Giant ragweed (Ambrosia trifida, L. henceforth referred to as GR), an annual non‐native invasive weed, may cause health problems and can reduce agricultural productivity. Chemical control of GR in grasslands may have irreversible side effects on herbs and livestock. In an attempt to propose a solution to the harmful effects of GR on grasslands, this study explores the fate of its soil seed bank (SSB) and considers the physical control of its SSB reduction.By studying GR distributed in grasslands of the Yili Valley, Xinjiang, China, we measured the spatial and temporal changes in seed density, seed germination, dormancy, and death. We analyzed seed germination, dormancy, and death following different storage periods. The study analyzed population characteristics over time, including seed fate, and examined physical control methods for reducing the SSB density.The SSB of GR occurs in the upper 0–15 cm of soil in grasslands. Seed density in the SSB decreased by 68.1% to 82.01% from the reproductive growth period to the senescence period. More than 98.7% of the seeds were rotten, eaten, germinated, dispersed, or died within one year after being produced. The seed germination rate of the SSB decreased with the number of years after invasion. When stored for 0.5 or 3.5 years, seed germination rates fell by 40%, during which time seed death rate increased by almost 40%. When GR was completely eradicated for two consecutive years, the SSB and population densities decreased by >99%.The vast majority of GR seeds germinated or died within one year; the germination rate decreased significantly if the seeds were stored dry at room temperature for a long time. Newly produced seeds are the main source of seeds in the SSB. Therefore, thoroughly eradicating GR plants for several years before the seeds can mature provides an effective control method in grasslands.

Introduction: The harvest index (HI), a crucial agronomic trait that measures the ratio of grain yield to aboveground biomass, serves not only as a vital indicator for assessing wheat yield but also as a core parameter for predicting straw resource. It reflects the "source-sink" relationship and biomass allocation strategies in crops. However, the spatial distribution patterns of wheat HI and their key driving factors in arid regions remain unclear.Methods: This study was conducted in Xinjiang, a typical arid region of China, during 2022–2023, involving two years of large-scale systematic sampling. By integrating multidimensional factors such as geographical and climatic conditions, agronomic management practices, and soil nutrient status, methods including correlation analysis, random forest models, structural equation modeling, and linear regression analysis were employed to systematically investigate the spatial distribution characteristics and driving mechanisms of wheat HI under different irrigation regimes in arid regions.Results: The results revealed that: (1) Wheat HI in arid regions exhibited significant spatial heterogeneity (0.43–0.67), with an overall distribution pattern of "central high, peripheral low" and "northern high, southern low." (2) The importance rankings of influencing factors differed between irrigation regimes. For irrigated wheat, the order of importance was: Geographic-climatic factors, soil nutrient factors, agronomic management factors. Comprehensive analysis identified longitude (lon), plant height (H), latitude (lat), and bulk density (BD) as the key drivers of the Harvest Index (HI) in irrigated wheat. In contrast, for rainfed wheat, the order was: soil nutrient factors, Geographic-climatic factors, agronomic management factors, with total nitrogen (TN), available phosphorus(AP), total potassium(TK), and total phosphorus (TP) emerging as critical drivers of HI.Discussion: Irrigation significantly enhanced wheat HI (p < 0.01), and irrigated wheat demonstrated significantly higher HI, yield, and aboveground biomass (AGB) compared to rainfed wheat (p < 0.01). Optimizing phosphorus management could enhance HI in both systems, while irrigation infrastructure development remains vital for yield stability. This study provides a theoretical basis and practical guidance for the synergistic multi-objective approach of “yield increase-irrigation-sustainability” in arid regions wheat production.

Plant functional traits are indicative of the long-term responses and adaptations of plants to their environment. However, the specific mechanisms by which desert plant functional groups (PFGs) adjust their ecological adaptation strategies to cope with harsh environments remain unclear, particularly in ecologically fragile farming–pastoral zones. To address this gap, this study investigates and analyzes the morphological and chemical characteristics of 13 desert plant species in the farming–pastoral zone of the northern Tarim Basin. Through cluster analysis, these desert plants were categorized into distinct PFGs to elucidate their ecological response strategies at a higher organizational level. The results were as follows: (1) Based on plant functional traits, the 13 desert plant species were classified into acquisitive, medium, and conservative PFGs. These groups exhibited significant differences in chemical element content and proportion, as well as morphological adjustments (p < 0.05). (2) The acquisitive functional group maintained high resource acquisition and turnover through high specific leaf area and leaf phosphorus content; the medium functional group occupied limited resources through greater plant height and canopy width, whereas the conservative functional group exhibited low growth rates but high morphological investment to ensure survival. Moreover, these differences in ecological adaptation strategies led to the selection of divergent central traits by different PFGs. (3) Low soil nutrient availability and soil salinization, rather than groundwater depth, were identified as the primary environmental factors driving the differentiation of PFGs in the farming–pastoral zone. These findings suggest that desert plants in arid regions employ diverse ecological adaptation strategies to cope with environmental pressures. This research study provides valuable insights and recommendations for the conservation and restoration of desert plant communities.

This study examines the impact of continuous cropping (short-term: 1–8 years; medium-term: 9–15 years; long-term: 16–30 years) on soil microbial community diversity, co-occurrence networks, and assembly processes in Xinjiang’s cotton region, a globally recognized arid zone. The results are as follows. Soil physicochemical analyses showed that as continuous cropping duration increased, soil organic matter and total nitrogen significantly decreased, whereas available phosphorus and potassium increased, and the soil’s aggregate structure degraded. Microbial community analysis indicated that long-term continuous cropping notably increased the richness of bacterial species (Chao1 index) and altered fungal communities’ diversity and composition, especially increasing the relative abundance of Cladosporium and Alternaria in the long term (GY30). Co-occurrence network analysis revealed higher complexity in bacterial and fungal networks in the short term. As cropping duration increased, bacterial network complexity significantly decreased, while fungal networks partially recovered in the long term, indicating greater fungal adaptability to environmental changes. Assembly process analysis revealed that the assembly of bacterial and fungal communities was jointly regulated by stochastic and deterministic processes, but with increasing cropping duration, deterministic processes weakened while stochastic processes intensified. Soil available phosphorus, potassium, and pH were identified as key factors influencing microbial community succession and assembly. This study highlights the significance of co-occurrence networks and assembly processes for understanding the dynamics of continuous cropping’s impact on soil microbial communities, offering a theoretical foundation for improving agricultural management.

Water scarcity in arid regions poses a severe threat to agricultural sustainability, necessitating optimized irrigation strategies. This study investigates the cumulative impacts of long-term irrigation deficits on soil quality, microbial diversity, and maize yield in the arid maize fields of Xinjiang, China, where consistent irrigation patterns have been maintained over multiple years. Seven sites were monitored from April 2023 to March 2024, with a single end-of-cycle sampling in March 2024. Using the Irrigation Water Deficit Index (IWDI), the sites were classified into low (LD, 16.37–22.30%), moderate (MD, 30.54–38.10%), and high drought (HD, 47.49–50.00%) categories. The findings reveal that long-term consistent irrigation deficits exacerbate soil salinization, compaction, and nutrient loss, with organic matter declining significantly under HD conditions. Bacterial richness increased by ~6% under HD, driven by stress-tolerant taxa, while fungal diversity decreased by 14–50%, impairing nutrient cycling functions critical for soil health. The Soil Quality Index (SQI) and maize yield declined with drought severity (LD > MD by 26.18% and 21.05%; LD > HD by 45.02% and 13.13%), highlighting the pivotal role of sustained irrigation patterns in maintaining productivity. These results underscore the need for tailored irrigation management in arid regions, such as precision drip irrigation, to mitigate soil degradation and sustain maize yields, providing a scientific foundation for optimizing water use efficiency in water-scarce agroecosystems under long-term irrigation regimes.

The development of smart agriculture has created an urgent demand for efficient and accurate weed recognition and detection technologies. However, the diverse and complex morphology of weeds, coupled with the scarcity of labeled data in agricultural scenarios, poses significant challenges to traditional supervised learning methods. To address these issues, a weed detection model based on a semi-supervised diffusion generative network is proposed. This model integrates a generative attention mechanism and semi-diffusion loss to enable the efficient utilization of both labeled and unlabeled data. Experimental results demonstrate that the proposed method outperforms existing approaches across multiple evaluation metrics, achieving a precision of 0.94, recall of 0.90, accuracy of 0.92, and mAP@50 and mAP@75 of 0.92 and 0.91, respectively. Compared to traditional methods such as DETR, precision and recall are improved by approximately 10% and 8%, respectively. Additionally, compared to the enhanced YOLOv10, mAP@50 and mAP@75 are increased by 1% and 2%, respectively. The proposed semi-supervised diffusion weed detection model provides an efficient and reliable solution for weed recognition and introduces new research perspectives for the application of semi-supervised learning in smart agriculture. This framework establishes both theoretical and practical foundations for addressing complex target detection challenges in the agricultural domain.