Climate‐adapted variety selection and optimal planting density for improving maize yield in the Guanzhong region, China

Abstract

BACKGROUND

Suboptimal densities and regional environmental heterogeneity often constrain maize yield intensification. Increasing planting density and selecting climate‐adapted varieties are therefore critical for improving grain yield.

RESULTS

The study was conducted at 33 sites in the Guanzhong region from 2017 to 2021, addressing uncertainty in density–variety selection for local maize production. Six varieties (including widely grown and newly introduced types) were tested at four planting densities (45 000–90 000 plants ha⁻¹), and their combined effects on yield and agronomic traits were assessed. Results showed that increasing planting density improved grain yield and economic returns. The experimental results identified 75 000–90 000 plants ha⁻¹ as the optimal density range for the region; within this range, yields increased by 6.24–27.31% compared with traditional lower densities (45 000–60 000 plants ha⁻¹). All six varieties responded positively to increased planting density, with Shandan 650 and Shandan 8806 showing the greatest yield gains. Specifically, adopting a planting density of 75 000 plants ha⁻¹ with varieties such as Shandan 650 could maximize economic returns while minimizing seed‐cost risks to local farmers in the Guanzhong region. Grain yield was closely associated with climatic conditions during the growing season, with higher precipitation in eastern Guanzhong contributing to superior yields. Furthermore, optimizing 100‐kernel weight and the ear‐to‐plant height ratio may help stabilize yields in medium‐ and low‐yielding fields.

CONCLUSION

Ultimately, these findings provide a practical yield‐enhancement strategy, offering useful references for transforming low‐productivity fields into high‐yield systems in similar monsoon‐influenced rotation zones worldwide. © 2026 Society of Chemical Industry.

INTRODUCTION

Global food demand is projected to rise sharply as population growth and economic development continue. ¹ By 2050, the global requirement for food crop production is expected to reach 5.499 billion tons, representing an increase of 46.8% compared to 2020. ² Among staple crops, maize (Zea mays L.), China's most important cereal crop, is experiencing rapidly increasing demand, particularly for animal feed and industrial processing. Ensuring a stable maize supply is therefore vital for national food security.^{3, 4} However, maize production in the Guanzhong region of Shaanxi Province, a key maize‐growing area in northwest China, faces multiple challenges.^{5, 6, 7}

By the end of the 21st century, global mean surface temperatures are projected to increase by 0.3–4.8 °C, ⁸ while extreme rainfall events have become increasingly frequent in recent years.^{9, 10} Climate warming has destabilized precipitation patterns, increasing uncertainty in food production.^{11, 12} In China, future climate scenarios are expected to reduce maize yields significantly.^{13, 14} Specifically, temperatures in the Guanzhong region are projected to rise by 2.3–5.3 °C between 2070 and 2099 relative to the 1950–1999 baseline. ¹⁵ Such long‐term warming may adversely affect maize growth and development.^{16, 17, 18} Meanwhile, average precipitation in north‐western China is predicted to increase by about 10% between 2046 and 2065 relative to 1986–2005 levels. ¹⁹ Although moderate short‐term rainfall may benefit maize growth, sustained increases in precipitation may cause soil waterlogging, reduced fertility, and more frequent extreme weather events, ultimately suppressing grain yield. ¹²

Addressing climate‐related risks through adaptive management strategies and natural adaptation has become a global priority.^{20, 21, 22} Agriculture is particularly vulnerable to changes in temperature and rainfall,^{23, 24} which can shift crop phenology and disrupt farming practices.^{25, 26, 27} Rational agronomic management, including appropriate variety selection and optimal planting density, is essential for matching crop performance to local environmental conditions and maximizing varietal potential. Aligning varietal selection with regional temperature and rainfall patterns is especially important for minimizing climate‐related yield losses.^{28, 29} Integrated crop management strategies are increasingly recognized as effective approaches for mitigating the adverse effects of climate change on agricultural productivity. ²⁷

China's maize yield per hectare remains only about 60% of that achieved in the United States. In the Guanzhong region, maize yields remain particularly low because planting densities are substantially lower than the US average.^{30, 31, 32} Optimizing planting density can improve yield stability and resilience, particularly under prolonged extreme rainfall conditions.^{33, 34} Increasing planting density can also narrow the gap between actual and potential yields.^{35, 36, 37} In addition, varieties with compact plant architecture, extended growth duration, and strong local adaptability are more favorable for achieving higher yields.^{38, 39} Overall, integrating appropriate planting density, suitable variety selection, and climate‐adapted management practices may further improve maize productivity.

This study analyzed data collected from 33 field sites across the Guanzhong region between 2017 and 2021, incorporating four planting densities and six maize varieties. The objectives were to: (i) provide a theoretical basis for selecting maize varieties and planting densities suitable for the Guanzhong region based on agronomic traits and yield responses; (ii) explore yield variation in relation to effective accumulated temperature and total rainfall during the growing season; (iii) identify the key factors influencing maize yield to support targeted yield improvement in low‐performing fields. The findings provide a scientific basis for optimizing maize cultivation in the Guanzhong region and may also serve as a reference for broader efforts to improve maize productivity under changing climatic conditions.

MATERIALS AND METHODS

Study site

The Guanzhong region, located in middle Shaanxi Province, China (33°41′ N–35°39′ N, 106°42′ E–110°35′ E), is a major agricultural area in northwest China. The region includes five cities – Baoji, Tongchuan, Xianyang, Xi'an, and Weinan – covering a total area of 56 000 km², which accounts for 19% of Shaanxi Province. Elevation ranges from 400 to 700 m above sea level. The region has a continental monsoon climate, with precipitation concentrated mainly from June to September. Spring is typically dry, whereas winter occasionally experiences short‐duration, high‐intensity rainfall events. The dominant cropping system is a winter wheat–summer maize rotation. Winter wheat is sown in October and harvested the following June, while summer maize is planted immediately after wheat harvest and harvested by early October.

The field experiments were conducted from 2017 to 2021 across 33 representative sites in the Guanzhong region, China. To optimize the spatial allocation of agricultural production for the efficient utilization of light, thermal, and water resources, the Guanzhong region was delineated into three sub‐zones (western, middle, and eastern) based on longitude and local phenological conditions (Fig. 1). ⁴⁰ The precise quantitative longitudinal ranges for the western, middle, and eastern zones were 106.52°–107.90° E, 108.08°–109.32° E, and 109.18°–110.15° E, respectively. Within these sub‐regions, the climate parameters specifically across the maize growing season at the respective experimental sites demonstrated distinct hydro‐thermal gradients: the middle zone received the highest growing‐season rainfall (470.88 mm), followed by the eastern (450.41 mm) and western (445.45 mm) zones, whereas the eastern zone exhibited the highest effective accumulated temperature (1722.24 °C), followed by the middle (1633.98 °C) and western (1441.68 °C) zones (Supporting Information, Fig. S2).

Figure 1.

Location of the Guanzhong region in Shannxi Province, China. Based on topography and climate conditions, we divided the Guanzhong region into three parts: the western part (Baoji), the middle part (Tongchuan, Xianyang, Xi'an), and the eastern part (Weinan). The red triangles on the map on the right represent the 33 experiment sites.

Field experiments

Field experiments were conducted at 33 sites across the Guanzhong region from 2017 to 2021 (Fig. 1). The average annual temperature was 13.40 °C, with average annual precipitation of 572.19 mm, of which approximately 75.97% fell during the maize growing season. The study used six maize varieties widely cultivated in the region: Shandan 650, Zhengdan 958, Shandan 636, Shandan 8806, Shandan 609, and Dongdan 60. Zhengdan 958 was used as the reference variety due to its widespread cultivation, stable performance, and consistently high yields. Other varieties were included to evaluate their performance under various planting densities.

Low planting density, often below 60 000 plants ha⁻¹, is considered a major constraint to maize productivity in the Guanzhong region. Therefore, four planting densities were evaluated: 45 000, 60 000, 75 000, and 90 000 plants ha⁻¹. Sowing dates ranged from 7 June to 24 June, and harvest dates ranged from 16 September to 23 October. The average growth duration ranged from 93 to 153 days. Agronomic management practices, including fertilization, weed and pest control, and irrigation, adhered to standard local practices and were standardized across all density treatments within each site to ensure consistency. Basal fertilization included urea (46% nitrogen), calcium superphosphate (12% P₂O₅), and potassium sulfate (50% K₂O), applied at rates of 225, 120, and 90 kg ha⁻¹, respectively, before sowing. Irrigation was applied once at the large bell (12‐leaf) stage using a sprinkler irrigation system, with a water application volume of 60–70 m³ ha⁻¹. Weeds were controlled after sowing using a pre‐emergent herbicide (96% S‐metolachlor) applied at 1–1.25 L ha⁻¹. Pests, primarily corn borers, were controlled from the large bell stage to silking using 5% chlorantraniliprole applied at 0.3–0.45 L ha⁻¹ or equivalent locally recommended pesticides. All management practices followed the high‐yield cultivation standards of the Guanzhong region.

Climate data

Meteorological data from 2017 to 2021 were obtained from the ECMWF ERA5 dataset through the Xiaozuanfeng platform (https://meteo.agrodigits.com/home/index). The dataset included daily average temperature and cumulative precipitation data.

Data analysis

Effective accumulated temperature

Effective accumulated temperature (°C) during the maize growing season was calculated as:

$$T_{\text{total}}=\sum \left(T_i-T_{\text{base}}\right)$$ (1)

where T _total is the cumulative effective temperature from sowing to day i, T _i is the average daily temperature, and T _base is the biological base temperature for maize growth, set at 10 °C.

Total precipitation

Cumulative precipitation during the maize growing season (in millimeters) was calculated as:

$$R_{\text{total}}=R_1+R_2+R_3+\dots +R_i$$ (2)

where R _i is the rainfall on day i, and i ranges from 1 to the total number of days from sowing to harvest.

Plant height and lodging rate

Five representative plants were selected from each plot at the grain‐filling stage to measure plant height and ear height. At harvest, lodged plants in the two central rows of each plot were counted, and lodging rate was calculated as the proportion of lodged plants relative to the total number of plants in those rows.

Yield and yield components

At physiological maturity, 15 uniformly growing maize plants were sampled from each treatment plot. Five ears per replicate were used to determine row number per ear and kernel number per row. Three subsamples from each replicate were used to measure fresh 100‐kernel weight. Samples were subsequently oven‐dried at 75 °C to constant weight to determine the dry 100‐kernel weight. Grain moisture content was measured using a portable moisture meter (PM8188; Kett Electric Laboratory, Tokyo, Japan). Total dry weight was calculated based on total fresh weight and grain moisture content, and grain yield was standardized to 14% moisture content.

Density‐driven yield gain

To evaluate the effect of increasing planting density on grain yield, the following equation was used: ⁴¹

$$\ln \left(R\right)=\ln \frac{X_{\mathrm{H}}}{X_{\mathrm{L}}}$$ (3)

where X _H is the average yield under higher planting densities (60 000–90 000 plants ha⁻¹), and X _L is the average yield under the lowest planting density (45 000 plants ha⁻¹). Percentage change was calculated as follows:

$$D=\left(R-1\right)\times 100$$ (4)

where positive values of D indicate yield increases under higher planting densities, negative values indicate decreases, and D = 0 indicates no change.

Economic benefit from seed input and output

Seed quantity (SQ) for each planting density was calculated as: ⁴²

$$\mathrm{SQ}=W/\left(e\times \mathrm{GP}\right)$$ (5)

where W is the average seed weight (in grams), e is the establishment rate (typically 95%), and GP is seed germination percentage.

Since machinery, fertilizer, pesticide, and irrigation costs remained constant across densities, seed input cost was the only variable cost considered. Net economic return (NE, in CNY kg⁻¹ ha⁻¹) was calculated as:

$$\mathrm{NE}=Y\times \mathrm{MP}-\mathrm{SQ}\times \mathrm{SP}-\mathrm{AI}\times \mathrm{CS}$$ (6)

where Y is grain yield (in kg ha⁻¹), MP is maize grain price (in CNY kg⁻¹), and SP is seed price (in CNY kg⁻¹). AI represents average agricultural inputs, including machinery, fertilizer, and herbicide costs, whereas CS represents the average economic cost of these inputs. NE reflects the return per unit of seed investment.

Statistical analysis

All statistical analyses were performed in R version 4.1.0. Multifactorial analysis of variance (ANOVA) was conducted using the ‘car’ and ‘agricolae’ packages to assess the effects of multiple independent variables. Correlation analyses were performed with the ‘Hmisc’ and ‘corrplot’ packages to evaluate and visualize relationships. Linear regression analyses were conducted using the ‘ggpmisc’ package to display regression equations and associated statistics.

The ‘rfPermute’ package was used to quantify the relative importance of climatic variables, agronomic traits, and planting density, with significance determined through permutation testing. Path analysis was conducted using the ‘semPlot’ package to visualize direct and indirect effects among variables. Figures were generated using the ‘ggplot2’ package.

RESULTS

Effects of planting density and variety on maize yield and agronomic traits

Planting density, variety, and year significantly influenced maize yield, ear row number, kernel number per row, and 100‐kernel weight. Planting density did not significantly affect plant height or ear height, whereas both traits were significantly affected by variety and year. The interaction between variety and year significantly affected plant height, ear row number, and kernel number per row, but not ear height, yield, or 100‐kernel weight. No significant three‐way interaction among planting density, variety, and year was observed (Table 1).

Table 1.

Multivariate analysis of variance for maize yield and agronomic traits

	Plant height (cm)	Ear height (cm)	Yield (kg ha−1)	Rows per ear	Kernels per row	100‐Kernel weight (g)
Dongdan 60	279.90a	118.18a	8915.58a	17.37a	33.78a	26.10d
Shandan 609	254.20b	105.32b	9010.41a	16.20bc	31.26c	30.54b
Shandan 636	231.52d	78.02d	8055.65b	15.04d	31.86bc	29.57bc
Shandan 650	234.90d	80.87d	8838.77ab	16.25b	33.25a	26.97d
Shandan 8806	243.46c	85.78c	8813.58ab	15.93c	32.57abc	28.65c
Zhengdan 958	243.25c	101.47b	9361.03a	14.52e	33.08ab	32.35a
P Value
Density	NS	NS	**	**	**	**
Year	**	**	**	**	**	**
Variety	**	**	**	**	**	**
Density*Year	NS	NS	NS	NS	NS	NS
Density*Variety	NS	NS	NS	NS	NS	NS
Year*Variety	*	NS	NS	**	**	NS
Density*Year*Variety	NS	NS	NS	NS	NS	NS

Note: **P < 0.01; *P < 0.05; NS: not significant. Values with different lowercase letters within a variety differ significantly at the 5% level.

The average maize yield across all sites from 2017 to 2021 was 8373.75 kg ha⁻¹. The lowest planting density (45 000 plants ha⁻¹) exhibited the greatest yield variability. At the treatment level, 75 000 and 90 000 plants ha⁻¹ did not differ significantly from each other but produced significantly higher yields than 60 000 and 45 000 plants ha⁻¹ (Fig. 2(b)). Specifically, yield at 75 000 plants ha⁻¹ increased by 6.24% and 24.30% compared with 60 000 and 45 000 kg ha⁻¹, respectively. At 90 000 plants ha⁻¹, yield increased by 8.82% and 27.31% relative to the same two lower densities (Fig. 2). Moreover, increased seed input at higher densities did not reduce economic returns (Fig. S1), with the highest net returns observed at 90 000 plants ha⁻¹.

Figure 2.

Maize yield distribution (a) and box plot (b) under different planting densities (45 000 plants ha⁻¹, 60 000 plants ha⁻¹, 75 000 plants ha⁻¹, and 90 000 plants ha⁻¹) in Guanzhong region. Values with different lowercase letters within a variety differ significantly at the 5% level.

All varieties showed increased yield with increasing planting density, indicating a generally positive response to intensification. However, significant differences in absolute yield were observed among genotypes. Zhengdan 958, Shandan 609, and Dongdan 60 consistently produced higher mean yields than Shandan 636 across the density gradient (Fig. 3(a)). These differences reflect distinct yield component configurations. Zhengdan 958 achieved higher yield primarily due to a significantly greater 100‐kernel weight. Shandan 609 produced more rows per ear than Shandan 636, whereas Dongdan 60 had a higher ear row number and kernel number per row than both Shandan 609 and Shandan 636. Consequently, the lower yield of Shandan 636 can be attributed to its weaker performance in these key yield components (Table 1). Compared with 45 000 plants ha⁻¹, marginal yield gains decreased as planting density increased across all varieties. Shandan 8806 consistently yielded higher than Dongdan 60 at all densities. Notably, Shandan 650 and Shandan 8806 still showed yield potential when density increased from 60 000 to 75 000 plants ha⁻¹ (Fig. 3(b)).

Figure 3.

The yield of different varieties in Guanzhong region (a) and the yield increase effect of different varieties of maize after increasing planting density relative to 45 000 plants ha⁻¹ (b). The colors and shapes of the symbols represent different varieties as shown in the legend, and the sizes represent different planting densities. The letters after the formula represent whether there is a significant difference in the average yield of different varieties.

Impact of climate on maize yield in the Guanzhong region

Across the Guanzhong region, total precipitation and effective accumulated temperature during the maize growing season positively correlated with yield at all planting densities (Fig. 4). The positive relationship between precipitation and yield (P < 0.01) weakened as planting density increased and was weakest at 90 000 plants ha⁻¹. Effective accumulated temperature also showed a significant positive correlation with yield (P < 0.01), with the weakest relationship observed at 45 000 plants ha⁻¹ (P = 0.05) and the strongest at 60 000 and 75 000 plants ha⁻¹ (P < 0.01).

Figure 4.

Correlation analysis of total rainfall (a), effective accumulated temperature (b) and yield during growth period under different planting densities. Different colors and shapes of lines, symbols and formulas in the figure represent different planting densities.

Regional variation in climate and agronomic conditions significantly influenced maize yield responses to planting density (Fig. 5(b)–(d)), with clear differences among subregions and varieties (Fig. 6). In subregions where yield differences among varieties were not significant, lower lodging rates were associated with greater yield stability and better adaptability. In the western Guanzhong region, Zhengdan 958 exhibited lodging rates 79.38% and 59.46% lower than those of Shandan 8806 and Shandan 609, respectively, at 90 000 plants ha⁻¹, indicating stronger suitability for high‐density planting. In the middle region, Zhengdan 958 showed a 71.96% reduction in lodging rate at 75 000 compared with 90 000 plants ha⁻¹, suggesting that 75 000 plants ha⁻¹ is the more appropriate density. In the eastern region, lodging rates at 90 000 plants ha⁻¹ were lowest for Shandan 650, followed by Shandan 8806 and Shandan 609, indicating that Shandan 650 is more suited for high‐density planting in this zone (Supporting Information, Table S1).

Figure 5.

The importance analysis of influencing factors of maize yield in all sites (a), the western regions (b), the middle regions (c), and the eastern regions (d). Significant codes: ** means significant difference at 0.01 level; * means significant difference at 0.05 level; ns means no significant difference.

Figure 6.

Yield (kg ha⁻¹) of different varieties of maize under different planting density in different parts of the Guanzhong region. The different letters in the figure indicate whether there are significant differences between different planting densities and variety combinations in the same part of the Guanzhong region. Values with different lowercase letters within a variety differ significantly at the 5% level.

Correlation and path analysis of maize yield

Maize yield positively correlated with planting density, plant height, ear height, 100‐kernel weight, total precipitation, and effective accumulated temperature during the growing season. In contrast, ear row number negatively correlated with planting density and 100‐kernel weight, but positively correlated with plant height and ear height. Kernel number per row negatively correlated with planting density, ear height, 100‐kernel weight, and effective accumulated temperature, but positively correlated with precipitation. Hundred‐kernel weight negatively correlated with planting density, plant height, ear row number, and kernel number per row, but positively correlated with ear height and precipitation. These findings highlight the complex interactions among yield‐related traits (Fig. 7).

Figure 7.

Correlation analysis of each factor with yield and yield components. The lower left corner represents the correlation coefficient between any two indicators, the blue represents the positive correlation in the upper right corner, the red represents the negative correlation in the upper right corner. Significant codes: *** means significant difference at 0.001 level; ** means significant difference at 0.01 level; * means significant difference at 0.05 level; blank square means no significant difference.

Random forest analysis identified planting density as the most important predictor of yield, increasing the mean squared error by more than 40% and substantially exceeding all other variables. Total precipitation and effective accumulated temperature were the next most influential factors, whereas the remaining variables had comparatively smaller effects (Fig. 5(a)). Path analysis confirmed that planting density had the strongest direct effect on yield. A 1% increase in planting density was associated with a 51% increase in yield. Planting density also indirectly influenced yield through its effects on other traits, although it negatively correlated with 100‐kernel weight (r = −0.26). In turn, 100‐kernel weight had a strong positive effect on yield, with a 1% increase associated with a 57% yield increase. However, 100‐kernel weight negatively correlated with ear row number and kernel number per row (r = −0.25 and − 0.22, respectively), both of which positively contributed to yield, increasing it by 19% and 22% per 1% increase, respectively. Ear row number and kernel number per row were also positively correlated (Fig. 8).

Figure 8.

The path map of planting density, yield and its components. Significant codes: *** means significant difference at 0.001 level; * means significant difference at 0.05 level.

Yield classification and key factors for improvement

The yield thresholds (9000 and 12 000 kg ha⁻¹) were established by integrating local farmer surveys with our experimental data distribution (Fig. 2(a)), aligning with regional extension benchmarks. Surveys showed that while local farmers target 12 000 kg ha⁻¹, their actual yields mostly range between 9000 and 12 000 kg ha⁻¹, with a portion still falling below 9000 kg ha⁻¹. Conversely, our experimental histogram (Fig. 2(a)) revealed that approximately half of the treatment yields hovered below 9000 kg ha⁻¹, peaking near 8000 kg ha⁻¹. This yield gap justifies our threshold logic, where < 9000, 9000–12 000, and ≥ 12 000 kg ha⁻¹ define conventional baselines, attainable yields under intensive management, and regional high‐yield potential, respectively.

Compared with the low‐yield group, the high‐yield group had significantly higher planting density (17.65%), 100‐kernel weight (12.98%), plant height (7.40%), and ear height (4.23%) (Fig. 9). Improving these traits could therefore substantially enhance maize yield in low‐performing fields. In the low‐yield group, 45 000 plants ha⁻¹ accounted for 34.1% of observations. In contrast, 90 000 plants ha⁻¹ dominated the medium‐ and high‐yield groups, accounting for 32.0% and 38.0% of observations, respectively. These results highlight the importance of increasing planting density to transition low‐yield fields into higher‐yielding systems. In terms of varietal distribution, Shandan 636 was most frequently associated with the low‐yield group (21.1%), whereas Zhengdan 958 dominated the medium‐ and high‐yield groups (20.3% and 25.4%, respectively), further confirming its superior yield performance (Fig. 10).

Figure 9.

The distribution of planting density (a), rows per ear (b), kernels per row (c), 100‐kernel weight (d), plant height (e), and ear height (f) during the growing period under different yield groups (Low: ≤ 9000 kg ha⁻¹; Medium: 9000–12 000 kg ha⁻¹; High: ≥ 12 000 kg ha⁻¹). The dashed lines in different colors represent the average of the different groups, and the dashed lines in dark gray represent the total average of all groups. Values with different lowercase letters within a yield group differ significantly at the 5% level.

Figure 10.

Different planting density (a) and variety (b) proportion under different yield groups (Low: ≤ 9000 kg ha⁻¹; Medium: 9000–12 000 kg ha⁻¹; High: ≥ 12 000 kg ha⁻¹).

DISCUSSION

Optimizing planting density and variety selection for maize yield improvement

Increasing planting density can significantly enhance maize yield, although this benefit tends to plateau beyond a certain threshold.^{36, 38, 43} In the Guanzhong irrigation area, maize is typically planted at densities below 60 000 plants ha⁻¹. Our 5‐year study across 33 sites demonstrated that increasing density to 75 000 and 90 000 plants ha⁻¹ significantly improved yield and economic benefits (Fig. 2).

Yield improvements are primarily associated with changes in ear number per unit area, kernel number per ear, and 100‐kernel weight. ³² Higher planting densities increase ear number per unit area and improve canopy light interception and photosynthetic efficiency, thereby increasing yield.^{39, 44, 45, 46} However, at high densities, intensified interplant competition reduces the availability of light, nutrients, and water, which negatively affects ear development and kernel filling.^{47, 48} In this study, planting density negatively correlated with ear row number, kernel number per row, and 100‐kernel weight (Fig. 7). Integrating moderate increases in planting density with optimized practices – such as precision fertilization, irrigation scheduling, and integrated crop management – can stabilize yield components and further improve yield.^{49, 50}

Previous research reports that the optimal planting density across China's five major maize‐producing regions ranges from 70 500 to 117 300 plants ha⁻¹. ³⁷ In our study, yield plateaued between 75 000 and 90 000 plants ha⁻¹, indicating that at ultra‐high densities, the benefits of increased plant population are often offset by mutual shading and intensified intraspecific competition for resources.^{51, 52} Although higher densities can maximize total canopy light interception, they often reduce per‐plant carbon assimilation and decrease kernel weight. ⁵³ We identified 75 000 plants ha⁻¹ as an optimal trade‐off (Fig. 2 and Table S1). This density maintains efficient canopy resource capture while preserving individual plant vigor. It also improves stalk strength and ear uniformity, which are critical for reducing field losses and improving grain quality under mechanized harvesting. ⁵⁴ Consequently, 75 000 plants ha⁻¹ represents a more sustainable density for maximizing both yield efficiency and operational stability in the Guanzhong region. When combined with appropriate irrigation, this density alleviates intra‐population water stress, improves 100‐kernel weight, and enhances yield efficiency.^{37, 55}

Maize varieties differ in growth characteristics, environmental adaptability, tolerance to high planting density, stress resistance, and yield potential.^{56, 57} Selecting high‐yielding, locally adapted varieties is therefore essential for yield optimization. ⁵⁸ In this study, variety significantly affected plant height, ear height, yield, and yield components (Table 1), indicating that genetic background strongly determines performance and adaptability.^{59, 60} All six maize varieties responded positively to increased planting density, with the greatest yield gains observed in Shandan 8806 and Shandan 650 (Fig. 3). This response is likely associated with compact leaf architecture that enhances light interception and reduces ear barrenness under crowding stress. In contrast, the poorer performance of Shandan 636 suggests lower canopy efficiency and/or greater sensitivity to shading.

Higher planting densities also increase the risk of lodging. Dense stands often promote internode elongation at the expense of stalk diameter and mechanical strength, creating a genotype‐dependent trade‐off between yield and lodging resistance. ⁶¹ Varieties such as Zhengdan 958 and Shandan 8806 appear better able to balance plant height with stalk strength (Tables 1 and S1), whereas other genotypes are more vulnerable to lodging‐induced yield loss at high density. ⁶² Therefore, selecting hybrids that combine compact plant architecture with strong lodging resistance is essential for yield intensification in the Guanzhong region.

Plant height and ear height are key traits influencing lodging resistance and photosynthetic efficiency. ⁶³ Optimal maize yields are generally associated with plant heights of 240–300 cm and ear heights of about 120 cm.^{64, 65} An ear height‐to‐plant height ratio of 36.6–39.4% is considered optimal for balancing light interception and lodging resistance.^{66, 67} In this study, all six varieties fell within the optimal plant height range, although ear height was generally lower than ideal. Medium‐ and high‐yielding systems had ear height‐to‐plant height ratios closer to the optimal range than low‐yielding systems (Fig. 9). Lodging incidence increased with planting density across all varieties. The observed lodging trend was: Shandan 609 > Shandan 8806 > Shandan 636 > Zhengdan 958 > Shandan 650 > Dongdan 60 (Table S1). This trend is likely driven by increased interplant competition at high density, which weakens stem strength.^{33, 68} Among all varieties, Shandan 650 showed the strongest lodging resistance and was best suited for high‐density planting.

Optimizing planting density under changing climate conditions

Optimizing planting density requires integrating agronomic management with local environmental conditions to reduce climate‐related risks and convert low‐yield systems into high‐yielding systems.^{69, 70, 71} Climate is a key determinant of maize yield, with adequate temperature and rainfall during the growing season essential for stable productivity.^{72, 73} Over the 5‐year study period, the average annual temperature in the Guanzhong region was 13.40 °C (Fig. S2). Research indicates that when regional temperatures range from 7 to 14 °C, an optimal planting density of 103 500 plants ha⁻¹ yields the best outcomes, ³⁷ suggesting that planting density in the Guanzhong region may still be increased under improved agronomic management.

From June to October, the average temperature was 21.0 °C. Our multi‐site data showed that higher effective accumulated temperature during the growing season supports higher maize yields (Fig. 4(b)). Located north of the Qinling Mountains, the Guanzhong region is prone to low temperatures and frost damage; therefore, higher growing‐season temperatures generally benefit crop development. However, projected climate warming may disrupt this balance. ⁷⁴ Temperatures in the region are expected to increase by 2.3–5.3 °C between 2070 and 2099 relative to 1950–1999. ¹⁵ Such warming may shorten the maize growing period,^{68, 75} thereby limiting biomass accumulation and reducing yield.^{17, 29} Under these conditions, high‐density planting may either accelerate phenological development or improve the efficiency of effective accumulated temperature use. ⁶⁸ Selecting heat‐tolerant and late‐maturing varieties aligned with local climate trends will therefore be critical for maintaining maize yield.^{29, 68, 76}

Rainfall is another critical determinant of maize yield in water‐limited regions, where drought stress frequently constrains productivity. ⁷⁷ In this study, yield positively correlated with total precipitation during the growing season, confirming that the Guanzhong region remains water‐limited. Average annual rainfall across sites was 572.2 mm (Fig. S2). Based on a previous meta‐analysis, the optimal planting density at this rainfall level is approximately 104 200 plants ha⁻¹. ³⁷ In addition, precipitation in northwest China is projected to increase by about 10% between 2046 and 2065 relative to 1986–2005. ¹⁹ Although the marginal yield response to increasing density decreases under higher rainfall (Fig. 4(a)), improved water availability may still support higher yields under dense planting systems.

We also observed regional differences in yield response to planting density across the western, middle, and eastern Guanzhong areas. The middle region received the highest growing‐season rainfall (470.88 mm), followed by the eastern (450.41 mm) and western (445.45 mm) regions. In contrast, the eastern region had the highest effective accumulated temperature (1722.24 °C), followed by the middle (1633.98 °C) and western (1441.68 °C) regions (Fig. S2). These climatic differences contributed to spatial variation in yield, with higher yields in the middle and eastern regions and lower yields in the western region (Fig. 6). Random forest analysis confirmed that rainfall exerted a stronger influence on yield than effective accumulated temperature in the Guanzhong region (Fig. 5(a)), highlighting the dominant role of water availability, particularly under dry conditions. Improving maize productivity in the western region will therefore require targeted interventions, including enhanced irrigation, drought‐tolerant varieties, and optimized planting density. At low planting density, yield showed a stronger correlation with rainfall than with accumulated temperature (Fig. 4), because reduced competition allows individual plants greater access to soil resources, making water availability the primary limiting factor. ⁶⁰ In contrast, at high planting density, yield showed a stronger association with accumulated temperature than with rainfall (Fig. 4). Under dense canopies, increased competition reduces uniformity in light and nutrient distribution, and accumulated temperature more strongly regulates growth rate and developmental progression, thereby influencing yield formation. ⁶⁸

Overall, these results improve the understanding of how climate and management interact to determine maize yield in the Guanzhong region and provide a scientific basis for region‐specific agronomic optimization strategies.

CONCLUSION

Optimizing planting density and selecting suitable maize varieties are key strategies for increasing yield and economic returns in the Guanzhong region. This study showed that a planting density of 75 000 plants ha⁻¹ achieved the highest overall yield, while Shandan 8806 and Shandan 650 exhibited strong yield potential across environments. Enhancing maize yield under higher planting densities depends on improving 100‐kernel weight, maintaining an optimal ear‐to‐plant height ratio, and managing soil moisture and temperature to reduce lodging risk. These factors are crucial for converting low‐ and medium‐yield fields into high‐yield systems. Effective accumulated temperature and rainfall during the growing season positively correlated with yield across all planting densities, highlighting the importance of climate adaptation in maize production. These findings provide practical guidance for increasing maize yield in underperforming areas of the Guanzhong region and offer useful insights for other semi‐arid and semi‐humid regions in northwest and north China with a continental monsoon climate, annual precipitation of 500–600 mm, and a typical winter wheat–summer maize double‐cropping system, such as the northern Huang‐Huai‐Hai Plain. The identified density–yield–climate relationships provide a decision‐support framework for farmers to adapt to increasingly variable monsoon rainfall and thereby improve the resilience of the wheat–maize rotation system. However, the analysis is limited by the lack of detailed data on additional field management practices (e.g., mulching and straw returning) and on more comprehensive soil and microclimate variables. Future research should incorporate these factors to improve the understanding of how planting density interacts with agronomic and environmental conditions to influence maize yield.

CONFLICT OF INTEREST

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.

Supporting information

Figure S1. Average economic returns (CNY/ha yr) for each maize variety under different planting densities from 2017 to 2022. Different lowercase letters within a variety indicate significant differences at the 5% level.

Figure S2. Monthly rainfall (green bars) and average temperature (red line) across subregions of Guanzhong, Shaanxi Province, from 2017 to 2021.

Table S1. Lodging rate (%) of various varieties under different planting densities across subregions of the Guanzhong region (2017–2021).

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.