This study was conducted in the Lhasa Agroecosystem Research Station (91°20′E, 29°41′N, 3688 m above sea level), Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Dazi County, Lhasa City, Tibet Autonomous Region during the 2018- 2019 and 2019-2020 growing seasons. The mean annual temperature is 7.7 ℃, with the hottest month occurring in July (16.3 ℃) and the coldest month occurring in December (-1.5 ℃). The frost-free period is 120 to 130 days. Mean annual precipitation is 425 mm, more than 90% of which is distributed in the rainy season from mid-May to late September. In 2018, the average annual temperature of the study area was 7.5 ℃, the average temperature of the hottest month was 16.5 ℃, and the average temperature of the coldest month was -2.7 ℃. The average annual precipitation in the study area was 432 mm. In 2019, the average annual temperature of the study area was 7.9 ℃, the average temperature of the hottest month was 17.1 ℃, the average temperature of the coldest month was -2.4 ℃, and the average annual precipitation was 434 mm. In 2020, the average annual temperature of the study area was 16.9 ℃, the average temperature of the hottest month was 7.7 ℃, the average temperature of the coldest month was -1.4 ℃, and the average annual precipitation was 496 mm. The meteorological conditions of the two growing seasons are not significantly different from those of the multi-year growing seasons, so they represent normal years.

(1) Setting up the warming system: Free Air Temperature Increase (FATI) was mainly set up with reference to the method of Tian et al. (2011), and the infrared radiator was used to achieve the daytime, nighttime and all-day warming throughout the whole growth period (Fig. 1). The size and installation height of the infrared radiator, soil temperature monitoring, winter highland barley seeding amount and number of rows, and the size and spacing of the sample squares were consistent with those given in our previous study (Zhong et al., 2019). In addition, considering the characteristics of high wind speed, low canopy and sparseness in the study area, infrared radiators cannot directly increase the air temperature (Niu et al., 2007), thus only the soil temperature was measured during the whole growth period of winter highland barley in this experiment.

(2) Experimental treatments: In this study, winter highland barley was used as the experimental material, and the variety was Dongqing No.1. Four temperature treatments were set up: All-day warming (AW), daytime warming (DW), nighttime warming (NW), and control (CK), which are the same as indicated below. All-day warming refers to the 24 h warming every day during the whole growth period of winter highland barley from sowing to harvesting. Daytime warming was only implemented from 08:00-20:00 every day, and nighttime warming was applied every day from 20:00 to 08:00 the next day during the same period.

The diurnal variation curves of the soil (5 cm) temperature in winter highland barley during the growing season of 2019-2020 for the AW, NW, and DW treatments and the control treatment (Fig. 2) showed that the soil temperature experienced significantly different increases under the different warming treatments, but the daily change of soil temperature (5 cm) under the different warming treatments was the same as that of the conventional control area, indicating that the warming system did not change the daily change characteristics of field temperature. In addition, the three warming treatments all increased the average daily temperature of the soil (5 cm), and the AW, NW and DW treatments increased soil temperature by 3.91 ℃, 1.81 ℃ and 1.01 ℃, respectively, and the effective warming area was found to reach 4 m². The average daytime temperature, average night temperature, daily maximum temperature and daily minimum temperature all increased the most under the AW heating treatment. At the same time, the daily variation of the soil temperature decreased due to the all-day warming and the night warming, while the daily variation of the soil temperature increased due to the daytime warming, forming three different warming scenarios which can fully simulate the future climate change trend (Table 1 and Fig. 3).

(3) Plot planning: Each treatment had four repetitions; each repetitive plot area was 4 m² (2 m×2 m). Each plot was divided into four parts, which were used as observation and sampling plots, respectively.

(4) Field cultivation management: The three treatments were consistent in the seeding amount (150 kg ha^-1), seeding method (drill), seeding date (for both growing seasons on October 7th) and water and fertilizer management of the control. The basic seeding number was 4 million ha^-1.

(1) Soil temperature measurement: In this experiment, an automatic temperature recorder, HOBO, was used to record the soil temperature during the whole growth period. A probe was installed in the center of each main area, and the probe was set at a soil depth of 5 cm. The probe value was recorded every 5 min and these data were used for calculating the average value.

(2) The relevant calculation formula for the soil effective accumulated temperature is as follows:

where EAT is the effective accumulated temperature (℃); Tm is the daily average soil temperature; Tb is the biological starting temperature of winter barley, and its value is 0 ℃. When Tm<Tb, the effective accumulated temperature of the soil is recorded as 0 ℃.

(3) Phenological period: Starting from the beginning of sowing, the occurrence dates of key phenological periods (sowing, emergence, tillering, revival, jointing, heading, flowering, filling and mature) of winter highland barley were observed and recorded, following the standard criterion that more than 50% of the winter highland barley in the square had reached the corresponding phenological period.

(4) Biomass and plant height: In the different growth stages of highland barley, field plant sampling was carried out in the sampling area of each plot sample, and the section size was 0.125 m² before harvest. All plants were taken completely, the residual soil was washed with water, and then the plant samples were dried for 48 hours at a constant temperature of 65 ℃. The total biomass of highland barley and the biomass of each organ (root, stem, leaf, ear) were measured with a balance, and the allocation proportion of the biomass for each organ was calculated. Ten plants were randomly selected from each plot, and the absolute plant height was measured with a tape measure (as accurate to 0.1 cm) after straightening.

(5) Grain yield and its components: In the mature stage, 1 m² (1 m×1 m) was selected from each plot for yield measurement. After natural air drying, biological yield and grain yield were measured, and the number of effective panicles in the 1 m² plot was counted. In addition, 20 plants with uniform growth were selected from each plot for indoor seed testing at harvest time. After hand-threshing, the full grains and infertile spikelets were separated by washing, and then the number of grains per panicle, seed setting rate (100% × number of filled grains/total number of grains), 1000-grain weight, spike length, and harvest index (grain yield/biological yield) were determined.

Excel 2010 was used for data organization. Statistical software R was used for the statistical analysis of the data and drawing the figures, and the significance of the differences between the average values of each index were obtained by one-way ANOVA. The changes in the above indexes were analyzed under the conditions of warming.

The AW, NW and DW treatments had significant effects on the phenological phase of winter highland barley, all of which resulted in advancements of the phenological phase and shortening of the whole growth period (Table 2). The magnitudes of the influences can be expressed as AW>NW>DW. Through the analysis of the data for the two growing seasons, it was found that compared with the control group, the AW, NW and DW treatments advanced the emergence stage of winter barley by an average of 2.2 d, 1.1 d and 1.3 d; the tillering stage was advanced by 6.5 d, 1.8 d and 2.0 d; the revival stage was advanced by 10.5 d, 4.6 d and 2.6 d; the jointing period was on average 18.0 d, 9.6 d and 4.5 d earlier; the heading stage was advanced by 18.0 d, 8.3 d and 4.7 d; the flowering stage was advanced by 20.7 d, 9.5 d and 6.5 d; the filling stage was advanced by 22.9 d, 10.5 d and 6.9 d; and the maturity period was advanced by 12.5 d, 5.9 d and 2.5 d, respectively. The AW treatments all reached a significant level of difference (P<0.05).

All the treatments produced an effective accumulated soil temperature >0 ℃ during the phenological phase of the winter highland barley, leading to increases in the effective accumulated temperature during the whole growth period (Table 3). Compared to the CK, the AW, NW, and DW treatments increased the effective accumulated soil temperatures by 833.9 ℃, 468.8 ℃ and 205.1 ℃ during the whole growth period of winter highland barley in the two years, respectively. Among them, during the two years, the average increases of the emergence-tillering stage were 182.5 ℃, 112.4 ℃ and 87.7 ℃, respectively; for the tillering-revival period, the average increases were 360.7 ℃, 182.4 ℃ and 66.0 ℃; and for the filling-mature stage, the average significant increases were 267.9 ℃, 128.4 ℃ and 49.1 ℃, respectively (P<0.05).

In addition, the three different warming treatments mainly shortened the number of days from sowing to heading, and the effects were AW>NW>DW. The number of days from heading to full maturity even increased, and the effect of AW was the most obvious (Fig. 4). The two-year data show that times from the sowing to the tillering stage of winter barley under the AW, NW, and DW treatments were shortened by 6.5 d, 1.9 d and 2.0 d on average compared with the control, respectively; the times from tillering to heading were shortened by 11.6 d, 6.4 d and 2.7 d, respectively; and the heading to the mature stage was increased by 5.5 d, 2.4 d, 2.2 d, respectively. The AW treatments all reached a significant level of difference (P<0.05).

The three different warming scenarios all had positive effects on the total above-ground biomass of winter barley at different growth stages, but the effects varied among the different stages. The warming during the generative growth phase tended to increase the total above-ground biomass, affecting performance as AW>DW>NW, but as NW>DW> AW in the late growth stage (Fig. 5). On the 30th day of the growing season, the above-ground biomass of the AW, NW and DW treatments were 38.7%, 2.7% and 23.0% higher than that of CK, and the AW (t=-2.991, P=0.012) treatment reached a significant level in 2019-2020. On the 176th day of the growing season, the biomass of AW, NW and DW increased by 227.1%, 47.0% and 63.2%, respectively, compared to the CK. The AW (t=-8.427, P=0.000) and DW (t=-2.977, P=0.013) treatments in 2018-2019 were significantly different, as was the AW (t=-6.565, P=0.000) treatment during 2019-2020. On the 216th day of the growing season, the above-ground biomass of the AW, NW and DW treatments increased by 82.4%, 33.5% and 41.2% on average compared with CK in the two years. AW (t=-8.977, P=0.000), NW (t=-2.679, P=0.028) and DW (t=-2.913, P=0.020) all reached a significant level in 2018-2019, but did not reach a significant level in 2019-2020. On the 266th day of the growing season, the above-ground biomass of the AW, NW and DW treatments increased by 8.6%, 37.0% and 16.6% on average compared with CK in the two years. There was a significant difference between the NW and CK treatments in 2018-2019 (t=-4.339, P=0.002), but neither the AW nor DW treatments reached a significant level in the two years. The effects of the three different warming scenarios on the plant height of winter highland barley in the early growth stage were the same as the trends of the above-ground biomass. The plant height was increased, in the order of AW >DW>NW, and AW >NW>DW in the middle growth stage, but showed a decreasing trend in the late growth stage except for the NW treatment (Fig. 6).

The results of the two years showed that the three warming treatments changed the dry matter distribution of winter barley, in which the proportions of leaf and stem tended to decrease, and the proportions of spike and root tended to increase. The proportion of the spike showed the following pattern: NW>AW>DW>CK, while the root proportion was in the order of AW>NW>DW>CK (Fig. 7). Compared to the CK, the proportions of leaf under the AW, NW and DW treatments decreased by 2.3%, 2.0% and 0.7%, respectively, on average, and the proportions of stem decreased by 1.8%, 3.4% and 1.8%, respectively, without reaching the significant level among all treatments (P<0.05). In addition, the proportions of panicle under the AW, NW and DW treatments increased by 2.9%, 4.9% and 2.3%, respectively, on average, and the proportions of root increased by 1.2%, 0.4% and 0.2%, respectively, which did not reach the statistically significant level among all treatments (P<0.05).

The correlations between above-ground biomass and plant height, and the average daily soil temperature, average daytime soil temperature, average night soil temperature, daily range soil temperature and effective accumulated soil temperature of winter barley at different growth stages were further analyzed (Table 4). The results showed that the above-ground biomass of highland barley at each growth stage was positively correlated with the average daily soil temperature, average night soil temperature and effective accumulated soil temperature, and the correlations increased at first but then decreased as the growth period progressed. The above-ground biomass of winter barley during 2018-2019 was significantly positively correlated with the average daily soil temperature during 0-176 d after sowing (t=4.822, P=0.040). The positive correlations between above-ground biomass and average daily soil temperature and daily range soil temperature decreased gradually with the advancement of the growth period and showed a negative correlation at the later growth stage. There was an extremely significant positive correlation between the above-ground biomass on the 30th day after sowing (t=10.547, P=0.009) and the 0-30 d average daily soil temperature after sowing (t=10.547, P=0.026), which was also significantly positive with the 0-176 d average daytime soil temperature in 2019-2020. Plant height was positively correlated with average daily soil temperature, average daytime soil temperature and effective accumulated soil temperature during the early and mid-term growth periods, and the correlation increased as the growth period progressed, but was negatively correlated in the later growth period. In addition, the 216th-day plant height was positively and significantly correlated with daily average soil temperature in 2018-2019; and it was positively and significantly correlated with average daily soil temperature, average night soil temperature and effective accumulated soil temperature during 2019-2020. The correlations between plant height and average daytime soil temperature and the daily range were similar to that of biomass, which gradually decreased during the growth period, and showed a negative correlation at the later growth stage. Moreover, the 30th-day plant height in 2018-2019 was positively and significantly correlated with the average daytime soil temperature.

The results of the two years showed that the three different warming scenarios had different effects on the yield and yield components of winter barley. Under the conditions of warming, the yield and yield components of winter barley tended to increase in the number of effective spikes and 1000-grain weight. The grain number per spike increased, except for the AW treatment, and the harvest index also increased. Spike length and seed setting rate tended to derease, except for the NW treatment. The effect on the number of effective spikes and 1000-grain weight was in the order of AW>NW>DW>CK, and for grain number per panicle the order was NW>DW>CK>AW (Table 5). The yields of AW, NW and DW increased by 16.4%, 24.6% and 9.5% on average in the two years, respectively, and the NW treatment reached a significant level (t=-2.541, P=0.026) compared with the control in 2019-2020. The number of effective spikes increased by 22.4%, 11.3% and 9.9% on average in the two years, respectively, and there was a significant difference between the AW treatment and control in 2018-2019 (t= -2.251, P=0.044). The 1000-grain weight increased by 9.9%, 2.8% and 0.4% on average in the two years, respectively. The harvest index increased by 1.3%, 5.2% and 1.4% on average in the two years, but did not reach a significant level compared with the control. The grain number per panicle under the NW and DW treatments increased by 10.0% and 3.8% on average in the two years, respectively, and decreased under the AW treatment by 10% on average in the two years, but did not reach a significant level in the two years. The spike length and seed setting rate under the AW and DW warming treatments decreased by 3.0% and 3.3%, and by 0.8% and 0.7%, respectively, in the two years, while the NW treatment increased these two components by 5.2% and 1.8%, respectively, in the two years.

The higher the average daily soil temperature during the whole growth period, the higher the 1000-grain weight and the number of effective spikes per unit area. The number of grains per spike and yield increased at first and then decreased with the increase of temperature (Fig. 8). In the 2018-2019 growing season, when the average daily soil temperature reached 10.00 ℃, the number of effective ears was expected to reach a maximum of 3.4244 million spikes ha^-1. In the 2019-2020 growing season, when the average daily temperature was 9.82 ℃, the number of spikes reached the highest value of 3.4379 million spikes ha^-1, and the increase in average daily soil temperature may be an important reason for the increase in yield of each warming treatment. There was also a significant positive correlation between the average daily soil temperature and the 1000-grain weight of winter highland barley in the two growing seasons. In the growing season of 2019-2020, the 1000-grain weight increased by 0.96 g for every 1 ℃ increase of the average daily soil temperature. In the year with a low 1000-grain weight, the effect of the average daily soil temperature increase on the 1000-grain weight was weaker. Higher soil environmental temperatures can also increase the number of grains per panicle of winter barley. In the 2018-2019 and 2019-2020 growing seasons, when the average daily soil temperatures reached 8.54 ℃ and 8.13 ℃, the highest numbers of grains per panicle were obtained. However, if the temperature exceeds a certain limit, the temperature will have a negative effect on the number of grains per ear. In addition, the increase in the effective accumulated temperature of the soil significantly increased the yield of winter barley. An effective accumulated temperature in the growth period of 2300-2500 ℃ may be the critical threshold for the suitable planting of highland barley. A low temperature affects the growth rate of crops, and a high temperature may cause more droughts and also affect the effect of temperature on increasing production.

Increasing the temperature will accelerate the growth and development of crops, advance each phenological phase, and thus shorten the whole growth period. Many relevant studies have reached this conclusion based on either models or field experiments. For example, Liu et al. (2018) found that warming by either 4 ℃ or 2 ℃ shortened the growth period of highland barley by 7 d or 5 d, respectively, and mainly shortened the seedling stage and filling stage. During 1981-2004, the phenological characteristics of winter wheat in Northwest China were negatively correlated with climate warming. When the canopy temperature was increased by 0.9-1.7 ℃, the beginning heading stage moved forward by 7.7-10.6 d ℃^-1, and the whole growth period was shortened by 5.6-7.1 d ℃^-1 (Wang et al., 2008). In this study, the all-day, nighttime and daytime warming increased the effective accumulated temperature of the soil during the whole growth period of winter barley and advanced the phenological stage. According to the increase in the accumulated temperature, the effect was in the order of all-day>nighttime>daytime, especially during the period of vegetative growth (from sowing to heading). One possible reason is that the continuous increase in the outside temperature meets the accumulated temperature required for growth and development earlier, which shortens the wintering period of winter barley. Furthermore, considering that the Tibetan Plateau is the initiating and sensitive area of global climate change, the temperature during the day is relatively high, while the temperature is low at night, and so the temperature difference between daytime and nighttime is large. Therefore, a farmland ecosystem on the Tibetan Plateau is more sensitive to climate warming. In the vegetative growth period, the average temperature from October to April of the following year was lower, which had an obvious warming effect and accelerated the growth of the crops. In the later reproductive growth period, the temperature was higher from May to June, thus the effect of warming in the daytime was significantly less than that in the night, and the effect of warming in the whole day was the most significant.

In addition, this study also found that after entering the reproductive growth period, the number of days from heading to maturity tends to be prolonged. Increasing the temperature accelerated the emergence and jointing stage of the vegetative growth period, but did not accelerate the heading to the maturity stage of reproductive growth. The main reason was that the control area was affected by the dry hot wind, and the resulting high temperature forced the ripening period to shorten the filling period. While warming can make the highland barley mature in advance, and effectively avoid the high-temperature stress of the dry hot wind on crops in the later reproductive period, this prolongs the time of filling, which is conducive to the high yield of highland barley (Wang et al., 2011).

This study showed that the above-ground biomass and plant height of winter barley had the same response trends to different warming levels in day and night. However, the effect of warming in different growth stages was different, in that warming had positive effects on biomass accumulation and plant height in the early growth stage of winter barley. Meanwhile, in the late growth stage of winter barley, unlike in previous studies (Shi et al., 2015; Su, 2016), the effect of nighttime warming was significantly higher than those of the other treatment treatments. The reason for this difference may be related to the different environments in the study areas. Many studies have confirmed that an appropriate temperature increase can promote the growth of crops. However, when the temperature exceeds the optimal temperature for crop growth, the temperature will become a limiting factor for growth (Ma et al., 2010; Fu et al., 2018a). During the generative growth phase, the average temperature of the daytime in the study area was generally low, and it was lower than the optimum temperature for winter barley growth. Daytime temperature is the leading factor, so the three different temperature increases can promote crop growth. The phenological period was advanced and the wintering period was shortened, which was conducive to the vegetative growth of highland barley. According to the increase in temperature during the daytime, the higher the increase in temperature during the daytime, the more obvious the effect of promoting the growth of highland barley. After the late growth stage, unlike in the other research areas (Fang et al., 2010; Tian, 2011), the daytime temperature in the Tibet was generally high in May and June. Thus, the increase in daytime temperatures for the all-day and daytime treatments may have exceeded the optimum temperature for crop growth, which would inhibit the growth of winter barley, but the temperature of the nighttime warming treatment was still at a moderate level, and it remained close to the optimum temperature after warming. Besides, the decrease in the diurnal range enhanced nighttime respiration, increased carbohydrate consumption in leaves, and thus improved the leaf photosynthetic rate. Therefore, the effect of nighttime warming was higher than the other treatments (Turnbull et al., 2002; Fu et al., 2018a, 2018b). In addition, although the all-day warming treatment included nighttime warming, it still affected the normal growth of crops after warming, which may have been because the inhibitory effect of daytime warming was greater than the promoting effect of nighttime warming, so the all-day warming may have the interaction effect of nighttime warming and daytime warming, although this still needs further research.

This study also indicated that the different warming modes all decreased the proportion of leaves and stems of winter barley at the mature stage, while the proportions of roots and spikes increased. The two years of research have also shown that the increase in the proportion of spikes for the nighttime warming treatment was greater than the increases for the daytime and all-day warming. This may be because the temperature increase range at night was more suitable for the growth of Tibetan winter barley, while the temperature increase for the all-day and daytime treatments did not reach the optimal temperature. As the main organs for the movement of substances, leaves and stems are important tissues for the transport of assimilates. The decrease in this resource allocation pattern indicates that moderate warming was conducive to the re-transport of assimilates from leaves, stems and other vegetative organs to the spikes. Moreover, the proportion of dry matter allocated to the spike at the full maturity stage was significantly higher than that of the control treatment, which compensated for the yield loss caused by the small amount of stored dry matter transferred to the grain after flowering and was beneficial to the increase of the yield (Zheng et al., 2017). The increase in the root proportion can allow the crop to absorb more water and nutrients from the soil, thus promoting the accumulation and growth of plant biomass. On the one hand, soil moisture may decrease due to a temperature increase, as had been confirmed in our previous experiments (Zhong et al., 2016), making the plants grow deeper into the soil, thus promoting an increase in the proportion of root distribution in the plants (Ma et al., 2017). On the other hand, the increase in underground root distribution may be related to the resistance of plants to the dry and hot environment. Plants put more energy into roots to control water loss by reducing leaf growth, thereby maintaining the water balance of the plants and adapting to the future warming climate (Shi et al., 2010; Zhang et al., 2016). The root distribution ratio under the all-day warming treatment was higher than those under the nighttime and daytime warming treatments. The reason might be that the all-day warming treatment had a longer warming time and a larger temperature increase, which caused greater soil drought and improved the crop’s performance in resistance to that stress.

Most preceding research suggests that warming causes crops to be affected by high-temperature stress, leading to different levels of crop failure, but considering the differences in different climate backgrounds in the study areas, there have been few production phenomena reported. In addition, the model forecast analysis had great deal of uncertainty, due to the lack of field experiments combined with the model. There have been few reports on the differences in crop yield due to different daytime and nighttime warming strategies, so this needs to be further verified by field experiments (David, 2007; Dong et al., 2011).

The results of this study showed that the yield components such as 1000-grain weight, effective spike number and grain number per spike tended to increase after increasing the temperature. In addition, the different temperature increases during daytime and nighttime made the Tibetan winter barley yield increase to different degrees, which might be caused by the low temperature and insufficient heat in the study area (Yang et al., 2017; Zhong et al., 2019). The relative effects of stimulation were NW>AW>DW>CK, which indicates that the night temperature increase in this study area is more suitable for crop production. This effect may be due to the higher daytime temperature in Tibet and the large temperature difference between daytime and nighttime. Therefore, the demand for heat by the crops during the daytime was less than that during the nighttime. Meanwhile, the fact that the second growing season was wetter than the first growing season may be one of the reasons for the better growth of winter highland barley in the second year.

It is known that the crop yield level mainly depends on the strength of the source, sink and flux, and the degree of coordination between them. Opening the source (improving light and production), expanding the sink (improving storage capacity) and saving flux (controlling transportation distribution) are the fundamental ways to improve crop yield (Pan et al., 2012; Dai et al., 2020). In terms of source, the increasing temperature led to advances in the growth period of winter barley, and increases in aboveground biomass accumulation and plant height during the early growth period, which promoted the growth of leaves and leaf area. Concurrently, the greater plant height improved the crop’s ability to intercept light, which is optimized in this type of highland barley, so it was more conducive to competition. This also increased the ability of the lower leaves to capture light, which was conducive to the process of photosynthesis and ultimately improved the yield. At the later growth stage, the biomass and plant height of the nighttime warming treatment were significantly higher than those of the other treatments, which may be an important reason for its higher yield. From the aspect of the sink, the warming treatment resulted in an increase in the daily minimum temperature of the soil during the winter barley tillering-revival stage, which reduced the frost damage. The increase in effective accumulated temperature accelerated the growth process and the filling period was completed earlier, effectively avoiding the phenomenon of high temperature forcing ripeness. Therefore, the filling period was prolonged, which increased the number of effective spikes per unit area, 1000-grain weight, and grains per spike. However, the relationship between temperature and grain number per panicle was not a simple linear relationship. The decrease of grain number per panicle under the all-day warming indicated that a temperature that is too high will also produce negative effects. The “source” of photosynthate in the growing season of crops comes from the leaves, while the formation of the cotyledon is based on the stored photosynthate (the nutrients stored in the seeds) as the source, that is, the sink into the source. Therefore, the enhancement of the source increased the material base of the reservoir, increased the capacity of the sink, and improved the grain fullness. Meanwhile, the sink had an obvious feedback effect on the size and activity of the source, and the increase in the need for assimilates from the sink can lead to an increase in the transport speed of assimilates from the source to the reservoir, and finally to a yield increase. In terms of translocation, both the size and quantity of sink and source influenced the direction and quantity of flux and played a role of “pulling force”. Each warming treatment increased the harvest index to different degrees, which reflected the ability of the photosynthate from the source organs to be transported to the sink organs to some extent, namely, the fluency of the flux was improved. In this experiment, the proportion of dry matter allotment to the spike at the mature stage was significantly higher than that of control, indicating that warming promoted the transport of photosynthate to the spike at a later growth stage, which may also be one of the reasons for the increase in yield (Daniel and Mark, 2003; Li, 2009; Tian et al., 2011).