Introduction

Tomato (Solanum lycopersicum L.) is an important fresh market vegetable crop grown in the US, with 1 225 870 t harvested from 38 160 ha in 2015 (USDA 2016), and is an important crop for farmers’ markets, farm stands, restaurants, and community supported agriculture in many regions, including northwest Washington. The optimum temperature for tomato growth is 18 to 24 °C, and temperatures below 10 °C negatively affect fruit set (Lovatt et al. 1998). The Skagit Valley is the major vegetable production region in northwestern Washington, and has a 20-yr average temperature of 16 °C with an average daily minimum of 11 °C and an average daily maximum of 22 °C during the summer growing season (June through Sept.; AgWeatherNet 2016). Consequently, tomato production is challenging in this region, and while high tunnels are popular worldwide for tomato production, there has been limited adoption in northwestern Washington (Carey et al. 2009; Lamont 2009). However, high tunnels can provide an environment for successful production of tomato in the area, especially open-ended high tunnels that exclude precipitation, providing protection against late blight (Phytopthora infestans) (Powell et al. 2014). A study in northwest Washington’s Skagit Valley found that Tomato cv. Celebrity yielded fruit 15 d earlier in open-ended high tunnels, harvest was 20 d longer, total and marketable fruit weight was 5 to 8-fold greater, and fruit lycopene content was 32% greater compared with open field production (Miles et al. 2012; Cowan et al. 2014). High tunnels could be used to increase summer production of fresh market tomato in relatively cool conditions of northwest Washington, but more information is needed to optimize production.

High tunnel conditions support intensified crop production requiring increased input applications to maximize yield. Applying large amounts of fertilizers and composts in conjunction with drip irrigation, can elevate soluble salt and nutrient levels in soil that would be leached out with precipitation in the open field (Blomgren and Frisch 2007; Montri and Biernbaum 2009). Elevated soil temperatures in the high tunnel, compared with the open field, may reduce soil moisture and affect soil biological activity and organic matter content (Montri and Biernbaum 2009). Thus, accelerated plant growth and distinct soil conditions, due to elevated temperatures within the high tunnel, creates a need for soil fertility strategies specific to the high tunnel environment. Many fresh market tomato growers in northwest Washington use organic or sustainable production practices, including manure and (or) compost applications to improve soil health and crop yield by enhancing the physical and chemical characteristics of the soil (Stone and Elioff 1998). Maintaining fertility in soil-based high tunnel production systems using organic fertilizers can be a challenge as these fertilizers release nutrients slowly, which can impact crop productivity relative to readily available nutrients in chemical fertilizers (Eghball et al. 2002; Alizadeh et al. 2012; Collins et al. 2013). Use of chemical and organic fertilizers together (integrated fertility management) might provide the benefits of both chemical and organic approaches.

Limited research has been carried out to measure plant growth in response to soil fertility management strategies in high tunnel tomato production. Specifically, little attention has been given to selecting appropriate fertilizer sources and amounts for crop production in protected environments such as high tunnels. In North Logan, Utah, composted poultry manure (CPM) and a polymer-coated slow-release urea were compared for production of tomato cv. Sunbrite grown in a high tunnel system. Both fertilizer sources were applied at three rates (112, 168, and 224 kg total N·ha⁻¹) and compared with a control treatment (no fertilizer applied). Pruning was conducted once per week for a total of five to six times after transplanting, and fresh weight of the pruning biomass was recorded. In two out of 3 yr of the study, total pruning fresh biomass was significantly greater for plants grown with urea than with CPM (Reeve and Drost 2012).

Given the relative lack of published research reports on the effects of fertility management in high tunnel tomato production, studies under field conditions may be informative. Demir et al. (2010) investigated the effect of poultry manure (PM) on plant growth and N leaf concentration of tomato cv. Natura at flowering and at final harvest in a greenhouse study in Turkey. Using three rates of PM (10, 20 and 40 g·kg⁻¹ soil) and a no PM control, the authors reported that plant dry biomass increased significantly with all three levels of PM as compared with the control, but there was no difference among the PM application rates; additionally, leaf N was significantly greater for all PM treatments at final harvest (Demir et al. 2010). A study in Turkey found that tomato fruit yield was greater with CPM than with other organic fertilizers (Ormin, Coplex, and composted cattle manure) or chemical fertilizer (Yanar et al. 2011). A field study carried out in Monticello, AR, compared production of several heirloom tomato cultivars using chemical and organic (poultry litter) fertility sources (Francis and Stark 2012). The authors found no differences in marketable fruit number and weight between the two production systems; however, the authors found that petiole sap nitrate-N ranged from 400 to 600 ppm for most samples for both production systems, which was within the sufficiency range for tomato [400–600 ppm (Hochmuth et al. 1991)] (Francis and Stark 2012). Another study carried out in Nigeria compared the effect of PM at four application rates, a chemical fertilizer, and an integrated chemical-PM fertilizer on yield of a local tomato cultivar (Adekiya and Agbede 2009). Leaf N content increased as PM rate increased, and the combination of PM plus chemical fertilizer resulted in higher leaf N than PM alone. The number of leaves per plant, plant height, and leaf area significantly increased with all study treatments as compared with the control (no fertilizer added), and all growth parameters were greatest for the chemical fertilizer plus PM treatment. The authors reported that the combination of PM plus chemical fertilizer resulted in higher yield than the other treatments (Adekiya and Agbede 2009). It appears clear that both chemical and organic fertilizers can provide tomato plants with required nutrients for optimal growth and yields, but results may vary depending on specific fertilizer programs and production system. Thus, further research is needed to determine the potential efficacy of organic fertilizer sources relative to chemical fertilizers in the context of high tunnels for those growers who prefer organic options.

Grafting has been shown to be a means to avoid biotic and abiotic stress, promote plant growth, and increase crop productivity (Lee 1994; Kubota et al. 2008; Lee et al. 2010). Vegetable grafting is used extensively within protected environment production systems, such as greenhouses and high tunnels. While vegetable grafting has been employed predominantly to manage plant diseases (Peregrine and Bin Ahmad 1982; Lee 1994, 2003, 2007; Paroussi et al. 2007; Sakata et al. 2007; Kubota et al. 2008), it also has been utilized to enhance nutrient uptake and increase nutrient-use efficiency (Ruiz et al. 1997; Santa-Cruz et al. 2002; Martinez-Ballesta et al. 2010). Grafting provides a rapid response to production constraints and is considered an ecological approach for growers to manage biotic and abiotic stress and provide high-quality tomato fruit (Rivard and Louws 2008). Growers and grafting specialists are attempting to identify rootstocks that are suitable for managing regional issues that affect plant growth and productivity (Kubota et al. 2008; Louws et al. 2010). A study in Magnesia, Greece, compared non-grafted, self-grafted, and grafted tomato in greenhouse (grown in soil) and open field conditions and found no significant effect on plant height under the greenhouse environment due to grafting; however, under open field conditions, plants grafted on ‘Heman’ rootstocks were significantly taller than non-grafted plants or plants grafted on ‘Primavera’ at 130 d after transplanting (Khah et al. 2006). Two open field studies of grafted tomato in northwest Washington, found that grafting ‘Cherokee Purple’ onto ‘Beaufort’ or ‘Maxifort’ rootstocks did not provide any advantages for tomato yield and fruit quality parameters such as firmness, total soluble solids (TSS) and lycopene (Buller et al. 2013); and, ‘Stupice’ grafted onto six different rootstocks showed no advantage over non-grafted ‘Stupice’ (Miles et al. 2015). While no differences were found between grafted and non-grafted tomato in open field conditions in these two studies, the dramatic differences in tomato production between high tunnel and open field systems in Washington warrants further investigation to determine the suitability of grafted tomato production in high tunnels generally, and to identify candidate rootstocks for those conditions.

With demand for vegetable grafting increasing each year in the United States, and as tomato is commonly grown in high tunnels, this study may help growers make better-informed decisions about fertilizer source and tomato rootstocks for grafted tomato production in high tunnels in northwest Washington and surrounding regions. Therefore, the goal of the current study was to compare plant growth, nitrate-N uptake, yield, and fruit quality of grafted and non-grafted tomato grown with either chemical fertilizer or an integrated fertilizer treatment (CPM and urea) in high tunnel conditions in northwest Washington.

Materials and Methods

This study was conducted at Washington State University (WSU) Northwestern Washington Research and Extension Center (NWREC) in Mount Vernon, WA in 2015 and 2016. The field site has Skagit silt loam soil that is a fine-silty, mixed, nonacid, mesic Typic Fluvaquent recently formed of alluvium and volcanic ash (National Resource Conservation Service 2010). Soil at the location contains 2.6% organic matter, and a pH value of 6.2 (Ghimire et al. 2018).

Climate measurements

Air and soil temperature (1.5 m height, and 5 and 20 cm depth, respectively), relative humidity (RH), soil water content and leaf wetness were measured (Hobo U30-NRC weather station, Onset Computer, Bourne, MA) at every 1 h with the respective sensors installed in the center of the high tunnel during the two growing seasons. Growing degree days at base temperature of 10 °C (GDD₁₀) (McMaster and Wilhelm 1997) were calculated from the daily minimum (T_min) and maximum (T_max) temperatures utilizing the following equation

(1)

Fruit measurements

Fruits were harvested when at least 75% ripe [corresponding to the “light red” ripeness stage (USDA 1975)], once each week starting at 82 d after transplanting (DAT) in 2015 (21 July 2015), and 98 DAT in 2016 (5 Aug. 2016), with final harvest after the first killing frost 179 DAT in 2015 (26 Oct. 2015) and 178 DAT in 2016 (24 Oct. 2016). At each harvest, weight and number of both marketable and unmarketable (total) fruit were recorded. Unmarketable fruit were classified according to the main reason for unmarketability, then discarded. Sunscald, cracking, bruising, catfacing, and zippering were the main defects causing tomato fruit in this study to be classified unmarketable.

Two to three marketable fruit, with total weight of 250–500 g, were randomly selected from each subplot at 88, 95, 102 and 109 DAT in 2015 (27 July, 3, 10 and 17 Aug. 2015), and 105, 117, 123 and 139 DAT in 2016 (12, 24, 30 Aug. and 15 Sept. 2016), for a total of four sampling times per subplot each year. Fruits were allowed to fully ripen in the laboratory for 3–7 d at room temperature (23 °C). Red-ripe fruit were rinsed in a water bath for 30 s, gently sprayed with cool water, then set on a towel and dried gently so as to not bruise. Fruit were cut in half about 3 mm from center (equator) and a 6 mm slice was cut from the larger segment. The slice was reserved for firmness testing while one of the halves was placed in a drying tray to measure water content and the other half was placed in the blender for other fruit quality measurements.

Firmness

Each tomato slice (6 mm thick) was placed below a penetrometer (L-500, Ametek, Hunter Spring Division, Hatfield, PA), then the tip (8 mm diameter, flat) was centered over the outer pericarp between septa and gradually lowered until it punctured the pericarp of the slice. The value was recorded to the nearest 5 gf with normal rounding, and the slice was rotated and the process repeated until four locations were measured. The average per subplot was calculated for each sample time and units of gram force were converted to Newtons by multiplying the average gram force by 9.807 × 10⁻³.

Water content

Fresh weight (FW) of each fruit half was recorded, fruit were then placed in a drying tray and dried in the oven for 48 h at 80 °C and dry weight (DW) was recorded. Percent water content was calculated as

(2)

Fruit homogenate

For each subplot, the second tomato half and the slice used for the firmness measurement were placed into a blender (Oster 6832; Jarden Consumer Solution, Boca Raton, FL) and homogenized at highest speed for 2 min. Fruit homogenate was used for total soluble solids, pH, titratable acidity, lycopene and β-carotene measurements.

Total soluble solids

One 50 mL centrifuge tube was filled completely with homogenate and centrifuged for 10 min at 3000 rpm (Beckman TJ-6, Beckman Instruments, Inc., Palo Alto, CA). Using a transfer pipette, 3–4 drops of tomato juice from the middle layer as separated by centrifugation were placed onto the clean window of a refractometer (Misco, Cleveland, OH) and the total soluble solids (TSS, measured as °Brix) was measured and recorded.

pH and titratable acidity

Two 50 mL centrifuge tubes were filled completely with homogenate and centrifuged for 10 min at 3000 rpm (Beckman TJ-6, Beckman Instruments, Inc., Palo Alto, CA). For each sample, 30–40 mL of tomato juice extracted from the middle layer was placed into a 50 mL beaker and the pH value was measured and recorded (Beckman ∅ 32 pH meter; Beckman Instruments, Inc., Fullerton, CA). Next, 4 mL of the extracted tomato juice was combined with 50 mL de-ionized water and manually titrated to pH 8.1 ± 0.1 with 0.2 N NaOH. The value of the titratable acidity was calculated by the following equation

(3)

In this equation, N is the normality of the titrant = 0.2; V₁ is the volume of titrant (mL); Eq. wt. = mg/mEq of the predominant acid, in this case citric acid (CA) = 64 mg/mEq; and V₂ is the volume of the sample (mL) = 4 mL in this case (Sadler and Murphy 2003).

Lycopene and β-carotene

One gram of well-blended homogenate was placed in a micro centrifuge tube, then tubes were immediately placed in the freezer (−30 °C) for a minimum of 2 wk. Samples were removed from the freezer, thawed for approximately 3 h, transferred to a 50 mL centrifuge tube, wrapped with aluminum foil to prevent light degradation, and 16 mL of cold 2:3 acetone:hexane reagent was added. The cap was closed tightly, and the tube was placed in the freezer (−30 °C) for 1 h to dissolve all the carotenoid. The tubes were removed from the freezer and placed at room temperature (23 °C) for 5 min to ensure separation into two specific layers; the upper layer was collected for analysis. Lycopene and β-carotene were measured spectrophotometrically (Navarro et al. 2006). The spectrophotometer (Beckman DU-640 spectrophotometer, Beckman Instruments, Inc., Fullerton, CA) was calibrated to zero using 3 mL of cold hexane solution and absorbance was measured at 663, 645, 505, and 453 nm (A₆₆₃, A₆₄₅, A₅₀₅, A₄₅₃, respectively). Lycopene [μg/100 g fruit (fresh weight basis)] was determined by the following formula

(4)

β-carotene (μg/100 g fruit) was determined by the following formula

(5)

Results

Climate conditions

The average air temperature in the high tunnel was 1.2 °C higher than the adjacent open field (Table 1). The average accumulated growing degree days (GDD₁₀) in the high tunnel from May to Oct. was 1394 in 2015 and 1238 in 2016, compared with 1058 and 971 in the open field in these years, respectively. This represents a 30% gain in GDD₁₀ in the high tunnel on average. The average soil temperature in the high tunnel at 5 and 20 cm depth was 18.9 and 18.8 °C, respectively, each year, which was 1.4 °C greater on average than in the open field (17.2 and 16.7 °C, respectively).

Table 1.

Environmental conditions in the open field (OF) and high tunnel (HT) at Washington State University Mount Vernon Northwestern Washington Research and Extension Center during the 2015 and 2016 summer growing seasons.

Plant growth

The total number of leaves per plant was not affected by fertilizer, rootstock, or year, or the interactions among those factors (P > 0.05) except between fertilizer and year at the first sampling time (21 DAT each year) (P = 0.04). When analyzed separately by year and sampling time, the number of leaves was greater when plants were fertilized with chemical fertilizer than with integrated fertilizer both early and mid-season (21 and 42 DAT, respectively) in 2015, and mid-season (45 DAT) in 2016 (Table 2). There was no effect on number of leaves per plant due to grafting at any sampling time in either year.

Table 2.

Mean number of leaves per plant of ‘Panzer’ tomato grown with two fertilizer treatments and four grafting treatments in a high tunnel at Washington State University Mount Vernon Northwestern Washington Research and Extension Center.

Plant height differed due to fertilizer at the first and second sampling times (P = 0.04 and P = 0.02, respectively), due to rootstock at the second, third and fourth sampling times (P = 0.05, P = 0.01 and P < 0.0001, respectively), and due to year at the second, third and fifth sampling times (P = 0.02, P = 0.0008 and P = 0.02, respectively); a significant interaction between rootstock and year was detected at all but the fifth sampling time (P < 0.05). When analysed separately by year and sampling time, plant height was greater for plants grown with conventional fertilizer than with integrated fertilizer at 21 and 42 DAT in 2015, and on all but the last sampling date (107 DAT) in 2016 (Table 3). There was no effect on plant height due to grafting with the three rootstocks in 2015, but in 2016 grafted plants were taller than non-grafted plants on all but the last sampling date, and ‘Estamino’ tended to have greater plant height at all sampling dates.

Table 3.

Mean plant height (cm) of ‘Panzer’ tomato grown with two fertilizer treatments and four grafting treatments in a high tunnel at Washington State University Mount Vernon Northwestern Washington Research and Extension Center.

Cumulative fresh biomass of pruned suckers differed due to fertilizer, rootstock, year, and sampling time (P < 0.0001, P = 0.0003, P = 0.008, and P < 0.0001, respectively). When analyzed separately by sampling time, cumulative fresh biomass of pruned suckers differed due to fertilizer at two sampling times, and due to rootstock and year at most sampling times, and significant interactions were detected among fertilizer and rootstock or fertilizer and year at some sampling dates, and no other interaction was found (Table 4). Cumulative fresh biomass of pruned suckers was greater with chemical fertilizer than with integrated fertilizer at all sampling times in both years (P ≤ 0.01) (Fig. 1). Except at 35 DAT in 2015 (P = 0.21), cumulative fresh biomass of pruned suckers was greater for grafted rootstocks than non-grafted plants (P ≤ 0.01); and, no differences in cumulative fresh biomass of pruned suckers were detected between rootstocks of grafted plants.

Table 4.

Overall P values from analysis of variance of the main factors (fertilizer, rootstock, and year) and their interactions for cumulative fresh biomass of pruned suckers (g·plant⁻¹) of ‘Panzer’ tomato grown with two fertilizer treatments and four grafting treatments in a high tunnel at Washington State University Mount Vernon Northwestern Washington Research and Extension Center in 2015 and 2016.

Fig. 1.

Mean cumulative fresh biomass of pruned suckers (g·plant⁻¹) of ‘Panzer’ tomato grown with two fertilizer treatments (A, B) and four grafting treatments (C, D) in a high tunnel at Washington State University Mount Vernon Northwestern Washington Research and Extension Center at 35, 42, 53, 60, 67, 74, 82, 102 and 116 d after transplanting in 2015 (A, C), and 39, 52, 60, 74, 84, 94, 102, 115 and 137 DAT in 2016 (B, D) during the summer growing seasons. Data values represent non transformed means. Mean ± standard error.

Plant nitrogen

Nitrate-N in plant tissue differed due to fertilizer (P = 0.049) 7 wk after transplanting, due to rootstock at 7 and 9 wk after transplanting (P = 0.001, P = 0.004), and due to year at 7, 9, and 11 wk after transplanting (P < 0.0001, P < 0.0001, and P = 0.03, respectively). There was a significant interaction between fertilizer and rootstock (P = 0.03) at 5 wk after transplanting and between rootstock and year (P = 0.01) at 9 wk after transplanting, but no other interactions were significant. When analysed separately by year and sampling time, there was no effect on nitrate-N in plant tissue due to fertilizer source at any sampling time in either year. There was also no effect on nitrate-N in plant tissue due to grafting with the three rootstocks in 2015; however, in 2016 all grafted plants had higher nitrate-N in plant tissue than non-grafted plants at 7 and 9 wk after transplanting, except for ‘DRO138TX’ at 9 wk (Table 5).

Table 5.

Mean nitrate-N in plant tissue and fresh petiole sap of ‘Panzer’ tomato grown with two fertilizer treatments and four grafting treatments in a high tunnel at Washington State University Mount Vernon Northwestern Washington Research and Extension Center.

Nitrate-N in fresh petiole sap was different due to rootstock at all sampling times (P ≤ 0.003) except 9 wk after transplanting (P = 0.11), and due to year at 5 and 9 wk after transplanting (P = 0.0003 and P < 0.0001, respectively). There was a significant interaction between fertilizer and year at 5 and 7 wk after transplanting (P = 0.03 and P = 0.049, respectively) and between all main factors 9 wk after transplanting (P = 0.03). When analyzed separately by year and sampling time, fertilizer source did not affect nitrate-N content of fresh petiole sap at any sampling time in either year (Table 5). A significant increase in nitrate-N of fresh petiole sap was observed 5 wk after transplanting as a result of grafting with ‘Estamino’ in 2015, and at 7 wk after transplanting in 2016; additionally, nitrate-N of fresh petiole sap was greater for all three rootstocks at 9 and 11 wk after transplanting in 2016.

Fruit yield

Total and marketable fruit number and weight were not affected by fertilizer source (P > 0.05). Total and marketable fruit number differed due to rootstock (P = 0.007 and P = 0.02, respectively). Total and marketable fruit number and weight also differed due to the year (all P < 0.0001). No interactions between the main factors were significant for total and marketable fruit number and weight. When analysed separately by year, total and marketable fruit number and weight were not affected by fertilizer source either year (Table 6). In 2015, there was no difference in total and marketable fruit weight due to grafting; however, the number of total and marketable fruit was greatest for non-grafted ‘Panzer’, and lower for grafted ‘Maxifort’ (Fig. 2). In 2016, the number of total and marketable tomato fruit did not differ due to grafting, but the weight of total and marketable fruit was greater for grafted plants as compared with non-grafted plants. The first fruit harvest was at 82 and 98 DAT in 2015 and 2016, respectively, and the last harvest was 179 and 178 DAT in 2015 and 2016, respectively, and harvest did not occur earlier or later due to fertilizer source or the use of grafted plants either year. Although, the non-grafted plants had a higher number of fruits at the first and last harvests in 2015, and at the first harvest in 2016. The most prevalent cause of unmarketability in both growing seasons was sunscald, accounting for 43% and 58% of unmarketable fruit in 2015 and 2016, respectively. There was a greater incidence of bruising and cracking in 2015 compared with 2016 (data not shown).

Table 6.

Mean total and marketable fruit weight (t·ha⁻¹) and number of ‘Panzer’ tomato grown with two fertilizer treatments in a high tunnel at Washington State University Mount Vernon Northwestern Washington Research and Extension Center during 2015 and 2016 summer growing seasons.

Fig. 2.

Mean total and marketable fruit number (A) and weight (B) of ‘Panzer’ tomato grown with four grafting treatments in a high tunnel at Washington State University Mount Vernon Northwestern Washington Research and Extension Center during 2015 and 2016 summer growing seasons.

Marketable fruit quality

Lycopene content differed due to fertilizer source (P = 0.0005), firmness, juice content and lycopene content differed due to rootstock (P = 0.05, P = 0.006 and P = 0.01, respectively), and firmness, TSS, pH and titratable acidity differed due to year (P < 0.0001, P < 0.0001, P = 0.01 and P < 0.0001, respectively), and there was no other effect on fruit quality parameters due to the main factors detected in this study, significant interactions between fertilizer and year and between rootstock and year were detected for TSS content (P = 0.0009 and P < 0.0001, respectively), but no other significant interactions were found. When analysed separately by year, fruit quality parameters were not affected by fertilizer source in 2015, but TSS and lycopene content were greater for integrated fertilizer than chemical fertilizer in 2016. Grafting with each of three rootstocks enhanced fruit water content in 2015 and TSS in 2016, though no differences in water content and TSS were detected among the grafted rootstocks (Table 7).

Table 7.

Mean fruit quality parameters of ‘Panzer’ tomato fruit grown with two fertilizer treatments and four grafting treatments in a high tunnel at Washington State University Mount Vernon Northwestern Washington Research and Extension Center during 2015 and 2016 summer growing seasons.

Discussion

Application of a chemical fertilizer increased growth of tomato plants (number of leaves per plant, plant height, and cumulative pruning fresh biomass) compared with plants grown with an integrated fertilizer treatment, especially early in the growing season. The increased growth of tomato plants in the current study with chemical fertilizer relative to plant growth under the integrated fertilizer regimen was in contrast to the findings of Adekiya and Agbede (2009), where all tomato growth parameters including number of leaves per plant and plant height were greatest for chemical fertilizer with PM. Results of the current study are similar to Reeve and Drost (2012) who reported in two out of 3 yr total pruning fresh biomass was significantly higher for plants grown with chemical fertilizer than with CPM. In the current study, plant growth measures differed between fertilizer treatments early in each season, and were similar later in the season, except for cumulative pruning fresh biomass which was greater with chemical fertilizer throughout the growing season. The greater early plant growth using chemical fertilizer relative to the integrated treatment in the current study is likely due to the more readily available nutrients from the chemical fertilizer compared with the integrated fertilizer treatment (Isah et al. 2014). Adding a source of readily available nutrients early in the season may ensure sufficient plant nutrition until organic sources become available to plants later in the season.

Results of the current study also indicate that grafted tomato plants tended to have greater plant growth than non-grafted plants throughout the growing season. For example, ‘Panzer’ grafted on ‘Estamino’ tended to have the greatest plant height throughout the growing season. These results differ from Khah et al. (2006) who found no significant effect due to grafting on plant height in a greenhouse environment (average daily temperature was 27.7 °C). However, those authors also found that under open field conditions (average daily temperatures was 23.5 °C), plants grafted with ‘Heman’ were significantly taller than non-grafted plants or plants grafted with ‘Primavera’ at the end of the growing season. In the current study, the average daily temperature was 16.6 °C in the high tunnel, which is less than optimal for tomato plant growth (optimal is 18–24 °C). Growth of grafted plants in the current study tended to be significantly greater than non-grafted plants. This result suggests that grafting onto select rootstocks can improve plant growth in sub-optimal conditions, perhaps by enhancing absorption and translocation of nutrients (Han et al. 2009; Nawaz et al. 2016; Ceylan et al. 2018). Further research is needed to better understand the relationship between environmental conditions and growth of grafted plants. Regardless, it appears that. rootstock selection may be a valuable consideration for developing an annual management plan and addressing long-term soil fertility for tomato production in regions with low temperatures during the growing season, as in northwest Washington, and also warrants further investigation.

Fertilizer source had no significant effect on nitrate-N in plant tissue or fresh petiole sap in this study. This finding differs from Adekiya and Agbede (2009) who found PM plus mineral fertilizer resulted in higher leaf N than PM alone, suggesting that the addition of chemical fertilizer to PM provided enhanced plant nutrition. In the current study, while there was no significant difference in nitrate-N between fertilizer treatments, there was a trend toward higher nitrate-N in plants grown with integrated fertilizer than those grown with chemical fertilizer. This trend may be somewhat suppressed in the present report due to the slightly lower (about 4.5%–9%) amount of total N able to be applied in the integrated treatments relative to the chemical fertilizer treatment. A lower nitrate-N in plant tissues may be caused by an overall greater plant growth with chemical fertilizer. As plant biomass increases, N concentration decreases due to an increase in the relative proportion of structural materials and storage tissues, which contain low levels of N (Greenwood et al. 1990). Thus, integrated fertilizer may be an adequate fertilizer regime, relative to nitrate-N accumulation in vegetative tissues, for tomato plants in high tunnel production in northwestern Washington; and under such a regime, additional readily available nutrients early in the growing season may be necessary for crop productivity. Grafted plants had higher nitrate-N concentrations than non-grafted plants, suggesting that the more vigorous root system of the rootstock may have been capable of absorbing and (or) translocating plant nutrients more efficiently than the scion root system (Nawaz et al. 2016). Nitrate-N in fresh petiole sap differed slightly due to the fertilizer treatment, confirming that sap nitrate-N content is a sensitive indicator of plant nutrient status, and therefore a useful and efficient tool to evaluate tomato plant N status. Plant nutrients in chemical fertilizers are more readily available than nutrients in organic sources such as compost (Eghball et al. 2002). Rapid plant growth due to grafting may require more available nutrients early in the season, potentially increasing the compatibility of chemical fertilizers with grafted plants. Organic fertilizers could be supplemented with chemical fertilizers in an integrated approach to meet high nutrient needs early in the season. Organic fertilizers could also be applied to soil in the previous fall to allow time for organic nutrients to mineralize, becoming available to plants early in the following season (Collins et al. 2013). However, in the current study, both chemical and integrated fertilizer treatments produced similar concentrations of nitrate-N in plant tissue of grafted plants. More work in this area could help optimize the use of organic fertilizers with grafted tomato plants in high tunnel systems.

Number and weight of total and marketable fruit did not differ due to fertilizer source either year. These results differ from Adekiya and Agbede (2009) who found that the combination of PM plus chemical fertilizer resulted in higher yield of a local tomato cultivar than PM or chemical fertilizer alone. Likewise, the current results differ from Yanar et al. (2011) who reported that fruit yield of tomato ‘Alida’ was greater with CPM than with several organic fertilizers (Ormin, Coplex and composted cattle manure) or an inorganic fertilizer (459N–153P–498K–200S–17Ca–50Mg kg·ha⁻¹). Although high tunnels enhance environmental conditions, the relatively low temperatures in northwestern Washington provided a sub-optimal environment for tomato growth and fruit development, subsequently affecting overall N-use efficiency from either fertilizer source, as compared with the previously mentioned studies where temperatures were likely not a limiting factor. Additionally, the lack of any significant impact of CPM in the present study could be attributed to management practices we did not evaluate, including timing and method of fertilizer application, plant architecture (i.e., pruning and trellising), and variety selection (Ilupeju et al. 2015).

Total and marketable fruit weight was greater for grafted plants than for non-grafted plants, but only in the second growing season. The average air temperature in the high tunnel in 2016 was lower than in 2015 (16.1 and 17 °C, respectively), suggesting that grafting may have increased nutrient-use efficiency or other production constraints in the lower temperature conditions of the second growing season. These results seem to contrast with the findings of Buller et al. (2013) where grafting ‘Cherokee Purple’ onto ‘Beaufort’ or ‘Maxifort’ rootstocks did not provide any advantages for tomato yield. Likewise, Miles et al. (2015) reported that grafted heirloom tomato ‘Stupice’ showed no advantage over non-grafted ‘Stupice’. Both of these studies were carried out in the open field in northwest Washington, where lower temperatures may have been too low for the more vigorous rootstocks to have adequately compensated. In the high tunnel environment of the current study, grafted plants tended to exhibit increased total and marketable fruit weight, though only statistically significant in the colder year. Previous studies found that rootstock genotypes present broad differences in low temperature tolerance (Kacjan-Marsic and Osvald 2004; Pogonyi et al. 2005). Given the relatively small number of rootstocks evaluated in this study and the prior open field studies in Washington, it is reasonable to expect differences between performance of grafted tomato in high tunnels compared with open field production, especially in a region where summer temperatures do not reach the optimal range for tomato growth and development.

There was no significant effect due to fertilizer source on fruit quality parameters measured in this study, except TSS and lycopene content were higher for fruit grown with integrated fertilizer than with chemical fertilizer in 2016. Application of CPM could enhance soil physical properties (Van Noordwijk et al. 1993; Celik et al. 2004), as well as improve root extension, which can promote higher water and nutrient uptake by plants and subsequently enhance starch production via photosynthesis (Chen and Avnimelech 1986). Starch is accumulated in green tomato, and levels decline with the onset of fruit ripening with an accompanying rise in soluble solids (Davies and Cocking 1965; Eskin 2000). The positive effect of applying PM is reflected in the increased levels of TSS and lycopene. Likewise, increased TSS in fruit from grafted plants in 2016 relative to non-grafted plants may have been due to improved nutrient absorption in the rootstocks. Fruit from grafted tomato plants had higher water content than fruit from non-grafted plants, though only statistically significant in 2015 despite a greater relative difference in 2016. An inverse relationship exists between water content and TSS in tomato fruit (Mitchell et al. 1991). It is likely that the relationship between fruit water content and TSS could be affected by grafting especially where temperatures might impact transpiration and translocation of photosynthates to the fruit; and, we recommend further inquiry into this relationship.

Conclusion

While the integrated fertilizer treatment produced less growth than chemical fertilizer, especially early in the growing season, both treatments produced comparable nitrate-N concentrations in tomato plant tissues, comparable total and marketable yields; and, the integrated fertilizer improved TSS and lycopene content of tomato fruit. Given that organic fertilizers can often take more time for nutrients to become available, an integrated fertility regimen, incorporating both chemical and organic fertilizer sources, early in the season may provide adequate plant nutrition, increasing soil organic matter without negatively impacting yield or fruit quality. Grafting enhanced plant growth, nitrate-N concentrations, and tomato fruit yield in high tunnel production in northwestern Washington.