More to Climate Change than Just Temperature: Recent Weather Patterns Led to Smaller Harvests

Weather conditions typical of the last 30 years have significantly limited harvests of winter wheat and spring barley, according to analyses of 124 years of crop yields obtained from long term experiments.

Looking at monthly summaries of daily temperatures, rainfall, and sunshine hours, for one site in Southern England, researchers have identified 10 distinct ‘clusters’ of yearly weather patterns that occurred from 1892 to 2016.

The annual weather cluster most dominant after 1991 generally had higher summer temperatures and more intense winter rainfall distributions but with a particularly dry June - and it is this weather pattern which was also associated with the lowest cereal yields.

Hay yields from pasture under the same weather pattern were not impacted to the same extent, however.

The scientists from Rothamsted Research used data collected from the weather station and crop experiments first set up at their Hertfordshire institute in the 19th Century, and say that their analysis of weather patterns and associations with crop yields is more relevant and realistic than solely looking at the effect of average temperature.

They found that five annual weather patterns which occurred frequently during much of the 20th Century have not recurred in more recent years, according to lead author Dr John Addy.

“This provides unambiguous evidence of climate change in several more dimensions than mean temperature alone.

“Climate change is about more than just increasing temperature. The long-term trend of annual mean temperature is only one aspect of recent climate change. Other changes in climate, seen in changes to within-year weather patterns relevant to crop production, have also occurred since the late-19th Century.”

Understanding simultaneous changes across multiple weather variables provides an understanding of how the climate has changed for agriculture, and potentially how this influences yield, he added.

The team used multivariate statistical analysis approaches to provide insight into the different patterns of variation in multiple weather variables during a year, the variation in these patterns amongst years, and how the frequency of the different patterns has changed, particularly in the last 25 years.

The frequency of occurrence of years within each weather cluster varied considerably during the study period, with the late 20th and early 21st Century having distinctly different distributions across clusters to earlier in the 20th Century.

Three clusters (‘Cold Winter and early-Spring’; ‘Cold August to September’; ‘Cool and dry March’) contained 64% of years during the 20th Century whilst one cluster, ‘High temperatures, drier June’ contained 63% of the 21st Century years included in the analysis.

The data for the study came from the Rothamsted Meteorological Station, which has recorded daily rainfall, temperature and sunlight since 1892, and Rothamsted's famous Long-Term Experiments with winter wheat, spring barley and pasture grasses, which began even earlier in the mid-19th Century.

Analysis of the annual yields of winter wheat, spring barley, and grassland across the different weather clusters showed that, on average, years in clusters typical of the 20th Century climate provided greater off-take of cereals than those from the early-21st Century.

Years with a cooler winter and early-spring provided the greatest winter wheat yields, whilst drier years with a warm early-summer provided the highest spring barley harvests.

However, the impact of weather was smaller for the grassland pasture than for the two cereal crops implying herbage production was more resilient to the changing climate at this site.

The three experiments analysed were the Broadbalk winter wheat experiment, Hoosfield continuous spring barley experiment, and the Park Grass continuous herbage experiment.