Mechanisms of selective mechanical weed control by harrowing

Keywords: weed harrowing, mechanical weed control, selectivity, biomechanics, uprooting, soil-covering, recovery, plant damage, methodology Current weed control problems in organic farming and minor crops show that alternatives for selective herbicides need further development. Mechanical control methods such as weed harrowing are attractive because of the high capacity, wide applicability and low cost. However, the variable effectiveness and limited selectivity at early crop growth stages are major limitations for reliable weed control. In contrast to herbicides, there is little fundamental knowledge of processes and factors that influence selectivity and effectiveness of mechanical weeding. To provide a more basic understanding, the uprooting, covering and regrowth of three model plant species were studied in detailed laboratory harrowing experiments on sandy soil.Uprooting appeared to contribute much more to weeding effectiveness than is commonly assumed. Although harrowing predominantly covered plants, covering killed few. In contrast, on average 47-61% of the uprooted plants were killed after six days without irrigation. Drier soil (5% as compared to 16% w/w) increased mortality of uprooted plants from 36 to 91%.Soil moisture content greatly affected uprooting and covering selectivity. Covering selectivity could be manipulated by working depth and working speed, thus exploiting differences in plant flexibility and height between weeds and crop. Uprooting selectivity could be improved by precise working depth control and by precisely steering the tines along the crop rows. Small-scale spatial patterns and within-population variability of plant sensitivity proved to be very important.This study indicated several opportunities for improving field experiment methodology and for modelling the efficacy and selectivity of mechanical weeding. A method to predict the relationship between crop and weed uprooting from measured plant anchorage forces was developed. This method could be used to quantify the selective potential of a crop-weed situation and the selective ability of harrowing independent of harrowing intensity.