**Test of the CERES-RAPE model for winter oilseed rape : effect of the plant nitrogen status on yield’s components**

^{1} Institut Supérieur d’Agriculture de Lille, F59046 Lille cedex __Colnenne@isa.fupl.asso.fr__^{
2} Institut National de la Recherche Agronomique, F78850 Thiverval-Grignon, __Roche@bcgn.grignon.inra.fr__ , __Meynard@jouy.inra.fr__

**ABSTRACT**

*In the winter oilseed rape model CERES CECOL, the nitrogen status of the crop after the beginning of flowering (F1) is of little importance in the yield production simulation, and specially in the number of flowers/m² determination.*

*Six field experiments have been carried out with different spring nitrogen nutrition levels. The results have showed that the simulated number of flowers/m² by CERES CECOL were systematically under estimated, and we have proposed a new model, which gave importance to the nitrogen status of the crop both before and after the stage F1.*

**KEYWORDS : **Number of flowers/m², nitrogen nutrition index, nitrogen status in spring

**INTRODUCTION**

The model CERES CECOL simulates the growth and the production of winter oilseed rape. It arises from the association of two modules. The first one describes the soil compartment, and the other one desscribes the phenology and the growth and the yield components of the crop (Husson, 1997). From emergence to the beginning of the flowering (i.e. in Autumn and in Spring), the growth is simulated according the model developed by Monteith (1972) which takes into account the nitrogen status of the crop.

During the yield components elaboration, the nitrogen nutrition occurs only on the flowers to pods succes rates (Leterme, 1991), and on the pods R.U.E. (Gabrielle et al., 1998).

However, different authors have already showed the effects of pedoclimatic conditions or agricultural practices on the yield components determination and specifically on the numbers of flowers/m² (Husson and Leterme, 1997). The objective of this study is (1) to measure the consequences of different nitrogen limiting growth conditions in Spring on the number of flowers/m², (2) to compare these observations to the results simulated by CERES CECOL.

**1. MATERIALS AND METHODS**

Six field experiments with the winter oilseed rape cultivar Goeland were carried out during two winter growing seasons (1995, 1996) in different pedoclimatic conditions. In Autumn, the plants were eventually fertilized to prevent a lack of nitrogen during this period. To obtain various nitrogen deficiencies at the beginning of flowering, different increasing nitrogen fertilizer levels have been spread (between 0 and 300 kg N /ha amonium nitrate) at the regrowth period.

In Spring we took regular measurements of dry matter and total nitrogen concentration (Dumas Chemical analysis) on the aerial parts, on three samples of 0.8 m² from each plot. These data allowed us to calculate the Nitrogen Nutrition Index (Lemaire et al., 1987) from the regrowth period until the stage G4 (development stage characterized by the ten first pods baterred). We measured the L.A.I. at the beginning of flowering (stage F1) to calculate the simulated numbers of flowers/m² with CERES CECOL, and the numbers of flowers/m² were counted on 50 plants from each plot.

**2. RESULTS**

*2.1. Characterization of the Nitrogen deficiencies and number flowers/m² observed*

We chose to present the results from one experiment : Saint Florent sur Cher in 1995. We can see in Table 1 some characteristics of the nitrogen deficiencies (dates of the beginning, intensities at stages F1 and G4, integrated N.N.I. between F1 and G4) and the observed numbers of flowers/m². From the end of winter (i.e. regrowth period) to the stage G4, the nitrogen deficiencies were variable in duration and in intensity. The lower the nitrogen supply in Spring was, the earlier and stronger the nitrogen deficiencies were. The nitrogen deficiency of the treatment T40 (40 kg N /ha) became approximately 600-800° (base 0) earlier than the T240 (240 kg N/ha), and was stronger. At stage F1, the N.N.I. values were 0.84 and 1.56 respectively for these two treatments. Between stages F1 and G4, the N deficiencies were increased and the last observation were 0.51 and 0.84 for T0 and T240. These N stresses could be described by the integrated N.N.I., which were for those two specific treatments 0.67 and 0.98 respectively.

The numbers of flowers/m² measured in field were correlated to the N supply, or to the integrated N.N.I. values. Earlier and stronger the N deficiencies were (i.e. the N supply is low) smaller the numbers of flowers were, and a statistic test showed significant differences between treatments (cf. Table 1).

*Table 1 : Description of the Nitrogen deficiencies in spring and the number of flowers/m² : Saint Florent/ Cher in 1995 *

Treatments |
Beginning of nitrogen deficiency(1) |
N.N.I. at stage F1 (2) |
N.N.I. at stage G4 (3) |
I.I.F.G. (4) |
Number of flowers/m² |

T0 |
Ante 1676 |
0.84 |
0.51 |
0.67 |
5370 d |

T40 |
1779-2125* |
0.75 |
0.52 |
0.63 |
8232 c |

T80 |
1676-2287 |
1.07 |
0.54 |
0.80 |
10961 b |

T120 |
2389 |
1.18 |
0.59 |
0.86 |
10665 b |

T160 |
2491 |
1.40 |
0.59 |
0.90 |
11453 b |

T200 |
2569 |
1.37 |
0.78 |
0.96 |
14153 a |

T240 |
2654 |
1.56 |
0.84 |
0.98 |
13088 a |

(1) expressed in sum of temperatures after emergence (base 0°)

(2) F1 : development stage of the beginning of flowering

(3) G4 : developement stage when the ten first pods are baterred

(4) I.I.F.G. : integrated N.N.I. between the stages F1 and G4

a, b , c… statistic groups (test Newman et Keuls, at 5%)

* between

*2.2. Comparison between the observed and simulated numbers of flowers/m².*

In the CERES CECOL model, the number of flowers/m² is simulated according to the following equation : FL/M² = 3000 * L.A.I.F1, where FL/M² is the simulated number of flowers/m² and L.A.I.F1 the L.A.I. measured at stage F1.

The comparison between the observed and simulated numbers of flowers/m² is presented in Figure 1. In all the situations, the simulated values were smaller than those observed. One of the explanations for these differences is that the nitrogen status of crops after the beginning of flowering was not taken into account in the model. This has been confirmed by many observations which concluded that the nitrogen nutrition modified the duration of flowering and the numbers of flowers/m² (Leterme, 1997).

*Figure 1 : Comparison between the numbers of flowers/m² observed and simulated*

**DISCUSSIONS – CONCLUSION**

To take into account both the period ante flowering (phase of the potential number of flowers/m² determination) and the period between F1 and G4 (phase of this yield component expression), a multiple regression between INF, IMF and IIFG has been calculated as follow : INF = 0.2787 IMF + 1.0714 I.I.F.G. – 0.3930 (r² = 0.885 ; n = 18).

INF is the number of flowers/m² expressed in relative value to take free of the pedoclimatic conditions variability, IMF the maximum intensity of the nitrogen deficiencies before the beginning of flowering and I.I.F.G. the integrated N.N.I. between the stages F1 and G4.

According to this model, the N status of crop after stage F1 has a higher importance than the ante flowering period on the number of flowers/m² determination.

This new model has to be improved again, but its advantage is to show the importance of the nitrogen status of crops during the phase F1 to G4. This introduces a new variable which has to be taken into account to determine the numbers of flowers/m² in the model CERES CECOL.

**ACKNOWLEDGEMENTS**

We thank the members of CETIOM for theirs participations in data collecting, advice, and the CETIOM which finances the experiments.

**REFERENCES**

1. Gabrielle B., Denoroy P., Gosse G., Justes E., Andersen M.N., 1998. Development and evaluation of a CERES-type model for winter oilseed rape. Field Crops Research, 57 (95-111).

2. Husson F., 1997. Validation et introduction d’aléas dans un modèle déterministe complexe : application à un modèle de croissance du colza , CECOL. Thèse de Docteur en Sciences, INA-PG, Paris (194p).

3. Husson F., Leterme P., 1997. Construction et validation d’un modèle de prédiction de la date de floraison du colza d’hiver. O.C.L., 5 (379-384).

4. Lemaire G., Gastal F., Salette J., 1989. Analysis of the effect of N nutrition on dry matter yield of a sward by reference to potential yield and optimum N content. Proceedings XVI International Grassland Congress, Nice, France (179-180)

5. Leterme P., 1991. Vers une modelisation globale du fonctionnement du peuplement d’une culture de colza d’hiver. 8^{th} International Rapeseed Congress, Saskatoon (Canada), C(03) (671-676).

6. Monteith J.L., 1972. Solar radiation and productivity in tropical ecosystems. Journal of Applied Ecology, 9 (747-766).