Abstract
An entropy minimax analysis to forecast the annual hydrologic status of central California has been re-performed using 96 water years for model training and reserving the 30 consecutive water years 1948–77 for model validation. On the validation set, the model exhibits a 67% accuracy. Statistical tests establish a 98% confidence that this success is not due to chance. Results of the analysis suggest that dry years may be more predictable than wet years for central California.