N-fertilizer postponing application improves dry matter translocation and increases system productivity of wheat/maize intercropping

Intercropping increases the grain yield to feed the ever-growing population in the world by cultivating two crop species on the same area of land. It has been proven that N-fertilizer postponed topdressing can boost the productivity of cereal/legume intercropping. However, whether the application of this technology to cereal/cereal intercropping can still increase grain yield is unclear. A field experiment was conducted from 2018 to 2020 in the arid region of northwestern China to investigate the accumulation and distribution of dry matter and yield performance of wheat/maize intercropping in response to N-fertilizer postponed topdressing application. There were three N application treatments (referred as N1, N2, N3) for maize and the total amount were all 360 kg N ha−1. N fertilizer were applied at four time, i.e. prior to sowing, at jointing stage, at pre-tasseling stage, and at 15 days post-silking stage, respectively. The N3 treatment was traditionally used for maize production and allocations subjected to these four stages were 2:3:4:1. The N1 and N2 were postponed topdressing treatments which allocations were 2:1:4:3 and 2:2:4:2, respectively. The results showed that the postponed topdressing N fertilizer treatments boosted the maximum average crop growth rate (CGR) of wheat/maize intercropping. The N1 and N2 treatments increased the average maximum CGR by 32.9% and 16.4% during the co-growth period, respectively, and the second average maximum CGR was increased by 29.8% and 12.6% during the maize recovery growth stage, respectively, compared with the N3 treatment. The N1 treatment was superior to other treatments, since it increased the CGR of intercropped wheat by 44.7% during the co-growth period and accelerated the CGR of intercropped maize by 29.8% after the wheat had been harvested. This treatment also increased the biomass and grain yield of intercropping by 8.6% and 33.7%, respectively, compared with the current N management practice. This yield gain was primarily attributable to the higher total translocation of dry matter. The N1 treatment increased the transfer amount of intercropped wheat by 28.4% from leaf and by 51.6% from stem, as well as increased the intercropped maize by 49.0% of leaf, 36.6% of stem, and 103.6% of husk, compared to N3 treatment, respectively. Integrated the N fertilizer postponed topdressing to the wheat/maize intercropping system have a promotion effect on increasing the translocation of dry matter to grain in vegetative organs. Therefore, the harvest index of intercropped wheat and maize with N1 was 5.9% and 5.3% greater than that of N3, respectively. This demonstrated that optimizing the management of N fertilizer can increase the grain yield from wheat/maize intercropping via the promotion of accumulation and translocation of dry matter.

www.nature.com/scientificreports/ Crop growth rate of wheat. The average CGR of wheat was significantly affected by cropping system, N-fertilizer treatment, and the two factors' interaction effect (except from early to end of May and early to end of Jul). The 3-year average CGR of intercropped wheat was higher than sole wheat in whole growth period. There was no difference between each treatment at early growth stage but this trend changed with the growth stage developed (Fig. 2). The CGR of wheat increased rapidly and reached a maximum value when wheat was at early grain-filling stage. At this stage, intercropping significantly increased it by 13.4-57.9%, 6.0-60.9% and 13.5-62.5% than sole system in 2018, 2019 and 2020, respectively. The IN 1  Crop growth rate of maize. The average CGR of maize was significantly affected by cropping system and N-fertilizer postponing application, but the two factors' interaction effect had no influence. The growth of intercropped maize was influenced by component wheat (Fig. 3). Before wheat harvest, the CGR of maize in sole cropping was higher than that in intercropping. The SN 1-m and SN 2-m treatments increased average CGR of sole maize by 15.6-40.9% and 9.0-21.6% compared with SN 3-m treatment. After wheat harvest, the CGR of intercropped maize was higher than sole maize. The maximum CGR of maize was occurred at the end of July to middle of August, i.e., at anthesis to early grain filling stage. At this stage, the average CGR of intercropped maize under IN 1  Biomass yield of wheat and maize. The biomass yield (BY) was significantly affected by cropping system, N-fertilizer treatment, and their interaction. On average of 3 years, the BY of intercropped wheat was 35.9-48.7% higher than that of sole wheat. The BY of intercropped maize was 12.8-31.1% higher than that of sole maize. Furthermore, the BY in intercropping was 24.7-32.9% higher than the weighted means of sole cropping (Fig. 4) Distribution characteristics on aboveground dry matter of wheat. The transfer amount (DTA), transfer rate (DTR), and contribution rate to grain yield (GCR) of wheat during three experimental years were significantly influenced by cropping system, but not by N treatment and their interaction ( Yield components of maize. The SN, and TKW of maize were significantly affected by cropping system, but not by N management and their interaction ( Table 4). The SN of intercropped maize was 6.5-15.7% higher than that of sole maize under the same land area. Whereas, the KNS and TKW of intercropped maize was were     Harvest index of wheat and maize. Harvest index (HI) of wheat and maize was significantly affected by cropping system and N-fertilizer treatment (except for wheat), but not by their interaction (Fig. 7). Path analysis. The correlation coefficients between the grain yield and yield components were used to separate into direct and indirect effects via path analysis (Fig. 8A). The spike number (SN) and thousand-kernel weight (TKW) of wheat had the highest direct path coefficient and correlation coefficient than kernel number per spike (KNS). In addition, TKW had a positive indirect path coefficient with SN and SN had a positive indirect path coefficient with TKW, indicating that yield was influenced by the interaction between them. Although  www.nature.com/scientificreports/ KNS has the lowest direct path coefficient (0.064), the indirect path coefficient of KNS to SN is − 0.240, which is 3.75 times for its direct path coefficient. The SN of maize was significantly correlated with grain yield (Fig. 8B). Furthermore, SN had the highest the correlation coefficient than TKW and KNS, indicating SN had direct influence on grain yield. Nevertheless, KNS and TKW could indirectly affect grain yield via SN, with TKW contributing more than KNS.

Discussion
The crop growth rate and biomass yield. Numerous studies had reported one important factor affecting the obtain of grain yield was dry matter accumulation 32,33 . This is mainly because dry matter accumulation and distribution in reproductive organs of crops, can reflect indirectly grain yield response to the availability of resource 34 . In the present study, the CGR of wheat/maize intercropping presents an obvious double-peak curve in each studied year. Before wheat harvest, the CGR of intercropping reached the maximum value. After wheat harvest, it decreased and reached the second maximum value when maize was at early grain filling stage. The reason was that intercropped wheat was earlier planted and created a competitive advantage over the later planted intercropped maize for resources uptake during co-growth period, resulted in a strong suppression of intercropped maize 5,9 . Thus, the CGR of the intercropped wheat was higher than sole wheat. Owing to the high  www.nature.com/scientificreports/ light intensity, wheat (C3 crop) may use light more efficiently in the intercropping than in sole crop during co-growth period 35,36 . The CGR of intercropped maize was lower than sole cropping before wheat harvest, but higher after wheat harvest. This result was consistent with previous studies, with aboveground dry matter of maize showing recovery growth after wheat harvest 5 . Therefore, the weighted means of BY in intercropping was 24.7-32.9% higher than that of monocropping. That means the intercropping can accumulate more dry matter than the corresponding sole system. In this study, IN 1 and IN 2 treatment, which postponed 20% and 10% of total N fertilizer from maize jointing stage to 15 days post-silking stage, had a significant effect on boosting the maximum CGR of intercropped wheat, which was boosted by 44.7% and 22.7% compared to IN 3 treatment. That may because the IN 3 treatment used excessive N fertilizer at maize jointing stage which is not suitable for intercropped wheat growing. Numerous studies have shown that nutrients play a crucial role in recovery growth of late-maturing crops after the early-maturing crops harvest 11,20,37 . Compared to the IN 3 treatment, the IN 1 and IN 2 treatment increased the CGR of intercropped maize by 29.8% and 12.6%. The reason might be the postponed topdressing N fertilizer greatly intensified the interspecific competition in co-growth period but eventually generated a substantial complementarity 19 . This was similar to previous research that adequate N supply plays a pivotal role in recovery growth of intercropped maize after wheat harvest 20 .

The transfer of vegetative products to ear. The proportion of photosynthetic products stored in leaves
and stems is relatively small and most dry matter accumulation during the grain-filling period is accumulated in grain 33 . In this study, intercropping had a significant effect on aboveground dry matter translocation. It increased DTA of leaf by 65.0%, DTR by 28.2%, and GCR by 69.3% and 89.5%, 60.6%, and 84.6% from stem compared with sole wheat, respectively. One main reason is that during the co-growth period, wheat has a competitive advantage, and can obtain more light and heat resources. Meantime, wheat shaded the adjacent maize, thus reducing solar radiation received by maize 5,38 . Interspecific competition not only includes aboveground competition but also contains belowground competition. Belowground competition was mainly for growth space, water, and nutrients. As shown in this study, IN 1 and IN 2 treatments increased the DTA of intercropped wheat by 10.7-28.4%, DTR by 3.4-8.1%, and GCR by 11.6-29.6% from leaf, compared to IN 3 treatment, and 14.9-51.6%, 10.6-34.1%, 19.9-55.0% from stem, respectively. Therefore, IN 1 treatment showed the best effect on optimizing dry matter distribution of aboveground tissue in intercropped wheat. Previous research suggested that adequate N supply directly affects the production, partitioning, and translocation of dry matter 22 . An increasing in wheat transferring amount, transferring rate, and contribution rate to grain might because wheat has a higher competitive ability for N. Intercropped wheat having much greater root length density, and roots spreading laterally into the maize strip during the co-growth period 39 , and then competing for N from the adjacent maize strip. www.nature.com/scientificreports/ However, late-maturing crops could form the compensatory effect of time and space when early-maturing crops were harvested. In this study, intercropping increased the DTA by 38.7%, DTR by 29.1%, and GCR by 53.6% from leaf compared to sole maize, by 27.4%, 20.4%, and 40.6% from stem, and by 51.4%, 61.2%, and 64.5% from husk, respectively. That means the increasing in maize aboveground dry matter translocation probably resulted from compensatory effect, which late-maturing crops (such as maize and soybean) root gradually expand to the underground space of early-maturing crops (like wheat) after it harvest, absorb more nutrient and water, thereby accelerated the growth rate of late-maturing crops 11 . It has been confirmed that recovery growth is fundamentally related to the supplemental N 40 . In wheat/maize intercropping, the IN 1 treatment increased the DTA, DTR, and GCR by 49.0%, 32.6%, and 48.4% from leaf, by 36.6%, 8.6%, and 39.1% from stem, and by 103.6%, 36.8%, and 105.7% from husk compared to IN 3 treatment, respectively. In this study, maize performed the highest compensatory intensity during the third recovery stage (i.e., from grain filling to maturity), which was similar to previous research 41 . That is to say, suitable fertilizer N management at this stage is the key to enhance recovery growth. The IN 1 treatment transferred 20% of total N at this stage can well match fertilizer N supply with crop N requirement.
Yield performance and yield components. The common advantages of intercropping are (i) efficient use of nutrients, light, and water 42,43 , (ii) achieving agricultural biodiversity, and (iii) increasing yield 28,44 . In northwest China, wheat/maize intercropping, an old cropping practice that aims to match efficient crop demands to the available growth resources and labor, has been widely used by farmers 45 . In the present study, the grain yield in intercropping was 19.1-30.7% higher than the weighted means of corresponding sole cropping. It was because intercropped wheat had a strong competition relative to the accompanying maize, more resources in the adjacent vacant area were available to intercropped wheat 28 , thus intercropped wheat obtained greater yield components and higher grain yield than sole wheat under the same area. After wheat harvest, expansion of absorption space for light, heat, and gas resources on the ground coupled with the expansion of absorption scope for water and nutrients underground gave intercropped maize a chance to compensate, which is the basis for high yields of intercropped maize 11 . It has been discovered that coordinated development among yield components is the foundation for achieving high grain yield for cereal crops 46,47 . In present study, intercropping increased the yield components of wheat and maize. Under the same land area, intercropping with the three N fertilizer postponed topdressing treatment increased SN of wheat by an average 18.0% and by 11.2% of maize compared to sole cropping, across the 3 years. Similarly, intercropping increased KNS of wheat by 15.0%. This is mainly because that favorable interspecific competition and compensation effect is beneficial to improve yield components and crop grain yield, thus obtaining the higher harvest index 47 .
In present study, IN 1 and IN 2 treatments boosted the mixed yield by 33.7% and 15.9% compared with IN 3 treatment. It had been reported that the N 1 treatment, where 45 kg N ha −1 was applied at the first topdressing plus 135 kg N ha −1 at the third topdressing, can boost the grain yield of intercropped pea and maize compared to the N 3 treatment which 135 kg N ha −1 was topdressing at the first plus 45 kg N ha −1 at the third topdressing 19 . Mainly because the competitive ability of legumes was improved in planting mixtures so as to enhance the yield of intercropping 48 . N application could not only boost the grain numbers per unit areas, but also improve grain protein concentration 49 . The IN 1 and IN 2 treatment increased the spike number (by 13.8 and 5.0%), kernel number per spike (by 16.0% and 10.0%), and the thousand-kernel weight (by 7.3% and 3.2%) of intercropped wheat; similarly enhanced the spike number (by 24.8% and 15.5%), the kernel number per spike (by 12.0% and 7.7%), and the thousand-kernel weight (by 11.9% and 9.5%) of intercropped maize, respectively. One reason for this phenomenon might be N-fertilizer postponed topdressing is an effective approach to match fertilizer N supply with crop N requirement which is crucial to achieving high productivity 10 . Path analysis showed that grain yield of wheat was mainly derived from spike number and thousand-kernel weight, and while kernel number per spike indirectly influences spike number so as to affect the grain yield. The grain yield of maize was mainly derived from spike number, while thousand-kernel weight and while kernel number per spike indirectly influences spike number so as to affect the grain yield. In this experiment, IN 1 and IN 2 treatments increased the average HI of intercropped wheat by 5.9% and 2.6%, and by 5.3% and 3.6% of intercropped maize compared to IN 3 treatment. This mainly because intercropped wheat can capture more resources during the co-growth stage and intercropped maize attributed to more transfer of aboveground dry matter to vegetative organs to ear during the late-growth stage 50 . Furthermore, the total LER of wheat/maize intercropping averages 1.28, which indicated the intercropping system used less land but produced more grain than their corresponding monocultures. This means intercropping system can more efficiently use the resources to product than monocultures.

Conclusions
The N-fertilizer postponed topdressing treatments, which transferred 20% or 10% of the total amount N from the jointing stage to 15 days post-silking stage, boosted the crop growth rate of intercropping wheat during the co-growth stage and simultaneously accelerated the crop growth rate of intercropping maize crops during their recovery growth stage, respectively. They also increased the biomass yield of intercropping by 8.6% and 5.0%, compared with traditional N management practices, respectively. The N fertilizer postponed topdressing optimized the transfer of dry matter from vegetative organs to grain and increased the proportion postponed that boosted the amount of transportation. The postponed topdressing applications at 20% and 10% enhanced the mixed grain yield by 33.7% and 16.0%, compared with traditional N management practices, respectively. The harvest index of intercropped wheat increased by 5.9% and 2.6%, respectively, and that of intercropped maize by 5.3% and 3.6%, compared with traditional N management practices, respectively. Our results showed that N fertilizer postponed topdressing, particularly postponing the application at 20%, can increase the accumulation www.nature.com/scientificreports/ of photosynthetic products and optimize the translocation of dry matter, which improved the productivity of intercropping systems. Experimental design. The experimental design was a factorial design with seven treatments and three replications. Cropping systems were sole maize, sole wheat, and wheat/maize intercropping. Three N-fertilizer postponed top-dressing treatments (N 1 , N 2 and N 3 ) were designed according to key growth stage of maize that was jointing stage (V6), pre-tasseling stage (V12), and 15 days post-silking (R2) (Fig. 9). The N 3 treatment is the local N management practice in this region. N fertilizer rate for sole maize was 360 kg N ha −1 , in which 20% and 40% of total N application were applied pre-plant and top-dressed at pre-tasseling stage, respectively. The remaining 40% was divided into jointing stage and 15 days post-silking stage and the allocations were, respectively: 10% and 30% for N 1 ; 20% and 20% for N 2 ; and 30% and 10% for N 3 , thus formed postponing application of 20% (N 1 ), 10% (N 2 ), and without postponing application (N 3 ). The total amount of N fertilizer was 285 kg N ha −1 for wheat/maize intercropping, which was calculated by the bandwidth ratio. N fertilizer rate for sole wheat was 180 kg N ha −1 , in which 108 kg N ha −1 was base applied at sowing and 72 kg N ha −1 at booting stage (i.e. pre-tasseling stage of maize). Crops in sole and intercropping received an equivalent N rate at specific area. The detailed treatment code and N-fertilizer management were presented in Table 5. The amount of phosphorus was 180 kg P ha −1 and applied in all plots before sowing. The types of N and P were urea (46-0-0, N-P-K) and superphosphate (11-51-0) fertilizers. The topdressing fertilizer in maize strips was achieved by the drip irrigation method. The plot size for intercropping was 5.7 m length × 6 m width, and for sole cropping was 6 m length × 6 m width, with every neighboring plot had a 50 cm wide by 30 cm high ridge built to eliminate potential water movement. In intercropping plots, wheat and maize were alternated in 190 cm wide strips, in which, wheat strip was 80 cm wide consisting of six rows with a row space of 12 cm, and maize strip was 110 cm wide consisting of three rows with 40 cm row (Fig. 10). Thus, in the wheat/maize intercropping, wheat account for 42% of the plot area and maize account for 58%. The planting density of sole wheat was 6,750,000 plants ha −1 and sole maize was 90,000 plants ha −1 . For each crop, the same area-based planting density was employed in intercropping and sole cropping. Intercropped wheat was at 2,840,000 plants ha − The use of maize and wheat seeds in the present study was permitted by Gansu Agricultural University and it complies with local and national guidelines and legislation. Maize was mulched by plastic film (polyethylene film 0.01 mm thick and 120 cm wide), which made in Lanzhou Green Garden Corporation of China, Lanzhou. It is an innovative technology largely adopted in arid areas to improve maize productivity 52 . There is low precipitation at the testing areas (< 155 mm annually), so that supplemental irrigation was applied. Before soil freezing, 120 mm of

Plant sampling and analysis
Aboveground dry matter. The sole and intercropped components were collected for aboveground dry matter determination at 15 days intervals before wheat harvest, and at 20 days intervals after wheat harvest. The first sampling was conducted at 15 days after maize emergence. For the sake of minimizing the influence of destructive sampling on yield formation, 2/3 of the plot in width was used to measure dry matter accumulation, and the remaining 1/3 were used to measure grain yield at physiological maturity. At each sampling date, 20 wheat plants in the same row were randomly selected to determine wheat aboveground dry matter (DM). For maize, 10 individual plants were randomly selected before jointing stage and 5 plants after jointing stage to determine maize DM. Samples were separated into leaf, stem, and ear of wheat and leaf, stem (include sheath), husk, and ear of maize per plant. All samples were oven-dried at 105 °C for 30 min and weighed after further drying at 80 °C until a constant weight was attained. Finally, the aboveground biomass was used to calculate the transportation amount, and transportation rate of dry matter in vegetative organs to grain, and the contribution rate of vegetative organs to grain according to Yin 50 . The equation was following: (1) DTA = LDW − DWM (2) DTR = DTA LDW × 100% Table 5. N fertilizer allocation amount (kg ha −1 ) and percentage in each treatment. a For sole maize, N 1 , N 2 and N 3 represent N-fertilizer applied at 36, 72, and 108 kg N ha −1 as first top-dressing plus 108, 72, and 36 kg N ha −1 at third top-dressing, respectively. For sole wheat, N 1 represents the N-fertilizer applied at 108 kg N ha −1 as base fertilizer at sowing plus 72 kg N ha −1 top-dressed at pre-tasseling. b Intercropped components (i.e., maize and wheat) received the same area-based N fertilizer rate as the corresponding sole crops. c The postponed percentage applied only for maize.  Figure 10. Information of (a) the spatial arrangement of wheat/maize intercropping with wheat strip of 80-cm (six rows) alternated with maize strip of 110-cm (three rows) and (b) the field planting diagrammatic representation of wheat/maize intercropping at Wuwei experimental station in northwestern China. www.nature.com/scientificreports/ where DTA (kg ha −1 ) represents transportation amount of dry matter in vegetative organ, LDW (kg ha −1 ) represents the largest dry weight of the vegetative organ, DWM (kg ha −1 ) represents the dry weight of the same vegetative organ in maturity, DTR represents transfer rate of dry matter in vegetative organ (%), GCR represents contribution rate of vegetative organs to grain (%) and GDW (kg ha −1 ) represents the dry weight of grain.
Crop growth rate. The crop growth rate was calculated (CGR) (kg ha −1 day −1 ) using the following equation: where W 2 and W 1 are the aboveground biomass accumulation sampled at T 2 and T 1 .
Grain yield, biomass yield, yield components, and harvest index. Grain yield (GY) and biomass yield (BY) were measured after air-drying, cleaning of the sole and intercropped systems from all plots. At the maturity stage, 30 wheat plants and 10 maize plants in the undisturbed natural strip were randomly selected to test kernel number per spike (KNS) and thousand-kernel weight (TKW); measure 2.5 × 0.8 m = 2 m 2 (wheat), 5 × 1.0 m = 5 m 2 (maize) square area to count the spike number (SN) and calculate the grain yield per unit area by threshing and weighing. Harvest index (HI) was determined by dividing GY by aboveground BY at physiological maturity: Land use efficiency. The land equivalent ratio (LER) was calculated as follows: where Y im and Y sm are the grain yield of intercropped maize and sole maize, respectively, and Y iw and Y sw are the grain yield of intercropped wheat and sole wheat, respectively. A value of LER > 1.0 indicates a yield advantage of intercropping over sole cropping and vice versa.
Statistical analysis. Data were analyzed at P < 0.05 level using Statistical Analysis Software (SPSS software, 21.0, SPSS Institute Ltd, Chicago, USA). Analysis of variance was conducted by using Duncan's multiple range tests at P < 0.05 level to test for the significance of cropping system, N-fertilizer postponed topdressing effects, and their interactions.