Main

However, the effect of any climate variability signal also depends on the temporal scale (for example, annual versus decadal) of such variability. For wheat crop yields, the relevant timescale for climatic variability is over years or even days, rather than decades, simply because wheat is an annual crop. Hulme et al. decided not to alter either inter-annual or inter-daily climatic variability.

By using an earlier global climate model (UKTR; ref. 2) than Hulme et al., we have analysed the response of a wheat simulation model3 to scenarios of climate change with and without changes in the inter-annual variability of precipitation (intensity and occurrence) and temperature. We introduced changes in climate variability derived from UKTR using a stochastic weather generator4. To measure the effect of inter-annual climate variability, we studied mean grain yield and its coefficient of variation (CV) from 30 individual years of simulation (the same total period as used by Hulme et al.).

We looked at two of the same sites as Hulme et al. For Spain, the simulated mean for the baseline climate wheat yield was 5.6 tonnes per hectare, with a CV of 0.24 (Table 1). In the climate scenario without a change in inter-annual variability, mean yield fell slightly to 5.2 t ha−1, with a CV of 0.23. With changes in climate variability, simulated mean yield dropped to 3.9 t ha−1 but its CV more than doubled to 0.48, largely because there were more prolonged dry spells over the vegetation period. The probability of producing yields of less than 3.5 t ha−1 in the ‘with variability’ scenario was nearly 0.50, compared with about 0.10 for the baseline and ‘without variability’ scenarios. Such changes in annual yield variability would make wheat a risky crop to grow in Spain and have important economic and social consequences.

Table 1 Mean grain yield with and without changes in inter-annual climate variability
Full size table

For the United Kingdom, changes in climate variability had little effect on either mean grain yield or its CV (Table 1). In contrast, Hulme et al. found no change in the range (and hence the variability) of simulated yields for these sites as a result of either multi-decadal natural or anthropogenic variability. The timescale of the imposed variability can therefore alter qualitatively whether or not variability has an effect.

We conclude that it is important when assessing the impact of climate change to differentiate between natural climate variability and anthropogenic climate change, as highlighted by Hulme et al., but also to apply changes in climate variability at appropriate timescales.

Reply