Growing Processing Tomatoes in the Po Valley Is More Sustainable Under Regulated Deficit Irrigation

Abstract

The Po valley (northern Italy) is the leading European region for processing tomato (Solanum lycopersicum L.) production. Although historically characterized by abundant water availability, this area is now increasingly affected by drought risk. This study presents a two-year evaluation of regulated deficit irrigation (RDI) on processing tomatoes in northern Italy. In 2019 (Parma) and 2022 (Piacenza), full irrigation (IRR, restoring 100% crop evapotranspiration) and RDI (100% IRR until the color-breaking stage, followed by 50% IRR) strategies were compared within a completely randomized block design. Overall, RDI resulted in a 25% reduction in water use without compromising yield, which was maintained through unchanged plant fertility and fruit size compared to IRR. Remote sensing data from PlanetScope imagery confirmed the absence of water stress in RDI-treated plants. Furthermore, increased soluble solids and dry matter contents under RDI suggest a physiological adaptation of processing tomatoes to late-season water deficit. Remarkably, environmental and economic sustainability indicators—namely water productivity and yield quality—were enhanced under RDI management. This study validates a simple, sustainable, and readily applicable irrigation approach for tomato cultivation in the Po valley. Future research should refine this method by investigating plant physiological responses to optimize water use in this key agricultural region.

1. Introduction

Processing tomato (Solanum lycopersicum L.) is a globally important crop, with major production centers located in China (10.4 million tons [t]), California (10.0 million t), and Italy (5.2 million t) [1]. Within the European and Italian context, the Emilia-Romagna region (Po valley, northern Italy) is a leading area for yield performance, with over 1.7 million t harvested in 2024 [2].

Tomato-based products derived from industrial processing—such as paste, sauce, and ketchup—are considered functional foods due to their high content of antioxidants, particularly carotenoids (e.g., lycopene) and polyphenols [3,4,5], which play a protective role against cardiovascular diseases, diabetes, and various forms of cancer [5].

As a high-yielding crop, processing tomato is associated with substantial evapotranspiration, leading to a seasonal water demand ranging between 400 and 600 mm [6] typically met through irrigation. However, leading production areas worldwide are increasingly exposed to the effects of climate change, which are expected to intensify crop water requirements [7]. In recent years, the Po valley has experienced repeated drought events affecting local agriculture [8], and projections indicate that key Italian production districts, such as Emilia-Romagna, are likely to face increasing drought risks [9]. This is particularly concerning Italy, where agriculture accounts for the largest share (exceeding 50%) of national water consumption [10].

The widespread adoption of drip irrigation has significantly enhanced water use efficiency in processing tomato production by delivering water directly to the root zone [11,12]. Nonetheless, in the face of growing water scarcity, the adoption of effective water-saving strategies is critical to ensure the resilience of high-water-demanding crops [8]. While structural measures such as water retention ponds could increase water availability for agriculture, their implementation requires medium-term planning and appropriate financial incentives [13,14]. In the short term, several strategies can help optimize water use, including decision support systems, biodegradable mulching films, and water-saving irrigation methods.

Although biodegradable mulch films can increase crop resilience under stressful conditions, their adoption in more humid climates may hinder rainwater infiltration and increase costs, limiting their widespread use [15,16,17,18]. Water-saving irrigation strategies therefore represent the most viable option in such contexts. Jensen et al. (2010) highlighted that deficit irrigation and partial rootzone drying can reduce water use by 9.2–19.7% compared to full irrigation management, without compromising yield [19]. However, besides increasing the costs for irrigation systems, partial rootzone drying irrigation may impair root–soil interactions in newly formed roots, potentially resulting in yield losses [20]. An alternative approach is represented by regulated deficit irrigation (RDI), which consists of strategically reducing water supply during phenological stages that are considered to be less sensitive to stress [21,22]. Several studies have shown that RDI is effective in maintaining yield while improving water productivity under Mediterranean conditions and is easily adoptable by farmers [23,24,25,26,27,28]. By targeting specific phenological phases, RDI can modulate plant metabolism, enhancing tolerance to stress and promoting the accumulation of beneficial compounds in fruits [29].

Digital tools and decision support systems can improve water use efficiency by delivering the effective crop water demand [30,31]. However, their use remains limited due to perceived technical complexity and high costs, particularly in regions like the Po valley, where irrigation water is relatively affordable [12,32]. In this context, remote sensing technologies have gained attention for their potential to monitor crop physiological status on a large scale, through cheap and non-destructive methods. Satellite-derived vegetation indices provide valuable information on plant vigor, biomass accumulation, and stress responses, especially in water management trials. The normalized difference vegetation index (NDVI) is the most widely used index due to its simplicity, while the modified soil-adjusted vegetation index (MSAVI) and optimized soil-adjusted vegetation index (OSAVI) improve sensitivity under low vegetation cover. The renormalized difference vegetative index (RDVI), soil-adjusted vegetation index (SAVI), and its derivatives SAVI2 and Transformed SAVI (TSAVI) better isolate vegetation signals from soil background. Structure intensive pigment index 2 (SIPI2) integrates additional spectral bands to enhance pigment detection. Collectively, these indices serve as robust indicators for assessing physiological crop parameters across different environmental and management conditions.

While RDI trials in Italy have focused on the southern regions, which are historically prone to drought [23,24,33,34], climate change is expanding its pressures northward, including the Po valley [9,35]. Given that crop responses to RDI depend not only on soil water deficit but also on soil and climatic conditions [24,36], site-specific validation is needed.

The present study aimed to evaluate the feasibility of RDI under northern Italian conditions compared to the farmer’s full irrigation practice (IRR, restoring 100% of crop evapotranspiration [ET_C] throughout the season). The two-year on-field experiment was carried out in two sites located in the Po valley to assess the fruit yield and technological quality of processing tomato cv ‘H1534’. Building upon successful trials previously conducted in Foggia (Capitanata district, Apulia, southern Italy) and Naples (Campania, southern Italy) [23,34], this work seeks to promote the adoption of a simple, sustainable, and easily adoptable irrigation strategy in a key district of processing tomato cultivation.

2. Materials and Methods

2.1. Field Trial Design and Crop Management

A randomized complete block design with three replicates was adopted to compare full irrigation (IRR) and regulated deficit irrigation (RDI) strategies over the two cropping seasons (2019 and 2022). Trials were conducted in different locations each year, and topsoil samples (0–40 cm) were collected one month prior to transplanting by averaging five samples per site. The hydrometer method was used for soil texture analysis, and Richards chambers for hydrological properties. Key physical and chemical characteristics of the fields are reported in

Table 1. Mean and standard deviation of topsoil characteristics (0–40 cm) in Parma and Piacenza over two cropping seasons (2019 and 2022).

	Parma (2019)	Piacenza (2022)
GPS coordinates	44°45′25.7″ N 10°22′11.7″ E	45°04′26.4″ N 9°37′33.7″ E
Elevation (m a.s.l.)	68.7	51.0
Skeleton (%)	0.0 ± 0.0	17.9 ± 1.0
Sand (g kg−1)	180.0 ± 10.6	280.0 ± 16.9
Silt (g kg−1)	510.0 ± 26.7	440.0 ± 19.1
Clay (g kg−1)	310.0 ± 13.4	280.0 ± 14.7
Texture	Silty clay loam	Clay loam
pH	7.6 ± 0.3	7.8 ± 0.2
Total limestone (%)	12.0 ± 0.5	13.0 ± 0.3
Active limestone (%)	5.5 ± 0.2	6.8 ± 0.2
Organic carbon (%)	1.8 ± 0.1	1.1 ± 0.0
Organic matter (%)	3.1 ± 0.1	1.9 ± 0.0
Total N (‰)	1.6 ± 0.0	1.4 ± 0.1
C/N ratio	11.3 ± 0.5	7.9 ± 0.2
P2O5 (mg kg−1)	54.1 ± 2.7	77.3 ± 3.6
K2O (mg kg−1)	299.0 ± 4.2	363.0 ± 12.6
Field capacity (m3 m−3)	0.35	0.37
Wilting point (m3 m−3)	0.17	0.20

. In Parma, soil texture was classified as silty clay loam, while in Piacenza, it was clay loam [37].

Mechanical transplanting of processing tomato cv ‘H1534’ (Furia Seed, Monticelli Terme, PR, Italy) was carried out at the four-leaf stage on 15 May 2019 and 10 May 2022. Crop rotation with wheat was implemented. Seedlings were spaced in twin rows at final densities of 2.78 (2019) and 3.42 (2022) plants m⁻². Fertilization and weed control were conducted according to regional guidelines, with nutrient application based on soil analysis and crop needs: 179 kg N, 105 kg P₂O₅, and 63 kg K₂O ha⁻¹ in 2019; 176 kg N, 71 kg P₂O₅, and 0 kg K₂O ha⁻¹ in 2022.

All plots (n = 12) were harvested simultaneously when approximately 90% of fruits were marketable (~1700 growing degree days), on 3 September 2019 and 23 August 2022. Growing degree days (GDD) were calculated using the non-linear method of Zhou and Wang (2018) [38], linking BBCH (Biologische Bundesanstalt, Bundessortenamt und CHemische Industrie) scale stages to GDD [39]. Base, optimal, and upper temperatures (7.0, 23.5, and 40.0 °C, respectively) were calibrated using Excel Solver (Microsoft Corporation, Redmond, WA, USA) to minimize standard deviation at harvesting time based on previously recorded data on the ‘H1534’ hybrid, imposing lower and upper constraints.

2.2. Irrigation Strategies

Drip irrigation was applied using a single drip line per twin row, with total flow rates of 2.6 mm and 1.3 mm h⁻¹ in 2019 and 2022, respectively. Soil water content (SWC, g kg⁻¹) was monitored with capacitive probes (10HS, Meter Group Inc., Pullman, WA, USA) installed horizontally at depths of 0.20 and 0.40 m at three locations per treatment on 4 June 2019 and 27 May 2022 before the beginning of the irrigation schedule.

Irrigation strategies followed those of Burato et al. (2024) [23]. Irrigation was applied when readily available soil water was depleted. The FAO dual-crop coefficients (Kcini = 0.15; Kcmed = 0.90; Kcend = 0.20) were adopted, and a depletion fraction value of 0.45 was used. Corrections of the crop coefficients Kcini (for precipitation events), Kcmed, and Kcend (for climatic conditions and crop height) were carried out following the methodology outlined by Allen et al. (1998) [40]. IRR fully restored crop evapotranspiration, while RDI mirrored IRR until the BBCH 702 stage for Solanaceous crops (the color-breaking stage fully reached on the second truss) [39], after the which irrigation volume was halved (onset on 20 July 2019 and 11 July 2022). This stage (~1000 GDD) is considered to be less sensitive to water stress in tomatoes [36].

2.3. Plant Growth Monitoring

Crop growth was monitored biweekly. Aboveground dry biomass was categorized into vegetation (stems and leaves) and fruits (t ha⁻¹) components. In each treatment, three plants were sampled, oven-dried at 65 °C, and weighed [23]. Additionally, remote monitoring was conducted via PlanetScope satellite imagery (level 3 B, 3 m spatial resolution) [41]. Due to the unavailability of 8-band Super Dove (PSB.SD) imagery in 2019, 4-band Dove R (PS2.SD) imagery was used for both years. The following vegetation indices were calculated on all available dates throughout the season to monitor plant growth and potential stress:

• Modified soil-adjusted vegetation index [42]:

$$\mathrm{M}\mathrm{S}\mathrm{A}\mathrm{V}\mathrm{I}={\displaystyle \frac{\sqrt{{(2\times \mathrm{N}\mathrm{I}\mathrm{R}+1)}^2-8(\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R})}}{2}}$$

(1)

• Normalized difference vegetation index [43]:

$$\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}={\displaystyle \frac{\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}}{\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}}}$$

(2)

• Optimized soil-adjusted vegetation index [44]:

$$\mathrm{O}\mathrm{S}\mathrm{A}\mathrm{V}\mathrm{I}={\displaystyle \frac{\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}}{\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}+0.16}}$$

(3)

• Renormalized difference vegetative index [45]:

$$\mathrm{R}\mathrm{D}\mathrm{V}\mathrm{I}={\displaystyle \frac{\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}}{\sqrt{\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}}}}$$

(4)

• Soil-adjusted vegetation index [46]:

$$\mathrm{S}\mathrm{A}\mathrm{V}\mathrm{I}=(1+0.5){\displaystyle \frac{\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}}{\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}+0.5}}$$

(5)

• Soil-adjusted vegetation index 2 [47]:

$$\mathrm{S}\mathrm{A}\mathrm{V}\mathrm{I}2={\displaystyle \frac{\mathrm{N}\mathrm{I}\mathrm{R}}{\mathrm{R}+\frac{0.0183}{1.2344}}}$$

(6)

• Structure intensive pigment index 2 [48]:

$$\mathrm{S}\mathrm{I}\mathrm{P}\mathrm{I}2={\displaystyle \frac{\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{B}}{\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}}}$$

(7)

• Transformed soil-adjusted vegetation index [49]:

$$\mathrm{T}\mathrm{S}\mathrm{A}\mathrm{V}\mathrm{I}={\displaystyle \frac{\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}}{\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}+0.16}}$$

(8)

where NIR = near-infrared, R = red, and B = blue reflectance. These indices are sensitive to water status in field-grown tomatoes [50,51,52]. Calculations were performed on each plot, excluding a 2 m external buffer. Two time points were chosen to detect potential variations related to RDI [36]: 2 weeks before (~800 GDD) and 2 weeks after RDI triggering (~1250 GDD).

2.4. Fruit Yield, Defects, and Sustainability Indicators

For each plot, total yield (TY, t ha⁻¹) was determined by manually harvesting 10 representative and contiguous plants from the central paired row and obtained by summing the marketable yield (MY, fully ripe fruits) and unripe fruit yield (green and color-breaking fruits). The percentage of each yield component on TY was calculated (%). Average fruit weight (g) was determined from 100 randomly selected marketable fruits [53]. The number of ripe fruits per plant was derived from the MY and average fruit weight [23].

Fruit defects were also recorded. Rotten fruit counts (overripe and blossom-end rot) were collected from 10 plants per plot, and sun-scalded incidence on MY was assessed on 100 ripe fruits per plot.

Two key indicators were selected to estimate the environmental and economic benefits of irrigation practices. Irrigation water productivity (WP_I, kg m⁻³, Equation (9) [54] and yield quality (YQ, t ha⁻¹, Equation (10)) [55] were calculated as follows:

$${\mathrm{W}\mathrm{P}}_{\mathrm{I}}={\displaystyle \frac{\mathrm{T}\mathrm{Y}}{\mathrm{I}}}$$

(9)

$$\mathrm{Y}\mathrm{Q}=\mathrm{M}\mathrm{Y}\times \mathrm{P}\mathrm{I}$$

(10)

where I = seasonal irrigation volume supplied to the crop, and PI = unitless price index based on soluble solids content (SSC, °Brix) in fruits. The price index (PI) followed 2023 payment guidelines for processing industries in northern Italy (D. Babini, pers. comm.). A base PI = 1.000 was assigned to SSC between 4.75 and 4.85 °Brix, while penalizations and bonuses were applied for SSC below (5 thresholds) and above (8 thresholds) this interval, respectively, with PI ranging from 0.825 to 1.175 (Supplementary Table S1).

2.5. Technological Traits

Thirty disease-free, fully ripe fruits (~2.5 kg) per plot were selected from the MY, washed, dried, and homogenized using a Waring blender (2 L capacity, Model HGB140, PartsTown, Addison, IL, USA). Homogenized samples (50 mL) were rapidly frozen at −20 °C and later split into three aliquots. The first (~5 g) was used to determine pH and titratable acidity (g citric acid per 100 g purée, or g% CA) through a pH-Matic 23^® titroprocessor equipped with a pH electrode model 5011 T (Crison Instruments, Barcelona, Spain). The second (~5 g) was analyzed for SSC using a Refracto 30PX digital refractometer (Mettler-Toledo, Novate Milanese, Milan, Italy). The third (~20 g) was dried at 65 °C until constant weight to assess dry matter content (g dry matter on 100 g purée, or g%) [56]. The soluble solids-to-titratable acidity ratio (SAR) was calculated by dividing SSC by titratable acidity [57].

2.6. Statistical Analysis

Data were analyzed using RStudio (2025.05.0 + 496 “Mariposa Orchid”) for Windows. Linear mixed models (LMMs) and generalized linear mixed models (GLMMs) were implemented via the ‘lme4’ package [58] to evaluate the effect of two distinct treatments. To account for spatial and temporal variability, years and blocks (nested within years) were incorporated as random factors, with treatments considered a fixed factor. Continuous data underwent the LMM procedure after verifying model assumptions (Shapiro–Wilk’s and Levene’s tests and residual plot analysis). Non-normal or heteroscedastic data were normalized using the ‘bestNormalize’ package [59]. GLMMs with gamma distribution were applied to normalized data violating LMM assumptions (i.e., mean fruit weight, number of ripe fruits per plant, color-breaking fruits, pH, titratable acidity, SAR and yield quality). Count (fruit defects) and proportion (percentage yield fractions) data were analyzed using GLMMs with a negative binomial distribution and binomial distributions, respectively. Treatment effects were considered significant at α = 0.05. Results were reported as means ± standard error for continuous and proportion data, or medians [95% confidence interval] for count data.

3. Results

3.1. Weather Conditions

In 2019, approximately 110 mm of rainfall was recorded during the early crop phase (from transplanting to the end of the May), with average temperatures around 15 °C (Figure 1). Thereafter, only two rainfall events exceeded 5 mm, while temperatures increased to an average of 25 °C. Throughout the season, average minimum and maximum temperatures were 18 °C and 30 °C, respectively. In 2022, total seasonal rainfall accounted for 117 mm, with only three events exceeding 10 mm. Average minimum and maximum temperatures were 17.1 and 32.4 °C, respectively, with daily maximum temperatures often exceeding 35 °C from one month post transplantation through early August.

3.2. Irrigation Management and Soil Water Content

In 2019, RDI resulted in a substantial reduction in irrigation volume (−34%, 163 mm less than IRR), while more contained water savings were obtained in 2022 (−15%, 74 mm less;

Table 2. Irrigation management under IRR (full irrigation) and RDI (regulated deficit irrigation) in 2019 and 2022.

Site	Year	N	I in IRR (mm)	I in RDI (mm)	IRR Depth 1 (mm)	RDI Depth 1 (mm)	Rain (mm)	Total IRR Volume (mm)	Total RDI Volume (mm)	Turn 2 (d)
Parma	2019	24	483	319	20	13	166	649	485	5
Piacenza	2022	23	479	405	21	18	117	596	522	5

¹ Depth: seasonal irrigation volume (I) divided by the total number of irrigation events (N); ² Turn: cycle duration divided by N.

). On average, RDI reduced seasonal irrigation volume by 25% compared to IRR over the two years.

In 2019, SWC initially exceeded the field capacity due to the early rainfall (Figure 2). Subsequently, SWC stabilized around field capacity under both treatments until early August. In 2022, higher sand content caused greater SWC fluctuations than in 2019. Nonetheless, up to the onset of RDI, SWC values were comparable between treatments. After RDI was triggered, SWC rapidly decreased in RDI, but converged with IRR values at harvesting due to the occurrence of a rainfall event.

3.3. Plant Growth

Plant growth under both treatments is illustrated in Figure 3. In both years, IRR plots showed greater vegetation biomass than RDI at the third sampling date. Fruit biomass trends diverged: in 2019, the biomass of early fruits was higher under IRR, while in 2022, RDI outperformed IRR. However, at harvest, both treatments were comparable in terms of fruits and vegetation biomass, as was biomass partitioning. Overall, aboveground biomass was lower in 2019 than in 2022.

Remote sensing analysis revealed no significant differences between treatments over the two years (Figure 4; Supplementary Table S2). At key time points (i.e., two weeks before and after the onset of RDI), all vegetation indices (MSAVI, NDVI, OSAVI, RDVI, SAVI, SAVI2, SIPI2, and TSAVI) were comparable across treatments.

3.4. Fruit Yield, Defects, and Water Productivity

Selected agronomic traits are shown in Figure 5, with full statistics in Supplementary Table S3. Over the two years, the mean TY and MY were 100.8 and 96.8 t ha⁻¹, respectively. No significant differences were observed between IRR and RDI in the TY (p = 0.686), MY (p = 0.686), and total unripe (p = 0.888), color-breaking (p = 0.971), and green fruit yield (p = 0.758). Similarly, irrigation treatments did not affect the mean fruit weight (p = 0.658, 65.7 g, on average), or the number of ripe fruits per plant (p = 0.980, 47 ripe fruits per plant, on average). Given the water savings ensured by RDI and the stability in fruit yield, WP_I significantly increased under RDI (p = 0.003), showing a 37% increase over IRR. On average, mean WP_I was 20.8 ± 2.5 kg m⁻³ under IRR and 27.7 ± 1.7 kg m⁻³ under RDI.

The incidence of sun-scalded fruits was not affected by treatments (p = 0.109). However, the number of overripe (p = 0.004), blossom-end rotten (p = 0.001), and total rotten fruits (p = 0.001) were, respectively, increased by 10%, 350%, and 25% under RDI with respect to IRR. Overall, rotten fruits accounted for the majority of fruit defects (IRR: 42 [26, 51]; RDI: 52 [36, 73]), primarily due to overripe fruits [IRR: 40 [23, 51]; RDI: 44 [27, 73]).

No differences were observed for the percentage of green (p = 0.851), color-breaking (p = 0.847), total unripe (p = 0.690), and fully ripe (p = 0.690) fruits on the TY (

Table 3. Relative yield fractions (% on total yield) under IRR (full irrigation) and RDI (regulated deficit irrigation) over two years (2019 and 2022). Different letters indicate significant differences between treatments assessed by GLMM (α = 0.05).

Yield Fraction		IRR	RDI
Unripe	Green	2.5 ± 0.9 a	2.6 ± 0.9 a
	Color-breaking	1.0 ± 0.4 a	1.1 ± 0.6 a
	Total	3.5 ± 1.2 a	3.6 ± 1.2 a
Ripe		96.5 ± 1.2 a	96.4 ± 1.2 a

; Supplementary Table S3). On average, red ripe fruits accounted for 96.4% of the TY.

3.5. Technological Traits

Most technological traits were affected by irrigation management, except for titratable acidity, which averaged 0.39 g% CA over the two years (p = 0.473; Figure 6; Supplementary Table S3). RDI led to a slight but significant decrease in fruit pH by 2% (4.41 ± 0.09, p = 0.010), while SSC, dry matter content, and SAR were boosted by + 10% (5.6 ± 0.1 °Brix, p < 0.001), + 8% (6.6 ± 0.1 g%, p = 0.002), and + 8% (14.2 ± 0.2, p < 0.001), respectively, compared to IRR (5.1 ± 0.1 °Brix, 6.0 ± 0.1 g%, and 13.2 ± 0.3, respectively). Consequently, YQ recorded a significantly improved under RDI (114.0 ± 11.6 t ha⁻¹, p = 0.025) when compared to IRR (102.2 ± 10.7 t ha⁻¹), corresponding to a + 12% increase.

4. Discussion

Our findings demonstrate that regulated deficit irrigation (RDI) effectively optimized water use by reducing the overall seasonal irrigation volume by 25% compared to full irrigation (IRR), despite the observed interannual meteorological variability. This outcome aligns with the global imperative to reduce water consumption in agriculture [10]. Notably, under more arid conditions, the same RDI strategy guaranteed even higher water savings (−31%, on average) [34], highlighting the role of rainfall events—particularly those occurring in the reproductive phase of 2022—in limiting irrigation frequency, and thus water savings, during RDI.

Despite the significant reduction in irrigation (–163 mm in 2019 and –74 mm in 2022), RDI had no detrimental effects on TY or MY. Yield stability under RDI has been widely observed in Mediterranean regions when applying similar strategies post flowering [23,24,27,28,34,60]. In spite of the greatly variable environmental conditions and soil textures, comparable results were obtained in previous research performed under southern Italian conditions adopting the same RDI approach on the ‘H1534’ tomato hybrid [23,34]. Burato et al. (2024) attributed tolerance to prolonged water scarcity occurring in the late crop cycle to the accumulation of drought-responsive metabolites (e.g., proline, polyphenols, and GABA) in ripe tomato fruits [23]. In a related study, Burato et al. (2025) highlighted that increased alanine levels, essential for buffering the pH status, alongside stable proline synthesis, can play a key role in maintaining osmotic balance and providing protection from stress under RDI [34].

In the present work, yield maintenance was ascribable to unchanged plant fertility and fruit size, suggesting the timely onset of RDI during a phenological stage that is less sensitive to water stress [24,28]. Early-stage water restrictions may result in flower abortion or the early drop of small fruits [61,62,63]. Based on these premises, the unchanged aboveground biomass likely supported the reproductive performances under RDI, in full accordance with previous research [23]. However, it should be noted that the present findings are context-specific, limited to the tested soil textures (i.e., silty clay loam and clay loam) and timing of RDI onset (BBCH 702), since variable responses were observed in the Mediterranean basin according to the pedoclimatic conditions and the triggering phase [25,26,33,64].

Vegetation indices derived from PlanetScope imagery supported the physiological stability of RDI-treated plants. The eight indices were carefully selected for their sensitivity to water variations in tomato crop under open-field conditions [50,51,52]. Alordzinu et al. (2021) observed that the NDVI and RDVI effectively detect water stress conditions at various crop stages in tomato grown in sandy loam and silty loam soils [52]. The absence of significant differences in these indices two weeks after RDI onset suggests adequate watering status under RDI, supporting yield maintenance (Figure 5) [23,34].

The enhanced WP_I (+34% under RDI compared to IRR) highlights the effectiveness of this irrigation strategy in maximizing yield per unit of water applied. This indicator is a widely utilized, simplified proxy for assessing the environmental impact of irrigation strategies [54]. Our findings are consistent with several works performed in the Mediterranean area [23,24,27,28,34] supporting RDI as a sustainable strategy for processing tomatoes across multiple climatic conditions.

The application of RDI, however, led to a higher incidence of defective fruits, particularly overripe fruits, likely due to accelerated fruit ripening under reduced water availability [34,65,66]. Blossom-end rot incidence was higher, although relatively low (<0.5 fruits per plant), under RDI, suggesting moderate stress during the fruit-sizing phase [67]. Despite the reduced fruit coverage by foliage that was visually assessed at harvesting, no increase in sun-scalded fruit incidence was observed, possibly because sunlight exposure was not intense or prolonged enough to cause damage [36,68]. Conversely, under southern Italian conditions, RDI did increase sun-scalded fruits with the ‘H1534’ hybrid, likely due to greater canopy reduction and higher irradiance exposure [34].

Technological traits play a pivotal role in determining the safety and the processing efficiency of canned products [69]. A lower pH was observed in RDI-treated fruits, in agreement with results from Valcárcel et al. (2020) [25]. Nevertheless, previous trials carried out under southern Italian conditions did not observe this change, suggesting an environmental modulation of metabolic responses [23,34]. Both dry matter content, which is mostly related to sugars and acids levels [69], and SSC were boosted under RDI by approximately 10%, resulting in potential improvements in the transformation efficiency of raw processing tomatoes by canning industries. The application of a mild water deficit might have promoted carbohydrate allocation towards reproductive organs via stomatal closure and reduced leaf transpiration [36,70,71]. However, contrasting results were reported, particularly under southern Italian climates, where a positive [23] or null [34] effect was recorded on the total and soluble solids contents of the ‘H1534’ hybrid. Under more arid conditions, processing tomatoes may excessively reduce leaf expansion (i.e., lower transpiration rates), limiting sugar accumulation via photosynthesis [34]. Remarkably, average SSC exceeded the minimum level recommended for processing [72] and the base threshold (4.85°Brix) established for northern Italy in 2023, ensuring favorable pricing. In the Po valley, where SSC is a key determinant of product value [17], the increase observed under RDI has remarkable economic implications on the production of processing tomatoes. Indeed, the YQ index, which was particularly developed for this market to incorporate SSC into MY estimation, increased by 12% under RDI, suggesting potentially higher profitability for farmers when adopting this strategy. It is worth emphasizing that RDI can further provide economic benefits by reducing extra costs related to labor and water consumption for irrigation practices [72]. Given the unchanged levels in terms of titratable acidity, the higher SAR observed under RDI may also enhance the sensory quality of RDI-derived canned products [69].

As a water-saving technique, RDI, while generally aimed at water conservation, may contribute to broader environmental goals. Recent studies suggest that reduced irrigation lowers global warming potential via improved nitrogen turnover and carbon dynamics, resulting in decreased greenhouse gas emissions [73,74]. Moreover, limiting groundwater withdrawal under RDI may help protect aquifer quality and long-term sustainability and mitigate soil salinization [75]. Ecosystem services could also be affected by altered water regimes, warranting further investigation into the broader environmental implications of RDI practices [76].

5. Conclusions

The present study aimed to provide farmers with an easily applicable and readily implementable water-saving technique in the Po valley, northern Italy’s leading production district, which is experiencing more frequent drought events. The tested RDI approach, which halved irrigation volumes after the BBCH 702 phenological stage, resulted in considerable water savings (−25%) compared to full irrigation, without requiring complex and dynamic approaches. Remarkably, RDI did not compromise yield performance, as plant fertility and fruit size remained unchanged, suggesting timely RDI onset during a stage that is less sensitive to water variations. This strategy significantly enhanced water productivity (+33%) and boosted key fruit technological traits, including soluble solids (+10%) and dry matter content (+8%), resulting in a substantial increase in yield quality (+12%)—an economic benefit for farmers in regions where SSC is a key determinant of market price. These findings, consistent with previous research, further suggest that processing tomatoes can physiologically adapt to late-stage water deficits, supporting the adoption of RDI not only in arid areas but also in major northern Italian production districts characterized by very different climatic and edaphic conditions.

Future research should focus on optimizing RDI onset timing on processing tomatoes across diverse pedoclimatic contexts in the Po valley, elucidating the specific physiological mechanisms underlying the observed changes in fruit quality and defects incidence, to further refine sustainable irrigation practices in this key district.