Developing parasitic weed control strategies for rainfed rice production environments

Abstract

Rice is a staple food crop of increasing importance in many countries in sub-Saharan Africa. Its production is often hampered by parasitic weed infestations, like Striga asiatica (an obligate parasitic weed) in rainfed uplands, and, more recently, Rhamphicarpa fistulosa (a facultative parasitic weed) in rainfed lowlands. Facultative parasites are able to complete their life cycle without a host plant. Obligate parasites need a host plant and have developed a synchronization mechanism: their seeds will only germinate after sensing root exudates of a host plant. This study partly focussed on farmer perception of the parasitic weed problem. Surveys and workshops with farmers and other stakeholders in three rice growing areas in Tanzania revealed that farmers are aware of the locally occurring parasitic weed species. Even though they know of a wide range of potential control options, they mainly practise hand weeding to control the parasitic weeds. Adoption of other methods is hampered by trade-offs (e.g. crop rotation reduces the area grown with rice; resistant varieties lack the preferred aroma, taste and grain size). Additionally, alternatives are often not available or affordable (e.g. fertilisesr, herbicides), or information on control efficacy under field conditions is simply missing. Based on informal discussions with farmers, altered sowing time, short-duration rice varieties and organic/inorganic fertilisers were selected for field testing. These measures are partially based on knowledge on the ecology of Striga species. In the absence of a suitable host, seeds will go back into a state of secondary dormancy. Delayed sowing might thus be a feasible escape mechanism. As a result of delayed sowing, the rice crop might encounter drought stress during grain filling. For that reason, this measure was tested for three rice varieties with a different maturity time (early, intermediate, late). Use of fertilisers was chosen since parasitic weeds are identified as indicators of poor soil fertility. Several hypotheses for the relation between fertility and the occurrence of parasitic weeds exist. Well-fertilised plants might simply be more vigorous and are generally producing a lower amount of root exudates. In pot experiments this was shown to result in fewer Striga infections, but whether such a reduced infection level will also be observed under field conditions is not known. Furthermore, organic fertilisers and better growing plants will over time result in a higher soil organic matter content. The associated soil microbial life might shorten the life time of the seeds of the parasite, thereby reducing the weed seed bank. Whether these mechanisms are equally effective against the facultative R. fistulosa, is difficult to foresee. All three measures were tested during three to four seasons in both S. asiatica and R. fistulosa infested fields in Kyela district, south-west Tanzania. The trials undertook farmer-participatory evaluation of the suitability of the management options. Striga asiatica number and biomass decreased with a delay in sowing time. This strategy proved most suitable when combined with an improved early-maturing variety, through which drought stress during grain filling could be avoided. Contrastingly, R. fistulosa biomass was increased, and concomitant rice yields decreased with delayed sowing times. Planting rice at the onset of the rainy season resulted in reduced R. fistulosa infection and higher rice grain yields. Clearly, manipulating rice sowing time showed to be an effective strategy to reduce parasitic weed infection and associated yield losses, but the manner in which this strategy should be employed strongly depends on parasitic weed species and ecosystem. Only in years with low parasitic weed infestation levels and in the absence of ordinary weeds, fertilisation with either organic or inorganic soil amendments had a positive effect on rice grain yield. Fertilisation frequently stimulated R. fistulosa infection, and had a modest but inconsistent decreasing effect on S. asiatica. Obviously, the addition of soil amendments, even for a period of four years, is not the silver bullet approach to minimize parasitic weed problems. In a final year, farmer groups selected their own preferred strategies and evaluated them at the end of the cropping season. Sowing time was adjusted according to the earlier findings. The early sowing time under lowland conditions favoured the traditional late maturing rice variety. Under upland conditions, farmers were more impressed with a newly developed, early maturing, resistant rice variety. Organic fertilizers, particularly rice husks, in combination with mineral fertilizers were the preferred soil amendment. The poor access to newly developed varieties, the absence of preferred quality traits in these new varieties and the lack of financial means of smallholder farmers prevent a rapid and wide-spread adoption of most measures. It is argued that it takes a wider effort, including innovations at institutional level, to sustainably solve the parasitic weed problem in rainfed rice.