The improvement of mechanical weed management requires a broad research agenda, including biological, agronomic, and engineering studies.
Much work is needed on the mechanisms whereby tillage affects perennial weed populations. How do the size distributions of root and rhizome fragments compare following tillage with different implements, and how does this affect a fragment's survival probability and subsequent rate of growth.? Do particular species survive better when buried with the shoot attached or with the shoot severed from the roots. How do these factors interact with soil temperature and moisture conditions in determining the degree of control by tillage.
Knowledge is equally sparse on the mechanisms whereby tillage affects seed banks. In particular, it is not even clear whether most depletion of seed banks by tillage is due primarily to additional seedling emergence, as implied by Figure 4.6b, or whether the tillage-related mortality factors listed in Figure 4.7 play an important role for some weed populations. Ultimately, the emergence models discussed in the section "Effect of the timing of tillage on weed seedling density" need to be extended to predict not just when emergence will occur, but how many seedlings will emerge, and how this relates to the tillage regime.
The destruction of weed seedlings during cultivation also requires mechanistic investigation. When using tools that disturb only a shallow surface layer, which species must be killed in the white thread stage, which remain susceptible after emergence, and how does this relate to development of the weeds' root systems.? Which species can recover from shallow burial? How does the rate of kill for particular species relate to soil moisture status at the time of disturbance? To what extent are seedlings of particular species killed by burial vs. dismemberment vs. desiccation, and does the cause of death vary with the implement? Fogelberg & Dock Gustavsson (1999) showed that a vertical-axis brush weeder killed more weeds by uprooting than by burial, though both mechanisms contributed significantly to weed mortality. Similar quantitative data are unavailable for most implements.
Much highly applied work is also needed to provide a scientific basis for cultivation recommendations. Many implements have not been studied systematically for performance under different speeds, angles, depths, or position relative to the row for even a single crop, but such data are needed for a variety of crops, each in a range of phenological stages. Quantitative European work on weeding harrows, flame weeders, and brush weeders provides models for the type of research that is needed (Rasmussen, 1991; Ascard, 1994b; Melander, 1997). Similarly, anecdotal observations and arguments from first principles have led to recommendations here and elsewhere regarding the role of tilth, soil texture, and clod size on the performance of cultivation implements, but scientific documentation to support these recommendations is largely lacking. This lack of information is analogous to making herbicides available to growers without supplying information on crop tolerance, appropriate rate, or timing of application.
A more difficult subject in need of investigation is root pruning during cultivation. Few data are available on how close or how deep various implements can be used without damage to crop root systems. Obviously, this depends on crop species and size, and probably also depends on growing conditions and soil properties. Further work is needed that builds on Russel, Fehr & Mitchell's (1971) study of soybean response to root damage by cultivators under various environmental conditions.
Regarding the engineering of mechanical weeders, the critical challenge at present is not to create implements based on new principles, but rather to make the present equipment more usable. In particular, most tools must be set for depth and distance from the row, and often angle as well. Provision of cranks, rulers, and angle gauges on supporting shanks and brackets could greatly simplify these adjustments (Mattsson, Nylander & Ascard, 1990). Ideally, a cultivator should be tuned to the crop size and soil conditions each time it is used, and this is essential for precision implements working in or close to the row. However, the nut-and-bolt work required to adjust most machines is frustrating and potentially interferes with timely use of the implement.
Guidance systems that allow controlled cultivation of small plants are also sorely needed. An important application of these will be for cultivation of crops planted in close rows (Olsen, 1995).
Finally, further research is needed on the integration of mechanical weed management with other ecological and chemical weed management tactics. The complexity of the potential interactions indicates that this work will need to extend beyond multifactorial empirical studies. A promising approach may be to link crop-weed competition models, which typically take weed density as an external forcing function, with weed population models, which focus on weed density, but generally take a simplified approach to competition. Despite the long history of tillage and cultivation, much biological and agronomic research is still needed to achieve the full potential of this ancient approach to weed management.
Was this article helpful?