Page 110 - Climate Change and Food Systems

climate change and food systems: global assessments and implications for food security and trade
Results show that household socio-economic characteristics like farming experience, access to free extension services, credit, mixed crop and livestock cropping systems, private property and perception of climate change are expected to have a significant positive impact on use of adaptation measures at the farm level.
Asfaw et al. (2014) use a multivariate probit model at household level data to examine
maize farmers’ adaptation strategies and crop productivity under climate variability in Malawi. The authors distinguish between exposure, sensitivity and adaptive capacity vis-a-vis climate disruptions and find that exposure to delayed onset of rainfall and to greater climate variability (measured through the coefficient of variation of rainfall and temperature) is positively associated with the choice of risk-reducing agricultural practices such as tree planting, legume intercropping, and soil and water conservation. However, the application of inorganic fertilizer
in maize plots (already high due to government subsidies) is reduced due to uncertain risk reduction benefits. Farm wealth and more secure land tenure are positively associated with higher adaptive capacity. The use of both modern
and sustainable land management practices is positively correlated with higher maize yields.
An important insight from the study is that community (system-level) adaptive capacity is also important and highlights the key role of
rural institutions, social capital and supply-side constraints in governing selection decisions for all farm practices examined.
Among the disadvantages of the econometric approach is the inability to trace transmission mechanisms of specific adaptation measures or
to isolate the marginal effect of these strategies
or measures. Moreover, the findings are not easily transferable to other contexts (e.g. an African study does not apply elsewhere), and the statistical results can be difficult to interpret under multiple possible outcomes (Schlenker et al., 2005).
The simulation approach, by contrast, traces costs and benefits of adaptation strategies through particular mechanisms of interest, typically through
climate-biophysical-economic linkages. The economic component of the simulation analysis take one of two pathways:
(i) decision makers are rational actors who consider the benefits and cost consequences of their choices and pursue economically efficient adaptation outcomes (optimization models); or
(ii) application of a decision-rule characterization of the response of actors to climate stressors (scenario-based models) (Dinar and Mendelsohn, 2011; Schlenker et al., 2006). Van Wijk et al. (2014) reviewed a large number of simulation models that apply rule-based management implemented either through rules or through model parameter settings and found that for these scenario (“what if”) type models, adaptation can mean very different things depending on the goals, scale and scope of the model (Bell et al., 2014). In the case of optimization models, adaptations (as well as extreme events) may be modelled implicitly (as production responses to shifts in input costs and/or output prices) or explicitly (as choice variables with empirically derived cost functions), but without treatment of adaptations as discrete responses to discrete events or perceptions (De Bruin et al., 2009; Patt et al., 2010).
A major disadvantage of simulation modelling is the high demand for data inputs and calibration. Where data and models are available, farm-level simulation models work well and can perform a wealth of simulation options, such as estimating the incremental change in crop output and
water supply in response to changes in climatic conditions and agricultural and water resource management techniques. Another advantage is the opportunity for stakeholder involvement at several stages of the analytic process: designing scope, adjusting parameters, selecting inputs, calibrating results and incorporating adaptation measures of specific local interest (Dinar and Mendelsohn, 2011).
90