The state of Florida ranked first in the US in tomato (Solanum lycopersicum L.) production area planted (13,350 ha) and harvested (13,030 acres) for fresh market in 2015 (USDA, 2016). Florida. Total fresh tomato production area for the US was 38,500 ha planted and 37,300 ha harvested. Tomatoes are grown on sandy soils in Florida. As a result of low organic matter and large porosity, the sandy soils in Florida have low cation exchange and water holding capacities, requiring growers to apply large quantities of fertilizer to supply all the nutrients that the crop needs for optimum yields [1]. Tomato has a high market value and growers usually apply more fertilizer than the crop requires avoiding the risk of yield reductions [2]. The University of Florida, Institute for Food and Agricultural Sciences (UF/IFAS) fertilizer recommendations for a crop under irrigation on a mineral soil in Florida testing low in phosphorus (P) and potassium (K) are: 224 nitrogen (N), 168 P₂O₅ and 252 K₂O kg/ha/crop season [3]. It is important to apply nutrients and water to the crop in the right amount and at the right time because high fertilizer application, low nutrient extraction and high quantities of water through irrigation and rainfall, can consequently result in N leaching [2,4,5].

With the low water holding capacity of Florida sandy soils, another part of tomato production vital for high yields is irrigation. The most common irrigation system used in south Florida for tomato production is seepage [6]. Irrigation application efficiency can be affected by many aspects such as soil characteristics, climate conditions, method of application, the kind and stage of crop among others [6]. Seepage irrigation has low water application efficiencies (25 to 60%) compared with 80% to 95% for drip irrigation [6,7].

Drip irrigation is known for reduced water use, improved nutrient management by fertigation, enhanced water distribution, improved nutrient uptake and simplified scheduling the irrigation of extensive areas of production [8-11].

Roots play important roles including absorption of nutrients and moisture [12]. If the development of the root system is affected by improper water and nutrient management, the nutrient, water and salt concentration and distribution will be affected in the plant [12]. Thus, growers must be concerned about nutrient applications, because the lack or excess of nutrients can make large difference in yields. Consequently, it is important to evaluate the efficacy of fertigation rates for tomato production under drip irrigation system to maximize nutrient and water use efficiency, and obtain improved yields in Florida. Fertigaiton rates can vary during the growing season because of unforeseen errors in scheduling such as fertigation missed during heavy rainfall, failures in pumping equipment, delays in fertilizer delivery, and miscommunications among field staff (personal communications with growers). Considering all these points, the current study had following two main objectives: 1) The effect of different fertigation rates on tomato harvest index; and 2) Distribution pattern of fertigation in different plant parts and the soil layers.

Experimental location and experimental design

The experiments were conducted at University of Florida, Southwest Florida Research and Education Center (SWFREC) in Immokalee, Florida (Latitude 26.42° N and Longitude 81.43° W, at 10.41 m above sea level). The soil was Immokalee fine sand classified as sandy, siliceous, hyperthermic, Arenic Haplaquods [13]. The three experiments were conducted with tomato grown on raised plastic mulch covered beds in a randomized complete block design, with 4 replications of each fertigation treatment applied twice per week. Three experiments were conducted in fall 2013, spring 2014, and fall 2014. Three beds, 20 cm high and 914 cm long, were constructed per plot with a spacing of 183 cm between bed centers in the fall 2013. The bed for the spring and fall 2014 seasons were 20 cm high and 1524 cm long with 183 cm between bed centers. Two drip irrigation lines (30.5 cm emitters spacing, Jain Irrigation Inc.) were installed in each bed (one drip line for irrigation and the other for fertigation). The beds were fumigated with methyl bromide at 180 kg ha^-1 in fall 2013. Approximately 67 kg ha^-1 of Pic-Chlor60 was applied to fumigate the beds in the spring and fall 2014 seasons. There was no herbicide applied in the beds in any of the fall seasons, however in the spring 2014 season 0.77 kg ha^-1 of Dual Magnum was applied pre-plant as an herbicide. The tomato plants were transplanted in a single row next to the fertigation and irrigation lines. The varieties used were RFT 6153 (Syngenta, USA) in fall 2013 and Tygress (Monsanto, USA) for the spring and fall 2014 seasons.

Fertilizer rates

All the plots were fertigated twice a week. The quantity of fertilizer applied varied with plant growth phase based on week after transplanting (WAT) as specified in Table 1. Total N and K fertilizer rates were 25%, 50%, 100%, 150% and 200% of UF/IFAS recommendations. In all seasons, 25% of the UF/IFAS recommended N rate, 12.5% of recommended K rate, and 50% of the recommended P rate were incorporated in the beds with 56 kg ha^-1 N, 14 kg ha^-1 P₂O₅ and 28 kg ha^-1 K₂O as soluble fertilizer. All beds were irrigated with a surface drip irrigation system under plastic mulch cover. As a result of several rain events after the beds were formed (recorded rainfall was 9.3 cm for 2 days after bedding) but prior to planting in the fall season 2013, the field was flooded and a large amount of nutrient was assumed to be leached from the beds. As a result of the flooding, N, P and K rates equal to the target total season fertilizer amounts were applied to each treatment by fertiation (Table 1).

Table 1: Mean dry weight of leaves, stems and fruit, 12 weeks after transplanting (WAT) in fall 2013, 11 WAT spring 2014 and 13 WAT fall 2014.

Soil sampling

Random soil samples were collected at 12 locations across the experimental field prior to for the fall 2013, spring 2014, and fall 2014 seasons to determine the initial nutrient status of the field prior to bed formation. Soil samples were taken with a 2.5 cm auger at three depths (0-15, 15-30, and 30-45 cm) below the soil surface. After bed formation, two soil samples were taken from each treatment plot with a 2.5 cm auger at the same three depths (0-15, 15-30, and 30-45 cm) below the bed surface. One sample was collected 7 cm perpendicular to the fertigation line (position 1) and in the planted row 15.2 cm between plants (position 2). Soil samples were taken prior to initial fertigation and once a month to harvest. In order to prevent transformation of N from NH₄ to NO₃, the samples were taken and stored in a cooler containing ice. The samples were kept in a freezer at <4°C until analysis.

Soil nitrogen, phosphorus and potassium analysis

Ammonium-N (NH₄⁺-N) and nitrate (NO₃^-_N) determination was conducted using 2 M KCl extraction method [14] from each soil sample (wet), 5.0 g were weighed and placed into a centrifuge tube. Forty ml of 2 M KCl solution was added to the centrifuge tube, capped and then shaken for 30 min at low speed. The solution was allowed to settle for 20 to 30 min, and filtered (Whatman paper number 1, 90 mm). The filtrate was poured in plastic vials and stored in a freezer at <4°C until analysis. After the solutions were thawed, NH₄⁺-N and NO₃^--N analysis was performed using a Flow Analyzer (QuichChem 8500, Lachat Co.) at 660 nm and 520 nm, respectively. The results were expressed on an oven-dry soil basis by determining the soil moisture content by gravimetric method at time of analysis.

The K and P analysis was conducted using Mehlich-3 (M-3) soil extraction [15]. Approximately 2.5 g of soil oven-dried for 24 h at 105°C was weighed and placed into centrifuge tubes and 25 ml of M-3 solution was added. The samples were capped and shaken for 5 min at a high speed. The sample solution was allowed to settle for 20 to 30 min., then, the extracts were filtered (Whatman paper number 1, 90 mm). The filtrate was stored in plastic sample vials in a refrigerator. Analysis for M-3 P and K was conducted by Inductively Coupled Plasma (ICP)-Optimal Emission Spectrometer Optima 7000DV (Perkin Elmer Co.) at 213.6 nm and 766.5 nm for P and K, respectively.

Biomass sampling and analysis

Crop biomass were collected at 5, 9 and 12 WAT in fall 2013; at 5, 9; 11 WAT in Spring 2014 and at 4, 8 and 13 WAT in fall 2014. One representative plant per plot was sampled for biomass determination. The samples were separated into leaf, stem and fruit sub-samples, weighed fresh and then placed in an oven at 65°C for three days (leaf samples) and one week (stem and fruit samples). The dry weight was determined and each sample was run through a stainless steel 20 mesh sieve and later through a 60 mesh sieve [16]. Total N in the leaves, fruit and stems were analyzed according to methods described by Hanlon et al. [16] using the C/N Analyzer NA2500 (Thermoquest CE Instruments). The determination of total tissue P and K contents was conducted by dry ash digestion method followed by Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) analysis [16,17].

Tomato harvest and yield

Fruits were harvested three times in each season; 11, 12 and 13 WAT for fall 2013, 10, 11 and 12 WAT spring 2014 and 11, 13 and 16 WAT for fall 2014 seasons. The fruit harvested was divided into marketable color (i.e. green and red) and size (i.e., extra-large, large, medium, small and un-marketable culls) according to USDA standards (1997).

Statistical analysis

The PROC GLM procedure from the SAS program [18] was used for determining treatment effects on nutrient concentration, biomass, fruit yields and soil nutrient concentration. In the cases when the F-value was significant, a second analysis was done using Duncan’s multiple range tests at a P-value of 0.05.

Biomass accumulation and nutrient concentration

Tomato is well known for its response to fertilizer application [4,19]. The common methods used for the application of the fertilizers (such as broadcasting, and banding) are not efficient, unlike fertigation where the nutrients are applied directly to the root zone and fertilizer use efficiency can be enhanced [19]. Significant differences were found among treatments for leaf, stem and fruit dry weight in the fall 2013 season indicating significantly greater biomass for treatment 3 (280 kg ha^-1 N, 253 kg ha^-1 K), but no significant increase in biomass at fertilizer rates greater than that rate (Table 2). In the spring 2014 season, the plants were affected by bacterial leaf spot, a severe wide-spread disease that affects tomato [20]. Yields during the spring 2014 season were limited by the disease and was lower in quality than expected for a commercial acceptable crop. Nonetheless, leaf and stem tissue biomass accumulation in the spring 2014 season (fruit data not included) increased with fertilizer rate but did not increase above the rate equal to treatment 3 (224 kg ha^-1 N, 196 kg ha^-1 K). In the fall 2014 season, there were no significant differences among treatments for leaves, stems and fruit dry biomass (kg ha^-1) (Table 2).

Table 2: Mean percent of nutrient (N, P and K) in the leaves, stems and fruit at 12 weeks after transplanting (WAT) in fall 2013, 11 WAT in spring 2014 and 13 WAT in fall 2014.

Concentration (%) of N in the leaves and stems increased significantly with fertilizer rates in the fall 2013 and spring 2014 seasons, and for fruit in fall 2013 and stems in fall 2014 seasons (Table 3). However, similar to biomass, no significant increases in N concentrations were found above treatment 4 (393 kg ha^-1 N in 2013 and 337 kg ha^-1 N in 2014). Phosphorus concentrations increased in the leaves in the spring and fall 2014 seasons at the lowest P fertilizer rate. Potassium concentrations in the leaves occurred for all seasons with increased K concentrations in stems, and fruit for both seasons of 2014. Plant tissue accumulation, as expected, was a function of biomass weight and tissue concentration. Thus N, P, and K content increased significantly among treatments for all tissues with the exception of leaf N, leaf and stem P in fall 2014 (Table 4).

Table 3: Mean accumulation of nutrient (N, P and K) in the leaves, stems and fruit, 12 weeks after transplanting (WAT) in fall 2013, 11 WAT in spring 2014 and 13 WAT in fall 2014.

Table 4: Total marketable yield at first harvest by fruit categories and total marketable yield in fall 2013, spring 2014 and fall 2014.

Nitrogen and K accumulation was greater in fruit than leaves and stems in fall 2013 while in fall 2014 N and K accumulation in fruit and leaves were similar to accumulation in stems. Hartz and Bottoms [21] made similar observations in N, though with lower N rate of 200 kg ha^-1. According to Hartz and Bottoms [21], the rate of tomato growth and N uptake peak during fruit setting and slow after fruit ripening begins, thus, N applications should coincide with peak requirements to prevent excessive application and leaching.

Tissue N accumulation in leaves, stems and fruits in the ranged of 27-38%, 10-16%, and 46-62% in fall 2013 and 40-48%, 12-18% and 32-46% in fall 2014, respectively. Phosphorus accumulation in leaves, stems and fruits in the ranges of 23-31%, 15-27%, and 44-56% in fall 2013 and 36-54%, 16-25% and 20-45% in fall 2014, respectively. Potassium was distributed in leaves, stems and fruit in the range of 13-31%, 9-21% and 47-75% in the two fall seasons, respectively. Thus, most potassium was allocated in the order fruit>leaves>stems across the two seasons.

Plant growth in fall 2014 was reduced by bacterial spot; therefore, increased fertilizer did not significantly increase yields. This disease can cause tremendous negative effects to the tomato plants in some cases may result in a complete loss of the crop in areas characterized for warm and humid weather these climatic conditions favor the survival of the disease [20]. The spring 2014 season of the tomato was characterize by several rain events which may have favored disease development which resulted in extremely damaged plants that did not grow well, some plants died. As a result of the bacterial spot, the production of tomato per plant was extremely low. Foliar applications of bactericide were made to control the disease but unfortunately the plants were extremely affected. The disease resulted in reduced yields for all the treatments compared with the fall 2013 and 2014 seasons, however differences in yields among treatments were observed with total harvest for treatments 3, 4 and 5 greater than treatments 1 and 2 (Table 5).

Table 5: Analysis of variance (ANOVA) for NH₄⁺-N, NO₃^-–N, P and K concentration in the soil in fall 2013, spring 2014 and fall 2014.

Marketable yield and number of boxes per hectare

It is important to apply the right rate of nutrients because the excessive application of fertilizer can lead to luxurious consumption [10]. Thus, the results of this study show that the application of more fertilizer than 224 kg ha^-1 N for tomato grown on sandy soils under drip irrigation did not result in a significant increase in yields. However, Motis et al. [22] found no significant differences among the treatments for marketable yield when they evaluated five levels of drip- or band-applied N at 0, 25, 50, 75, and 100% applied at 224 kg ha^-1, 196 kg ha^-1 K. Mixed-results in two years were obtained by researchers in Florida under seepage-irrigation where highest marketable yields were obtained with 172 and 298 kg ha^-1 of N in 2007 and 2008, respectively [23].

In the fall 2013 and fall 2014 seasons, there were no significant differences in total fruit (all marketable size categories combined) weight among fertilization rates at the first harvest (Figure 1 and Table 5). Total marketable yield and number of boxes per hectare for treatments 3, 4 and were higher than the average yield and number of boxes per hectare reported for fresh market in 2012. Whereas, the total marketable yields and number of boxes for treatments 1 and were lower than the state average.

Figure 1: Tomato fruit yield in fall 2013 at first, second and third harvests in Immokalee, Florida.

There were significant differences among fertilization rates at first harvest (all size marketable categories combined) for only the spring 2014 season (Table 5). Total harvest yield for treatments 3, 4 and 5 were significantly greater than two lower fertilization rates. Thus, the fertilization rates that received more fertilizer than 393 kg ha^-1 N and 365 kg ha^-1 K in 2013 and 337 kg ha^-1 N and 309 kg ha^-1 K in 2014 were characterized as having excessive leaf and stem biomass compared with the treatment 3 rates, but had statistically the same yield in fruit per plant or area. Yield data collected during the fall 2013 and spring 2014 season indicated the 280 kg ha^-1 N (2013) and 224 kg ha^-1 N (2014) nitrogen rate produced significantly greater total yields than the lower nitrogen rates but not significantly different than the two higher nitrogen rates. Zotarelli et al. [11] and Scholberg et al. [2] found similar results indicating that applying more fertilizer than the recommended rate does not typically result in an increase in tomato yields. Santos [24] reported similar total marketable fruit yields among N rates ranging between 224 and 336 kg ha^-1 irrigated with the seepage plus drip combination. However, there was a significant N effect in plots receiving only seepage irrigation with marketable fruit weight almost doubling from 12.0 to 22.7 tons/acre when applying 224 and 336 kg ha^-1 N, respectively.

Santos [25] demonstrated that K rates in the order of 336-560 kg ha^-1 did not result in yield differences using seepage irrigation. This indicates that the irrigation system might have a bearing on potential nutrient uptake. Importantly, Zotarelli et al. [5] found that applying N in excess of the UF/IFAS recommendation on Florida sandy soil increased N leaching by 64, 59, and 32%, respectively, for pepper, tomato, and zucchini crops using drip irrigation. Likewise, other researchers in Florida found that an application greater than 168 kg ha^-1 of N in drip irrigated tomatoes did not result in greater yield [11,26], agreeing with the yield results for this study for N rates >224 kg ha^-1.

Soil nutrient distribution in the soil

Soil samples were collected at the experimental site four times during each season (once each month) in order to determine the nutrients status in the soil. Interaction of sampling dates by fertilization rate and soil depth was significant for NH₄⁺-N for the fall 2013 season (Table 6). Soil NH+ concentration increased with fertilizer N rate with significantly greater concentrations for treatments 4 and 5 and at the 15-30 cm depth (Table 7). Significant differences for sampling dates were found for P (Table 6), where the lowest concentration of P was observed at 9 weeks after transplanting (WAT) and highest at 4 WAT, but 0, 4 and 13 WAT were similar in P concentrations (Tables 8-10). Season-long mean NO₃^-N and P had significant differences among depths (Table 6). The soil concentration of NO₃^--N increased with fertilizer rates for all sampledates whereas soil K concentrations was greatest at 4 and 9 WAT for the treatment 2 rate (Table 8). Soil NH₄⁺ was greater in the top 15 cm than in the 15-30 and 30-45 cm soil depth for all threeseasons, while P concentration were greatest in the 15-30 cm depth and the lowest at 30-45 cm soil depth for the fall 2013 and fall 2014 seasons (Table 9), Season-long mean K and P had significant differences among depths. The highest concentration of K was in the top 0-15 cm mean while there were not significant differences between the depths 15-30 and 30-45 cm (these depths presented the lowest concentrations), for P the highest concentration was observed at 0-15 and 15-30 cm (there were not significant differences between these depths) and the lowest concentration was at 30-45 cm. These data indicate that applications of fertilizer through fertigation help to keep the nutrients in the root zone and can lead to improve the uptake of nutrients through the roots system of the tomatoes. In fall 2013 season, the lowest concentration of NO₃^-N were observed at 15-30 and 30-45 cm. Hebbar et al. [19] found similar results where the lowest concentration of NO₃^-N was observed at the 30-45 cm in the treatment of normal fertilizer fertigation. Also, NH₄⁺-N highest concentration was in the top 0-15 cm in the spring 2014 season. On the other hand, the highest concentration of NO₃^-N was found in the top 0-15 cm for the present study, which shows how the use of fertigation helps to keep the NO₃^-N in the root zone area reducing the risk of leaching. Similar observations were found by Kadyampakeni et al. [27,28] in a study on citrus trees.

Table 6: Interaction of fertilization rate by weeks after transplanting (WAT) and depth for means of soil NH₄⁺-N in fall 2013.

Table 7: Means of fertilizations rates with weeks after transplanting (WAT) of concentration of NO₃^-–N and K in fall 2013, NH₄⁺-N, NO₃^-–N in spring 2014 and NH₄⁺-N, NO₃^-–N, K and P in fall 2014 in the soil.

Table 8: Mean NO₃^- N and P concentration in fall 2013, NH₄⁺-N concentration in spring 2014, K and P concentration in fall 2014 at 0-15, 15-30 and 30-45-cm soil depth.

Table 9: Mean soil phosphorus concentration at different weeks after transplanting (WAT) in fall 2013 and spring 2014.

In the fall 2013 season the greatest concentration of P was found at 15-30 cm similar results were found in the fall 2014 season were the highest concentration of P in the soil was found at 0-15 and 15-30 cm. In the same season, the highest concentration of K was found in the top 0-15 cm and the lowest at the 15-30 and 30-45 cm soil depth. Soil P concentrations were significantly lower at 9 WAT for both fall 2013 and spring 2014, potentially indicating reduced availability at suggest application timings (Table 10). Soil K concentrations were significantly greater for treatments 3, 4 and 5. In fall 2013 season the lowest concentration of P was at 9 WAT and highest at 4 WAT but 0, 4 and 13 WAT were similar in P concentrations, meanwhile at spring 2014 season the highest concentration of NH4-N at 11 WAT was found using 200% of UF/IFAS recommendation.

Table 10: Mean potassium concentration in the soil among fertilization rates in spring 2014.

The tomato plants that received more fertilizer than 280 kg ha^-1 N and 253 kg ha^-1 K (2013) or 224 kg ha^-1 N and 196 kg ha^-1 K (2014) had excessive leaf and stem biomass compared with lower N and K rates, but had statistically no increase in fruit yield per plant or area. Fruit yield of the same N and K rates (treatment 3) was not significantly different than the two higher N and K rates. The lack of yield increase shows that even though plants received more fertilizer, these fertilization rates did not produce significantly more fruit than moderate N and K rates. These results confirm the hypothesis that applying more fertilizer than recommended does not result in greater yields. Thus, the application of more fertilizer for tomato grown on sandy soils of Florida under drip fertigation does not typically result in a significant increase in yields. Soil nitrate concentrations highest soil concentration in top 0-15 cm at 7 cm distance from the fertigation line in fall 2013 and in the spring 2014. Highest phosphorus concentration was found in the 0-30 cm soil depth in both fall seasons. The highest K concentrations were found in the top 0-15 cm in the fall season 2014. This study showed how the implementation of drip irrigation system helps to maintain the nutrients in the root zone for enhanced uptake as a function of soil depth and drip per placement.