Modelling the Effect and Variability of Integrated Weed Management of Phalaris minor in Rice-Wheat Cropping Systems in Northern India

Phalaris minor Retz. (littleseed canarygrass) is the most problematic and herbicide-resistant weed in the rice-wheat cropping system in India. As such, it poses a severe threat to wheat yield and food security. A number of herbicidal and agronomic practices have been identified for the effective control of P. minor. These include crop rotation, crop establishment methods, herbicide spray technology, sowing time, weed seed harvest and effective herbicide mixtures. A population model of P. minor was built based on the life cycle of the species, herbicide resistance mechanisms and the effects of weed control practices. The model simulated the interactions of these factors and provided the best management recommendations for sustainably controlling this noxious weed species. Model results indicate that integration of chemical and non-chemical control methods was the most effective and sustainable strategy. For example, the integration of a happy seeder (a tractor-mounted mulching and sowing machine) with an effective post-emergence herbicide reduced the probability of weed control failure by 32% compared to the scenario with a rotavator and the same herbicide. Similarly, more conventional crop establishment methods such as a rotavator and conventional tillage could be accompanied by pre- or post-emergence applications of herbicide mixtures. Adoption of good herbicide spray technology and weed seed harvest delayed the onset of resistance evolution by up to four years. Furthermore, effective crop rotation such as the inclusion of sugarcane in place of rice in the summer season reduced the risk of resistance evolution by 31% within the 10 year simulation period. In addition to the scenarios using representative parameter values, the variability of model predictions was investigated based on some field experiments. The model provided a powerful tool for promoting Integrated Weed Management and the sustainable use of herbicides. Pragmatic ways of dealing with uncertainty in model prediction are discussed.