Abstract
Three on-farm and three net-house experiments were conducted at the Durbachara village of Gouripur upazila under Mymensingh district and at the net-house of the Department of Agronomy, Bangladesh Agricultural University of Bangladesh during 2013 to 2016 with a view to (i) finding out the effect of weed control practices and crop residue retention on crop performance and economics and weed dynamics in T. aman rice-mustard-Boro rice and T. aman rice-wheat-mungbean cropping patterns under strip tillage system, (ii) examining the any residual effect of herbicides applied in rice and wheat on the succeeding mungbean in T. aman rice-wheat-mungbean cropping pattern and (iii) assessing the soil weed seed bank dynamics under short-term conservation agriculture (CA) practices under above mentioned two cropping patterns and long-term CA practice at four different locations of Bangladesh. In the above mentioned two patterns, T. aman rice, Boro rice, mustard and wheat were grown under the treatments: six weed control practices viz. (i) Conventional tillage (CT)+ three hand weeding (HW) (Control), (ii) Glyphosate (Gly)+strip tillage (ST)+ one HW, (iii) Gly+ST+pre-emergence (PE) herbicide, (iv) Gly+ST+post-emergence (PO) herbicide, (v) Gly+ST+PE+PO, (vi) Gly+ST+weed-free (WF); and two levels of crop residue retention viz. no-residue (Ro) and 50% residue (R50) (by height). The treatments were replicated four times in randomized complete block design. CT consisted of two primary tillage operations by two-wheel tractor and ST was done by versatile multi-crop planter in single pass operation. A non-selective pre-plant herbicide, glyphosate was applied in all crops at three days before ST operation; the PE herbicide, pendimethalin was applied at three DAT/DAS in rice and wheat but immediately after sowing of mustard; the PO herbicide, ethoxysulfuron in rice and carfentrazon-ethyl in wheat was applied at 25 DAT/DAS; oxadiazon in mustard was applied at 15 DAS at recommended doses. Results revealed that, in ST, glyphosate applied before ST operation followed by sequential application of PE and PO herbicides with the retention of 50% crop residue had controlled weeds most effectively and economically in terms of both density and biomass by exerting the highest weed control efficacy, as well as gave the highest yield and economic returns of all crops without any phyto-toxicity. The performance of mungbean in T. aman rice- wheat-mungbean pattern in terms of plant population and height, number of branches and pods plant1, pod length, seed yield and germination of harvested seeds was not hampered by the residual toxicity of any herbicide applied in rice and wheat. After 2- year field trials, in short-term conservation agriculture practice under two cropping patterns, ST decreased annual weeds but increased perennial weeds while CT increased both annual and perennial weeds. Retention of 50% crop residue reduced both of them. The composition and density of weeds was the highest at 0-5 cm soil depth followed by 5-10 cm and 10-15 cm soil depth. CT decreased weed species diversity as well as introduced fewer new dominant species but ST increased weed species diversity and many new dominant species in the soil weed seed bank. Long-term ST with the retention of 40/50% crop residue reduced soil weed seed bank than CT with 20% residue. This reduction was even higher than bed planting and zero tillage at Mymensingh, Durgapur, Godagari and Baliakandi areas of the country.