Authors: Sowmya Koduru, Amir Sadeghpour, Alexandre-Brice Cazenave, Christian Aguilar, Joseph Oakes, Benjamin Davies, Hillary L. Mehl, Maria Balota
Categories: Article, Biomass sorghum, Economics, Harvest time, Planting date, Seeding rate, Plant development, Plant sciences, Biofuels
Source: Scientific Reports
Authors: Sowmya Koduru, Amir Sadeghpour, Alexandre-Brice Cazenave, Christian Aguilar, Joseph Oakes, Benjamin Davies, Hillary L. Mehl, Maria Balota
Seeding rate is an important factor in optimizing dry matter yield of biomass sorghum [Sorghum bicolor (L.) Moench]. It is vital to understand how seeding rate (S rate) varies across environments, planting dates, and whether optimum S rate changes by date of harvest in the Mid-Atlantic region. From 2017 to 2019, a study was conducted at Virginia Tech’s Tidewater Agricultural Research & Extension Center in Suffolk, VA, across two locations (Holland and Hare) with six seeding rates (120, 160, 200, 240, 260, and 310 thousand seeds ha^− 1^), two planting dates [early planting (EP) date as in a full season cropping system versus late planting (LP) mimicking a double-crop system] and three harvesting dates (August – October) to determine the economic optimum seeding rate of biomass sorghum. Dry matter yield, plant height, stem diameter, disease, and lodging severity were evaluated. Results showed that compared to LP, EP increased dry matter yield by 69%, plant height by 21%, and stem diameter by 15%. Dry matter yield decreased with higher S rates under EP, whereas under LP, higher S rates increased yield in 3 of the 6 environments. In contrast, plant height and stem diameter consistently decreased with increase in S rate. In general, a S rate of 120,000 ha^− 1^ was optimum under EP, and 310,000 ha^− 1^ under LP, when considering the disease and lodging resistance as well as economics. An October harvest always had greater yield than August or September. Planting biomass sorghum at these recommended S rates and harvesting in October could improve farm profitability while decreasing disease and lodging susceptibility.
The online version contains supplementary material available at 10.1038/s41598-026-41418-1.
Growing concerns about climate change and the anticipated scarcity of non-renewable energy by 2060 necessitate seeking sustainable and carbon-neutral energy alternatives such as plant-based biofuels^1–5^. Plants can mitigate rising atmospheric carbon dioxide concentrations by fixing carbon through photosynthesis, offering a promising approach to reduce industrial emissions^6–8^. Additionally, the abundance of plant biomass worldwide positions this energy source as more accessible and sustainable compared to fossil fuels. Various crops, including sugarcane (Saccharum officinarum L.), maize (Zea mays L.), and sugarbeet (Beta vulgaris L.), have been studied as bioethanol energy crops. However, producing biofuels from these crops has proven inefficient, as they compete with food production for arable land and requires substantial resources to achieve yields that meet biofuel demands. In contrast, biomass sorghum (Sorghum bicolor L.) does not directly compete with food or feed use, although its cultivation inevitably occupies land that could support other crops. Nevertheless, biomass sorghum has emerged as a promising bioethanol energy crop, due to its high biomass yield, significant sugar content, and adaptability to diverse environments and soil types^9–13^.
Despite extensive research on biomass sorghum, studies examining the effect of cultural practices on sorghum performance in the Mid-Atlantic U.S. remain limited. For instance, the impacts of seeding rate (S rate), planting date, and harvest date on dry matter yield of biomass is poorly understood and requires further investigation. The impact of S rate on biomass sorghum yield in previous studies has been inconsistent. For example, increasing the S rate improved yield at only one of three study locations in Italy^14^. In Louisiana^15^, higher S rates reduced individual plant size but increased overall biomass yield. Conversely, in Alabama there was a notable yield reduction when S rates increased from 218,000 to 393,000 ha^− 1^, while S rates from 116,000 to 291,000 ha^− 1^ had no significant impact on yield^16^. While yield is of primary concern, S rate also influences plant height and stem diameter, traits closely linked to lodging risk and overall harvest efficiency. This study also highlighted that S rates influenced plant height and stem diameter, with effects varying by location^16^. In China it was reported that increasing plant density enhanced both dry matter yield and plant height while reducing stem diameter, and it was concluded that higher S rates produce taller plants with narrower stems^17^. These contrasting results highlight that the optimum S rate varies across environments and must be determined locally to balance yield and plant growth.
In addition to influencing yield, increasing the S rate in biomass sorghum fields has significant implications for disease severity, particularly in dense canopies that maintain high relative humidity, creating favourable conditions for disease development^18–20^. Foliar anthracnose, caused by Colletotrichum sublineola (Henn.) ex Sacc. and Trotter, is one of the most prevalent diseases affecting sorghum in the Mid-Atlantic region. This pathogen can infect all foliar parts of the plant, including stalks, and cause damage at any developmental stage throughout the growing season^21^. While anthracnose is well-documented for its yield-limiting effects on grain sorghum, its impact on lignocellulosic biofuel production remains less explored^21,22^.
Seeding rate also influences lodging risk in biomass sorghum. Higher S rates result in taller plants with weak and narrower stems, increasing susceptibility to lodging in both sorghum and maize^13,17,23,24^. While factors such as wind, nitrogen application, and temperature play prominent roles in lodging, increased S rates amplify this risk by producing taller plants^25^. Additionally, the presence of anthracnose stalk rot can further compromise stem integrity, thereby indirectly elevating the likelihood of lodging. Although some agronomic recommendations advocate for high plant densities to maximize yield and land use efficiency^17^, these practices must be carefully balanced against the increased lodging risk and disease severity. Therefore, based on the lack of clear information on the effect of S rate on biomass sorghum, it is crucial to explore and identify optimum S rates and how they affect agronomic traits in various environments or soils in the Mid-Atlantic region.
Harvest date is another important factor that substantially influences the dry matter yield of biomass sorghum. Studies on bioenergy crops have reported that biomass suitability for energy conversion is closely related to harvesting date^2,15,26^. For instance, one study showed biomass increased by 42% between early and late August harvest^27^ and another study showed that sweet sorghum could be harvested from late July through early November with continuous yield increase^28^. However, while delayed harvest generally increases biomass yield, it can also heighten the risk of disease incidence, as prolonged field exposure creates favourable conditions for pathogens to infect stressed or senescent plants, compromising growth and biomass yield^22^. Additionally, as plants mature, their stems may weaken due to increased leaf senescence and reduced photosynthetic area, which depletes carbohydrates in the stalk, ultimately increasing susceptibility to lodging^28^. This not only reduces the harvestable biomass but also complicates mechanical harvesting, leading to increased losses. Thus, optimal harvest time depends on the specific end-use, highlighting the need for targeted research to determine the appropriate time of harvest.
In the Mid-Atlantic region, biomass sorghum can be planted from May (early planting [EP]) through late June or early July (late planting [LP]). This planting flexibility allows growers to incorporate biomass sorghum into their farm management systems, such as through double cropping after winter wheat (Triticum aestivum L.). Double cropping offers several benefits, including reducing the risk of crop failure by diversifying rotations and optimizing land use year-round. Additionally, maintaining continuous soil cover helps reduce soil erosion, enhance soil carbon levels, and minimize residual nitrogen losses to the environment^29,30^. This practice can improve farm profitability^31^. However, research comparing double cropping with full-season systems has produced mixed results, with studies reporting higher, similar, or lower biomass yields in double cropping systems^32–35^. Early planting of sorghum was previously reported to significantly increase yield, while seeding rate and its interaction with planting date had no significant effect^36^. To the best of our knowledge, there is no study in the literature specific to the Mid-Atlantic region that evaluates a three-way interaction of planting date, S rate, and harvest timing. Therefore, research is needed to better understand the performance of these integrated practices and the role of S rate in optimizing biomass sorghum production. The overarching objective of this study was to determine the optimum S rate for biomass sorghum under EP and LP considering economics of crop production and seed prices. Specific objectives were to (i) assess the effect of S rate and planting date (EP versus LP) on biomass sorghum dry matter yield, plant height and stem diameter when grown in the Mid-Atlantic; (ii) assess disease and lodging severity of plants under different S rates; (iii) determine the appropriate date of harvest for environments in 2017 and 2019 and (iv) identify relationships among factors impacting biomass sorghum yield and most economical seeding rate.
Field experiments were conducted at the Virginia Polytechnic Institute and State University’s Tidewater Agricultural Research and Extension Center (TAREC) in Suffolk, VA (36° 41’ 06’’ N, 76° 26’ 01’’ W) from 2017 to 2019. Each year replicated plots were established at two locations designated by adjacent road names, Holland and Hare. These locations were chosen for their differences in soil type, precipitation, and wind exposure (Holland is located near a tree line). The Holland site was characterized by loamy soils (Eunola loamy fine sand, fine-loamy, siliceous, semiactive, thermic Typic Paleaquults), while the Hare site had loamy sand soils (Suffolk loamy sand, fine-loamy, siliceous, semiactive, thermic Typic Hapludults; and Eunola loamy fine sand, fine-loamy, siliceous, semiactive, thermic Aquic Hapludults). These two locations, along with the three-year duration of the study, created a total of six distinct environmental conditions.
Data were collected each year from a weather station located within 5 km of both the Hare and Holland sites (Fig. 1, Table S1). The data included average air temperature and monthly precipitation. For each environment, cumulative in-season growing degree-days (GDD) were calculated using a base temperature of 10°C^37^. We calculated a crop stress index (CSI) to incorporate precipitation and GDD into one simple unit representing environment as suggested by^38^ using the following
\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\mathrm{CSI}},=,{{\mathrm{D}}_{10}}/{\text{ }}\left( {{\mathrm{P}},+,1} \right)
where D~10~ is GDD summed over a month-long period, and P is the amount of precipitation (mm) accumulated during the same month. ### Experimental design The experiment utilized a randomized complete block design with four replicates at each field location. Plots were 6 m long and 3.6 m wide, allowing for four biomass sorghum rows, each spaced 91 cm apart. The number of seeds per kg was determined using the test weight of sorghum (25 g), and S rate treatments were 120, 160, 200, 240, 260, and 310 thousand seeds ha^− 1^. The seed was certified as a seed with at least 85% germinability, each year and no adjustments were made with respect to S rates. Planting date treatments were (i) early planting (EP) in May/June, representing full-season biomass sorghum, and (ii) late planting (LP) in July, mimicking a double cropping system. For EP, planting at the Holland site occurred on 9 June 2017, 8 June 2018, and 24 May 2019; and at the Hare site on 9 June 2017, 7 June 2018, and 23 May 2019. The LP planting dates at both Holland and Hare were 5 July 2017, 2 July 2018, and 29 June 2019. For the EP treatments at four environments (excluding 2018 environments due to lack of enough harvesting dates), we also evaluated three harvesting dates (August, September, and October), thereby creating an environment × S rate × harvesting date study. NexSteppe biomass sorghum hybrid NS212 was used, based on its good dry matter yield in our preliminary trials. Field preparation involved disking to a depth of 25 cm, disk ripping to a depth of 50 cm, and field cultivating to a depth of 25 cm in the spring before planting. Weeds were controlled using pre- and post-emergence herbicides. Each year, soil samples were collected and analyzed at the Virginia Tech Soil Testing Lab for corrections of pH and N-P-K based on Virginia Cooperative Extension guidelines for a sorghum crop^39^. Soil pH was adjusted using agricultural lime applied in the end of March or early April, and N-P-K using a dry bulk blend broadcasted first, then pre-plant incorporated. Supplemental nitrogen was also applied each year at the V6 or V7 growth stages using liquid UAN ammonium sulphate 24-0-0–2. ### Dry matter yield, plant height, and stem diameter Dry matter yield, plant height, and stem diameter were collected in each environment. As explained above, for the sites sampled in 2017 and 2019, dry matter yield was measured three times during the growing season (August, September and October). For each S rate treatment, two-30 cm Sect. (0.54 m²) of plants were harvested from two inner-rows of each plot by cutting at the stem base. The maximum stem length of each plant was recorded using a rolling tape, measuring from the base to the highest point of the head, and the average was calculated for each plot to determine plant height. Stem diameter was measured 2.5 cm above the base, and the average for each plot was recorded. All plants harvested from the plots were combined and weighed in the field to obtain the fresh weight. The plants were then shredded using a garden branch chopper (3-inch chipper shredder, Dirty Hand Tools, Colorado, USA). To determine moisture content, 1-kg subsamples were taken from each plot and dried in a forced-air dryer at 60 °C until constant mass. The dry matter yield was subsequently calculated for each plot. ### Disease and lodging severity assessments Foliar anthracnose and lodging severity assessments were conducted during the four weeks preceding biomass harvest (approximately 82–141 days after planting, DAP). Foliar anthracnose severity was visually evaluated on a per-plot basis, with the percentage of leaf area affected by the disease recorded^21^. In 2017, an ordinal scale was used to rate the disease severity, where 0 = no disease, 1 = trace amounts, 2 = 1–10%, 3 = 11–25%, 4 = 26–50%, 5 = 51–75%, 6 = 76–90%, and 7 = 91–100% of the leaf area showing anthracnose lesions. To more precisely assess differences in disease severity in 2018 and 2019, the proportion of total leaf area with lesions was visually estimated. In order to compare proportion data to 2017 ratings, it was necessary to convert ordinal data to percentage values using the previously described mid-point method^40^. Though this introduces some bias for the 2017 disease assessments compared to the other two years, relative treatment differences can still be assessed. Lodging severity was visually estimated as the percentage of plants with stalks bent at an angle of 30° or greater^41^. ### Economic analysis A sensitivity analysis was conducted to assess the economic impact of price fluctuations in feedstock, seed cost, and dry matter yield on the profitability of biomass sorghum production. Feedstock prices ranging from $50 to $90 Mg^− 1^, and S rates in kg ha^− 1^ were used with sorghum seed price of $10 kg^− 1^. Gross income was calculated by multiplying dry matter yield (Mg ha^− 1^) by the respective feedstock price ($ Mg^− 1^), while net income ($ ha^− 1^) was derived by subtracting seed cost ($ kg^− 1^) from the gross income ($ ha^− 1^). This analysis was conducted for each S rate across six environments for both EP and LP scenarios to identify the most economically viable S rate. Additionally, a separate analysis was performed for each harvesting date at both locations in 2017 and 2019, where income ($ ha^− 1^), cumulative $ loss, and % loss was calculated with a fixed feedstock price of $70 Mg^− 1^. ### Statistical analysis Data were tested for normality using the Shapiro-Wilk test. Disease and lodging severity values were converted to proportions and subjected to the variance-stabilizing angular transformation (arcsine square root transformation) used for data following a binomial distribution^42^. Dry matter yield was log-transformed due to non-normal residuals, while other parameters did not require transformation as their residuals were normally distributed. Analyses of variance (ANOVA) was performed using PROC Mixed in SAS^43^ to explore the fixed effects of the six S rates, two planting dates, six environments, and their interactions on the dry matter yield, plant height, stem diameter, foliar anthracnose, and lodging severity. For years with repeated harvests (2017 and 2019) under EP, dry matter yield data from the final harvest date were used in the analysis. Block was treated as a random effect. Additionally, a separate ANOVA was performed on dry matter yield using Mixed models in SAS to examine the effects of date of harvest and S rate for the environments in 2017 and 2019. Since samples were repeatedly harvested over the season, date of harvest was considered as a repeated measure, and an autoregressive covariance was used (AR1). Block and treatments nested within block (indicating plots) were assigned as random effects. The fixed effects in the model were date of harvest, environment, planting date and S rates and their interactions. When main effects or their interactions were significant, Tukey mean separation was used to evaluate significance level at α *=* 0.05. For regression analysis, the linear and quadratic model fit was evaluated using R² (the higher the better), RMSE (the lower the better) and *P* values (*P* < 0.05) in JMP Pro 18.0.2 software^43^. Agronomic and disease related parameters thought to be influenced by weather conditions were visualized using a principal component analysis in JMP Pro 18.0.2 software^43^. Principal component analysis biplot was analysed on correlations and two principal components were chosen based on eigenvalues and the scree plot outputs (Fig. S1). Data dry matter yield (DMY); plant height (PHT); stem diameter (SD); disease severity (DS); lodging severity (LS); cumulative rainfall (C~RF~); cumulative growing degree days (C~GDD~); cumulative crop stress index (C~CSI~) for both EP and LP. The cumulative indices (C~RF~, C~GDD~, C~CSI~) were calculated as sum of monthly values across the respective growing season (May-October for EP and July-November for LP). In addition, June rainfall (June~RF~); August rainfall (Aug~RF~); October rainfall (Oct~RF~); June growing degree days (June~GDD~); June crop stress index (June~CSI~), and June temperature (June~T~) were considered for EP. For LP, July rainfall (July~RF~); September rainfall (Sep~RF~); November rainfall (Nov~RF~); July growing degree days (July~GDD~); July crop stress index (July~CSI~), and July temperature (JulyT) were considered. Cluster analysis was performed to group the variables. ## Results ### Weather conditions Weather variables likely to affect sorghum growth and stress levels varied among years of the study (Fig. 1, Table S1). In 2018, the highest rainfall was recorded, with 768 mm for EP and 654 mm for LP, corresponding to lower CSI values of 3 for EP and 2 for LP. In contrast, 2017 had the lowest rainfall, with 636 mm for EP and 454 mm for LP, leading to the highest CSI values of 5 for EP and 4 for LP. Meanwhile, 2019 exhibited intermediate rainfall levels of 731 mm for EP and 579 mm for LP, with a consistent CSI of 4 for both planting dates. During the growing season (May-October for EP and July-November for LP), temperature and GDD remained relatively stable across the three years (2017, 2018 and 2019), ranging from 20 °C to 24 °C and from 1611 to 2097, respectively. These findings suggest that rainfall played a significant role in influencing sorghum growth and stress levels, particularly in 2018 and 2019 (Fig. 1). Fig. 1Monthly rainfall and GDD from planting to harvest of early (EP, May-October) and late (LP, July-November) plantings of biomass sorghum hybrids at Hare and Holland in Suffolk, VA, in 2017 through 2019. Bars represent rainfall and lines represent GDD. ^†^GDD is growth degree days calculated from max and min daily temperature and using 10 °C as base temperature. ### Agronomic traits Sorghum dry matter yield was significantly influenced by environment (E) (*P* ≤ 0.0001), planting date (P) (*P* ≤ 0.0001), E × P (*P* = 0.0002), P × S rate (S) (*P* = 0.0074), and E × P × S (*P* = 0.0049) (Table 1; Fig. 2 A; Tables 2 and 3). In 2017, the response of sorghum dry matter yield to S rate at each planting date was best explained by a quadratic model, except at the Hare site with LP where no significant relationship was observed (Table 3). In 2017, at the Holland site, the S rate of 120,000 ha^− 1^ yielded 33% higher dry matter compared to 240,000 ha^− 1^ with EP. Under LP, the S rate of 120,000 ha^− 1^ yielded 36%, higher dry matter compared to 260,000 ha^− 1^. In 2017 at the Hare site, S rates of 120,000ha^− 1^ yielded 62%, higher dry matter compared to 310,000 ha^− 1^ with EP. There was no significant relationship to explain the response of dry matter yield to S rate with LP. The highest dry matter yield (23.4 Mg ha^− 1^) was observed with a S rate of 160,000 ha^− 1^ while the lowest (16.9 Mg ha^− 1^) was observed with 260,000 ha^− 1^ (Table 3). Table 1Effect of environment, planting date, and seeding (S) rate on dry matter, plant height, stem diameter, disease severity and lodging severity of biomass sorghum.Source of variationDry matter yield(Mg ha^− 1^)Plant height(cm)Stem diameter(mm)DiseaseSeverity(%)Lodging severity (%)Environment2017-Holland21.0 e437 c18.0 c52 d40 c2017-Hare23.6 b436 c19.6 b38 e17 d2018-Holland17.6 f410 d17.5 d62 b86 a2018-Hare22.7 c469 b16.7 e61 b20 d2019-Holland31.9 a519 a19.7 a54 c11 e2019-Hare21.0 d465 b19.6 a76 a60 bPlanting Date†EP28.8 a499 a19.7 a59 a39 aLP17.1 b413 b17.2 b55 b39 aS rate120,00024.3 a479 a21.3 a56 c34 c160,00023.2 c472 b19.5 b55 c30 c200,00024.0 b463 c18.9 c56 c39 b240,00022.6 d446 d17.3 d57 b44 a260,00022.3 e441 d16.8 e58 b45 a310,00021.4 f435 e16.9 e60 a42 aTrendL/QLLQQR^2^0.82/0.840.960.900.960.69†EP: early planting; LP: late planting; L: linear; Q: quadratic; same letters indicate no treatment differences at *P* ≤ 0.05. In 2018 at the Holland and Hare sites, there were no significant relationship to explain the response of dry matter yield to S rate with EP, however, under LP the quadratic model provided the best fit at both the sites (Table 3). At the Holland site, a S rate of 200,000 ha^− 1^ produced the highest dry matter yield of 34.5 Mg ha^− 1^ while the lowest (15.6 Mg ha^− 1^) was recorded at a S rate of 310,000 ha^− 1^ with EP. In contrast under LP, the S rate of 310,000 ha^− 1^ yielded 72% higher dry matter compared to 120,000 ha^− 1^. At the Hare site, the highest dry matter yield (35.6 Mg ha^− 1^) was observed with S rate of 200,000 ha^− 1^ while the lowest (24.0 Mg ha^− 1^) was recorded at a S rate of 310,000 ha^− 1^ with EP. Under LP, the S rate of 310,000ha^− 1^ produced 79% higher sorghum dry matter yield compared to 120,000 ha^− 1^. In 2019 across all the environments, the response of sorghum dry matter yield to S rate at each planting date was explained by a quadratic model whereas at the Hare site with EP linear model also provided the best fit (Table 3). At the Holland and Hare site with EP, a S rate of 120,000 ha^− 1^ produced the highest dry matter yields (51.3 and 25.9 Mg ha^− 1^, respectively), whereas the lowest dry matter yields were recorded at a S rate of 310,000 ha^− 1^ (36.8 and 20.1 Mg ha^− 1^, respectively). However, with LP at Holland and Hare, S rates of 200,000 and 120,000 ha^− 1^ produced the highest dry matter yields (25.8 and 19.7 Mg ha^− 1^, respectively), while the lowest dry matter yields were recorded for S rates of 310,000 and 240,000 ha^− 1^ (19.5 and 14.8 Mg ha^− 1^, respectively). Table 2Interaction of planting date and seeding (S) rate on dry matter yield (Mg ha^− 1^) of biomass sorghum.S ratePlanting DateEPLP120,00031.7 a17.3 c160,00028.9 c17.4 b200,00031.4 b16.6 e240,00028.0 e17.2 d260,00028.2 d16.4 f310,00025.1 f17.7 aTrendLnsR^2^0.72-L: linear; not significant at *P* ≤ 0.05; same letters indicate no treatment differences at *P* ≤ 0.05. A separate analysis of the environments in 2017 and 2019 (four in total) under EP revealed that sorghum dry matter yield was significantly influenced by date of harvest (D) (*P* ≤ 0.0001), E (*P* ≤ 0.0001), S (*P* = 0.0337), D × E (*P* ≤ 0.0001) (Figs. 3 and 4). A quadratic model provided the best fit for the response of dry matter yield to S rate. Dry matter yield significantly decreased with increase in S rate from 120,000 to 310,000 ha^− 1^. The S rates of 120,000 and 160,000 ha^− 1^ produced similar dry matter yields. The S rates of 310,000 and 260,000 ha^− 1^ produced lower dry matter yield of 24.9 and 24.5 Mg ha^− 1^, respectively (Fig. 3). The D × E interaction showed that across all environments, delaying harvest from August to September and then to October significantly increased sorghum dry matter yield (Fig. 4). The Holland site in 2019 with September and October harvest produced the highest dry matter yield. The lowest dry matter yield was observed with August harvest in 2017 at the Holland site. Delaying the harvest from August to October increased the dry matter yield by 76%. October harvest always yielded higher than the August harvest suggesting that to maximize yield, harvest should be delayed until October. Table 3Interaction of environment, planting date and seeding (S) rate on dry matter yield (Mg ha^− 1^) of biomass sorghum using data from six environments.201720182019HollandHareHollandHareHollandHareS rateEPLPEPLPEPLPEPLPEPLPEPLP120,00031.5 a17.7 a33.2 a21.8 b24.9 b8.4 f29.5 b13.1 f51.3 a23.0 b25.9 a19.7 a160,00028.5 b16.7 b29.0 b23.4 a20.6 e9.9 e25.0 e17.3 c42.2 b21.9 e25.5 b15.1 d200,00026.4 d13.3 d25.3 e19.3 e34.5 a10.0 d35.6 a16.2 d40.4 c25.8 a22.9 c15.0 e240,00023.7 f15.6 c26.4 d21.1 c22.0 d12.5 c28.4 c16.1 e39.0 d22.8 c22.9 c14.8 f260,00025.2 e13.0 f27.2 c16.9 f24.9 c13.1 b25.5 d17.7 b40.9 e22.1 d21.5 d15.6 c310,00027.2 c13.1 e20.4 f19.5 d15.6 f14.5 a24.0 f23.5 a36.8 f19.5 f20.1 e16.0 bTrendQQQnsnsQnsQQQQ/LQR^2^0.930.710.80--0.96-0.780.850.650.950.81Q: quadratic; L: linear; not significant at *P* ≤ 0.05; same letters indicate no treatment differences at *P* ≤ 0.05. Fig. 2Interaction of environment and planting date on mean **(A)** dry matter yield, **(B)** plant height, and **(C)** stem diameter of biomass sorghum. EP: early planting; LP: late planting. Same letters indicate no treatment differences at *P* ≤ 0.05. Bars represent standard error. Fig. 3Effect of seeding (S) rate on dry matter yield (Mg ha^− 1^) of biomass sorghum using data from four environments (Holland and Hare in 2017 and 2019; 2018 environments excluded). Same letters indicate no treatment differences at *P* ≤ 0.05. Sorghum plant height was significantly influenced by E (*P* ≤ 0.0001), P (*P* ≤ 0.0001), S (*P* ≤ 0.0001), E × P (*P* ≤ 0.0001) and E × S (*P* = 0.0001) (Table 1; Fig. 2B; Table 4). The response of plant height to S rate was best described by a linear model, where plant height decreased by 3 cm for every 10,000 increase in S rate (Table 1). The E × P interaction showed that across all the environments, Holland site had higher plant height compared to Hare, and EP produced higher plant height compared to LP. The Holland and Hare environments with EP produced taller plants compared to LP in 2017 (17 and 12%, respectively), 2018 (39 and 26%, respectively), and 2019 (20 and 14%, respectively) (Fig. 2B). The E × S interaction showed that at the Holland site in 2017, the S rates of 160,000; 260,000; 120,000 and 200,000 ha^− 1^ produced similar plant height followed by S rate of 310,000 and 240,000 ha^− 1^. At the Hare site, S rates of 120,000 and 160,000 ha^− 1^ produced taller plants, while the shortest plants were observed with S rate of 260,000 ha^− 1^ (Table 4). In 2018, at the Holland site, both quadratic and linear models provided the best fit for the response of plant height to S rate, where plant height decreased by 18 cm for every 10,000 increase in S rate. At the Hare site, a linear model provided the best fit, where plant height decreased by 26 cm for every 10,000 increase in S rate. In 2019 at the Holland site, both quadratic and linear model provided the best fit where plant height decreased by 13 cm for every 10,000 increase in S rate. At the Hare site, the S rate of 120,000 and 200,000 ha^− 1^ produced similar plant height and the lowest plant height was observed with the 260,000 ha^− 1^ S rate (Table 4). Table 4Influence of the interaction between environment and seeding (S) rate on plant height (cm) of biomass sorghum.Plant height201720182019S rateHollandHareHollandHareHollandHare120,000442 a458 a446 a507 a536 a485 a160,000447 a447 a430 b510 a539 a458 b200,000442 a424 c424 b485 b527 b478 a240,000421 b443 b388 d460 c502 c464 b260,000447 a405 d406 c421 d518 b450 c310,000425 b440 b367 e429 d492 c459 bTrendnsnsL/QLL/QnsR^2^--0.900.850.81/0.83-L: linear; Q: quadratic; not significant at *P* ≤ 0.05; same letters indicate no treatment differences at *P* ≤ 0.05. Stem diameter of sorghum was significantly influenced by E (*P* ≤ 0.0001), P (*P* ≤ 0.0001), S (*P* ≤ 0.0001), and E × P (*P* ≤ 0.0001) (Table 1; Fig. 2 C). The response of stem diameter to S rate was best described by a linear model, where stem diameter decreased by 0.06 mm for every 10,000 increase in S rate (Table 1). The E × P interaction showed that, across all the environments EP produced greater stem diameters than LP. At the Holland and Hare site, across all the years EP produced an average of 16% and 18%, respectively greater stem diameter than LP (Fig. 2 C). Fig. 4Influence of the interaction between harvest date and environment on mean dry matter yield of biomass sorghum under early planting. Same letters indicate no treatment differences at *P* ≤ 0.05. Bars represent standard error of the mean. ### Disease and lodging severity Sorghum foliar anthracnose severity was influenced by E (*P* ≤ 0.0001), P (*P* ≤ 0.0001), S (*P* = 0.0491) and E × P (*P* ≤ 0.0001) (Table 1; Fig. 5 A). The relationship between disease severity to S rate was best described by a quadratic model, where disease severity slightly increased with increases in S rate from 120,000 to 310,000 ha^− 1^ (Table 1). Lower disease severity was observed with S rates of 200,000 ha^− 1^ or less (Table 1). The greatest disease severity, 60%, occurred with the S rate of 310,000 ha^− 1^ (Table 1). The E × P interaction showed that in 2017, at both Holland and Hare, EP had lower disease severity (41% and 23%, respectively) than LP. However, in 2018 and 2019, LP had lower disease severity than EP at both sites. In 2018, at both Holland and Hare, LP had lower disease severity (55% at both the sites) than EP. Similar trend was observed in 2019 Holland and Hare, where LP had lower disease severity (41% and 65%, respectively) compared to EP (Fig. 5 A). Lodging severity of sorghum was significantly influenced by E (*P* ≤ 0.0001), S (*P* = 0.0024) and E × P (*P* ≤ 0.0001) (Table 1; Fig. 5B). The response of lodging severity to S rate was best described by a quadratic model, where lodging severity increased with increase in S rate from 120,000 to 310,000 ha^− 1^ (Table 1). Reduced lodging severity (30% and 34%) was observed with S rates of 160,000 and 120,000 ha^− 1^, respectively (Table 1). The greatest lodging severity occurred with higher S rates (240,000 to 310,000 ha^− 1^). The E × P interaction showed that in 2017 at both Holland and Hare, EP had less lodging severity than LP. In 2018 at the Holland site, EP had less lodging than LP, however the opposite was true for the Hare site. At both Holland and Hare in 2019, LP had lower lodging severity than EP (Fig. 5B). Fig. 5Interaction of environment and planting date on **(A)** disease severity and **(B)** lodging severity of biomass sorghum. EP: early planting; LP: late planting. Same letters indicate no treatment differences at *P* ≤ 0.05. Bars represent standard error of the mean. ### Principal component analysis Principal component analysis was limited to two axes because the first two components captured most of the variance (eigenvalues > 1). (Figure 6 A and B, S1). Early planting, principal component 1 (PC1) accounted for 53.8% of the variation and was driven by DS, June~RF~, Oct~RF~, C~RF~, PHT, and DMY, with strong negative loadings from June~GDD~, C~GDD~, June~CSI~, and June~T~. Principal component 2 (PC2) explained 22.1% of the total variation and was alternatively driven by Aug~RF~, C~CSI~, DMY, SD, and C~GDD~, with negative contributions from LS, DS, and June~T~ (Table S4, Fig. 6 A). Cluster analysis grouped the variables into three distinct clusters. Cluster 1 included June~RF~, Oct~RF~, June~GDD~, C~GDD~, June~CSI~, June~T~, and DS. Cluster 2 comprised of C~CSI~, Aug~RF~, C~RF~, and LS and cluster 3 consisted of DMY, PHT, and SD. Late planting, PC1 accounted for 49% of the total variation and was driven by July~GDD~, July~T~, CSI, C~CSI~, SD, and DMY, with strong negative loadings from Nov~RF~, C~RF~, and July~RF~. The PC2 explained 25.9% of the variation and was alternatively driven by C~GDD~, Sept~RF~, DS, LS, and CSI, with negative contributions from PHT, SD, and DMY (Table S5, Fig. 6B). Cluster analysis grouped the variables into three distinct clusters. Cluster 1 included Nov~RF~, C~RF~, July~GDD~, July~T~, July~RF~, CSI, and C~CSI~, cluster 2 comprised of Sept~RF~, C~GDD~, PHT, and SD. Cluster 3 consisted of LS, DMY, and DS. Fig. 6Principal components analysis on environments and S rates influence **(A)** early planting (EP) on dry matter yield (DMY); plant height (PHT); stem diameter (SD); disease severity (DS); lodging severity (LS); June rainfall (June~RF~); August rainfall (Aug~RF~); October rainfall (Oct~RF~); cumulative rainfall (C~RF~); June growing degree days (June~GDD~); cumulative growing degree days (C~GDD~), June crop stress index (June~CSI~), C~CSI~ (cumulative crop stress index); and June temperature (June~T~). **(B)** late planting on DMY; PHT; SD; DS; LS; July rainfall (July~RF~); September rainfall (Sep~RF~); November rainfall (Nov~RF~); C~RF~; July growing degree days (July~GDD~); C~GDD~, July crop stress index (July~CSI~), C~CSI~; and July temperature (July~T~). ### Economic analysis A sensitivity analysis was conducted to evaluate changes in net income with variations in biomass selling prices, ranging from $50 to $90 Mg^− 1^, across different S rates within each environment for both EP and LP using a fixed seed price of $10 kg^− 1^. Across all the environments, EP consistently produced higher net income compared to LP for all seed rates and feedstock prices, since dry matter yield under EP was greater compared to LP (Table S2). In 2017 at the Holland and Hare sites, the S rate of 120,000 ha^− 1^ produced the highest net income, except for 2017-Hare with LP where 160,000 ha^− 1^ produced higher net income. In 2018 at the Holland and Hare sites, S rate of 200,000 ha^− 1^ under EP, and S rate of 310,000 ha^− 1^ under LP produced the highest net income. In 2019, at the Holland and Hare site, the S rate of 120,000 ha^− 1^ produced the highest net income with EP as compared to other S rates in both EP and LP. A separate economic analysis (four environments) for date of harvest (August, September and October) showed that October harvest always produced greater income compared to August and September harvests (Table S3). In 2017, the August and September harvests in both locations resulted in losses of 64 and 19%, respectively, compared to the October harvest. In 2019, the August and September harvests resulted in losses of 34 and 5%, respectively at the Holland site, and 31 and 22%, respectively at the Hare site compared to the October harvest (Table S3). ## Discussion ### Agronomic traits Environmental conditions varied considerably across the three years of the study period (Fig. 1), which directly correlated with significant differences in biomass sorghum performance (Table 1). Previous studies^44–51^ also reported that the environment had major influence on productivity of biomass sorghum. Environmental factors such as soil type, rainfall, temperature, and agronomic management practices partly explain variations in biomass sorghum yield (Table 1, S1; Fig. 1). The year 2019 produced the highest dry matter yield (31.9 Mg ha^− 1^) at the Holland site which can be explained with uniform distribution of rainfall throughout the growing period (Fig. 1). In contrast, 2018 experienced high intensity rainfall during July and August, coinciding with the rapid vegetative growth and transition to reproductive stages, the two critical stages of biomass sorghum growth. The excessive rainfall might have contributed to high rate of crop lodging observed at the Holland site which resulted in the lowest dry matter yield (Table 1). In 2017, comparatively low rainfall was observed, which likely contributed to reduced dry matter yield, suggesting timely rainfall is a major driver in biomass production. Additionally, dry matter yield showed a poor to negative correlation with temperature under both EP and LP (Table S4, S5). This indicates that temperature during the early stages of crop growth could significantly influence seed germination and establishment, potentially affecting final plant population and, consequently, dry matter yield (Table S1). It is evident from the results that even distribution of rainfall in 2019 likely provided the optimal balance of moisture availability, allowing the sorghum to maximize its biomass production without the negative effects of either drought or excess water while also resulting in less lodging. Furthermore, excessive rainfall, as observed in 2018, likely contributed to poor seed emergence, reduced solar radiation and increased lodging all of which significantly decreased the sorghum dry matter yield (Table 1; Fig. 1). Consistent with our findings, poor sorghum establishment under waterlogged conditions has been previously reported^52,53^. These inter-annual differences in rainfall patters underscore the importance of developing management strategies, such as adjusting planting date and S rate, to enhance sorghum resilience under variable weather conditions. Sorghum dry matter yield was 69% higher for EP compared to LP (Table 1). The results from this study are supported by findings^36^ in sweet sorghum, biomass sorghum^54^, grain sorghum^55^, maize and soybean^56^, soybean^57^, and sweet corn^58^. The last study also reported that earlier planting dates produced the highest yields. For both EP and LP, in 2019 and 2017, in most cases, low S rate of 120,000 ha^− 1^ produced higher dry matter yield at both sites compared to high S rates (Table 3). These findings are in line withAlderfasi et al. (2016) who observed decrease in biomass yield with increase in planting density. Similarly, Jahanzad et al.^59^ reported forage sorghum yields were similar between S rates of 150,000 and 250,000 ha^− 1^. When it comes to 2018, which experienced high rainfall events S rate of 200,000 ha^− 1^ produced the highest dry matter yield under EP at Holland and Hare sites, likely because most plants lodged, and slightly higher S rate might have compensated for the dry matter yield. Consistent with this, dry matter yield in 2018 showed weak negative correlation (*r* = −0.02) with lodging. However, under LP conditions high S rate of 310,000 ha^− 1^ produced the highest dry matter yield at both sites as this S rate might have counterbalanced the short vegetative growth time, helping to capture more resources quickly and efficiently. Increased plant density ensures the canopy closes quicker, optimizing photosynthesis and mitigating yield losses due to the shorter growing time. This was not the case with early planting as it seems early planting allows plants to have enough time to fully develop and establish and thus, have less competition for resources, including sunlight, nutrients, and water. Low plant population, thus, help those plants to become larger and thicker (higher stem diameter) which provides high yields. These results are supported by those of Villar et al. (1989) and Williams et al. (1999), who found that LP requires higher S rate compared to EP. In addition, the higher S rate requirement for LP could be due to overlap of planting time or early growth stages of biomass sorghum with warm temperatures (Table S1). A separate analysis (four environments) also showed that dry matter yield significantly decreased with increases in S rates (Fig. 2) indicating that planting biomass sorghum with 120,000 ha^− 1^ could produce optimum dry matter yield. In EP, although S rates differed widely, higher plant population did not translate into higher yields reflecting greater intra-specific competition as well as decreased stem diameter resulting in thinner plants which also lodged and thus, experienced yield reduction^60^. In contrast, under LP, increase in stand density improved light interception during the shortened season, resulting in higher yields^61,62^. These results also suggest that new research should focus on lower seeding rates than 120,000 ha^− 1^ to assess whether biomass sorghum respond positively to lower seeding rates in which growers could not only achieve high production but also reduce their production costs. As with the date of harvest among the environments, delayed harvest up to the end of October produced the maximum dry matter yield compared to early harvest, this could be attributed to the extended duration of photosynthesis and continued biomass accumulation over the longer growing period. (Fig. 4). These findings are in line with Atis et al.^63^, Hassan et al.^26^. Our results contribute to future research direction suggesting a new trial design to assess a combination of S rates below 120,000 ha^− 1^ and harvesting dates beyond October to identify management scenarios that maximize yield and farm profit. However, it is also important to note that harvest timing is strongly influenced by weather conditions and the requirements of the subsequent crop in rotation. Morphological parameters form a basis for interpreting plant growth, and the parameters such as sorghum plant height and stem diameter were positively correlated with dry matter yield under EP and LP, where optimum conditions as in 2019 produced higher plant height and stem diameter, while the wetter conditions in 2018 produced the least (Table 1). It is also evident from PCA correlations that under LP conditions, rainfall during the months of July, September and November were negatively correlated with plant height and stem diameter (Table S5, Fig. 6B). Between the planting dates, EP resulted in 21 and 15% higher plant height and stem diameter, respectively clearly indicating the advantage of EP in improving the plant performance with extended growing period and optimum temperature. These findings are in line with^45,56–58^. The S rate of 120,000 ha^− 1^ resulted in the highest plant height and stem diameter. While it is generally understood that increasing the S rate leads to greater within-species competition for light, resulting in taller plants, our results could be explained by more optimal plant spacing at low S rates, which might have minimized competition for resources such as nutrients, water, and sunlight. Additionally, at this S rate, plants likely experienced better root development and resource allocation, leading to improved agronomic traits. In contrast, the higher S rate of 310,000 ha^− 1^ produced the lowest plant height and stem diameter as overcrowding could increase competition for resources^64^(Table 1). Furthermore, the dense canopy may have reduced air circulation, creating a microenvironment with higher humidity and conditions more conducive for foliar anthracnose development. Consistent with our findings^16,17,49,65,66^, also reported that higher S rates resulted in reduced plant height and stem diameter. Furthermore, results of PCA also confirmed that stem diameter was negatively correlated with disease and lodging severity, under both EP and LP. These findings clearly shows that plants with narrower stem diameters were more susceptible to disease and lodging (Table S4, S5; Fig. 6 A and B,). Narrow stem diameters might have increased lodging and disease severity either indirectly, by weakening stems due to reduced photosynthetic area caused by increased foliar anthracnose severity or directly through an increased incidence of anthracnose stem rot. ### Disease and lodging severity The lower sorghum anthracnose severity observed in 2017 coincided with drier conditions that are less conducive for disease development^67,68^. Furthermore, the drier conditions likely reduced the leaf wetness duration, which is crucial for *C. sublineola* spore germination and infection. The relatively high rainfall in 2018 and 2019 likely provided a more humid, and thus conductive, microclimate for pathogen establishment and spread (Table 1; Fig. 1). These findings align with previous studies^68,69^. In 2017 at both sites, EP had less disease severity, which can be explained by drier conditions during crop developmental stages (~ 60–90 DAP) that are susceptible to fungal infection, whereas more moisture was present during the corresponding developmental stages for the LP resulting in greater infection and disease development. Moreover, in 2018 and 2019, LP had less disease severity compared to EP (Fig. 5 A), as consistent rainfall during the months of June, July and August might have caused the EP sorghum leaves to remain wet for longer periods (Fig. 1). Foliar anthracnose severity increased with increase in S rate from 120,000 to 310,000 ha^− 1^. These could be explained with several interconnected factors, as dense plant canopies reduce air circulation, creating a more humid microclimate that favours disease development by prolonging the leaf wetness duration. Moreover, increased competition for nutrients and light might have weakened plant defence mechanisms, making them more susceptible to the disease. These findings align with reports by^19,20^, who observed increased disease severity in response to higher seeding rates. Lodging severity followed the same trend as disease severity where the drier environmental conditions in 2017 had lower lodging severity compared to wetter conditions as in 2018 and 2019 (Table 1; Fig. 1). Planting at an earlier date resulted in less lodging in drier environments as seen in 2017, as EP might have allowed the biomass sorghum to develop and establish fully compared to the LP conditions. However, in 2018 and 2019, LP had less lodging compared to EP (Fig. 5B), and this could be explained with high rainfall during the months of July and August which might have increased lodging under EP (Figs. 1 and 6 A). These results are supported by^70,71^, who also reported that rainfall played a significant role in lodging. Lodging severity increased with increase in S rate, likely because of increased plant density and reduced stem strength (Table 1). This is supported by the observed negative PCA correlations of lodging severity with plant height and stem diameter, under both EP and LP, respectively (Table S4, S5; Fig. 6 A and B). This might be likely due to higher plant densities intensifying competition for resources, especially sunlight, prompting sorghum plants to grow taller as they strive to capture more light. This increased plant height, coupled with thinner stems, reduces mechanical strength and weakens structural integrity, thereby heightening susceptibility to lodging. This aligns with the findings of^65,72^, who reported reduced culm strength and altered culm morphology at higher seeding rates, resulting in an increased lodging index. Also as noted by^35,73^, thin-stemmed plants resulting from high S rates are more prone to lodging, suggesting that high S rate should be avoided in regions where lodging is a significant concern. ### Economic analysis In Virginia, each year, 20% of the soybean (*Glycine max* L.) acreage planted in the state is double crop soybean planted behind wheat. This is because the state’s climate can accommodate two crops in one year for the economic benefit of the growers. However, in absence of rotation, double cropped soybean’s yields can suffer over years. We envisioned that LP sorghum in this research could mimic a rotation solution for double crop soybean and address some critical agronomic and economic questions for the Virginia growers. The economic analysis showed that for EP, lower S rate of 120,000 ha^− 1^ generally resulted in high net income due to reduced input costs and high dry matter yield (Table S2). A seeding rate below 200,000 ha^− 1^ was always economical and increasing the S rate above that always resulted in a profit loss for growers. In contrast, for LP, higher S rate of 310,000 ha^− 1^ produced greater net income, likely due to high plant density compensating for the reduction in individual plant dry matter yield. These findings highlight the critical need to fine-tune S rate based on planting date and environmental conditions to maximize profitability while minimizing the cost of production. By staying within the optimal S rate range of 120,000 ha^− 1^ for EP and 310,000 ha^− 1^ for LP, farmers can achieve a balance between targeted returns and cost efficiency (Table S2). Since our results suggested increased S rate resulted in increased biomass sorghum yield for LP, and this was economical, future research should focus on finding whether exceeding 310,000 ha^− 1^ can further improve the biomass sorghum production with increasing growers’ profit. The economic analysis performed for date of harvest of biomass sorghum showed that delaying the harvest up to October resulted in higher economic returns, compared to early harvest in August or September (Table S3). Reduced income from August and September harvest dates were due to trade-off between early maturity and biomass yield. Delaying the harvest up to a later date in October would provide enough time for the biomass sorghum to fully express its genetic potential, thereby producing greater yields with improved income (Table S3). It is important to mention that our economic analysis was based on fixed seed price and did not account for other variable costs. Incorporating partial budgeting in future studies would provide a more comprehensive assessment of profitability. ## Conclusions The results of this study indicate that S rate significantly affects sorghum dry matter yield, plant height, stem diameter, disease and lodging severity. Planting date dictates the optimum S rate, highlighting the importance of fine-tuning management practices to specific planting date. Under EP, increasing the S rate above 120,000 ha^− 1^ generally led to diminishing returns and could negatively impact sorghum growth and dry matter yield. Conversely, under LP, the highest tested S rate of 310,000 ha^− 1^ showed a positive influence on sorghum dry matter yield. Harvesting later in October produced significantly higher dry matter yield compared to the early harvest indicating to maximize production, harvest must be delayed until October. These results clearly indicated that the optimum S rate for biomass sorghum is not universal but rather depends critically on the planting date and also environment. Adhering to these planting date-based S rate and time of harvest recommendations, farmers can balance the trade-offs between individual plant performance and overall field productivity with improved disease and lodging resistance. We propose that future research should evaluate seeding rates less than 120,000 plants ha^− 1^ for EP, higher than 310,000 plants ha^− 1^ for LP, and harvesting dates beyond October to determine maximum economic return for growers in the Mid-Atlantic region. ## Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1