3.6.1S
CHANGING PRODUCTION SITUATIONS IN RICE AND IMPLICATIONS FOR PLANT PATHOLOGY

S SAVARY, FA ELAZEGUI, L WILLOCQUET and PS TENG

ORSTOM-IRRI Project on Rice Pest Characterisation, IRRI, Los Banos, The Philippines

Background and objectives
Agricultural changes are often reported as important factors determining the importance of diseases. A study was conducted to document this hypothesis in the case of rice. A methodology was developed to simultaneously analyse the considerable diversity of lowland rice production situations and disease profiles in tropical Asia.

Methods
This methodology involves the following steps:
(i) Selecting a number of sites that encompass such diversities; (ii) Developing a survey procedure at the field level; (iii) Determining clusters of production situations; (iv) Determining clusters of diseases, and more broadly, of pest injuries; (v) Determining if some production situations and some injury profiles are shared by geographically different sites; (vi) Testing the existence of linkages between production situations and injury profiles; (vii) Conducting yield loss experiments where rice yield is seen as a response surface to yield- determining and yield-limiting factors, reflecting production situations, and yield-reducing factors, reflecting injury profiles; and (viii) Using key, common production situations as a basis for an empirical, risk-analytical approach.

Results and conclusions
Several clusters of production situations and of injury profiles were found to be shared by different sites across Asia, providing grounds for the extrapolation of results of the study. This result shows that the regular, site-specific approach to disease management should be complemented by asystems, production situation-based approach. A strong link was also found between production situations and injury profiles, suggesting that the former determines to a large extent the latter. Yield loss experiments indicated a significant interaction of some combinations of injuries (in the form of factors generated by principal component analysis) and attainable yield, measured in uninjured plots. Some particular injuries appear to play a prominent role in these interactions, suggesting that a given level of such particular injuries lead to increased (or decreased) relative yield losses when the attainable yield is increased (i.e., the production situation is improved). Among these injuries that represent a specific risk under some production situations, some are caused by diseases, such as sheath blight and rice tungro.