Hulme et al. reply — Porter and Semenov point out the importance of changes in climate variability on different timescales when assessing the significance of climate-change impacts. They are correct to assert that changes in inter-daily, inter-annual and multi-decadal climate variability are important, but our study1 was intended to draw attention to natural multi-decadal climate variability and to indicate how it might obscure the identification of significant effects of anthropogenic climate change. It was therefore primarily a study of signal-to-noise ratios in impacts indicators, something that has not previously been attempted using low-frequency climate variability. The effects on crop yields of changes in inter-annual and inter-daily climate variability have been considered previously2,3,4.
We do not agree with Porter and Semenov that the primary climate-variability timescale of relevance for wheat yields is the inter-annual. Changes in the daily variability of climate and in the frequency of extreme events can be very important for crop performance5, and daily climate variability can change independently of annual or decadal timescales. Porter and Semenov did not perturb daily climate variability in their study2. We think that decadal climate variability is also important for wheat yields: such low-frequency changes in climate (whether or not they are accompanied by changes in higher climate frequencies) can profoundly affect crop suitability and economic viability6 and alter the competitiveness of different agricultural products on these timescales.
The effect of anthropogenic changes in inter-annual and/or inter-daily climate variability on an impact indicator will depend on their magnitude with respect to natural changes (for example, for Europe, in relation to the North Atlantic Oscillation). This is also a signal-to-noise problem that can be explored using our methodology, but was not addressed by Porter and Semenov.
The challenge for those studying the impacts on climate is to examine the effects of changes in the full spectrum of climate variability, clearly distinguishing between anthropogenic and natural changes, and interpreting these effects in the light of other forces of environmental change. If these can be quantified more comprehensively than in the past, then we can provide decision-makers and resource managers with information that can better assist them to manage future risks arising from climatic change.
Hulme, M. et al. Nature 397, 688–691 (1999).
Semenov, M. A. & Porter, J. R. Agric. Forest. Meteorol. 73, 265–283 (1995).
Riha, S. J., Wilks, D. S. & Simeons, P. Clim. Change 32, 293–311 (1996).
Mearns, L. O., Rosenzweig, C. & Goldberg, R. Clim. Change 35, 367–396 (1997).
Harrison, P. A., Butterfield, R. E. & Downing, T. E. ECU Res. Rep. 9 (ECU, Oxford, 1995).
Parry, M. L. & Carter, T. R. Clim. Change 7, 95–110 (1985).
About this article
Cite this article
Hulme, M., Harrison, P. & Arnell, N. reply Climate variability and crop yields in Europe. Nature 400, 724 (1999). https://doi.org/10.1038/23388