Understanding and assessing lodging risk in winter wheat
A detailed literature review was done which indicated that there was still some `uncertainty' as to the exact cause of lodging and also identified the need for a quantitative method to assess lodging risk. Separate published models for determining the windthrow of trees and the anchorage of wheat roots were combined, and a model was developed in order to integrate all aspects of lodging in wheat, and express them in terms of risk. The model was broadly sectioned into three components; the plant canopy, stem base, and roots. The aerial force imposed on the stem base and roots was calculated (using both plant measurements and randomly selected weather conditions in July). By comparing this to the failure moment of the stem base and roots, the model then predicted the probability of lodging occurring. At the same time, wheat was grown in a series of field experiments at ADAS Rosemaund, Herefordshire between 1994-1996 and the effects of various agronomic factors on the crop and its yield were found to be similar to those reported in previous work and from practical experience. Lodging was most prevalent during the 1995-96 season causing yield losses of up to 1.3 t/ha, while very little lodging occurred in the previous two seasons. Reductions in grain quality were also associated with lodging in both the 1994-95 and 1995-96 seasons. Lodging was observed to occur by both stem failure and root failure, with root failure being the predominant cause in 1996. Lodging was influenced most of all by 'crop structure', as affected by agronomy, which was clearly illustrated by 93% lodging occurring in the early-sown, high seed rate, high residual nitrogen treatment compared to only 8% lodging in the late-sown, low seed rate, low residual nitrogen treatment. It was found that the latter treatment described, with the addition of full PGR, had no lodging in any season, as a direct result of lodging resistance gained from optimising plant structure due to crop husbandry. The actual 'type' of lodging was also influenced by the state of the soil; in the dry summer of 1995, soil strength was very high (average c. 100 kPa) and lodging occurred by stem failure, whereas during 1996 when root lodging occurred, the surface soil was moist and soil strength was much lower (average c. 20 kPa). Agronomic practices greatly affected lodging risk in the field experiments. A low seed rate (250 seeds/m2) provided the most consistent and effective method of reducing lodging in all seasons, by significantly increasing the stem diameter (by up to 0.35 mm) and improving root structure (by producing up to 7 more crown roots and increasing the size of the root cone diameter by up to 7 mm). Early sowing (late-September) increased crop height (by up to 6 cm) in all seasons except 1994-95 and resulted in increased lodging in the 1995-96 season. High soil residual nitrogen increased lodging slightly but its effect on `crop structure' was much less than seed rate or sowing date in all seasons. Plant growth regulators (PGR) reduced lodging compared to the nil `control' by significantly reducing stem height (average c. 10cm), but not through increasing stem failure moment or thickening the stem wall width. The reduced nitrogen `canopy management' treatment also generally reduced lodging across seasons compared to the nil `control' although, not by as much as or as consistently as with PGRs. Two PGR-use schemes currently available were examined and recommendations given were found to be poor when compared to the experimental findings and much less `comprehensive' than the modelling approach used here. Other results from a range of winter wheat varieties tested found that variation in basal stem structure and crown root structure was large, which was shown by the model to have implications in terms of lodging risk. These findings indicate the need for improved information and better targeting of varietal lodging resistance in the future. Other findings showed that the use of 'Baytan' seed treatment significantly decreased lodging risk by producing a deeper crown root anchorage and a larger root cone diameter. It was also found that severe stem base disease (fusarium and sharp eyespot) reduced the stem failure moment causing up to 40% greater lodging risk compared to uninfected stems. The model was then used to support the experimental findings by converting the large differences caused in plant structure into estimates of lodging risk and results showed that model probabilities matched reasonably well with the actual lodging in the various experimental treatments. Certain measurements such as plant height and angle of root spread were found to be unimportant. In contrast, a model sensitivity analysis found that plant natural frequency, stem base diameter and root cone diameter were crucial `indicators' of lodging risk. It was also found that wind speed and field altitude were less influential than rainfall in increasing lodging risk. A handheld lodging device (torquemeter) was field tested and provided important results which found the relationship assumed in the model between soil strength and root failure to be flawed, so that root lodging was underestimated. This finding has allowed considerable improvements to be made to the below-ground model. The identification of various `indicators' of lodging risk have successfully provided the basis for further development of the model, to enable a more quantitative, predictive lodging risk assessment scheme for use by farmers and consultants.