Evaluation of sesame (Sesamum indicum L.) varieties for drought tolerance using agromorphological traits and drought tolerance indices

Description of the experimental area

The experiment was conducted at Liben senior secondary and preparatory school located in Amhara Region, West Gojjam zone, North Achefer district (Fig. 1). The school is located at 11°41′51″ north latitude and 36°56′35′ east longitude. The area has arid and semi-arid climatic conditions with a soil type suitable for irrigation. The altitude of the district ranges from 1,500 to 1,800 meter above sea level (m.a.s.l). It is also characterized by unimodal rainfall with an average annual rainfall ranging from 1,000 to 1,500 mm. The minimum and maximum daily temperatures are 25 °C and 30 °C, respectively (NADoA, 2013 cited in Demeke, Mekuriaw & Asmare, 2017).

Study plant materials

A total of four sesame varieties were used in this study. All of them are released varieties. These are T-85, Kelafo-74, Mehado-80 and abasena. The former had indeterminate form of growth, while the remaining varieties had determinate growth (Mawcha et al., 2020). The former two were released in 1976 and the later in 1989. They are the main export varieties and well known by their market names as humera-1, gondar-1, wollega and abasena, respectively. Then here after the market names were used. Humera-1 is characterized by whitish, large and sweet taste seeds, high productivity, high shattering property, 45–50% oil content, maturity time of 110–115 days and adapted to the Humera plains (Jaleta, 2012). Gondar-1 has light brown coloration and good uniformity of seed, and a maturity date ranging from 110–120 days. It was released and adapted to the Gode region (Endale, 2017). Similarly, wollega is characterized by its small gray seeds, high oil content (49–56%), low sweetness, maturity date of 100–110 days and released and adapted to irrigated areas of Awash valley (Jaleta, 2012; Endale, 2017). Abasena, on the other hand is characterized by gray, large and sweet taste seeds, high productivity, high shattering property, 44–48% oil content, maturity time of 110–120 days and adapted to high rainfall (Bekora, 2013). Seeds of humera-1, gondar-1 and abasena were collected from Amhara Agricultural Research Institute, Gondar Agricultural Center, and seed of wollega variety was collected from local farmers around the study area.

Experimental design and treatments

The experimental field was tilled using human power. All the weeds and large plant debris were removed manually from the experimental field. Then, it was exposed to sunshine for six days prior to the next tilling. The land was ploughed again and again until smooth soil particles were obtained as this is necessary for the better growth of the crop. The field was ploughed, leveled, ridged and divided into plots before sowing.

The field experiment was carried out using factorial design laid down in a randomized complete block design (RCBD) with three replications. The size of the experimental field was 39 m² (9.75 m × 4 m) consisting of 24 plots each with 1 m² area. The replications were 0.5 m far apart. Each replication consisted of eight plots with 0.25 m spacing to prevent lateral penetration of water. Each plot had three rows with 37 cm spacing. There was 18.5 cm spacing between each sesame plant. Around each plot, 13 cm area was left to avoid edge effect. Four holes were prepared along each row and three seeds were hand sown in each. The sesame varieties were assigned to plots randomly by a lottery method. Thinning to one plant/hole was carried out after complete emergence. N and P inorganic fertilizers were applied at the rate of 100 kg/ha (Zenawi & Mizan, 2019) and 46 kg/ha (Gebremariam, 2015), respectively. All the P and half of the N fertilizers were applied at the time of sowing and the remaining half N at branching stage. The common management practices such as weeding, pesticide and other parasites monitoring and control measures were carried out as needed.

Treatments consisted of four sesame varieties grown under two water levels. The water levels were determined gravimetrically as described by Turner (2018). Bottom perforated pots filled with dry soil were weighed and then watered for adequate period of time. The saturated soil was left overnight to drain the excess water by the force of gravity and then was weighed. At this condition, the soil water was at its field capacity equivalent to pot capacity. The water level at field capacity was considered as well-watered and 50% of the field capacity was taken as water-stressed treatment. The water level of all treatments was maintained to field capacity throughout the vegetative stage, until the imposition of water stress. Water stress was imposed 30 days after sowing (DAS), which represents flower initiation stage. This is because flowering starts 35–45 days after sowing in sesame and continues for 75 to 85 days for early types and 150 days for some varieties to mature (Terefe et al., 2012). Sesame plants grown under well-watered treatment in each plot were irrigated with 333 mL of water for every three days and those in the water-stressed received half of this volume at the same time interval.

Soil analysis

Soil samples representative of the experimental field were randomly collected from different points and thoroughly mixed together to form a composite soil sample. The sample was dried, ground and passed through a two mm sieve and used for further soil property analysis. The pH was determined in 1:2.5 soil/water ratio using pH meter. The total percent of nitrogen was determined by Kjeldahl digestion method, organic carbon content and organic matter were measured by volumetric method as described by Roper et al. (2019). The analysis result revealed that the soil had a pH of 6.5, total percent nitrogen (0.084), percent organic carbon content (0.49) and organic matter (0.85).

Data collection

Growth and physiological data of the study sesame varieties were gathered from three randomly selected and tagged plants in each plot. When the tagged plants were lost due to various reasons, data were collected from other plants in the same plot. Growth data were collected 30 and 60 days after water stress imposition, while physiological data were gathered 30 days after water stress imposition

Growth parameters

Plant height of tagged plants was measured from the ground level to the tip of the youngest leaf and expressed in centimeter (cm). Number of leaves and branches of the same plants were counted and recorded. The length of leaves was measured from the leaf base to the tip and width at the maximum area of the blade. Relative growth rate (RGR) was assessed based on height and determined according to Dodig et al. (2021) as RGR_H = (logH_t2 –logH_t1)/t₂ –t₁ where, H_t1 is height of the plant measured at t₁ and H_t2 is height of the plant measured at t₂, t₁ and t₂ are times of the first and second height measurements, respectively.

Physiological parameters

Relative water content

Relative water content (RWC) was estimated according to Gonzalez & Gonzalez-Vilar (2003). The fully expanded third leaf from the top was collected from tagged plants and weighed producing fresh weight (FW). These leaves were preserved in plastic bags and transported to laboratory. Each leaf was floated on distilled water in a Petri dish for 24 hr. The leaves were then blotted gently with tissue paper and weighed again to get the turgid weight (TW). The leaves were then oven dried at 70 °C for 24 hr and the dry weight (DW) was recorded. Finally, the leaf relative water content was calculated as (FW-DW)/(TW-DW) ×100.

Yield and biomass parameters

Yield, yield related and biomass data were collected after the complete maturation of the crop. The numbers of capsules/pods per plant in each plot were counted from three tagged plants and recorded during data collection period. All the pods from each plant and plot were collected and dried in an oven at 80 °C to constant weight and threshed plot wise. Grain yield was determined for each plot (1 m²) and reported in terms of kg/ha as described by Nadeem et al. (2015).

After harvesting the pods, each plant was uprooted and parts were separated into shoot (stem and root) and root. The roots were carefully detached from soils and thoroughly washed with tap water. The shoot and root parts were dried in an oven at 80 °C to constant weight and weighed separately. Harvest index was calculated from biological yield (shoot and root dry weight) and grain yield in kg/ha according to Nadeem et al. (2015) as harvest index = seed yield/biological yield ×100.

Drought tolerance indices

Drought tolerance indices are important tools that provide better opportunities to select genotypes with good performance under normal and stress conditions. Several of the tolerance indices were calculated as follow:

Stress tolerance index (STI) = Ys*Yp/Y¯p²………………..Ahmed et al. (2020)

Mean productivity (MP) = Ys+Yp/2 ……………Sun et al. (2023)

Stress tolerance (TOL) = Yp-Ys ………………. Anwaar et al. (2020)

Yield index (YI) = Ys/ Y¯p …………Shahrokhi, Khorasani & Ebrahimi (2020)

Geometric mean productivity (GMP) = ……………Adhikari et al. (2019)

Yield stability index (YSI) = Ys/Yp …………………………Kouighat et al. (2023)

Stress susceptibility index (SSI) = 1-(Ys/Yp)/1- (Y¯s/Y¯p)….Anwaar et al. (2020)

% reduction = Yp-Ys/Yp*100 ……………………………Choukan et al. (2006)

Yield potential score index (YPSI) =0.5 (MP + STI) –0.5 (SSI + TOL) …(Malinowska, Donnison & Robson, 2020)

Stress intensity (SI) = 1- (Y¯s/Y¯p ……………….. Zare (2012)

Yield stress score index (YSSI) = 0.5 (STI + SSI) …………………Malinowska, Donnison & Robson (2020)

where, Ys is the yield of each genotype under water-stressed condition; Yp is the yield of each genotype under well-watered condition; Ys¯ is the mean yield of each genotype under water-stressed condition and Yp¯ is the mean yield each genotype under well-watered condition.

Ranking of sesame varieties

Different drought tolerance indices discriminate different sesame varieties as drought tolerant. Thus, identifying drought tolerant genotypes based on a single index does not produce clear results. To identify desirable drought tolerant varieties, the mean, standard deviation of ranks, and rank sum of all indices were calculated. For screening drought tolerant varieties, a rank sum (RS) was calculated by using the following formula:

Rank sum = Mean ranks (MR) + standard deviation of ranks (SDR)…Noorifarjam, Farshadfar & Saeidi (2013)

Data analysis

All the data were analyzed using the Statistical Package for Social Sciences (SPSS, version 23). Multiple comparisons of means were carried out with Tukey HSD test to see variations between treatments at 30 and 60 days after drought imposition. Factorial Analysis of Variance (ANOVA) was conducted for each trait under WW and WS conditions and results were considered significant at p < 0.05. Furthermore, principal component analysis (PCA) was used to characterize trait variation. PCA and correlation analysis between grain yield and drought tolerance indices were analyzed with R software. The biplot was generated using factor analysis and data processing with R package.

Yield and its attributes of sesame varieties

The sesame varieties demonstrated a statistically significant variation in biomass production, yield and yield related traits under the two water levels (WW and WS) at p < 0.05. A significant reduction in shoot biomass, root biomass, biological yield, the number of pods per plant and seed yield was recorded under WS condition among the sesame varieties. These traits showed an average reduction of 52.2, 72.5, 54.0, 51.9, and 52.8%, respectively compared to WW condition. The highest reduction in yield and yield related traits between WW and WS conditions was found in wollega followed by humera-1, while the lowest was found in abasena followed by gondar-1 (Table 3, Table S3). There was insignificant difference in shoot biomass, harvest index and seed yield between water levels in abasena variety (Table 3, Table S3). Significant differences were also recorded among varieties both under similar and different water levels (Table 3, Table S3). Accordingly, wollega, gondar-1 and humera-1 varieties produced considerably greater shoot biomass and biological yield than abasena under WW condition. The variation in shoot biomass between the varieties under WS condition was insignificant at p < 0.05. All varieties under the WS condition had significantly lower root biomass than those in WW condition, implying that enhanced root growth and the consequently increased root/shoot ratio is not a drought tolerance strategy in the sesame varieties in the area under study. In this regard, abasena variety had a relatively greater root biomass than the others under WS condition (Table 3, Table S3). Sesame plants of wollega variety produced a relatively higher biological yield followed by gondar-1 both under WW and WS conditions than others (Table 3, Table S3).

In terms of yield and related traits, gondar-1 variety had greater number of pods per plant followed by humera-1 and wollega varieties under WW and WS conditions (Table 3, Table S3). Wollega sesame variety produced a significantly greater seed yield (kg/ha) under WW condition followed by humera-1 and gondar-1 varieties. Correspondingly, sesame variety of gondar-1 had better seed yield than others followed by humera-1 under WS condition. However, the variation among the varieties in seed yield (kg/ha) under WS condition was insignificant. Wollega sesame variety had higher harvest index under WW condition attributed to the highest seed yield (kg/ha), while a relatively higher harvest index was recorded under WS condition in abasena and gondar-1 varieties, which might be due to the lower biological yield caused by the WS condition. Plants of abasena performed the least in most parameters under both water levels (Table 3, Table S3).

When variety, water levels and their interactive effects were analyzed, they produced a statistically significant difference in biological yield, number of pods per plant, Ys, Yp, shoot and root biomass at p < 0.01 (Table 4). Harvest index was not significantly affected by variety and water levels. Similarly, the interactive effect (variety * water levels) had no remarkable effect on number of pods per plant. The partial eta squared value revealed that water levels had greater effect on Ys, Yp, biological yield, number of pods per plant, shoot and root biomass followed by variety (Table 4).

Drought tolerance indices

Analysis of variance of Ys, Yp and drought tolerance indices showed a highly significant difference among sesame varieties under the different water levels, indicating the presence of high genetic variability. Drought tolerance indices were calculated on the basis of seed yield (kg/ha) under WW (Yp) and WS (Ys) conditions. The mean seed yield under WW conditions ranged from 2,144.44 to 5,177.78 kg/ha, while 1,344.44 to 1,811.11 kg/ha was obtained under WS condition. Wollega sesame variety had the highest and the second lowest mean seed yield (kg/ha) under WW and WS conditions, respectively (Fig. 2). Gondar-1 and humera-1 varieties achieved the highest seed yield (kg/ha) under WS condition. In both water levels, plants of abasena performed the least and the variation in yield (kg/ha) between WW and WS conditions was found to be significantly lower than the differences seen in other sesame varieties (Fig. 2).

The sesame varieties in this study showed a significant difference in drought tolerance indices (Table 5). Wollega sesame variety had significantly higher MP, GMP and TOL than others followed by humera-1 variety. According to these indices, wollega and humera-1 varieties were found to be more desirable and comparatively drought tolerant. Unlike this, wollega variety had lower STI than gondar-1 and humera-1 varieties (Table 5). The mean STI values demonstrated that gondar-1 and humera-1 had better drought tolerance characteristics. In the present study, abasena sesame variety had the lowest TOL, SSI and the highest YI and YSI of all the study varieties (Table 5). The difference between Ys and Yp values in abasena variety was the lowest indicating the lower sensitivity of the variety to drought stress. Varieties with lower TOL, SSI, and higher YI, YSI values are more stable and tolerant, respectively under drought conditions. Based on these indices thus, abasena could be identified as stable and tolerant variety. Plants of wollega variety demonstrated a significantly higher percent reduction than abasena that showed the lowest reduction. The studied varieties revealed a significantly different stress intensity of which the weakest was seen in abasena variety. Similarly, plants of gondar-1 and humera-1 had higher YSSI and YPSI values compared to others (Table 5).

Correlation

A strong association between Yp, Ys and drought tolerance indices is an important indicator to select the most desirable drought tolerance indices. Ys had positive and significant correlation with Yp. Moreover, Ys revealed a significantly positive correlation with MP, STI, GMP, YSSI and YPSI at p < 0.01 (Table 6). Similarly, Yp had significantly positive association with MP, STI, GMP, TOL, percent reduction, SI and YSSI at p < 0.01 probability level (Table 6). This showed that MP, STI, GMP and YSSI had significantly higher correlation with yield under WW and WS conditions so that these indices can be used as criteria to select varieties for drought tolerance. According to MP and GMP values, wollega and humera-1 were found to be the most desirable and drought tolerant varieties, respectively, while gondar-1 and humera-1 had higher STI and YSSI indicating higher yield performance under drought conditions. Furthermore, TOL had positive and significant correlation with MP, GMP, percent reduction, SSI and SI at p < 0.01. Its correlation with YI, YSI and YPSI was negative and significant (Table 6).

Principal component analysis

The minimum number of principal components that accounted for most of the variations in drought tolerance indices was determined based on eigen value. Principal components with eigen values greater than 1 were selected. Accordingly, two principal components fulfilled the acceptable level/variance of the dataset. The first and second principal components (PC1) and (PC2) explained 86.2% and 12.7% of the variation, respectively. The PC1 and PC2 together covered 98.9% of the variation in the dataset (Table 7). The results of the analysis showed that PC1 had large positive association with Yp, TOL, MP and GMP. Others such as YI, YSI and YPSI had negative correlation with PC1. PC2 had large positive association with YPSI, Ys, STI, GMP, MP and YSSI, while negatively correlated with TOL (Table 7). The large positive and negative values of drought tolerance indices indicate the strong effect of each index on each principal component.

The biplot analysis showed that Yp, TOL, GMP, MP, Ys and YSSI orient in similar direction and had tight angles indicating a positive correlation and strong effect on PC1. Similarly, YPSI, Ys, STI and GMP contributed higher proportion to PC2 (Fig. 3). The biplot analysis clustered the study varieties roughly into two groups as low yielding (abasena) and relatively above average performing varieties (humera-1, gondar-1 and wollega) under WW and WS conditions (Fig. 3). This was corroborated with the values of STI, MP and GMP where the varieties in the second group had higher average yield values under WW and WS conditions (Table 7 and Fig. 3).

Ranking of sesame varieties

Rank sum was calculated due to the fact that identification of genotypes for drought tolerance using individual indices is difficult. Different indices identified different sesame varieties for drought tolerance. Rank sum for all indices was used as an indicator to select the best varieties. Lower and higher rank sum values indicate high drought tolerant and susceptible genotypes, respectively. Accordingly, humera-1 sesame variety had the lowest rank sum value followed by gondar-1 variety. These varieties represented the best varieties with the highest performance under WW and WS conditions (Table 8). The rank sum result identified abasena sesame variety as the most susceptible and low yielding variety (Table 8).