*Article* **Influence of Phosphite Supply in the MS Medium on Root Morphological Characteristics, Fresh Biomass and Enzymatic Behavior in Five Genotypes of Potato (***Solanum tuberosum* **L.)**

**Richard Dormatey 1,2,†, Chao Sun 1,2,†, Kazim Ali 1,3, Tianyuan Qin 1,2, Derong Xu 1,2, Zhenzhen Bi 1,2 and Jiangping Bai 1,2,\***


**Citation:** Dormatey, R.; Sun, C.; Ali, K.; Qin, T.; Xu, D.; Bi, Z.; Bai, J. Influence of Phosphite Supply in the MS Medium on Root Morphological Characteristics, Fresh Biomass and Enzymatic Behavior in Five Genotypes of Potato (*Solanum tuberosum* L.). *Horticulturae* **2021**, *7*, 265. https://doi.org/ 10.3390/horticulturae7090265

Academic Editors: Agnieszka Hanaka, Jolanta Jaroszuk-Sciseł and ´ Małgorzata Majewska

Received: 11 July 2021 Accepted: 23 August 2021 Published: 26 August 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Abstract:** Crop production is threatened by low phosphorus (P) availability and weed interference. Obtaining plant genotypes that can utilize Phosphite (Phi) as fertilizer can supplement phosphates (Pi) while providing an environmentally friendly means of weed control. The study was conducted to determine the tolerance and enzymatic behavior of five potato genotypes to PO3. Explants were regenerated in vitro from two nodal cuttings and cultured on Murashige and Skoog (MS) medium under controlled conditions for 30 days. Matured plantlets were subcultured for 20 days in MS medium containing (0.25, 0.5 mM) Phi and Pi and No-P (-Phi + -Pi). The results showed significant genotypic variation in tolerance indices among the five genotypes. Atlantic showed greater tolerance to Phi, with highest total root length (50.84%), root projected area (75.09%), root surface area (68.94%), root volume (33.49%) and number of root forks (75.66%). Phi induced an increasing trend in the levels of hydrogen peroxide in the genotypes with the least effect in Atlantic. The comprehensive evaluation analysis confirmed the tolerance of Atlantic genotype with this ranking; Atlantic, Longshu3, Qingshu9, Longshu6 and Gannong2. Antioxidant enzyme activities and proline content also increased significantly under Phi and No-P treatments. The results suggested that potato genotypes with larger root systems may be more tolerant to Phi than genotypes with smaller root systems.

**Keywords:** phosphite stress; antioxidant enzyme; hydrogen peroxide; root morphology; potato; genotypes

#### **1. Introduction**

Phosphorus (P) is an important macronutrient required by all living organisms and a very important cellular component that plays a crucial role in biological activities [1,2]. Phosphorus is involved in the signaling of target proteins through phosphorylation and dephosphorylation that determine various cellular performances for optimal plant growth [3]. The major P forms include phosphate and phosphite, which are used in agriculture [4]. Phosphate anions (H2PO4 −, HPO4 <sup>2</sup><sup>−</sup> and PO4 <sup>3</sup>−) are certainly the main forms of P used by plants for metabolic processes and development, while phosphite is a reduced form of Pi that can be readily taken up by plants through Pi transporters [3]. More than 90% of the Pi required by plants is supplied by the soil, which provides adequate storage [5]. However, an estimated 80% of Pi fertilizer applied to soil worldwide is lost through immobilization and conversion to inorganic forms that plants cannot utilize directly [6]. Since Pi is highly reactive and rapidly transformed by soil microbes, only 20–30% is effectively

utilized by plants [7]. According to Gianessi [8], in most soils, weeds and crops compete for the available Pi, resulting in Pi deficiency to meet plant growth requirements. As global demand for food increases, the overuse of PO4 fertilizers and herbicides has become inevitable [9]. This can accelerate the depletion of non-renewable phosphorus reserves, and increase production costs and prices of agricultural products; it also has significant environmental impacts, such as runoff into water bodies, leading to algal blooms, eutrophication, etc. [10,11]. Overuse of herbicides in cultivation has led to the emergence of herbicide-resistant superweeds in recent years [12]. Thus, low soil phosphorus availability and herbicide-resistant weeds have been identified as major threats to the long-term sustainability of agriculture [7,12,13], for which an effective long-term solution is urgently needed.

Phosphite anions (H2PO3 − and HPO3 <sup>2</sup>−) have high solubility and low soil reactivity. Although plants and various microorganisms cannot utilize Phi, it can be used as a potential target to enhance germplasm for phosphorus utilization by plants [14,15]. Phosphite has an inhibitory effect on plant growth with similar properties to herbicides [16]. The phosphite salt does not pose any risk to human and animal health and is therefore massively used as an effective fungicide in crop production [2]. Thus, phosphite has a direct effect on phytopathogenic fungi by inhibiting mycelial proliferation and reducing conidiogenesis of *Fusarium* sp. isolated from the rhizosphere of plants [17]. In addition, Phi can act indirectly by stimulating the inherent defense mechanisms of plants to limit pathogen growth [18], and also activate host defense genes that help plants defend against disease [19] and directly suppress the growth of pathogens such as *Phytophthora* [20–22]. According to Mehta et al. [23] and Thao and Yamakawa [24], Phi anions cannot be utilized by plants as a phosphorus nutrient, although Phi is well taken up by plant leaves and roots. The supply of Phi to plants as a sole source of P fertilizer can hinder plant growth and a higher dose can completely destroy plants [25,26]. Moreover, McDonald, Grant and Plaxton [16] claimed that Phi-treated plants accumulate Phi rapidly in their cells. Phosphite is phloem-mobile and accumulates in sink tissues [27]. Since Phi is not metabolized by plants, it remains in tissues for a long time and consequently disrupts the signal transduction chain that allows plants to detect and respond to Pi deficiency at the molecular level, thereby amplifying the negative effects of Phi [23,28].

On the contrary, the stimulatory effect of Phi mediates structural and biochemical changes in potato periderm and rind [29]. Phi application improved fruit set and yield of *Persea americana* (avocado) and also restored optimal growth of Pi deficient *Citrus* species [20]. Again, several reports indicated impressive results of Phi on plant P nutrition, which ultimately increased crop yields [25,30]. When Phi is added to the soil, it comes into contact with microorganisms that help Phi to oxidize to Pi [2]. Thus, following microbial oxidation reactions, Phi may become available to the plant as a P nutrient through this indirect pathway. Interestingly, efforts to generate transgenic plants with microbial genes (*ptxD*) that allow plants to use Phi as a sole P source have opened new possibilities for the use of this P-containing compound for plant nutrition [30]. In contemporary agriculture, Phi is emerging as a unique biostimulant that improves crop productivity and quality, through direct antibiotic effects on microorganisms and inhibition via enhanced plant defense responses. In addition, Phi induces a variety of abiotic stress tolerance mechanisms, including heat tolerance [31,32]. Obtaining a potato genotype that is tolerant to Phi will enable us to conduct more advanced genetic studies to understand gene functions for subsequent molecular work. In this experiment, we investigated the effects of Phi concentrations, MS medium without P nutrient (No-P) and sufficient Pi under in vitro cultures to determine the tolerance of potato genotypes. Nevertheless, there is insufficient information to test the assumption that Phi can stimulate antioxidant enzyme activities and hydrogen peroxide levels. Therefore, the experiment aims to: (1) determine the tolerance of five potato genotypes to Phi stress using tolerance indices and (2) evaluate the responses of antioxidant enzymes in *Solanum tuberosum* L. plantlets grown in different concentrations of Phi.

#### **2. Materials and Methods**

#### *2.1. Place of the Experiment and Materials*

The experiment was conducted in Gansu Providential key laboratory of Aridland Crop Science, College of Agronomy, Gansu Agricultural University Lanzhou, China (36◦03 N; 103◦40 E). Potato genotypes; Qingshu9, Longshu6, Longshu3, Atlantic and Gannong2, were used in this experiment. The Atlantic genotype is reported to be droughtsusceptible [33,34], the Longshu6 genotype is classified as moderately drought-tolerant [35], while Qingshu9 and Longshu3 were designated as drought-tolerant genotypes.

#### *2.2. Source of Genotypes and Preparation of In Vitro Explants*

The five potato genotypes were obtained from the laboratory of Crop Improvement and Germplasm Enhancement, Gansu Agricultural University, Lanzhou. Two years of field trials have yielded a wealth of germplasm traits with a large number of sterile tissue culture seedlings for genetic screening. Uniform explants from two nodal cuttings were cultured on the potato growth medium of Murashige and Skoog [36], which contained sucrose 30 gL−<sup>1</sup> and agar 5 gL−1. The pH was adjusted to 5.8 and the medium was autoclaved at 121 ◦C at 15 1b psi for 25 min. Cultures were maintained in a growth room at 25 ± 1 ◦C, 16 h photoperiod, active photosynthetic radiation of 45 μmol photons m<sup>−</sup>2s−1, and relative humidity of 55–66% for a 30-day growth period. Plantlets were harvested after 30 days and used for subsequent experiments. Uniform cuttings of each with two axillary buds were subcultured in an ethanol-sterilized chamber with laminar air flow and propagated in MS medium supplemented with final concentrations of (0.25, 0.5 mM) Phi and Pi and No-P supply, in sterilized glass vials (120 × 50 mm). Five cuttings of explants were cultured in each vial, which was tightly sealed with the lids and kept at 25 ± 1 ◦C in the growth room.

#### *2.3. Experimental Design and Treatments*

A 5 × 5 factorial trial in a completely randomized design with 3 replicates was conducted in a controlled growth room. Treatments included five potato genotypes, two concentrations (0.25 and 0.5 mM) each of phosphite and phosphate, and a medium without P fertilizer, representing (No-P). We used 0.5 mM Pi as a control. Fifty vials per genotype were cultured, with each vial containing five explants. After 20 days, the plantlets were examined for physio-morphological indices. The rest of the plantlets were plunged into liquid nitrogen and immediately preserved at −80 ◦C for biochemical analysis.

#### *2.4. Measurements of Data*

#### 2.4.1. Physio-Morphological Parameters

Physiological parameters, such as shoot and root length (cm), were determined with a ruler by randomly selecting three plants from each replication and averaging these data. The number of roots and leaves were counted on each selected plant. Fresh stem weight, fresh root weight and total plant weight (g) were also measured using an electronic balance, with subsequent calculation of root to shoot ratio and tolerance biomass index for each genotype.

The roots of the sampled plantlets were carefully detached from the stems, washed in distilled water and scanned using a root scanner (STD) 4800, EPSON, Quebec City, QC, Canada and the root morphological indices such as total root length (TRL), root projected area (RPA), root surface area (RSA), root volume (RV), number of root tips (NRT), number of root forks (NRF) were calculated using root image analysis software Win RHIZO version 5.0 (Regent Instruments, Inc., Quebec City, QC, Canada).

#### 2.4.2. Tolerance Indexes Determination

Based on the formula of Wilkins [37], the Phi tolerance indices (TIs) of the root system, which clearly indicate the tolerance of the root systems to Phi stress, were calculated according to the modifications of Dawuda et al. [38]. Since Phi uptake has direct effects on the different measured root morphological indices and plant tolerance to Phi stress, Phi TI

was determined at the end of the experiment for each root index. Considering the score of TI, we classified the genotype with the largest TI for most of the calculated indices as the most tolerant genotype among the five potato genotypes studied. The formula of TI is as follows: TI = index under Phi stress/index without Phi stress × 100.

#### 2.4.3. Determination of Hydrogen Peroxide and Malonaldehyde Contents in Samples

The content of hydrogen peroxide (H2O2) in the shoot samples was determined as described by Junglee et al. [39], with minor modifications. A 0.1 g fresh shoot sample was crushed using a mortar and pestle in liquid nitrogen. The homogenate was transferred to a 2 mL centrifuge tube and kept in an ice bath. An amount of 1.5 mL of 0.1% Trichloroacetic acid (TCA) was added and the uniform mixture was centrifuged at 12,000× *g* for 15 min at 4 ◦C. The supernatant 0.5 mL was carefully mixed with 0.5 mL Phosphate Buffer Saline (PBS) and 1 mL KI (1M) at 7.0 pH. The mixture was kept at 28 ◦C for one hour. The absorbance was measured using a spectrophotometer (model U-5100, Seya-Namioka, Hitachi Hightechnologies, Minato-ku, Tokyo, Japan) at 390 nm. The contents of H2O2 were determined with reference to the standard curve (0, 1, 2, 3, 4 and 5 mmol−1). Malonaldehyde (MDA) Content Assessment Lipid peroxidation was measured by calculating the amount of MDA emitted using the technique for thiobarbituric acid (TBA) as presented in Hodges et al. [40]. Preserved fresh shoot samples of 0.15 g were crushed using a mortar and pestle and 4.5 mL of 10% TCA was added. Then, the homogenized substance was centrifuged at 500× *g* for 15 min at 4 ◦C. The supernatant was transferred to a centrifuge bottle and the volume (V) was recorded. Two mL of the supernatant was then mixed with 2 mL of 0.6% TBA. The homogenized mixture was warmed in boiling water for 20 min, the reaction stopped in an ice bath, and centrifuged at 5000× *g* for 10 min. The supernatant (2 mL) of V1 was transferred to a cuvette. The absorbance of the supernatant was measured at 450, 532, and 600 nm, respectively. The MDA content was estimated according to the following formula: MDA concentration (μmol/L) = 6.45 × (A532 – A600) − 0.56 × A450. MDA content (μmol/g FW)=C(μmol/L) × V (L) × V1 (mL)/2 mL × M (g FW).

#### 2.4.4. Determination of Antioxidant Enzymes Activities and Proline Contents in Shoots

Enzyme samples were prepared from frozen tissue preserved at −80 ◦C. Each shoot sample (approximately 0.5 g) was crushed in liquid nitrogen using a mortar and pestle and homogenized in 5 mL of 0.1 M phosphate buffer (pH 7.8) containing 0.5 mM ethylenediamine tetraacetic acid (EDTA). Each homogenate was centrifuged at 12,000× *g* for 15 min at 4 ◦C. The supernatant was collected for determination of enzymatic activity. The activities of catalase (CAT), peroxidase (POD) and superoxide dismutase (SOD) in the shoot homogenate were determined using a reagent kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) following the manufacturer's instructions. The principles of these kits are summarized as follows:

Catalase activity was determined by the spectrophotometric ammonium molybdate method, in which ammonium molybdate rapidly stops the H2O2 degradation reaction, catalyzed CAT, as the remaining H2O2 reaction produces a yellow compound that can be examined by absorbance at 405 nm. A catalase unit activity was classified as the amount of enzyme in 1 g of fresh tissue that reduces 1μmol of H2O2 per minute at 37 ◦C.

SOD activity was calculated according to the method of Dhindsa et al. [41]. The method is based on photochemical reduction of SOD-motivated Nitrotetrazolium Blue Chloride (NBT) at 560 nm. A unit of SOD activity was well defined as the amount of enzyme that inhibits 50% of the oxidation.

Peroxidase activity was determined by catalysis of hydrogen peroxide by POD, observing absorbance changes at 420 nm and estimating the activity of POD. One unit of POD activity was defined as the amount of enzyme in 1 g of fresh plant tissue reducing 1 μg of H2O2 at 37 ◦C per min. Proline content was determined according to the method of Bates et al. [42]. Fresh shoot samples weighing 0.15 g were crushed with 4.5 mL of 3% (*w*/*v*) sulfosalicylic acid homogenization and the homogenate was heated in boiling water

for 30 min. This was then filtered through 0.2 μm filter paper. The extract and the volume of extract were designated as Vt. The supernatant was used to determine the amount of proline. The reaction mixture consisted of 2 mL of plant extract and an appropriate amount of ninhydrin glacial acetic acid. The test tubes containing the substance were heated in boiling water for 30 min. The reaction was quenched with the addition of toluene in an ice bath. The substance was shaken vigorously on vortex mixer for 15–30 s and divided into two phases (upper and lower chromophase). The upper chromophase (toluene) was carefully aspirated with a pipette, and absorbance was taken at 520 nm. The amounts of proline were measured from the standard curve and expressed as <sup>μ</sup>g·g<sup>−</sup>1FW. The amount of proline was calculated as: Proline content (μg/gFW) = C × Vt/(V × W).

#### *2.5. Analyses of Data*

All data collected were analyzed using SPSS software 22.0 version (IBM Corp., Chicago, IL, USA). Means of treatments were separated by Duncan's multiple range tests with a probability of 5%. The distribution of means was presented in the figures using standard deviations. All graphs were created using GraphPad Prism version 8.0 (GraphPad Software, Inc., San Diego, CA, USA). Principal component analysis (PCA) was performed using the software PAST-PAlaeontological Statistics, version 1.34. To confirm the tolerance status of potato genotypes to Phi, comprehensive evaluation analysis based on PCA was carried out using R software package Statistics, version 3.5.3, considering the following formula:

$$\mu(\text{Xi}) = (\text{Xi} - \text{Xmin}) / (\text{Xmax} - \text{Xmin}) \,\text{i} = 1, 2, 3 \dots \text{n} \tag{1}$$

In Formula (1), μ(Xi) refers to the membership function value of the i-th comprehensive index, and Xi refers to the i-th. Comprehensive index value, Xmax refers to the maximum value of the i-th, comprehensive index, Xmin refers to the i-th, the minimum value of a comprehensive index. Calculation of the weighting of each comprehensive index:

$$\text{Wi} = \frac{\text{Pi}}{\sum\_{i=1}^{\text{m}} \text{Pi}} \quad \text{i} = 1, \text{ 2, 3} \dots \text{n} \tag{2}$$

In Formula (2), Wi represents the importance of the i-th comprehensive index in all comprehensive indices. In terms of degree and weight, Pi is the contribution rate of the i-th comprehensive index of each genotype. Calculation of comprehensive evaluation value (D).

$$\mathbf{D} = \sum\_{\mathbf{i}=1}^{\mathbf{m}} [\mu(\mathbf{X}\mathbf{i})\mathbf{x}\mathbf{W}\mathbf{i}] \quad \mathbf{i} = 1, \ 2, \ 3 \dots \mathbf{n} \tag{3}$$

In Formula (3), D represents the phosphite tolerance of different potato genotypes. The D value is obtained by calculating the weighted membership function value. The larger the value of D, the more the Phi tolerance.

#### **3. Results**

#### *3.1. Influence of Phosphite and Phosphate on Physiological Parameters of Five Potato Genotypes after 20 Days' Growth Period*

The results showed significant (*p* < 0.01) genotype x phosphorus source and rates interaction effect on all physiological parameters (Figure 1a–d). Phosphite stress significantly reduced the growth of potato genotypes in this study. Reduction in growth was observed in both roots and shoots resulting in reduction in fresh biomass and fresh root to shoot ratio. Mostly all physiological indices across all genotypes were decreased by PO3 (0.25 and 0.5 mM) and No-P, but Atlantic genotype was least affected. At Phi 0.5 mM, the decrease in leaf number was least (40.67%) in Longshu6 and greatest (57.16%) in Qingshu9. The least (61.93%) and greatest (81.73%) decrease in the number of roots occurred in Atlantic and Gannong2, respectively. The least (58.18, 6.22%) and greatest (69.95, 91.33%) decrease in shoot and root length was observed in Atlantic and Gannong2 genotypes, respectively. Moreover, similar results were observed among the five genotypes with

respect to PO3 (0.25 mM), but the decrease in physiological parameters due to Phi stress was more pronounced at 0.5 compared to 0.25 mM. In the treatment without P supply, physiological growth was also decreased in all the five potato genotypes except Atlantic and Longshu3, which recorded increase in root length as compared to their respective control. Compared to the control, the application of PO3 at (0.25 and 0.5 mM) and No-P had a negative effect on the fresh biomass indices of the genotypes (Table 1). Application of PO3 (0.25 mM) to genotype Gannong2 caused greater reductions in FRW (80.95%), FSW (82.48%) and TPW (82.00%). In addition, smaller reductions were observed in FSW (43.24%) and TPW (30.20%), and there was a slight increase in FRW (7.41%) of Atlantic genotype. The application of Phi (0.5 mM) gave similar results with a slight difference in severity compared to the results obtained with 0.25 mM. Thus, application of higher rates of Phi could be lethal to potato plantlets. Maximum decrease in FRW (87.21%), FSW (83.94%) and TPW (80.50%) and minimum decrease in FRW (14.81%), FSW (60.81%) and TPW (48.51%) were observed in potato genotypes Gannong2 and Atlantic, respectively.

**Figure 1.** Effect of Phi, Pi and No-P on (**a**) number of leaves, (**b**) number of roots, (**c**) shoot length (cm), and (**d**) root length (cm) of five potato genotypes grown in modified MS media containing 0.25 and 0.5 mM Phi and Pi and No-P supply for 20 days. Values denote the mean of 3 replicates, ±standard deviation (SD). Means obtained with the same letter in the minuscule do not differ by Duncan Multiple range's test (*p* ≤ 0.05).


**Table 1.** Effect of Phi, Pi and No-P supply on fresh biomass, root/shoot ratio and biomass tolerance index of five potato genotypes studied under in vitro conditions.

Data indicate the mean ± SD of 3 biological replications and were tested for significance using Duncan's multiple range tests. Individual columns marked with different lowercase letters indicate significant differences (*p* < 0.01). \* The BTI was not subjected to an analysis of variance.

> Fresh biomass indices of No-P treated potato genotypes were decreased due to the absence of Pi in the growth media. The effect of Pi deficiency was more pronounced in the shoots than in the roots of the five potato genotypes. The greatest decrease in fresh biomass indices (FRW, FSW and TPW) was recorded in genotype Gannong2, while the least occurred in genotype Longshu3. Moreover, the negative effects of Phi and No-P on growth of all genotypes were measured in biomass tolerance index (BTI). The result showed that Atlantic genotype had the highest (70.39 and 51.49%) BTIs, followed by LS3 (51.79 and 36.92%), while the lowest (34.16 and 15.00%) BTIs were recorded in Qinshu9 and Gannong2 genotypes at Phi (0.25 and 0.5 mM). In No-P treatment, genotype Longshu3 recorded the highest value (62.56%) followed by Atlantic (56.93%) and the lowest value (34.00%) was observed in Gannong2. Among Phi-treated genotypes in terms of root to shoot ratio (RSR), Atlantic recorded the highest value (0.79 and 0.69), and the lowest value (0.26 and 0.28) was recorded in Longshu6 and Qingshu9 at (0.25 and 0.5 mM). In No-P treatment, Longshu3 had the highest value (1.22), while Longshu6 had the lowest (0.40). However, in Pi-sufficient genotypes, the number of leaves, number of roots, shoot length and root length were increased compared to their respective controls. The maximum increase was recorded in the Atlantic genotype. Fresh biomass indices showed similar results in all the five genotypes as compared to the respective controls. The lowest root, shoot, and total plant weights were recorded in Qingshu9 and Longshu3, while maximum increase was recorded in Gannong2 and Longshu6 genotypes, respectively.

#### *3.2. Effects of Phosphite, Phosphate and No-P Supply on Root Morphological Characteristics of Five Potato Genotypes for 20 Days Growth Period*

There were significant (*p* < 0.01) genotype x phosphite interaction effects on total root length (TRL), root projected area (RPA), root surface area (RSA), root volume (RV), number of root tips (NRT), and number of root forks (NRF) (Figure 2). In general, root morphological indices were decreased in all five genotypes, with the Atlantic genotype being the least affected. The decrease in TRL due to Phi effects was least (32.86 and 43.17%) in Atlantic and greatest (80.85 and 80.93%) in Longshu6 and Gannong2 at Phi (0.25 and 0.5 mM), respectively. The least decrease in RPA (15.62 and 24.81%) was recorded in Atlantic while the greatest (68.57 and 75.85%) was observed in Gannong2 at (0.25 and 0.5 mM). Moreover, the least (16.86 and 30.85%) and greatest (85.57 and 84.50%) decrease in RSA was observed in Atlantic, Longshu3 and Longshu6 and Gannong2 genotypes, respectively. The least (53.85 and 66.51%) decrease in RV occurred in Atlantic, while the greatest (82.08 and 84.70%) was observed in Longshu6 and Gannong2. Moreover, the least decrease (29.49 and 24.34%) of NRT occurred in Atlantic and Longshu3 genotypes, while the greatest (85.02 and 89.86%) was observed in Longshu6 and Gannong2, at 0.25 and 0.5 mM, respectively. The least (16.57 and 31.06%) decrease in NRF was observed in Atlantic, while the greatest (83.94 and 80.01%) was measured in Longshu6 and Gannong2, at Phi 0.25 and 0.5 mM, respectively. The decrease in root morphological parameters observed in the five genotypes was due to Phi effects, which caused a decrease in the size of the root and shoot systems of the genotypes. The root morphological indices of the treatment without P supply were also decreased in all five genotypes. Mainly due to the absence of Pi in the No-P treatment, this reduced the root growth of the susceptible genotypes, while the tolerant genotypes developed longer root systems. The least decrease in morphological indices was observed in Longshu3 and Atlantic, while the greatest decrease was observed in Gannong2 and Longshu6 genotypes. However, Pi-sufficient potato genotypes exhibited much higher root morphological indices than genotypes grown under Phi and No-P. Root morphological indices were slightly increased in Pi-sufficient plants of the five potato genotypes. Among the genotypes: Longshu3 and Longshu6 recorded the highest increase, while Atlantic and Qingshu9 had the least increase in all root morphological indices.

**Figure 2.** Effect of PO3, PO4 and No-P on (**a**) Total root length (**b**) Root projected area (**c**) Root surface area (**d**) Root volume (**e**) Number of root tips and (**f**) Number of root forks of five potato genotypes grown for 20 days in medium containing concentrations of 0.25, 0.5 mM PO3 and PO4 and No-P. Values symbolize the mean of 3 replicates ± standard deviation (SD). Means followed by the same lowercase letter do not differ among treatments by Duncan Multiple range's test (*p* ≤ 0.05).

#### *3.3. Tolerance of the Five Potato Genotypes to Phosphite Stress*

There were significant differences (*p* < 0.01) in the tolerance indices (TIs) among the five potato genotypes at 0.25, 0.5 mM and No-P (Figure 3). The result showed that after the plantlets were grown for 20 days in the Phi media, Atlantic had the highest TI for majority of the measured indices such as TRL (56.84%), RPA (75.09%), RSA (68.94%), RV (33.49%) and NRF (75.66%) at Phi (0.5 mM). Considering this result, among the five potato genotypes tested, the Atlantic genotype proved to be the most tolerant, except TI for NRT, which was highest in Longshu3 (69.16%). However, compared to Atlantic, the other genotypes obtained lower TI values and were found to be susceptible to PO3 stress.

**Figure 3.** Root PO3 tolerance indexes (%) of five potato genotypes under 0.5 mM Phi treatment. Qingshu9 = QS9; Longshu6 = LS6; Longshu3 = LS3; Atlantic = ATL and Gannong2 = GN2. Total root length = TRL; Root projected area = RPA; Root surface area = RSA; Root volume = RV; Number of root tips = NRT and Number of root forks = NRF. Values represent the mean of 3 replicates ± standard deviation (SD). Bars associated with different lowercase letters show significant differences by Duncan Multiple Range's test (*p* ≤ 0.05).

#### *3.4. Content of H2O2 and MDA in the Shoots*

There was a significant (*p* < 0.001) genotype x phosphite x rate effect on H2O2 and MDA content in shoots of potato genotypes (Figure 4). Compared to the control plants, Phi increased H2O2 and MDA content in the shoots of the five genotypes by 50.96 to 68.84% and 43.63 to 70.17% at 0.25 mM, 56.01 to 71.64% and 48.32 to 71.48% at 0.5 mM, respectively. The highest H2O2 and MDA values were observed in GN2, followed by genotypes Longshu6, Qingshu9 and Longshu3. The lowest H2O2 and MDA levels were observed in Atlantic genotypes at Phi 0.25 and 0.5 mM, respectively. Moreover, similar results were obtained with respect to treatment without P supply in the five potato genotypes. However, Pisufficient plants in all five potato genotypes exhibited low H2O2 and MDA contents, compared to genotypes under Phi stress.

**Figure 4.** Effects of Phi, Pi and No-P on H2O2 and MDA content; (**a**) Shows H2O2 content and (**b**) shows MDA content in the shoots of potato genotypes. Values denote the mean of 3 replicates ± standard deviation (SD). Duncan multiple range test (*p* ≤ 0.05) shows large differences between bars assigned to different lowercase letters.

#### *3.5. Antioxidant Enzymes Activities and Content of Proline*

The effect of genotype x Phi interaction on antioxidant enzyme activities as well as the levels of proline was also significant (*p* < 0.01) in the shoots of the five potato genotypes (Figure 5). Compared with the individual control plants, Phi stress increased the proline contents in the shoots of all genotypes by 80.11 to 81.09% and 76.79 to 79.21% at Phi 0.5 and 0.25 mM, respectively. Moreover, the activities of CAT, POD and SOD increased by 46.07 to 55.54%, 50.77 to 53.14% and 54.39 to 63.09%, respectively, in all genotypes at 0.5 mM. The greatest increase in CAT activity was observed in the ATL genotype, while the LS6 genotype had the greatest POD and SOD activities. Similar increases in CAT, POD and SOD activities were observed in all genotypes at Phi (0.25 mM) and in No-P supplied potato genotypes.

**Figure 5.** Phi and Pi effects on enzyme activities and contents of proline; (**a**) Shows the activities of SOD; (**b**) Shows the activities of POD; (**c**) Shows the activities of CAT; (**d**) Contents of proline. Values denote the mean of 3 replications ± standard deviation (SD). Bars associated with different lowercase letters show significant differences by Duncan Multiple Range's test (*p* ≤ 0.05).

#### *3.6. Relationships between Root Morphological Characteristics, Fresh Biomass and Biochemical Responses of the Five Potato Genotypes under Phi Stress*

The correlation matrix between root morphological indices, fresh biomass, antioxidant enzyme activities, MDA, H2O2 and proline of the five potato genotypes under Phi stress were significantly negative (Table 2). All six root morphological indices were negatively correlated with MDA, H2O2, CAT, SOD, POD and Pro., e.g., TRL showed a significant negative correlation with MDA, H2O2, CAT, SOD, POD and Pro (r = −0.87 \*\*, r = −0.94 \*\*, r = −0.80 \*\*, r = −0.83 \*\*, r = −0.87 \*\* and r = −0.89 \*\*). The fresh biomass, i.e., FRW, FSW and TPW were also negatively correlated with MDA, H2O2, CAT, SOD, POD and Pro, e.g., FRW was negatively correlated with MDA, H2O2, CAT, SOD, POD and Pro (r = −0.67 \*, r = −0.75 \*\*,r= −0.62 \*, r = −0.69 \*, r = −0.69 \* and r = −0.72 \*\*). Principal component analysis (PCA) was used to determine the effects of Phi and No-P on root morphological indices, fresh biomass, antioxidant enzymes, MDA, H2O2 and proline (Table 3). The cumulative contribution percentage of the two principal components associated with the

response is 90.16%, the eigenvalue of PC1 is 8.818 and the contribution percentage is 62.98%. The eigenvectors include TRL, FRW, RPA, RSA, FRSR, NRF, RV and NRT. The eigenvalue of PC2 is 3.971, which corresponds to 27.18%, where the higher charges are CAT, SOD, MDA, POD and Pro show significant separations between the treatments. Genotype LS3 had the highest score followed by ATL, QS9, LS6 and GN2 at No-P, along PC1. On the other hand, ATL recorded the highest score followed by LS3, while GN2 had the lowest score at Phi (0.25 and 0.5 mM), along PC2. The treatments No-P and Phi (0.25 and 0.5) were far from the origin, implying that Phi and No-P strongly affected the root morphological characteristics and fresh biomass of potato genotypes (Figure 6). The comprehensive evaluation according to the principal component analysis of Phi tolerance was calculated using Formula (1) to calculate different products based on the two independent comprehensive indicators to obtain the membership function value μ(Xi) of each comprehensive index. Table 4 shows that, using the higher concentration of Phi (0.5 mM) treatment, in the same comprehensive index CI(1), Atlantic μ(X1) is the largest, with 0.937, indicating that Atlantic has the highest Phi tolerance on the CI(1) comprehensive index. According to the contribution rate of each comprehensive index, the weights of the two comprehensive indicators in terms of Phi tolerance were calculated using Formula (2) as follows: 0.699, 0.301. In addition, Formula (3) was used to calculate the Phi tolerance value according to the D value, and the Phi tolerance among the five genotypes was ranked as: Atlantic > Longshu3 > Qingshu9 > Longshu6 > Ganannong 2, respectively.

**Table 2.** Correlation matrix describing the relationship between root morphological characteristics, fresh biomass and activities of antioxidant enzymes, MDA, H2O2 and proline in potato plants under Phi stress at 20 days after treatments.


MDA = Malonaldehyde; H2O2 = Hydrogen peroxide; CAT = Catalase; SOD = Superoxide dismutase; POD = Peroxidase; Pro = Proline; TRL = Total root length; RPA = Root projected area; RSA = Root surface area; RV = root volume; NRT = Number of root tips; NRF = Number of root forks; FRW = Fresh root weight; FSW = Fresh shoot weight and TPW = Total plant weight. \* = significant difference at 5% probability level; \*\* = significant difference at 1% probability level.

> **Table 3.** Principal Component, loadings related to Phi treatments variable and explained variance for final sampling.


**Figure 6.** Scatter diagrams of arranged first two main components; PC1/PC2.



#### **4. Discussion**

In vitro regeneration studies are considered an efficient technique for defining the regulatory mechanisms of stress tolerance in potato explants. Several studies have shown the importance of in vitro shoot regeneration in plant breeding programs for economic drives [43]. To select more resistant cultivars for production, plant breeders need knowledge on how to improve regeneration of explants grown under stress conditions [44,45]. Five genotypes with different drought stress tolerance were selected based on sound data on their drought tolerance availability. These genotypes are widely grown and consumed in China. Therefore, it is important for farmers, consumers and breeders to understand their tolerance to PO3. In addition, we need to investigate whether the mechanism of

drought tolerance in potato differs from the mechanism that may support PO3 tolerance. In this study, we investigated the effects of Phi, Pi deficiency and Pi sufficiency on potato plant growth. Our data showed that Phi at concentrations of 0.25 and 0.5 mM severely inhibited plant growth, whereas Pi adequate plants grew normally across the five genotypes. We found that the number of leaves, roots, shoot length and root length of five potato genotypes cultured in Phi media were significantly reduced compared to those propagated in Pi-adequate media. Similar negative growth effects of Phi on *B. nigra* seedlings and *B. napus* cell suspension cultures have been documented previously [46]. Plants cultured in Phi media are very sensitive to Phi and show signs of toxicity such as leaf chlorosis and stunted growth [16,24,30]. These findings coincided with our observation in the present study. The observed decrease in plant growth in the presence of Phi could be related to two important morphological changes in the plants. Reduced internode length was an identifying feature of Phi-treated plants. Plant height was significantly reduced in Phi-treated plants in all five genotypes, whereas P-sufficient plants appeared normal in the presence of Pi concentrations (0.25 and 0.5 mM). The second important cause was a significant decrease in root development in Pi deficiency-treated plants and Phi treated plants. However, potato genotypes Atlantic and Longshu3 showed a modest increase in root development rate compared with the other genotypes studied at 0.25 and 0.5 mM Phi.

We further found that Phi at 0.25 and 0.5 mM altered the fresh biomass parameters such as fresh root weight, fresh shoot weight, total plant weight and root to shoot ratio in all genotypes as compared to Pi-sufficient plants. The reduction in fresh biomass of the five genotypes caused by Phi toxicity was more pronounced in fresh root weight than fresh shoot weight at both concentrations. This reduction had a large effect on total plant weight, resulting in reduced plant growth and development in Phi-treated plants, as previously documented for *B. nigra* plants [47]. The reduced root to shoot ratio observed in the majority of genotypes is consistent with the deleterious effect of Phi on root hair production, anthocyanin accumulation in the shoot, and stimulation of enzymes induced by Pi deficiency [48]. In summary, our experiment in potato is consistent with previous reports on the biological effects of Phi in *Brassica* sp. [49] and in tomato [50]. Under Phi supply conditions, even a low dose (0.25 mM) was sufficient to induce a decrease in fresh biomass. The majority of the potato genotypes showed a decrease in root/shoot ratio compared to Pi-sufficient plants, except Atlantic and Longshu3, which showed a marginal increase. Furthermore, the effect of Pi deficiency on potato genotypes was studied using a medium containing -Phi + -Pi, which corresponds to No-P. All genotypes under this treatment showed a significant decrease in fresh biomass except Atlantic and Longshu3 which showed an increase in root weight compared to the other genotypes. The considerable increase in root length and root weight of these genotypes resulted in a slight increase in root to shoot ratio compared to the respective controls. This observation corroborates the findings of Lambers and Plaxton [51], who observed that the absence of Pi in the growth media altered root phenotypes such as increased root hair density and root length as well as metabolism (e.g., release of carboxylates and Pi scavenging enzymes into the rhizosphere). The mechanism for overcoming the P deficit in growth media containing No-P plants is explained by an increase in root length [4,46,52]. Moreover, several studies have shown that plants under water stress increase root length in deeper profiles and that the main difference between shallow and deeper rooting genotypes is manifested in the stress conditions imposed in each case [53,54]. Biomass tolerance index (BTI) was calculated for each genotype in relation to the stress variables (0.25, 0.5 mM and No-P) to clearly define the tolerance of potato genotypes to Phi. The result showed that in No-P treatment, Longshu3 genotype had the highest BTI, followed by Atlantic, and Gannong2 had the lowest BTI. At Phi 0.25 and 0.5 mM, the Atlantic genotype had the highest BTI followed by Longshu3 genotype, while Gannong2 genotype had the lowest BTI.

Root growth and shoot growth are correlated with each other. According to Polania et al. [55], shoot growth supplies carbon and some hormones to the roots, while root growth supplies water, nutrients and hormones to the shoot. Despite the fact that no previous study has explicitly investigated the utility of root morphological traits for plant performance under Phi stress, the results of the current study suggest that Phi interference reduces root morphological traits in all genotypes except the Atlantic genotype. It has been suggested that the ability of Phi to limit *Arabidopsis* development is due to the competitive inhibition of Pi uptake and the inability of plants to readily utilize Phi through oxidation to Pi [46]. Phi cannot enter P metabolic pathways unless it is converted to Pi [16,56]. Moreover, the growth of plants cultivated in the Phi treatments was comparable to that of plants grown under the No-P supply treatment in terms of root morphological indices. These results confirm the findings of Lee et al. [57] for *Ulva lactuca*, Schroetter et al. [58] for *Zea mays*, Thao et al. [59] for *Brassica rapa*, Avila et al. [60] for *Zea mays*, Zambrosi et al. [61] for *Citrus* spp., and Hirosse et al. [62] for *Ipomoea batatas*. These researchers observed that Phi anion does not replace Pi anion in P nutrition of plants. They added that the use of Phi as the sole source of P resulted in a significant reduction in plant development compared to treatments with insufficient Pi fertilization. Root morphological traits such as architecture, branching, root volume, root hair length and density were found to have reflective effects on nutrient uptake from nutrient sources, which could be used to determine plant tolerance under stress conditions. According to Dawuda et al. [38], the size of root system of lettuce plants influenced their tolerance to cadmium stress in nutrient solution. The results of this study showed that the addition of Phi (0.25 and 0.5 mM) to MS media decreased the TRL, RPA, RSA, RV, NRT and NRF of all potato genotypes studied. This consequently reduced the size of root systems of most genotypes except Atlantic. Other researchers have found that plant uptake of P and other nutrients depends on root surface area, root system length and lateral roots to capture a large volume of nutrients in the soil/growth medium [63–65]. In addition, our results show that the Atlantic genotype, which had the larger root system, had root tolerance indices to Phi stress at both concentrations compared with the rest of the genotypes, which had the smallest root systems. The tolerance of the Atlantic genotype to Phi stress was confirmed by the largest tolerance indices for TRL, RPA, RSA, RV, and NRF. The larger root system possessed by Atlantic probably enhanced the uptake of other trace nutrients in the growth media, which contributed to its tolerance. The results of this study are in agreement with those of Wang et al. [66], who postulated that soybean genotypes with larger root systems are more tolerant to cadmium stress. Pandey et al. [67] also reported that genotype PDM-139, a green Gram genotype, with larger root surface area and root volume was the more tolerant genotype to low P compared to those with smaller root surface area and root volume. Several other reports suggest that larger root surface area, root volume, and root hair length of plants under P-stressed growth conditions are characteristics of tolerant genotypes [68,69].

In addition, plant tolerance to Phi-stress was also determined by the hydrogen peroxide (H2O2) content formed during stress exposure. Increased formation of reactive oxygen species (ROS) such as H2O2 induces oxidative stress due to the toxicity of Phi [70]. Zhang et al. [71] postulated that *Vicia sativa*, which had the lowest H2O2 content was more tolerant than *Phaseolus aureus* which had higher H2O2 content when both were exposed to cadmium stress. According to Oyarburo et al. [72], H2O2 accumulation in leaves is reduced, antioxidant enzyme activities are increased, gene expression is upregulated and accumulation of glucanases and chitinases is induced, which correlates positively with stress tolerance of the plant. In our current study, Phi interference was observed to increase the hydrogen peroxide content in the shoots of the five potato genotypes. Nevertheless, the Atlantic genotype had the lowest increase in H2O2 content, indicating that Atlantic is more tolerant to Phi stress than the other genotypes tested. These results indicate that potato genotypes with larger root systems are more tolerant to Phi stress than genotypes with smaller root systems. Antioxidant enzymes play a critical role in plant cell defense against stress-induced cell damage caused by the formation of free radicals, mainly in the form of

ROS. As a result, it has been suggested that increasing antioxidant enzyme activity may improve plant growth and yield. This result is in agreement with Ramos et al. [73], who indicated that increased SOD and CAT activity induced by low selenium concentrations increased leaf yield of *Lactuca sativa* L. Avila et al. [4] noted that CAT activity was low in Pi-sufficient plants but high by 71% in Phi-treated plants. Our current experiment provided similar results: the activities of antioxidant enzymes (e.g., SOD, POD and CAT) increased in all five genotypes in the presence of Phi and No-P treated plants, but decreased in Pi-sufficient plants. These antioxidant enzymes are involved in the detoxification of reactive oxygen species. Recent research has shown that Phi-anion can induce molecular changes that promote stress tolerance, such as activation of guaiacol peroxidase activity and lignin biosynthesis in maize [60], and structural and biochemical changes in periderm and cortex of potato tubers [20]. On the other hand, studies on the effects of Phi on antioxidant enzymes are still rare. All morphological variables of roots were significantly negatively correlated with antioxidant enzyme activities, MDA, H2O2 and Pro. Fresh biomass indices also showed significant negative correlation with antioxidant enzyme activity, MDA, H2O2 and Pro. The negative correlations observed under Phi stress indicated that Phi had a deleterious effect on potato root morphology and fresh biomass indices. To further confirm the tolerant potato genotype to Phi stress, a comprehensive evaluation analysis was conducted based on principal component analysis. The results evaluated the genotypes in this order: Atlantic, Longshu3, Qingshu9, Longshu6 and Gannong2.

#### **5. Conclusions**

Roots are responsible for the uptake of water and inorganic nutrients and are the primary organs affected by phosphite stress. Therefore, adaptation of roots to phosphite stress affects shoot response, physiological functions and plant growth. In the present study, the responses of five potato genotypes to 0.25, 0.5 mM Phi and Pi, and No-P were investigated. The results showed that Phi stress and No-P supply significantly reduced the size of root and shoot systems of the five potato genotypes tested. Nevertheless, the Atlantic genotype with the largest root system showed the highest tolerance to Phi stress by exhibiting the highest tolerance index values for total root length, root projected area, root surface area, root volume, number of root forks, and fresh biomass tolerance. H2O2 and MDA levels increased in shoots of all genotypes, but Atlantic genotype showed the least increase, indicating greater tolerance to Phi stress. Major antioxidant enzymes such as CAT, POD and SOD activities and proline content increased under stress conditions. Greater tolerance parameters and lower H2O2 contents were obtained from the Atlantic potato genotype under Phi stress, suggesting that potato genotypes with larger root systems may be more tolerant to Phi stress. The tolerance character of the Atlantic genotype was confirmed by comprehensive evaluation analysis using principal component analysis. The obtained results may be very useful for the selection of the genetically modified potato plants using the *ptxD* selection marker gene. However, the concentration and accumulation of P in shoot and root of Pi starved plants were not determined in the present study. The determination of this index will contribute to a better understanding of the mechanism of the negative effect of Phi anion on the physiological and morphological growth of potato genotypes. Therefore, future research should focus on the concentration and accumulation of P in shoot and root of plants grown under Phi stress and also determine the details of the molecular and genetic mechanisms of Phi tolerance in potato genotypes, especially in genotypes with relatively large root systems.

**Author Contributions:** Conceptualization, R.D. and C.S.; methodology, R.D. and C.S.; soft-ware, T.Q. and D.X.; validation, K.A. and Z.B.; formal analysis, R.D. and C.S.; investigation, R.D. and C.S.; resources, J.B.; data curation, D.X.; writing—original draft preparation, R.D., K.A. and C.S.; writing—review and editing, R.D., K.A. and C.S.; visualization, Z.B.; supervision, J.B.; project administration, J.B.; funding acquisition, J.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Natural Science Foundation of China (Grant No. 32060502 and 31960442), the Special Fund for Discipline Construction of Gansu Agricultural University (GAU-XKJS-2018-085, GAU-XKJS -2018-084), the special Fund for Talents of Gansu agricultural University (Grant No. 2017RCZX-44) and the Gansu Provincial Department of Education Fund (2019B-073).

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Twenty-Years of Hop Irrigation by Flooding the Inter-Row Did Not Cause a Gradient along the Row in Soil Properties, Plant Elemental Composition and Dry Matter Yield**

**Sandra Afonso 1,2, Margarida Arrobas <sup>1</sup> and Manuel Ângelo Rodrigues 1,\***


**Abstract:** In hops (*Humulus lupulus* L.), irrigation by flooding the inter-row can carry away suspended particles and minerals, causing gradients in soil fertility. The effect of more than 20 years of flooding irrigation on soil and plants was evaluated in two hop fields by measuring soil and plant variables in multiple points along the rows. In a second experiment 1000 kg ha−<sup>1</sup> of lime was applied and incorporated into the soil to assess whether liming could moderate any gradient created by the irrigation. At different sampling points along the rows, significant differences were recorded in soil properties, plant elemental composition and dry matter yield, but this was not found to exist over a continuous gradient. The variations in cone yield were over 50% when different sampling points were compared. However, this difference cannot be attributed to the effect of irrigation, but rather to an erratic spatial variation in some of the soil constituents, such as sand, silt and clay. Flooding irrigation and frequent soil tillage resulted in lower porosity and higher soil bulk density in the 0.0–0.10 m soil layer in comparison to the 0.10–0.20 m layer. In turn, porosity and bulk density were respectively positively and negatively associated with crop productivity. Thus, irrigation and soil tillage may have damaged the soil condition but did not create any gradient along the row. The ridge appeared to provide an important pool of nutrients, probably caused by mass flow due to the evaporation from it and a regular supply of irrigation water to the inter-row. Liming raised the soil pH slightly, but had a relevant effect on neither soil nor plants, perhaps because of the small amounts of lime applied.

**Keywords:** *Humulus lupulus* L.; soil porosity; soil bulk density; liming; hop ridges

#### **1. Introduction**

Hop plants (*Humulus lupulus* L.) require an adequate supply of water during the growing season to sustain their huge canopy [1]. In most of the hop producing regions of the world, the crop needs to be irrigated, particularly in lower latitudes of reduced precipitation in summer. Although hop fields have started to be drip irrigated all over the world, there is a long tradition of surface watering of this crop, by flooding the space between rows [1,2]. In this kind of surface or furrow irrigation system, water is applied at the top end of each furrow (in hops to the inter-row space) and flows down the field under the influence of gravity [3]. This is still the most commonly used irrigation method for hops in northern Portugal [4]. The water use efficiency with this irrigation technique is highly dependent on the field gradient and water infiltration rate, which can vary considerably, inducing spatial and temporal variability in the main soil properties [5]. In addition, flood irrigation can affect the spatial distribution of soil physicochemical properties which may exacerbate the spatial variability in crop growth and yield [6].

Flood irrigation can have a major impact on soil properties by varying salinity, redox potential, compaction and/or porosity [7–10]. Furthermore, hop fields which are

**Citation:** Afonso, S.; Arrobas, M.; Rodrigues, M.Â. Twenty-Years of Hop Irrigation by Flooding the Inter-Row Did Not Cause a Gradient along the Row in Soil Properties, Plant Elemental Composition and Dry Matter Yield. *Horticulturae* **2021**, *7*, 194. https://doi.org/10.3390/ horticulturae7070194

Academic Editors: Jolanta Jaroszuk-Sciseł, Małgorzata ´ Majewska and Agnieszka Hanaka

Received: 18 June 2021 Accepted: 13 July 2021 Published: 15 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

flood-irrigated need to be frequently tilled to control summer weeds and to reduce soil compaction and superficial crusts in the short term. This allows a better infiltration of water, but that can also have a negative impact on the soil in the long term [11,12]. Soil compaction, increased by furrow irrigation, may also reduce soil drainage and aeration, contributing to the reduction of soil redox potential which influences soil chemistry and plant nutrient availability [10,13]. The degree of compaction of a soil can be assessed by measuring some physical properties, such as bulk density and porosity [12,13]. As the soil becomes more compact, bulk density increases and soil porosity decreases, which reduces water and air diffusion into the soil [11,14]. In some hop fields in northern Portugal it was found that the decrease in soil redox potential, associated with an excess of water and/or poor drainage, was the main cause of the spatial variability found in crop growth and yield [4].

Soil pH is another relevant issue in hop production. The range of pH most suited for growing hops is considered to be between 5.7 and 7.5 [15,16]. The application of lime is recommended for acidic soils, and a positive relationship has been found between the increase in soil pH and hop yield [15,17]. However, the effect of liming on crops can also vary with the irrigation system. Some researchers have studied the influence of liming in rice under flooding conditions, since great interactions between flooding, soil acidity and nutritional disorders are usually found [18–20]. In hops, these interactions are less well known, or the response to liming, but it is believed that it may be relevant enough to be studied, since the crop continues to be irrigated by flooding in several parts of the world.

This study evaluated the variation in soil properties and nutritional status and the productivity of hop plants created along the rows by flooding irrigation. As a second line of study, the effect of the application of lime on soil properties and on hop nutritional status, growth and yield was evaluated, to ascertain if the application of lime could compensate for the variability created by the irrigation system. Both lines of study were carried out in commercial hop fields which had been flood-irrigated for over 20 years.

#### **2. Materials and Methods**

#### *2.1. General Experimental Conditions*

The field experiments were carried out during two growing seasons (2017 and 2018) on a commercial farm located in Pinela (41◦40 33.6" N; 6◦44 32.7" W), Bragança, north-eastern Portugal. A detailed location of field experiments is shown in Figure 1.

**Figure 1.** Map of Portugal indicating Pinela (**left**) and the two hop fields identified in this study as field 1 and field 2 (**right**). Images from https://www.google.com/maps/place/Pinela (accessed on 7 July 2021).

The region benefits from a Mediterranean-type climate, with an annual average temperature and accumulated precipitation of 12.7 ◦C and 772.8 mm, respectively. The

average monthly temperatures and precipitation recorded during the experimental period are shown in Figure 2.

**Figure 2.** Average monthly temperatures and precipitation recorded during the experimental period at a weather station located in Bragança, north-eastern Portugal.

The hop plots where the study was undertaken are ~2 ha in size each, with the rows having a length ranging from 150 to 180 m, and established with the cultivar Nugget. The fields are arranged in a 7 m conventional high trellis system, with concrete poles connected with steel cables, in a "V" design system. The farmer has managed the fields by flooding irrigation since the hop crop was installed more than 20 years ago. Several tillage passes (3 to 4) are performed every year to remove the crusts and facilitate water infiltration. The fertilization programme includes the application of a compound nitrogen (N): phosphorus (P): potassium (K) fertilizer (7:14:14, 7% N, 14% P2O5, 14% K2O) early in the spring, followed by two applications of N fertilizer (ammonium nitrate, 27% N) as a side-dressing, totalling ~150, 44 and 83 kg ha−<sup>1</sup> of N, P and K, respectively. The farmer also follows a phytosanitary programme for crop protection against pests and diseases.

#### *2.2. Field Experiments and Soil and Plant Sampling*

The first experiment (Experiment 1) was carried out during the growing season of 2017 in two hop fields. It consisted of the evaluation of soil properties, plant nutritional status and crop yield, searching for any gradient along the rows created by the irrigation system. The rows used in this experiment were divided into nine segments of equivalent length, creating nine positions (P1, P2, ... , P9) for soil and plant sampling. The soil was sampled between rows and on the ridges to a depth of 0.0 to 0.2 m. Three rows and inter-rows of hops were used to create three replicates for each position. Each soil sample for analysis resulted from six sampling points (composite samples). The soil was sampled by using an open-face auger.

For the determination of soil bulk density and porosity, a different approach to soil sampling was followed. It was found unnecessary to sample in the ridges since no compaction was expected in this part. Instead, the soil was sampled at two different depths, 0.0 to 0.10 m and 0.10 to 0.20 m. Due to the increased difficulty of sampling, particularly in the 0.10 to 0.20 m layer, only five positions were considered (P1, P3, P5, P7 and P9) and sampled in three replicates. For these analyses, undisturbed soil cores were taken by using appropriate cylinders of 100 cm3. Soil samplings were carried out on 10 March 2017.

The plants used in this experiment for the evaluation of their nutritional status and crop productivity were randomly selected and marked when plant height was close to 3 m (to avoid using very atypical plants) and close to each of the positions used for soil sampling. Leaf sampling for crop nutritional status assessment was done at ~2 m in height, on 17 July 2017. At harvest (1 September 2017), plant biomass was cut at ground level. Subsequently, the aboveground biomass was separated into leaves, stems and cones and weighed fresh. Subsamples of each plant part were weighed fresh again and then oven-dried at 70 ◦C and weighed dry for determination of dry matter yield.

The second experiment (Experiment 2) consisted of the application of 1000 kg ha−<sup>1</sup> of lime (55% CaCO3, 28% CaO and 20% MgO) in February 2017, to assess the liming effects on soil properties and plants in comparison to the untreated control. This experiment was also carried out in two hop plots. The general methodology for soil and plant sampling was similar to that reported for Experiment 1, consisting of marking nine positions along the rows. The soil was sampled on 4 January 2019, only between the rows, at 0.0–0.20 m soil depth, using an open-face auger. Leaf samples were taken at ~2 m in height, on 17 July 2017 and 18 July 2018. At harvest (1 September 2017 and 31 August 2018), plant biomass was cut at ground level and treated as reported for Experiment 1.

#### *2.3. Laboratory Analyses*

The undisturbed soil samples from Experiment 1 were oven-dried at 105 ◦C and weighed. Soil bulk density was estimated from the weight of dry soil divided by the volume of the cylinder. Soil porosity was determined as the ratio of nonsolid volume (soil particle density—bulk density) to the total volume of soil (soil particle density) [21]. The other soil samples from Experiments 1 and 2 were oven-dried at 40 ◦C and sieved in a mesh of 2 mm. The samples were analysed for pH (H2O and KCl), electrical conductivity (soil:solution, 1:2.5), exchangeable complex (ammonium acetate, pH 7.0) and organic carbon (C) (Walkley−Black method). Extractable P and K were determined by a combination of ammonium lactate and acetic acid buffered at pH 3.7. Soil boron (B) was extracted by hot water and the extracts analysed by the azomethine-H method. More details of these analytical procedures are given in Van Reeuwijk [22]. Other micronutrients [copper (Cu), iron (Fe), zinc (Zn), and manganese (Mn)] were determined by atomic absorption spectrometry after extraction with ammonium acetate and EDTA, following the methodology reported by Lakanen and Erviö [23].

Tissue samples (leaves, stems and cones) from both experiments were oven-dried at 70 ◦C and ground. Elemental tissue analyses were performed by Kjeldahl (N), colorimetry (B and P), flame emission spectrometry (K) and atomic absorption spectrophotometry (calcium (Ca), magnesium (Mg), Cu, Fe, Zn and Mn) methods after nitric digestion of the samples [24].

#### *2.4. Data Analysis*

Data was subjected to analysis of variance, according to the experimental designs, using SPSS program version 25. When significant differences were found between the experimental treatments, the means were separated by the Tukey HSD (sampling position) and Student's-*t* (field, sampling site, lime treatment) tests (α = 0.05). Linear regression analysis was performed to understand the effects of gradient on soil properties and plant nutritional status and productivity in Experiment 1 and the relationship between soil pH and plant variables in Experiment 2. The relation between the variables was obtained through correlation analysis with the Pearson coefficient, when the assumption of normality and linearity was accomplished; when this was not the case, the Spearman coefficient was used.

#### **3. Results**

#### *3.1. Gradients in Soil and Plants along the Rows*

#### 3.1.1. Soil Properties

The silt and sand contents varied significantly between the sampling positions (Table 1). The two fields also differed significantly in clay and sand content. The soil bulk density and soil porosity varied significantly between the sampling positions and fields but in the opposite way. The interaction between sampling position and field was significant for soil porosity, which means that the effect of the irrigation on this variable depended on the field.


**Table 1.** Soil separates and soil bulk density and porosity from samples collected at 0.0–0.20 m depth, in March 2017, as a function of sampling position (1, ... , 9), and field. Means followed by the same letter are not significantly different by Tukey HSD (sampling position) or Student's *t* (field) tests (α = 0.05).

> The soil bulk density was higher in the soil surface (0.0–0.1 m) when compared to the deeper (0.1–0.2 m) layer (Figure 3). The soil bulk density did not vary significantly along the rows for both soil depths. The soil porosity, in turn, was lower in the surface layer, and the gradient found along the rows was not significant for any of the soil layers. The soil bulk density and porosity varied significantly between the two fields, but the gradients found along the rows were not statistically significant.

**Figure 3.** Soil bulk density and porosity from soil samples taken at different sampling positions along the gradient of irrigation (1, . . . , 9), as a function of depth (D1, 0.0–0.10 m; D2, 0.10–0.20 m) and field (F1, field 1; F2, field 2).

Some other soil properties determined from the samples collected at 0.0–0.20 m depth varied significantly between sampling sites, sampling positions and fields (Table 2). Extractable P and K, conductivity, organic C, CEC and extractable Zn and B showed significantly higher values in the samples collected in the ridges. However, soil pH (H2O and KCl), base saturation and extractable Mn were significantly higher in the samples collected in the inter-rows. Most of the soil properties varied significantly between the sampling positions, the exceptions being soil pH, conductivity and extractable K. Soil properties also differed significantly between fields, except for soil conductivity. Significant

interaction between the sampling site and the field was found for extractable P, conductivity, exchangeable Ca and extractable Fe. Significant interaction between the sampling position and the field was found for organic C and extractable Fe, Mn, Zn, Cu and B. No significant interaction was found between the three factors of this experiment.

#### 3.1.2. Hop Dry Mater Yield and Leaf Nutrient Concentration

Aboveground dry biomass (stems, leaves, cones and total) in Field 1 showed a clear tendency for a decrease along the rows (Figure 4). However, the decrease was only statistically significant for stem dry matter yield (DMY). For all plant parts, the coefficients of determination (R2) were not particularly high, which helps to explain the lack of significant correlation between the two variables. In Field 2, no clear tendency was found in aboveground DMY.

**Figure 4.** Dry matter yield (DMY) from plants collected at harvest in September 2017, at different sampling positions along the gradient of irrigation (1, . . . , 9), and as a function of field (F1, field 1; F2, field 2).

N concentration in the leaves taken at 2 m height did not vary significantly along the rows in any of the fields (Figure 5). Leaf P also did not vary significantly along the rows but the values in Field 1 were lower than in Field 2. Leaf K levels did not vary significantly along the rows in Field 1 but increased significantly in Field 2. Leaf Ca and Mg levels showed a slight tendency to increase in both fields but without statistical significance. In general, the micronutrients showed even more erratic tendencies when the values of the two fields were compared and only the values of leaf Cu showed a significant decrease along the rows in Field 1.


**Figure 5.** Leaf nutrient concentration from samples taken at 2 m height and at different sampling positions along the gradient of irrigation (1, . . . , 9), as a function of field (F1, field 1; F2, field 2).

3.1.3. Correlation Analysis between Soil Properties and Plant Dry Matter Yield

Soil bulk density and porosity correlated in a different way with soil pH (H2O and KCl), leaf P and total DMY (Table 3). That is, the correlations of soil pH were positive for soil bulk density and negative for soil porosity at 0.0–0.10 m depth. Leaf P concentration was significantly and negatively correlated with soil bulk density at 0.10–0.20 m depth, in contrast to the positive correlation found with soil porosity. Leaf Fe concentration was found significant and negatively correlated only with soil porosity at 0.10–0.20 m depth. The strongest correlations were found for total DMY with soil bulk density (r = −0.706) and soil porosity (r = 0.714), both at 0.10–0.20 m depth.



Significant correlations were found for soil clay content, positive for soil pH and leaf K, and negative for leaf P, leaf Cu, total DMY and cone DMY. In contrast, soil sand content correlated significantly and negatively with soil pH (H2O) and leaf K, and positively with leaf Cu and B. Soil silt content correlated significantly and negatively with leaf B.

#### *3.2. Liming Experiment*

#### 3.2.1. Soil Properties

Most soil properties, such as extractable K, P, Mn, Zn, Cu, B, conductivity and pH presented significantly higher values in the limed plot in comparison to the untreated control (Table 4). Exchangeable Ca and CEC showed higher values in the limed plot but not significantly different to those observed in the control. Significant differences between the two fields used in this experiment were also found for most of the soil properties, the values of extractable K, P, Zn, Cu and B, conductivity, pH, exchangeable Ca, CEC and base saturation being significantly higher in Field 1. Only extractable Fe was significantly higher in Field 2. The interaction between liming and field was significant for extractable K, conductivity, and pH.

#### 3.2.2. Plant Response to Liming

The concentration of nutrients in the leaves taken at 2 m height showed significant differences between treatments for leaf P in 2017 and for leaf Fe and B in 2018 (Table 5). The values reported for P and Fe were significantly higher in the limed plots, and those reported for B were significantly higher in the control. Total and cone DMY were significantly lower in the limed plots with the exception of total DMY in 2017, whose differences between treatments were not statistically significant. When comparing fields, significant differences were found for some nutrients and total and cone DMY. However, only leaf concentrations of K, Cu and B, and total and cone DMY, maintained the same trend in both years and fields. In 2017, significant interaction between the liming treatment and the field was only found for leaf N and Mn and in 2018 for leaf P and total DMY.

#### 3.2.3. Correlation Analysis between Soil pH and Plant Variables

Significant correlations between the soil pH (H2O and KCl) and leaf nutrient concentration were found for several nutrients, but a similar trend over the two years was found only for leaf Cu and B, both presenting negative correlations with soil pH (Table 6). Soil pH and leaf P, for instance, showed a negative correlation in 2017 and a positive correlation in 2018. Significant and negative relations between soil pH and total and cone DMY were also found for the first year of plant sampling.



181


**Table 5.** *Cont.*

pHKCl

0.269

 0.526 \*\*

 0.585 \*\*

 0.436 \*

−0.321

Significant correlations at the

−0.317 correspondent

 levels of \* 0.05 and \*\* 0.01.

−0.670 \*\*

−0.127

−0.477 \*

 0.188

−0.066

−0.717 \*\*

#### **4. Discussion**

The results of Experiment 1 showed significant differences in some soil properties at different positions along the rows, but not over a continuous gradient. Thus, the results cannot be attributed to the flooding irrigation, but they were probably caused by heterogeneity in spatial variability of important soil constituents such as clay, sand and silt, since it is well-known that soil texture determines many other soil physical and chemical properties [25]. Variations in soil properties were also found when comparing different soil layers. The soil bulk density was higher in the soil surface layer (0.0–0.1 m), and porosity was found to be higher in the deeper (0.1–0.2 m), layer. The soil bulk density and porosity in agricultural fields are influenced not only by soil texture but also by external loads which cause soil compaction [13,26,27]. In this particular case, it seems that the effects of frequent irrigation and soil tillage prevailed, which may have prevented a proper soil aggregation, leading to an increase in soil bulk density and a reduction in soil porosity on the surface layer which was directly impacted by the cultivator. The variation in soil properties was also significant when comparing fields. The field higher in clay and lower in sand presented significantly higher soil bulk density. Usually, clayey soils tend to have a lower bulk density and higher porosity than sandy soils [28]. However, these results indicate an opposite trend, probably because of the negative effect on soil aggregation and compaction caused by frequent soil tillage. Other studies have also found spatial variability in bulk density and water infiltration on flooded fields caused mainly by tillage practices, particularly when heavy machinery is used [8,29].

The soil samples collected from the ridges showed significantly higher values of extractable P and K. In the ridges, the conditions for nutrient uptake were poor since they are created every year by soil pushed from the inter-rows, which means that they contain nutrients barely taken up by the plant due to the limited expansion of roots in this position. In addition, in this irrigation system, the water flows from the inter-row to the ridge due to the gradient of water potential caused by the evapotranspiration from the latter and the continuous water supply to the inter-row. This means that nutrients tend to accumulate in the ridge, carried by mass flow, in contrast to what happens in the inter-row, from which nutrients tend to be leached out. Mass flow is the main driving force causing the movement of most nutrients in the soil [30–32]. Thus, soil conductivity was higher in the ridge, due to the increased presence of salts as demonstrated by the increase in CEC. Organic C also appeared higher in the ridge, probably because this zone is not tilled so frequently, which reduces the exposure of organic matter to the heterotrophic microorganisms that cause its oxidation [33]. This zone also contains the remaining bines (those that do not climb) and weeds, which are incorporated into the soil when the ridge is created, which usually represent more debris than that incorporated in the inter-row. B also increased in the ridge, perhaps due to higher levels of organic C, which have the ability of retaining B in the soil [34,35].

Soil pH (H2O and KCl), base saturation and extractable Mn were significantly higher in the samples collected in the inter-rows. These results are probably related to the decrease in the potential redox, which may have increased the pH of the soil [36]. The increase in soil pH in the inter-rows was probably also related to the increase in the concentration of cation ions, such as Ca and Fe [37]. Base saturation increased in the inter-rows probably due to the presence of the divalent cations, less available to move into the ridge by mass flow. The higher concentration of Mn in the inter-rows might have also been due to the reduction of Mn that occurred at the beginning of the reduction process. This can occur when the redox potential is still positive [38].

A clear gradient along the rows was not observed for total and cone DMY. These results did not corroborate the hypothesis that flood irrigation is creating a spatial variation in plant performance along the rows. The differences detected in the plants seem to be due to spatial variability in the soil constituents, namely the soil separates which, in turn, influence soil bulk density and porosity. The results from the correlation analysis showed significant and negative relations between total DMY and soil bulk density (r = −0.706) and

between total DMY and clay content (r = −0.676). In contrast, total DMY and soil porosity at 0.10–0.20 m correlated significantly and positively (r = 0.714). The soil surface layer presented a higher bulk density, which has already been explained by the effect of irrigation and frequent soil tillage, which reduces the stability of soil aggregates, increasing bulk density and decreasing porosity [10,26]. On the other hand, it seems that the higher porosity in the 0.10–0.20 m layer was an important factor affecting DMY, likely because in the surface layer the diffusion of oxygen to ensure the biological processes of the soil is always easier. Soils with a higher clay content tend to retain more water, decreasing soil aeration which negatively affects the function of root and plant metabolism [13,39]. Under the conditions of this experiment, the clay content in the soil seemed to be negatively associated with hop DMY, mainly because clay is a determinant factor of soil bulk density and porosity, which were identified in this study as determinant factors in crop productivity.

Irrigation also did not cause any relevant gradient in tissue nutrient concentration as detected by the analysis of variance. However, correlation analysis provided some data that deserves to be commented on. Leaf P was significantly and positively correlated with soil porosity at 0.10–0.20 m, but was negatively correlated with soil bulk density at 0.10–0.20 m and clay content. Leaf P did not show any consistent gradient along the rows, but was lower in the field presenting a higher soil bulk density and clay content. This reveals that P uptake was enhanced by the increased porosity of the soil at the deeper layer and by the lower clay content. Similarly, on barley (*Hordeum vulgare* L.) there was reported a reduction in P uptake and yield associated with heavy soil compaction [13]. The higher porosity of soil may have facilitated P root uptake from the deeper layer, which is richer in P, probably due to the increase in the vertical movement of P as the result of fertilization and flooding as reported by [40]. In turn, the higher clay content may have resulted in higher P adsorption and lower P availability. In contrast, leaf Fe was significantly and negatively correlated with soil porosity at 0.10–0.20 m depth. Leaf Fe also presented an opposite tendency between fields, decreasing along the rows in the field with a lower clay content and higher soil porosity. This result is probably related to soil reduction conditions, as the availability of Fe decreases when soil oxygen and redox potential increases [37]. Leaf K showed a significant and positive correlation with soil clay content and a negative one with sand content. The availability of K in the soil is not directly affected by redox potential, but its fixation in 2:1 clay minerals is facilitated by the increase in soil pH [36]. There has also been reported an antagonistic effect between Fe and K in paddy fields [41,42], an aspect that may also have influenced these results.

In Experiment 2, the application of lime increased several variables of soil fertility, including pH, but did not significantly increase exchangeable Ca and CEC. In fact, the rate of lime applied in this experiment was too low to cause important changes to soil properties, as is usually achieved when using high rates of lime [32]. In a previous study, Ceh and ˇ Cremožnik [ ˇ 17] applied 2.3 t lime ha−<sup>1</sup> and reported similar results, that is, a reduced effect on soil properties due to the application of lime.

The main effect on the elemental composition of the leaves resulting from the application of lime would have been the significant increase of leaf P in the first growing season after the lime application. This raised the soil pH contributing to a reduction in P fixation, which in acidic soils is due to reactions with Al and Fe oxides, which precipitate P as AlPO4 and FePO4 [43].

Total and cone DMY did not increase with the application of lime, but rather showed a decreasing trend. It is generally considered that the optimal pH for hop growth is between 5.7 and 7.5 [15,44]. In this study, soil pH was below the lowest value of the reported range, which would have favoured a positive effect on the vegetation. However, the lime application influenced some soil properties, but not enough to have a high impact on the elemental composition of the leaves. In general, the nutrient content of the leaves was found to be within the sufficiency ranges established for hops [45], both in the limed and in the control treatments. Regarding total DMY, a significant interaction between lime treatment and field was recorded, which may also have contributed to difficulties in the interpretation of these results.

Correlation analysis, in turn, also did not show coherent trends over the two years of the study. Perhaps the most relevant result was the negative correlation between soil pH and biomass production in the first year, which again refers to diverse interactions which may have occurred between environmental variables (year) and factors under study (field and liming). The effect of environmental variables on the performance of the hop plant is well known [46–48], although in this study it was not possible to clarify the isolated effects of any of them.

#### **5. Conclusions**

Irrigation by flooding the space between rows over more than 20 years was not responsible for any gradient in soil properties, plant elemental composition and plant performance, although variations in those variables were found at different positions in the row caused by erratic spatial variability of some constituents of the soil, such as sand, silt and clay. However, irrigation followed by soil tillage on repeated occasions during the growing season seems to have reduced soil porosity and increased soil bulk density in the surface 0.0–0.1 m soil layer. These variables were found to be related to crop productivity in positive and negative ways, respectively.

This study also showed that the ridge is a point of nutrient accumulation, particularly for those that move more easily in the soil by mass flow, thereby showing also higher conductivity and CEC. The reduced water potential in the ridge created by evapotranspiration is the driving force causing the water flow from the inter-row. Organic C was also higher in the ridge in comparison with the inter-row, probably due to the annual incorporation of weeds and weaker hop bines (those that did not climb) when the ridge is created in early spring.

Although the original soil was acidic, and the application of 1000 kg ha−<sup>1</sup> of lime caused a small increase in pH, this did not lead to other relevant changes in soil properties, nor in plant nutrition status or total and cone DMY. The liming effect might not have been enough to nullify the effects of the interaction between factors that always occur in field experiments.

**Author Contributions:** Conceptualization, M.Â.R.; methodology, M.A. and S.A.; formal analysis, S.A.; investigation, S.A.; resources, M.Â.R. and M.A.; data curation, S.A.; writing—original draft preparation, S.A.; writing—review and editing, M.Â.R.; supervision, M.Â.R. and M.A.; project administration, M.Â.R.; funding acquisition, M.Â.R. and M.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors are grateful to the Foundation for Science and Technology (FCT, Portugal) for financial support from national funds FCT/MCTES, to CIMO (UIDB/AGR/00690/2020) and for Sandra Afonso's doctoral scholarship (BD/116593/2016).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** No new data were created or analyzed in this study. Data sharing is not applicable to this article.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Article* **Signal Intensity of Stem Diameter Variation for the Diagnosis of Drip Irrigation Water Deficit in Grapevine**

**Chen Ru, Xiaotao Hu \*, Wene Wang, Hui Ran, Tianyuan Song and Yinyin Guo**

Key Laboratory of Agricultural Soil and Water Engineering in Arid and Semiarid Areas, Ministry of Education, Northwest A&F University, Yangling 712100, China; chenru1024@nwafu.edu.cn (C.R.); wangwene@nwsuaf.edu.cn (W.W.); huiran@nwsuaf.edu.cn (H.R.); tianyuansong1014@nwafu.edu.cn (T.S.); guoyinyin0513@nwafu.edu.cn (Y.G.)

**\*** Correspondence: huxiaotao11@nwsuaf.edu.cn; Tel.: +86-138-9281-6133; Fax: +86-29-8708-2117

**Abstract:** Precise irrigation management of grapevines in greenhouses requires a reliable method to easily quantify and monitor the grapevine water status to enable effective manipulation of the water stress of the plants. This study describes a study on stem diameter variations of grapevine planted in a greenhouse in the semi-arid area of Northwest China. In order to determine the applicability of signal intensity of stem diameter variation to evaluate the water status of grapevine and soil. The results showed that the relative variation curve of the grapevine stem diameter from the vegetative stage to the fruit expansion stage showed an overall increasing trend. The correlations of MDS (maximum daily shrinkage) and DI (daily increase) with meteorological factors were significant (*p* < 0.05), and the correlations with SWP, RWC and soil moisture were weak. Although MDS and DI can diagnose grapevine water status in time, SIMDS and SIDI have the advantages of sensitivity and signal intensity compared with other indicators. Compared with MDS and DI, the R2 values of the regression equations of SIMDS and SIDI with SWP and RWC were high, and the correlation reached a very significant level (*p* < 0.01). Thus, SIMDS and SIDI are more suitable for the diagnosis of grapevine water status. The SIMDS peaked at the fruit expansion stage, reaching 0.957–1.384. The signal-to-noise ratio of SIDI was higher than that of MDS across the three treatments at the vegetative stage. The value and signal-to-noise ratio of SIDI at the flowering stage were similar to those of SIMDS, while the correlation between SIDI and the soil moisture content was higher than that of SIMDS. It can be concluded that that SIDI is suitable as an indicator of water status of grapevine and soil during the vegetative and flowering stages. In addition, the signal-to-noise ratio of SIMDS during the fruit expansion and mature stages was significantly higher than that of SIDI. Therefore, SIMDS is suitable as an indicator of the moisture status of grapevine and soil during the fruit expansion and mature stages. In general, SIMDS and SIDI were very good predictors of the plant water status during the growth stage and their continuous recording offers the promising possibility of their use in automatic irrigation scheduling in grapevine.

**Keywords:** grapevine; maximum daily shrinkage; daily increase; stem water potential; leaf relative water content; signal intensity

#### **1. Introduction**

Fruit tree orchards are common in arid and semi-arid areas where water for irrigation is scarce. This, together with an increasing world population that has to be fed and with other water-using sectors competing for the limited water resources, makes the use of precise irrigation techniques in those orchards unavoidable. The response of the scientific community to this challenge has been to invest a substantial amount of research in the development of deficit irrigation approaches [1,2] and of new irrigation technologies based on more-precise, user-friendly water-stress indicators. Some can be continuously and automatically recorded, having a great potential for irrigation scheduling [3].

**Citation:** Ru, C.; Hu, X.; Wang, W.; Ran, H.; Song, T.; Guo, Y. Signal Intensity of Stem Diameter Variation for the Diagnosis of Drip Irrigation Water Deficit in Grapevine. *Horticulturae* **2021**, *7*, 154. https:// doi.org/10.3390/horticulturae7060154

Academic Editors: Agnieszka Hanaka, Jolanta Jaroszuk-Sciseł and ´ Małgorzata Majewska

Received: 16 May 2021 Accepted: 9 June 2021 Published: 15 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

In recent years, it has become very popular to study the relationship between plant and water based plant water status indicators [4]. There are many ways to monitor and diagnose crop water status. From the perspective of plant physiology, the short-term microchange dynamics of plant organs (stems, leaves, fruits, etc.) are closely related to the water status of plants and have been widely focused on by many scholars [2–4]. The most widely used approach for evaluating plant water status has been to determine leaf [5,6], stem water potential [6] and leaf water content [7–10]. Plant water potential is a significant and reliable indicator of plant water status for scheduling the irrigation of plants. Argyrokastritis et al. [5] established the relationship between leaf water potential and water stress index, which can well characterize the water deficit status of plants. Wang et al. [6] used leaf and stem water potential to characterize the water status of grapevine, which can well judge whether the grapevine is in deficit state and determine when to need irrigation. However, the main disadvantage of plant water potential is the relatively cumbersome measurement procedure, including the necessity of frequent trips to the field and a considerable input of labor. The robustness of the sensors used to measure stem diameter fluctuations have renewed interest in using these parameters as plant water status indicators [1,6,11,12]. Apart from being capable of an early detection of water stress, even if this is mild, these techniques permit continuous and automated recordings of the plant water status and an immediate, consistent and reliable response to water deficit [13–15].

Monitoring crop moisture conditions using the stem diameter microchange method has been popular since the mid-1980s. The microchange of plant stem diameter is closely related to its water status, when the root system absorbs enough water, the stem expands gradually, and when the water is deficient, the stem shrinks gradually. Therefore, it is possible to diagnose the water status of the plant with the microchange of stem diameter, which can provide an index for the real-time prediction of crop water shortage and precision irrigation [12,16,17]. Today, in fact, fruit tree orchards and vineyards are being irrigated based on changes in the stem diameter [18,19]. Among them, MDS and DI indexes are commonly studied today as the indicators of plant water content [20–22]. According to some previous studies, the ability of an index to be suitable for use as a water diagnosis indicator was mainly evaluated in terms of three qualities: sensitivity, signal intensity and variability. An appropriate indicator should have better sensitivity and signal intensity to water stress and exhibit less variability [4,14,23]. However, the disadvantage of this method is that it cannot determine whether the critical value is independent of crop species or growth stage. In addition, meteorological factors have a great influence on MDS under the same water conditions [24]. The maximum stem diameter over time can be used to diagnose water deficit, but the growth rate of the crop is different at different growth stages. Under high evaporation intensity at the mature stage, crop stems may also shrink even if the crop does not lack water, and the variation in daily MXSD (maximum stem diameter) has no more significance [25]. Therefore, it may be difficult to apply the MDS of stem diameter as a crop moisture stress signal in practice.

The observation of stem diameter and its dynamic changes is beneficial to the study of plant moisture changes under interlaced internal and external factors. However, the observation of stem diameter is easily influenced by meteorological factors [26–28]. How to eliminate the interference of external environmental factors on the variation in stem diameter is always the difficulty in determining the most appropriate indicator. Due to the influence of other factors, it was difficult to diagnose plant water content. Only the observed values of stem diameter variation and the prediction values of stem diameter variation under no water stress, should be calculated to assist in the diagnosis. The comparison of MDS and DI with reference factors (relative values) can be used to directly reflect crop water status. The signal intensity is obtained by standardizing the reference value of the stem diameter indicator under fully irrigated conditions and the measured value of actual growth conditions, which can effectively eliminate the influence of meteorological factors [29]. Thus, the accuracy of these equations is very important to future studies, as these equations are the foundation for diagnosing plant water content and making irrigation schemes based

on plant stem diameter variations. Currently, existent studies concentrate mostly on the grapevine stem diameter variations in outdoor conditions and seldom did it to grapevines planted in a greenhouse [22,30,31]. In addition, the past research on the variation in stem diameter has mainly focused on the feasibility of moisture diagnosis, maximum daily shrinkage, daily increase, and other stem diameter indicators, which have been verified for use in the moisture status diagnosis of different crops [11,24,32–34]. The variation in stem diameter is influenced by environmental factors and the crop growth characteristics. The difference in crop growth at different growth stages may significantly affect the potential of stem diameter variation indicators for use in determining irrigation regimes. Therefore, different indicators should be applied in different growth stages [26]. There have been few reports on this topic, and further research is needed.

For these reasons above mentioned, this research selected the greenhouse grapevine as the study object. Because of its rich nutrition and delicious taste, grapevine has become a kind of world major fruit. In China, grapevines have been widely cultivated as well, and a considerable part among them are planted in sunlight greenhouses. Due to the present water shortage in the Northwest China, it is important to acquire accurate crop water content information and timely plan water-saving irrigation schemes, which benefit the sustainable development of local agriculture. Thus, the main aims of this study were as follows: (1) to explore the relative variation in the changes in MDS and DI in stem diameter during different stages; (2) to clarify the correlation of microchanges in stem diameter with stem water potential, leaf relative water content, and soil water content; (3) to evaluate whether SIMDS and SIDI can be applied to diagnose grapevine moisture and soil moisture status; (4) to analyze the sensitivity of signal intensity indicators and to determine the suitability of SIMDS and SIDI under different stages.

#### **2. Materials and Methods**

#### *2.1. Study Area*

The experiment was carried out in greenhouse of Yuhe Farm, Shaanxi Province, from March to July 2018 (108.58◦ E, 37.49◦ N). The annual average rainfall in this area is 365.7 mm, the annual average temperature is 8.3 ◦C, the annual relative humidity is 69.37%, and the annual average duration of sunshine is 2893.5 h, which is representative of the typical continental marginal monsoon climate of the area. Table 1 shows the meteorological data (cultivation stage averages) recorded over the experimental year. The test soil was an aeolian sandy soil. The chemical properties of the soil were as follows: the soil ammonium nitrogen was 7.48 mg·kg−1, the nitrate nitrogen was 22.91 mg·kg−1, the available phosphorus was 4.07 mg·kg<sup>−</sup>1, and the available potassium was 163.47 mg·kg<sup>−</sup>1. The physical properties of the soil are shown in Table 2.

#### *2.2. Experimental Design*

Six-year-old grapevine (early-maturing variety 6–12, which was selected from the scarlet bud transformation in 1998) were planted in greenhouse, and grapevines with good growth and similar shapes were selected for the experiments. The entire growth period of grapevines can be divided into four main growth stages: the vegetative stage, the flowering stage, the fruit expansion stage, and the coloring mature stage, the cultivation period was 121 days during the growth season. The greenhouse was oriented east-west and was 70 m long and 9 m wide. The grapevine row width and row spacing were 0.8 m and 1.5 m, respectively, and the plant spacing was 0.6 m, with 14 grapevines per row. Artificial warming was carried out in greenhouse to ensure the growth temperature of grapevines on 11 March 2018. Drip irrigation was used in the experiment. A single-wing labyrinth drip irrigation belt (produced by Xinjiang Dayu Water Saving Company, Xinjiang, China) was adopted. The inner diameter and wall thickness of drip irrigation belt was 0.02 m and 0.018 m, respectively. the distance between the drippers was 0.3 m, the design flow of the dripper was 4.0 L·h<sup>−</sup>1, and the laying mode of the drip belt was one row of two pipes.


**Table 1.** Meteorological data of different cultivation stages in greenhouse.

Note: Ta (°C): air temperature; Ra (w·m−2): solar radiation; RH (%): relative humidity; VPD (kpa): vapor pressure deficit.

**Table 2.** Physical properties of the soil.


Note: FC: field capacity; PWP: permanently wilting point.

The experiment was conducted with drip-irrigated grapevines under three irrigation treatments: a full irrigation treatment (T1:100% M) and two regulated deficit irrigation treatment (T2: 80% M; T3: 60% M), M represents the irrigation quota. There were three treatments in total and three plots per treatment (each plot had a length of 8 m, a width of 4.5 m, and an area of 36 m2), with a random block arrangement. The irrigation dates and irrigation amount is shown in Table 3, the grapevines were irrigated 12 times during the entire growth period. The irrigation quota was calculated by Equation (1). The irrigation time was determined according to whether or not T1 reached the lower limit of the water quantity, which was 65% of *β*<sup>1</sup> at the vegetative and coloring mature stages and 70% of *β*<sup>1</sup> at the flowering and fruit expansion stage. The predicted wet layer depth of the soil was 0.8 m. The total amount of fertilization during the entire growth period was 0.84 t·ha−1, and the proportion of N:P:K was 1.0:0.6:1.2. Fertilization was carried out over three periods: the germination stage accounted for 20% of the total amount of fertilization, the flowering and fruit expansion stages accounted for 60%, and the coloring mature stage accounted for 20%. The drip irrigation and fertilization were controlled by integrated irrigation and fertilization equipment.

$$M = 0.1\gamma\_s HP(\beta\_1 - \beta\_2) \tag{1}$$

where *M* represents the irrigation quota, mm; *γ<sup>s</sup>* represents the soil dry bulk density, 1.64 g·cm−<sup>3</sup> in 0–40 cm soil depth, 1.46 g·cm−<sup>3</sup> in 40–80 cm soil depth; *<sup>H</sup>* is the predicted wet layer depth of soil, 0.8 m; *P* is the designed wet soil ratio, 0.8; *β*<sup>1</sup> is the field water holding capacity, 13.18% in 0–40 cm soil depth, 17.45% in 40–80 cm soil depth; and *β*<sup>2</sup> is the lower limit of the soil moisture content, 65% of *β*<sup>1</sup> at the vegetative and coloring mature stages and 70% of *β*<sup>1</sup> at the flowering and fruit expansion stage.

#### *2.3. Observation Indicators*

#### 2.3.1. Meteorological Factors

All meteorological data are automatically measured and recorded every 30 min using a Watchdog micro series (Spectrum Technologies Inc., Chicago, IL, USA) meteorological station in the middle of the greenhouse. The monitoring indicators includes air temperature (Ta), relative humidity (RH), solar radiation (Ra) and other meteorological parameters. The vapor pressure deficit (VPD) was estimated by the RH and Ta and was calculated by the modified Penman formula. The formula [35] was as follows (2):

$$VPD = 0.6108 \times EXP (\frac{17.27 \times T\_a}{T\_a + 237.3}) \times (1 - \frac{RH}{100}) \tag{2}$$


**Table 3.** Irrigation amount of grapevine under different cultivation stages.

#### 2.3.2. Soil Moisture Content

The soil moisture automatic monitoring system consisted of an EM50 data recorder (Environmental Logging System, Decision Devices, Inc., Pullman, WA, USA) and four ECH2O5TE sensors (Decision Devices, Pullman, WA, USA). The soil moisture automatic monitoring system was installed 30 cm from the base of the grapevines and perpendicular to the planting row. Three representative grapevines were selected, and three monitoring systems were installed for each treatment. A sensor was installed every 20 cm, the buried depth was 80 cm, the soil volume moisture content was recorded every 30 min. Before the beginning of the growing season, in order to ensure the accuracy of the ECH2O5TE sensor, soil samples were taken every 20 cm with a soil drill until 80 cm, and the soil moisture content was calculated by drying method. At the same time, the data recorded by the ECH2O5TE sensor in different soil layers were recorded. Three days of soil moisture data were used to calibrate ECH2O5TE by drying method. The regression equation was established by regression analysis between the soil water content calculated by drying method and the soil water content monitored by ECH2O5TE. In addition, the same method was used to calibration ECH2O5TE every 10 days during grape growth.

#### 2.3.3. Stem Diameter Microchanges

The stem diameter microchanges were automatically monitored continuously using a DEX20 (Dynamax, USA, 0.050 mm) instrument. The instrument was installed at the stem 10 cm above the ground, with a maximum displacement of 5 mm and a recording interval of 30 min. The relative variation (RV) in stem diameter was defined as the change value at the time of probe installation, and the grapevine stem diameters were inconsistent when the sensor was installed due to differences among different plants. To explain the difference in plant growth caused by different water treatments, the initial value of stem diameter at the time of sensor installation in each growth period was set to 1 mm, so that the three treatments could be compared easily. The maximum daily shrinkage (MDS) was calculated by subtracting the minimum stem diameter (MNSD) from the maximum stem diameter (MXSD). Periodic changes in stem diameter were observed daily; MXSD usually occurred in the early morning and MNSD occurred at noon. The daily increase (DI) of stem diameter was obtained by subtracting the daily MXSD from that of the day before.

#### 2.3.4. Stem Water Potential and Relative Water Content of Leaves

The pressure chamber (TP-PW-II, Top Cloud-agri Technology Company, Zhejiang, China) was used to measure the stem water potential (*ϕs*) every 5–7 days, and the *ϕ<sup>s</sup>* is measured at 9:00 to 10:00 BJS. Three grapevines were selected for each treatment, and

one branch under good growth conditions was selected as the sample on the sunny side outside the crown. The sample was put into a plastic bag containing moist gauze and quickly brought into the laboratory. The sample was clamped in a pressure chamber and pressurized by gas (compressed nitrogen), the pressure used for exudation of tissue fluid was observed. At this time, the pressure value was the stem water potential.

The relative water content (RWC) of leaves was determined by the drying method. The selection and determination of leaves were the same as those used for leaf water potential measurement. After weighing the fresh weight, the leaves were immersed in water for 12 h and then taken out. The water on the surface of the leaves was wiped with absorbent paper and weighed. Then, the leaves were immersed in water for 1 h, taken out, wiped dry and weighed until reaching a consistent weight. After 0.5 h of dehumidification at 105 °C for 0.5 h, the leaves were dried to constant weight at 80 °C. Leaf relative water content (RWC) = (initial fresh weight − dry weight)/(saturated fresh weight − dry weight) × 100%.

#### 2.3.5. Signal Intensity Calculation of Stem Diameter Indicator

The reference value is usually calculated by the stem diameter indicator under nonwater stressed conditions or by substituting the meteorological indicator into the reference equation [21]. The calculation formulas of SIMDS and SIDI are as follows:

$$\text{SI}\_{\text{MDS}} = \text{Measured MDS} / \text{Reference MD} \tag{3}$$

$$\text{SI}\_{\text{DI}} = \text{Measured DI}/\text{References} \text{ DI} \tag{4}$$

In this study, the regression equation between the MDS and DI of stem diameter and meteorological factors showed that the correlation between the MDS and DI of three treatments and Ta was the best. The soil moisture of the W1 treatment always remained above 65% of the field capacity at the vegetative and mature stages and 70% of the field capacity at the flowering stage and fruit expansion stage. The reference equation for stem diameter was established between the MDS and DI values of the W1 and Ta to calculate the reference value of MDS and DI for each growth stage.

#### 2.3.6. Flexible Evaluation of Signal Intensity

High variability indicators need to be measured many times to reduce error, which increases the costs of such methods. Therefore, the intensity and variability (coefficient of variance, CV) of indicators should also be considered. When soil moisture changes, the ratio of signal intensity (SI) to noise is greater in the short term, indicating that the indicator is more suitable for moisture status diagnosis [36]. The formula for calculating the signal-to-noise ratio is shown in formula (5).

$$\text{Signal-to-noise ratio} = \text{signal intensity} / \text{coefficient of variation} \tag{5}$$

#### *2.4. Data Analysis*

The correlation and regression analysis were carried out using SPSS 21.0 software (SPSS Inc., Chicago, IL, USA). Multiple comparisons were performed by least significant difference tests, with a significance level of 0.05. Microsoft Excel 2010 Software was used for processing data. The graphs were created by using Origin 2018. Correlation analysis was conducted between MDS, DI, SIDI, SIMDS and meteorological factors; The relationships between MDS, DI, SIDI, SIMDS and SWP, RWC as well as between MDS, DI, SIDI, SIMDS and soil water content were analyzed through regression analyses. In all cases, the coefficient of determination (R2) was used to assess the goodness of fit of the associations among variables.

#### **3. Results**

#### *3.1. The Relative Variation of Stem Diameter under Different Stages*

The relative variation (RV) curve of stem diameter under different stages showed a 24 h up and down cycle, and different irrigation amounts had different influences on the stem diameter of grapevine (Figure 1). The total increase in stem diameter was 0.128 mm under W1, while those of W2 and W3 fluctuated and decreased. The stem diameter of W3 began to decrease sharply after 7 April, and the total increase in stem diameter under W2 and W3 was −0.143 and −0.570 mm, respectively (Figure 1a). There was a certain difference in the RV curve between the vegetative and flowering stages, and the RV curves of stem diameter under three treatments showed an up and down growth trend (Figure 1b). In terms of the total increase in stem diameter, W1 was the largest (0.555 mm), and W2 and W3 showed values 69.55% and 32.79% of that under W1, respectively.

**Figure 1.** Relative variation curve of stem diameter during different stages. (**a**) Represents the vegetative stage, (**b**) represents the flowering stage, (**c**) represents the fruit expansion stage, and (**d**) represents the coloring mature stage.

There were significant differences in the RV curve of stem diameter among different treatments during the fruit expansion stage (Figure 1c). The RV curve of W1 showed an upward trend, and the RV curve of W2 fluctuated by approximately 1 mm. Before irrigation on 20 May, the RV curve of the W3 treatment showed a decreasing fluctuating trend, the diameter of the stem recovered after irrigation, and the recovery effect gradually weakened when the soil moisture content gradually decreased. The stem diameter recovered values under W2 and W3 were 86.28% and 72.79% of that under W1, respectively. The stem diameter RV under W3 remained stable at approximately 0.4 mm after 25 May. The RV curve of W1 and W2 showed a decreasing trend during the mature stage (Figure 1d), the RV curve of W3 fluctuated at 1 mm, but the three treatments still had significant shrinkage. The contractions under W1 and W2 were more pronounced than those under W3. The total increase in stem diameter among the three treatments was negative.

The daily change of stem diameter was the same under three treatments (Figure 2), it showed a trend of first increasing, then decreasing, and then increasing over 24 h. The MDS of the stem diameter showed significant differences under three treatments. The MDS of W3 was the largest at 0.138 mm, and W1 was the lowest at 0.051 mm. The stem diameter of W1 could recovered to the maximum of the previous day and continued to grow. However, the stem diameter under W2, W3 could not recover to the maximum of the previous day owing to moisture stress, and growth of both W1 and W2 appeared negative. The MXSD and MNSD of three treatments appeared at the same time on rainy days, which included that rainy weather had no significant effect on the occurrence time of MXSD and MXSD under different treatments. The variation in Ra, Ta and RH on rainy days was smaller than that on sunny days, and the degree of stem diameter contraction was also lower on rainy days than that on sunny days, which showed meteorological factors may be the main force affecting the variation in stem diameter. We can conclude that MDS decreased with increasing irrigation amount, the influence of meteorological factors and soil moisture on the variation in stem diameter was interactive.

**Figure 2.** The daily change in stem diameter of grapevine under different weathers. Sunny days: 6/23, 6/24; Rainy day: 6/25.

#### *3.2. Evaluation of Applicability as a Moisture Diagnosis Indicator*

#### 3.2.1. The Correlation of MDS and DI with Meteorological Factors

The MDS and DI are the main components of stem diameter variation and are affected by soil moisture and meteorological factors [22]. Correlation analysis was carried out between MDS (Figure 3) and DI (Figure 4) in stem diameter and meteorological factors. The change in MDS was similar to those in meteorological factors (Ra, Ta and VPD). The MDS increased with increasing Ra, Ta and VPD and decreased with increasing RH. The correlation of DI with meteorological factors was opposite to that of MDS. The correlations between the MDS and DI of the three treatments and meteorological factors were significant (*p* < 0.05). The correlation between MDS and meteorological factors decreased with increasing irrigation, and the correlation between the DI and meteorological factors increased with increasing irrigation. It can be seen from Table 4 that the correlation coefficients of MDS, DI and Ta were the highest, at 0.601–0.692 and 0.683–0.723, respectively.

#### 3.2.2. The Correlation of MDS, DI with Stem Water Potential and RWC of Leaves

Stem water potential (SWP) and relative water content (RWC) of leaves are important indicators to characterize plant water status and exhibit the most direct response to drought during crop growth [37,38]. The models of MDS, DI and SWP and RWC were established. The coefficient of determination (R2) is shown in Table 5. The regression equations of MDS and DI with SWP and RWC generally meet the significance test, which shows that both MDS and DI have obvious correlation with SWP and RWC, but R2 values are not high. We found that, compared with W1 and W2, the R<sup>2</sup> values of MDS with SWP and RWC in W3 treatment were below 0.20, and the significance was weak. Therefore, MDS and DI were

easily disturbed by meteorological factors and could not be directly used for the diagnosis of grapevine water status.

**Figure 3.** The relationship of MDS with Ta (**a**), Ra (**b**), RH (**c**) and VPD (**d**).

**Figure 4.** The relationship of DI with Ta (**a**), Ra (**b**), RH (**c**) and VPD (**d**).


**Table 4.** Correlation analysis of MDS and DI in different stages under different treatments.

Note: The correlation between MDS and meteorological factors. \*\* Correlation is significant at the 0.01 level.

**Table 5.** Stem water potential and leaf relative water content model with MDS and DI.


\* indicates a significance level of *p* = 0.05; \*\* indicates a significance level of *p* = 0.01.

#### 3.2.3. The Correlation of MDS and DI with Soil Moisture under Different Stages

Soil moisture has been widely used as an indirect index of crop water deficit. It can be used to diagnose crop moisture status if stem diameter variation is sensitive to soil moisture. Regression analysis was carried out between MDS and DI in stem diameter and soil moisture (Figure 5). It can be concluded that the response of DI to soil moisture was more sensitive during the flowering stage, while that of MDS was more sensitive during the fruit expansion stage. In general, the R2 values of the MDS and DI models were low under different treatments. The reason for this result is that the MDS and DI in grapevine stem diameter were easily affected by meteorological factors. Therefore, combined with the above conclusion, it is necessary to eliminate the interference of meteorological factors on the stem diameter variation.

**Figure 5.** The correlation of MDS and DI with soil moisture content under different stages. (**a**) Represents the correlation of MDS with soil moisture content; (**b**) represents the correlation of DI with soil moisture content. \* indicates a significance level of *p* = 0.05; \*\* indicates a significance level of *p* = 0.01.

#### 3.2.4. Signal Intensity of Stem Diameter Indicator

The signal intensities of MDS (SIMDS) and DI (SIDI) under different treatments at different stages were significantly different (Figure 6). The SIMDS fluctuated by approximately 1 mm from the vegetative stage to fruit expansion stage under W1, while it mostly fluctuated below 1 mm at the mature stage. The SIDI fluctuated approximately 1 under W1 during the whole growth period. The degree of change in SIMDS and SIDI under W2 and W3 increased sequentially, and the signal values of the same stage were larger than those of the W1. It can be seen that with the increase in irrigation, the SIDI and SIMDS values of stem diameter tended to become stable. The SIMDS of the three treatments peaked during the fruit expansion stage (Figure 6e), followed by the flowering stage (Figure 6c). The SIDI of the W3 dropped sharply to below 0 on 26 May and was restored after rehydration on 27 May. Compared with those under the other growth stages, the SIMDS and SIDI of each treatment decreased to different degrees at the mature stage (Figure 6g,h), and that of W3 treatment decreased most significantly. The SIMDS of the W3 decreased rapidly to below the W1 level within one week after irrigation stopped and then remained stable. The SIDI was difficult to use to distinguish the moisture status of grapevines at the mature stage due to the large fluctuation and instability of the signal values. In conclusion, SIMDS was preliminarily determined to be an appropriate indicator of plant moisture status, and SIDI can be used as an indicator of moisture status at stages other than the mature stage.

3.2.5. The Correlation of SIMDS and SIDI with Meteorological Factors, Stem Water Potential and RWC

The correlation coefficients of the SIMDS and SIDI with meteorological factors during different growth stages are shown in Figure 7. In addition, while the individual correlation coefficient reached a significant level, the correlation of the SIDI and SIMDS of three treatments with meteorological factors during the whole growth period was not significant (Figure 7), which indicated that the influence of meteorological factors on SIDI and SIMDS had been excluded.

It can be seen from Table 5 that the correlations between MDS and DI and SWP and RWC are relatively low. Therefore, after eliminating the interference of meteorological factors, the fitting diagrams and equations of SIMDS and SIDI with SWP and RWC are established (Figure 8 and Table 6). With the increase in SIMDS, SWP under W3 decreased most significantly, RWC under W1 decreased most significantly; with the increase in SIDI, SWP and RWC under W1 had the most significant increase. On the whole, the R2 of each equation was high, and the correlation reached a very significant level (Table 6). We concluded that the ability of SIMDS and SIDI to represent the deficit status of plants is greatly improved after eliminating the interference of meteorological factors.


**Table 6.** Stem water potential and leaf relative water content model with SIMDS and SIDI.

Note: \*\* indicates a significance level of *p* = 0.01; \*\*\* indicates a significance level of *p* < 0.001.

**Figure 6.** The change in SIMDS and SIDI under different stages of grapevine. SIMDS represents signal MDS of stem diameter, and SIDI represents signal DI of stem diameter. (**a**,**c**,**e**,**g**) Represents SIMDS at the vegetative, flowering, fruit expansion and mature stage, respectively. (**b**,**d**,**f**,**h**) Represents SIDI at the vegetative, flowering, fruit expansion and mature stage, respectively.

3.2.6. Adaptable Evaluation of Signal Intensity under Different Stages

In addition to using signal intensity of stem diameter to characterize the water status of plants, the signal intensity of stem diameter was also explored to monitor soil water content and diagnose whether grapevines were under water stress in real time. Therefore, the regression model between the SIMDS and SIDI of stem diameter and soil moisture was established (Table 7). It can be seen from Table 7 that the determination coefficient (R2) of the SIMDS model under three treatments were high during the growth stage, which indicated that SIMDS showed a good diagnostic effect on soil moisture status. The R2 of the SIDI model was higher at the vegetative and flowering stages, and R2 decreased to 0.022–0.232 (*p* > 0.05) at the mature stage. We can conclude that SIMDS and SIDI were more suitable for diagnosing soil water status than MDS and DI. However, the diagnostic effect of the two indicators was quite different under different stages, so it is necessary to further consider the moisture sensitivity under different stages to select the optimum indicator.

**Figure 7.** The correlation coefficients between SIMDS, SIDI and meteorological factors under different treatments at different stages.

**Figure 8.** Relationships between SIMDS and SIDI and SWP and RWC under different treatments. (**a**) Represents the relationship between SWP and SIMDS, (**b**) represents the relationship between SWP and SIDI, (**c**) represents the relationship between RWC and SIMDS, and (**d**) represents the relationship between RWC and SIDI.


**Table 7.** The model of SIDI and SIMDS with soil water content at different growth stages.

Note: θ represents the soil moisture content. \* indicates a significance level of P0.05; \*\*\* indicates a significance level of *p* < 0.001.

The signal intensity of MDS and DI (SIMDS, SIDI) and the coefficient of variation of signal intensity under different stages are shown in Figure 9. The signal-to-noise ratio of SIMDS and SIDI under different treatments at different stages can be seen in Figure 10. The SIMDS and SIDI increased first and then decreased over the whole growth stage (Figure 9). We found that the average values of SIMDS and SIDI under different treatments were similar at the vegetative stage, but the variability of SIMDS was greater than that of SIDI (Figure 10), so SIDI was a more suitable diagnostic indicator of grapevine water status and soil water status during the vegetative stage. The average signal intensity, sensitivity and signal-tonoise ratio were similar at the flowering stage, but SIDI had a better correlation with soil moisture at the flowering stage (Table 7); thus, SIDI should be selected as the diagnostic indicator of grapevine water status and soil water status at the flowering stage. The signalto-noise ratio of SIMDS during the fruit expansion and the mature stages was higher than those of SIDI (Figure 10), and the sensitivity of SIMDS to soil moisture was better than that of SIDI. Therefore, SIMDS was selected as the most suitable indicator of grapevine water status and soil water status during the fruit expansion and the mature stages.

**Figure 9.** The average value of SIMDS and SIDI and coefficient of variation of signal strength under different growth stages. (**a**) Represents the vegetative stage, (**b**) represents the flowering stage, (**c**) represents the fruit expansion stage, and (**d**) represents the mature stage. Different lowercase letters indicate that there is a statistical difference at P0.05 under different treatments.

**Figure 10.** The signal-to-noise ratio of SIMDS and SIDI under different treatments at different stages. (**a**) Represents the vegetative stage, (**b**) represents the flowering stage, (**c**) represents the fruit expansion stage, and (**d**) represents the mature stage. Different lowercase letters indicate that there is a statistical difference of signal-to-noise ratio at P0.05 in the same column.

#### **4. Discussion**

#### *4.1. Relative Variation in Grapevine Stem Diameter*

To achieve sustainable water use and efficient cultivation of crops, the moisture condition of crops is an important factor. Under both high and low soil moisture conditions, the grapevine stems shrunk in the daytime and recovered or expanded at night, and the microchange in stem diameter was closely related to the water status of the plant [39,40]. The present study showed that the stem diameter under the W3 treatment began to decrease sharply after 7 April, and the total increase in stem diameter was −0.570 mm. Because the growth of new shoots mainly depends on the absorption of soil moisture and nutrients by the root system and transport of these nutrients to the new shoots, the soil moisture under the W3 treatment was low. When transpiration stopped at night, the moisture absorbed by the root system was not sufficient to make up for the transpiration loss during the day, so the increase in stem diameter stopped or stems even exhibited negative growth, similar to the results of Xiong [41].

The difference in the MDS among the three treatments was not significant during the vegetative and flowering stages. The reason for this phenomenon is that the rapid growth rate conceals the short-term variation in stem diameter caused by water deficit, which indicates that MDS is not a suitable indicator of moisture status during early grapevine development [42]. During the mature stage, the growth of the stems slowed with the seasonal process. At the fruit growth stage, the stems also ceased growth or shrunk without water stress [43,44]. The results showed that the relative variations in stem diameter among the three treatments decreased gradually, similar to the results of Intrigliolo and Castel [44] and Girón et al. [44]. With the increase in water stress, the MDS under the W3 treatment was the lowest at the flowering and fruit expansion stages. This effect may be the result of the combination of the degree of water stress on the plant and the ability of the tissue to hold water against the water potential gradient [12]. Further research is needed to more accurately explain these findings.

#### *4.2. The Correlations of MDS and DI with Meteorological Factors*

Numerous studies have shown that MDS is sensitive to soil and plant moisture status, and this measure has been applied in production as a key indicator to guide fruit tree irrigation [31,45–47]. The DI reflects the growth rate of the stem, which is affected by the water supply in the root zone and the intensity of transpiration. The DI is also

sensitive to the plant moisture status during the rapid growth stage of crops. Therefore, it is particularly important to analyze stem diameter microchanges for diagnosing crop water deficit. Previous studies have shown that the key meteorological factors affecting stem diameter of fruit trees under outdoor conditions are daily mean water pressure deficit (VPD) or daily maximum temperature (Tmax) [16,23,48]. In the current study, we found that Tmax and VPD are the key meteorological factors affecting MDS and DI, which is consistent with previous studies on outdoor fruit trees. In addition, the plastic film on the top of the greenhouse has good light transmittance, which can well transmit sunlight into the grapevine of the greenhouse, which also explains the reason why the indoor and outdoor results are similar.

Our results showed (Figures 3 and 4) that the positive correlation between MDS and VPD was extremely significant (*p* < 0.01), and the negative correlation between DI and VPD was significant (*p* < 0.05). Goldhamer et al. [16] indicated that VPD was the main factor affecting the stem diameter variation of almond trees. Moriana et al. [49] showed that Ta had the best correlation with MDS, followed by VPD, which was different from the results of this study, and these discrepancies may have been caused by the differences in plant type and test sites. Therefore, when the crop growth environment is changed, the relationship between the indicator of stem diameter and meteorological factors may change, so the reference equation obtained under a specific condition cannot be used to calculate the reference value, which needs to be further referenced with local meteorological data.

In addition, under the same experimental conditions, due to the different responses of different grapevine varieties to water stress, there may be differences in the stomatal resistance, transpiration rate, and photosynthetic rate under different varieties following water deficit, which may cause changes in sap flow in the stem, resulting in differences in the stem diameter indicators of different grapevine varieties [50,51]. Therefore, the reference equation obtained on a certain grapevine variety and the SIMDS and SIDI irrigation threshold values may no longer be applicable. When calculating the crop reference value, it is better to use the same crop variety under the same growth conditions as the reference crop.

#### *4.3. Signal Value and Signal-to-Noise Ratio of the Stem Diameter*

The stem diameter indicator is greatly affected by external meteorological factors. The model of the correlations between MDS and DI and stem water potential, RWC and soil moisture cannot exclude the interference of external factors [14]. According to the experimental results, the SIMDS and SIDI had better sensitivity, signal intensity and reliability in diagnosing the grapevine water content. In addition, the diagnostic applicability of SIMDS and SIDI was different in different growth stages of grapevine. A possible explanation might be the great coefficient of variation among grapevine plants growth rates masked the differences created by water stress on SIMDS at early and middle growth stage. Thus, the SIDI is a more appropriate indicator of the grapevine water content than SIMDS during the vegetative and flowering stages, but as important plant growth indexes, the variation of SIMDS and SIDI should be taken into full consideration in practice as well. The signal value of the stem diameter variation indicator can eliminate the interference of meteorological factors, so an accurate reference value is critical for the use of signal intensities for guiding crop irrigation [52,53]. The SIMDS and SIDI values of the W1 treatment fluctuated up and down by approximately 1 (Figure 6). The SIMDS and SIDI maximum of the W2 and W3 treatments under water stress was approximately 2.0, and the water deficit led to a significant increase in SIMDS and SIDI compared with the full irrigation treatment. The reason for these findings may be related to the water absorption by the root system and the ability of the stem to transport water following water stress, but the details of these mechanisms should be researched in the future. The SIDI at the mature stage showed an irregular curve with large fluctuations, which may have been caused by the large fluctuation of meteorological factors at the mature stage. However, the variation curve of SIMDS at the mature stage was relatively stable, which may have been related to

the stem growth characteristics of grapevine at the later growth stage; the relevant internal physiological mechanism needs to be further explored.

Although stem water potential and RWC measurements require frequent trips to the field and a considerable input of labor, these parameters are reliable plant-based water status indicators and have been used for irrigation scheduling in fruit trees [43,54–56]. The present results showed that there were weak correlations between MDS, DI and stem water potential and RWC. These results were due to the unstable meteorological factors during this period, which affected the short-term grapevine stem growth to a certain extent, resulting in the unstable changes in stem diameter. However, when the interference of meteorological factors on MDS and DI were eliminated, SIMDS and SIDI were not sensitive to meteorological factors. According to the fitting equations of grape stem water potential and RWC with SIMDS and SIDI, the fitting effect of each equation is good (Figure 8). The coefficient of determination (R2) is approximately 0.7, which indicates that it is feasible to use the signal intensity of stem diameter variation to characterize crop water status. This approach can not only substantially reduce the labor required but also can be used to continuously and nondestructively carry out index observation, which is consistent with the research results of Badal et al. [17]. Furthermore, when using this index to diagnose the plant water content, the best time is on a fine day, as the stem diameter microchange indexes were small and not significantly different (*p* > 0.05) between stress and full irrigation treatments under poor weather conditions in this experiment [6].

It can be seen from Figures 9 and 10 that the larger coefficient of variation (CV) of SIMDS during the vegetative and flowering stages, which resulted in lower signal-to-noise ratio. While SIDI has a greater variability in the late growth stage, the larger CV of SIMDS and SIDI increased the uncertainty of judging water stress in grapevine. In order to further explore the applicability of SIMDS and SIDI at different growth stages, more sensors should be installed to acquire the real water content message in practice. Previous studies have shown that some factors could affect the plant CV of MDS such as the crop load, location of sensor installation, and so on [43,48,49,57]. The relative researches should be studied further in future experiments to facilitate practical operation of this technic.

#### **5. Conclusions**

This work shows that there were significant differences in stem diameter variation under different irrigation levels. Water shortage resulted in larger maximum daily shrinkage and smaller daily increase. The stem water potential and leaf relative water content of stress plants (W2 and W3 treatment) were significantly lower than that of the W1 treatment. Regression analysis between MDS, DI and meteorological factors revealed that the MDS of stem diameter was positively correlated with Ra, Ta and VPD and negatively correlated with RH, and the DI among three treatments decreased with the increase in Ra, Ta and VPD and increased with the increase in RH. The key meteorological factors influencing grapevine stem diameter variation in a greenhouse were VPD and Ta. The MDS and DI had a weak correlation with stem water potential and RWC, thus these measures cannot be directly applied as indicators of the moisture status of grapevine and soil. SIMDS and SIDI can distinguish the differences in the grapevine stem diameter indicators under different soil moisture conditions, eliminate the interference of meteorological factors, and were highly correlated with stem water potential, RWC and soil moisture content. At the vegetative and flowering stages, SIDI has less variability and greater reliability than SIMDS, it is more suitable for the diagnosis of grapevine water status in these two periods. At the fruit expansion and the mature stages, the signal-to-noise ratio of SIMDS is significantly higher than that of SIMDS, so it is more suitable to be used as a diagnostic index of water status in late growth stage of grapevine. In sum, compared with other plant water diagnosing indexes, the SIMDS and SIDI indexes had the advantages of sensitivity, signal intensity and reliability and were good indicators of the grapevine water content.

**Author Contributions:** C.R., X.H., W.W. and H.R. designed the experiments, C.R., Y.G. and T.S. performed research and date analysis, C.R. wrote the manuscript with contributions from all authors. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Key Research and Development Program of China (2017YFD0201508), the National Natural Science Foundation of China (51179163), Shaanxi Key Science and Technology Innovation Team Project (2016KTZDNY-01–05).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data is available upon request to the corresponding author.

**Acknowledgments:** The authors would like to acknowledge all the team members of key Laboratory of Agricultural Soil and Water Engineering in Arid and Semiarid Areas, Ministry of Education.

**Conflicts of Interest:** The authors declare that they have no conflict of interest.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Horticulturae* Editorial Office E-mail: horticulturae@mdpi.com www.mdpi.com/journal/horticulturae

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com