Abstract ID: 171
Reliability indices for S2S model for societal application
Lead Author: Ranjeet Singh Bais Bais
IIT Kharagpur, India
Keywords: Reforecast, Subseasonal, Forecast, Monsoon , S2S prediction database
Abstract: Prediction in the subseasonal to seasonal (S2S) timescale has taken a big leap forward during the first phase of the S2S prediction project which includes the skillful forecast of MJO and associated events across the world, development of the S2S prediction database, and spreading awareness about the usefulness of the S2S forecast product. Though the skill of the S2S forecast has increased in the past few years, the same has not been translated to the end users of the forecast. One of the reasons why the application communities have not fully realized the potential of the S2S prediction is due to inherent variability in the weather system which the S2S models could not well capture. This prevents the users to build up on the forecast information for making strategies to deal with the consequences of the coming high-impact weather events. This study proposes a framework for developing a reliability index for the S2S model forecast using reforecast data for the selected long-lived weather extreme events. The application of the developed reliability index is demonstrated through a case study in predicting active and break spells over the Core Monsoon Zone of India two to three weeks in advance using reforcast data of ERPv1 developed by IITM, Pune. The result shows that the model has good reliability for break events but poor reliability in the case of active events. This analysis not only renders the user about the forecast reliability but also provides flexibility in making informed decisions for the optimum utilization of the resources. In the future, this approach could be integrated with other high-impact weather events by the S2S modeling agencies so as to make the forecast in this timescale amiable.
Co-authors:
Amey Pathak (Assistant Professor, IIT Kharagpur, India)
