² INRA Unité d'Agronomie – Rue Fernand Christ – F-02007 Laon CEDEX - France
email : recous@laon.inra.fr

ABSTRACT

For environmental purposes, very early sowing of winter rapeseed may reduce winter nitrate leaching thanks to high N uptake capacities of rapeseed in autumn. However, freezing could lead to high losses of leaf nitrogen, amounting to more than 100 kg N ha^-1 (Dejoux et al., 1999).

Here we investigated agronomic and environmental consequences of the decomposition of fallen leaves, based on both field and laboratory studies with ¹⁵N labelled leaves (C/N=9). Potential rate of decomposition was measured by incubation in the laboratory. In the field, all leaves were removed at beginning of winter and replaced by labelled leaves, artifically frozen at –15°C , which were laid on the soil surface.

Compared on a thermal time basis, decomposition rate proceeded as quickly in field as in the incubation. This was explained by the high N and the soluble C concentrations of the leaves. One third of the N was absorbed again by rape plants before flowering. This strong reabsorption is explained by the temporal synchronization of decomposition and of active N absorption by rape in spring. Leaf decomposition did not increase soil N mineral content at flowering or at harvest, but we observed a 40% loss of ¹⁵N. As no leaching was simulated, this loss was supposed to be gaseous. This large percentage could be explained by the placement of leaves during decomposition, on soil surface, and conducive climatic conditions.

For environmental purposes, the quantities and nature of N gaseous emissions have to be studied for other climatic conditions and types of leaves. As a proportion of N is reabsorbed, N fertilization rate could be accordingly reduced.

Introduction

Very early sowing of winter rapeseed is very efficient for N uptake in autumn as the crop may absorb as much as 300 kg ha^-1 and also for reduction of soil mineral N before winter drainage. However many leaves may freeze during winter and fall on soil surface (Dejoux et al., 1999). The N lost by frozen leaves may represent up to 100 kg ha^-1 (Dejoux et al., 1999). Much research is available concerning the fate of N of crop residues or incorporated catch crops (Watkins et Barraclough, 1996), but very little work deals with the fate of dead leaves (frozen or senescent) under an established crop (Berger et al., 1996). However there are major differences between the 2 situations. Dead rapeseed leaves lie on the soil surface and are whole whereas crop residues are generally chopped and incorporated. Moreover, there is no delay between start of decomposition and the subsequent crop establishment as the rape crop is already present.

Our aim was to characterize the fate of frozen leaves N, i.e. spring N recovery by the rapeseed crop and risks of N losses to the environment, which are determined by the decomposition rate and net N mineralization of the leaves. Indeed, the proportion of N which is mineralized but non absorbed by the crop presents environmental risks : gaseous emissions, leaching of the additionnal soil mineral N resulting from leaves mineralization, until and after harvest. Moreover, net N mineralization may be high since these leaves are still green and have a higher N content than senescent leaves (Trinsoutrot et al., 1999). We associated a field experiment approach with laboratory incubations in controlled conditions. We intended to characterize with the latter experiment the kinetics of leaf mineralization under known conditions (particularly soil temperature and moisture) in order to define the potential N availibility. This is necessary to extrapolate our results, particularly to other climatic conditions than those observed during the field experiment.

Materials ans methods

In the field experiments, rape was grown in greenhouse to produce homogeneously ¹⁵N labelled leaves at 2.53 (± 0.09) atom %¹⁵N excess. In the field, rape was sown on 9 August 1996 on a deep loamy soil in Grignon in the Paris Basin (France, 48.5° latitude North). The preceding crop was winter barley whose residues were burnt in July 1996. The crop was given 100 kg N ha^-1 at sowing and 90 kg N ha^-1 in spring as non-labelled ammonium nitrate and was fully protected against pests and diseases. Open-ended cylinders (13 cm deep and 40 cm internal diameter) were pressed on 15 October 1996 into the established winter oilseed rape crop. All leaves were removed on 11 December. Both labelled and unlabelled leaves were frozen in a freezer at -15°C to simulate natural frost and were brought back to the field on 27 December. ¹⁵N labelled leaves were placed within the cylinder, and non-labelled outside them. The equivalent of 74.5 kg N ha^-1 was added by leaves. Some of the leaves were preserved for incubation study.

Within and outside the cylinders, we measured on 6 replicates per sampling date: dry matter, N content and atom ¹⁵N% excess, on shoot, root, spring dead leaves and soil surface residues. In the soil we measured moisture, inorganic N (nitrate and ammonium content) and atom ¹⁵N% excess in the organic matter and the bulk soil. Soil and plants were sampled at 4 dates: just before defoliation at the beginning of winter (11 December 1996), at the end of winter (12 February 1997), at the onset of flowering (2 April) and at maturity (3 July). Sampling dates are named S0, S1, S2, S3. Recovery of ¹⁵N, expressed as percentage of applied ¹⁵N, is called Nitrogen Use Efficiency (NUE).

Laboratory experiments consisted in incubating during 24 weeks at 15°C frozen labelled and unlabelled leaves samples which were oven-dried (at 80°C), ground (to 1mm) and incorporated to soil sample. Water and mineral N were not limiting during all incubations. The control treatment received no leaves. CO₂ accumulation and soil N content were regularly measured in all treatments.

Results and discussion

Leaf characteristics

Labelled leaves were essentially made of soluble compounds which totalled about 80% of carbon and nitrogen in the leaves, as measured with the Van Soest method, (Table 1). Therefore, cellulose, hemi-cellulose and lignin were at very low concentrations. For example, lignin content was only 1 to 1.4% of dry matter, in agreement with results of Trinsoutrot et al. (1999) on rape leaves and lucerne aerial parts. Soluble carbon in 20°C water is the form of carbon that is the most available for microorganisms, and it represented 45% of leaves total carbon.

Table 1– Some chemical characteristics of the applied labelled leaves. Analyses were performed on frozen/oven-dried samples.

	% Carbone	% Nitrogen	C/N
	% dry matter	% dry matter
Whole leaves	41.7	4.5	9.2
	% C-residue	% N-residue
Water soluble fraction, 20°C	44.5	8.91	5
	% C-residue	% N-residue
Soluble *	79.3	81.6	8.9
Hemicellulose *	8.6	14.0	5.6
Cellulose *	10.7	1.3	75.6
Lignin *	1.4	3.0	4.3

* according to the Van Soest method

The fate of leaves N: ¹⁵N tracing appreciation

The major ¹⁵N uptake by rape from the late winter (S1) to early flowering (S2). Rape recovered 3% of the initially applied ¹⁵N at S1, 28% at S2 and 24% at S3 (Table 2). Therefore, ¹⁵N Use Efficiency (NUE) is relatively high compared to those of other authors: NUE by a subsequent crop of residues N are generally lower (Mary, 1987) except for N rich residues. The respective input of C and N in soil determines at short term the equilibrium between microbial immobilization and mineralization process, on which the net mineralization depends. The low C:N ratio of the leaves (9) has surely contributed to the high NUE. In our experiment, rapeseed crop which has lost its leaves is the subsequent crop, whereas rapeseed N demand after winter is indeed very high and early and its rooting system is already established. Therefore there is a good synchronization between the time courses of leaf N mineralization and N uptake by rape. This may favour absorption at the expense of immobilization or gaseous losses. Likewise Jensen (1994) measured for late summer incorporated residues a NUE of 15% when a winter crop is sown (winter wheat or barley), whereas NUE drops to 6% when a spring crop is sown (spring wheat or barley).

Table 2. ¹⁵N balance, expressed as the percentage of leaves applied ¹⁵N.

Variance analysis for each compartment, on sampling date evolution of ¹⁵N %

Sampling date	Residues 15N	Soil mineral 15N	Soil organic 15N	15N use efficiency (NUE)	Total 15N recovery
S0: theoritical point	100	0	0	0	100
S1	46 ± 2.6 a	24 ± 2.6 a	18 ± 2.5 a	3 ± 0.7 c	91 ± 3.8 a
S2	18 ± 1.8 b	5 ± 1.3 b	23 ± 3.2 a	28 ± 0.8 a	74 ± 4.8 ab
S3	6 ± 1.2 c	16 ± 4.2 ab	15 ± 2.2 a	24 ± 0.4 b	60 ± 6.0 b
Proba > F	0.00	0.00	0.12	0.00	0.01

Soil organic ¹⁵N was determined after soil surface residues were manually removed and independently analysed. It represents at most 23% of applied ¹⁵N at S2 whereas soil mineral ¹⁵N represents at most 24% at S1 (Table 2). Total nitrogen recovery (Table 2) is 91% at the end of winter (S1), then 74% at flowering onset (S2) and 60% at harvest (S3). As no leaching has been neither detected by ¹⁵N measurements nor simulated by a leaching model, the ¹⁵N inbalance should result from gaseous losses, which would be relatively high compared to other ¹⁵N studies with a monitoring of leaching. The major difference between our study and these lies in non-incorporation of frozen leaves, whereas surface decomposition would be favorable to gaseous losses (Breland, 1994). It leads in particular to reduce implicated N in organization processes and to increase surface mineral N; but an important gradient of soil mineral N is favourable to ammoniac volzatilization, derived from decomposition of rape leaf protein applied on soil.

Comparison of laboratory and field decomposition rate

Leaf decompositon was very quick in the incubations, since we estimated that leaf C was totally decomposed within 50 days (Figure 1). This is consistent with numerous studies performed on vegetal organs decomposition (Watkins et Barraclough, 1996) and may be explained by the leaves biochemical composition: the recalcitrant fractions were indeed in very low proportions (4 to 5% of lignin and cellulose) whereas 80% was in soluble form, and 45% was water soluble. Trinsoutrot et al. (1999) studied decomposition of seven field cultivated leaf samples and their water soluble C content was only 28% on average. Leaf freezing (at –15°C) contributed to increase leaf soluble C content and therefore accelerated decomposition. According to Breland (1994), freezing accelerates the liberation of soluble compounds by physical fragmentation of cells and structure attack (cellular walls). Berger et al. (1996) had effectivily measured on rape leaves that the water soluble compounds increases after frost.

Figure 1. Decomposition kinetics of N, estimated in field from ¹⁵N net mineralization, and of C, estimated from incubation C mineralization.

Field decomposition was indirectly calculated from ¹⁵N measurements on soil surface residues (cf. Table 1) and field experiment duration was converted in days at 15°C (normalised days for temperature) according to Aita et al. (1997), and super-imposed to laboratory decomposition kinetics, calculated from cumulated C mineralization of incubations (Figure 1). Both kinetics are very close: field decomposition rate was almost as quick as in incubation. However, incubation decomposition was theoretically more favorable than in the field because leaf grinding and homogeneous mixture of soil and residues are favorable to decomposition (Recous et al., 1998). It means that for leaves with high contents of soluble compounds and nitrogen, soil contact plays a minor role as observed by Bremer et al. (1991) for very fermentescible and N-rich residues.

We extrapolated this decomposition scheme via laboratory decomposition kinetics. For 9 of the 10 years in Grignon, more than 81% of N leaves would be decomposed between 1 February and the end of April. It means that frozen leaves N mineralization would almost always occur early enough relative to the active N uptake period of rape in spring.

Conclusion

Frozen rapeseed leaves decomposition is original because it occurs on soil surface, probably leading to high gaseous N losses, and reduced N immobilization compared to incorporation (data not shown). Moreover, these leaves decompose very rapidly, because of their initial biochemical composition and effects of the frost, and N is highly recovered, because of the synchronization of decomposition and rape N uptake. For environmental purposes, quantities and nature of N gaseous emissions have to be studied for other climatic conditions and types of leaves. As a proportion of N is reabsorbed, spring N fertilization has to be reduced, by 30% in this study. This percentage may vary with leaf quality, especially N content, but not with climatic conditions as decomposition should occur early enough.

Acknoledgements

We thank I. Trinsoutrot who has performed the incubation experiment, J. Troizier, G. Grandeau, JF. Leblond and E. Fovart for the field experiment, O. Delfosse and F. Lafouge for analyses, P. Leterme and B Mary for their scientific comments.

References

1. Aita C., Recous S. et Angers D., 1997. Short term kinetics of residual wheat straw C and N under field conditions : characterisation by 15N13C tracing and soil particle size fractionation. Eur. J. Soil Sci., 48, 283-294.

2. Berger G., Schmaler K. et Richter K., 1996. Effects of catch-crops on the dynamics of mineral nitrogen in the soil in winter and on the nitrogen conservation under special conditions of sandy soils. Arch. Agron. Soil Sci., 40, 217-229.

3. Breland T. A., 1994. Enhanced mineralization and denitrification as a result of heterogeneous distribution of clover residues in soil. Plant Soil, 166, 1-12.

4. Bremer E., van Houtum W. et Van Kessel C., 1991. Carbon dioxide evolution from wheat and lentil residues as affected by grinding, added nitrogen and the absence of soil. Biol. Fertil. Soils, 11, 221-227.

5. Dejoux J. F., Meynard J. M. et Reau R., 1999. Rapeseed new crop management with very early sowing in order to reduce N-leaching, N-fertilization. In "'New horizons for an old crop', Proc. of the 10th Inter. Rapeseed Congress", Canberra-Australia, 26-29/09/99. GCIRC (this proceeding).

6. Jensen E. S., 1994. Availability of nitrogen in 15N-labelled mature pea residues to subsequent crops in the field. Soil Biol. Biochem., 26, 465-472.

7. Mary B., 1987. Effets du précédent cultural sur la disponibilité du sol en azote minéral. C.R. Acad. Agric. Fr., 73, 57-69.

8. Recous S., Richard G., Fruit L., Chenu C. et Angers D. A., 1998. Factors affecting the contact between soil and incorporated crop residues : short-term effects on C evolution. In "16th Congrès Mondial de Science du Sol", AFES (ed.), Montpellier, France, 20-26/08/1998. 753.

9. Trinsoutrot I., Recous S., Bentz B., Linères M., Chèneby D. et Nicolardot B., 1999. Relationships between biochemical characteristics and C and N mineralization of crop residues grown in temperate agrosystems. Soil Sci. Soc. Am. J., soumis.

10. Watkins N. et Barraclough D., 1996. Gross rates of N mineralization associated with the decomposition of plant residues. Soil Biol. Biochem., 28, 169-175.