Weed pressure level and the correlation between weed competitiveness and rice yield without weed competition: An analysis of empirical data

Abstract

Improving rice ( Oryza spp.) competitiveness against weeds would provide a low-cost and safe tool for an integrated weed management strategy. This paper addresses the underlying causes of tradeoff between yield without weed competition and weed competitiveness. Rice yield or weed biomass under weedy conditions are used as indicators of weed competitiveness. For this analysis, a common database was compiled from the results of 45 concurrent field trials comparing the performance of four to 64 genotypes in weed-free and weedy conditions in Asia, West Africa, North America and Latin America. Cyperus spp., Echinochloa spp. and Eleusine indica were the most frequent dominant weed species, being found in 9, 20 and 7 trials, respectively. Mean relative yield reduction [(yield under weed-free conditions − yield under weedy conditions)/yield under weed-free conditions] across genotypes tested for each trial was defined as weed pressure level. Mean yield without weed competition across genotypes ranged from 1.8 to 11.6 t ha −1 with mean relative yield reduction from almost 0 to 91%. Correlations for rice yield between weed-free and weedy conditions were generally positive, and significant in 27 trials. The correlations were related to weed pressure level and dominant weed species, but not to ecosystem (upland or lowland) or yield level under weed-free conditions. Relative yield reduction had a more pronounced effect than dominant weed species. Correlation decreased as weed pressure level increased, and became negative when relative yield reduction exceeded 80%, suggesting that different morpho-physiological mechanisms are responsible for high yields under weed-free conditions or severe weed competition. Correlations between rice yield under weed-free conditions and weed biomass varied, giving 17 and 19 for positive and negative ones, respectively. These correlations were related neither to weed pressure level, ecosystem and yield level under weed-free conditions, nor to dominant weed species. These results indicate that correlations between rice yields under weed-free and weedy conditions can be strongly affected by weed pressure level, and, unless severe weed competition occurs, there appears to be no tradeoff between them. Association of morpho-physiological mechanisms with weed competitiveness under conditions differing in weed pressure levels and dominant weed species deserves further investigation.