The water control experiment for maize in present study was conducted at the Jinzhou Agricultural Meteorological Experimental Station in Liaoning Province, which has a temperate monsoon continental climate with average temperature and precipitation during 1981–2010 of 9.9°C and 568 mm, respectively. The study area has a typical brown soil with a pH 6.3 and nutritional composition including soil organic matter content of 15.24 g·kg⁻¹, nitrogen of 1.04 g·kg⁻¹, phosphorus of 0.50 g·kg⁻¹ and potassium of 22.62 g·kg⁻¹. The average field capacity, the wilting point and bulk density for the top 50 cm soil layer are 22.64%, 5.64% and 1.426 g·cm⁻³, respectively.

The experiment pots with sealed bases were made of PVC pipe with diameter 40 cm and height 100 cm, and were closely fixed row by row on the aboveground with a homemade metal fence, which formed a row and column interval of 40 cm among the plant. In the autumn before the experiment year, surface soil (0–20 cm) was evenly mixed, weighed and loaded into the pots. The soil water content was measured, and dry soil weight in pots was calculated. The maize variety was ‘Xianyu 335’. Three seeds were manually planted into the pots at the soil depth of 5 cm and a strong plant was remained when the corn had its fifth leaf. This research consisted of different experiments reflecting respectively the real-time and prolonged effects of drought. The experimental treatments and their abbreviations are shown in Table 1 .

The real-time effect experiment of drought included VPWC, RPWC and QCWC, as well as their corresponding control treatments: VPCK, RPCK and QCCK. More specifically, each treatment had six replicate pots. Natural precipitation was allowed before jointing, and appropriate amounts of water were added when precipitation was insufficient to ensure normal growth of maize plants. At jointing stage, six samples were selected to determine soil moisture: the average relative moisture of the soil column was 39.1 ± 4.0% and 49.4 ± 3.8% in 2019 and 2020, respectively. Then, water was replenished according to the difference between the measured and the optimum water content, i.e. relative soil moisture of 75%. After entering jointing stage, a large mobile waterproof shelter was used to prevent natural precipitation reaching the ground (Mi et al., 2018). According to the growth stage, weather and soil water conditions, the control treatment was irrigated with appropriate water to ensure the normal growth of maize. Treatments VPWC and RPWC reduced water supply from jointing to silking stage and from tasseling to milk ripening stage, respectively, to build drought episodes. Based on RPWC treatment, QCWC was conducted by wrapping female panicles in plastic bags to limit pollination and then to inhibit grain growth, in order to analyze the response of DMP in different organs under inhibited ear growth. Specifically, the DMWs of different organs of maize for each treatment were measured at the end of VPWC, RPWC and QCWC.

The prolonged effects of drought were reflected with comparisons of DMWs and DMPs of different organs after rewatering between drought and control treatments, and the experiment was designed as follows. After the end of drought process, some treatments for VPWC, RPWC and QCWC adopted the same irrigation measures as the control treatment until maturity, and were defined as VPAWC, RPAWC, and QCAWC, respectively, with the corresponding control treatments of VPAWC and RPAWC named CKA, and that of QCAWC named QCACK. Similarly, each treatment had six replicate pots.

Tables 2 , 3 show the dates of growth periods of maize and irrigation regimes for the different treatments. In 2019, for the VPWC, water supplementation was not conducted on June 27, and was half of the amount of CKA from June 27 to July 15, and was the same as that of CKA after July 20. The DMWs for VPWC were observed on July 16. For the RPWC, water supplementation was not conducted on July 20, and was half of that of CKA from July 22 to August 14, and was the same as that of CKA after August 22. The DMWs for RPWC were observed on 15 August and harvesting was conducted on 15 September. In 2020, the water supplementation of VPWC was one-third of CKA for July 9–15, half of CKA on July 20 and consistent with CKA from July 24. The DMWs for VPWC were observed on July 21. For the RPWC, water supplementation was half of CKA from July 24 to August 11, and consistent with CKA after August 20. The DMWs for RPWC were observed on August 20 and harvesting was conducted on September 18.

The observation data each year were statistically analyzed using SPSS 20.0 software (SPSS Inc., Chicago, IL, USA, URL:(https://www.ibm.com/cn-zh/spss?lnk=flatitem)) separately. The differences among experimental treatments were calculated using Duncan’s multiple comparison test and a one-way ANOVA at the 0.05 significance levels.

Figure 2 shows the green leaf area per plant, plant height and stalk diameter for WC and CK after different treatments. Different letters represent significance level (P < 0.05) of differences of the variables between WC and CK treatments; no letter indicates insignificant difference; also applies in the other figures. The green leaf area of RPWC in 2020 was 0 because the leaves were all dry and not green. In 2019 and 2020, the leaf areas of most WC treatments were significantly smaller than those of the controls except that of VPWC in 2019.

For plant height, there were no significant differences between WC and CK for RP, RPA, QC and QCA treatments in 2019, while those for VP and VPA were significant. In 2020, RPWC resulted in a significant small reduction in plant height, and rewatering narrowed the difference at later growth stages. The QCWC significantly decreased plant height, and its effect increased in later growth stages. Both VPWC and VPAWC induced significant and sharp decreases in plant height.

In 2019, stalk diameter was insignificantly reduced after VPWC, and significantly increased after rewatering. The RPWC significantly reduced stalk diameter, but the reduction was insignificant after rewatering. There were no significant variations in stalk diameter after QCWC and following rehydration. In 2020, stalk diameter was insignificantly affected by VPWC and RPWC, and reduced significantly after QCWC, but insignificantly increased for QCAWC.

The RSRs of VPWC and VPAWC were significantly 32% and 132% higher than those of CKs in 2019, respectively, and correspondingly in 2020 were insignificantly higher by 12% and 13% ( Figure 5 ). The RSR of RPWC increased significantly by 33% and insignificantly by 22%, while those of RPAWC varied slightly and significantly increased by 14% compared with the control in 2019 and 2020, respectively. However, the RSRs of QCCK were 82% and 20% larger than those of RPCK in 2019 and 2020, respectively. The RSRs of QCACK were 125% and 112% larger than those of CKA in 2019 and 2020, respectively. Whereas, the RSRs of QCWC and QCAWC were very similar and significantly 37% smaller compared with the controls in 2019, respectively, and correspondingly insignificantly 31% and 40% less in 2020.

The DMWs of grain for different treatments in 2020 were calculated with the ratio of grain to ear DMWs from the experiment in 2019 based on small variability for these ratios, and further HIs in 2020 were obtained. Comparing HIs between WCs and CKs showed that those for VPAWC and RPWC were significantly lower than the controls in 2019 and 2020. The HIs of RPAWC in 2019 and 2020 were insignificantly and significantly lower than the controls, respectively. The reduction of HIs in descending order was VPAWC, RPWC and RPAWC for both years.

At the end of VPWC, maize was at silking stage in 2019 and 2020. The ear, leaves, stalk and roots DMPRs of VPCK in 2019 were 15%, 21%, 38% and 25%, respectively, and the leaves, stalk and roots DMPRs of VPCK in 2020 were 21%, 52% and 27%, respectively, when the ear and stalk DMWs were considered together as stalk DMW ( Figure 6 ). After VPWC, the roots and leaves DMPRs significantly increased by 21% and 33%, respectively, the ear DMPR decreased significantly by 70%, and the change of stalk DMPR was not obvious in 2019; the leaves and stalk DMPRs increased and decreased significantly by 38% and 20%, respectively, and the root DMPR increased insignificantly by 10% in 2020. We speculate that the decrease of stalk DMPR in 2020 was mainly due to the reduction of ear DMW.

At the end of RPWC, the maize was at 16 and 2 days after milk ripening in 2019 and 2020, respectively. The DMPR of each maize organ for RPCK was 4% in bracts, 44% in ear, 15% in leaves, 22% in stalk and 15% in roots in 2019, and correspondingly 6%, 45%, 11%, 24% and 14% in 2020. The DMPRs of most of maize organs were significantly changed by drought stress. Specifically, ear DMPR significantly decreased by 24% and 36% in 2019 and 2020, respectively, and leaves, stalk and roots DMPRs increased by 17%, 19% and 27% in 2019, respectively, and by 27%, 69% and 37% in 2020; the increase in leaves DMPR in 2020 was insignificant. In addition, bracts DMPR was unaltered in 2019 and increased significantly by 42% in 2020 relative to the control.

Comparing the DMPR between QCCK and RPCK treatments showed that QC significantly reduced ear DMPR in 2019, but significantly increased the bracts, stalk and roots DMPRs. In 2020, the reduction in ear DMPR of QCCK relative to RPCK was less than that in 2019, and the DMPRs of other organs increased slightly. The DMPR of ear increased slightly for QCWC compared with QCCK, that of stalk decreased insignificantly, that of roots did not change, that of bracts significantly decreased and that of leaves significantly increased in 2019. The DMPRs of bracts and ear decreased insignificantly, that of roots decreased significantly and those of leaves and stalk increased significantly in 2020.

The effect of previous drought on following maize growth can be shown by comparing the DMPRs of each organ between WC and CK at maturity. For VPAWC, the DMPR of ear decreased significantly, of stalk and roots increased significantly, and of bracts and leaves did not change significantly relative to the control in 2019 ( Figure 7 ). In 2020, the DMPR of ear decreased, those of leaves and stalk increased significantly and that of roots was almost unchanged relative to the controls. Compared with the effects of VPWC, the difference between VPAWC and CKA in leaves DMPR decreased, while those in the DMPRs of other organs increased obviously in 2019. In 2020, the differences in DMPRs for ear and stalk increased, and those of other organs were unchanged.

For RPAWC, there was no significant difference in the DMPR of maize organs between WC and CK in 2019, and the DMPRs of ear, stalk and other organs were significantly smaller, larger and unchanged relative to the controls in 2020, respectively. Compared with the effects of RPWC, in 2019, the differences in DMPRs of each organ between RPAWC and CKA were obviously reduced. In 2020, the differences in bracts and roots DMPRs obviously decreased, and were unchanged for ear, leaves and stalk DMPRs.

In terms of QCACK, compared with CKA, the DMPRs of bracts, leaves and roots increased significantly, that of stalk increased insignificantly and that of ear significantly decreased in 2019. The DMPRs of bracts and leaves did not change, of ear decreased significantly, of stalk increased insignificantly and of roots increased significantly in 2020. Compared with QCACK, the DMPRs of bracts and leaves for QCAWC had no significant change, of ear and roots significantly decreased and of stalk significantly increased in 2019. The DMPRs of bracts and roots significantly decreased, of ear decreased insignificantly and of leaves and stalks significantly increased in 2020.

Relative to difference between RPAWC and CKA, the differences in the DMPRs of ear, stalk and roots between QCAWC and QCACK increased in 2019. However, the differences in the DMPRs of bracts, leaves and roots increased in 2020. Compared with the effects of QCWC, the differences between QCAWC and QCACK in the DMPRs of ear and roots changed from being inconspicuous to significantly decreasing, that of stalk changed from decreasing to significantly increasing, and those of leaves and bracts changed from significantly increasing and decreasing, respectively, to no significant difference in 2019. The differences in the DMPRs of different organs in 2020 were invariable.

The change of morphological characteristics is the most direct manifestation of maize affected by drought (Welcker et al., 2011), which lowers the physiological function of leaves and further influences crop growth and yield (Chaves et al., 2009; Lobell et al., 2014). Droughts during different periods of maize growth affected green leaf area. As the plant growth center was transferred to the reproductive organ and the leaves gradually senesced, RPWC accelerated senescence of leaves and decreased green leaf area sharply compared to VPWC, related to decreasing photosynthesis capacity with the progress of growth (Song et al., 2018). In addition, drought affected two aspects of stalk growth: plant height and stalk diameter. Specifically, plant height was mainly influenced by drought in VP (Song et al., 2018). After rewatering, stalk growth was compensated by increasing diameter, reflecting that the plant adapted to environmental stress through varying its morphology (Lamers et al., 2020). RPWC slightly affected plant height, but significantly decreased stalk diameter ( Figure 2 ), resulting in a decrease in stalk DMW ( Figure 4 ).

Drought induces the change in the DMP pattern, which is conducive to drought resistance (Shipley and Meziane, 2002). Studies on dry matter have mostly focused on plant aboveground parts, with very limited attention to total DMW due to the difficulty in acquiring whole roots (Mccormack et al., 2015; Julia et al., 2016; Komainda et al., 2016). In this study, the drought regime was artificially manufactured based on an experiment conducted throughout the whole growth process of maize using large experimental pots to allow whole roots sampling. The aboveground and total DMWs decreased significantly due to drought in different periods, and the DMWs of the various organs decreased to different degrees, reflecting differences in response to drought. The DMWs of stalk, roots and leaves were affected by VPWC in order from large to small, and were close to or larger than the controls after rehydration, indicating the compensation effect of rewatering after drought (Zhen and Wang, 2018). The effects of RPWC in descending order were leaves, roots and stalk, and the compensative growth of each organ to rehydration was not obvious, which was related to the short duration from rehydration to maturity and the gradual senescence. Notably, there was no significant difference between QC and non-QC treatments in total DMW, indicating that QC had no effect on total maize biomass. The DMWs of roots, stalk, leaves and bracts under QCCK were significantly higher than those under non-QC treatment, especially for roots and stalk, indicating that inhibition of ear growth could increase DMWs of vegetative organs. For QCWC, the DMW of each organ was significantly lower than the control, with rehydration playing a limited role in decreasing drought influence.

The DMPs of different plant organs vary with growth stage and environment (Yin et al., 2016; Hu et al., 2022). In this study, VPWC and RPWC both increased RSR, an important indicator of crop yield (Liu et al., 2017), showing that drought promoted DMP to roots, consistent with the functional balance theory (Shipley and Meziane, 2002). The RSRs of QCCK and QCACK were larger than of RPCK and CKA, respectively ( Figure 5 ), meaning that the dry matter originally allocated to ear would be allocated to other organs, especially roots, and the response of DMP to drought would be different from non-QC treatment, that is, the RSR decreased instead of increased. The HI, also known as reproductive effort in ecology, is an important parameter to measure crop productivity in agricultural production, and is often used as a constant to estimate yield by multiplication with aboveground DMW (Maddonni, 2012), but in fact it varies with plant growth (Bonelli et al., 2016). The RPWC reduced HI in different degrees, and rehydration had a recovery effect on HI. The HI of VPAWC was smaller than that of RPAWC, meaning that VPWC had a greater aftereffect on HI than RPWC.

Further study showed that mild water stress had a limited effect on the DMP pattern of maize during VP, reflecting that DMP of maize has adaptability to some level of water stress. When VPWC was aggravated, the DMPRs of roots and leaves increased, DMP of ear was inhibited (Yu et al., 2009). For RPCK, there was also good interannual consistency in the DMP pattern, but drought reduced the DMPR of ear, and increased the DMPRs of other organs because, on one hand, drought suppressed photosynthates partitioning into ear and, on the other hand, DMRD from vegetative organ was inhibited (Dang et al., 2014). In addition, the interannual differences in ear and stalk DMPRs in response to drought were probably caused by the difference in degree and occurrence stage of drought (Wang et al., 2016), further reflecting the variation in response of allometry growth among organs to the environment (Zhang et al., 2019). The DMP for QCCK showed that the dry matter originally allocated into ear will be allocated to stalk and roots. QC also changed the effect of drought during RP on DMP pattern. Specifically, QCWC significantly increased leaf DMP, and affected the DMPs of other organs differently due to the difference of the extent of ear inhibition between years.

Deep understanding of the aftereffect of previous drought on maize growth has an important role in drought impact prediction (Luo et al., 2016). The previous drought during different growth periods had various significant aftereffects on DMP of maize after rewatering. The aftereffect of VPWC had temporal difference among maize organs and varied with the growth process in different years (Mi et al., 2017). However, rehydration alleviated the response of DMP of each organ to RPWC. Thus, the persistence of drought impact was related to the occurrence stage (Cai et al., 2020) and degree of drought (Alam et al., 2014; Cai et al., 2022), and was weaker in RP than in VP, which is also associated with the duration during rewatering to maturity. Regrettably, soil water content was not continuously measured and so the influence of drought degree could not be evaluated in this study. Furthermore, rehydration after QCWC caused remarkable reduction in DMPR of ear and roots and increased stalk DMPR. Besides, the QC intensified the aftereffect of drought in RP on the DMP. Notably, the effect of drought on DMW was inconsistent with that on DMPR for each organ (Shi et al., 2021), meaning that the two variables should be discussed separately.

The DMRD of vegetative organ is an important source of DMA of reproductive organ during late maize growth period (Weiner et al., 2009). Potential DMRD is the key parameter in the DMP process of crop models, and plays a crucial role in accurately estimating crop yield (Dang et al., 2014). Ma and Zhou (2016) set the redistribution potential of stalk, leaves and roots as 30%, 10% and 10%, respectively, while those in this study were relatively higher. The DMRD potential of stalk was significantly higher than those of leaves and roots, and the latter two were similar to each other. Under normal growth conditions, from tasseling to maturity, the contributions to ear DMA were in descending order of stalk, roots and leaves. Whereas, the DMRDs of stalk and roots decreased sharply and slightly under drought, respectively, probably due to plant senescence and leaf abscission, indicating that drought had an inhibitory effect on the DMRD of stalk and roots. Liu et al. (2006) found that the DMPRs of maize roots and leaves at maturity significantly decreased by more than 1 times relative to those in silking stage, which is consistent with results in this study.