Introduction
Ensuring world food security without degrading the environment and compromising the critical ecosystem services is one of the biggest global challenges. Horticultural production systems are intensively managed which can compromise soil functioning (Norris and Congreves, Reference Norris and Congreves2018). Given the high economic and nutritive value and global demand, the importance of food security is elevated. Additionally, many horticultural crops, specifically vegetable crops, leave little residue in the soil after harvest, have short growing season thus leaving the soil exposed for extended periods of time and contribute very little organic matter to the soil. Cover crops (CC) have been shown to sustain soil health and maintain agricultural productivity (Blanco-Canqui et al., Reference Blanco-Canqui, Shaver, Lindquist, Shapiro, Elmore, Francis and Hergert2015; Jian et al., Reference Jian, Lester, Du, Reiter and Stewart2020), specifically in horticultural production systems. In the temperate humid climate such as southwestern Ontario, most of the vegetable crops are harvested from the end of August to early September. As such, there is a large opportunity to successfully establish and grow CC in the fall season after the harvest of vegetable crops (Norris and Congreves, Reference Norris and Congreves2018).
Globally, tomato (Solanum lycopersicum L.) is the second most important horticultural crop, with a production area of about 5,000,000 ha (FAOSTAT, 2020) and a very high economic value. In Ontario, Canada processing tomato is grown on 5169 ha with a gross farm value of $91,758,000 (OPVG, 2023). In the past 5 years, tomato production has increased in Ontario and Canada (OPVG, 2023; Statistics Canada, 2024) and is expected to continue to increase in the future mainly due to the high nutritive value, associated health benefits, and antioxidant properties of tomato (Bhowmik et al., Reference Bhowmik, Kumar, Paswan and Srivastava2012).
There is a growing consumer demand for vegetables that offer higher nutritional benefits, which are linked to an improved chemical composition of the produce (Li et al., Reference Li, Deng, Liu, Loewen and Tsao2012; Sridhar et al., Reference Sridhar, Ponnuchamy, Kumar, Kapoor, Vo and Prabhakar2021). Processing tomatoes are rich in phytochemicals, particularly phenolics and carotenoids along with essential nutrients such as Na, K, Ca, Mg, P, Fe, Cu, and Zn that are beneficial to human health (Dorais, Ehret and Papadopoulos, Reference Dorais, Ehret and Papadopoulos2008; Hazewindus et al., Reference Hazewindus, Haenen, Weseler and Bast2012). Lycopene, one of the most effective antioxidants among carotenoids, makes up approximately 80–90% of the carotenoid content in red ripe tomatoes (Li et al., Reference Li, Deng, Liu, Loewen and Tsao2012). Likewise, beta-carotene, a precursor of vitamin A, accounts for around 7% of tomato carotenoid content (Li et al., Reference Li, Deng, Liu, Loewen and Tsao2012). Both lycopene and beta-carotene are known to reduce the risk of cancer and cardiovascular diseases (Ali et al., Reference Ali, Sina, Khandker, Neesa, Tanvir and Kabir2021; Yang, Zhang and Tsao, Reference Yang, Zhang and Tsao2020). In addition to carotenoids, vitamin C and polyphenols, primarily flavonoids, are also significant antioxidant components in tomato that are important for human health. Plant phenolic compounds enhance human health as they function as antiatherogenic, anti-inflammatory, antiallergenic, and cardioprotective agents (Shahidi et al., Reference Shahidi, Varatharajan, Oh and Peng2019; Zhang, Liu and Tsao, Reference Zhang, Liu and Tsao2016). Due to these health benefits, polyphenols have received increasing attention by investigators (Rasouli, Farzaei and Khodarahmi, Reference Rasouli, Farzaei and Khodarahmi2017; Xiao, Reference Xiao2022; Zhang and Tsao Reference Zhang, Liu and Tsao2016). Research on land use management practices is, therefore, needed to develop sustainable vegetable production systems with minimum environmental degradation and support chemical composition.
While links between soil health and crop productivity are implied, studies on soil health assessment in vegetable crops are not as comprehensive and prevalent as in field crops (Norris and Congreves, Reference Norris and Congreves2018). Moreover, very few published studies have considered the CC impacts on tomato fruit quality (Belfry et al., Reference Belfry, Trueman, Vyn, Loewen and Van Eerd2017; Chahal and Van Eerd, Reference Chahal and Van Eerd2021; Price et al., Reference Price, Duzy, Balkcom, Kelton, Kornecki, Sarunaite and Petr2016). Fruit quality is particularly an important parameter for tomato crop where financial incentives are provided to Ontario growers based on color and natural tomato soluble solids (OPVG, 2023). Research studies on vegetable crop attributes and soil characteristics are, therefore, needed to understand and identify if or to what extent agricultural management practices promote soil, plant, and human health.
Although CC research on tomato fruit quality is limited, CC-induced effects on various agroecosystem services have been comprehensively documented in the literature. For instance, CC reduced water and wind erosion (Blanco-Canqui et al., Reference Blanco-Canqui, Shaver, Lindquist, Shapiro, Elmore, Francis and Hergert2015), conserved soil moisture (Sharma, Irmak and Padhi, Reference Sharma, Irmak and Padhi2018), enhanced weed suppression (Masilionyte et al., Reference Masilionyte, Maiksteniene, Kriauciuniene, Jablonskyte-Rasce, Zou and Sarauskis2017), enhanced carbon sequestration (Chahal and Van Eerd, Reference Chahal and Van Eerd2020), increased N availability, improved insect pest management (O’Reilly et al., Reference O’Reilly, Robinson, Vyn and Van Eerd2011, Reference O’Reilly, Lauzon, Vyn and Van Eerd2012), and increased the subsequent crop yield (Belfry et al., Reference Belfry, Trueman, Vyn, Loewen and Van Eerd2017; Chahal and Van Eerd, Reference Chahal and Van Eerd2020). Nonetheless, recent studies suggest that CC effects on soil health indicators (physical, chemical, and biological characteristics), the soil functions that the indicators perform, and crop yield are inconsistent and are dependent on agronomic and environmental variables (Blanco-Canqui et al., Reference Blanco-Canqui, Ruis, Koehler-Cole, Elmore, Francis and Shapiro2023; Ruis et al., Reference Ruis, Blanco-Canqui, Koehler-Cole, Jasa, Slater, Elmore and Ferguson2020). The concept of soil functions highlights the multifunctionality of the soil system (Drobnik et al., Reference Drobnik, Greiner, Keller and Grêt-Regamey2018) and is closely linked with the various ecosystem services that the soil provides. One or more soil health indicators are usually aggregated together and are associated with one or more critical soil functions, such as erosion control, water quality and supply, nutrient cycling, climate regulation, biodiversity conservation, and biomass production (Table 1). Additionally, CC effects on soil health and crop productivity are observed in the medium to long term (Basche et al., Reference Basche, Kaspar, Archontoulis, Jaynes, Sauer, Parkin and Miguez2016; Blanco-Canqui et al., Reference Blanco-Canqui, Mikha, Presley and Claassen2011; Chahal and Van Eerd, Reference Chahal and Van Eerd2019; Olson, Ebelhar and Lang, Reference Olson, Ebelhar and Lang2014). Furthermore, despite a huge body of research on CC, the direct (or indirect) linkage among soil, plant, and human health, and the various ecosystem services provided by CC is not comprehensively explored.
Table 1. Assigning soil function(s) to various soil health indicators and the method/protocols employed
A long-term CC experiment, established in 2007, at Ridgetown, Ontario, Canada was used to assess the relationship of soil health with plant productivity (tomato fruit yield) and human health (tomato fruit quality). Previous studies at the same experimental site have demonstrated a positive effect of CC on crop productivity (Chahal and Van Eerd, Reference Chahal and Van Eerd2021; Trueman et al., Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023), soil health (Chahal and Van Eerd, Reference Chahal and Van Eerd2019), and disease suppression (Trueman et al., Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023) in the medium to long term. The main objectives of the study were (i) to evaluate the long-term (12-year) effect of CC on a suite of soil physical, chemical, and biological indicators (n = 56), and tomato marketable yield and quality, and (iii) determine the association between soil functionality and tomato fruit yield and quality parameters. We hypothesized that long-term CC treatments would have greater soil health, fruit yield, and fruit quality than no-CC control (i.e., null hypothesis of no difference). This research will advance our understanding of the potential soil mechanisms through which long-term CC increase crop productivity and contribute to enhancing knowledge of soil functionality.
Materials and methods
Site description and experimental design
As previously described (e.g., Belfry et al., Reference Belfry, Trueman, Vyn, Loewen and Van Eerd2017; Chahal and Van Eerd, Reference Chahal and Van Eerd2018, Reference Chahal and Van Eerd2019), a long-term CC experiment was established in 2007 (Site A) at the Ontario Crops Research Center, Ridgetown, Ontario, Canada (42°46’N, 81°96’W, elevation 200 m). Note the adjacent site initiated in 2008 (Site B) was not used for this study. The soil at the site is a Normandale sandy loam (Matschke, Reference Matschke2019) (Orthic Humic Gleysol) with 71:16:13% sand:silt:clay. The experiment had a crop rotation of the processing vegetables grown in the region and included grains and oilseeds (Table 2).
Table 2. Crop rotation and postharvest activities in the long-term cover crop experiment since establishment in 2007 at Ridgetown, Ontario, Canada
a All vegetable crops were for the processing market.
b Represents the year of soil health sampling and crop yield.
The experimental design was the split–split plot randomized complete block with four replications. Cover crop treatments (winter cereal rye (Secale cereale L.), radish (Raphanus sativus var. longipinnatus L.), mixture of radish and rye (radish + rye), and no-CC control) were the permanent main plot effect. The main plot size was 16 m × 6 m. Winter wheat straw management (straw retained and removed) was the split plot effect (6 × 8 m) in 2018. Details of winter wheat straw management are described in Trueman et al. (Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023). The N fertilizer application (none or application of calcium ammonium nitrate 27:0:0 at 140 kg N ha−1) was the split–split plot effect (6 × 4 m; i.e., n = 80). The fertilizer N was hand broadcasted before transplanting tomato crop in 2019.
Most recently, CC were planted on August 10, 2018, after harvesting the winter wheat, removing wheat straw in split plot, and tillage to approximately 15 cm depth by disking and cultivating the field. Cover crops were planted with a seed drill (rows 19 cm apart) at seeding rates of 67 kg ha−1 (rye), 14 kg ha−1 (radish), and the mix of 9 kg ha−1 and 34 kg ha−1 (radish and rye, respectively). On August 30, 2018, weeds were chemically controlled with glyphosate (2.47 L ha−1 Factor® 540) in no-CC plots (Trueman et al., Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023).
In the following spring, the entire experimental area was sprayed with glyphosate to terminate rye and control weeds. Fertilizer P and K (0-23-30) was surface applied to the entire experiment at 560 kg ha−1 and incorporated up to 15 cm with disc and cultivator. One day prior to tomato transplanting, herbicide (2.47 L ha−1 Boundary® LQD, 2.47 L ha−1 Roundup WeatherMAX®) was applied to all plots. On June 4, 2019, plots were cultivated once and processing tomatoes (cultivar CC337 provided by Conagra Brands, Inc.) were transplanted at a plant density of 26,337 plants ha−1. Pest control and other field management operations were done according to recommended Ontario processing tomato production practices (OMAFRA, 2008).
Tomato yield and quality
Processing tomatoes were hand harvested on September 9, 2019 from each split–split plot (n = 80), as previously described (Trueman et al., Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023). Briefly, harvest was targeted when fruit was 70% red in the most advanced plots. Tomato fruit was sorted into red (<5% orange), orange (>90% yellow or orange), breaker (>1% but <90% yellow or = orange), green and rots, for which fresh fruit weight was recorded. Tare penalty of damage and lesions was assessed on 100 red fruits. Fresh weights of all fruit except rots and with the tare penalty applied and harvest area (2 m × two tomato rows) were used to determine fresh marketable fruit yield expressed as Mg ha−1. From each split–split plot, delta yield was calculated to determine the difference in marketable tomato yield of the control treatment (no cover crop [main-plot factor], straw retained [split-plot factor: winter wheat applied in 2018], 0 N [split–split plot factor: fertilizer N applied to tomato crop in 2019]) from the remaining treatments.
During harvest, a 2-kg subsample of red fruit was collected from each split–split plot, rinsed with water to remove soil, surface dried, and stored at room temperature (~20°C) until analyzed (within 2 days). The fruit quality assessment methods were consistent with Van Eerd, Loewen and Vyn (Reference Van Eerd, Loewen and Vyn2015). Each subsample was blended at medium speed for 40 s, in a vacuum (88 kPa) using a Waring CB6 commercial blender (Waring Commercial, Torrington, CT). Subsequently, each subsample was sieved through a 27-mesh screen to separate the seeds from the pulp, which was used to quantify color, pH, and soluble solids for each subsample. To evaluate color, a subsample of pulp was put in a small dish and analyzed using an Agtron E-5M spectrophotometer (calibrated at 48) and the filtrate was used to measure pH, using a calibrated Orion pH meter (Thermo Fisher Scientific, Nepean, ON). Simultaneously, another subsample (~10 mL) of pulp was filtered through Fisher Scientific P8 filter paper and soluble solids were measured by means of a Palette PR101 temperature compensated digital refractometer (Atago USA, Inc., Bellevue, WA).
Phytochemical content and antioxidant activities were assessed using a sample of six red tomatoes from each split–split plot collected at harvest. Directly following harvest, the samples were washed, sliced vertically (~2-cm thick), placed flat in a plastic bag, and stored in a freezer at −18°C. Samples were kept frozen till analyzed using methods outlined in Li et al. (Reference Li, Deng, Liu, Young, Zhu, Loewen and Tsao2011, 2012, 2013), Zhang et al. (Reference Zhang, Deng, Ramdath, Tang, Chen, Liu, Liu and Tsao2015). The following were quantified: total phenolics (gallic acid equivalent, mg g−1 dry weight), total flavonoids (catechin equivalent, mg g−1 dry weight), 2, 2 diphenyl-1- picrylhydrazyl (DPPH, trolox equivalent, μmol g−1 dry weight), lutein (mg g−1 dry weight), lycopene (mg g−1 dry weight), beta-carotene (mg g−1 dry weight), and total carotenoids (mg g−1 dry weight). The plant phenolic data were expressed as content (kg ha−1) for data presentation based on the tomato fruit biomass. For each of the split–split plot, tomato fruit quality index was the mean of standardized (z score) fruit quality parameters: Agtron color, natural tomato soluble solids, fruit pH, total phenolics content, total flavonoids content, DPPH content, lutein content, lycopene content, beta-carotene content, and total carotenoids content.
Soil sampling and analysis
Prior to spring field activities, soil sampling (0–15 cm) was done in mid-May 2019 by the Soil Health Institute (SHI) as a part of North American Project to Evaluate Soil Health Measurements (NAPESHM) (Norris et al., Reference Norris, Bean, Cappellazzi, Cope, Greub, Liptzin, Rieke, Tracy and Morgan2020). A comprehensive description of soil sample collection is available at Liptzin et al. (Reference Liptzin, Norris, Cappellazzi, Mac Bean, Cope, Greub, Rieke, Tracy, Aberle, Ashworth and Tavarez2022) and Norris et al. (Reference Norris, Bean, Cappellazzi, Cope, Greub, Liptzin, Rieke, Tracy and Morgan2020). For soil health analysis, the CC treatments of interest in the NAPESHM project excluded oat and were sampled from the straw retained split plots (n = 16; note split–split plot treatment had not yet been applied). Briefly, using spade (38 × 15 cm blade), six holes (15 × 15 cm) were dug to a 15-cm depth in a zigzag pattern within a plot covering row and inter-row locations, 1 m away from edges (Norris et al., Reference Norris, Bean, Cappellazzi, Cope, Greub, Liptzin, Rieke, Tracy and Morgan2020). A slice of soil (4-cm wide, 1.5-cm thick, 15-cm depth) was removed from three sides of the hole using a soil knife. Soils from three sides of six holes were placed in a plastic bag to produce a composite sample for analysis, and the bag was placed in an ice-filled cooler for transport to the laboratory. The bulk soil was combined in an isopropyl alcohol-sterilized container and passed through an 8-mm sieve. Soil (about 400 g) from each plot was sealed in a plastic bag, placed in an ice-filled cooler, and delivered to soil testing laboratories for biological properties analysis. The remainder of the soil was shipped to other laboratories, where soil was air dried, and sieved to 2 mm before being analyzed for physical and chemical properties. To measure soil bulk density (BD), four soil cores to a depth of 7.6 cm were collected by cylinder from each experimental unit. Two cores were stored individually, while the other two were combined. Bulk density was determined by dividing the weight of oven-dried soil by the volume of the core, adjusted for coarse fragments (Blake, Reference Blake, Hartge and Klute1986). Overall, these sampling depths were consistent with soil health sampling protocols (e.g., Norris et al., Reference Norris, Bean, Cappellazzi, Cope, Greub, Liptzin, Rieke, Tracy and Morgan2020) and were not intended to be reflective of crop rooting depth but reflect the expected greatest density of roots and microbial communities that influence soil function.
A total of 54 soil health indicators consisting of physical, chemical, and biological parameters were analyzed from 16 experimental units (i.e., four CC treatments in four replicates). The samples for soil health analysis were collected from no-CC, radish, rye, and radish + rye treatments applied to all winter wheat straw retained and straw removed plots. The soil health indicators were measured according to the methods followed in three different laboratories, that is, the Ohio State University lab (OSU), Cornell University lab, and Haney soil health test as previously described by Norris et al. (Reference Norris, Bean, Cappellazzi, Cope, Greub, Liptzin, Rieke, Tracy and Morgan2020) and Sainju et al. (Reference Sainju, Liptzin, Dangi and Ghimire2021) and outlined in Table 1. Consistent with the idea outlined in Blum (Reference Blum2005), we used the concept of ‘soil functions’ to better understand the CC induced effects on the critical tasks that the soil provides. One or more of the soil health indicators were categorized and grouped based on the soil functions that they perform. Six critical soil health functions identified in our study were (a) erosion control, (b) water quality and supply, (c) nutrient cycling, (d) climate regulation, (e) biodiversity conservation, and (f) biomass production (Table 1).
Data analysis
The CC-induced effects on the soil health indicators were analyzed using R (version 4.1.1). For the soil health analysis, four CC treatments were available (i.e., oat CC treatment was not sampled). As indicated earlier, some soil health indicators were measured in duplicate or triplicate as multiple labs (OSU, Cornell, and Haney) quantified the same indicator. Likewise, some nutrient concentrations were measured using different extraction methods (H3A, Mehlich 3, and Modified Morgan). Therefore, these data were subjected to two-way ANOVA (conducted with built-in ‘stats’ package in R) to determine if lab or extraction method impacted the interpretation of CC effects. Linear mixed model (lme4 package) was used where fixed effects were CC and method with replicate as random effect. If there was no interaction of CC and method, data obtained from the methods commonly employed in Ontario, commercial laboratories or at the University of Guelph, were chosen for further analyses. Therefore, the analysis focused only on CC effects and data were analyzed as one-way ANOVA. To determine the impact of CC treatment on soil health indicators, a linear mixed model (lme4 package) was used with CC treatment as fixed effect and replication as the random effect.
For crop delta yield and fruit quality parameters, all five CC treatments were used and analyzed using the PROC GLIMMIX in SAS (SAS Institute version 9.4, Cary, NC). The statistical model consisted of CC, straw management, N fertilizer, and their two-way and three-way interactions as the fixed effects. The random effects were replication, replication by CC, and replication by CC by straw management. The interaction term was added to account for the split–split plot effect. The initial analysis revealed no interaction between the main effects; hence the analysis focused on CC effects only. All the assumptions of ANOVA were met (i.e., homogeneity and errors were random, independent, normally distributed, mean of zero). Assumptions were tested by plotting residuals of predicted by fixed effects and assessing normality with a Shapiro–Wilk test (Bowley, Reference Bowley2015). Cover crop treatment means were compared using Tukey–Kramer’s HSD test. The level of statistical significance was set to P < 0.1 to be less conservative in detecting CC treatment differences as the number of soil health samples were low (n = 16); hence, we prioritized understanding relationships among many different indicators of soil health (n = 56). To better understand the relationship of soil functionality with the crop productivity and quality in 2019, principal component analysis (PCA) was conducted using PROC PRINCOMP in SAS.
Results
Soil functions
Erosion control
Among erosion control indicators, long-term cover cropping significantly influenced aggregate stability (Table 3). Cereal rye, radish + rye, and radish had the greatest (average 91.6 ± 1.15%), while no-CC had the least aggregate stability (83.8 ± 1.15%; Table 3). No differences in the remaining erosion control indicators due to long-term CCing were observed (Table 3).
Table 3. Effect of long-term cover cropping on surface soil (15 cm) mean (SE) indicators of functions related to erosion control or water quality and supply in 2019 at Ridgetown, Ontario, Canada
a-b Within each row, means followed by a different letter indicate statistically significant difference at P < 0.1 per Tukey–Kramer.
Water quality and supply
Cover cropping influenced two indicators associated with water regulation (Table 3). Bulk density was significantly lower in radish + rye (1.14 mg m−3) compared to no-CC (1.25 mg m−3). Mesoporosity was significantly greater in radish + rye (0.31 cm3) compared to no-CC (0.26 cm3), indicating improved soil structure. Intermediate bulk density and mesoporosity values were observed in the two other CC treatments. However, no differences in the remaining soil indicators due to long-term CCing were observed (Table 3).
Nutrient cycling
Compared to other functions, a greater number of soil properties associated with nutrient cycling (18 of 31 indicators) were statistically significantly different among CC treatments (Table 4). Soil pH was greater with radish + rye and radish than rye and no-CC treatments. Water extractable N, electrical conductivity, S, and B were significantly influenced by CC treatment, where values followed in descending order of radish + rye = radish ≥ rye ≥ no-CC. Similarly, nitrate-N and K followed the same descending order. Potentially mineralizable N was 17 mg kg−1 greater with radish + rye than no-CC, but the other two treatments were not different from all treatments. Radish + rye and radish had 373 mg kg−1 greater Ca concentrations than rye, with the no-CC having intermediate concentrations. Water extractable C and Na concentrations were greater with radish than no-CC, with the other two treatments being intermediate. Soil Mn concentrations were approximately 3 mg kg−1 greater with radish than no-CC and rye, but radish + rye was not different. (Table 4). In contrast, no-CC had the greatest and rye the lowest concentration of total P, and inorganic P with the other two treatments as intermediary. All planted CCs had lower sodium adsorption ratio than no-CC. Soil Al concentrations were greater with no-CC than radish and radish + rye, where rye was intermediate. Soil Ni concentrations were greater in no-CC and rye plots compared to radish + rye plots, with radish being intermediary (Table 4). Overall, radish + rye and radish tended to have better indicators of nutrient cycling than no-CC and rye tended to be intermediary.
Table 4. Effect of long-term cover cropping on surface soil (15 cm) mean (SE) indicators of functions related to nutrient supply in 2019 at Ridgetown, Ontario, Canada
a–c Within each row, means followed by a different letter indicate statistically significant difference at P < 0.1 per Tukey–Kramer.
Climate regulation
Soil carbon and nitrogen-related indicators were assessed to evaluate the impact of cover cropping on climate regulation (Table 5). No significant differences were observed in soil organic carbon (SOC), total nitrogen, or organic matter content across treatments. However, radish + rye had greater inorganic N content, than rye and no-CC, but similar amounts as radish. Microbial respiration, measured as CO₂-C after 1-day or 4-day incubation, showed no significant variation across treatments (Table 5).
Table 5. Effect of long-term cover cropping on surface soil (15 cm) mean with standard error (SE) indicators of functions related to climate regulation and biodiversity conservation in 2019 Ridgetown, Ontario, Canada
a–c Within each row, means followed by a different letter indicate statistically significant difference at P < 0.1 per Tukey–Kramer.
Biodiversity conservation
Only one soil biological indicators with function relevant to biodiversity conservation were influenced by cover cropping (Table 5). Alkali phosphatase enzyme activity values followed in descending order of radish ≥ no-CC = radish + rye ≥ rye. No differences in the remaining soil indicators due to long-term CCing were observed (Table 5).
Tomato fruit quality and yield
Delta yield and all fruit quality parameters did not have a significant interaction between CC, straw, and N management; hence, only CC effects were shown. A significant CC effect was observed (P = 0.003). Delta yield was greater for all CC than the no-CC and followed descending order of radish ≥ radish + rye = rye ≥ oat> > no-CC (Table 6). Agtron color and pH were not significantly different among CC treatments (Table 6). Natural total soluble solids were significantly greater in radish than no-CC and rye, but the remaining two CC treatments were not different than any treatment (Table 6). Oat had the least lycopene, total carotenoid content, total phenolic content, and DPPH (Fig. 1). Radish, rye, radish + rye either had equal or greatest content of phytochemicals and antioxidants (Fig. 1). The phytochemical characteristics of tomato fruit were not significantly different among CC treatments for total flavonoid content and two carotenoids (lutein and beta-carotene) (Fig. 1). Tomato fruit quality index was also significantly different among CC treatments where cereal rye and radish had the greatest values, while oat had the least values and the other two treatments were intermediary (Table 6).
Table 6. Effect of long-term cover cropping on mean with standard error (SE) processing tomato fruit delta yield and quality parameters in 2019 at Ridgetown, Ontario, Canada
a–c Within each column, means followed by a different letter indicate statistically significant difference at P < 0.1 per Tukey–Kramer.
d Delta yield represents the difference between marketable tomato yield of treatment and control (51.8 Mg ha−1; no cover crop [main-plot factor], straw retained [split-plot factor: winter wheat applied in 2018; P = 0.177], 0 N [split–split plot factor: fertilizer N applied to tomato crop; P = 0.157] plots. Two-way and three-way interactions were not significant P ≥ 0.592.
Figure 1. Effect of long-term cover cropping on the phytochemical characteristics of processing tomato fruit in 2019 at Ridgetown, ON (means with standard error, n = 16). The phytochemical properties measured were (a) carotenoid content (lupein, lycopene, β-carotene, and total carotenoid content) as measured by high-performance liquid chromatography, (b) total phenolic and flavonoid content as measured by chlorogenic acid equivalents, and (c) antioxidant capacity as measured by 2,2-diphenyl-1-picrylhydrazyl (DPPH) on Trolox equivalents. Means followed by a different letter indicate statistically significant difference at P < 0.1 per Tukey–Kramer.
Relationships between soil functions and tomato yield and quality
To better understand relationships and soil impact of tomato yield and quality, soil health indicators were grouped based on the soil functions that they provide (Table 1). A considerable amount of the variance was accounted for in each functional datasets; PC1 explained from 37.2% to 59.5% and PC2 was 11.4% to 38.1% depending on soil function (Fig. 2). Principal component analysis showed a clear difference between no-CC and CC species specifically for the erosion control, nutrient supply, and climate regulation services provided by the CC (Fig. 2). As demonstrated by PCA, the concentration of soil health indicators was lesser from no-CC plots than the CCs (Fig. 2), suggesting no-CC had a lesser capacity to sustain soil functions than CC species.
Figure 2. Principal component analysis revealing associations among dependent variables: soil health indicators, tomato marketable yield, and quality from four long-term cover crop treatments (no cover crop, radish, rye, and mixture of radish and rye) in 2019 at Ridgetown, ON (n = 16). The soil health indicators were grouped based on the soil functions (a–f). See Table 1 for list of indicators.
As expected, tomato marketable yield and fruit quality index were closely associated with each other (Fig. 2a–f). Tomato yield and fruit quality index were closely associated with soil organic matter, aggregate stability, total C, inorganic N, ACE protein index, and microbial C (Fig. 2), where PC1 explained 35.4% and PC2 explained 16.8% of the variance in the entire dataset (Supplementary Fig. S1). Tomato yield had a negative relationship with sand and bulk density (Fig. 2a), sodium adsorption ratio, Al (Fig. 2c), and available water holding capacity (Fig. 2e). Tomato yield was positively associated with all the fruit quality and antioxidant and phytochemical parameters such as fruit color, DPPH, β-carotene, lycopene, total carotenoid content, total flavonoid content, and lutein, but not with natural total soluble solids (Fig. 2b).
Discussion
In our study, 22 of 56 indicators were significantly different due to long-term CCing. The soil health indicators had greater values or better soil quality with CC than without. Our findings of a trend of a greater soil health with long-term CCing than no-CC was consistent with a previous medium-term study conducted at the same site (Chahal and Van Eerd, Reference Chahal and Van Eerd2018, Reference Chahal and Van Eerd2019) and other long-term CC research (Basche et al., Reference Basche, Kaspar, Archontoulis, Jaynes, Sauer, Parkin and Miguez2016; Klopp et al., Reference Klopp, Blanco-Canqui, Jasa, Slater and Ferguson2025; Liptzin et al., Reference Liptzin, Rieke, Cappellazzi, Bean, Cope and Greub2023; Olson et al., Reference Olson, Ebelhar and Lang2014; Peng et al., Reference Peng, Rieke, Chahal, Norris, Janovicek, Mitchell, Roozeboom, Hayden, Strock, Machado and Sykes2023). Relative to the medium-term study at the same site, this long-term research had fewer soil health indicators which were significantly different among CC treatments that was attributed to smaller sample size (n = 16) in our study than previously (n = 40; Chahal and Van Eerd, Reference Chahal and Van Eerd2019).
The soil physical indicators were mainly assigned to functions of erosion control and regulating water quality and supply. Aggregate stability, bulk density, and mesoporosity were significantly different among the CC treatments, where values indicate better or equivalent soil health with than without long-term CC. The 10% greater bulk density in the no-CC than radish + rye (but no difference for other CC) was somewhat surprising given the sandy loam soil that is less prone to compaction than fine-textured soils. This was not consistent with other findings at this site (Chahal and Van Eerd, Reference Chahal and Van Eerd2019; Peng and Van Eerd Reference Peng and Van Eerd2024) and pooled with other North American long-term CC experiments (Bagnall et al. Reference Bagnall, Morgan, Cope, Bean, Cappellazzi, Greub, Liptzin, Norris, Rieke, Tracy and Aberle2022a and Reference Bagnall, Morgan, Bean, Liptzin, Cappellazzi and Copeb; Peng et al., Reference Peng, Rieke, Chahal, Norris, Janovicek, Mitchell, Roozeboom, Hayden, Strock, Machado and Sykes2023), which may be partially attributed to different sampling methods (e.g., bulk density rings vs. intact cores, surface vs. 7.6 cm depth). Consistent with bulk density, the no-CC had 19% lower mesoporosity than the radish–rye. Tillage most recently occurred in August of the previous year, thus tap and fibrous roots of the radish + rye mix likely contributed to the observed bulk density and mesoporosity differences. Regardless, all values were within expected range for sandy loam soil, do not indicate compaction, nor would be expected to limit root growth.
Aggregate stability finding aligns with one of the three indicators of soil health recommended by the Soil Health Institute (Bagnall et al., Reference Bagnall, Rieke, Morgan, Liptzin and Cappellazzi2023). Our result of significantly greater aggregate stability with radish + rye than the no-CC was consistent with our medium-term (Chahal and Van Eerd, Reference Chahal and Van Eerd2019) and other long-term research (Blanco-Canqui and Jasa, Reference Blanco-Canqui and Jasa2019). Blanco-Canqui and Jasa (Reference Blanco-Canqui and Jasa2019) reported an increase in the mean weight diameter of water-stable aggregates with grass CCs grown in a long-term no-till system on a silty clay loam in Nebraska, USA. Likewise, a recent study based on long-term CC experiments in North America reported a significant increase in aggregate stability by ~5–22% (Rieke et al., Reference Rieke, Bagnall, Morgan, Flynn, Howe, Greub, Mac Bean, Cappellazzi, Cope, Liptzin and Norris2022). One potential mechanism through which CC increase aggregate stability might be attributed to enhanced soil C concentration (2.03–2.13%) from CC residues and roots which would encourage microbial activity and produce extracellular mucilaginous materials that stabilize aggregates (Blanco-Canqui et al., Reference Blanco-Canqui, Mikha, Presley and Claassen2011; Liu, Ma and Bomke, Reference Liu, Ma and Bomke2005). Another mechanism by which CC enhance aggregate stability is through physical protection of surface soil and mitigating erosion by wind and water (Blanco-Canqui et al., Reference Blanco-Canqui, Shaver, Lindquist, Shapiro, Elmore, Francis and Hergert2015).
Among the indicators related to the soil function of nutrient supply, pH, electrical conductivity, Mn, S, K, and nitrate had greater values with CCs than without. As a measure of soluble salts, excessive soil electrical conductivity can lower plant water availability, nutrient supply, and reduce soil structure. Cover crops significantly increased the EC by 14.3–35.7% compared to the no-CC control, but all values were in the acceptable range, that is, 0–1.5 dS m−1 for plant growth and microbial activity (Smith and Doran, Reference Smith and Doran1997). Long-term CC increased the concentration of soil nutrients such as K, S, Mn, and nitrate-N which was consistent with prior reviews and meta-analysis (Koudahe, Allen and Djaman, Reference Koudahe, Allen and Djaman2022; Farmaha, Sekaran and Franzluebbers, Reference Farmaha, Sekaran and Franzluebbers2022). Nascente and Stone (Reference Nascente and Stone2018) reported increased concentration of soil K with CC in soybean-upland rice rotation. In contrast, no-CC plots had greater SAR, total P, and inorganic P than CC plots. Consistent with our findings, a study by Villamil et al. (Reference Villamil, Bollero, Darmody, Simmons and Bullock2006) in Illinois, USA, showed that corn grown after hairy vetch (Vicia villosa Roth.) and rye CC had lower amounts of soil available P. Similarly, Almeida et al. (Reference Almeida, Menezes-Blackburn, Zhang, Haygarth and Rosolem2019) found that a CC of ruzigrass (Urochloa ruziziensis) decreased soil available P in a long-term soybean-CC system in Brazil. The lower P levels with CC than no-CC in our study might be due to the immobilization of P in CC biomass, which may be beneficial from an environmental perspective as it reduces the likelihood of P loss through runoff (Villamil et al., Reference Villamil, Bollero, Darmody, Simmons and Bullock2006). Additionally, the higher crop yields observed with CC than no-CC in our study might have led to increased P removal from the soil with the harvested crops, as historical fertilizer P (and K) applications were not adjusted based on soil P levels.
Soil with high SAR values is susceptible to clay dispersion (Frenkel, Goertzen and Rhoades, Reference Frenkel, Goertzen and Rhoades1978), decreased infiltration (Suarez, Wood and Lesch, Reference Suarez, Wood and Lesch2008), and reduced aggregate stability (Rahimi, Pazira and Tajik, Reference Rahimi, Pazira and Tajik2000). The exact reason for low SAR values in CC plots compared to no-CC is not fully understood but might be related to the enhanced aggregate stability in our study. Enhancing aggregate stability helps to prevent clay dispersion and improves water infiltration. In general, SAR values greater than 13 are considered sodic and are detrimental to plant growth (Carter and Gregorich, Reference Carter, Gregorich, Carter and Gregorich2008), which was not the case in our study.
Among the indicators associated with climate regulation and biodiversity conservation soil function, only inorganic N and alkali phosphatase enzyme activity, respectively, were significantly greater with CC than no-CC. Similarly, acid phosphatase activity was lowest in the no-CC, but not significantly. Phosphatase enzymes have a key role in P cycling (García-Ruiz et al., Reference García-Ruiz, Ochoa, Hinojosa and Carreira2008). The lower phosphatase enzyme activity in the no-CC plots is likely a reflection of the greater soil total P and inorganic P concentrations minimizing the need for P mineralization. While the opposite was observed in CC plots. Phosphatase enzymes are sensitive to sustainable soil management practices with a significant increase in phosphatase enzymes with CC (Adetunji et al., Reference Adetunji, Ncube, Mulidzi and Lewu2020; Fernandez et al., Reference Fernandez, Sheaffer, Wyse, Staley, Gould and Sadowsky2016; Lupwayi et al., Reference Lupwayi, Larney, Blackshaw, Kanashiro, Pearson and Petri2017; Tyler, Reference Tyler2020). The increased alkali–phosphatase enzyme activity might be attributed to the larger microbial communities in CC than no-CC plots as demonstrated in our medium-term CC experiment at the same site (Tosi et al., Reference Tosi, Drummelsmith, Obregón, Chahal, Van Eerd and Dunfield2022), and in the meta-analysis by Kim et al. (Reference Kim, Zabaloy, Guan and Villamil2020).
The main-plot mean tomato marketable yield ranged from 58.2 to 74.6 Mg ha−1 in 2019, which was near provincial average yields on production field (Trueman et al., Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023). In both years, tomato marketable yield in our study was not negatively impacted by the CC species. When comparing the CC species with the no-CC, none of the CC species reduced yield, all were statistically greater than the no-CC + straw retained +0 N treatment. This result was consistent with the previous research at the same long-term CC experiment by Belfry et al. (Reference Belfry, Trueman, Vyn, Loewen and Van Eerd2017), Chahal and Van Eerd (Reference Chahal and Van Eerd2021), and Trueman et al. (Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023) and confirms the role of CC in improving or maintaining crop productivity in the long term.
All the tomato fruit quality parameters were within the acceptable range of the tomato industry standards and were consistent with our previous research at the same site (Belfry et al., Reference Belfry, Trueman, Vyn, Loewen and Van Eerd2017; Chahal and Van Eerd, Reference Chahal and Van Eerd2021). The tomato fruit color readings were comparable to other studies (Garcia and Barrett, Reference Garcia and Barrett2006; Seliga and Shattuck, Reference Seliga and Shattuck1995; Warner, Zhang and Hao, Reference Warner, Zhang and Hao2004). The average natural tomato soluble solids (3.93°Brix) was comparable with other research (Lenzi et al., Reference Lenzi, Antichi, Bigongiali, Mazzoncini, Migliorini and Tesi2009; Belfry et al., Reference Belfry, Trueman, Vyn, Loewen and Van Eerd2017; Chahal and Van Eerd, Reference Chahal and Van Eerd2021), but was lower than the Ontario processing vegetable industry average of 4.76°Brix in 2019 (pers. comm. Tim Suitor, Highbury Canco Corp.). Cover crop differences in natural tomato soluble solids with no-CC and cereal rye being the lowest and radish being the greatest suggest the possible effects of long-term CCing on plant health and the allocation of soluble solids in the fruit and greater anthracnose disease on foliage and fruits in the no-CC (Trueman et al., Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023). Tomato fruit pH was not significantly different among CC treatments. Belfry et al. (Reference Belfry, Trueman, Vyn, Loewen and Van Eerd2017) reported that fruit pH may be influenced by tomato cultivar rather than the management strategies. Additionally, fewer number of CC studies on tomato fruit pH further limits our understanding of the CC induced effects on tomato fruit pH (Belfry et al., Reference Belfry, Trueman, Vyn, Loewen and Van Eerd2017; Chahal and Van Eerd, Reference Chahal and Van Eerd2021; Creamer et al., Reference Creamer, Bennett, Stinner and Cardina1996; Lenzi et al., Reference Lenzi, Antichi, Bigongiali, Mazzoncini, Migliorini and Tesi2009).
Cover crop induced effects on the tomato fruit phytochemical contents and antioxidant activity were observed, where the CC had either equal or greater phytochemical and antioxidant characteristics than the no-CC. It is possible that improvements in soil health in CC plots might have reduced the water stress to the tomato plants than the no-CC plots which perhaps would have led to less antioxidant activity with the long-term CC usage. While significant, caution should be used in interpreting the phytochemical and antioxidant data because these results might be confounded by fruit maturity. In 2019, no-CC control had significantly greater red fruit yields than the radish treatment (Trueman et al., Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023). As the tomato fruit matures and ripens, the chlorophyll breaks down and leads to the accumulation of carotenoids and lycopene. Lycopene is one of the most important tomato antioxidants (Brandt et al., Reference Brandt, Pék, Barna, Lugasi and Helyes2006). Another possible explanation is that production of plant secondary metabolites, such as phenolics, and the antioxidant activity is increased when the plant is under stress. In 2019, the severity of anthracnose fruit rot was significantly greater in no-CC plots than the other CC treatments (Trueman et al., Reference Trueman, Awrey, Delaporte, Kerr, Weersink and Van Eerd2023), which perhaps activated plant defense and contributed to an increased antioxidant activity in no-CC plots. Similar results were reported by Mikulic-Petkovsek et al. (Reference Mikulic-Petkovsek, Schmitzer, Jakopic, Cunja, Veberic, Munda and Stampar2013) where the infection by Colletotrichum coccodes increased the accumulation of total phenolic content in pepper fruit.
Our PCA results suggest that tomato marketable yield and fruit quality index were closely associated with the soil functions of nutrient supply and biodiversity conservation. This highlights the major role of soil nutrients, organic matter, and biological characteristics on the crop productivity, with positive implications of improving soil health on enhancing the plant health. Close association of crop yield with soil organic matter, total C, total N, microbial active C, ACE protein index confirms that increased concentration of soil C and N perhaps stimulates soil microbial activity, enhances soil C and N cycling, and increases the crop productivity.
With the entire soil health dataset, PC1 and PC2 explained a relatively high amount of the data variance (52.2%). For instance, at this site in the medium term, only 44.3% was explained (Chahal and Van Eerd, Reference Chahal and Van Eerd2019). As expected, when the 56 indicators were separated into soils functions more of the variability in the data were explained, except for erosion control. Interestingly, among the tested soil functions, the greatest variability in the soil health data was explained by the PC1 for biomass production (59.5%) and climate regulation (58.9%). This further confirms the strong and positive association of soil organic matter, respiration, soil C and N, and the tomato fruit quality parameters with tomato productivity in our study. Results of our study also demonstrated a positive association between crop yield and clay content. Soil having higher clay content can hold more water and nutrients required for plant growth than coarse textured soils which perhaps contribute to increasing the crop yield. This is particularly important for our study on sandy loam soil with only 13% clay. For instance, in a study by Jiang and Thelen (Reference Jiang and Thelen2004), clay content accounted for 32% of the variation in corn yield and 20% in soybean yield in a humid, temperate climate. Likewise, clay concentration was related to mean crop yield and accounted for 24% of the variability in crop yield in the long-term experiments at two dryland farming sites with sandy loam texture in eastern Montana, USA (Sainju, Liptzin and Jabro, Reference Sainju, Liptzin and Jabro2022). Negative association between crop yield and porosity was expected. Soil texture at the study site was sandy loam (coarse textured) which is characterized with low soil porosity. Low soil porosity was perhaps associated with less drought stress and contributed to an increase in crop yield (Sainju et al., Reference Sainju, Liptzin and Jabro2022). Further investigation is, however, needed to understand the underlying mechanisms and interactions with factors like nutrient availability, water retention, and root penetration as related to texture.
Conclusions
This study provided a unique opportunity to better understand the long-term effects of CC on soil health indicators, associated soil functions, tomato marketable yield, and quality, while gaining insights into the soil health drivers of crop yield. None of the planted CC had a negative effect on the tomato fruit yield and quality. Tomato delta marketable yield with CC was greater than the no-CC; therefore we reject the null hypothesis. Our results suggested that processing tomato grown succeeding long-term CC on fertile soil did not demonstrate human health benefits as there was not strong evidence that CC adoption would influence phytochemical contents and antioxidant activities of processing tomato.
In our study, 22 of 56 soil health indicators had greater values were with rather than without long-term CC, thus supporting our hypothesis. Our results confirmed a strong association of tomato yield with soil chemical and biological indicators (such as soil organic matter, ace protein index, total C and N, microbial C), mainly linked with the soil functions of nutrient supply, biodiversity conservation, and climate regulation. Although the exact mechanism is not clear, it is possible that enhanced soil C and N cycling and nutrient availability to the crop associated with high soil organic matter might have positively impacted the crop productivity.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S1742170525100148.
Acknowledgments
Authors express their gratitude to the technical support of Sean Vinkand funding agencies of Grain Farmers of Ontario and Ontario Ministry of Agriculture, Food, and Agribusiness, through the Ontario Agri-Food Innovation Alliance. We are grateful to the NAPESHM project for collecting the soil health data. The NAPESHM project is part of a broader effort titled, ‘Assessing and Expanding Soil Health for Production, Economic, and Environmental Benefits’. The project is funded by the Foundation for Food and Agriculture Research (grant ID 523926), General Mills, and The Samuel Roberts Noble Foundation. The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of the Foundation for Food and Agriculture Research nor other listed organizations.
Competing interests
The authors declare none.
Open access
