A huge amount of climate research has explored how global warming will affect the crop harvest. The general outlook is that global agricultural production will decline; in Africa and South Asia, for instance, studies have predicted that the yield of eight major crops will fall significantly by mid-century. Combine that with rising populations, and some scientists have predicted that in the future much more land will need to be devoted to farming.

In reality, the situation could be more nuanced. Research has already shown that irrigation can, to an extent, assuage the impact on crops of a warming climate. On the other hand, climate change is also expected to bring extremes of weather, the effects of which are much harder to predict.

"One of the difficulties in determining impacts is that climate extremes are rare, so you often don't have many data points to analyse," said Tara Troy at Lehigh University, US.

Troy and colleagues at Columbia University and the University of Colorado Denver looked at how climate extremes and irrigation will affect future crop yields. To counter the usual lack of data, they pooled 50 years of historical data on weather and crop yields from more than 3000 US counties. The researchers standardized the data to accommodate climate differences that occur naturally between different locations. In the final step, they analysed the data two ways: once by taking into account whether the crops had been irrigated or rain-fed, and once ignoring the method of watering altogether.

For the analysis, the team turned to "conditional-density" plots. These are a type of statistical data presentation that allowed the researchers to visualize how crop yields varied with the climate, and in so doing avoid making easy assumptions about underlying relationships. Indeed, they revealed that the relationship between crop yields and the climate is far from the simple linear one to which it's sometimes approximated.

For the soya crop, for example, yields decreased steadily with maximum temperatures until a certain threshold value, when they plummeted. Such nonlinearity manifested when yields were plotted against rainfall, too.

It was not all bad news, however. Troy and colleagues also found that irrigation could shift the critical threshold when yields plummeted to higher temperatures, and in some cases "decouple" yields from temperature altogether. Returning to the soya example, yields of the rain-fed crop dropped with average growing-season temperature, whereas the yields of the irrigated crop were unaffected.

"Simple linear models of the relationship between climate and yields are too simplistic," said Troy, "particularly when looking at extremes rather than growing-season averages. The existing historical data can be used to better understand these relationships, so that we can plan accordingly for projected changes in climate."

Troy says her group's study also stresses the importance of irrigation. "As was already known, irrigation is an important adaptation mechanism in agriculture," she said, "but what this study does is show how it benefits agriculture under a wide range of climate indices."

The findings are published in Environmental Research Letters (ERL).

Related links

Related stories