Driving factors of nitrate leaching in arable organic cropping systems in Northern France

Stockless organic cropping systems are increasing in regions of intensive arable farming. A key component of their environmental performance is nitrogen leaching. Here, we aim to quantify nitrate leaching in organic systems and to understand its driving factors, related to crop rotation and management, in an on-farm context. Aided by an organic farmers-network, we monitored 35 organic fields representing 11 cropping systems (CS) during three years, to acquire data on crop production, N uptake, soil water and mineral nitrogen contents. Combined with the LIXIM model, these data allowed us to calculate water drainage and nitrate leaching below the rooting depth. Calculated nitrate-N leaching ranged from 3 to 46 kg N ha(-1) yr(-1), with a mean of 15 kg N ha(-1) yr(-1), similar to conventional CS in the same region. N-leaching varied highly between fields and was related to the preceding crop and to the autumn land use, but not to N surplus. Splitting up crop rotations into 9 types of crop sequence proved useful to characterize the variability of nitrate leaching at the field scale. The highest leaching (33-37 kg N ha(-1) yr(-1)) was found in the sequence 'grain legume-winter crop' and 'alfalfa destruction-winter crop' and the lowest losses (2-8 kg N ha(-1) yr(-1)) occurred under living alfalfa or clover cover crop. Thus, this study revealed the equivocal role of legumes as preceding crop, with results strongly depending on their date of destruction. Further, an ad hoc meta-model was calibrated to upscale the prediction of N-leaching to the CS-level, represented by three fields during three years. This model uses easily available inputs: mean amount and distribution of soil mineral nitrogen in the soil profile at mid-November (at the crop sequence level), and soil water turnover rate during autumn and winter (at the soil-year level). The upscaled meta-model predicted the N leaching of CS with a Root Mean Square Error of 4 kg N ha(-1) and a model efficiency of 0.85 (Nash coefficient). This meta-model result revealed two main interactive driving factors of N-leaching at the CS scale: frequency of crop sequence and pedoclimate. In conclusion, to improve the N use efficiency at CS scale, it appears necessary to optimize the succession of crop sequences in a multicriteria approach including the soil sensitivity to leaching, under the rotation constraints defined at the farm scale.