Vineyard protection with rain-shelter: relationships between radiometric properties of plastic covers and table grape quality

The rise of unpredictable, heavy rainfalls increases incidence of diseases, berry rot, berry cracking, and enhance the need for rain shelters. The influence of the type of rain shelter on microclimate and grape quality is still little explored. The aim of this trial was to investigate the radiometric properties of a new type of cover, its effect on vineyard microclimate and on table grape quality, as compared with a reference cover, during a vintage with unusually abundant precipitation for Southern Italy. The new cover was characterized by lower transmissivity to IRshort radiation, and by a very higher transmissivity to UV radiation. It showed a tendency to decrease air temperature by 1 °C, increase RH by 2-3%, reduce VPD, and improve vine water status with respect to the reference. At harvest, grapes covered with the new sheet had higher TSS concentration, TSS / TA ratio, total skin polyphenol and flavonoids content, and berry skin resistance to rupture. Those berries, moreover, increased the level of volatile compounds such as linalool, typical of ‘Italia’ muscat aroma, and terpineol.


Introduction
The rise in unpredictability of heavy rainfalls enhances incidence of diseases, berry rot, berry cracking, and consequent use of pesticides.Rain shelters can be used to mitigate these effects. Covering table grape vineyards with plastic sheets, in order to protect grapes and to allow a harvest delay up-to 30-90 days, has been done in Italy since the mid-1960s, and was improved as 'late covering technique' [1]. Since the sheets are usually set-up at veraison, when berries become more sensitive to rot, and they are rolled out only above the foliage, these covers were originally considered only as a mechanical protection from rain. However, plastic sheet coverings inevitably filter the incoming solar radiation and alter its radiometric traits, modifying the microenvironment and influencing vine ecophysiological function and grape quality [2,3,4]. Nevertheless, the influence of the type of plastic sheet on microclimate of 'late covered vineyards' and on their grape quality is still little explored. The aim of the present trial was to investigate the features of a new type of rain shelter, and to compare it with a reference cover, by analysing the sheet radiometric properties and their effects on the canopy microenvironment and on the quality attributes of table grapes during a vintage with unusually abundant precipitation for Southern Italy.

Materials and methods
The trial was carried out on a 1-hectare commercial vineyard of Laporta farm, located in Trinitapoli (BT province, Apulia region, Italy, 41° 18' N, 16° 00 E). It was established in 2007 grafting cv. Italia onto 140 Ru rootstock planted 2.4 x 2.4 m apart on clay-loam soil. Vines were trained to overhead tendone trellis (Puglia type), pruned to 4 canes with 10 buds/cane, and protected with white hail net during the growing cycle. The year 2014 was characterized by unusually heavy rainfalls during grape ripening (Fig. 1), with 209 mm between July and mid-October (equal to 38% of the average annual regional rainfall, namely 550 mm). In particular, in the 1 st and in the 3 rd ten-day periods of September, and in the 1 st ten-day period of October, there were 54, 29 and 52 mm of rains, respectively. In that year, some vineyard rows were covered with a new type of plastic sheet consisting of a polyethylene fabric, 200 m thick, provided by Beaulieu Technical Textile (Belgium). This treatment was labelled as NEW. The other rows were covered with a commercial polyethylene film 80 m thick, very popular in the region, as a reference; this treatment was labelled as REF. Both cover types were transparent to solar radiation and equipped with additives including antistatic, anti-drop and UV stabilizers. They were set up on August 10 th (veraison), over the hail net, in alternate groups of 4 adjacent rows. Three replicates of 10 representative vines/replicate were considered. Before the cover set-up, in order to estimate the influence of the hail net on the amount of photosynhetically active radiation (PAR) available for vines, the photosynthetic photon flux (PPF) filtering though the net over canopy was measured at 15 vine positions/treatment (solar bars AccuPar LP80, Decagon), and compared with that available in open air. The number of shoots/vine and of bunches/shoot was assessed. Farm viticultural practices were applied uniformly, either before or after covering, including cluster selection (1.5 bunch/cane), leaf thinning to improve light penetration through the foliage, sub-irrigation and fertigation. Cover radiometric properties were analyzed according to Vox et al. [5]. In the solar wavelength range (200-2500 nm), direct transmissivity was measured (10 nm steps) using a double beam UVVIS-NIR spectrophotometer (Lambda 950 Perkin Elmer) with perpendicular ray incidence, while for total transmissivity an integrating sphere (diam. 60 mm) was used; diffuse transmissivity was calculated as the difference. Transmissivity coefficients, as weighted average values, were calculated [6]. Within the solar radiation, the following wavelength ranges were also considered: photosynthetically active radiation (PAR, 400-700 nm), short infrared radiation (IR s , 700-2500 nm); ultraviolet radiation (UV 220-320 nm), UVB (280-320 nm), and UVA (320-380). In two consecutive typical days of August (19 th and 20 th ) and of September (16 th and 17 th ), at mid-morning (from 11:30 to 12.30 hrs.), microclimatic parameters were measured in open air (2 m above the ground) and under the covers. PPF was taken on horizontal plain; under covers, it was measured at 3 levels: over canopy, at ground, and at the fruiting zone pointing down the solar bar sensors to intercept light reflection from soil to grapes. Air temperature and relative humidity (T air , RH, thermo-hygrometers HD 8501 H, Delta Ohm) were taken over canopy (2.45 m above the soil) and at bunch level (1.80 m above the soil). The air vapour pressure deficit (VPD air ) was calculated. The vine water status was assessed as stem water potential ( stem , 2 leaves/vine). In 2 days, measurements were taken at each vine position. From August 19 th to harvest, at intervals of 8-10 days, berries were sampled from several bunch portions in order to monitor their quality attributes. On 50 berries/replicate, the following parameters were assessed: berry weight, skin colour (CIE L*a*b*, chroma meter Konica Minolta CR400), juice total soluble solids concentration (TSS, refractometer VM-7 Atago), juice titratable acidity (TA, as tartaric acid) by neutralization with NaOH 0.1 N (Titralyzer Laboratories Dujardin-Salleron T11-107). On 10 berries/replicate, peeled skins were extracted in EtOH:H 2 O:HCl (70:30:1, pH1) for 24 hrs. to evaluate: index of total polyphenol content (TP) and of flavonoid content (TF). For TP, 0.1 mL, added with 1 mL of Folin reagent and 4 mL Na 2 CO 3 (10%), was brought up to 20 mL volume with distilled H 2 O, and read at 750 nm using a spectrophotometer (Shimadzou UV-1700). For TF, extracts were 50 times diluted with EtOHCl, and read from 230 to 400 nm. Calculations were done [7]. On the skin extracts, antioxidant activity (AA) was assessed by ABTS· + method [8]. ABTS· + solution was diluted in ethanol (1:88) whose absorbance, at 30 °C and 734 nm, is 0.70 (± 0.02), against ethanol as blank. 200 µL of diluted sample were added to 2 mL of chromogen into a 10 mm pl cuvette; after 15 min, the absorbance at 734 nm was read. Calculation of percentage inhibition: Inhib.(%)=(1-Abs s /Abs b )*100 (s=sample, b=blank). The equivalent concentration (µmol/L Trolox µM) was obtained by replacing the inhib.(%) in Trolox standard curve. Grape aromas were analyzed by headspace solid phase microextraction, by using an 85 μm carboxen /polydimethylsiloxane fibre (Supelco). Grapes were defrosted at 5 °C, pedicels and seeds were removed. 100 g of berries, 2 g of CaCl 2 , 20 g of NaCl and 100 μL of internal standard solution (100 μg/g 2-methylpentanol methanolic solution) were homogenized. For each measurement, 8 g of puree was put into a 15 mL capped SPME vial. After 20 min of sample stirring at 40 °C, the fibre was exposed to the capped vial headspace for 30 min. Then it was inserted into the gas chromatograph (GG) in splitless mode (temperature 250 °C) for 4 min to allow the volatile desorption. Volatile determination was done by a 6890 N GS equipped with a 5975C mass spectrometer (Agilent Technologies). Analytes were separated on a DB-WAX capillary column (60m× 250 μm× 0.25 μm) with the following program: 40 °C for 4 min, 40-180 °C at 3 °C/min (total run 51 min). The transfer line temperature was 280 °C. Mass detector conditions were electronic impact mode at 70 eV, source temperature 230 °C, scanning rate 2.88 scan/s and mass scanning range m/z 30-400. The carrier gas was helium at 1 mL/min. Identification of volatile compounds was achieved comparing their mass spectra with reference spectra contained in a library (NIST 02) of reference data (matching probability P>80) and comparing their retention times and mass spectra with those obtained by standard injections of pure compounds. Compounds were quantified as μg 2-methylpentanol equivalent/g. Data collected from replicates were statistically processed by one-way ANOVA, at 95 percent probability level, in order to assess the effects exerted by the two covers. Data are presented as average of the replicates ± standard error.

Results and discussion
The analysis of radiometric coefficients pointed out that the two covers had quite similar transmittance to photosynthetically active radiation (PAR), while, as for the IRs wavelength range, the NEW cover had lower transmittance to the direct component (-22%), namely the most powerful in heating. Nonetheless, the NEW cover showed a much higher transmittance to Ultraviolet (UV) radiation, that modulates the biosynthesis of secondary metabolites essential for berry quality and healthy properties, such as phenols and flavours [9]. As for the UVB wavelength range, that is the most active, the increase was 300% for the direct component, and 544% for the diffuse component, that has more chance to penetrate throughout canopy gaps.  'Italia' berry weight after veraison increased at the same rate in both treatments until the end of September (10 g), when it was quite constant for one week, and finally slightly decreased (Fig. 2a), showing that, at that point of berry evolution, transpiration overcame the water influx. In fact, the rains of that period, different than those occurring in the 1 st ten-day period of September, had a very low influence on parameters related to berry growth and to metabolite concentration. The more humid microclimate of NEW cover slowed this last process: at harvest, the berry was 3% heavier than the REF one.
However, despite this result, the final NEW bunch weight (642.8±45.6 g) was lower by 4% than that of REF grape (669.9±60.2 g), since the former had fewer berries (68.4±6.2) than the latter (74.0±6.1). The harvest operations include the removal of rotten berries from bunches: the more humid microenvironment of the NEW cover seems to have slightly increased the mass of berries needed to be removed. Juice of grapes under the NEW cover tended to increase TSS concentration and lower TA (Fig. 2b). In this treatment 'Italia' grape reached the minimum maturity requirement suggested by OIV Resolution Viti 1/2008 (TSS / TA ratio  20) on 29 th August (Fig. 2c); under REF cover, this threshold was reached within the next 10 days. The cover did not influence berry skin colour. At harvest, colour coordinates were: L*29.2±0.9, a*-3.2±0.1, b*8.3±0.1 for the NEW berry, L*38.6±1, a*-3.1±0.1, b*7.9±0.6 for the REF berry. However, the treatment influenced the resistance of berry skin, that was significantly higher for the NEW grape (0.32±0.00 kg) than for the REF one (0.26±0.02 kg). This means that the former grape was likely less sensitive to mechanical damages occurring at harvest and during post-harvest. Total skin phenol accumulation was similar or higher when compared to that found in 'Italia' grapes sampled in the same vineyard during a trial run two years prior [11]. In the present trial, grapes under the NEW cover maintained, at harvest, high values of skin TP and TF, that decreased in the grapes under the REF cover (Fig.  3a and 3b). For TF, this difference between treatments was significant when expressed either per kilogram of grapes (NEW = +22%), or per single berry (NEW = +15%). Accumulation of TP, and even more of TF, being processes associated to photoprotection, are diminished by a low availability of UV radiation; moreover, its final amount is negatively affected by temperatures greater than 30 °C [12,13]. The NEW cover likely favored TP and TF accumulation because of its much greater transparency to UV radiation; this effect was more evident after mid-September; that is, when the last leaf thinning increased the light penetration through the canopy gaps to the ground which reflected light upwards. Skin flavonoids, together with phenolic acids, carotenoids, and anthocyanins in red cultivars, are major compounds-although not the only ones-related to grape antioxidant activity, and therefore its health benefits [14]. Antioxidant activity of grape skin extracts is known to change with cultivar, vintage and ripening level [15]. In the present trial, AA index of 'Italia' skin extracts was significantly higher in the NEW treatment at any date except harvest (Fig. 3c). Clearly, other compounds (e.g. other phenols, vitamins, etc.) were involved in this response. At harvest, 12 volatile compounds were detected (Tab. 3). Grapes covered with the NEW sheet were particularly rich in linalool compounds, that are widely known as the most involved in the typical muscat aroma of cv. 'Italia', and in α-terpineol, that confers sweet, floral, rose, and fruity aromas. Linalool, and the expression levels of the representative genes, may be higher in berries receiving more UV radiation [16]. Grapes covered with the REF sheet were particularly rich in hexanal (associated with green aroma), and ethanol.

Conclusions
The radiometric properties of the rain shelter proved to influence both the canopy microclimate and grape quality. Hence, those properties must not be overlooked when choosing the plastic cover for 'late covered vineyards'. A modest reduction of transmissivity to IR short radiation, with a consequent decrease in air temperature, increase of relative humidity and lowering of air VPD, had a positive effect on the final berry weight. On the other hand, a more humid microclimate likely stimulated berry rot. The high transmissivity of the new cover to UV radiation seemed responsible for main differences concerning berry phenol accumulation, antioxidant activity and some aromas. Providing a proper canopy management, UV radiation can reach bunches, also as reflected from the ground.