The experimental site was located in Jiangsu Province, China (31°30′N, 120°33′E). The soil is formed on clayey lacustrine deposits as Gleyic Stagnic Anthrosol. Initial soil analysis was 19.2 g organic C kg^-1, 1.3 g total N kg^-1, 0.9 g total P kg^-1, 15.0 g total K kg^-1and pH of 7.0. The soil at the experimental site is classified as loam with sand of 33.8%, silt of 38.6%, and clay of 27.6%. The experimental site belongs to a humid subtropical climate with an average annual temperature of 16 °C and annual precipitation of 1100-1200 mm (Wang et al., 2018b).

The facility operation followed the procedures described by Wang et al. (2019). The treatments were randomly arranged in three blocks, with each block having four treatments (rings). The area of each ring is 50 m². The four treatments included ambient condition (CK), elevated [CO₂] (500 ppm, CE), canopy warming (+2°C, WA) by infrared heaters, and combined treatment of elevated [CO₂] and canopy warming (CW). An interval of 28 m was set up between rings to avoid any potential contamination across treatments.

Winter wheat (Triticum aestivum L.) of Yangmai No.14 was planted at a row spacing of 20 cm in November 2013 and harvested in May 2014. Based on local practice, the basal fertilizer in the form of urea (46% N) was applied at a rate of 187.5 kg ha^-1, and top-dressed at a rate of 150 kg ha^-1 at the elongation stage. The topdressing fertilizer was provided by a compound fertilizer (15 N: 15 P₂O₅: 15 K₂O) at a rate of 375 kg ha^-1 after the heading stage. The weed control and insecticide application were carried out according to local agronomic management. The wheat was cultivated under rain-fed conditions.

Plant and soil samples were taken at the elongation, heading, and ripening. Since climatic change treatment altered wheat development, soil and plant sampling was conducted based on the phenological stage. Wheat plants were randomly collected 1 m² from each plot, plant sample was separated into shoot and root. Root samples were rinsed with water to get rid of the soil. Meanwhile, the shoot and root samples were rinsed and de-enzyme at 105°C for 0.5 h, and then oven-dried at 70°C for 48 h. All plant samples were ground to 0.25 mm. For soil samples, five soil cores (0-15 cm) were taken and then homogenized to form a mixed soil sample. After removing visible residues and stones, soil samples were passed through a 2 mm sieve and maintained at 4°C before analysis.

Soil available P was extracted with NaHCO₃ and analyzed by a spectrophotometer (TU-1810, China). Soil available K concentrations were estimated using a flame photometer (FP6410, China) after extraction with 1 M ammonium acetate (NH₄OAc). Soil NH 4 + -N and NO 3 − N concentrations were determined using a subsection flow analysis instrument (Skalar, Netherlands).

Plant shoot and root samples were pretreated with H₂SO₄-H₂O₂, and analyzed for N concentrations by the Kjeldahl digestion method. Phosphorus concentrations were determined by a spectrophotometer (TU-1810, China), and K concentrations by a flame photometer (FP6410, China).

Nutrient transfer coefficients were used to evaluate the impact of elevated [CO2] and canopy warming on the capacity of plant nutrient uptake (Wang et al., 2016). Nutrient transfer coefficients were estimated as:

general linear mixed model (GLM) was used to detect the [CO₂], warming and growth stage (main factor), with blocks treated as a random variable. Accordingly, a three-way analysis of variance was used to test the differences between treatments and the growth stage. The probability level (P< 0.05) was considered to be statistically significant. Statistical analyses were conducted by SPSS v.22.0 (IBM Corp., Armonk, NY, USA).

When averaged across three growth stages, elevated [CO₂] did not alter plant N or P concentrations ( Figure 1 ), but reduced K concentrations by 11.0% (P = 0.001) in the shoot and 11.5% (P = 0.001) in the root ( Figure 1 and Table 1 ). By contrast, canopy warming significantly increased shoot N, P, and K concentrations by 8.9% (P = 0.016), 7.5% (P = 0.054), and 15.0% (P< 0.001), but reduced root N and K concentrations by 12.3% (P< 0.01), and 31.6% (P< 0.001), respectively. However, plant N, P, and K concentrations varied among stages and reached the peak at the elongation stage. Significant interactions of [CO₂] × warming and [CO₂] × warming × stage were observed in the N and P concentrations: elevated [CO₂] aggravated the positive effect of canopy warming on shoot N and P concentrations, especially at the elongation stage, but mitigated the adverse impact of canopy warming on root N and P concentrations at the heading stage.

Elevated [CO₂] and canopy warming had significant effects on nutrient transfer coefficients, and these effects varied with wheat growth stages ( Table 2 ). When averaged across three growth stages, canopy warming significantly increased N, P, and K transfer coefficients by 22.2% (P = 0.001), 27.9% (P = 0.001), and 84.3% (P< 0.001), respectively. Elevated [CO₂] enhanced P and K transfer coefficients by 27.9% (P = 0.058) and by 10.1% (P< 0.05), but did not affect the N transfer coefficient. A significant interaction of [CO₂] × warming was detected for P (P< 0.001) and K (P< 0.01) transfer coefficients: the increases in P and K transfer coefficients by canopy warming were aggravated by elevated [CO₂]. The impact of canopy warming appeared differently among stages as evidenced by significant warming × stage interaction.

Elevated [CO₂] and canopy warming did not alter soil nutrient status ( Table 3 ), although elevated [CO₂] slightly increased soil available P concentrations by 13.2% (P< 0.05). However, soil nutrient availability varied across stages with the peaking soil NO 3 − N concentrations occurring at the heading stage and the peaking P and K concentrations at the ripening stage. There was a significant interaction of [CO₂] × warming × growing stages on soil K concentrations. A significant effect of elevated [CO₂] × warming interaction was observed for soil NO 3 − N and available P concentrations: warming increased soil NO 3 − N and available P concentrations under ambient [CO₂] but not under elevated [CO₂].

Elevated [CO₂] alone decreased plant K concentrations, but the responses varied among growth stages ( Table 1 ). The dilution effect in the K concentrations has been widely described in plants under elevated [CO₂] (Han et al., 2015; Wang et al., 2019). Kanowski (2001) found that elevated [CO₂] (790 ppm) reduced K concentrations in Flindersia. By contrast, elevated [CO₂] did not alter the N and P concentrations for both shoot and root ( Figure 1 and Table 1 ). This contrasts with other studies and indicates that elevated [CO₂] is associated with the dilution of nutrient concentrations in wheat grain under sufficient fertilizer input (Lam et al., 2012b). Our previous study argued that elevated [CO₂] did not affect plant nutrient concentrations under adequate fertilizer supply in the rice paddy field (Wang et al., 2018a). On the other hand, the reasons were ascribed to the levels of [CO₂] elevation (500 ppm) in this study, which was much lower than in other studies (more than 550 ppm). Whereas, the P and K transfer coefficients were significantly increased by elevated [CO₂], which was ascribed to an increase in nutrient demand by crop aboveground biomass (Wang et al., 2019). Indeed, we observed that elevated [CO₂] significantly increased grain yield by 29.6% ( Figure S1 ). Elevated [CO₂] did not affect the N transfer coefficient, which is due to the inhibition of NO 3 − N assimilation or lower investment in Rubisco in the shoots of wheat (Bloom et al., 2010; Ainsworth and Long, 2021). These results indicated that the mechanisms for nutrient translocation from root to shoot varied with plant nutrient demands.

Canopy warming significantly increased nutrient concentrations in plant shoots ( Figure 1 and Table 1 ). Warming-induced increase in plant N concentrations (29.8-32.7%) was also observed in a tallgrass prairie ecosystem (An et al., 2005). Trueman and Gonzalez-Meler (2005) have indicated that higher air temperature would increase the vapor pressure deficit of the canopy and leaf transpiration, thereby increasing nutrient translocation from root to shoot. Our previous study showed a significant increase in evapotranspiration under canopy warming conditions in this winter wheat field (Wang et al., 2018b). Moreover, we also found that canopy warming significantly reduced nutrient concentrations in roots and increased nutrient transfer coefficients ( Tables 1 , 2 ).

Our results observed that combined treatment of elevated [CO₂] and canopy warming did not affect shoot nutrient concentrations ( Figure 1 ). This is consistent with a previous study conducted by Cheng et al. (2010), who observed no significant change in rice N concentrations under concurrent elevated [CO₂] (680 ppm) and high night temperature (+10°C). However, Jauregui et al. (2015) has shown that the interaction of elevated [CO₂] (700 ppm) and temperature (+ 4°C) significantly decreased N and K concentrations, but did not affect the P concentrations in wheat leaf in a greenhouse study. In contrast, elevated [CO₂] (550 ppm) and elevated air temperature (+ 2°C) significantly increased P uptake in rice in an open top chamber study (Bhattacharyya et al., 2014). The inconsistent results were attributed to differences in experimental designs and variations in crop cultivars, and the low statistical power of individual studies. Reich et al. (2016) found that temperature determined the response of plant N assimilation and mineral nutrient composition to elevated [CO₂]. Similarly, our results found that canopy warming altered nutrient uptake response to elevated [CO₂], with elevated [CO₂] significantly increasing P and K transfer coefficients under canopy warming, but decreasing the P transfer coefficient under ambient temperature. Therefore, further studies are needed to reveal the mechanisms of plant nutrient assimilation under future concurrent elevated [CO₂] and warming conditions.

As mentioned above, elevated [CO₂] and canopy warming significantly altered nutrient uptake. However, opposite to our hypothesis, elevated [CO₂] or canopy warming did not affect soil nutrient status ( Table 3 ). Previous studies reported that elevated [CO₂] did not change soil N or P availability in paddy fields (Ma et al., 2007; Cheng et al., 2016), although it is generally accepted that elevated [CO₂] increased nutrient demand through increasing plant biomass (Lam et al., 2012b; Wang et al., 2018a). However, a recent study demonstrated that [CO₂] fertilization increased N and P availability in a P-limited forest ecosystem (Hasegawa et al., 2016). The present study was not constrained by nutrients due to the frequent fertilizer applications, which suggests that soil nutrient availability can be replenished by fertilizer input in an intensively managed agricultural ecosystem under future climate scenarios.

Our results showed that canopy warming increased soil NO 3 − N and available P concentrations under ambient conditions ( Table 3 ). Warming greatly affects soil microbial and enzyme activity, which stimulates soil nutrient availability (Liu et al., 2015; Osanai et al., 2015; Wang et al., 2016). Warming increased nutrient mineralization, which leads to the stimulation of nutrient availability and increases nutrient assimilation by plants (Zuccarini et al., 2020; Iversen et al., 2022). In contrast, canopy warming did not significantly alter soil nutrient availability under elevated [CO₂] ( Table 3 ). The acceleration of soil nutrient availability is counteracted by the plant demand and soil moisture, and was even reduced under warming. Elevated temperature decreased soil moisture, resulting in a limitation in soil nutrient mineralization under dry conditions (Borken and Matzner, 2009; Wang et al., 2016). Our results demonstrated that both canopy warming alone and combined with elevated [CO₂] generally increased nutrient transfer coefficients ( Figure 1 and Table 2 ). Therefore, the long-term climatic change probably increases soil nutrient consumption, which has a negative impact on food production. However, the responses of soil nutrient status to future concurrent elevated [CO₂] and canopy warming are complicated, which warrants further studies for developing adaptation strategies to future climate change.