Evaluation of PNU CGCM Ensemble Forecast System for Boreal Winter Temperature over South Korea

PNU CGCM 앙상블 예보 시스템의 겨울철 남한 기온 예측 성능 평가

Abstract

The performance of the newly designed Pusan National University Coupled General Circulation Model (PNU CGCM) Ensemble Forecast System which produce 40 ensemble members for 12-month lead prediction is evaluated and analyzed in terms of boreal winter temperature over South Korea (S. Korea). The influence of ensemble size on prediction skill is examined with 40 ensemble members and the result shows that spreads of predictability are larger when the size of ensemble member is smaller. Moreover, it is suggested that more than 20 ensemble members are required for better prediction of statistically significant inter-annual variability of wintertime temperature over S. Korea. As for the ensemble average (ENS), it shows superior forecast skill compared to each ensemble member and has significant temporal correlation with Automated Surface Observing System (ASOS) temperature at 99% confidence level. In addition to forecast skill for inter-annual variability of wintertime temperature over S. Korea, winter climatology around East Asia and synoptic characteristics of warm (above normal) and cold (below normal) winters are reasonably captured by PNU CGCM. For the categorical forecast with $3{\times}3$ contingency table, the deterministic forecast generally shows better performance than probabilistic forecast except for warm winter (hit rate of probabilistic forecast: 71%). It is also found that, in case of concentrated distribution of 40 ensemble members to one category out of the three, the probabilistic forecast tends to have relatively high predictability. Meanwhile, in the case when the ensemble members distribute evenly throughout the categories, the predictability becomes lower in the probabilistic forecast.

Fig. 1. Schematic diagram of PNU Global Long-term Seasonal Ensemble Prediction System of which produce 40 ensemble members on monthly basis.

Fig. 2. Temporal correlation coefficients (Y axis) between ASOS and ensemble mean according to ensemble size (X axis). All possible ensemble combinations are calculated without duplication within 40 ensemble members.

Fig. 3. Spatial distribution of the ratio of ensemble numbers to 40 ensembles (shaded) which have statistically significant TCC. Dashed areas denote where ENS have statistically significant TCC. Value at Upper-right corner is area averaged TCC of ENS over S. Korea region (Boxed). Statistical significance is evaluated at 95% confidence level.

Fig. 4. a) Taylor diagram and box plots, b) Temporal Correlation Coefficient and c) Root Mean Square Error (unit: ℃) for temperature over S. Korea. Each ensemble member is denoted in grey open circle (a) and box (b, c) and ensemble mean is marked in red closed circle.

Fig. 6. Climatology of sea level pressure (shaded), wind at 850 hPa (vector) (a, b) and geopotential height at 500 hPa (shaded), zonal wind at 300 hPa (contour) (c, d) from R2 (a, c) and ENS (b, d) during boreal winter.

Fig. 7. Normalized anomaly composite map of sea level pressure (shaded) and 850 hPa wind (vector) for the case of Warm (a, c) and Cold (b, d) winter in observation (a, b) and PNU CGCM (c, d).

Fig. 8. Normalized anomaly composite map of 500 hPa geopotential height (shading with light grey contour) and 300 hPa zonal wind (bold black contour) for the case of Warm (a, c) and Cold (b, d) winter in observation (a, b) and PNU CGCM (c, d).

Fig. 9. Hit rate (%) of probabilistic prediction (bar) with respect to the distribution of 40 ensemble members and deterministic prediction (line).

Fig. 5. Box plot of Hit rate, Heidke skill score, False Alarm Rate for the 40 ensemble members for surface temperature compared with ASOS over South Korea.

Table 1. Warm, normal and cold winters classified by the 3 × 3 contingency table based on the deterministic forecast of PNU CGCM with threshold ± 0.43σ.

Table 2. Warm, normal and cold winters classified by the 3 × 3 contingency table based on the probabilistic forecast of PNU CGCM with threshold ± 0.43σ.