A Multi-Criteria Evaluation of the Effectiveness of Nitrogen and Sulfur Fertilization in Different Cultivars of Winter Rapeseed—Productivity, Economic and Energy Balance

This article presents the results of a three-year experiment involving a multi-criteria evaluation (productivity, economic and energy balance) of the effectiveness of nitrogen (N) and sulfur (S) fertilization in different cultivars of winter oilseed rape (open-pollinated, semi-dwarf hybrid, long-stem hybrid) grown in north-eastern (NE) Poland. The yield of the semi-dwarf cultivar was 11% lower than the yield of the long-stem hybrid cultivar and 18% higher than the yield of the open-pollinated cultivar. In all cultivars, N fertilization improved yields up to a rate of 180 kg ha−1 and up to a rate of 230 kg ha−1 in years with low precipitation in spring and summer. Seed yield increased in all cultivars in response to S fertilization at 40 kg ha−1. Higher rates of N fertilizer decreased the content of crude fat and glucosinolates (GLS) and increased the concentration of total protein in all cultivars. Sulfur fertilization increased the content of total protein (in long-stem cultivars) and GLS (in all cultivars). Production costs ranged from €542–624 ha−1 (≤130 kg N ha−1) to €619–697 ha−1 (≥180 kg N ha−1). The demand for energy in the production of winter rapeseed ranged from 14.5–19.3 GJ ha−1 (≤130 kg N ha−1) to 22.4–27.0 GJ ha−1 (≥180 kg N ha−1).


Introduction
The global production of oilseed crops is estimated at 550 million Mg yr −1 . The leading oilseed crops in the world are soybeans (Glycine hispida (Moench) Maxim.) (63%), rapeseed (Brassica napus (L.) ssp. oleifera (Metzg)) (12%) and sunflower (Helianthus annuus L.) (9%) [1]. Rapeseed became a major oilseed crop mainly due to biological progress [2]. The discovery of genetic sources of variability in the main seed quality traits, that is, the content of erucic acid (EA) and glucosinolates (GLS), were the milestones that contributed to advances in rapeseed production and an increase in the area under rapeseed. The first EA-free variety was derived from the forage spring rapeseed cultivar Liho in 1961 [3]. In the late 1950s, Dr. Krzymański discovered that spring rapeseed cv. Bronowski contained ten-times less GLS than the remaining varieties of B. napus [4]. Owing to these breakthrough discoveries, contemporary rapeseed cultivars grown for the food and feed processing industries contain practically no EA in oil and are characterized by low levels of GLS in fat-free seed residues. These rapeseed lines are referred to as double-low, double-zero or canola-quality cultivars [5]. The production and commercialization of canola-quality rapeseed has led to a rapid increase in the demand for these cultivars in the food and feed processing industries [6]. In the last 15 years (2003-2017), the area under Oilseed rape, in particular winter cultivars, requires intensive farming and its yields are strongly determined by agricultural inputs [98,99]. Production costs, yields and, consequently, the economic performance of rapeseed are determined by the prices of seeds and agricultural inputs [100][101][102] as well as the applied production technology [103]. High-input technologies are characterized by higher input costs but also higher productivity. In Europe, rapeseed production costs in high-input technologies range from €500 ha −1 to even €1000 ha −1 . Production costs are around 12% lower in integrated technologies [104] and up to 40% lower in low-input technologies [105]. The economic performance of winter oilseed rape is also largely determined by genetic factors (cultivar) because the yields of various cultivars can differ by 0.7-1.1 Mg ha −1 [21,106] to 1.8 Mg ha −1 [9].
The production of agricultural biomass is a strategic sector from the perspective of global food security, which remains to be the main but no longer the only priority in agricultural production. Biomaterials are also being converted to bioproducts and bioenergy to counteract the threats to global energy security [107][108][109][110][111][112][113][114][115]. Environmental problems and the rising cost of fossil fuels have increased the significance of research into the energy efficiency of agricultural crops. Efforts are being made to improve the energy balance of biomass [116]. Energy consumption in agriculture is directly linked with technological progress. High-input production technologies have increased the availability of food but have also made the agricultural sector more dependent on fossil fuels [117]. Rapeseed production in high-input systems entails higher energy consumption and lower energy efficiency but higher energy output per unit area [108]. Therefore, production systems characterized by different levels of agricultural inputs should be evaluated not only in terms of agricultural productivity (yield and yield quality) and economic performance but also in terms of energy efficiency [108]. A multi-criteria approach can be applied to evaluate the effectiveness of various technologies in the production of new crop cultivars.
The objective of this study was to identify N and S fertilization systems characterized by the highest agricultural productivity, economic performance and energy efficiency in the production of various cultivars of winter oilseed rape (open-pollinated, long-stem hybrid and semi-dwarf hybrid) in north-eastern (NE) Poland. The effect of nitrogen and sulfur fertilization on the biomass yield, harvest index, nitrogen use efficiency, marginal nitrogen use efficiency, seed quality (crude fat content, total protein content, fatty acid composition, lower heating value, qualitative and quantitative composition, GLS content), production value, revenue per hectare, profitability index, energy inputs, energy output, energy gain and the energy efficiency ratio of various winter oilseed rape cultivars was also evaluated.
Each year, the experiment was established on Haplic Luvisol (LV-ha) developed from boulder clay [119]. In order to determine the chemical properties of soil, composite samples consisting of 8 to 10 cores were collected annually (before fertilization and sowing) from each plot to a depth of 0-20 cm (Table 1).

Plant Materials
The biomass yield of winter rapeseed (seeds and straw) from each plot was determined by weight after threshing and it was expressed in DM per hectare, taking into account the moisture content of the samples after oven-drying at 105 • C (FD 53 Binder GmbH, Tuttlingen, Germany). The harvest index (HI) was calculated with the following formula (Equation (1)): Seed and straw yield (Mg DM ha −1 ) .
The below formula was used to calculate nitrogen use efficiency (NUE) (Equation (2)) [8]: Marginal nitrogen use efficiency was calculated with the below formula (Equation (3)) [121] where: MNUE-Marginal nitrogen use efficiency (kg seeds per 1 kg N) ∆Y-increase in seed yield (kg) for the applied nitrogen rates ∆R N -increase in nitrogen rate (kg).

Seed Quality Analysis
The procedure for determining the concentrations of crude fat, total protein and GLS in winter oilseed rape seeds and the FA profile of oil was described previously by Jankowski et al. [122].

Economic Analysis
The analysis of rapeseed production costs included the costs associated with labor, seeds, fertilizers, crop protection, the operation of tractors and farming machines and fuel. The costs of input materials and the operation of tractors and agricultural machines were estimated based on actual consumption, market prices (4Q 2019) and the technical parameters of agricultural machines (measured by the authors, Table 2) and labor (€4.60 h −1 ). The economic efficiency of winter oilseed rape production was determined based on the following indicators-production value, revenue per hectare and the profitability index [120].

Energy Analysis
The relevant data were obtained based on direct measurements of diesel oil consumption, labor and the field capacity of farming machines and equipment in the Agricultural Experiment Station in Bałcyny for standard agricultural operations in large-area farms ( Table 2). The energy equivalents of the agricultural inputs were determined based on literature data [123][124][125].
The lower heating value (LHV), energy value of seeds (energy output), energy gain and the energy efficiency ratio were calculated based on the methods described by Sokólski et al. [120].

Statistical Analysis
The data relating to seed and straw yield, nitrogen use efficiency, harvest index and seed quality traits were processed by split-split-plot mixed model variance, where cultivar, N and S fertilization were the fixed effects and experimental years and replications nested within years were the random effects. Post-hoc multiple comparisons were performed using Tukey's test (HSD). The results were considered statistically significant at α = 0.05. All analyses were conducted with the use of Statistica ver. 13.3 software (Tibco Software Inc., Palo Alto, USA) [126]. The F-values of ANOVA are presented in Table 3.   Table 3. F-test statistics in ANOVA. Total precipitation during the autumn growing season approximated the long-term average. Mild winters in the first and third year of the study (2011/2012 and 2013/2014) contributed to overwintering success. Below-zero temperatures (−1.9 • C to −7.2 • C) were noted only in January and February (year 1) or only in January (year 3). In the second year of the experiment (2012/2013), below−zero temperatures were observed during the entire winter dormancy period, that is, in December (−3.4 • C), January (−4.4 • C), February (−0.8 • C) and March (−4.0 • C). The amount and distribution of precipitation was most favorable for the spring and summer growth of winter oilseed rape in year 1 (2012). In the first year of the study, precipitation was determined at 307 mm between April and July and it was 25% higher than the long-term average. Precipitation exceeded the long-term average by 32-50% in April (45 mm), June (107 mm) and July (112 mm) of 2012. In the second year of the experiment (2013), spring precipitation exceeded the long-term average by 13% but its distribution was less favorable. Most rainfall (approx. 59%) occurred only during ripening (July), whereas precipitation in the preceding months (April-June) was below the long-term average (by 22-38%). In the third year of the study (2014), total precipitation during the spring growing season amounted to only 153 mm and it was equivalent to 62% of the long-term average. Rainfall levels approximating the long-term average were noted only in June. The average daily temperature in spring and summer was similar to (years 1 and 2) or somewhat higher (year 3) than the long-term average (Table 4).

Biomass Yield and the Harvest Index
The yields of winter oilseed rape were stable and relatively high during the experiment (6.22−6.36 Mg ha −1 ). The seed yield of the semi−dwarf cultivar (PR44 D06) was higher (by 0.48 Mg ha −1 ) than the seed yield of the open−pollinated cultivar (Adriana) and lower (by 0.79 Mg ha −1 ) than the seed yield of the long−stem hybrid cultivar (Artoga) ( Table 5). The yield of the long−stem hybrid cultivar was particularly high (7.96 Mg ha −1 ) in the first year of the study, when rainfall was abundant and evenly distributed during the spring and summer growing seasons ( Figure 1). In all studied cultivars, N fertilizer rate of up to 180 kg ha −1 exerted a yield−forming effect ( Table 5). The yield−forming effect of higher N rates (up to 230 kg ha −1 ) was observed only in years with lower precipitation in the growing season (years 2 and 3) ( Figure 2). The application of S at 40 kg ha −1 led to a small (3%) but significant increase in seed yield in all analyzed cultivars ( Table 5). The genetic factor (cultivar) did not affect the effectiveness of S fertilizer. No significant interaction of N × S fertilization was found for this factor (Table 3). Nitrogen use efficiency was 10-24% higher in long-stem (Artoga) and semi-dwarf (PR D06) hybrid cultivars than in the open-pollinated cultivar (Adriana) and 12% higher in the long-stem hybrid cultivar than in the semi-dwarf hybrid cultivar (Table 5). In the long-stem hybrid cultivar, NUE was particularly high in the first year of the study, characterized by above-average precipitation in the spring and summer growing seasons ( Figure 3). The NUE of winter oilseed rape decreased with a rise in N rate from 80 kg ha −1 to 230 kg ha −1 (69.4 kg seeds per 1 kg N vs. 29.4 kg seeds per 1 kg N) ( Table 5). The above decrease was particularly pronounced in hybrid cultivars (Artoga, PR D06), regardless of their growth habit (indeterminate, determinate) ( Figure 4). cultivars, N fertilizer rate of up to 180 kg ha −1 exerted a yield−forming effect ( Table 5). The yield−forming effect of higher N rates (up to 230 kg ha −1 ) was observed only in years with lower precipitation in the growing season (years 2 and 3) ( Figure 2). The application of S at 40 kg ha −1 led to a small (3%) but significant increase in seed yield in all analyzed cultivars ( Table 5). The genetic factor (cultivar) did not affect the effectiveness of S fertilizer. No significant interaction of N × S fertilization was found for this factor (Table 3).   cultivars, N fertilizer rate of up to 180 kg ha −1 exerted a yield−forming effect ( Table 5). The yield−forming effect of higher N rates (up to 230 kg ha −1 ) was observed only in years with lower precipitation in the growing season (years 2 and 3) ( Figure 2). The application of S at 40 kg ha −1 led to a small (3%) but significant increase in seed yield in all analyzed cultivars ( Table 5). The genetic factor (cultivar) did not affect the effectiveness of S fertilizer. No significant interaction of N × S fertilization was found for this factor (Table 3).   Nitrogen use efficiency was 10-24% higher in long-stem (Artoga) and semi-dwarf (PR D06) hybrid cultivars than in the open-pollinated cultivar (Adriana) and 12% higher in the long-stem hybrid cultivar than in the semi-dwarf hybrid cultivar (Table 5). In the long-stem hybrid cultivar, NUE was particularly high in the first year of the study, characterized by above-average precipitation in the spring and summer growing seasons ( Figure 3). The NUE of winter oilseed rape decreased with a rise in N rate from 80 kg ha −1 to 230 kg ha −1 (69.4 kg seeds per 1 kg N vs. 29.4 kg seeds per 1 kg N) ( Table 5). The above decrease was particularly pronounced in hybrid cultivars (Artoga, PR D06), regardless of their growth habit (indeterminate, determinate) ( Figure 4).  Regardless of cultivar, N rates higher than 180 kg ha −1 exerted a yield-forming effect only in combination with soil-applied S. In treatments supplied with high rates of N (180-230 kg N ha −1 ), an increase in the N rate by 1 kg without supplementary S fertilization was not productive in any of the studied cultivars ( Figure 5).  Regardless of cultivar, N rates higher than 180 kg ha −1 exerted a yield-forming effect only in combination with soil-applied S. In treatments supplied with high rates of N (180-230 kg N ha −1 ), an increase in the N rate by 1 kg without supplementary S fertilization was not productive in any of the studied cultivars ( Figure 5). As expected, straw yield was lowest (10.12 Mg ha −1 ) in the semi-dwarf hybrid cultivar, moderate (10.54 Mg ha −1 ) in the open-pollinated cultivar and highest (11.80 Mg ha −1 ) in the long-stem hybrid cultivar ( Table 5). The straw yield of the long-stem hybrid cultivar was particularly high (11.92 Mg ha −1 ) in the first year of the experiment with above-average precipitation and a favorable rainfall pattern during the spring and summer growing seasons ( Figure 6). Straw yield continued to increase up to the N rate of 130 kg N ha −1 ( Table 5). Higher N rates (up to 230 kg N ha −1 ) ( Figure 7) increased straw yield only when the water requirements of plants were well met (year 1). Sulfur fertilization had no influence on straw yield (Table 3).  Table 5). The straw yield of the long-stem hybrid cultivar was particularly high (11.92 Mg ha −1 ) in the first year of the experiment with above-average precipitation and a favorable rainfall pattern during the spring and summer growing seasons ( Figure 6). Straw yield continued to increase up to the N rate of 130 kg N ha −1 (Table 5). Higher N rates (up to 230 kg N ha −1 ) ( Figure 7) increased straw yield only when the water requirements of plants were well met (year 1). Sulfur fertilization had no influence on straw yield ( Table 3).
The ratio of seed yield to total biomass yield (harvest index) was higher in the semi-dwarf cultivar (0.347) than in long-stem cultivars (0.321-0.339). The harvest index of winter rapeseed increased up to the N rate of 180 kg ha −1 ( Table 5). The influence of N fertilization on the harvest index was determined by weather conditions (Y × N interaction) and cultivar (Cv. × N interaction) ( Table 3). In the dry year (year 3), an increase in N rates up to 230 kg ha −1 contributed to an increase in the harvest index ( Figure 8). In the long-stem hybrid cultivar, the ratio of seed yield to straw yield was most desirable at the N rate of 130 kg ha −1 . In the open-pollinated cultivar and the semi-dwarf hybrid Energies 2020, 13, 4654 13 of 36 cultivar, the highest value of the harvest index was achieved at the N rate of 180 kg N ha −1 (Figure 9). The harvest index was not significantly affected by S fertilization (Table 3).
As expected, straw yield was lowest (10.12 Mg ha −1 ) in the semi-dwarf hybrid cultivar, moderate (10.54 Mg ha −1 ) in the open-pollinated cultivar and highest (11.80 Mg ha −1 ) in the long-stem hybrid cultivar ( Table 5). The straw yield of the long-stem hybrid cultivar was particularly high (11.92 Mg ha −1 ) in the first year of the experiment with above-average precipitation and a favorable rainfall pattern during the spring and summer growing seasons ( Figure 6). Straw yield continued to increase up to the N rate of 130 kg N ha −1 (Table 5). Higher N rates (up to 230 kg N ha −1 ) (Figure 7) increased straw yield only when the water requirements of plants were well met (year 1). Sulfur fertilization had no influence on straw yield (Table 3).  The ratio of seed yield to total biomass yield (harvest index) was higher in the semi-dwarf cultivar (0.347) than in long-stem cultivars (0.321-0.339). The harvest index of winter rapeseed increased up to the N rate of 180 kg ha −1 ( Table 5). The influence of N fertilization on the harvest index was determined by weather conditions (Y × N interaction) and cultivar (Cv. × N interaction) ( Table  3). In the dry year (year 3), an increase in N rates up to 230 kg ha −1 contributed to an increase in the harvest index ( Figure 8). In the long-stem hybrid cultivar, the ratio of seed yield to straw yield was most desirable at the N rate of 130 kg ha −1 . In the open-pollinated cultivar and the semi-dwarf hybrid cultivar, the highest value of the harvest index was achieved at the N rate of 180 kg N ha −1 (Figure 9). The harvest index was not significantly affected by S fertilization (Table 3).  The ratio of seed yield to total biomass yield (harvest index) was higher in the semi-dwarf cultivar (0.347) than in long-stem cultivars (0.321-0.339). The harvest index of winter rapeseed increased up to the N rate of 180 kg ha −1 ( Table 5). The influence of N fertilization on the harvest index was determined by weather conditions (Y × N interaction) and cultivar (Cv. × N interaction) ( Table  3). In the dry year (year 3), an increase in N rates up to 230 kg ha −1 contributed to an increase in the harvest index ( Figure 8). In the long-stem hybrid cultivar, the ratio of seed yield to straw yield was most desirable at the N rate of 130 kg ha −1 . In the open-pollinated cultivar and the semi-dwarf hybrid cultivar, the highest value of the harvest index was achieved at the N rate of 180 kg N ha −1 (Figure 9). The harvest index was not significantly affected by S fertilization (Table 3).

Seed Quality
The crude fat content of seeds was higher in the open-pollinated cultivar and in the semi-dwarf hybrid cultivar (506-509 g crude fat kg −1 DM seeds). An increase in N rate decreased (by 2-5%) the crude fat content of seeds (Table 6) and the greatest decrease (5%) was noted in the semi-dwarf hybrid cultivar ( Figure 10). The adverse effect of N fertilization on crude fat concentration was less pronounced when precipitation was low during the spring and summer growing seasons (year 3) ( Figure 11). Spring S fertilization had no influence on the crude fat content of seeds in any of the analyzed cultivars (Table 3). The concentrations of major fatty acids (FAs) in the oil of winter oilseed rape reached 60-64% C 18:1 , 19-23% C 18:2 and 10-11% C 18:3 . The long-stem hybrid cultivar (Artoga) was least abundant in C 18:1 (59.5%) and most abundant in C 18:2 (22.5%) and C 18:3 (11.4%). Nitrogen fertilization contributed to an increase in C 18:1 (up to 130 kg N ha −1 ) as well as C 18:2 and C 18:3 (up to 180 kg N ha −1 ) levels ( Table 6). The content of major FAs in the oil of winter oilseed rape was not significantly affected by S fertilization (Table 3).
denotes the non-significance of main effects.

Seed Quality
The crude fat content of seeds was higher in the open-pollinated cultivar and in the semi-dwarf hybrid cultivar (506-509 g crude fat kg −1 DM seeds). An increase in N rate decreased (by 2-5%) the crude fat content of seeds (Table 6) and the greatest decrease (5%) was noted in the semi-dwarf hybrid cultivar ( Figure 10). The adverse effect of N fertilization on crude fat concentration was less pronounced when precipitation was low during the spring and summer growing seasons (year 3) ( Figure 11). Spring S fertilization had no influence on the crude fat content of seeds in any of the analyzed cultivars (Table 3).  The concentrations of major fatty acids (FAs) in the oil of winter oilseed rape reached 60-64% C18:1, 19-23% C18:2 and 10-11% C18:3. The long-stem hybrid cultivar (Artoga) was least abundant in C18:1 (59.5%) and most abundant in C18:2 (22.5%) and C18:3 (11.4%). Nitrogen fertilization contributed to an increase in C18:1 (up to 130 kg N ha −1 ) as well as C18:2 and C18:3 (up to 180 kg N ha −1 ) levels ( Table 6). The content of major FAs in the oil of winter oilseed rape was not significantly affected by S fertilization ( Table 3).
The semi-dwarf hybrid cultivar (PR44 D06) accumulated significantly less total protein (by 7.6-8.8 g kg −1 DM) than long-stem cultivars (Adriana and Artoga). Nitrogen applied in spring up to the rate of 180 kg ha −1 increased total protein concentration in the seeds of winter oilseed rape by 6% ( Table 6). The effect of N and S fertilization on total protein content varied across cultivars (Cv. × N and Cv. × S interactions) ( Table 3). Nitrogen contributed to a greater increase in the total protein content of seeds in hybrid cultivars (13%) than in the open-pollinated cultivar (5%) (Figure 12). Protein synthesis increased (by 3%) under the influence of S fertilization only in long-stem cultivars (Adriana and Artoga). In the semi-dwarf hybrid cultivar (PR44 D06), S fertilization decreased total protein concentration by 2% (Figure 12). The semi-dwarf hybrid cultivar (PR44 D06) accumulated significantly less total protein (by 7.6-8.8 g kg −1 DM) than long-stem cultivars (Adriana and Artoga). Nitrogen applied in spring up to the rate of 180 kg ha −1 increased total protein concentration in the seeds of winter oilseed rape by 6% ( Table 6). The effect of N and S fertilization on total protein content varied across cultivars (Cv. × N and Cv. × S interactions) ( Table 3). Nitrogen contributed to a greater increase in the total protein content of seeds in hybrid cultivars (13%) than in the open-pollinated cultivar (5%) (Figure 12). Protein synthesis increased (by 3%) under the influence of S fertilization only in long-stem cultivars (Adriana and Artoga). In the semi-dwarf hybrid cultivar (PR44 D06), S fertilization decreased total protein concentration by 2% (Figure 12).
Long-stem cultivars (Adriana and Artoga) accumulated 34-39% more GLS than the semi-dwarf hybrid cultivar (PR44 D06) ( Table 6). In long-stem cultivars, N rates up to 180 kg ha −1 decreased GLS content by 7-8%, whereas higher N rates increased GLS content by 7-11%. The GLS content of seeds in the semi-dwarf hybrid cultivar was not significantly modified by N fertilization (Figure 13). Sulfur fertilization rates of 40 kg ha −1 and 80 kg ha −1 increased GLS levels by 19% and 29%, respectively ( Table 6). The GLS content of seeds in long-stem cultivars increased up to the S rate of 80 kg S ha −1 and it was 1.4-fold higher in the hybrid cultivar than in the open-pollinated cultivar (2.6 µM g −1 DM vs. 3.7 µM g −1 DM). In the semi-dwarf cultivar, GLS levels continued to increase up to the S rate of Energies 2020, 13, 4654 16 of 36 40 kg S ha −1 and were 30% lower than in long-stem cultivars ( Figure 13). A significant interaction of N × S fertilization was found for the GLS content of winter oilseed rape seeds ( Table 3). The S-induced increase in GLS levels was enhanced by growing rates of N fertilization ( Figure 14). Therefore, N fertilization exacerbated the negative effects of S fertilization on the GLS content of all studied cultivars.
The content of major FAs in the oil of winter oilseed rape was not significantly affected by S fertilization ( Table 3).
The semi-dwarf hybrid cultivar (PR44 D06) accumulated significantly less total protein (by 7.6-8.8 g kg −1 DM) than long-stem cultivars (Adriana and Artoga). Nitrogen applied in spring up to the rate of 180 kg ha −1 increased total protein concentration in the seeds of winter oilseed rape by 6% ( Table 6). The effect of N and S fertilization on total protein content varied across cultivars (Cv. × N and Cv. × S interactions) ( Table 3). Nitrogen contributed to a greater increase in the total protein content of seeds in hybrid cultivars (13%) than in the open-pollinated cultivar (5%) (Figure 12). Protein synthesis increased (by 3%) under the influence of S fertilization only in long-stem cultivars (Adriana and Artoga). In the semi-dwarf hybrid cultivar (PR44 D06), S fertilization decreased total protein concentration by 2% (Figure 12).

Figure 12.
The effect of N and sulfur (S) rates on the total protein content of winter oilseed rape seeds, across years. Means with the same letters do not differ significantly at p ≤ 0.05 in Tukey`s test.
Long-stem cultivars (Adriana and Artoga) accumulated 34-39% more GLS than the semi-dwarf hybrid cultivar (PR44 D06) ( Table 6). In long-stem cultivars, N rates up to 180 kg ha −1 decreased GLS content by 7-8%, whereas higher N rates increased GLS content by 7-11%. The GLS content of seeds in the semi-dwarf hybrid cultivar was not significantly modified by N fertilization (Figure 13). Sulfur fertilization rates of 40 kg ha −1 and 80 kg ha −1 increased GLS levels by 19% and 29%, respectively ( Table 6). The GLS content of seeds in long-stem cultivars increased up to the S rate of 80 kg S ha −1 and it was 1.4-fold higher in the hybrid cultivar than in the open-pollinated cultivar (2.6 μM g −1 DM    The content of alkenyl GLS in total GLS ranged from 60% in the semi-dwarf cultivar to 74-75% in long-stem cultivars. Long-stem cultivars accumulated 66-75% more alkenyl GLS than the semi-dwarf cultivar (9.3-9.8 µM g −1 DM vs. 5.6 µM g −1 DM). An increase in the N fertilizer rate to 180 kg ha −1 decreased the content of alkenyl GLS by 8% in all cultivars of winter oilseed rape (Table 6). An increase in the N fertilizer rate from 180 kg ha −1 to 230 kg ha −1 increased (by 6-11%) the content of alkenyl GLS in long-stem cultivars and induced a further drop (by 3-4%) in the analyzed parameter in the semi-dwarf cultivar ( Figure 15). Sulfur fertilization rates of 40 kg ha −1 and 80 kg ha −1 increased the content of alkenyl GLS by 28% and 40%, respectively ( Table 6). The increase in the S fertilizer rate from 40 kg ha −1 to 80 kg ha −1 did not cause an increase in alkenyl GLS levels only in the semi-dwarf cultivar ( Figure 15). The S-induced increase in the concentrations of alkenyl GLS was further exacerbated by a rise in N rates in all cultivars ( Figure 16).  The lower heating value (LHV) of winter rapeseed was weakly modified by the experimental factors. The above parameter was marginally higher (by 2%) in years with lower precipitation during the spring and summer growing seasons (years 2 and 3). Hybrid (long-stem and semi-dwarf) cultivars were characterized by significantly higher LHV (by 1%) than the open-pollinated cultivar ( Table 6). Nitrogen and S fertilization did not differentiate LHV values in any of the studied cultivars (Table 3).

Economic Efficiency of Winter Rapeseed Production
The costs associated with the production of the open-pollinated cultivar ranged from €542-599 (≤130 kg N ha −1 ) to €619-672 ha −1 (≥180 kg N ha −1 ). Sulfur fertilization increased the production costs of the open-pollinated cultivar by up to ±3%. In hybrid cultivars, production costs were €25-28 (long-stem) and €21-25 ha −1 (semi-dwarf) higher on average than in the open-pollinated cultivar. The difference (3-5% on average) between the production costs of hybrid cultivars and the open-pollinated genotype had no significant effect on cost structure. Mineral fertilizers accounted for 37-43% (≤130 kg N ha −1 ) to 47-49% (≥180 kg N ha −1 ) of total production costs in all cultivars (Tables 7-10).
All cultivars of winter oilseed rape generated revenues and high profits in NE Poland. The long-stem hybrid cultivar (Artoga) was most economically effective, followed by the semi-dwarf hybrid cultivar (PR44 D06) and the open-pollinated cultivar (Adriana) ( Table 10). In each of the analyzed cultivars, seed production value and revenue were highest in the production technology with fertilizer rates of 230 kg N ha −1 and 40 kg S ha −1 . The influence of N and S fertilizer rates varied across cultivars. The production technology involving 180 kg N ha −1 and 40 kg S ha −1 maximized the profitability of the open-pollinated cultivar (358%) and the semi-dwarf hybrid cultivar (368%). In the long-stem hybrid cultivar (Artoga), profitability peaked (419%) in response to a lower N rate (130 kg N ha −1 ) supplemented with 40 kg S ha −1 (Table 10).

Energy Efficiency of Winter Rapeseed Production
The demand for energy in the production of winter rapeseed ranged from 14.5-19.3 (≤130 kg N ha −1 ) to 22.4-27.0 GJ ha −1 (≥180 kg N ha −1 ) (Tables 11-13). The energy demand of all winter oilseed rape cultivars increased by 26% (130 kg N ha −1 ), 52-53% (180 kg N ha −1 ) and 78-79% (230 kg N ha −1 ) when N fertilizer was applied at rates higher than 80 kg ha −1 . Cultivar differentiated energy demand by up to ±0.5% in all production technologies. The variation in energy demand in response to S fertilization did not exceed ±4%. Mineral fertilization accounted for 75-86% of total energy inputs in the production of the evaluated winter rapeseed cultivars. The proportions of the remaining agronomic operations in the structure of energy inputs were determined at-tillage-6-11%, harvest-4-8%, weed and pest control-3-5% and sowing-1-2% (Tables 11-13).
The energy output and energy gain in the production of open-pollinated and semi-dwarf cultivars were highest in the production technology where N and S fertilizers were applied at 230 kg N ha −1 and 40 kg S ha −1 (Table 14). In the production of the long-stem hybrid cultivar, energy output peaked in response to the highest fertilization levels (230 kg N ha −1 , 80 kg S ha −1 ), whereas the highest energy gain was noted in response to the lowest N rate (180 kg N ha −1 ) without S fertilization. In all tested cultivars, energy efficiency was highest in the low-input production technology with an N fertilizer rate of 80 kg N ha −1 and no S fertilization (Table 14). All winter oilseed rape cultivars were characterized by a positive energy balance and high energy efficiency. However, energy output (173.2 GJ ha −1 ), energy gain (149.0 GJ ha −1 ) and energy efficiency (10.14) were highest in the long-stem hybrid cultivar. The average energy efficiency of the semi-dwarf hybrid cultivar and the open-pollinated cultivar was 10% and 18% lower, respectively (Table 14).    373  373  373  373  373  373  373  373  373  373  373  373  Pest control  324  324  324  324  324  324  324  324  324  324  324  324  Harvest and transport  1107  1132  1087  1132  1157  1157  1201  1201  1201  1201  1201  1201  Table 13. Energy inputs in the production of semi-dwarf hybrid cultivar of winter oilseed rape (cv. PR44 D06), across years.

Biomass Yield and the Harvest Index
Rapeseed cultivars represent genetic progress and they are the most important and the cheapest factors in the crop production technology. The agricultural performance of winter oilseed rape cultivars differs significantly. In Europe, the yields of hybrid cultivars are approximately 7-9% [8] or even 13% higher compared with the yields of open-pollinated cultivars [9]. The introduction of cultivars resistant to clubroot [127] and imazamox (CL) [128], and the development of semi-dwarf cultivars of winter oilseed rape [129], were important milestones in agricultural production in the last decade. The yields of semi-dwarf cultivars of winter oilseed rape are generally 1-7% [9,10,130,131] to 9-18% [10,130,132,133] lower in comparison with cultivars with a conventional growth habit. In semi-dwarf cultivars, lower yields result from a smaller (by 24-31%) number of seeds per silique [132] or a smaller (by 6-18%) number of seeds per silique and lower (by 3-6%) 1000-seed weight [10,134]. The yields of semi-dwarf cultivars approximate those noted in morphotypes with a conventional growth habit under conditions that do not support the achievement of their full yield potential [21] and under unfavorable thermal conditions in winter [10]. In the present experiment conducted in NE Poland, weather conditions promoted high and stable yields (6.2-6.4 Mg ha −1 ). Under favorable conditions, the yield of the semi-dwarf hybrid cultivar was 11% lower (0.79 Mg ha −1 ) than in the long-stem hybrid cultivar but 8% higher (0.48 Mg ha −1 ) than in the open-pollinated cultivar. Similar relationships between the yields of open-pollinated and hybrid cultivars with an indeterminate and determinate growth habit were reported by Gugała et al. [133] in central-eastern Poland.
Semi-dwarf cultivars of winter rapeseed are less susceptible to lodging and are characterized by a higher harvest index [23]. In a study conducted by Miersch et al. [23] in Central Germany (Lower Saxony), the seed yield to straw yield ratio (harvest index) of 108 hybrids (54 normal-type and 54 semi-dwarf hybrids) was significantly higher in semi-dwarf hybrids (0.410-0.460) than in normal-type hybrids (0.340-0.410). In the present experiment, the harvest index was also higher in the semi-dwarf cultivar (0.347) than in long-stem cultivars (0.321-0.339). Contrary results were reported in a study by Jankowski et al. [10], where semi-dwarf cultivars produced shorter but more branched stems and formed a higher number of productive siliques than the long-stem cultivar [10]. The harvest index was also higher in long-stem cultivars than in semi-dwarf cultivars of winter oilseed rape in the work of Gugała et al. [133] and Dresbøll et al. [135].
Research into double-low cultivars of conventional winter oilseed rape demonstrated the positive effects of spring N fertilization at 150-180 kg N ha −1 [10,136,137]. In high-input production technologies, yields improved in response to N fertilization rates of 200-250 kg N ha −1 [10,26,44,[138][139][140][141] or even 300 kg ha −1 [70,142]. Miersch et al. [23] demonstrated that yields were 9% higher in semi-dwarf hybrids than in normal-type hybrids (2.24 Mg ha −1 vs. 2.05 Mg ha −1 ) when nitrogen supply was low, whereas no differences in yield were observed between cultivars when nitrogen supply was high. Kessel et al. [30] also reported differences in NUE between open-pollinated and hybrid cultivars. Despite the above, a significant interaction of cultivar and N rate for seed yield is difficult to prove [30,44,[143][144][145]. Differences in the development of long-stem and semi-dwarf cultivars of winter oilseed rape are manifested mainly in fall and early spring [146]. At the end of flowering, differences between cultivars generally disappear and their harvest index reaches similar values [10,133,135]. For this reason, long-stem and semi-dwarf cultivars can respond similarly to N fertilization [10,130,132,133]. In the current study, N fertilizer applied at a rate of up to 180 kg ha −1 exerted a yield-forming effect in all winter oilseed rape cultivars. In all tested cultivars, the yield-forming effect of the highest N rate (230 kg ha −1 ) was observed only during the spring and summer growing seasons with low precipitation. The strong impact of environmental and climatic conditions on the yield-forming potential of N fertilizer applied in spring was also described by Jankowski et al. [10]. Nitrogen rates exceeding 180 kg ha −1 exerted no yield-forming effect when the water requirements of winter oilseed rape were fully met in spring. In years with low rainfall in spring, N fertilizer rates up to 240 kg ha −1 improved yields in all analyzed Energies 2020, 13, 4654 23 of 36 cultivars [10]. In a study by Litke et al. [26], weather conditions also significantly differentiated the yield-forming effects of N fertilization across growing seasons. The seed yield of winter oilseed rape supplied with 240 kg N ha −1 ranged from 1.8 Mg ha −1 to 6.8 Mg ha −1 .
According to Cheema et al. [147], the harvest index increases with a rise in N fertilization rate until the achievement of maximum seed yield. The harvest index is lower when the N rate exceeds the rate required for maximum seed yield. In a study by Dresbøll et al. [135], an increase in the N fertilization rate from 120 kg N ha −1 to 280 kg N ha −1 increased the harvest index in the semi-dwarf cultivar (from 0.340 to 0.375) and the long-stem cultivar (from 0.355 to 0.385). In the work of Miersch et al. [23], the harvest index of both semi-dwarf and normal-type hybrids was higher in production technologies characterized by high (0.410-0.460) than low N supply (0.340-0.410). Jankowski et al. [10] found that the harvest index of winter oilseed rape peaked in response to the N rate of 180 kg ha −1 . A further increase in N rate (240 kg ha −1 ) caused a marginal decrease in the value of the harvest index (N rates higher than 180 kg N ha −1 were more likely to stimulate straw yield than seed yield) [10]. In the present experiment, the harvest index of winter oilseed rape also increased up to the N rate of 180 kg ha −1 . However, the yield-forming effect of N fertilization varied across cultivars. The harvest index of the long-stem hybrid cultivar peaked in response to the N rate of 130 kg N ha −1 . The highest increase in the harvest index of the open-pollinated (long-stem) cultivar and the semi-dwarf hybrid cultivar was noted when the N rate was increased to 180 kg ha −1 . These findings indicate that N rates exceeding 130 kg ha −1 exerted more stimulatory effects on seed yield than straw yield only in the open-pollinated cultivar and the semi-dwarf hybrid cultivar.
The S requirements of oilseed rape range from 1.5 to 2.0 kg S 100 kg −1 of seeds [148,149]. Seed yields increase considerably in response to S fertilization when S levels are low in soil and plants.
The lower the availability of S for plants, the greater the S-induced increase in yields [32,43,45,150,151]. In the experiments conducted by Withers et al. [36] and Bilsborrow et al. [37], oilseed rape yields increased by 15-74% and were highest (65% higher than in the control treatment) in soils deficient in SO 2− 4 and at S fertilizer rate of 40 kg ha −1 . In soils with moderate SO 2− 4 levels, S fertilization increased oilseed rape yields by up to 10%. Jankowski [44] demonstrated that the yield-forming effect of S fertilization was strongly correlated with environmental and climatic conditions. The yield-forming effect of S fertilization was weak in soils relatively abundant in SO 2− 4 and in years with above-average precipitation in fall. High S rates (90 kg ha −1 ) improved yields in soils deficient in SO 2− 4 and in a year with heavy precipitation in late fall [44]. In the present experiment, established on soil with low and moderate levels of SO 2− 4 , S fertilization exerted a weak yield-forming effect (seed yield increased by 3% in response to the S rate of 40 kg ha −1 ).
Zukalová et al. [152] demonstrated that the genetic factor significantly influences the effectiveness of S fertilization. The demand for S fertilizer was higher in low-yielding cultivars. The optimal S rate in the cultivation of a genetically modified (highest yielding) cultivar was determined at 40 kg ha −1 . The S-induced increase in the seed yield of less productive (open-pollinated and hybrid) cultivars was 2-4 lower than the corresponding increase in the seed yield of the transgenic cultivar and it required higher S supply (80 kg ha −1 ). In a study by Jankowski [44], S fertilization induced a significant increase in seed yield only in the cultivar with a lower yield potential (open-pollinated). The yield of the more productive hybrid cultivar did not increase in response to S fertilization, In contrast, Wielebski and Wójtowicz [153] found that S fertilization increased yields in both open-pollinated and hybrid cultivars. In the current study, all of the tested winter oilseed rape cultivars responded similarly to S fertilization.
Sulfur plays a particularly important role in N metabolism by accelerating the conversion of plant-absorbed N into protein [154]. Sulfur increases seed yields mainly by improving the N balance in plants [42,85]. Balanced fertilization of winter oilseed rape is essential for maximizing productivity (seed yield) and profits [42]. The interaction effects of S and N on rapeseed yields have been long recognized [42,45,80]. In a study by Jankowski [44], high N rates (240 kg ha −1 ) exerted a yield-forming effect only in combination with an S fertilizer rate of 30 kg ha −1 . A correlation between the agronomic effectiveness of high N rates and S fertilization was also noted in alternative Brassica crops [50,54,80].
According to Fismes et al. [55], N and S exert synergistic effects if both elements are optimally balanced and antagonistic effects if one of the elements occurs in excess. Seed yields can decrease in S-deficient plants supplied with high N rates because non-protein N exerts toxic effects on plants [87]. In the present experiment, N rates higher than 180 kg ha −1 exerted a yield-forming effect only in combination with S fertilization. In production technologies with high N rates (180-230 kg N ha −1 ), the increase in the N rate by 1 kg without supplementary S fertilization was unproductive in all cultivars of winter oilseed rape.
Sulfur fertilization generally improves the ratio of seed yield to total biomass yield in Brassica crops. In the work of Ahmad et al. [155] and Fazili et al. [41], the S rate of 40 kg S ha −1 increased the harvest index of field mustard (Brassica rapa L.) from 0.317 to 0.335 and from 0.265 to 0.345, respectively. In a study by Piri [156], the harvest index of black mustard (Brassica nigra L.) increased from 24% to 29% in response to the application of 45 kg S ha −1 . In experiments conducted in NE Poland, S fertilization did not differentiate the harvest index of camelina [50]. In the current study, S fertilization did not cause differences in the harvest index of the tested winter oilseed rape cultivars.

Seed Quality
In a study of registered crop varieties conducted by the Polish Research Center for Cultivar Testing (COBORU), the average content of crude fat and total protein in the seeds of 25 open-pollinated cultivars grown in 2016-2018 was determined at 466-500 g kg −1 DM and 354-396 g kg −1 DM, respectively [9]. Nutrient concentrations in the seeds of 94 hybrid cultivars grown in the same period were 3 g kg −1 DM (crude fat) and 4 g kg −1 DM (total protein) lower than in the seeds of open-pollinated cultivars. In 2016-2018, the semi-dwarf cv. Thure (the only semi-dwarf variety examined by COBORU) accumulated 11 g kg −1 DM less crude fat than long-stem hybrids [9]. The semi-dwarf cultivar was also less abundant in crude fat in the present experiment. In a study conducted in Czechia, the crude fat content of semi-dwarf cultivars of winter oilseed rape was determined at 432-472 g kg −1 DM and was 1% and 2% higher on average than in hybrid and open-pollinated cultivars, respectively [157]. In Central Germany (Lower Saxony), semi-dwarf and normal-type hybrids of winter oilseed rape accumulated similar amounts of crude fat (496-500 g kg −1 DM) and total protein (183-184 g kg −1 DM) [23].
Glucosinolate levels are an important parameter in winter rapeseed processing, in particular in feed production. Open-pollinated cultivars registered in Poland accumulate around 8.6-14.4 µM of GLS g −1 seeds. Glucosinolate concentrations are around 5-6% higher in hybrid than in open-pollinated cultivars. The semi-dwarf cultivar of winter oilseed rape tested by COBORU contained approximately 1% more GLS in seeds than long-stem hybrid cultivars [59]. Research conducted in Poland outside COBORU stations also demonstrated that GLS content was 2-3% higher in semi-dwarf than in long-stem cultivars of winter oilseed rape [158,159]. In contrast, Ratajczak et al. [129] found that GLS levels were 13% lower in the semi-dwarf cultivar than in long-stem hybrid and open-pollinated cultivars. In the present study, both long-stem cultivars of winter oilseed rape also accumulated 34-39% more GLS than the semi-dwarf hybrid cultivar. The content of undesirable alkenyl GLS was significantly higher in long-stem cultivars than in the semi-dwarf cultivar (74-75% vs. 60%). The concentration and qualitative composition of GLS in the seeds of the semi-dwarf hybrid indicate that this cultivar is more suitable for the production of feed than long-stem hybrid and open-pollinated cultivars.
High N fertilization rates in spring decrease the crude fat content and increase the total protein content of winter oilseed rape. The above relationships have been observed in all rapeseed types, including long-stem cultivars with high EA and GLS content [160], canola-quality cultivars with low EA and GLS content [61,64,80,147] and HEAR lines with high EA content and low GLS content [161]. Similar correlations between N fertilization vs. crude fat and total protein content were reported in many Brassica crops, including camelina [50], brown mustard [68] and white mustard [53,54]. In the present study, N fertilization also increased the crude fat content and decreased the total protein content of winter oilseed rape. The N-induced changes in the concentrations of the above nutrients varied across cultivars. Higher N rates provoked a greater decrease in the crude fat content of the semi-dwarf cultivar than long-stem cultivars (5% vs. 3%). The N-induced increase in total protein concentration was nearly three-fold higher in both hybrid cultivars than in the open-pollinated cultivar (13% vs. 5%). In a study by Miersch et al. [23], crude fat accumulation was highly similar in semi-dwarf and long-stem hybrids (495 g kg −1 DM and 500 g kg −1 DM, respectively) supplied with low N rates. Semi-dwarf cultivars accumulated less crude fat than long-stem cultivars (466 g kg −1 DM vs. 476 g kg −1 DM) in response to high N supply.
Nitrogen fertilization also influences the content and qualitative composition of GLS. In the works of Bilsborrow et al. [70] and Narits [61], the total GLS content of rapeseed increased up to 140-150 kg N ha −1 and was stabilized at higher rates. Zhao et al. [74] and Zhao et al. [139] reported an increase in the GLS content of rapeseed up to N rates of 240-300 kg N ha −1 . However, Wójtowicz et al. [162] and Wójtowicz et al. [163] found that an increase in the spring rate of N from 60 to 220 kg ha −1 did not differentiate GLS levels in oilseed rape. In the present experiment, N fertilization exerted a positive effect on the content and qualitative composition of GLS in the semi-dwarf cultivar. In long-stem cultivars (hybrid and open-pollinated), total GLS concentrations and the content of alkenyl GLS continued to decrease up to the N rate of 180 kg ha −1 . In the above cultivars, total GLS concentrations and the content of alkenyl GLS increased in response to higher rates (>180 kg N ha −1 ) of N fertilization.
The absence of plant-available S generally inhibits the biosynthesis of total protein in oilseed rape, which was observed in traditional cultivars (high-EA and high-GLS) [164], canola-quality cultivars [80,81,83] and alternative Brassica crops, including B. juncea canola [151], B. juncea mustard [151,165], C. sativa [48] and S. alba [54,166]. In the current study, the S-induced increase in total protein content was noted only in long-stem cultivars (open-pollinated and hybrid). Sulfur fertilization decreased total protein concentration in the seeds of the semi-dwarf cultivar.
Sulfur fertilization exerts varied effects on the crude fat content of winter oilseed rape seeds [34]. Differences in the crude fat content of Brassica crops, including B. campestris [167], C. sativa [48,50], B. juncea and S. alba [34], were also noted in response to S fertilization. In the present experiment, the application of S did not affect crude fat concentrations in winter oilseed rape. Sulfur fertilization had no influence on the crude fat content of rapeseed in the works of Filipek-Mazur et al. [47], Ijaz and Honermeier [57], Šiaudinis and Butkutė [38] and Jankowski et al. [34]. In turn, Jankowski et al. [83], Ceh et al. [168] and Sardana and Atwal [169] observed a decrease in the crude fat content of rapeseed supplied with S. Crude fat levels in rapeseed increased in response to S fertilization in Pakistan [80], Iran [81], central and southern Poland [82], China [170] and France [64].
Sulfur fertilizers can compromise the quality of seeds in canola cultivars of winter rapeseed by increasing their total GLS content by 7% [171], 14-37% [34,56,83] (Table 6) or even 50-80% [56,80]. The adverse impact of S fertilization on GLS levels is generally stronger in hybrids than in open-pollinated cultivars [44]; Figure 13 and in plants supplied with high rates of N [44]; Figure 14. Sulfur stimulates primarily the biosynthesis of alkenyl GLS [34,79,172], which compromises the nutritional value of non-fat seed residues [173]. Sulfur enhances the synthesis of methionine which acts as a precursor for alkenyl GLS [60]. In the present experiment, higher S rates increased the content of alkenyl GLS by 28-41% only in long-stem cultivars (with a higher concentration of alkenyl GLS). In the semi-dwarf cultivar (with a low content of alkenyl GLS), the accumulation of GLS was low even in response to a high S fertilization rate (80 kg ha −1 ). The detrimental effects of S on winter rapeseed cultivars with a higher content of alkenyl GLS were also observed by Wielebski [174]. In turn, Rotkiewicz et al. [175] and Jankowski [44] reported a stronger correlation between S fertilization and the concentration of alkenyl GLS in rapeseed cultivars with a lower GLS content.

Economic Efficiency of Winter Rapeseed Production
Economic efficiency, defined as the difference between agricultural inputs and outputs, is a very important consideration in crop production. Only farms where revenues exceed costs are able to grow and build a competitive advantage on the agricultural market [103]. A study conducted by Nilsson et al. [104] in Northern and Central Europe (United Kingdom, Germany, Poland, Estonia, Sweden) demonstrated that traditional (high-input) winter rapeseed production technologies were characterized by the highest economic efficiency but also by the highest production costs. The costs associated with the production of winter rapeseed in the high-input technology were estimated at €800-1000 ha −1 in Germany, €700-800 ha −1 in the United Kingdom, €600-700 ha −1 in Sweden and €550-580 ha −1 in Poland. The relevant costs were lower by €100-150 ha −1 on average in the integrated (medium-input) technology [104]. Sokólski et al. [120] also argued that the production technology of winter oilseed rape should be optimized for specific cultivars (morphotypes) to maximize profitability. According to Chiriac et al. [21], cultivars with different yield potential generate different costs in identical production technologies. In the work of Sokólski et al. [120], the costs associated with the high-input production technology of winter oilseed rape seeds ranged from €861-910 ha −1 (semi-dwarf cultivars) to €956 ha −1 (long-stem cultivar). In the current study, the costs associated with the production of the open-pollinated cultivar (Adriana) ranged from €542-599 (≤130 kg N ha −1 ) to €619-672 ha −1 (≥180 kg N ha −1 ). Production costs were 4% higher in hybrid cultivars (semi-dwarf and long-stem) than in the open-pollinated cultivar. Sulfur fertilization increased production costs by up to ±3% in all analyzed cultivars of winter oilseed rape. Economic efficiency was highest in the long-stem hybrid cultivar (Artoga), followed by the semi-dwarf hybrid cultivar (PR44 D06) and the open-pollinated cultivar (Adriana). Long-stem hybrids should be grown in high-input technologies (230 kg N ha −1 and 40 kg S ha −1 ) in small-area farms where production is limited mainly by access to land (highest revenue per ha). However, when the key limiting factor is the availability of energy resources rather than land (large-area farms), the energy efficiency of long-stem hybrids is expected to be higher in low-input technologies (130 kg N ha −1 and 40 kg S ha −1 ).

Energy Efficiency of Winter Rapeseed Production
In Europe, energy consumption in rapeseed production ranges from 13 35 GJ ha −1 to 35 GJ ha −1 , depending on agricultural intensity [108,[176][177][178]. In a study conducted in a large-area farm in NE Poland, the energy inputs associated with the cultivation of winter oilseed rape reached 29-32 GJ ha −1 in the high-input technology, 21-24 GJ ha −1 in the medium-input technology and 14-19 GJ ha −1 in the low-input technology [44,108]. In the present experiment, energy consumption ranged from 14-19 GJ ha −1 (≤130 kg N ha −1 ) to 22-27 GJ ha −1 (≥180 kg N ha −1 ) regardless of cultivar. In western Asia (Turkey, Iraq), the energy inputs associated with rapeseed production were estimated at 18 GJ ha −1 [179,180]. In the Mediterranean Region (Italy, Greece, Croatia), energy consumption in the production of winter oilseed rape ranged from 9-10 GJ ha −1 to 17-23 GJ ha −1 , subject to the intensity of agricultural inputs [176,181,182]. In large-area farms in the northern USA, energy inputs in rapeseed production reached 9.5 GJ ha −1 [125].
The energy generation potential of biomass depends mainly on its energy efficiency ratio (ratio of energy outputs to energy inputs). In Europe, the energy efficiency ratio of rapeseed production ranges from 1.1-2.4 [176,177,183] to 3.5-5.4 [108,109,184]. In the present experiment, the most energy-efficient crop was the long-stem hybrid cultivar (Artoga), followed by the semi-dwarf cultivar (PR44 D06) and the open-pollinated cultivar (Adriana). The long-stem hybrid was characterized by the highest energy gain in the production technology with an N rate of 180 kg ha −1 and no S fertilization. This production technology is particularly recommended in small and medium-sized farms. In large-area farms where the availability of energy resources is the main limiting factor, the long-stem hybrid cultivar (Artoga) should be cultivated in a lower-input technology with an N rate of 80 kg ha −1 and no S fertilization (highest energy efficiency ratio). Sokólski et al. [120] also observed a relationship between cultivar and the energy efficiency of the production technology of winter oilseed rape. In their study, the optimal production technology of semi-dwarf cultivars was also most energy efficient. In turn, the cultivation of the long-stem cultivar of winter oilseed rape was most energy efficient in the low-input production technology [120].