The Expert Seasonal Prediction System for Seasonal Outlook in Korea (ESPreSSO-KR) v1.0 was developed in 2017 to provide reliable seasonal predictions of monthly mean temperature. It is a hybrid dynamical-statistical system based on seasonal predictions from the APEC Climate Center Multi-Model Ensemble as well as its statistical downscaling techniques and expert knowledge. We have added several new features and sub-modules to the ESPreSSO and proudly present the new ESPreSSOv2.1. It has been completely re-written in Python, a more general language than the original GrADS + FORTRAN. The new ESPreSSOv2.1 provides extended seasonal forecasts with a lead time of up to three-months and sub-regional (provincial) forecasts as requested by the Korea Meteorological Administration. Furthermore, a probabilistic categorical forecast module was developed to generate tercile seasonal forecasts. It also automatically runs in the background, generates new seasonal forecasts, and sends seasonal outlooks to subscribers. We also improved maintenance modules so that an administrator can check the status of ESPreSSOv2.1 at a glance. Each module is self-contained and replaceable so that an administrator can easily switch computational methodologies. We compare the technical details of ESPreSSOv1.0 vs. ESPreSSOv2.1 and provide the practical basis for upgrading ESPreSSO.

1. INTRODUCTION

2. ESPreSSOv2.1

2.1 프로그래밍 언어의 변경(포팅)

2.2 권역별 예보 기능

2.3 확률 삼분위 예보 기능

2.4 예측 선행 기간의 확장

2.5 자동화 기능 및 인터페이스 개선 8

3. SCIENCE behind the ESPreSSO (December) 9

4. RESEARCH RESULTS

4.1 ESPreSSOv2.1의 예측 결과

4.2 ESPreSSO의 예측 성능 평가 및 검증

4.3 프로젝트의 기타 성과물

5. CONCLUDING REMARKS

REFERENCES