Environmental risk assessment of GM crops in Europe proceeds by step-wise estimation of effect, first in the plant, then the field plot (e.g. 10-100 m-2), field (1000-10,000 m-2) and lastly in the environment in which the crop would be grown (100-10,000 km2). Processes that operate at large scales, such as cycling of carbon (C) and nitrogen (N), are difficult to predict from plot scales. Here, a procedure is illustrated in which plot scale data on yield (offtake) and N inputs for blight resistant (both GM and non-GM) and blight-susceptible potato are upscaled by a model of crop resource use to give a set of indicators and metrics defining N uptake and release in realistic crop sequences. The greatest potential damage to environment is due to loss of N from the field after potato harvest, mainly because of the large quantity of mineral and plant matter, high in N, that may die or be left in the field. Blight infection intensifies this loss, since less fertiliser N is taken up by plants and more (as a proportion of plant mass) is returned to the soil. In a simulation based on actual crop sequences, N returns at harvest of potato were raised from 100 kg ha-1 in resistant to 150 kg ha-1 in susceptible varieties subject to a 40% yield loss. Based on estimates that blight-resistant types would require ~20% of the fungicide applied to susceptible types, introduction of resistant types into a realistic 6-year cropping sequence would reduce overall fungicide use to between 72 and 54% depending on the inputs to other crops in the sequence.
Keywords: Biogeochemical; Carbon; Cisgenic; Crop model; Environmental risk assessment; GM; Late blight; Nitrogen; Pesticide; Phytophthora infestans; Potato.