The APEC Climate Center produces seasonal predictions of prominent climatological variables using a 2.5 degree grid system. However, agriculture and water resource management are concerned with a specific site or a narrow basin, which are only covered roughly with this grid system, and demand daily values for data such as relative humidity, solar radiation, and surface wind speed instead of large-scale circulation variables. These gaps mean that seasonal prediction data is too limited to be used widely and popularly. Therefore, we developed a statistical downscaling method to enhance the utility of the seasonal predictions. Our proposed downscaling method is based on a weather generator called “acidWG,” and it is divided mainly into precipitation and temperature models. The precipitation model consists of several parametric sub-models whose parameters are adjusted for seasonal prediction. The temperature model decomposes the daily variation in the temperature and introduces low-frequency oscillation terms that are simulated and downscaled according to seasonal predictions. Finally, we propose an algorithm for downscaling seasonal prediction based on those models. The proposed method is applied to the Nakdong river basin during MAM and SON to evaluate its performance in terms of reproducibility, validity, and predictability. Results show that the proposed method produces suitable seasonal prediction scenarios.

1. 서론

2. 연구자료와 대상지역

3. 계절예측 상세화 방법

3.1 APCC 계절확률예측과 관측기간 월강수/기온상태의 정의

3.2 acidWG 모형과 모의

3.2.1 강수모형과 모의

3.2.2 기온 모형과 모의

3.3 상세화 방법

4. 연구결과

4.1 재현성 평가

4.1.1 습윤일 강수량 분포

4.1.2 습윤일 강수량의 공간상관성

4.1.3 일최고/최저기온의 분포

4.1.4 기온의 공간상관성

4.1.5 시간적 자기상관 및 교차상관

4.2 계절확률예측 상세화 정보의 적절성과 예측성 평가

5. 결론 및 토론

5.1 결 론

5.2 토 론

REFERENCES