Project Topics
Get Complete Project Material File(s) Now! »
Improvement of crop nitrogen nutrition
Legumes are usually included in cropping systems with the expectation that they will provide nitrogen to the system that will benefit the principal crop. Indeed, N fertility, obtained from either fertilizer or legumes in monoculture or rotation systems, is probably one of the most, if not the most important leverage to reduce yield variability (Varvel, 2000). Many studies have quantified the pre-crop value of legume species in terms of fertilizers reduction while introducing a legume prior to cereal cultivation. Legume introduction in the rotation could reduce the economic optimum N application rate for cereal crops up to 43–153 kg fertilizer-N ha−1 (St. Luce et al., 2015; Oliveira et al., 2019). Similar results can be found in intercropping. Cereal-pea intercropping can improve grain and straw quality of the intercropped cereal and allow reduced N inputs compared with sole cropping (Lauk and Lauk, 2008; Yu, Xue and Yang, 2014; Monti et al., 2016; Kakraliya et al., 2018). Although both types of diversification could improve N balance, the mechanisms supporting this apparent benefit seem rather different.
The positive response of the cereal in rotation with legumes is attributed to increased levels of readily available nitrogen. Some studies have observed an increase in mineral N in root-zone of cereal preceded by a legume compared to cereal in monoculture (Dalal et al., 1998; Jensen et al., 2004; López-Bellido, Muñoz-Romero and López-Bellido, 2013; Plaza-Bonilla et al., 2016; Peoples et al., 2017). Increase in soil N also occurs for non-legume preceding crop but was found greater from legumes than non-legume crop (Iannetta et al., 2016). Incremental changes in soil N associated with the pulse crops were found strongly correlated to N2 fixation and were highly variable, particularly when a cropping system is evaluated over a long term (Walley et al., 2007; Iannetta et al., 2016). Nitrogen mineralized from grain legume residues was estimated to contribute to only 15-40% of N acquired by the following wheat crop (Evans et al., 2001). The remaining percentage of cereal N must be obtained from pre-existing soil sources. Net soil N balances in grain legume-cereal rotations may be positive or negative depending on the legume species (Lauk and Lauk, 2008; St. Luce et al., 2015; Guinet et al., 2019), symbiotic performance, and agronomic factors (such as crop residues management) (Chalk, 1998; Walley et al., 2007). For example, pulse crops that typically achieve only modest levels of N2 fixation such as Desi and Kabuli chickpea or common bean are more likely to be either N neutral or contribute to a soil N deficit (Walley et al., 2007). The management of legume stubble and straws (exported, buried, or let on the surface of the soil) is also, obviously, a major determinant of the contribution of legumes to N soil balance. Moreover, when management of N balance is not well understood, incorporation of legume in a system could lead to nitrate leaching (Hauggaard-Nielsen, Mundus and Jensen, 2009; Plaza-Bonilla et al., 2016).
In intercropping, any species utilizing the same combination of resources will be in direct competition. However, based on differences in phenological characteristics of species in mixtures, the interaction among them may lead to an increased capture of a limiting growth resource (Willey, 1990). Most annual crop mixtures such as those involving cereals and legumes are grown almost at the same period, and develop root systems that explore the same soil zone for resources (Ndakidemi, 2006). Thus even if they might have slight N transfer through mycorrhizal interaction (Šarūnaitė, Deveikytė and Kadžiulienė, 2010), improvement of the N nutrition of the overall cropping system is mainly due to the ability of the grain legume to rely on another source of N. Results highlighted that intercropping consistently stimulates complementary N use between legumes and cereals by increasing N2 fixation by grain legumes and increasing soil N acquisition in cereals (Rodriguez et al., 2020).
Improvement of soil properties
While some of legume’s rotational advantages may be associated with improvements of the availability of N in soils, factors unrelated or indirectly related to N also play an important role (Osterholz, Liebman and Castellano, 2018).
High phosphorus (P) availability presumably improves biomass and grain yield in intercropping (Latati et al., 2016). Excessive applications of P fertilizers and P-rich manures have led to an accumulation of P in many soils, which lowers P-fertilizer-use efficiency and leads to P losses via runoff and eutrophication of surface waters (Zhang, Chen and Vitousek, 2013). Increased P uptake in intercropping can be achieved through complementarity or facilitation interaction (Xue et al., 2016). Niche complementarity can occur when two intercropped species tap into different pools of soil P, such as inorganic and organic pools (Li et al., 2008) or different chemical forms of each of these pools (Wang, Marschner and Zhang, 2012). Interspecific facilitation in P uptake occurs when one species increases soil P availability and the intercropped companion species can take advantage of that. Intercropped cereals benefit from legumes in terms of P acquisition mostly in alkaline and neutral soils in which rhizosphere acidification in response to N2 fixation raises P availability (Xue et al., 2016). Most cereal/legume intercropping studies implicitly assume that cereals will benefit from the legumes because the roots of legumes are known to secrete larger amounts of protons (Tang, Barton and McLay, 1997), carboxylates (Pearse et al., 2006) and phosphatases (Makoi, Chimphango and Dakora, 2010) per unit of rhizosphere soil than those of cereals.
Moreover, the production of hydrogen gas (H2) as a by-product of nitrogen fixation have recently been shown to impact significantly microbial communities. Legume symbioses with rhizobia lacking hydrogenase enzymes (which can recycle hydrogen) have traditionally been viewed as energetically inefficient. However, recent studies suggest that hydrogen release to soil may be beneficial to increasing soil carbon sequestration and promoting growth of hydrogen-oxidizing bacteria beneficial to plant growth both in intercropping and rotation (Golding, 2009; Peoples et al., 2009; Golding and Dong, 2010; Angus et al., 2015).
A higher crop diversity would also result in higher soil microbial biomass and richness (Bartelt-Ryser et al., 2005; Yu, Xue and Yang, 2014; Venter, Jacobs and Hawkins, 2016; Qin et al., 2017; Borase et al., 2020). Change in microbial population may induce differences in the catabolic capability of soil (Aschi et al., 2017; Borase et al., 2020). Legumes also seem to induce a change in the surrounding habitat of microbial communities by increasing the active (labile) carbon pool (Hernanz, Sánchez-Girón and Navarrete, 2009; Nath et al., 2019; Sánchez-Navarro et al., 2019) and achieving a suitable soil pH (Qin et al., 2017; Borase et al., 2020).
The benefits of intercropping are also highly subjected to microbial communities and particularly to the mycorrhizal symbiosis establishment and its impact on the soil microflora functionalities (Wahbi, Prin, et al., 2016). In particular, it is well known that plant species, highly dependent on the mycorrhizal symbiosis for their growth (i.e., legumes) will promote the development of the mycorrhizal fungal growth (Duponnois, Plenchette and Bâ, 2001; Ingraffia et al., 2019). Mychorrizal fungi facilitate the uptake and the transport of less mobile soil nutrients (such as P) (Tang et al., 2014) by plants, enhance N fixation and drive biological interactions among neighboring plants (Ingraffia et al., 2019).
Finally, in the short term, plant diversity effects, including legume diversity effects on soil are mediated by a general stimulation of soil microbes and nutrient availability which mostly beneficiate intercropping whereas the longer-term effects of particular plant species are more likely due to compositional shifts in soil microbial communities (Bartelt-Ryser et al., 2005).
Weed, pest and auxiliaries
Among the advantage of crop diversification with legumes, a lot of biotic interactions between the different system components are expected. Weed and pest control is often referred to as an important concern, especially when farmers are aiming at reducing phytosanitary inputs. It has been highlighted that direct weed control can only be successful where preventive and cultural weed management is applied to reduce weed emergence (e.g. through appropriate choice of crop sequence, tillage, cover crops) (Bàrberi, 2002). As a consequence, weed management must be approached on a long-term basis and requires an integration of many cultural practices. Pulses, as break-crop or as part of a rotation, can contribute to weed or pest control (Seymour et al., 2012). They represent a source of diversification in order to break disease, pest and weed cycles and optimize nutrient management in standard crop rotations (Stagnari et al., 2017). Weed biomass (Midya et al., 2005; Banik et al., 2006) and pest abundance (Wahbi, Maghraoui, et al., 2016) can also be significantly reduced in cereal-legume intercropping systems compared with pure stands. However, crop species and varieties express various shading ability, and therefore different competitiveness against weeds (Gollner, Starz and Friedel, 2019).
Another biotic interaction which have received less attention is the impact of legume usage on pollinators. Legumes are also known as bee-pollinated species (Marzinzig et al., 2018). The importance of plant-pollinator interaction in flowering crops such as legumes is based both on the dual roles of pollination for production services and breeding strategies, as well as the increasing need to mitigate the decline of pollinators. Pollinator-friendly legumes can significantly contribute to the habitat zone. Due to their specific flower type and the unusual color of the petals, they can attract, for example, wild bees, which are also effective for many other crops (Suso et al., 2016). Among strategies generally proposed in favor of pollinators, flowering crops are much more appreciated by farmers than flower strips or other unproductive habitats (Suso et al., 2016).
Variability of pulse crops agro-ecosystem properties
The first three principal components (PC) altogether explained 74% of the variability of agro-ecosystem properties with the first PC (33%) strongly positively associated with biomass yield (BY) and grain yield (GY) and to a lesser extent %Ndfa (percentage of nitrogen derived from atmosphere) (Figure 2a). This axis opposed high yield producers like Canavalia ensiformis (GY 2.53 ± 2.8 t ha-1, BY 9.79 ± 2.29 t ha-1) to poor yield providers like Cymopsis tetragonoloba (GY 1.01 ± 1.3 t ha-1, BY 1.42 ± 0.5 t ha-1), respectively representative of the phaseolids clade (warm season legume) and sister the indigoferoid clade.
The second PC axis (24%) opposed LAI (leaf area index) to percentage of yield reduction due to weeds (%YR) and GY (Figure 2a), but the correlations between the second PC axis and respectively, LAI and %YR were both low (r = -0.15, P = 0.5; Supplementary Table S2). %YR was also found to be only slightly correlated to GY (r = 0.27, P = 0.22). This could suggest that LAI alone is not enough to predict species ability to produce high yield under strong weed competition. The second axis opposed Lathyrus ochrus (LAI 4.1 ± 0.3, %YR 5%) and Lens culinaris (LAI 3.9 ± 2, %YR 2 ± 1 %), both representative of Galegoid clade, to Cyamopsis tetragonoloba (LAI 1.4, %YR 79 ± 6 %) and Glycine max (LAI 2.7 ± 1, %YR 60 %), representative of the phaseolids clade. The third axis explained 16% of the variability and opposed WUE (water use efficiency) and %Ndfa (Figure 2b). As expected, GY and BY were significantly correlated (r = 0.69, P < 0.001; Supplementary table S 2). PCA results associated %Ndfa to yield on the first axis, yet it was only slightly correlated to BY (Figure 2; r = 0.22; Supplementary Table S2). This property (%Ndfa) along with WUE is well represented on axis 3 which opposed (Figure 2c) Vigna unguiculata (%Ndfa 24.7
± 0.5 %, WUE 8.3 ± 10), representative of phaseolids clade, to Lathyrus sativus (%Ndfa 93.3 ± 3.5 %, WUE 6.3 ± 3.1), representative of Galegoid clade. Pearson’s correlation coefficient showed a very slight positive, but not significant, correlation of WUE with LAI (r = 0.35, P = 0.11). Quality of the representation of WUE in the PCA was low; the property was slightly associated with the three axes with low explain of variation explain on each axes, and it was also slightly but not significantly correlated with BY (r = 0.27, P = 0.22).
The projection of individual species in the space of covariation showed that species characterized by high LAI and good tolerance to weed infestation (Lathyrus ochrus, Lens culinaris, Lathyrus sativus, Vicia sativa), are relatively unproductive species with GY varying from 0.6 to 1.1 t ha-1 and BY from 2.3 to 3.4 t ha-1. Moreover, they seemed relatively adapted to nitrogen deficiency and drought. Except for Lens culinaris, which showed a low nitrogen fixation efficiency, %Ndfa varied from 69% to 93%, while WUE varied from 6.05 to 7.5 kg ha-1 mm-1. However, relatively unproductive species could be sensitive to weed infestation, especially as they show a lower WUE (Vigna radiata, Vigna mungo) or a lower %Ndfa (Vigna unguiculata).
Productive species (GY from 1.1 to 2.5 t ha-1 and BY from 3.9 to 10.6 t ha-1) can be sensitive or insensitive to weed infestations (28 to 83 % of yield reduction), and this ability cannot be related to nitrogen or water adaptations. However, productive species, except for Phaseolus lunatus, show a high value of WUE (from 5.2 to 10.4 kg ha-1 mm-1) and %Ndfa (from 61 to 82%).
Cyamopsis tetragonoloba is found apart from every other species in the space of covariation as it showed low values on every agro-ecosystem property.
(Co-)variations of functional traits
The first three PC axes of the PPCA (probabilistic principal component analysis) performed on functional traits collected for the 43 species explained 61% of total variance (Figure 3).
PC1 (33%) was associated with morphological traits, especially leaf traits (e.g. leaflet length, leaflet width, leaflet number, leaf area, plant height, and to a lesser extend seed diameter and seed weight). This axis opposed small plants such as Astragalus sp. (i.e. A. hamosus: seed diameter 2.1 ± 0.4 mm; TSW 17.7 ± 18.9 g; 21 ± 4.2 leaflets; leaflet length 9.8 ± 1.0 mm; plant height 55.9 ± 4.1 cm, LA 48.3 ± 19.6 cm2) to large plants such as Canavalia sp. (i.e. C. gladiata: seed diameter 23.3 ± 10.4 mm; TSW 1007.5 ± 382.3 g; tri-foliate; leaflet length 137.5 ± 72.2 mm; plant height 432.9 ± 186.8 cm; LA 143.7 ± 15.9 cm2). The number of leaflets is characteristic of the species’ origin; more than half (n =
26) of the species are trifoliate, most of them being tropical legumes except for a small number of Mediterranean species (n = 5) with a number of leaflets ranging from 1 to 4 leaflets per leaf. The remaining (n = 17) have many leaflets, ranging from 5 to 24; this was the case of Mediterranean and European species only. Number of leaflets showed significant negative correlation to leaflet length (r = -0.81, P < 0.001; Supplementary Table S3) which ranged from 9.8 ± 1.0 mm (Astragalus hamosus) to 137.5 ± 72.2 mm (Canavalia gladiata). Plant height (PH) ranged from 11.3 cm (Cicer reticulatum) to 432.9 ± 186.8 (Canavalia gladiata) and averaged 103.1 ± 91.8 cm. LA ranged from 21.11 cm2 (Vicia ervilia) to 6885.4 ± 5324.5 cm2 (Psophocarpus tetragonolobus) and was positively correlated to all leaflet traits and PH; leaflet number (r = -0.56, P < 0.001), leaflet length (r = 0.61, P
< 0.001), leaflet width (r = 0.55, P < 0.01) and PH (r = 0.44, P < 0.01; Supplementary Table S3). PC2 (15%) was strongly associated with seed traits such as thousand seed weight (TSW), seed diameter and seed oil and protein content (Figure 3a). This axis opposed species with small, low quality seeds (i.e. Vigna radiata oil content 0.97 ± 0.2 %; protein content 25.2 ± 1.8 %; seed diameter 2.75 ± 1.8 mm; TSW 39.9 ± 11.1 g) to species producing larger seeds with higher oil and protein content (i.e. Lupinus mutabilis oil content 18.2± 5.2 %; protein content 45.4 %; seed diameter 11 ± 4.5 mm; TSW 286.75 ± 171.2 g) (Figure 3c). Seed diameter ranged from 2.1 ± 0.3 mm (A. annularis) to 23.3 ± 10.4 mm (C. gladiata) with a mean of 7.8 ± 4.5 mm. TSW was closely correlated with seed diameter (r = 0.87, P < 0.001; Supplementary Table S3) and ranged from 12.8 ± 3.9 g (Trigonella foenum-graecum) to 1178 g (Phaseolus coccineus) and averaged 255.4 ± 294.1 g. Seed protein content ranged from 10.9 % (Canavalia ensiformis) to 45.4 % (Lupinus mutabilis) and was moderately correlated to seed oil content (r = 0.45, P < 0.01; Supplementary Table S3) which varied greatly from 0.4 % (Vigna angularis) to 47.1 ± 1.9 % (Arachis hypogaea) but with a dominance of low values (mean: 4.8 ± 8.4 %).
Relationships between functional traits and agro-ecosystem properties
CART analysis showed that grain legume species with SLA above 274 cm2 g-1 (average 245.5 ± 105.4) produce more biomass (Figure 4). BY is maximal (8 t ha-1) for high SLA combined with TSW above 37 g (which concerns most of the species as only nine species have smaller seeds), and PH above 83 cm (not far from the average 103.1 ± 91.8). BY around 4-5 t ha-1 can be achieved by small species (< 83cm) with high SLA, and doesn’t seem to be related to seed weight. Low SLA species can still achieve intermediate BY if they have small seeds. Among the lowest biomass producers (small SLA, big seeds) erect species and tall species have higher BY.
GY is maximal (2.2 t ha-1) for species which take more than 168 days to mature (average 135.0 ± 38.4) (Figure 5). In early maturing species, GY is maximized when LNC is higher than 40 mg g-1 (average 38.4±12.1) in species that produce big seeds (TSW > 164 g). In small-seeded species (TSW < 164 g), there is no relation between LNC and GY whereas erect habit generally leads to higher GY than climbing habit. Rapid crop cycle combined with small seeds generally leads to low GY with one exception for tall species (PH >= 61 cm) with erect habit.
The number of days to maturity (DM) was also discriminant for %Ndfa. This agro-ecosystem property is indeed maximized (66-78%) for late maturing (DM >= 142 days) species, among which small-seeded species (TSW < 147 g) fix more nitrogen from the atmosphere than large-seeded species (TWS >= 147 g; Figure 6). A slightly smaller amount of nitrogen fixation could also be reached by early maturing species with smaller leaflets length (< 101 mm) and S/R specifically above 4.8. Yet, the advantage of a large S/R ratio was only observed in this situation, while early maturing species with long leaflets exhibit low %Ndfa.
In our data analyses, LAI of pulses is sequentially smaller in small-seeded species that have high protein content in their seeds, an epigeal germination and small leaflets (Figure 7).
Yield of species with very small leaflets (length < 30 mm) is less negatively impacted by weeds (Figure 8). However, leaflet length under 30 mm is only encountered in four species over the 25 and two of them are under 10% YR which is extremely low compared to other species. For the remaining 21 species, DM was again a crucial trait. Short cycle (< 103 days to maturity) leads to reduced yield losses due to weeds (under 60%). Yet, a long crop cycle duration combined with hypogeal germination and a better harvest index (HI) than average (>= 0.29) also resulted in a similar percentage of yield loss due to weeds (43%). Among early maturing species, yield of those that have the highest oil content in seeds is less affected by the presence of weeds. WUE (8.3 kg ha-1 mm-1) is maximized in species with high LNC (>= 50 mg g-1) (Figure 9). While in species with low LNC those with small leaves have the lowest WUE, it is noticeable that in species with larger leaves could have a low WUE is significantly higher for those that have the lowest seed oil content.
Table of contents :
CHAPTER 1: GENERAL INTRODUCTION
A. DIVERSITY TO SUPPORT RESILIENCE AND ECOSYSTEM SERVICES
Biodiversity in agro-ecosystem
Design and evaluation of crop diversification strategies
Trait-based approach for mixture design
B. LEGUMES: A CRUCIAL ROLE TO PLAY IN THE DIVERSIFICATION OF CROPPING SYSTEM
An underused diversity
Context in Europe
Properties of legumes and expected benefit for agro-ecosystems
1. Agro-ecosystem productivity
2. Improvement of crop nitrogen nutrition
3. Improvement of soil properties
4. Weed, pest and auxiliaries
5. Effect on climate change mitigation
C. PROBLEMATIC AND APPROACH OF THE THESIS
Research questions and general methodology
Hypotheses
Approach
D. REFERENCES
CHAPTER 2. EVALUATION OF PULSE CROPS’ FUNCTIONAL DIVERSITY SUPPORTING FOOD PRODUCTION
A. INTRODUCTION
B. MATERIAL AND METHODS
Approach
Data sources
Data analysis
C. RESULTS
Variability of pulse crops agro-ecosystem properties
(Co-)variations of functional traits
Relationships between functional traits and agro-ecosystem properties
D. DISCUSSION
Plant production strategies among pulse species
Patterns of trait covariations and plant production strategies
From trait profiles to services
Limits
E. CONCLUSION
F. SUPPLEMENTARY MATERIALS
G. REFERENCES
CHAPTER 2. FUNCTIONAL DIVERSITY OF PULSES’ SPECIES IN RELATION WITH AGROECOSYSTEM
PROPERTIES
A. INTRODUCTION
B. MATERIAL AND METHODS
Plant material and growing conditions
Measurements of individual plant traits
Measurement of agro-ecosystem properties
Data analysis
C. RESULTS
Consistency of functional trait space
Characterization of functional trait diversity
Trade-offs between agro-ecosystem properties
Relevant traits for prediction of agro-ecosystem properties and structuration of functional
diversity
D. DISCUSSION
Trait related to resource acquisition are less stable across growing conditions and
varieties than architectural traits
Intraspecific variation contributed only marginally to global variability
Leaf morphological traits confirmed as variability drivers among pulses
LMR and LAR reflect diverging biomass investment strategies
Yield is inversely related with Ndfa and soil cover
Trait-based predictions of agro-ecosystem properties
E. CONCLUSION
F. SUPPLEMENTARY MATERIALS
G. REFERENCES
CHAPTER 4. AN EXPLORATION OF THE VARIABILITY OF PHYSIOLOGICAL RESPONSES TO SOIL DRYING IN RELATION WITH C/N BALANCE ACROSS THREE SPECIES OF THE UNDER-UTILIZED GENUS VIGNA GENUS
A. INTRODUCTION
B. MATERIAL AND METHODS
Plant material and growing conditions
Dry-down experiment and N fixation activity
Data analysis
C. RESULTS
NARA and NTR variation along FTSW gradient
Variation of leaf traits
Relation between leaf traits and parameters of transpiration and N fixation dynamics
D. DISCUSSION
A very late response of transpiration to soil drying for all Vigna sp.
A high tolerance of N2 fixation to soil drying in Vigna sp.
Leaf traits as indicators of C and N metabolism changes under drought
E. CONCLUSION
F. ADDITIONAL RESULTS: C. ARIETINUM
Introduction
Material and Methods
Results
1. NARA and NTR variation along FTSW gradient
2. Variation of leaf traits
3. Relationships between leaf traits and parameters of transpiration and N fixation dynamics
Discussion
1. A very early response of transpiration to soil drying for C. arietinum
2. A high sensitivity of N2 fixation to soil drying in C.arietinum.
3. Leaf traits as indicators of C and N metabolism changes under drought
Conclusion
G. REFERENCES
CHAPTER 5. GENERAL DISCUSSION
A. OVERVIEW AND MAIN RESULTS OF THE THESIS
Synergies or trade-offs between water stress resistance, crop production and desired
support agro-ecosystem services
1. Pulses species and varieties incidence on agro-ecosystem property suites
2. Water deficit influence on agro-ecosystem properties rate and stability
Patterns of trait covariations and plant production strategies
1. Traits structuring pulses diversity
2. Trait co-variations and trade-offs
3. Traits variations between cultivars and growing conditions
From traits to agro-ecosystem services
1. Important traits for prediction of agro-ecosystem properties
2. Traits to agro-ecosystem properties
3. Trait variability as an indicator of the sensitivity of physiological process to environmental conditions
B. LIMITS
Trait choice
An unsatisfactory relation to agro-ecosystem services
Intra-specific variability
Trait-based approach applied to the study of diversified agro-ecosystem
C. PERSPECTIVES
D. REFERENCES
CHAPTER 6: GENERAL CONCLUSION
ANNEXES
