Potential for Enhancing Forage Sorghum Yield and Yield Components in a Changing Pannonian Climate

Abstract

Climatic variability, particularly fluctuating precipitation and rising temperatures, poses a significant threat to crop productivity and stability. Forage sorghum hybrids are a promising alternative for fodder and bioenergy due to their high level of drought tolerance. This study evaluated genotypic variation and environmental adaptability of 60 forage sorghum genotypes: 13 parental lines, their 40 crosses and seven commercial hybrids, to identify high-yielding, stable hybrids for biomass production under changing agroecological conditions. Field trials conducted over two contrasting years revealed significant genotype-by-environment interactions (p < 0.05), highlighting the need for multi-year evaluations. While favorable rainfall in 2020 enhanced vegetative traits (plant height, stem diameter, leaf area), biomass yield variability increased, emphasizing that favorable vegetative development does not necessarily correlate with yield stability. Principal component analysis indicated that plant height, stem diameter and leaf-related traits contributed most to genotypic differentiation. However, no single trait emerged as a reliable predictor of yield, suggesting complex trait interaction. These findings underscore the importance of integrative breeding strategies that combine phenotypic trait assessment with environmental adaptability to ensure sustainable biomass production. Sorghum’s drought tolerance and resilience make it a promising crop for future food and feed security in regions prone to climatic stress.

1. Introduction

Sorghum (Sorghum bicolor L. Moench) is an annual plant species with multiple uses. Besides grain sorghum, which is considered the fifth most important cereal [1], sorghum can be cultivated for fodder (green fodder, hay, haylage, silage) and as broomcorn [2]. According to recent data, in 2023, sorghum was grown in an area of almost 40,000,000 ha in the world, 274,000 ha in Europe [3]. The production of sorghum in Serbia in the same year was sown on approximately 2600 ha, with a high yield potential, an average yield higher than 3100 kg ha⁻¹ of grain. One of the most important advantages of growing sorghum is its easy adaptation to different agroecological conditions. It is an excellent addition to ruminant nutrition and an important bioenergy crop. Forage sorghum has high yield potential and less resource requirements, because of its high level of drought and high temperature tolerance [4,5]. Mahmood et al. [6] concluded that sorghum can be used as an alternative to maize in energy production due to its better adaptation to drought conditions and more reliable plant production.

Climate change poses a major threat in the 21st century, with rising global temperatures and more frequent, intense extreme weather events like heatwaves, droughts, floods and storms expected in the coming years [7]. In response to climate challenges, growing crops like sorghum, which require less water, tolerate diverse environmental conditions and provide high yields, is essential for sustainable farming [8,9,10,11]. One of the main goals in improving forage sorghum breeding programs is the selection of high-yielding hybrids suitable for pure cropping or intercropping with maize to produce high-quality silage and serve as a valuable raw material for biogas production. Discovery of cytoplasmic male sterility in sorghum advanced the exploitation of the effect of heterosis for yield and yield components and enabled the creation of forage sorghum commercial hybrids by crossing grain sorghum as a female parent and Sudan grass as a male parent [12]. The selection of forage sorghum is primarily based on yield as an indicator of environmental adaptation, while understanding the correlation between key morphological traits is essential for guiding improvement strategies and predicting selection outcomes.

The most important traits found to be associated with higher fodder yield included plant height, number of leaves, leaf area, stem diameter and biomass yield per unit area [13]. The author also suggests a strong association between plant height, leaves number, leaf area and dry matter yield. Principal component analysis (PCA) is widely used to identify traits with the greatest contribution to genetic variation, providing a foundation for effective trait prioritization in forage sorghum breeding programs [14,15,16,17,18]. The expression of plant height is significantly influenced not only by genotype but also by environmental factors. Pataki [19] states that as forage sorghum plant height increases, the dry matter content also rises, with precipitation amount and distribution during development playing a major role. Numerous studies indicate that breeding efforts aimed at increasing dry matter yield in forage sorghum should focus on enhancing the number of leaves and plant height [20,21]. A greater number of leaves and a larger total leaf area positively contribute to increased biomass yield. Sorghum leaves vary in length depending on their position on the plant; those near the top, aside from the flag leaf, are characterized by reduced width and length, while leaves in the middle part of the stem tend to be longer [22]. The stem constitutes the largest proportion of fresh fodder and stem diameter also directly influences yield [23]. Precipitation levels and accumulated temperatures are crucial factors for the accumulation of fresh biomass in forage sorghum [24].

Although numerous studies have explored the associations between major agronomic traits and yield components of forage sorghum, there is still a notable gap in understanding how these traits respond to diverse fluctuating environmental conditions, particularly under combined abiotic stresses. In breeding programs, this lack of knowledge limits the predictability of final plant performance and limits the identification of genotypes with stable and desirable traits.

This study addresses the limited knowledge on trait stability and genotype-by-environment interactions in forage sorghum under increasingly variable agroecological conditions of Pannonian climate, offering novel insights into hybrid performance across contrasting seasons, aiming to define a trait-based selection framework for the development of high-yielding, climate-resilient genotypes.

The aim of this research was to contribute to the advancement of forage sorghum breeding by examining the stability of trait relationships and the dependency of trait associations on climatic fluctuations. This will contribute to a better understanding of hybrid performance and adaptability, ultimately supporting the development of more resilient and high-yielding varieties suited to extreme climatic challenges. Consequently, the objectives of this study were: (1) to test experimental hybrids and a diverse panel of their parental components for biomass yield and the most important yield components in unstable agroecological conditions of the Pannonian climate, (2) to examine the association between parameters that affect green fodder and dry matter yield and (3) to test combining abilities which will provide a genetic framework for interpreting the observed variation in yield and correlations between traits and identify the most promising hybrid combinations for fodder production under challenging climatic conditions.

2. Materials and Methods

A field trial was set up during 2019 and 2020 at Rimski šančevi site, Novi Sad, Serbia (45°20′ N, 19°51′ E), on chernozem soil type, clay loam texture according to the USDA classification system [25]. The elementary plot consisted of two rows, 5 m long, at 50 cm row-to-row space. The trial design was a resolvable incomplete block design–lattice design, alpha form, with three replications.

The study was conducted on 60 different genotypes (Table S1), which where (1) 40 experimental forage sorghum hybrids; (2) 13 parental lines (eight restorer lines and five male sterile testers); (3) seven commercial hybrids available at the Serbian market.

The tested parental components are part of the gene collection of the Institute of Field and Vegetable Crops, National Institute of the Republic of Serbia, Novi Sad, Serbia. Experimental hybrids were created by manual crosses between male parents (Sudan grass) and female parents (grain sorghum) during 2018 and 2019.

Sorghum genotypes were sown on 22 May 2019 and 8 May 2020. Sampling and cutting of all genotypes were performed at soft-dough grain maturity stage (BBCH 85), which is recommended as the most optimal for cutting when the production aim is silage [23]. In 2019, cutting of all genotypes was conducted in two dates, 29 August and 6 September, due to differences in the timing of the appropriate developmental stage. In contrast, in 2020, cutting was completed in a single day, on 7 September, owing to more uniform maturation compared to the first year of the study.

Data on temperatures and precipitation were taken from the Republic Hydrometeorological Service of Serbia’s electronic publications “Meteorological yearbook–climatological data” (https://www.hidmet.gov.rs/latin/meteorologija/klimatologija_godisnjaci.php, 30 June 2025) [26], obtained from their official meteorological station located at Rimski šančevi site. The data are presented in Figure 1.

Although the field trials were conducted in two consecutive growing seasons (2019 and 2020), the climatic conditions differed substantially between them, particularly in terms of temperature regimes and the distribution of precipitation. This natural environmental contrast provided a valuable opportunity to assess hybrid performance and adaptability under distinct agroecological stress scenarios. In both years, the growing season was characterized by high temperatures with an uneven rainfall distribution. A significant difference between the two years was the considerably higher temperatures during the first year of the study, as well as extreme events of intense and unevenly distributed rainfall during the second year. The total precipitation in 2019 was 632.1 mm, which is in line with the long-term average, while the amount during the growing season of sorghum (May–August) was 311.4 mm. In 2020, the total precipitation was 733.2 mm, with 424 mm during the growing season, significantly higher than the long-term average of 632.8 mm and 281 mm, respectively (for the period from 1964 to 2018). Both growing seasons also had significantly higher average temperatures compared to the long-term average of 11.4 °C, reaching 13.4 °C and 21.4 °C in 2019 and 12.8 °C and 20.6 °C during the growing season in 2020.

In this research, the analyzed traits were (1) biomass yield–GFY, green fodder yield and DMY, dry matter yield; (2) yield components–PH, plant height; SD, stem diameter; SDCP, stem diameter at cutting point; IN, internodes number; LN, leaves number; LA, leaf area; PL, panicle length; NL, neck length; (3) most important phenological stages–DH, number of days to heading (BBCH 51–55) and DSD, number of days to soft-dough grain maturity stage (BBCH 85) in sorghum genotypes.

During both growing seasons, a sample of 10 plants per plot was analyzed for yield components according to UPOV guidelines [27]. Plant height was measured from the cutting point to the tip of the panicle for each plant. Stem diameter, besides at the cutting point, was measured above the fourth leaf, including the flag leaf, or above the third true leaf. Neck length in sorghum was measured on the part of the stem from the end of the panicle to the flag leaf. Measurements of leaf length and leaf width at the widest part of the fourth leaf from the top of the plant, including the flag leaf, or the third true leaf, were used to calculate the parameter leaf area using the following formula:

$$LA=LL\times LW\times 0.71,$$

(1)

where

• LA—leaf area (cm²);
• LL—leaf length (cm);
• LW—leaf width (cm);
• 0.71—constant for estimating sorghum leaf area [28,29,30].

Green fodder yield was measured from the entire plot. To determine dry matter yield, the mass of 10 plants was measured at the moment of cutting, as well as after drying (oven-dried to absolutely dry state).

Phenological stages were recorded according to the BBCH scale [31].

Line × tester method [32] was used for analysis of combining abilities (GCA, general combining ability and SCA, specific combining ability).

Associations between quantitative traits (plant height, stem diameter, leaves number, leaf area, green fodder yield and dry matter yield) and categorical variables (genotype groups and years) were analyzed using principal component analysis with mixed data (PCAmix). Prior to conducting the PCA, Bartlett’s test was performed to assess the suitability of the dataset, confirming that the variables were sufficiently correlated for multivariate analysis. PCA was applied as a statistical method to reduce the dimensionality of complex agronomic data and to improve the interpretability of trait relationships by generating uncorrelated components that explain the maximum variance. This enabled the identification of key traits contributing to genetic variation, supporting effective genotype selection and breeding planning. Genotype groups included three categories: stable, superior in one year and atypical genotypes. PCAmix was performed using XLSTAT in Microsoft Excel (XLSTAT Statistical and Data Analysis Solution, 2022.5.1) [33]. The package ‘ggplot2’ in R software, version 4.2.2 was used for the visualization of the PCAmix analysis (R Core Team, 2022) [34]. ANOVA was performed using Minitab 17 Statistical Software (trial version). For visualization of the combining analysis results using packages ‘tidyverse’ and ‘pheatmap’ in R software [34], a heatmap is created. It displays special combining ability value matrices for all possible parent combinations, allowing quick identification of the best combinations by using darker colors that indicate stronger effects (positive or negative).

3. Results

3.1. Sorghum Genotypes Showed Variability in Biomass Yield and Yield Components Responses to Climatic Factors in the Pannonian Region

The comparative analysis of sorghum genotypes across 2019 and 2020 revealed notable differences in phenotypic expression among parental lines, hybrids and commercial hybrids, largely attributed to fluctuations in agroecological conditions between the two growing seasons (Figure 2). The 2020 growing season was overall more favorable for biomass and yield traits such as GFY, DMY and PH, potentially due to better environmental conditions. The year 2019 showed earlier flowering and higher leaf number, though yields were generally lower. Noticeable year-to-year differences confirm strong genotype-by-environment interaction (p < 0.05), which should be considered in the selection of stable and high-performing sorghum genotypes.

Plant height showed relatively stable distributions across years, suggesting that this trait was less affected by year-to-year environmental fluctuations (p = 0.960, Table S2). During 2020, plants were taller on average (241.21 cm,

Table 1. Descriptive statistical parameters for yield components and yield evaluated in sorghum genotypes.

	2019						2020
	Min	Max	Mean	SD	CV (%)	SE	Min	Max	Mean	SD	CV (%)	SE
DH	47.00	79.00	56.72	6.30	11.11	0.47	57.00	94.00	70.39	7.34	10.43	0.55
DSD	15.00	33.00	25.51	3.40	13.33	0.25	28.00	36.00	32.78	1.24	3.77	0.09
PH	66.23	372.30	237.16	55.75	23.51	4.16	82.80	342.50	241.21	45.56	18.89	3.40
IN	2.50	12.80	8.23	1.57	19.04	0.12	4.70	13.00	8.65	1.41	16.28	0.11
SDCP	4.70	21.93	13.26	2.72	20.48	0.20	6.27	25.15	13.86	2.79	20.15	0.21
SD	4.79	17.56	10.43	2.17	20.77	0.16	4.22	15.74	9.41	1.83	19.44	0.14
LN	4.90	14.10	8.15	1.83	22.46	0.14	3.20	11.90	6.99	1.28	18.36	0.10
LA	145.26	522.94	291.47	66.05	22.66	4.92	138.63	500.03	278.44	64.55	23.18	4.81
PL	16.71	46.16	28.41	6.12	21.54	0.46	12.40	42.00	27.13	4.98	18.35	0.37
NL	0.00	22.87	9.08	5.31	58.48	0.40	0.00	13.50	5.22	2.93	56.02	0.22
GFY	24.08	114.54	51.95	16.23	31.25	1.21	12.05	137.70	61.33	23.38	38.11	1.74
DMY	6.17	47.26	18.78	7.71	41.06	0.58	2.78	79.00	23.97	12.42	51.81	0.93

DH, number of days to heading (BBCH 51–55); DSD, number of days to soft-dough grain maturity stage (BBCH 85); PH, plant height (cm); IN, internodes number; SDCP, stem diameter at cutting point (mm); SD, stem diameter (mm); LN, leaves number; LA, leaf area (cm²); PL, panicle length (cm); NL, neck length (cm); GFY, green fodder yield (t ha⁻¹); DMY, dry matter yield (t ha⁻¹); SD, standard deviation; CV, coefficient of variation; SE, standard error.

). Stem diameter showed greater variability in 2019 (ranging from 4.79 to 17.56 mm), while in 2020 average value was slightly lower (9.41 mm). The number of leaves per plant was higher in 2019 (8.15 leaves) and exhibited greater variability (ranging from 4.90 to 14.10 leaves), compared to a narrower range observed in 2020 (3.20 to 11.90). Both of these traits (stem diameter and leaf number) showed very significant variation across two years (p = 0.000). Leaves had larger areas in 2019 (291.47 cm²) and higher variability. Leaf area displayed moderate variation between years (p = 0.018). Green fodder yield and dry matter yield exhibited the most pronounced differences between years, making these traits more dependent on environmental conditions (p = 0.000). Tested genotypes had higher fresh biomass yield in 2020 (61.33 t ha⁻¹ GFY and 23.97 t ha⁻¹ DMY), suggesting improved growing conditions.

In 2020, both days to heading (BBCH 51–55) and days to soft-dough stage (BBCH 85) were noticeably higher (70.39 DH and 32.78 DSD) than in 2019 (56.72 DH and 25.51 DSD), suggesting a delay in crop development during the second year of the study (p = 0.000). This delay can be attributed to the increased precipitation at the beginning of the growing season in 2020, which likely extended the vegetative period. The number of internodes was slightly higher on average in 2020 (8.65), indicating prolonged stem elongation, which is often promoted by favorable water availability (p = 0.019). The stem diameter at the cutting point was higher in 2020 (13.86 mm), suggesting stronger and more robust stem bases, which may have developed as a result of reduced water stress and more uniform growth conditions during the vegetative phase (p = 0.020). Panicle length showed slight reduction in 2020 (27.13 cm; p = 0.000). Neck length exhibited a more pronounced decrease in 2020 (5.22 cm), which could be a result of excess moisture and fluctuating conditions during the reproductive phase, potentially affecting the development of this trait (p = 0.000).

A Tukey pairwise comparison was conducted to assess the significance of differences in trait values between the two growing seasons (2019 and 2020), using Tukey’s method with 95% confidence (Figure 2). The results revealed that DH, DSD, IN, SDCP, GFY, DMY showed significantly higher values in 2020 (group A), while SD, LN, LA, PL, NL were significantly higher in 2019 (group A in 2019 vs. group B in 2020), suggesting a differential response of traits to seasonal climatic variation. Only PH showed no significant difference between years (group A in both growing seasons), indicating its relative stability across environmental conditions.

3.2. Key Agronomic Traits Associated with Biomass Yield Improvement

To identify which key traits were most strongly associated with the yield performance of genotypes with differing productivity levels across the two years, a PCAmix analysis was conducted. The biplot of a PCAmix analysis (Figure 3) integrates both quantitative traits (plant height, stem diameter, leaves number, leaf area, green fodder yield and dry matter yield) and categorical variables (genotype groups and years). Through this analysis, the performance of three genotype groups (stable, superior in one year and atypical), across two years based on key traits, was tested. The first three principal components, with eigen values greater than unity, contributed 60.0% to the overall variability. Dim1 (39.0%) captures the highest variability in the dataset. Traits and genotypes with high absolute values along this axis contribute significantly to overall variation. Dim2 (19.6%) represents additional variation that explains different genotype responses. The colors on the biplot represent the contribution of each variable to the total variability in the presented dimensions. According to these results, PH, LA, SD and LN, highlighted in red, indicate the highest contribution, primarily influencing the first principal dimensions. In contrast, the remaining traits, including the group of superior genotypes in one year, GFY and DMY, were more associated with the subsequent dimensions, while the two tested years (2019 and 2020), shown in blue, displayed the lowest contribution and did not represent key factors in the variability explained by the PCAmix model. The years 2019 and 2020 are distinctly positioned, indicating differences in environmental conditions and genotype performance between them. The year 2020 is associated with higher values for green fodder yield (GFY) and dry matter yield (DMY), suggesting that conditions in this year favored biomass production, as confirmed earlier. GFY and DMY are strongly correlated, as expected, indicated by their close association. Plant height (PH), stem diameter (SD), leaf number (LN) and leaf area (LA) showed significant influence on variation among genotypes. This suggests that these traits may serve as effective selection criteria for improving biomass yield and overall adaptability under variable environmental conditions.

A group of genotypes with stable performances is located near the biplot origin, indicating that they exhibit consistent values across both years without extreme deviations. Genotypes superior in one year are positioned closer to GFY and DMY, meaning they showed high productivity but only under specific conditions. Atypical genotypes are located in the upper left quadrant, suggesting that their performance significantly differs from the main trends observed in other groups. This PCAmix biplot showed how different sorghum genotypes performed across two years, highlighting key trait associations and genotype stability.

Notably, the parental components with positive high general combining ability (GCA) for most traits across both growing seasons predominantly belonged to the group of stable performers. This confirms the appropriate selection of parental lines, as their consistent combining potential reflects good adaptation to the agroecological conditions of the Pannonian region.

Since PCAmix showed only associations between quantitative traits and categorical variables (genotype groups and years), ANOVA was performed to test which predictor variables significantly influenced response variables (plant height, stem diameter, leaves number, leaf area, green fodder yield and dry matter yield).

The genotype group affected all trait effects significantly (p < 0.01), confirming the presence of genetic variability across the tested groups (

Table 2. Analysis of variance for tested sorghum genotype groups (stable, superior in one year and atypical performance of genotypes) across the two years.

Source of Variation	df	PH	SD	LN	LA	GFY	DMY
Genotype group	2	**	**	**	**	**	**
Year	1	n.s.	**	**	n.s.	n.s.	n.s.
Genotype group × Year	2	n.s.	n.s.	n.s.	n.s.	**	**

PH, plant height (cm); SD, stem diameter (mm); LN, leaves number; LA, leaf area (cm²); GFY, green fodder yield (t ha⁻¹); DMY, dry matter yield (t ha⁻¹); df, degrees of freedom; n.s., not statistically significant; **, statistically significant at the 0.01 level.

). The effect of year was significant only for stem diameter (SD) and leaves number (LN), while the interaction between year and genotype group was significant for green fodder yield (GFY) and dry matter yield (DMY). These findings support the PCAmix-based interpretation, particularly highlighting that variation in yield traits was driven by both genotype group and environmental conditions, whereas other traits were predominantly influenced by genetic background only. Despite the contribution of individual traits, none of them significantly influenced yield performance on their own, confirming the complexity of biomass yield determination and the importance of trait interactions and environmental factors.

3.3. Promising Hybrid Combinations for Biomass Production Under Challenging Climatic Conditions Are Identified Based on Combining Ability Effects

After identifying patterns of trait variation across years and groups of genotypes through PCAmix analysis and confirming the statistical significance of these differences using ANOVA, we focused on evaluating the performance of the newly developed experimental hybrids. This approach allowed integration of multivariate trends (PCAmix), statistical significance (ANOVA) and hybrid performance (SCA) in order to propose a subset of the most promising hybrid combinations for future selection and breeding across multiple traits. To assess their specific combining ability (SCA), the 40 experimental hybrids for six key agronomic traits over two years were analyzed. The combining ability analysis results were shown in a heatmap (Figure 4), providing insights into the average sorghum SCA values over two years (Table S3). Through hierarchical clustering, the hybrids were grouped and those with the most favorable performance across traits and years were highlighted. It is noted that experimental hybrids 23, 14, 40, 31, 2, 1, 21, 35 and 8 predominantly performed well with positive high SCA values (represented in red), indicating the highest potential for most of the tested traits and should be prioritized for further selection. These hybrid combinations represent valuable genetic resources for further breeding efforts aimed at developing high-yielding hybrids adapted to the Pannonian climate, ensuring stable biomass production under variable environmental conditions. When comparing these results with previous analysis, most of the hybrids selected for favorable performance and superior SCA performance belonged to the stable genotype performance category, with the exception of the hybrids superior in one year, 23, 40 and 21.

To evaluate the practical relevance of the hybrids selected for superior SCA performance, average trait values across two years were compared with those of the commercial hybrids included in the trial (Table S4). The mean values were used as a basis for comparison, providing a reliable indication of the performance of hybrids under the tested conditions. The results indicated that experimental hybrids 2, 21, 23 and 31 not only demonstrated favorable specific combining ability, but also exceeded the performance of two commercial hybrids in biomass yield (54 and 59) and one in key agronomic traits–stem diameter and leaf area (54), underscoring their potential for future breeding and commercialization.

Experimental hybrids 33, 24 and 7, all with stable performances throughout the two-year research, are characterized by highly negative SCA values according to the combining ability analysis, represented in blue, suggesting less favorable hybrid combinations (Figure 4). In our study, these hybrids exhibited the lowest dry matter yields, combined with reduced leaf area values. Identifying both promising and less favorable hybrids provides a foundation for refining selection strategies and improving breeding efficiency, ultimately contributing to the advancement of forage crop production.

4. Discussion

4.1. Genotypic Variation in Biomass Yield and Its Climatic Adaptability in the Pannonian Region: Implications and Interpretation

Climatic variability, especially changes in precipitation patterns and increased temperatures compared to the long-term average, significantly affects crop production stability and presents a major challenge to global food security [35]. One effective strategy to enhance water-use efficiency and reduce yield vulnerability to water stress is the cultivation of drought-tolerant species such as sorghum. Particularly, genotypes with superior drought tolerance and growth rates can produce more fodder, thus supporting improved water-use efficiency and sustainable agricultural production [36,37,38,39]. Variation in agronomic performance can often be attributed to rainfall quantity and distribution, especially concerning moisture sufficiency and temperature favorability during panicle development and flowering [40]. However, drought stress can negatively affect the crop if it occurs during the early vegetative stage, panicle development before flowering or the period from pollination to maturity, with post-flowering drought stress being particularly detrimental to yield [41,42,43]. The optimal temperature ranges for sorghum development are 26–34 °C during the vegetative phase and 25–28 °C during the reproductive phase [44]. Although sorghum has diverse uses, each of them requires specific traits, such as moderate plant height and thick stems for biomass production due to their contribution to biomass yield and lodging resistance, making it essential to breed customized varieties.

However, since many agronomic traits are influenced by environmental factors, final adjustments to each variety must be made according to the specific growing conditions and intended use [45]. Our results highlight the complex genotype-by-environment interaction in forage sorghum and underscore the need for selection strategies tailored to specific agroecological conditions. While 2019 was characterized by high temperatures and limited biomass accumulation, it showed more consistent trait expression. In contrast, 2020 experienced increased and irregular rainfall, which enhanced vegetative growth (PH, SD, LA) but introduced greater variability in biomass yield (GFY, DMY) and negatively impacted certain reproductive traits such as panicle and neck length. These findings demonstrate that favorable environmental conditions for vegetative growth do not necessarily ensure reproductive success or yield stability. Therefore, evaluating multiple traits across diverse environments is essential for identifying high-performing, stable genotypes. This is consistent with previous research identifying temperature and water availability as key abiotic stressors in forage sorghum production [40,46,47,48].

The pronounced differences in green fodder and dry matter yields between the two years underscore the importance of multi-year evaluations to accurately capture genotypic performance and adaptability under variable agroecological conditions. Previous studies from different localities support these findings. Despite the sandy soil, higher rainfall at the Gross-Gerau location in Germany may have enhanced sorghum dry matter yield, with a significant site × cultivar interaction observed. Additionally, the biogas and methane yield of certain sorghum cultivars were comparable to those of maize, making sorghum a promising alternative for energy production [6]. Sorghum’s potential as a high-quality silage and biofuel crop is due to its high yield and resistance to arid climates. Despite limited annual precipitation, forage sorghum can complement other energy crops [48]. The same results were published in research where sorghum yield and nutritional quality exhibit greater stability under climate change compared to major cereal crops such as maize and rice [49]. The strong influence of agroecological factors on forage sorghum yield is also evidenced by the significant differences in rainfall amount and distribution between 2013 and 2014, which likely affected the average stem height of sorghum hybrids (276 cm vs. 433 cm) and contributed to the higher green fodder yield observed under the more favorable conditions of 2014 [50]. The research in the Mykolaiv region, Ukraine, demonstrated that a sorghum yield of 40.6 t ha⁻¹ could be achieved with 350 mm of annual precipitation, while fluctuations in rainfall over three years significantly affected biomass yield. Southern Ukraine is increasingly impacted by higher temperatures, droughts and other extreme climate events, with a 17% rise in average annual temperature and a 12% decrease in precipitation over the past 50 years. Under favorable weather conditions, it is predicted that sweet sorghum yield could increase by 0.5–2.1 t ha⁻¹ [48]. Accumulated temperature and precipitation significantly influence sorghum yield on the Korean Peninsula, where concentrated summer rainfall during the crop’s extended growth period increases production stress and threatens fodder supply stability under climate change [51]. Although sorghum does not produce fully mature seeds under Polish climate conditions, its ability to generate high green biomass underscores its strong potential as a valuable forage crop [52]. Negri et al. [53] reported that sorghum yield in 2019 matched the national average, declined under water scarcity in 2020 and partially recovered in 2021, highlighting the crop’s resilience and supporting its promotion in Mediterranean regions for sustainable food and feed production under climate uncertainty.

Ultimately, sorghum is highly adaptable to climate variability and water scarcity, thriving in diverse environments, including those with low input and salinity, serving both as food and feed. Its ability to maintain high productivity under changing climate conditions positions it as a crucial crop for ensuring food security, particularly in regions with limited water resources and adverse environmental factors [8]. Furthermore, analysis in Germany indicates that future climate conditions could extend the sorghum growing season, enabling earlier sowing, later harvesting and greater yield potential through the selection of later-maturing cultivars [54]. Significant differences among 338 lines of different origin were observed for all traits across environments, with year effects influenced by earlier sowing in 2014. For successful use of sorghum as a bioenergy and forage crop in temperate Central Europe, improvements in cold tolerance, early maturity and optimal plant architecture (plant height, stem diameter and brix) are essential [55]. Similarly, analysis of sorghum genotypes in Nigeria showed a low phenotypic coefficient of variation for flowering time, maturity, plant height and panicle length, while moderate to high genetic variability was detected for 1000-grain weight, panicle diameter, panicle weight, number of seeds per panicle, grain yield and seedling emergence count [56]. Research conducted in Italy suggests that developing novel sorghum hybrids adapted to Mediterranean conditions could promote its cultivation and encourage European farmers to adopt it, thanks to its high drought tolerance and resilience to environmental changes [57]. Atique-ur-Rehman et al. [58] confirmed that climate change, characterized by rising mean maximum and minimum temperatures and fluctuations in rainfall patterns, has adversely affected both the quality and fodder yield of sorghum. Their research shows that high temperatures and reduced rainfall in 2016 significantly decreased sorghum fodder quality, crude protein content and overall fodder production.

Taken together, our results, alongside evidence from other studies, underscore the importance of evaluating multiple traits across diverse conditions to identify genotypes with stable and desirable agronomic performance. These insights support the refinement of breeding strategies aimed at improving yield stability and adaptability under current and future agroecological challenges.

4.2. Relevance of Key Agronomic Traits in Enhancing Biomass Yield

Yield is a complex quantitative trait governed by multiple interacting factors, where individual yield components, such as stem diameter, number of leaves and plant height, represent isolated traits within a broader physiological and environmental context. According to previous research [13,19,20,21,23], the most important traits, found to be associated with higher fodder yield, were selected for this research. Traits such as days to flowering, plant height and stem diameter have proven to be key for increasing biomass yield. These traits indirectly influence the improvement of bioethanol production when that is the breeding goal, further highlighting their significance in sorghum breeding programs [43,59,60,61].

Although these components contribute to the overall performance, their individual effects in our research did not significantly explain the variability in the main yield traits (GFY and DMY), suggesting that no single trait emerged as a dominant predictor. Instead, the final yield outcome appears to result from the complex interplay between genotype, environment and multiple physiological traits. Consequently, the evaluated set of agronomic traits alone does not sufficiently differentiate high-yielding genotypes, highlighting the need for a more integrative approach when selecting superior hybrid combinations under variable environmental conditions. According to Chaudhary et al. [62], traits such as days to 50% flowering, days to milking stage, plant height, leaf breadth, leaf length, leaf area, stem girth, number of leaves per plant, leaf stem ratio, key quality traits and green fodder yield, with the highest heritability in broad sense, present the greatest potential for further improvement. Identifying genotypes with higher yield potential and stability under different environmental conditions is crucial for their potential use in sorghum breeding programs. Given that most breeding efforts are time-consuming and costly, it is essential to simultaneously select for multiple traits and examine the interrelationships among phenotypic characteristics [63]. Moreover, comprehending how plants adapt to changing environments, particularly through traits like phenotypic plasticity, is key for predicting their survival, ecological impact and productivity under global climate change pressures. Expanding long-term studies on wild populations will therefore be essential to fully understand how plants respond to environmental fluctuations, including those driven by climate change [64].

In our results, the PCAmix analysis highlighted that plant height, stem diameter, leaf number and leaf area had the greatest contribution to variability among genotypes. These findings are consistent with those reported earlier [18]. Genotypes with stable performance were positioned near the biplot origin, indicating consistent values across both years without extreme deviations. In contrast, genotypes superior in one year exhibited high productivity only under specific environmental conditions. The observed separation between years suggests notable agroecological differences, influencing genotype performance. Additionally, the positioning of genotype groups provides valuable insights into their adaptability and response to environmental conditions. These findings underscore the importance of understanding how genotypes respond to varying environmental conditions, which is fundamental in identifying hybrids with superior resilience and yield stability, particularly for breeding programs aimed at improving sorghum performance under unpredictable climatic conditions. Our results are in agreement with earlier findings [17].

The study conducted on sorghum accessions in Eritrea revealed moderate to high genetic variability for most traits, with plant height, harvest index, panicle length and seedling vigor showing the highest variation. Conversely, days to flowering, days to maturity and number of leaves exhibited low variability. These results suggest substantial potential for effective selection and genetic improvement, with minimal environmental influence enhancing the reliability of these traits for breeding programs [43]. Since biomass has the strongest effect on ethanol production of forage sorghum, selection should focus on tall, thick-stemmed, late-maturing, high-biomass genotypes with favorable traits such as plant height, stem diameter, stalk brix and days to flowering [63].

4.3. Combining Ability as a Tool for Identifying Climate-Resilient Biomass Hybrids

Morphological traits with high stability across diverse climatic conditions provide a fast and effective means for the preliminary evaluation of genetic diversity. These traits also serve as reliable selection criteria in breeding programs, supporting the development of a sorghum ideotype with enhanced yield potential [65].

Our results emphasize the need for further selection of genotypes with enhanced adaptability and the optimization of agronomic practices aimed at stabilizing yield performance. After identifying patterns of trait variation across years and genotype groups through PCAmix analysis and confirming the statistical significance of these differences with ANOVA, we focused on evaluating the performance of newly developed experimental hybrids. Combining the ability analysis of grain sorghum and Sudan grass lines provides a foundation for selecting genotypes that are good general combiners, capable of combining well on average with many different lines. The best specific combinations have also been identified, which exhibit transgressive segregation of traits. This process improves and facilitates breeders’ work, reducing the number of hybridizations required to create new promising hybrids in future breeding efforts.

This analysis highlighted hybrids with the most favorable specific combining ability (SCA) values, indicating their potential for high yield across multiple traits. Notably, hybrids 23, 14, 40, 31, 2, 1, 21, 35 and 8 exhibited positive high SCA values, suggesting their suitability for further selection in breeding programs aimed at stable biomass production under variable environmental conditions. Experimental hybrids such as 2, 21, 23 and 31 exceeded the performance of commercial hybrids in biomass yield and key agronomic traits like stem diameter and leaf area. This highlights their potential for future breeding and commercialization, suggesting that they can offer higher productivity compared to established varieties. Mohammed et al. [66] demonstrated that plant height, grain yield, days to 50% flowering, inflorescence exertion and panicle shape exhibited greater specific combining ability variance, indicating the predominance of non-additive gene action in the post-rainy season. Furthermore, G × E interactions influenced traits like plant height and flowering time, suggesting the need for season-specific breeding strategies for sorghum improvement.

5. Conclusions

These results provide a practical basis for assessing breeding progress and the potential of newly developed genotypes, while also highlighting the need for comprehensive selection strategies that account for climatic variability and trait integration to support the development of resilient, high-yielding forage sorghum hybrids.

Testing a wide set of experimental hybrids along with their parental components under unstable agroecological conditions of the Pannonian climate revealed significant genotypic variability in biomass yield and its components, emphasizing the importance of multi-environmental evaluation in identifying high-performing, adaptable genotypes.

The comparison of the traits between the two growing seasons revealed clear differences that reflect the influence of varying climatic conditions. These findings underscore the importance of integrating multivariate (such as PCA) and univariate (such as ANOVA) statistical approaches with practical hybrid performance assessments, to refine selection strategies and improve breeding efficiency.

Combining ability analysis provided insight into the genetic control of yield-related traits, allowing the identification of superior hybrid combinations with both high specific combining ability and desirable agronomic performance, thereby offering valuable guidance for the selection of genotypes suited to future breeding efforts under climate stress. Specifically, experimental hybrids 2, 21, 23 and 31 demonstrated both superior performances and stability across contrasting years, indicating their potential for cultivation under increasingly unpredictable climate scenarios. Their consistent performance also highlights their value as model genetic material for future studies aimed at uncovering the genetic basis and environmental factors contributing to yield stability.