Assessing the Representativeness of Data on Yield Losses Due to Rice Diseases in Tropical Asia

The representativeness of information on yield losses due to rice diseases in tropical Asia was studied. Published studies involving different groups of diseases (viral, bacterial, and fungal) and conducted in different rice production ecosystems were compared to help identify research gaps, the filling of which could improve current disease management in rice and help in developing strategies that fit the management needs of fast-evolving rice production systems in the future. Four criteria of representativeness of yield loss information were used: representativeness over time (the proportion of studies conducted over more than one crop cycle), representativeness over space (the proportion of studies conducted in more than one location), representativeness of scale (the proportion of studies conducted on the scale of plots or fields), and representativeness of injury (the standard deviation of the proportion of studies using inoculation, spontaneous infection, or chemical control). A strong imbalance in both the number and the representativeness of studies dealing with fungal, viral, and bacterial diseases was found. Most of the few studies of yield loss due to viral diseases (mainly rice tungro disease) were conducted on the scale of individual (potted) plants or were based on one-year data sets, often reflecting strong epidemics only. Studies of bacterial diseases were conducted in single locations only, and whether such results can be extrapolated still needs to be addressed. There is an acute need to better document yield losses in rice ecosystems other than the irrigated ecosystem. While studies conducted in the upland, rain-fed lowland, and deep-water rice ecosystems seem to have a high degree of representativeness, this cannot compensate for their small number in view of the great diversity of these environments. Studies of irrigated rice tend to concentrate on one year and one location. This approach may be based on the erroneous view that the irrigated ecosystem is homogeneous, and possible extrapolation of data from these studies needs to be examined.