WEED ECOLOGY AND CONTROL IN RAINFED RICE CROPPING SYSTEMS.

摘要

Factors responsible for shifting weed populations in rainfed rice (Oryza sativa L.) cropping systems have not been adequately identified or investigated. This thesis focuses on the effect of soil moisture content, handweeding, landscape position, soil oxygen diffusion rate, and length of cropping on weed shifts. In wet-seeded rice, soil moisture contents from 70 to 100% (w/w) favored the emergence of Echinochloa glabrescens Munro ex Hook f. Leptochloa chinensis (L.) Nees emerged best when soil moisture contents were 70 to 90%. Herbicide performance in wet-seeded rice is affected by soil moisture content. In irrigated rice, residual herbicides are usually applied 6 to 8 days after wet-seeding. These results show that preplant applications may be better in rainfed wet-seeded rice. Soil oxygen diffusion rates were negatively correlated with soil water table depths in bunded fields. Aeration, combined with handweeding, regulated the composition of the weed flora in dry-seeded bunded rice. In a dryland rice cropping pattern, weeding intensity regulated the weed flora. Handweeding reduced the carryover of Cyperus rotundus L. in corn (Zea mays L.) grown after rice. Following two years of intensive weeding in rice followed by corn, weeding in corn reduced the density of C. rotundus in rice. In contrast, under bunded conditions weeding in one crop had no effect on weeds in the subsequent crop. Weed weights increased and grain yields decreased as cropping continued in a rainfed wetland double-cropping rice system. Other factors which determined the composition of the weed flora in rainfed rice cropping systems were land preparation, previous crop management, and flooding data.