연구보고서
- 저자
- 손수진 박사
- 작성일
- 2019.06.12
- 조회
- 635
- 요약
- 목차
Recent progress in predicting seasonal climate variability has attributed to a major scientific advancement in dynamical seasonal prediction (DSP). The Asia-Pacific Economic Cooperation (APEC) Climate Center (APCC) has been devoted to producing and disseminating the operational multi-model ensemble (MME) seasonal prediction with a state-of-the-art climate models or systems since 2005. However, there has been a relative lag in the effective utilization of seasonal forecasts due to the limitation of available prediction data. In addition to the fact, this project has been also simulated by the recent advent of new operational communities for MME-based seasonal climate prediction. Therefore, a multi-year project to develop a prediction system for new content based on APCC MME seasonal forecasts was launched in 2017. In the first year, the prediction skill of the potential content to be selected for the new climate data services has been evaluated comprehensively.
The first objective of this project in 2018, the second year, was to develop an operational prediction system for forecasts of ENSO type (or phase) and strength (or intensity) events. The probabilistic approach to ENSO type (El Niño, ENSO-neutral, or La Niña) and strength (strong/moderate/weak El Niño or La Niña) events is based on an uncalibrated MME with equal weighting using a parametric Gaussian fitting method, which is the most appropriate for use in an operational prediction system based on the first year of research results from this project. The system has been fully tested for stability and coherence with the current operation system for MME seasonal forecasts and is currently in operation. The forecast probabilities of ENSO type and strength events with a six month lead will be disseminated to APEC members via our website by the end of this year.
The second objective is to develop an MME-based global drought prediction system and evaluate its prediction skills in terms of probabilistic forecasting. The major driver of climate and source of predictability for terrestrial precipitation, accumulated over six months, have been analyzed and its temporal and spatial structures have been identified. Investigations of local MME-based drought predictions were conducted based on drought-prone regions and periods selected using these results. The (deterministic) MME-based global drought prediction system (with a six-month lead) developed as prototype in 2017, in particular, has this year been expanded to a probabilistic forecast system. The usefulness of probabilistic forecasts in the test-bed has also been verified based on various measures. Finally, it was found that MME-based drought predictions can provide useful seasonal climate information and implies further possibilities of being one of the new content in the climate service.