The Effect of Irrigation on the Vineyard Canopy and Individual Leaf Morphology Evaluated with Proximal Sensing, Colorimetry, and Traditional Morphometry

Abstract

The high number of grapevine (Vitis vinifera L.) cultivars grown world-wide are described and identified according to detailed morphological and morphometric descriptor lists. The grapevine leaf is of utmost importance in characterization, despite its traits being very sensitive to environmental factors. In this study, the effect of irrigation/drought stress on the individual leaf morphology and morphometry of the ‘Hárslevelű’ grapevine (Vitis vinifera L.) cultivar was examined. To verify the effect of the applied irrigation methods (drip and subsoil irrigation) on the plant’s water status, water potential measurements were carried out during the 2022 season. The effect of the applied treatments on the vegetative growth was evaluated according to point quadrat and a multichannel LiDAR analysis in order to describe the width of the canopy area, row volume, and area coverage index. The individual leaf morphology was assessed via traditional morphometry and colorimetry. Our results showed that rainfed plants had a significantly lower stem ψ compared to the drip- and subsoil-irrigated plants at all examined dates. The point quadrat results indicate that the leaf layer number was significantly (p < 0.05) influenced by the position, while the treatment showed no effect on the leaf layer number. The leaf colorimetry showed a difference among the samples, as significant alterations were found in 28 out of the 32 examined color properties. Within the traditional morphometric analysis, 54 traits were evaluated, and 14 of the traits were significantly altered due to the different water management systems.

1. Introduction

Grapevine (Vitis vinifera L.) is one of the most widespread horticultural crops. It is grown on more than 7.3 million hectares under a wide range of growing and environmental conditions. In the last decades, the changing climate has caused a significant alteration in yield and quality. All of that is mainly triggered by the uneven rainfall that usually causes drought stress in the vineyards, making irrigation systems widely established. The impact of irrigation on vineyards is multifaceted, encompassing both physiological responses within the grapevine as well as broader environmental repercussions. At the physiological level, water availability directly influences vine growth, canopy development, and grape maturation [1]. Water stress, whether excessive or deficient, can alter the balance of vegetative and reproductive growth, affect the nutrient uptake, and modulate the synthesis and accumulation of secondary metabolites in grape berries, including sugars, acids, phenolics, and aroma compounds [2]. Consequently, irrigation regimes have the potential to shape not only grape yield but the sensory attributes and commercial value of the produced wines as well. Irrigation can improve the stability of the yield year to year, as well as its quality [3], increase berry anthocyanin content [4] and size [5], and optimize the water balance in the grapes [6]. Several studies have examined the effects of irrigation on the yield and quality parameters, while less information is available about the detailed canopy architecture and individual leaf morphology.

Grapevine cultivars and genotypes are described by morphological and morphometric traits to ensure the true-to-typeness of the wine products, which is a key element of the appellation. The International Organisation of Wine and Vine (OIV) provides more than 150 descriptors for the [7] characterization of the grapevine genotypes. Among these, large numbers of traits are linked to the leaf petiole and lamina, including size, shape, lobature, coloration, and hair types. Grapevine leaf morphometry is usually referred to as ampelometry. The study includes the description and evaluation of linear and angular traits defined by Ravaz [8] and later implemented by Galet [9] and Chitwood et al. [10]. See Bodor-Pesti et al. [11] and citations therein. Among the leaf morphological characteristics, lamina coloration has multiple significances. The color could be typical to the cultivar [7], which provides the possibility for identification [12], while the coloration and the calculated vegetation indices (v.i.) give valuable information about the nutrient and water status [13].

Although the ampelometric and colorimetric traits are typical to the genotypes, former studies have shown that the leaf has a morphological/morphometric plasticity induced by several factors. For example, the bud load/pruning method (length of the pruning elements) has a significant effect on the size of the venation and the individual leaf area [14]. The result is then in line with Intrieri et al. [15], who found that the leaf sizes of conventional and minimally pruned plants differ significantly. Later, leaf samples of the ‘Furmint’ grapevine cultivar obtained from rows with different orientations and elevations above sea level were compared by Bodor et al. [16]. Their findings showed that both factors alter linear and angular leaf traits. Recent studies investigated the effect of the irrigation/drought stress on leaf morphology and found that the individual leaf area was modified by the water status [17,18].

In line with the morphometric traits, colorimetric properties also vary within the same genotype and are influenced by environmental factors, such as chlorosis [19,20] or water stress [21]. Recently, leaf morpho-colorimetric traits of ‘Sauvignon blanc’, ‘Chardonnay’, ‘Carmenère’, ‘Merlot’, and ‘Pinot noir’ samples were compared, and differences caused by different water management techniques were found [22].

Although the above-mentioned studies detail the main effects of irrigation/drought stress on morphological and colorimetric traits, they do not provide deep enough insights into the ampelometric features and color properties, such as color indices. Therefore, the aim of our study was (i) to set up an experiment with different water management methods, namely rainfed, drip-irrigated, and subsoil-irrigated parcels, (ii) verify the effect of water management on the canopy architecture with manual monitoring and proximal sensing, and (iii) investigate morphometric and colorimetric traits of the grapevine cultivar ‘Hárslevelű’ to identify those descriptors that are influenced by the treatments. The used descriptors in the treatments must therefore be taken into consideration during true-to-typeness characterization.

2. Materials and Methods

2.1. Experimental Vineyard and Irrigation

The experiment was carried out on the ‘Hárslevelű’ grapevine (Vitis vinifera L.) cultivar at the Mikóczy Family Estate (Tata, Hungary). Cordon-trained vines were pruned to 5 spurs, leaving 2 to 3 buds on each spur. The spacing was 3.0 m within rows and 0.9 m within the vines. The length of the rows was 270 m, with 300 plants in each row. In this experiment, three treatments were compared: (i) rainfed (no irrigation was applied), (ii) subsoil irrigation (plastic tubes were placed about 40–60 cm below the soil surface to deliver water directly to the roots of the plants), and (iii) drip irrigation (drip tubes were attached to the support system of the plantation above the surface, from which the water was delivered directly to the plants). Each treatment was established in 3 rows, and the middle row was investigated in this study. During the vegetation period (from March to September 2022) altogether, 102 mm of water divided 18 times was supplied to the irrigated rows, according to the meteorological data provided by the meteorological station in the vineyard. The precipitation was 233.8 mm in total (between 1 January 2022 and 30 September 2022), the total heat sum from April 2022 to September 2022 was 3036.43 °C (with 2945 °C active heat sum), the leaf wetness hour (time) was 2 h on average, and the relative humidity was 57% on average.

2.2. Water Potential Measurement

To verify the effect of the irrigation, the stem water potential (stem ψ) was measured during August and September (5, 19 August, 2, 21 September, and 5 October 2022) to verify the differences in the plant water status. For each measurement, 10 leaves were randomly chosen along the investigated rows and enclosed in plastic bags lined with aluminum foil for 45 min according to Williams and Araujo [23]. Water potential was determined at midday between 11:00 and 12:00 a.m. each day, with a pump-up pressure chamber (PMS Instrument Company, 1725 Geary Street SE, Albany, OR 97322 USA).

2.3. Canopy Structure Analysis

Point Quadrat Evaluation

The evaluation of the canopy structure was carried out in August 2022 via the point quadrat method according to Smart et al. [24]. Compared to the original methodology, 3 heights of the canopy were evaluated: (i) the bunch zone (120 cm), (ii) the middle third of the canopy (9th to 12th nodes of the shoots—150 cm), and (iii) the top of the canopy (170 cm). Ten insertions were applied in 5 replications along the investigated rows at 3 heights. Data were recorded at each insertion point: leaves (L), bunches (B), and gaps (G). Based on the primary data, the gap percentage, the average leaf layer number, and the average bunch number were calculated. Leaf layer maps were created with the Past software Matrix plot (Past version 4.12b) [25].

2.4. LiDAR Record

2.4.1. Hardware

In pursuit of developing a comprehensive digital representation of the vineyard environment, our methodology included the integration of advanced sensor technologies. These included a multichannel LiDAR system, an IMU, a thermal camera, and an Intel RealSense D435 (Intel Corporation, 2200 Mission College Blvd. Santa Clara, CA 95054-1549) depth sensor for capturing color imagery, which was set up on a moving vehicle run between the rows. The 3D LiDAR and IMU were of paramount importance for the synthesis of an accurate digital twin. The IMU served a critical role in mitigating vibrations, operating at a frequency of 100 Hz, while the 3D LiDAR facilitated the acquisition of high-resolution point cloud data at a frequency of 10 Hz. All sensor components were strategically housed within a protective enclosure.

2.4.2. Measurements

We conducted a thorough investigation of each of the three rows in two separate runs, capturing data from both sides of the canopy. This dual-run approach allowed for a comprehensive analysis. Point clouds generated from the recorded data were meticulously matched to enhance accuracy (Figure 1a). Data recording occurred continuously along 100 m of each row, with calculations performed at 70 cm intervals to extract key canopy metrics. Canopy width, canopy area, and row volume were determined at 70 cm intervals. For the calculation of the area coverage index (ACI), the point cloud was delimited to 160 cm height sections (see Figure 1b).

2.4.3. Spatial FieldSLAM

The study centered on processing raw data obtained from a multichannel VLP-16 LiDAR sensor in order to facilitate spatial mapping for a complex wine environment. A pivotal initial step involved the application of Movella Xsens Inertial Motion Unit (IMU) measurements along the roll and pitch axes to align the LiDAR recordings parallel with the ground and effectively filter or mitigate vibration-induced errors. The positioning and mapping of spatial points was based on the FieldSLAM algorithm [26], which accurately adds each iteration of the readings to the whole point cloud.

2.5. Individual Leaf Morphology

2.5.1. Ampelometry

For the leaf morphometric evaluation, 15 to 20 leaf samples were collected from the middle third of the shoots from both sides (NE and SW) of the canopy, between berry set and veraison, all in accordance with the International Organisation of Vine and Wine [7] guidelines. Samples were stored in plastic bags in a cooling box until digitalization. Each leaf lamina (without the petiole) was scanned with a Canon Pixma Mg3650S (30-2, Shimomaruko 3-chome, Ohta-ku, Tokyo 146-8501, Japan) at a 300 dpi resolution. The single leaf area and 53 morphometric traits (Table 1) were evaluated in accordance with the procruste coordinates of 31 biometric landmarks on the lamina (Figure 2a) via the GRApevine LEaf Digitalization (GRA.LE.D.) [27] software (version 2.04). Linear and angular morphometric traits were included according to the OIV (2009) [7] Descriptor List from OIV601 to OIV618 (Table 1). The mean values (n = 15 per treatment) of each investigated trait were converted into the OIV categories according to the Descriptor List [7], with a slight modification. Bilateral linear and angular OIV traits were investigated on both sides of the leaf lamina and indicated as R (right) and L (left). Traits labeled with “D” indicate the difference measured in “mm” between the landmarks.

2.5.2. Leaf Colorimetry

Samples obtained for the ampelometric evaluation were also used for the leaf colorimetric characterization. The color evaluation was carried out with ImageJ [28] according to 5 regions with 200,000 pixels on each leaf sample (Figure 2b). The average RGB values were recorded from each region, and then data were transformed to CIE-L*a*b*, from which hue (hue = 180 + arctan(b*/a*) and chroma (c = ($\sqrt{a^{*2}+b^{*2}}$) values were calculated. Chroma, as defined by CIE (International Commission on Illumination), is “the colorfulness of an area judged as a proportion of the brightness of a similarly illuminated area that appears white or highly transmitting”, and hue is defined as the “attribute of a visual perception according to which an area appears to be similar to one of the colors: red, yellow, green, and blue, or to a combination of adjacent pairs of these colors considered in a closed ring” [29].

These RGB values and further indices that showed a correlation to water stress in previous studies were included (Table 2) according to Sang et al. [13] and Sánchez-Sastre et al. [30].

2.6. Statistical Analysis

The mean, standard deviation, and coefficient of variability were calculated for summary statistics. Data were subjected to an ANOVA and a Tukey’s post hoc test. Visual Studio Code was used for the statistical analyses. The programming language was Python.

3. Results

3.1. Water Potential

The stem water potential (stem ψ) was evaluated five times during the ripening period from the 5th of August to 5th of October 2022 to verify the differences in the plant water status of the investigated rows (Figure 3). The statistical analysis showed that rainfed plants had a significantly lower stem ψ compared to the drip- and subsoil-irrigated plants across all dates, except for two, where the rainfed plants differed significantly only from subsoil-irrigated (2 September 2022) or from the drip-irrigated plants (21 September 2022). During the evaluation of the water status, the stem ψ of the rainfed plants ranged from −0.92 MPa to −1.43 MPa, and in the subsoil- and drip-irrigated plants, the range was −0.66 MPa to −1.21 MPa and −0.73 MPa to −1.27 MPa, respectively.

3.2. Point Quadrat Measurement

Measurements were taken at three canopy heights: bunch zone, middle third, and top of the canopy. The results showed that the leaf layer number was significantly (p < 0.05) influenced by the positions, but the treatment had no effect if the same heights were compared (Figure 4 and Figure 5). The densest canopy was obtained in the middle of the subsoil-irrigated rows, with four leaves per insertion, while the thinnest canopy occurred at the top of the rainfed row (2.4 leaves per insertion). The vertical trend of the density was different among the treatments. In the case of the rainfed row, no significant differences were observed between the three heights. In contrast, the irrigated row showed a significantly denser canopy at the middle than those in the fruit zone. If the leaf layer number data and position of the insertion are not considered, there are no significant differences between the treatments. The mean fruit number in the bunch zone was 1.38, 1.18, and 1.36 bunches per insertion, and the totals were 69, 59, and 68 bunches per 50 insertions in the rainfed, drip-irrigated, and subsoil-irrigated rows, respectively, without significant differences. The gap number showed low values in all rows. The highest value was recorded in the top position of the canopy in the rainfed row, where the value was 12% (6 gaps out of the 50 insertion).

3.3. LiDAR Monitoring of the Canopy

Irrigation had a significant effect on the mean canopy width (p < 0.05). The lowest value was recorded in the rainfed row, where the mean was 49 cm (min.: 40 cm; max.: 66 cm). The drip-irrigated and subsoil-irrigated rows had 57 cm (min.: 41 cm; max.: 83 cm) and 72 cm (min.: 51 cm; max.: 95 cm) of row width, respectively. The coefficient of variability showed that the rainfed row had the most uniform canopy width, with 8.7% of variability, compared to the drip- and subsoil-irrigated rows (13.4% and 12.6%, respectively).

The canopy area was evaluated according to each of the 0.7 m sections and for the full 100 m of the rows. The results showed that the mean area of the section was higher in the case of the drip-irrigated plants (1.03 m²), which had a significantly higher area than those of the subsoil-irrigated (1.00 m²) and rainfed rows (0.96 m²), which also differed from each other at p < 0.05.

Concerning the full row (100 m) length, the highest value was recorded in the drip-irrigated row, followed by the subsoil-irrigated and rainfed rows, with 148.45 m², 144.96 m², and 138.43 m² respectively. The variability of the canopy area was almost uniform in all cases, as the coefficient of variability ranged from 2.4% to 2.8% for the subsoil irrigation and rainfed rows, respectively.

Row volume was the highest in the subsoil-irrigated row (0.72 m³/0.7 m) and the lowest in the rainfed row (0.48 m³/0.7 m). The rainfed row showed the lowest variability (9.53%), while the highest value was recorded in the drip-irrigated row (14.26%). The area coverage index showed similar values in the three treatments, with 0.92, 0.93, and 0.94 in the subsoil-irrigated, drip-irrigated, and rainfed rows, respectively.

3.4. Individual Leaf Morphology

3.4.1. Ampelometry—OIV Descriptors

In terms of the OIV (2009) [7] Descriptor List, the irrigation had an effect on the leaf morphometric traits (Table 1). The length of the N₁ vein (OIV601) was the largest (5) in the samples obtained from the subsoil-irrigated row, while the rainfed and drip-irrigated samples belonged to the same lower category (3). The length of the N₂ distal or superior lateral vein (OIV602) and the N₃ proximal or inferior lateral vein (OIV603) showed no differences among the samples, while the length of the N₄ vein was the shortest (7) in the rainfed plants. The distance from the petiole sinus to the lower lateral leaf sinus on the right side of the rainfed leaves was the shortest (5), while all other samples and their sides belonged to the seventh category. The length of the vein in N₅ showed no difference, as all samples belonged to the first category. The length of the serration on the tip N₂ vein was the longest (5) on the right side of the leaves obtained from the rainfed plants, while all other samples and their sides belonged to the third category. OIV613 (width of tooth N₂) and OIV604 (length of tooth N₄) showed no difference among the samples. In contrast, the width of the tooth at the tip of N₄ (OIV615) showed variability influenced by the irrigation. The angular traits (OIV607: angle between the tip of N₁ and the tangent between petiole point and the first branch of the N₂; OIV608: angle between N₂ and N₃ measured on the tangents formed before these veins first branch; OIV609: angle between the tooth tip of N₃ and the tangent between the first branch of N₃ and the tooth tip of N₄) showed no variability. In OIV610 (the angle between the tooth tip of N₃ and the tangent between the petiole point and the tooth tip of N₅), however, this was not uniform among the samples obtained from the different rows.

3.4.2. Ampelometry—Traditional Morphometry

Within the traditional morphometric analysis, 54 traits were evaluated (Table 1). The results indicated that 14 out of the 54 traits were significantly altered due to the different water management systems. The individual leaf area was the smallest in the rainfed plants (180 cm²) and the highest in the subsoil-irrigated row (208.78 cm²). The differences were significant between the subsoil-irrigated and the control samples. We obtained the same results in OIV 601. In OIV 603R, the values were the highest in the subsoil-irrigated row (81.55 mm). The control row showed significant differences compared to the other two treatments. The OIV 604L values were the highest in the subsoil-irrigated row (54.97 mm) and the lowest in the control row (49.92 mm). We also experienced significant differences between the subsoil-irrigated row and the control row. In OIV 605L, only the subsoil-irrigated row showed significant differences between the control rows. The OIV 606R values were the highest in the subsoil-irrigated row (74.13 mm) and the lowest in the control row (67.40 mm). The control row showed significant differences compared to the two other treatments. In the case of D 01-02, the drip-irrigated row showed the highest results (13.76 mm), and the control row showed the lowest results (11.06 mm). The control row thus showed significant differences compared to the two other rows. In D 01-03, the subsoil-irrigated row showed the highest values (26.76 mm), and the control row showed the lowest values (23.83 mm). The subsoil-irrigated row showed significant differences compared to the other two treatments. In D 01-04, only the subsoil-irrigated row showed significant differences compared to the drip-irrigated row. Values of D 01-29 were the highest in the case of the subsoil-irrigated row (66.83 mm), and the lowest were in the control row (60.66 mm). We found significant differences between the subsoil-irrigated row and the control row. In D 01-11, only the drip-irrigated row showed significant differences compared to the control row. The D 25-15 values were the highest in the subsoil-irrigated row (166.58 mm), and the lowest were in the control row (154.73 mm). The control row showed significant differences compared to the other two treatments. In D 23-17, the values were the highest in the subsoil-irrigated row (147.71 mm) and the lowest in the control row (135.70 mm). Only the subsoil-irrigated row showed significant differences compared to the control row. In D17-20, the control row showed significant differences compared to the other two treatments.

Table 1. Statistics summary of the linear and angular ampelometric traits, with the corresponding OIV categories.

Treatment		Rainfed			Subsoil Irrigation			Drip Irrigation
Trait *		mean	st. dev.	OIV	mean	st. dev.	OIV	mean	st. dev.	OIV
Surface, cm2		180.00	21.90 a		208.78	30.31 b		197.95	36.42 ab
OIV 601	Length of vein N1	119.78	15.97	3	127.10	13.10	5	118.86	17.10	3
OIV 602R	Length of vein N2	99.44	9.43 a	5	105.48	8.88 b	5	102.21	10.56 ab	5
OIV 602L		101.45	10.27	5	107.86	11.19	5	102.42	11.96	5
OIV 603R	Length of vein N3	74.91	5.10 a	5	81.55	6.28 b	5	80.63	8.05 b	5
OIV 603L		75.91	6.18	5	81.26	7.38	5	80.64	7.62	5
OIV 604R	Length of vein N4	49.45	3.83	7	53.92	5.91	9	52.82	7.44	9
OIV 604L		49.92	5.31 a	7	54.97	6.50 b	9	52.79	6.04 ab	9
OIV 605R	Length of petiole sinus to upper lateral leaf sinus	84.27	9.59	7	87.49	8.84	7	88.01	12.15	7
OIV 605L		84.82	11.17 a	7	91.16	7.38 b	7	88.58	11.31 ab	7
OIV 606R	Length of petiole sinus to lower lateral leaf sinus	67.40	4.70 a	5	74.13	5.26 b	7	72.65	7.01 b	7
OIV 606L		69.26	5.70	7	73.86	6.84	7	73.21	6.72	7
OIV 611R	Length of vein N5	17.96	4.48	1	19.12	7.04	1	19.46	4.59	1
OIV 611L		18.44	5.05	1	19.98	5.60	1	18.27	4.40	1
OIV 612R	Length of tooth N2	16.89	32.65	5	8.19	2.10	3	7.21	1.70	3
OIV 612L		8.40	2.70	3	8.97	5.78	3	7.64	2.05	3
OIV 613R	Width of tooth N2	15.30	2.17	5	15.68	3.50	5	14.16	2.35	5
OIV 613L		15.11	3.63	5	16.88	9.55	5	14.98	3.11	5
OIV 614R	Mature leaf: length of tooth N4	7.40	2.13	1	6.87	1.33	1	7.19	1.36	1
OIV 614L		6.78	1.47	1	7.70	1.79	1	7.12	1.49	1
OIV 615R	Width of tooth N4	12.76	3.75	5	12.21	1.78	5	12.57	2.05	5
OIV 615L		11.65	2.29	3	12.87	2.58	5	12.61	2.30	5
OIV 617R	Length between the tooth tip of N2 and the tooth tip of the first secondary vein of N2	46.41	9.97	5	47.05	6.44	5	45.46	7.35	5
OIV 617L		45.84	8.84	5	49.58	9.45	5	44.71	8.34	3
OIV 618R	Distance between the tooth tips of N4	86.87	13.73		98.88	11.76		93.43	12.46
OIV 618L	Distance between the tooth tips of N5	21.68	8.18		25.16	11.03		23.29	11.36
OIV 607R	Angle between the tip of N1 and the tangent between petiole point and the first branch of the N2	57.81	8.31	7	57.41	6.31	7	60.37	7.28	7
OIV 607L		57.30	6.06	7	56.61	6.94	7	59.48	5.38	7
OIV 608R	Angle between N2 and N3 measured on the tangents formed before these veins first branch	60.17	5.57	7	61.16	8.85	7	58.15	5.97	7
OIV 608L		59.81	6.89	7	57.70	5.76	7	57.25	6.70	7
OIV 609R	Angle between the tooth tip of N3 and the tangent between the first branch of N3 and the tooth tip of N4	51.39	4.56	5	50.10	5.52	5	50.73	4.03	5
OIV 609L		50.52	7.51	5	49.88	6.76	5	50.23	5.45	5
OIV 610R	Angle between the tooth tip of N3 and the tangent between the petiole point and the tooth tip of N5	73.67	9.05	9	72.02	9.87	9	68.56	8.30	7
OIV 610L		71.30	9.33	9	69.69	9.76	7	70.26	8.28	9
OIV 618aR	Angle between the tooth tip of N5 and the tangent between the petiole point and the tooth tip of N5	93.07	15.48		96.41	12.22		92.59	14.25
OIV 618aL	Angle between the tooth tip of N4 and the tangent between the petiole point and the tooth tip of N4	34.24	14.46		37.38	15.58		36.27	17.09
D 01-02	Distance between the points 01 and 02	11.06	2.49 a		13.63	3.29b		13.76	2.90 b
D 01-03	Distance between the points 01 and 03	23.83	3.35 a		26.76	4.59b		24.34	2.30 b
D 01-04	Distance between the points 01 and 04	33.74	7.40 ab		37.86	8.16b		31.78	6.86 a
D 01-05	Distance between the points 01 and 05	25.46	3.98		27.01	3.84		26.44	4.29
D 01-06	Distance between the points 01 and 06	11.37	3.18		12.64	3.66		13.12	2.62
D 01-31	Distance between the points 01 and 31	34.17	5.66		36.93	7.52		34.06	6.97
D 01-09	Distance between the points 01 and 09	32.17	6.42		36.08	7.80		35.71	6.32
D 01-29	Distance between the points 01 and 29	60.66	6.38 a		66.83	7.40 b		65.11	5.24 ab
D 01-11	Distance between the points 01 and 11	59.80	4.36 a		66.81	6.64 ab		65.76	7.38 b
D 27-13	Distance between the points 27 and 13	149.91	10.13		161.96	11.72		160.49	13.94
D 25-15	Distance between the points 25 and 15	154.73	12.24 a		166.58	15.52 b		164.46	15.12 b
D 23-17	Distance between the points 23 and 17	135.70	11.87 a		147.71	16.30 b		146.91	16.50 ab
D 11-13	Distance between the points 11 and 13	51.71	5.51		54.78	7.60		53.95	7.86
D 13-17	Distance between the points 13 and 17	74.40	11.45		76.66	8.41		74.85	11.47
D 17-20	Distance between the points 17 and 20	83.56	10.05 a		91.52	11.47b		89.83	12.88 b
D 20-23	Distance between the points 20 and 23	81.12	9.93		88.24	12.42		87.33	13.62
D 23-27	Distance between the points 23 and 27	74.61	11.16		78.32	11.85		74.40	12.42
D 27-29	Distance between the points 27 and 29	51.85	8.44		54.29	8.09		53.76	9.42

* Bilateral traits were evaluated on both sides of the leaf lamina, indicated with R (right) and L (left). Different letters mean significant differences between the treatments at p < 0.05.

Table 1. Statistics summary of the linear and angular ampelometric traits, with the corresponding OIV categories.

Treatment		Rainfed			Subsoil Irrigation			Drip Irrigation
Trait *		mean	st. dev.	OIV	mean	st. dev.	OIV	mean	st. dev.	OIV
Surface, cm2		180.00	21.90 a		208.78	30.31 b		197.95	36.42 ab
OIV 601	Length of vein N1	119.78	15.97	3	127.10	13.10	5	118.86	17.10	3
OIV 602R	Length of vein N2	99.44	9.43 a	5	105.48	8.88 b	5	102.21	10.56 ab	5
OIV 602L		101.45	10.27	5	107.86	11.19	5	102.42	11.96	5
OIV 603R	Length of vein N3	74.91	5.10 a	5	81.55	6.28 b	5	80.63	8.05 b	5
OIV 603L		75.91	6.18	5	81.26	7.38	5	80.64	7.62	5
OIV 604R	Length of vein N4	49.45	3.83	7	53.92	5.91	9	52.82	7.44	9
OIV 604L		49.92	5.31 a	7	54.97	6.50 b	9	52.79	6.04 ab	9
OIV 605R	Length of petiole sinus to upper lateral leaf sinus	84.27	9.59	7	87.49	8.84	7	88.01	12.15	7
OIV 605L		84.82	11.17 a	7	91.16	7.38 b	7	88.58	11.31 ab	7
OIV 606R	Length of petiole sinus to lower lateral leaf sinus	67.40	4.70 a	5	74.13	5.26 b	7	72.65	7.01 b	7
OIV 606L		69.26	5.70	7	73.86	6.84	7	73.21	6.72	7
OIV 611R	Length of vein N5	17.96	4.48	1	19.12	7.04	1	19.46	4.59	1
OIV 611L		18.44	5.05	1	19.98	5.60	1	18.27	4.40	1
OIV 612R	Length of tooth N2	16.89	32.65	5	8.19	2.10	3	7.21	1.70	3
OIV 612L		8.40	2.70	3	8.97	5.78	3	7.64	2.05	3
OIV 613R	Width of tooth N2	15.30	2.17	5	15.68	3.50	5	14.16	2.35	5
OIV 613L		15.11	3.63	5	16.88	9.55	5	14.98	3.11	5
OIV 614R	Mature leaf: length of tooth N4	7.40	2.13	1	6.87	1.33	1	7.19	1.36	1
OIV 614L		6.78	1.47	1	7.70	1.79	1	7.12	1.49	1
OIV 615R	Width of tooth N4	12.76	3.75	5	12.21	1.78	5	12.57	2.05	5
OIV 615L		11.65	2.29	3	12.87	2.58	5	12.61	2.30	5
OIV 617R	Length between the tooth tip of N2 and the tooth tip of the first secondary vein of N2	46.41	9.97	5	47.05	6.44	5	45.46	7.35	5
OIV 617L		45.84	8.84	5	49.58	9.45	5	44.71	8.34	3
OIV 618R	Distance between the tooth tips of N4	86.87	13.73		98.88	11.76		93.43	12.46
OIV 618L	Distance between the tooth tips of N5	21.68	8.18		25.16	11.03		23.29	11.36
OIV 607R	Angle between the tip of N1 and the tangent between petiole point and the first branch of the N2	57.81	8.31	7	57.41	6.31	7	60.37	7.28	7
OIV 607L		57.30	6.06	7	56.61	6.94	7	59.48	5.38	7
OIV 608R	Angle between N2 and N3 measured on the tangents formed before these veins first branch	60.17	5.57	7	61.16	8.85	7	58.15	5.97	7
OIV 608L		59.81	6.89	7	57.70	5.76	7	57.25	6.70	7
OIV 609R	Angle between the tooth tip of N3 and the tangent between the first branch of N3 and the tooth tip of N4	51.39	4.56	5	50.10	5.52	5	50.73	4.03	5
OIV 609L		50.52	7.51	5	49.88	6.76	5	50.23	5.45	5
OIV 610R	Angle between the tooth tip of N3 and the tangent between the petiole point and the tooth tip of N5	73.67	9.05	9	72.02	9.87	9	68.56	8.30	7
OIV 610L		71.30	9.33	9	69.69	9.76	7	70.26	8.28	9
OIV 618aR	Angle between the tooth tip of N5 and the tangent between the petiole point and the tooth tip of N5	93.07	15.48		96.41	12.22		92.59	14.25
OIV 618aL	Angle between the tooth tip of N4 and the tangent between the petiole point and the tooth tip of N4	34.24	14.46		37.38	15.58		36.27	17.09
D 01-02	Distance between the points 01 and 02	11.06	2.49 a		13.63	3.29b		13.76	2.90 b
D 01-03	Distance between the points 01 and 03	23.83	3.35 a		26.76	4.59b		24.34	2.30 b
D 01-04	Distance between the points 01 and 04	33.74	7.40 ab		37.86	8.16b		31.78	6.86 a
D 01-05	Distance between the points 01 and 05	25.46	3.98		27.01	3.84		26.44	4.29
D 01-06	Distance between the points 01 and 06	11.37	3.18		12.64	3.66		13.12	2.62
D 01-31	Distance between the points 01 and 31	34.17	5.66		36.93	7.52		34.06	6.97
D 01-09	Distance between the points 01 and 09	32.17	6.42		36.08	7.80		35.71	6.32
D 01-29	Distance between the points 01 and 29	60.66	6.38 a		66.83	7.40 b		65.11	5.24 ab
D 01-11	Distance between the points 01 and 11	59.80	4.36 a		66.81	6.64 ab		65.76	7.38 b
D 27-13	Distance between the points 27 and 13	149.91	10.13		161.96	11.72		160.49	13.94
D 25-15	Distance between the points 25 and 15	154.73	12.24 a		166.58	15.52 b		164.46	15.12 b
D 23-17	Distance between the points 23 and 17	135.70	11.87 a		147.71	16.30 b		146.91	16.50 ab
D 11-13	Distance between the points 11 and 13	51.71	5.51		54.78	7.60		53.95	7.86
D 13-17	Distance between the points 13 and 17	74.40	11.45		76.66	8.41		74.85	11.47
D 17-20	Distance between the points 17 and 20	83.56	10.05 a		91.52	11.47b		89.83	12.88 b
D 20-23	Distance between the points 20 and 23	81.12	9.93		88.24	12.42		87.33	13.62
D 23-27	Distance between the points 23 and 27	74.61	11.16		78.32	11.85		74.40	12.42
D 27-29	Distance between the points 27 and 29	51.85	8.44		54.29	8.09		53.76	9.42

* Bilateral traits were evaluated on both sides of the leaf lamina, indicated with R (right) and L (left). Different letters mean significant differences between the treatments at p < 0.05.

3.4.3. Colorimetry

Significant differences were found in 28 out of the 32 color properties. The values of L* were the highest in the drip-irrigated and control row, and they showed a significant difference (p < 0.05) compared to the other samples. a* also showed a significant difference in that drip-irrigated samples had significantly higher values (p < 0.05) when compared to the other samples. For b*, we found a significant difference between all three treatments. In chroma, the control row (13.71) and subsoil-irrigated row (12.88) showed the highest results, and they also showed a significant difference compared to the drip-irrigated row. Hue also showed significant differences between the tree treatments. In the case of R, the rainfed row showed the highest results (61.56) but only a significant difference compared to the subsoil-irrigated row (59.38). For G, we experienced that the drip-irrigated row (72.18) and the control row showed the highest value (72.38) and showed a significant difference compared to the subsoil-irrigated row (70.17). In the case of B, the drip-irrigated row showed the highest value (59.08) and significantly differed when compared to the other treatments. The SAVI values only showed significant differences compared to the drip-irrigated row and the control row. In r, all of the tree treatments showed significant differences. We experienced a significant difference for g in the drip-irrigated row when compared to the other treatments. In b, all of the tree treatments showed significant differences among each other. Looking at the results in the case of R−G, only the drip irrigation row showed significant differences compared to the other two treatments. In R−B, the control row had the highest results (9.87), and the drip-irrigated row had the lowest (2.80). All tree treatments showed significant differences among each other. The G−B values were the highest in the control row (20.69), and all of the tree treatments showed significant differences among each other. In (R−G)/(R+G), only the drip-irrigated row showed significant differences compared to the control row. Looking at the results of (R−B)/(R+B), all tree treatments showed significant differences among each other. In (G−R)/(G+R), only the drip-irrigated row and the control row showed significant differences among each other. In (G−B)/(G+B), the drip-irrigated row significantly differed from the other two treatments. In (R−B)/(R+G+B), all tree treatments showed significant differences among each other. In (G−B)/(R+G+B), only the subsoil-irrigated row showed a significant difference compared to all of the other treatments. Looking at the results of RGRI, only the drip-irrigated row showed a significant difference compared to the control row. In GLI, the drip-irrigated row was significantly different from the two other treatments. In VARI, the drip-irrigated row showed significant differences when compared to the control row. The IPCA values were the highest in the control row (39.58) and significantly differed among each other in all tree treatments. In ExB, the drip-irrigated row had significant differences compared to the two other treatments. In ExG, all tree treatments showed significant differences among each other, and the same was observed for ExGR.

Table 2. Statistical summary of the RGB, L*a*b values, and calculated color indices.

	Subsoil		Drip Irrigation		Rainfed
Index	mean	st. dev.	mean	st. dev.	mean	st. dev.
L*	28.32	2.48 a	29.30	2.16 b	29.24	2.84 b
a*	−8.36	1.76 b	−7.38	1.49 a	−8.61	1.66 b
b*	9.77	2.89 b	5.97	2.25 a	10.62	3.42 c
Chroma	12.88	3.31 b	9.53	2.56 a	13.71	3.66 b
Hue	179.15	0.05 b	179.34	0.09 c	179.12	0.06 a
R	59.38	5.11 a	60.80	4.28 ab	61.56	6.33 b
G	70.17	6.20 a	72.18	5.45 b	72.38	7.00 b
B	51.03	2.82 a	59.08	2.11 b	51.69	3.34 a
SAVI	0.12	0.02 ab	0.13	0.02 b	0.12	0.02 a
r	0.33	0.01 c	0.32	0.00 a	0.33	0.01 b
g	0.39	0.01 b	0.38	0.01 a	0.39	0.01 b
b	0.28	0.01 b	0.31	0.01 a	0.28	0.02 c
R−G	−10.80	2.17 b	−11.39	1.82 a	−10.83	1.92 b
R−B	8.35	3.91 b	1.72	2.80 a	9.87	5.36 c
G−B	19.14	5.41 b	13.10	4.21 a	20.69	6.27 c
(R−G)/(R+G)	−0.08	0.01 ab	−0.09	0.01 a	−0.08	0.01 b
(R−B)/(R+B)	0.07	0.03 b	0.01	0.02 a	0.09	0.04 c
(G−R)/(G+R)	0.08	0.01 ab	0.09	0.01 b	0.08	0.01 a
(G−B)/(G+B)	0.16	0.04 b	0.10	0.03 a	0.17	0.04 b
(R−G)/(R+G+B)	−0.06	0.01	−0.06	0.01	−0.06	0.01
(R−B)/(R+G+B)	0.05	0.02 b	0.01	0.01 a	0.05	0.02 c
(G−B)/(R+G+B)	0.11	0.02 b	0.07	0.02 a	0.11	0.03 b
RGRI	0.85	0.03 ab	0.84	0.02 a	0.85	0.02 b
GLI	0.12	0.02 b	0.09	0.02 a	0.12	0.02 b
VARI	0.14	0.02 b	0.15	0.02 a	0.13	0.02 b
IPCA	36.56	10.44 b	24.70	8.11 a	39.58	12.17 c
ExR	−0.08	0.02	−0.08	0.01	−0.08	0.01
ExB	−0.15	0.03 b	−0.09	0.03 a	−0.15	0.04 b
ExG	−0.69	0.02 b	−0.73	0.02 a	−0.68	0.03 c
ExGR	−0.60	0.02 b	−0.65	0.02 a	−0.59	0.03 c
Gray	0.36	0.01	0.35	0.00	0.36	0.01
CIVE	18.72	0.01	18.73	0.01	18.72	0.01

Different letters mean significant differences between the treatments at p < 0.05.

Table 2. Statistical summary of the RGB, L*a*b values, and calculated color indices.

	Subsoil		Drip Irrigation		Rainfed
Index	mean	st. dev.	mean	st. dev.	mean	st. dev.
L*	28.32	2.48 a	29.30	2.16 b	29.24	2.84 b
a*	−8.36	1.76 b	−7.38	1.49 a	−8.61	1.66 b
b*	9.77	2.89 b	5.97	2.25 a	10.62	3.42 c
Chroma	12.88	3.31 b	9.53	2.56 a	13.71	3.66 b
Hue	179.15	0.05 b	179.34	0.09 c	179.12	0.06 a
R	59.38	5.11 a	60.80	4.28 ab	61.56	6.33 b
G	70.17	6.20 a	72.18	5.45 b	72.38	7.00 b
B	51.03	2.82 a	59.08	2.11 b	51.69	3.34 a
SAVI	0.12	0.02 ab	0.13	0.02 b	0.12	0.02 a
r	0.33	0.01 c	0.32	0.00 a	0.33	0.01 b
g	0.39	0.01 b	0.38	0.01 a	0.39	0.01 b
b	0.28	0.01 b	0.31	0.01 a	0.28	0.02 c
R−G	−10.80	2.17 b	−11.39	1.82 a	−10.83	1.92 b
R−B	8.35	3.91 b	1.72	2.80 a	9.87	5.36 c
G−B	19.14	5.41 b	13.10	4.21 a	20.69	6.27 c
(R−G)/(R+G)	−0.08	0.01 ab	−0.09	0.01 a	−0.08	0.01 b
(R−B)/(R+B)	0.07	0.03 b	0.01	0.02 a	0.09	0.04 c
(G−R)/(G+R)	0.08	0.01 ab	0.09	0.01 b	0.08	0.01 a
(G−B)/(G+B)	0.16	0.04 b	0.10	0.03 a	0.17	0.04 b
(R−G)/(R+G+B)	−0.06	0.01	−0.06	0.01	−0.06	0.01
(R−B)/(R+G+B)	0.05	0.02 b	0.01	0.01 a	0.05	0.02 c
(G−B)/(R+G+B)	0.11	0.02 b	0.07	0.02 a	0.11	0.03 b
RGRI	0.85	0.03 ab	0.84	0.02 a	0.85	0.02 b
GLI	0.12	0.02 b	0.09	0.02 a	0.12	0.02 b
VARI	0.14	0.02 b	0.15	0.02 a	0.13	0.02 b
IPCA	36.56	10.44 b	24.70	8.11 a	39.58	12.17 c
ExR	−0.08	0.02	−0.08	0.01	−0.08	0.01
ExB	−0.15	0.03 b	−0.09	0.03 a	−0.15	0.04 b
ExG	−0.69	0.02 b	−0.73	0.02 a	−0.68	0.03 c
ExGR	−0.60	0.02 b	−0.65	0.02 a	−0.59	0.03 c
Gray	0.36	0.01	0.35	0.00	0.36	0.01
CIVE	18.72	0.01	18.73	0.01	18.72	0.01

Different letters mean significant differences between the treatments at p < 0.05.

4. Discussion

Climate-change-induced precipitation changes show an uneven trend in Hungary, and they cause either drought stress or large fractions of total precipitation in other periods. In 2022, the annual main precipitation was 497 mm, resulting in the 17th driest year since 1901, in contrast with 2023, where the annual precipitation was 767 mm. That makes 2023 the eighth most precipitated year since the beginning of the 20th century [31]. The total yearly precipitation, as well as the monthly sum of rainfall, indicate a large variability between the years, meaning that irrigation in Hungary is more frequently applied not only in table grapes but in winegrape growing too [32]. Irrigation/drought stress has a notable effect on the plant’s physiology and morphological traits, making a regular evaluation including relevant categories necessary. According to the water deficit threshold provided by van Leeuwen et al. [2] and Deloire et al. [33], the water stress in our study was moderate to weak and the water deficit was moderate to severe irrespective of the treatment, even if the stem ψ values significantly differed. According to our morphometric and colorimetric results, we think that even if the plants belong to the same water deficit category, the traits could differ because of the water status.

Former studies have shown that irrigation/water stress has a noticeable influence on the grapevine canopy architecture. For example, de Azevedo et al. [34] and later Flexas et al. [35] demonstrated that water management significantly influenced the canopy size and the LAI (leaf area index). Fasiolo et al. [36] applied 2D LiDAR sensors for detecting the structure and density of the canopy in a row with different water management systems and found significant differences in the projected area. Meanwhile, the canopy volume and surface were not altered. This, in line with the LiDAR observation records, showed a reduction in the canopy width, canopy area, and volume in the rainfed row. Interestingly, the point quadrat evaluation did not verify the canopy reduction, as no significant differences in leaf layer number were observed. The same result was observed by Castellarin et al. [37], who reported a lower leaf layer number caused by the water stress, but only at the end of the vegetation period. The study emphasized that the numbers of the leaves did not change but the size did, and this could cause the size differences at the canopy level. This theory is confirmed by our results, which indicate that compared to the rainfed row, the individual leaf area (LA) increased by 15% and by 9% in the subsoil-irrigated and rainfed rows, respectively. This is in line with Gómez-del-Campo et al. [17] and Sabir [18], who found that irrigation has a significant effect on LA. Later, Briglia et al. [38] demonstrated that early drought stress had no effect on the LA, while 11 days after the drought imposition, the difference was statistically verified, as the stressed plants had smaller leaves. The individual LA leaf elongation (length and width) was investigated by Romero and Martinez-Cutillas [39], who found that neither regulated deficit irrigation (RDI) nor partial root-zone irrigation (PRI) had any effects on the individual LA, but they did find differences in the leaf growth. In this study, leaf petiole length was not investigated, but former investigations have shown that this length could also change due to external factors [14], which could cause a wider canopy.

The larger leaf lamina obtained in our study was possibly caused by the vein length increase, which was caused by water management. In terms of the OIV (2009) [7] Descriptor List, it was found that 9 out of the 31 traits were altered, with 1 category between the samples obtained from the different treatments. Among these, the length of the N₁ and N₄ was altered, resulting in a longer lamina. Parallel with the evaluation of categorical variables, we examined the primary data of the ampelometric investigation and then subjected the data to a variance analysis. In total, 14 out of the 54 traits were found to be significantly altered, including the individual leaf area and some of the linear traits. Identification of the cultivars and especially clones are usually limited by molecular genetic methods, therefore still making morphological identification an important methodology involved in research [40] and in practice to provide a phenotypic background for patents (for details, see Bodor-Pesti et al. [11]). For this reason, it is essential to understand morphometric plasticity.

Grapevine leaf has a bilateral symmetry where the main vein N₁ is considered the axis. In this study, leaf morphology was described on both sides of the lamina, and the sides were also evaluated individually. The results indicate that water management has a significant effect on the leaf traits, and this effect influences the symmetry of the leaf. Some of the bilateral traits (present on both sides of the lamina), for example, OIV603 or OIV604, showed a significant alteration caused by the treatments, but only on one side of the lamina. According to Benitez et al. [41], fluctuating asymmetry (FA) could be caused by environmental stress, meaning that water stress could influence the leaf symmetry. Further research could be conducted to evaluate the effect of water management on the leaf FA.

Leaf coloration is an important trait in grapevine identification (OIV, 2009) [7] and a useful indicator of environmental stress; therefore, simple RGB images and calculated color indices are widely applied in agriculture. In this study, only five regions of the leaf lamina, with 200,000 pixels each, were subjected to the RGB evaluation. This method provides deeper insight into the colorimetric variability compared to those that investigate the lamina individually. We found that except for (R−G)/(R+G+B), ExR, Gray, and CIVE, all color attributes showed a significant difference between the samples obtained from the treatments. Former studies have reported that RGB-based vegetation indices showed a high correlation with chlorophyll concentration [13,30,42]. Moreover, the preliminary results of the later report emphasized that RGB-based indices would be appropriate to monitor environmental stress in a more particular water deficit. As the concentration of the chlorophyll is highly influenced by water deficits [43], discoloration is a possible consequence. Our results showed that chlorophyll concentration in the rainfed row was lower than in both of the irrigated rows, as many of the indices showed a significant alteration. For example, G (green) had a higher value in the rainfed plants, and this color property showed a significant negative correlation with chlorophyll content in our former study [11] obtained using the same grapevine cultivar.

5. Conclusions

In this study, we used proximal sensing methods to describe the effects of water management on the canopy architecture and individual leaf morphometry and colorimetry of the ‘Hárslevelű’ grapevine cultivar. The results showed that canopy size and biomass were significantly altered, which was caused by the irrigation. Leaf layer number was not influenced, but the width of the canopy might have increased due to the individual leaf size increase. Leaf morphometric traits, mainly linear characteristics linked to the lamina size, were also influenced by the treatments. Among with the colorimetric traits, many of those that are linked to the chlorophyll concentration were changed. These results highlight the effects environmental factors have on the morphological and morphometric traits of grapes linked to the canopy and individual leaves.