A predictive model of water quality variables has been developed for SPOT-1 imagery, and applied to Lake Patzcuaro, Mexico. By using canonical and principal component analysis it is shown that, at most, two water quality variables, suspended solids and chlorophyll-a concentration, can be derived from the SPOT data. Having established the independent predictiveness of a set of empirical relations between the SPOT data and the water quality variables, the whole lake was analysed to reveal the spatial distribution of suspended solids and chlorophyll-a. It was revealed that a very serious algal bloom had occurred. No bloom of such intensity has ever been observed before or suspected to have occurred in Lake Patzcuaro. The applicability of SPOT imagery to water quality monitoring is clearly demonstrated.