Experimental units consisted of mesocosms (Ø=50 cm pots filled with almost 49 L of soil). These were sown on May 18^th with Felina 32 hemp (C. sativa L.) genotype aiming for a target plant density of 200 pt m^-2, which equates to approximately 40 plants per mesocosm. An open field experiment was carried out in the facilities of the University of Naples, Dept. of Agricultural Sciences (Portici, Southern Italy; 70 m a.s.l.) using a medium fertile sandy soil (91.0% sand, 4.5% silt, and 4.5% clay, pH of 6.6, organic matter 2.6%, a total N of 1.1 g kg^-1, C:N of 13, 127.2 mg kg^-1 of Olsen P₂O₅, and 471.8 mg kg^-1 of assimilable K₂O) managed to achieve a bulk density of 1.36 Mg m^-3 within each mesocosm. Saline solutions, prepared by adding 15.7, 35.2, and 54.8 mmol L^-1 of NaCl to tap water and labeled as S1, S2, and S3, respectively, were used for crop irrigation. These solutions corresponded to electrical conductivities (EC) of 2.0, 4.0, and 6.0 dS m^-1 and were compared with a tap water control (S0) having an EC of 0.4 dS m^-1. All irrigation treatments were applied in combination with two levels of biostimulant application: untreated – NoB; and treated – B with Trainer^®, a legume protein hydrolysate (Hello Nature, Italy) mainly composed of amino acids and soluble peptides previously characterized by Rouphael et al. (2018). A total of eight treatments (4 levels of water salinity-S x 2 levels of biostimulant application-B) were arranged in triplicate over a randomized complete block design consisting of 24 experimental units (4S x 2B x 3 replicates).

Due to the good soil P and K availability, only N fertilization was performed, adding ammonium nitrate (26% N), according to an ordinary rate of 80 kg N ha^-1. The foliar application of the legume-derived PH biostimulant was carried out four times, approximately every 12 days starting from June 10^th, at a dose of 3 mL per liter, based on the manufacturer’s recommendations. Simultaneously, the untreated plants were sprayed with tap water.

The amounts of NaCl required for each saline solution were calculated according to the following Equation 1:

where NaCl concentration represents the quantity of NaCl to be dissolved in 1 L of irrigation water; EC represents the electrical conductivity of the irrigation water, and the coefficient 0.64 is an empirical value derived by past experiments (El-Nakhel et al., 2022). The electrical conductivity (EC) of the irrigation solutions was calibrated using a conductimeter before each application.

For all treatments, the first three waterings were made with tap water to favor good germination and the emergence of seedlings. Then, from June 17^th to July 16^th, mesocosms were irrigated 10 times with saline water according to the experimental design supplying 35.8, 71.7, and 107.5 g of NaCl per mesocosm, for S1, S2, and S3, respectively, with a cumulative water input of 28 liters. Reference ET was estimated by the Hargreaves formula (Hargreaves and Samani, 1985), while crop ET was calculated using the crop coefficients (Kc) proposed by Cosentino et al. (2013) (0.4 from plant emergence to 6^th fully expanded leaf, between 0.4 and 1.1 from 6^th fully expanded leaf to complete plot covering). Rainfall events were subtracted in the daily calculations. Soil moisture at field capacity (FC) and at wilting point (WP) was 19.1% v/v and 9.1% v/v, respectively, which was estimated using the pedotransfer function by Saxton and Rawls (2006).

The water was manually distributed, and the amount of each irrigation volume was aimed at restoring the readily available water fraction within each mesocosm according to Doorenbos and Pruitt (1979). The refill point for unrestricted growth was fixed according to Di Bari et al. (2004) and Cosentino et al. (2013), who fixed readily available water for fiber hemp at 66% of plant available water.

Bulk soil was carefully mixed and then sampled for chemical and physical characterization before the preparation of the experimental units.

The following physicochemical properties were determined on the soil fine fraction (< 2 mm):

soil texture analysis was done using the Boycous hydrometric method; soil pH and EC were measured on 1:2.5 and 1:5 soil (g) water (mL) suspensions, respectively; organic carbon was determined according to the Walkley & Black method (Walkley and Black, 1934); soil organic matter (SOM) was calculated as 1.726 x OC; total nitrogen was determined according to the Kjeldahl method (Bremner, 1965); carbonate was determined according to the Dietrich–Fruhling calcimeter method (Loeppert and Suarez, 1996); and finally, soil P₂O₄ content was measured according to the Olsen method, while the ammonium acetate method was adopted for assimilable K₂O.

The harvest was made on July 19^th at full flowering in order to simulate the double use-destination of hemp: inflorescence for medical and nutraceutical purposes and biomass production. For each experimental unit, plants were cut at soil level and weighed in order to obtain the total fresh weight. Total biomass was split into inflorescences, leaves, and stems and weighed; then, for each treatment and replicate, a sub-sample of each part of the plant was oven-dried until constant weight, both for dry matter determination and the analyses reported in the following paragraphs. Plant height and stem diameter were measured over 10 individuals for each experimental unit.

Dried plant tissues were ground separately using a Wiley Mill to achieve a size finer than 20 µm. The Kjeldahl method (Bremner, 1965) was utilized to evaluate total nitrogen in soil and dried plant tissues, which involves mineralization with 96% sulfuric acid in conjunction with potassium sulfate and a trace amount of copper. Separation and quantification of the macro and micronutrients in the dried plant tissues were conducted using ion chromatography (ICS-3000, Dionex, Sunnyvale, CA, USA). This process was paired with a conductivity detector, employing an IonPac CG12 A guard column and an IonPac analytical column for K⁺, Ca²⁺, Mg²⁺, and Na⁺. Moreover, an IonPac AG11-HC guard column and an IonPac AS11-HC analytical column were used for NO₃ ^-, PO₄ ^-, Cl^-, and SO₄ ^2-, as detailed by Rouphael et al. (2017b).

Leaves and flowers of fiber hemp plants were collected and dried under shadow natural conditions at 25°C. The dried material was ground in a rotary hammer mill and homogenized obtaining an average particle size below 4 mm. Once obtained, the material was stored in sellable and hermetic plastic bags that were stored under dry and dark conditions prior to use.

Powdered leaves and flowers (100 mg for each) were extracted, separately, with 10 mL of methanol ultra LC-HRMS grade, placed in an ultrasound bath at 37 kHz and 800 W, for 15 min, and centrifuged for 10 min at 6000 g. After filtration (PTFE 0.22 µm), the pooled filtrates were evaporated to dryness under vacuum with a rotary evaporator to obtain the dried extracts (S0 L: 13.2 mg; F: 20 mg – S1 L: 21.6 mg; F: 17.7 mg – S2 L: 20.4 mg; F: 18.9 mg – S3 L: 22.7 mg; F: 23.0 mg). The extracts dissolved in methanol ultra LC-HRMS grade (concentration 2 mg/mL) were analyzed by LC-HRMS/MS analysis.

All LC-HRMS and LC-HRMS/MS analyses were performed on an LTQ-Orbitrap mass spectrometer equipped with an ESI interface and Excalibur data system (Thermo Fisher Scientific Spa, Rodano, Italy) coupled to a Thermo Ultimate 3000 HPLC system (Agilent Technology, Cernusco sul Naviglio, Italy). The LC-HRMS was carried out on a Kinetex 2.6 µ POLAR C18 100Å (100x3mm) column (Phenomenex, Torrance, CA, USA), using 0.1% v/v of HCOOH in H₂O (solvent A) and CH3CN (solvent B) as a mobile phase. The gradient elution was optimized as follows: 50% B for 3 min, 50% to 95% B for 20 min, hold for 2 min, followed by 5 min of initial conditions. The total run time, including column wash and equilibration was 28 minutes, with a flow rate of 0.5 mL/min and injection volume of 5 µL. The MS and MSn spectra, in positive and negative mode, were recorded in Data Dependent Acquisition mode, inducing fragmentation of the most intense five peaks for each scan. Source conditions: spray voltage at 3.5 kV (positive mode) and 2.9 kV (negative mode); capillary voltage: 25 V; source temperature: 320°C; normalized collision energy: 25. The acquisition range was m/z 150–1500. Although the spectra were recorded in positive and negative mode, only the data obtained in positive mode were taken into account.

Plant dry biomass was calculated by multiplying fresh biomass (leaves, stems, and inflorescences) measured for each experimental unit by dry matter percentage measured oven drying the plant sub-samples. Biomass production was recorded as g D.W. m^-2, taking as reference the mesocosm area, and then converted in Mg D.W. ha^-1. The N uptake of aboveground biomass was calculated by multiplying dry aboveground biomass by the N content of tissues and then reported as kg N ha^-1.

A field control, grown in large experimental plots (1000 m²), served as a benchmark to compare the crop yields from mesocosms with those achievable under standard conditions in the experimental area.

The following statistical analyses were performed using MS Excel 2013 and SPSS 21 (SPSS Inc. Chicago, USA). A two-way analysis of variance (ANOVA) was performed for biometric parameters using a general linear model. Means were separated according to the LSD test (p< 0.05). A one-way ANOVA was performed to compare the effect of saline irrigation and biostimulant application on cannabinoid-relative concentration in flowers and leaves. In this case, each combination of NaCl salinity level and biostimulant application was considered as a single treatment, accounting for a total of eight treatments. Kolmogorov-Smirnov and Levene tests were carried out to attest normality of distribution and homogeneity of variance, respectively. Logarithmic transformation was applied to studied variables, when necessary, to ensure normality of distribution.

Principal Component Analysis (PCA) (Bro and Smilde, 2014) and ANOVA-simultaneous component analysis (ASCA) (Smilde et al., 2005) were employed to explore the data relative to the phytocannabinoids quantification performed by LC-MS. Both analyses were computed using the PLS toolbox version 8.9 (Eigenvector Research, Manson, USA) under MATLAB environment, version R2018b (MathWorks Inc., Massachusetts, USA). Prior to PCA, the data matrix, made of 16 rows (samples) and 13 columns (measured phytocannabinoids), was submitted to some pre-processing steps. A normalization step was carried out in order to minimize non-sample-related variations. In particular, the norm1 approach was used. Briefly, the area of each cannabinoid was divided by the sum of the areas of all the measured phytocannabinoids within a sample. Then, data was mean-centered and scaled to unit variance (autoscaling). The latter step employs the standard deviation as a scaling factor, thus giving all the phytocannabinoids the same chance to affect the model independently by their absolute values.

The mean effect of water salinity and biostimulant application was significant for most of the monitored parameters, while the interaction between these two main factors was not significant. Freshwater (S0) and low NaCl saline (S1) irrigations were not different in terms of total biomass (12.6 Mg DW ha^-1 on average), stems (6.6 Mg DW ha^-1 on average), leaves (3.9 Mg DW ha^-1 on average), and inflorescences (2.4 Mg DW ha^-1 on average) ( Table 1 ). The highest water NaCl salinity level (S3) halved total biomass production, impacting all plant organs with the same magnitude (49% average decrease for stems, leaves, and inflorescences). The S2 irrigation was associated with a clear decrease pattern in plant growth, yet only total and leaf biomass were significantly less than freshwater control. Both plant height and stem diameter were more sensitive to water NaCl salinity compared to other biometric parameters. A pronounced reduction of both parameters occurred at an EC value of 4 dS m^-1 (S2), and plant height further decreased, shifting from S2 to S3.

The total biomass yields for S0 and S1 treatments were in line with the results of other open field experiments carried out in Italy under standard irrigation management. For example, Struik and colleagues (2000) recorded an average biomass yield of 13.7 Mg DW ha^-1 for the Felina hemp variety. However, the same authors observed a higher stem yield, which accounted for 60-68% of the total biomass contrasting with S0 results from our mesocosm experiment (45-50% of the total biomass). Such a difference could be attributed to varying environmental factors and differences in the experimental setups, as confirmed also by comparing yield data from our S0 pots with our open field control (see MM section). It is worth noting that Tang and colleagues (2017) reported an open-field Felina growth consistent with our findings (5.5 Mg ha^-1 stems and 2.6 Mg ha^-1 for inflorescences).

Despite observing significant changes in biomass accumulation across different plant organs, NaCl salinity did not impact the relative allocation of assimilates with stems, leaves, and inflorescences, representing 52%, 31%, and 17% of total biomass (data not shown). Yet, plants from our mesocosm experiment watered with S0 showed a slightly lower relative abundance of stems and inflorescences compared to the field control, likely due to the constraints imposed by the experimental setup.

The dry matter (DM) content of total biomass (average value of 31%) and inflorescences (average value of 29%) remained unaffected by salinity levels. However, the DM contents of stems and leaves were affected differently by increasing water salinity ( Table 2 ). Specifically, saline water significantly led to a reduction in the DM content of stems from 35 to approximately 33%. In contrast, leaves showed the highest DM content with S3, which was comparable to S1, and the lowest with S2.

There is limited knowledge about hemp’s resistance to salinity, though recent studies have begun to illuminate its germination process (Sun et al., 2023) and physiological responses at the seedling stage (Zhang et al., 2023) under saline conditions. Our results do not show any major differences between treatments regarding the plant count per square meter (with an average value of 200 ± 22 pt m^-2). However, as shown in Table 1 , S1, S2, and S3 limited crop growth by 7%, 30%, and 48%, respectively.

In Figure 1 , we plotted the average values of total plant biomass and individual organs (stems, leaves, and inflorescences) against levels of water salinity. The regression models fitted to each variable accounted for more than 95% of the data variability and, in all cases, the correlation coefficients were above the critical values for Pearson’s product-moment correlation with 2 degrees of freedom, with p-values ranging from 0.01 to 0.05. Additionally, the slope coefficients of all tested regression functions were highly significant, with p-values between 0.003 and 0.024. This evidence provides a reference point for estimating hemp response to water salinity using the regression function to make some comparisons with other fiber crops. For instance, FAO Paper 29 (Ayers and Westcott, 1989) indicates that flax biomass drops by 25% and 50% occurs with a water EC value of 2.5 and 3.9 dS m^-1. In contrast, our regression function shows that hemp growth drops by similar percentages at EC values of 3.3 and 6.2 dS m^-1, respectively. These findings offer initial insights into hemp response to water salinity conditions, suggesting that hemp, being a medium-low tolerance crop, can adapt better to heightened saline levels than flax. It must also be noted that, in addition to osmotic stress, our experiment is subjecting hemp plants to NaCl toxicity, which can cause toxicity and nutrient imbalances (see Figure 2 ), compared to other forms of saline stress. This enhances the value of the information we provided, supporting the idea of identifying hemp as a suitable fiber crop candidate for saline-degraded land.

Biostimulant treatment increased all the productive parameters, raising total biomass, stems, leaves, and inflorescences by 40%, 38%, 42%, and 50%, respectively ( Table 1 ). As far as we are aware, the impact of plant-derived protein hydrolysates on the growth dynamics of hemp remains largely uninvestigated. However, these biostimulants proved to be effective in enhancing crop performance also under limiting conditions such as saline and draught stress (Rouphael et al., 2017a), or in low-fertility soils (Di Mola et al., 2020). Our results were coherent with those reported by Di Mola et al. (2021), who showed a 32% boost in total biomass for oilseed hemp (var. Eletta Campana) after the application of protein hydrolysate biostimulants. According to the limited literature on hemp and cannabis cultivation augmented with the addition of plant or animal-derived protein hydrolysates or amino acid mixtures, there are other examples consistent with our research. Wise et al. (2024) demonstrated that fish-derived hydrolysates, either in combination with other compounds or alone, can enhance root growth and nutrient uptake in hemp. Amino acid supplementation under hydroponic conditions (Malík et al., 2022) improved the nitrogen accumulation pattern in the flowers and leaves, also altering the development pattern of cannabis. However, diminished effectiveness of protein hydrolysates on hemp growth was observed under conditions of high metal stress (Ofori-Agyemang et al., 2024).

Crop N nutrition was affected by both water salinity and biostimulant application, though no interaction between these factors was recorded ( Table 3 ). Only leaves displayed a variance in N content, with a noticeable dip in N adsorption with S2 and S3 irrigations, bringing down the N content from 2.5% (average value of S1 and S0 treatments) to 2.3% (as average value of S2 and S3 treatments).

Plant N uptake expressed as kg N ha^-1 mirrors the patterns seen in plant growth. An average total N uptake of 195 kg N ha^-1 was recorded for S0 and S1 irrigations, with a sharp decrease observed at S2 and S3 (-39% on average). A similar trend was shown by leaves, which contributed to nearly half of the total N allocation, whereas inflorescence and stems showed a more gradual reduction.

Nitrogen uptake values aligned with those recorded by Idilko and colleagues (1997), revealing an N uptake ranging from 142 to 256 kg N ha^-1 for the cultivar Kompolti. In comparison with our field control, there were no differences in the N content of plant tissues and total N uptake. However, there were notable discrepancies in the N uptake of the different plant organs due to the different biomass partitioning under field conditions.

Biostimulant application significantly increased plant N uptake regardless of water NaCl salinity levels. This was consistently paired with an increase in biomass ( Table 1 ) and N concentration ( Table 3 ), except for inflorescences. Their N concentration remained consistent between biostimulated plants and the control. These results strongly suggest that the application of protein hydrolysate can enhance nitrogen use efficiency in hemp, even under adverse conditions. The increased assimilatory capacity in biostimulated hemp is likely due to the recognized effect of biostimulants on root development as recently reported by Sifola et al. (2023) in rocket treated with both legume-derived protein hydrolysate and tropical plant extract.

Plant nutrient status was also monitored assessing the mineral profile of hemp leaves. No effect of tested treatments was recorded for several monitored elements, such as for K⁺, Mg⁺, Ca²⁺, and SO₄ ^- that showed average values in leaves of 33, 3.5, 16, and 3.4 g kg^-1 (d.w.), respectively, across the experimental units.

As shown in Figure 2 , Chloride (a) and Sodium (d) were significantly affected by water EC. The first element showed a gradually increasing trend, with S3 displaying the highest Cl^- content [almost 20 g kg^-1 (d.w.)] followed by S1 and S2 (15 g kg^-1 (d.w.) on average). Conversely, sodium exhibited a significant increase only with S3. These results suggest that the reduction in plant growth under NaCl salinity conditions was not solely due to a decrease in water potential but also due to potential toxicity arising from excessive Chloride assimilation. We can also postulate this process as a possible cause of nutrient imbalance for other anions, specifically Nitrate ( Figure 2C ) and Phosphate ( Figure 2B ). In fact, all EC levels resulted in a lower content of these elements in foliar tissues compared to the S0 control. Our findings are consistent with those reported by Hajiboland (2013) and Niu et al. (2018), which describe how an excess of chloride in irrigation water can disrupt the uptake and translocation of nutrients in hemp plants. Chloride accumulation may competitively inhibit the absorption of essential anions such as nitrates and phosphates, thereby altering the nutritional balance within the plant. High levels of Cl^- can also negatively affect plant metabolism by impeding nutrient transporter functions and reducing photosynthetic efficiency.

According to Figure 2 , the application of biostimulants is linked to a significant decrease in Cl^- uptake, which is counterbalanced by an increase in phosphate and nitrate assimilation. This outcome supports the observed trend of enhanced plant growth after biostimulant treatment across various salinity levels and affirms the widely acknowledged effect of biostimulants in ameliorating abiotic stress caused by NaCl salinity. Through mechanisms that involve boosting antioxidant defenses (Calvo et al., 2014), supplying osmoprotective compounds (Sharma et al., 2014), aiding in nutrient uptake, modulating stress hormones, and preserving ion balance, biostimulants enhance plant resilience and counteract nutritional imbalances in saline environments.

This section focuses on the qualitative assessment of phytocannabinoids in hemp samples cultivated under varying salt stress conditions. The aim is to explore the potential of different parts of hemp as sources of bioactive compounds for industrial, nutraceutical, and medical applications. The 16 samples submitted to LC-HRMSMS and chemometric analyses were named according to their origin: a group named with the prefix F, which includes eight samples obtained from inflorescences and group L consisting of leaf samples.

An HPLC-HRMS method was developed to be capable of reliably and efficiently identifying a total of 13 compounds with an accuracy error below 10 ppm. HPLC retention times and MS data of compounds in the methanol extracts from leaves and inflorescences of C. sativa are listed in Supplementary Table S1 . As expected, the compound reporting the most intense chromatographic peak ( Supplementary Figure S1 ) in all the samples analyzed was CBD, m/z 315.2308 [M + H]+, for which the MS/MS fragmentation pattern was in agreement with the literature data (Chianese et al., 2022), including the fragments corresponding to resorcinol core (m/z 181.12 [M + H]⁺), the p-menthane moiety (m/z 135.15 [M + H]⁺), and partial cleavage of the terpene moiety (m/z 259.10, 235.20 and 193.17 [M + H]⁺) ( Supplementary Table S1 ).

The majority of the annotated compounds, identified using standards from prior studies, were identified in the non-polar region, with special emphasis on neutral cannabinoids, including cannabidiol (CBD), cannabidiolic acid (CBDA), delta-9-tetrahydrocannabinol (Δ⁹-THC), cannabinol (CBN), cannabigerol (CBG), cannabielsoin (CBE), cannabielsoic acid (CBEA), cannabichromene (CBC), cannabicyclol (CBL), cannabicitran (CBT), and other phenolic derivatives such as cannflavin A and B and canniprene.

LC-HRMS analysis enabled the identification of 13 phytocannabinoids in the 16 investigated samples. The relative concentrations of the identified cannabinoids in the flower and leaf samples, together with the ANOVA output, are reported in Figure 3 .

A data matrix, consisting of 16 rows (samples) and 13 columns (measured cannabinoids), was analyzed using an explorative multivariate technique called Principal Component Analysis.

The resulting PC1/PC2 biplot ( Figure 4 ) shows a clear separation between the leaves and inflorescence samples along PC1, accounting for 57.04% of the total variance. This indicates that the origin of the samples is by far the main feature influencing their cannabinoid composition. Specifically, the leaf samples (green squares) exhibit higher levels of CBE, CBEA, canniprene, and CBDA (situated in the left part of the plot), while they presented lower levels of CBT, CBG, CBN, Δ⁹-THC, CBC, cannflavin B, and cannflavin A (located in the right part of the plot), compared to the inflorescence samples.

Upon further examination of subsequent PCs and coloring the samples based on salinity levels, a clear trend regarding this feature was detected in the PC2/PC3 biplot and, especially along PC3, representing 10.81% of the total variance of the dataset ( Figure 5 ).

From this plot, it is evident that lower salinity conditions, such as S0 and S1, lead to an increased content of Δ⁹-THC and CBC (located in the upper section of the plot) compared to the group of samples treated with higher water salinity (S2 and S3) which lie, instead, at the bottom of the plot.

Finally, a further inspection of the next PCs facilitated the identification of an interesting trend along PC7 (1.97% of the total variance), as reported in Figure 5 . In particular, it seems that samples that have been treated with the biostimulant (red diamonds at the bottom of the plot) are characterized by higher levels of CBG and CBDA and lower levels of cannflavin B and CBE, compared to the samples that did not receive the biostimulant.

Then, in order to partition the variation of the LC-HRMS phytocannabinoids dataset according to the study’s factors (origin, water salinity, and biostimulant), a multivariate extension of the analysis of variance (ANOVA) called ANOVA-simultaneous component analysis (ASCA) (Smilde et al., 2005) was employed. The first factor (Xorigin) corresponds to the different origin of the samples and contains two levels (inflorescence and leaves); the second one (Xsalinity) corresponds to the irrigation water salinity level and it is characterized by 2 levels (low and high salinity). Even though the original classes were S0, S1, S2, and S3, the previously performed PCA model ( Figure 6 ) clearly showed only two main clusters corresponding, indeed, to low and high salinity groups. The third factor (Xbiostimulant) is related to the presence or absence of the biostimulant and it also is characterized by 2 levels. The ASCA model also includes three interaction factors (Xorigin x salinity; Xorigin x biostimulant; X salinity x biostimulant) in order to evaluate the presence of potential synergic effects among the three investigated factors.

The ASCA results are reported in Table 4 and they show, as expected, that the origin of the samples had the biggest effect and explained 54.42% of the total variance, implying that there is a statistically significant variation in the cannabinoid composition of samples obtained from the leaves compared to those obtained from the inflorescences. Furthermore, the salinity effect turned out to be statistically significant even though it explained only 8.62% of the total variance, confirming the results obtained by the PCA. No statistically significant effect was instead found in relation to the presence of the biostimulant. Moreover, no significant interaction effects were found among the investigated factors, meaning they can safely be interpreted independently of each other.

The comprehensive understanding of the impact of salt-induced stress on C. sativa crops remains incomplete. A recent study investigated salinity sensitivity in an organic potting mix with C. sativa in hydroponic and aquaponic solutions during the flowering stage through the addition of up to 40 mM NaCl to a nutrient solution of 1.8 mS cm^-1. The evaluation focused solely on the Δ⁹-THC/CBD content, revealing a linear decrease in total cannabinoid content with increasing NaCl concentration (Yep et al., 2020). This trend was further validated in 2021 by Anderson et al., who identified optimal fertilizer rates at a low level of 3.5 mM N, observing reduced growth and total cannabinoid concentrations at higher fertilizer concentrations. Their findings highlighted the impact of soil EC and fertilizer rates on CBD concentrations in essential oil hemp, indicating that lower levels of soil EC and fertilizer rates affect CBD more than Δ⁹-THC and CBG (Anderson et al., 2021).

Our study, to the best of our knowledge, presents empirical evidence of how increased salinity levels in C. sativa crops can affect not only Δ⁹-THC and CBD but also the content of other minor phytocannabinoids and polyphenols. Notably, our results indicate that CBD concentration is negatively affected by lower soil EC levels (S0 and S1) compared to Δ⁹-THC, while it increases with higher salinity levels (S2-S3). Moreover, the variation in phytocannabinoid concentrations across salinity treatments sheds light on the content of other minor phytocannabinoids with significant bioactivities. Among them, CBC emerges as one of the most abundant phytocannabinoids in cannabis extract, exhibiting a selective affinity for the CB2 receptor and demonstrating modest antinociceptive and anti-inflammatory effects (Pollastro et al., 2018; Udoh et al., 2019). Therefore, our study suggests that implementing effective management practices, combined with phytocannabinoid profiling and testing, could lead to the development of optimized cultivars for industrial or medical purposes.

The raw data for LC-MS analysis have been deposited in the Harvard Dataverse and can be found at https://doi.org/10.7910/DVN/JAADNB.