| dc.description.abstracteng | Oilseed rape is the third most important oil crop worldwide. Under many conditions it is characterised by relatively low nitrogen efficiency. In Germany a nitrogen supply of 240 kg N ha 1 is aimed at which derives from soil mineral nitrogen and nitrogen fertiliser. Thus, large amounts of fertiliser are applied. But only about 50 % of fertilised nitrogen are recovered by the crop. Furthermore nitrogen losses appear with leaf shedding that starts during flowering. Only little amounts of nitrogen are taken up between end of flowering and maturity. With a nitrogen harvest index of 0.7 to 0.8 at least 20 % to 30 % of plant nitrogen remain on the field after harvest. As a result nitrogen surpluses of 90 kg N ha 1 to 100 kg N ha 1 were reported after cultivation of oilseed rape. EU legislative restrictions address greenhouse gas emissions in production of biodiesel made of oilseed rape (EU directive 2009/28/EG) and nitrogen surpluses in agricultural production (EU nitrate directive). These restrictions have moved nitrogen efficiency into focus of rapeseed breeders.
Nitrogen use efficiency can be defined as seed yield per unit available nitrogen. Nitrogen use efficiency is the product of nitrogen uptake efficiency, i.e. the amount of nitrogen which can be taken up per unit available nitrogen, and nitrogen utilisation efficiency measured as grain yield which is produced per unit nitrogen which was taken up. Nitrogen efficiency is a difficult trait to select for as seeds and straw need to be harvested and analysed for nitrogen content. This is laborious and time consuming. Indirect selection methods would facilitate selection. Reflectance of plants is reported to be related to chemical composition and structural features particularly of leaves. Also the nitrogen status of plants and crop stands were shown to be predicted by reflectance.
Roots are important for nitrogen uptake and thus, for nitrogen efficiency. Therefore, it may also be worth to consider the root when selecting for nitrogen efficiency. Phenotyping of roots is difficult, often destructive and only possible for a very limited number of genotypes. Electrical capacitance has been discussed for decades to be related to root characteristics like root surface area or root mass. If so, it may be also related to nitrogen uptake and thus nitrogen efficiency. In contrast to other methods for root phenotyping it is a flexible, non-destructive method that allows quick phenotyping of large numbers of genotypes in the field at any time and any place.
Field experiments were conducted to explore nitrogen efficiency of winter oilseed rape and its prediction by hyperspectral canopy reflectance and electrical capacitance in the field. A diversity set consisting of 29 genotypes was tested at five Central and Northern German environments in seasons 2011/12 and 2012/13. Genotypes were grown at high (160 kg N ha 1 to 180 kg N ha 1) and without nitrogen supply. Two parallel trials were conducted – one was harvested at end of flowering, the other one at maturity. Aboveground biomass at end of flowering, nitrogen uptake at end of flowering, nitrogen uptake efficiency at end of flowering, seed yield, nitrogen uptake at maturity, nitrogen uptake efficiency at maturity, nitrogen utilisation efficiency, nitrogen use efficiency, nitrogen harvest index and nitrogen uptake between end of flowering and maturity were analysed for genetic variation and heritability. The genotype by nitrogen level interaction was examined to answer the question whether selection for nitrogen efficiency parameters should be conducted at different nitrogen levels.
Two populations of 15 DH lines each and their test hybrids were tested at three environments in season 2013/14. They were grown at high (N1) and without nitrogen (N0) fertilisation and harvested at maturity. Seed yield, nitrogen uptake at maturity, nitrogen uptake efficiency at maturity, nitrogen utilisation efficiency, nitrogen use efficiency and nitrogen harvest index were analysed to test variation, difference between DH lines and test hybrids, interactions between variety type and nitrogen level, heritability and mid-parent heterosis. For the diversity set and DH lines and test hybrids the contributions of variances of nitrogen uptake efficiency and nitrogen utilisation efficiency to nitrogen use efficiency were computed.
Hyperspectral canopy reflectance was measured with a portable field spectrometer in each plot before flowering and during fruit development. Reflectance from 305 nm to 1800 nm was used to develop calibrations for nitrogen uptake at end of flowering and at maturity and for seed yield. Calibrations were developed across and within nitrogen levels. Calibrations were validated in tenfold cross validations and external validations.
Electrical capacitance of the diversity set, DH lines and test hybrids was measured in the field at end of flowering and during fruit development. Its relation to nitrogen efficiency and agronomic parameters was tested. To examine the relationship between electrical capacitance and root characteristics ten genotypes of the diversity set selected for differences in electrical capacitance were tested in the field and under controlled conditions. Next to electrical capacitance stem diameter and root masses in three horizons were determined in the field trials. Under controlled conditions single plants were grown in plastic tubes in the greenhouse. Electrical capacitance, stem diameter and fresh mass of whole root system, taproot and lateral roots were measured directly. Image-base analysis was used to analyse further root characteristics like root diameter, root area, root system width and root tips.
Field trials with the diversity set revealed high heritabilities from 0.67 to 0.92 for aboveground biomass at end of flowering, nitrogen uptake at end of flowering, nitrogen uptake efficiency at end of flowering, seed yield, nitrogen uptake at maturity, nitrogen uptake efficiency at maturity, nitrogen utilisation efficiency, nitrogen use efficiency, nitrogen harvest index and nitrogen uptake between end of flowering and maturity. Thus, these traits can be used as selection criteria for nitrogen efficiency. Except nitrogen uptake efficiency all traits were significantly affected by the interaction between genotype and nitrogen level. Therefore, selection environments should resemble nitrogen supply of target environments. At both nitrogen levels nitrogen utilisation efficiency contributed more to the variation in nitrogen use efficiency than nitrogen uptake efficiency.
All traits but nitrogen harvest index showed significant variation among pairs of DH line and descending test hybrid. Heritabilities ranged from 0.31 to 0.82. Most traits were not affected by nitrogen level. Only nitrogen uptake efficiency at maturity and nitrogen use efficiency were significantly higher at N0 than at N1. Significant differences between DH lines and test hybrids were observed only for nitrogen utilisation efficiency and nitrogen use efficiency – both higher for test hybrids than for DH lines. Interactions between nitrogen level and variety type and between nitrogen level and descent (describes the pair of a DH line and the test hybrid derived from this DH line) were not significant for any trait.
Mid-parent heterosis at both nitrogen levels was detected for seed yield, nitrogen utilisation efficiency, nitrogen use efficiency and nitrogen harvest index. Hybrids surpassed the parental mean for nitrogen uptake and nitrogen uptake efficiency at N1 but at N0 the heterosis was negative, i.e. hybrids performed worse than the parental mean. For seed yield, nitrogen uptake, nitrogen uptake efficiency at maturity and nitrogen use efficiency heterosis was higher at N1 than at N0. But for nitrogen utilisation efficiency and nitrogen harvest index higher heterosis was expressed at N0. In contrast to the diversity set nitrogen use efficiency of DH lines and test hybrids was dominated by nitrogen uptake efficiency while nitrogen utilisation efficiency only contributed to a small portion to nitrogen use efficiency. DH lines and test hybrids were only grown in one season which was characterised by an extraordinary warm winter and early spring. Therefore, the findings for DH lines and test hybrids need to be confirmed in further field trials.
Best calibrations with hyperspectral reflectance showed coefficient of determinations up to 0.87 for calibration and up to 0.85 for cross validation though lower for seed yield than for nitrogen uptake. That suggests the application of hyperspectral reflectance as indirect selection method. Calibrations based on spectral data before flowering resulted in better predictions than calibrations based on spectral data during fruit development. There was no general pattern when calibrations across nitrogen levels were compared with separate calibrations within nitrogen levels. For nitrogen uptake best calibrations across nitrogen levels outperformed best calibrations within nitrogen levels. Best calibrations for seed yield within low nitrogen supply outperformed best calibrations within high nitrogen supply and across nitrogen levels.
Most calibrations lost their predictive ability when tested with external datasets. Thus, they need to be further improved before they can be applied in breeding programs.
Electrical capacitance revealed significant genetic variation and high heritabilities in the diversity set (h² = 0.81) and for the ten genotypes tested for root characteristics in the field (h² = 0.91) and under controlled conditions (h² = 0.95). Thus, electrical capacitance can in principal be used as selection criterion. But only few significant phenotypic correlations were found between electrical capacitance and nitrogen efficiency parameters in field trials with the diversity set. At N1 electrical capacitance at end of flowering correlated negatively with dry matter content of aboveground biomass at end of flowering ( r = -0.70) and positively with nitrogen uptake (r = 0.57) and nitrogen uptake efficiency at maturity (r = 0.57). At N0 electrical capacitance at end of flowering correlated positively with oil content (r = 0.57). It cannot be suggested to employ electrical capacitance as selection criterion for nitrogen efficiency parameters. The ten selected genotypes did not show significant differences in root masses and stem diameter in the field trial. Accordingly, they did not reveal significant phenotypic correlations between electrical capacitance and root masses or stem diameter. On plot level electrical capacitance correlated weakly with root mass in the middle (r = 0.46) and bottom (r = 0.34) horizon and strongly with stem diameter (r = 0.78). Under controlled conditions only the phenotypic correlation between electrical capacitance and stem diameter was significant (r = 0.78). Genetic correlations between electrical capacitance and root characteristics and stem diameter were detected. Correlation coefficients ranged from 0.55 for root fresh mass to 1.11 for root tip diameter. The same root traits that were related to electrical capacitance also correlated with stem diameter. Genetic correlation coefficients between stem diameter and root characteristics were higher than those between electrical capacitance and root characteristics. Although electrical capacitance might be related to root characteristics, stem diameter, which is much easier to measure, also correlates to root characteristics and often with higher correlation coefficients.
By the current study it could be shown that nitrogen efficiency and related parameters can be implemented as traits in plant breeding as they revealed high heritabilities. Selection should be conducted at nitrogen levels that resemble target nitrogen supply. Hyperspectral canopy reflectance measured before flowering can be applied to facilitate selection. Nevertheless, calibrations must be further improved. Electrical capacitance cannot be suggested as selection criterion for nitrogen efficiency parameters. It remained open which trait is captured by electrical capacitance. Yet, the high heritabilities confirm that it is a heritable trait. | de |