Page 152 - Climate Change and Food Systems

climate change and food systems: global assessments and implications for food security and trade
figure 10
Current European (EU-27) total demand and production of cereals (FAOSTAT, 2014)
and projected potential future cereal demand and production (by 2050) illustrating production surpluses and export potential. Future demand is projected to increase by 35% relative to the year 2000 (see von Braun, 2008; Msangi and Rosegrant, 2011) and future production is projected to increase by 50% until 2050
4. Discussion of uncertainties
In this section we first provide an overview of
the various uncertainties involved in CC impact projections. This is followed by a discussion of their relative importance, and examination of uncertainty that is numerical and can be quantified, with a couple of known “unknowns”.
Figure 1b attempts to illustrate schematically the propagation of uncertainties and errors in CC impacts along the impact modelling chain. Figure 1 does not explicitly include “socio-economic scenario uncertainty,” which is usually large and increases more rapidly into the future the longer the time horizon. Such scenario uncertainty in turn results in uncertainty in GHG emission scenarios, which are used in GCMs that have their own inherent uncertainties (Räisänen and Räty, 2012; Rummukainen, 2012, 2014). As GCMs (usually with grid boxes of 150 to 200 km resolution) are too coarse for agricultural impact assessments, statistical or dynamic downscaling methods (e.g. RCMs) must be applied in order to produce climate
scenario data as input for impact models such as crop simulation models or other techniques for estimating suitability and productivity for agricultural crops and livestock. However, these impact models come with their own uncertainties (Walker et al. 2003; Palosuo et al., 2011;
Asseng et al., 2013). Eventually, results from biophysical impact models are fed into global/ regional economic and trade models or bio- economic farm type or regional land use models, creating a further dimension of uncertainty, finally resulting in a considerable uncertainty range. This has previously been labelled the “uncertainty cascade” (Jones, 2000).
Summarizing the scientific-technical challenges for crop modelling, Rötter et al. (2011a) identified impact uncertainty analysis as one of four research areas that should be addressed to overcome deficiencies of current impact assessment methodologies. One of the main goals of CC impact assessments is to give a thorough account to decision-makers (risk managers) of the level
of certainty of model-based impact simulations.
132