Comments on a report of regression-based evidence for impact of recent climate change on winter wheat yields

In their 2008 paper entitled “Impact of recent climatic change …”, Xiao et al. [Xiao, G., Zhang, Q., Yao, Y., Zhao, H., Wang, R., Bai, H., Zhang, F., 2008. Impact of recent climatic change on the yield of winter wheat at low and high altitudes in semi-arid northwestern China. Agric. Ecosyst. Environ. 127, 37–42] presented evidence for increasing grain yield of winter wheat (Triticum aestivum L.) from 1981 to 2005 at two locations in China, as well as for non-linear effects of rainfall and mean temperature on yield. Inspection of their graphs suggested that the responses were overstated. Reanalysis of five graphs indicated that yield had increased at just one site, that there was no effect of rainfall, and that mean temperature only affected yield at the higher site. This temperature effect was described with an exponential function (r2 = 0.32), but the response could equally be considered linear (r2 = 0.27). Overall, the reanalyses suggest that four described relations were incorrect due to errors in analysis or reporting. The results highlight the difficulties with using regression with historic data to detect effects of climatic change on agriculture.