It shows that, on average, the amount of nitrogen lost by leaching in grassland (mowed or grazed, permanent grassland, mowed temporary grassland or lucerne) is low, even very low (20 to 30 kg/ha) compared to annual crops. It remains relatively important compared to forests where nitrogen losses by leaching are nil.
Two elements of management are particularly important. On one hand, there is a level of nitrogen fertilisation, and on the other hand, the harvesting method.
The graph alongside, proposed by Benoît and Simon in 20042) on grass crops managed either as pastures or by mowing, shows that the amount of nitrogen leached by grazed grassland is still higher than by mown grassland. It also shows that loss of nitrogen by leaching becomes significant in pasture when fertilisation reaches 200 kg/ha/year.
In the case of mowing crop, the leaching becomes significant beyond 300 kg/ha/year.
The cause for this difference is the level of nitrogen restitution by animals at the time of use as pasture. While with mowing, the nitrogen is expelled into the air away from the plot.
The difference of nitrogen loss when working in a grassland combining grasses and legumes should be stressed. In this case, nitrogen fertilisers are absent, the legume ensuring the nitrogen nutrition of the whole cover.
In this context, very little nitrogen leaching is observed, even in the case of pasture. In fact, in this configuration, the nitrogen restitution by animal manure induces an inhibition of the symbiotic fixing of legumes.
An illustration of this process of mineralisation was proposed (on lucerne) by Justes and Thiébeau (graph alongside). During the experimentation, lucerne was returned with or without new growth.
The control in this experiment was a beetroot. The study shows that:
in the case of lucerne returned after a re-growth of 30 cm, the cumulated mineralisation is of approximately one hundred units of control more than the control beetroot
the reversal of lucerne immediately after the harvest, thus without re-growth, reduces by approximately 50 units the quantity of mineralised nitrogen, this effect occurring shortly after the reversal (young re-growth tissues mineralise very fast)
the consequences on the mineralisation of grassland reversal must be considered over a period of two years after the reversal, the nitrogen mineralisation of all root structures being particularly long
There are other examples quantifying the amounts of mineralised nitrogen during the reversal of grassland.
Orders of magnitude reaching 500 units of nitrogen were measured in the case of rye-grass/white clover grassland mixtures in Brittany.
In fact, it is on this level that it is possible, by a valorisation of effluents and use of all food resources, to optimise the nitrogen balance of farms.
A synthesis conducted by Simon et al. in 20003) proposed an estimate of the excess nitrogen to a range of 200 farms with different farming systems. This study demonstrated that:
Farms based solely on cereal production have the most favourable nitrogen balance with low surplus. This is related to their capacity to precisely adjust the fertilisation of this type of environment.
Among livestock farms, mono-gastric farms (pigs and poultry) have the highest nitrogen surpluses.
Herbivorous livestock farms are located in an intermediate position. In these configurations, the farms involving livestock and cereal crops (or livestock and mixed farming) have a more favourable nitrogen balance. In fact these configurations allow for the best use of livestock effluents and a production of additional sources of animal feed on the farm.
To reflect the limitation of nitrogen losses, it is also necessary to take into account the scale of agricultural land.
These scales are particularly difficult to address but this can be done through two different methods: modelling and experimentation.
Modelling can be used to simulate a virtual Drainage basin. The work published by Ruiz et al. in 20024) was used to test the consequences of various organisations of production at the drainage basin level on the exportations of various crops and the losses at the emissary (on the river level).
Three drainage basins were simulated, associating two types of cultivation: highly fertilised crops (N +) and low fertilised land occupancy (N-) which can be considered as grasslands.
The three simulated drainage basins include:
for the first, fertilised crops along the river and less fertilised space on top of the basin
for the second, land occupancy in checkerboard
for the third, the potential source of pollution (N +) is on top and low fertilised areas along the river.
The analysis of nitrogen outputs at the level of the drainage basin shows that, when the source is on top, the total amount of nitrogen exported, and thus measured, are greater (180 units / ha / year) and losses at the emissary are significantly reduced (from 150 to 98 units/ha/year).
It is also makes sense to underline the very great reduction of nitrogen losses by denitrification.
In parallel with this analysis at the virtual drainage basin level, studies have been conducted on the drainage basin of Fontaine du Theil (35, France). This limited-size basin is heavily equipped and located on an impermeable Schist substrate.
In this drainage basin, the nitrogen content at the emissary is, in this study, at 58 ppm.
To reduce this nitrate content, three scenarios of basin occupancy were simulated and tested:
Optimising fertilisation to adjust it to the needs of different crops and plants of the basin
Conducting a thorough optimisation, which means fertilisation at 25% below the needs of the crops
Converting a part of the annual crop surfaces into grassland.
The third option leads to a significant decrease in the nitrate content of the water at the emissary. The important point of this study is to highlight the time required for the development of water quality at the level of a drainage basin.