Adding value to variety trial data: a performance rating for wheat varieties for dry conditions


Cereals & Oilseeds
Project code:
01 April 2012 - 31 December 2012
AHDB Cereals & Oilseeds.
AHDB sector cost:
Total project value:
Project leader:
Eric S. Ober, Plant Biology and Crop Sciences Department, Rothamsted Research, Harpenden, Herts AL5 2JQ


pr517-final-project-summary pr517-final-project-report

About this project


The wet conditions of 2012 notwithstanding, on average 10% of the UK wheat yield is lost each year due to insufficient soil moisture. Record-breaking dry conditions, most recently in 2011, reinforce the need to find ways to maximise productivity when water is limiting. Growers want to know what variety to plant on light land, which varieties are more likely to yield better in dry conditions, and which varieties show better stability of yield across a range of conditions.

The aim of this project was to provide quantitative data to help guide these decisions by mining the information contained in the multi-location variety trials conducted each year for the Recommended List (RL).

The objectives of this project were to evaluate the 2011 RL trial yield data by assigning a drought stress index to each test site using site-specific soil and weather data, then to score each variety according to how well it performed relative to other varieties along a gradient of sites from unstressed to stressed.

In addition, using various statistical methods, the data were analysed to show which varieties tend to be more stable than others across locations, and which varieties show the best combination of yield potential and yield stability.

In 2011, test sites varied in the level of drought stress and variety rankings changed from site to site. A regression analysis showed that some varieties showed relatively better yields as conditions became drier (Cocoon, Delphi, SY-Epson, KWS-Gator), while others showed greater sensitivity to water availability (Chilton, Denman, Gallant, Grafton) and yielded poorly compared with the tolerant varieties at the stressed sites. Other varieties showed little response to changing water availability and also yielded well across all sites (e.g. Conqueror). There were small differences in the stability of varieties across sites and variety rankings for yield, adjusted for stability, changed little.

It is important to note that the data were from only one year and, therefore, insufficient to provide a robust picture of variety performance. Nevertheless, the results show how stability, yield potential and drought tolerance can be evaluated to provide more information on variety performance.

The real stability of varieties would have to be judged from a larger dataset drawn from multiple years and sites. Drought symptoms of different varieties were scored in a survey of 300 random fields, but did not reveal strong varietal differences and, therefore, could not be used to corroborate drought rankings derived from the RL trial data.

The results show that there is valuable, untapped information inherent in multi-location variety trial data that can be used to add value to those data and current variety recommendation procedures. Furthermore, when these data are combined with specific environmental variables for each trial, additional information about varieties and test locations can be obtained with little extra cost.