Requirements and system design for the new CIS MLR forecasting system (SIPS)

• Main Authors: Serrer, Martin, Tivy, Adrienne
• Format: Report
• Language: English
• Published: National Research Council of Canada 2013
• Subjects: software system; sea ice forecasting; Sea ice
• Online Access: https://doi.org/10.4224/21271017; https://nrc-publications.canada.ca/eng/view/ft/?id=16989651-7a0f-472a-a4c5-a8a0387c3d2b; https://nrc-publications.canada.ca/eng/view/object/?id=16989651-7a0f-472a-a4c5-a8a0387c3d2b; https://nrc-publications.canada.ca/fra/voir/objet/?id=16989651-7a0f-472a-a4c5-a8a0387c3d2b

SIPS will replace the current regression-based forecasting system, MLR-Model, that is run as a research tool and experimental operational tool at the Canadian Ice Service. The existing system is a large and complex collection of Fortran programs and scripts written in numerous languages (python, bsh, ksh, sed, etc.) that rely on a very rigid configuration both from an operating system as well as from a data organization point of view. The purpose of developing a new software system is to ease portability between computers, simplify the pre-processing and post-processing of input/output data, simplify the data structure, expand the application of the model, improve visualization, eliminate dependence on the proprietary NAG library and make the system accessible to other ice centers. All functionality found in the existing system will be retained. In addition to the above stated improvements the SIPS system will include the following new functionality.  the ability to forecast on different time-scales  a mechanism to easily integrate new predictor datasets  visualization and analysis tools NRC publication: Yes