Agronomy Journal最新文献

Under Mediterranean irrigated conditions cover cropping (CC) and double cropping (DC) are diversification/intensification strategies that can increase grain yield and resource use efficiency of the traditional winter fallow-maize (Zea mays L.) system. Four cropping systems were evaluated in terms of productivity, water and nitrogen use efficiency (WUE and NUE) under sprinkler-irrigated conditions during three growing seasons in the Ebro Valley, Spain: (1) long-season maize with winter fallow (F-LSM), (2) long-season maize after a leguminous cover crop (common vetch, Vicia sativa L.) (CC-LSM), (3) short-season maize after a cereal crop (barley, Hordeum vulgare L.) (B-SSM), (4) short-season maize after a leguminous crop (winter peas, Pisum sativum L.) (P-SSM). The introduction of the vetch winter cover crop required an additional 5% irrigation water but allowed to reduce the nitrogen fertilizer applied by 20%, increasing the system grain yield synthetic nitrogen use efficiency (NUEsynt-g) by 27% without affecting the cropping system grain yield and WUE. DC systems required 12% more irrigation water than the traditional F-LSM but produced more grain. The B-SSM was the most productive system (21.9 Mg grain ha⁻¹) and increased the WUE by 32% compared to the F-LSM system, but required a 39% more nitrogen fertilizer. Compared to the traditional F-LSM system, the P-SSM cropping system increased the grain yield (+16%), protein yield (+66%), NUEsynt-g (+20%), and the WUE (+10%). The diversification and intensification of the traditional F-LSM system increased yield (with the DC systems) and resource use efficiency (WUE with the DC systems; NUE with CC-LSM and P-SSM cropping systems).

McNally, B. C., Elmore, M. T., Kowalewski, A. R., Braithwaite, E. T., & Cain, A. B. (2025). Irrigation frequency and mowing height influence annual bluegrass in perennial ryegrass. Agronomy Journal, 117, e70232. https://doi.org/10.1002/agj2.70232

North Brunswick, NJ was mistakenly placed in quotes throughout the article. It has now been corrected in the following places: in the second sentence of the abstract; in the caption of Table 1, and in the second sentence under Table 1's first footnote; in the fourth sentence under Section 3.1; in the captions of Tables 2, 3, 4, and 5; in the last sentence of the second paragraph under Section 3.3; and in the last sentence of first paragraph under Section 3.4.

Rutgers University Horticulture Farm No. 2 in North Brunswick, NJ was mistakenly placed in quotes in both the third sentence in Section 2.1 and the first sentence in Section 3.1.

In addition, in the last sentence in the fifth paragraph in the Introduction, in the fifth sentence under Section 3.1, and in the second to last sentence in the second paragraph in Section 3.4, New Jersey should not have been in quotes.

We apologize for these errors.

Soybean (Glycine max L. Moench) yield is influenced by fluctuations in weather throughout the growing season and across the planting dates. Therefore, for growers, predicting soybean yield early in the season and addressing yield variability is essential for strategic decisions and resource utilization. The objective of our study was to capture soybean yield variability under three different planting dates (early, mid, and late), for two seeding rates (low, ∼247,100 and high, ∼370,650 seeds ha⁻¹) and two maturity groups (MGs 3 and 4), using machine learning models to predict soybean yield. Agronomic and meteorological data from the Kansas Mesonet and high-resolution (3 m) PlanetScope satellite imagery were used to predict and address soybean yield variability. Results showed that early-planted soybeans have demonstrated higher mean yield potential with a higher coefficient of variation than mid- and late-planted soybeans. Therefore, to quantify and model this variability, four models, including Random Forest (RF), Adaptive Boosting (AdaBoost), K-Nearest Neighbor, and Least Absolute Shrinkage and Selection Operator, were evaluated. The RF and AdaBoost algorithms performed comparatively better (R²: 0.79–0.80; root mean square error: 0.38–0.39 Mg ha⁻¹; mean absolute error: 0.31 Mg ha⁻¹; mean squared error: 0.14–0.15 Mg ha⁻¹; mean absolute percentage error: 0.08%). Moreover, we have observed that the accuracy percentage (10% error threshold) and R² were relatively higher as the crop matured, with the highest during the late vegetative and reproductive stages. This highlights the importance of in-season monitoring of the resources and market planning.

Nearly 1 billion ha of soils affected by salinization have been identified worldwide (8.7% of the planet's soils). These soils are mainly found in naturally arid or semi-arid environments. The map also shows that 20%–50% of irrigated soils across all continents are too saline. Thus, soil salinity is one of the most critical threats to food security. It adversely affects the growth and productivity of agricultural crops. Tomato is the most important horticultural plant and an essential annual crop for human food worldwide. The effects of salinity on tomato (Solanum lycopersicum L.) plants have been studied in recent years by several researchers. Attempts to improve tomato salinity tolerance through conventional breeding programs have had limited success due to the complexity of the trait. Thus, various cultural techniques, in addition to varietal selection, are applied to mitigate the harmful effects of salinity, such as seed pretreatments through priming methods, chemical fertilizers, and organic amendments like the use of beneficial soil microorganisms, including plant growth-promoting rhizobacteria and arbuscular mycorrhizal fungi. This review paper provided valuable information on the behavior of tomato cultivars under saline conditions. The review also provides a synthetic overview of current and relevant scientific advances allowing the improvement of salinity tolerance of tomato plants. However, natural seed or soil treatments to combat salinization have not been widely developed. Nevertheless, the strategies developed in this review, combined with recent advances in emerging biotechnological solutions, could allow mitigating the effects of salinity on tomato plants.

While vegetation indices (VIs)-based machine learning (ML) techniques have been developed for predicting crop yield, limited research has focused on how VI selection impacts ML model predictions or on identifying optimal VI combinations. In this study, three ML models, including Distributed Random Forest (DRF), Gradient Boosting Machine (GBM), and Deep Neural Network (DNN), were established to predict rice (Oryza sativa L.) yield using eight VIs: difference vegetation index (DVI), land surface wetness index, normalized difference vegetation index (NDVI), normalized difference (red − blue)/(red + blue) vegetation index, ratio vegetation index (RVI), soil adjusted vegetation index (SAVI), transformed vegetation index (TVI), and Keetch–Byram drought index (KBDI), extracted at five key growth stages: re-greening, tillering, stem elongation, preliminary heading, and full heading. The feature attribution method was used to quantify the relative contributions of input variables to yield predictions. The results are as follows: (1) The three ML models produce accurate rice yield predictions using DVI, NDVI, RVI, and SAVI, with root mean square error (RMSE) ranging from 174.80 to 291.83 kg/ha, R² from 0.56 to 0.84, and Nash Sutcliffe efficiency (NSE) from 0.56 to 0.84. But three models produce poor predictions with KBDI, with RMSE ranging from 344.01 to 404.73 kg/ha, R² from 0.31 to 0.44, and NSE from 0.14 to 0.38. (2) The DNN model performs better than the GBM and DRF models for rice yield prediction. (3) Note that 80% of the most important input variables are associated with the rice preliminary heading stage for the DNN models, whose importance values ranged from 0.65 to 1.00, and the average TVI at this growth stage is the most important variable. Therefore, the DNN technique, when integrated with VIs from the preliminary heading stage, is recommended for rice yield prediction.

This study presents a comprehensive review of literature focused on technological interventions that enhance traditional agricultural farming and rural practices. It aims to highlight how modern technologies can be integrated with traditional farming methods to build more efficient, climate-resilient, and context-appropriate agricultural practices. The review is divided into two primary sources: (a) peer-reviewed research articles and (b) granted patents related to relevant technologies. A systematic keyword-based search was conducted using terms such as “Traditional farming practices” and “Agri-rural processes” across high-ranking academic directories. The selection of journals was based on their reputational ranking and subject relevance. A similar strategy was applied to the Derwent and XL Scout patent databases to track innovation trends supporting traditional agricultural systems. Each selected paper and patent was examined in detail to assess its significance, application scope, and contribution to sustainable rural development. The analysis identifies a range of technological innovations that can complement and enhance traditional agricultural farming practices. These interventions provide opportunities for improved efficiency, resilience, and sustainability, particularly in rural terrace farming. By analyzing both scientific literature and patented innovations, this work offers actionable insights for policymakers, researchers, and practitioners. It highlights how innovation can be balanced with tradition in rural transformation strategies. This study uniquely combines a comparative review of scientific research and patent analysis to explore the coexistence of tradition and technology. It contributes to the understanding of how innovation can be tailored to local practices and cultural heritage to foster inclusive and resilient agricultural development.

On-farm research is important for estimating the performance of crop management practices under real-world conditions, offering localized insights that drive adoption. However, conventional research trial designs such as the randomized complete block design often fail to capture spatial variability and can be complex to implement on commercial farms. To address these limitations, the single-strip spatial evaluation approach (SSEA) was developed, allowing farmers to test treatments using a single-strip design while leveraging spatial yield data collected by harvester-mounted sensors for corn (Zea mays L.) grain and silage. In this approach, yield stability zones, generated from multi-year interpolated yield data, enable evaluation of treatment effects across different zones within the field. While the single-strip design may introduce spatial bias, this can be mitigated by replicating treatments across multiple fields. To improve accessibility, a web-based tool was developed that automates the analysis, generates confidence charts, and produces downloadable reports for farmer use. We describe the process and resources used for building the tool and present its functionality through a real-world single-strip case study. Developed with input from a statewide advisory committee, the tool includes zone distribution donut plots and a color-coded confidence chart with interpretations of spatial responses. By streamlining spatial data analysis and reporting, the SSEA web tool empowers farmers, farm advisors, and crop consultants to independently conduct on-farm trials, interpret treatment effects by zone, and make informed management decisions. The SSEA web tool represents a significant step toward spatially informed on-farm research and supports broader adoption of data-driven, site-specific agricultural practices.

This study explores how scientists can support on-farm experiments using analytical methods that align with farmers’ endogenous learning processes to inform their management decision. Four maize (Zea mays L.) farmers across 10 site-years in New York participated in this study to evaluate the effectiveness of a nitrogen-fixing inoculant (NFI) applied with a reduced side-dress nitrogen rate. Farmers designed and implemented their own experiments using a range of layouts, including side-by-side comparisons and strip trials. Two analytical approaches were compared: a quantitative yield analysis using spatial regression, and a causal pathway analysis based on mechanistic steps informed by field sampling (e.g., quantitative polymerase chain reaction detection of NFI organisms, nitrogen nutrition index, and yield). While yield data suggested positive or neutral treatment effects at all sites when simply comparing yield average, the spatial regression analysis and causal pathway analysis identified positive outcomes in only seven or four of 10 site-years, respectively, reflecting a more conservative interpretation of efficacy. Both methods provided consistent conclusions at four out of 10 site-years, demonstrating the contribution of metrics other than yield in the interpretation process. Findings suggest that simple causal diagrams can structure data collection and interpretation in ways aligned with farmers' goals. Supporting farmer experiments with digital agronomy, mechanistic reasoning, and site-specific data enhances learning outcomes and scientific rigor without requiring formal replication. This work contributes to the development of collaborative, scalable methodologies that integrate farmer knowledge and scientific analysis in on-farm experimentation.

Farmers often conduct unreplicated on-farm experiments (OFE) to evaluate management practices such as the application of plant growth regulators (PGR) in winter wheat (Triticum aestivum L.). Traditional methods of comparing strip average yields, such as using weigh wagons or yield monitors, lack error estimates and are causally confounded by field variability. Prescription (Rx) maps with randomization and replication may reduce causal confounding but are not always feasible. We propose a methodology to improve causal inference from unreplicated strip trials using propensity score matching (PSM). PGR strip trials were implemented using growers’ fields and equipment at two sites. Yield data, topographic covariates, and soil properties were collected. Propensity scores were calculated and used to create weights for covariate balancing. Next, treatment effect estimates and 95% confidence intervals were calculated for each site using G-computation. Various benchmark models were included to compare the results of commonly implemented spatial models to the results from PSM. Spatial benchmark models showed evidence of spatial confounding, a purely statistical artifact rather than a causal effect. This artifact may alter treatment estimates and test statistics in strip trials where experimental units are not randomized throughout the field. PSM has potential to address the lack of replication and randomization in simple two-treatment strip trials. PSM can potentially increase accessibility to rigorous OFE and improve decision-making in agricultural practices, particularly in contexts where traditional experimental designs present barriers to participation.

Understanding the long-term impacts of crop rotation systems on rice (Oryza sativa L.) yield and stability is key to redesigning agroecosystems, optimizing management, and refining sustainable intensification strategies. This study evaluated the impacts of the rotation system and the previous crops on irrigated rice yield and its stability over 9 years, using an RCB design experiment in Uruguay. Rotations were (1) Rice1-Rice2-Perennial Pasture (R-PP); (2) Rice-Biannual Pasture (R-BP); (3) Rice1-Soybean1-Soybean2-Rice2-Perennial Pasture (R-Sy-PP); (4) Rice1-Soybean-Rice2-Sorghum (R-Crops); (5) Rice-Soybean (R-Sy); and (6) continuous rice (CR), all with winter cover crops between grain crops. The highest yields were obtained in rotations including soybean (R-Sy, R-Sy-PP, R-Crops: 11.03 Mg ha⁻¹), which were 7% and 15% higher than those including only pastures (R-BP and R-PP) and CR, respectively. However, the highest effect on yield and yield stability was observed by previous crops. Independently of rotation, rice following soybean had the greatest productivity (11.33 Mg ha⁻¹), followed by rice after pastures (10.60 Mg ha⁻¹), and rice after rice (9.46 Mg ha⁻¹). These differences were amplified in high-yielding years, with rice after soybean (12.72 Mg ha⁻¹) yielding 5%, 17%, and 22% more than after perennial pastures, biannual pastures, and rice, respectively. Soybean as a previous crop increased rice yield in all rotations but decreased yield stability as demonstrated by an environmental index combining four parameters. For rice-pasture systems in temperate climates, rotation intensification integrating soybean offers a viable strategy for increasing rice productivity, particularly in high-yielding years, despite lower yield stability.