ABSTRACT

Our research program aims to demonstrate that changing the sowing date could be a powerful means to improve the environmental balancesheet of a winter rape crop. The hypothesis is that earlier sowing dates could enable rape to absorb more nitrogen in autumn and decreased soil N mineral before leaching. We compared the effects of very early (end of july) and normal sowing dates (beginning of september) on N uptake and N leaching risk. This was assumed on experimental plots and under farmer conditions.

N potential absorption was very high, almost 200 kg N ha^-1 before winter with normal sowing dates and up to 300-350 kg N ha^-1 with very early sowing dates. This absorption was strongly correlated with leaf area index. In farmer field-trials network, there were no major limitations to sow in dry conditions and absorption was dependent of N availibility. Consequently, very early sowing decreased systematically soil N mineral content before leaching, to a level of less than 30 kg N ha ^-1 for 90 cm depth, whereas it was very variable with normal sowing dates : 20 to 280 kg N ha^-1. We simulated the same drainage for both sowing dates, but simulated water NO₃^- content was systematically lower than 50 mg l^-1 only for early sowing.

N lost by leaves fall was more important in very early sowing and during cold winter. Therefore, differences of plant N accumulated after winter was less than before. N spring fertilizer is reduced accordingly to N accumulated in plants after winter (Reau et al., 1996). Therefore, reduction of N fertilizer was dependant of leaves fall intensity. In conclusion, very early sowing are of great interest to reduce N leaching, especially when N availibility is high, like after sludge or slurry spreading.

Introduction

Winter oilseed rape is currently sown in France from 20 August to 10 September depending on region and it is known that the earliest sowings are of most interest for high N absorption in autumn. The preceding crop is almost always a cereal crop, harvested in july, 4 to 10 weeks before rape sowing. Very early sowing, one month before the current date, could be of interest to improve economical, energetic and environmental balancesheets of the crop. With a longer duration of growth from sowing to winter, a higher potential of growth and N uptake may be expected. Thus, very early sowing may be used with more efficiency than normal sowing as a N catch-crop before winter and this could reduce soil mineral N before winter drainage and so the risk of N leaching. Moreover, N accumulated in plants after winter has to be deduced from the N fertilizer rate in spring according to Reau et al. (1996). Therefore, very early sowing may be interesting to spare N fertilizer as they absorb more N in autumn.

We conducted a field experiment in controlled conditions to measure the potential N uptake in autumn of winter oilseed rape. We also conducted experiments under farmer field conditions in order to verify the interests of very early sowing for N uptake and reduction of N leaching risk in farmers fields, and to make a diagnosis of limiting factors of very early sowing technique in order to define situations and regions of most interest.

Materials ans methods

A field experiment was conducted in Grignon (France) with the cultivar Goëland (Brassica napus). We present results only from 1996/97, the second of a 2 year experiment. We compared 3 sowing dates, named Sowing 1 (S1: 15 July), Sowing 2 (S2: 03 August) and Sowing 3 (S3: 03 September), in a randomized blok lay-out with 3 replicates. Sowing 1 and 2 are very early sowing dates wheras sowing 3 is a normal sowing date. The crop was given 300 kg N ha^-1 at sowing. We verified N was not limiting for N uptake in autumn, according to the critical and maximum N content curves, determined for rape by Colnenne et al. (1998). The crop was fully protected against pests and diseases, and irrigated from sowing and early autumn to ensure rapid and homogeneous emergence and no water limitations. Each week, we gathered the dead leaves which have fallen on a plastic wire netting. We regularly measured the dry matter and N content of root, shoot and gathered dead leaves, and the green leaf area index (GLAI).

During 2 years we also conducted 28 experiments in a farmer field network with the same cultivar Goëland. We compared 2 different crop managements based on very early sowing and normal sowing dates respectively, in large plots (1000 m²) without repetition. Normal sowing were conducted according to regional recommandations. The goal with the very early sowing treatment was to sow 1 month before the normal one. Spring N fertilization rates were adjusted for each treatment to N accumulated in plants after winter, (Reau et al., 1996). Experiments were made in 4 french regions with contrasting climate and with different crop rotations, soil types and depth. Four experiments were receiving slurry or sludge at the end of July, just before the very early sowing; in this paper we present more results on these experiments because they have high risk of N leaching in winter. We measured soil mineral N, accumulated plant N and plant density before winter, after winter, and just before harvest, and yield.

Results and discussion

Leaves production and losses and autumn N uptake in experimental plots

We measured very high values of GLAI during autumn, which reached 4 for sowing 1 and 3 and more than 5 for sowing 2 (Figure 1). For each sowing date, values above 3 were reached very quickly after sowing and they remained high for very early sowing during autumn. At the end of autumn and especially in winter, which was very cold (minimum temperature of –13.3°C), GLAI strongly decreased and was almost nil after winter for all of the 3 sowing dates.

Accumulated shoot N reached very high levels before winter, more than 200 kg ha^-1. It was strongly correlated with GLAI before leaves began to fall (Figure 2). This relation has often been observed for initial vegetative growth because leaves, and especially limb, is the major constituant of plants and there are few roots, stems and petioles. After leaves began to fall, there was no more relation between GLAI and accumulated shoot N because as plants became bigger, they accumulated more N in stems and petioles and their leaves became thicker, leading to lower values of GLAI with larger amounts of N in the rest of the plant.

In this study, leaves fall appeared to be an important characteristic of rape functioning in autumn and winter, as shown by the GLAI decrease (Figure 1), therefore we studied precisely the quantity of N lost by dead leaves (Figure 3). The first losses were very early after sowing (800 °day). Before the beginning of winter, losses were important only for the 2 early sowings as they reached around 100 kg ha^-1 against 25 kg ha^-1 for sowing 3. During winter, losses were high for all sowing dates, ranging from 75 to 100 kg ha^-1. In fact, leaves fallen in autumn were senescent and had low N content whereas leaves fallen in winter were frozen and still green, their N content was therefore higher since they could not remobilize N to the rest of the plant before falling.

When estimating the total N uptake it is therefore needed to take into account not only root and shoot N but also the cumulated N of dead leaves (Figure 4). Sowings 1 and 2 reached as high values of 260 kg ha^-1 when the sowing 3 was still emerging. As N uptake curves of sowing 2 and 3 were almost parallel, the delay of N uptake of sowing 3 was not make up. In the beginning of winter , sowing 3 could absorb 200 kg ha^-1, which 120 to 150 kg less than the very early sowing. In conclusion, we have demonstrated the ability to absorb very high quantity of N in autumn, and the interest of earlier sowings to strongly increase this ability. This rapeseed characteristic may be explained by its ability to produce large amount of leaves in early autumn (Vos et van der Putten, 1997). We have also analysed if the interest of very early sowing could be of value under farmer field conditions in various climates in the second experimental lay-out.

Farmer field-trials: diagnosis for implantation and pests in autumn

Planned sowing dates were generally respected, in very early sowing particularly. But in Poitou, most of the normal sowing dates were delayed for sowing. The lack of rain in the late summer in both experimental years, followed by abundant rainfalls in September 1995 caused these delays of the sowing dates and in 5 experiments, 'normal sowing' were made in the end of September. The very early sowings were always made under dry conditions and germination never started before a rain. In some of the experiments, seeds remained a long time in dry soil, but without reduction in emergence rates. The duration from sowing to emergence and the emergence rates were very variable within experiments, equally variable in both very early sowing and in normal sowings. In very early sowing, plants rapidly reached a stage of 2/3 leaves after emergence and they were able to survive even 3 or 4 weeks of high temperatures and without rains after emergence, probably thanks to a rapid deep rooting. Thus, plant density was never limiting for growth in autumn.

In the first year with very high slug pressure, there were almost no treatments made against slugs and no damages observed in very early sowings whereas in normal sowings, many treatments were necessary. However, important damages could not be avoided in 2 fields, making re-sowings of rape necessary for them. Indeed, when slugs began to be active, rape was not sensitive anymore in very early sowings, contrary to normal sowings. On the other hand, it was necessary to apply 20% more insecticides in very early sowing, but there were no insect damages. Last, more weeds emerged in very early sowings, especially summer weeds and cereal volunteers, but it led to increased post-emergence herbicide applications in only 5% of the experiments. Moreover, in high available N situations we observed that rape was rapidly covering the soil and was able to limit weeds emergence and to increase weed mortality.

Farmer field-trials: autumn N uptake, soil mineral N and fertilization rate

In the 4 experiments receiving slurry or sludges in summer (Table 1), very early sowing (S1) have accumulated large amounts of N and have systematically reduce soil mineral N below 35 kg ha^-1. The highest value of N accumulation is consistent with the one measured in experimental plots. On the contrary, N accumulation was much more variable for normal sowing (S2). In West part of France (region Poitou), late sowing or emergence were associated to especially low N accumulation, lower than 30 kg ha^-1, leading to large amount of soil mineral N, ranging from 130 to 250 kg ha^-1. As demonstrated in experimental plots, measurements of accumulated N in the beginning of winter underestimates N absorbed because of leaf fall, especially in very early sowing.

That may explain why the normal sowing had a higher accumulated N in the South. The lower amount of soil mineral N in very early sowing may proove that it has absorbed the same amount of N or even more than the normal sowing. Simulated N leaching was almost nil for very early sowing whereas it ranged from 20 to 130 kg ha^-1 for normal sowing.

As N accumulated in plants after winter has to be deduced from the N fertilizer rate in spring according to Reau et al. (1996), it was possible to spare N fertilizer in 2 very early sowings. However, there were no differences of accumulated N after winter and consequently no fertilization reduction in the others experiments because of winter leaf fall. As fertilization was adjusted, we measured no increase in soil mineral N at harvest nor in residues N content (data not shown). Last, yields were egal or higher with very early sowing. For the others experiments without organic matter spreading in summer, soil mineral N before winter was always below 30 kg ha^-1 in very early sowing, egal or slightly less than normal sowing (20 experiments), except for 5 normal sowings, with values ranging from 40 to 80 kg ha^-1. Therefore simulated N leaching was generally low and always lower for very early sowing. Moreover, there were few differences in spring N fertilization rates and yields.

Tableau 1. Some results for experiments with summer slurry or sludge spreading
(S1 = normal sowing; S2 = very early sowing)

Conclusion

In experimental conditions as in farmer fields, very early sowing are very interesting to improve the N absorption capacity of rapeseed. Therefore, it seems possible to strongly reduce risk of winter N leaching thanks to earlier sowing, what may be particularly interesting for situations with high N availibility in summer and autumn, as after summer slurry application. The South experiment shows that it is possible to get very efficient N catch-crop with normal sowing. But it was sown relatively early and got favorable conditions for emergence and autumn growth, i.e. rain and warmth. Therefore, these results demonstrate that the environmental interest for N is less hazardous with very early sowing. We verified that there were no major limiting factor of implantation, pests or yield elaboration in very early sowing, but some risk have to be more extensively investigated, as risk of frost.

Characteristics of winter rapeseed crop in autumn is to produce and to loose large amount of leaves over autumn and winter. Consequently, plants have high GLAI for a long period in autumn, which seems to be interesting to compete with weeds; which was studied specifically in another experiment (Dejoux et al., 1999a). On the other hand, as leaves losses are very large, their decomposition was also investigated because their N mineralization may influence rape spring N nutrition and eventually increase mineral soil N and risk of leaching (Dejoux et al., 1999b).

References

1. Colnenne C., Meynard J. M., Reau R., Justes E. et Merrien A., 1998. Determination of a critical nitrogen dilution curve for winter oilseed rape. Ann. Bot., 81, 311-317.

2. Dejoux J. F., Ferré F. et Meynard J. M., 1999a. Effects of sowing date and nitrogen availibility on competitivity of rapeseed against weeds in order to develop new strategies of weeds control with reduction of herbicides use. In "'New horizons for an old crop', Proc. of the 10th Inter. Rapeseed Congress", Canberra-Australia, 26-29/09/99. GCIRC (this proceeding).,

3. Dejoux J. F., Recous S. et Meynard J. M., 1999b. Decomposition of 15N-labelled rapeseed leaves, frozen in winter : evaluation of N mineralization and rape N-uptake in spring under field and laboratory conditions. In "'New horizons for an old crop', Proc. of the 10th Inter. Rapeseed Congress", Canberra-Australia, 26-29/09/99. GCIRC (this proceeding).,

4. Reau R., Colnenne C. et Wagner D., 1996. Nitrogen fertilizer needs of rapeseed in autumn. In "4th congress ESA", Wageningen, The Netherlands, ESA, 2, 588-289.

5. Vos J. et van der Putten P. E. L., 1997. Field observations on nitrogen catch crops. I. Potential and actual growth and nitrogen accumulation in relation to sewing date and crop species. Plant Soil, 195, 299-309.