[Interview] Harvesting adzuki beans with combines instead of sickles… National Institute of Crop Science to distribute Neoul and Yeoul in 2027 after Arari and Hongmiin

Kim Dae-gyu, head of the Cheongchun Adzuki Bean Growers’ Group, who cultivates about 10 ha (30,000 pyeong) of adzuki beans in Gongju, South Chungcheong Province, picks up a sickle when harvest season comes. If it does not rain, he dries the cut adzuki beans for 3 to 4 days, then brings in a thresher to thresh them. It takes about a week to harvest 30,000 pyeong, and during that time he calls in 4 to 6 seasonal workers. The growers’ group, made up of 24 young farmers, cultivates a total area of about 23 ha (more than 70,000 pyeong).

When the same work is done with a combine, the situation changes. According to a 2015 analysis by the Rural Development Administration’s National Institute of Crop Science, harvest labor time per 10 a (1,000㎡) was 17.44 hours for manual harvesting—cutting with sickles, drying, and then threshing—and 0.86 hours for combine harvesting. In other words, harvest labor time can be reduced by 95%. However, sickles are still used in most adzuki bean fields in Korea. According to the National Institute of Crop Science, Korea’s self-sufficiency rate for domestically produced adzuki beans remains around 20%. With rural aging and rising labor costs added to this, harvest mechanization has become the first hurdle the domestic adzuki bean industry must overcome.

On the 15th, the National Institute of Crop Science held a “Field Demonstration of Mechanical Harvesting of New Adzuki Bean Varieties” in Gongju, South Chungcheong Province, and showed the variety “Hongmiin,” developed for mechanical harvesting, being harvested with a combine. The Korea Agricultural Technology Newspaper met with Researcher Ju Young-kwang and Senior Researcher Lee Seok-ki of the Central and Northern Area Crop Research Center at the National Institute of Crop Science, who prepared the demonstration, to hear how far adzuki bean mechanization has come and what remains to be done. We also heard from farmers who took part in the demonstration.
 

Manual sickle harvesting takes 17 hours per 10 a…Using a combine cuts labor by 90% and raises income by 50%

The effects of mechanization are not limited to harvesting work. According to a National Institute of Crop Science analysis (2015) included in the 2021 Rural Extension Project Implementation Guidelines, total labor time per 10 a fell by 23.06 hours, from 25.66 hours for manual hill seeding and manual harvesting to 2.60 hours for drill seeding and combine harvesting. A major factor was that the 17.44 hours spent on harvesting (11.78 hours) and threshing (5.66 hours) were replaced by 0.86 hours of combine harvesting. In the same data, labor costs per 10 a fell from 279,000 won to 28,300 won, and income rose 50%, from 1.058 million won to 1.595 million won.

The need to lower production costs is also evident in prices. According to what the researchers heard from the industry, with production recently declining, the price of adzuki beans sold from production areas to processors has risen to 15,000 to 16,000 won per kg. This is far above the 10,000 won per kg level that the industry regards as appropriate for commercialization. Competition with imports is also a challenge. Researcher Ju said, “Consumers prefer domestic products as traditional crops, and companies also say they have better taste and aroma than Chinese products,” but added, “The quality of imports is also improving, so domestic varieties must continue to be improved.” 

Non-lodging Arari, Hongda, and Hongmiin open the way for combine harvesting

The two researchers pointed to “Arari,” bred in 2011, as the starting point for adzuki bean mechanization. Until then, the main varieties grown, such as “Chungju Adzuki Bean” and local landraces, were vining types whose stems kept growing and easily lodged. They had a structure that combines could not enter. Senior researcher Lee said, “The biggest change in the varieties now being distributed is that they do not vine and maintain an upright form,” adding, “At the same time, yield per 10 a has also increased from 150 kg to about 200 kg.”

According to the National Institute of Crop Science, Arari currently accounts for about 60% of domestic adzuki bean cultivation. About 15% is Chungju Adzuki Bean, for which farmers save and use their own seed. “Hongda,” released in 2017, and “Hongmiin,” released in 2020, have been distributed through the Korea Agriculture Technology Promotion Agency (KOAT), and their share has exceeded 10%.

Each variety has different uses. Arari is dark red and has a high paste yield, making it suitable for bean paste. Hongda is a small-seeded type with a 100-seed weight of 14.1 g, suited for adzuki bean tea and whole adzuki beans. Hongmiin is a large-seeded type with a 100-seed weight of 20.5 g, and when processed it has a bright red color, making it suitable for whole adzuki beans. Explaining why Hongmiin was demonstrated at this event, Researcher Ju said, “Its grains are large and bright red, so processors have relatively high preference for it,” adding, “Rice cake shops in the Gongju area wanted a variety with excellent appearance.”

Along with variety, cultivation method determines losses. In the same implementation guideline data, when harvested with a standard combine, the cutting-unit loss rate in drill-seeded cultivation was 3.39%, about one-quarter the level in high-ridge cultivation (12.23%). The cutting height was also lower in drill seeding, at 6.9 cm, than in high ridges (10.5 cm). The method recommended by the researchers is as follows. Using a tractor-mounted drill seeder, make a wide bed 150 cm across. Sow two seeds per hill at a depth of 5 cm, with a ridge width of 40 cm and 15 cm between plants. In a 2016 demonstration in the Gangwon Provincial Agricultural Research and Extension Services test field (Chuncheon), where Arari was harvested with a multipurpose combine, the cutting and discharge loss rate in drill-seeded cultivation was 4.61%, lower than the 5.70% for high-ridge hill seeding.

Harvest timing is also important. Researcher Ju advised, “Adzuki beans should be harvested when the leaves have fallen and the stems are well dried so that the crop can be harvested without clogging,” adding, “If you go in before they are fully ripe, the combine may stop while moving, so you need to check maturity carefully.” According to the researchers, there was almost no loss at the Gongju demonstration. Kim, who watched the demonstration, said, “Hongmiin was really good,” and said he plans to continue expanding its cultivation. Regarding losses, he said, “Some did spill, but it was similar to running a thresher,” adding, “If you dry and thresh them, even if you run the thresher gently, they still bounce out and spill.” 

Pod height and simultaneous maturity remain challenges…Equipment costs and cultivation scale are also obstacles

However, according to the researchers, only large-scale farming corporations in some regions harvest directly with combines. Most farms, which cultivate small areas, only use combines to thresh adzuki beans after cutting them with sickles. Ultimately, cutting is still left to human hands.

The researchers first pointed to the remaining challenges in varieties. Researcher Ju cited “pod height,” that is, the height from the ground to the lowest pod. He explained, “The lower part of the pods must be at a certain height above the ground so they do not get caught during combine harvesting,” adding, “If it is uniformly at least 5 cm, a skilled operator can harvest, but for anyone to do it without losses, it needs to be 10 to 15 cm.” For farmers, income comes first, so rather than risk pods getting caught in the combine and reducing marketability, they choose to cut with sickles. Researcher Ju added, “Raising the podding position is still at the line-testing stage,” adding, “There is high variability in podding position, so it is not easy to make it uniform, but we expect results within a few years.”

Another challenge is “simultaneous maturity.” Adzuki beans ripen sequentially from the lower pods, so the stems and pods do not mature all at once. Researcher Ju said, “The trait of ripening at the same time is an area that has been less improved in the varieties developed so far.”

There are also barriers beyond varieties. The two researchers agreed that what blocks mechanization is structure rather than technology. The machines needed from seeding to harvesting have already been developed in forms that can be used together with soybeans. Multipurpose combines can also be used for soybeans, adzuki beans, and mung beans simply by changing the sorting sieve, and they have been distributed considerably through paddy-field soybean projects. The problem is that cultivation scale is not large enough for individual farms or cooperatives to own the equipment.

The barrier felt in the field is also equipment cost. Kim expressed his desire to use a combine to harvest adzuki beans, but said it would be difficult to make the numbers work with combines costing 150 million to 200 million won per unit. He said, “I wish there were machinery subsidies for young farmers like me, but they are too limited,” adding that as for whether he will harvest with a combine in 2027, he will decide after weighing the income and expenditure from sales of the 2026 adzuki bean crop.

Research and equipment foundations are also not sufficient. Senior researcher Lee said, “Technology development is needed for mechanization of the entire adzuki bean production process, but there are only 3 people working on adzuki bean variety development, so more research support is needed.” He emphasized, “Agricultural technology centers run many rental programs, but not many have proper equipment suited to adzuki beans,” adding, “If a system is created to handle post-harvest cleaning and drying quickly, farmers will be able to supply raw materials much more easily.” 

Neoul yields 15% more than Arari, while Yeoul has 2.5 times more resistant starch

The next candidates expected to address the remaining challenges are “Neoul” and “Yeoul,” bred in 2025.

Neoul yields 247 kg per 10 a, which is 15% more than Arari (214 kg), and has 33 pods per plant, also 22% more. Researcher Ju said, “Because bean-paste processors prefer dark red, Arari cultivation exceeds 60%,” adding, “As Arari has been cultivated for a long time and its yield and performance have declined, we developed this variety to keep the color the same while increasing yield.” Yeoul is a variety intended to replace the early-maturing Hongda. It is bright red and yields 236 kg per 10 a, which is 10% more than Arari. Researcher Ju introduced it by saying, “Yeoul focuses on functionality, with a resistant starch content of 15%, which is 2.5 times higher than Arari.”

The National Institute of Crop Science said it plans to distribute the two varieties through a new technology dissemination demonstration project in 2027, supplying 5 ha each to 3 regions centered on main production areas, and that the target regions are under discussion. Field demonstrations will also continue. Senior researcher Lee said, “In Gyeongju, famous for Hwangnam bread made with domestic adzuki beans, the agricultural technology center is very interested in linking with processing, so we are preparing a mechanical harvesting demonstration for Neoul.” 

Cultivation methods to avoid high temperatures to be established by 2027…High-temperature-adapted varieties also being developed

Developing varieties to respond to climate change is also a task. Senior researcher Lee said, “As adzuki beans suffered damage from high temperatures in 2023 and 2024, research also faced great difficulties,” adding, “Unexpected high temperatures and climate change are a burden both in research fields and for farmers.” In the short term, the researchers plan to establish cultivation methods that avoid high temperatures by adjusting sowing time by 2027. They are also analyzing at which growth stage exposure to high temperatures causes sterility. Based on this, Researcher Ju said they will select genetic resources and develop varieties with high adaptability to high temperatures step by step.

Regarding future research directions, Senior researcher Lee said, “For adzuki beans, there are still areas that need work, such as podding position, and I also want to do more research to increase yield,” adding, “We are considering ways to strengthen disease resistance by combining various research approaches, including genomic analysis, and to shorten the breeding period through generation acceleration.” He emphasized, “Since the self-sufficiency rate is only about 20%, our first goal is to develop new varieties and expand seed distribution.”