To investigate ammonia volatilization characteristics in paddy fields under the combined regulation of bio-organic fertilizer application and controlled irrigation, a rice plot experiment was conducted. The experiment involved two irrigation modes: flooded irrigation (F) and controlled irrigation (C). These were combined with three fertilization strategies: 100% chemical fertilizer (A); 15% of chemical nitrogen replaced by bio-organic fertilizer on an equal-nitrogen basis (B); and 30% of chemical nitrogen replaced by bio-organic fertilizer on an equal-nitrogen basis (C). The results showed that under the same irrigation regime, bio-organic fertilizers with chemical nitrogen fertilizers reduced the mean ammonia volatile flux from paddy fields and ammonia volatile accumulation from paddy fields by 11.34% ~ 31.47% and 13.74%~31.12%, respectively, during the rice growth period, bio-organic fertilizer substitution treatments decreased average ammonium nitrogen and total nitrogen concentrations in surface water by 7.92%~38.12% and 2.39%~14.92%, respectively. The 30% substitution treatment (C) was the most effective in this regard. A significant positive correlation (p<0.05) was observed between the paddy water NH??-N concentration and ammonia volatilization flux. Under the same fertilization conditions, controlled irrigation treatment showed lower ammonia volatilization flux, cumulative ammonia volatilization, and average concentrations of total nitrogen and ammonium nitrogen in surface water compared with flooded irrigation throughout the rice. The comprehensive analysis concluded that nitrogen replacement of 30% chemical nitrogen fertilizers by bio-organic fertilizers and other nitrogen under controlled irrigation mode can significantly reduce ammonia volatilization emission and improve the utilization rate of nitrogen fertilizers in paddy fields.

This study investigates the effects of sluice and dam control on water quantity and quality in main farmland drainage ditches. A four-year in-situ observation experiment was conducted in five main drainage ditches within a typical plain agricultural area, where monitoring sections for water level and quality were established. The temporal and spatial variation characteristics of water level and water quality of ditches were analyzed with and without sluice dam. The results showed that sluice and dam controls effectively raised upstream water levels and regulated the spatiotemporal distribution of water resources. The water level difference between the upstream and downstream of the sluice was 1.27 m, and the water level at 3 km and 6 km upstream were 0.96 m and 0.22 m higher than the average water level of the adjacent control ditch, respectively. The water level difference between the upstream and downstream of the dam was 0.64 m. Sluice and dam controls negatively affected water quality, particularly near the control structures. The average concentrations of TN, TP and COD_Mn of the CT were 3.01、0.20 and 5.34 mg/L, respectively, and the value of CK were 2.15、0.15 and 4.21 mg/L. The comprehensive Water Quality Index (WQI) of controlled ditches was 20% lower than uncontrolled ditches. The control structures also significantly altered the seasonal response pattern of water quality. In the controlled ditches, water quality improved during the flood season compared to the non-flood season, primarily due to a decrease in TN concentration from 3.78 to 2.70 mg/L, which resulted in the WQI score increasing from 44.79 to 46.67. The water quality of the uncontrolled canal deteriorated during the flood season compared to the non-flood season, with the WQI changed from 6.22 to 52.17.

To reveal the water requirements and resilience to climate change of different plant types in karst peak-cluster depression areas, this study integrates the unique rock-soil structure of such regions to explore and establish a sustainable development model for three-dimensional ecological agriculture that harmonizes ecology and economy. The ultimate goal is to provide scientific guidance for local agricultural practices. In a typical karst peak-cluster depression in Pingguo, Guangxi, we selected 15 dominant tree, shrub, and grass species as our study subjects. Photosynthetic physiological and ecological parameters were measured during the same time period to analyze their variation patterns and responses to environmental changes. Based on the unique rock-soil structure of the peak-cluster depression, a plant optimization configuration model that harmonizes "ecology and economy" was proposed. The results showed that native plants generally exhibit strong adaptability to environmental changes in the study area, and appropriate human interventions can enhance their resilience to extreme weather. There are also significant differences in the photosynthetic physiological and ecological characteristics of introduced plants. Suitable species should be selected for cultivation based on the soil and water conditions of different geomorphic areas. Invasive plants cause considerable damage to the ecosystem and warrant attention. For the poor soil and water conditions on peak slopes, two herbaceous plants, Ficus pumila and Cichorium intybus, are recommended for rapid revegetation and effective control of rocky desertification. On gentle slopes, the strategy of interplanting Dalbergia odorifera with Millettia reticulata was proposed to develop an under-forest economy, simultaneously conserving soil and water while increasing local household incomes. The depressions, having the best water and soil conditions, are also prone to flooding. Thus, two economic crops, Pennisetum purpureum and maize, were selected for their high transpiration and photosynthetic rates to improve water use efficiency while mitigating flood risks effectively.

Based on MODIS remote sensing data, NDVI、Wet、LST、NDBSI、BCSI and PM2.5 were extracted and standardized. The standardized remote sensing ecological index (SRSEI) was constructed by principal component analysis on the Google Earth Engine cloud platform. The Theil-Sen median trend analysis, Mann-Kendall test, Hurst index, and Geodetector were used to evaluate the ecological environment of Fenhe River Basin and analyze the driving factors. The results indicate the following: ① The contribution rate of the first principal component is greater than 67%, and the SRSEI of the mountainous land with forest land as the main land type is greater than 0.6, and the SRSEI of the basin with impervious surface and farmland as the main land type is less than 0.4. ② From 2000 to 2020, the average SRSEI of the Fenhe River Basin increased from 0.447 9 to 0.559 3. An improving trend in ecological quality was observed across 92.29% of the basin. The areas with deteriorating trend of the ecological environment are mainly distributed in Taiyuan, Linfen and Yuncheng basins, of which 40.24% of the ecological environment deterioration trend is persistent. ③ For the driving force of the ecological environment in the Fenhe River Basin: BCSI、NDVI、NDBSI and LST are the four factors with the largest driving force, and LST ∩ BCSI and Wet ∩ BCSI can have a more significant impact on the ecological environment. It is feasible to evaluate the ecological environment of Fenhe River Basin by SRSEI model. During the 21 years, the overall ecological environment of the watershed has improved; the ecological environment of the basin is fragile; ecological governance policies should focus on the basin ecological governance while maintaining the mountain ecological environment. The main ecological governance measures should be to improve vegetation coverage, control and repair saline-alkali land, increase soil moisture and reduce surface temperature. The main ecological governance measures should be to improve vegetation coverage, control and repair saline-alkali land, increase soil moisture and reduce surface temperature.

This study investigates the impact of polyethylene (PE) floating spheres with different diameters on evaporation suppression efficiency (ESE) under the same surface coverage conditions, considering their varying sensitivities to radiation and aerodynamic factors. Experiments were conducted under three environmental conditions—greenhouse, wind tunnel, and outdoor settings—to evaluate the ESE performance of floating spheres with varying diameters. Gaussian Process Regression (GPR) was applied to fit the ESE data and determine the optimal floating sphere diameter for each environment. Results showed that due to their high net radiation absorption capacity, the surface temperatures of the floating spheres were significantly higher than that of the water surface. In outdoor experiments, wind-induced surface cooling of the water led to the largest temperature differences between the floating spheres and the underlying water, with surface temperatures of 60、80 and 100 mm spheres reaching 65.7、61.5 and 64.3 ℃, respectively—representing increases of 52.7%、44.2% and 49.6% relative to the water surface temperature. Moreover, floating coverage enhanced thermal stratification in the water body, with a distinct temperature inflection observed at depths of 40~60 cm under both calm and outdoor conditions. A seasonal inversion of stratification was also observed, with decreased temperatures at these depths in summer and increased temperatures in autumn. Among the three environments, the 80 mm diameter PE spheres consistently exhibited stable ESE performance, achieving suppression rates of 72.97%、71.74% and 73.44%, respectively. GPR fitting identified the optimal diameters for static wind, isolated wind, and outdoor environments as 84.6、80.2 and 77.0 mm, respectively. Based on both experimental results and model analysis, a floating sphere diameter of approximately 80 mm is recommended to achieve optimal evaporation suppression performance.

To improve the accuracy of monthly precipitation time series prediction, this study proposes a hybrid modeling framework combining Wavelet Packet Transform (WPT), the Elk Herding Optimization (EHO) algorithm, and machine learning. The framework aims to enhance the generalization performance of six base models: Hybrid-kernel Relevance Vector Machine (HRVM), Hybrid-kernel Least Squares Support Vector Machine (HLSSVM), Hybrid-kernel Support Vector Machine (HSVM), and their single-kernel counterparts (RVM, LSSVM, SVM). A total of 18 combined models were developed and validated using monthly precipitation data from two rainfall stations in Dali Prefecture, Yunnan Province. The methodology involved three main steps. First, the original time series were decomposed into sub-series using WPT at one, two, and three levels (WPT1、WPT2、WPT3). Second, for each base prediction model, the EHO algorithm found the optimal hyperparameters by optimizing a fitness function built on the training set. Finally, the optimized models predicted each decomposed sub-series, and the results were reconstructed to form the final precipitation forecast. The results indicated that: ① All 18 proposed models achieved good fitting and prediction accuracy. Among them, the WPT3-EHO-HRVM, WPT3-EHO-HLSSVM, and WPT3-EHO-HSVM models performed best, yielding the lowest prediction errors, with Mean Absolute Error (MAE) values between 0.81 and 1.70 mm and a coefficient of determination (R2) between 0.999 6 and 0.999 9. ② For any given decomposition level, the hybrid-kernel models consistently outperformed their single-kernel counterparts, as their ability to combine different kernels allowed them to better adapt to diverse data distributions and significantly improve prediction accuracy. ③ Prediction accuracy increased with the WPT decomposition level; the decomposition effect was ordered WPT3 > WPT2 > WPT1. ④ The EHO algorithm effectively improved the prediction accuracy and efficiency of all base models by optimizing their hyperparameters.

To identify an accurate model for estimating evapotranspiration(ET) in semi-arid regions, this study was conducted on a maize farmland ecosystem at the Huining experimental station in Gansu, China. We employed a new and improved Parallel Two-Source Energy Balance (P-T-SEB) model coupled with four different stomatal conductance models: the Leuning (RL), Jarvis-Stewart (JS), Stannard (ST), and Irmak-Mutiibwa (IM) models. The model parameters were calibrated using a Bayesian-based differential adaptive algorithm to estimate ET during the peak growing season of maize. Model performance was then validated by comparing the simulated ET against observational data measured using the eddy covariance method. The results showed that the P-T-SEB model significantly outperformed the Shuttleworth-Wallace (SW) model, reducing the root mean square error (RMSE) by 9%. The RL model performed the best among the stomatal conductance models, with its RMSE being 75%, 66%, and 43% lower than those of the JS, ST, and IM models, respectively. Based on all evaluation indices, the performance ranking of the four stomatal conductance models was: RL > IM > ST > JS. We concluded that coupling the RL stomatal conductance model with the P-T-SEB model provides the most accurate estimates of evapotranspiration for the studied semi-arid region.

This study aimed to investigate the physiological effects of sediment-laden flooding on double-cropping rice and to reveal the mechanisms of its yield reduction in the Poyang Plain. The goal is to provide a theoretical basis for developing measures to mitigate yield losses from such events. This study used a combination of field experiments and indoor analysis. Based on the results of previous flooding experiments and the average sediment concentration of various rivers in the Poyang Lake Basin (0.07~0.73 kg/m³) over the years, two types of flooding depths were set: 2/3 plant height and full flooding. Three different sediment contents, S1 (0.0 kg/m³), S2 (0.5 kg/m³), and S3 (1.0 kg/m³), were used to observe and analyze the root activity, sword leaf membrane lipid peroxidation, antioxidant enzyme activity, chlorophyll content, and other indicators of early and middle rice at 6 and 9 days after flooding. The results showed that the root system of early rice was well-developed during the heading and flowering stage, but severe damage was caused by complete flooding. In addition, flooding during the jointing and booting stage of middle rice severely inhibited root function, reducing root absorption area and root activity by 37.00% and 48.94% (early rice fully flooded), and 63.54% and 79.36% (middle rice) (p<0.05), respectively. This stress triggered a rapid response from the antioxidant enzyme system; however, membrane lipid peroxidation still intensified as a result of cell membrane damage. Chlorophyll degradation and complete flooding lead to severe hypoxia in rice, causing rapid changes. Compared to 2/3 flooding, complete submergence in early rice increased SOD and POD activities by 43.56% and 34.33%, and MDA and Proline (Pro) content by 95.03% and 119.79%, respectively. The content of Chl a and Chl b in middle rice sword leaves significantly decreased by 8.06% and 28.02%, respectively. The submergence and increase in sediment content of double cropping rice lead to a further decrease in root absorption area and root activity under full submergence. SOD activity first increases and then decreases, POD activity continues to increase, membrane lipid peroxidation intensifies, MDA and Pro content rapidly increase, and photosynthesis is severely inhibited. During full flooding, the functional blades attach sediment to block stomata and provide shading for turbid sediment water, resulting in a significant impact of sediment during full flooding.

Sediment deposition in channels is a bottleneck affecting the normal operation and safety of pumping irrigation districts along the Yellow River. Studying the properties and transport patterns of sediment entering the channels, and determining the channel's scour and deposition under different water and sediment inflow conditions, is key to solving this problem. This study analyzed the changes in sediment grain size and concentration along the main channel of the Madihao Pumping Irrigation District during different irrigation periods. It also discussed the incipient velocity of sediment and the sediment-carrying capacity of the flow using different theoretical formulas. A water and sediment transport model was used to simulate the scour-deposition equilibrium points in the main channel under different sediment gradations, concentrations, and discharges. The results show that: ① In the main channel, sediment with a particle size of less than 100 μm accounts for over 75% of the total suspended load. The suction and disturbance from the pumps cause bottom sediment from the riverbed near the pumping station's forebay to enter the main channel. ② The energy input from the pumps causes the sediment concentration in the main channel to first increase and then decrease. Compared with other formulas, Zhang Luohao's incipient velocity formula and Zhang Hongwu's sediment-carrying capacity formula were found to be more suitable for the hydraulic calculations in this irrigation district. Under current flow and boundary conditions, the channel is prone to siltation when the median particle size (d??) of the incoming sediment is ≥ 90 μm. ③ By simulating the channel scour and deposition processes under various sediment gradations, concentrations, and discharges, 17 scour-deposition equilibrium points were determined. These findings will help reveal the mechanisms of the channel scour and deposition process in the main channel of the irrigation district and provide a scientific basis for solving the channel deposition problem.

To address the complexity and lack of portability of traditional open-channel flowmeters, this paper introduces a novel, portable device for rectangular channels, termed the 'plate flow measurement device. Its measurement performance and hydraulic characteristics were investigated through a combination of theoretical analysis, physical experiments, and numerical simulation. The results showed that the measured channel flow rate (Q) is a function of the force exerted on the measuring plate (F_R ) and the plate's width (B). Consequently, a functional relationship, Q = f(F_R, B), was established. The velocity profiles upstream and downstream of the plate exhibited significant spatial heterogeneity, with the downstream flow being more affected by the device than the upstream flow. The velocity distribution in each cross-section was symmetrical about the channel's central axis. Flow velocity was greatest near this central axis, and the point of maximum velocity was observed to be below the free water surface, rather than at it. Both the vertical and transverse velocity distributions within the rectangular cross-section conformed to a parabolic profile.

To establish the application and promotion of low flowrate and high frequency smart drip irrigation technology for cotton, and to determine the appropriate discharge rate of emitter and irrigation quota, experimental studies were carried out in Changji, Xinjiang between 2022 and 2024. An experimental field was divided into four equal plots. All plots shared the same headworks, water source, fertilization rates, and agronomic management under a unified intelligent control system. Four types of drip tapes with different emitter flow rates (0.8、1.4、1.8 and 2.2 L/h) were used. Key indicators, including irrigation amount, frequency, uniformity, and yield, were monitored and analyzed for each plot. The results show that, the average irrigation quota are 195.0、225.0、270.0 and 330.0 m3/hm²，irrigation quotas are 3 985.5、3 955.5、3 955.5 and 4 015.5 m3/hm²，irrigation uniformities are 93.27%、92.86%、92.55% and 90.08%，yields are 6 645、7 455、7 380 and 7 155 kg/hm². Through comprehensive analysis, it is recommended that the optimal discharge rate of emitter with smart drip irrigation of cotton is 1.4 L/h, corresponding irrigation frequency is 17 times and irrigation cycle is 4 days.

In order to understand the response pattern of alfalfa (Medicago sativa) roots to different slopes and soil moisture differences in the Loess Region of North China, and to provide a scientific basis for the design of ecological slope protection projects and irrigation management, this study took alfalfa as the research object, with slope gradients (0°、20°、45°) and soil water content (15%、20%) as independent variables. The completely randomized design and root scanning technology were adopted to analyze the response patterns of root biomass, root morphology, root architecture, and fractal characteristics of alfalfa. The results showed that the soil water content and slope gradient of ecological slopes in the Loess Region of North China would have an interactive effect on the root biomass accumulation, morphological characteristics of alfalfa roots, as well as the topological structure and fractal characteristics of the roots. Under the experimental conditions, increasing the soil water content could promote the biomass accumulation of alfalfa roots, as well as the increase in the total root length, total surface area, total volume, and average root diameter. The average increases were 26.42%、17.14%、11.91%、57.60% and 11.77%, respectively. The effect of increasing the slope gradient varied significantly depending on the soil water content. Under the condition of low soil water content, all these indicators gradually decreased with the increase of the slope. Under the condition of high soil water content, the root biomass, root volume, and root diameter first increased and then decreased, while the total root length and total root surface area gradually increased with the increase of the slope. While the treatments did not alter the roots' fundamental dichotomous branching pattern (topological index≈0.5), they did affect other topological and fractal parameters. However, there were differences in the root topological parameters and fractal indices. The root branching structure was more complex under the condition of low soil water content, and the root space exploration ability was higher under the condition of high soil water content. The increase of the slope gradient would inhibit the root space exploration ability. For practical engineering applications on steep slopes, these findings suggest that moderately increasing moisture in the root zone—without compromising overall slope stability—can enhance the soil-reinforcing effect of the alfalfa root system.

This study aimed to identify the optimal types and concentrations of organic water-soluble fertilizers to improve the yield and quality of tomatoes grown in facilities in Shanxi Province. Using the 'strawberry tomato' cultivar as the test variety, a comparative experiment was conducted in a winter solar greenhouse. Four organic water-soluble fertilizers—a potassium fulvic acid fertilizer (T1), a humic acid fertilizer (T2), a chitin-seaweed extract fertilizer (T3), and a compound organic fertilizer (T4)—were applied via drip irrigation. A no-fertilizer treatment served as the control (CK). Each fertilizer type was tested at three concentrations: low (L), medium (M), and high (H). We then evaluated their effects on tomato growth, quality, and yield. The results showed that the T4-M treatment was the most effective. Compared to the CK treatment, T4-M increased plant height and stem diameter by 42.46% and 57.14%, respectively, at 120 days post-planting. At 60 days post-planting, it increased the leaf SPAD value and average leaf area by 10.56% and 102.07%, respectively. Furthermore, this treatment significantly improved fruit quality—increasing soluble sugar, sugar-acid ratio, Vitamin C, and lycopene content by 32.30%、141.15%、82.65% and 52.47%, respectively—while also boosting the total yield by 65.91%. Coupling coordination analysis further confirmed that the T4-M treatment was optimal for enhancing tomato quality. In conclusion, the T4-M fertigation scheme is recommended for facility-grown tomatoes in Shanxi to optimize growth, quality, and yield. This approach also holds potential for improving fertilizer and water use efficiency.

Hyperspectral technology helps in the non-contact and efficient inversion of soil moisture content in frozen soils. However, the spectral reflectance characteristics of soil in a frozen state have a significant impact. Therefore, adopting more effective spectral data processing and analysis methods is crucial for improving the accuracy of moisture content inversion in frozen soils. This study focuses on typical seasonal frozen soils in the black soil region of the Heshan Farm in Heilongjiang Province. Hyperspectral reflectance and soil moisture data were obtained, and the raw spectral reflectance was processed using fractional-order differentiation (FOD) with orders ranging from 0 to 1 (step size of 0.2) and continuous wavelet transform (CWT) at scales ranging from 4 to 1024 (increasing by powers of 4). After extracting the characteristic bands through Variable Importance in Projection (VIP) analysis, four algorithms—Partial Least Squares Regression (PLSR), Support Vector Machine (SVM), Random Forest (RF), and Extreme Learning Machine (ELM)—were used to construct soil moisture prediction models. The results showed that preprocessing methods with low-gradient fractional-order differentiation and continuous wavelet transform yielded better modeling performance, with the coefficient of determination exhibiting a unimodal distribution across different fractional orders and scales. The inversion accuracy varied significantly across different models, with the SVM model achieving the highest accuracy, followed by the ELM model, while the PLSR and RF models performed the worst. For frozen soils, the SVM model combined with CWT pre-processing at scales of 4 and 16 yielded the highest prediction accuracy (R2 _P >0.59; RMSE=0.291 and 0.299, respectively). In unfrozen conditions, model accuracy was higher than in frozen conditions, with the SVM model based on 0.2、0.4、0.6 and 0.8-order differentiation providing the best performance (R2 _P greater than 0.83, the Ratio of Prediction to Deviation values were greater than 3.6). The low-scale CWT-VIP-SVM model developed in this study offers technical support for hyperspectral monitoring of soil moisture in seasonal frozen black soil.

The planting structure of crops is an important reference basis for planning irrigation systems and estimating crop yields. Accurately identifying crops using remote sensing is a challenging task. In this study, time series images from May to September and individual monthly images were utilized to extract spectral reflectance and indices, and employed algorithms such as RF, SVM, and DT to conduct thorough classification of corn, soybeans, and rice. The classification accuracy differences between time-series images and single-month images were compared. The results showed that, compared to using the best single-month image, time-series imagery improved the classification accuracy by: ① 6.4% using the RF algorithm, ② 3.33% using the SVM algorithm, and ③ 6.25% using the DT algorithm. When using time-series imagery, the RF algorithm's accuracy was 2.92% higher than that of SVM and 6.25% higher than that of DT. In conclusion, this study demonstrates that using time-series imagery improves crop classification accuracy, and the RF model was the best-performing algorithm among those tested.

In response to heterogeneous water resource endowments and differentiated agricultural water-saving needs across Ningbo's districts, this study constructed a regional evaluation index system using the PSR framework. Through Moran's I spatial analysis, we identified evolving spatiotemporal patterns: from 2018 to 2022, water-saving agriculture demonstrated progressive improvement but formed a three-tier spatial hierarchy with significant disparities. Development exhibited spatial fragmentation, featuring isolated high-value zones and weak regional coordination. To address these imbalances, strategic priorities include establishing cross-regional technology-sharing platforms and collaborative governance mechanisms to enhance sustainable agricultural water management.