A TECHNIQUE FOR MAPPING THE SPATIAL-DISTRIBUTION OF ELYMUS REPENS, WITH ESTIMATES OF THE POTENTIAL REDUCTION IN HERBICIDE USAGE FROM PATCH SPRAYING

The spatial distribution of Elymus repens L. was mapped in five cereal fields during the 1994 season. Weed maps were created using a semi-au tomated system mounted on a vehicle that travelled up the tramlines, c entred on a 12-m-wide spray boom. Two operators detected weeds visuall y and recorded their presence and density by means of a simple push-bu tton system that recorded data from (2x1) m(2) areas, or cells, across the boom section on a portable computer. The position of the vehicle along the tramline was monitored by integrating output from a wheel se nsor. Eight tramlines of one field were mapped three times consecutive ly, to assess detection and navigational error. Pair-wise comparisons of the three runs gave approximately 85% repeatability on presence/abs ence data, 80% on zero/low- or high-density data and 85% on low/high-d ensity data when weeds were definitely present. Simultaneous compariso ns of all three runs gave 78%, 69% and 75% respectively. Repeated runs of tramlines up to 550 m long recorded mean differences of approximat ely 2 m with a maximum error of 14 m (2.5%). The spatial distribution of E. repens within the five fields was visibly patchy, but the size a nd morphology of patches varied both within and between fields. Thus, the potential reduction in herbicide usage as a result of patch sprayi ng varied with patch morphology and infestation level. The effect of l owering the resolution from six (2x1) m(2) cells to three (4x1) m(2) c ells across the 12-m boom width showed little change in potential redu ction in herbicide usage.