Spatial and temporal dynamics of gene movements arising from deployment of transgenic crops

Abstract

Many of the pressing questions about whether or not to release a genetically modified crop can be resolved into population-dynamics questions concerning invasion and persistence of the transgenic crop itself, or of hybrids with other crops and wild relatives. Progress in assessing risk demands a coherent theoretical framework within which questions such as the reversibility of breakdown in pest resistance or ‘escape’ of herbicide-tolerance genes can be phrased and tested. Here we begin to sketch out a framework that provides some answers while taking account of the inherent spatial and temporal variability of agricultural and semi-natural systems. After a brief summary of the risks associated with the deployment of novel crop varieties, we define what we mean by invasion and persistence in variable environments. This leads to a discussion of stochasticity and spatial scales in which we distinguish between different sources of variability that affect the probability of invasion and persistence times. We illustrate this with the distributions of global extinction times of crop plants or hybrids that persist as local patches in the landscape following the introduction of a new variety. Empirical studies of persistence of transgenes are next discussed. We begin with pollen dispersal, which leads to the dispersal profiles from point sources from which separation distances can be calculated to minimize contamination between transgenic and conventional crops. Three important aspects that emerge from some work on the dynamics of feral populations of transgenic crops are identified: feral patches of some transgenic crops are transient with short persistence times; there is large environmentally-driven variability in ecological performance; the seed bank is an important reservoir from which plants may emerge even after a patch appears to have gone extinct. Next we show how such variability can be incorporated into stochastic life history models, from which it is possible to identify intrinsic, genetically controlled properties of crop plants that favour persistence. These analyses suggest how to design new crop varieties with characters that minimize the risk of spread and persistence in the landscape. Finally, we introduce landscape-dynamic models in which we show, using individual-based percolation and metapopulation models, how the spatial arrangement of crop plants and feral patches of ‘escaped’ plants in the landscape influence invasion and persistence