From Food Production to Food Security - Global Environmental ...

provide sufficient representation of, for instance, (i) different governance arrangements (e.g.land tenure); (ii) different land management (e.g. principal farming systems); and (iii) keydrivers in regional scenarios (e.g. demography).Third, establishing institutional buy-in at the regional level can be difficult, as it includesnegotiating with a potentially wide range of research partners (government labs, universities,NGOs and private sector) and donors. Research on food security, and especially whencoupled with GEC issues, is complex and full of uncertainties. It can take time for results tobecome apparent, which challenges timeframes for funding and the information needs of the‘client’. Co-planning helps all partners grasp these realities, and take an active interest in theproject. An example of how and why different communities ‘buy-in’ to a research agenda ingiven in Paper 6 (Box 5).Encouraging regional research networks helps overcome many of these challenges asmembers of the network bring different skills and contacts to the other stakeholders. Teambuilding can be promoted through the adoption of standardized methods, and the use of theparticipatory scenario exercises, as discussed above. The use of scenarios methods can alsohelp stakeholders to see the links between spatial levels (global to regional; regional to local),although ‘downscaling’ and ‘upscaling’ scenarios both present their own methodologicalchallenges.The scientific contribution of integrating the approaches: from traditional agronomy toproduction ecology and agroecology to ‘food system ecology’Lessons from production ecology, agroecology and human ecologyAs discussed in the Introduction and Paper 2, much of the work on food security stems fromagricultural science generally, and empirical agronomy in particular. Such work has made amassive contribution to enhancing food production and removing the threat of famine formany, and its importance both historically and into the future is clear. The interest in gaininga more mechanistic understanding of food production (and crop growth in particular) hashowever led to the development of ‘production ecology’(van Ittersum and Rabbinge, 1997)and the broader concept of ‘agroecology’(Dalgaard et al., 2003). Both have emerged overrecent decades as key areas of research and much energy has been directed towards thedevelopment of mechanistic models to both explore the impact on crop growth of scenariosof changing environmental conditions and also to challenge theories thereby enhancingmodel structure.Production ecology initially considered limitations to crop growth, i.e. what determinesproductivity, or ‘yield’. Attention was first placed on the ‘crop system’ and in particular oncrop characteristics, radiation, temperature and CO 2 as factors that fundamentally define the‘potential’ crop growth. Water and nutrients were then incorporated as factors that limit cropgrowth (i.e. the ‘attainable’ yield within the ‘cropping system’). Pests, diseases, weeds and113