Experimental site, experimental design, and cultural practices

This research was performed at the Horticulture Field Research Station in Saskatoon, SK (52°8′N, 106°38′W; 515 m asl) on a Dark Brown Chernozem (Haplic Kastanozem) of the Sutherland clay association. The field trial was initiated in 2017 (Field A) and repeated on an adjacent field in 2018 (Field B). To characterize each field site, a composite soil sample (0–30 cm depth; made up of 7–9 individual cores) was collected using a Dutch auger (4-cm diameter) from across the entire plot area prior to establishing the trials. Field A and Field B, respectively, had a pH of 7.6 and 7.3, organic matter of 5.3% and 5.1%, cation exchange capacity of 29.2 and 31.6 m_eq 100 g⁻¹, Bray-P of 508 and 443 mg·kg⁻¹, extractable K of 1253 and 1461 mg·kg⁻¹, extractable S of 35 and 27 mg·kg⁻¹, nitrate-N of 89 and 93 kg·ha⁻¹, and ammonium-N of 17 and 15 kg N·ha⁻¹. Field A in the previous year produced a mixture of bell pepper (Capsicum annuum L.), eggplant (Solanum melongena L.) and pumpkin (Cucurbita pepo L.) grown on plastic mulch, whereas Field B in the previous year produced romaine lettuce (Lactuca sativa L. var. longifolia), head lettuce (Lactuca sativa L. var. capitata ‘Iceberg’), onion (Allium cepa L.), cabbage (Brassica oleracea L. var. capitata) and brussels sprouts (Brassica oleracea L. var. gemmifera). Both fields did not receive organic manure for at least 2 yr prior to our study. The temperature and precipitation conditions during the experimental period are shown in Table 1, relative to the 30-yr normal.

Table 1.

Weather data from 2017 to 2019 and 30-yr mean in Saskatoon, SK.^a

For the experiment, vegetables were produced according to a 3-yr annual sequence of broccoli (Brassica oleracea L. var. italica) – sweet corn (Zea mays var. saccharata) – root crop, where the root crop was carrot (Daucus carota L. subsp. sativus) for Field A, and garden beet (Beta vulgaris L. subsp. maritima) for Field B. We followed this sequential crop rotation for a multiyear study (from 2017 to 2019 for Field A; from 2018 to 2019 for Field B). All crop phases were present each year, where each crop type was produced in one of three sections of the field and arranged next to each other according to the rotation sequence. Each crop type section had a randomized complete block, split-plot design with three replicates. Each replicate had five main plots to test a range of N fertilizer rates (n = 5). The main plots were 3-m wide by 12-m long. The main-plots were subdivided into two split-plots (3 m by 6 m) to compare the cover crop vs. no cover crop (n = 2).

We tested five N fertilizer treatments that incrementally increased from no fertilizer (0-N control) to a high rate (i.e., 220–300 kg N·ha⁻¹ depending on the crop; Table 2). The fertilizer N source was granular urea, broadcast and incorporated into the soil via rototilling (10-cm depth) prior to planting. Shortly after urea application, broccoli (cv. Green magic) was mechanically transplanted (as 4-wk old transplants, which were produced in a greenhouse) on 30 May 2017, 29 May 2018, and 29 May 2019. Sweet corn (cv. Vision), carrot (cv. Bolero) and beet (cv. Merlin) were directly seeded on 25–28 May 2017, 26–29 May 2018, and 28 May 2019, respectively. Broccoli was planted with 50-cm between-row spacings and 30-cm in-row spacings. Sweet corn was seeded with 1-m between-row spacings and with 20-cm in-row spacings. Root crops were seeded with 50-cm between-row spacings and with 2-cm in-row spacings. During the growing season, the crops were irrigated with over-head irrigation on a weekly basis, receiving at least a total of 2.5 cm of irrigation and (or) precipitation per week.

Table 2.

Aboveground rye cover crop production and quality at the time of termination, as influenced by the amount of fertilizer N that was applied to vegetable crops the previous year.

Crops were harvested between late July and early September, depending on the species and weather (broccoli on 24 July 2017, 25 July 2018, and 22 July 2019; sweet corn on 28 Aug. 2017, 27 Aug. 2018, and 5 Sept. 2019; carrot on 14 Aug. 2017, 14 Aug. 2018, and 8 Aug. 2019; beet on 31 Aug. 2018, and 31 July 2019). After harvest samples were collected (as described below) the harvestable vegetable component was removed from the field. Then, the vegetable crop residues were flailed and rototilled into the soil to a depth of approximately 15–20 cm. Immediately after incorporating the vegetable crop residues each year (on 2 Aug. 2017 and 20 Aug. 2018 after broccoli; on 24 Aug. 2017 and 17 Aug. 2018 after carrot; and on 11 Sept. 2018 after beet), a shoulder-season rye cover crop was seeded. The rye seed was hand broadcast at a rate of 123 kg·ha⁻¹, the typical recommended rate for this region into one of the split-plots and the other split-plot was left unplanted. In the following spring, the rye cover crop was terminated before the booting stage around the middle of May each year (22 May 2018 and 16 May 2019) via an application of 1000 g a.e.·ha⁻¹ glyphosate. The fall rye was then incorporated into the soil by rototilling prior to fertilizing and planting the subsequent vegetable crop, according to the rotation. To ensure nutrients other than N were not limiting, pre-plant soil nutrient tests were performed and recommended rates of P and (or) S fertilizer applied as needed.

Plant and soil measurement and analysis

Vegetable yield data was collected by recording the fresh weight of the harvested portion collected from a 3-m row near the center of each plot: broccoli heads were cut approximately 4 cm from the top, ripe sweet corn cobs were clipped from stalks, and carrot or beet roots were lifted from the soil. Also, whole plants were collected by clipping three random plants at the soil surface or by digging up the root crops. The plant was separated into the harvestable portion and the crop residue portion. For these samples, fresh weights were recorded prior to oven drying at 60 °C for 48 h or until a constant dry weight was achieved, and dry weights recorded. Fall rye biomass was measured in April or May the year after it was planted by clipping biomass at the soil surface from within two 0.25 m² areas per split-plot. Biomass samples were weighed, dried, reweighed and ground (1 mm) prior to nutrient analysis. Plant tissue C and N concentrations was determined using a LECO CN628 elemental analyzer (LECO Instruments, Michigan, USA).

Soil samples were collected in the spring of each year before planting from 0–15 and 15–30 cm depths. Soils from three random cores for each split-plot were composited for each depth. Soil samples were homogenized and frozen (−10 °C) until analysis. Soil samples were thawed at 22 °C ± 2 for 4 h, sieved (<2 mm), and a sub-sample (∼5 g) analyzed for inorganic N (Maynard et al. 2008). For analysis, the extracts were thawed and a sub-sample (∼1 mL) analyzed for NO₃⁻ and NH₄⁺ concentrations, using an air segmented, continuous flow colorimetric method with a SEAL AA3 HR chemistry analyzer (SEAL analytical Kitchener Ontario). Soil mineral N content was converted to kg N·ha⁻¹ using the average bulk density measured by the cylinder method at each soil depth increment (Blake and Hartge 1986).

Nitrogen use efficiency calculations

The amount of N mineralized over the growing season was estimated from the unfertilized control treatments as follows (eq. 1):

(1)

in which N_min is the soil N mineralized in the 0–30 cm depth, N_res is the residual soil N at harvest in the 0–30 cm depth, N_plant is the N uptake by plants, and N_ini is the initial soil inorganic N prior to planting/seeding in the 0–30 cm depth. This provides an estimate of inorganic N available in the soil for plant uptake throughout the growing season.

Nitrogen use efficiency indices were calculated using plant dry weights in kg·ha⁻¹ for each crop and year, separately (Van Eerd 2007). Nitrogen content (kg N·ha⁻¹) in the plant (P_n) and yield (Y_n), was determined by multiplying N percentage by dry matter content. Soil mineral N content (S_n) as sum of NO₃⁻-N and NH₄⁺-N contents was calculated to 30-cm depth and the data converted to kg N·ha⁻¹ based on soil bulk density. Nitrogen harvest index (NHI, %) is the proportion of N in the harvested plant part (i.e., the head, ear, or root) relative to total plant N content (Van Eerd 2007), and is calculated as follows (eq. 2):

(2)

Nitrogen uptake efficiency (NUpE, %) is the proportion of N taken up by the crop, relative to soil available N (Van Eerd 2007), and is calculated as follows (eq. 3):

(3)

Apparent N recovery (ANR, %) is the proportion of N taken up by the crop yield, relative to soil available N (Van Eerd 2007), and is calculated as follows (eq. 4):

(4)

Statistical analysis

Prior to analyzing the data, the residuals were tested for normality and homogeneity using the Shapiro–Wilk and Bartlett tests, respectively (p > 0.05). Each crop, year, and field were considered as different experiments and analysed separately; data was subjected to an analysis of variance (ANOVA) using PROC MIXED in SAS (SAS version 9.1, SAS Institute Inc., Cary, NC, USA). To determine the impact of N fertilizer and the rye cover crop on plant and soil parameters, the ANOVA was conducted according to a split-plot design. The fixed effects were N level, rye cover crop, and their interaction. Replicates and the interaction between replication and the whole-plot factor (N levels) were set as random effects. Treatment means were compared using a Tukey’s HSD test (α = 0.05).

Weather conditions

Average monthly air temperature during the vegetable growing seasons generally tracked the 30-yr mean (Table 1). Monthly rainfall during the study period was considerably below the 30-yr mean on multiple occasions, for example from June–August in 2017, May–August in 2018, and March–May in 2019 (Table 1). However, a few months received more than normal rainfall, i.e., June and July of 2019 when rainfall was 17% and 39% higher than the 30-yr mean (Table 1). As common practice for vegetable production in semi-arid regions, irrigation was supplied to supplement low rainfall on an as-needed basis.

Rye cover crop biomass, C/N ratio and N content

By the time that the rye cover crop was terminated in the spring prior to planting the vegetable crops, it produced 506–2344 kg·ha⁻¹ of aboveground biomass dry matter (DM), depending on the year and position in the rotation (Table 2). Sometimes the fertilizer N rate did not affect rye DM but other times it did, for example, after all crop phases in Field B in 2019, and after sweet corn in Field A in 2019 (Table 2). For these cases, higher N rates tended to increase the rye DM accumulation (Table 2).

The C/N ratio of the rye cover crop ranged from 10.3 to 13.7 and was relatively stable across years, fields, and position in the rotation (Table 2). While differences in the C/N ratio were often not significant, there was a general tendency for the ratio to decrease with increasing N rate, when N fertilizer was added to the soil (Table 2).

The rye cover crop contained 18.5–101.5 kg N·ha⁻¹ prior to its termination — with variability between years, fields, and position in the rotation (Table 2). More often than not, the quantity of N contained in the rye biomass was influenced by N rate; generally, a higher amount of N was associated with higher N rates (Table 2).

Few N rate by cover crop interactions on crop and soil parameters

There were limited or no two-way interactions (N rate by cover crop) on crop productivity parameters and NUE indices ( Supplementary Table S1 (cjps-2021-0032suppla.docx) ¹[1]). Likewise, there were only a few N rate by cover crop interactions on soil inorganic N contents ( Supplementary Table S2 (cjps-2021-0032suppla.docx) ¹). Consequently, we hereafter focus on the main effects.

Cover crop effects

Rye cover crop maintained or increased vegetable crop productivity

The rye cover crop frequently influenced vegetable crop productivity ( Supplementary Table S1 (cjps-2021-0032suppla.docx) ¹). Vegetable crop yields were 10%–26% greater in the rotation with vs. without the cover crop (Table 3). There was only one case where the cover crop did not influence yield — that being sweet corn in 2018 (Table 3). Further, vegetable crop harvest indexes (HIs) were generally (by 4%–24%) higher in the rotation with vs. without the cover crop (Table 3).

Table 3.

Productivity of vegetable crops grown in rotation with and without shoulder-season rye cover crop (CC).

Rye cover crop increased vegetable crop N contents

The cover crop routinely influenced the amount of N contained in the vegetable crop yield and crop residue ( Supplementary Table S1 (cjps-2021-0032suppla.docx) ¹), resulting in greater N contents compared with the without cover crop plots (Fig. 1). Including the rye cover crop in the rotation increased the amount of N in broccoli yield by 27%–38%, sweet corn yield by 43%–49%, and root crop yield by 29%–49% (Fig. 1). Furthermore, the rye cover crop increased the amount of N in the vegetable crop residues by 12%–37% for broccoli, 7%–41% for sweet corn, and 14%–33% for root crops (Fig. 1).

Fig. 1.

The amount of N removed with yield and returned as crop residues of broccoli, sweet corn and root crop when grown in rotation with vs. without the shoulder season rye cover crop. For each site-year and each vegetable crop type, the means followed by different letters are significantly different at α = 0.05 (the top letters with black font apply to yield data; the bottom letters with white font apply to crop residue data).

Rye cover crop lowered soil inorganic N levels at vegetable planting but increased levels at harvest

The cover crop regularly influenced soil inorganic N contents at vegetable crop planting and at harvest ( Supplementary Table S2 (cjps-2021-0032suppla.docx) ¹). Prior to planting all vegetables, soil NO₃⁻-N and NH₄⁺-N contents were consistently lower with vs. without the rye cover crop in the rotation (Fig. 2). However, by the time that vegetables were harvested, soil NO₃⁻-N and NH₄⁺-N contents were higher with vs. without the rye cover crop in the rotation (Fig. 2).

Fig. 2.

Increase or decrease in soil nitrate-N and ammonium N as a result of growing a rye cover crop before the vegetable crop. Panel (A) and (B) are soil inorganic N values at the time of planting the veg crop, and (C) and (D) are soil inorganic N values after the vegetable crop was harvested. The 1:1 line represents no difference in nitrate or ammonium between vegetable grown with vs. without cover crop. The green, blue, and pink markers represent broccoli, sweet corn, and root crops, respectively. Different shapes represent different site-years: the squares, triangles, and circles correspond to Field A 2018, Field A 2019, and Field B 2019, respectively. Data for the soil depths of 0–15 and 15–30 cm are indicated by the solid and dotted markers, respectively.

Rye cover crop maintained or improved vegetable crop N use efficiency

The cover crop frequently increased vegetable crop NUE indices compared with no cover crop ( Supplementary Table S1 (cjps-2021-0032suppla.docx) ¹; Fig. 3). For broccoli, including the rye cover crop increased the crop NUpE by 17%–34%, the ANR by 15%–31%, and the NHI by 11%–30% (Fig. 3). Similar changes were observed for sweet corn where the rye cover crop increased the crop NUpE by 18%–24%, the ANR by 34%–49%, and the NHI by 28%–38% (Fig. 3). For root crops, the rye cover crop increased NUpE by 23%–24%, ANR by 20%–49%, and NHI by 9%–20% (Fig. 3).

Fig. 3.

The N uptake efficiency (NUpE), apparent N recovery (ANR) and N harvest index (NHI) for broccoli (A, D and G), sweet corn (B, E and H) and root crop (C, F and I) in rotation with and without the shoulder season rye cover crop. For each row and within each productivity measure, means followed by different letters are significantly different due to the cover crop effect, at α = 0.05.

N fertilizer rate responses

Increasing the N fertilizer rate did not consistently increase vegetable crop productivity

Since there were few/no N fertilizer rate by cover crop interactions, the results of N application rates are presented as an average response (with and without cover crops). Generally, vegetable crop yields increased with N application but not always in a consistent and incremental fashion (Table 4). For example, the point at which a higher N rate no longer increased yield was obtained when only a small amount of N was added to broccoli in Field B in 2019 (i.e., 75 kg N·ha⁻¹), when a medium amount of N was added to sweet corn in Field B in 2019 (i.e., 150 kg N·ha⁻¹), and when a high amount of N was added to broccoli in Field A in 2019 (i.e., 225 kg N·ha⁻¹). Application of any amount of N fertilizer reduced yield of sweet corn on Field A in 2018 (Table 4). The crop HI was frequently influenced by N rate but the effect was inconsistent across crop types, fields, and years in that sometimes higher N rates resulted in higher HI, and at other times lower HI (Table 4).

Table 4.

Mean crop yield and harvest index as affected by N fertilization.

Increasing the N fertilizer rate did not consistently increase vegetable crop N contents

After inclusion of cover crop, the N rate often impacted the quantity of N removed from the field in crop yields ( Supplementary Table S1 (cjps-2021-0032suppla.docx) ¹). In general, higher N rates corresponded to greater vegetable crop N contents, but the response tended to plateau prior to the highest N rate applied (Table 5). The amount of N returned to the soil as crop residues, where affected by N rate, ranged from 88 to 226 kg·ha⁻¹ for broccoli, 85 to 126 kg·ha⁻¹ for sweet corn, and 38 to 84 for carrot (Table 5).

Table 5.

Mean crop yield and residue N content as affected by N fertilization.

Increasing the N fertilizer rate increased the soil inorganic N levels at vegetable planting and harvest

After inclusion of cover crop, the N rate significantly affected soil NO₃⁻-N and NH₄⁺-N contents in the 0–15 and 15–30 cm depths prior to planting vegetable crops (Table 6). As a general pattern, higher soil NO₃⁻-N and NH₄⁺-N contents at planting corresponded to the treatments that received higher N fertilizer applications year after year (Table 6). At harvest, the N rate significantly influenced soil NO₃⁻-N and NH₄⁺-N contents (0–15 and 15–30 cm depths) in many cases — such that soil NO₃⁻-N and NH₄⁺-N contents were generally higher with higher amounts of N fertilizer (Table 6). On average, total amount of soil inorganic N in the top 0–30cm (0–15 + 15–30) that remained after harvest of broccoli, sweet corn and root crops were 73, 69 and 87 kg N·ha⁻¹, respectively (Table 6).

Table 6.

At planting and harvest, average soil NO₃⁻-N and NH₄⁺-N levels as affected by N fertilization.

Lower N fertilizer rates improved vegetable crop N uptake and recovery efficiency but reduced N harvest index

After the establishment of the cover crop, the vegetable crop NHI generally increased with N fertilizer rate, but not always in an incremental manner as the NHI tended to plateau prior to reaching the highest N rate (Table 7). The NUpE and ANR were significantly impacted by N rate in nearly all cases (Table 7). Overall, NUpE decreased with increasing N fertilization rate for all vegetable crops, except for broccoli in Field B, 2019 (Table 7). Similarly, ANR values decreased with increasing N rate for most of the cases in Field A, but just for one case in Field B (Table 7).

Table 7.

The N harvest index (NHI), N uptake efficiency (NUpE) and apparent N recovery (ANR) for sweet corn, broccoli, carrot and beet as affected by N fertilization.

Rye cover crop as part of a vegetable crop rotation in a cold climate

The rye cover crop was successfully established each fall prior to the onset of winter, and most of the growth occurred during the early fall and spring periods. However, the aboveground rye biomass observed in our study (conducted in a cold climate on the Canadian prairies) was generally low compared with other studies in more temperate climates and longer growing seasons (Clark et al. 1997; Feyereisen et al. 2006; West et al. 2020).

Cover crop maintained or improved vegetable crop yield and increased N use efficiency by influencing soil N cycling

The vegetable crop yields observed in our study were consistent with and somewhat higher than the provincial yield averages (Statistics Canada 2019). Vegetable crops grown after the rye cover crop had higher yields and HI compared with crops grown without a prior cover crop (Table 3). This might be explained by a closer synchrony between vegetable crop N uptake and N released from the rye biomass and the previous year’s vegetable crop residues, via mineralization during the early part of the growing season (Griffin et al. 2000; Restovich et al. 2012; Radicetti et al. 2017).

Cover cropping not only increased the amount of N in the vegetable crop yields but also in the crop residues, compared with the no cover crop control (Fig. 1). The N returned to the soil as vegetable crop residues at harvest can serve as a nutrient source for the next vegetable crop if it is not lost from the system before then — a period which can be 8–10 mo in cold climates such as the Canadian prairies. Due to low C/N ratios [8–13 for broccoli, 17–30 for sweet corn, 11–17 for carrot, and 7–11 for beet (data not shown)] the vegetable crop residues likely mineralized shortly after harvest in late summer and early fall. Other researchers have reported rapid mineralization of residues with similar C/N ratios to ours (13–25), when conditions were favourable for decomposition (Radicetti et al. 2016a, 2016b). Therefore, a large amount of available N would be present in the soil during the post-harvest period throughout the fall and winter. The previous year’s cover crops would have increased this amount by way of increasing the vegetable crop residue N contents. However, by repeatedly including the cover crop during the shoulder-season period, it may help retrieve NO₃⁻-N from the soil in the fall and subsequent spring. These processes facilitate the cycling of N within the soil-crop system year after year and reducing the risk of N loss (Griffin et al. 2000; Campiglia et al. 2014; Radicetti et al. 2017; Snapp and Surapur 2018). We recommend future research determines the threshold whereby biomass accumulation translates into reduced N losses by directly measuring the N losses.

The rye cover crop reduced soil inorganic N prior to planting of the vegetable crops (Fig. 2) indicating that it served as a catch crop, lowering the levels of environmentally reactive N in the soil during a key period — when the soil would otherwise be fallow and a thaw or leaching event would risk N loss prior to planting crops and crop N uptake (West et al. 2016). After vegetable crops were produced and harvested, a cover crop effect was still apparent, but with the opposite result in that more soil inorganic N remained in the soil in the rotation with the cover crop vs. without the cover crop (Fig. 2). This in-season rebound of soil inorganic N in the cover cropped rotation suggests a net release of N from the cover crop biomass (West et al. 2020). Net N mineralization during the early part of the growing season would have been favoured by the low C/N ratio (10.3–13.7) of the rye cover crop biomass (Table 2). Plausibly, the timing of N release benefitted the vegetable crop, due to the higher yields, N contents, and NUE indices observed with vs. without the cover crop (Table 3, Figs. 1 and 3). The amount of mineralized N in cover cropped plots was estimated as 43%–217% higher than the amount of N mineralized in the plots with no cover crop across fields, growing seasons and vegetable crops. It is possible that the N immobilized by the rye cover crop was re-mineralized during the vegetable growing season when the crops were actively taking up N. This would reduce the potential for further N leaching loss later in the growing season.

However, cover crops may not always synchronize soil N supply with crop demand and may not impart crop yield or NUE benefits (Restovich et al. 2012; Snapp and Surapur 2018; Wang et al. 2018). The outcome is regulated by the rate and timing of cover crop decomposition as it compares to crop N demand. Environmental benefits associated with improving crop NUE sometimes come with agronomic trade-offs (Pantoja et al. 2015; Martínez et al. 2017; West et al. 2020), but in our case, the rye cover crop did not negatively affect vegetable crop yields nor NUE; rather, increased vegetable yield and NUE.

A little goes a long way: lowering N fertilizer rates improve vegetable crop NUE and maintain yield

Depending on the site and year, crop yields did not increase with N rates above 75–225 kg N·ha⁻¹ for broccoli and sweet corn, and above 110–165 kg N·ha⁻¹ for root crops (Table 4). However, crop NUpE and ANR decreased with higher N rates (Table 7). Together, these results point towards unnecessary N application when fertilizer exceeds these thresholds. There were no significant increases in NHI at N rates higher than 75–150 kg N·ha⁻¹ for broccoli and sweet corn, and at N rates higher than 110 kg N·ha⁻¹ for root crops across growing seasons, whereas the NUpE and ANR values generally decreased with increasing N rates (Table 7). Like others (Zhang et al. 2011; Pantoja et al. 2015), our results show that the most efficient N use is achieved with relatively low fertilizer rates, and that luxury N consumption at higher N rates does not improve NUE. Although N supply is necessary to support vegetative growth and yield, increasing the N supply beyond the crop needs will not result in further yield increases, and may actually reduce yield (Long et al. 2004; Pokluda et al. 2018). Evidence of this effect was observed for sweet corn yields and HIs, and carrot HIs (Table 4). These reductions in yield and HI might be ascribed to excessive soil available N, resulting in more vegetative growth at the expense of yield, a delay in time taken to reach the maturity and luxury N uptake (Griffin et al. 2000; Westerveld et al. 2007; Yoldas et al. 2008). Our results indicate that the recommended N rates should compensate for high background soil N especially when cover crops are not used in the rotation.