Optimization of a wheat small red bean double cropping system in South Korea

Wheat (Triticum aestivum L.) and small red bean [Vigna angularis (Willd. Ohwi & Ohashi)] are the main ingredients of walnut-shaped “Hodugwaja”. An innovative wheat small red bean double cropping system was evaluated in a rice field in the Cheonan region (Korea) to determine its effect on land use. The effects of different nitrogen levels, sowing dates, and density on growth, yield, and quality of wheat and small red bean were also investigated using selected wheat (‘Keumgang’, ‘Sooan’, and ‘Goso’) and small red bean (‘Hongeon’, ‘Chungju’, and ‘Arari’) varieties. The effect of different fertilizer treatments [N1 (50%, 6.6 kg/10a), N2 (100%, 8.8 kg/10a), and N3 (200%, 13.2 kg/10a)] were investigated for wheat, while the effect of sowing date and density were investigated for the small red beans. Our findings revealed that the best variety, sowing date, and nitrogen level combination for wheat small red bean double cropping system in Cheonan area is ‘Goso’ sown on October 26, N3 nitrogen application, and ‘Chungju’ sown on July 10 with high ridge cultivation, at a density of 60 × 15 cm. This system was the most ideal yielding 521.6 kg/10a (1000 m2) and 275 kg/10a of ‘Goso’ and ‘Chungju’, respectively. This pioneering research provides a reliable cultivation plan and theoretical basis for implementing the double cropping system of wheat small red beans in central Korea. Undeniably, this study also provides a basis for future field experiments on wheat planting patterns and small red bean fertilization.

Data collection of wheat growth and yield. The heading date was recorded when 40% of plants (per variety) per replication reached the heading based on the Agricultural Science and Technology Research Survey and Analysis Standard of the Rural Development Agency 30 . The culm and spike lengths were also recorded on the heading date. The yield per 10a (1000 m 2 ) was calculated based on yield components including, the number of ears, the thousand-grain weight, and the number of grains per ear. The unit replication area was recorded following the standards of the Rural Development Agency 28 .
The number of spikes was recorded from 10 plants from each replication block, 25 days after heading. In addition, the number of grains per spike was recorded for three spikes per replication and replicated three times.
The leaf color was also measured after nitrogen fertilizer treatment. A Chlorophyll Meter (SPAD 502, Minolta, Japan) was used to detect chlorophyll 20 days after heading from the mid-position of the leaf blade of the uppermost leaf collected between 10:00 and 11:00 am (20 blades). The leaves were dried and ground using a mortar and pestle before determining their nitrogen content in an elemental analyzer (CN elementary analyzer, Vario Max, Germany). Grains were harvested when moisture content reached 20%. A grain moisture meter (GMK-303RS, G-won, Korea) was used to measure the moisture content.
Data collection of small red bean growth and yield characteristics. Flowering date, stem length, and the number of pods were recorded from 10 plants per replication in triplicates following the Agricultural Science and Technology Analysis Standard 30 . According to the yield components per square meter, the yield per10a (1000 m 2 ) was calculated based on yield components including, the number of plants, the number of pods per plant, the number of seeds per pod, and the weight of 1000 seeds.
The flowering date was recorded when 40% of the plants in a block bloomed. The plant height was measured from the ground to the shoot spike tip, and the number of pods per plant was recorded, excluding empty pods. Ten plants in the second and third rows were selected for the survey following the Agricultural Science and Technology Survey and Analysis Standards 30 . The small red bean was harvested once the tip of the pod turned yellow, and moisture content was reduced to 14%. Besides, the weight of 100 grains was recorded when the moisture content was less than 13%.
Climate conditions during wheat small red bean double cropping cultivation. The global average temperature was highest in 2016, with Korea experiencing the hottest summer 17 Table 1. Global warming affects the flowering of many crops and global ecosystems 31 .
Meteorological data between 1981 and 2010 (Normal year) and from 2015 to 2017 (experiment period) were obtained from the Cheonan Meteorological Center ( Table 1). The minimum, maximum, and average temperatures were higher from 2015 to 2017 than from 1981 to 2010 throughout the year. Daily minimum temperatures were higher from 2015 to 2017 than in 1981-2010 (0.5-1.2 °C from November to May and not more than 0.2 °C from July to September). The precipitation in the experiment period was higher than the Normal year throughout the year, especially from October to December when wheat emerges and tillers. However, the sunshine duration in the experiment period was slightly lower than the Normal year throughout the year.
Since arable land has become limited due to the rapid urbanization in the late twentieth century, a double cropping system could achieve efficient land use. In Cheonan, wheat is planted in October and harvested in June the following year 28 due to the high temperatures in the region during winter (Table 1). Meanwhile, the small red bean is sown in June and harvested in October of the same year 32 . Therefore, the double cropping system of wheat and small red beans could succeed in central Korea, especially utilizing the rice field during reduced rice consumption.
Soil conditions during wheat small red bean double cropping cultivation. The chemical properties of the soil samples were within the optimum ranges established by Standards for fertilization of soil 33  www.nature.com/scientificreports/ for the exchangeable potassium (K + ) and pH level ( Table 2). For instance, Organic matter was 22 g/kg, phosphorous was 225 g/kg, and electric conductivity was 1.0 dS/m. However, the exchangeable K + level was 0.66 cmol/ kg, slightly higher than the acceptable range, while the pH value was slightly lower (6.0) than the optimal pH between 6.5 and 7.0 (Table 2). Therefore, the soil used was suitable for wheat and small red bean cultivation 30 .
Statistical analysis. SAS software (Version 9.2, Statistical Analysis System, SAS Institute, Cary, NC, USA) was used for data analysis. ANOVA (Analysis of Variance) was used to compare means via the least significant difference (LSD) at 1% and 5% significance levels.
Experimental statement. We confirm that the seeds in this manuscript have been authorized and licensed.All experimental methods are carried out according to the relevant guidelines of Korea Academy of Agricultural Sciences. All data generated or analysed during this study are included in this published article.

Results and discussion
Wheat growth. Late October to mid-June. The average, maximum and minimum temperatures increased by 0.8 °C, 0.7 °C, and 0.6 °C, respectively, during the wheat cultivation period between 2015 and 2017, compared with the previous years . Similarly, the IPCC report (3rd and 5th) pointed out that global warming increases the average earth temperature 34,35 . The effective accumulated temperature of the days when the daily average temperature is above 0 °C increased by 173 °C during the wheat cultivation period in 2015-2017, higher than in previous years . The monthly average temperature decreased only in October during the sowing period and increased from November (Table 1) (Fig. 1). The suitable period for wheat sowing in South Korea is between mid and late October based on the standard agricultural recommendations 28 . Wheat was sown on October 26 in the Cheonan area, considering the harvest time of the two crops. Furthermore, the average monthly meteorological factors and changing trends were investigated until June 25, the harvest period of the second year.   www.nature.com/scientificreports/ In addition, in autumn wheat, the average minimum temperature in January reached -15°C to safely overwinter for wheat 16 , which is lower than − 5.0 and − 7.9 °C in mid-to-late October 28 . The low temperature was due to increased precipitation (over 212 mm) and reduced sunshine duration (less than 0.3 h) in the experiment period. The precipitation between October and December is critical for ensuring the increased number of wheat grains per year (Fig. 1e, Table 1). The precipitation increased by 15, 48, and 41 mm, respectively, between 2015 and 2017, higher than in previous years. However, the precipitation during the harvest period (June) decreased www.nature.com/scientificreports/ by 1.6 times (35 mm). The average temperature increased from late October to November over the experiment period. Late October was a suitable planting period due to the 23 mm precipitation. The average temperature increased in June during the harvest period, while the precipitation gradually decreased, indicating that wheat can be cultivated over the winter in the Cheonan area ( Fig. 1, Table 1). Heading of three wheat varieties occurred between 24 and 27 April (Table 3), coinciding with the nationwide heading date 36 . However, since the daily average temperature in mid-February 2017 in the central region where Cheonan is located was 0.4 °C higher than in 2016, the heading dates of three varieties in 2017 were all two days earlier than in 2016. Therefore, the difference is related to the high temperatures 37 . The precipitation and sunshine duration were lower (79.7 mm and 6.7 h, respectively) in 2017 than in 2016 due to the recent global warming in South Korea 38 .
The effect of additional nitrogen fertilization on SPAD and plant N content. Photosynthesis peaked in mid-April and gradually decreased 21 days after heading, before significantly declining during the wheat growth period 28 . In this study, the effects of different fertilization treatments on the SPAD value and leaf nitrogen content 20 days after heading were investigated. The SPAD was associated with nitrogen content in wheat leaves, consistent with previous results in rice (Oryza sativa L.) 39,40 .
Therefore, Nitrogen fertilizer levels significantly affect the yield. Kim et al.indicated that the leaf color concentration should be determined before wheat cultivation for proper fertilization 39 . The leaf SPAD in each variety was highest under the N3 treatment but significantly lower under the N1 treatment relative to N3. 'Keumgang' leaves had the highest nitrogen concentration (4.41%). The yield per 10a was also highest under N3 treatment and lowest under N1 treatment, with 'Goso' having the highest yield (521.6 kg/10a) ( Table 4).  www.nature.com/scientificreports/

Effect of additional nitrogen fertilizers on the growth characteristics of wheat. The Rural
Development Administration recommends 9.4 kg of nitrogen fertilizer per 10a as a food quality standard 27 for wheat in South Korea. However, most farmers do exceed the recommended quantity. In this experiment, the growth and yield of wheat were analyzed at different nitrogen fertilizer levels.
In the central region of South Korea, nitrogen is traditionally applied once before the elongation of wheat internode to avoid delay in wheat maturity. Therefore, three different nitrogen fertilizer treatments were used according to the existing cultivation method. The wheat growth characteristics, such as culm and spike lengths, were recorded during the nitrogen experiments in a two-year wheat small red bean double cropping system in the Cheonan area ( Table 5).
The average spike lengths of 'Keumgang' , 'Sooan' and 'Goso' were 7.5, 7.4, and 8.3 cm, respectively, and did not differ significantly among the nitrogen fertilization levels (N1, N2, N3) (  11 , where spike length was not directly associated with nitrogen levels, while culm length increased with higher nitrogen fertilization levels. Although the culm and spike lengths were significantly different among the three varieties in our study, the difference was not significant among the three treatments. The differences could be due to climate and soil conditions during internode elongation ( Table 5).
The temperature and sunshine duration comparison between Cheonan and southern regions (Gwangju, Jeon-ju) from March to April during the internode growth period of wheat is shown in Table 6. The average and maximum temperatures were lower by 1.2-1.8 °C and 0.8-1.4 °C, respectively, in the Cheonan area compared with the southern area. However, the sunshine duration was similar in both regions.
The internode growth stage was consistent among reports related to meteorological factors, such as the highest temperature, average temperature, and sunshine duration between March and April 41 .
The effect of nitrogen fertilizers on wheat yield. Cook and Baten reported that nitrogen fertilization significantly increases the number of wheat ears 42 . Nitrogen fertilizer significantly increases the number of ears and yield than 1000-grain weight in barley, thus increasing the yield per unit area 43 . Besides the number of ears, other components are also associated with increased yield 44,45 . Kim et al. also reported that the number of ears and 1000 grain weight substantially affect barley yield 46 . Moreover, increasing fertilization was more effective than increasing sowing seeds in wheat 47,48 . However, the high yields of some high-yielding wheat regions are due to genotype improvement, mechanization and the application of large amounts of nitrogen fertilizer and other pesticides. This intensification level depends largely on fossil fuels and may not be sustainable 49 .
The yield-related characteristics of wheat, based on the varieties and nitrogen fertilization levels, are shown in Table 7. The average number of grains per ear of the three varieties 'Keumgang' , 'Sooan' and 'Goso' were 38.1, 37.2, and 42.3, respectively.
The number of grains per panicle was highest under N3 treatment, contrary to Kim et al. 11 . In their study, the characteristics of the variety significantly affected the number of grains per ear than the fertilizer type and levels.  50 , indicating the influence of nitrogen fertilizer levels.
Furthermore, the average 1000 grain weight of 'Keumgang' , 'Sooan' , and 'Goso' was 40.3, 40.0, and 38.9 g, respectively, exhibiting no significant difference among the varieties. This finding could be due to the influence of the unique characteristics of varieties, consistent with Kim et al. 51 . Hobbs et al.reported that nitrogen treatment significantly increases the number of ears per unit area and grains per ear than the 1000-grain weight 52 , indicating that the characteristics of the crop variety affect the weight of 1000 grains.
The average yields of 'Keumgang' , Sooan' and 'Goso' were 452.4, 487.3, and 496.7 kg per 10a, respectively. Besides, the yields were significantly different at the three nitrogen fertilizer levels (Table 7), with 'Goso' having the highest yield (521.6 kg/10a) under N3 treatment. The yields increased with higher nitrogen fertilizer application. Therefore, nitrogen fertilizer substantially influences wheat yield 42,53,54 . Ayoub et al.reported that final wheat yield increases with higher nitrogen fertilizer levels, similar to this study 55 .

Growth characteristics of small red bean. Optimum sowing time of small red bean as a second crop after
wheat. Early July to mid-October. The average, maximum and minimum temperatures increased by 0.5 °C, 0.3 °C, and 0.6 °C, respectively, during the small red bean breeding period between 2015 and 2017, higher than the normal year (1981-2010). The temperature increase in winter was greater than that in summer during the  www.nature.com/scientificreports/ red bean growth period compared with the wheat growth period, similar to the previous research result 56 . The precipitation was 485 mm higher in 2015-2017 than in the normal year. This could be attributed to the torrential rainfall of 233 mm in the Cheonan area on July 16, 2017. However, the overall precipitation and the sunshine duration increased in the experiment period. In the past three years, the sunshine duration has decreased by 0.7 h compared with the normal year (Table 1) (Fig. 2).  57 . However, the wheat has to be harvested after mid-June if small red beans are planted after wheat. Therefore, it is necessary to explore the possibility of planting in July.
Besides, delayed harvesting affects the small red bean varieties in Korea, including the 'Chungju' , since over 90% of the small red bean have an intermediate growth type 58 . However, the recently developed varieties, such as 'Hongeon' and ' Arari' , have a determinate and semi-determinate growth type 59 . Therefore, the sowing period can be extended based on the chosen variety.
The number of days from sowing to flowering in 'Chungju' was 49 days, 47 days, and 46 days, on July 1, July 10, and July 20, respectively. For ' Arari' , it took 47 days, 46 days, and 46 days, respectively (Table 8). 'Hongeon' flowered earliest (33 to 34 days) among the varieties used in this study. Furthermore, our findings revealed that the unique characteristics of the varieties influenced the number of days to the flowering stage, consistent with the previous studies 60 .
The growth temperature was suitable up to the flowering period when the three varieties were sown between July 1 and July 20 (Table 8). Besides, the number of days to flowering was also not significantly different between July and June sowing.
Climate during small red bean growing period (early July to mid-October). The average monthly meteorological elements from July 1, the first sowing date of small red bean grown after harvesting wheat, to October 21, the final harvesting date, are shown in Table 7.
The average, maximum, and minimum temperatures increased by 0.5 °C, 0.3 °C, and 0.6 °C, respectively, in the experiment period (2015-2017), higher than in the normal year (1981-2010). The temperature increase in winter was higher than in summer during the experiment period, consistent with a previous Korean study 56 . The precipitation was 484.4 mm in the experiment period was higher than in the normal year (1981-2010), with 232.7 mm of rainfall per day on the Cheonan area on July 16, 2017. However, overall precipitation increased between 2015 and 2017 except in September (2016). Sunshine duration also reduced by 0.7 h in the last three years compared with the normal year (Fig. 2, Table 1).
In the entire small red bean growth period, the cumulative temperature has been reported to be good for flowering and fruiting at 1000 °C or higher 61,62 . For the July 1, July 10, and July 20 sowing dates, 'Chungju' had cumulative temperatures from 1168.3 to 1257.1 °C, with an average of 47 days to flowering date (Table 8). However, ' Arari' had an average of 46 days. In both varieties, the cumulative temperature decreased with a delay of the sowing date.
'Hongeon' had an average of 33 days from sowing to flowering date for the three sowing dates of July 1, July 10, and July 20, with cumulative temperatures of 833.9 °C, 863.2 °C, and 899.3 °C, respectively. The cumulative temperature required for 'Hongeon' growth is also 1000 °C or higher. Therefore, the cumulative temperature does not influence growth and harvest when sowing before July 20 except for the 'Hongeon' in Cheonan.

Effect of sowing date and density on small red bean yield. Several studies on yield characteristics
based on sowing date have been reported on soybeans. Cha and Lee indicated that culm length increases in the dense planting regardless of the sowing time of the soybean when planted after harvesting barley 63 . Furthermore, the number of branches, pods per plant, and seeds per pod increases in the sparse planting plot.
Park et al. showed that the stem length increases with an increased number of plants per hill due to plant competition 64 . However, the stem length does not elongate enough to induce lodging since the cultivation period is short.
Besides, Rho et al. reported 100-seed weight, number of pods per plant, and number of seeds per pod as the yield characteristics of small red beans 12 . However, stem length and flowering date are indirect factors 65 . www.nature.com/scientificreports/ The growth and yield characteristics of the small red bean based on the sowing date and density are shown in Tables 9 and 10. The analysis was conducted for three years (2015-2017) in the wheat small red bean double cropping system in the Cheonan area, the central area of South Korea.
However, 'Chungju' , 'Hongeon' , and ' Arari' stem lengths were shortest (51.7, 39.4, and 52.0 cm, respectively), at 60 × 25 cm on July 20. The stem length increase with higher planting density could be attributed to the competition for nutrients between plants. Similarly, the narrower the interval between plants, the longer the soybean stems, from late sowing with second cultivation after wheat harvesting 66 . Other studies also reported similar results, where the culm length decreased with the delayed sowing date of the small red bean 62,67 .
The average number of pods per plant in 'Chungju' , 'Hongeon' , and ' Arari' were 34.1, 23.6, and 31.4, respectively. The number of pods was significantly different depending on the cultivation method, such as sowing dates and sowing density (Table 9), highest on July 1, and lowest on July 20. The pods were also highest at 60 × 25 cm Table 9. Growth characteristics of small red bean according to seeding date and seed spacing. NS nonsignificant. ANOVA's test; **: significant at P < 0.01. z Days from seeding to flowering. Therefore, the number of pods per plant increases with decreased sowing density. Furthermore, the number of pods per plant in 'Chungju' , 'Hongeon' , and ' Arari' were lowest (23.8, 18.7, and 21.9, respectively) at 60 × 15 cm sowed on July 20, indicating that the number of pods increases with increased branches due to the long period of vegetative growth and the wider density between individual plants. Similarly, the number of seeds per plant in soybean increases with early sowing date and broader sowing density 68 .
The average number of seeds per pod was 8.18, 8.62, and 8.57 in 'Chungju' , 'Hongeon' and ' Arari' , respectively, and showed significant difference with sowing dates (Table 10). However, sowing density did not affect the number of seeds per pod. Overall, the average number of seeds per pod was highest (8.61) on July 1 and the lowest (8.31) on July 20. The total number of seeds per pod was about 8, similar to Hong et al.indicating that the number of seeds per pod is similar regardless of the sowing date 67 . However, the average number of seeds per pod of 'Chungju' , 'Hongeon' and ' Arari' was 6.0, 6.5, and 6.8, respectively, as presented in NICS 69 , the average score of the national adaptation experiment conducted for over three years. In this study, the number of seeds per pod was about 2 seeds higher than the national average. www.nature.com/scientificreports/ The 100-grain weight was 15.1, 14.8, and 15.4 g for 'Chungju' , 'Hongan' , and ' Arari' , respectively, significantly different at various sowing dates (Table 10). The 100-grain weight was highest on July 20 and least on July 1. The weight was also highest at 60 × 15 cm and least at 60 × 25 cm. The 100-grain weight was highest in ' Arari' (16.7 g) at 60 × 15 cm on July 20 and least in 'Chungju' at 60 × 15 cm on July 1. The average 100-grain weight was 14.0 g, and it increased with delayed sowing. Similarly, Kang et al.indicated that the weight of the seeds increases with delayed sowing of the small red bean 70 .
Therefore, a high yield was obtained at an earlier sowing date and shorter sowing density. Besides, the three varieties had good yield even when sowed on July 20. For instance, 'Hongeon' sowing was delayed due to the bad weather conditions, such as rainfall after harvesting wheat, increasing the sowing period compared to 'Chungju' or ' Arari' , thus high yield.

Conclusion
The study results demonstrated that 'Goso' (wheat) and 'Chungju' (small red bean) are the best varieties for Hodugwaja production. In this study, 'Goso' and 'Chungju' had high yields, good growth characteristics for cookies, and superior characteristics of bean sediment. Our findings reveal that in a double cropping system, 'Goso' sown on October 26 with additional nitrogen fertilization at 200% of the standard (N3), and 'Chungju' sown on July 10 at a sowing density of 60 × 15 cm in a high ridge (25 cm) cultivation method, produces the highest yields. However, more studies are necessary for the standard fertilization level of nitrogen in wheat cultivation.

Data availability
All data generated or analyzed during this study are included in this published article. The original datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.