Predicting lodging in winter wheat
Lodging, the permanent displacement of plant stems from the vertical, occurs on average once every four years in UK wheat crops, when it reduces the yield and bread making quality of grain. Lodging usually occurs in the summer and control is most commonly sought by applying growth retarding chemicals in the spring. This thesis develops a method of predicting which crops are prone to lodging so that spring lodging controls can be targeted most effectively. A model was developed in association with this study which calculates stem and root lodging risk from summer-time plant, weather and soil factors. The spread of the plant's root plate, the stem diameter and the number of shoots per plant were shown to have very strong influences on lodging. Structural rooting depth, stem failure yield stress, height at centre of gravity and the rate at which stems oscillate in wind (natural frequency) were also important, but less influential. Methods of predicting the most important lodging-associated plant characters from crop observations in spring were developed from the literature. These were then tested through experiments in 1995 and 1996 with factorial combinations of crops sown in late September and late October, at 500 seeds m-2 and 250 seeds m-2 and with large and small levels of residual soil nitrogen. Early sowing, dense seed rates and fertile soils all increased stem and root lodging, with sowing date having the greatest influence. Plants sown at high densities had small root plates and were poorly anchored; they also had fewer shoots causing a smaller leverage. Early sown plants had shoots with a high centre of gravity and slow natural frequency, causing a greater leverage. Plants sown early on fertile soils had narrow, weak stems. Final shoot number per plant was predicted with good precision (R2=094) from spring plant number m-2 and maximum shoot number m-2 using a model of tiller survival. Stem diameter was predicted with moderate precision (R2=057) from spring canopy size and shoot number m-2, via a calculation of the amount of dry matter partitioned to each stem base. Root plate spread showed a linear and inverse relationship to spring plant density (R2=0.48), mainly as a result of variation in the length of the rigid roots and in the width of the plant base. It is concluded that early season crop observations have the potential to predict the values of the most influential lodging-associated plant characters, from which a model of lodging can calculate the proneness of crops to stem or root lodging in time for remedial action. The next steps would be to develop prediction schemes for other plant characters which influence lodging and test all the predictions in a wider range of crops, sites and seasons.