search
Contact Us
HOME

  • Crossref logo
  • Crossref Similarity Check logo
Journal Search Engine

to

Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 2287-5824(Print)
ISSN : 2287-5832(Online)
Journal of The Korean Society of Grassland and Forage Science Vol.46 No.3 pp.202-212
DOI : https://doi.org/10.5333/KGFS.2026.46.3.202

Effects of Sowing Date on Forage Yield and Quality of Summer-Sown Oat (Avena sativa L.) in the Central Region of Korea

Ja Hwan Ku, Sujung Kim, Keunchun Park, Jeong Ju Kim, Su Jeong Kim*
Central-Northern Region Crop Research Center, National Institute of Crop and Food Science, Rural Development Administration, Suwon 16429, Republic of Korea
*Corresponding author: Su Jeong Kim, Central-Northern Region Crop Research Center, National Institute of Crop and Food Science, Rural Development Administration, Suwon 16429, Republic of Korea. Tel: +82-31-695-4113, Fax: +82-31-695-4170, E-mail: sjkim30@korea.kr
September 1, 2026 September 18, 2026 September 21, 2026

Abstract


Late-summer sowing of forage oat (Avena sativa L.) allows use of the short autumn growing period after summer-crop harvest, but the optimal sowing date for central Korea remains uncertain. Two forage oat cultivars ('Rayeon' and 'Highspeed') were grown at four sowing dates (7, 18 and 27 August and 5 September 2025) in a randomized complete block design with four replications at Suwon, Korea. Data were analyzed by factorial analysis of variance and Tukey's honestly significant difference test, and orthogonal polynomial contrasts were used to characterize the sowing-date response. Sowing date significantly affected growth and heading traits, dry matter yield and total digestible nutrient yield, but not the concentrations of forage-quality traits. Dry matter yield peaked at the 27 August sowing (10.53 t/ha), 113% above the 7 August sowing (4.95 t/ha), and declined to 9.34 t/ha on 5 September; total digestible nutrient yield peaked on the same date (6.81 t/ha). 'Rayeon' produced more culms (635 vs. 555 culms/m2) and a higher dry matter yield (8.61 vs. 7.85 t/ha) than 'Highspeed', whereas 'Highspeed' had lower concentrations of neutral detergent fiber (48.9 vs. 52.1%) and acid detergent fiber (29.9 vs. 32.0%) and a higher total digestible nutrient concentration (65.3 vs. 63.6%); the cultivars therefore did not differ significantly in total digestible nutrient yield. Days to heading showed a significant sowing date × cultivar interaction: 'Rayeon' headed within 48–53 days at every sowing date, whereas 'Highspeed' required 54–64 days and headed on 8 November when sown on 5 September, almost coinciding with harvest. Under the conditions of this single-year experiment, dry matter and total digestible nutrient yields were highest at the 27 August sowing, and earlier-heading cultivars such as 'Rayeon' may offer greater harvest flexibility when sowing is delayed to early September.


Days to heading , Dry matter yield , Forage oat , Forage quality , Late summer sowing

초록

    Ⅰ. INTRODUCTION

    Global instability has disrupted grain supply chains, and rising compound-feed prices have increased production costs for livestock farms. Total domestic demand for forage was approximately 5.10 million tonnes in 2024 and domestic supply was 4.03 million tonnes, giving a self-sufficiency rate of about 79.0%; however, rice straw, which is of poor quality, accounted for 65.7% of the forage supplied (MAFRA, 2026). One important constraint is that the production of high-quality forage in the central and northern regions remains limited relative to regional livestock demand. Increasing the self-sufficiency of high-quality forage therefore requires the development and dissemination of forage crops and cultivation methods suited to these regions.

    Oat has a high crude protein content and is highly palatable to livestock (Thompson and Day, 1959), and it can be grown in regions with a short crop growing season or an early onset of cold weather (Burgess et al., 1972). In Korea, research on oat cultivation as a high-quality forage crop began in the 1980s (Kim et al., 2020), and cultivars suitable for summer sowing have been selected (Han et al., 2012). The National Institute of Crop and Food Science has developed and distributed 'Highspeed' as a cultivar for summer sowing and 'Rayeon' as a cultivar for spring sowing (NICS, 2025); because 'Rayeon' heads early, it may also be suited to late summer sowing, but it has not been evaluated for that purpose. Oat grows rapidly and can be incorporated into diverse cropping sequences (Song et al., 2014a;2014b). In Korea, late-summer-sown oat can utilize the autumn growing period following the harvest of summer crops such as forage maize or early-maturing rice (Han et al., 2012). The effects of sowing season and harvest date on forage yield and feed value have been evaluated in the mountainous areas of the Yeongnam region (Han et al., 2018), and a triple-cropping system comprising a winter forage crop (e.g., rye), a summer crop (forage maize or early-maturing rice), and late-summer-sown oat harvested in autumn has been established in paddy fields of the southern region to increase forage productivity (Song et al., 2014a;2014b). Compared with spring- or autumn-sown oat, late-summer-sown oat has been reported to have lower neutral detergent fiber (NDF) and acid detergent fiber (ADF) contents, a higher crude protein content and a higher proportion of total digestible nutrients (TDN), and thus superior forage quality (Han et al., 2012;Song et al., 2014a). Lactic acid bacteria inoculation (Ilavenil et al., 2015) and the use of multiple layers of silage film (Song et al., 2018) have also been examined as methods for improving oat silage quality. To our knowledge, however, no results have been reported on the forage yield and quality of late-summer-sown oat in the central region. The objectives of this study were therefore to quantify the effects of late-summer sowing date on forage yield, heading characteristics, and forage quality of oat and to identify the best-performing sowing date under the experimental conditions at Suwon.

    Ⅱ. MATERIALS AND METHODS

    1. Experimental site and design

    The experiment was conducted in 2025 at the Central-Northern Region Crop Research Center, National Institute of Crop and Food Science, Rural Development Administration, Suwon, Republic of Korea (37°16′ N, 126°59′ E). The soil was a medium-textured loam. Two forage oat cultivars bred by the National Institute of Crop and Food Science were compared. 'Rayeon' is registered for spring sowing and typically heads around 11 May when sown in spring, whereas 'Highspeed' is registered for summer sowing and typically heads around 10 October when sown in mid-August (NICS, 2025). The nominal sowing-date treatments were 7, 18 and 27 August and 5 September 2025. For the second sowing-date treatment, replication 1 was sown on 18 August and replications 2 to 4 on 19 August; days to heading were calculated from the actual sowing date of each plot. The eight treatment combinations were arranged in a randomized complete block design with four replications, giving thirty-two plots. Each plot measured 2 m by 1.8 m (3.6 m2). Seed was drilled at 150 kg/ha with a single-row hand-pushed plot seeder (model JP-1; JANG SEEDER, Cheongju, Republic of Korea) at a row spacing of 10 cm. The 10-cm row spacing was used intentionally to establish a dense and uniform stand under summer-sowing conditions and to facilitate weed suppression. Fertilizer was applied at rates of 120, 90 and 70 kg/ha of N, P2O5 and K2O, respectively. Nitrogen was split equally, half applied as basal fertilizer and half when the plants reached a height of 20 cm.

    2. Growth and yield measurements

    Culm number was counted in a 1.5 m2 quadrat per plot and expressed as culms per square meter. Plant height was measured on 10 plants per plot as the distance from the ground surface to the tip of the uppermost leaf or the panicle tip. The start of heading was recorded as the date on which panicles had emerged in approximately 3% of plants in a plot and the heading date as the date on which approximately 40% of plants had headed; the interval between the actual sowing date of each plot and its heading date was recorded as days to heading. Plots from the first sowing-date treatment were harvested on 11 November, whereas those from the remaining three treatments were harvested on 12 November, cut at a stubble height of 5 cm. Fresh weight was recorded for the whole plot, and a subsample of 1,000 g was dried at 70°C for 72 h to determine the dry matter content, from which dry matter yield was calculated and expressed as tonnes per hectare.

    3. Forage quality and free sugar analysis

    Dried samples were ground to pass a 30-mesh screen and analyzed for crude protein, NDF and ADF. Crude protein was calculated as nitrogen content multiplied by 6.25, with nitrogen determined using a combustion nitrogen analyzer (TruMac N; Leco, St. Joseph, MI, USA). ADF and NDF were determined using a fiber analyzer (ANKOM 2000; ANKOM Technology, Macedon, NY, USA) according to the method of Van Soest et al. (1991). TDN was calculated from ADF as 88.9 − 0.79 × ADF (Schroeder, 2013), and TDN yield was obtained for each plot as the product of dry matter yield and TDN concentration. Free sugars were determined by high-performance liquid chromatography (HPLC). One gram of each dried sample was mixed with 0.1 g of activated carbon, after which 10 mL of 80% ethanol was added. The mixture was vortexed and shaken at 180 rpm for 1 h at 40°C. After centrifugation at 1,000 × g for 20 min, the supernatant was filtered through a 0.45 μm membrane filter and the filtrate was stored under refrigeration until analysis. The extracts were analyzed using an HPLC system (e2695; Waters, Milford, MA, USA) fitted with a Carbohydrate ES column (5 μm; Grace, Deerfield, IL, USA) and an evaporative light scattering detector (2420 ELS Detector; Waters, Milford, MA, USA). The mobile phase was a mixture of acetonitrile and deionized distilled water (77:23, v/v), delivered isocratically at 1.0 mL/min; the column temperature was 50°C and the injection volume was 20 μL. Fructose, glucose, maltose and sucrose were identified and quantified against external standards (Sigma-Aldrich, St. Louis, MO, USA), and total sugar was calculated as the sum of the individual sugars. All quality and sugar values are expressed on a dry matter basis.

    4. Weather conditions

    Meteorological data for the experimental period were obtained from the Suwon weather station (station No. 119) of the Korea Meteorological Administration (KMA, 2025), and the thirty-year normals (1991–2020) for the same station were used for comparison (Fig. 1).

    5. Statistical analysis

    Data were analyzed using a factorial analysis of variance for a randomized complete block design, with block, sowing date, cultivar, and the sowing date × cultivar interaction included in the model, leaving 21 degrees of freedom for error. Treatment means were separated by Tukey's honestly significant difference test at a significance level of 0.05. To characterize the shape of the response to sowing date, the sowing date sum of squares was partitioned into linear, quadratic and cubic single-degree-of-freedom contrasts. Because the nominal sowing-date treatments were unequally spaced, orthogonal polynomial coefficients were computed for the treatment spacing of 0, 11, 20, and 29 days after the first nominal sowing date by Gram-Schmidt orthogonalization and scaled to a maximum absolute value of one, giving -1.000, -0.267, 0.333 and 0.933 for the linear contrast, 0.871, -1.000, -0.807 and 0.936 for the quadratic contrast, and 0.254, -0.909, 1.000 and -0.345 for the cubic contrast. Dry matter yield and TDN yield were additionally fitted to quadratic regression models using days after the first nominal sowing date as the explanatory variable, and the date corresponding to the maximum fitted value was estimated for each response. For the start of heading and the heading date, calendar dates were converted to day of year before analysis and the resulting means were converted back to calendar dates for presentation. All analyses were performed using R (version 4.6.0; R Core Team, 2026).

    Ⅲ. RESULTS AND DISCUSSION

    1. Weather conditions

    From early August to early November 2025, the mean air temperature across the ten-day periods was 21.0°C, 1.4°C above the thirty-year normal of 19.6°C. Only early August (-0.1°C) and late October (-0.4°C) fell below normal, and the largest positive anomalies occurred in late August (3.1°C), early October (3.0°C) and mid-October (2.7°C). Temperature then fell from 17.4°C in mid-October to 11.7°C in late October, a decline of 5.7°C within a single ten-day period (Fig. 1).

    From early August to early November, total precipitation was 711.5 mm, 142% of the thirty-year normal of 500.5 mm, but highly uneven in distribution. Rainfall was well below normal around sowing in early August (17.5 vs. 110.7 mm) and again in late August (55.1 vs. 103.3 mm), whereas the four consecutive ten-day periods from mid-September to mid-October received 479.3 mm, 3.9 times the normal of 121.5 mm. Conditions then became exceptionally dry, with only 2.7 mm in late October and 5.0 mm in early November. The 2025 growing season was therefore warmer than normal overall, with rainfall concentrated from mid-September to mid-October, followed by an abrupt shift to cool, dry conditions.

    2. Growth and dry matter yield

    Culm number, plant height, and dry matter yield were all significantly affected by sowing date (Table 1). Dry matter yield increased by 113%, from 4.95 t/ha at the 7 August sowing to 10.53 t/ha at the 27 August sowing, and then decreased to 9.34 t/ha at the 5 September sowing; Tukey's test separated all four sowing-date means. Culm number followed a similar pattern, increasing from 350 culms/m2 at the first sowing to 716 culms/m2 at the third, with no significant difference among the last three sowing dates. When averaged across cultivars, plant height was highest at the 27 August sowing (130 cm) and lowest at the 5 September sowing (123 cm).

    Early August was warm (ten-day mean temperature of about 27°C, close to normal), and rainfall was much lower than normal (17.5 vs. 110.7 mm; Fig. 1). Sulaiman et al. (2023) showed that oat germination is sensitive to temperature and reduced water potential. Although seed-zone water potential was not measured in the present study, the low culm number at the 7 August sowing may reflect poorer establishment under the hot, low-rainfall conditions around sowing.

    'Rayeon' produced more culms (635 vs. 555 culms/m2), had a greater overall mean plant height (128 vs. 124 cm), and yielded more dry matter (8.61 vs. 7.85 t/ha) than 'Highspeed'. In a two-year experiment in the mountainous areas of the southern region, Shin et al. (2021) evaluated summer oat sown from early August to early September and found no significant effect of sowing period on dry matter yield (11.2, 10.9, 11.2, and 8.0 t/ha for early-August, mid-August, late-August, and early-September sowing, respectively) or forage quality. In contrast, harvest date had a large effect: dry matter yield increased from 7.1 t/ha in late October to 12.5 t/ha in early November. The authors recommended sowing by about 25 August and harvesting from early November. The sowing-date response in the present experiment was larger and statistically significant. Conditions around the first sowing were markedly drier than normal, and the first sowing subsequently had the lowest culm density (Table 1), observations that are consistent with poorer establishment. However, germination rate and initial seedling density were not measured directly. In Shin et al. (2021), by contrast, the numerical reduction observed at the early-September sowing did not reach statistical significance. For still later sowing, Song et al. (2014b) reported that dry matter and TDN yield were halved when sowing was delayed beyond early September. Dry matter yield here peaked at the 27 August sowing, close to the sowing limit of about 25 August recommended for the southern region. The sowing date × cultivar interaction was not significant for culm number or dry matter yield, providing no evidence that the cultivar differences for these traits varied among sowing dates. For plant height, however, the interaction was significant (p = 0.044), indicating that the magnitude of the cultivar difference varied among sowing dates. 'Rayeon' was numerically taller than 'Highspeed' at all four sowing dates, but the observed difference ranged from 1 cm on 18 August (127 vs. 126 cm) to 9 cm on 7 August (129 vs. 120 cm; Table 1). Because no simple-effect comparisons were performed, these treatment-combination differences are interpreted descriptively.

    The smaller decline from 27 August to 5 September, relative to the increase from 7 to 27 August, is consistent with a trade-off between establishment conditions and the shortening of the remaining growing period. Culm number did not differ significantly among the last three sowing dates, whereas plant height at the 5 September sowing was the lowest among the four (Table 1), suggesting that the lower yield at the latest sowing was associated more closely with reduced growth per culm than with lower culm density. Delaying sowing from 27 August to 5 September removed nine relatively warm growing days in early September; both treatments were then exposed to the late-season decline in temperature, with ten-day means of 11.7°C in late October and 10.6°C in early November (Fig. 1). Because dry matter accumulation in oat declines at low temperatures (Smith, 1974;Mantai et al., 2017), the loss of warmer early-season growing days was unlikely to be compensated for by biomass accumulation under the cool conditions near harvest.

    The shape of the response was examined by partitioning the sowing-date sum of squares into orthogonal polynomial contrasts (Table 1). For dry matter yield, the linear and quadratic contrasts were both highly significant (p < 0.001) and together accounted for 96% of the sowing-date sum of squares. The negative quadratic contrast estimate indicated a predominantly unimodal response, whereas the much smaller but significant cubic component (p = 0.007) indicated only a modest departure from a purely quadratic pattern. TDN yield showed the same pattern (linear and quadratic p < 0.001, cubic p = 0.006), whereas culm number showed significant linear and quadratic components (p < 0.001) but no cubic component (p = 0.901). This pattern is consistent with a balance between avoiding unfavorable establishment conditions at the earliest sowing and retaining sufficient thermally productive growing time before harvest. Descriptive quadratic regressions fitted to dry matter and TDN yields estimated peak dates of approximately 29 and 30 August, respectively; both estimates were between the third and fourth sowing dates and closer to the third. Because the corresponding cubic contrasts were also significant, these dates should be interpreted as approximate summaries of the response rather than as exact optimum sowing dates. Plant height, by contrast, showed no linear component (p = 0.664) but significant quadratic (p < 0.001) and cubic (p = 0.005) components, reflecting the increase from 125 cm at the 7 August sowing to 130 cm at the 27 August sowing, followed by a decrease to 123 cm at the 5 September sowing.

    3. Heading characteristics

    Sowing date, cultivar and their interaction all significantly affected heading characteristics (Table 2). Heading was delayed with each successive delay in sowing, the mean heading date shifting from 4 October at the 7 August sowing to 31 October at the 5 September sowing. Days to heading, however, did not increase monotonically: it was longest at the 7 August sowing (58 days), shortest at the 27 August sowing (51 days) and longer again at the 5 September sowing (57 days). 'Rayeon' headed within 48 to 53 days at all sowing dates, whereas 'Highspeed' required 54 to 64 days; 'Highspeed' sown on 5 September did not head until 8 November, a few days before the harvest on 12 November, whereas 'Rayeon' sown on the same date headed on 24 October. Temperature strongly affects oat development, with cool conditions delaying heading and maturity (Smith, 1974). Oat sown in early September or later has been reported to fail to head before harvest, possibly because of insufficient photoperiodic induction (Shin et al., 2014;Song et al., 2014a). Shin et al. (2021) likewise found that summer oat sown on 7 September in the southern mountainous area did not head at all, and attributed this to a photoperiod shorter than 12 h from early in the growth period; in the present experiment the 5 September sowing did head, but in 'Highspeed' only four days before harvest. 'Rayeon' showed less delay in heading than 'Highspeed' at the later sowing dates, suggesting greater suitability for maintaining timely heading under late-summer sowing conditions.

    The convex response of days to heading cannot be explained by photoperiod alone, because the longest period to heading occurred at the 7 August sowing, when the photoperiod was longest. Mantai et al. (2017) determined the cardinal temperatures for oat development from emergence to anthesis as 4°C (minimum), 22°C (optimum) and 30°C (maximum), the rate of development declining as temperature departs from the optimum in either direction. Ten-day mean temperatures during early growth following the 7 August sowing were 27–28°C (Fig. 1), above the reported optimum of 22°C and close to the upper threshold of 30°C; development may therefore have been slower than under optimal temperatures. The low culm density at this sowing (Table 1) may also have delayed canopy closure. At the 27 August sowing, development proceeded while temperatures fell through the near-optimum range of 20 to 25°C and days to heading was accordingly shortest, whereas at the 5 September sowing it was displaced into the cooling phase of the season, so thermal accumulation declined and days to heading increased again (Peltonen-Sainio and Rajala, 2007). Averaged across cultivars, the polynomial partition supports this description: for days to heading, the quadratic contrast was highly significant (p < 0.001), whereas the linear contrast was not (p = 0.088), indicating a pronounced quadratic response with an intermediate minimum rather than a monotonic response to delayed sowing. The wider range shown by 'Highspeed' (54 to 64 days) than by 'Rayeon' (48 to 53 days) suggests that 'Highspeed' may have been more sensitive than 'Rayeon' to deviations from favorable developmental conditions.

    4. Forage quality and free sugar content

    None of the measured forage-quality concentrations was affected by sowing date, and the sowing date × cultivar interaction was not significant for any trait in Table 3. Crude protein ranged from 9.7 to 11.0% of dry matter across sowing dates and did not differ among treatments. For the concentrations of NDF, ADF, and TDN, cultivar was the only significant source of variation (all p < 0.001): 'Highspeed' showed lower NDF (48.9 vs. 52.1%) and lower ADF (29.9 vs. 32.0%) than 'Rayeon' and consequently a higher TDN concentration (65.3 vs. 63.6%). Smith (1975) reported that the yields of in vitro digestible dry matter and of total nonstructural carbohydrates in oat fall as temperature rises and increase as it falls, and Han et al. (2012) reported that spring-sown oat lignifies faster as the season warms, whereas in summer-sown oat the temperature falls toward harvest so that the accumulation of cell wall constituents is suppressed. These previous findings suggest that the cooling conditions associated with late-summer growth may limit fiber accumulation; however, sowing date did not significantly affect NDF, ADF, or TDN concentration in the present experiment.

    The coefficient of variation for crude protein (18.0%) was by far the largest among the quality traits, being 4.3–11.3 times the corresponding values for NDF (3.3%), ADF (4.2%), and TDN (1.6%) (Table 3). Leaf proportion was not measured in this experiment; therefore, the following interpretation should be regarded only as a possible explanation. Crude protein is concentrated in the leaf blades, and the leaf proportion of whole-crop oat is generally expected to decline around heading; if that applies here, plots that had headed earlier by the common harvest date might have carried a lower leaf proportion, an effect that would oppose that of a longer growing period. Given these potentially opposing influences and the relatively large residual variation, the absence of a statistically significant sowing-date effect on crude protein should be interpreted cautiously. The proposed mechanism should be tested directly by measuring leaf proportion and the crude protein concentration of the relevant plant fractions.

    Because the two cultivars ranked in opposite directions for dry matter yield and TDN concentration, digestible nutrient production was evaluated as TDN yield (Table 3). Sowing date had a highly significant effect on this trait (p < 0.001), whereas neither the cultivar effect (p = 0.067) nor the interaction was significant. TDN yield followed the same inverted U-shaped response as dry matter yield, rising from 3.17 t/ha at the 7 August sowing to 6.81 t/ha at the 27 August sowing and declining to 6.06 t/ha at the 5 September sowing. The late August sowing therefore produced 115% more digestible nutrients than the earliest sowing and 12% more than the early September sowing. Thus, the sowing-date pattern observed for dry matter yield was also evident when productivity was expressed as TDN yield. The 0.76 t/ha advantage of 'Rayeon' in dry matter yield was not carried through to TDN yield, because the higher TDN concentration of 'Highspeed' offset much of its lower dry matter production; the difference between the best and the poorest sowing date (3.64 t/ha) was about ten times the difference between the cultivars. Song et al. (2014b) reported a halving of dry matter and TDN yield when sowing was delayed beyond early September, and the smaller decline observed here may partly reflect the warmer-than-normal autumn of 2025 (Fig. 1).

    Total sugar content was affected by neither sowing date nor cultivar and averaged 19% of dry matter, ranging from 18.4 to 19.6% across sowing dates (Table 4). Among the individual free sugars, only glucose differed significantly between cultivars, being higher in 'Highspeed' than in 'Rayeon' (7.0 vs. 6.0%, p = 0.026). In this study, “total sugar” denotes the sum of fructose, glucose, maltose, and sucrose quantified by HPLC. This measured free-sugar fraction is not equivalent to WSC, which includes additional soluble carbohydrates, and the two measurements should therefore not be compared directly. Because WSC concentration, dry matter content at ensiling, and fermentation characteristics were not measured, the present data do not establish whether the harvested forage contained sufficient fermentable substrate for satisfactory ensiling without additives. This question, together with the potential effect of lactic acid bacteria inoculation (Ilavenil et al., 2015), should be evaluated directly in an ensiling experiment.

    For the cropping system that motivated this study, the practical question is whether sowing oat on 27 August, the best-performing date among those tested, can be reconciled with the harvest schedule of the preceding summer crop. In the central region, forage maize is generally cleared in the second half of August and early-maturing rice from late August to early September, so a 27 August sowing of oat is feasible after forage maize but may be difficult after early-maturing rice. Nevertheless, averaged across the two cultivars, the 5 September sowing produced a dry matter yield of 9.34 t/ha and a TDN yield of 6.06 t/ha, both approximately 89% of the corresponding yields at the 27 August sowing. These results indicate that, under conditions similar to those of the present experiment, an earlier-heading cultivar such as 'Rayeon' may provide greater harvest flexibility when sowing is delayed until early September. Conversely, the 7 August sowing produced the lowest dry matter and TDN yields in this experiment, indicating that the earliest sowing date tested did not confer a yield advantage under the 2025 conditions at Suwon.

    A further constraint concerns the crop that follows. In the central plains, the recommended sowing dates are approximately 25 September for Italian ryegrass and 15 October for rye and triticale (Ahn et al., 2019). The same study reported that, among the three sowing dates tested, sowing on 11 October produced the highest dry matter yields for both triticale (16.6 t/ha) and Italian ryegrass (10.5 t/ha); sowing on 31 October reduced the corresponding yields to 11.8 and 7.9 t/ha, respectively. Harvesting oat on 11 or 12 November would therefore generally preclude establishment of a subsequent winter forage crop within its recommended sowing window in the same field. Under this harvest schedule, late-summer-sown oat would function more realistically as an alternative to a winter forage crop within a two-crop sequence than as the third crop of the triple-cropping systems developed in the southern region (Song et al., 2014a;2014b). Where a winter forage crop must follow, oat would need to be harvested by late September to allow Italian ryegrass sowing or by about mid-October to allow rye or triticale sowing. At the 27 August sowing, even the latter schedule would place harvest around heading for 'Rayeon' but before the heading date of 'Highspeed' (Table 2). The associated yield trade-off could be substantial: Shin et al. (2021) obtained 7.1 t/ha from a late-October harvest and 12.5 t/ha from an early-November harvest. Quantifying this trade-off under central-region conditions, including whether improved forage quality at an earlier harvest compensates for reduced biomass yield, will require experiments incorporating multiple harvest dates.

    Ⅳ. CONCLUSIONS

    In this single-site, single-year experiment, sowing date significantly affected dry matter and TDN yields, whereas cultivar significantly affected dry matter yield and NDF, ADF, and TDN concentrations. Neither sowing date nor cultivar significantly affected crude protein or total sugar concentrations. Dry matter and TDN yields showed predominantly unimodal responses to sowing date; among the four dates tested, both were highest at the 27 August sowing. Descriptive quadratic regressions placed the approximate maxima near 29 and 30 August, respectively, but these estimates should not be interpreted as exact optima because the corresponding cubic contrasts were also significant.

    'Rayeon' had a higher overall dry matter yield and, when sowing was delayed to 5 September, headed earlier than 'Highspeed' while showing less variation in days to heading across sowing dates. 'Highspeed', however, had lower NDF and ADF concentrations and a higher TDN concentration. These results suggest that earlier-heading cultivars may provide greater harvest flexibility under similar conditions. The experimental year was warmer than normal and had unevenly distributed rainfall. In addition, all plots were harvested on essentially the same date (11 or 12 November), and only two cultivars were compared. Therefore, multi-year, multi-site trials incorporating multiple harvest dates and a wider range of maturity classes are needed before a regional sowing recommendation can be made.

    Ⅴ. ACKNOWLEDGEMENTS

    This study was supported by the Professional Researcher Program of the Rural Development Administration (National Institute of Crop and Food Science), Republic of Korea, in 2026, as part of the research project "Development of New Varieties of High-Quality Forage Cereal Crops (Phase 4)" (Project No. PJ017257022026).

    Figure

    KGFS-46-3-202_F1.jpg

    Ten-day mean air temperature (lines) and ten-day precipitation totals (bars) at Suwon from early August to early November 2025, together with the corresponding 30-year normals (1991–2020). E, M, and L indicate the early, middle, and late ten-day periods of each month, respectively.

    Table

    Effects of sowing date and cultivar on culm number, plant height, and dry matter yield of forage oat grown at Suwon in 2025
    Treatment-combination values are means ± SD (n = 4); sowing-date and cultivar means are based on n=8 and 16, respectively. Letters compare sowing-date and cultivar means separately (Tukey’s HSD, p < 0.05); treatment-combination means were not compared. Because S × C was significant for plant height (p = 0.044), the corresponding combination means are descriptive. Linear, quadratic, and cubic are 1-df orthogonal contrasts for S based on 0, 11, 20, and 29 d after the first sowing. SEM, standard error of a treatment-combination mean; CV, coefficient of variation from the error mean square. The second sowing occurred on 18 August in replicate 1 and 19 August in replicates 2–4.
    Effects of sowing date and cultivar on heading characteristics of forage oat grown at Suwon in 2025
    Treatment-combination values are means ± SD (n = 4); sowing-date and cultivar means are based on n = 8 and 16, respectively. Letters compare means separately within treatment combinations, sowing dates, and cultivars (Tukey’s HSD, p < 0.05). Linear, quadratic, and cubic are 1-df orthogonal contrasts for S based on 0, 11, 20, and 29 d after the first sowing. SEM, standard error of a treatment-combination mean; CV, coefficient of variation from the error mean square. Calendar dates were analyzed as day of year and reconverted for presentation; their SD and SEM values are in days. The second sowing occurred on 18 August in replicate 1 and 19 August in replicates 2–4, and days to heading were calculated from the actual sowing dates.
    Effects of sowing date and cultivar on forage quality and TDN yield of forage oat grown at Suwon in 2025
    Treatment-combination values are means ± SD (n = 4); sowing-date and cultivar means are based on n = 8 and 16, respectively. Letters compare sowing-date and cultivar means separately (Tukey’s HSD, p < 0.05); treatment-combination means were not compared. Linear, quadratic, and cubic are 1-df orthogonal contrasts for S based on 0, 11, 20, and 29 d after the first sowing. SEM, standard error of a treatment-combination mean; CV, coefficient of variation from the error mean square. CP, crude protein; NDF, neutral detergent fiber; ADF, acid detergent fiber; TDN, total digestible nutrients. TDN yield was calculated as dry matter yield × TDN concentration.
    Effects of sowing date and cultivar on total sugar and individual free-sugar concentrations of forage oat grown at Suwon in 2025
    Treatment-combination values are means ± SD (n = 4); sowing-date and cultivar means are based on n = 8 and 16, respectively. Letters compare sowing-date and cultivar means separately (Tukey’s HSD, p < 0.05); treatment-combination means were not compared. Linear, quadratic, and cubic are 1-df orthogonal contrasts for S based on 0, 11, 20, and 29 d after the first sowing. SEM, standard error of a treatment-combination mean; CV, coefficient of variation from the error mean square. Total sugar is the sum of fructose, glucose, maltose, and sucrose determined by HPLC. All concentrations are expressed on a dry matter basis.

    Reference

    1. Ahn, E.K., Jeong, E.G., Park, H.M., Jung, K.H., Hyun, U.J. and Ku, J.H. 2019. Double cropping productivity of main whole-crop silage rice and winter feed crops in the central plains of Korea. Korean Journal of Crop Science. 64(4):311-322.
    2. Burgess, P.L., Grant, E.A. and Nicholson, J.W.G. 1972. Feeding value of 'forage' oats. Canadian Journal of Animal Science. 52:448-450.
    3. Han, O.K., Park, T.I., Park, H.H., Song, T.H., Hwang, J.J., Baek, S.B., Kim, D.W. and Kwon, Y.U. 2012. Effect of seeding dates on yield and quality of various oat cultivars for year-round forage production. Journal of the Korean Society of Grassland and Forage Science. 32(3):209-220.
    4. Han, O.K., Ku, J.H., Min, H.G., Lee, H.J., Joo, Y.H., Lee, S.S., Oh, J.S., Jung, K.H. and Kim, S.C. 2018. Effect of sowing and harvest time on forage yield and feed value of spring and fall oats at Youngnam mountain area. Journal of the Korean Society of Grassland and Forage Science. 38(2):126-134.
    5. Ilavenil, S., Srigopalram, S., Park, H.S., Kim, W.H., Lee, K.D. and Choi, K.C. 2015. Beneficial effects of lactic acid bacteria inoculation on oat based silage in South Korea. Journal of the Korean Society of Grassland and Forage Science. 35(3):207-211.
    6. Kim, K.M., Kang, C.S., Kim, Y.K., Kim, K.H., Park, J.H., Yoon, Y.M., Park, H.H., Jeong, H.Y., Choi, C.H., Park, J.H., Kim, Y.J., Cheong, Y.K., Han, O.K. and Park, T.I. 2020. Past and current status, and prospect of winter cereal crops research for food and forage in Korea. Korean Journal of Breeding Science. 52(S1):73-92.
    7. Korea Meteorological Administration (KMA) 2025. Open MET data portal. (Accessed 15 July 2026).https://data.kma.go.kr
    8. Ministry of Agriculture, Food and Rural Affairs (MAFRA) 2026. Guidelines for implementing agricultural, forestry, livestock and food policy project. Ministry of Agriculture, Food and Rural Affairs. Sejong, Korea.
    9. National Institute of Crop and Food Science (NICS) 2025. 2025 major food crop varieties guide. National Institute of Crop and Food Science. Wanju, Korea.
    10. Mantai, R.D., Silva, J.A.G. da, Marolli, A., Mamann, A.T.W. de, Sawicki, S. and Krüger, C.A.M.B. 2017. Simulation of oat development cycle by photoperiod and temperature. Revista Brasileira de Engenharia Agrícola e Ambiental. 21(1):3-8.
    11. Peltonen-Sainio, P. and Rajala, A. 2007. Duration of vegetative and generative development phases in oat cultivars released since 1921. Field Crops Research. 101:72-79.
    12. R Core Team 2026. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria.https://www.R-project.org/
    13. Schroeder, J.W. 2013. Forage nutrition for ruminants. Publication AS1250 (Revised). NDSU Extension Service. Fargo, ND, USA. pp. 1-16.
    14. Shin, C.H., Ko, K.H. and Kim, J.D. 2014. Effect of different seeding dates on agronomic characteristics and productivity of Sudangrass hybrid and oat in cropping after corn for silage in Kyeongbuk. Journal of the Korean Society of Grassland and Forage Science. 34(2):81-86.
    15. Shin, S.H., Lee, H.J., Ku, J.H., Park, M.R., Rha, K.Y. and Kim, B.J. 2021. Effects of planting and harvest times on the forage yield and quality of spring and summer oats in mountainous areas of southern Korea. Korean Journal of Crop Science. 66(2):155-170.
    16. Smith, D. 1974. Influence of temperature on growth of Froker oats for forage. I. Dry matter yields and growth rates. Canadian Journal of Plant Science. 54:725-730.
    17. Smith, D. 1975. Influence of temperature on growth of Froker oats for forage. II. Concentrations and yields of chemical constituents. Canadian Journal of Plant Science. 55:897-901.
    18. Song, T.H., Park, T.I., Kang, H.J., Park, H.H., Han, O.K., Cho, S.K., Oh, Y.J., Jang, Y.W., Roh, J.H. and Park, K.G. 2014a. Forage productivity and feed value in triple cropping systems with winter forage crops-silage corn-summer oat cultivation at paddy field in southern region of Korea. Journal of the Korean Society of Grassland and Forage Science. 34(2):87-93.
    19. Song, T.H., Park, T.I., Park, H.H., Cho, S.K., Oh, Y.J., Jang, Y.W., Roh, J.H., Park, K.G. and Kang, H.J. 2014b. Study of the use of winter forage crops, early maturing rice and summer oats in triple cropping systems at paddy field in southern region. Journal of the Korean Society of Grassland and Forage Science. 34(4):227-233.
    20. Song, T.H., Park, J.H., Yoon, C., Kang, C.S., Son, J.H., Yoon, Y.M., Kim, Y.K., Cheong, Y.K. and Park, T.I. 2018. Effect of number of film layers and storage duration on the fermentation quality of whole crop oat silage. Journal of the Korean Society of Grassland and Forage Science. 38(1):1-6.
    21. Sulaiman, Ullah, S., Saud, S., Liu, K., Harrison, M.T., Hassan, S., Nawaz, T., Zeeshan, M., Nasar, J., Khan, I., Liu, H., Adnan, M., Kumar, S., Ali, M.I., Jamal, A., Zhu, M., Ali, N., Ali, S., El-Kahtany, K. and Fahad, S. 2023. Germination response of oat (Avena sativa L.) to temperature and salinity using halothermal time model. Plant Stress. 10:100263.
    22. Thompson, R.K. and Day, A.D. 1959. Spring oats for winter forage in the southwest. Agronomy Journal. 51:9-11.
    23. Van Soest, P.J., Robertson, J.B. and Lewis, B.A. 1991. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. Journal of Dairy Science. 74:3583-3597.
    Copyright © The Korean Society of Grassland and Forage Science. All rights reserved.