Timely and accurate monitoring of winter wheat yields is beneficial for the macro-guidance of agricultural production and for making precise management decisions throughout the winter wheat reproductive period. The accuracy of crop yield prediction can be improved by combining unmanned aerial vehicle (UAV)-based multispectral data with deep learning algorithms. In this study, 16 yield-sensitive vegetation indices were constructed, and their correlations were analyzed based on UAV multispectral data of winter wheat at the heading, flowering, and filling stages. Seven input variable sets were obtained based on the combination of data from these three periods, and four generalized machine learning algorithms (Random Forest (RF), K-Nearest Neighbor (KNN), Bagging, and Gradient Boosting Regression (GBR)) and one deep learning algorithm (1D Convolutional Neural Network (1D-CNN)) were used to predict winter wheat yield. The results showed that the RF model had the best prediction performance among the generalised machine learning models. The CNN model achieved the best prediction accuracy based on all seven sets of input variables. Generalised machine learning models tended to underestimate or overestimate yields under different irrigation treatments, with good prediction performance for observed yields < 7.745 t·ha⁻¹. The CNN model showed the best prediction performance based on most input variable groups across the range of observed yields. Most of the differences between observed and predicted values (Y_i) for the CNN models were distributed between −0.1 t·ha⁻¹ and 0.1 t·ha⁻¹, and the model was relatively stable. Therefore, the CNN model is recommended in this study for yield prediction and as a reference for future precision agriculture research.

Timely and accurate monitoring of the nitrogen levels in winter wheat can reveal its nutritional status and facilitate informed field management decisions. Machine learning methods can improve total nitrogen content (TNC) prediction accuracy by fusing spectral and texture features from UAV-based image data. This study used four machine learning models, namely Gaussian Process Regression (GPR), Random Forest Regression (RFR), Ridge Regression (RR), and Elastic Network Regression (ENR), to fuse data and the stacking ensemble learning method to predict TNC during the winter wheat heading period. Thirty wheat varieties were grown under three nitrogen treatments to evaluate the predictive ability of multi-sensor (RGB and multispectral) spectral and texture features. Results showed that adding texture features improved the accuracy of TNC prediction models constructed based on spectral features, with higher accuracy observed with more features input into the model. The GPR, RFR, RR, and ENR models yielded coefficient of determination (R²) values ranging from 0.382 to 0.697 for TNC prediction accuracy. Among these models, the ensemble learning approach produced the best TNC prediction performance (R² = 0.726, RMSE = 3.203 mg·g⁻¹, MSE = 10.259 mg·g⁻¹, RPD = 1.867, RPIQ = 2.827). Our findings suggest that accurate TNC prediction based on UAV multi-sensor spectral and texture features can be achieved through data fusion and ensemble learning, offering a high-throughput phenotyping approach valuable for future precision agriculture research.

Drought stress is one of the key factors restricting crop yield. The beneficial effects of exogenous proline on crop growth under drought stress have been demonstrated in maize, rice, and other crops. However, little is known about its effects on wheat under drought stress. Especially, the water-holding capacity of leaves were overlooked in most studies. Therefore, a barrel experiment was conducted with wheat at two drought levels (severe drought: 45% field capacity, mild drought: 60% field capacity), and three proline-spraying levels (0 mM, 25 mM, and 50 mM). Meanwhile, a control with no stress and no proline application was set. The anatomical features, water-holding capacity, antioxidant capacity, and proline content of flag leaves as well as grain yields were measured. The results showed that drought stress increased the activity of catalase and peroxidase and the content of proline in flag leaves, lessened the content of chlorophyll, deformed leaf veins, and decreased the grain yield. Exogenous proline could regulate the osmotic-regulation substance content, chlorophyll content, antioxidant enzyme activity, water-holding capacity, and tissue structure of wheat flag leaves under drought stress, ultimately alleviating the impact of drought stress on wheat yield. The application of proline (25 mM and 50 mM) increased the yield by 2.88% and 10.81% under mild drought and 33.90% and 52.88% under severe drought compared to wheat without proline spray, respectively.

The rice waterlogging stress test was conducted at the experimental base of the College of Agriculture, Yangtze University, using Yangxian You 418 as the test subject, in order to investigate the impact of waterlogging on rice growth during the period from July to August each year. Six waterlogging stress tests with different waterlogging depth (1/4 plant height (1/4PH), 2/4 PH, and 3/4 PH) and duration (5 d and 7 d) were set up at the jointing–booting stage of rice (T1: 1/4 PH, 7 d; T2: 2/4 PH, 7 d; T3: 3/4 PH, 7 d; T4: 1/4 PH, 5 d; T5: 2/4 PH, 5 d; T6: 3/4 PH, 5 d;) with shallow water irrigation (CK) as control. The plant height, population leaf area, above-ground dry matter, and the yield of rice were measured. The correlation between the waterlogging depth and rice yield reduction was analyzed, and the flood disaster threshold index of rice was established. The results showed that at the end of stress, the plant height of all waterlogged treatments exceeded CK, and the plant height of T3 and T6 treatments significantly increased by 31.90% and 15.93%, respectively. The leaf area of rice treated with T1, T3, T4, and T5 was higher than CK (p < 0.05), and the above-ground dry matter of rice treated with T2, T3, T4, T5, and T6 was higher than CK (p < 0.05). When normal irrigation was restored to the maturity stage, the plant height of all rice treated with waterlogging was still higher than CK (p < 0.05). However, as the degree of waterlogging increased, rice yield decreased significantly, with a notable reduction of 31.68% observed in the T3 treatment compared to CK. Assuming a drainage index based on a 20% decrease in rice yield, it is imperative that the ratio of flooded depth to plant height remains below 37% when waterlogging persists for 7 days in rice cultivation. These research findings offer crucial scientific insights for implementing effective drainage management measures during flood disasters in rice paddies.

In recent years, the alternating occurrence of high temperature and waterlogging disasters in South China has seriously reduced the yield of single cropping rice. Studying the changes in anatomical structure of the rice root system could provide theoretical basis for understanding the mechanisms of high temperature and waterlogging stress. To examine interactions between temperature and waterlogging stress, an experiment was set up in a growth chamber consisting of two temperatures (moderate, 30–34 °C and high, 35–38 °C) with three depths of flooding (0–5, 10 and 15 cm). Treatments commenced at the booting stage and lasted five days, after which all treatments were returned to a 0–5 cm flooding depth and the same temperature regime. Observations were made immediately after cessation of treatments, then after 5, 10 and 20 d to test the effect of treatments on subsequent root anatomical development. The low-stress control (0–5 cm, medium temperature) showed no change with time in aerenchyma area, thickness of the outer root, stele diameter, and the number nor diameter of xylem vessels. Root diameter and stele diameter under the high-stress control (0–5 cm, high temperature) were decreased by 29.09% and 15.28%, respectively, at the booting stage, whereas the high stress control (15 cm, high temperature) affected only the vessel diameter, reducing it by 14.11% compared with that in the low-stress control (0–5 cm, medium temperature). Compared to the high-stress control (0–5 cm, high temperature), the interaction of high temperature and waterlogging stress alleviated the inhibiting effect of the changes in the root system, especially after the end of the stress. We thought that waterlogging could reduce the damage of high temperature on rice root growth. Low water depth waterlogging has little effect on rice root system and aerenchyma area root diameter at 0, 5, 10 and 20 d after the stress ended, and the thickness of the outer root, stele diameter and the number and diameter of vessels at 0 d under M15 (15 cm, medium temperature) had no significant difference compared with M5 (0–5 cm, medium temperature). However, the increase in rice root diameter, stele diameter, thickness of the outer root cortex depth and vessel diameter were inhibited under high temperature stress at the booting stage. Root diameter and stele diameter under H5 (0–5 cm, high temperature) were decreased by 29.09% and 15.28%, respectively, at the booting stage, whereas H15 (15 cm, high temperature) affected only the vessel diameter, reducing it by 14.11% compared with that in the M5. Compared to H5, the interaction of high temperature and waterlogging stress alleviated the inhibiting effect of the changes in the root system, especially after the end of the stress. We thought that waterlogging could lighten the damage of high temperature on rice root growth.

As one global issue, soil salinization has caused soil degradation, thus affecting the sustainable development of irrigated agriculture. A two-year study was conducted in 2018 and 2019 to identify the effects of subsurface drainage spacing on soil salinity and groundwater level, the latter of which is in a high-water table in drip irrigation cotton fields in the Tarim Basin oasis in southern Xinjiang, China. Three subsurface drainage treatments, with a drain spacing of 10 m (W10), 20 m (W20), and 30 m (W30), respectively, and a drainage-absent treatment (CK), are tested. With CK, soil salinity in the 0–60 cm layer was accumulated within a year. In contrast, the subsurface drainage reduced the soil salinity at a leaching rate of 10–25%. When decreasing the drain spacing, it was found that the soil desalination rate increased significantly (p < 0.05) with good repeatability. Experimental results showed that the fitting equation of the soil salinity leaching curve could accurately describe the soil salinity leaching pattern of drip irrigation, and thus could be further used to inversely determine the theoretical drip irrigation leaching quota for those soils with different salinity degrees. As such, subsurface drainage could effectively control the groundwater table. Compared with CK, subsurface drainage deepened the groundwater table and mitigated the fluctuation of the groundwater level. These effects were strengthened by reducing the drain spacing. Correspondingly, the influence of the fluctuation of the groundwater table was reduced.