The application of linear decision rules to reservoir management and design is extended in two directions. First, an improved linear decision rule is derived for reservoirs suffering evaporation losses. Release commitments generated by this rule are based on the storage at the beginning of the period, the predicted evaporation depth in the period, and the linearized storage area curve of the reservoir. A clearer and more rigorous derivation of the chance constraints, including the effects of both evaporation losses and possible failures to meet previous commitments, is presented. Second, several reservoir performance measures are formulated as objective functions that can be optimized by linear programing. These performance measures include expected and reliable values of storages and releases, deviations from targets, and reliabilities of achieving stated goals.