Mechanisms On Weed Seed Dispersal By Irrigation Water And Sustainable Weed Management Strategy Under Rice-Wheat/Oilseed Rape Double Cropping System

Weed seed bank is the primary source of subsequent weed occurrence and infestations, and the control of weed seed dispersal is one of the important measures of weed management strategy. The middle and lower reaches of the Yangtze river valley is an important region for growing cereal crops in China. In this region, rice-wheat/oilseed rape double cropping system is prevailing, where irrigation is regularly carried out in the growing season of lowland rice. Weed seeds have adapted well to dispersal by irrigation water flow and are inputted and outputted the fields. Thus it is important to gain a sustainable weed management strategy to control weed seed dispersal by irrigation water flow in this region. The objectives of this study are:1) to reveal the mechanisms on weed seed dispersal by irrigation water flow by determining the size, morphological and anatomical structure, weight, buoyancy of weed seeds and their relationships; 2) to track weed seeds dispersed by agronomic practices of irrigating, ploughing and harrowing; 3) to find out the diversity and composition of floating weed seed communities and associated cropping systems across the main rice-growing regions in China; and 4) to measure the effects of long-term fertilization regimes on weed communities under rice-oilseed rape double cropping system. Except for continuing the plot experimental research on influence of the different control measures on weed community dynamics in our laboratory (2000～2006), a scale-enlarged experiment was established and more ecological measures were additionally adopted to verify a supposed sustainable weed management strategy focusing on depleting the potential weed community, and to provide a basis for the extensive extension of this strategy.The size of 134 weed seed species belonging to 30 families and weight of 104 weed seed species belonging to 27 families under rice-wheat/oilseed rape double cropping system were measured to study the relationship between the size, weight and buoyancy of those seeds. The results showed that the relative volume and weight of different weed seed speceis are significantly different. The relative volume of Xanthium sibiricum is the biggest (1455.03 mm3), and that of Aster subulatus is the smallest (0.04mm3). The weight per seed of X. sibiricum is 106.22 mg, the heaviest, whereas that of Roippa cantoniensis is 0.02 mg, the lightest.Seeds of 104 weed species belonging to 27 families were selected to determine their buoyancy. The results showed that the buoyancy of weed seed from Poaceae, Asteraceae, Polygonaceae, Leguminosae, Amaranthaceae, Umbelliferae, Euphorbiaceae and Ranunculaceae was higher (the persisting floating time on water surface of most weed seeds> 500 hours), whereas that of weed seed from Cruciferae, Cyperaceae, Caryophyllaceae, Scrophulariaceae, Labiatae, Rubiaceae, Alismataceae, Convolvulaceae, Commelinaceae, Pontederiaceae, Malvaceae, Boraginaceae and Acanthaceae was relatively lower (the persisting floating time on water surface of most weed seeds<300 hours).The weed seeds of Beckmannia syzigachne and Euphorbia helioscopia had very high buoyancy; their persisting floating time on water surface was 840 hours. The persisting floating time of some other weed seeds on water surface was also more than 800 hours, such as Artemisia lavandulaefolia, Eclipta prostrata, and Aeschynomene indica. Weed seeds sank more quickly in the higher degree of muddy water. The results suggested that the structure of the weed seeds dertermined the buoyancy, but the degree of muddy water and the environment also affected it.Seeds of 48 weed species were anatomized using cryo-sectioning and good micrographes directly indicating the microstructures adaptive to floating were obtained. The results showed that weed fruits with high buoyancy (persisting floating time on water surface> 700 hours) had developed appendages (glume or perianth) or pericarp tissue which were composed of highly lignified cells or parenchyma.Through tracking the Beckmannia syzigachne and Alopecurus japonicus seeds dyed by safranine T which were released in the irrigation ditch and the field, the dispersal distance and range of these two weed seed species by irrigating, ploughing and harrowing were revealed. The results showed that the Beckmannia syzigachne and Alopecurus japonicus seeds could be dyed very well by safranine T. The buoyancy, weight and germinating rate of these dyed seeds did not significantly change. In a 350-meter long irrigation ditch, one hour of irrigating with water flow speed of 0.5 m/s and rate of 0.13 m3/s could disperse 16.80% of the dyed Beckmannia syzigachne seeds and 34.49% of the dyed Alopecurus japonicus seeds which were released at the water entrance of irrigation ditch. The dispersal speed of Alopecurus japonicus seeds was faster than the Beckmannia syzigachne seeds by the irrigation water flow. The irrigation water flow also could disperse the dyed Beckmannia syzigachne and Alopecurus japonicus seeds in the field. The agronomic practices of ploughing and harrowing could disperse the two weed seed species well, and the dispersal distance and range of the dyed Beckmannia syzigachne and Alopecurus japonicus seeds released at the center of the field were larger than those released at the water entrance of the field. The ploughing and harrowing could disperse the dyed Beckmannia syzigachne released at the center of the field across the whole field. The effect of the wind on dispersing the weed seeds was small. The method of safranine T dyeing could be relatively widely used in the hydrochory research of pale weed seeds.The diversity and composition of floating weed seed communities and associated cropping systems were surveyed in 27 sites across the main rice-growing regions in China. The aim was to better understand weed seed dispersal via irrigation water in fileds. Seed of 74 species, belonging to 20 families, were identified from floating matter on the water surface. Thirty-five species from three families:Poaceae (15), Asteraceae (11), and Polygonaceae (9), accounted for 47% of all species identified. The composition of floating weed seed communities varied in different regions. Weed species with seed maturing in the summer accounted for 64% of the weed seed community. Species richness scores, Shannon-Wiener index and Pielou evenness index were significantly different among the floating weed seed communities. The diversity of weed seed communities in the Yangtze river valley was higher than that in other sites, and some sites were dominated by a few weed species, such as Beckmannia syzigachne, Alopecurus aequalis, A. japonicus, and Polypogon fugax. The 27 sample sites of weed seed communities can be clustered into two groups on the basis of previous crop:sites with a previous crop of lowland rice and sites with previous crops of winter fallow, winter wheat or oilseed rape. Canonical correspondence analysis (CCA) revealed that irrigation frequency, previous crop, and latitude, but not soil type or longitude, significantly affected species composition. The number and diversity of weed seed species decreased with increasing irrigation frequency but increased in lowland rice fields whose previous crop changed from lowland rice, winter wheat or oilseed rape to winter fallow. Filtering irrigation water and collecting and removing floating weed seeds from the water surface could be integrated into weed management practices to control weeds in lowland rice fields before planting rice.The effects of long-term fertilization regimes on weed communities were evaluated in a rice-oilseed rape double cropping system. The fertilization treatments were:no fertilizer (NoF), N fertilizer (NF), chemical fertilizer (CF), chemical fertilizer plus pig manure (CFM), chemical fertilizer plus oilseed rape stalk (CFO), chemical fertilizer plus rice straw (CFR), and chemical fertilizer plus oilseed rape stalk and rice straw (CFOR). The treatments were maintained for 20 years. In oilseed rape field, a total of 17 weed species from 11 families were recorded. Of these, Alopecurus aequali, A.japonicas, Beckmannia syzigachne, Malachium aquaticum, and Geranium carolinianum had high population density, and were dominant in different plots. The weed species richness varied significantly among different fertilization treatments, ranging from 5 species in CFOR to 12 in NF. Shannon-Wiener index in NF was significantly higher than that in any other fertilization treatment while Simpson dominance index was the lowest. Pielou evenness index in NoF and NF were significantly higher than those in other fertilization treatments. As suggested by Whittaker index, CFO and CF had the most significant effects on weed community composition, CFM and CFR ranked the second, while NF and CFOR had no significant effect on weed community composition. In lowland rice field, a total of 11 weed species from 6 families were recorded. Of these, Rotala indica, Lindernia procumbens, Monochoria vaginalis and Cyperus difformis had high population densities under all fertilization treatments. The sedge weed Eleocharis yokoscensis was found only in the NoF and NF plots, while Pycreus globosus and Scirpus juncoides were found only in NoF plots. The weed species richness varied significantly among different fertilization treatments, ranging from 7.3 in CFOR to 10.7 in NoF. The composition and abundance of weed species also differed in different fertilization treatments. The Shannon-Wiener index for the NoF, CFM and CFR treatments was significantly higher than that of the other fertilization treatments. The Simpson dominance index in the NoF, CFM treatments was significantly lower than that of the other fertilization treatments. The Pielou evenness index in the NoF and NF treatments was significantly lower than that of the other fertilization treatments. Based on the Whittaker index, the CF and CFM treatments had the greatest, and the CFO and CFOR treatements the second greatest, effects on weed community composition. Whereas, the NF and CFR treatments had no significant effect on community composition as compared to the NoF treatment. The results indicate that the application of a balanced N, P, K fertilizer with or without organic manure (pig manure and/or crop straw) allowed for all weed species present to thrive, thus maintaining weed species diversity as a result of no one species becoming dominant.A plot field experiment was continuously conducted to evaluate the effects of five different weed management strategies:manual weeding (MA), chemical weeding (CH), weed free (WF1), weed free without water interception (WFO) and a check with water interception (CK), on the dynamics of weed seed bank under rice-wheat cropping system. The results showed that under different weed management strategies, the species richness index and the Shannon-Wiener index decreased and the Simpson dominance index icreased steadily when soil seed bank managed by the above five weed management strategies, but the Pielou evenness index of weed seed bank showed no significant changes. Except for CK, the density of weed seed bank declined annually, and the sequence for the effects of weed seed depletion was WF1>WFO>CH>MA>CK. The density of weed seed bank in the treatment of WF1 decreased about 85% after 8 years. There were evident changes in relative abundance of weed species among the potential weed communities, and differences between community similarity indices within management strategies increased with the year prolonging. The density and species composition of aboveground weed commuities also exhibited marked changes.A scale-enlarged field experiment was conducted to evaluate the effects of five different weed management strategies:chemical weeding without water interception (CHO), chemical weeding with water interception (CH1), chemical weeding with water interception plus mechanical blowing weed seeds (MB), chemical weeding with water interception plus net gaining weed seeds (NG) and double chemical weeding (CH2), on the dynamics of weed seed bank under rice-wheat double cropping system. The results showed that under different weed management strategies, the species richness index, the Shannon-Wiener index and the Pielou evenness index decreased and the Simpson dominance index icreased steadily when soil seed bank managed by the above five weed management strategies. Except for CHO, The density of weed seed bank declined annually, and the sequence for the effects of weed seed depletion was CH2> NG> MB> CH1. The density of weed seed bank in the treatment of CH2 and NG decreased about 44% and 35% after 2 years, respectively. There were evident changes in relative abundance of weed species among the weed seed banks, and differences between community similarity indices within management strategies increased with the year prolonging, being similar as the plot experiments.