Many lessons were learned from transitioning six forecasting systems to operations in 10 years. On the technical aspects which include hardware, software, systems -- we found that keys to successful transition are on 1) methods to harden the software infrastructure to make a robust, automated system; 2) backup and alternative procedures for handling missing or corrupted input data; 3) standardized validation and skill assessment metrics; 4) preparation of complete documentation including validation test reports, standard operating procedures (SOP), and software user guides; 5) adequate near-real-time observations of discharge, and water levels to provide LBCs for the system and 6) field projects in the Great Lakes (i.e. IFYGL) to provide surface and subsurface data for the evaluation of the forecast models during development and testing. In particular, program source codes need to be frozen during the testing, validation and the transition period with proper version control.
In addition to the technical aspects, a successful system transition from the research/development stage into operations also involves non-technical aspects, such as commitment from senior leadership, frequent communications among all involved parties on progress and milestones, training sessions for the system operators and user engagement workshops for the end users.