F. Vertès and J.C. SimonStation d'Agronomie INRA, 4 rue Stang Vihan 29000 Quimper, France
Introduction
Materials and methods
Representativeness of the lysimeters
Water drainage and nitrate leaching losses
Nitrogen balance
Discussion
Conclusion
References
Grassland soils have pools of inorganic nitrogen (nitrate and/or ammonia) derived from fertiliser, legume symbiotic fixation, atmospheric deposition, excretal deposition and soil organic matter mineralisation. It is rare to observe large amounts of nitrate in grassland soils, unless in the small areas on which animal excreta has been deposited where there can be the equivalent of 50 to more than 800 kg N/ha locally.
Part of this N returned to the sward that is not uptaken by herbage growth can be volatilised or quickly nitrified, and so become available for loss by denitrification or leaching. The lysimeter experiment described here investigates some of these N pools, and fluxes between them, under grazed pastures.
The lysimeter experiment established in Quimper in 1989 compared the leaching losses under pure ryegrass receiving 250 kg N/ha/year and a grass-clover sward without nitrogen. In each plot (400 m2) 13 lysimeters equally spaced were buried (0.358 m2 x 85 cm height, 2 controls and 11 grazed). The number was a compromise between statistical constraints (Table I) and resources available.
Table 1. Dung and urine voided by dairy cows (Peyraud, pers. comm.) and by heifers (experiment described). Hypothesis: Grassland production: 10 t/ha, daily intake: 15 kg/cow, 8 to 11 kg/heifers.
|
|
Dairy cows (mean) |
Heifers (1991-92) |
||
|
Dung |
Urine |
Dung |
Urine |
|
|
VOLUME (litres/ha/year) |
|
26680 |
|
10450 (rg) |
|
|
|
|
15700 (ass) |
|
|
WEIGHT (t/ha/year) |
20 |
|
11 (rg) |
|
|
|
|
12 (ass) |
|
|
|
N CONTENT (p. 1000) |
4 |
6.2 |
3.86 (rg) |
7.4 (rg) |
|
|
|
3.77 (ass) |
7.4 (ass) |
|
|
N TOTAL (kg/ha/year) |
80 |
165 |
42 (rg) |
80 (rg) |
|
|
|
45 (ass) |
115 (ass) |
|
|
NUMBER/ha/year |
9340 |
|
7180 (rg) |
|
|
|
|
7370 (ass) |
|
|
|
|
8000 |
|
6600 (rg) |
|
|
|
|
|
9350 (ass) |
|
|
LYSIMETERS (amount/m2) |
2 kg |
2.61 |
0.62 kg |
1.251 |
The plots were grazed by heifers (7 heifers, 6 times per year). The observers collected data on plants (yield, clover content, mineral composition), soils (mineral nitrogen content), animal returns (distribution, quantity and chemical composition) and water (drained volumes and mineral composition). 15N is used to quantify N2 fixed by clover, to investigate how fixation is modified by urine or dung deposition, and to estimate indirectly the amount of mineral nitrogen available for the plants during each regrowth (Vertès, 1992).
The total lysimeters surface is 1.05% of the total area. They received 2.8% (1991) and 2.5% (1992) of the total number of depositions (more than the 1% expected according to Murphy, 1986). All dung and urine depositions were recorded on maps, and their distribution appeared quite homogeneous. As shown in Figure 1, the representativeness of lysimeters during 1991 and 1992 was good except for urine deposition on pure ryegrass: the lysimeters received more urine (1991), or received it close to the drainage period (November 1992) and then quickly leached.
DUNGS on PURE RYEGRASS cumulated 1991-92
DUNGS on MIXTURE, cumulated 1991-92
URINE on PURE RYEGRASS cumulated 1991-92
URINE on MIXTURE, cumulated 1891-92
Concerning all of the inputs, the two types of swards received the same amount of dung but the grass/clover plot received 40% more urine than the pure grass plot (nevertheless the lysimeters received the same quantity of urine for both swards during 1991 and 1992).
During the first winter, drainage was measured under bare soil in the 26 lysimeters, to control homogeneity: variation coefficients were low 5% for pure grass (583 mm) and 3% for mixture (521 mm). Table II illustrates the results of the following years.
Table II: Comparison of water drainage and nitrate leaching under the two swards (1990 to 1993, mean and s)
|
|
1990-91 |
1991-92 |
1992-93 |
|||
|
rg |
ass |
rg |
ass |
rg |
ass |
|
|
DRAINAGE (mm/ha/year) |
495 (21.7) |
498 (16.8) |
398 (37.4) |
392 (15) |
499 (17.4) |
512 (20.9) |
|
N LEACHED (kg/ha/year) |
20.7 (9.7) |
28.3 (5.1) |
23 (54.3) |
15.8 (31.9) |
68.3 (100.6) |
11.6 (32.3) |
The variation coefficients are low for drainage data (3.4% to 9.5%) and N losses data, in grazed lysimeters without urine deposition (5 to 32%). On the other hand they are quite high for the mean leaching losses (all lysimeters), due to the very different amount of urine deposition (20 to 280%).
Nitrogen balance is calculated as follows:
N balance = + fertilisation (known) + animal deposition (known or estimated) + fixation (known with 15N analyses) + atmospheric deposition (known) - N uptake by grazed grass (known with low precision) - N leached (known) - gaseous N losses (unknown).
Without urine deposition, the pure ryegrass lysimeters had the following balance (1992): + 250 (fen) -270 (+-27) (N upt) - 3.1 (+-3.3) (N leach). With urine the results vary between + 7 and -120 kg/ha/year (for pure grass). The mixed sward results will not discussed here.
Few lysimeter measurements are carried out in real grazing conditions in Europe. The calculation and interpretation of all the data has not the same accuracy and possibility to be extrapolated. Generally the methodology provided reliable data for i) drainage, mean leaching losses under the 2 plots ii) improvement of knowledge concerning animal returns (amount, composition and distribution in a grazed plot) iii) good measurement of the effect of animal returns on clover N fixation (15N) iv) good indirect estimation of mineral nitrogen available for pure grass during one regrowth (fertiliser + urine + mineralisation, 15N).
The effects of animal returns on plants, soil and water vary with type (dung or urine), vegetation (pure grass or grass/clover), amount and period of the year. Dung depositions result mainly in ungrazed areas and changes in mixture equilibrium. No short term effect was observed on N losses or on clover N fixation. The effect of urine results in a significant increase of grass growth and nitrogen content as well as a significant decrease in clover N fixation, mainly in early spring or summer and autumn. During the more favourable growing period in spring, the grass can use available nitrogen very efficiently; no significant increase in leaching losses was measured in lysimeters having received a quantity of urine less than 250 kg/ha.
Edaphic and climatic caracteristics may help to interpret the results: free draining soils with low denitrification, drainage important: 550 mm mean for winter. The whole profile is usually leached (beyond 350 mm) and no residual mineral nitrogen is observed at the end of winter.
Nevertheless, annual nitrogen balance on each lysimeter appears difficult to interpret:
- inaccurate measurement of nitrogen uptake: the microplots cut before each grazing period overestimate the animal uptake; mortality of some leaves or stolons may occur during long rest period (rejected areas, difficult to manage). A more accurate comparison is in progress with the total amount of dung and the analysis of herbage digestibility in both swards. Few studies concern the characteristics and evolution of the areas affected by one deposition (Hippolyte, 1984).- methodological problems in 15N data interpretation (Witty, 1983)
- urine deposition in lysimeters is not always caught (and analysed) and thus input is not always known accurately (large variation of urine N content with animals, hour and sward type).
- lack of knowledge on the gaseous exchanges, mainly volatilisation (Jarvis, 1992).
- effect of animal returns on mineralisation/immobilisation are not measured (soil mineral nitrogen content is measured with auger sampling throughout all the year).
Until now, the losses under grazed pastures are low, in relation to moderate fertilisation, good use of nitrogen by young grass and regulation of fixation by clover. Local losses due to excretal returns may be high, due to the amount and the climatic conditions (use by grass, involvement in microbiomass, washed down).
The experiment attains its objects and asks new questions that imply the need for more precise measurements under more controlled conditions. These measurements could concern the gaseous losses, the soil biological response through microbiological activity to animal treading and poaching, the variation in growth potential of plants and the grassland "buffer capacity" characterisation.
JARVIS S.C. (1992). Nitrogen cycling and losses in clover based pastures. Herba 5, pp. 56-59
HIPPOLYTE C.(1984) Bull. CRZV Theix, pp. 11-17
MURPHY W.E. (1986). Nutrient cycling under grazing. BGS Symposium 19, pp. 14-19
VERTES F. (1992) - Nitrate leaching in grazed pastures: a lysimeter experiment using 15N enrichment as a means of determining nitrogen fixation and animals returns. "Third research conference British Grassland Society, pp. 161-162
WITTY J.F., (1983). Measurement of N2 fixation by 15N fertiliser dilution. Problems of declining soil enrichment. In Temperate legumes". DG Jones and DR Davies (eds)., Pitman Publishing INC, pp. 253-269
