Management zoning is one of the most important procedures in precision agriculture. Among them, the fuzzy C-means (FCM) and K-means have been commonly used to divide the management zones. However, the conventional methods cannot treat the high-dimensional and non-linear data, due to the data scale and the attribute complexity. This study aims to introduce spectral clustering and deep clustering methods, compare them with FCM and K-means, and evaluate their performance in farmland management zoning, with the goal of providing more scientific decision-making support for precision agriculture. The research object was taken as the farmland at the park scale. According to the field data collection, the soil fertility index (SFI) was constructed to combine the comprehensive weight and multiple soil attributes. The spatial distribution of soil attributes was obtained to quantify the soil fertility level, which was input into the clustering model as a zoning indicator. The Kappa consistency test and analysis of variance were carried out to compare spectral clustering and deep clustering. A systematic evaluation was also made to explore their potential for the management of zoning in precision agriculture. The results showed that the soil fertility index better represented the spatial distribution of the soil fertility in the study area. The fertility level was relatively high in the central area, while the lower was observed in the northeast. The fertility level above the medium level (SFI≥0.5) accounted for 84.06% of the cultivated land, indicating the high soil fertility. The Kappa consistency test showed that there was the highest consistency between the deep clustering and the FCM (the Kappa coefficient was 0.767), followed by that between the spectral clustering and the deep clustering (the Kappa coefficient was 0.657). The consistency between the K-means and the rest was the lowest (the Kappa coefficient was only 0.21-0.23). The significance analysis shows that there were significant differences among the categories that were divided by the four clusterings when the number of clusters was 2 and 3. While there was no significant difference among some categories when the number of clusters was 4. The averaged silhouette coefficient (ASC) and the Calinski Harabasz index (CHI) showed that spectral clustering performed best when the number of clusters was 3. Its zoning was used to significantly distinguish the indicators, such as the soil nutrients, apparent electrical conductivity, vegetation index, and previous crop yield within different management zones. There was a decrease in the coefficient of variation of each variable within each zone. The spectral clustering effectively captured the spatial variability of the soil attributes and crop growth. There was consistency with the spatial distribution of the soil attributes after zoning. In conclusion, the spectral clustering was performed better than before in the spatial data zoning. The map of the management zoning was generated to more accurately reveal the spatial heterogeneity for the farmland management zoning. The finding can provide a more scientific and efficient zoning decision for precision agriculture.

A two-year field experiment was conducted to measure the effects of densification methods on photosynthesis and yield of densely planted wheat. Inter-plant and inter-row distances were used to define rate-fixed pattern (RR) and row-fixed pattern (RS) density treatments. Meanwhile, four nitrogen (N) rates (0, 144, 192, and 240 kg N ha⁻¹, termed N0, N144, N192, and N240) were applied with three densities (225, 292.5, and 360 × 10⁴ plants ha⁻¹, termed D225, D292.5, and D360). The wheat canopy was clipped into three equal vertical layers (top, middle, and bottom layers), and their chlorophyll density (ChD) and photosynthetically active radiation interception (FIPAR) were measured. Results showed that the response of ChD and FIPAR to N rate, density, and pattern varied with different layers. N rate, density, and pattern had significant interaction effects on ChD. The maximum values of whole-canopy ChD in the two seasons appeared in N240 combined with D292.5 and D360 under RR, respectively. Across two growing seasons, FIPAR values of RR were higher than those of RS by 29.37% for the top layer and 5.68% for the middle layer, while lower than those of RS by 20.62% for the bottom layer on average. With a low N supply (N0), grain yield was not significantly affected by density for both patterns. At N240, increasing density significantly increased yield under RR, but D360 of RS significantly decreased yield by 3.72% and 9.00% versus D225 in two seasons, respectively. With an appropriate and sufficient N application, RR increased the yield of densely planted wheat more than RS. Additionally, the maximum yield in two seasons appeared in the combination of D360 with N144 or N192 rather than of D225 with N240 under both patterns, suggesting that dense planting combined with an appropriate N-reduction application is feasible to increase photosynthesis capacity and yield.

The organ-specific critical nitrogen (N_c) dilution curves are widely thought to represent a new approach for crop nitrogen (N) nutrition diagnosis, N management, and crop modeling. The N_c dilution curve can be described by a power function (N_c = A₁·W^−A2), while parameters A₁ and A₂ control the starting point and slope. This study aimed to investigate the uncertainty and drivers of organ-specific curves under different conditions. By using hierarchical Bayesian theory, parameters A₁ and A₂ of the organ-specific N_c dilution curves for wheat were derived and evaluated under 14 different genotype × environment × management (G × E × M) N fertilizer experiments. Our results show that parameters A₁ and A₂ are highly correlated. Although the variation of parameter A₁ was less than that of A₂, the values of both parameters can change significantly in response to G × E × M. Nitrogen nutrition index (NNI) calculated using organ-specific N_c is in general consistent with NNI estimated with overall shoot N_c, indicating that a simple organ-specific N_c dilution curve may be used for wheat N diagnosis to assist N management. However, the significant differences in organ-specific N_c dilution curves across G × E × M conditions imply potential errors in N_c and crop N demand estimated using a general N_c dilution curve in crop models, highlighting a clear need for improvement in N_c calculations in such models. Our results provide new insights into how to improve modeling of crop nitrogen–biomass relations and N management practices under G × E × M.

Establishing the universal critical nitrogen (N_C) dilution curve can assist in crop N diagnosis at the regional scale. This study conducted 10-year N fertilizer experiments in Yangtze River Reaches to establish universal N_C dilution curves for Japonica rice based on simple data-mixing (SDM), random forest algorithm (RFA), and Bayesian hierarchical model (BHM), respectively. Results showed that parameters a and b were affected by the genetic and environmental conditions. Based on RFA, highly related factors of a (plant height, specific leaf area at tillering end, and maximum dry matter weight during vegetative growth period) and b (accumulated growing degree days at tillering end, stem–leaf ratio at tillering end, and maximum leaf area index during vegetative growth period) were successfully applied to establish the universal curve. In addition, representative values (most probable number [MPN]) were selected from posterior distributions obtained by the BHM approach to explore universal parameters a and b. The universal curves established by SDM, RFA, and BHM-MPN were verified to have a strong N diagnostic capacity (N nutrition index validation R² ≥ 0.81). In summary, compared with the SDM approach, RFA and BHM-MPN can greatly simplify the modeling process (e.g., defining N-limiting or non-N-limiting groups) while maintaining a good accuracy, which are more conducive to the application and promotion at the regional scale.