Mapping metabolic modules to chemotypic variation in Iranian Anethum graveolens L. ecotypes under harvest times

Abstract

The current research investigated three Iranian dill (Anethum graveolens L.) ecotypes from Ardabil, Bushehr, and Kerman to evaluate the effects of harvesting times on essential oil yield, composition, and chemotypic variation. Aerial parts of the ecotypes were collected at three diurnal times (6 am, 12 noon, and 6 pm). Following grinding, essential oils were extracted from the dried powder (50 g in 1:10 w/v distilled water) using a Clevenger-type apparatus for 4 h, then collected, dehydrated over anhydrous sodium sulfate, and stored in amber vials at 4 °C prior to GC-MS analysis. Factor analysis revealed four factors explaining of the variation, reflecting distinct chemotypic patterns: (i) a thujene-terpinene-germacrene-myristicin versus pinene-dominated axis, (ii) an oxygenated monoterpene and bioactive compound axis, (iii) a sesquiterpene versus monoterpene/phenylpropanoid trade-off, and (iv) a pinene-type monoterpene axis. These factors demonstrated coordinated and independent regulation of monoterpene, sesquiterpene, and phenylpropanoid pathways, highlighting both metabolic trade-offs and synergistic accumulation within biosynthetic modules. The treatment-by-trait biplot explained 74% of total variance and revealed strong treatment-by-trait interactions, indicating that ecotype and harvest time jointly determine essential oil profiles. Ardabil ecotype favored monoterpene hydrocarbons and carvacrol, especially at 6 am and 12 noon, while Kerman ecotype exhibited higher levels of sesquiterpenes and phenylpropanoids depending on collection time. Positive and negative correlations among compounds reflected coordinated biosynthetic fluxes and antagonistic relationships, enabling chemotype differentiation and identification of functionally enriched oils. The results emphasize the importance of integrating genetic, environmental, and temporal factors for optimizing essential oil composition in A. graveolens L.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-58098-6.

Introduction

The global demand for safe, healthy food products drives interest in medicinal and edible plants, valued for nutritional, sensory, and bioactive properties^1,2. Also, plants that produce bioactive secondary metabolites are particularly significant because these compounds contribute to human health, food preservation, and industrial applications, while generally being safer and more environmentally sustainable than synthetic chemicals. Dill (Anethum graveolens L.), as a leafy aromatic herb, originated from Europe and Asia and is commonly found across the Mediterranean region³. It is cultivated extensively for its culinary, medicinal, and industrial uses, owing to its diverse phytochemical composition. Dill produces a wide array of secondary metabolites, particularly essential oils, which have multiple pharmacological capabilities, like antioxidant, antimicrobial, anti-inflammatory, and preservative effects. Extracts from dill seeds have also demonstrated antimicrobial activity and food-preserving properties, making them valuable for industrial applications^4–6. The plant also provides proteins, fiber, vitamins, and minerals, enhancing its nutraceutical value^7,8. Importantly, A. graveolens-based products are increasingly preferred as biodegradable and safe alternatives to synthetic additives⁹.

Multiple factors, including plant part, developmental stage, harvesting time, environmental conditions, genetic background, and cultivation practices, influence the composition and yield of dill essential oils. According to reports, the vegetative stage is characterized by the dominance of α-phellandrene, β-phellandrene, limonene, and p-cymene, whereas the flowering stage is characterized by carvone, p-cymene, and dill ether^10,11. These variations indicate the importance of harvesting at appropriate growth stages to optimize the concentration and quality of essential oils. Among these factors, harvesting time has a particularly significant impact on essential oil yield and composition. Diurnal variations affect metabolic activity, enzyme function, and the accumulation of volatile compounds, resulting in measurable differences in essential oil concentration and profile depending on whether harvesting occurs in the morning, at noon, or in the evening. Optimizing harvest timing is therefore critical not only for maximizing aromatic intensity and bioactive potential but also for enhancing industrial value and therapeutic efficacy¹². Several studies have shown that harvesting time can influence both the quantity and chemical profile of secondary metabolites, affecting antioxidant capacity, antimicrobial activity, and the concentration of bioactive monoterpenes and sesquiterpenes in medicinal plants^13,14. Despite existing knowledge, systematic evaluation of native ecotypes under controlled harvest schedules remains limited. This study hypothesizes that essential oil yield, and essential oil compounds quality and quantity in native A. graveolens ecotypes is significantly influenced by diurnal harvest time, with specific timing optimizing bioactive compound concentrations. The novelty lies in correlating discrete harvest times (6 am, 12 noon, and 6 pm) with essential oil profiles across three distinct Iranian ecotypes (Ardabil, Bushehr, and Kerman), providing targeted data for cultivation and industrial use.

Materials and methods

Plant materials and experiments

Seeds of three distinct A. graveolens L. ecotypes, originating from the Iranian provinces of Ardabil, Kerman, and Bushehr, were obtained from the Iranian Seed and Plant Improvement Institute. The experimental field was fallow in the year preceding cultivation. Taxonomic identity of each ecotype was verified by the Iranian Seed and Plant Improvement Institute experts and by the botanists at the University of Mohaghegh Ardabili. Voucher specimens were deposited under accession numbers TN-Ag-430 (Ardabil ecotype), TN-Ag-421 (Kerman ecotype), and TN-Ag-470 (Bushehr ecotype) at Department of Horticultural Science of the University of Mohaghegh Ardabili (Ardabil, Iran). The regional climate featured a mean annual temperature of 9 °C, 303.4 mm of annual precipitation, 127 frost-free days per year, and an average relative humidity of 70% (data obtained from the Ardabil Provincial Meteorological Department). Soil sampling and analysis were conducted prior to sowing. Soil samples were collected from the top 0–30 cm layer and analyzed for texture, pH, electrical conductivity (EC), and iron (Fe) content.

Experimental design and cultivation practices

After viability testing, uniform seeds were sown directly into prepared rows. Initial seeding employed appropriate plant densities, after which seedlings were thinned post-germination to retain three to four robust plants per experimental block, ensuring adequate spacing. Standard agronomic practices were applied, including irrigation twice per week. No chemical pest, weed, or pathogen control was necessary during the growth period. The experiment was arranged in a factorial design based on a randomized complete block design (RCBD) with three replications. The use of three replications is standard in RCBD for agricultural experiments, providing a robust foundation for statistical analysis of variance while accounting for field variability in a logistically feasible manner. The experimental factors were ecotype (Ardabil, Kerman, and Bushehr) and harvesting time (6 am, 12 noon, and 6 pm).

Harvest and post-harvest processing

Plants were harvested at the flowering stage. The harvesting method involved cutting aerial parts above the crown at the designated times. Harvested material was immediately placed in specialized paper bags and transferred to a drying room, where it was dried in shade without exposure to direct sunlight to prevent photodegradation.

Essential oil extraction and yield determination

Dried aerial parts were ground and separated into small particle-sized plant segments. Essential oils were extracted from 50 g of powder per replicate using a Clevenger-type apparatus with hydrodistillation for 4 h. The essential oil yield was calculated and reported as a critical parameter for assessing the agricultural and industrial viability of the studied ecotypes under different harvest times. The extracted oils were collected, dehydrated over anhydrous sodium sulfate, and stored at 4 °C in amber vials until analysis. Each extraction was performed in triplicate.

Gas chromatography-mass spectrometry (GC-MS) analysis

Chemical profiling was conducted using an Agilent 7890B GC system coupled with an Agilent 5977 Series MSD detector, equipped with an HP-5MS capillary column (30 m × 0.25 mm i.d., 0.25 μm film thickness). Helium was the carrier gas at a flow rate of 1.0 mL min⁻¹. The injection volume was 1 µL with a split ratio of 1:50. The applied temperature program is detailed in Table 1. Component identification involved a dual approach: mass spectra and GC-MS chromatograms (Supplementary Fig. 1) were matched against Wiley 7 and NIST 05 libraries, and retention indices (RI) were calculated using a homologous series of n-alkanes (C₈-C₂₀). Final identification was confirmed by comparing experimental RI values with reference data from the NIST Chemistry WebBook and published literature.

Table 1.

Temperature programming conditions for the GC-MS analysis of essential oil compounds in studied Anethum graveolens L. ecotypes harvested at different times.

Initial		50 °C	Hold Time/1 min
Ramp 1	0	100 °C	8 °C/min
Ramp 2	2 min	110 °C	2 °C/min
Ramp 3	0	185 °C	5 °C/min
Ramp 4	5 min	280 °C	30 °C/min
Ramp 5	3 min	290 °C	20 °C/min

Statistical analysis

Factor analysis

To identify the key volatile compounds responsible for variability among ecotypes and harvesting times, factor analysis was performed on the standardized essential oil data derived from GC-MS. It was based on the correlation matrix, allowing all variables to contribute equally regardless of their original scale. Factors with eigenvalues greater than 1.0 were used based on Kaiser’s criterion. To enhance interpretability and minimize the overlap among correlated variables, factor analysis was subsequently conducted using the varimax rotation method. This orthogonal rotation technique redistributes the variance to achieve a simpler structure, emphasizing variables that strongly load on a single factor. The rotated factor loadings were used to identify clusters of compounds representing distinct chemical pathways or biosynthetic relationships within the essential oil composition. The varimax approach allowed for a more biologically meaningful interpretation of the data, distinguishing between groups of volatiles associated with specific harvesting times or ecotypic origins. Factors were interpreted based on the magnitude and direction (positive or negative) of compound loadings.

Treatment-by-trait biplot analysis

Following factor analysis, a treatment-by-trait biplot (TT biplot) was constructed to visualize and interpret the relationships between the experimental treatments (combinations of ecotypes × harvesting times) and the measured traits. Thus, the treatments were considered as entries, and the essential oil traits as testers. The principal component axes of the biplot were derived from singular value decomposition of the centered and standardized data matrix. The first two principal components, which explained the majority of the total variance, were used to generate a two-dimensional biplot using the GGEbiplot software¹⁵. The fitted model is computed by the following formula¹⁶:

In the above model, Y_ij denotes the mean value of treatment i for trait j, while represents the overall mean of the trait. The term SD_j corresponds to the standard deviation of trait j across all treatments. For the n-th principal component, Φ_n indicates its eigenvalue, and Ψ_in and Ω_jn represent the scores of treatments i and trait j on that component, respectively. R_ij accounts for the residual error in the model. To achieve symmetrical scaling between ecotypes and traits, eigenvalues are adjusted through vector absorption, enabling a standardized and balanced representation of both entities. These symmetric scales form the basis for constructing TT interaction biplots, where each treatment and trait is depicted by a distinct point or symbol. The resulting biplot provides a visual framework to interpret the relationships and interactions between treatments and traits, allowing simultaneous evaluation of treatment performance across multiple traits and highlighting patterns of association among traits themselves.

Results

Prior to sowing seeds, the experimental field was plowed and a soil analysis was conducted. The soil at the site was loam-textured with a pH of 6.85, an electrical conductivity of 2.38 dS/m, and contained 4.8 mg/kg iron, comprising 43.2% sand, 43.1% silt, and 13.7% clay. The geometric mean values of essential oil components from plants grown in soil with these characteristics are presented in Table 2. The suitability of the dataset for factor analysis was first assessed, revealing that the Kaiser-Meyer-Olkin (KMO) index exceeded the recommended threshold of 0.51, with a value of 0.64. Also, the sphericity test of Bartlett was meaningful, confirming the presence of sufficient interrelationships among the traits and further supporting the appropriateness of the dataset for factor analysis. The first factor explained 46.6% of the variation and was primarily characterized by high positive loadings of β-thujene (BT), α-terpinene (AT), trans-pinane (PI), germacrene D (GD), myristicin (MY), and high negative loadings of α-pinene (AP). This factor represented a contrast between thujene–terpinene-trans-pinane-germacrene-myristicin-enriched chemotypes as well as pinene-dominated chemotypes (Table 3), and it could be named as terpene and aromatic compounds factor. Biochemically, this factor reflected differential flux between GPP-derived monoterpenes (pinene branch) and combined p-menthane, sesquiterpene (FPP), and shikimate-derived pathways, suggesting genotypic variation in terpene synthase activity and precursor allocation (Fig. 1).

Table 2.

Geometric means of the main components (%) of dill (Anethum graveolens L.) essential oil in three Iranian ecotypes under three harvest times.

Code	Compound name	Retention index			Ardabil			Bushehr			Kerman
		RI cal	RI ref	Reference	6 am	12 noon	6 pm	6 am	12 noon	6 pm	6 am	12 noon	6 pm
AP	α-Pinene	938	942	24	2.68	1.77	3.42	2.93	2.78	2.8	2.83	3.34	2.8
CA	Camphene	950	951	25	0.26	0.06	0.07	0.07	0.06	0.06	0.06	0.08	0.06
BT	β-Thujene	962	960	26	-	10.75	-	0.22	-	0.23	0.21	0.29	0.23
BP	β-Pinene	979	980	27	0.78	0.27	0.34	0.18	-	0.17	-	0.95	0.17
AT	α-Terpinene	1017	1015	27	0.14	0.21	0.12	0.12	0.13	0.12	0.14	0.15	0.12
TE	Terpinolene	1083	1089	27	0.67	0.3	0.14	-	0.23	0.24	0.27	-	0.17
PI	trans-Pinane	1089			-	1.48	0.58	0.26	0.16	0.22	0.11	0.2	0.17
EX	3,9-Epoxy-p-menth-1-ene *	1182	1178	28	2.18	2.95	4.95	7.47	9.1	10.12	8.86	11.11	11.55
CV	Carvacrol	1299	1304	27	0.42	0.24	0.25	0.24	0.23	0.31	0.32	0.3	0.3
BC	β-Cubebene	1390	1383	25	4.3	5	4.42	-	1.03	1.28	-	0.73	1.21
GD	Germacrene D	1484	1476	29	1.98	6.2	0.17	1.19	-	-	0.8	0.74	-
MY	Myristicin	1519	1519	28	0.07	0.64	0.02	-	-	0.12	-	0.19	0.16
AI	Apiol	1676	1681	28	2.31	3.47	4.29	3.35	3.79	6.15	4.8	6.67	6.92
EOY	Essential oil yield (% (w/v))	-	-	-	0.04	0.17	0.05	0.46	0.41	0.26	0.64	0.81	0.18

RI _cal: Calculated retention index; RI _ref: Reference retention index; * Another name of 3,9-Epoxy-p-menth-1-ene is 3,6-Dimethyl-2,3,3a,4,5,7a-hexahydrobenzofuran as deposited in NIST Chemistry WebBook.

Table 3.

Scores of factor analysis were rotated by the Varimax method in this study.

Code	Compound name	F1	F2	F3	F4
BT	β-Thujene	0.96	0.20	-0.16	-0.06
CA	Camphene	-0.11	0.43	0.81	0.34
AP	α-Pinene	-0.83	-0.19	-0.22	0.44
BP	β-Pinene	0.07	-0.01	0.35	0.92
AT	α-Terpinene	0.98	0.06	0.05	0.09
TE	Terpinolene	0.14	0.50	0.80	-0.12
PI	trans-Pinane	0.83	0.34	-0.44	0.02
EX	3,9-Epoxy-p-menth-1-ene	-0.38	-0.83	-0.26	-0.10
CV	Carvacrol	-0.18	0.02	0.90	0.26
BC	β-Cubebene	0.42	0.81	0.06	0.32
GD	Germacrene D	0.93	0.29	0.09	0.02
MY	Myristicin	0.95	0.06	-0.12	0.15
AR	Apiol	-0.16	-0.71	-0.29	0.22
EOY	Essential oil yield	-0.02	-0.86	-0.06	0.07
Eigenvalue		6.52	3.74	1.56	1.03
Proportion		46.6	26.7	11.2	7.4
Cumulative		46.6	73.3	84.5	91.8

Fig. 1.

Schematic representation of the plastidial 2-C-methyl-d-erythritol 4-phosphate (MEP) and cytosolic mevalonate (MVA) biosynthetic pathways for the characterized Anethum graveolens L. essential oil compounds (Ag: Anethum graveolens; DXS: 1-Deoxy-d-xylulose-5-phosphate synthase; DXR: 1-deoxy-D-xylulose-5-phosphate reductoisomerase; MCT: 2 C-methyl-d-erythritol 4-phosphate cytidylyltransferase (or 4-diphosphocytidyl-2-methyl-D-erythritol synthase (CMS)); CMK: 4-(cytidine 5′-diphospho)-2-C-methyl-D-erythritol kinase; MDS: 2 C-methyl-d-erythritol 2, 4-cyclodiphosphate synthase; HDS: (E)-4-hydroxy-3-methylbut-2-enyldiphosphate synthase; HDR = IDS: 1-hydroxy-2-methyl-2-(E)-butenyl 4- diphosphate reductase; IDI: Isopentyl diphosphate isomerase; GDS: Geranyl diphosphate synthase (geranyl pyrophosphate synthase (GPS=GPPS)); MTPS: Monoterpene synthase; DTPS: Diterpene synthase; AACT: acetyl-CoA c-acetyltransferase 2; HMGS: Hydroxymethylglutaryl-CoA synthase; HMGR: 3-hydroxy-3-methylglutaryl coenzyme A reductase; MK: Mevalonate kinase (or MVK); PMK: 5- phopspho-mevalonate kinase; MDD (or PMD): Mevalonate-5-diphosphate decarboxylase (Mevalonate-5- phosphate decarboxylase (MDC) or Diphosphomevalonate decarboxylase); FPPS: Farnesyl pyrophosphate synthase; STPS: Sesquiterpene synthase; and TTPS: Triterpene synthase).

The second factor explained 26.7% of the variation (Table 3), and contrasted β-cubebene (high positive loading) with 3,9-epoxy-p-menth-1-ene, apiol, and essential oil yield (high negative loadings), representing a mix of aromatic compounds like camphene and β-cubebene, but also includes essential oil yield. The negative loading of apiol and essential oil yield suggested a relationship with compounds that could be less dominant in terms of yield but still play an important role. This factor could be named as aromatic and yield-related compounds. The second factor contrasted β-cubebene (high positive loading) with 3,9-epoxy-p-menth-1-ene and apiol (high negative loadings), representing a sesquiterpene vs. monoterpene/phenylpropanoid trade-off axis.

The third factor exhibited high positive loadings for terpinolene and carvacrol as well as negative loadings for α-pinene, which accounts for 11.2% of the observed variation (Table 3). This factor is strongly associated with monoterpenes and aromatic compounds like α-pinene and carvacrol. Finally, the third factor was characterized by high positive loadings for camphene, terpinolene, and carvacrol, indicating a shift toward oxidative monoterpene metabolism and enhanced bioactivity.

The fourth factor was defined by high positive loading of β-pinene, which accounts for 7.4% of observed variation (Table 3). This factor is associated with β-pinene, a compound known for its presence in the essential oil of various plants, including dill. It could be labeled as β-pinene factor. High scores of the fourth factor identified pinene-rich chemotypes desirable for essential oil production, emphasizing freshness and resinous notes, as well as potential antioxidant and antimicrobial activity.

The first two principal components of the TT biplot model accounted for 74% of the total variation (Fig. 2), the first component explained 47% and the second component 27% of the total variance. This magnitude of explained variability indicates remarkable treatment-by-trait interactions, reflecting both crossover and non-crossover types of interactions that altered treatment rankings across traits. In this current research, the TT biplot approach was used to visualize the relationships between treatments and traits. As shown in Fig. 2, the biplot effectively distinguished performance patterns of traits, whereas carvacrol, camphene, β-pinene, and terpinolene, were grouped around treatment A-M (Ardabil + 6 am sampling). In contrast, β-thujene, α-terpinene, trans-pinane, β-cubebene, germacrene D, myristicin, were associated with treatment A-N (Ardabil + 12 noon sampling). Treatment K-N (Kerman + 12 noon sampling) exhibited superiority for 3,9-epoxy-p-menth-1-ene, apiol, α-pinene, and essential oil yield, while treatment K-E (Kerman + 6 pm sampling) was not the best for any of the recorded characteristics. Ecotype Ardabil favored the accumulation of monoterpene hydrocarbons (BP, CA, and TE) and carvacrol, while 12 noon sampling enhanced oxygenated monoterpenes and phenylpropanoids (BT, AT, PI, GD, and MY). In contrast, the Kerman ecotype was associated with higher levels of EX, AI, EOY, and α-pinene, depending on the sampling time.

Fig. 2.

Exploring the best treatments for essential oil components of dill (Abbreviations of traits are: β-Thujene (BT), Camphene (CA), α-Pinene (AP), β-Pinene (BP), α-Terpinene (AT), Terpinolene (TE), trans-Pinane (PI), 3,9-Epoxy-p-menth-1-ene (EX), Carvacrol (CV), β-Cubebene (BC), Germacrene D (GD), Myristicin (MY), Apiol (AI), and essential oil yield (EOY). Abbreviations treatments are: A-M (Ardabil + 6 am sampling), A-N (Ardabil + 12 noon sampling), A-E (Ardabil + 6 pm sampling), B-M (Bushehr + 6 am sampling), B-N (Bushehr + 12 noon sampling), B-E (Bushehr + 6 pm sampling), K-M (Kerman + 6 am sampling), K-N (Kerman + 12 noon sampling), and K-E (Kerman + 6 pm sampling)).

The biplot model described a satisfactory magnitude of variability among traits, so trait associations were determined via the angle cos of vectors (cos 0° = +1 as positive correlation, cos 90° = 0 as no correlation, cos 180° = -1 as negative correlation), while longer vectors denoted greater variability (Fig. 3). Most notable relationships included: strong positive correlations among 3,9-epoxy-p-menth-1-ene (EX), essential oil yield (EOY), and apiol (AI); among PI, MY, AT, BT, and ED; and among carvacrol (CV), β-pinene (BP), and camphene (CA).

Fig. 3.

Interrelationships among the components of dill essential oil (Abbreviations of traits are: β-Thujene (BT), Camphene (CA), α-Pinene (AP), β-Pinene (BP), α-Terpinene (AT), Terpinolene (TE), trans-Pinane (PI), 3,9-Epoxy-p-menth-1-ene (EX), Carvacrol (CV), β-Cubebene (BC), Germacrene D (GD), Myristicin (MY), Apiol (AI), and essential oil yield (EOY). Abbreviations treatments are: A-M (Ardabil + 6 am sampling), A-N (Ardabil + 12 noon sampling), A-E (Ardabil + 6 pm sampling), B-M (Bushehr + 6 am sampling), B-N (Bushehr + 12 noon sampling), B-E (Bushehr + 6 pm sampling), K-M (Kerman + 6 am sampling), K-N (Kerman + 12 noon sampling), and K-E (Kerman + 6 pm sampling)).

Clear negative relationships were evident for EX, EOY, and AI, with TE, as well as for α-pinene (AP), with PI, MY, AT, BT, and ED (Fig. 3). The observed negative correlations emphasize the presence of two major biosynthetic directions within current dill ecotypes: a monoterpene hydrocarbon-rich chemotype (α-pinene, β-thujene, terpinolene, and camphene), Near-zero correlations occurred between α-pinene (AP) with CV, CA, and BP, and among TE, with PI, MY, AT, BT, and ED. Most of the above-mentioned graphical interrelationships are well supported by the numerical correlations (Table 4), with minimal inconsistencies, and some discrepancies are expected, because the TT biplot model was able to describe only 74% of the variation, leaving 26% of the variability unexplained by the first two principal components.

Table 4.

Simple coefficients of correlations for essential oil components of dill (Anethum graveolens L.) in this study.

BT	CA	AP	BP	AT	TE	PI	EX	CV	BC	GD	MY	AI	EOY
	-0.17	-0.83	-0.05	0.94	0.12	0.93	-0.48	-0.32	0.54	0.94	0.94	-0.25	-0.21
CA		-0.03	0.61	-0.02	0.78	-0.31	-0.57	0.81	0.42	0.12	-0.14	-0.51	-0.37
AP			0.26	-0.76	-0.45	-0.63	0.44	0.06	-0.38	-0.82	-0.74	0.37	0.30
BP				0.17	0.15	-0.09	-0.21	0.50	0.30	0.13	0.15	0.02	0.09
AT					0.20	0.81	-0.47	-0.14	0.49	0.94	0.90	-0.26	0.01
TE						-0.06	-0.61	0.71	0.54	0.30	0.09	-0.52	-0.55
PI							-0.49	-0.50	0.61	0.83	0.86	-0.22	-0.29
EX								-0.18	-0.83	-0.66	-0.32	0.85	0.60
CV									0.12	-0.10	-0.20	-0.05	-0.15
BC										0.59	0.52	-0.47	-0.76
GD											0.86	-0.46	-0.21
MY												0.00	-0.18
AI													0.38

Significant coefficients at 7 degrees of freedom are 0.67 and 0.80 at the 0.05 and 0.01 probabilities, respectively. BT = β-Thujene, CA= Camphene, AP = α-Pinene, BP = β-Pinene, AT = α-Terpinene, TE= Terpinolene, PI= trans-Pinane, EX = 3,9-Epoxy-p-menth-1-ene, CV= Carvacrol, BC = β-Cubebene, GD= Germacrene D, MY= Myristicin, AI= Apiol, and essential oil yield = EOY.

The concept of an ideal treatment was applied to evaluate proximity of each treatment to the optimum performance point (Fig. 4). Treatments closest to the ideal were considered superior, whereas those farther away are less desirable, so treatment A-M (Ardabil + 6 am sampling), following to A-N (Ardabil + 12 noon sampling), were nearest to the ideal position, while most treatments were farthest and thus least favorable. Trait discriminative ability was assessed via the magnitude of standard deviation, with larger values indicating stronger differentiation among ecotypes (Fig. 5). Traits positioned nearer the ideal point exhibit greater discriminative power, so β-cubebene (BC), following TE, GD, and AT were the most discriminative traits. Except for EX, AP, EOY and AI, all traits showed above-average discriminative potential. The typical representativeness of each trait is measured by the angle between its vector and the average axis, so it showed that BC, and TE had small angles and thus high typical potential, whereas PI, MY, and BT displayed larger angles and lower representativeness. Examining of 3,9-epoxy-p-menth-1-ene (EX), one of the most important components of dill essential oil, is presented in Fig. 6. The horizontal axis represents EX magnitude, and the arrow denotes its direction, so all treatments (ecotype + harvesting timing), except A-M (Ardabil + 6 am sampling), and A-N (Ardabil + 12 noon sampling), exhibited the highest EX amounts. All treatments except A-M and A-N combined high, with low variability, making them ideal candidates.

Fig. 4.

Determining the position of treatments regarding the position of ideal treatment in the components of essential oil in dill. Abbreviations treatments are: A-M (Ardabil + 6 am sampling), A-N (Ardabil + 12 noon sampling), A-E (Ardabil + 6 pm sampling), B-M (Bushehr + 6 am sampling), B-N (Bushehr + 12 noon sampling), B-E (Bushehr + 6 pm sampling), K-M (Kerman + 6 am sampling), K-N (Kerman + 12 noon sampling), and K-E (Kerman + 6 pm sampling).

Fig. 5.

Determining the position of essential oil components of dill regarding the position of THE ideal trait (Abbreviations of traits are: β-Thujene (BT), Camphene (CA), α-Pinene (AP), β-Pinene (BP), α-Terpinene (AT), Terpinolene (TE), trans-Pinane (PI), 3,9-Epoxy-p-menth-1-ene (EX), Carvacrol (CV), β-Cubebene (BC), Germacrene D (GD), Myristicin (MY), Apiol (AI), and essential oil yield (EOY).

Fig. 6.

Examination of 3,9-epoxy-p-menth-1-ene (EX), according to the applied treatments (Abbreviations treatments are: A-M (Ardabil + 6 am sampling), A-N (Ardabil + 12 noon sampling), A-E (Ardabil + 6 pm sampling), B-M (Bushehr + 6 am sampling), B-N (Bushehr + 12 noon sampling), B-E (Bushehr + 6 pm sampling), K-M (Kerman + 6 am sampling), K-N (Kerman + 12 noon sampling), and K-E (Kerman + 6 pm sampling).

Discussion

This study confirms the hypothesis that diurnal harvest times (6 am, 12 noon, and 6pm) significantly influence essential oil yield and composition in native A. graveolens ecotypes, with specific timing optimizing bioactive compound concentrations. Factor analysis and TT biplot modeling revealed distinct chemotypes structured along key biochemical pathways, differentiating monoterpene hydrocarbon-rich profiles from phenylpropanoid–oxygenated monoterpene profiles. This metabolic partitioning reflects underlying flux differences between the MEP-derived monoterpenes and the shikimate/FPP-derived sesquiterpenes and phenylpropanoids, a pattern documented in other aromatic herbs^17,18.

The primary finding is that harvesting at midday maximizes oxygenated monoterpenes and phenylpropanoids, whereas morning or evening harvesting favors monoterpene hydrocarbons such as α- and β-pinene. This divergence is attributable to lower thermal stress and reduced oxidative degradation during cooler morning and evening periods, which preserve heat-sensitive monoterpene hydrocarbons. In contrast, midday light intensity drives the biosynthesis of oxygenated monoterpenes via light-dependent terpene synthase activity, but also promotes oxidative stress that can degrade volatile hydrocarbons. The coordinated clustering of oxygenated monoterpenes and phenylpropanoids suggests shared enzymatic or transcriptional controls, as reported in other Apiaceae species¹⁹. Negative correlations between monoterpene hydrocarbons and oxygenated or phenylpropanoid compounds indicate biosynthetic trade-offs due to precursor competition and differential gene expression^13,20. Ecotype differences further modulate these patterns. Kerman ecotype, adapted to higher temperatures and aridity, accumulated higher levels of specific sesquiterpenes and phenylpropanoids, consistent with the ecological role of these compounds in stress mitigation¹⁹. Ardabil ecotypes, from cooler temperate conditions, showed enrichment in monoterpene hydrocarbons and carvacrol. This indicates that genotypic predisposition interacts with microenvironmental factors to shape chemotypic identity²¹. The observed diurnal rhythms in essential oil accumulation align with demonstrated seasonal and diurnal fluctuations in camphene and carvacrol, firmly supporting the present findings related to oxygenated monoterpene enrichment during midday harvests²². The functional potential of A. graveolens essential oils extends beyond aroma and flavor. Oxygenated monoterpenes and phenylpropanoids, including carvacrol, apiol, and myristicin, confer antioxidant, antimicrobial, and anti-inflammatory properties, enhancing the value of these oils in food preservation, pharmaceuticals, and cosmetics^6,23. Monoterpene hydrocarbons such as α- and β-pinene provide resinous, fresh aromas desirable in perfumery and culinary applications. Therefore, strategic harvesting based on both ecotype and time of day can optimize essential oil quality and bioactivity, providing a practical framework for industrial and agronomic applications. The identification of metabolically independent groups allows targeted improvement of specific compounds without inadvertently altering hydrocarbon components, enabling the development of cultivars with multi-functional essential oil capability.

Conclusion

Ardabil ecotypes favored monoterpene hydrocarbon and carvacrol enrichment, particularly during morning (6 am) and midday harvests (12 noon), while Kerman ecotypes accumulated higher levels of sesquiterpenes and phenylpropanoids depending on sampling time. For maximum accumulation of oxygenated monoterpenes and phenylpropanoids (e.g., carvacrol, myristicin, and apiol), which contribute to bioactivity and aroma complexity, harvest aerial parts during midday. Positive and negative correlations among compounds revealed coordinated biosynthetic modules and metabolic trade-offs, providing mechanistic insights into chemotype differentiation. For essential oils rich in monoterpene hydrocarbons (e.g., α- and β-pinene, camphene, and terpinolene), which provide fresh, resinous aromas, harvest in the morning or evening to minimize oxidative degradation.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 ^{(612.5KB, docx)}

Author contributions

S.D.D. : Investigation, Methodology; K.F.K. : Investigation, Writing – original draft; N.S. : Writing – original draft, Data curation, Formal analysis, Software, Supervision, Validation; M.M. : Writing – review & editing, Formal analysis, Funding acquisition, Methodology, Investigation, Software, Supervision, Validation.

Funding

This research received no external funding.

Data availability

The data that support the findings of this study are available from the corresponding authors, Dr. Naser Sabaghnia and Dr. Mehdi Mohebodini, upon reasonable request. Inquiries should be directed to [sabaghnia@maragheh.ac.ir] or [mohebodini@uma.ac.ir].

Declarations

Competing interests

The authors declare no competing interests.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The authors alone are responsible for the content of the paper.

Ethics declaration

Ethics committee review and/or approval was not required for this study as it does not fall under the category of research conducted.

Permission statement for plant material collection

All plant material was collected by the co-ordination of Department of Horticultural Science of University of Mohaghegh Ardabili with Iranian Ministry of Agriculture, Iranian Natural Resources Organization, Iranian Department of Environment, and Iranian Research Institute of Forests and Rangelands.