Factors in conservation farming that reduce erosion

A medium-term (10 years) stubble x tillage field experiment was established in 1984 on a red-brown earth at Tarlee, 70 km north of Adelaide, to develop a suitable system for sustaining the soil resource. Measurements of infiltration capacity, soil detachment rate, and erosion were taken in summer, autumn, winter, and spring 1989-90. The rotation was wheat-barley-grain legume, and treatments included 3 levels of stubble retention (0.5, 3.0, 5.0 t/ha.year) and 4 types of tillage [no-tillage (NT), direct drill (DD), reduced tillage (RT), conventional cultivation (CC)]. NT was seeded with narrow points (30 mm) and the other tillage treatments with wide shares (150 mm). The Northfield rainfall simulator with an erosive rainfall of 100 mm/h and an energy of 28.6 J/m(2).mm was used to measure runoff and soil and nutrient loss. This paper reports on erosion from this experiment. The results show that runoff was reduced through farming practices such as the retention of adequate stubble residue (about 3-5 t/ha.year of cereals), NT, or a combination of these factors. Increasing the average annual stubble retention decreased runoff and soil loss linearly. The greater the amount of stubble retained annually, the less the runoff and soil loss, whether or not the soil surface was protected. The significant sediment release and soil loss from bared soil is inversely related to soil stability measured by a reduction in soil organic matter. Crop type also influenced erosion; for example, soil was more vulnerable to erosion after peas than after cereal. The amount of stubble after harvest was usually greater with cereals than with grain legumes. Runoff as a percentage of applied rain, and soil loss, ranged from 26 to 60% and 0.52 to 1.1 t/ha for 0.5 t/ha.year stubble (means of all treatments) for April 1989 and August 1990, respectively, and from 5 to 35% and 0.03 to 0.8 t/ha for 5.0 t/ha.year stubble for the same simulation period. Runoff rates in the last 3 min of 18 min simulation ranged from 0.4 to 1.1 mm/min for 5.0 t/ha.year stubble and from 0.8 to 1.7 mm/min for 0.5 t/ha.year stubble. The runoff rates recorded at the 18th minute of simulation ranged from 0.5 to 1.2 mm/min for NT and 0.7 to 1.5 mm/min for CC.