an irrigation mode. The ¹⁵N loss increased obviously as irrigation quota increased. Moreover, the correlation analysis between ¹⁵N loss and the possible impact factors indicated that the soil mineral ¹⁵N content after irrigation was one important factor influencing the ¹⁵N loss. Among the three irrigation modes, spray irrigation caused the lowest ¹⁵N loss of 10.3–13.1% when using the same irrigation quota. It was concluded that the irrigation modes have profound impacts on the fate of urea-¹⁵N. Irrigation could be used as a regulation pathway of plant N absorption and agricultural N output.

Irrigation has a significant impact on NxO (N2O and NO) emissions from cultivated land, yet the N2O or NO emission among the irrigation lower limits under different irrigation modes has not been well compared. In an irrigated lettuce field, three DR (drip irrigation) lower limits were designed, including 75% (DR1), 65% (DR2) and 55% (DR3) field capacity, and one FI (furrow irrigation) lower limit (65% field capacity). The N2O and NO emission fluxes and soil nitrogen (N) forms were determined, and the linear correlation between these indicators was analyzed. Results showed that under the same irrigation regime, the N2O and NO emissions from furrow irrigation treatment increased by 36.8% and 45.2% respectively compared to that from drip irrigation treatment. The cumulative N2O and NO emissions under DR3 were 30.2% and 28.6% higher than under DR1, respectively. Moreover, DR1 was also the lowest among the four treatments in soil NO3--N concentration. The N2O and NO emission fluxes were more correlated to soil NH4+-N (r=0.88 and 0.76) or NO2--N (r=0.90 and 0.80) concentration than soil NO3--N and soluble organic N, indicating that N2O and NO were mainly produced by the soil nitrification process. When the irrigation regime was the same, N2O and NO emissions were lower with drip irrigation than with furrow irrigation. Besides, drip irrigation with small quota but high frequency reduced N2O and NO emission compared to that with large quota but low frequency. Keywords: nitrous oxide, nitric oxide, drip irrigation, furrow irrigation, soil nitrogen DOI: 10.25165/j.ijabe.20231604.7961 Citation: Hou M M, Xiao Y, Zhu Q Y, Chen J N, Huang H, Jin Q, et al. NxO emissions in response to the irrigation lower limits under different irrigation modes in a lettuce field. Int J Agric & Biol Eng, 2023; 16(4): 159–167.

Soil residual N is a potential factor threatening the environment, but it is also an N fertilizer resource. Few studies have evaluated the fate of soil residual N under agronomic practice. The objective of this study was to investigate the distribution of residual N and its possible influencing factors with different irrigation regimes. Under three N residual situations created by the previous season using the ¹⁵N labeled urea, we employed lettuce as the plant material and three lower limits of drip irrigation including 75% (DR1), 65% (DR2), and 55% (DR3) accounting for the field water capacity as experimental treatments. A furrow irrigation treatment (FI) with the same irrigation regime as DR2 was used as control. Results showed that 2.1–4.8% of the residual ¹⁵N from the previous season was absorbed by the succeeding lettuce, 78.0–84.4% was still remained in the 0–80 cm soil, and 10.9–20.0% was unaccounted for. After harvest of succeeding lettuces, the soil residual ¹⁵N mainly existed in the mineral form. Moreover, the lettuce reuse efficiency for¹⁵N was positively correlated with the total residual ¹⁵N amount (p < 0.01) and the mineral ¹⁵N amount (p < 0.01). The overall results indicated that an appropriate irrigation regime (DR2) was conducive to promoting absorption of residual N by succeeding crop.

Early and accurate identification of tomato leaf diseases constitutes a key safeguard for mitigating economic losses in tomato production. Conventional tomato leaf disease detection methodologies are constrained by inherent limitations, such as low operational efficiency, inadequate detection precision, and limited adaptability to environmental fluctuations. In contrast, the integration of deep learning techniques has yielded improvements in this research domain. Consequently, the development of deep learning-based approaches for the rapid and precise detection of tomato leaf diseases holds considerable theoretical significance and practical application value. To improve the detection accuracy of tomato leaf diseases, this study proposes a transfer learning-based DenseNet disease recognition model named DenseNet-BiFPN-ECA Fusion Network. The bidirectional feature pyramid network (BiFPN) is introduced at the terminal of DenseNet121 to achieve multi-scale feature fusion, while the efficient channel attention (ECA) mechanism is applied to enhance the discriminative capacity of fused features. Classification is ultimately completed via a global average pooling layer and a fully connected layer. The experimental results demonstrate that the improved model achieves an accuracy of 90.63% on the small-sample tomato leaf dataset collected from complex greenhouse environments, representing an improvement of 20.32 percentage points over the original DenseNet121 model. On the large-scale open-source Plant Village dataset, the model attains an accuracy of 98.47%, significantly outperforming the baseline models. Furthermore, a comparative analysis shows that the highest accuracy achieved by DenseNet, ResNet101, and VGG16 models on the same dataset is only 83.59% (within ±0.5%). This result validates the effectiveness of DenseNet-BiFPN-ECA Fusion Network in disease recognition tasks. The model provides a reliable technical reference for the intelligent diagnosis of tomato leaf diseases.