Optimizing sowing time and weather conditions for enhanced growth and seed yield of chia (Salvia hispanica L.) in semi-arid regions

Weather details of the study location

Field trials were conducted for two consecutive years (2021–22 and 2022–23) at ICAR–National Institute of Abiotic Stress Management (NIASM), Baramati, Pune, Maharashtra, India (Fig. S1). The study site is positioned at 18.15850556°N and 74.50085556°E at an elevation of 570 m above sea level (MSL). This region falls within the hot and semi-arid zone of the Deccan Plateau region, which is known as the water scarcity zone. The mean maximum and minimum temperature of the region was 31.2 °C and 21.9 °C, respectively. The region receives an average annual precipitation of 576 mm, a major portion (75%) is received between August and October (Harisha et al., 2023). The annual open-pan evaporation rate of the region is 1,965 mm, which is three times more than annual rainfall. The detailed weather parameters for the cropping seasons (2021–22 and 2022–23) are outlined in Fig. 1 and Table S1. The weather data on maximum temperature (Tmax), minimum temperature (Tmin), bright sunshine hours (BSS), open pan evaporation, rainfall and relative humidity (RH) for the location during the cropping season was obtained from the weather observatory of ICAR-NIASM, Baramati.

Soil details of the experimental site

The soil type of the experimental site was shallow basaltic with 81.9% sand, 10.4% silt, and 7.5% clay exhibiting low water holding capacity (Rajagopal et al., 2018). The chemical properties of the soil were pH (7.48), an electrical conductivity (0.21 dS m⁻¹), a moderate level of organic carbon (6.5 g kg⁻¹), low available nitrogen (81.2 kg ha⁻¹), phosphorus (3.6 kg ha⁻¹ as P₂O₅), and potassium (80.0 kg ha⁻¹ as K₂O).

Experimental details and crop management

The experiment consists of two factors; dates of sowing and chia morphotypes were laid out in split-plot design with three replications. Fifteen dates of sowing (S1; 1st July, S2; 15th July, S3; 1st August, S4; 15th August, S5; 1st September, S6; 15th September, S7; 1st October, S8; 15th October, S9; 1st November, S10; 15th November, S11; 1st December, S12; 15th December, S13; 1st January, S14; 15th January, and S15; 1st February) were treated as main factor and two chia types (white and black) as subfactor. The plots of size 3 m × 2.5 m were prepared for sowing the seeds of chia types (2.5 kg ha⁻¹) after mixing with sand in 60 cm wider rows. Subsequently, excess and weak plants were removed by retaining one healthy and maintained a uniform distance of 20 cm between plants within rows. Recommended nutrients (N:P₂O₅:K₂O at 90:60:75 kg ha⁻¹) were applied through fertilizers such as urea, di-ammonium phosphate, and muriate of potash (Harisha et al., 2023). The full dose of P₂O₅ and K₂O, and 50% of N was applied during field preparation as a basal, whereas the remaining 50% of N (45 kg/ha) was top dressed through urea in three splits at 30 days after sowing (DAS) (15 kg/ha), 45 DAS (15 kg/ha), and 60 DAS (15 kg/ha). The urea was applied in band placement five cm away from plants and light raking was done to mix it with soil. The chia was cultivated under rainfed conditions with supplemental irrigation. During extended dry spells, supplemental irrigation was scheduled based on soil drying and visible crop moisture deficit symptoms during both the rainy and winter seasons, at a depth of five cm at weekly intervals using a drip system. Weeds were controlled by manual hand weeding. However, the crop remained unaffected by pests and diseases during both cropping periods.

Measurement of chia morphological parameters and phenology

Chia growth attributes such as plant height and dry biomass production were recorded at harvest from five randomly selected plants, separately in each treatment. Floral characters such as days to flower bud appearance (FBA), completion of flowering, and maturity were recorded from randomly selected five plants as per the procedure outlined by Brandan, Curti & Acrechea (2019). A day to 50% flowering was recorded treatment wise when 50% of plants opened their first flower. Likewise, growing degree days (GDD), also called heat unit, accumulated up to maturity were calculated for each sowing date as suggested by Nuttonson (1957). ${\displaystyle \mathrm{GDD}=\frac{{\mathrm{T}}_{\mathrm{Max}}+{\mathrm{T}}_{\mathrm{Min}}}{2}-{\mathrm{T}}_{\text{base}}.}$

T_max is maximum temperature, T_min is minimum temperature, T_base is base temperature (10 °C) Ayerza & Coates (2009).

Likewise, heat use efficiency (HUE) indicates the capacity of a plant to produce yield per unit of heat used. HUE of the chia crop was calculated using the formula suggested by Singh & Khushu (2012). ${\displaystyle \mathrm{HUE}\left(\mathrm{kg}{\mathrm{ha}}^{-1}{°\mathrm{C}}^{-1}{\mathrm{day}}^{-1}\right)=\frac{\text{Grain yield}\left(\mathrm{kg}{\mathrm{ha}}^{-1}\right)}{\text{Accumulated GDD}\left(°\mathrm{C day}\right)}}$

Seed yield and yield attributes of chia

Yield determinants of chia such as the number of spikes per plant, spike length, seed yield per spike and 1,000 seed weight were recorded from five randomly selected plants from each treatment (Harisha et al., 2024). Then, seed yield was determined by recording the seed weight from fifty plants in the plot of 7.5 m² and sun dried for 3–4 days to attain moisture content of 7 ±  0.5% and expressed in kg ha⁻¹. Likewise, plot wise dry biomass yield was determined from randomly selected five plants after sun drying for 2–3 days followed by oven drying at 63 °C for 72 h to attain constant weight and expressed as dry biomass kg ha⁻¹. Later, the harvest index (HI) was calculated based on the seed and biological yield of chia. ${\displaystyle \mathrm{HI}\left(\text{\%}\right)=\frac{\text{Grain yield}\left(\mathrm{kg}{\mathrm{ha}}^{-1}\right)}{\text{Dry Biomass}\left(\mathrm{kg}{\mathrm{ha}}^{-1}\right)}}$

Later, grain filling duration (GFD) was calculated considering the number of days between 50% flowering and physiological maturity. Similarly, the grain filling rate was calculated by dividing seed yield with grain filling duration as explained by Sattar et al. (2023) in wheat.

Statistics

Before conducting an analysis of variance, the data recorded on various growth, phenology, and yield parameters of chia during both years was tested for normality by the Shapiro–Wilk test using the PROC UNIVARIATE procedure in SAS 9.3 (SAS Institute, Inc., Cary, NC, USA). Then, normal data was subjected to analysis of variance (ANOVA) using the mixed model (proc GLIMMMIX in SAS v 9.3). Chia morphotypes, year, and sowing dates were considered as fixed effects and replications as random effects. Post-hoc test was conducted to compare the difference (α = 0.05) using Tukey’s honest significant difference (HSD) test. Further, Pearson’s correlation coefficient was used to describe the association between weather parameters (T_max, T_min, RH, accumulated GDD, bright sunshine hours, and rainfall), vs grain yield, days to flower bud appearance, flowering duration, and maturity (Gomez & Gomez, 1984). To interpret multi-environment (sowing dates × weather parameters × chia types) interaction, principal component analysis (PCA) biplot analysis was carried out using R software (version 4.2.3) (Gopinath et al., 2021).

Chia growth and floral phenology

Growth determinants such as plant height and biomass accumulation in chia morphotypes differed significantly (p < 0.05) across sowing dates (Figs. 2A–2B). Among chia types, black seeded plants were found to be more vigorous with greater height (119.6 cm) and biomass accumulation (2,883.9 kg ha⁻¹) compared with white seeded plants (117.3 cm, and 2,662.2 kg ha⁻¹ respectively). Regarding fifteen sowing dates, early sowing (S1: 1st July, S2: 15th July, S3: 1st August) demonstrated the highest plant height (199.1 cm, 195.1 cm, and 185.3 cm respectively,), and biomass production (4,294.2–4,021.9 kg ha⁻¹) compared to other sowing dates (Figs. 2A–2B). Whereas delayed sowing after S3 up to S15 conspicuously reduced the plant height and biomass accumulation (1,735.1–1,899.4 kg ha⁻¹) in chia types.

Similarly, floral phenological events such as days to flower bud appearance (FBA), days to 50% flowering, days to completion of flowering, days to maturity and flowering duration have responded (p < 0.05) to dates of sowing (Table 1 and Table S2). Notably, flowering phenology did not differ among white and black seeded chia morphotypes. Whereas, early sown plants (S1 and S2) took more days to FBA (70.5–78.2), and it was drastically reduced to 35.0 days in late sown conditions (S7: 1st October). Further delay in sowing after S8: 15th October to S15: 1st February gradually delayed the FBA (54.8 days). Similarly, days to 50% flowering, days to complete flowering and days to maturity followed a similar trend as that of FBA (Table 1). The flowering duration was significantly delayed in late sown conditions (S13 to S15; 63.6 to 77.5 days) over other sowing dates. The shortest flowering duration of 47 days was observed in S8 and S9 sowing conditions. Moreover, early sown conditions (S1 to S4) enhanced the grain filling duration (39.8 to 41.8 days) with a decreasing trend up to S11 and a subsequent increase up to S15. Across years of cultivation, the second year (2022–23) noticed maximum plant height (122.4 cm), and biomass accumulation (2,934.7 kg ha⁻¹) with delayed flowering duration, grain filling duration and maturity (115.3 days).

Relation between prevailing weather parameters and flowering phenology of chia

Weather conditions during the vegetative phase (germination to bud appearance) strongly influenced the flowering phenology of chia (Fig. 3A). The Pearson’s correlation suggested that FBA exhibited a positive correlation with day length (r = 0.7), accumulated GDD (r = 0.87), T_min (r = 0.42), and RH (r = 0.38). While FBA was negatively related to diurnal temperature difference (T_diff) (r =  − 0.38) and bright sunshine hours (BSS) (r =  − 0.43). The flowering duration had a positive correlation with day length (r = 0.85), T_max (r = 0.79), T_min (r = 0.59), and accumulated GDD (r = 0.84) prevailed during flowering phase (flower initiation to completion). However, flowering duration was negatively correlated with RH prevailing during the flowering phase (r =  − 0.64) (Fig. 3B).

Yield attributes and seed yield of chia

Yield attributes of chia morphotypes responded to sowing dates during two years of investigation (Table 2). Black seeded chia types produced more spikes per plant (30.3), longer spike length (17.99 cm), higher 1,000 seeds weight (1.15 g), HUE (0.37 kg ha⁻¹ °C⁻¹ day⁻¹), grain filling rate (17.8 kg ha⁻¹ day⁻¹), and seed yield (564.6 kg ha⁻¹) compared to white types. While white seeded morphotypes maintained a greater harvest index (21.32%) across sowing dates. Within sowing dates, treatments (S3–S6; 1st August–15th September) maintained a greater number of spikes, spike length (20.1–21.71 cm), 1,000 seeds weight (1.15–1.16 g), and seed yield (741.0–811.0 kg ha⁻¹) with greater HUE, grain filling rate, and HI. In contrast, delayed sowing after S7 to S15 adversely influenced the HUE, grain filling rate, and seed yield of chia morphotypes. Regarding year effect, the first year (2021–22) recorded a superior number of spikes, 1,000 seed weight, HUE, grain filling rate, and seed yield (579.2 kg ha⁻¹) over 2022–23 (Table 2). Further, the triple interaction between sowing dates, chia types, and year showed a significant effect on seed yield and HUE. In 2021–22 black seeded chia produced higher seed yield in S3–S6, whereas in 2022–23 it was higher in sowings S3–S4 (Table S3). Therefore, sowing up to 15th September could favour the seed yield and heat use efficiency of chia morphotypes in semi-arid conditions.

Weather parameters vs yield attributes of chia

Weather parameters across the growing period up to maturity established a significant (p < 0.05) relation with yield attributes of chia. The seed yield was positively influenced by RH (r = 0.93 and R² = 0.856), HUE (r = 0.9), and rainfall (RF) (r = 0.76 and R² = 0.529). However, T_max (r =  − 0.82 and R² = 0.674), T_diff (r =  − 0.87 and R² = 0.856), accumulated GDD (r =  − 0.31), BSS (r =  − 0.84) were negatively influenced the seed yield (Figs. 4 and 5A–5D). Besides, T_max is negatively related to chia yield attributes; seed yield per spike (r =  − 0.76), spike length (r =  − 0.65), and 1,000 seed weight (r =  − 0.58) (Fig. 4). Notably, RH during the entire cropping period displayed strong positive associations with chia yielding traits; (r = 0.71 to 0.85). Analysis of diurnal temperature difference revealed a negative correlation with all growth and yield-related traits of chia. Moreover, the relation between seed yield and plant traits was also found to be significant (Fig. 4). Seed yield exhibited positive correlations with plant height (r = 0.73), spike length (r = 0.78), number of spikes per plant (r = 0.80), number of branches per plant (r = 0.76), seed yield per spike (r = 0.89), and 1,000 seed weight (r = 0.74). Conversely, seed yield showed a negative correlation with flowering duration (r =  − 0.66) and crop duration (r =  − 0.28). Hence, prevailing weather parameters had a considerable role in determining the growth and yield of chia morphotypes.

Interaction between yield traits and weather parameters due to sowing dates and chia morphotypes

Multivariate analysis was conducted to elucidate the interaction, inter-relationship and variations among various yield traits and weather parameters that prevailed during the entire chia duration and chai types. Principal component analysis (PCA) revealed that the first two components (PC1 and PC2) captured 94.1% of the total variability (Fig. 6A). In PC1, traits such as seed yield, spike length, number of spikes, seed weight per spike, plant height, and biomass production demonstrated strong positive associations as indicated by the narrow angles between their vectors. Similarly, weather parameters RH, RF, and T_min showed strong positive associations with seed yield and yield-related traits. These variables explained the maximum total variability as evidenced by the length of their vectors. Conversely, BSS, flowering duration, and T_max exhibited negative associations with each other. In PC2, variables such as accumulated GDD, day length, flowering duration, days to 50% flowering, days to maturity, and FBA were positively associated with each other but negatively influenced the seed yield (Fig. 6A). Furthermore, sowing times (S4, S5, and S6) in both black and white seed types were closely related to higher seed yield, and yield traits as favoured by weather parameters like RH and rainfall. Conversely, delayed sowing times (S13, S14, and S15) in both varieties coincided with intense sunshine hours, poor RH, and higher T_max resulting in longer flowering duration, and accumulated GDD negatively determined the seed yield and yield traits of chia (Fig. 6B).

Growth parameters of chia

Black seeded chia morphotype was found to be more vigorous over white seeded owing to greater plant height and biomass accumulation. This might be due to its superior genetic characteristics and adaptation as described by Grimes et al. (2018) and Guttedar et al. (2023). Early sowing during the rainy season (July S1–S2) resulted in higher plant height and biomass accumulation because of long day conditions (average day length; 12.5 h and accumulated GDD; >2,000 °C) (Fig. 1C), led to more vegetative growth and delayed reproductive growth over subsequent sowing dates (Guttedar et al., 2023). This was also due to the receipt of sufficient rain, and the prevailing ideal temperature around ∼30 °C during the vegetative phase favoured the growth and biomass accumulation in both types of chia, thus increasing the risk of lodging. Our findings corroborate the results of Goergen et al. (2018) in chia, where higher GDD and longer photoperiod increase plant height and biomass. Similarly, Silva et al. (2018) reported enhanced vegetative growth in chia due to a greater number of branches during early sowing. Whereas shorter plants with reduced biomass accumulation in case of delayed sowings after December (S11) to February (S15) were attributed to prevailing dry weather (high temperature and low RH) with the least rainfall during active growth stages (Figs. 1A–1C). The crop biomass production is closely associated with dominant environmental factors such as temperature, RH, and rainfall, which together determine crop duration (Guttedar et al., 2023). Thus, chia is very sensitive to day length, RH, and temperature, which determines its biomass accumulation and yield.

Flowering phenology and maturity in chia

The delayed FBA in chia during early (S1–S3) and delayed sowings (S14–S15) was possibly due to longer day length conditions (>12.5 h) compared to intermediate sowings (S4–S13) with shorter day lengths (<12 h). As a result, flowering duration was extended (56.6 to 77.5 days) owing to more number of days between FBA and completion of flowering. The positive correlation between the flowering duration and T_max during flowering phase indicates the potential cause for delayed flower opening due to high temperatures (Fig. 4). It is important to note that hot weather (high temperature; >34 °C, low RH; <50%, and no rainfall; Fig. 1), along with the commencement of long days during flowering phase (March–April) in delayed sowing resulted in delayed FBA, and also the conversion of floral structures into vegetative parts in chia (Guttedar et al., 2023). Whereas, delayed flowering in early sowing (S1–S2) was probably related to long day conditions associated with higher RH, rainfall and accumulated GDD during the vegetative stage. Similarly, Grimes et al. (2018) and Benetoli da Silva et al. (2020) highlighted that alterations in chia phenology are primarily linked to fluctuations in RH and higher GDD. Brandan, Curti & Acreche (2020) also reported that the growing degree days are longer than the pre-flowering phase. Therefore, aligning chia flowering with optimal RH and rainfall conditions, synchronizing flower opening and ensuring a shorter flowering duration are crucial for efficient resource utilization and mitigating high temperatures and long days with higher accumulated GDD (Foulkes et al., 2011; Sylvester-Bradley, Riffkin & O’Leary, 2012).

Days taken for flower opening and its completion decide the duration of crop maturity. In the present study, early sowings as well as delayed sowings extended the chia maturity (125 to 143 days) compared to intermediate sowings (93 to 114 days), primarily due to delayed FBA, and flowering duration in chia. Similarly, Lobo et al. (2011) in Tucumán, Argentina and Baginsky et al. (2016) in Las Cruces, Chile, demonstrated that January sowing resulted in delayed flowering (105–111 days) and crop maturity (160–170 days respectively). Subsequently, grain filling duration (between 50% flowering and maturity) was extended with early and delayed sowing dates. A similar trend was observed with the completion of flowering (Jamboonsri et al., 2012; Sattar et al., 2023). Further, chia maturity was slightly delayed in 2022–23. This delay was likely attributed to increased accumulated GDD, higher RH and the occurrence of rainfall, which fostered enhanced vegetative growth. Thus the crop has taken more days to complete flowering and extended grain filling duration, as a positive correlation was observed between FBA, RH, rainfall, and day length (Fig. 3A). A similar pattern of extended maturity and grain filling duration was found in lentil (Jamboonsri et al., 2012; Maphosa, Preston & Richards, 2023). Both white and black seed chia types did not differ with respect to flowering phenology and maturity. Therefore, chia, being a short-day tropical plant, thrives well under photoperiods of less than 12.5 h of light.

Seed yield and yield attributes of chia

Black seeded chia morphotypes produced greater seed yield (564.6 kg ha⁻¹, 10.8% higher) than the white type (509.2 kg ha⁻¹). This improvement in seed yield with black types could be attributed to improved biomass accumulation and yield-contributing parameters such as number of spikes, spike length, and 1,000 seed weight (1.05 g). Previous researchers have noticed the genetic variation and superiority of black seeded chia types for yielding characters because of their wider adaptability (Ayerza & Coates, 2009; Guttedar et al., 2023). In this investigation, the seed yield of chia varied from 47.6 to 811.0 kg ha⁻¹ across fifteen sowing dates. The higher seed yield with mid sowing dates (S3–S5) was mainly due to improved yield contributing parameters (Table 2 and Fig. 3B). Similar associations between seed yield and traits such as the number of spikes, spike length, and harvest index have been reported in both black and white types of chia (Baginsky et al., 2016). The positive relation between flower dry weight and seed yield in chia was also reported by Brandan, Curti & Acreche (2020). Despite congenial RH, rainfall and temperature, early sowing dates (S1 and S2) produced lower seed yield because the plants produced a lower number of spikes because of more height and canopy spread due to prolonged vegetative phase under long days. This might hinder the production of branches, inflorescences, and subsequent translocation of photosynthates towards seed filling. This concurs with the finding of Han et al. (2006) in soybean, where overcrowding canopy leads to poor branching with less number of pods. Interestingly, delayed sowing after S5 to S15 drastically reduced the seed yield of chia, ranging between 8.63–94.13% (Table 2). The poor chia seed yield was primarily due to the underdevelopment of yield governing traits, as reported by Guttedar et al. (2023), that delayed sowing (October) reduced the seed yield in Indian conditions. Therefore, it is crucial to complete sowing by 1st August (S3) to 1st September (S5) to achieve higher seed yield (790–811 kg ha⁻¹) in chia.

Weather and yield attributes of chia

This study among a few, clearly deciphered the impact of weather parameters in determining the chia yield across various (fifteen) sowing windows in a year. The biplot analysis confirmed that prevalence of optimum temperature (30–31 °C), rainfall (200–350 mm), and RH (67–72%) during S3–S5 sowing (1st August to 1st September) resulted in higher seed yield attributes (Fig. 6B). Our findings are in conformity with the results of Grimes et al. (2018) and Benetoli da Silva et al. (2020) that climatic requirements of moderate to high temperature, minimum temperature (<10 °C) with adequate rainfall enhanced the chia yield in Germany and Brazil. Meanwhile, delayed sowings (S10–S15) reduced the 1,000 seed weight of chia, mainly due to higher temperatures during the flowering phase, leading to prolonged flowering and grain-filling durations. This might also affect pollination, resulting in grain shrinkage due to the production of reactive oxygen species, reduced pollen tube development, increased pollen mortality, and grain abortion (Nawaz et al., 2013; Dubey et al., 2019). Thus, prolonged flowering and maturity durations negatively influenced the seed yield, owing to non-synchronized flowering, resulting in poor seed setting and seed yield per spike, as evidenced by a negative correlation between seed yield per spike and flowering duration (r =  − 0.56). Therefore, the increased temperature during the grain filling period increases the percentage of chaffy seed formation, as found in cases of soybean and rice (Borowska & Prusiński, 2021; Sanwong et al., 2023).

Similarly, the reduced grain filling rate in chia might be due to increased thermal load, which is manifested in terms of higher accumulation of GDD leading to higher grain filling duration in early (S1–S4) and delayed sowings (S8–S15). These sowing dates also decreased the HUE in chia, despite higher seed yield in S3 and S4, poor HUE was possibly related to more accumulation of GDD. Similar findings were reported in wheat, where very early and delayed sowings significantly reduced the thermal use efficiency (Kaur & Pannu, 2008; Singh, Kingra & Singh, 2016). It’s also found that exposing crops to higher temperatures at critical growth phases tends to affect the phenophase duration, HUE and yield (Parya et al., 2010). The sowing dates; S4–S6 result in a medium flowering phase (50–51 days), and maturity (98.8–114.4 days), leading to ideal plant height and biomass accumulation in chia over early and delayed sowing dates. Therefore, chia sowing from 1st August to 15th September coincides with optimum RH, lower diurnal temperature difference, and rainfall, favoured the yield attributes and seed yield of chia in semi-arid conditions.