Meta-analytic modelling of the incidence-yield and incidence-sclerotial production relationships in soybean white mould epidemics

White mould (Sclerotinia sclerotiorum) is a destructive disease of soybean worldwide. However, little is known of its impact on soybean production in Brazil. A meta-analytic approach was used to assess the relationship between disease incidence and soybean yield (35 trials) and between incidence and sclerotia production (29 trials) in experiments conducted in 14 locations across four seasons. Region, site elevation and season included as moderators in random-effects and random-coefficients models did not significantly explain the variability in the slopes of the incidence-yield relationship. The Pearson's r, obtained from back-transforming the Fisher's Z estimated by an overall random-effects model, showed that incidence of white mould was moderately and negatively correlated with yield (r = -0.76, P < 0.0001). A random-coefficients model estimated a slope of -17.2 kg ha(-1)% (-1), for a mean attainable yield of 3455 kg ha(-1), indicating that a 10% increase in white mould incidence would result in a mean yield reduction of 172 kg ha(-1). White mould incidence and production of sclerotia were strongly and positively correlated (r = 0.85, P < 0.0001). For every 10% increase in white mould incidence, 1 kg ha(-1) of sclerotia was produced. The relationship between disease incidence and production of sclerotia was stronger in southern regions and at higher elevation. In the absence of management, economic losses associated with white mould epidemics, assuming 43% incidence in 22% of the soybean area, were estimated at approximately US $1.47 billion annually within Brazil.