Abstract/Summary

Spatial variation in seed and grain quality was investigated in uniformly-managed commercial winter barley crops in the south-east of England in order to assess scope for precision agriculture (PA) and management of seed crops. Within-field spatial variation in barley seed quality, yield and protein content was investigated at different spatial scales to determine scale-dependence of variation and relationships with environmental factors. Two fields were studied, one in 2013 and one in 2014, using unbalanced nested sampling designs with five spatial scales between 1 and 81 m and 138-150 sampling points per field. Spatially-correlated variation in seed moisture content, thousand grain weight (TGW) and protein content was moderate and the variograms of most variables increased with lag distance to a sill. Scale-dependent correlations of seed and grain quality with pedo-climaticbiotic variables were stronger over long (> 20 m) than over short (< 20 m) distances. Interestingly, some correlations changed from positive to negative or vice versa at different spatial scales implying scale-dependent processes. Although spatial variability in seedling vigour was unlikely to be of agronomic importance for PA in either field, it provides an example of different scale dependent processes. In one field, high seedling vigour was associated with high canopy temperature over an 81 m scale, but it was correlated with lower temperatures over a 3 m spatial scale. It is suggested that in general higher canopy temperatures promoted better seed development, more rapid drying and reduced seed dormancy accounting for the long distance correlation. At high temperatures, however, too high a temperature could become supra-optimal and detrimental to seed viability and vigour, and in locations with higher canopy temperatures, there could be a negative correlation over short distances. Spatial dependence was, however, low for seed germination, partly because germination percentage was high, averaging 95-97% across the fields and at least 90% at all sampling points. No PA interventions, therefore, needed to be considered for germination. Such variation in germination as did occur, was at a shorter spatial scale than the minimum sampled (i.e. <1m) and is probably accounted for by intrinsic variation in seed development between individual plants and tillers. By contrast, PA for seed quality could be envisaged in the two fields studied for seed moisture content (SMC), thousand grain weight (TGW) and grain protein content, which varied highly over longer distances (> 20 m). Management strategies of zones of high and low TGW is, however, unlikely to be of economic value; for protein it might be relevant for malting barley. For seed moisture content, zones >17% could be harvested separately either for dedicated drying of moister seeds or else postponing their harvest. An economic analysis suggests drying costs would be reduced through such a strategy, but un costed advantages are likely to include optimal use of drying facilities and avoiding heat damage to seeds which do not need drying. For such zones to be usable by farmers, they need to be predictable. Spatially-correlated variation in seed moisture occurred such that high seed moisture was associated with areas of higher soil moisture from anthesis to maturation. Higher soil moisture areas were associated with and might be predictable from soil texture, soil organic matter, soil EC and to some extent canopy variables, but further research is needed both to validate these recommendations in other fields and seasons and to assess factors likely to influence adoption of zone management at harvest.