Sulphur nutrition of cereals in Britain: Yield responses and prediction of likely deficiency

Abstract

Shortage of sulphur has become a problem for arable crops in some areas of the U.K., mainly as a result of decreased inputs from atmospheric deposition. The aims of this project were to assess those areas in Britain where cereals are at the most risk of S deficiency, to investigate the change in S status of wheat grain during the past decade, and to identify sites giving a response to S fertiliser and relate that response to atmospheric inputs, soil S and crop S status.

A qualitative computer model using soil, atmospheric deposition and meteorological data predicts that currently 11% of the British land area is at high risk of S deficiency for cereals, and a further 22% at medium risk. The high risk areas are in south-east Scotland, the Scottish Borders, East Anglia, the Welsh Borders and south-west England. These areas are characterised by small inputs of S from the atmosphere, low content of soil organic matter and light soil texture. The SO2 emissions in the U.K. are set to decrease to 40% of the 1980 level by 2003. The model predicts that such decreases would increase the medium and high risk areas to 27 and 22%, respectively.

The concentrations of S in British wheat grain have decreased considerably during the past decade. Grain samples with low S concentrations were located mainly in Scotland, northern England, and the west and southwest of England, whereas samples from central England tended to have higher S concentrations. Breadmaking varieties had significantly higher grain N and S concentrations than other varieties, but little difference was found between varieties in grain N:S ratio. For the varieties Mercia in both 1992 and 1993 and Hereward in 1993, grain S concentration correlated better with loaf volume than grain N concentration.

The effects of S application at a rate of 40 kg/ha on grain yield were tested at 21 and 19 sites around Great Britain in the 1992/93 and 1993/94 seasons, respectively. Yield responses were obtained at three sites at a significance level of p<0.05 in both 1992/93 and 1993/94 seasons, and at two further sites at p<0.10 in 1992/93. The yield increases due to S application ranged from 4.2 to 18.4%. The responsive sites were located in Scotland, Scottish Borders, East Anglia and Southwest England. The geographical distribution of the responsive sites agreed well with the model prediction and therefore validated the modelling approach. In contrast, soil extractable S in spring was not a good indicator of S supply. Also, there was no clear relationship between either grain S concentration or N:S ratio with yield responses.